article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
positron emission tomography ( pet ) is a nuclear imaging technique used to monitor metabolic activity of the body .
the pet system detects gamma rays generated when electrons collide with positrons emitted by radiolabelled biologically active molecules , such as fluorodeoxyglucose . in oncology , pet is widely used in diagnostics of cancer metastases , in monitoring of progress in course of the cancer treatment , and also in planning radioterapheutic interventions . in the last case , pet scans , often in combination with computer tomography ( ct ) scans ,
are used to delineate biologically active parts of tumors , which are then treated with ionizing radiation .
accurate delineation of the tumor is crucial for delivering appropriate radiation dose and minimizing adverse side - effects of the therapy .
however , the problem is not trivial as evidenced by substantial intra- and interobserver variations when delineation is performed manually by experts@xcite
. therefore it is not surprising that automated methods for accurate and reproducible delineation are sought .
in this piece of research we evaluated 3d implementations of spatial distance weighted fuzzy c - means ( sdwfcm ) of guo et al .
@xcite , dictionary - based model ( dict ) of dahl and larser @xcite and convolutional neural network ( cnn ) @xcite , and compared them to classical approaches clustering approaches : k - means ( km ) @xcite and gaussian mixture models ( gmm ) @xcite . in order to assure consistency of the input data ,
images detected to be sharper than an arbitrary threshold ( after convolution with the 3d laplacian kernel ) were slightly blurred using the gaussian filter .
the step was omitted in the gmm and cnn case . in order to clean resulting labels from small artifacts ,
the labels were subject to subsequent morphological opening and closing , except the cnn method .
we used @xmath0 implementations of km and gmm methods .
the method is a spatial information - aware extension of the fuzzy c - means clustering algorithm ( fcm ) @xcite .
both methods assign to each point a coefficient of belonging to one of @xmath1 clusters .
similarly to fcm , sdwfcm starts by calculating cluster centers and euclidean distances based on regular features ( such as image intensity ) .
however , in the next step the distance matrix is modified with weight based on spatial distance of each pixel to the cluster center .
then , coefficients of belonging are calculated based on the modified distance matrix .
the procedure repeats until convergence or maximum runtime .
finally the tumor label is assigned to @xmath2 most intense clusters .
parameters for the sdwfcm algorithm ( as well as for km and gmm ) were optimized for the training set ( see tabl .
[ results ] and ref .
@xcite ) .
the general idea of the method is to assign each patch of the test image , the label which corresponds to the most similar patch in the dictionary .
the dictionary , which consists of a set of patch - label pairs is built in two steps .
first , a sample of training patches is used to seed the dictionary .
based on the label similarity threshold , similar patches are clustered and averaged . then , a version of the vector quantization technique is used to adjust the dictionary patches such that real labels of its matches are as close to binarized dictionary labels ( also adjusted in the training ) .
finally , the labeling algorithm assigns binary labels of the most similar dictionary patch . in our implementation
, the labeling window walks pixel by pixel and each pixel gets the label of the best fitting patch or an average label , which is then binarized using the final labeling threshold @xmath3 .
dict was set up with the following parameters : 3x3x3 patches were utilized , 10% of the training patches were used to seed the dictionary using similarity threshold 0.5 ; all training patches were used for adjusting the dictionary in 10 iterations with the step coefficient @xmath4 0.05 ( see ref .
@xcite ) .
deep convolutional neural networks have proved to be one of the most efficient tools in image processing . for the tumor delineation problem we tested many cnn architectures consisting of up to dozen layers , including three or four convolutional layers .
the input of the network was a batch of 2d or 3d patches from the pet tumor images .
the network was trained using the adagrad stochastic gradient descent algorithm @xcite , which aimed at minimizing the euclidean distance between the output of the final layer of the network and the ground truth . to account for a relatively small number of training samples , the set of available data
was artificially augmented with rotationally transformed samples .
the best architecture , selected in the 5-fold cross - validation process , was the 3d convolutional neural network with 3 convolutional layers followed by relu activations with ascending number of filters ( fig . 1 ) .
[ figcnn ] , activ @xmath5 - activation layer with function @xmath5.,width=432 ]
the training dataset provided by organizers of the petseg challenge consisted of 4 scans from 3 clinical samples , 9 scans from 3 phantoms and 6 simulated scans in 2 sets .
we assessed efficiency of the methods on the training set in terms of the dice coefficient . as the scores were calculated on the training set , they are rather upper estimates of the real performance of the methods .
.*dice coefficients for training data . *
notations : km - k - means , gmm - gaussian mixture model , sdwfcm - spatial distance weighted fuzzy c - means , dict - dictionary - based model , cnn - convolutional neural network ; @xmath6/@xmath7/@xmath1 - number of clusters , @xmath2 - number of selected most intense clusters , @xmath8 - degree of fuzzy classification , @xmath9 - weight of spatial features , @xmath10 - number of neighbours . [
cols="<,<,>,>,>,>,>,>",options="header " , ] the best results were obtained using cnn ( dice score mean : 0.86 , median : 0.89 ) .
the remaining methods performed slightly worse than cnn and similarly to each other ( dice score mean : 0.82 - 0.83 , median : 0.81 - 0.83 ) . in general , simulated data proved to be the easiest ( mean dice score 0.88 - 0.91 ) , while clinical data were the hardest ( mean dice score 0.76 - 0.78 ) , regardless the method . a single most difficult case was _
clinical_1 _ for which none of the algorithms exceeded dice score of 0.69 .
it is of note , that the gaussian blurring had a positive impact on segmentation of this case .
clinical_1 , clinical_3 _ and _ phantom_2.1 _ were the only cases where cnn did not reach dice score of 0.80 .
interestingly , cnn was always better than other methods for phantom data , with average margin over the second best result of almost 7% .
looking more into details , cnn and dict achieved the highest sensitivity ( 0.89 and 0.85 , respectively ) , while sdwfcm achieved the highest precision ( 0.85 ) . in accordance with these results ,
cnn and dict were more prone to overestimate the total volume of the tumor ( on average by 10% ) , while sdwfcm lined towards underestimated volumes ( on average by 3% ) .
all methods overestimated the volume of _
( by 30 - 49% ) , _ phantom_2 _ ( 9 - 58% ) and _ phantom_3 _ ( 8 - 47% ) , while the volume of _
clinical_3 _ was generally underestimated ( 12 - 26% ) .
interestingly , while cnn was the method which overestimated the volume of clinical samples the most ( 22% ) , its volume predictions were closest to the ground truth for the phantoms ( + 9% ) .
accordingly , the contour mean distance of cnn predictions was the worst for clinical samples ( 2.28 in comparison to 1.81 by dict ) , and the best for phantom samples ( 0.72 in comparison to 1.30 by dict ) .
overally , our early results suggest that deep convolutional neural networks can enable a step forward also in the pet image segmentation .
this claim should be further validated with more testing , especially on clinical samples , for which performance gain using cnn was less impressive than for phantom data .
this research has been conducted under the support programme of the partnership between higher education and science and business activity sector financed by city of wroclaw .
the authors thank jakub szewczyk for his help with preparing the manuscript , and piotr krajewski for continuous support of the project . | in oncology , positron emission tomography imaging is widely used in diagnostics of cancer metastases , in monitoring of progress in course of the cancer treatment , and in planning radiotherapeutic interventions .
accurate and reproducible delineation of the tumor in the positron emission tomography scans remains a difficult task , despite being crucial for delivering appropriate radiation dose , minimizing adverse side - effects of the therapy , and reliable evaluation of treatment . in this piece of research we attempt to solve the problem of automated delineation of the tumor using 3d implementations of the spatial distance weighted fuzzy c - means , the deep convolutional neural network and a dictionary model .
the methods , in diverse ways , combine intensity and spatial information . |
the authors are grateful to the useful discussion with hsien - chung kao and shoichi kawamoto in the early stage of this work .
wyw would like to thank the warm hospitality received in the ustc , china when part of this work was presented in the east asia joint workshop on fields and strings and insightful discussion with sang - jin sin .
this work is supported in parts by the chung yuan christian university , the taiwan s ministry of science and technology ( grant no .
102 - 2112-m-033 - 003-my4 ) and the national center for theoretical science .
in this appendix , we will derive the relation among parameters @xmath74 in the warping factors : one can freely absorb the combination @xmath80 into redefinition of @xmath14 , that is to set @xmath81 .
furthermore one can eliminate @xmath82 by a choice of regularization scheme , such that all the divergent terms to be _
@xmath14-independent _ in the onshell action of gravity sector . in the other words
, one would use the same counterterm in both confinement and deconfinement phases . to be specific
, the onshell action reads j. m. maldacena , int .
j. theor .
phys . * 38 * , 1113 ( 1999 ) [ adv .
* 2 * , 231 ( 1998 ) ] doi:10.1023/a:1026654312961 [ hep - th/9711200 ] .
s. s. gubser , i. r. klebanov and a. m. polyakov , phys .
b * 428 * , 105 ( 1998 ) doi:10.1016/s0370 - 2693(98)00377 - 3 [ hep - th/9802109 ] .
e. witten , adv .
theor . math .
* 2 * , 253 ( 1998 ) [ hep - th/9802150 ] .
s. s. gubser , class .
. grav . * 22 * , 5121 ( 2005 ) doi:10.1088/0264 - 9381/22/23/013 [ hep - th/0505189 ] .
s. s. gubser , phys .
d * 78 * , 065034 ( 2008 ) doi:10.1103/physrevd.78.065034 [ arxiv:0801.2977 [ hep - th ] ]
. s. a. hartnoll , c. p. herzog and g. t. horowitz , phys .
* 101 * , 031601 ( 2008 ) doi:10.1103/physrevlett.101.031601 [ arxiv:0803.3295 [ hep - th ] ] .
s. s. gubser , phys .
* 101 * , 191601 ( 2008 ) doi:10.1103/physrevlett.101.191601 [ arxiv:0803.3483 [ hep - th ] ] .
e. nakano and w. y. wen , phys .
d * 78 * , 046004 ( 2008 ) doi:10.1103/physrevd.78.046004 [ arxiv:0804.3180 [ hep - th ] ] . t. albash and c. v. johnson , jhep * 0809 * , 121 ( 2008 ) doi:10.1088/1126 - 6708/2008/09/121 [ arxiv:0804.3466 [ hep - th ] ] . s. a. hartnoll , c. p. herzog and g. t. horowitz , jhep * 0812 * , 015 ( 2008 ) doi:10.1088/1126 - 6708/2008/12/015 [ arxiv:0810.1563 [ hep - th ] ] . m. kruczenski , d. mateos , r. c. myers and d. j. winters , jhep * 0405 * , 041 ( 2004 ) doi:10.1088/1126 - 6708/2004/05/041 [ hep - th/0311270 ] .
t. sakai and s. sugimoto , prog .
phys . * 113 * , 843 ( 2005 ) doi:10.1143/ptp.113.843 [ hep - th/0412141 ] .
j. erlich , e. katz , d. t. son and m. a. stephanov , phys .
* 95 * , 261602 ( 2005 ) doi:10.1103/physrevlett.95.261602 [ hep - ph/0501128 ] .
l. da rold and a. pomarol , nucl .
b * 721 * , 79 ( 2005 ) doi:10.1016/j.nuclphysb.2005.05.009 [ hep - ph/0501218 ] .
o. aharony , j. sonnenschein and s. yankielowicz , annals phys . * 322 * , 1420 ( 2007 ) doi:10.1016/j.aop.2006.11.002 [ hep - th/0604161 ] .
c. p. herzog , phys .
lett . * 98 * , 091601 ( 2007 ) doi:10.1103/physrevlett.98.091601 [ hep - th/0608151 ] . s. w. hawking and d. n. page , commun .
* 87 * , 577 ( 1983 ) .
doi:10.1007/bf01208266 y. nambu , phys . rev . * d10 * , ( 1974 ) 4262 j. w. chen , y. j. kao and w. y. wen , phys . rev .
d * 82 * , 026007 ( 2010 ) doi:10.1103/physrevd.82.026007 [ arxiv:0911.2821 [ hep - th ] ] .
t. faulkner , g. t. horowitz , j. mcgreevy , m. m. roberts and d. vegh , jhep * 1003 * , 121 ( 2010 ) doi:10.1007/jhep03(2010)121 [ arxiv:0911.3402 [ hep - th ] ] .
s. s. gubser , f. d. rocha and p. talavera , jhep * 1010 * , 087 ( 2010 ) doi:10.1007/jhep10(2010)087 [ arxiv:0911.3632 [ hep - th ] ] .
g. t. horowitz and m. m. roberts , phys .
d * 78 * , 126008 ( 2008 ) doi:10.1103/physrevd.78.126008 [ arxiv:0810.1077 [ hep - th ] ] .
r. hofmann , int .
j. mod .
a * 20 * , 4123 ( 2005 ) doi:10.1142/s0217751x06035129 , 10.1142/s0217751x05023931 [ hep - th/0504064 ]
. t. faulkner , h. liu , j. mcgreevy and d. vegh , phys .
d * 83 * , 125002 ( 2011 ) doi:10.1103/physrevd.83.125002 [ arxiv:0907.2694 [ hep - th ] ] .
d. b. kaplan , j. w. lee , d. t. son and m. a. stephanov , phys .
d * 80 * , 125005 ( 2009 ) doi:10.1103/physrevd.80.125005 [ arxiv:0905.4752 [ hep - th ] ] .
w. y. wen , phys .
b * 707 * , 398 ( 2012 ) doi:10.1016/j.physletb.2011.12.053 [ arxiv:1009.3952 [ hep - th ] ] .
m. unsal , phys .
d * 80 * , 065001 ( 2009 ) doi:10.1103/physrevd.80.065001 [ arxiv:0709.3269 [ hep - th ] ] . | we show that a hairy black hole solution can provide a holographically _ dual _ description of quark confinement .
there exists a one - parameter sensible metric which receives the backreaction of matter contents in the holographic action , where the scalar and gauge field are responsible for the condensation of chromomagnetic monopoles .
this model features a preconfining phase triggered by second - order monopole condensation and a first - order confinement / deconfinement phase transition . to confirm the confinement ,
the quark - antiquark potential is calculated by probing a qcd string in both phases . at last ,
contribution from kaluza - klein monopoles in the confining phase is discussed .
the holographic correspondence between a gravitational theory in the bulk and a quantum field theory on the boundary , first emerged under the anti - de sitter / conformal field theory ( ads / cft ) correspondence @xcite , has been proved useful to study condensed matter phenomena . in particular , the authors in @xcite proposed a gravity model in which abelian symmetry of higgs is spontaneously broken by the existence of black hole , also known as hairy black hole .
this mechanism was incorporated in the model of superconductivity and critical temperature was observed @xcite .
this model was later studied in the presence of magnetic field @xcite and with full backreaction @xcite .
on the other hand , the ads / cft analogy has also explained many qualitative and quantitative features of confining gauge theories like qcd , either by engineering branes and strings in the ten dimensional string theory @xcite , or by bottom - up construction of a randall - sundrum like background @xcite .
in particular , holographic models of deconfinement were first discussed in the sakai - sugimoto model at finite temperature@xcite and the hard / soft wall model@xcite . in all cases , the confinement / deconfinement transition is identified as the first - order hawking - page phase transition , where the thermal ads is energetically favored at low temperature , while a black hole solution at high temperature @xcite . in this paper
we propose the first holographic confinement model without using the hawking - page phase transition .
we recall an old but appealing proposal of dual superconductor @xcite , stating that the confining phase can be regarded as a dual type ii superconducting state due to the condensation of chromomagentic charges ( monopoles ) , such that the chromoelectric lines between a quark - antiquark pair are squeezed by dual meissner effect into flux tubes , as known as the qcd strings .
one may wonder if the hairy black hole scenario in the holographic model of dual superconductor@xcite works just fine by trivially replacing cooper pairs by monopole condensate , and the backreacted metric in the @xcite might be the confinement geometry in desire .
however , a second thought and straightfoward computation will show this is not the case .
the backreaction of condensate may have resulted to a gapped excitation as shown in the single electron spectral function @xcite , but the attractive force mediated by phonons is coulomb - like , rather than a linear potential for flux tube in the dual picture .
then the challenging task is to find out the appropiate ansatze which can produce the linear potential while condensation occurs . in this letter
, we consider a @xmath0-dimensional dual supercondcutor model which has a holographic description of eisntein - maxwell - dilaton gravity theory in the ads@xmath1 .
the action is composed of a gravity sector @xmath2 , a matter sector @xmath3 for massive scalars and a dual @xmath4 gauge field sector @xmath5 as follows : @xmath6 we will study a class of metric of following general form : @xmath7 with @xmath8\nonumber\\ & g_{1}(z)=\frac{exp(\beta ( p z / z_{h})^2)}{z^2}\frac{1}{1-(\frac{z}{z_{h}})^3}\nonumber\\ & g_{2}(z)=\frac{exp(\gamma ( p z / z_{h})^2)}{z^2}\end{aligned}\ ] ] the metric corresponds to a hairy black hole with a flat horizon at @xmath9 and approaches asymptotically to ads as @xmath10
. one can show by explicit calculation that the oversimplified ansatz @xmath11 , which was adopted in the backreacted holographic superconductor such as one in @xcite , is unable to produce the linear potential in the confining phase .
as shown in the appendix , the warping factors can be further pinned down to @xmath12 and @xmath13 after imposing the condition that the onshell action of gravity sector is independent of the choice of @xmath14 after regularization .
the total free energy is given by summing up onshell action of all sectors and the value of @xmath14 is determined by minimizing the total free energy at a given temperature .
the variable @xmath15 , on the other hand , is a free parameter , which can be fine tuned to match the tension of flux tube . in practice ,
the desirable onshell action of gravity sector @xmath2 should have local minimum only at @xmath16 , which happens as @xmath15 is within some range as shown in the figure [ alpha_value ] . for different @xmath15 .
the blue , green , purple , yellow curves correspond to @xmath17 - 0.6 , 0 , 0.2 and 1 , respectively .
it is obvious that the blue ( @xmath18 ) and yellow ( @xmath19 ) curves are undesirable since that they have additional local minimum at some @xmath14 other than @xmath20.,width=302 ] in the deconfining phase , the solution ( [ ansatz ] ) reduces to the ads schwarzschild black hole , while confining parameter @xmath14 is zero and the oneshell actions of scalar and gauge sectors are trivially zero .
as temperature decreases , one expects a transition to the confining phase as the the black hole develops hairs after receiving backreaction from condensate . the scalar field , which represents the order parameter of chromomagnetic charge condensate , poccesses the asymptotical form : and confining parameter @xmath14 at different temperatures .
those curves associated to temperature above @xmath21 usually exhibit a cusp at @xmath22 and a cutoff at @xmath23 , in between the monopole condensation occurs .
the blue , purple , yellow and green curves correspond to the scaled temperatures @xmath24 at 1.0 , 1.1 , 1.2 , 1.3 , respectively .
one observes competition between local minima @xmath25 ( confining phase ) and @xmath16 ( deconfining phase ) at the temperature @xmath26 , which is between @xmath27 and @xmath28 , signaling a preconfining phase .
there is only a confining phase when temperature is below @xmath21 .
here we set @xmath29 , @xmath30 and @xmath31 for simulation .
, width=302 ] @xmath32 with @xmath33 , where @xmath34 and @xmath35 are interpreted as source and condensate , respectively .
we remark that one can also switch the role of source and condensate for a scalar field in asymptotical ads@xmath1 .
that would correspond to the condensate of operators with scaling dimension @xmath36 .
the chemical potential @xmath37 and charge density @xmath38 can be read out from the asymptotic form of gauge field : @xmath39 in our simulation , the same boundary conditions adopted in the @xcite are imposed , that is , @xmath40 and @xmath38 is fixed at the asymptotic boundary and @xmath41 at the horizon .
the condensate @xmath42 and chemical potential @xmath37 are obtained numerically by the shooting methods .
the phase transition is justified by evaluation of the total onshell action against confining parameter @xmath14 at different temperatures . in the figure [ onshell_action ]
, we see that the confining phase ( local minimum at @xmath25 ) is energetically favored at and below the critical temperture @xmath21 .
the preconfining phase appears when two local minima @xmath16 and @xmath25 competes with each other during @xmath43 .
the confinement / deconfinement phase transition occurs around @xmath26 when the local minimum @xmath16 starts to dominate .
each onshell curve with temperature above @xmath21 exhibits a cusp at @xmath44 and a cutoff .
the cusp is generated by nontrivial condensate @xmath42 , which occurs when @xmath45 .
there is no ground state solution for the scalar @xmath46 above the cutoff . since confinement appears only when monopoles condensate
, one can conclude that the former phase transition is directly triggered by the latter in the dual superconductor model .
our model shows that @xmath47 and predicts the existence of preconfining phase with a fraction window @xmath48 .
we remark that the preconfing phase did not appear in the previous model@xcite and our result agrees with observation in some nonperturbative approach to thermal @xmath49 yang - mills theory @xcite .
although the dual superconductor model is the an effective ginzburg - landau theory which describes a second - order phase transition for the monopole condensate , we argue that the confinement / deconfinement phase transition in our hairy black hole model is of first order .
this is because temperature in our model is not only determined by the horizon but also the hair , that is @xmath50 the sudden shift of parameter @xmath14 at the transition causes a discontinuous jump of horizon position @xmath51 and therefore a first - order phase transition .
is plotted against the seperation between string end points on the boundary . in the static configuration
, the action can be regarded as the potential between a heavy quark - antiquark pair . in the deconfinement phase ,
the coulomb potential ( red , lower branch ) is terminated as the string is _ sucked _ into the black hole ( red , upper branch ) . in the confinement phase , the linear potential ( blue ) dominates as seperation distance increases .
, title="fig:",width=302 ] is plotted against the seperation between string end points on the boundary . in the static configuration
, the action can be regarded as the potential between a heavy quark - antiquark pair . in the deconfinement phase ,
the coulomb potential ( red , lower branch ) is terminated as the string is _ sucked _ into the black hole ( red , upper branch ) . in the confinement phase
, the linear potential ( blue ) dominates as seperation distance increases . , title="fig:",width=302 ] in the imhomogenous background where monopole condensate occurs locally , one may expect both confinement and deconfinement coexist in different regions and the color force lines are squeezed thanks to the dual meissner effect .
one can further confirm both the confinement and deconfinement background by probing a string of profile @xmath52 , given by the nambu - goto action @xmath53 and @xmath38 ( black curve ) in the @xmath54-@xmath46 plane . by calculating the wavefunction @xmath55 at each intersecting point , we find they correspond to the ground state ( @xmath56 , blue curve ) , first excited state ( @xmath57 , pruple curve ) , and second excited state ( @xmath58 , brown curve ) , respectively . from viewpoint of boundary theory ,
those excited modes are the kaluza - klein ( kk ) monopoles with heavier mass .
we use contour curves @xmath59 and @xmath30 , parameters @xmath31 , and @xmath60 for this simulation.,width=377 ] the string profile in the gravity bulk is plotted against the seperating distance in the left figure [ string ] . with increasing seperation , the u - shaped string eventually touches the horizon and breaks in the presence of hairless black hole , indicating the deconfinement phase . in the confining phase
, the string can not enter the infared core of bulk in the presence of hairy black hole and stay near @xmath61 .
the horizontal segment of string is responsible to the linear potential .
we remark that string in the confining phase can not break without introducing additional dynamic degrees of freedom .
the onshell string action can be regarded as the free energy or potential of static heavy quark - antiquark pair , which is plotted against the seperation distance @xmath62 in the right figure [ string ] . in the deconfining phase , though it exhibits a coulomb potential for small seperation , no force exists for seperation @xmath63 in the unit @xmath64 because the string is eventually pulled into the black hole . in the confining phase , on the other hand
, a linear potential has already manifested for seperation @xmath65 .
-excited state can be approaximated by the relation @xmath66 .
( right ) the plot on the right shows the condensate ratio @xmath67 for each kk monopole , title="fig:",width=302 ] -excited state can be approaximated by the relation @xmath66 .
( right ) the plot on the right shows the condensate ratio @xmath67 for each kk monopole , title="fig:",width=302 ] at last , we would like to address a possible role played by excited modes of scalar hair in the confining phase . as demonstrated in the figure [ excitemode ] ,
multiple solutions exist as temperature well below the critical temperature of condensation .
those solutions correspond to those wavefunctions @xmath55 with @xmath68 nodes , or excited states . from the point of view of ads
bulk , the tower of excited modes might be related to the efimov states due to loss of conformality @xcite . from viewpoint of dual theory on the boundary ,
those excited states are the kaluza - klein ( kk ) monopoles with effective mass which grows with mode number @xmath68 . in the figure [ monopole ]
, we plot the critical temperature @xmath69 and condensate @xmath70 for the first five kk modes .
we observe that the a @xmath68-order kk monopole can only be excited at temperature lower than @xmath69 .
the measured monopoles mass @xmath71 would be proportional to the expectation value of condensate @xmath70 from all avaliable kk modes up to @xmath72 at some temperature , that is @xmath73 . as a result , an increase in monopole mass @xmath71 is expected as temperature moves away from the critical point in the confining phase thanks to the growing @xmath72 .
we remark that the kk monopole was condisered as part of a magnetic bion condensation in some qcd - like theory@xcite . in conclusion , we are among the first to construct a holographic model of quark confinement by chromomagnetic monopole condensation .
this toy model features a second - order monopole condensation as well as first - order confinement / deconfinement phase transition , and it predicts a preconfining phase , which is mostly unseen in the previous holographic construction but observed in other nonperturbative approach .
finally we hypothesize that excited modes of condensate might contribute to the increasing monopole mass in the confining phase .
our noval result has provided supportive evidence for the application of holographic method and serves as a toy model toward the quark confinement in realistic qcd . |
the universe is filled by light .
there are two principal sources of such photons the cosmic microwave background ( cmb ) and the extragalactic background light ( ebl ) .
while cmb properties are quite well measured and theoretically understood for more than 20 years ( e.g. @xcite ) , many ebl models with widely different parameters were developed @xcite@xcite ; for instance , the total ebl intensity for these models differs by more than a factor of two @xcite . comparatively recently
, however , some of these ebl models started to converge , at least in the 0.510 @xmath2 wavelength region ( e.g. , @xcite and @xcite ) .
high - energy @xmath0-rays with primary energy @xmath3/\epsilon ) [ tev]$ ] get absorbed @xcite@xcite ( @xmath4 ) on ebl and cmb photons with energy @xmath5 , reaching maximum cross section at @xmath6/\epsilon ) [ tev]$ ] @xcite .
therefore , the improved reliability of ebl models allows for more detailed studies of this fundamental quantum electrodynamics process in the @xmath0-ray energy range of about 500 @xmath1 10 @xmath7 , and of extragalactic @xmath0-ray propagation effects in general .
using the cmb for such studies is , in principle , also possible @xcite , but at present is not feasible due to several factors .
indeed , the typical mean free path @xmath8 of a @xmath9 @xmath10= 1 @xmath11 @xmath0-ray on the cmb is @xmath1210 @xmath13 ; for such high energies there is still no discovered sources , while using lower energies is not convenient either as in the @xmath14100 @xmath7 energy region the @xmath15 dependence is very strong : @xmath8 falls for about an order of magnitude for every 10 % of @xmath16 decrease . therefore
, extremely well knowledge of experimental systematics on primary @xmath0-ray energy would be crucial while using cmb photons at @xmath14100 @xmath7 as a probe of the @xmath4 process . in the present conference contribution we emphasize on the potential importance of the development of electromagnetic cascades on the ebl / cmb to the interpretation of observations in the high energy ( he , @xmath16=100 @xmath17100 @xmath1 ) and the very high energy ( vhe , @xmath16=100 @xmath1100 @xmath7 ) ranges .
this short paper does not pretend to claim great originality ; most of its basic ideas were already discussed and most of its main results were already presented in @xcite@xcite .
soon after the discovery of the first @xmath7 extragalactic @xmath0-ray source @xcite , the first @xmath0-astronomical constraints on the ebl density were obtained @xcite@xcite , accounting for only the pair - production process as the main cause of transformation of the primary @xmath0-ray spectrum ( for very distant sources , it is necessary to include adiabatic losses ) .
this model we call `` the absorption - only model '' . since 1993 , almost every year several papers appear , assuming the absorption - only model , sometimes at a rate of a dozen a year or more . in what follows , any significant deviations from the absorption - only model are called `` anomalies '' , even if they do not fall beyond the conventional physics .
several such anomalies were reported @xcite@xcite , with a statistical significance not overwhelming in every single case , but together they indicate that the absorption - only model is incomplete and must be modified in some way ( see @xcite for more details ) .
secondary electrons and positrons ( hereafter simply `` electrons '' ) produced in the @xmath4 process radiate cascade @xmath0-rays by means of the inverse compton ( ic ) process ; these @xmath0-rays may contribute to the observable spectrum of an extragalactic source .
therefore , besides the properties of background photon fields ( ebl / cmb ) , another very important factor is the strength and structure of the extragalactic magnetic field ( egmf ) .
some selected constraints on the egmf strength @xmath18 in voids of the large scale structure ( lss ) , most of them assuming correlation length of the field @xmath19= 1 @xmath20 , are shown in figure [ f1 ] .
there are two regions of @xmath18 values that satisfy most of these constraints : @xmath21 @xmath222@xmath23 @xmath24 and @xmath25 @xmath223@xmath26 @xmath24 .
the first case corresponds to the absorption - only model regime : for such a strong egmf secondary electrons are , as a rule , strongly deflected and delayed so that cascade photons do not contribute to the point - like image of the source . on the other hand , for the case of the second option
such a contribution is still possible , at least in the vhe energy range .
the lower bound on the @xmath18 value , @xmath25 @xmath24 , is highly uncertain @xcite-@xcite . very recently
a paper @xcite appeared , disfavouring a narrow range of values around @xmath18= @xmath21 @xmath24 for @xmath19= 1 @xmath20 .
let us assume for a moment the following simple two - phase model of the egmf .
lss voids with @xmath27 fill a fraction of the total volume @xmath28 , while the rest is occupied by the comparatively strong ( @xmath29 @xmath24 ) egmf with @xmath30 @xmath20 .
figure [ f2 ] shows several fits to the observed spectral energy distribution ( sed ) of blazar 1es 0229 + 200 ( @xcite for the four lowest - energy bins shown in black and @xcite for other bins ) assuming this model with various values of @xmath31 .
cascade component dominates at low energies ( @xmath32300500 @xmath1 ) .
details of simulations correspond to the case of figure 15 of @xcite .
reasonable fits in the energy range 30 @xmath1 10 @xmath7 may be obtained for the case of @xmath31 between 0.4 and 0.6 .
however , below 30 @xmath1 the model intensity exceeds the observed one .
this may be explained either by the deflection and/or delay of cascade electrons while travelling the egmf @xcite@xcite or by additional ( with respect to ic ) electron energy losses @xcite .
figure [ f3 ] depicts the main spectral signatures of the electromagnetic ( em ) cascade model : 1 ) high - energy cutoff , similar to the absorption - only model , 2 ) an ankle at the intersection of the primary and cascade components ( in fact , this signature is also visible in figure [ f2 ] ) , 3 ) a low - energy cutoff for the case of a non - zero egmf value ( `` magnetic cutoff '' ) , and 4 ) a `` second ankle '' at the low - energy region of the spectrum .
we argue that all above - mentioned anomalies may be interpreted within the framework of the em cascade model , namely : + 1 .
the apparent excess of observed @xmath0-rays in the highest - energy bins claimed in @xcite@xcite was discussed by @xcite .
a prominent ankle ( signature 2 in figure [ f3 ] ) may account for this effect .
the unusual spectral hardening towards lower energies @xcite is explained by the existence of a `` magnetic cutoff '' ( signature 3 in figure [ f3 ] ) , as the authors of @xcite themselves note ( however , see @xcite and @xcite itself for other possible explanations ) .
the same spectral feature ( a prominent magnetic cutoff ) may account for the result of @xcite , as discussed in @xcite .
+ 4 . finally
, `` halos '' observed around some blazars also may be explained in the context of the em cascade model @xcite .
+ thus , the em cascade model appears to be the simplest extragalactic @xmath0-ray propagation scenario that coherently explains all these anomalies .
other , more exotic models that could account for a part of these anomalies do exist @xcite@xcite ; the main difficulties of these scenarios were discussed by us in @xcite,@xcite .
recent observations of some blazars in the 1 @xmath1 10 @xmath7 energy region seem to indicate that the secondary component of cascade @xmath0-rays may constitute a considerable contribution to the observable flux in the @xmath331500 @xmath1 energy range .
this work was supported by the rfbr grant 16 - 32 - 00823 .
t.d . acknowledges the support of the students and researchers exchange program in sciences ( steps ) , the re - inventing japan project , jsps , and the hospitality of the university of tokyo icrr .
peebles , `` principles of physical cosmology '' , princeton university press ( 1993 ) t.m .
kneiske et al .
, a&a 413 ( 2004 ) 807 f.w .
stecker et al . , apj 648 ( 2006 ) 774 a. franceschini et al . , a&a 487 ( 2008 ) 837 j.r .
primack et al . , aipcp 1085 ( 2008 ) 71 j.d .
finke et al . , apj 712 ( 2010 ) 238 t.m .
kneiske & h. dole , a&a 515 ( 2010 ) a19 a. dominguez et al .
, mnras 410 ( 2011 ) 2556 r.c .
gilmore et al . , mnras 422 ( 2012 ) 3189 y. inoue et al . , apj 768 ( 2013 ) 197 e. dwek & f. krennrich , aph 43 ( 2013 ) 112 a.i .
nikishov , sov .
jetp 14 ( 1962 ) 393 r.j .
gould & g. shreder , phys .
155 ( 1967 ) 1408 j.v .
jelley , phys .
( 1966 ) 479 t. dzhatdoev et al .
, preprint astro.ph/1609.01013 ( 2016 ) t.a .
dzhatdoev , j. phys .
( 2015 ) 012035 t. dzhatdoev , izvestiya rossiiskoi akademii nauk 79 ( 2015 ) 363 ( preprint astro - ph/1501.00259 ) t. dzhatdoev et al .
, proc . of the isvhecri-2016 conference , to appear in epj web of conf .
( 2017 ) t. dzhatdoev et al . , proc . of the 2@xmath34 icppa conference , to appear in j. phys .
( 2017 ) m. punch , nature 358 ( 1992 ) 477 f.w .
stecker & o.c .
de jager , apj 415 ( 1993 ) 71 o.c . de jager et al . ,
nature 369 ( 1994 ) 294 d. horns & m. meyer , jcap ( 2012 ) 033 d. horns , preprint astro - ph/1602.07499 ( 2016 ) a. neronov et al .
, a&a 541 ( 2012 ) a31 a. furniss et al .
, mnras 446 ( 2015 ) 2267 w. chen et al .
( 2015 ) 211103 t.c .
arlen et al . , apj 796 ( 2014 ) 18 a.e .
broderick et al . , apj 752 ( 2012 ) 22 s. archambault et al . ,
preprint astro - ph/1701.00372 ( 2017 ) ie .
vovk et al . , apj lett .
747 ( 2012 ) l14 e. aliu et al .
, apj 782 ( 2014 ) 13 a. neronov & ie .
vovk , science 328 ( 2010 ) 73 c.d .
dermer et al . , apj lett . 733 ( 2011 ) l21 t. dzhatdoev et al . , proc .
of the 34@xmath35 russian cosmic ray conference , to appear in izvestiya rossiiskoi akademii nauk ( 2017 ) m.l .
ahnen et al . , preprint astro - ph/1612.09472 ( 2016 ) w. essey et al . , phys .
104 ( 2010 ) 141102 a. mirizzi et al .
d 76 ( 2007 ) 023001 | we present a very brief overview of some recent @xmath0-ray observations of selected blazars to reveal an indication for a considerable or even dominant contribution of secondary @xmath0-rays from electromagnetic cascades to the observable spectra in the 1500 @xmath1 energy range . |
it has been recognized that exotic magnetic excitations known as skyrmions may exist @xcite in a two - dimensional electron gas in a strong homogeneous magnetic field ( quantum hall system ) near spin polarized groundstates .
these are excitations of a two - dimensional spontaneous ferromagnet , the physics of which is relevant to this system ( despite the presence of a strong magnetic field ) , because of the small land @xmath0-factor in gaas systems ( where most experiments take place ) , which makes the zeeman coupling very small compared to other energy scales ( coulomb interaction , cyclotron energy ) in the problem .
skyrmions are spin configurations with a non - trivial winding number ( pontryagin index ) .
they were first discussed in the context of four - dimensional field theories @xcite , and were later recognized as states occurring in the non - linear sigma model description of two - dimensional ferromagnets @xcite . for filling factors @xmath1 close to one ( @xmath2 is the number of electrons and @xmath3 the number of magnetic flux quanta penetrating the system ) , these turn out to be the lowest energy quasiparticles under typical experimental circumstances .
skyrmions can thus be introduced into the groundstate by adding or removing charge from the system .
@xcite . experimentally , the case for the existence of skyrmions in a system close to @xmath4 is quite strong .
nmr experiments show a degrading of the spin polarization with deviation of filling factor from one @xcite that is in remarkably good agreement with hartree - fock theory @xcite .
the quasiparticle spin measured in transport experiments @xcite are also reasonably well accounted for by hartree - fock calculations @xcite .
electromagnetic absorption experiments @xcite further support that doping away from @xmath4 injects skyrmions into the system . in weaker magnetic fields , near filling factor @xmath5 , early experiments @xcite suggested that spin - polarized quasiparticles are lower in energy than skyrmions , so the effects seen near @xmath4 would not be present at higher filling .
this is consistent with calculations of skyrmion energies near @xmath5 that include finite thickness corrections @xcite , which indicate that skyrmions will be present only at much smaller zeeman couplings than realized in typical experiments .
the size of the skyrmion may be quantified by a number @xmath6 , the difference in the spin component @xmath7 between the skyrmion and the spin - polarized quasiparticle . because of the necessarily small zeeman coupling , stable skyrmions close to @xmath5 have large values of @xmath6 .
they also become unstable with respect to spin - polarized quasiparticles at a finite value of @xmath6 .
( for @xmath4 , @xmath8 as the zeeman coupling reaches the maximum value for which the system supports skyrmions ; i.e. , the skyrmion state smoothly goes into the spin - polarized quasiparticle state . ) for a two - dimensional electron gas ( 2deg ) with width of about 2@xmath9 , where @xmath10 is the magnetic length , the minimum @xmath6 expected @xcite is approximately 4 .
recently however , nmr experiments @xcite have uncovered evidence that some anomalous degrading of spin polarization @xmath11 occur as one dopes away from @xmath5 at relatively high zeeman couplings .
these experiments further indicate that the number of overturned spins per quasiparticle is quite small , @xmath12 .
the simple models usually considered @xcite are inconsistent with this , and one is naturally led to inquire as to what other ingredients might change the critical zeeman coupling and smallest @xmath6 observable near @xmath5 .
two possible answers are landau level mixing and screening by filled landau levels . it should be noted that these are not distinct effects : screening by filled landau levels occurs because they may admix high ( unoccupied ) landau levels to smooth fluctuations due to external potentials and/or inhomogeneous electron densities in partially filled levels .
conversely , the states which may be used for landau level mixing in a partially filled level are limited to those that are not occupied by electrons in other levels , due to pauli exclusion .
thus , a correct treatment of either screening or landau level mixing near @xmath5 must include both these effects . in this work
, we present a method by which these may be incorporated into the hartree - fock description of skyrmion states .
our principal conclusions may be summarized as follows : ( i ) for skyrmions near @xmath4 , landau level mixing tends to lower the quasiparticle energy , although not enough to quantitatively explain the activation energies seen in experiment @xcite .
introduction of a finite width of one magnetic length lowers the energy of the skyrmion by approximately 40% , and inclusion of landau level mixing lowers the energy by approximately another 20% for @xmath13 , where @xmath14 is the cyclotron frequency of the electrons , and @xmath15 is the dielectric constant of the host crystal .
the resulting quasiparticle gap is approximately a factor of 2 larger than what is found in experiment @xcite .
this result agrees qualitatively with that of another study of landau level mixing effects on @xmath4 skyrmions @xcite .
( ii ) for @xmath4 , landau level mixing tends to suppress quasihole - like skyrmions ( i.e. , lowering the maximum zeeman coupling for which they are stable ) , while enhancing the stability of quasielectron ( anti-)skyrmions .
( iii ) for @xmath5 and higher , we find that a sufficiently realistic model of the effective electron - electron interaction , as modified by the finite thickness of the electronic wavefunctions , is necessary to obtain reliable results .
use of a simple potential due to zhang and das sarma @xcite grossly overestimates the stability of @xmath5 skyrmions ; more realistic potentials @xcite allow skyrmions only for very small zeeman couplings .
( iv ) screening and landau level mixing for @xmath5 and @xmath16 tend to lower the energy of spin - polarized quasiparticles more than that of skyrmions , making the latter even less stable .
( v ) the results of ref . can not be understood solely on the basis of hartree - fock states for skyrmions .
the remainder of this article is organized as follows . in section
ii below , we discuss the method used to allow screening and landau level mixing to be included in our calculations .
section iii gives details of our results , and we conclude with a summary of our findings in section iv .
most previous hartree - fock studies of skyrmions have relied on landau level representation of the single particle states @xcite .
we choose instead to construct the wavefunctions in real space .
this enables us to include in the model landau level mixing occurring in weak magnetic fields , without having to expand over the large number of landau levels necessary in the former approach .
we thus trade the calculational convenience of working with the functions given in closed analytic form ( landau levels ) for a closer description of the single - particle states by representing them on a real - space grid . in this calculation
we aim to model all the participating particles , including the ones in the filled levels .
our hartree - fock wavefunction is a slater determinant composed of single - particle states which have @xmath17 as a good quantum number @xcite but whose radial form is to be determined self - consistently : @xmath18 e^{im\theta } .\end{aligned}\ ] ] here @xmath19 and @xmath20 are polar coordinates , @xmath21 is the angular momentum quantum number , and @xmath22 labels different states of the same @xmath21 .
the sign @xmath23 corresponds to two families of solutions , @xmath24 for antiskyrmion ( or quasielectron spin structured solution ) and @xmath25 for skyrmion ( quasihole ) . in very strong magnetic fields , the functions @xmath26 and
@xmath27 take the form expected for landau level states .
when the strength of the magnetic field is lowered to bring the ratio of cyclotron and coulomb energy scales close to 1 , the form of the radial part relaxes toward some modified form , as dictated by the interactions in the system . using the trial form of the wavefunction ( [ trial ] ) , the many - body schrdinger equation with the hamiltonian @xmath28 \nonumber\\ & + & \frac { 1}{2 } \sum_{\sigma \sigma ' } \int d^2r d^2r ' \psi_\sigma^\dagger ( \vec r)\psi_{\sigma ' } ( \vec r ' ) v(\vec r - \vec r ' ) \psi_{\sigma'}^\dagger ( \vec r')\psi_\sigma ( \vec r)\end{aligned}\ ] ] ( where @xmath29 denotes spin and @xmath30 the coulomb interaction ) , upon variation with respect to the functions @xmath31 and @xmath32 , gives a system of mean - field single - particle equations : @xmath33 f_{im}(r)\nonumber \\
& -&\int _ { 0}^{\infty } r'\,dr'\sum _ { m'}v^{ex}_{m - m'}(r , r ' ) \rho _ { m'}^{\uparrow \uparrow } ( r',r ) f_{im}(r ' ) \nonumber \\ & -&\int _ { 0}^{\infty } r'\,dr'\sum _ { m'}v^{ex}_{m - m'}(r , r ' ) \rho _ { m'}^{\downarrow \uparrow } ( r',r ) g_{im\pm 1}(r ' ) \nonumber \\ & = & \epsilon_i f_{im}(r)\end{aligned}\ ] ] together with the analogous equation for the function @xmath27 .
here is a dictionary of the notation accompanying eq .
[ master ] : the operator @xmath34 is given by @xmath35 \nonumber \\ & - & m\frac { \hbar \omega _ c}{2 } + \frac { ( m^*)^2}{4}\frac { \omega _ c^2r^2}{2m^*}\end{aligned}\ ] ] with @xmath36 , @xmath37 the magnitude of the external magnetic field , and @xmath38 the effective mass of the electron .
@xmath0 is the land g - factor , and @xmath39 is the bohr magneton .
@xmath40 s denote generalized densities : @xmath41,\end{aligned}\ ] ] and @xmath42 is the uniform background density . @xmath43 and @xmath44 are the following integrals of the coulomb potential over the azimuthal variable : @xmath45 and , finally , @xmath46 stands for the single - particle hartree - fock energy .
the finite width of the sample is modeled using the form of the in - plane potential due to cooper @xcite : we replace the coulomb interaction @xmath47 in the eq .
( [ potentials ] ) by @xmath48 the symbol @xmath49 denotes the width of system in the direction perpendicular do the plane of the system , and @xmath50 , @xmath51 are coordinates in that direction . to handle the boundaries of the system , we assume the electron states with angular momentum @xmath52 have the ferromagnetic groundstate form ( i.e. landau levels with well defined spin ) . the states with @xmath53 are explicitly included in the calculation . for @xmath6
not too large we find it is sufficient to allow variations of the states with @xmath21 of up to 30 for @xmath54 and up to 50 for @xmath55 . in practice , including boundary electrons from the states with @xmath21 between 31 and 100 ( @xmath54 ) and between 51 and 120 ( @xmath55 ) describes the effect of the system edge with precision matching the rest of the calculation .
understanding eq .
( [ master ] ) as a system of coupled eigenproblems , we look for the self - consistent single particle solutions .
( discretization will turn each eigenequation into a matrix diagonalization problem which can be handled using standard methods . )
the results we thus obtain will be largely presented as comparisons between energies of the spin - polarized quasiparticle and energies of the corresponding skyrmion . to assess the energy of the skyrmion in the region of parameters where it is not stable , we add to the hamiltonian a term of the form @xmath56 , @xmath57 being the spin operator , and @xmath58 a tunable parameter .
this term favors a state with total spin @xmath59 , but is insensitive to the detailed form of the wavefunctions .
this allows the variational scheme to pick out the lowest energy slater determinant of the form given in eq.([trial ] ) within the space of states with the same fixed value of @xmath6 .
based on the calculation described in the previous section , we present some of the results the method allows us to obtain ; we focus mainly on the singly charged excitations in the first three landau levels .
consideration of higher landau levels is also possible , but computation of the potential lookup tables becomes prohibitively expensive , and , as the results so far indicate , leads to no new insight . in the following we shall take the unit of energy to be @xmath60 , and the unitless zeeman splitting to be @xmath61 , where @xmath62 is the electron charge , @xmath15 is the dielectric constant of the host material , @xmath9 is the magnetic length in the field @xmath37 and @xmath63 is the bohr magneton . in fig .
[ lowest ] we show the energy difference between the spin - polarized quasiparticle and the skyrmion of size @xmath6 , @xmath64 .
this quantity is a pure interaction energy ( i.e. zeeman energy is not included ) , and represents the energy gained or lost in deforming a spin - polarized quasiparticle into a skyrmion when zeeman coupling is absent .
of particular importance is the slope ( negative slopes indicate that skyrmion is stable for some value of @xmath0 ) and the curvature ( concave curves will support small - sized skyrmions ) . for concave curves
the largest zeeman splitting that supports skyrmions is the negative of the initial slope of the curve @xcite . for large cyclotron energies our results
are essentially identical to those obtained using the single landau level method @xcite .
note the quasielectron and quasihole excitations are precisely degenerate in this case , due to particle - hole symmetry . for smaller values of @xmath65 ,
the two curves split ; surprisingly , the quasihole skyrmion is suppressed by landau level mixing , whereas the quasielectron skyrmion is enhanced .
( the former result is in agreement with ref . ) .
= 2.5 in the energy gaps ( fig .
[ gap ] ) which result from creation of skyrmion - antiskyrmion pairs when landau level mixing and finite thickness corrections are included are considerably smaller than what is found for two - dimensional layers and no mixing @xcite .
however , the resulting energies are still almost a factor of two larger than what is found in experiment @xcite .
the discrepancy is likely to be due to disorder .
= 2.5 in figs .
( [ first ] ) and ( [ second ] ) present analogous results for @xmath5 and @xmath16 . note the considerably smaller energy scales in these figures , indicating that skyrmions can only be stable ( if ever ) for small values of @xmath66 @xcite .
it is apparent that the introduction of landau level mixing and screening destabilizes the skyrmion .
evidently , spin - polarized particles are better able to take advantage of the admixture of higher landau levels than skyrmions .
= 2.5 in = 2.5 in quasiparticle energies depend on the well width , as illustrated in figs .
[ first173 ] and [ first100 ] . as expected @xcite
, we find that for narrower wells the difference in energy is less favorable for the skyrmion ( fig .
[ first100 ] ) .
note that the width used in fig .
[ first173 ] is close to an experimentally reported value @xcite .
= 2.5 in = 2.5 in it is worth remarking at this juncture that a reasonably realistic model of the electron - electron interaction with finite sample thickness corrections is needed to obtain qualitatively correct results .
[ sarma ] shows that the use of a simpler model potential ( zhang and das sarma @xcite ) @xmath67 which is commonly used in studying quantum hall systems ( see for example refs . and ) gives substantially different results then those presented above ( fig . [ first ] ) .
= 2.5 in the principal difference between @xmath68 and @xmath69 is the behavior at small r ; the former diverges logarithmically , whereas the latter is regular .
other divergent potentials give results consistent with figs . [ first ] and [ second ] ; it is likely that the oversimplified behavior at short distances is responsible for the poor performance of @xmath69 in this problem .
based on results in figs .
[ lowest ] , [ first ] , and [ second ] we can construct the phase diagrams of skyrmion stability for the filling factors of @xmath70 and @xmath71 .
large @xmath72 and small @xmath73 is the region favoring the spin structured excitations .
we see that the region `` shrinks '' as one moves to the higher filling factors . also , according to this calculation the breaking of symmetry between the quasihole and quasielectron excitations upon lowering the cyclotron energy is quite spectacular in the lowest landau level , whereas it plays no significant role in the higher ones .
= 2.5 in in the paper by song _
et al_. @xcite the reported excitation at the parameter values of @xmath74 , and @xmath75 falls well outside the boundary expected from the hartree - fock calculation .
it is tempting to speculate that inclusion of impurity effects can stabilize the skyrmions at values of @xmath0 bigger than allowed in a pure sample . to test this idea
we can include a simple model of an impurity in our calculation : a point charge ( impurity ) at a distance @xmath76 above the central plane of the system .
it is replaced by an effective non - uniform charge density in the plane producing the same potential , @xmath77 the effective density can be found to be @xmath78 for @xmath79 .
results with and without such an impurity are illustrated in fig .
[ imp2nu1 ] .
as may be seen , the impurity favors spin - polarized quasiparticle over skyrmion .
a similar result is expected for a short - range ( e.g. delta - function ) impurity potential .
apparently the simplest models of disorder are not likely to explain the results of ref . .
it is probable that more complicated impurities ( e.g. multiply charged or magnetic ones ) could stabilize the small - spin skyrmions at @xmath54 .
however , in the absence of data indicating such types of disorder in real samples , an investigation of this phenomenon is left for future work .
= 2.5 in = 2.5 in
in this paper we have presented a real - space method for computing hartree - fock states and energies of two - dimensional systems in magnetic fields , appropriate for systems with circular symmetry in which landau level mixing may be important .
the method was applied to compute the effects of landau level mixing and screening on skyrmion states .
it was found that in most cases these tend to destabilize skyrmions , with a notable exception occurring for the case of the quasielectron ( antiskyrmion ) around @xmath4 .
the calculations indicate that hartree - fock states can not account for the results of ref .
( reporting skyrmions at @xmath5 ) .
this is in agreement with earlier studies where landau level mixing and screening were not included .
this work was supported by nsf grant nos.dmr98-70681 and phy94 - 07194 , the research corporation , and the center for computational sciences of the university of kentucky .
lee and c.l .
kane , phys .
lett . * 64 * , 1313 ( 1990 ) .
s.l.sondhi , a.karlhede , s.a.kivelson , and e.h.rezayi , phys . rev .
b * 47 * , 16419 , ( 1993 ) .
h.a.fertig , l.brey , r.ct , and a.h.macdonald , phys .
b * 50 * , 11018 ( 1994 ) . t.h.r .
skyrme , proc .
roy . soc . * a262 * , 233 ( 1961 ) . | we present a hartree - fock study that incorporates the effects of landau level mixing and screening due to filled levels into the computation of energies and states of quasiparticles in quantum hall ferromagnets .
we use it to construct a phase diagram for skyrmion stability as a function of magnetic field and zeeman coupling strengths .
we find that landau level mixing tends to favor spin - polarized quasiparticles , while finite thickness corrections favor skyrmions .
our studies show that skyrmion stability in high landau levels is very sensitive to the way in which electron - electron interactions are modified by finite thickness , and indicate that it is crucial to use models with realistic short distance behavior to get qualitatively correct results .
we find that recent experimental evidence for skyrmions in higher landau levels can not be explained within our model .
0.08 in |
the origin of the martian methane ( ch@xmath1 ) is still poorly understood . despite the fact that the presence of ch@xmath1 remains under debate ( zahnle et al .
2011 ; zahnle 2015 ) , detections have been claimed at the 1060 parts per billion by volume ( ppbv ) level in mars atmosphere from space and ground - based observations at the end of the 90s and during the following decade ( formisano et al . , 2004 ;
krasnopolsky et al . 2004 ; mumma et al .
2009 ; fonti and marzo 2010 ) .
recent observations suggest a ch@xmath1 atmospheric abundance of @xmath010 ppbv , and in some cases no or little ch@xmath1 with an upper limit of @xmath07 ppbv in 20092010 , during mars northern spring ( krasnopolsky 2012 ; villanueva et al .
more recent _ in situ _
measurements performed by mars science laboratory ( msl ) have evidenced variations in the methane detection at the location of gale crater . despite a background level of methane remaining at 0.69 @xmath2 0.25 ppbv , an elevated level of methane of 7.2 @xmath2 2.1 ppbv was evidenced during a timespan of @xmath06 months ( see table 1 of webster et al .
2015 ) , a range of values comparable to the levels observed remotely during the last decade .
because local methane enhancements such as those measured by msl require ch@xmath1 atmospheric lifetimes of less than 1 yr ( lefvre and forget 2009 ) , its release from a subsurface reservoir or an active primary source has widely been discussed in the literature .
a plausible explanation is that ch@xmath1 could have been produced either by hydrothermal alteration of basaltic crust ( lyons et al . 2005 ) or by serpentinization of ultramafic rocks producing h@xmath3 and reducing crustal carbon into ch@xmath1 ( oze and sharma 2005 ; atreya et al .
2007 ; chassefire and leblanc 2011 ; chassefire et al .
2013 ; holm et al .
this hypothesis is supported by the fact that ultramafic and serpentinized rocks have been observed on mars , in particular in the nili fossae region ( brown et al .
2010 ; ehlmann et al .
2010 ; viviano et al .
once formed , methane storage on mars is commonly associated with the presence of hidden clathrate reservoirs .
martian clathrates would form an intermediate storage reservoir in the subsurface that regularly releases methane into the atmosphere ( prieto - ballesteros et al .
2006 ; chastain and chevrier 2007 ; thomas et al . 2009 ; gainey and elwood madden 2012 ; herri and chassefire 2012 ; mousis et al . 2013 , 2015 ) .
however , because clathrates are more likely thermodynamically stable in the martian subsurface and at depths depending on the soil s porosity ( mousis et al . 2013 ) , their existence has never be proven by remote or _ in situ _ observations .
interestingly , it has been recently proposed that halite or regolith could also sequestrate ch@xmath1 on the martian surface ( fries et al .
2015 ; hu et al .
2015 ) , but these mechanisms still need to be thoroughly investigated . here , because of their ability to trap substantial amounts of gases , we suggest that zeolites may form an alternative plausible storage reservoir of methane in the martian subsurface .
spectral evidence for the presence of zeolite has been found on the martian surface ( ruff 2004 ; ehlmann et al .
2009 ; carter et al .
2013 ; ehlmann 2014 ) and there is strong geological case arguing for the presence of this aluminosilicate as part of the martian regolith . in sec .
2 , we explain why chabazite , analcime and clinoptilolite are good candidates to account for the widespread occurrence of zeolites on mars .
we also provide an estimate of the amount of zeolites potentially existing on the planet .
is dedicated to the description of the adsorption properties of chabazite , analcime and clinoptilolite .
the amount of methane potentially trapped in these zeolites in martian conditions is estimated in sec .
is devoted to discussion .
zeolites have been first detected by ruff ( 2004 ) on martian dust using the mars global surveyor ( mgs ) tes spectroscopic observations .
fialips et al .
( 2005 ) then suggested that the water - equivalent hydrogen observed by mars odyssey could be partially stored by zeolite minerals present in the first meters in the martian regolith .
indeed , dickinson and rosen ( 2003 ) observed up to 18 wt% of authigenic chabazite in frozen soils of antarctica ( equivalent to martian conditions ) .
recently , both omega and crism instruments onboard the esa mars express and nasa mars reconnaissance orbiter ( mro ) detected zeolite minerals on the rocky outcrops of several places on mars ( ehlmann et al .
2009 ; carter et al .
2013 , ehlmann 2014 ) . while the first observations on dust and soils suggested a grossly zeolite mineral distribution at mid - latitude , we now have detailed observations revealing the geological / morphological context of zeolite outcrops ( 152 occurrences were detected by carter et al . 2013 ) .
for instance , ehlmann et al .
( 2009 ) claimed the identification of pure analcime ( si - al - na form ) in the deposits in and around the central peaks of two 25-km impact craters nearby nili fossae and isidis .
these peaks would then reflect post - impact hydrothermal alteration ( osinski and pierazzo 2013 ) .
carter et al .
( 2013 ) had a detailed discussion of the issue of their timing of formation and concluded that most hydrous minerals , including zeolites , were formed during the noachian period .
however , they also noticed the presence of zeolites in the younger northern lowlands , probably resulting from ice - volcano interaction . in summary ,
both tes and omega instruments were able to remotely differentiate zeolite spectra from other alteration minerals , namely opal a and saponite formed under similar conditions .
such secondary zeolites result from low temperature aqueous alteration by alkaline brines ( or ice ) of volcanic glass included in pyroclastic or volcanic sedimentary rocks and form authigenic cements in volcanoclastic sandstone .
note that volcanic ash and tephra , the common contributor to sedimentary material on mars , should be widespread , as explosive volcanism on mars is the rule rather than the exception ( grott et al .
however , the resolution of existing infrared spectra remains insufficient to constrain the variety of zeolites that really crystallized on mars . among the possible zeolites ,
chabazite is a good candidate to account for their widespread occurrence on mars .
this mineral is the end product of weathering sequences in a wide range of chemical context ranging from silica - rich to silica - poor volcanic rocks .
chabazite typically forms in chemically open systems , in which transports of soluble ions take place efficiently by flowing vadose water or near - surface ground water ( sheppard and hay 2001 ) . on the other hand , in the closed systems in the martian subsurface ,
more alkali analcime and clinoptilolite should be the major zeolites due to limitation of transports of soluble ions .
also , there are several terrestrial locations where nearly pure analcimes form thick bedding ( several tens of meters ) with wide special extent ( hundreds of kilometers ) ( sheppard and gude 1973 ; whateley et al .
1996 ; deer et al .
2004 ) .
one can provide an estimate of the amount of zeolites potentially existing on mars .
using noachian estimates for the martian crustal thermal flux ( 1220@xmath4c / km ) and thermodynamic data of low - grade metamorphic facies ( @xmath0160220 @xmath4c and from 0 to 3 @xmath510@xmath6 kpa ) , zeolites may be formed at depths ranging from approximately 8 to 1520 km ( e.g. mcsween et al .
this estimate is confirmed by the detection of zeolites near central peaks , independently suggesting that those minerals are indeed present at depths of several kilometers in the crust . assuming this depth range ( 815 km ) , it corresponds to a global volume of 10@xmath7 km@xmath8 and a global equivalent layer ( gel ) reaching @xmath07 km of martian zeolites .
however , the maps of carter et al .
( 2013 ) show that the area where zeolites ( and all hydrous minerals ) were detected by remote sensing is equivalent to the surface of the 045@xmath4s latitudinal band , i.e. about 35% of the surface of mars . if we do not consider temperature constraints but only different thicknesses ( 0.001 to 10 km ) of a 100% zeolite layer at all 045 @xmath4s latitudes , the total volume and gel are in the @xmath05 @xmath5 10@xmath95 @xmath5 10@xmath10 km@xmath8 and @xmath00.353500 m ranges , respectively .
the smallest values may be considered as reasonable estimates ( in the range of @xmath01% of zeolite in crystal clays ; ehlmann et al .
2011 ) , but other geological settings or models can be considered .
for instance , the total volume of possible isolated cylindrical zeolite layers located beneath @xmath0150 impact craters ( carter et al .
2013 ) ranging from 5 to 200 km of diameter ( zeolite layer thickness from 0.1 km to unrealistic 20 km ) may reach @xmath010@xmath8 to 10@xmath10 km@xmath8 ( 0.07 to 700 m gel ) .
these isolated layers may correspond to zeolite minerals formed by post - impact hydrothermal alteration ( osinski and pierazzo 2013 ) .
therefore , it seems that any scenario of zeolite geological generation ( sparse post - impact hydrothermal alteration or crustal global alteration ) can lead to important ranges of volumes / gel .
these values , in particular the most optimistic ones , should not be taken as true quantities , but only as starting reasonable estimates .
indeed crustal porosity and fluids surely decrease the efficiency of zeolite formation at depths : the low water - to - rock ratio would prevent any alteration while secondary mineralizations fill the pores .
in this section , the adsorption selectivity of ch@xmath1 with respect to co@xmath3 is investigated on chabazite , analcime and clinoptilolite .
the common chemical formula of a hydrated chabazite is [ ca@xmath11al@xmath12si@xmath13o@xmath14,(h@xmath3o)@xmath15 .
there exists several forms of chabazite zeolites that differ in their si / al ratio and the nature of cations ( ba@xmath16 , ca@xmath16 , sr@xmath16 , k@xmath17 , na@xmath17 ) , which counterbalance the electric charges .
the framework structure of chabazite is composed of sio@xmath1 and alo@xmath1 tetrahedrons joined by their oxygen atoms
. this arrangement forms primary building units interconnected by secondary building units , as shown in fig .
[ chab1 ] .
the unit cell of chabazite thus contains one large ellipsoidal cavity accessed by six 8-ring windows ( pascale et al .
2002 ) .
the chabazite zeolite gets specific adsorption properties for various molecules having a size smaller than the 8-ring apertures , which gives them access to the ellipsoidal cages . owing to the presence of compensation alkali cations ,
the chabazite is a hydrophilic material .
its adsorption capacity of water is around 0.2 cm@xmath8/g at 257 k ( jnchen et al . 2006 ) .
this zeolite is also able to adsorb co@xmath3 and ch@xmath1 , two molecules of interest for the martian atmosphere .
figure [ chab2 ] shows the adsorption isotherms of co@xmath3 and ch@xmath1 on chabazite at 300 k calculated by monte carlo simulations in the grand canonical ensemble ( gcmc ) ( garcia - perez et al .
2007 ) for pressure ranging below 10@xmath8 kpa .
these simulations are in a very good agreement with some experimental data reported in the literature ( watson et al . , 2012 ; jensen et al . 2012 ) . as would expected , owing to the presence of a quadripolar moment in the carbon dioxide molecule , the chabazite adsorbs more co@xmath3 than ch@xmath1 . in spite of this , the amount of methane adsorbed is not insignificant at 300 k and even better at lower temperature
. it can reach @xmath02 mol / kg against more than 6 mol / kg in the case of co@xmath3 .
this result suggests that the chabazite will selectively adsorb methane and carbon dioxide , with an adsorption in favor of the latter molecule .
the adsorption selectivity of methane with respect to carbon dioxide is defined by the relation : @xmath19 where @xmath20 and @xmath21 are the mole fractions of component i in the adsorbed phase and in the gas phase at equilibrium , respectively .
@xmath22 can be predicted from the adsorption isotherms of single components by means of the ideal adsorbed solution theory ( ias theory ; myers and prausnitz .
when the gas pressure converges towards zero , each single adsorption isotherm exhibits a linear part ( see fig .
[ chab2 ] ) , which corresponds to the henry s law region . in this domain ,
the adsorbed amount of each single component i is proportional to the gas pressure : @xmath23 = @xmath24 . by applying the ias theory to the henry s law region ,
this allows us to derive the adsorption selectivity from the ratio between the henry constants : @xmath25 the ratio of the adsorbed amounts of ch@xmath1 and co@xmath3 can then be related to the ratio of their partial pressures at equilibrium via the relation : @xmath26 henry constants determined from experimental adsorption isotherms at 300 k are given in table [ table_chab ] with the corresponding selectivities . with the values of the adsorption enthalpies found in the literature , the henry constants can be estimated at any temperature relevant to mars conditions by using the vant hoff relation : @xmath27 once the values of @xmath28 have been determined for ch@xmath1 and co@xmath3 at given temperature , it is possible to derive the corresponding adsorption selectivity from eq .
[ eq2 ] .
the theoretical chemical formula of analcime is [ naalsi@xmath3o@xmath11],(h@xmath3o ) .
the structure of analcime , represented in fig .
[ chab1 ] , is very constricted ; the basic sio@xmath1 and alo@xmath1 tetrahedra mutually link to form 4 or 6 membered rings .
the maximum diameter of a sphere that can diffuse throughout this structure is @xmath02.4 .
because the kinetic diameters of co@xmath3 and ch@xmath1 are 3.3 and 3.8 respectively , these two molecules can not be adsorbed in analcime .
only water can be adsorbed in analcime , due to its smaller diameter ( @xmath02.6 ) .
the common chemical formula of clinoptilolite is [ m@xmath11al@xmath11si@xmath30o@xmath14,(h@xmath3o)@xmath12 , where m is a compensation cation easily exchangeable which can be na , k , ca , sr , ba and mg according to the source of minerals ( sand et al .
clinoptilolite has the same framework structure as heulandite .
however it presents a better thermal stability .
the porosity is composed of three sets of intersecting channels , all in the same plane ( fig .
1 ) : a channels with 8-membered rings ( aperture 3 @xmath5 7.6 ) , b channels parallel to a channels with 8-membered rings ( aperture 3.3 @xmath5 7.6 ) and c channels quasi perpendicular to the two others with 8-membered rings ( aperture 2.6 @xmath5 4.7 ) ( baerlocher et al . 2007 ) .
the microporous volume determined by water adsorption is around 0.16 cm@xmath8/g at 298 k. owing to the presence of compensation cations , clinoptilolite exhibits good adsorption affinity towards water , carbon dioxide and methane which , unike analcime , can enter its micropores despite a pore aperture close to the kinetic diameter of these molecules .
as chabazite , this zeolite preferentially adsorbs carbon dioxide compared to methane . at 298 k
, some varieties of clinoptilolite can adsorb more than 3.6 mol / kg of co@xmath3 at room temperature under 10 kpa ( breck et al . 1974 ) while only 0.25 mol / kg of ch@xmath1 under the same conditions ( kouvelos et al .
2007 ) . here
, the determination of the adsorption selectivity of ch@xmath1 with respect to co@xmath3 in clinoptilolite is calculated following the same approach as for chabazite .
adsorption capacities , adsorption enthalpies , henry constants and adsorption selectivities for co@xmath3 and ch@xmath1 have been derived from the experiments of arefi pour et al .
in sec . 3 , we have shown that the application of the henry s law allows to extrapolate the amounts of ch@xmath1 and co@xmath3 trapped in chabazite or clinoptilolite at low pressure range . because the current martian surface atmospheric pressure ( 0.6 kpa ) is located in the validity domain of henry s law ( see the example of fig .
[ chab2 ] given at 300 k ) , this enables us to investigate the amount of ch@xmath1 that would be potentially trapped in martian chabazite or clinoptilolite in contact with an older martian atmosphere at various temperatures , assuming that the methane abundance was higher than today s value at that time .
the adsorption selectivity of methane with respect to carbon dioxide @xmath22 represents the ratio of the ch@xmath1 abundance in chabazite or clinoptilolite to its abundance in the coexisting gas phase at low pressure range .
the evolution of @xmath22 as a function of temperature is illustrated by fig .
[ chab3 ] in the cases of the two zeolites . with values between 2.5 @xmath5 10@xmath31 and 0.169 ( chabazite ) and between 2.5 @xmath5 10@xmath32 and 0.094 ( clinoptilolite ) in the 150300 k range , we find that the ch@xmath1/co@xmath3 ratio increases with higher temperatures in both zeolites , regardless of the initial ch@xmath1-co@xmath3 gaseous mixture
. figure [ chab4 ] represents the evolution of the ch@xmath1/co@xmath3 ratio in the two zeolites as a function of the ch@xmath1/co@xmath3 ratio in the coexisting gas at three temperatures of interest , namely the coldest winter temperature reached in the south pole region ( 150 k ) , and the average night ( 200 k ) and day ( 300 k ) surface temperatures at mid - latitudes .
it shows that the ch@xmath1/co@xmath3 ratio must be in the @xmath010@xmath313 @xmath5 10@xmath33 range at 150 k in the coexisting gas phase to give a value in chabazite matching the ch@xmath1 abundance range measured by msl ( in the @xmath00.257.2 @xmath5 10@xmath34 range ) .
ch@xmath1/co@xmath3 ratios must also exceed @xmath05 @xmath5 10@xmath35 and 4 @xmath5 10@xmath36 in gas to give values in chabazite higher than those measured by msl at 300 and 200 k , respectively .
on the other hand , the ch@xmath1/co@xmath3 ratio in the coexisting gas must be in the @xmath010@xmath373 @xmath510@xmath36 range in clinoptilolite to match the ch@xmath1 abundance range measured by msl at 150 k. interestingly , because @xmath22 is higher in clinoptilolite than in chabazite at temperatures lower than @xmath0270 k , smaller ch@xmath1/co@xmath3 ratios ( @xmath07 @xmath5 10@xmath35 at 300 k and @xmath05 @xmath5 10@xmath37 at 200 k ) are needed to allow this zeolite to match the msl values .
comparisons with models depicting the composition of clathrates potentially existing in the martian subsurface show that chabazite or clinoptilolite can be comparable methane sinks ( i.e. , methane trapping from a methane - containing atmosphere on early mars or from an abiotic source in the crust on early / present mars ) at significantly higher temperatures .
for example , @xmath38 in chabazite or clinoptilolite becomes greater or equal to 0.1 at temperatures reaching 300 k ( see fig . [ chab3 ] ) , whatever the initial ch@xmath1 atmospheric mole fraction .
similar selectivities are achieved in clathrates for ch@xmath1 mole fractions in the 10@xmath3310@xmath39 range but the existence of these structures requires temperatures lower than @xmath0150 k at 0.6 kpa of atmospheric pressure ( mousis et al .
2013 ) , namely the coldest temperature reached during winter in the south pole region .
therefore , scenarios advocating a substantial trapping of volatiles in martian clathrates argue that these ices are buried in the soil at sufficient depth , allowing them to be isolated from the atmosphere and remain stable over long time periods ( chastain and chevrier 2007 ; thomas et al .
2009 ; herri and chassefire 2012 ; chassefire et al .
2013 ; mousis et al .
this scenario also applies to martian chabazite or clinoptilolite , allowing ch@xmath1 to be extracted either from a potentially methane - rich ancient atmosphere or directly from an abiotic source localized in the crust .
indeed , because of their burial in the soil , these zeolites could have preserved the trapped methane over long time periods and create the sporadic releases observed in the atmosphere over the last decade due to impacts , seismic activity or erosion .
an alternative possibility would be to assume that zeolites did continuously remain in equilibrium with the martian atmosphere during the course of its evolution . in this case
, zeolites would not be able to supply any methane to the atmosphere : because the value of @xmath38 is in the @xmath010@xmath3210@xmath40 range between 200 and 300 k , the amount of ch@xmath1 trapped in chabazite and clinoptilolite would be lower than the measured atmospheric levels .
for the same reason , chabazite and clinoptilolite could not act as ch@xmath1 sinks if they remain in contact with the atmosphere . similarly to the proposed trapping scenarios in clathrates , the methane stored in these zeolites could have been produced earlier either via hydrothermal alteration of basaltic crust ( lyons et al . 2005 ) or via serpentinization reactions ( oze and sharma 2005 ; atreya et al .
2007 ; chassefire and leblanc 2011 ; chassefire et al .
2013 ; holm et al .
otherwise , alternative methane sinks should be considered on mars . in order to quantitatively
test the link between the current time presence of ch@xmath1 in the atmosphere and the possible destabilization of zeolites , we can estimate the total gel of zeolites that must be destabilized each second , assuming an initial quantity of trapped ch@xmath1 . in the following ,
we make the assumption that chabazite or analcime are the dominant zeolites on mars . if , instead , analcime is the dominant form , then this material can not be at the origin of the atmospheric ch@xmath1 , due to the small size of its porous network .
two case studies can be envisaged . in the first case
, we assume that ch@xmath1 is trapped in chabazite or clinoptilolite in contact with an ancient martian atmosphere at a surface pressure and temperature of 0.6 kpa and 150 k , respectively .
matching the upper msl value at 150 k requires a ch@xmath1/co@xmath3 ratio of 3 @xmath5 10@xmath33 and 3 @xmath5 10@xmath36 in the gas phase released by chabazite and clinoptilolite , respectively .
these ratios correspond to amounts of trapped ch@xmath1 of @xmath010@xmath39 mol kg@xmath40 in chabazite and @xmath010@xmath36 mol kg@xmath40 in clinoptilolite ( see fig . [ chab4 ] ) .
the total injection flux of ch@xmath1 in the martian atmosphere has been estimated to be 85100 kg s@xmath40 ( mischna et al .
2011 ; holmes et al . 2015 ) .
assuming a mean density of 2000 kg m@xmath32 for chabazite and clinoptilolite , this flux corresponds to at least @xmath05.3 @xmath5 10@xmath41 kg s@xmath40 of zeolites and 5 @xmath5 10@xmath42 m yr@xmath40 as gel .
a more precise calculation would be based on the realistic value of localized surface flux 10@xmath4310@xmath34 kg m@xmath39 s@xmath40 derived by holmes et al .
( 2015 ) , assuming a source within a homogeneous 5@xmath4x5@xmath4 region . in this case
, the corresponding mass of destabilized zeolites would be 6 @xmath5 10@xmath44 kg m@xmath39 s@xmath40 ( chabazite ) and 6 @xmath5 10@xmath45 kg m@xmath39 s@xmath40 ( clinoptilolite ) .
these values correspond to localized layers with thicknesses of @xmath010@xmath3210@xmath40 m for chabazite and @xmath0101000 m for clinoptilolite . in the second case
, we assume that chabazite or clinoptilolite are directly filled by pure methane produced from an abiotic source localized at depth in the crust . here , given the high lithostatic pressure , the henry s law does not apply anymore and single adsorption temperature isotherms of ch@xmath1 derived from experiments or gcmc computations must be used . as a toy example
, we consider the figure [ chab2 ] which shows that at 300 k and 10@xmath8 kpa of gas pressure ( corresponding to a depth of @xmath090 m ) , the amount of ch@xmath1 trapped in chabazite reaches @xmath02 mol kg@xmath40 . assuming a total ch@xmath1 injection flux of 85100 kg s@xmath40 in the martian atmosphere , we find that it corresponds at least to @xmath02650 kg s@xmath40 of chabazite and 2.9 @xmath5 10@xmath37 m yr@xmath40 as gel . using the localized surface flux 10@xmath4310@xmath34 kg m@xmath39 s@xmath40 sourced from a 5@xmath4x5@xmath4 region
, the corresponding mass of destabilized chabazite would be 3 @xmath5 10@xmath46 kg m@xmath39 s@xmath40 , representing a localized layer of @xmath05 @xmath5 10@xmath365 @xmath5 10@xmath33 m yr@xmath40 . in the case of clinoptilolite , assuming an amount of trapped ch@xmath1 10 times smaller than the one estimated for chabazite ( see sec .
3.3 ) , the aforementioned values of gel and localized layer would be increased by the same factor .
when based on realistic values of the methane flux , our calculations show that the second case appears more plausible than the first because it requires amounts of chabazite or clinoptilolite well below those independently quantified from geological constraints .
if the martian methane present in chabazite or clinoptilolite is directly sourced from an abiotic source in the subsurface , the destabilization of a zeolite localized layer of a few millimeters per year at worst may be sufficient to explain the current observations .
our study suggests that if the methane outgassing from excavated chabazite or clinoptilolite prevails over any other source on mars , then the presence of these minerals around gale crater could explain the variation of the ch@xmath1 level observed by msl .
an interesting follow - up of this work would be to investigate the adsorption / desorption efficiencies of other gases in zeolites .
coupling all the data together might lead to predictions of other observable effects in the martian atmosphere that could be used to test the present hypothesis . finally , it is interesting to note that the zeolites adsorption properties depend on their si / al ratios .
the smaller is this ratio , the greater are the compensating cations and the hydrophilic and organophilic properties of the zeolites .
the adsorption selectivity of ch@xmath1 with respect to co@xmath3 also varies according to this ratio and the nature of the cation .
however , in the cases of chabazite and clinoptilolite , there will always be a preferential adsorption of co@xmath3 at the expense of ch@xmath1 , whatever the si / al ratio and the nature of the cation .
this preferential adsorption results from the specific interactions induced by the quadrupole moment caused by the presence of pi electrons in co@xmath3 , which is not the case with ch@xmath1 which puts into action non - specific interactions .
the work contributed by o.m . was carried out thanks to the support of the a*midex project ( n^o^ anr-11-idex-0001 - 02 ) funded by the `` investissements davenir '' french government program , managed by the french national research agency ( anr ) .
we acknowledge support from the `` observatoire des sciences de lunivers '' ( osu ) institut pythas , the `` institut national des sciences de lunivers '' ( insu ) , the `` centre national de la recherche scientifique '' ( cnrs ) and `` centre national detude spatiale '' ( cnes ) and through the `` programme national de plantologie '' and mex / pfs program .
we also thank two anonymous reviewers who made very interesting suggestions that helped us improve our first manuscript .
arefi pour , a. , sharifnia , s. , neishaborisalehi , r. , ghodrati , m. 2015 .
performance evaluation of clinoptilolite and 13x zeolites in co@xmath3 separation from co@xmath3/ch@xmath1 mixture .
j. natural gas sci .
26 , 1246 - 1253 .
brown , a. j. , hook , s. j. , baldridge , a. m. , crowley , j. k. , bridges , n. t. , thomson , b. j. , marion , g. m. , de souza filho , c. r. , bishop , j. l. 2010 .
hydrothermal formation of clay - carbonate alteration assemblages in the nili fossae region of mars .
earth and planetary science letters 297 , 174 - 182 .
carter , j. , poulet , f. , bibring , j .-
p . , mangold , n. , murchie , s. 2013 .
hydrous minerals on mars as seen by the crism and omega imaging spectrometers : updated global view .
journal of geophysical research ( planets ) 118 , 831 - 858 .
fries , m. d. , steele , a. , hynek , b. m. 2015 .
halite as a methane sequestration host : a possible explanation for periodic methane release on mars , and a surface - accessible source of ancient martian carbon .
lunar and planetary science conference 46 , 3017 .
ehlmann , b. l. , mustard , j. f. , murchie , s. l. , bibring , j .-
, meunier , a. , fraeman , a. a. , langevin , y. 2011 .
subsurface water and clay mineral formation during the early history of mars .
nature 479 , 53 - 60 .
ehlmann , b. l. , and 11 colleagues 2009 .
identification of hydrated silicate minerals on mars using mro - crism : geologic context near nili fossae and implications for aqueous alteration .
journal of geophysical research ( planets ) 114 , e00d08 .
fialips , c. i. , carey , j. w. , vaniman , d. t. , bish , d. l. , feldman , w. c. , mellon , m. t. 2005 .
hydration state of zeolites , clays , and hydrated salts under present - day martian surface conditions : can hydrous minerals account for mars odyssey observations of near - equatorial water - equivalent hydrogen ? .
icarus 178 , 74 - 83 .
garca - prez , e. , parra , j. b. , ania , c. o. , garca - snchez , a. , van baten , j. m. , krishna , r. , dubbeldam , d. , calero , s. 2007 . a computational study of co@xmath3 , n@xmath3 , and ch@xmath1 adsorption in zeolites .
adsorption 13 , 469476 .
herri , j .-
m . , chassefire , e. 2012 .
carbon dioxide , argon , nitrogen and methane clathrate hydrates : thermodynamic modelling , investigation of their stability in martian atmospheric conditions and variability of methane trapping .
planetary and space science 73 , 376 - 386 .
holm , n. g. , oze , c. , mousis , o. , waite , j. h. , guilbert - lepoutre , a. 2015 .
serpentinization and the formation of h@xmath3 and ch@xmath1 on celestial bodies ( planets , moons , comets ) .
astrobiology 15 , 587 - 600 .
jensen , n. n. , rufford , t. e. , watson , g. , zhang , d. k. , chan , k. i. , may , e. f. 2012 .
screening zeolites for gas separation applications involving methane , nitrogen , and carbon dioxide .
journal of chemical & engineering data 57 , 106113 .
mumma , m. j. , villanueva , g. l. , novak , r. e. , hewagama , t. , bonev , b. p. , disanti , m. a. , mandell , a. m. , smith , m. d. 2009 . strong release of methane on mars in northern summer 2003 .
science 323 , 1041 .
pascale , f. , ugliengo , p. , civalleri , b. , orlando , r. , darco , p. , dovesi , r. 2002 .
hydrogarnet defect in chabazite and sodalite zeolites : a periodic hartree - fock and b3-lyp study .
journal of chemical physics 117 , 5337 - 5346 .
prieto - ballesteros , o. , kargel , j. s. , fairn , a. g. , fernndez - remolar , d. c. , dohm , j. m. , amils , r. 2006 .
interglacial clathrate destabilization on mars : possible contributing source of its atmospheric methane .
geology 34 , 149 .
sheppard , r. a. , hay , r. l. 2001 .
formation of zeolites in open hydrologic systems . in bish ,
d.l . and ming , d.w .
( eds ) natural zeolites : occurrence , properties , applications , reviews in mineral . and geochem . ,
45 , 261276 .
sheppard , r. a. , gude , a. j. iii 1973 .
zeolites and associated authigenic silicate minerals in tuffaceous rocks of the big sandy formation , mohave county , arizona .
u. s. geological survey , professional paper 830 , 36 pp .
villanueva , g. l. , mumma , m. j. , novak , r. e. , radeva , y. l. , kufl , h. u. , smette , a. , tokunaga , a. , khayat , a. , encrenaz , t. , hartogh , p. 2013 . a sensitive search for organics ( ch@xmath48 , ch@xmath49oh , h@xmath47co , c@xmath47h@xmath50 , c@xmath47h@xmath47 , c@xmath47h@xmath48 ) , hydroperoxyl ( ho@xmath47 ) , nitrogen compounds ( n@xmath47o , nh@xmath49 , hcn ) and chlorine species ( hcl , ch@xmath49cl ) on mars using ground - based high - resolution infrared spectroscopy .
icarus 223 , 11 - 27 .
viviano , c. e. , moersch , j. e. , mcsween , h. y. 2013 .
implications for early hydrothermal environments on mars through the spectral evidence for carbonation and chloritization reactions in the nili fossae region .
journal of geophysical research ( planets ) 118 , 1858 - 1872 .
watson , g. c. , jensen , n. k. , rufford , t. e. , ida chan , k. , may , e. f. 2012 .
volumetric adsorption measurements of n@xmath3 , co@xmath3 , ch@xmath1 , and a co@xmath3 + ch@xmath1 mixture on a natural chabazite from ( 5 to 3000 ) kpa .
journal of chemical & engineering data 57 ( 1 ) , 93101 .
/co@xmath3 ratio in chabazite and clinoptilolite as a function of its ratio in coexisting gas represented at @xmath52 = 150 , 200 and 300 k and 0.6 kpa of total pressure .
the grey area corresponds to the range of ch@xmath1 measurements made so far by msl in the martian atmosphere .
bottom : amount of ch@xmath1 trapped chabazite and clinoptilolite as a function of the ch@xmath1/co@xmath3 ratio in coexisting gas calculated at @xmath52 = 150 , 200 and 300 k and 0.6 kpa of total pressure.,width=340 ] | the origin of the martian methane is still poorly understood .
a plausible explanation is that methane could have been produced either by hydrothermal alteration of basaltic crust or by serpentinization of ultramafic rocks producing hydrogen and reducing crustal carbon into methane .
once formed , methane storage on mars is commonly associated with the presence of hidden clathrate reservoirs . here
, we alternatively suggest that chabazite and clinoptilolite , which belong to the family of zeolites , may form a plausible storage reservoir of methane in the martian subsurface .
because of the existence of many volcanic terrains , zeolites are expected to be widespread on mars and their global equivalent layer may range up to more than @xmath01 km , according to the most optimistic estimates .
if the martian methane present in chabazite and clinoptilolite is directly sourced from an abiotic source in the subsurface , the destabilization of a localized layer of a few millimeters per year may be sufficient to explain the current observations .
the sporadic release of methane from these zeolites requires that they also remained isolated from the atmosphere during its evolution .
the methane release over the ages could be due to several mechanisms such as impacts , seismic activity or erosion .
if the methane outgassing from excavated chabazite and/or clinoptilolite prevails on mars , then the presence of these zeolites around gale crater could explain the variation of methane level observed by mars science laboratory .
mars , mars , atmosphere , mars , surface , mineralogy , astrobiology |
let @xmath8 be a field of characteristic @xmath9 .
let @xmath10 be a reducive algebraic group over @xmath11 , @xmath12 a borel subgroup and @xmath13 the weyl group and its simple reflections .
denote by @xmath14 the bounded derived category of sheaves of @xmath8-vector spaces on @xmath1 constructible along @xmath15-orbits . in @xmath14 there
exist the intersection cohomology sheaves @xmath16 .
these sheaves are a certain extension of the constant sheaf in degree @xmath17 on the bruhat cell @xmath18 to its closure .
let @xmath19 be the hecke algebra of @xmath20 over @xmath21 $ ] normalised so as to satisfy @xmath22 and let @xmath23 be the kazhdan - lusztig basis of @xmath19 .
it satisfies @xmath24 h_x$ ] .
given a finite dimensional graded vector space @xmath25 let @xmath26 be its poincar polynomial .
the character of a sheaf @xmath27 is the element of @xmath19 given by @xmath28 if @xmath8 is of characteristic zero , a theorem of kazhdan and lusztig @xcite says that @xmath29 .
thus the poincar polynomials of the stalks of the intersection cohomology sheaves are given by kazhdan - lusztig polynomials .
it then follows that the same is true in almost all characteristics , however for any given characteristic almost nothing is known .
it is a difficult question to determine over which fields one has @xmath29 and , if not , what these characters are .
it has been known since the original papers of kazhdan and lusztig ( @xcite and @xcite ) that in non - simply laced cases the intersection cohomology complexes may have a different character in characteristic 2 .
( this happens , for example , in the only non - smooth schubert variety in the flag variety of @xmath30 . ) in 2002 braden discovered examples of schubert varieties in simply laced types @xmath31 and @xmath32 where the character of the interesection cohomology sheaf in characteristic 2 is different to all other characteristics ( see the appendix ) . in this article
we define combinatorially a certain subset @xmath33 of _ separated elements _ and show : [ thm - sep ] suppose that @xmath34 , then @xmath29 for any field @xmath8 .
the determination of the characters of @xmath16 is closely related to the decomposition theorem in positive characteristic .
given a simple reflection @xmath35 let @xmath36 be the corresponding standard minimal parabolic subgroup and consider the quotient map @xmath37 if @xmath8 is of characteristic zero , the decomposition theorem implies that @xmath38 is a direct sum of shifts of intersection cohomology sheaves .
this need not be true if @xmath8 is of positive characteristic . given @xmath39 and @xmath35
let @xmath40 be all the parameters of kazhdan - lusztig basis elements that appear in the product @xmath41 .
then we have : [ thm - dec ] suppose that @xmath42 and @xmath43 lie in @xmath44 .
then the decomposition theorem holds for @xmath45 .
whilst being of considerable intrinsic interest , these questions are also important in representation theory .
assume that @xmath8 is algebraically closed , that @xmath46 is greater than the coxeter number of @xmath0 and let @xmath47 be a split semi - simple and simply connected algebraic group over @xmath8 whose root system is dual to that of @xmath10 , and let @xmath48 be a borel subgroup and maximal torus in @xmath47 . fix positive roots
@xmath49 so that the roots corresponding to @xmath15 are those lying in @xmath50 . to each weight
@xmath51 one may associate a standard module @xmath52 which is non - zero if and only if @xmath53 is dominant , in which case it contains a unique simple submodule @xmath54 .
a conjecture of lusztig @xcite expresses the characters of the simple modules @xmath54 in terms of the ( known ) characters of standard modules .
a particular case of the conjecture is the following : let @xmath55 denote the half - sum of the positive roots , and @xmath56 the steinberg weight , then it is conjectured that , for all @xmath57 , @xmath58 = h_{x , y}(1)\ ] ] where @xmath59 $ ] is the kazhdan - lusztig polynomial indexed by @xmath57 .
a theorem of soergel @xcite says that ( [ eq : lusztig ] ) is equivalent to the semi - simplicity of @xmath60 for all @xmath61 and @xmath35 . of course
, in order to apply theorems [ thm - sep ] and [ thm - dec ] it is necessary to know the set @xmath44 .
the essential ingredient in the calculation of @xmath44 is the @xmath0-graph of the coxeter system @xmath20 .
unfortunately , even in simple situations the @xmath0-graph can be very complicated and no general description is known .
however , using fokko du cloux s program _ coxeter _
@xcite it is possible to use a computer to determine the set @xmath44 for low rank weyl groups .
the simplest situation is when @xmath62 .
this only occurs in type a in low rank : [ thm : a6 ] if @xmath10 is of type @xmath2 for @xmath63 then @xmath62 .
hence , in all characteristics the intersection cohomology complexes have characters given by kazhdan - lusztig basis elements and the decomposition theorem holds for @xmath64 for all @xmath35 and @xmath61 .
it also follows that ( [ eq : lusztig ] ) holds for @xmath65 with @xmath5 if @xmath8 has characteristic at least @xmath66 . using results of braden in the appendix , we are also able to extend this theorem to cover @xmath47 of type @xmath32 . in other types and type @xmath2 for @xmath67 our techniques are not as effective . in most examples that we have computed
@xmath44 is not the entire weyl group .
however , we are able to confirm the characters and decomposition theorem for many intersection cohomology complexes in ranks @xmath68 .
it also seems that the elements @xmath69 for which our methods fail will provide an interesting source of future research .
indeed in the appendix braden shows that , both in type @xmath32 and @xmath31 , if @xmath8 is of characteristic 2 , then for minimal elements in @xmath70 one has @xmath71 .
these two examples , combined with the case of dihedral groups , leads one to suspect a close relationship between @xmath70 and those intersection cohomology complexes for which one has @xmath72 for some field of coefficients @xmath8 .
let us briefly mention that , in @xcite braden and macpherson give an algorithm for the calculation of the stalks of the intersection cohomology complexes , using only the fixed points and one - dimensional orbits of a maximal torus on the flag variety .
( this data is encoded in the so - called `` moment graph '' of the flag variety . )
the forthcoming paper @xcite of fiebig and the author extends this result , showing that the moment graph of the flag variety can be used to calculate the characters of parity sheaves , and hence determine those intersection cohomology complexes for which @xmath71 .
thus the results of this paper could ( at least in principle ) be deduced from the moment graph .
in fact , the computations of torsion in the appendix translate easily into the moment graph language and give a proof that the moment graph sheaves corresponding to certain bott - samelson varieties do not split as expected unless @xmath73 is invertible in the coefficient ring
lastly , in this article we only consider field coefficients
. however versions of the intersection cohomology complex exist with coefficients in @xmath74 and the statement that @xmath75 for a field of characteristic @xmath76 is equivalent to the absence of @xmath76-torsion in the stalks and costalks of an intersection cohomology complex over @xmath74 .
the reader is referred to @xcite for more details .
the structure of the paper is as follows . in section [ sec - hecke ]
we review the hecke algebra and kazhdan - lusztig basis in more detail and recall the @xmath0-graph associated to @xmath20 . in section [ sec - sheaves ]
we discuss the _ parity sheaves _
, introduced in @xcite , which are our main theoretical tool . in section [ sec - sep ]
we define the subset @xmath33 and prove theorems [ thm - sep ] and [ thm - dec ] . in section [ sec - comp ]
we discuss the calculation of the sets @xmath44 via computer and give some examples of the sets @xmath44 for low rank weyl groups .
i would like to thank tom braden for useful correspondence , pointing out errors in previous versions , and contributing the appendix .
i would also like to thank daniel juteau and carl mautner for comments on a previous version of this paper .
in this section we recall the hecke algebra and kazhdan - lusztig basis in slightly more detail .
let @xmath20 be a coxeter system with bruhat order @xmath77 and length function @xmath78 .
given @xmath79 we define the _ left _ and _ right descent set _ to be @xmath80 recall that the hecke algebra is the free @xmath81$]-module with multiplication given by @xmath82 the elements @xmath83 are invertible and there is an involution @xmath84 on @xmath19 which sends @xmath83 to @xmath85 and @xmath86 to @xmath87 .
we will call elements fixed by this involution _ self - dual_. there exists a basis @xmath88 of @xmath19 called the _ kazhdan - lusztig basis _ which is uniquely determined by requiring : 1 .
the @xmath89 are self - dual ; 2 .
@xmath90 where @xmath91 and @xmath92 $ ] for @xmath93 .
the polynomials @xmath94 are ( up to a renormalisation ) the _ kazhdan - lusztig polynomials_. one may check , for example , that @xmath95 .
the action of @xmath96 for @xmath35 on the kazhdan - lusztig basis has a particularly simple form .
we denote by @xmath97 the coefficient of @xmath87 in @xmath94 . then :
@xmath98 ( a similar formula describes multiplication by @xmath96 on the left ) .
thus all the information about the action of @xmath96 on the left and right on the kazhdan - lusztig basis may be encoded in a labelled graph , known as the _ @xmath0-graph_.
the vertices correspond to the elements of @xmath0 and are labelled with the left and right descent sets .
there is a directed edge between @xmath99 and @xmath100 if @xmath101 , in which case the edge is labelled with the value of @xmath102 . for more details on the kazhdan - lusztig basis and @xmath0-graphs the reader
is referred to @xcite , @xcite or @xcite .
in this section we briefly recall some basic properties of `` parity sheaves '' introduced in @xcite and motivated by @xcite .
these are our main technical tool .
we recall briefly the setting of @xcite .
we fix a field of coefficients @xmath8 .
all spaces will be complex algebraic @xmath103-varieties , for @xmath103 a complex linear algebraic group .
given an @xmath103-space @xmath104 , we write @xmath105 for the bounded derived category of constructible @xmath8-sheaves on @xmath104 and @xmath106 for the bounded @xmath103-equivariant derived category of constructible sheaves of @xmath8-vector spaces on @xmath104 ( see @xcite ) . by abuse of language , we call objects in @xmath107 complexes .
we denote by @xmath108 the forgetful functor ( see @xcite ) . if @xmath103 has finitely many orbits on @xmath104 then the image of the forgetful functor is contained in @xmath109 , the full subcategory of @xmath105 consisting of sheaves whose cohomology is locally constant along @xmath103-orbits .
the endomorphism ring of any indecomposable object in @xmath107 is local , and hence the krull - schmidt theorem holds in @xmath107 .
all maps will be equivariant morphisms of complex algebraic varieties . given a map @xmath110 we have functors @xmath111 , @xmath112 from @xmath107 to @xmath113 and @xmath114 , @xmath115 from @xmath113 to @xmath107 .
similar functors exist between @xmath105 and @xmath116 . on the categories @xmath117 and @xmath116 we have the verdier duality functor , which we denote by @xmath118 .
we have isomorphisms of functors @xmath119 and @xmath120 .
all functors @xmath121 commute with the forgetful functor .
now let @xmath10 denote a reductive complex algebraic group and @xmath122 a borel subgroup and maximal torus .
let @xmath0 denote the weyl group and @xmath123 the set of simple reflections . throughout @xmath124 , where @xmath125 is either @xmath15 or a minimal standard parabolic subgroup @xmath36 corresponding to @xmath35 ( i.e. @xmath126 ) .
we regard @xmath104 as a @xmath15-variety .
each @xmath15-orbit is isomorphic to an affine space and the strata are classified by @xmath0 if @xmath127 and @xmath128 if @xmath129 . given @xmath39 ( resp .
@xmath130 ) we denote by @xmath131 ( resp .
@xmath132 ) the stratum @xmath18 ( resp .
@xmath133 ) , by @xmath134 ( resp .
@xmath135 ) its inclusion and by @xmath136 ( resp .
@xmath137 ) in the @xmath15-equivariant constant sheaf on @xmath131 ( resp .
@xmath138 ) . for brevity , @xmath139 . a complex @xmath140 is _ @xmath141-even _ if @xmath142 ( resp .
@xmath143 ) is isomorphic to a direct sum of constant sheaves concentrated in even degrees , for all strata @xmath144 ( resp .
@xmath145 ) .
a complex is _ even _ if it is both @xmath146- and @xmath147-even .
a complex @xmath148 is _ ( @xmath141- ) odd _ if @xmath149 $ ] is ( @xmath141- ) even .
a complex @xmath140 is ( ? )
-_parity _ if we have an isomorphism @xmath150 with @xmath151 ( ? ) -even and @xmath152 ( ? ) -odd .
note that direct sums and summands of ( ? ) -parity sheaves are ( ? ) -parity .
an indecomposable parity complex is called a _
parity sheaf_. the following theorem shows that one may classify parity sheaves on the flag variety in a similar way to intersection cohomology complexes : [ thm - clas ] there exists ( up to isomorphism ) a unique parity sheaf @xmath153 with support contained in @xmath154 and @xmath155 $ ] .
each @xmath156 is self - dual and any indecomposable parity complex is isomorphic to @xmath157 $ ] for some @xmath39 and @xmath158 .
given a @xmath146-parity complex @xmath159 and @xmath39 we may write @xmath160 for some finite dimensional graded @xmath8-vector space @xmath161 . we define the character of @xmath162 in the hecke algebra to be : @xmath163 given a @xmath146-parity complex @xmath162 it is easily seen that @xmath164 agrees with the character of @xmath165 as defined in the introduction .
a similar character map exists for @xmath166 for a simple reflection @xmath35 .
let @xmath167 denote the left ideal @xmath168 in @xmath19 .
then @xmath167 is free with basis @xmath169 for @xmath170 , where @xmath171 denotes the subset of elements @xmath39 such that @xmath172 . given a @xmath146-parity complex @xmath173 and @xmath170 we can write @xmath174 for some graded @xmath8-vector space @xmath175 . we define the character of @xmath162 to be : @xmath176 for any @xmath35 we have obvious maps given by inclusion and multiplication : @xmath177^ { \cdot \h_s } & \ar@(dl , dr)[l]^{\operatorname{inc } } \he_s}\ ] ] the quotient map @xmath178 induces functors : @xmath179^ { \pi_{s * } } & \ar@(dl , dr)[l]^{\pi_s^ * } d^b_{b}(g/{p_s})}\ ] ] the following lemma is well - known ( see @xcite , lemme 2.6 ) : [ lem - parchar ] 1 . if @xmath159 is @xmath146-parity , then so is @xmath180 and @xmath181 2 .
if @xmath182 is @xmath146-parity , then so is @xmath183 and @xmath184 ) = \operatorname{inc}(\operatorname{ch}(\se)).\ ] ] 3 . if @xmath159 or @xmath185 is @xmath146-parity then @xmath186 ) \cong v^{-1 } \operatorname{ch}(\se).\
] ] we first show the character relations .
statement ( 3 ) is a straightforward consequence of the definitions and ( 2 ) follows from the definitions and the fact that @xmath187 for @xmath39 .
it remains to show ( 1 ) .
we prove ( 1 ) by induction on the number of @xmath39 for which @xmath188 .
if this number is one , then ( by definition of @xmath146-parity ) @xmath162 is necessarily a direct sum of shifts of @xmath189 , for some @xmath79 .
we may assume that @xmath190 $ ] .
let us write @xmath191 for the image of @xmath42 in @xmath192 . if @xmath193 then @xmath194 restricts to an isomorphism @xmath195 .
hence @xmath196.\ ] ] if @xmath197 then the restriction of @xmath194 to @xmath131 induces a ( trivial ) @xmath198-bundle over @xmath132 , hence @xmath199.\ ] ] a simple calculation in the hecke algebra then shows that in both cases @xmath200 as claimed .
we now turn to the general case .
we may assume without loss of generality that @xmath162 is @xmath146-even .
choose @xmath39 so that @xmath131 is open in the support of @xmath162 and let @xmath201 denote the inclusion .
then @xmath202 and we have a distinguished triangle of @xmath146-even sheaves @xmath203}{\to}\ ] ] by induction @xmath204 applied to the first or third terms is @xmath146-parity and ( 1 ) holds .
it follows that the same is true of @xmath162 because @xmath205 and @xmath206 it remains to see that @xmath207 and @xmath208 preserve the classes of parity sheaves .
however this follows immediately because @xmath209 ( as @xmath210 is proper ) and @xmath211 ) \mathbb{d } \cong \mathbb{d } ( \pi^*[1])$ ] ( because @xmath210 is a smooth fibration with fibres of complex dimension 1 ) .
consider @xmath10 as a @xmath212-space via @xmath213 .
as the second copy of @xmath15-acts freely on @xmath10 , the quotient equivalence ( @xcite ) yields an equivalence of triangulated categories : @xmath214 consider the inversion map @xmath215 .
then this is @xmath216-equivariant with respect to the swap map @xmath217 .
this induces an equivalence @xmath218 consider the functor @xmath219 .
then @xmath220 commutes with @xmath118 .
it is easy to see @xmath220 preserves parity complex and that , for a parity complex @xmath159 , @xmath221 where @xmath222 is the anti - involution @xmath223 .
define endofunctors on @xmath224 by @xmath225 \text { and } \ver_s ( - ) : = \iota^ * \pi_s^ * { \pi_s } _ * \iota^*(-)[1].\ ] ] then the functors @xmath226 and @xmath227 preserve parity sheaves ; the shift is chosen so that @xmath228 and @xmath229 . by and the above lemma , @xmath230 for parity sheaves @xmath159 .
the first result about the characters of parity sheaves is the following : [ cor - dual ] for all @xmath39 , @xmath231 is self - dual .
we proceed via induction on @xmath232 with the base case being trivial .
let us fix @xmath42 and choose @xmath35 with @xmath233 .
by theorem [ thm - clas ] we may write @xmath234 where @xmath235 ^{\oplus m_{x,\eta}}.\ ] ] the verdier self - duality of @xmath236 and each @xmath237 for @xmath238 together with krull - schmidt implies @xmath239 by induction , the @xmath240 for @xmath238 are self - dual .
hence both @xmath241 and @xmath242 are self - dual .
thus so is @xmath243 .
let @xmath35 be a simple reflection .
the next proposition relates parity sheaves on @xmath244 to those on @xmath1 : [ cor : ps ] let @xmath39 be such that @xmath197 and denote by @xmath191 the image of @xmath42 in @xmath128 .
we have isomorphisms @xmath245 \cong \se(w)\ ] ] and @xmath246 \oplus \se({\overline{w}})[1].\ ] ] as @xmath156 is a direct summand of @xmath247 $ ] we have @xmath248 it follows ( by considering @xmath249 ) that @xmath250 \oplus \se({\overline{w}})[-1 ] \oplus \sg\ ] ] for some parity complex @xmath251
. we may also decompose @xmath245 \cong \se({w } ) \oplus \sg^{\prime}.\ ] ] hence @xmath252 \cong \se({\overline{w}})[1 ] \oplus \se({\overline{w}})[-1 ] \oplus \sg \oplus { \pi_s } _ * \sg^{\prime}.\ ] ] however , by lemma [ lem - parchar ] , @xmath253 ) = ( v + v^{-1 } ) \operatorname{ch } ( \se(\overline{w}))\ ] ] and so @xmath254 .
hence @xmath251 and @xmath255 are zero .
in @xmath14 there exist the intersection cohomology sheaves @xmath16 which may be defined as @xmath256 $ ] ( see @xcite ) .
the intersection cohomology complexes admit equivariant lifts @xmath257 which are uniquely determined ( up to isomorphism ) by requiring @xmath258 ( see @xcite )
. we will need the following in the next section : [ prop : ic ] for @xmath39 , @xmath259 if and only if @xmath260 .
hence if @xmath261 then @xmath262 .
recall that @xmath16 is the unique complex in @xmath14 satisfying the four conditions : 1 .
@xmath16 is verdier self - dual ; 2 .
@xmath263 for @xmath264 ; 3 . @xmath265 $ ] ; 4 .
@xmath266 is concentrated in degrees @xmath267 for @xmath238 .
note that 1 ) , 2 ) and 3 ) are always satisfied by @xmath268 .
now , if @xmath269 then ( 4 ) is also satisfied and so @xmath270 .
on the other hand , if @xmath271 then @xmath243 is a self - dual and positive combination of kazhdan - lusztig basis elements containing more than one term ( again by proposition [ cor - dual ] ) .
it is straightforward to see that if this is the case then @xmath268 can not satisfy ( 4 ) .
let @xmath272 be a set of representatives of isomorphism classes of indecomposable parity sheaves on @xmath1 .
we would like to investigate when their characters are equal to the kazhdan - lusztig basis .
the set @xmath273 yields a self - dual basis of @xmath274 with certain positivity properties which are shared by the kazhdan - lusztig basis . in this section
we investigate to what extent these properties already determine the basis . as motivation ,
let us consider some examples : 1 . _
( simplistic ) _ let @xmath61 and suppose that @xmath275 , @xmath276 and @xmath277 . we know that @xmath278 is a direct summand of @xmath279 $ ] with self - dual character .
hence @xmath280 and so @xmath281 by the uniqueness of the kazhdan - lusztig basis .
( more realistic ) _ fix @xmath61 and suppose that we can show that @xmath282 for all @xmath283 .
suppose further that the only kazhdan - lusztig basis element that appears with non - zero coefficient in all expressions @xmath284 with @xmath285 is @xmath286 .
then , using the fact that @xmath237 occurs as a direct summand in @xmath287 $ ] for all @xmath283 , it follows that @xmath288 . in the above examples it is essential to know how the elements @xmath96 for @xmath35 act on the kazhdan - lusztig basis
this is precisely the information provided by the @xmath0-graph discussed in section [ sec - hecke ] .
we start with some definitions .
given an element @xmath289 we define the _ kl - support _ to be the set : @xmath290 we say that @xmath291 is _ kl - supported in degree 0 _ if , in addition , all @xmath292 .
given @xmath293 we may write the difference @xmath294 in the standard basis as @xmath295 if all @xmath296 $ ] we write @xmath297 . note that if @xmath162 is a direct summand of a parity complex @xmath251 then @xmath298 .
[ support - lemma ] suppose @xmath162 is a direct summand of a parity complex @xmath251 whose character is kl - supported in degree 0 .
then the character of @xmath162 is also kl - supported in degree zero and @xmath299 because @xmath241 is kl - supported in degree 0 , it is isomorphic to a direct sum of indecomposable parity sheaves @xmath156 without shifts . moreover , each @xmath156 that occurs is kl - supported in degree zero .
hence @xmath162 is also isomorphic to a direct sum of such indecomposable parity sheaves and the lemma follows . a more conceptual proof of the above lemma is provided by the following ( which will not be needed below ) .
the character of a self - dual parity complex @xmath162 is kl - supported in degree zero if and only if @xmath162 is perverse .
now if @xmath251 is as in the lemma it is perverse and splits into a direct sum @xmath300 ^{\oplus m_{x,\eta}}.\ ] ] however we must have @xmath301 if @xmath302 as each direct summand must be perverse .
hence any direct summand is self - dual and perverse .
given a subset @xmath303 we denote by @xmath304 we now define a function @xmath305 from some subset @xmath306 ( to be defined below ) to the power set of @xmath0 .
this function and its domain are defined inductively as follows : 1 .
suppose we have defined @xmath308 on all @xmath309 .
then @xmath99 belongs to @xmath310 if there exists @xmath311 or @xmath312 such that either @xmath313 in this case we define @xmath314 to be the set : @xmath315 [ ex - dih ] let @xmath0 be a dihedral group : @xmath316 if @xmath317 ( resp .
@xmath318 ) is not the longest element then : @xmath319 for the longest element @xmath320 one has @xmath321 .
if @xmath308 is defined on @xmath61 and @xmath322 we say that @xmath99 is _
separated_. the set of all separated elements will be denoted by @xmath44 . if @xmath0 is a dihedral group @xmath323 then it follows from above that the separated elements are @xmath324 .
in particular , @xmath325 and @xmath326 are the only rank two weyl groups in which @xmath62 .
one may show ( for example using the results of @xcite ) that if @xmath10 is a kac - moody group having weyl group @xmath0 a dihedral group @xmath323 , then the elements @xmath327 are precisely those for which @xmath29 in all characteristics .
theorem [ thm - sep ] in the introduction is an immediate consequence ( using proposition [ prop : ic ] ) of the following : [ prop - sup ] if @xmath308 is defined on @xmath61 then @xmath240 is kl - supported in degree 0 and @xmath328 in particular , for all @xmath34 we have @xmath277 .
clearly @xmath329 and so we may assume by induction that @xmath330 for all @xmath331 , where @xmath61 is some element on which @xmath308 is defined . without loss of generality we may assume , by the inductive definition of @xmath308 above , that there exists some @xmath311 so that @xmath332 .
hence @xmath333 is kl - supported in degree zero .
we may now apply lemma [ support - lemma ] to conclude that @xmath240 is kl - supported in degree zero and that @xmath334 however , as @xmath240 is kl - supported in degree zero and is a direct summand of all @xmath335 for all @xmath336 we conclude that @xmath337 an identical argument applies on the right .
the proposition then follows by intersecting these conditions .
the following remains from the introduction : under the assumptions of the theorem , @xmath259 and @xmath338 splits as a direct sum of parity sheaves , all of which are isomorphic to equivariant intersection cohomology complexes .
we are then done by the above proposition , combined with proposition [ prop : ic ] and proposition [ cor : ps ] .
in this section we give some examples of the sets @xmath33 for low rank weyl groups . as is clear from the definition of @xmath308 , the only piece of information needed to calculate the @xmath44 and @xmath308 is the @xmath0-graph of @xmath20 .
however , no general description of the @xmath0-graph is known ( for descriptions of some subgraphs see @xcite and @xcite and for a description of the computational aspects of the problem see @xcite and @xcite ) .
thus , in order to calculate @xmath308 and @xmath44 we have to restrict ourselves to examples .
this involves two steps : 1 .
calculation of the @xmath0-graph of @xmath20 , and 2 .
calculation of the function @xmath308 using the @xmath0-graph .
step 1 ) is computationally quite difficult , especially when the weyl group is large .
luckily there exists the program _ coxeter _ written by fokko du cloux @xcite , which calculates the @xmath0-graph very efficiently .
step 2 ) is then relatively straightforward .
a crude implementation in magma ( whose routines for handling coxeter groups proved very useful ) as well as the @xmath0-graphs obtained from _ coxeter _ are available at : http://people.maths.ox.ac.uk/williamsong/torsion/ this site also contains a complete description of the sets @xmath44 and @xmath308 for all weyl groups of ranks less than 6
. we will now describe examples of the sets @xmath44 .
the values of @xmath308 on values not in @xmath44 may be found on the above web page . here
thus all intersection cohomology complexes with coefficients of any characteristic have the same characters as in characteristic zero and the decomposition theorem is always true .
this is the statement of theorem [ thm : a6 ] .
let @xmath340 with coxeter generators @xmath341 with @xmath342 corresponding to the simple transpositions @xmath343
. in @xmath0 , 38 of the 40 320 elements do not belong to @xmath44 .
the elements which do not lie in @xmath44 break up naturally into five groups , which we now describe .
consider the following elements of @xmath0 : @xmath344{perm7a } \end{array } , \quad w_2 = \begin{array}{c } 67823451 \\
\includegraphics[width=2cm]{w2 } \end{array } , \quad w_3 = \begin{array}{c } 84567123 \\
\reflectbox{\includegraphics[width=2cm]{w2 } } \end{array } \\
w_4 = \begin{array}{c } 62845173 \\
\includegraphics[width=2cm]{perm4a } \end{array } \ ; \text { and } \ ; w_5 = \begin{array}{c } 84627351 \\
\includegraphics[width=2cm]{perm1a } \end{array}. \end{array}\ ] ] the first group consists of @xmath345 the second group consists of @xmath346 ( note that @xmath347 is maximal in @xmath348 ) .
the third group @xmath349 is obtained from @xmath348 by inversion ( or by applying the automorphism @xmath350 ) .
it contains @xmath351 as a maximal element .
the fourth group consists of @xmath352 the fifth group consists of @xmath353 it would be interesting to investigate the intersection cohomology complexes corresponding to the minimal elements in @xmath354 for @xmath355 directly .
note that the set @xmath356 has already arisen in kazhdan - lusztig combinatorics ; these are the so - called `` hexagon permutations '' of billey and warrington ( see @xcite , the name refers to a characteristic hexagon shape appearing in their heap representation ) .
we say that a permutation @xmath357 contains the pattern of a permutation @xmath358 if there is a collection of indices @xmath359 so that @xmath360 are in the same relative order as @xmath361 .
otherwise , we say that @xmath42 avoids the pattern @xmath362 . for a more general notion of pattern avoidance which works for general coxeter groups ,
. a geometric interpretation of pattern avoidance is given in @xcite .
the significance of hexagon permutations is explained by the following result : let @xmath363 be a reduced word for an element @xmath357 .
then @xmath364 is a small map if and only if the permutation @xmath42 avoids the pattern @xmath365 and the four hexagon permutations .
the bott - samelson resolutions of the hexagon permutations are semi - small , but not small . in the appendix
, braden treats the hexagon permutations in detail and shows that @xmath271 if @xmath8 is of characteristic 2 .
consider @xmath367 with generators @xmath368 & t\ar@2{-}[r ] & v}.\ ] ] this is the first group on which @xmath308 is not defined everywhere ; it is not defined on the element @xmath369 .
this means that there may be a characteristic in which the parity sheaf corresponding to @xmath369 is not perverse .
of the 48 elements of @xmath0 , 20 do not lie in @xmath44 .
they are : @xmath370 . in @xmath371 , which contains 384 elements , @xmath308 is not defined on 8 elements , and 180 elements do not lie in @xmath44 . in @xmath372 , which contains 3840 elements ,
@xmath308 is not defined on 26 elements and 1696 elements do not lie in @xmath44 .
we label our generators @xmath373 , @xmath374 , @xmath375 and @xmath86 of @xmath0 as follows : @xmath376 \\ s \ar@{-}[r ] & t\\ & & v\ar@{-}[ul]}\ ] ] here , @xmath308 is defined on all of @xmath0 and 7 elements do not belong to @xmath44 .
let @xmath377 be the automorphism of @xmath0 mapping @xmath378 .
the elements not in @xmath44 are @xmath379 , @xmath380 and @xmath381 as well as @xmath382 , @xmath383 , @xmath384 and @xmath385 . in the appendix
, braden discusses the case of @xmath347 in more detail .
in fact , in this example one may extend the techniques of section [ sec - sep ] to deduce combinatorially that one always has @xmath386 clearly it is enough to see this for @xmath387 . as @xmath388 is separated , @xmath389 occurs as a direct summand of @xmath390 .
but as @xmath391 is separated and @xmath392 we conclude that @xmath393 . in @xmath394
, @xmath308 is not defined on one element , and 176 of the 1920 elements in @xmath0 do not lie in @xmath44 . in @xmath395 , which contains 23040 elements ,
@xmath308 is not defined on 33 elements and 3259 elements of @xmath0 do not lie in @xmath44 . in @xmath396 , @xmath308 is not defined on 23 elements and 621 of the 1152 do not lie in @xmath44 . in this case
we have already calculated @xmath44 in example [ ex - dih ] . here
we obtain nothing new . if @xmath398 then @xmath399 .
however these schubert varieties are smooth , and so the @xmath400 in any characteristic if @xmath327 .
let us briefly describe how the above algorithm may be taken further with some of the geometric input contained in the appendix .
for example , if @xmath0 is of type @xmath31 then , with notation as in [ subsec : a7 ] , the calculations in the appendix allow one to conclude that @xmath401 if @xmath402 .
then the above algorithm may be used to deduce that @xmath269 for all hexagon permutations @xmath42 . similarly in @xmath32 if one knows , with notation as in [ subsec : d4 ] , that @xmath403 then it follows that @xmath404 , @xmath405 and @xmath406 all have characters given by kazhdan - lusztig basis elements . by the calculations of the appendix
this occurs if and only if @xmath402 .
it follows that lusztig s conjecture is true around the steinberg weight in type @xmath32 .
one can also turn these arguments around to deduce that one has @xmath271 ( and hence @xmath71 ) for families of elements , once one knows that it occurs once .
for example , if @xmath0 is of type @xmath32 and one knows that @xmath407 ( as is the case in characteristic 2 ) one can not have @xmath408 for @xmath409 .
indeed , assume that @xmath410 .
then , as @xmath411 is separated @xmath412 occurs as a direct summand of @xmath413 .
but @xmath414 and so the calculations of [ subsec : d4 ] show that @xmath415 .
similar arguments apply in the other cases .
similarly one may show in type @xmath31 that if @xmath416 for one hexagon permutation then this is true for all of them . a geometric explanation for this is given in remark [ rem : otherhex ] of the appendix .
_ _ by tom braden _
_ let @xmath10 be a semisimple complex group , and fix a choice of a borel subgroup @xmath15 , maximal torus @xmath417 , and opposite borel subgroup @xmath418 .
let @xmath419 be the corresponding sets of roots and positive roots , chosen so that the weights of @xmath420 acting on @xmath421 are @xmath422 .
let @xmath423 denote the weyl group and @xmath123 the set of simple reflections .
let @xmath424 be the length function , and @xmath77 the bruhat - chevalley order on @xmath0 .
consider the flag variety @xmath425 ; @xmath10 acts on @xmath104 by left multiplication . the set of @xmath417-fixed points is in bijection with @xmath0 by @xmath426 , where @xmath427 is any lift of @xmath42 to @xmath10
. using this bijection , we abuse notation and refer to points in @xmath428 and elements of @xmath0 by the same symbols .
the flag variety @xmath104 has two decompositions by bruhat cells and dual bruhat cells @xmath429 , where @xmath430 and @xmath431 .
the notation @xmath432 is chosen to indicate that it should be thought of as a normal slice at @xmath42 to the stratification @xmath433 .
it is also easy to describe the one - dimensional @xmath417-orbits in @xmath104 .
the closure of a one - dimensional orbit is a closed irreducible @xmath417-invariant curve ; we will refer to such curves as `` @xmath417-curves '' for short . for any positive root @xmath434 ( simple or not ) ,
let @xmath435 denote the corresponding reflection .
[ flag 1d orbits ] for any @xmath434 and any @xmath39 , there is a unique @xmath417-curve @xmath436 which contains @xmath42 and @xmath437 , and all @xmath417-curves are of this form .
the weight of the action of @xmath417 on the tangent space @xmath438 is @xmath439 . since @xmath440 , the above formula for the tangent weight
can also be given , up to sign , by saying that if @xmath436 is the @xmath417-curve joining @xmath42 and @xmath441 for some @xmath442 , then the @xmath417-weight of @xmath443 is @xmath444 .
the sign can then be specified by noting that the weight is in @xmath445 if and only if @xmath446 .
let @xmath447 be a sequence of simple roots , not necessarily distinct , and let @xmath448 be the corresponding sequence of simple reflections : @xmath449 . put @xmath450 .
then @xmath363 is a reduced word for @xmath42 if @xmath451 . for each simple reflection @xmath341 ,
let @xmath452 be the corresponding minimal parabolic containing @xmath15 , whose lie algebra is @xmath453 .
the bott - samelson variety @xmath454 associated to @xmath363 is defined to be the quotient @xmath455 where @xmath456 acts on @xmath457 on the right by @xmath458 let @xmath459 $ ] denote the point of @xmath460 corresponding to @xmath461 .
we wish to describe the set @xmath467 of @xmath417-fixed points .
let @xmath468 .
for each @xmath469 , define @xmath470 and @xmath471 by @xmath472 , \ ; { \mathbf{w}}^{{\varepsilon}}= s_1^{{{\varepsilon}}(1)}\cdots s_l^{{{\varepsilon}}(l)}.\ ] ] for any @xmath473 and any @xmath462 , define @xmath474 = ( { { \varepsilon}}(1),\dots,{{\varepsilon}}(k),0\dots,0 ) \in d_l$ ] . as usual , we will refer to elements of @xmath0 and points of @xmath428 by the same symbols .
let @xmath482 .
let @xmath483 denote the longest element in @xmath0 .
then there are five @xmath417-fixed points in @xmath484 , namely @xmath485 , where @xmath486 let us denote these five elements of @xmath394 by @xmath487 .
there exists a @xmath417-curve containing @xmath488 and @xmath489 if and only if @xmath490 . the tangent weights of the curves joining @xmath488 to @xmath491 for @xmath492 are @xmath493 , @xmath494 , @xmath480 , @xmath495 , and @xmath481
. the composition of @xmath496 with the projection @xmath497 restricts to a birational map @xmath498 which identifies @xmath484 with the blow - up of @xmath499 at two points .
the exceptional fibers are the @xmath417-curves joining @xmath500 to @xmath501 and @xmath502 to @xmath503 .
the one - dimensional @xmath417-orbits of @xmath504 are more difficult to classify than the fixed points . unlike the flag variety @xmath1 , bott - samelson varieties generally have infinitely many @xmath417-curves . we will describe a collection of @xmath417-curves which span the tangent space at each fixed point , but there are in general many other @xmath417-curves . denote the standard basis of @xmath476 by @xmath505 , where @xmath506 is the kronecker @xmath507-function .
for any @xmath473 and @xmath508 , we have a @xmath417-curve joining @xmath485 and @xmath509 , namely @xmath510\mid x'_j = x_j , j\ne i\},\ ] ] where @xmath511 $ ] .
this curve projects under @xmath210 to the @xmath417-curve in @xmath1 which joins @xmath512 and @xmath513 , and so the tangent weight of this curve at @xmath485 is @xmath514}(\alpha_i)$ ] . note that @xmath417-curves which project down to fixed points , such as the ones in example [ sl3 example ] , are not of this type
besides their definition as @xmath15-orbits , the bruhat cells @xmath433 in the flag variety @xmath104 can can also be described as bialynicki - birula cells for the action of a dominant cocharacter @xmath515 . for any @xmath39
, we have @xmath516 the bott - samelson variety @xmath504 will not in general have finitely many @xmath15-orbits , but we can still consider its biaynicki - birula cells ( for the same cocharacter @xmath517 ) . given @xmath473 , we define @xmath518 [ bb cells ] the dimension of the bialynicki - birula cells is given by @xmath519}(\mu_k ) \in
-\phi^+\}\\ = \#\{1 \le k \le l \mid \ell({\mathbf{w}}^{{{\varepsilon}}[k ] } ) > \ell({\mathbf{w}}^{{{\varepsilon}}[k]}s_k)\}. \end{split}\ ] ] the cell @xmath520 fibers linearly over @xmath131 , @xmath521 , with fiber of dimension @xmath522}(\mu_k ) \in -\phi^+\}\\ = \#\{1 \le k \le l \mid \ell({\mathbf{w}}^{{{\varepsilon}}[k-1 ] } ) > \ell({\mathbf{w}}^{{{\varepsilon}}[k-1]}s_k)\}. \end{split}\ ] ] these cells give a paving by affines of the fiber @xmath523 .
fix a word @xmath363 as above , and let @xmath524 , an object in @xmath525 .
we want to understand when does the decomposition theorem hold for @xmath526 , i.e. when does it split as a direct sum of ( shifted ) intersection cohomology sheaves . put @xmath527
. we will be interested in the natural homomorphism @xmath528 of cohomology groups . because @xmath210 is proper ,
this is the same as the map obtained by applying hypercohomology to the adjunction morphism @xmath529 , where @xmath530 is the inclusion . on the other hand @xmath531
is an isomorphism if @xmath532 and is zero if @xmath533 , so if the decomposition theorem holds for the projection @xmath210 with @xmath534 coefficients , then for any @xmath535 the cokernel of @xmath536 will have no torsion . when the resolution @xmath210 is semi - small , this map was previously considered by @xcite , in the guise of an intersection form on the top borel - moore homology of @xmath537 with @xmath534 coefficients . in
the semi - small case @xmath536 is an isomorphism modulo torsion , and it is an isomorphism if and only if the decomposition theorem holds .
when @xmath210 is not semi - small we do not know to what extent the lack of torsion in @xmath538 implies the splitting of @xmath526 . because there is a cocharacter of @xmath417 which contracts @xmath541 onto @xmath362
, we have an isomorphism between @xmath542 and @xmath543 . using the properness of @xmath210 ,
this is the same as the cohomology of the fiber @xmath537 . since this fiber has a paving by affines , its cohomology is torsion free .
the torsion freeness of the domain of the map @xmath539 follows from lefschetz duality and the following lemma .
let @xmath545 , @xmath546 be the unipotent parts of @xmath15 , @xmath418 .
let @xmath547 .
then @xmath548 is a @xmath417-invariant open neighborhood of @xmath362 , and the map @xmath549 , @xmath550 is an isomorphism .
this map identifies @xmath551 with the cell @xmath552 and @xmath553 with the slice @xmath541 .
since @xmath554 is @xmath555-equivariant and @xmath556 , it follows that @xmath557 .
since @xmath558 is an open subvariety of a smooth variety , @xmath544 must be smooth .
if the fiber @xmath559 is smooth , then we can replace the pair @xmath560 by @xmath561 , where @xmath545 is the total space of the normal bundle @xmath562 to @xmath310 in @xmath544 , and @xmath563 is the complement of the the zero section in @xmath545 . in this case , by the thom isomorphism theorem we have @xmath564 , @xmath565 , and the map @xmath536 can be identified with multiplication by the euler class @xmath566 .
the spaces involved are all @xmath417-invariant , so we can also consider the corresponding map in equivariant cohomology : @xmath567 the same arguments as the non - equivariant case , together with standard arguments about equivariant formality and localization , prove the following .
we let @xmath572 , so @xmath0 is the symmetric group on the set @xmath573 . taking the torus @xmath417 to be the diagonal matrices , the lattice @xmath574 of characters is naturally identified with @xmath575
let @xmath576 be the @xmath577th standard basis vector of @xmath575 .
the roots of @xmath10 are then the vectors @xmath578 , for @xmath579 , @xmath580 .
choose the borel subgroup @xmath15 to be the lower triangular matrices .
with this choice , the positive roots are @xmath581 with @xmath582 , and the simple roots are @xmath583 , @xmath584 .
the simple reflection @xmath585 corresponding to @xmath586 is the transposition of @xmath577 and @xmath587 .
we now fix @xmath588 where @xmath589 is the shortest of the `` hexagon permutations '' introduced in [ subsec : a7 ] .
the one - line notation of @xmath42 is @xmath590 .
a reduced word @xmath363 for @xmath42 can be given by the sequence of simple roots @xmath591 reduced words for the other three hexagon permutations are obtained from this by appending @xmath592 at the beginning or the end , or both . to prove this , we look more closely at the fiber @xmath597
. we will compute the map @xmath598 by working with @xmath417-equivariant cohomology and localization .
first we describe the @xmath417-fixed points in @xmath310 .
@xmath310 has two irreducible components .
the first , call it @xmath609 , is isomorphic to @xmath610 , where @xmath611 is isomorphic to @xmath612 blown up at two points .
the @xmath417-fixed points in @xmath609 are the ones given by ( a ) above .
let @xmath617 be the ( non - reduced ) word corresponding to the sequence of simple roots @xmath618 it is a subword of @xmath363 ; specifically , the roots @xmath480 , @xmath592 and @xmath619 have been omitted .
there is an embedding @xmath620 which fills in the coordinates for the missing roots with the identity element of @xmath621 .
the simple reflections @xmath622 that appear in @xmath617 generate the weyl group of the group @xmath623 , embedded into @xmath624 as block diagonal matrices acting on the middle two factors in the decomposition @xmath625 .
it follows that we have an isomorphism @xmath626 , where the factors are the bott - samelson varieties for @xmath627 and @xmath628 , respectively .
both @xmath629 and @xmath630 are isomorphic to the bott - samelson variety in example [ sl3 example ] , and it is easy to see that @xmath631 is a product of two copies of the fiber from that example . to see that @xmath632 is an irreducible component of @xmath310 , note that a computation with theorem [ bb cells ] shows that the paving by affines of @xmath310 given by intersecting with the biaynicki - birula cells has only one cell of dimension four , namely @xmath633 , and all other cells are of smaller dimension .
so the closure of this cell must be a component of @xmath310 , and it is the only four - dimensional component , so it is equal to @xmath609 . to understand the other component , note that there is only one cell in @xmath634 of dimension three , namely @xmath635 , and all other cells are of smaller dimension
. if we let @xmath636 be the smallest subword of @xmath363 containing the nonzero entries of @xmath637 , and @xmath638 , then the @xmath417-fixed points of @xmath639 are the ones given in ( b ) in the statement of the theorem .
it is easy to check that @xmath640 , @xmath641 , and @xmath642 each map these fixed points onto a pair of fixed points in @xmath104 , and so they each map @xmath613 onto a @xmath417-curve .
the map @xmath643 gives the required isomorphism of @xmath613 with @xmath614 .
the fiber @xmath310 is not smooth , but there is only one component of dimension four , so we can reduce the computation of @xmath598 to the smooth case : it is isomorphic to the restriction map @xmath644 , where @xmath545 is the total space of the normal bundle @xmath562 to @xmath609 in @xmath645 .
as remarked earlier , @xmath646 can be identified with multiplication by the euler class @xmath647 , so the image of @xmath598 is spanned by @xmath566 .
we will compute this class by computing the equivariant euler class @xmath648 and then finding its image in ordinary cohomology . to do this , we split the normal bundle @xmath562 into line bundles .
we have seen in the proof of proposition [ y components ] that @xmath649 ; we claim that this intersection is transverse .
since @xmath650 and @xmath210 is @xmath15-equivariant , @xmath210 is a submersion over the cell @xmath552 , and the required transversality follows from the transversality of @xmath541 and @xmath552 in @xmath104 .
let @xmath651 be the subwords of @xmath363 of length @xmath652 which contain all of the simple reflections in @xmath617 , together with one of the remaining simple reflections @xmath653 .
if @xmath654 is the restriction to @xmath609 of the normal bundle to @xmath655 in @xmath656 , then we have @xmath657 .
to compute the equivariant euler classes @xmath658 , we compute their restrictions to the fixed point set @xmath659 .
the restriction @xmath660 is just the @xmath417-weight of the tangent space to the unique @xmath417-curve containing @xmath661 , contained in @xmath656 and not contained in @xmath655 .
this curve is the curve joining @xmath661 and @xmath662 , where @xmath663 , @xmath664 , @xmath665 , @xmath666 . using propositions [ flag 1d orbits ] and [ bs fixed points ]
, we can compute that its @xmath417-weight is the sum of the entries under @xmath667 and @xmath668 in the following tables : the equivariant class @xmath671 induces the same class in @xmath672 as @xmath673 where we abuse notation and write a weight @xmath674 instead of its pullback under the map @xmath675 . after a little computation one sees that this class restricts to zero at every point of @xmath676 except @xmath677 and @xmath678 , where it has the same restriction as @xmath679 , the equivariant euler class of the tangent bundle to @xmath609 .
( to compute the localization of @xmath679 to the fixed points , use the identification of the weights of @xmath417-curves in example [ sl3 example ] .
note that the labeling of the fixed points @xmath680 in that example corresponds to the labeling of the fixed points @xmath681 and @xmath682 of @xmath629 and @xmath630 . )
then the atiyah - bott localization formula shows that @xmath566 is twice a generator of @xmath672 , as desired .
[ rem : otherhex ] the other hexagon permutations can be shown to have @xmath73-torsion by a similar computation ; we give the main points .
let @xmath683 and @xmath684 for @xmath685 . the fiber @xmath686 is still four - dimensional , but now the union of the components of maximal dimension is isomorphic to @xmath687 , where @xmath688 and @xmath689 , the union taken so that @xmath690 is identified with a @xmath417-curve in @xmath611 with trivial normal bundle .
the excess intersection formula then implies that matrix of @xmath598 is diagonal under the natural bases given by the components of @xmath691 ( in other words , the components are orthogonal under the intersection form ) . the normal bundle to the component
@xmath610 is the same as before , so we have @xmath692 .
let @xmath693 .
we follow the notation of section [ subsec : d4 ] : the simple reflections in @xmath0 are @xmath694 where @xmath695 all commute with each other .
let @xmath363 be the word @xmath696 , put @xmath697 , and let @xmath698 .
the @xmath417-fixed points in @xmath559 are @xmath699 the fiber @xmath310 is the transverse intersection of @xmath700 and @xmath544 .
it is @xmath417-equivariantly isomorphic to @xmath614 , where the @xmath417-weights on the three factors are @xmath701 , @xmath702 , and @xmath703 , respectively . since @xmath310 is smooth , we have @xmath704 , and the map @xmath705 can be identified with multiplication by @xmath706 on @xmath707 , where @xmath708 is the normal bundle to @xmath310 in @xmath544 .
as in the previous example we compute this by computing the localization of the equivariant class @xmath709 to the fixed points @xmath616 .
we have @xmath710 so @xmath711 , where @xmath712 are the pullbacks of a generating class of @xmath713 by the three projection maps . multiplication by this class from @xmath714 to @xmath715 is given by the matrix @xmath716\ ] ] with respect to the natural monomial basis in @xmath717 , @xmath718 , @xmath719 .
this matrix has determinant @xmath720 , so @xmath721 has @xmath73-torsion .
a. belinson , j. bernstein , and p. deligne .
faisceaux pervers . in _
analyse et topologie sur les espaces singuliers , i ( luminy , 1981 ) _ , volume 100 of _ astrisque _ , pages 5171 .
france , paris , 1982 . f. du cloux
. coxeter . a computer program available from http://math.univ - lyon1.fr/~ducloux / coxeter / coxeter3/english / coxeter3_% e.html[math.univ-lyon1.fr/ ducloux / coxeter / coxeter3/english / coxeter3_e.html ] .
d. kazhdan and g. lusztig .
schubert varieties and poincar duality . in _ geometry of the laplace operator ( proc .
pure math .
hawaii , honolulu , hawaii , 1979 ) _ , proc .
pure math . , xxxvi , pages 185203 .
soc . , providence , r.i . , 1980 .
a. lascoux and m .-
polynmes de kazhdan & lusztig pour les grassmanniennes . in _ young tableaux and schur functors in algebra and geometry ( toru , 1980 ) _ , volume 87 of _ astrisque _ , pages 249266 .
france , paris , 1981 .
g. lusztig . some problems in the representation theory of finite chevalley groups . in _ the santa cruz conference on finite groups ( univ .
california , santa cruz , calif . , 1979 ) _ , volume 37 of _ proc .
pure math .
_ , pages 313317 .
soc . , providence , r.i . | we present a combinatorial procedure , based on the @xmath0-graph of the coxeter group , which shows that the graded dimension of the stalks of intersection cohomology complexes of certain schubert varieties is independent of the characteristic of the coefficient field .
our procedure exploits the existence and uniqueness of parity sheaves . in particular
we are able to show that the characters of all intersection cohomology complexes with coefficients in a field on the flag variety @xmath1 of type @xmath2 for @xmath3 are given by kazhdan - lusztig basis elements . by results of soergel
, this implies a part of lusztig s conjecture for @xmath4 with @xmath5 .
we also give examples where our techniques fail . in the appendix by tom
braden examples are given of intersection cohomology complexes on the flag varities of @xmath6 and @xmath7 whose stalks have different graded dimension in characteristic 2 . |
recent years have seen a growing interest in the application of the geometrical ideas originally developed for studying einstein s theory of general relativity to other areas of physics , such as condensed matter physics .
the motivation is both the theoretical aim of developing the mathematical tools capable of dealing with as wide a range of physical problems as possible , and the desire to construct laboratory analogues of the exotic conditions which general relativity allows , but which are likely ever to remain inaccessible to direct experimental investigation .
this in turn may provide a stimulus for further laboratory investigations , possibly resulting in the discovery of new physical effects . in the present article we shall pursue this direction by demonstrating how the mathematical formalism of lie groups , which is of widespread use in general relativity and high energy physics @xcite , can be harnessed to study optical properties of symmetrical phases of matter .
we develop a formalism allowing the symmetries of an electromagnetic medium to be directly exploited in solving maxwell s equations . when considering symmetries in classical mechanics , the discussion is simplified by passing to the lagrangian or hamiltonian picture .
we present a formalism which similarly makes symmetries manifest for maxwell s equations . in order to motivate and illuminate the development of this formalism , we consider the example of light propagation in chiral nematic liquid crystals .
we believe this to be the first application of lie group techniques to such a problem .
the tools we develop , however , are more widely applicable to media with a continuous symmetry group . they may also be considered a generalisation of ` transformation optics ' , extending those ideas to allow for the possibility that the dielectric tensor and magnetic susceptibility differ . we begin in section [ sec2 ] with a brief discussion of the helical ground state for a chiral nematic liquid crystal . in section [ sec3 ]
we introduce the geometry of the optical metric of joets and ribotta @xcite , describing the propagation of the extraordinary ray in a uniaxial birefringent material .
in particular we study this metric and its geodesics for the helical ground state of a chiral nematic .
the resulting metric is invariant under the simply transitive action of three - dimensional group of isometries which is locally isometric to the euclidean group @xmath2 of the plane , which we discuss in detail in section [ sec31 ] .
in fact the isometry is the universal cover @xmath1 and the lie algebra is of type @xmath0 in bianchi s classification .
we extend the discussion to @xmath3 in an appendix .
the identification of the symmetry group for the chiral phase permits a fully geometrical discussion of electromagnetic phenomena , an approach we exploit .
the high degree of symmetry allows us to solve the hamilton - jacobi and wave equations up to one quadrature in the former and up to solutions of mathieu s equation in the latter case .
this opens up the possibility of a detailed analytic investigation of the type of caustics and optical singularities which should be observable in such systems @xcite .
we then go beyond the geometric optics approximation to consider the full maxwell equations in section [ sec4 ] . for a uniaxial material whose director field takes on a helical configuration
we show how the theory of lie groups leads to separation of variables for these equations .
the resulting equations take the form of a generalised mathieu equation .
this result is similar to others in the literature @xcite , but the derivation is fully motivated from the inherent symmetries of the problem . throughout the paper ,
we shall make use of the machinery of differential geometry , in particular tangent vectors , differential forms and the lie derivative .
references @xcite provide readable accounts of these concepts .
a nematic liquid crystal has an order parameter given by a _ director _ or direction field specified by a unit vector @xmath4 , defined up to a sign @xmath5 , with @xmath6 .
a _ chiral _
nematic has a built in twist , specified by a parameter @xmath7 , which can be realised as a _
torsion _ , which alters the usual derivative operator acting on a vector by ^q_i n_j= _
i n_j + q _ ijkn_k .
the frank - oseen free energy functional in the one - constant approximation is equivalent , up to a boundary term , to : f [ ] = ( | ^q |^2 - ( -1 ) ) d^3 x , where we have added a lagrange multiplier field @xmath8 to enforce the constraint that @xmath9 .
the free energy would be minimised if ^q_i n_j= 0 .
[ bog ] however , as can be seen by taking another @xmath10 derivative and skew symmetrising , this is not possible over an extended region so the system is _ frustrated _ and must adopt some compromise configuration @xcite .
one such configuration is the helical phase for which = ( pz ) + ( pz ) . for more details about the liquid crystals
the reader may consult @xcite .
for the helical phase , = 0,=-p , ^2 = -p^2 , this configuration is a stationary point of the free energy , satisfying the second order euler - lagrange equation resulting from extremising @xmath11 $ ] : -^2 + 2 q = ( -2q^2 ) .
provided we choose = p^2 - 2 pq+2q^2 . among these solutions of the euler - lagrange equations ,
the one minimising the free energy density has @xmath12 .
the only non - vanishing components of @xmath13 are ^q _ 2n_3=qn_1^q_1 n_3=-q n_2 .
it follows that the helical ground state is not a solution of the first order frustrated `` bogomolnyi equation '' ^q_i n_j= 0 .
another means of relieving the frustration is the ` double twist ' structure , given in cylindrical polar coordinates @xmath14 by @xmath15 along the @xmath16 axis , this configuration has @xmath17 , so inside a sufficiently small cylinder , the free energy density is in fact _ lower _ than that for the helical phase .
a structure composed of these tubes can fill space , but there will necessarily be defects where the tubes meet @xcite .
such a configuration gives the so called ` blue phase ' . whether the blue or the helical phase is thermodynamically preferred depends on the energetic cost associated to accommodating the defects of the blue phase .
if @xmath18 is the director , and @xmath19 , with @xmath20 is the unit tangent vector , then then the inverse speed or slowness of an extraordinary ray is is given by @xcite n= , where @xmath21 is the refractive index of the _ ordinary ray _ and @xmath22 that of the _ extra - ordinary ray_. fermat s principle reads n ds = 0 .
thus the rays are geodesics of the _ joets - ribotta metric _
o^2 = n_e^2 d ^2 + ( n_o^2 -n_e^2 ) ( .d ) ^2 .
assuming the refractive indices are constants , we can write down the metric for the helical ground state given above @xmath23 as another example , consider a particular case where the director field , @xmath24 , is in a ` hedgehog ' configuration ( cf . @xcite ) and where in addition the refractive indices @xmath25 , vary with position inside the ball @xmath26 according to = r , n_e= 1 , n_o= 1 .
[ hyp ] the resulting joets - ribotta metric is ds_o^2 = d ^2 1-r^2 + ( .d ) ^2 ( 1- r^2 ) ^2 .
which one recognises as that of hyperbolic space @xmath27 in beltrami coordinates .
remarkably , because hyperbolic space is projectively flat in these coordinates , the light rays are straight lines in this case .
one could consider the 3-sphere @xmath28 , by changing the sign in front of @xmath29 .
then one has a model related to maxwell s fish - eye lens by a coordinate transformation , but whose rays are straight lines . other examples may be found in @xcite .
optical properties such as ( [ hyp ] ) my appear unnatural , but modern meta - materials are increasingly able to mimic such refractive indices , at least within a certain range of frequencies for the electromagnetic field .
the aim of the present section is to obtain the isometry group of the apparently quite complicated metric ( [ metric ] ) .
an enormous simplification results if we use the formalism of metrics on lie groups
. we will start with a brief discussion , tailored to the euclidean group @xmath2 of isometries of the plane , consisting of translations and rotations in two dimensions .
those familiar with the construction of left- and right - invariant forms on lie groups may wish to skip to the summary at the end of this subsection .
we can realise elements of @xmath2 as matrices as follows .
we first fix a number @xmath31 . to any point @xmath32 in the plane , we associate the column vector @xmath33 we first note that a general isometry of the plane can be decomposed into a clockwise rotation of angle @xmath34 about the origin , followed by a translation of @xmath35
. we can write down a matrix @xmath36 depending on parameters @xmath37 which performs this operation as follows @xmath38 corresponding to any three numbers @xmath37 , we have a unique isometry and conversely each isometry corresponds to a unique @xmath37 , provided that we regard @xmath16 and @xmath39 as the same obtained by dropping the identification of the @xmath16 coordinate .
this is a slightly technical point which we shall not labour . ] .
we refer to @xmath37 as coordinates on the group @xmath2 .
our aim is to construct a set of vector fields which are invariant under an action of the group @xmath2 .
we ll consider for the moment a matrix lie group @xmath40 , i.e. a group whose elements are @xmath41 matrices for some appropriate @xmath42 where the group action is by matrix multiplication .
we note that any element @xmath43 of @xmath40 gives rise to two natural transformations on the group itself . acting on @xmath44 , a general element of @xmath40 , they give : @xmath45 and
are known respectively as left and right translation , or the left and right action of @xmath43 .
we note that the left action and the right action commute : @xmath46 now suppose that @xmath43 is infinitesimally close to the identity matrix , @xmath47 for an infinitesimal parameter @xmath48 .
we have @xmath49 where the infinitesimal generator @xmath50 can be interpreted as a tangent vector to @xmath40 at the point @xmath44 , where we think of @xmath40 as a submanifold , i.e. a surface - dimensions , merely that it is of lower dimension than the space in which it lives . ] in the space of all @xmath51 matrices .
this is a _
right - invariant _ vector field , since @xmath52 because of equation ( [ comact ] ) above . following a similar procedure , we find that the vector fields @xmath53 are _ left - invariant_. in order to construct the left- and right - invariant vector fields of @xmath2
, we must first find all suitable @xmath54 such that @xmath55 is an element of @xmath2 .
we can do this by using the coordinate representation @xmath36 above . since @xmath56 , we can find the most general @xmath54 by taylor expanding @xmath36 for small @xmath37 .
we find that the general @xmath54 is a linear combination of the matrices : @xmath57 we d like to express the vector fields in terms of the @xmath37 coordinates .
to do this , we define a basis of vector fields for @xmath2 as follows @xmath58 we can either think of @xmath59 as a concrete matrix tangent to @xmath2 as a surface in the space of @xmath60 matrices , or more abstractly as the vector field which generates a shift from @xmath37 to @xmath61 in the coordinate space .
this notation captures the fact that under a change of variables for the coordinate space , @xmath62 , the vector fields transform in the same way as differential operators following the chain rule . we can readily calculate the left invariant vector field corresponding to @xmath63 : @xmath64 in a similar way
, we can find the rest of the left- and right - invariant vector fields , @xmath65 : @xmath66 now that we have the left- and right - invariant vector fields , we can construct the left- and right - invariant one - forms which are dual to them .
taking @xmath67 to be the one - forms dual to @xmath68 , the left - invariant forms @xmath69 and right invariant forms @xmath70 are : @xmath71 the matrices @xmath72 have a natural commutator algebra , the lie algebra @xmath73 .
this determines in a natural way the lie algebra of the vector fields @xmath74 and @xmath75 : @xmath76 & = & 0 , & \qquad & \bigl [ r_1 , r_2 \bigr ] & = & 0,\\ \bigl [ l_3 , l_1 \bigr ] & = & + l_2 , & \qquad&\bigl [ r_3 , r_1 \bigr ] & = & -r_2 , \\
\bigl [ l_3 , l_2 \bigr ] & = & -l_1 , & \qquad&\bigl [ r_3 , r_2 \bigr ] & = & + r_1 , \end{array}\ ] ] and the maurer - cartan algebra of the one - forms @xmath77 the fact that left and right actions commute ( [ comact ] ) is reflected in the fact that @xmath78=0 $ ] .
we can check the claimed invariance explicitly .
first we note the matrix identity @xmath79 from which we deduce that the element @xmath80 acting by right translation takes @xmath37 to @xmath81 where x&= & x+p z -p z , + y&= & y+p z + p z , + z & = & z+ . making these substitutions into @xmath70 , treating @xmath82 as constants , we find dx + py dz&=&dx + py dz , + dy - px dz & = & dy - px dz , + p dz & = & p dz , so that the @xmath70 are indeed invariant under right translations . interchanging the roles of @xmath83 and @xmath84 in ( [ lraction ] ) we deduce that the element @xmath80 acting by left translation takes @xmath37 to @xmath81 where x&= & x p -y p + , + y&= & yp + x p + , + z & = & z+ . substituting into @xmath69 , again treating @xmath82 as constants
, we can check that ( pz ) dx + ( pz ) dy&=&(pz ) dx + ( pz ) dy , + ( pz ) dy - ( pz ) dx&=&(pz ) dy - ( pz ) dx , + p dz & = & p dz , so the @xmath69 are invariant under left translations .
armed with these invariant one - forms , we are now in a position to construct metrics which are invariant under an action of @xmath2 . for example , the flat metric can be written in terms of the left - invariant one - forms as : ds ^2 = p^-2 ( ^3 ) ^2 + ( ^1)^2 + ( ^2 ) ^2 , and is hence manifestly left - invariant .
now for the helical ground state of the nematic liquid crystal , d = ^1 , thus the _ joets - ribotta metric _ of the helical phase may be written as ds ^2_o & = & n_e^2 ( p^-2 ( ^3 ) ^2 + ( ^1)^2 + ( ^2 ) ^2 ) + ( n_o^2-n_e^2 ) ( ^1 ) ^2 , + & = & n_o^2 ( ^1)^2 + n_e^2 ( ^2)^2 + ( ^3)^2 [ jr ] which is a left invariant metric on @xmath1 .
in fact , any left invariant metric may be brought into this form by a global right action of @xmath2 .
cartan s formula for a p - form reads _
x = i_x d + d(i_x ) , and so the non - vanishing lie derivatives are _
^1 & = & - ^2 , + l _
^2 & = & ^1 , + l _
^2 & = & -^3 , + l _
^1 & = & ^3 .
thus while @xmath85 is an additional symmetry of the flat metric none of the @xmath74 are symmetries of the joets - ribotta metric . to summarise then , @xmath86 may be considered as coordinates on @xmath1 , the universal cover of the two - dimensional euclidean group @xmath2 with @xmath87 left - invariant one - forms and @xmath74 left - invariant vector fields .
if we were to identify the coordinate modulo @xmath88 , @xmath89 the group would be the @xmath42 -fold cover of the euclidean group @xmath2 , which corresponds to @xmath90 . with this identification ,
the joets - ribotta metric of the helical phase is a left - invariant metric .
its symmetry algebra , the lie algebra @xmath91 , is of bianchi type @xmath0 . as an aside , this may be obtained from the rotation group algebra @xmath92 by means of a wigner - inn contraction . if @xmath93 are the generators of @xmath92 , one sets @xmath94 and takes the limit @xmath95 . under this contraction ,
the direction field @xmath96 arises as the image of the hopf fibration generated by the right action of @xmath97 and the joets - ribotta metric as the image of the berger - sphere . for the relevance of the hopf fibration to chiral nematics
see @xcite . in the ray approximation
we are looking at geodesics with respect to a left - invariant metric on the universal cover of the euclidean group @xmath1 . now
the euclidean group @xmath2 is the configuration space for a rigid body in two - dimensional euclidean space @xmath98 .
the motion of a rigid body moving in a homogeneous , incompressible , inviscid fluid @xcite is known to correspond to geodesic motion with respect to a left - invariant metric on the euclidean group @xcite .
the present situation corresponds to a cylinder with its axis in a plane @xcite which may be reduced to the quadrantal pendulum . to see this in detail note that the eikonal equation is ( l_3 w)^2 + ( l_2 w ) ^2 + ( l_1 w ) = ^2 , and it separates . if @xmath99 , then ( ) ^2 + ( k_x(pz ) + k_y ( pz ) ) ^2 + ( k_x(pz ) - k_y ( pz ) ) ^2 = ^2 .
the killing vectors @xmath75 give rise to three constants of the motion of the form p_i = g _ r^_i .
of which two , @xmath100 and @xmath101 mutually commute . ) .
the curves are shown projected into the @xmath102 plane .
one is unbounded in @xmath16 , whereas the other is bounded .
not shown is the motion in the @xmath103 direction which gives both these curves a cork - screw motion .
[ geodfig],width=288 ] we may immediately find the equations for rays in a first order form by making use of the relation : @xmath104 we find : @xmath105 here @xmath106 , @xmath107 and @xmath108 .
let us first consider ( [ zeq ] ) . introducing new constants @xmath109 and defining @xmath110 , we find @xmath111 which is the so - called _ quadrantal pendulum _
equation @xcite .
the pendulum has two different types of behaviour , depending on the constants @xmath112 and @xmath113 .
if @xmath114 , then @xmath115 and hence @xmath116 will increase without bound .
this corresponds to a pendulum swinging through complete revolutions .
if @xmath117 , then @xmath118 will oscillate about @xmath119 for some integer @xmath42 .
this corresponds to the standard libratory motion of a pendulum .
thus we find two behaviours for the rays .
either the rays can penetrate in the @xmath16-direction or else they are trapped to move between two planes perpendicular to the @xmath16-axis .
finally , we can consider the other equations of motion , ( [ xeq ] ) and ( [ yeq ] ) .
we can interpret these as saying that the tangent vector of the ray oscillates around an average direction .
for rays which are not bounded in @xmath16 , the result is a ` cork - screw ' curve , similar to a helix .
the ` tightness ' of the spiral is determined by how close @xmath120 is to @xmath121 .
figure [ geodfig ] shows some examples . for an approximate description of light propagation beyond the ray approximation one may use the scalar wave equation .
the scalar wave equation captures some of the wave aspects of light , whilst ignoring the complications relating to polarisation which arise in the full maxwell equations .
we may expect the wave equation on the joets - ribotta metric to share some features with one of the two polarisations of the full maxwell equations .
it takes the form : 0 = - + l_3 l_3 + l_2 l_2 + l_1 l_1 .
[ waveeqn ] it separates .
that is if @xmath122 , then + ( ^2 - ( k_x(pz ) + k_y ( pz ) ) ^2 - ( k_x(pz ) - k_y ( pz ) ) ^2 ) f = 0 . recalling @xmath123 from the previous section , this is of the form + ( + ( 2 ) ) f=0 , [ mathieu ] which is mathieu s equation . by the floquet - bloch theorem ,
the general solution of ( [ mathieu ] ) is of the form f=
c_1 e^i f ( ) + c_2 e^-i f(- ) where @xmath124 and @xmath125 depends on @xmath112 and @xmath113 . expanding @xmath126 as a fourier series , we deduce the laue - bragg conditions that an incoming wave with wave vector @xmath127 incident on some region where propagation is described by ( [ waveeqn ] ) is reflected / diffracted with wave vector @xmath128 where p ( _ out- _ in ) _ z = m .
we may think of @xmath129 as defining a dispersion relation , averaged over the period in the vertical direction .
when @xmath125 is real we expect propagating waves , whereas when @xmath125 has an imaginary component the solutions either decay or grow exponentially in @xmath118 .
it can be shown that the marginal cases between propagation and damping occur when @xmath130 ( note @xmath125 is only defined up to multiples of @xmath131 ) .
this defines a set of surfaces in the @xmath132 space which separate out the regions where the wave propagates and where it is damped . to determine these surfaces ,
we can ( for @xmath133 , the other case follows similarly ) expand @xmath134 in fourier series : f = _
-^ c_n e^in and obtain a three term recurrence relation -n^2 c_n + c_n + 2 ( c_n-2 + c_n+2 ) = 0 . the condition that this relation admits a non - trivial solution may be related to the vanishing of an infinite determinant , a procedure known as hill s method , see e.g. @xcite .
before we discuss maxwell s equations for the helical phase of a nematic liquid crystal , we will first formulate maxwell s equations for a general medium in the language of differential forms
. this will be the most convenient language in which to discuss how to apply the machinery of lie groups to the problem in hand .
we will work directly with the fields rather than introducing potentials as this avoids tackling the issue of gauge invariance .
we ll work on a @xmath135-dimensional manifold @xmath44 , but this restriction is not necessary .
we begin by noting that we can define two @xmath136-forms : @xmath137 these forms encode the fields contained in the antisymmetric @xmath135-tensors with components @xmath138 where @xmath139 , @xmath140 are the electric field and magnetic displacement , @xmath141 is the electric displacement field and @xmath142 is the magnetic field .
the advantage of packaging the fields as two - forms is that maxwell s equations become the simple pair of relations @xmath143 with @xmath144 the current @xmath145-form .
of course , to close this system of equations for a prescribed @xmath144 , we must specify a relation between @xmath134 and @xmath146 , the _ constitutive relation_. this is nothing more than the usual relations one requires relating @xmath147 and @xmath148 . in the language of forms ,
we require a map from the space of sections of @xmath149 to itself will be a @xmath150 form , but similar considerations apply ] .
in many materials , the constitutive relation is local and linear , so may be represented by a section of the bundle @xmath151 , i.e. a possibly space dependent linear map @xmath152 such that @xmath153 where @xmath152 acts pointwise .
this tensor @xmath152 , together with the differentiable structure of the manifold , is the minimal data required to define maxwell s equations - it has not thus far been necessary to introduce a metric or other structure to @xmath44 . in index notation
the maxwell equations take the form : _ [ f _ ] = 0 , _ [ ( c_]^ f _ ) = 0 .
we note that defining @xmath152 as an endomorphism , i.e. with two indices up and two down , ensures that it is not necessary to define a connection in order to take derivatives covariantly . in order that ( [ max2 ] ) defines a suitable hyperbolic system of partial differential equations , restrictions are required on @xmath152 .
we shall assume that @xmath152 satisfies some such suitable conditions , without specifying what those might be . as a simple example , we may take @xmath152 to be the hodge map induced by a lorentzian metric @xmath154 , i.e. we take @xmath155 . if @xmath154 is the flat metric , this gives the classical maxwell equations in the vacuum .
if @xmath154 is not flat , we may interpret the field as an electromagnetic field propagating in a gravitational background . in the case
that @xmath154 is static , we can alternatively interpret the field as propagating through some material with a position dependent dielectric tensor , @xmath156 and magnetic susceptibility @xmath157 .
this is the basis of _ transformation optics _ @xcite . for a material in which maxwell s equations have a gravitational interpretation it must be the case that @xmath158 in suitable units , i.e. the material is _
impedance matched _
this need not be the case for a general material .
we will take @xmath152 to have the following form : @xmath159 note that if @xmath160 for some lorentzian metric , we have @xmath161 so that @xmath162 , justifying our assertion that materials with a gravitational analogue are impedance matched . the liquid crystals in which we are interested are _ uniaxial _ , so that at each point , we can assume that the quadrics defined by @xmath48 and @xmath125 are _ spheroidal _ ( i.e. ellipsoids with an axis of symmetry ) with a common axis . in other words , there is locally a basis in which the tensors have the form @xmath163 if we assume that the axis of the material lies along @xmath24 , this can be written in a more covariant form as @xmath164 before we discuss the consequences of such a constitutive relation in the case of a nematic liquid crystal in the helical ground state , let us first consider for a moment the geometric optics approximation .
let us consider the maxwell equations described above in a geometric optics limit .
we consider a field which takes the form @xmath165 where by assumption @xmath166 are @xmath167 as @xmath168 .
we assume that there are no currents or charges , so that maxwell s equations become @xmath169 .
we also assume that @xmath152 varies slowly by comparison to the wavelength of the field . inserting our ansatz and collecting terms in @xmath112
, we find @xmath170 let us first consider the @xmath171 terms .
this is a system which asserts that @xmath172 is in the kernel of a linear operator which maps from one @xmath173-dimensional space to another @xmath173-dimensional space .
the condition that a non - trivial @xmath172 exists gives a differential condition on @xmath174 involving @xmath152 which we interpret as the eikonal equation .
associated to a solution of the eikonal equation is a @xmath136-form @xmath172 which gives the polarisation of the wave . in general , there will be only one polarisation associated to each solution of the eikonal equation .
once we have solved for @xmath174 and @xmath172 , we can inductively construct @xmath175 by solving the equations @xmath176 presumably the well posedness of this system is a necessary condition that @xmath152 be an acceptable constitutive map . in the case where @xmath160 , the eikonal equation can be shown to reduce to @xmath177 which is the hamilton - jacobi equation for geodesics of the metric .
in this case , there is a two dimensional space of possible polarisation tensors .
they take the form @xmath178 the hamilton - jacobi equation requires that @xmath179 be null .
suppose for example that at a point , @xmath179 is parallel to @xmath180 , then the space of polarisations at that point is spanned by @xmath181 and @xmath182 . in the case where @xmath152 has the uniaxial form introduced above , the eikonal equation reduces to the form : @xmath183 the medium is thus _
birefringent_. we see straight away that the condition on @xmath174 factors into two separate hamilton - jacobi equations associated to the two metrics @xmath184 these are both of the joets - ribotta form we have previously considered . it can be checked that the polarisation tensor associated to a solution of the hamilton - jacobi equation of @xmath185 has @xmath186 , whereas for a solution of the hamilton - jacobi equation of @xmath187 , the polarisation tensor has @xmath188 .
note that we do not require that @xmath24 remains constant for this derivation , provided it varies slowly compared to the wavelength of the light . in the case that @xmath24 varies from point to point ,
the polarisation will also change so that to leading order in @xmath112 , either the magnetic or electric field parallel to the director will vanish , depending on which type of ray we consider .
often , one takes @xmath189 in which case , @xmath187 is simply the minkowski metric and its geodesics are the _ ordinary rays_. the rays of the metric @xmath185 are the _ extraordinary rays _ and @xmath185 is the joets - ribotta metric , where we identify @xmath190 and @xmath191 . if @xmath152 is of the form ( [ standardc ] ) , but with no uniaxial assumption , then the rays will typically be geodesics of a finsler geometry .
so far , we have re - cast familiar results into the notation of differential forms . whilst this is a satisfying exercise ,
it is not clear that it introduces any benefits beyond putting the equations in a manifestly coordinate invariant form . for our purposes ,
the great advantage is that this form of the equations permits a concise discussion of the symmetries of the system and allows the machinery lie groups to be brought to bear .
we start by defining a killing vector @xmath192 to be a vector which satisfies @xmath193 recall that @xmath152 is simply a tensor , so the lie derivative is defined as a consequence of the differentiable structure of @xmath44 . making use of this and cartan s relation , we deduce that if a @xmath136-form @xmath134 obeys maxwell s equations : @xmath194 then so will @xmath195 and in particular , the diffeomorphism induced by @xmath192 will map solutions of the equations into solutions of the equations .
an important example occurs when @xmath160 and @xmath192 is a killing vector of @xmath154 .
suppose that we have a group which acts simply transitively on @xmath44 by left actions and which preserves the material configuration , as is the case for the @xmath196 symmetry of helical ground state of the nematic liquid crystal
. then it must be that @xmath152 may be written in terms of the left invariant one - forms and their duals as : @xmath197 where @xmath198 are some _ constant _ coefficients . here
, indices run over @xmath199 .
we can make use of this to write down maxwell s equations for a nematic liquid crystal in its helical state .
we take @xmath200 this choice of basis is very similar to the rotating basis chosen by peterson , who investigated the electromagnetic field propagating through a nematic liquid crystal in its ground state @xcite . in our case , this choice of basis arises naturally from the group structure of underlying symmetries .
we assume further that @xmath201 where @xmath202 , @xmath125 have the uniaxial form we previously assumed ( [ biax ] ) .
maxwell s equations for the electric and magnetic fields take the form @xmath203 the matrix @xmath204 has non - zero components p_11 = p_22 = 1 .
these equations can be separated with the ansatz @xmath205 the components @xmath206 are given by a linear combination of other components , so that the maxwell equations reduce to a system of differential equations of the form : f(z ) + ( + _ 1 e^2
i p z+ _ 2 e^-2
i p z ) f(z ) = 0 .
[ genmat ] here @xmath207 is a @xmath135-vector and @xmath208 are @xmath209 matrices , given by : @xmath210 and @xmath211 where we have introduced @xmath212 .
we see that the euclidean symmetry of the original problem is still manifest since a rotation in the @xmath213@xmath103 plane sends @xmath214 , which is cancelled by a suitable shift in the @xmath16 coordinate .
we may view ( [ genmat ] ) as a generalised mathieu equation .
mathieu s equation itself may be written in this form with @xmath215 matrices . by floquet
s theorem , the general solution of ( [ genmat ] ) will take the form : f(z ) = e^i _ 1 z h_1(z ) + e^i _ 2 z h_2(z ) + e^i _ 3 z h_3(z ) + e^i _ 4 z h_4(z ) where @xmath216 are @xmath135-vectors . making use of discrete symmetries of the equations
, one may show that if @xmath125 is a floquet exponent , then so is @xmath217 and @xmath218 , implying relations amongst the @xmath219 .
this equation may be studied using the infinite determinant techniques of hill , an approach similar to that of @xcite , but that takes us beyond the scope of the current paper .
we hope to address this issue in future work .
since we have retained the independence of the magnetic susceptibility and the permittivity , this analysis applies equally well to magnetic materials with helical phases @xcite .
we have shown that certain properties of the chiral phase of a nematics liquid crystal are intimately tied to the symmetries it possesses .
we have shown that the joets - ribotta metric , which describes the propagation of extraordinary rays , is a left - invariant metric on @xmath220 and we have shown how the underlying symmetry group can be practically used to understand properties of waves in such a medium . we have separated the hamilton - jacobi equation and the wave equation for this metric . the wave equation can be reduced to mathieu s equation and the hamilton - jacobi equation to the quadrantal pendulum equation .
we have also seen how maxwell s equations for a general uniaxial material whose director field lies in a helical configuration can be reduced to coupled ordinary differential equations generalising mathieu s equation via a novel application of the theory of lie groups .
this new formalism is applicable to the macroscopic maxwell equations whenever the medium has a continuous symmetry group .
the approach taken generalises transformation optics to permit non - impedance matched media .
as we have seen even in this simple example , the extraordinary light rays propagating through a liquid crystal explore a much richer geometry than the usual flat geometry of light rays in the vacuum .
this opens up the possibility of constructing analogues for the propagation of light in a gravitational field . in this case
the light rays in the liquid crystal may be mapped onto light rays propagating in a bianchi @xmath0 cosmology @xcite whose spatial sections have a fixed geometry , but one may imagine more ambitious possibilities .
it is interesting to ask whether the set up above generalises to the bianchi type @xmath3 group . for this section
we set @xmath221 , in order not to clutter up the formulae .
we now define left - invariant one - forms and dual vector fields by ^3&= & dz , l_3= , + ^1 & = & e^hz ( z dx + z dy ) , l_1= e^-hz ( z + z ) , + ^2 & = & e^hz ( z dy - z dx ) , l_2 = e^-hz ( z - z ) .
the right - invariant one - forms and vectors fields are ^3&= & dz , r_3 = + x -y - h ( x + y ) , + ^1&= & dx + ( 1+h)y dz , r_1= , + ^2&= & dy - ( 1-h)x dz , r_2= . the metric and @xmath21 will be taken to be constant , that is position independent . ]
n_e^2 ( _ 1 ^ 2 + _ 2 ^ 2 + _ 3 ^2 ) = n^2_e ( dz ^2 + e^2hz ( dx ^2 + dy ^2 ) ) is in fact that of hyperbolic three space . in the upper half space or poincar patch space model .
setting e^hz = , x= , y= it becomes ( dz^2 + dx^2 + dy^2 ) , and we see that optically we can think of a vertically stratified isotropic medium with cartesian coordinates @xmath222 and refractive index .
rays are now circles orthogonal to the plane @xmath223 .
the metric ds ^2_o= n_e^2 ( _ 1 ^ 2 + _ 2 ^ 2 + _ 3 ^2 ) + ( n_o^2-n_e^2 ) _ 1 ^ 2 may thought of as describing a vertical stratified anisotropic medium with extraordinary and ordinary refractive indices varying with height @xmath224 in the same way , i.e. as , .
such a variation might be due to temperature variation within the material , for example .
as before , the wave equation separates but @xmath225 now satisfies + 2h + ( ^2 n_e^2 - ( 1 + ) ( k_x^2 + k_y^2 ) + ( k_x^2 + k_y^2 ) ( 1- ) ( 2z- ) ) f = 0 . | we show that the optical structure of the helical phase of a chiral nematic is naturally associated with the bianchi @xmath0 group manifold , of which we give a full account .
the joets - ribotta metric governing propagation of the extraordinary rays is invariant under the simply transitive action of the universal cover @xmath1 of the three dimensional euclidean group of two dimensions .
thus extraordinary light rays are geodesics of a left - invariant metric on this bianchi type @xmath0 group .
we are able to solve by separation of variables both the wave equation and the hamilton - jacobi equation for this metric .
the former reduces to mathieu s equation and the later to the quadrantal pendulum equation .
we discuss maxwell s equations for uniaxial optical materials where the configuration is invariant under a group action and develop a formalism to take advantage of these symmetries .
the material is not assumed to be impedance matched , thus going beyond the usual scope of transformation optics .
we show that for a chiral nematic in its helical phase maxwell s equations reduce to a generalised mathieu equation .
our results may also be relevant to helical phases of some magnetic materials and to light propagation in certain cosmological models . |
the mass distribution of dark matter haloes provides a direct probe of the nature of the dark matter particle , as the inner structure of dark matter haloes is particularly sensitive to the dark matter properties .
for example , warm dark matter should produce lower density halo cores than cold dark matter ( cdm ) because of the phase density ceiling introduced by the non - zero thermal velocity of warm particles ( e.g. tremaine & gunn 1979 ) .
the variation of peak halo phase density with halo mass is also dependent on the `` coldness '' of the dark matter particle ( e.g. lake 1989 and references therein ) .
spectroscopic observations of stellar motions in galaxies , lensing properties and x - ray temperature maps of cluster cores each can provide a measurement of the central dark matter distribution in haloes , albeit with some uncertainty inherent in inferring the dark matter distribution from properties of baryons or baryon dominated regions .
cdm haloes in n - body simulations consistently have a steep central cusp where the density rises as @xmath17 ( navarro , frenk , & white 1996 , 1997 , nfw hereafter ; huss , jain , & steinmetz 1999 ; power et al .
2003 ) , @xmath18 ( moore et al . 1998 , 1999 ,
m99 hereafter ; taylor & navarro 2001 ; governato , ghigna , & moore 2001 ; fukushige & makino 2001 , 2003 ) , or somewhere in between ( e.g. klypin et al . 2001 ; fukushige , kawai , & makino 2003 ; hayashi et al . 2003 ; navarro et al .
these numerical findings appear in conflict with the most direct observational results .
rotation curves of low surface brightness ( lsb ) dwarfs consistently yield density profiles with nearly constant density cores ( e.g. flores & primack 1994 ; moore 1994 ; salucci & burkert 2000 ; de blok et al .
studies of more luminous galaxies imply similar problems ( e.g. salucci & burkert 2000 ; salucci 2003 ; gentile et al .
the disagreement with rotation curves may indicate an insurmountable problem with cdm models , may be due to uncertainties in measuring accurate stellar curves at just @xmath24 of the virial radii ( van den bosch et al .
2000 ; van den bosch & swaters 2001 ; see however e.g. simon et al .
2003 ) , or may perhaps reflect some systematic bias common to all high resolution n - body simulations . alternatively , the disagreement may be due to a problem with common assumptions made when reconstructing mass profiles from circular velocity data ( hayashi et al .
strong gravitational lensing in clusters can potentially provide a direct measurement of the halo mass profile , and indeed central mass profiles for several lensing clusters have been calculated ( tyson , kochanski , & dellantonio 1998 ; shapiro & iliev 2000 ; sand , treu , & ellis 2002 ; gavazzi et al . 2003 ; sand et al .
2004 ) , but have yielded conflicting results .
cluster density profiles inferred from chandra luminosity - temperature mapping have steep cusps that are inconsistent with the flat cores observed from lsb rotation curves ( lewis , buote & stocke 2003 ) , though this method is sensitive to models of the intracluster gas . in sum
, many observational studies suggest a flatter profile than predicted by cdm models , but observations have not yet converged upon a basic shape of the density profile ( e.g. , jimenez et al .
2003 ) .
nfw found that cdm haloes have a `` universal '' density profile that is independent of mass , cosmological parameters , and the initial density fluctuation spectrum with significant scatter from halo to halo , @xmath25 where @xmath26 and @xmath27 is a characteristic inner radius ( the concentration radius ) and inner density , respectively .
fukushige & makino ( 1997 ) , based on a single halo with @xmath28 particles , as opposed to the @xmath29 particles of the nfw study , found a profile with slope between @xmath17 and @xmath30 .
m99 , using results from a series of six @xmath28 particle haloes , also found a profile steeper than @xmath17 , and proposed the following profile : @xmath31}. \label{eqm99}\ ] ] the nfw and m99 profiles are both specific cases of a three parameter profile family proposed by hernquist ( 1990 ) , and further developed by zhao ( 1996 ) .
the highest resolution haloes to date are a series of eight clusters , several with @xmath32 particles ( fukushige , kuwai , & makino 2004 ) , which have central slopes steeper than nfw and shallower than m99 .
the level of `` universality '' of the density profile is a matter of debate ( e.g. tatitsiomi et al .
jing & suto ( 2000 , 2002 ) , found central density cusps of @xmath33 , @xmath34 , and @xmath18 for a simulated halo with cluster , group , and galaxy mass , respectively .
a similar range of inner slope values was found in a recent set of high resolution haloes ( hayashi et al . 2003 ;
navarro et al .
the central cusp is especially sensitive to flattening caused by poor resolution or other numerical effects ( e.g. moore et al . 1998 ) , so slightly different numerical techniques might produce significantly different density profiles , making it sometimes difficult to compare results of different authors
. the appearance of near universality in many previous studies could be due to the fact that most simulations have modelled objects in the range of galaxies to clusters ; in this mass - range the effective slope of the linear power spectrum ( @xmath22 , where @xmath35 ) is @xmath36 -2 , implying cusps of roughly nfw slope ( syer & white 1998 ; subramanian , cen , & ostriker 2000 ) .
ricotti ( 2003 ) , using a large set of haloes with @xmath37 to @xmath38 particles , found considerably flatter cusps for high - redshift low - mass haloes , with @xmath39 at @xmath40 for @xmath41 .
flat , low mass haloes match model predictions ( syer & white 1998 ; subramanian , cen , & ostriker ; 2000 ) , wherein the central slope varies as @xmath42 .
see however , moore et al .
( 2001 ) , who find a @xmath34 cusp in a @xmath41 halo at redshift four . in cdm models ,
any dependence on @xmath22 would be manifested as a dependence on halo mass . in the @xmath0cdm model ,
@xmath22 asymptotically approaches -3 for low mass haloes .
as @xmath22 nears -3 , @xmath43 , the characteristic mass of collapsing haloes as a function of scale factor ( see section 4.3 ) diverges , so haloes of all masses collapse nearly simultaneously .
if one models halo formation as the assembly of spherically symmetric shells of material whose density is largely determined by the scale factor of the universe , then the density profile should be shallower when @xmath22 nears -3 .
the characteristic radius , @xmath26 , indicates the size of the central density region and is usually defined in terms of the concentration parameter , @xmath44 .
nfw , as well as a number of other authors found that the concentration radius decreases with halo mass , even for power law cosmologies where @xmath22 is constant .
this agrees with other simulations ( e.g. bullock et al .
2001a ; eke , navarro & steinmetz 2001 ) , and also with predictions based on the halo model wherein nonlinear properties are predicted starting from press & schechter ( 1974 ) theory ( e.g. huffenberger & seljak 2003 ) .
the concentration dependence on mass becomes weaker when @xmath22 is closer to -3 ( eke , navarro , & steinmentz 2001 ) . [ cols="<,<,<,<,<,<,<,<,<,<",options="header " , ] in order to gain an understanding of the physical effects that set the inner slope of halo density profiles , we plot the evolution of the profile slope in terms of the @xmath45 virial radius in non - comoving coordinates for our four best haloes .
this allows us to ignore the effects that expansion of the universe or evolving @xmath46 have on the power law slope of the profile .
we see in fig .
[ slotilezphys ] that the density profile inner slope for each halo evolves very little in non - comoving coordinates , plotted as @xmath47 , a result also found by fukushige & makino ( 2001 ) and fukushige et al .
( 2004 ) . in our cluster cl1
, the slope slowly steepens with time at @xmath482 , but in our other less massive haloes , the inner density profile slope shows no significant change .
the large fluctuation in the profile slope at @xmath49 for the galaxy halo and at @xmath50 for the dwarf halo are most likely due to the presence of subhaloes that are disrupted earlier .
whatever physical mechanism is responsible for setting the density profile , it must have largely occurred at very high redshift .
the lack of evolution in the physical densities means that the profile of an individual halo is substantially shallower at high redshift in terms of @xmath51 .
such apparent steepening of the slope with time is merely a scaling issue due to the growth of the virial radius as mass is added to the outer regions of the halo .
here we note that the profile slope is almost never shallower than the nfw asymptotic value of @xmath17 . only in the very inner region of our largest cluster cl1 at @xmath49 , where the slope trends toward @xmath52 at the innermost resolved radius ( not plotted ) ,
is there a slight hint that nfw slope may be too steep at high redshifts , but this could be simply a result of a subhalo just beyond that radius creating a local density minima , or it could be due to artificial numerical resolution effects at high redshift where the halo resolution is lower .
the profile concentration and thus the inner slope at a given radius are predicted by nfw and others to be a function of mass and redshift , when considered in terms of @xmath51 , with characteristic radius @xmath53 increasing with increasing @xmath15 .
here , @xmath13 is the characteristic mass of collapsing haloes defined by the scale at which the rms linear density fluctuation equals the threshold for non - linear collapse ( i.e. @xmath54 ) .
we have measured the nfw concentration parameter , @xmath55 , by forcing our profiles to an nfw profile , and performing a least squares fit . in fig .
[ conccomb ] , we plot the concentration parameter for our set of haloes , and we show the eke , navarro , & steinmetz ( 2001 ) prediction for @xmath56 .
the concentration dependence on @xmath15 for our haloes is significantly steeper than predicted by nfw for @xmath57 , and within the scatter of our haloes for @xmath58 .
measured values of the inner slopes also show clear trends with mass and redshift , shown in fig .
[ slopemstar ] , though this is due at least in part , to the degeneracy between slope and concentration . here
, we have measured the average value of the power law slope @xmath59 between 2@xmath4 and 5@xmath60 .
the asymptotic inner slope parameter , @xmath6 , from a two - parameter fit of @xmath6 and concentration as given by equation 3 , is plotted in fig .
[ gammamstar ] , and has a weak dependence on @xmath15 for our haloes , with a large scatter .
the median value of @xmath6 in our sample trends toward shallower inner slopes with increasing @xmath15 , given by : @xmath61 with a scatter of @xmath62 for our haloes , and valid for haloes of @xmath10 to @xmath11 .
[ conc2parammstar ] shows that the concentration parameter in this two - parameter fit , c@xmath63 , shows a significant trend toward higher values as @xmath15 decreases .
the trend is weaker and has significantly more scatter than the forced nfw fit @xmath55 dependence on @xmath15 ( see fig .
[ conccomb ] ) . adopting a power - law parameterisation as in e.g. huffenberger & seljak ( 2003 )
, the median @xmath7 for our haloes is : @xmath64 with a @xmath15 dependent scatter roughly equal to @xmath16 . in our cubehi simulation
, we examine the stability in the profiles of our set of ten clusters over timescales separated by @xmath65 for @xmath66 ; fig .
[ cubeslo ] . in the outer regions ,
orbiting substructure creates substantial scatter , however , the inner density profile is relatively stable .
thus , differences between the central density profiles of haloes of similar mass reflect different inherent properties of the halo , rather than temporal effects of orbiting substructure . in the hopes of understanding what physical processes are responsible for setting the central density profile of each halo
, we examine the evolution of the main halo and its progenitors .
we construct merger histories for each cluster and examine a number of properties , including : evolution of the central cusp mass concentration ; cluster accretion history ; total collapsed progenitor mass ; and angular momentum profiles . to follow the mass accretion history of the cluster and its progenitors
, we use the friends - of - friends algorithm ( fof , davis et al . 1985 ) . in order to follow the evolution of the mass concentration that makes up the cluster cusp , we use skid ( stadel 2001 ) , which is able
to identify bound mass concentrations independently of environment . in fig .
[ cubeskidcen ] , we plot the mass evolution of the central skid progenitor .
the two haloes with the earliest forming central skid halo are also those that have density profiles most closely matching the m99 profile ( cube4 and cube6 ; see fig .
[ slonfwmoorez0 ] , table 2 ) , both with slopes that steepen slowly toward r@xmath67 , beginning at radii of roughly 10@xmath60 , implying large concentration radii .
if not simply a coincidence , then this implies that the central cusp material is assembled earlier in haloes with steeper central slope parameters .
we then consider the effects of the mass accretion history of the cluster on the final density profile , and find that accretion history correlates with halo concentration , not with cusp slope .
[ cubefofcen ] shows that the three haloes with the highest concentration undergo a phase of rapid growth at @xmath68 , making their normalised mass temporarily @xmath69 times larger than the other seven clusters , which experience nearly uniform accretion rates .
we have also examined the evolution of the total collapsed progenitor mass , and find a similar correlation with cluster concentration . at redshifts of @xmath70 ,
the same three highly concentrated haloes have @xmath69 times more total mass in collapsed progenitors ( not plotted ) , where we have only considered progenitors of mass greater than @xmath71 percent of the final cluster mass .
the correlation of the concentration parameter with halo and progenitor collapse time is not surprising ; in fact , as have shown in section 4.3 , lower mass haloes , which are assembled earlier in a hierarchical model , have smaller concentration radii .
the concentration trends also qualitatively agree with correlations of formation epoch and concentration found in numerical simulations by wechsler et al .
however , a cautionary note is needed here .
even though in the cubehi simulation , the halo masses differ by only a factor of @xmath69 and thus cover a narrow range in median concentration parameter ( see eq .
5 ) , the two least massive and hence most poorly resolved clusters , are also the two most concentrated . a larger set of higher resolution haloes is needed to conclusively rule out the possibility that haloes resolved with fewer particles ( and identical softening ) lead to higher concentrations . in fig .
[ cubemom ] , we plot the angular momentum parameter @xmath72 , from bullock et al .
( 2001b ) , where @xmath73 is the specific angular momentum .
we see no correlation of profile concentration or central slope with @xmath74 in either the inner or the outer regions .
we also have examined the angular momentum profiles at high redshift , but find no clear correlation with halo concentration or central slope .
this is puzzling given that merger histories , which correlate with the density profiles , should also correlate with the angular momentum distribution . in order to understand the effects of the power spectral slope index , @xmath22
, we have simulated the renormalized volume cluster cl1 in scale free cosmologies with a range of values for @xmath22 . here
the initial density fluctuation power spectrum is given by @xmath75 , and is normalized to @xmath76 with @xmath77 and @xmath78 .
[ plawpro ] shows the density profile for a cluster with power law initial conditions given by @xmath210 , -1 , -2 , and -2.7 , followed by the corresponding density profile slopes in fig .
[ plawslo ] .
here we also include plot the power et al .
density criteria as we have not done convergence tests specifically for @xmath20 cosmologies .
note that the @xmath79 run has significantly less mass because of the @xmath76 normalization .
there is a clear trend that steeper power spectra yield density profiles with flatter slopes and larger concentration radii .
visually , the @xmath80 cluster displays much more prominent substructure , due to its proportionally stronger small scale power , with many nearly spherical haloes and few filaments .
a two parameter profile fit yields @xmath81 for @xmath80 and @xmath82 for @xmath79 ( see table 2 ) . a number of authors ( hoffman & shaham 1985 ; crone , evrard , & richstone 1994 ; cole & lacey 1996 ; syer & white 1998 ; subramanian et al .
2000 ; huffenberger & seljak 2003 ; nfw 1997 ; eke , navarro & steinmetz 2001 ; ricotti 2002 ) have suggested a power law dependence of the density profile that qualitatively agrees with our power law simulations .
however , syer and white ( 1998 ) predict that an @xmath79 power law cosmology should have an inner slope of @xmath83 , which is much shallower than seen in our haloes , though not necessarily inconsistent if the slope flattens only at very small radii .
the same model predicts an inner slope of @xmath84 for the @xmath80 cosmology , which is very close to the inner slope at the minimum resolved radius of our @xmath80 halo .
\(1 ) _ density profile trends and the cdm `` cusp problem '' . _
we follow the evolution of sixteen haloes over a large range in parameter space , with masses corresponding to dwarfs through clusters back to @xmath49 for our highest resolution cluster and @xmath50 for two dwarfs . by using identical haloes at varying resolution
, we have shown that our haloes are likely to be free from biases related to numerical resolution .
none of our simulated haloes has a density profile slope significantly shallower than @xmath17 down to the minimum resolved radii of 0.5@xmath4 to 2@xmath4 .
these steep cusps are similar to those found in previous cdm simulations , and appear to be in conflict with reported rotation curves from dwarfs and lsbs .
our resolution experiments confirm that steep cusps are formed regardless of numerical resolution , down to the minimum resolved radius of each halo .
previous high resolution simulations discussed earlier have mostly modelled haloes with masses of @xmath85 or higher , leaving the possibility that a mass dependence on the density profile might be able to solve the cusp problem . however , our results show that the observed conflict with @xmath0cdm density profiles gets worse when considering simulated dwarfs .
the enclosed mass at a given radius near the halo core is higher for simulated haloes of lower mass , since the concentration radius is smaller and the measured slope is steeper .
\(2 ) _ convergence of profile slopes . _
we confirm that the physical slope of haloes remains stable over time .
the inner regions of haloes appear to be composed of mass assembled at very high redshift , implying that present - epoch density - profiles are determined by the high redshift merger history .
we do see some apparent evolution in the physical slope at very high redshifts for our highest resolution cluster and group .
this suggests that for group and clusters - size haloes , events occurring at @xmath86 are partially responsible for determining the final shape of the density profile , while galaxies and smaller mass haloes have their inner density profile shape almost entirely determined at higher redshifts .
the progenitor region of the host halo , because of its large scale density enhancement , is the likely site of early - forming small haloes , many of which will merge to form the cusp .
after the universe has expanded by a few factors since the cusp material is assembled , the characteristic densities of haloes merging into the main halo should be lower than typical densities within the main cusp , because of their generally later assembly epoch .
consequently , merging subhaloes will likely be disrupted before dynamical friction can bring them near the centre of the main halo , where they can affect the central density profile . in this scenario
, halo density profiles could converge to roughly flat profiles ( in terms of @xmath51 ) at extremely high redshifts .
cdm haloes at low redshift might then be expected to have a very small flat inner core , but at radii much less than the inner @xmath24 that can be currently probed by simulations or observations .
\(3 ) _ lessons from the power law cosmologies . _
the dependence of the concentration parameter on @xmath15 is most likely the result of a mechanism analogous to what sets the dependence of concentration on the spectral slope in our power law cosmologies . in the power law cosmologies ,
the ratio of local small - scale to large - scale power depends primarily on the spectral index . when there is lots of small scale power , the central density profile is generally steeper at any specific radius than when there is little small scale power . in the case of a shallow spectral index ,
subhaloes form early , which implies that they have small characteristic radii and thus large characteristic densities , whereas steep spectral indices have late - forming subhaloes .
similar formation epoch arguments can also explain the steeper profiles of small @xmath15 haloes . for smaller @xmath15 haloes , many subhaloes should form earlier relative to their hosts , since the subhaloes lie at lower @xmath87 fluctuations in the density field relative to the local fluctuation on the scale of the host halo . alternatively ,
smaller @xmath15 haloes , which generally form earlier , and hence when the mean density of the universe is evolving more rapidly , will have subhaloes and cusp material assembled over a wider range of densities than for large @xmath15 haloes .
haloes with small @xmath15 are thus similar to haloes with a shallow spectral index , because they have subhaloes that form earlier at higher densities relative to the host halo . in models where the inner slope is formed from disrupted subhaloes infalling via dynamical friction , lots of small - scale power should give the main halo a larger concentration parameter .
in fact , when there are lots of dense subhaloes more of them will be able to reach closer to the core , steepening the halo nearer to the centre .
\(4 ) _ is there an asymptotic cusp ? _
a strength of the nfw profile , is that varying just the concentration radius can yield reasonable fits to haloes with very different inner slopes at scales currently resolved by simulations .
however , there are no compelling theoretical arguments that the halo profile converges to a slope of @xmath17 at smaller radii .
the analytical models mentioned in the introduction imply that the core slope should be a function of @xmath15 or power spectral slope .
it is possible , and theoretically motivated , that haloes of higher @xmath15 or steeper spectral slope than what we have simulated here would actually have a shallower slope than @xmath17 , which would be inconsistent with the nfw profile .
most of our haloes have density profiles that get ever shallower with decreasing radius down to our minimum resolved radius , although they are still consistent with asymptotic cusp slopes of @xmath17 or steeper . additionally , a few of our haloes appear to have converged to slopes at @xmath18 or steeper , which would be inconsistent with the nfw profile , though because of the degeneracy between central slope and concentration parameter , the nfw profile is not definitively ruled out .
simulations of haloes with very high @xmath15 , very steep power spectral indices , or orders of magnitude more particles able to probe much farther inward , should be able to test whether or not the r@xmath88 nfw central slope corresponds to a minimum possible slope set by the physical processes of cdm halo formation .
navarro et al .
( 2004 ) recently proposed a new profile form based on fits to a set of high resolution haloes .
the navarro et al .
profile differs from ours in that it has no asymptotic cusp , but instead continually flattens with decreasing radius .
it does not become shallower than r@xmath88 until radii smaller than currently resolvable , so it is generally consistent with most of our haloes . however , our profile form is more flexible in that it is able to better match haloes that have steep asymptotic cusps with large concentration radii , and is still a good match to those haloes that continuously flatten down to the minimum resolved radii in our simulations .
future simulations will likely need to probe below @xmath230.001 r@xmath89 in order to determine whether dark haloes have an asymptotic central cusp .
the mass distribution at such small radii has a strong effect on the flux of hypothesized dark matter annihilation signals that may be detectable via @xmath6-rays from the galactic centre ( e.g. stoehr et al .
2003 ; evans et al . 2004 ) .
\(5 ) _ profile scatter and the two parameter profile _ when we plot our haloes with the best fit concentration parameters , neither the nfw nor the m99 function provide a good fit to all of the haloes .
however , because each halo has a unique profile determined by poorly understood and complex high redshift events , any possible single parameter profile will be unlikely to fully describe all haloes . our two parameter fit describes the general trends with @xmath15 , providing a significant improvement over nfw and m99 , but it still does not account for the large halo to halo scatter at a given mass and redshift .
our evidence in the cubehi clusters for a likely correlation between the cusp material assembly epoch and cusp slope , and between halo accretion history and concentration radius , suggests that the density profiles might be at least partly determined from the evolutionary history of the progenitor haloes and substructures clumps .
larger sets of high resolution haloes should be able to quantify the role of mergers and other stochastic processes in shaping the density profile .
we are grateful to the anonymous referee for helpful and constructive comments and suggestions .
we thank lucio mayer for assistance with one of the runs .
dr has been supported by the nasa graduate student researchers program .
dr was partially supported by pparc .
fg is a david e. brooks research fellow .
fg was partially supported by nsf grant ast-0098557 at the university of washington .
trq was partially supported by the national science foundation .
simulations were performed on the origin 2000 at ncsa and nasa ames , the ibm sp4 at the arctic region supercomputing center ( arsc ) and at cineca ( bologna , italy ) , the nasa goddard hp / compaq sc 45 , and at the pittsburgh supercomputing center .
we thank chance reschke for dedicated support of our computing resources , much of which were graciously donated by intel .
abramson , i. s. 1982 , ann .
10 , 1217 bardeen , j.m . , bond , j.r . ,
kaiser , n. , szalay , a.s .
, 1986 , apj , 305 , 15 .
bennett , c. l. et al . , 2003 , apjs , 148 , 97b bentley j. l. , 1975 , communication of the acm 18 , 9 bullock j. s. , kolatt t. s. , sigad y. , somerville , r. s. , kravtsov a. v. , klypin a. a. , primack j. r. , dekel a. , 2001a , mnras , 321 , 559 bullock j. s. , dekel , a. , kolatt t. s. , kravtsov a. v. , klypin a. a. , porciani , c. , primack , j. r. , apj , 2001b , 555 , 240 cole s. , lacey c. , 1996 , mnras , 281 , 716 crone m. , evrard a. , richstone d. , 1994 , apj , 434 , 402 davis , m. , efstathiou , g. , frenk , c.s . ,
white , s.d.m .
, 1985 , apj , 292 , 381 de blok w. j. , mcgaugh s. s. , bosma a. , rubin v. c. , 2001 , apj , 552 , l23 diemand j. , moore b. , stadel j. , kazantzidis s. , 2004 , mnras , 348 , 977 eke v. r. , navarro j. f. , steinmetz m. , 2001 , apj 554 , 114 eke , v.r . , cole , s. , frenk , c.s .
, 1996 , mnras , 282 , 263 evans n. w. , ferrer f. , sarkar s. , 2003 , phrvd , 69 , 123501 flores r. a. , primack j. r. , 1994 , apj , 427 , l1 fukushige t. , kawai a. , makino j. , 2004 , apj , 606 , 625 fukushige , t. , & makino , j. 1997 , apj , 477 , l9 fukushige , t. , & makino , j. 2001 , apj , 557 , 533 fukushige , t. , & makino , j. 2003 , apj , 588 , 674 gavazzi r. , fort b. , mellier y. , pello r. , dantel - fort m. , 2003 , a&a , 403 , 11 gentile g. , salucci p. , klein u. , vergani d. , kalberla p , 2004 , mnras , 351 , 903 ghigna s. , moore b. , governato f. , lake g. , quinn t. , stadel j. , 1998 , mnras , 300 , 146 ghigna s. , moore b. , governato f. , lake g. , quinn t. , stadel j. , 2000 , 544 , 616 governato , f. , ghigna , s. , & moore , b. 2001 , asp conf .
ser . 245 : astrophysical ages and times scales , 469 hayashi , e. , et al . , 2003 , astro - ph/0310576 hernquist l. , 1990 , apj , 356 , 359 hoffman y. , shaham j. , 1985 , apj , 297 , 16 huffenberger k. m. , seljak u. , 2003 , mnras , 340 , 1199 huss a. , jain b. , steinmetz m. , 1999 , apj , 517 , 64 jimenez r. , verde l. , oh p. , 2003 ,
mnras , 339 , 243 jing , y. p. , & suto , y. 2000 , apj , 529 , l69 jing , y. p. , & suto , y. 2002 , apj , 574 , 538 katz n. , white s. , 1993 , apj , 412 , 478 klypin , a. , kravtson , a. v , , bullock , j. s. , & primack , j. r. 2001 , apj , 554 , 903 lacey c. , cole s. , 1994 , mnras , 271 , 676 lake , g. , 1989 , aj , 98 , 1253 lewis a. d. , buote d. a. , stock j. t. , 2003 , apj , 586 , 135 merritt d. , tremblay b. , 1994 , aj , 108 , 514 merritt , d. 1994 , http://www.rit.edu/@xmath23drmsps/inverse.html moore b. , 1994 , nature , 370 , 629 moore b. , governato f. , quinn t. , stadel j. , lake g. , 1998 , aj , 499 , l5 ( m98 ) moore b. , quinn t. , governato f. , stadel j. , & lake g. , 1999 ( m99 ) , mnras , 310 , 1147 moore b. , calcaneo - roldan c. , stadel j. , lake g. , sebastiano g. , governato g. , 2001 , phys . rev .
d , 64 , 063508 navarro j. f. , frenk c. s. , white s. d. m. , 1996 , apj , 462 , 563 ( nfwa ) navarro j. f. , frenk c. s. , white s. d. m. , 1997 , apj , 490 , 493 ( nfwb ) navarro j. f. , et al . , 2004 , mnras , 349 , 1039 power , c. , navarro , j. f. , jenkins , a. , frenk , c. s. , white , s. d. m. , springel , v .
, stadel , j. , & quinn , t. , 2003 , mnras , 338 , 14 press w.h . , schechter p. , 1974
, apj , 187 , 425 reed , d. , gardner , j. , quinn , t. , stadel , j. , fardal , m. , lake , g. , & governato , f. , 2003 , mnras , 346 , 565 ricotti m. , 2003 , mnras , 344 , 1237 salucci p. , burkert a. , 2000 , apj , 537 , l9 salucci p. , 2003
, astro - ph/0310376 sand d. j. , treu t. , ellis r. s. , 2002 , apj , 574 , l129 sand d. j. , treu t. , smith g. p. , ellis r. s. , 2004 , apj , 604 , 88 scott , david w. 1992 , `` multivariate density estimation '' , john wiley and sons , new york , p. 130
shapiro p. r. , iliev i. t. , 2000 , apj , 542l , 1 simon j. , bolatto a. , leroy a. , blitz l. , 2003 , apj , 596 , 957 spergel d. , et al . , 2003 ,
apjs , 148 , 175 splinter r. , melott a. , shandarin s. , suto y. , 1998 , apj , 497 , 38 stadel , j , 2001 , phdt .
stoehr f. , white s. , springel v. , tormen g. , & yoshida n. , 2003 , mnras , 345 , 1313 subramanian k. , cen r. , ostriker , j. p. , 2000 ,
apj , 538 , 528 syer d. , white s. d. m. , 1998 , mnras , 293 , 337 tasitsiomi a. , kravtsov a. , gottlober s. , klypin a. , 2004 , apj , 607 , 125 taylor j. , navarro j. , 2001 , apj , 563 483 tremaine s. , gunn j. , 1979 , phys .
42 , 407 tyson j. a. , kochanski g. p. , dellantontio i. p. 1998 , apj , 498 , l107 van den bosch f. c. , robertson b. e. dalcanton j. j. , de blok w. j. g. , 2000 , aj , 119 , 1579 van den bosch f. , swaters r. , 2001 , mnras , 325 , 1017 wadsley j. , stadel j. , quinn t. , 2004 , newa , 9 , 137 wechsler r. h. , bullock j. s. , primack j. r. , kravtsov a. v. , dekel a. , 2002 , apj , 568 , 52 .
zhao h. s. , 1996 , mnras , 278 , 488
here we present the algorithms which we used to derive smooth estimates , @xmath90 and @xmath91 , of the particle number density and surface density profiles from the @xmath92-body positions .
the routines in mapel ( merritt 1994 ) allow one to derive maximally unbiased estimates of @xmath93 and @xmath94 using penalty functions that embody the approximate power - law nature of these functions .
however the mapel routines are relatively slow , and this fact presented difficulties when constructing estimates using the @xmath95 particle data sets .
kernel based algorithms are faster but potentially more biased ; however we found that density profiles produced with mapel and kernel methods are in excellent agreement down to our mininum resolved radii .
thus we have used the kernel estimator in our analyses .
our derivation follows that in merritt & tremblay ( 1994 ) . in the absence of any symmetries in the particle distribution ,
a valid estimate of the number density @xmath93 corresponding to a set of particle positions @xmath96 is @xmath97\ ] ] where @xmath98 is the window width and @xmath99 is a normalized kernel , e.g. the gaussian kernel ( see fig .
a1 ) : @xmath100 now imagine that each particle is smeared uniformly around the surface of the sphere whose radius is @xmath101 and whose origin is at ( 0,0,0 ) . if the density profile is actually spherically symmetric , this smearing will leave the density unchanged ; if not , it will produce a spherically symmetric approximation to the true profile .
the spherically - symmetrized density estimate is we want to vary the window width with position in such a way that the bias - to - variance ratio of the estimate is relatively constant .
let @xmath108 be the window width associated with the @xmath109th particle .
the density estimate based on a variable window width is @xmath110 the optimal way to vary @xmath108 is according to abramson s ( 1982 ) rule : @xmath111 since we do nt know @xmath112 a priori , we instead varied @xmath108 according to @xmath113 we found that @xmath114 gave good results , which is reasonable since density profiles are close to @xmath115 .
one could improve on this by first constructing a pilot estimate of @xmath93 then using abramson s rule .
the surface density profile could be computed via simple projection of @xmath90 .
instead , we computed @xmath91 directly from the coordinates projected along one axis .
the two - dimensional kernel estimate of @xmath116 in the absence of any symmetries is @xmath117\ ] ] where @xmath118 is the two - dimensional gaussian kernel ,
@xmath119 now smear each particle uniformly in angle @xmath120 at fixed @xmath121 .
the density estimate becomes | we use numerical simulations in a @xmath0cdm cosmology to model density profiles in a set of sixteen dark matter haloes with resolutions of up to seven million particles within the virial radius .
these simulations allow us to follow robustly the formation and evolution of the central cusp over a large mass range of 10@xmath1 to 10@xmath2 @xmath3 , down to approximately 0.5@xmath4 of the virial radius , and from redshift 5 to the present , covering a larger range in parameter space than previous works .
we confirm that the cusp of the density profile is set at redshifts of two or greater and remains remarkably stable to the present time , when considered in non - comoving coordinates .
motivated by the diversity and evolution of halo profile shapes , we fit our haloes to the two parameter profile , @xmath5^{3-\gamma}}$ ] , where the steepness of the cusp is given by the asymptotic inner slope parameter , @xmath6 , and its radial extent is described by the concentration parameter , @xmath7 ( with @xmath7 defined as the virial radius divided by the concentration radius ) . in our simulations , we find @xmath8log@xmath9 for haloes of @xmath10 to @xmath11 , with a large scatter of @xmath12 , where @xmath13 is the redshift dependent characteristic mass of collapsing haloes ; and @xmath14 , with a large @xmath15 dependent scatter roughly equal to @xmath16 .
our redshift zero haloes have inner slope parameters ranging approximately from @xmath17 ( i.e. navarro , frenk , & white ) to @xmath18 ( i.e. moore et al . ) , with a median of roughly r@xmath19 .
this two parameter profile fit works well for all types haloes in our simulations , whether or not they show evidence of a steep asymptotic cusp .
we also model a cluster in power law cosmologies of @xmath20 , with @xmath21 ( 0 , -1 , -2 , -2.7 ) . here
we find that the concentration radius and the inner cusp slope are a both function of @xmath22 , with larger concentration radii and shallower cusps for steeper power spectra .
we have completed a thorough resolution study and find that the minimum resolved radius is well described by the mean interparticle separation over a range of masses and redshifts .
the trend of steeper and more concentrated cusps for smaller @xmath15 haloes clearly shows that dwarf sized @xmath0cdm haloes have , on average , significantly steeper density profiles within the inner few percent of the virial radius than inferred from recent observations .
code to reproduce this profile can be downloaded from http://www.icc.dur.ac.uk/@xmath23reed/profile.html .
[ firstpage ] galaxies : haloes galaxies : formation methods : n - body simulations cosmology : theory cosmology : dark matter |
quantum phenomenon has been observed by numerous experiments on microscopic scales . however , on macroscopic scales , it is difficult to find quantum effects , such as quantum superpositions .
a lot of physicists have been looking up the root of quantum - to - classical transition for decades .
the reason can be divided into two categories : coarsened measurement and decoherence@xcite .
commonly viewpoint is that decoherence plays a prominent role in quantum - to - classical transition .
there are two routes to explain decoherence : one route is that system interacts with external environments , the other is taken in wave function collapse@xcite , which need not external environments .
the latter one is often inspired by general relativity and makes a fundamental modification on quantum theory .
recently , igor at al.@xcite demonstrated the existence of decoherence induced by gravitational time dilation without any modification of quantum mechanics .
this work motivates further study on decoherence due to time dilation .
spontaneous emission between two atomic levels inevitably occurs .
we research decoherence due to time dilation during spontaneous emission . without spontaneous emission
, decoherence will not occur in our model only by time dilation . as we all know
, spontaneous emission can induce decoherence .
we find that gravitational time dilation can reduce or increase the decoherence due to spontaneous emission in different reference frames ( different zero potential energy point ) .
it is attributed to the fact that in different reference frame , the distinguishability of emission photon from different positions is different .
the direction of emission light also influences the coherence of quantum superpositions in fixed direction of gravitational field . in order to make the decoherence due to time dilation stronger than due to spontaneous emission
, time - delayed feedback control@xcite is used .
the rest of paper is arranged as follows . in section
ii , we present the model about the decoherence of quantum superpositions due to time dilation during spontaneous emission .
coherence of particle s position in different reference frame is explored in section iii . in section iv , we discuss the influence of different directions of emission light . in section
v , a time - delayed feedback scheme is utilized to increase decoherence induced by gravitational time dilation .
we deliver a conclusion and outlook in section vi .
firstly , let us simply review the gravitational time dilation which causes clocks to run slower near a massive object . given a particle of rest mass @xmath0 with an arbitrary internal hamiltonian @xmath1 , which interacts with the gravitational potential @xmath2 .
the total hamiltonian @xmath3 is described by@xcite @xmath4,\end{aligned}\ ] ] where @xmath5 is external hamiltonian . for a free particle , @xmath6 . in eq.(1 ) , the last term , @xmath7 , is simply the velocity - dependent special relativistic time dilation .
the coupling with position , @xmath8 , represents the gravitational time dilation .
when we consider slowly moving particles , @xmath9 , the gravitational time dilation will be the main source of time dilation .
it will not be canceled by the velocity - dependent special relativistic time dilation .
we consider that an atom with two levels is in superposition of two vertically distinct positions @xmath10 and @xmath11 .
the atom is coupled to a single unidirectional light field , as depicted in fig .
1 . decoherence of an atom is induced by gravitational time dilation under the situation of spontaneous emission .
the dotted line represents a homogeneous gravitational field @xmath12 , where @xmath13 is the gravitational acceleration on earth .
initial atom is in the superposition state : @xmath14 .
the direction of emitting photon is along the solid line , x - direction , which is contrary with the direction of gravitational field . ]
the whole system interacts with a homogeneous gravitational field @xmath12 which generates the gravitational time dilation .
the total system - field hamiltonian is described by ( @xmath15 ) @xmath16|x_1\rangle\langle x_1|\nonumber \\+\sqrt{\kappa_1/2\pi}(1+g x_1/c^2)|x_1\rangle\langle x_1|\int dw [ a b^\dagger(w)\exp(-i w x_1/c)+h.c]\nonumber\\ + [ mc^2+m g x_2 + w_1(1+g x_2/c^2)|1\rangle\langle1|+
w_2(1+g x_2/c^2)|2\rangle\langle2|]|x_2\rangle\langle x_2|\nonumber\\ + \sqrt{\kappa_2/2\pi}(1+g x_2/c^2)|x_2\rangle\langle x_2|\int dw [ a b^\dagger(w)\exp(-i w x_2/c)+h.c]+\int dw w b^\dagger(w)b(w),\end{aligned}\ ] ] where @xmath17 and @xmath18 ( @xmath19 ) are eigenvalues for the atomic level 1 and 2 , respectively , and operator @xmath20 .
@xmath21 and @xmath22 denote the coupling constants in position @xmath10 and @xmath11 , respectively . without extra control
, the two coupling constants should be same : @xmath23 . the last term in eq.(2 ) represents the free field hamiltonian , and the filed modes , @xmath24 , satisfy @xmath25=\delta(w - w')$ ] . using pauli operator , @xmath26 , to simplify the eq.(2 )
, we can obtain the new form of system - field hamiltonian @xmath27|x_1\rangle\langle x_1|+\sqrt{\kappa/2\pi}(1+g x_1/c^2)|x_1\rangle\langle x_1|\int dw [ a b^\dagger(w)\exp(-i w x_1/c)+h.c]\nonumber\\ + [ e_2+w_0/2(1+g x_2/c^2)\sigma_z]|x_2\rangle\langle x_2|+\sqrt{\kappa/2\pi}(1+g x_2/c^2)|x_2\rangle\langle x_2|\int dw [ a b^\dagger(w)\exp(-i w x_2/c)+h.c]\nonumber\\+\int dw w b^\dagger(w)b(w),\end{aligned}\ ] ] where @xmath28 for @xmath29 and @xmath30 .
we consider the initial field in the vacuum state and the atom in the state @xmath31 .
then , the atom will spontaneously emit photon . according to there being only a single excitation conservation between system and field@xcite , the system state in any time @xmath32
can be solved analytically , see appendix .
the quantum coherence of particle s position state can be quantified by the interferometric visibility @xmath33 , as shown in eq.(27 ) in appendix . when the time @xmath32 satisfy @xmath34 and @xmath35 , the amplitude of excitation state @xmath36 and @xmath37 .
then , we arrive at @xmath38 ^ 2+[w_0(\lambda_2-\lambda_1)]^2 } } \exp[-\lambda_1 ^ 2\kappa\tau],\end{aligned}\ ] ] where @xmath39 for @xmath29 . from the above equation
, we can see that the decoherence comes from the spontaneous emission ( when @xmath40 ) and the gravitational time dilation .
spontaneous emission can generate the decoherence due to the fact that photon is emitted from different positions , which leads to having a phase difference @xmath41 , where @xmath42 denotes the frequency of photon .
and we achieve that coherence depends on the reference frame .
different zero potential energy point ( different value of @xmath43 ) will give different coherence strength .
the counterintuitive result occurs because in different frame the phase difference will become different so that the distinguishability of emitting photon from two positions is different
. reducing the zero potential point ( increasing the value of @xmath43 ) , the phase difference will increase because of time dilation . for a fixed position difference @xmath44
, the quantum coherence can be rewritten @xmath45 ^ 2+(w_0\delta)^2 } } \exp[-\lambda_1 ^ 2\kappa\tau].\end{aligned}\ ] ] there is an optimal value of @xmath43 , which can give the maximal quantum coherence , as shown in fig .
2 . diagram of quantum coherence @xmath46 changing with @xmath43 .
the quantum coherence depends on reference frame .
the parameters are given by : @xmath47 , @xmath48 , @xmath49 . ] in an optimal reference frame , one can obseve the maximal coherence : @xmath46 is close to 1 . in order to observe the decoherence induced by gravitational time dilation
, it need to satisfy that the decoherence effect from time dilation is stronger than only from spontaneous emission ( @xmath40 ) : @xmath50 ^ 2+(w_0\delta)^2}}\exp[-\lambda_1 ^ 2\kappa\tau ] \ll\exp[-\kappa\tau ] . _ { } \end{aligned}\ ] ] noting that the value of @xmath51 is generally small in experiment , the condition @xmath52\gg1 $ ] is necessary for observing decoherence mainly induced by gravitational time dilation .
when one changes the direction of emitting photon , the quantum coherence will change accordingly , @xmath53
^ 2+[w_0(\lambda_2-\lambda_1)]^2 } } \exp[-\lambda_2 ^ 2\kappa\tau].\end{aligned}\ ] ] it is due to that the phase difference changes with the direction of emitting photon , becoming @xmath54 .
different directions of emitting photon in the fixed gravitational field will generate different quantum coherence @xmath46 .
then , we consider general three - dimensional space : the emitting photon can be along any direction , as shown in fig .
diagram shows that the photon can spontaneously emit in any direction . here
@xmath55 denotes the angle between direction of photon and x - direction , which can change from 0 to @xmath56 . ]
we obtain the quantum coherence of particle s position state as following : @xmath57 ^ 2+[w_0(\lambda_2-\lambda_1)]^2}}| \int_0^{\pi/2 } d\theta\sin\theta\cos^2\theta \exp[(iw_0\lambda_1-\lambda_1 ^ 2\kappa)\tau\cos\theta]+\exp[-(iw_0\lambda_2+\lambda_2 ^ 2\kappa)\tau\cos\theta]|,\\ & = \frac{3\kappa\lambda_1\lambda_2}{\sqrt{[\kappa(\lambda_1 ^ 2+\lambda_2 ^ 2)]^2+[w_0(\lambda_2-\lambda_1)]^2}}|[-2 + \exp(k_1 ) ( 2 - 2 k_1 + k_1 ^ 2)]/k_1 ^ 3+[-2 + \exp(k_2 ) ( 2 - 2 k_2 + k_2 ^ 2)]/k_2 ^ 3|,\\ & \textmd{in which},\nonumber\\ & k_j=[(-1)^jiw_0\lambda_j-\lambda_j^2\kappa]\tau , \ \textmd{for } \
j=1,2,\end{aligned}\ ] ] where the coupling strength between atom and light field changes with the direction of emitting photon , becoming @xmath58@xcite . for @xmath59 and @xmath60 , @xmath61 ^ 2+[w_0(\lambda_2-\lambda_1)]^2}}(1-\lambda_1 ^ 2\kappa\tau-\lambda_2 ^ 2\kappa\tau)<v'<v.$ ] it means that the quantum coherence in general three - dimensional space is smaller than in one - dimensional space of fixed direction . for @xmath62 and @xmath63 , @xmath61 ^ 2+[w_0(\lambda_2-\lambda_1)]^2}}|\cos3\varphi|(3/[(w_0 ^ 2\lambda_1 ^ 2+(\lambda_1 ^ 2\kappa\tau)^2]^{3/2}+3/[(w_0 ^ 2\lambda_1 ^ 2+(\lambda_1 ^ 2\kappa\tau)^2]^{3/2})\geq v,$ ] with @xmath64 .
it means that in new condition the quantum coherence in general three - dimensional space is larger than in one - dimensional space of fixed direction .
the root of generating @xmath65 is the phase difference changing from @xmath41 to @xmath66 .
when one chooses the center of two positions as the zero potential point , the interferometric visibility reads @xmath67 ^ 2+(w_0\delta)^2 } } \exp[-(1-\delta/2)^2\kappa\tau].\end{aligned}\ ] ] in order to observe the decoherence from gravitational time dilation , not from the spontaneous emission , it is necessary to satisfy condition @xmath68 $ ] .
however , the value of @xmath69 is very small in experiment .
so , the condition is hard to meet . we can utilize the time delay feedback @xcite to increase the decoherence from gravitational time dilation . diagram of time delay feedback .
the center of two positions is chosen as zero potential point .
so , in the new reference frame , the position @xmath10 ( @xmath11 ) is transformed to be @xmath70 ( @xmath71 ) . here , we just consider that the photon emits along the fixed x - direction , which can be easily generalized to the case of three - dimensional space . at @xmath72
, a mirror is put to reflect the light field , leading to that a time - delay light field is fed back to system - field interaction . ] as shown in fig .
4 , the light field is reflected by a mirror .
the whole system - field hamiltonian can be described by @xmath73\cos(2w r)+h.c\}\nonumber\\ & + \sqrt{\kappa/2\pi}(1+g x_2/c^2)|x_2\rangle\langle x_2|\int dw\{a b^\dagger(w)2\exp[-i w(r+\delta c^2/2g)]\cos [ w ( 2r+2\delta c^2/g)/c]+h.c\}\nonumber\\ & + \int dw w b^\dagger(w)b(w)+[e_1+w_0/2(1+g x_1/c^2)\sigma_z]|x_1\rangle\langle x_1|+[e_2+w_0/2(1+g x_2/c^2)\sigma_z]|x_2\rangle\langle x_2|,\end{aligned}\ ] ] using the way in appendix , we can obtain the quantum coherence at time @xmath74 . with the feedback , the spontaneous emission is suppressed due to superposition effect . the total system - field wave function can also be described by eq.(13 ) in appendix . when the conditions @xmath75 and @xmath76 hold , for @xmath74 the amplitudes @xmath77 $ ] and @xmath78 , where @xmath79 .
when @xmath80 and @xmath81 , @xmath82 .
so , we achieve that the quantum coherence @xmath83 . without gravitational time
dilation , the quantum coherence is much larger than 0 .
so , utilizing time - delayed feedback scheme can satisfy that the decoherence induced by the gravitational time dilation is far less than by spontaneous emission .
we explore the decoherence of an atom s positions induced by the gravitational time dilation only in the situation of spontaneous emission . as the phase difference of photon emitted from two positions are different in different reference frames , the quantum coherence of superposition state of positions depends on the reference frame . so one can choose proper reference frame to observe the decoherence from the gravitational time dilation .
it is worth mentioning that the direction of emitting photon will influence the quantum coherence . so comparing the case of fixed emitting direction with the case of any direction
, there are some differences about quantum coherence . when one chooses the center of two positions as the zero potential point , the decoherence induced by the gravitational time dilation is difficult to be far larger than by spontaneous emission .
the time delay feedback can be used to increase the decoherence from the time dilation with proper conditions . in this article
, we only discuss the decoherence of an atom with two energy levels induced by gravitational time dilation .
it is interesting to research the decoherence of many particles with many energy levels induced by time dilation with spontaneous emission . in this case
we believe that it will increase the decoherence effect from the gravitational time dilation .
and considering extra drive is the further research direction .
in this situation , due to the fact that a single excitation conservation between system and field do not hold , the question will become complex and rich .
this work was supported by the national natural science foundation of china under grant no .
11375168 .
maximilian schlosshauer , rev .
mod . phys . * 76 * , 1267 ( 2005 ) .
w. h. zurk , rev .
phys . * 75 * , 715 ( 2003 ) .
s. raeisi , p. sekatski , and c. simon , phys .
lett . * 107 * , 250401 ( 2011 ) .
h. jeong , m. paternostro , and t. c. ralph , phys .
lett . * 102 * , 060403 ( 2009 ) . h. jeong and t. c. ralph , phys .
* 97 * , 100401 ( 2006 ) .
p. caldara , a. la cognata , d. valenti , and b. spagnolo , int . j. quantum . inf . * 9 * , 119 ( 2011 ) .
d. kast and j. ankerhold , phys .
. lett . * 110 * , 010402 ( 2013 ) .
hyunseok jeong , youngrong lim , and m. s. kim , phys .
* 112 * , 010402 ( 2014 ) .
r. penrose , gen .
* 28 * , 581 - 600 ( 1996 ) .
l. di@xmath84si , phys .
a * 40 * , 1165 - 1174 ( 1989 ) .
a. bassi , k. lochan , s. satin , t.p .
singh , rev .
471- 527 ( 2013 ) . i. pikovski , m. zych , f. costa , and @xmath85 .
brukner , nat . phys .
* 11 * , 668 - 672 ( 2015 ) .
s. shankar , m. hatridge , z. leghtas , k. sliwa , a.narla , u.vool , s. m. girvin , l. frunzio , m. mirrahimi , and m. h. devoret , nature * 504 * , 419 ( 2013 ) . h. m. wiseman and g. j. milburn , phys .
a * 49 * , 4110 ( 1994 ) .
i. pikovski , m. zych , f. costa , and @xmath85 .
brukner , arxiv : 1508.03296v1 ( 2015 ) .
p. w. milonni and p. l. knight , phys .
rev . a * 10 * , 1096 ( 1974 ) .
marlan o.scully and m.suhail zubairy , quantum optics , cambrige university press , 1997 .
julia kabuss , dmitry o. krimer , stefan rotter , kai stannigel , andreas knorr , alexander carmele , arxiv:1503.05722v1 ( 2015 ) .
arne l. grimsmo , phys .
lett . * 115 * , 060402 ( 2015 ) .
@xmath86 + in the single photon limit , the total system - field wave function is described by @xmath87 , @xmath88 , @xmath89 and @xmath90 denote the corresponding amplitudes of the four states at time @xmath32 . applying the schr@xmath91dinger equation in the rotating frame , we arrive at the following set of partial differential equations : @xmath92c_1+(1+g x_1/c^2)\sqrt{\kappa/2\pi}\int dw \exp[i(wx_1/c - w)t]c_{1k},\\ i\partial_tc_{1w}=[e_1-w_0/2(1+g x_1/c^2)]c_{1k}+(1+g x_1/c^2)\sqrt{\kappa/2\pi}\int dw \exp[i(-wx_1/c+w)t]c_{1},\\ i\partial_tc_2=[e_2+w_0/2(1+g x_2/c^2)]c_1+(1+g x_2/c^2)\sqrt{\kappa/2\pi}\int dw \exp[i(wx_2/c - w ) t]c_{2k},\\
i\partial_tc_{2w}=[e_2-w_0/2(1+g x_2/c^2)]c_{2k}+(1+g x_2/c^2)\sqrt{\kappa/2\pi}\int dw \exp[i(-wx_2/c+w)t]c_{2}.\end{aligned}\ ] ] substituting @xmath93c_1 , c'_{1k}=\exp[-i(e_1-w_0/2(1+g x_1/c^2))t]c_{1k},$ ] @xmath94c_2 , c'_{2k}=\exp[-i(e_2-w_0/2(1+g x_2/c^2))t]c_{2k}$ ] into above equations , we arrive at the following simplified equations : @xmath95c'_{1w},\\ i\partial_tc'_{1w}=(1+g x_1/c^2)\sqrt{\kappa/2\pi}\int dw \exp[i(-wx_1/c+(w - w_0(1+g x_1/c^2))t]c'_{1},\\ i\partial_tc'_2=(1+g x_2/c^2)\sqrt{\kappa/2\pi}\int dw \exp[i(wx_2/c+(w_0(1+g x_2/c^2)-w)t]c'_{2w},\\ i\partial_tc'_{2w}=(1+g x_2/c^2)\sqrt{\kappa/2\pi}\int dw \exp[i(-wx_2/c+(w - w_0(1+g x_2/c^2))t]c'_{2}.\end{aligned}\ ] ] eq.(15 ) and eq.(17 ) are integrated formally and inserted into eq.(14 ) and eq.(16),respectively . utilizing the integral @xmath96=2\pi\delta(t),\end{aligned}\ ] ]
we can analytically solve the set of partial differential equations . at time
@xmath32 , using the initial values @xmath97 and @xmath98 , we obtain @xmath99,\\ c'_{1w}=\frac{1-\exp[-1/2\lambda_1
^ 2\kappa t - i(\lambda_1w_0-w)t]}{\lambda_1 ^ 2\kappa+i(\lambda_1w_0-w)}\sqrt{\kappa/2\pi}\lambda_1\exp[iwx_1/c],\\ c'_2(t)=1/\sqrt{2}\exp[-1/2\lambda_2 ^ 2\kappa t],\\ c'_{2w}=\frac{1-\exp[-1/2\lambda_2 ^ 2\kappa t - i(\lambda_2w_0-w)t]}{\lambda_2 ^ 2\kappa+i(\lambda_2w_0-w)}\sqrt{\kappa/2\pi}\lambda_2\exp[iwx_2/c],\end{aligned}\ ] ] where @xmath39 for @xmath29 .
the quantum coherence of position state can be quantified by the interferometric visibility @xmath100{c'_1}^*c'_2+\exp[i(x_2-x_1)w_2g / c^2t]\int dwc'^*_{1w}\int dw'c'_{2w'}|,\end{aligned}\ ] ] where the term @xmath101 can be integrated by residue theorem .
we can arrive at @xmath102-\exp[-1/2\lambda_1 ^ 2\kappa t+i\lambda_1w_0t+i\xi(\tau - t)]-\nonumber\\ & \exp[-1/2\lambda_2 ^ 2\kappa t - i\lambda_2w_0t+i(w_0\lambda_1+i\lambda_1 ^ 2\kappa)(\tau+t)]+\exp[-1/2\lambda_1 ^ 2\kappa ( t+\tau)-1/2\lambda_1 ^ 2\kappa t+i(\lambda_2-\lambda_1)w_0t+i\lambda_1w_0\tau]\},\\ & \textmd{in which},\nonumber\\ & \tau=(x_2-x_1)/c\geq0,\\ & \textmd{for}\ t<\tau , \
\xi = w_0\lambda_1+i\lambda_1 ^ 2\kappa,\ \
\textmd{for}\ t\geq\tau,\ \xi = w_0\lambda_2-i\lambda_2 ^ 2\kappa.\end{aligned}\ ] ] | we investigate decoherence of quantum superpositions induced by gravitational time dilation and spontaneous emission between two atomic levels . it has been shown that gravitational time dilation can be an universal decoherence source . here , we consider decoherence induced by gravitational time dilation only in the situation of spontaneous emission .
then , we obtain that the coherence of particle s position state depends on reference frame due to the time dilation changing the distinguishability of emission photon from two positions of particle .
changing the direction of light field can also result in the difference about the coherence of quantum superpositions . for observing the decoherence effect mainly due to gravitational time dilation
, time - delayed feedback can be utilized to increase the decoherence of particle s superpositions . |
we are interested in answering two very basic questions about continuous - time , discrete - symbol stochastic processes : * what are their minimal maximally predictive models their ? * what are information - theoretic characterizations of their randomness , predictability , and complexity ? for shorthand , we refer to the former as _ causal architecture _ and the latter as _ informational architecture_. minimal maximally predictive models of discrete - time , discrete - state , discrete - output processes are relatively well understood ; e.g. , see refs .
some progress has been made on understanding minimal maximally predictive models of discrete - time , continuous - output processes ; e.g. , see refs . @xcite .
relatively less is understood about minimal maximally predictive models of continuous - time , discrete - output processes , beyond those with exponentially decaying state - dwell times @xcite .
the following is a first attempt at a remedy that complements the spectral methods developed in ref .
@xcite , as we address the less tractable case of uncountably infinite causal states .
we start by analyzing continuous - time renewal processes , as addressing the challenges there carries over to other continuous - time processes .
( elsewhere , we outline the wide interest and applicability of renewal processes in physics and the quantitative sciences generally @xcite . )
the difficulties are both technical and conceptual .
first , the causal states are now continuous or hybrid discrete - continuous random variables , unless the renewal process is poisson .
second , transitions between causal states are now described by partial differential equations . finally , and perhaps most challenging , most informational architecture quantities must be redefined . with these challenges
addressed , we turn our attention to a very general class of continuous - time , discrete - alphabet processes stateful renewal processes generated by unifilar hidden semi - markov models . we identify their and find new expressions for entropy rate and other informational architecture quantities , extending results in ref .
@xcite .
our main thesis is rather simple : minimal maximally predictive models of continuous - time , discrete - symbol processes require a wholly new calculus . to develop it , sec .
[ sec : background ] describes the required new notation and definitions that enable extending the framework which is otherwise well understood for discrete - time processes @xcite .
sections [ sec : causalstates]-[sec : infoarch ] determine the causal and informational architecture of continuous - time renewal processes .
section [ sec : uhsmms ] characterizes the and calculates the entropy rate and excess entropy of unifilar hidden semi - markov models .
we conclude by describing potential applications to bayesian inference algorithms using new enumerations of topologies and to information measure estimation using the formulae of sec .
[ sec : uhsmms ] .
a continuous - time , discrete - symbol time series @xmath0 is described by a list of symbols @xmath1 in a finite alphabet @xmath2 and dwell times @xmath3 for those symbols . in this representation
, we demand that @xmath4 to enforce a unique presentation of the time series .
sections [ sec : causalstates]-[sec : infoarch ] focus on point processes for which @xmath5 . and so , in this case , we label the time series only with dwell times : @xmath6 .
we view the time series @xmath7 as a realization of random variables @xmath8 .
when the observed time series is strictly stationary and the process ergodic , in principle , we can calculate the probability distribution @xmath9 from a single realization @xmath7 . demarcating the present splits @xmath10 into two parts : the time @xmath11 since first emitting the previous symbol and the time @xmath12 to next symbol .
thus , we define @xmath13 as the _ past _ and @xmath14 as the _ future_. ( to reduce notation , we drop the @xmath15 indices . ) the _ present _ @xmath16 itself extends over an infinitesimally small length of time .
continuous - time renewal processes have a relatively simple generative model . _ interevent intervals _ @xmath17 are drawn from a probability density function @xmath18 . the _ survival function _
@xmath19 is the probability that an interevent interval is greater than or equal to @xmath20 and , in a nod to neuroscience , we define the _ mean firing rate _ @xmath21 as : @xmath22 the minimal generative model for a continuous - time renewal process is therefore a single causal - state machine with a continuous - value observable @xmath23 ; as shown in fig .
[ fig : genmodel ] . of periods of silence ( corresponding to output symbol @xmath24 )
are drawn independently , identically distributed ( iid ) from probability density @xmath18 . ]
a process _ forward - time causal states _ are defined , as usual , by the _
predictive _ equivalence relation @xcite , written here for the case of point processes : @xmath25 it is straightforward to write the predictive equivalence relation for continuous - time , discrete - alphabet point processes using the notation .
this partitions the set of allowed pasts .
each equivalence class of pasts is a forward - time causal state @xmath26 , in which @xmath27 is the function that maps a past to its causal state .
the set of forward - time causal states @xmath28 inherits a probability distribution @xmath29 from the probability distribution over pasts @xmath30 .
_ forward - time prescient statistics _ are any refinement of the forward - time causal - state partition . by construction , they are a sufficient statistic for prediction , but not necessarily _ minimal _ sufficient statistics @xcite
. _ reverse - time causal states _ are essentially forward - time causal states of the time - reversed process . in short , reverse - time causal states @xmath31 are the classes defined by the retrodictive equivalence relation , written here for the case of point processes : @xmath32 it is , again , straightforward to write the predictive equivalence relation for continuous - time , discrete - alphabet point processes using the notation given above . and , similarly , reverse - time causal states @xmath33 inherit a probability measure @xmath34 from the probability distribution @xmath35 over futures .
reverse - time prescient statistics are any refinement of the reverse - time causal - state partition .
they are sufficient statistics for retrodiction , but not necessarily minimal .
the main import of these definitions derives from the _ causal shielding _ relations : @xmath36 the consequence of these is illustrated in fig .
[ fig : intuition2 ] .
that is , arbitrary functions of the past and future do not shield the two aggregate past and future random variables from one another .
so , these causal shielding relations are special to prescient statistics , causal states , and their defining functions @xmath27 and @xmath37 .
forward and reverse - time generative models do not , in general , have state spaces that satisfy eqs .
( [ eq : cs1 ] ) and ( [ eq : cs2 ] ) .
( left oval , red ) , the future @xmath38 ( right oval , green ) , the forward - time causal states @xmath39 ( left circle , purple ) , and the reverse - time causal states @xmath40 ( right circle , blue ) .
@xcite . ) the forward - time and reverse - time statistical complexities are the entropies of @xmath39 and @xmath40 , i.e. , the memories required to losslessly predict or retrodict , respectively .
the excess entropy @xmath41 $ ] is a measure of process predictability ( central pointed ellipse , dark blue ) and theorem @xmath42 of ref .
@xcite shows that @xmath43 $ ] by applying the causal shielding relations in eqs .
( [ eq : cs1 ] ) and ( [ eq : cs2 ] ) . ]
the _ forward - time _ is that with state space @xmath39 and transition dynamic between forward - time causal states .
the _ reverse - time _ is that with state space @xmath40 and transition dynamic between reverse - time causal states .
defining these transition dynamics for continuous - time processes requires a surprising amount of care , as discussed in secs .
[ sec : causalstates]-[sec : infoarch ] .
we are broadly interested in information - theoretic characterizations of a process predictability , compressibility , and randomness .
a list of current quantities of interest , though by no means exhaustive , is given in figs .
[ fig : intuition2 ] and [ fig : intuition3 ] .
curiously , many lose meaning when naively applied to continuous - time processes ; e.g. , see refs . @xcite .
this section , as a necessity , will redefine many of these in relatively simple , but new ways to avoid trivial divergences and zeros .
the _ forward - time statistical complexity _
$ ] is the cost of coding the forward - time causal states and the _ reverse - time statistical complexity _
$ ] is the cost of coding reverse - time causal states .
when @xmath39 or @xmath40 are mixed or continuous random variables , one employs differential entropies for @xmath46 $ ] .
the result , though , is that the statistical complexities are potentially negative or infinite or both ( * ? ? ?
8.3 ) , perhaps undesirable characteristics for a definition of process complexity .
this definition , however , allows for consistency with complexity definitions for discretized continuous - time processes .
@xcite for possible alternatives for @xmath46 $ ] . together , a process _ causal irreversibility
_ @xcite is defined as the difference between the forward and reverse - time statistical complexities : @xmath47 if the forward- and reverse - time process are isomorphic i.e . , if the process is temporally reversible then @xmath48 .
renewal processes are temporally symmetric : @xmath48 @xcite . as such
, we will refer to forward - time causal states and the forward - time as simply causal states or the , with the understanding that reverse - time causal states and reverse - time will take the exact same form with slight labeling differences .
we start by describing prescient statistics for continuous - time processes .
the lemma which does this exactly parallels that of lemma @xmath42 of ref .
the only difference is that the prescient statistic is the _ time _ since last event , rather than the number of @xmath24s ( count ) since last event .
the time @xmath11 since last event is a prescient statistic of renewal processes .
[ lem : ctrp_prescient ] from bayes rule : @xmath49 interevent intervals @xmath17 are independent of one another , so @xmath50 .
the random variables @xmath11 and @xmath12 are functions of @xmath10 and the location of the present . both @xmath11 and @xmath12 are independent of other interevent intervals .
and so , @xmath51 . this implies : @xmath52 the predictive equivalence relation groups two pasts @xmath53 and @xmath54 together when @xmath55 .
we see that @xmath56 is a sufficient condition for this from eq .
( [ eq : lemproof ] ) . the lemma follows .
some renewal processes are quite predictable , while others are purely random .
a poisson process is the latter : interevent intervals are drawn independently from an exponential distribution and so knowing the time since last event provides no predictive benefit
. a fractal renewal process can be the former . there
, the interevent interval is so structured that the resultant process can have power - law correlations @xcite .
then , knowing the time since last event can provide quite a bit of predictive power @xcite .
intermediate between these two extremes is a broad class of renewal processes whose interevent intervals are structured up to a point and then fall off exponentially only after some time @xmath57 .
these intermediate cases can be classified as either of the following types of renewal process , in analogy with ref .
@xcite s classification .
note that an eventually @xmath58-poisson process , but not an eventually poisson process , will generally have a discontinuous @xmath18 .
an _ eventually poisson process _ has : @xmath59 for some @xmath60 and @xmath61 almost everywhere .
we associate the eventually poisson process with the minimal such @xmath62 .
[ def : ep ] an _ eventually @xmath58-poisson process _ with @xmath63 has an interevent interval distribution satisfying : @xmath64 for the smallest possible @xmath57 for which @xmath65 exists .
[ def : edp ] a familiar example of an eventually poisson process is found in the spike trains generated by poisson neurons with refractory periods @xcite .
there , the neuron is effectively prevented from firing two spikes within a time @xmath62 of each other the period during with its ion channels re - energize the membrane voltage to their nonequilibrium steady state .
after that , the time to next spike is drawn from an exponential distribution . to exactly predict the spike train s future
, we must know the time since last spike , as long as it is less than @xmath62 .
we gain a great deal of predictive power from that piece of information .
however , we do not care much about the time since last spike exactly if it is greater than @xmath62 , since at that point the neuron acts as a memoryless poisson neuron .
these intuitions are captured by the following classification theorem .
a renewal process has three different types of causal state : 1 . when the renewal process is not eventually @xmath58-poisson , the causal states are the time since last event ; 2 . when the renewal process is eventually poisson , the causal states are the time since last event up until time @xmath57 ; or 3
when the renewal process is eventually @xmath58-poisson , the causal states are the time since last event up until time @xmath57 and are the times since @xmath57 mod @xmath58 thereafter .
[ the : renewcausalstates ] lemma [ lem : ctrp_prescient ] implies that two pasts are causally equivalent if they have the same time since last event , if @xmath66 . from lemma [ lem : ctrp_prescient ]
s proof , we further see that two times since last event are causally equivalent when @xmath67 . in terms of @xmath18
, we find that : @xmath68 using manipulations very similar to those in the proof of thm .
@xmath42 of ref .
so , to find causal states , we look for @xmath69 such that : @xmath70 for all @xmath71 . to unravel the consequences of this
, we suppose that @xmath72 without loss of generality .
define @xmath73 and @xmath74 , for convenience .
the predictive equivalence relation can then be rewritten as : @xmath75 for any @xmath71 , where @xmath76 .
iterating this relationship , we find that : @xmath77 this immediately implies the theorem s first case . if a renewal process is _ not _
eventually @xmath58-poisson , then @xmath78 for all @xmath71 implies @xmath66 , so that the prescient statistics of lemma [ lem : ctrp_prescient ] are also minimal . to understand the theorem s last two cases , we consider more carefully the set of all pairs @xmath79 for which @xmath80 for all @xmath71 holds .
define the set : @xmath81 and define the parameters @xmath57 and @xmath65 by : @xmath82 and : @xmath83 note that @xmath57 and @xmath65 defined in this way are unique and exist , as we assumed that @xmath84 is nonempty . when @xmath85 , then the process is eventually @xmath58-poisson .
if @xmath86 , then the process must be an eventually poisson process with parameter @xmath57 . to see this , we return to the equation : @xmath87 and rearrange terms to find : @xmath88 as @xmath89 , we can take the limit that @xmath90 and we find that : @xmath91 the righthand side is a parameter independent of @xmath92 .
so , this is a standard ordinary differential equation for @xmath18 .
it is solved by @xmath93 for @xmath94 .
theorem [ the : renewcausalstates ] implies that there is a qualitative change in @xmath39 depending on whether or not the renewal process is poisson , eventually poisson , eventually @xmath58-poisson , or not eventually poisson . in the first case , @xmath39 is a discrete random variable ; in the second case , @xmath39 is a mixed discrete - continuous random variable ; and in the third and fourth cases , @xmath39 is a continuous random variable .
identifying causal states in continuous - time follows an almost entirely similar path to that used for discrete - time renewal processes in ref .
the seemingly slight differences between the causal states of eventually poisson , eventually @xmath58-poisson , and not eventually @xmath58-poisson renewal processes , however , have surprisingly important consequences for continuous - time . as described by thm .
[ the : renewcausalstates ] , there are often an uncountable infinity of continuous - time causal states . as one might anticipate from refs .
@xcite , however , there is an ordering to this infinity of causal states that makes calculations tractable .
there is one major difference between discrete - time and continuous - time : transition dynamics often amount to specifying the evolution of a probability density function over causal - state space . as such
, a continuous - time constitutes an unusual presentation of a hidden markov model : they appear as a system of conveyor belts or , under special conditions , like conveyor belts with a trash bin or a second mini - conveyor belt . beyond the picaresque metaphor , in fact they operate like conveyor belts in that they transport the time since the last event , resetting it here and there in a stateful way .
unsurprisingly , the exception to this general rule is given by the poisson process itself .
the of a poisson process is exactly the minimal generative model shown in fig .
[ fig : genmodel ] . at each iteration
, an interevent interval is drawn from a probability density function @xmath95 , with @xmath96 . knowing the time since last event does not aid in predicting the time to next event , above and beyond knowing @xmath97 . and
so , the poisson has only a single state . in the general
setting , though , the dynamic describes the evolution of the probability density function over its causal states .
how to represent this ?
we might search for labeled transition operators @xmath98 such that @xmath99 , giving partial differential equations that govern the labeled - transition dynamics .
, tracking the time since last event and depicted as the semi - infinite horizontal line , are isomorphic with the positive real line .
if no event is seen , probability flows towards increasing time since last event , as described in eq .
( [ eq : mathcalo0 ] ) .
otherwise , arrows denote allowed transitions back to the reset state or `` @xmath24 node '' ( solid black circle at left ) , denoting that an event occurred . ]
the of a renewal process that is not eventually poisson takes the state - transition form shown in fig .
[ fig : nedp ] .
let @xmath100 be the probability density function over the causal states @xmath101 at time @xmath20 .
our approach to deriving labeled transition dynamics parallels well - known approaches to determining fokker - planck equations using a kramers - moyal expansion @xcite . here
, this means that any probability at causal state @xmath101 at time @xmath102 could only have come from causal state @xmath103 at time @xmath20 , if @xmath104 .
this implies : @xmath105 however , @xmath106 is simply the probability that the interevent interval is greater than @xmath101 , given that the interevent interval is at least @xmath107 , or : @xmath108 together , eqs .
( [ eq : probflow1 ] ) and ( [ eq : probflow2 ] ) imply that : @xmath109 from this , we obtain : @xmath110 hence , the labeled transition operator @xmath111 given no event takes the form : @xmath112 the probability density function @xmath100 changes discontinuously after an event occurs , though .
all probability mass shifts from @xmath113 resetting back to @xmath114 : @xmath115 in other words , an event `` collapses the wavefunction '' . the stationary distribution @xmath116 over causal states is given by setting @xmath117 to @xmath24 and solving .
( at the risk of notational confusion , we adopt the convention that @xmath116 denotes the stationary distribution and that @xmath100 does not . ) straightforward algebra shows that : @xmath118 from this , the continuous - time statistical complexity directly follows : @xmath119 this was the nondivergent component of the infinitesimal time - discretized renewal process statistical complexity found in ref .
@xcite .
are isomorphic with the real line only to @xmath120 $ ] , as they again denote time since last event .
a leaky absorbing node at @xmath57 ( solid white circle at right ) corresponds to any time since last event after @xmath57 .
if no event is seen , probability flows towards increasing time since last event or the leaky absorbing node , as described in eqs .
( [ eq : mathcalo0 ] ) and ( [ eq : mathcalo0b ] ) . when an event occurs the process transitions ( curved arrows ) back to the reset state
node @xmath24 ( solid black circle at left ) . ] as thm . [ the : renewcausalstates ]
anticipates , there is a qualitatively different topology to the of an eventually poisson renewal process , largely due to the continuous - time causal states being mixed discrete - continuous random variables . for @xmath121 ,
there is `` wave '' propagation completely analogous to that described in eq .
( [ eq : mathcalo0 ] ) of sec .
[ sec : nedp ] .
however , there is a new kind of continuous - time causal state at @xmath122 , which does not have a one - to - one correspondence to the dwell time .
instead , it denotes that the dwell time is _ at least _ some value ; viz . , @xmath57 .
new notation follows accordingly : @xmath100 , defined for @xmath123 , denotes a probability density function for @xmath121 and @xmath124 denotes the probability of existing in causal state @xmath125 .
normalization , then , requires that : @xmath126 the transition dynamics for @xmath124 are obtained similarly to that for @xmath100 , in that we consider all ways in which probability flows to @xmath127 in a short time window @xmath128 .
probability can flow from any causal state with @xmath129 or from @xmath122 itself .
that is , if no event is observed , we have : @xmath130 the term @xmath131 corresponds to probability flow from @xmath122 and the integrand corresponds to probability influx from states @xmath132 with @xmath133 .
assuming differentiability of @xmath124 with respect to @xmath20 , we find that : @xmath134 where @xmath135 is shorthand for @xmath136 .
this implies that the labeled transition operator @xmath111 takes a piecewise form which acts as in eq .
( [ eq : mathcalo0 ] ) for @xmath121 and as in eq .
( [ eq : mathcalo0b ] ) for @xmath122 . as earlier , observing an event causes the `` wavefunction collapse '' to a delta distribution at @xmath114 .
the causal - state stationary distribution is determined again by setting @xmath137 and @xmath138 to @xmath24 .
equivalently , one can use the prescription suggested by thm . [ the : renewcausalstates ] to calculate @xmath139 via integration of the stationary distribution over the prescient machine given in sec .
[ sec : nedp ] : @xmath140 if we recall that @xmath141 , we find that : @xmath142 the process continuous - time statistical complexity precisely , entropy of this mixed random variable is given by : @xmath143 this is the sum of the nondivergent @xmath144 component and the rate of divergence of @xmath144 of the infinitesimal time - discretized renewal process @xcite . -poisson renewal process : graphical elements as in the previous figure .
the circular causal - state space at @xmath57 ( circle on right ) has total duration @xmath65 , corresponding to any time since last event after @xmath57 mod @xmath65 . if no event is seen , probability flows as indicated around the circle , as described in eq .
( [ eq : mathcalo0 ] ) . ] probability wave propagation equations , like those in eq .
( [ eq : mathcalo0 ] ) , hold for @xmath121 and for @xmath145 . at @xmath122 , if no event is observed , probability flows in from both @xmath146 and from @xmath147 , giving rise to the equation : @xmath148 unfortunately , there is a discontinuous jump in @xmath100 at @xmath122 coming from @xmath147 and @xmath149 . and
so , we can not taylor expand either @xmath150 or @xmath151 about @xmath152 .
again , we can use the prescription suggested by thm .
[ the : renewcausalstates ] to calculate the probability density function over these causal states and , from that , calculate the continuous - time statistical complexity .
below @xmath121 , the probability density function over causal states is exactly that described in sec .
[ sec : nedp ] : @xmath153 . for @xmath154 ,
the probability density function becomes : @xmath155 recalling def .
[ def : edp ] , we see that @xmath156 and so find that for @xmath157 : @xmath158 altogether , this gives the statistical complexity : @xmath159 [ cols= " < , < " , ]
we define continuous - time information anatomy @xcite quantities as _ rates_. as mentioned earlier , the present extends over an infinitesimal time . to define information anatomy rates , we let @xmath160 be the symbols observed over an arbitrarily small length of time @xmath161 , starting at the present @xmath162
. it could be that @xmath160 encompasses some portion of @xmath163 ; the notation leaves this ambiguous .
the entropy rate is now : @xmath164}{d\delta } ~. \label{eq : hmupercs}\end{aligned}\ ] ] this is equivalent to the more typical random - variable `` block '' definition of entropy rate @xcite : @xmath165 / \delta$ ] .
similarly , we define the _ single - measurement entropy rate _ as : @xmath166}{d\delta } ~ , \label{eq : h0percs}\end{aligned}\ ] ] the _ bound information rate _ as : @xmath167}{d\delta } ~ , \label{eq : bmupercs}\end{aligned}\ ] ] the _ ephemeral information rate _ as : @xmath168}{d\delta } ~,\end{aligned}\ ] ] and the _ co - information rate _ as : @xmath169}{d\delta } ~.\end{aligned}\ ] ] in direct analogy to discrete - time process information anatomy , we have the relationships : @xmath170 so , the entropy rate @xmath171 , the instantaneous rate of information creation , again decomposes into a component @xmath172 that represents active information storage and a component @xmath173 that represents `` wasted '' information . ):
information diagram for the past @xmath174 , infinitesimal present @xmath160 , and future @xmath175 .
the measurement entropy rate @xmath176 is the rate of change of the single - measurement entropy @xmath177 $ ] at @xmath178 .
the ephemeral information rate @xmath179 $ ] is the rate of change of useless information generation at @xmath178 .
the bound information rate @xmath180 $ ] is the rate of change of active information storage . and , the co - information rate @xmath181 $ ] is the rate of change of shared information between past , present , and future .
these definitions closely parallel those in ref . @xcite . ] prescient states ( not necessarily _ minimal _ ) are adequate for deriving all information measures aside from @xmath182 . as such
, we focus on the transition dynamics of noneventually @xmath58-poisson and , implicitly , their bidirectional machines .
to find the joint probability density function of the time to next event @xmath183 and time since last event @xmath184 , we note that @xmath185 is an interevent interval ; hence : @xmath186 the normalization factor of this distribution is : @xmath187 so , the joint probability distribution is : @xmath188 equivalently , we could have calculated the conditional probability density function of time - to - next - event given that it has been at least @xmath184 since the last event .
this , by similar arguments , is @xmath189 .
this would have given the same expression for @xmath190 . to find the excess entropy ,
we merely need calculate @xcite : @xmath191 \\ & = \int_0^{\infty } \int_0^{\infty } \mu\phi ( { { \sigma } } ^+ , { { \sigma } } ^- ) \log \frac{\mu \phi ( { { \sigma } } ^+ , { { \sigma } } ^-)}{\phi ( { { \sigma } } ^+)\phi ( { { \sigma } } ^- ) } d { { \sigma } } ^+ d { { \sigma } } ^- ~.\end{aligned}\ ] ] algebra not shown here gives : @xmath192 unsurprisingly @xcite , this agrees with the formula given in ref .
@xcite , which was derived by considering the limit of infinitesimal time discretization .
now , we turn to the more technically challenging task of calculating differential information anatomy rates .
suppose that @xmath193 is a random variable for paths of length @xmath161 .
each path is uniquely specified by a list of times of events .
let @xmath194 be a random variable defined by : @xmath195 we first illustrate how to find @xmath176 , since the same technique allows calculating @xmath171 .
we can rewrite the path entropy as : @xmath196 = { \operatorname{h}}[x_{\delta } ] + { \operatorname{h}}[\gamma_{\delta}|x_{\delta } ] ~.\end{aligned}\ ] ] for renewal processes , when @xmath21 can be defined , we see that : @xmath197 straightforward algebra shows that : @xmath198 & = \mu\delta - \mu\delta\log ( \mu\delta ) + o(\delta^2\log\delta ) ~.\end{aligned}\ ] ] we would like to find a similar asymptotic expansion for @xmath199 $ ] , which can be rewritten as : @xmath200 & = { \pr}(x_{\delta}=0 ) { \operatorname{h}}[\gamma_{\delta}|x_{\delta}=0 ] \nonumber \\ & \qquad + { \pr}(x_{\delta}=1 ) { \operatorname{h}}[\gamma_{\delta}|x_{\delta}=1 ] \nonumber \\ & \qquad + { \pr}(x_{\delta}=2 ) { \operatorname{h}}[\gamma_{\delta}|x_{\delta}=2 ] ~.\end{aligned}\ ] ] first , we notice that @xmath193 is deterministic given that @xmath201the path of all silence .
so , @xmath202 = 0 $ ] .
second , we can similarly ignore the term @xmath203 $ ] since @xmath204 is @xmath205 and , we claim , @xmath206 $ ] is @xmath207 : by standard maximum entropy arguments , @xmath206 $ ] is at most @xmath208 , and by noting that trajectories with only one event are a strict subset of trajectories with more than one event but with multiple events arbitrarily close to one another , @xmath206\geq { \operatorname{h}}[\gamma_{\delta}|x_{\delta}=1]$ ] which , by arguments below , is @xmath207 .
thus , the term @xmath209 $ ] is @xmath210 at most . finally , to calculate @xmath211 $ ]
, we note that when @xmath212 , paths can be uniquely specified by an event time , whose probability is @xmath213 .
a taylor expansion about @xmath214 shows that @xmath215 for some @xmath216 in which @xmath217 for all @xmath218 .
so , overall , we find that : @xmath219 where @xmath220 for any @xmath193 with at least one event in the path .
the largest corrections to @xmath221 come from ignoring the paths with two or more events , rather than from approximating all paths with only one event as equally likely . in sum
, we see that : @xmath200 & = \mu\delta\log\delta + o(\delta^2\log\delta ) ~.\end{aligned}\ ] ] together , these manipulations give : @xmath196 & = \mu\delta - \mu\delta\log\mu + o(\delta^2\log\delta ) ~.\end{aligned}\ ] ] this then implies : @xmath222}{d\delta } \\ & = \mu-\mu\log\mu ~.\end{aligned}\ ] ] a similar series of arguments helps to calculate @xmath223 defined in eq .
( [ eq : hmupercs ] ) , where now , @xmath21 is replaced by @xmath224 : @xmath225 which gives : @xmath226 algebra ( namely , integration by parts ) not shown here yields the expression : @xmath227 as expected , this is the nondivergent component of the expression given in eq .
( @xmath228 ) of ref .
@xcite for the @xmath161-entropy rate of renewal processes . and
, it agrees with expressions derived in alternative ways @xcite .
we need slightly different techniques to calculate @xmath172 , as we no longer need to decompose a path entropy . from eq .
( [ eq : bmupercs ] ) , we have : @xmath229}{d\delta } ~.\end{aligned}\ ] ] let s develop a short - time @xmath161 asymptotic expansion for @xmath230 .
first , we notice that @xmath231 , so that : @xmath232 we already can identify : @xmath233 to understand @xmath234 , we expand : @xmath235 recall that @xmath236 is @xmath205 , that : @xmath237 and that : @xmath238 then , straightforward algebra not shown gives : @xmath239 this can be used to derive : @xmath240 in nats .
when @xmath241 , for instance , @xmath242 for all @xmath184 , confirming in a much more complicated calculation that poisson processes really are memoryless .
this allows us to calculate the total @xmath172 as : @xmath243 in nats . and , from this , we find @xmath173 using : @xmath244 continuing
, we calculate @xmath245 from : @xmath246 and , we calculate @xmath247 via : @xmath248 all these quantities are gathered in table [ tab:1 ] , which gives them in bits rather than nats .
the of discrete - time , discrete - symbol processes are well understood and , as we now appreciate from secs . [
sec : causalstates]-[sec : infoarch ] , the predictive equivalence relation defining them readily applies to continuous - time renewal processes .
this gives the latter s analogous maximally predictive models : continuous or hybrid discrete - continuous , when minimal .
here , we introduce a new class of process generators that are unifilar versions of ref .
@xcite s hidden semi - markov models , but whose dwell time distributions can take any form . ( note that general semi - markov models are a strict subset . ) roughly speaking , they are stateful renewal processes , but this needs to be clarified .
many of their calculations reduce to those in secs . [
sec : causalstates]-[sec : infoarch ] .
when appropriate , we skip these steps .
we start by introducing the minimal generative models in fig .
[ fig : uhsmm ] .
let @xmath249 be the set of states in this generative model . each state
@xmath250 emits a symbol @xmath251 and a dwell time @xmath252 for that symbol , and , based on the state @xmath253 and emitted symbol @xmath1 , transitions to a new state @xmath254 .
we assume that the underlying generative model is _ unifilar _ : that the new state @xmath253 is uniquely specified by the prior state @xmath253 and emitted symbol @xmath1 .
we introduce a perhaps unfamiliar restriction on the labeled transition matrices @xmath255 .
define @xmath256 and @xmath257 .
then , we focus only on generative models for which @xmath258 .
this simply ensures that there is no uncertainty in when one dwell time finishes and another begins .
for example , consider the generator in fig .
[ fig : uhsmm](bottom ) : if states @xmath259 and @xmath260 were both to emit a @xmath24 in succession , it would be impossible to tease apart when the process switched from state @xmath259 to state @xmath260 .
the restriction introduces no loss of generality for our purposes . ,
@xmath261 , @xmath262 , and @xmath263 are isomorphic to @xmath264 . during an event interval ,
symbol @xmath265 is emitted .
a transition occurs to a new event when the state dwell time is exhausted at @xmath266 , which is distributed according to @xmath267 .
( bottom ) generative model with three hidden states ( @xmath268 , @xmath259 , and @xmath260 ) emits symbols @xmath269 for dwell times @xmath270 drawn from probability density functions @xmath271 , @xmath272,@xmath273 , and @xmath274 , respectively .
( transition labels as in fig .
[ fig : genmodel ] . ) ] a prescient model of this combined process is shown in fig .
[ fig : uhsmm](top ) .
each state @xmath250 comes equipped with one or more renewal process - like tails ( semi - infinite spaces that act as continuous counters ) that generically take the form of fig .
[ fig : nedp ] . the leakiness of these ( dissipative ) counters is given by @xmath275 , the probability density function from which the dwell time is drawn .
this new form of state - transition diagram depicts the of these hidden semi - markov processes .
moreover , if one or more of the dwell - time distributions gives an eventually poisson or an eventually-@xmath58 poisson structure , the presentation in fig .
[ fig : uhsmm](top ) is a prescient machine , but not the .
more generally , any such unifilar minimal generative model has a prescient machine with a `` node '' for each underlying hidden state @xmath253 and as many counters as needed one for every almost - everywhere unique @xmath276 .
each counter leaks probability to the next underlying hidden state @xmath254 , which is completely determined by @xmath253 and @xmath1 .
the presentation in fig .
[ fig : uhsmm](top ) is a prescient machine for the process generated by the unifilar hidden semi - markov model of fig .
[ fig : uhsmm](bottom ) .
[ the:2 ] to show that this is a prescient machine , we need to show that the present model state consisting of hidden state @xmath253 , current emitted symbol @xmath277 , and dwell time @xmath270 is uniquely specified by the observed past almost surely . the observed symbol @xmath1 is given by the current symbol in the observed past .
the restriction on successive emitted symbols ( that @xmath278 ) implies that the observed dwell time @xmath270 is exactly the observed length of @xmath1 .
finally , the underlying hidden state @xmath253 is determined uniquely by a _ function _ of the past almost surely , in which all dwell - time information is removed , by assumption : the restriction mentioned earlier implies there is no uncertainty in when one dwell time finishes and another begins . and , the unifilarity of the dynamic on hidden states @xmath253 implies that the sequence of symbols in the observed past are sufficient to specify the hidden state @xmath253 almost surely .
hence , @xmath253 is determined uniquely from the observed past almost surely .
the theorem follows . theorem [ the:2 ] can be straightforwardly generalized to specify conditions under which the presentation is an , a minimal prescient machine , by incorporating the conditions of thm .
[ the : renewcausalstates ] .
the stationary distribution for @xmath279 directly follows the treatment for the continuous - time renewal processes in sec .
[ sec : nedp ] , and so : @xmath280 where @xmath281 .
then we note that : @xmath282 and so : @xmath283 to find @xmath284 , we again calculate the probability mass dumped at @xmath285 in terms of @xmath284 : @xmath286 after a straightforward substitution of eq .
( [ eq : foo ] ) and noting that @xmath287 , we find : @xmath288 so : @xmath289 let @xmath290 be the stationary distribution for the underlying discrete - state : @xmath291 where the eigenvector is normalized such that the sum of its entries is @xmath42 .
then : @xmath292 or , rewriting and normalizing , we have : @xmath293 altogether , we find that the steady - state distribution is given by : @xmath294 using the formulae for entropies of mixed random variables @xcite , we find a statistical complexity of : @xmath295 \\ & = \left\langle { \operatorname{h}}[\rho(\tau|g , { { x } } ) ] \right\rangle_{p(g , { { x } } ) } + { \operatorname{h}}[p(g , { { x } } ) ] \\ & = \left\langle \int_0^{\infty } \mu_{g , { { x } } } \phi_{g , { { x } } } ( \tau ) \log \frac{1}{\mu_{g , { { x } } } \phi_{g , { { x } } } ( \tau ) } d\tau \right\rangle_{p(g , { { x } } ) } \\ & \qquad + { \operatorname{h}}\left [ \frac{\pi(g ) t_g^ { ( { { x } } ) } /\mu_{g , { { x } } } } { \sum_{g ' , { { x } } ' } \pi(g ' ) t_{g'}^ { ( { { x } } ' ) } /\mu_{g ' , { { x } } ' } } \right ] ~.\end{aligned}\ ] ] note that @xmath296 $ ] is the statistical complexity of the underlying discrete - time and that @xmath297 $ ] is the statistical complexity of a noneventually @xmath58-poisson renewal process with interevent distribution @xmath298 , averaged over @xmath253 and @xmath1 .
hence , the statistical complexity of these unifilar hidden semi - markov processes differs from the statistical complexity of its `` components '' by : @xmath299 - { \operatorname{h}}[\pi(g ) ] ~.\end{aligned}\ ] ] whether this difference is positive or negative depends on both matrices @xmath300 . in general
, we expect the difference to be positive . since there are multiple observed symbols @xmath301 generated by these machines , @xmath176 ( and so @xmath247 ) and @xmath172 as defined in sec .
[ sec : infoarch ] diverge .
however , the entropy rate @xmath171 and excess entropy @xmath302 as defined in sec .
[ sec : infoarch ] do not diverge for processes generated by this restricted class of unifilar hidden semi - markov models . from the steady - state distribution given in eq .
( [ eq : steadystate ] ) and from the entropy rate expressions in eqs .
( [ eq : hmust])-([eq : hmufinal ] ) of sec .
[ sec : infoarch ] , we immediately have the entropy rate for these unifilar hidden semi - markov models : @xmath303}{d\delta } \nonumber \\ & = \sum_{g , { { x } } } \rho(g , { { x } } ) \left(-\mu_{g , { { x } } } \int_0^{\infty } \phi_{g , { { x } } } ( \tau)\log \phi_{g , { { x } } } ( \tau ) d\tau\right ) \nonumber \\ & = -\frac{\sum_{g , { { x } } } \pi(g ) t_g^ { ( { { x } } ) } \int_0^{\infty } \phi_{g , { { x } } } ( \tau)\log \phi_{g , { { x } } } ( \tau ) d\tau } { \sum_{g ' , { { x } } ' } \pi(g ' ) t_{g'}^ { ( { { x } } ' ) } /\mu_{g ' , { { x } } ' } } ~. \label{eq : hmu_hsmm}\end{aligned}\ ] ] to ground intuition , recall that each state in the underlying for semi - markov processes corresponds to a unique observation symbol .
hence , setting @xmath304 to @xmath305 and noting that each @xmath253 is uniquely associated to some @xmath1 in eq .
( [ eq : hmu_hsmm ] ) recovers the results of ref .
@xcite for the entropy rate of semi - markov processes , though the notation differs somewhat . ) to a semi - markov process in reverse - time so that the underlying model is unifilar rather than co - unifilar ; but entropy rate is invariant to time reversal @xcite . ] the process excess entropy @xmath43 $ ] can be calculated if we can find the joint probability distribution @xmath306 of forward- and reverse - time causal states .
to this end , we add an additional restriction on the generative model : we focus only on generative models for which @xmath307 . with this restriction on labeled transition matrices ,
the time - reversed of the process has the same form as the of the forward - time process , but with a different @xmath249 .
the latter is related to the forward - time @xmath249 via manipulations described in ref . @xcite . as such
, we can write down @xmath308 : @xmath309 where we obtain @xmath310 from standard methods @xcite applied to ( only ) the dynamic on @xmath249 . note that @xmath311 reduces to @xmath312 as @xmath254 and @xmath313 uniquely
specify the distribution from which @xmath314 is drawn and since @xmath315 .
we leave the the final steps to @xmath302 as an exercise .
though the definition of continuous - time causal states parallels that for discrete - time causal states , continuous - time and information measures are markedly different from their discrete - time counterparts .
similar technical difficulties arise more generally when describing minimal maximally predictive models of other continuous - time , discrete - symbol processes that are not the continuous - time markov processes analyzed in ref .
the resulting do not appear like conventional hmms recall figs .
[ fig : nedp]-[fig : edp ] and , especially , fig .
[ fig : uhsmm](top)and most of the information measures excepting the excess entropy are reinterpreted as differential information rates
. moreover , the continuous - time machinery gave us a new way to calculate these information measures .
traditionally , expressions for such information measures come from calculating the time - normalized path entropy of arbitrarily long trajectories ; e.g. , as in ref .
instead , we calculated the path entropy of arbitrarily short trajectories , conditioned on the past .
this allowed us to extend the results of ref .
@xcite for the entropy rate of continuous - time discrete - output processes to a previously untouched class of processes unifilar hidden semi - markov processes .
there are two immediate practical benefits to an in - depth look at the of continuous - time hidden semi - markov processes .
first , statistical model selection when searching through unifilar hidden markov models is significantly easier than when searching through nonunifilar hidden markov models @xcite , and these benefits should carry over to the case of continuous - time .
second , the formulae in table [ tab:1 ] and those in sec .
[ sec : uhsmms ] provide new approaches to binless plug - in information measure estimation ; e.g. , following ref .
@xcite .
the machinery required to use continuous - time is significantly different than that accompanying the study of discrete - time .
our results here pave the way toward understanding the difficulties that lie ahead when studying the structure and information in continuous - time processes .
the authors thank the santa fe institute for its hospitality during visits .
jpc is an sfi external faculty member .
this material is based upon work supported by , or in part by , the u. s. army research laboratory and the u. s. army research office under contract number w911nf-13 - 1 - 0390 .
sm was funded by a national science foundation graduate student research fellowship , a u.c .
berkeley chancellor s fellowship , and the mit physics of living systems fellowship .
v. girardin . on the different extensions of the ergodic theorem of information theory . in r.
baeza - yates , j. glaz , h. gzyl , j. husler , and j. l. palacios , editors , _ recent advances in applied probability theory _ , pages 163179 .
springer us , 2005 .
actually , we would apply eq .
( [ eq : hmu_hsmm ] ) to a semi - markov process in reverse - time so that the underlying model is unifilar rather than co - unifilar ; but entropy rate is invariant to time reversal @xcite . | we introduce the minimal maximally predictive models ( ) of processes generated by certain hidden semi - markov models .
their causal states are either hybrid discrete - continuous or continuous random variables and causal - state transitions are described by partial differential equations .
closed - form expressions are given for statistical complexities , excess entropies , and differential information anatomy rates .
we present a complete analysis of the of continuous - time renewal processes and , then , extend this to processes generated by unifilar hidden semi - markov models and semi - markov models .
our information - theoretic analysis leads to new expressions for the entropy rate and the rates of related information measures for these very general continuous - time process classes . |
recent surveys have found a significant population of low - mass stars , brown dwarfs and planetary - mass objects in young open clusters . since low - mass objects evolve little over the lifetime of the universe , the present day mass function of these objects is a good representation of the initial mass function ( imf ) .
the mass function in this low - mass regime is however poorly known due to faintness of these objects and also due to uncertainty in the mass - luminosity relations .
low - mass objects at or below the hydrogen burning mass limit ( hbml ) of 0.08 m@xmath2 are known to be warmer and hence more luminous when young although they cool rapidly and fade with age ( baraffe et al . 1998 ) .
the combination of youth and proximity in some open clusters make them ideal targets for searches of low - mass objects below the hbml particularly at infrared wavelengths . in the present study
we have adopted a statistical approach to determine the mass function ( dn / dm @xmath4 m@xmath5 ) of objects in the mass range @xmath6 to 0.0250.05 m@xmath2 using data of three open clusters namely ic 348 , @xmath0 orionis and pleiades .
we have used the data from two extended sky surveys 2mass second incremental release and the latest version of gsc on the three open clusters .
the limiting magnitudes of 16.5 , 15.5 , 15 and 21 in the @xmath7 , @xmath8 , @xmath9 and @xmath10 ( poss ii iiia - f ) passbands , respectively , enable us to probe down to 0.025 m@xmath2 in ic 348 and @xmath0 orionis and 0.04m@xmath2 in pleiades . unlike most other previous studies which rely on confirming candidate low - mass objects through spectroscopy we use a statistical approach to estimate the number of low - mass objects . in this approach
it is important to use several control fields close to each cluster to subtract the contribution of foreground and background objects .
the nature of these two extended surveys permitted us to use several control fields and variable field sizes .
table 1 lists the positions and radii of the fields chosen for the three clusters .
lllcl + fields&ra ( j2000)&dec ( j2000)&radius + & ( h m s ) & ( @xmath11 @xmath12 @xmath13)&(arcmin ) + ic 348&03 44 30&+32 17 00&20 + control 1&03 49 08&+31 19 08&20 + control 2&03 44 10&+33 19 26&20 + & + @xmath0 orionis&05 38 45&02 36 00&30 + control 1&05 58 29&04 29 48&30 + control 2&0.5 11 00&00 20 00&30 + & + pleiades&03 47 00&+24 07 00&90 + control 1&03 18 00&+26 41 00&90 + control 2&03 05 00&+24 42 00&90 + in the first stage of analysis we merged the 2mass and the gsc sources by taking 2mass coordinates and cross correlating them with the gsc catalogue .
a search radius of 2was found to be the optimum value for the cross correlation being small enough to reject spurious and multiple detections and large enough to include any minor positional uncertainties in the two catalogues . selection criteria to pick out cluster members were derived by inspecting various colour magnitude diagrams ( cmds ) .
the use of the f versus ( f r ) cmd overcomes the degeneracies between the mass and distance found in the k@xmath14 versus ( j k@xmath14 ) cmd , and minimizes the overlap between reddened background stars and low - mass members of the cluster .
this leads to a more efficient rejection criteria for non members which is accomplished using the empirical data from leggett ( 1992 ) and the theoretical isochrones of baraffe et al .
the selection criteria for low - mass members , derived using the suitably distance - scaled and extinction corrected model isochrones of baraffe et al .
( 1998 ) , are listed in table 2 .
the sources satisfying the criteria are further statistically filtered for possible contaminants using the control fields .
details of the procedure are discussed in tej et al .
finally , the masses of the selected candidates are determined by comparing observed magnitudes with those derived from evolutionary models ( baraffe et al .
1998 and chabrier et al .
2000 ) .
llll + ic 348&@xmath0 orionis&pleiades + @xmath15&@xmath16&@xmath17 + + 10.14 ] & + 10.26]&@xmath18 + & & @xmath19 $ ] + @xmath20&@xmath21&@xmath22 + @xmath23 & @xmath24 & @xmath25 + @xmath26 0.90&@xmath26 0.94&@xmath26 0.83 + @xmath27 0.20&@xmath27 0.20&@xmath27 0.22 + = 10truecm
we first carried out a detailed study for ic 348 . for this cluster , we derived the mass function using the solar metallicity models of baraffe et al ( 1998 ) and dusty models of chabrier et al .
( 2000 ) both of which gave similar results .
the location of low - mass members isolated by us and the spectroscopically confirmed low - mass members of luhman ( 1999 ) are in good agreement which builds the confidence in our selection criteria .
we then used the same methodology for @xmath0 orionis and pleiades .
the resulting exponents of the mass functions for ic 348 , @xmath0 orionis and pleiades are respectively 0.7 , 1.2 and 0.5 with an estimated error of @xmath28 0.2 . for the younger clusters ic 348 and @xmath0 orionis , there is some hint of a possible dip in the mass function at the position of the hbml . the results of this statistical approach ( table 3 )
imply that though the mass function continues to rise above the hbml it is appreciably flatter compared to the salpeter mass function . in figure 1 we have plotted the derived mass function for the three clusters . taking the canonical salpeter exponent of 2.35 in the mass range 110 m@xmath2 , the chabrier exponent of 1.55 in the mass range 0.51.0 m@xmath2 and the values obtained by us below 0.5 m@xmath2 we estimate the mass contribution to be about 40% for objects below 0.5 m@xmath2 and 4% for objects below the hbml of 0.08 m@xmath2 .
our results are consistent with previous studies ( e.g. bjar et al . , 2001 ) and suggest that although low - mass objects are at least as numerous as their stellar counterparts their contribution to the total mass of the cluster is small .
we would like to thank brian mclean and mario lattanzi ( gsc - ii project scientists ) for access to the development version of gsc - ii in advance of publication .
we are grateful to i. baraffe and f. allard for making electronic versions of the latest models available and generating model isochrones for the nonstandard f passbands .
baraffe , i. , chabrier , g. , allard , f . , & hauschildt , p.h .
1998 , a&a , 337 , 403 bjar , v.j.s . ,
martin , e.l . ,
zapatero osorio , m.r .
, et al . 2001 , apj , 556 , 830 chabrier , g. , baraffee , i. , allard , f. , & hauschildt , p.h .
2000 , apj , 542 , 464 leggett , s.k .
1992 , apjs , 82 , 351 luhman , k. l. 1999 , apj , 525 , 466 tej a. , sahu , k.c .
, chandrasekhar , t. and ashok , n.m .
2002 apj ( in press ) ( astro - ph/0206325 ) | we present a statistical method to derive the mass functions of open clusters using sky survey data such as the 2 micron all sky survey ( 2mass ) and the guide star catalogue ( gsc ) .
we have used this method to derive the mass functions in the stellar / substellar regime of three young , nearby open clusters , namely ic 348 , @xmath0 orionis and pleiades .
the mass function in the low mass range ( m@xmath1 ) is appreciably flatter than the stellar salpeter function for all three open clusters .
the contribution of objects below 0.5 m@xmath2 to the total mass of the cluster is @xmath340% and the contribution of objects below 0.08 m@xmath2 to the total is @xmath34% .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
in an earlier paper @xcite , hereafter referred to as i , we have explored the lhc implications of imposing @xmath0 quasi - yukawa unification ( qyu ) at the grand unification scale ( @xmath12 gev ) .
this modified approach to the third family ( @xmath0 ) yu @xcite can be motivated by the desire to construct realistic supersymmetric models of grand unified theories ( guts ) which also incorporate realistic masses and mixings observed in the matter sector .
for instance , the desired quarks and charged lepton masses for the second family fermion can be incorporated , following @xcite , by including higgs fields in the ( 15,1,3 ) representation of @xmath1 ( 4 - 2 - 2 ) @xcite , which develop a non - zero gut scale vacuum expectation values ( vevs ) .
the third family yukawa couplings receive , in this case , sizable new contributions , and the deviations from exact yu can be stated as follows @xcite : @xmath13 where @xmath2 measures the deviation from the exact yu . restricting the deviation to @xmath14
we refer eq.([qyu ] ) to qyu condition .
the 4 - 2 - 2 model has many salient features distinguishing it from other high scale theories @xcite .
the discrete left - right ( lr ) symmetry reduces the number of gauge couplings from three to two with @xmath15 .
it also requires gaugino masses of @xmath16 and @xmath17 to be equal at @xmath4 .
the matter fields of each family belong to @xmath18 and @xmath19 .
the lr symmetry requires the existence of right handed neutrino . in this paper
we reconsider qyu in the framework of 4 - 2 - 2 defined above , taking into account the fact that the mssm gaugino masses @xmath20 at @xmath4 can be non - universal .
in particular , we assume the following asymptotic relation @xcite @xmath21 which follows from the assumption of left - right symmetry at @xmath4 and the fact that @xmath22 derived from 4 - 2 - 2 is given as follows ; @xmath23 here @xmath24 , @xmath25 and @xmath26 are the asymptotic gaugino masses for @xmath22 , @xmath16 and @xmath27 , and @xmath28 and @xmath29 are diagonal generators of @xmath17 and @xmath30 respectively .
the setup of our paper is as follows . in section [ sec : scan ] we briefly explain our scanning procedure and list the experimental constraints that we impose on the data obtained from our scans . in section [ sec : spec ]
we show the fundamental parameter space that is allowed by the experimental constraints and qyu .
section [ sec : coan ] provides the implications for dark matter sector such as coannihilation channels and the resonance solution .
section [ sec : higgsinodm ] considers a higgsino - like lsp and emphasize the implications for direct detection experiments .
section [ sec : isasoft ] compares the isajet and softsusy , and the small variations in their phenomenology .
we also present the benchmark points obtained from the different scans to exemplify our results .
finally , we conclude our study in section [ sec : conclusion ] .
in our scan , we employ isajet 7.84 @xcite , softsusy-3.4.1 @xcite and superiso relic v3.3 @xcite to calculate the low scale observables .
the gauge and yukawa couplings are first estimated at the low scale .
isajet evolves the gauge couplings and the yukawa couplings of the third family up to @xmath4 , while softsusy performs the calculations in the three family approximation in evolution of yukawa couplings .
we do not strictly enforce the gauge unification condition @xmath31 , since a few percent deviation from the gauge coupling unification can be generated by unknown gut - scale threshold corrections @xcite .
hence , @xmath4 is calculated to be the scale where @xmath32 and @xmath33 deviates a few percent .
after @xmath4 is determined , the soft supersymmetry breaking ( ssb ) parameters determined with the boundary conditions defined at @xmath4 are evolved together with the gauge and yukawa couplings from @xmath4 to the weak scale @xmath34 .
the susy threshold corrections to the yukawa couplings @xcite are taken into account at the common scale @xmath35 in isajet , while softsusy evaluates them at the electroweak scale .
the entire parameter set is iteratively run between @xmath36 and @xmath4 using full 2-loop rges , and the ssb parameters are extracted from rges at the appropriate scales @xmath37 .
we have performed random scans over the following parameter space : @xmath38 taking eq.[gaugino422 ] into account .
we use the @xmath39 notation in which @xmath40 is the universal ssb mass term for the particles , and @xmath24 , @xmath25 , @xmath26 are ssb mass terms for the gauginos of @xmath22 , @xmath16 and @xmath27 respectively , @xmath41 is the universal ssb term for trilinear scalar interactions , @xmath42 is the ratio of vevs of the mssm higgs doublets , @xmath43 is coefficient of the bilinear higgs mixing term , and @xmath44 is the top quark mass . note that we set the top quark mass to 173.3 gev @xcite , and our results are not too sensitive to a @xmath45 variation in @xmath44 @xcite .
we employ the metropolis - hasting algorithm as described in @xcite , and require all collected points to satisfy radiative electroweak symmetry breaking ( rewsb ) with lsp neutralino .
the rewsb gives a crucial theoretical constraint on the parameter space @xcite .
after collecting data , we impose constraints from the mass bounds @xcite , rare decays of b - meson such as @xmath46 @xcite , @xmath47 @xcite , and @xmath48 @xcite .
after obtaining the region allowed by the lhc constraints , we also apply the wmap bound @xcite on the relic abundance of lsp neutralino .
isajet interfaces with isatools @xcite for b - physics and relic density observables , while we interface softsusy with superiso relic in order to calculate these observables .
the experimental constraints imposed in our data can be summarized as follows : @xmath49 we emphasize here the mass bounds on the higgs boson @xcite and gluino @xcite .
we allow a few percent deviation from the observed mass of the higgs boson , since there exist about a 2 gev error in estimation of its mass arising due to theoretical uncertainties in the calculation of the minimum of the scalar potential , and experimental uncertainties in @xmath44 and @xmath50 @xcite .
besides these constraints , we require our solutions to do no worse than the sm in comparing predictions for the muon anomalous magnetic moment . note that we relax the wmap bound to @xmath51 on the solutions obtained from superiso relic in order to take into account uncertainties in the calculation of the relic abundance of lsp neutralino .
@xmath52{422_nuhm1_quasi_c_m0.png } & \includegraphics[scale=1]{422_nuhm1_quasi_soft_c_m0.png } \\
\includegraphics[scale=1]{422_nuhm1_quasi_c_m10.png } & \includegraphics[scale=1]{422_nuhm1_quasi_soft_c_m10.png } \end{array}\ ] ] @xmath53{422_nuhm1_quasi_c_m2.png } & \includegraphics[scale=1.2]{422_nuhm1_quasi_soft_c_m2.png } \\
\includegraphics[scale=1.2]{422_nuhm1_quasi_c_m3.png } & \includegraphics[scale=1.2]{422_nuhm1_quasi_soft_c_m3.png } \end{array}\ ] ] @xmath54{422_nuhm1_quasi_c_tanb.png } & \includegraphics[scale=1.2]{422_nuhm1_quasi_soft_c_tanb.png } \\
\includegraphics[scale=1.2]{422_nuhm1_quasi_c_a0m0.png } & \includegraphics[scale=1.2]{422_nuhm1_quasi_soft_c_a0m0.png } \end{array } \label{c_ta}\ ] ] @xmath54{422_nuhm1_quasi_ul_glu.png } & \includegraphics[scale=1.2]{422_nuhm1_quasi_soft_ul_glu.png } \\
\includegraphics[scale=1.2]{422_nuhm1_quasi_stop_neut.png } & \includegraphics[scale=1.2]{422_nuhm1_quasi_soft_stop_neut.png } \end{array}\ ] ] in this section , we highlight the allowed regions in the fundamental parameter space of 4 - 2 - 2 given in eq.([paramspace ] ) and present the results for the supersymmetric particles and higgs boson mass spectrum that we obtain from the scans using isajet and softsusy which is interfaced with superiso relic .
fig.[c_mass ] shows the plots in @xmath55 and @xmath56 planes .
the left panel shows the results obtained by using isajet , while the right panels displays the results from softsusy .
all points are compatible with rewsb and lsp neutralino .
green points satisfy the mass bounds on the sparticles and the constraints from rare b - decays .
blue points form a subset of green and they are compatible with the qyu condition .
points in red are a subset of blue , and satisfy the constraint on relic abundance of lsp neutralino .
they are consistent with the wmap bound within @xmath57 uncertainty in isajet plots , while @xmath58 for those obtained from softsusy and superiso relic .
as seen from the @xmath55 panels , qyu ( blue ) requires @xmath59 tev , while @xmath60 is only loosely constrained .
similarly fig.[c_mass2 ] displays the plots in @xmath61 and @xmath62 planes .
color coding is the same as in fig.[c_mass ] .
we can see from the @xmath61 panel that @xmath25 can be as low as 300 gev .
such light @xmath25 solutions yield bino - wino mixing at the low scale which plays a role in reducing the relic abundance of lsp neutralino to the desired range .
the @xmath62 plane shows that @xmath63 gev is compatible with qyu which leads to a heavy gluino ( @xmath64 tev ) at low scale .
the results for the remaining parameters are shown in fig.[c_ta ] with plots in @xmath65 and @xmath66 planes .
color coding is the same as in fig.[c_mass ] .
the plots in the @xmath65 planes shows that qyu requires rather high @xmath67 values .
the top left panel shows that @xmath68 is compatible with qyu , while it is possible to find qyu solutions with softsusy for @xmath69 .
the @xmath66 panels from both isajet and softsusy show that @xmath70 can lie in the range ( @xmath71 ) .
we present the results for the mass spectrum of the colored particles in fig.[c_spec ] in @xmath72 and @xmath73 planes .
color coding is the same as in fig.[c_mass ] .
in addition , the blue points satisfy @xmath74 as well as qyu condition .
the gluino mass compatible with qyu and @xmath74 is found to be @xmath64 tev as stated above , and it can be tested in future experiments at the large hadron collider ( lhc ) .
similarly , the stop quarks satisfy @xmath75 tev .
isajet and softsusy are in good agreement regarding results of the mass spectrum .
in the previous section we have focused on the fundamental parameter space and the mass spectrum of the colored particles .
since we accept only those solutions which lead to lsp neutralino , it is worth investigating the implications of 4 - 2 - 2 on the dark matter observables . indeed ,
if the lsp neutralino is mostly a bino , its relic abundance is usually so high that it can not be consistent with the wmap observation .
however , one can identify various coannihilation channels that reduce the relic abundance of neutralino to the desired ranges .
4 - 2 - 2 has some rich phenomenological implications and allows various coannihilation channel scenarios at the low scale @xcite , since it allows asymptotically different masses for the gauginos as given in eq.([gaugino422 ] ) . on the other hand ,
if one imposes @xmath0 yu at @xmath4 with @xmath76 , only the gluino - neutralino coannihilation channel can survive @xcite . relaxing this to @xmath77
yu opens up , in addition , the stop - neutralino channel @xcite . in this section ,
we consider the phenomenological implications of qyu in 4 - 2 - 2 regarding the dark matter and the structure of lsp neutralino . besides the bino - like lsp neutralino , it is possible to find solutions with bino - wino mixture , bino - higgsino mixture , or mostly higgsino lsp neutralino which leads to different phenomenology .
fig.[dm ] displays the results in @xmath78 and @xmath79 planes .
color coding is the same as in fig.[c_spec ] .
as stated in the previous section , @xmath25 can be as low as 300 gev .
the line in @xmath78 plane indicates solutions for which @xmath80 and they yield bino - wino mixing at the low scale .
similarly , the line in @xmath81 plane corresponds to the solutions which have @xmath82 .
these solutions can lead to very interesting implications , since lsp neutralino is bino - higgsino mixture near this line .
moreover , the lsp neutralino is found to be mostly higgsino below the line .
@xmath54{422_nuhm1_quasi_m1_m2.png } & \includegraphics[scale=1.2]{422_nuhm1_quasi_soft_m1_m2.png } \\
\includegraphics[scale=1.2]{422_nuhm1_quasi_m1_mu.png } & \includegraphics[scale=1.2]{422_nuhm1_quasi_soft_m1_mu.png } \end{array}\ ] ] @xmath83{422_nuhm1_quasi_char1_neut.png } & \includegraphics[scale=0.9]{422_nuhm1_quasi_soft_char1_neut.png } \\
\includegraphics[scale=0.9]{422_nuhm1_quasi_stau_neut.png } & \includegraphics[scale=0.9]{422_nuhm1_quasi_soft_stau_neut.png } \\
\includegraphics[scale=0.9]{422_nuhm1_quasi_ma_neut.png } & \includegraphics[scale=0.9]{422_nuhm1_quasi_soft_ma_neut.png } \end{array}\ ] ] fig.[coan ] summarizes our results for the coannihilation channels compatible with qyu in @xmath84 , @xmath85 , and @xmath86 planes .
color coding is the same as in fig.[c_spec ] .
the solid lines in the plots correspond to the related coannihilation channel regions .
the isajet panel of @xmath84 shows that the neutralino and the lightest chargino of mass @xmath87 gev can be nearly degenerate as expected from the @xmath78 planes of fig.[dm ] .
we can find solutions with chargino - neutralino coannihilation channel for @xmath88 gev , if we relax the lsp neutralino relic density to @xmath58 , as seen in the softsusy and superiso relic panel .
besides chargino - neutralino coannihilation , the stau - neutralino channel is found to be compatible with qyu as seen from the @xmath85 plane .
there are plenty of solutions for @xmath89 gev in the isajet panel , while softsusy yields fewer solutions with light staus .
another solution allowed by qyu is the @xmath7resonance shown in the @xmath86 planes .
the solid line in these panels corresponds to @xmath90 in which two lsp neutralinos annihilate via the @xmath7boson .
the @xmath7resonance solutions can be found for @xmath91 gev in the data set obtained from isajet , while it can be realized for @xmath91 if one applies @xmath58 to the data set obtained with softsusy and superiso .
in the previous section we have identified various coannihilation channels and a resonance solution which reduce the relic abundance of lsp neutralino so that the dark matter phenomenology in the 4 - 2 - 2 framework can be consistent with the wmap experiment . in this section
we briefly explore an alternative scenario in which the lsp neutralino is a gaugino - higgsino .
this case opens up possibilities for direct detection experiments via relic lsp neutralino scattering on nuclei .
the case with bino - wino mixture , or equivalently those with chargino - neutralino coannihilation , yield moderate cross - sections in these scattering processes , since the lsp interacts with quarks in the nucleon also via @xmath92 interactions .
the scattering cross - section reaches its highest values when the lsp neutralino is a bino - higgsino mixture or mostly higgsino , since the yukawa interactions between quarks and the higgsino component of lsp neutralino take part in the scattering processes .
we present our results in fig.[directdetection ] for neutralino - nucleon scattering for both spin - independent and spin - dependent cases in @xmath93 and @xmath94 planes .
color coding is the same as in fig.[c_spec ] . in the @xmath93 plane
, the dashed ( solid ) red line represents the current ( future ) bound of the xenon1 t experiment , while the dashed ( solid ) black line shows the current ( future ) bound of the cdms experiment . in the @xmath94 plane ,
the dashed red line represents the bound from the super k experiment , while the dashed ( solid ) black line shows the current ( future ) reach of the icecube experiment .
only the results from isajet are shown in the panels . as seen from the @xmath93 and @xmath94 plane ,
the spin - independent cross - section for the lsp neutralino with bino - wino mixture is of order @xmath95 pb , while it rises by two orders of magnitude for bino - higgsino mixture .
furthermore , the spin - independent cross - section lies between @xmath96 pb if the lsp neutralino is mostly a higgsino within reach of the direct detection experiments such as xenon1 t and supercdms .
finally , we also check that qyu with higgsino(-like ) dark matter and mass @xmath97 tev is also realized in 4 - 2 - 2 and @xmath39 supersymmetric models with universal gaugino masses ( namely cmssm boundary conditions ) at @xmath4 .
@xcite gives a detailed analysis and comparison among several numerical codes including isajet and softsusy .
in sec.[sec : spec ] we show that qyu prefers regions with larger @xmath67 , and a @xmath98 difference between the yukawa couplings obtained from softsusy and isajet @xcite can lead to some quantitative differences in the results . in this section , we present two tables of benchmark points that exemplify the results obtained from our scans .
table [ isajet ] presents four benchmark points obtained from the isajet scan .
the points are chosen to be consistent with the constraints mentioned in sec.[sec : scan ] .
point 1 is an @xmath7resonance solution , and point 2 depicts a solution with higgsino dark matter .
points 3 and 4 display stau - neutralino and chargino - neutralino coannihilation solutions respectively .
point 4 also exemplifies the solution with the heaviest cp - even higgs boson mass we obtained .
similarly table [ softsusy ] displays four benchmark points consistent with the experimental constraints obtained from softsusy and superiso relic scan .
the points are chosen to be consistent with the constraints mentioned in sec.[sec : scan ] .
point 1 displays an @xmath7resonance solution .
points 2 and 3 depict solutions with the higgsino dark matter , while wmap bound on relic abundance of lsp neutralino is satisfied through the stau - neutralino coannihilation for point 2 , and chargino - neutralino coannihilation for point 3 .
point 4 also shows a solution with chargino - neutralino coannihilation and exemplifies the solution with the heaviest cp - even higgs boson mass obtained .
tables [ isajet ] and [ softsusy ] also summarize the differences between isajet and softsusy in the fundamental parameters that yield similar implications .
the gaugino masses differ a few hundred gev in the case of stau - neutralino coannihilation , while it rises up to about 1 tev in scalar masses @xmath40 and @xmath60 .
the largest difference can be realized at the point 4 s which displays the largest sm - like higgs boson mass from both scan .
@xmath40 is about 4 tev heavier in softsusy result , while @xmath60 is about 2 tev lighter . on the other hand , the phenomenological results obtained from the scans yield very similar results .
the fundamental parameter space allowed by the experimental results and qyu are quite similar , and the same coannihilation channels are identified from both scans .
we have employed isajet and softsusy interfaced with superiso relic to explore the lhc implications of quasi - yukawa unified ( qyu ) supersymmetric models based on @xmath1 . in these qyu models ,
the third family yukawa unification relations involving @xmath99 , @xmath100 and @xmath101 , is quantified by a parameter @xmath2 which takes values @xmath102 .
in contrast to earlier studies , the mssm gaugino masses at @xmath4 are non - universal but consistent with the gauge symmetry @xmath6 . the thermal relic abundance of the lsp neutralino is compatible with the wmap bounds through the chargino and stau coannihilation channels , as well as the a - resonance solution .
we also identify solutions with higgsino - like and pure higgsino dark matter ( @xmath103 tev ) which can be tested in the direct dark matter searches such as xenon1 t and supercdms . the predicted gluino mass ranges from 1 - 4 tev , while the stop masses are heavier than 2 tev or so .
it is reassuring to note that the low energy phenomenology obtained from isajet and softsusy are in good qualitative agreement .
we thank ilia gogoladze , m. adeel ajaib and dimitri i. kazakov for useful discussions .
ht would like to thank michael afanasev for his advice in programming .
this work is supported in part by doe grants de - fg02 - 91er40626 ( qs ) and russian foundation for basic research ( rfbr ) grant no .
14 - 02 - 00494-a ( ht ) .
this work used the extreme science and engineering discovery environment ( xsede ) , which is supported by the national science foundation grant number oci-1053575 .
part of the numerical calculations reported in this paper were performed at the national academic network and information center ( ulakbim ) of turkey scientific and technological research institution ( tubitak ) , high performance and grid computing center ( truba resources ) .
s. dar , i. gogoladze , q. shafi and c. s. un , phys .
d * 84 * , 085015 ( 2011 ) [ arxiv:1105.5122 [ hep - ph ] ] . b. ananthanarayan , g. lazarides and q. shafi , phys . rev .
d * 44 * , 1613 ( 1991 ; phys . lett .
b * 300 * , 24 ( 1993)5 ; q. shafi and b. ananthanarayan , trieste hep cosmol.1991:233 - 244 . m. e. gomez , g. lazarides and c. pallis , nucl .
b * 638 * , 165 ( 2002 ) [ hep - ph/0203131 ] ; m. e. gomez , g. lazarides and c. pallis , phys .
d * 67 * , 097701 ( 2003 ) [ hep - ph/0301064 ] ; c. pallis and m. e. gomez , hep - ph/0303098 ; j. c. pati and a. salam , phys . rev .
d * 10 * , 275 ( 1974 ) . g. lazarides and q. shafi , nucl .
b * 189 * , 393 ( 1981 ) ; s. f. king and q. shafi , phys .
b * 422 * , 135 ( 1998 ) [ hep - ph/9711288 ] ; g. lazarides , m. magg and q. shafi , phys .
b * 97 * , 87 ( 1980 ) ; t. w. kephart , c. a. lee and q. shafi , jhep * 0701 * , 088 ( 2007 ) [ hep - ph/0602055 ] ; t. w. kephart and q. shafi , phys .
b * 520 * , 313 ( 2001 ) [ hep - ph/0105237 ] ; r. n. mohapatra and j. c. pati , phys . rev .
d * 11 * , 566 ( 1975 ) ; g. senjanovic and r. n. mohapatra , phys .
d * 12 * , 1502 ( 1975 ) ; g. senjanovic , nucl .
b * 153 * , 334 ( 1979 ) ; m. magg , q. shafi and c. wetterich , phys .
b * 87 * , 227 ( 1979 ) ; m. cvetic , nucl .
b * 233 * , 387 ( 1984 ) ; t. w. b. kibble , g. lazarides and q. shafi , phys .
b * 113 * , 237 ( 1982 ) ; phys .
d 26 , 435 r. n. mohapatra and b. sakita , phys .
d * 21 * , 1062 ( 1980 ) .
i. gogoladze , r. khalid and q. shafi , phys .
d * 79 * , 115004 ( 2009 ) [ arxiv:0903.5204 [ hep - ph ] ] .
f. e. paige , s. d. protopopescu , h. baer and x. tata , hep - ph/0312045 ; for updates and changes in the current version , see http://www.nhn.ou.edu/~isajet/[isajet 7.84 ] b. c. allanach , comput .
phys . commun .
* 143 * , 305 ( 2002 ) [ hep - ph/0104145 ] .
f. mahmoudi , comput .
commun . * 180 * , 1579 ( 2009 ) [ arxiv:0808.3144 [ hep - ph ] ] ; f. mahmoudi , comput .
commun .
* 180 * , 1718 ( 2009 ) ; a. arbey and f. mahmoudi , comput .
commun . *
181 * , 1277 ( 2010 ) [ arxiv:0906.0369 [ hep - ph ] ] .
j. hisano , h. murayama , and t. yanagida , nucl . phys .
* b402 * ( 1993 ) 46 ; y. yamada , z. phys . *
c60 * ( 1993 ) 83 ; j. l. chkareuli and i. g. gogoladze , phys . rev .
d * 58 * , 055011 ( 1998 ) .
d. m. pierce , j. a. bagger , k. t. matchev and r. j. zhang , nucl .
b * 491 * , 3 ( 1997 ) [ hep - ph/9606211 ] .
[ atlas and cdf and cms and d0 collaborations ] , arxiv:1403.4427 [ hep - ex ] .
i. gogoladze , r. khalid , s. raza and q. shafi , arxiv:1402.2924 [ hep - ph ] .
i. gogoladze , r. khalid , s. raza and q. shafi , jhep * 1106 * ( 2011 ) 117 .
g. belanger , f. boudjema , a. pukhov and r. k. singh , jhep * 0911 * , 026 ( 2009 ) ; h. baer , s. kraml , s. sekmen and h. summy , jhep * 0803 * , 056 ( 2008 ) .
l. e. ibanez and g. g. ross , phys .
* b110 * ( 1982 ) 215 ; k. inoue , a. kakuto , h. komatsu and s. takeshita , prog .
phys . * 68 * , 927 ( 1982 ) [ erratum - ibid . *
70 * , 330 ( 1983 ) ] ; l. e. ibanez , phys . lett .
* b118 * ( 1982 ) 73 ; j. r. ellis , d. v. nanopoulos , and k. tamvakis , phys . lett . * b121 * ( 1983 ) 123 ; l. alvarez - gaume , j. polchinski , and m. b. wise , nucl . phys .
* b221 * ( 1983 ) 495 .
k. nakamura _ et al .
_ [ particle data group ] , j. phys .
g * 37 * , 075021 ( 2010 ) .
[ lhcb collaboration ] , phys .
lett . * 110 * , 021801 ( 2013 ) .
y. amhis _ et al . _
[ heavy flavor averaging group collaboration ] , arxiv:1207.1158 [ hep - ex ] .
d. asner _
[ heavy flavor averaging group collaboration ] , arxiv:1010.1589 [ hep - ex ] .
g. hinshaw _ et al . _
[ wmap collaboration ] , arxiv:1212.5226 [ astro-ph.co ] .
h. baer and m. brhlik , _ phys .
_ * d 55 * ( 1997 ) 4463 ; h. baer , m. brhlik , d. castano and x. tata , _ phys .
_ * d 58 * ( 1998 ) 015007 ; k. babu and c. kolda , _ phys .
* 84 * ( 2000 ) 228 ; a. dedes , h. dreiner and u. nierste , _ phys .
* 87 * ( 2001 ) 251804 ; j. k. mizukoshi , x. tata and y. wang , _ phys
. rev . _ * d 66 * ( 2002 ) 115003 .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 716 * , 1 ( 2012 ) [ arxiv:1207.7214 [ hep - ex ] ] . s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * , 30 ( 2012 ) [ arxiv:1207.7235 [ hep - ex ] ] .
the atlas collaboration , atlas - conf-2013 - 047 .
g. degrassi , s. heinemeyer , w. hollik , p. slavich and g. weiglein , eur .
j. c * 28 * , 133 ( 2003 ) [ hep - ph/0212020 ] ; t. hahn , s. heinemeyer , w. hollik , h. rzehak and g. weiglein , phys .
* 112 * , 141801 ( 2014 ) [ arxiv:1312.4937 [ hep - ph ] ] and references therein
. i. gogoladze , r. khalid and q. shafi , phys .
d * 80 * , 095016 ( 2009 ) [ arxiv:0908.0731 [ hep - ph ] ] . | we explore the dark matter and lhc implications of @xmath0 quasi yukawa unification in the framework of supersymmetric models based on the gauge symmetry @xmath1 . the deviation from exact yukawa unification is quantified by a dimensionless parameter @xmath2 ( @xmath3 ) , such that the yukawa couplings at @xmath4 are related by @xmath5 .
in contrast to earlier studies which focused on universal gaugino masses , we consider non - universal gaugino masses at @xmath4 that are compatible with the gauge symmetry @xmath6 .
we perform two independent scans of the fundamental parameter space , one of which employs isajet , while the other uses softsusy interfaced with superiso .
these scans reveal qualitatively similar allowed regions in the parameter space , and yield a variety of neutralino dark matter scenarios consistent with the observations .
these include stau and chargino coannihilation scenarios , the @xmath7resonance scenario , as well as higgsino dark matter solution which is more readily probed by direct detection searches .
the gluino mass is found to be @xmath8 tev , the stop mass is @xmath9 tev , while the first two family squarks and sleptons are of order @xmath10 tev and @xmath11 tev respectively . |
the broadband spectrum observed in blazars is dominated by beamed emission produced in relativistic jets that emanate from the central black hole .
these jets propagate through a dense radiative environment , and interact with seed photons that are supplied by extended radiation sources , notably the accretion disk around the black hole , gaseous clouds in the broad line region ( blr ) , and a dusty molecular torus located at larger scales ( see , e.g. , joshi et al .
2014 for a recent account ) .
this interaction should affect the dynamics and emission of the jet .
the primary concern of most previous works ( e.g. , dermer & schlickeiser 1993 , sikora et al .
1994 , blandford & levinson 1995 , ghisellini & madau 1996 ; see also levinson 2006 and references therein ) has been the effect of this interaction on the observed spectrum , and simple emission models that ignore dynamical effects and complex structures have been constructed for this purpose , although some recent works incorporate more realistic models for the dynamics of the emitting plasma ( e.g. , joshi & bottcher 2011 , joshi et al .
2014 , and references therein ) . in erc models ,
the high - energy component of the spectral energy distribution ( sed ) is attributed to inverse compton scattering of ambient seed photons by non - thermal electrons accelerated in dissipative regions inside the jet .
this process , unlike synchrotron emission , gives rise to a loss of linear momentum of emitting fluid elements and a consequent radiative drag that tends to decelerate the bulk flow .
in certain circumstances , explored below , this can lead to rapid , large amplitude flares .
rapid variability over the entire electromagnetic spectrum is a characteristic property of blazars .
episodic gamma - ray emission with flares durations of hours to weeks is quite typical to many gamma - ray blazars , with the most extreme activity recorded in the sub - class of tev agns , e.g. , mrk 421 , mrk 501 , pks 2155 - 304 .
the observed variability time imposes a stringent constraint on the maximum size of the emission region , @xmath0 , which in extreme cases is inferred to be of the order of the gravitational radius of the putative black hole .
the naive expectation is that the large amplitude short duration flares seen in gamma - ray blazars originate from small radii , as the fraction of jet energy that can be tapped for production of @xmath1 rays scales as @xmath2 , where @xmath3 is opening angle of the jet and @xmath4 the emission radius , and is small for @xmath5 .
some of the variable gamma - ray flux may be attributed to a sparking gap at the base of the jet , as proposed for m87 ( levinson 2000 , neronov & aharonian 2007 , levinson & rieger 2011 ) , however , in powerful blazars the spectrum emitted from the gap is unlikely to extend beyond a few gev , owing to the large pair - production opacity .
this , and the indications of correlated emission at much lower energies ( radio - to - x rays ) strongly suggest that the jets are important sources of episodic gamma - ray emission . it has been argued that in some cases the observations favor models in which the variable gamma - ray emission originates from small regions locate at large radii ( e.g. , sikora et al . 2008 , agudo et al . 2011 ) .
such events may be produced by a converging shock in a reconfinement nozzle ( bromberg & levinson 2009 ) or magnetic reconnection in minijets ( giannois 2013 ) .
however , recent analysis ( nalewajko et al . 2014 ) challenges the far dissipation scenarios , indicating typical emission radii in the range 0.1 - 1 pc .
as shown below , on these scales the radiative drag can be substantial .
in this paper we consider the effect of compton drag on the dynamics and emission of dissipative fluid shells .
we show that if dissipation of the bulk energy commences not too far out , the blob experiences strong deceleration that leads to large variation of the observed flux emitted from the blob by virtue of the change in its bulk lorentz factor .
the effect of compton drag on the dynamics of a relativistic jet has been considered earlier by several authors under different assumptions ( e.g. , phinney 1987 , li et al .
1992 , sikora et al .
our method is similar to that presented in sikora et al .
( 1996 ) , however , we focus on short events that may lead to rapid flares , and compute the lorentz factor profiles and the resulting gamma - ray lightcurves for a range of conditions . for typical ambient luminosities ,
the duration of flares produced by this mechanism is on the order of the characteristic size of the emitting blob .
thus , flare durations as short as the dynamical time of the central engine can naturally be accounted for in this model , provided that the bulk energy can be dissipated at a large enough rate .
a preliminary account of the decelerating shell model is given in levinson ( 2007 ) .
here , we present an elaborated analysis of the dynamics of the flow , and also compute the resulting light curves .
the construction of this model was originally motivated by the apparent discrepancy inferred in tev blazars between the relatively large doppler factors , @xmath6 , required to avoid strong attenuation of the vhe flux emitted from the inner regions , and the much lower values , @xmath7 a few , inferred from superluminal motions and source statistics ( georganopoulos & kazanas 2003 , levinson 2007 ) .
however , the question of how the compton drag affects the dynamics and emission of dissipative outflows is of general interest , and is relevant essentially to all blazars .
in the simplified model invoked here , a blob is ejected from a central engine of size @xmath8 and accelerated to a lorentz factor @xmath9 .
when it reaches some radius @xmath10 cm , dissipation suddenly starts , e.g. , via formation of shocks or explosive conversion of magnetic energy , leading to continuous particle acceleration .
inverse compton scattering of ambient photons then leads to a strong radiative drag that tends to decelerate the blob . in certain circumstances ,
this drag may be compensated by injection of energy and momentum into the blob by some external agent , delaying the deceleration of the emitting plasma .
the deceleration may be delayed also in cases where the dissipation commences very close to the central engine , where the radiation field is dominated by direct illumination from the disk and is highly unisotropic , as in this zone the blob may propagate at the equilibrium lorentz factor until reaching scales where the external radiation field is roughly isotropic ( e.g. , sikora et al . 1996 , vuillaume et al . 2014 ) . in the present model we shall refer to this case as delayed deceleration . in reality , the structure and velocity of the emitting plasma are expected to be non - uniform by virtue of confined dissipation and rapid cooling . in what follows
we ignore such complications and compute the dynamics and emission of the blob under simplifying assumptions about the microphysics of dissipation .
the energy and momentum fluxes of the emitting fluid are given by , @xmath11 where @xmath12 is the proper specific enthalpy , @xmath13 , @xmath14 and @xmath15 are the pressure , proper density and proper internal energy , respectively , @xmath16 is the proper magnetic field , and @xmath17 is the bulk lorentz factor . in terms of the total power , @xmath18 , where @xmath3 is the half opening angle of the flow , we have @xmath19 for @xmath20 , @xmath21 cm and @xmath22 erg s@xmath23 . for simplicity , we assume that the ambient radiation field is roughly isotropic in the frame of the central engine , and has a luminosity @xmath24
. this quasi - isotropic radiation field is contributed by scattering and reprocessing of the central uv radiation by gas in the broad line region , and by emission from a dusty torus ( e.g. , joshi et al .
2014 ) . here
, @xmath24 represents the sum of these components , and is a fraction of the total luminosity of the continuum source .
the radial profile of the ambient intensity intercepted by the jet depends on the geometries of the broad line region and the dusty torus ( e.g. , joshi et al . , 2014 ) . denoting @xmath25 , we express it as @xmath26 with the normalization @xmath27 , such that the total energy density at radius @xmath28 is @xmath29 . for central emission , @xmath30 . for a flat profile within some scale @xmath31 , @xmath32 at @xmath33 . for the spectrum of the ambient radiation field
we adopt @xmath34 that mimics a typical soft spectrum . in the rest frame of the blob the intensity is given by @xmath35 and the comoving energy density by @xmath36 from equation ( [ is - exact ] ) it is seen that the comoving intensity peaks sharply around the direction opposite to the blob velocity , viz . , @xmath37 .
to simplify our calculations we approximate the comoving intensity as a beam moving in the direction @xmath37 : @xmath38 with @xmath39
. it can be readily verified that with this choice the comoving energy density satisfies equation ( [ us - exact ] ) .
we further suppose that the electron distribution function is isotropic in the fluid rest frame and can be approximated as a power law : @xmath40 ; @xmath41 , where @xmath42 is the corresponding electron energy , as measured in the comoving frame .
we define @xmath43 to be the fraction of the total internal energy carried by the relativistic electrons , that is , @xmath44 , where @xmath45 , subject to the boundary condition @xmath46 . in terms of this parameter @xmath47 , and we note that @xmath48 in the limit @xmath49 . we suppose that @xmath50 reaches a maximum value over some characteristic length scale that depends on the microphysics of the specific dissipation mechanism .
this length scale is typically of the order of the gyroradius of accelerated electrons , which is shorter than the cooling time of the highest energy electrons .
thus , we take @xmath50 to be constant at @xmath51 .
the maximum lorentz factor of the electron distribution , @xmath52 , is likely to be limited by cooling . in the limit of bohm diffusion
the acceleration time is @xmath53 .
equating with the compton cooling time , @xmath54 , and adopting @xmath55 , we estimate @xmath56 in terms of the fraction @xmath57 , @xmath58 . here
@xmath59 is the magnetic field at the onset of dissipation .
we consider only situations in which dissipation is rapid enough to produce high enthalpy inside the blob on a timescale comparable to the radiative time .
then , one can adopt a simple prescription in which the dissipation is formally treated as an initial condition ( at @xmath60 ) of the flow equations .
the magnetization of the flow is then expected to be low , so that the effect of the magnetic field on the dynamics of the system can be neglected .
this may not apply to cases were the dissipation time is much longer than the light crossing time of the shell .
we suppose that the dissipated energy is redistributed in a way that a fraction @xmath50 of the total dissipation energy is injected in the form of a power law electron distribution , as explained above . under these assumptions
, the dynamics of the blob subsequent to the onset of dissipation is governed by the equation @xmath61 where the source terms @xmath62 account for the radiative drag acting on the blob , and are given explicitly in equations ( [ s_c^0-lab])- ( [ s_c^i - lab ] ) . for a conical expansion ,
the @xmath63 and @xmath28 components of equation ( [ eq : mot - gen ] ) can then be written , to order @xmath64 , in the form ( see appendix [ app : eq - mot ] ) @xmath65 in terms of the coordinates @xmath66 and @xmath67 , where @xmath68 is the lagrangian derivative , and @xmath69 denotes the trajectory of the blob s front .
the interior of the blob encompasses the range @xmath70 .
now , for the relativistic electron population invoked above @xmath71 , where @xmath72 and to order @xmath64 the source terms , equations ( [ s_c^0-lab ] ) - ( [ s_c^i - lab ] ) , are given by @xmath73 and @xmath74 we consider first a uniform blob ( @xmath75 ) having a fixed length @xmath76 in the frame of the central engine .
its motion is then characterized by a single lorentz factor at all times , @xmath77 .
we suppose that microphysical processes redistribute the dissipation energy in a way that a fraction @xmath50 of the total internal energy is uniformly injected in the form of a power law electron distribution inside the blob . for clarity , we ignore here the contribution of the thermal population to the compton drag . as shown below , this
is justified when @xmath78 is larger than the thermal energy , here @xmath52 is the upper cutoff of the electron distribution given in equation ( [ gmax ] ) , and @xmath79 is defined below .
the inclusion of the thermal electrons in the source terms does not alter our results significantly , and at any rate , if the dissipation produces a population of relativistically hot electrons with a thermal energy @xmath80 rather than a power law distribution , it can be readily accounted for by taking @xmath81 , @xmath82 in equation ( [ alpha ] ) below .
we find it convenient to use the parametrization @xmath83 , where @xmath84 is the average energy of nonthermal electrons .
for the power law energy distribution invoked above we have @xmath85}{[1-(\gamma_{2}/\gamma_{1})^{2-q}]}\frac{\gamma_1}{\gamma_2}. \label{chi}\ ] ] with @xmath86 and @xmath87 it gives @xmath88 .
for @xmath89 we have @xmath90^{-1}>0.1 $ ] , and for @xmath91 , @xmath92 .
a reasonable choice for the minimum electron energy , adopted henceforth , is @xmath93 .
we further define the fiducial coordinate @xmath94 , and write @xmath95 , denoting @xmath96 . equations ( [ mot - radial-0 ] ) and ( [ mot - radial - r ] ) with @xmath97 can then be re - expressed as @xmath98 in terms of the constant @xmath99 the solution of these coupled equations depend on the equation of state and the assumptions about @xmath100 .
approximate analytic solutions can be obtained in the case @xmath101 for which @xmath102 is a good approximation . for a relativistically hot blob
, we adopt the equation of state @xmath103 .
we then obtain @xmath104 as a first example , let us take the maximum electron energy to be constant during the deceleration phase , that is .
, @xmath105 .
the solution then reads : @xmath106},\label{g - sol1}\ ] ] where @xmath107dx^\prime$ ] , with @xmath108 for @xmath109 and @xmath110 for a flat profile , @xmath111 . as a second example
, we suppose that @xmath100 is given by equation ( [ gmax ] ) , and that the magnetic field evolution is dictated by the ideal mhd limit , viz .
, @xmath112 .
we then obtain the solution @xmath113 ^ 2},\label{g - sol2}\ ] ] with @xmath114^{1/2}dx^\prime$ ] .
it is worth noting that these analytic solutions hold only for times @xmath115 at which the blob is relativistically hot , and @xmath116 .
figure [ fig - f1 ] exhibits exact numerical solutions of equations ( [ eq - mot - v2])-([eq - state - v2 ] ) for @xmath117 , @xmath100 given by equation ( [ gmax ] ) with @xmath118 , intensity profile @xmath109 , and different choices of @xmath119 . as a check
, we obtained solutions also for different intensity profiles and different assumptions about @xmath100 , and found little differences in the lorentz factor profiles for a given @xmath119 in the regime of rapid deceleration .
the key parameter that determines the dynamics is @xmath119 .
the model outlined above assumes uniform acceleration of electrons to nonthermal energies at all times .
this assumption may hold in certain situations but not in general . in shocks , for instance , particle acceleration is confined to a region around the shock front of a characteristic size comparable to the gyroradius of the accelerated electrons , as measured in the shock frame . for the highest energy electrons
it is roughly equal to the cooling distance ( if not limited by escape ) . for lower energy electrons it may be even smaller . as the downstream fluid moves away from the shock , electrons of energy @xmath120 cool over time @xmath121 , where @xmath122 is the comoving energy density of the external radiation ( see eq .
[ us - exact ] ) .
the distance traversed by these electrons in the black hole frame before cooling down is @xmath123 . using equation ( [ alpha ] )
then yields @xmath124 for @xmath125 .
consequently , ic scattering off nonthermal electrons should not lead to significant deceleration of the blob , unless the entire shock energy is converted to non - thermal electrons ( @xmath81 ) .
however , about half the shock energy is carried by a population of thermal electrons having an average comoving energy @xmath126 just downstream of the shock , where @xmath127 denotes the shock lorentz factor ( that is , the lorentz factor of the upstream fluid measured in the shock frame ) .
these electrons cool as they propagate away from the shock
. the lagrangian rate of change of the thermal energy in the downstream flow is governed by the equation @xmath128 to compute the shock structure one needs to solve equations ( [ mot - radial-0])-([mot - radial - r ] ) coupled to equation ( [ gamm_t - cool ] ) .
such treatment is beyond the scope of this paper . to illustrate the effect of radiative drag on the shock we invoke the uniform shell approximation , that is , keep only the lagrangian derivatives in equations ( [ mot - radial-0])-([mot - radial - r ] )
then , adopting @xmath129 and @xmath130 we obtain the rate of change of the bulk lorentz factor : @xmath131 with @xmath95 , the solution of the coupled equations ( [ gamm_t - cool ] ) and ( [ gamm - shock ] ) reads : @xmath132^{4/7}},\label{g - sol - therm}\ ] ] here @xmath133dx^\prime$ ] , and @xmath134 since @xmath135 significant deceleration is expected on sub - parsec scales .
for instance , assuming a flat intensity profile below the radius @xmath136 cm , with @xmath137 erg @xmath138 at @xmath139 , we obtain @xmath140 .
thus , colliding shells having a lorentz factor @xmath141 will experience substantial deceleration .
now , the deceleration of the downstream plasma should lead to a gradual strengthening of the reverse shock and a weakening of the forward shock .
our preliminary calculations indicate substantial over - compression of the reverse shock already at @xmath142 a few . in the frame of the central engine
this translates to a deceleration of the entire shocked shell , as described qualitatively by the simple blob model outlined in the preceding section .
the strengthening of the reverse shock should lead to enhanced dissipation rate , whereby the bulk energy of the unshocked shell is ultimately radiated away with high efficiency .
consequently , as long as the beaming cone of emission is narrower than the angular extent of the shell , the total flux observed will remain roughly constant .
thus , the delayed - deceleration model is relevant for the evolution of the total flux .
however , the change in the doppler factor resulting from the deceleration of the shocked shell leads to a change in the observed energy of scattered photons and this , in turn , can significantly alter the evolution of the sed .
in particular , we anticipate different durations and times of peak emission of flares observed in different energy bands .
the non - uniformity of the emitting plasma downstream of the shock adds complexity .
a comprehensive analysis of emission from internal shocks in blazars is given in joshi & bottcher ( 2011 ) and joshi et al .
( 2014 ) ignoring the effect of compton drag on the dynamics of the shock .
the decay of the emission in their model is due to cooling of the emitting electrons following shock crossing .
this situation is well represented by the delayed deceleration model we adopt below .
the neglect of compton drag is justified only at radii where @xmath143 in equation ( [ alphat ] ) . as explained above
, the inclusion of compton drag and non - uniform particle acceleration should have a profound effect on the light curves .
a complete treatment of the dynamics and emission of internal shocks that are subject to a strong compton drag will be presented in a future publication . as will be shown below , the shape of the lightcurves reflects the geometry and dynamics of the emitting material .
in particular , time retardation associated with the curvature of the emitting surface sets a limit on the rise and decay times of the flare .
furtheremore , as mentioned above certain situations can be described by delayed deceleration of the blob , that can have an important effect .
to illustrate such effects , we shall consider also cases in which the lorentz factor remains constant at its initial value @xmath9 up to some radius @xmath144 , and only then deceleration commences .
in reality , the structure of the emitting zone is expected to be non - uniform in those cases , depending on the specific model .
those details may affect the resulting emission .
our purpose here is merely to illustrate how dynamical effects are imprinted in the lighcurves . for this purpose
our simple treatment of delayed deceleration is sufficient .
moreover , this prescription also describes situations in which the radiative drag is too small to affect the dynamics of the shell , and the decay of the emission at the end of the dissipation phase is due to cooling of the emitting electrons .
since we are mainly interested here in the high - energy emission from the blob , we consider only the contribution of inverse compton scattering . the spectral evolution measured by a distant observer
is computed as follows : for a given choice of parameters we first solve equations ( [ eq - mot - v2])-([eq - state - v2 ] ) numerically to obtain the lorentz factor profile @xmath145 in the frame of the black hole . adopting the beam approximation , the intensity of the background radiation at any radius @xmath28 is then transformed into the rest frame of blob using @xmath145 in equation ( [ i - beamed ] ) .
the comoving intensity thereby obtained is used to compute the comoving emissivity of the scattered radiation , @xmath146 .
the emissivity in the black hole frame is given by @xmath147^{-2 } j_{sc}^\prime(\epsilon^\prime,\mu^\prime , r)$ ] , where @xmath148 and @xmath149 .
the scattered intensity emitted by the blob is obtained upon integrating the emissivity across the blob , taking into account the time delay between emission of photons from different locations : @xmath150 where the distance @xmath0 is related to the blob s length @xmath76 through @xmath151dr^\prime,\ ] ] and is a function of @xmath28 . in deriving equation ( [ i - obs ] )
we assumed that the emissivity is uniform inside the blob and vanishes outside it .
note that @xmath152dr^\prime / c \label{t_ob}\ ] ] is the time measured by a distant observer viewing the blob at an angle @xmath153 relative to its direction of motion , so that formally @xmath154 .
high energy photons emitted by the blob will be attenuated by pair production on background photons . here
we model this attenuation by an exponential cutoff at the corresponding pair production optical depth @xmath155 .
the latter is computed using equations ( 3.1)-(3.3 ) in blandford & levinson ( 1995 ) , and plotted in figure [ fig - f2 ] . the spectral flux measured by a distant observer viewing the blob at time @xmath156 at an angle
@xmath157 satisfies @xmath158 where the radius @xmath28 is computed at the observed time @xmath159 using equation ( [ t_ob ] ) , that is @xmath160 , and the inner integration is over the emitting surface of the blob at the retarded time @xmath161 . in the immediate deceleration case ,
the rise and decay times of the observed flux are dominated by temporal delays associated with the curvature of the emitting surface . for a conically expanding shell
the emitting surface is spherical , and equation ( [ ob - flux ] ) reduces to : @xmath162 with @xmath163dy^\prime\right\},\\ & & \mu_1(r , y)=\min \left\{1,1-\frac{(1-\mu_2)(r - y)-\delta}{r - y-\delta}\right\},\end{aligned}\ ] ] where @xmath3 is the opening angle of the flow , and @xmath164 is determined from the condition @xmath165 .
the evolution of the sed is shown in figure [ fig - f3 ] in the case of immediate deceleration ( left panel ) and delayed deceleration with @xmath166 ( right panel ) .
the corresponding lightcurves are displayed in figure [ fig - f4 ] at a photon energy @xmath167 gev .
the right pannels exhibit the emissivity , and the left pannels the observed flux computed for these emissivities using equation ( [ ob - flux - curvt ] ) .
the effect of the jet opening angle on the shape of the lughtcurve is more prominent in the non - delayed case , as seen in the upper left panel of figure [ fig - f4 ] ; in the delayed case this effect is essentially negligible .
these lightcurves are rather typical in the regime where the deceleration time of the blob , @xmath168 , is not much larger than its light crossing time @xmath169 . in this regime ,
the duration of the flare measured by a distant observer viewing the source at an angle smaller than the opening angle of the flow and the overall shape of the lightcurve depend on details , as seen in figure [ fig - f4 ] . in the case of immediate deceleration
the lightcurve is asymmetric , with the rise and decay times determined by the curvature of the emitting surface . for a sufficiently large lorentz factor , such that the beaming angle is smaller than the opening angle of the flow ( i.e. , @xmath170 ) , these times , as measured by a distant observer , satisfy @xmath171 .
those times might be comparable to the light crossing time of the blob if dissipation commences at a radius @xmath172 , as in the case shown in [ fig - f4 ] and [ fig - f5 ] . in the case of delayed deceleration the rise and decay times are determined by the delay : @xmath173 .
the lightcurve tends to be more symmetric in this case ( see figs [ fig - f5 ] for a comparison ) .
the change in the emissivity during the coasting ( pre - deceleration ) phase , as seen in the lower right panel in figure [ fig - f4 ] , is due to the dependence of the intensity of target photons and the density of emitting electrons on @xmath28 ( @xmath174 for the conical flow considered in the above examples , with intensity profile @xmath109 in equation ( [ is - star ] ) ) .
this dependence may slightly affect the shape of the lightcurve .
the shape of the lightcurve displayed in the lower left panel of figure [ fig - f4 ] is in qualitative agreement with those computed in joshi et al .
( 2014 ) , and is quite typical to observed gamma - ray flares .
the rough symmetry of the light curve is due to the assumed uniformity of the emissivity . in cases where the acceleration of the electrons is confined to a small region inside the shell , e.g. , acceleration in shock fronts ,
we anticipate a faster rise and a slower decay , particularly at very high energies , at which the cooling time is much shorter than the light crossing time of the shell .
we considered the dynamics of a dissipative shell in the presence of a strong radiative drag .
our analysis indicates that for rapid dissipation substantial deceleration is anticipated in blazars on sub - parsec scales ( and even parsec scales for sufficiently luminous sources ) , that should give rise to rapid , large amplitude variability owing to changes in the beaming factor .
it is worth noting that even modest changes in the lorentz factor can lead to large amplitude variations of the emitted flux , owing to its sensitive dependence on the doppler factor . in principle
, this mechanism can produce flares with durations as short as the dynamical time of the central engine , at a very high efficiency .
the radiative drag exerted on the thermal electrons alone should lead to rapid deceleration at radii @xmath175 cm , where @xmath9 is the bulk lorentz factor at the onset of dissipation and @xmath176 erg s@xmath23 is the luminosity of the ambient radiation field intercepted by the flow , provided the dissipation is rapid enough to keep the specific enthalpy large ( that is , @xmath177 ) .
this is , for instance , the situation in internal shocks , as shown in section [ sec : shocks ] , but may also occur in other cases , e.g. , effective magnetic field dissipation in poynting flux dominated jets ( sikora et al . 1996 ) .
scattering of ambient photons by non - thermal electrons accelerated in situ will dominate the radiative force if the proper scale over which electrons are accelerated exceeds @xmath178 , where @xmath179 is the cooling time of the non - thermal electrons , as measured in the rest frame of the flow , and @xmath50 is the fraction of total energy carried by the non - thermal population . in general ,
the effect of radiative drag is expected to be more prominent in fsrqs than in bl lacs .
nonetheless , equation ( [ alpha ] ) indicates that deceleration of dissipative shells may also be relevant to low luminosity sources , provided that electrons can be effectively accelerated to the cooling cutoff .
this seems to be case in tev blazars .
it has been shown elsewhere ( levinson 2007 ) that if the tev spectrum emitted during strong flares extends to energies at which the pair production opacity exceeds unity deceleration should be effective .
it is naively anticipated that the gamma - ray emission produced through ic scattering of ambient photons by the nonthermal electrons accelerated in the blob will be correlated with lower energy emission generated by synchrotron cooling of the same electrons . however , synchrotron self - absorption may give rise to a strong suppression of the radio emission in cases where the deceleration occurs well below the radio core . in that case
, the gamma ray flare will either precede the ejection of a superluminal component , or not be accompanied by one at all , depending on the asymptotic bulk lorentz factor and the synchrotron cooling time . on the other hand , the onset of dissipation depends on the duty cycle ( the time interval between ejections of consecutive shells in the case of internal shocks ) , and is expected to occur over a range of scales , even in an individual object .
blobs that dissipate their energy at large enough radii will not experience strong deceleration . as a consequence ,
a variety in the behavior of flares is expected , as indeed revealed by recent multi - waveband studies ( e.g. , marscher et al .
2011 ) .
we thank noemie globus for help .
this research was supported by a grant from the israel science foundation no .
the general expression for the source term associated with compton drag is ( phinney 1982 ; sikora et al .
1996 , van putten & levinson 2012 ) : @xmath180\ ] ] here @xmath181 and @xmath182 denote the distribution functions of the photons and electrons , respectively . in the rest frame of the blob
the electron distribution @xmath182 is isotropic , and it is convenient to compute the source terms there , and then to transform .
we have @xmath183 where @xmath184 is the cosine of the angle between the photon and electron directions , and @xmath185 , @xmath186 .
the zeroth component reads : @xmath187\\ & = & \sigma_t\int d^3k'f'_r\int d^3p ' f'_e k^{'0}\left[(1-\beta\mu)-\gamma^2(1-\beta\mu)^2\right]\\ & = & \sigma_t\int d^3k ' k^{'0 } f'_r\int dp ' p^{\prime 2}f'_e \int d\mu\left[(1-\beta\mu)-\gamma^2(1-\beta\mu)^2\right].\\\end{aligned}\ ] ] now , @xmath188 , and using @xmath189 one has @xmath190 the spatial component reads : @xmath191\\ & = & \sigma_t\int d^3k'f'_r\int d^3p ' f'_e k^{\prime i}\left[(1-\beta\mu)-\gamma^2\beta(1-\beta\mu)^2\right]\\ & = & \sigma_t\int d^3k ' k^{'i } f'_r\int dp ' p^{'2}f'_e ( 2+\frac{4}{3}\gamma^2\beta^2)=\sigma_tn'_e<1 + 2\gamma^2\beta^2/3>t_r^{'0i},\end{aligned}\ ] ] where @xmath192 now , in the black hole frame the radiation field is isotropic and we have @xmath193 , @xmath194 , @xmath195 .
then @xmath196 thus , we finally have for the source terms in the comoving frame : @xmath197 to obtain the source term in the lab frame we perform a lorentz transformation , @xmath198.\label{s_c^0-lab } \\ & & s_c^{i}=\gamma\beta_\gamma^{i } s_c^{'0}+\gamma s_{c}^{\prime i}=-\frac{4}{3}\sigma_tn'_e\gamma^3\beta_\gamma^iu_s[1+(5+\beta^2_\gamma)<\gamma^2\beta^2/3 > ] \label{s_c^i - lab}\end{aligned}\ ] ] for cold electrons @xmath199 , and @xmath200 , as required , since in the thomson limit invoked here the scattering is fully elastic and there should be no energy loss in the rest frame of the blob if the electrons are cold . in the lab frame
we then recover the result @xmath201 .
in spherical coordinates , the radial expansion of a hydrodynamic shell is governed by the equations @xmath202 where @xmath203 and @xmath204 are source terms that account , respectively , for energy and momentum losses by the radiative drag .
the above equations can be combined to yield @xmath205 in terms of the lagrangian derivative @xmath68 .
we suppose that at time @xmath66 the dissipative shell is contained between the radii @xmath206 and @xmath207 . for simplicity we suppose that the length of the shell @xmath76 is constant in the black hole frame .
the velocity of the shell is then uniform with @xmath208 . transforming to the coordinates @xmath209 , @xmath210
, we have @xmath211 using the relation @xmath212 , equation ( [ eq - mot - appb ] ) gives @xmath213 to order @xmath64 .
equation ( [ mot - radial0-app ] ) yields 9 agudo , i. et al .
2011 , apj , 735 , l10 blandford , r. d. & levinson , a. 1995 , apj , 441 , 79 bromberg o. , & levinson a. 2009 , apj , 699 , 1274 dermer , c.d . & schlickeiser , r. 1993 , apj , 416 , 458 georganopoulos m. , & kazanaz d. 2003 , apj , 594 , l27 ghisellini g. & madau 1996 , mnras , 280 , 67 giannios , d. 2013 , mnras , 431 , 355 joshi , m. & bottcher , m. 2011 , apj , 727 , 21 joshi , m. marscher , a. p. & bottcher , m. 2014 , apj , 785 , 132 levinson a. 2000 , phys .
85 , 912 levinson a. 2006 , int .
phys . a , 21 , 6015 levinson a. 2007 , apj , 507 , 145 levinson a. & blandford r. d. 1996 , apj , 456 , l29 levinson a. & rieger , f. 2011 , apj , 730 , 123 li , z. begelman , m. c. chiueh , t. 1992 , apj , 384 , 567 marscher , a. jorstad , s. g. larionov , v. m. aller , m. f. & lahteenmaki , a. 2011 , j. astrophys .
astr . , 32 , 233 nalewajko , k. begelman , m. & sikora , m. 2014 , apj , 789 , 161 neronov , a. & aharonian , f. a. 2007 , apj , 671 , 85 phinney , e. , 1982 , mnras , 198 , 1109 phinney , e. , 1987 , in zensus j. a , pearson t. j. , eds , super ? `` ' ' luminal radio sources .
cambridge univ . press , cambridge ,
p.301 " sikora , m. begelman , m. c. & rees , m. j. 1994 , apj , 421 , 153 sikora , m. sol , m. begelman , m. c. & madejski , g. m. 1996 , mnras , 280 , 781 sikora , m. moderski , r. & madejski , g. m. 2008 , apj , 675 , 71 stawarz l. et al .
2006 , mnras , 370 , 981 stern , b. & poutanen , j. 2011 , mnras , 417 , 11 tavecchio , f. et al .
2013 , mnras , 435 , l24 van putten , m & levinson , a. 2012 , relativistic astrophysics of the transient universe ( cambridge : cambridge univ .
press ) vuilaume , t. henri , g. & petrucci , p - o , 2014 , proc .
( arxiv:1411.4571v1 ) | we compute the dynamics and emission of dissipative shells that are subject to a strong compton drag , under simplifying assumptions about the dissipation mechanism . we show that under conditions prevailing in blazars , substantial deceleration is anticipated on sub - parsec and parsec scales in cases of rapid dissipation
. such episodes may be the origin of some of the flaring activity occasionally observed in gamma ray blazars .
the shape of the light curves thereby produced reflects the geometry of the emitting surface if the deceleration is very rapid , or the dynamics of the shell if the deceleration is delayed , or initially more gradual , owing , e.g. , to continuous injection of energy and momentum . |
in a number of recent works [ 1 - 3 ] we have calculated various hadronic correlation functions and compared our results to results obtained in lattice simulations of qcd [ 4 - 6 ] .
the lattice results for the correlators , @xmath2 , may be used to obtain the corresponding spectral functions , @xmath3 , by making use of the relation g ( , t)=_0^d ( , t ) k ( , , t),where k ( , , t)=.the procedure to obtain @xmath3 from the knowledge of @xmath2 makes use of the maximum entropy method ( mem ) [ 7 - 9 ] , since @xmath2 is only known at a limited number of points . in our studies of meson spectra at @xmath4 and at @xmath5 we have made use of the nambu jona - lasinio ( njl ) model .
the lagrangian of the generalized njl model we have used in our studies is l&=&(i - m^0)q+_i=0 ^ 8 [ ( ^i q)^2 + ( i _ 5 ^i q)^2 ] + & -&_i=0 ^ 8 [ ( ^i_q)^2 + ( ^i_5_q)^2 ] + & + & + + l_conf . here , @xmath6 is a current quark mass matrix , @xmath6=diag@xmath7 .
the @xmath8 are the gell - mann ( flavor ) matrices and @xmath9 , with @xmath10 being the unit matrix .
the fourth term is the t hooft interaction and @xmath11 represents the model of confinement used in our studies of meson properties . in the study of hadronic current correlators
it is important to use a model which respects chiral symmetry , when @xmath12 .
therefore , we make use of the lagrangian of eq .
( 1.3 ) , while neglecting the t hooft interaction and @xmath11 .
thus , there are essentially three parameters to consider , @xmath13 , @xmath14 and a cutoff parameter @xmath15 , which restricts the momentum integrals so that @xmath16 .
when we use the njl model to study matter at finite temperature , we introduce the temperature - dependent parameters @xmath17 and @xmath18 .
( we have also used a gaussian cutoff for the momentum integrals in our earlier work . ) these parameters have been adjusted to obtain fits to the spectral functions @xmath3 for @xmath19 and @xmath20 , which are the values of @xmath21 studied in the lattice simulations of qcd [ 10 ] .
the temperature - dependent coupling constants and cutoff parameters of our work are analogous to the corresponding density - dependent parameters introduced in ref .
[ 11 ] and [ 12 ] . further study of models with temperature - dependent and density - dependent parameters are of interest and a general theoretical formalism for the introduction of such dependencies should be considered . in this work
we limit our study to the data for the vector correlator at @xmath0 [ 10 , 13 ] and therefore only need to specify @xmath18 and the momentum cutoff .
( we remark that the results for the scalar , vector , pseudoscalar and axialvector correlators are quite similar in the deconfined region . ) in figs . 1 and 2 we show the data obtained by the mem method at @xmath19 and @xmath20 for both pseudoscalar and vector correlators [ 10 , 13 ] .
the second peaks in these correlators are known to be a lattice artifacts [ 13 ] .
the organization of our work is as follows . in appendix
a we present the formalism for calculation of pseudoscalar and vector correlation functions . in appendix
a we discuss the calculation of the correlator in the case the quark and antiquark carry zero total momentum . in appendix b
we show how the formalism is modified for the correlator calculated at finite momentum @xmath22 . in section
ii we present the results of our calculation of the imaginary part of the correlator @xmath23 .
( since we place @xmath22 along the _
z_-axis this quantity may be written as @xmath24 in accord with the notation of ref . [
our results for @xmath23 will be presented for a series of values of @xmath25 in section ii . in section
iii we present our result for the coordinate - dependent correlator @xmath26 which is proportional to the correlator defined in eq .
( 1 ) of ref . [
10 ] , c(z)=12_-^dp_ze^ip_zz_0^d .
we may also use the form c(z)=14_-^dp_ze^ip_zz_0^dp^2 .
finally , in section iv we present our conclusion and further discussion .
we make use of the formalism presented in appendices a and b to obtain values of the vector correlator at @xmath0 . here
we take @xmath27 mev , since we have usually made comparison to lattice calculations made in the quenched approximation . in fig .
3 we present @xmath28 , for various values of @xmath25 , as function of @xmath29 .
comparison may be made to the lattice data shown in fig .
2 [ 10 , 13 ] .
( we again note that our calculation does not reproduce the second peak in the lattice data which is known to be a lattice artifact [ 13 ] . ) the curves shown in fig .
3 are given for values of @xmath25 ranging from 0.10 gev to 2.10 gev in steps of 0.20 gev .
for these calculations we have used @xmath30gev .
our results for the various values of @xmath25 given in fig.3 may be compared to fig.20 of ref.[14 ] .
we see that the results calculated by completely different methods are similar . for pseudoscalar states obtained by mem
are shown [ 10,13 ] .
the solid line is for @xmath19 and the dashed line is for @xmath20 .
the second peak is a lattice artifact [ 13 ] . ] for vector states obtained by mem are shown [ 10 , 13 ] . see the caption of fig .
1 . the second peak is a lattice artifact [ 13 ] . ]
is shown for various values of @xmath31 as a function of @xmath29 . starting with the topmost curve
the values of @xmath32 in gev units are 0.10 , 0.30 , 0.50 , 0.70 , 0.90 , 1.10 , 1.30 , 1.50 , 1.70 , 1.90 and 2.10 . here
we have used @xmath33 gev@xmath34 and @xmath30 gev . ]
defined in eq.(1.5 ) is shown .
the dotted line represents a fit using an exponential function . ]
we have used the results of our calculations which were presented in fig . 3 to calculate @xmath26 of eq . ( 1.5 ) .
the result of that calculation is shown in fig .
we note that the simple assumption for the behavior of this correlator that is usually made , @xmath35 $ ] , is born out in our calculation for 1gev@xmath36gev@xmath37 . for the study of charmonium on the lattice ,
values found for the screening mass are given in ref .
[ 14 ] . for the result shown in fig .
4 we obtain a screening mass of 1.02 gev which may be compared to the value of @xmath38@xmath39=1.27 gev .
the dotted line represents an exponential fit to our result .
recent theoretical work concerning the quark - gluon plasma has been discussed by shuryak @xcite .
he notes that the physics of excited matter produced in heavy - ion collisions in the region @xmath40 is different from that of a weakly coupled quark - gluon plasma .
that is due to the strong coupling generated by the bound states of the quasiparticles .
these bound states appear as resonant structures when the imaginary parts of the hadronic correlators are extracted from lattice data using the maximum entropy method ( mem ) [ 4 - 9 ] .
these resonances lead to very strong interactions between the quasiparticles and are , in part , responsible for very small mean free paths and collective flow in heavy - ion collisions . that flow may be described by hydrodynamics .
indeed , shuryak suggests that ... if the system is macroscopically large , then its description via _ thermodynamics _ of its bulk properties ( like matter composition ) and _ hydrodynamics _ for space - time evolution should work " @xcite . in our present work we have presented a simple chiral model that is able to reproduce the resonances extracted from the lattice data via the mem procedure .
we have calculated the imaginary parts of the vector correlator for various values of the total external momentum @xmath22 .
we have also calculated the correlation function , @xmath26 , and find the simple exponential behavior @xmath35 $ ] for a limited range of _ z_. we have shown in an earlier work that exponential behavior with the appropriate screening mass may be obtained for the full range of _ z _ values if quite small values of @xmath15 ( of the order of 0.4 gev ) are used @xcite .
finally , we note that an extensive discussion of screening masses appears in ref .
@xcite . for ease of reference
, we present a discussion of our calculation of hadronic current correlators taken from ref.[3 ] .
the procedure we adopt is based upon the real - time finite - temperature formalism , in which the imaginary part of the polarization function may be calculated .
then , the real part of the function is obtained using a dispersion relation .
the result we need for this work has been already given in the work of kobes and semenoff @xcite .
@xcite the quark momentum is @xmath41 and the antiquark momentum is @xmath42 .
we will adopt that notation in this section for ease of reference to the results presented in ref .
@xcite . ) with reference to eq.(5.4 ) of ref .
@xcite , we write the imaginary part of the scalar polarization function as j_s(__p__^2 , t)=12n_c_s(__p__^0)^k_max ( ) + \{[1-n_1(k)-n_2(k ) ] ( _ _ p__^0-e_1(k)-e_2(k ) ) + -[n_1(k)-n_2(k ) ] ( _ _ p__^0+e_1(k)-e_2(k ) ) + -[n_2(k)-n_1(k ) ] ( _ _ p__^0-e_1(k)+e_2(k ) ) + -[1-n_1(k)-n_2(k ) ] ( _ _ p__^0+e_1(k)+e_2(k))}.here , @xmath43^{1/2}$ ] .
relative to eq.(5.4 ) of ref .
@xcite , we have changed the sign , removed a factor of @xmath44 and have included a statistical factor of @xmath45 . in addition , we have used a sharp cutoff , @xmath15 , for the momentum integral .
we also note that n_1(k)=1e^e_1(k)+1,and n_2(k)=1e^e_2(k)+1.for the calculation of the imaginary part of the polarization function , we may put @xmath46 and @xmath47 , since in that calculation the quark and antiquark are on - mass - shell . in eq.(a1 )
the factor @xmath48 arises from a trace involving dirac matrices , such that _ s&=&-[(k+m_1)(k - p+m_2 ) ] + & = & 2p^2 - 2(m_1+m_2)^2,where @xmath49 and @xmath50 depend upon temperature . in the frame where @xmath51 , and in the case @xmath52 , we have @xmath53 . for the scalar case , with @xmath52 ,
we find j_s(p^2 , t)=(1-)^3/2 [ 1 - 2n_1(k)],where k^2=4-m^2(t),with @xmath16 . for pseudoscalar mesons , we replace @xmath48 by _ p&=&-[i_5(k+m_1)i_5(k - p+m_2 ) ] + & = & 2p^2 - 2(m_1-m_2)^2,which for @xmath52 is @xmath54 in the frame where @xmath51 .
we find , for the @xmath38 mesons , j_p(p^2,t)=(1-)^1/2 [ 1 - 2n_1(k)],where @xmath55 , as above , with @xmath16 .
thus , we see that the phase space factor has an exponent of 1/2 corresponding to a _ s_-wave amplitude . for the scalars ,
the exponent of the phase - space factor is 3/2 , as seen in eq.(a6 ) . for a study of vector mesons
we consider _ ^v=[_(k+m_1)_(k - p+m_2)],and calculate g^_^v=4[p^2-m_1 ^ 2-m_2 ^ 2 + 4m_1m_2],which , in the equal - mass case , is equal to @xmath56 , when @xmath52 and @xmath51 .
this result will be needed when we calculate the correlator of vector currents .
note that , for the elevated temperatures considered in this work , @xmath57 is quite small , so that @xmath58 can be approximated by @xmath59 , when we consider the vector current correlation functions . in that case , we have j_v(p^2,t ) j_p(p^2,t).at this point it is useful to define functions that are not for @xmath60 : _ p(p^2,t)=(1-)^1/2[1 - 2n_1(k)],and _ v(p^2,t)=(1-)^1/2[1 - 2n_1(k)],for the functions defined in eq.(a14 ) and ( a15 ) we need to use a twice - subtracted dispersion relation to obtain @xmath61 , or @xmath62 . for example , _ p(p^2,t)=_p(0,t)+ [ _ p(p_0 ^ 2,t)-_p(0,t ) ] + + _
4m^2(t)^^2 ds , where @xmath63 can be quite large , since the integral over the imaginary part of the polarization function is now convergent .
we may introduce @xmath64 and @xmath65 as complex functions , since we now have both the real and imaginary parts of these functions .
we note that the construction of either @xmath66 , or @xmath67 , by means of a dispersion relation does not require a subtraction .
we use these functions to define the complex functions @xmath68 and @xmath69 . in order to make use of eq.(a16 ) , we need to specify @xmath70 and @xmath71 .
we found it useful to take @xmath72 2 and to put @xmath73 and @xmath74 .
the quantities @xmath75 and @xmath76 are determined in an analogous function .
this procedure in which we fix the behavior of a function such as @xmath77 or @xmath77 is quite analogous to the procedure used in ref .
@xcite . in that work we made use of dispersion relations to construct a continuous vector - isovector current correlation function which had the correct perturbative behavior for large @xmath78 and also described the low - energy resonance present in the correlator due to the excitation of the @xmath79 meson . in ref .
@xcite the njl model was shown to provide a quite satisfactory description of the low - energy resonant behavior of the vector - isovector correlation function .
we now consider the calculation of temperature - dependent hadronic current correlation functions .
the general form of the correlator is a transform of a time - ordered product of currents , ic(p^2 , t)=d^4xe^ipx<>,where the double bracket is a reminder that we are considering the finite temperature case . for the study of pseudoscalar states
, we may consider currents of the form @xmath80 , where , in the case of the @xmath38 mesons , @xmath81 and @xmath82 .
for the study of scalar - isoscalar mesons , we introduce @xmath83 , where @xmath84 for the flavor - singlet current and @xmath85 for the flavor - octet current @xcite . in the case of the pseudoscalar - isovector mesons
, the correlator may be expressed in terms of the basic vacuum polarization function of the njl model , @xmath86 .
thus , c_p(p^2 , t)=j_p(p^2 , t),where @xmath87 is the coupling constant appropriate for our study of @xmath38 mesons .
we have found @xmath88 by fitting the pion mass in a calculation made at @xmath4 , with @xmath89 gev .
the result given in eq.(a18 ) is only expected to be useful for small @xmath90 , since the gaussian regulator strongly modifies the large @xmath90 behavior .
therefore , we suggest that the following form is useful , if we are to consider the larger values of @xmath90 .
= .(as usual , we put @xmath91 . )
this form has two important features . at large @xmath92
, @xmath93 is a constant , since @xmath94 is proportional to @xmath92 .
further , the denominator of eq.(a19 ) goes to 1 for large @xmath92 . on the other hand , at small @xmath92
, the denominator is capable of describing resonant enhancement of the correlation function .
as we have seen , the results obtained when eq.(a19 ) is used appear quite satisfactory .
( we may again refer to ref .
@xcite , in which a similar approximation is described . ) for a study of the vector - isovector correlators , we introduce conserved vector currents @xmath95 with i=1 , 2 and 3 . in this case
we define j_v^(p^2 , t)=(g^-)j_v(p^2 , t)and c_v^(p^2 , t)=(g^-)c_v(p^2 , t),taking into account the fact that the current @xmath96 is conserved .
we may then use the fact that j_v(p^2,t ) = 13g_j_v^(p^2,t)and j_v(p^2,t)&= & 23(1-)^1/2[1 - 2n_1(k ) ] + & & j_p(p^2,t).(see eq.(a7 ) for the specification of @xmath97 . )
we then have c_v(p^2,t)=_v(p^2,t)11-g_v(t)j_v(p^2,t),where we have introduced _
v(p^2,t)&= & 23(1-)^1/2[1 - 2n_1(k ) ] + & & _
p(p^2,t ) . in the literature
, @xmath98 is used instead of @xmath99 [ 4 - 6 ] .
we may define the spectral functions _
v ( , t)=c_v ( , t),and _ p ( , t)=c_p ( , t ) , since different conventions are used in the literature [ 4 - 6 ] , we may use the notation @xmath100 and @xmath101 for the spectral functions given there .
we have the following relations : _ p ( , t)=_p ( , t),and = _ v ( , t),where the factor 3/4 arises because , in refs .
[ 4 - 6 ] , there is a division by 4 , while we have divided by 3 , as in eq.(a22 ) .
here we extend the work of appendix a to consider case of finite three - momentum , @xmath102 .
we consider the calculation of @xmath103 .
the momenta @xmath104 and @xmath102 are the values external to the loop diagram .
internal to the diagram , we have a quark of momentum @xmath105 leaving the left - hand vertex and an antiquark of momentum @xmath106 entering the left - hand vertex .
it is useful to define e_1(k)&=&|+/2| + & = & ( k^2 + 4+kp)^1/2and e_2(k)&=&|-/2| + & = & ( k^2 + 4-kp)^1/2.here @xmath107 and @xmath108 .
we have j_v(p^0 , , t)=12n_c_v(p^0)^k_max ( ) + \{[1-n_1(k)-n_2(k ) ] ( p^0-e_1(k)-e_2(k ) ) + -[n_1(k)-n_2(k ) ] ( p^0+e_1(k)-e_2(k ) ) + -[n_2(k)-n_1(k ) ] ( p^0-e_1(k)+e_2(k ) ) + -[1-n_1(k)-n_2(k ) ] ( p^0+e_1(k)+e_2(k))}.here , n_1(k)=1e^e_1(k)+1,and n_2(k)=1e^e_2(k)+1.in eq .
( b5 ) , the second and third terms cancel and the fourth term does not contribute .
it is useful to rewrite @xmath109 using = 2||_x(-x),where x^2&=&^2 + & = & .we find ||=12kp||,and obtain j_p(p^0 , , t)=12n_c_p(p^0)(2)^2^k_max + 12e_1(k)e_2(k)[1-n_1(k)-n_2(k ) ] || + ( -x)d().we note there is a singularity when @xmath110 .
that occurs when @xmath111 or @xmath112 . for our calculations we eliminate the point with @xmath112 when evaluating the angular integral over @xmath113 in the last expression .
we obtain j_p(p^0 , , t)=n_c_p(p^0)^k_max k^2dk + .|_x , where _ x _ is obtained from eq .
( b9 ) , x=^1/2 99 bing he , hu li , c. m. shakin , and qing sun , phys .
d * 67 * , 014022 ( 2003 ) .
bing he , hu li , c. m. shakin , and qing sun , phys .
d * 67 * , 114012 ( 2003 ) .
bing he , hu li , c. m. shakin , and qing sun , phys .
c * 67 * , 065203 ( 2003 ) .
i. wetzorke , f. karsch , e. laermann , p. petreczky , and s. stickan , nucl .
suppl . ) * 106 * , 510 ( 2002 ) f. karsch , s. datta , e. laermann , p. petreczky , and s. stickan , and i. wetzorke , nucl .
a * 715 * , 701c ( 2003 ) f. karsch , e.laermann , p. petreczky , s. stickan , and i. wetzorke , phys .
b * 530 * , 147 ( 2002 ) .
m. asakawa , t. hatsuda and y. nakahara , nucl .
a * 715 * , 863 ( 2003 ) t. umeda , k. nomura and h. matsufuru , hep - ph/0211003 .
i. wetzorke , hep - ph/0305012 .
( invited talk at the seventh workshop on quantum chromodunamics , villefranche - sur - mer , france , jan . 6 - 10 , 2003 ) p. petreczky , j. phys.g * 30 * , s431-s440 ( 2004 ) .
m. ruggieri , hep - ph/0310145 .
r. casalbuoni , r. gatto , g. nardulli , and m. ruggieri , phys .
d * 68 * , 034024 ( 2003 ) .
p. petreczky , private communication .
s. datta , f. karsch , p. petreczky and i. wetzorke , hep - lat/0312037 .
a. das , _ finite temperature field theory _
( world scientific , singapore , 1997 ) .
e. shuryak , hep - ph/0312227 .
r. l. kobes and g. w. semenoff , nucl .
b * 260 * , 714 ( 1985 ) . c. m. shakin , wei - dong sun , and j. szweda , ann . of phys .
( ny ) * 241 * , 37 ( 1995 ) .
hu li and c.m .
shakin , hep - ph/0209136 .
xiangdong li , hu li , c.m .
shakin , and qing sun , nucl - th/0405035 .
w. florkowski , acta phys .
b * 28 * , 2079 ( 1997 ) . | we have calculated spectral functions associated with hadronic current correlation functions for vector currents at finite temperature .
we made use of a model with chiral symmetry , temperature - dependent coupling constants and temperature - dependent momentum cutoff parameters .
our model has two parameters which are used to fix the magnitude and position of the large peak seen in the spectral functions . in our earlier work ,
good fits were obtained for the spectral functions that were extracted from lattice data by means of the maximum entropy method ( mem ) . in the present work
we extend our calculations and provide values for the three - momentum dependence of the vector correlation function at @xmath0 .
these results are used to obtain the correlation function in coordinate space , which is usually parametrized in terms of a screening mass .
our results for the three - momentum dependence of the spectral functions are similar to those found in a recent lattice qcd calculation for charmonium [ s. datta , f. karsch , p. petreczky and i. wetzorke , hep - lat/0312037 ] . for a limited range
we find the exponential behavior in coordinate space that is usually obtained for the spectral function for @xmath1 and which allows for the definition of a screening mass . |
the search of new gauge interactions has been of great interest .
such efforts help us understand whether there is any other new force in nature and how it fits to the grand picture of particle physics . if found , there is also a possibility that the new force carrier provides a portal between the standard model ( sm ) visible sector and a hidden sector involving new dynamics and matter contents . in the simplest scenario ,
such a new vector boson may have an origin from some extra @xmath4 gauge symmetry , under which some particles in both visible and hidden sectors are charged .
it may even have kinetic or mass mixing with the sm photon or @xmath5 boson field to facilitate the mediation .
after its symmetry breaking , as is often assumed , the @xmath4 gauge boson acquires a mass and is commonly called the @xmath6 boson if it is heavier than the electroweak scale or the dark photon if it is lighter .
there have been a vast amount of studies on the neutral gauge boson over the years @xcite . for example , recent direct searches at the lhc have already pushed the lower bound on the sequential @xmath6 mass to about 3 tev @xcite .
further investigations rely on advances in the colliding energy and beam luminosity .
on the other hand , probes of the dark photon of mass below sub - gev have been done using nuclear transitions @xcite or from its effects on the magnetic dipole moment of electron and/or muon @xcite .
the dark photon can be radiated off from some particle by bremsstrahlung and then decay into a pair of leptons in the fixed - target and beam - dump experiments @xcite . with an appropriate coupling , it can be produced at @xmath7 colliders as well .
if the dark photon couples to quarks , one can consider its production in meson decays , if kinematically allowed .
recently , there is an elevated interest in the study of dark photon because of an experimental anomaly involved in isoscalar @xmath3be transitions reported by krasznahorkay et .
al . @xcite . in the transition from an excited state to the ground state ,
the nucleus emitted an electron - positron pair whose open angle and invariant mass were found to deviate from the sm expectation of internal pair creation ( ipc ) by @xmath8 .
it was shown that the distributions of open angle and invariant mass could be well fit by introducing a new particle with mass 16.7 mev produced in the transition . in refs .
@xcite , the authors claimed that the new particle could be a vector boson @xmath9 and provided the preferred ranges of its couplings with sm particles that were consistent with current dark photon search constraints .
there are several analyses on how to further test the model in ref .
@xcite using low - energy physical processes @xcite , as well as proposals of alternative models for the @xmath3be anomaly @xcite . in this work
, we propose to use rare leptonic decays of kaon and pion , @xmath10 and @xmath11 , as felicitous means to probe the light dark photon in the mass range of about @xmath1 mev , particularly in view of the putative gauge boson @xmath9 with vectorial couplings mentioned above .
this is because the final - state electron - positron pair production can be enhanced via the mediation of the @xmath9 boson over the sm background through a virtual photon .
we show that the sm background and the signal have very different spectra in the @xmath7 invariant mass : the sm background has a continuous spectrum , whereas the decay through a dark photon features in a sharp resonance peak around the dark photon mass @xmath12 . moreover , the @xmath13 and @xmath14 decays are able to probe most part of the preferred coupling space inferred from the @xmath3be anomaly , assuming specific production rates of kaons / pions and their energy resolutions .
we provide the projected experimental limits on the dark photon couplings based on these decay processes .
we show that the signal events and sm background events have different behaviours in the @xmath15 angular distribution in the leptonic kaon decay .
we also discuss the influence of the kaon / pion structure dependence ( i.e. , the form factors ) , and find their effects almost irrelevant in current considerations .
this paper is organized as follows . in section [ sec : x ] , we review the leptonic meson decays of interest to us .
we present their decay amplitudes , including both inner bremsstrahlung and structure - dependent parts , and provide the form factors involved in the latter part . in section [ sec : scen ] , we consider three different dark photon scenarios , depending on how it couples to the sm particles , and three different experimental schemes for the kaon and pion decays .
we estimate the projected reach in the dark photon couplings in section [ sec : proj ] .
numerical results are presented in section [ sec : resu ] , where we also discuss effects coming from the structure - dependent contributions .
our findings are summarized in section [ sec : summary ] .
there can be many different ways to realize a light @xmath4 gauge boson and let it couple with sm particles ( directly or via mixing with the photon and/or @xmath5 boson ) . in the following
, we will keep the formalism as general as possible , without explicitly referring to any particular new physics model .
suppose the gauge coupling @xmath17 of the dark photon @xmath9 has a ratio of @xmath18 to the sm electric coupling , _
i.e. _ , @xmath19 .
we denote the coupling charge of electron , muon , @xmath20 quarks , and @xmath21 mesons respectively by @xmath22 , whose values depend on the model
. the dark photon @xmath9 can contribute to the @xmath23 decay , where the variables in the parentheses denote the momenta of the corresponding particles .
the radiative kaon decay involves both inner bremsstrahlung ( ib ) and structure - dependent ( sd ) parts , and the total decay amplitude is given by @xcite @xmath24 ~,\ ] ] where @xmath25 is the fermi decay constant , @xmath26 mev is the kaon decay constant , the cabibbo - kobayashi - maskawa ( ckm ) matrix element @xmath27 , and @xmath28\,,\end{aligned}\ ] ] with the @xmath29 part being due to ib , the @xmath30 part containing the sd form factors , @xmath31 , @xmath32 , and @xmath33 denoting the total width of @xmath9 .
the expressions in eqs .
( [ eq1 ] ) and ( [ eq2 ] ) can be readily modified to give those for the sm background by replacing symbols associated with the dark photon by those for photon .
we follow the convention in ref .
@xcite for the kaon form factors , which is consistent with the one used in ref .
, as explicitly shown in eq .
( c.8 ) of ref .
we set @xmath34 , and the rest form factors from recent measurements are parametrized as @xcite @xmath35 with @xmath36 , @xmath37 , @xmath38 , @xmath39 mev , @xmath40 mev , and @xmath41 mev . in the case of signal events ,
the distribution in the electron - positron invariant mass is a sharp resonance peak around @xmath12 .
the partial width of the dark photon decay into a pair of fermions is @xmath42\,,\ ] ] where @xmath43 , @xmath44 is the color factor , and @xmath45 , @xmath46 are the vector and axial - vector couplings between @xmath9 and @xmath47 . taking @xmath48 mev and assuming vectorial couplings as in refs .
@xcite , the dominant decay channel should be @xmath7 , @xmath49 and @xmath50 . for the sm background , one should set @xmath51 , @xmath52 , @xmath53 and replace @xmath54 by @xmath55\,.\ ] ] in this case , the electron - positron invariant mass distribution gives a continuous spectrum .
we also consider the pion decay @xmath56 .
similar to the kaon decay , the pion decay amplitude is @xcite @xmath57 where @xmath58 where @xmath59 mev , @xmath60 mev is the pion decay constant , the ckm matrix element @xmath61 , and the form factors @xcite @xmath62 with @xmath63 , @xmath64 , @xmath65 , and @xmath66 .
the ib term in the @xmath67 decay amplitude is proportional to @xmath68 , therefore the ib contribution is more important than that in the @xmath69 decay amplitude .
the sd contribution of the @xmath69 becomes more important at the kinematic regime of large @xmath7 invariant mass @xcite .
to explain the @xmath70 anomaly , it is sufficient for the dark photon to have couplings with only fermions in the first family while satisfying gauge anomaly cancellation @xcite .
the preferred ranges of dark photon couplings are found to be @xcite @xmath71 decays measured at the na48/2 experiment . if the dark photon also couples to the second family and we require gauge anomaly cancellation within each family , then the @xmath4 charges for muon and strange quark have either @xmath72 or @xmath73 .
we will thus consider the following three dark photon scenarios : * * scenario 1 * : only @xmath74 .
* * scenario 2 * : add @xmath75 and @xmath76 to scenario 1 . * * scenario 3 * : add @xmath77 and @xmath78 to scenario 1 . with @xmath79 , @xmath80 , @xmath81 , and @xmath82 ,
the @xmath4 charge of kaon is @xmath83 for scenario 1 , @xmath84 for scenario 2 , and @xmath85 for scenario 3 .
the @xmath4 charge of pion is @xmath86 for all three scenarios .
here we consider three different experimental schemes for the estimation of projected limits in the determination of dark photon gauge coupling .
suppose we produce @xmath87 kaons and pions , respectively , and measure the @xmath7 invariant mass from the decays with an energy resolution of @xmath88 .
the schemes are : * * scheme 1 * : @xmath89 , @xmath90 , and @xmath91 mev , * * scheme 2 * : @xmath92 , @xmath93 , and @xmath91 mev , * * scheme 3 * : @xmath92 , @xmath93 , and @xmath94 mev .
the energy resolution @xmath88 used here is much larger than the dark photon width for @xmath95 , as preferred by the @xmath70 anomaly @xcite .
therefore , we can not determine its width from the resonance peak in the @xmath7 invariant mass distribution .
but in contrast , sm background gives a continuous spectrum .
.[t1 ] branching ratios of the sm background in the range of @xmath96 , where @xmath48 mev , for different experimental schemes . [ cols="<,^,^,^",options="header " , ] in this section , we estimate the projected limits for the @xmath4 gauge coupling or , equivalently , the @xmath18 parameter for the above - mentioned three experimental schemes , following the steps outlined in ref .
@xcite .
first , we use the simple definition of standard deviation @xmath97 where @xmath98 is the number of signal events , and @xmath99 is that number of background events . in the following
, we will consider @xmath100 , corresponding to about @xmath101 confidence limit ( c.l . ) .
base on the experimental schemes , we write down the ratio of @xmath98 and @xmath99 in terms of the branching ratios @xmath102 where @xmath103 and @xmath104 are the branching ratios of the leptonic kaon ( or pion ) decay through @xmath9 and @xmath105 , respectively , with the requirement that the @xmath7 invariant mass falls within the energy range : @xmath96 . with a fixed @xmath12
, @xmath103 is proportional to @xmath106 due to the breit - wigner description and @xmath107^{-1 } \simeq \frac{\pi}{m_{x}\gamma_{x}}\delta(q'^2-m^2_{x})$ ] . combining eqs .
( [ eq10 ] ) and ( [ eq11 ] ) , we obtain the 2@xmath2 probing limit on the ratio @xmath108 where @xmath109 can be any reasonable value , as long as @xmath110 .
the branching ratios of sm background in various schemes are listed in table [ t1 ] . in fig .
[ f1 ] , we show @xmath111 versus @xmath12 in colored curves for the three experimental schemes from kaon ( left plot ) and pion ( right plot ) decays by fixing @xmath112 and @xmath113 ( _ i.e. _ , scenario 2 ) .
we also show in black curves the effect of form factors on @xmath111 by removing the sd part . for the dark photon mass of interest to the @xmath3be anomaly , @xmath114 mev ,
the effects from kaon or pion form factors are basically negligible
. projected limits on @xmath18 with @xmath115 ( about 95% c.l . ) with @xmath112 , and @xmath113 ( scenario 2 ) .
the left panel is for the @xmath116 decay , and the right panel is for the @xmath117 decay .
the red , green , and blue curves , including both ib and sd contributions , are for schemes 1 , 2 , and 3 , respectively . for each colored curve ,
the adjacent black curves include only the ib contribution . , title="fig:",height=211 ] projected limits on @xmath18 with @xmath115 ( about 95% c.l . ) with @xmath112 , and @xmath113 ( scenario 2 ) .
the left panel is for the @xmath116 decay , and the right panel is for the @xmath117 decay .
the red , green , and blue curves , including both ib and sd contributions , are for schemes 1 , 2 , and 3 , respectively . for each colored curve ,
the adjacent black curves include only the ib contribution .
, title="fig:",height=211 ] before concluding this section , we comment on the decay length of the dark photon as produced from the kaon or pion decay .
the decay length of @xmath9 produced from @xmath116 is @xmath118 , where @xmath119 is the boost factor , @xmath120 is the lifetime , and @xmath121 is the speed of light .
the largest boost factor from @xmath122 is @xmath123 .
taking the smallest value of @xmath124 for the @xmath3be anomaly , the largest decay length of @xmath9 with @xmath48 mev is @xmath125 therefore , the dark photon should decay within the detector . on the other hand , a decay length of @xmath9 of @xmath126 1 cm corresponds to @xmath127 .
contours of @xmath128 for scenario 1 on the @xmath129-@xmath130 plane .
@xmath131 comes from the experimental scheme 1 .
the light green area is the region favored by the @xmath3be anomaly @xcite .
the black dotted line is for @xmath132 .
the colored lines indicate the contours for various values of @xmath128 ; for example , the red contour is for @xmath133 .
, height=336 ] the 2@xmath2 projected limits from the @xmath134 decay under the experimental scheme 1 ( red ) , scheme 2 ( green ) , and scheme 3 ( blue ) and for scenario 1 ( first row ) , scenario 2 ( second row ) , and scenario 3 ( third row ) . the light green area is the coupling region favored by the @xmath3be anomaly @xcite .
depending on the schemes , the region outside the red , green , or blue areas are experimentally probe - able .
the back dotted lines in the left - column plots indicate the case with @xmath135 .
the grey regions in the middle - and right - column plots indicate the possibility that the @xmath9 has a decay length larger than 1 cm .
, title="fig:",height=144 ] the 2@xmath2 projected limits from the @xmath134 decay under the experimental scheme 1 ( red ) , scheme 2 ( green ) , and scheme 3 ( blue ) and for scenario 1 ( first row ) , scenario 2 ( second row ) , and scenario 3 ( third row ) . the light green area is the coupling region favored by the @xmath3be anomaly @xcite .
depending on the schemes , the region outside the red , green , or blue areas are experimentally probe - able .
the back dotted lines in the left - column plots indicate the case with @xmath135 .
the grey regions in the middle - and right - column plots indicate the possibility that the @xmath9 has a decay length larger than 1 cm .
, title="fig:",height=144 ] the 2@xmath2 projected limits from the @xmath134 decay under the experimental scheme 1 ( red ) , scheme 2 ( green ) , and scheme 3 ( blue ) and for scenario 1 ( first row ) , scenario 2 ( second row ) , and scenario 3 ( third row ) . the light green area is the coupling region favored by the @xmath3be anomaly @xcite .
depending on the schemes , the region outside the red , green , or blue areas are experimentally probe - able .
the back dotted lines in the left - column plots indicate the case with @xmath135 .
the grey regions in the middle - and right - column plots indicate the possibility that the @xmath9 has a decay length larger than 1 cm .
, title="fig:",height=144 ] the 2@xmath2 projected limits from the @xmath134 decay under the experimental scheme 1 ( red ) , scheme 2 ( green ) , and scheme 3 ( blue ) and for scenario 1 ( first row ) , scenario 2 ( second row ) , and scenario 3 ( third row ) . the light green area is the coupling region favored by the @xmath3be anomaly @xcite .
depending on the schemes , the region outside the red , green , or blue areas are experimentally probe - able .
the back dotted lines in the left - column plots indicate the case with @xmath135 .
the grey regions in the middle - and right - column plots indicate the possibility that the @xmath9 has a decay length larger than 1 cm .
, title="fig:",height=144 ] the 2@xmath2 projected limits from the @xmath134 decay under the experimental scheme 1 ( red ) , scheme 2 ( green ) , and scheme 3 ( blue ) and for scenario 1 ( first row ) , scenario 2 ( second row ) , and scenario 3 ( third row ) .
the light green area is the coupling region favored by the @xmath3be anomaly @xcite .
depending on the schemes , the region outside the red , green , or blue areas are experimentally probe - able .
the back dotted lines in the left - column plots indicate the case with @xmath135 .
the grey regions in the middle - and right - column plots indicate the possibility that the @xmath9 has a decay length larger than 1 cm .
, title="fig:",height=144 ] the 2@xmath2 projected limits from the @xmath134 decay under the experimental scheme 1 ( red ) , scheme 2 ( green ) , and scheme 3 ( blue ) and for scenario 1 ( first row ) , scenario 2 ( second row ) , and scenario 3 ( third row ) . the light green area is the coupling region favored by the @xmath3be anomaly @xcite .
depending on the schemes , the region outside the red , green , or blue areas are experimentally probe - able .
the back dotted lines in the left - column plots indicate the case with @xmath135 .
the grey regions in the middle - and right - column plots indicate the possibility that the @xmath9 has a decay length larger than 1 cm .
, title="fig:",height=144 ] the 2@xmath2 projected limits from the @xmath134 decay under the experimental scheme 1 ( red ) , scheme 2 ( green ) , and scheme 3 ( blue ) and for scenario 1 ( first row ) , scenario 2 ( second row ) , and scenario 3 ( third row ) . the light green area is the coupling region favored by the @xmath3be anomaly @xcite .
depending on the schemes , the region outside the red , green , or blue areas are experimentally probe - able .
the back dotted lines in the left - column plots indicate the case with @xmath135 .
the grey regions in the middle - and right - column plots indicate the possibility that the @xmath9 has a decay length larger than 1 cm .
, title="fig:",height=144 ] the 2@xmath2 projected limits from the @xmath134 decay under the experimental scheme 1 ( red ) , scheme 2 ( green ) , and scheme 3 ( blue ) and for scenario 1 ( first row ) , scenario 2 ( second row ) , and scenario 3 ( third row ) . the light green area is the coupling region favored by the @xmath3be anomaly @xcite .
depending on the schemes , the region outside the red , green , or blue areas are experimentally probe - able .
the back dotted lines in the left - column plots indicate the case with @xmath135 .
the grey regions in the middle - and right - column plots indicate the possibility that the @xmath9 has a decay length larger than 1 cm .
, title="fig:",height=144 ] the 2@xmath2 projected limits from the @xmath134 decay under the experimental scheme 1 ( red ) , scheme 2 ( green ) , and scheme 3 ( blue ) and for scenario 1 ( first row ) , scenario 2 ( second row ) , and scenario 3 ( third row ) .
the light green area is the coupling region favored by the @xmath3be anomaly @xcite .
depending on the schemes , the region outside the red , green , or blue areas are experimentally probe - able .
the back dotted lines in the left - column plots indicate the case with @xmath135 .
the grey regions in the middle - and right - column plots indicate the possibility that the @xmath9 has a decay length larger than 1 cm .
, title="fig:",height=144 ] same as fig .
[ f4 ] , but with only the ib contribution included .
, title="fig:",height=144 ] same as fig .
[ f4 ] , but with only the ib contribution included .
, title="fig:",height=144 ] same as fig .
[ f4 ] , but with only the ib contribution included .
, title="fig:",height=144 ] same as fig .
[ f4 ] , but with only the ib contribution included .
, title="fig:",height=144 ] same as fig .
[ f4 ] , but with only the ib contribution included .
, title="fig:",height=144 ] same as fig .
[ f4 ] , but with only the ib contribution included .
, title="fig:",height=144 ] same as fig .
[ f4 ] , but with only the ib contribution included .
, title="fig:",height=144 ] same as fig .
[ f4 ] , but with only the ib contribution included .
, title="fig:",height=144 ] same as fig .
[ f4 ] , but with only the ib contribution included .
, title="fig:",height=144 ] same as fig .
[ f3 ] , but for the @xmath136 decay .
the back dotted lines of the left - column plots indicate the case with @xmath137 .
, title="fig:",height=144 ] same as fig .
[ f3 ] , but for the @xmath136 decay .
the back dotted lines of the left - column plots indicate the case with @xmath137 .
, title="fig:",height=144 ] same as fig .
[ f3 ] , but for the @xmath136 decay .
the back dotted lines of the left - column plots indicate the case with @xmath137 .
, title="fig:",height=144 ] same as fig .
[ f3 ] , but for the @xmath136 decay .
the back dotted lines of the left - column plots indicate the case with @xmath137 .
, title="fig:",height=144 ] same as fig .
[ f3 ] , but for the @xmath136 decay .
the back dotted lines of the left - column plots indicate the case with @xmath137 . , title="fig:",height=144 ] same as fig .
[ f3 ] , but for the @xmath136 decay .
the back dotted lines of the left - column plots indicate the case with @xmath137 . , title="fig:",height=144 ] same as fig .
[ f3 ] , but for the @xmath136 decay .
the back dotted lines of the left - column plots indicate the case with @xmath137 .
, title="fig:",height=144 ] same as fig .
[ f3 ] , but for the @xmath136 decay .
the back dotted lines of the left - column plots indicate the case with @xmath137 . ,
title="fig:",height=144 ] same as fig .
[ f3 ] , but for the @xmath136 decay .
the back dotted lines of the left - column plots indicate the case with @xmath137 .
, title="fig:",height=144 ] since the best - fit mass of dark photon for the @xmath3be anomaly is 16.7 mev @xcite , we will fix @xmath48 mev in the following discussions . refs .
@xcite worked out the ranges of dark photon couplings with neutron , proton , and electron that could explain the @xmath3be anomaly and evade other experimental constraints .
we will show how the leptonic kaon and pion decays can probe this region .
first , we take the experimental scheme 1 and consider the kaon decay in scenario 1 as an example .
the sm background branching ratio is @xmath138 .
then the signal branching ratio corresponding to @xmath139 upper limit of @xmath140 is @xmath141 using @xmath131 as a basic unit , we draw the contours of @xmath128 on the @xmath142 parameter plane in fig . [ f2 ] .
the black dotted line in the plot indicates @xmath132 and thus @xmath143 .
the region with @xmath144 is between the two red lines .
this is the region where the dark photon in scenario 1 can not be checked using the leptonic kaon decay under the experimental scheme 1 .
however , all the region outside the red contour , including most of the light green region , can be probed .
we perform a scan over the dark photon couplings , @xmath145 . in fig .
[ f3 ] , we show the @xmath139 projected limits for scenarios 1 , 2 , and 3 in experimental schemes 1 , 2 , and 3 . most part of the light green area , outside the red , green , or blue regions , can be probed by the @xmath134 decay . in the left column ,
the black dotted lines indicate the case with @xmath135 . in this case
, the leptonic kaon decay loses sensitivity to probing the dark photon . in the right column
, the leptonic kaon decay also has less sensitivity to the dark photon along the semi - major axis of the ellipses .
this is because when @xmath146 , a cancellation occurs in the ib amplitude and , therefore , @xmath147 is suppressed .
[ f4 ] is the same as fig .
[ f3 ] , except that we turn off the sd contribution in the amplitude of leptonic kaon decay .
there is almost no noticeable numerical difference between fig .
[ f3 ] and fig .
this demonstrates that uncertainties from the kaon form factors have little effects on the analysis , especially for @xmath148 mev . in fig .
[ f5 ] , we show the @xmath139 projected limits from the @xmath149 decay as another possible probe . comparing to fig .
[ f3 ] , the projected limits from the leptonic pion decay are not as good . but a good portion of the preferred parameter region for the @xmath3be anomaly ( light green area ) can still be probed .
angular distribution of the @xmath150 decays , where @xmath151 denotes the angle of @xmath152 with respect to the 3-momentum of @xmath9 , but measured in the @xmath7 center - of - mass frame .
the @xmath153 is the ratio between the number of events within each bin to the total number of events .
virtually no difference can be seen among the four scenarios .
, height=336 ] the @xmath154 distribution of the @xmath150 decays , where @xmath155 denotes the angle of @xmath156 with respect to the @xmath15 3-momentum sum , but measured in the @xmath15 center - of - mass frame .
both ib and sd contributions are included in the calculations .
, title="fig:",height=211 ] the @xmath154 distribution of the @xmath150 decays , where @xmath155 denotes the angle of @xmath156 with respect to the @xmath15 3-momentum sum , but measured in the @xmath15 center - of - mass frame .
both ib and sd contributions are included in the calculations .
, title="fig:",height=211 ] we now turn to the discussions of angular distributions of different subsystems in the @xmath157 final state .
first , we show in fig .
[ f6 ] the distribution of @xmath158 from the leptonic kaon decay , where @xmath151 denotes the angle of @xmath152 with respect to the boost direction of @xmath9 , but measured in the @xmath7 center - of - mass frame .
the parabolic shape of the angular distribution signifies the feature of vectorial couplings of the @xmath9 boson and the photon , and it is almost independent of @xmath4 charges and the sd part . in fig .
[ f7 ] , we show the distribution of @xmath154 , where @xmath155 is the angle of @xmath156 with respect to the boost direction of the @xmath15 system , but measured in the @xmath15 center - of - mass frame .
apparently , the sm background and signal from dark photon have different behaviours . from the sm background
, the @xmath156 tends to fly toward the direction of @xmath9 in the @xmath116 rest frame . for the signal events
, however , it has an isotropic distribution . in the left plot , when @xmath159 and @xmath160 , we are left with only the first term in the curly brackets of @xmath29 in eq .
( [ eq2 ] ) that has no angular dependence .
when @xmath132 and @xmath161 , on the other hand , the second term in the curly brackets survives and has some minor angular dependence . when @xmath162 , the angle - independent parts cancel , rendering a more dramatic angular dependence , as shown by the blue - dotted histogram .
the same cancellation also happens for the sm background , as the electric charges of @xmath116 and @xmath163 have opposite signs .
the result is shown by the black - dotted histogram in the right plot .
in this work , we propose to use the rare leptonic kaon and pion decays , @xmath164 to probe a light dark photon of @xmath1 mev mass .
this is particularly suitable for probing the putative @xmath9 gauge boson hinted at by the recent @xmath3be anomaly .
we consider three scenarios for the standard model particle couplings with the dark photon .
the first scenario assumes that the dark photon only couples to fermions in the first family , as directly relevant for the @xmath3be anomaly . in the second and third scenarios ,
the dark photon is allowed to also couple with fermions in the second family , with equal or opposite signs to the couplings in the first family .
we also consider three schemes for the estimation of experimental reach .
we estimate the projected limits by calculating the numbers of events from both dark photon and sm background .
we perform a scan of dark photon couplings , @xmath145 , and compare the results with the region favored by the @xmath3be anomaly .
moreover , we show the angular distributions of final - state @xmath7 and @xmath15 systems in their own center mass frames . in general , scheme 1 has the best sensitivity to probe the dark photon couplings due to a larger number of events as well as a better energy resolution .
most of the parameter space preferred by the @xmath3be anomaly can be probed by both @xmath10 and @xmath11 decays .
the effects from kaon and pion form factors are found to be negligible in the projected limits for @xmath48 mev . in each specific experimental scheme , the projected limits from the kaon decay are stronger than those from the pion decay . except for some special case where @xmath116 and @xmath156 have opposite couplings with the dark photon ,
the signal events and the sm background events have different behaviours in the @xmath15 angular distributions .
the @xmath7 angular distribution can directly reveal the vectorial nature of the dark photon interaction .
c.w.c would like to thank the hospitality of kavli ipmu during his visit in 2016 summer , where this project was initiated .
this research was supported in part by the ministry of science and technology of taiwan under grant no .
nsc 104 - 2628-m-002 - 014-my4 , and in part by world premier international research center initiative ( wpi ) , mext , japan . | rare leptonic kaon and pion decays @xmath0 can be used to probe a dark photon of mass @xmath1 mev , with the background coming from the mediation of a virtual photon .
this is most relevant for the 16.7-mev dark photon proposed to explain a 6.8@xmath2 anomaly recently observed in @xmath3be transitions by the atomki collaboration .
we evaluate the reach of future experiments for different scenarios of how the dark photon couples with the standard model particles , and show that a great portion of the preferred 16.7-mev dark photon parameter space can be decisively probed .
we also show the use of angular distributions to further distinguish the signal from the background . |
there has been conclusive evidence indicating that our universe is experiencing an accelerated expansion .
this acceleration is believed driven by a so called dark energy ( de ) in the framework of einstein s general relativity .
the simplest explanation of such de is the cosmological constant .
but it suffers serious problems such that its value is far below the prediction of any sensible quantum field theories and it inevitably sustains the coincidence problem with the mysterious same order energy density as the matter field today .
there exists an alternative way to explain the acceleration of the universe expansion by modifying the einstein s gravity .
one simplest attempt is called @xmath0 gravity , in which the scalar curvature in the lagrange density of einstein s gravity is replaced by an arbitrary function of @xmath2 @xcite .
however it is quite non - trivial to construct the viable @xmath0 model which can satisfy both cosmological and local gravity constraints @xcite .
only a few viable @xmath0 models with simple analytic forms have been found @xcite .
however , if we do not limit to these @xmath0 models with explicit expressions , in a more general case , @xmath0 gravity in cosmology can leave enough freedoms to accommodate any desired expansion histories of the universe @xcite @xcite .
a general phenomenological @xmath0 cosmological model were formulated numerically by specifying the effective de equation of state @xmath3 and the present value of the parameter @xmath4
@xcite .
on the other hand , it is always possible to transform the theory of @xmath0 gravity presented in the jordan frame to the einstein frame by using the conformal transformation @xcite . in the einstein frame ,
the model contains the coupling between the matter and an additional scalar field @xcite .
if the scalar field plays a role of de in the universe , de would interact with dm .
the phenomenological de and dm interacting model has been studied extensively in literatures , see @xcite and references therein .
more interestingly , it was found that a general @xmath0 gravity can be consistently constructed in a covariant form in terms of the mass dilation rate function in the einstein frame@xcite .
this mass dilation rate function marks the coupling strength between de and dm .
the condition for the @xmath0 cosmology to avoid the instability at high - curvature region and be consistent with cmb observation require the energy flow from de to dm in the framework of the interacting model @xcite .
this was found helpful to alleviate the coincidence problem@xcite . in this paper
, we will further examine the equivalence of the @xmath0 gravity presented in the jordan frame and the einstein frame to describe the cosmology . in the background jordan frame
, we will derive the analytic @xmath0 functions to give the same expansion history of the universe as the @xmath1cdm model .
this can provide further insights on how cosmology can be used to test the @xmath0 gravity at the largest scales .
in addition to modeling @xmath0 gravity in jordan frame using the effective de equation of state , we will present a systematic and self - consistent way to construct the @xmath0 model in terms of the mass dilation rate function in einstein frame .
moreover we will develop linear perturbation theories to further exhibit the equivalence of the @xmath0 gravity in the jordan frame and the einstein frame .
this paper is orgnized as follows : in section [ background ] , we will discuss the @xmath0 gravity in the background cosmology . in the jordan frame
, we will derive the general analytical solutions for the @xmath0 model which has the same expansion history as the @xmath1cdm cosmology . in the einstein frame
, we will argue that once the mass dilation is specified , the @xmath0 gravity can be constructed accordingly . in section [ sectionperturbedspacetime ] , we will extend our discussion to the perturbed space - time .
we will show that perturbation theories in jordan frame and einstein frame can be consistently connected by conformal transformations . in section [ subhoizion ]
, we will construct a viable @xmath0 cosmological model in terms of mass dilation rate specified in the einstein frame and discuss the subhorizon approximation of the constructed @xmath0 model
. our conclusion and discussion will be presented in the last section .
we start with the 4-dimensional action in @xmath0 gravity @xcite @xmath5 where @xmath6 . the field equation can be obtained by varying the above action with respect to @xmath7 , which leads to @xmath8 where @xmath9 and @xmath10 if we define the left hand side of eq .
( [ frfield ] ) as a tensor @xmath11 @xmath12 we can recast the modified einstein s equation into a similar form to the standard einstein equation @xmath13 the trace of the above equation gives rise to a scalar equation
@xmath14 taking the derivative of @xmath15 and further noting that @xmath16 we obtain @xmath17 we can see that @xmath18 is a free scalar , which can be used to characterize the @xmath0 gravity .
in the jordan frame , we consider the universe which is described by the flat friedmann - robertson- walker(frw ) metric @xmath19 from eq.([einsteineq ] ) , we can obtain the equation governing the dynamics in the background spacetime @xmath20 in the jordan frame , the matter field is conserved , which satisfies @xmath21 where dot denotes the derivative with respect to the cosmic time @xmath22 . from eq.([field ] ) , @xmath15 can be presented as @xmath23 where @xmath24 .
taking the derivative of the above equation and further noting that @xmath25 , we obtain @xmath26 where @xmath27 @xmath2 is in the unit of @xmath28 and @xmath29 . for illustrative purposes ,
we take an expansion history in the jordan frame that matches a de model with equation of state @xmath3 , @xmath30 in @xcite a family of numerical solutions of @xmath0 models that match the representative expansion histories of de model was obtained . here
we find that for the @xmath0 model to mimic the same expansion history as that of the @xmath1cdm cosmology in the background @xmath31 , we can analytically solve eq.([f ] ) and obtain the general analytical solution @xmath32\nonumber\\ & & + d(e^{3x})^{p_+}(1-\omega_m^0)^{p_+}{(\omega_m^0)}^{-p_+}{_2f_1}[q_+,p_+,r_+,z ] \nonumber\quad,\\\label{solution}\end{aligned}\ ] ] where @xmath33 is the hypergeometric function , which is defined as @xmath34 @xmath35 and @xmath36 are coefficients which will be determined by boundary conditions .
the indexes in the solution read , @xmath37 and @xmath38 since the index @xmath39 is negative , the fist term of eq.([solution ] ) will diverge when @xmath40 goes to the minus infinity , which is not acceptable in physics . besides , a viable @xmath0 model should be in a form of chameleon " type @xcite , which means that in the high curvature region it should go back to the standard einstein gravity to pass the local solar - system tests .
the boundary condition for the scalar field @xmath41 thus requires @xmath42 and we need to set @xmath43 .
the solution turns out to be @xmath44\label{fsolution}\quad.\ ] ] @xmath36 above is a free parameter which characterizes the different @xmath0 models which have the same expansion history as that of the @xmath1cdm model . with the solution of @xmath41 , we can easily figure out the analytical expression for @xmath0 by doing the integration @xmath45 we obtain @xmath46(1-\epsilon)+{\rm constant}\label{finallfr}\quad , \ ] ] where @xmath47}{\gamma_{e}[p_+]}(1-\omega_m^0)^{p_+}(\omega_m^0)^{-p_+}(e^{3x})^{p_+}{_2f_1}[q_+,s_+,r_+,z]\nonumber\quad,\ ] ] and @xmath48 @xmath49 is the complete euler gamma function , which is defined as @xmath50 when @xmath51 , eq.([finallfr ] ) should be consistent with eq.([fhubble ] ) and should go back to the @xmath1cdm model .
this puts the constant in eq.([finallfr ] ) into @xmath52 .
thus finally we have the analytic @xmath0 form @xmath46(1-\epsilon)+6(1-\omega_m^0)\label{finallfrwithconstant}\quad .\ ] ] models that have the same expansion history as the @xmath1cdm cosmology.,width=336,height=307 ] eq.([finallfrwithconstant ] ) is a general analytical expression for @xmath0 models that can have the same background expansion history as that of the @xmath1cdm cosmology .
the solution has only two free dimensionless parameters , namely , @xmath53 and @xmath36 . in order to avoid the short - timescale instability at high curvature @xcite
, it requires @xmath54 .
moreover for a viable @xmath0 model , it also requires @xmath55 @xcite . in fig [ fro ] , we illustrate the @xmath0 models obtained from eq.([finallfrwithconstant ] ) and the blues lines indicate the viable @xmath0 models which satisfy the condition mentioned above .
if we do not start from the concrete model describing the expansion history of the universe , there is a lot of freedom for choosing the scalar field @xmath41 except the boundary condition eq.([boundary ] ) .
the @xmath0 gravity can be constructed only when the scalar field @xmath41 can be reasonably chosen based on physical motivations . however , in the jordan frame , there is no clear physical motivation to provide a particular form for the field @xmath41 .
we need to turn to the einstein frame . in the einstein frame
, @xmath41 relates to a scalar @xmath56 representing the mass dilation rate @xmath57 in gravitational field @xcite and @xmath56 can be deemed as the coupling strength in the framework of interacting dark energy models @xcite .
@xmath41 can be determined by constructing @xmath58 which , in turn , the @xmath0 model can be constructed . in the next section
we will shift our attention to the einstein frame .
the equations of motion in the einstein frame can be obtained by using conformal transformations @xmath59 as shown in @xcite , after the conformal transformation the expansion history in the einstein frame is governed by @xmath60 where @xmath58 is the coupling strength between the scalar field and matter which can also be explained as the mass dilation rate function @xcite . from the relation between @xmath18 and @xmath58 and expressing @xmath18 into @xmath61 , @xmath62 , we can rewrite the continuity equation for the scalar field into @xmath63 once the dilation function is specified , we can explore the expansion history of the universe in the einstein frame .
@xmath58 has clear physical meaning , which represents the coupling strength of the scalar field to the matter field .
although @xmath58 is a free field , the choice of @xmath58 must be consistent with the viability condition of the @xmath0 gravity .
a viable @xmath0 model must pass the stringent local test , which requires that when @xmath64 , the model should go back to the standard einstein gravity , so that @xmath65 .
eqs.([1])-([3 ] ) governing the expansion history are not stable when @xmath66 at high curvature region , except that the coupling strength is extremely small , for example @xmath67 .
this phenomena is consistent with that discussed in @xcite .
thus a viable @xmath0 model requires that @xmath58 is positive and drops to sufficiently small values in the early time of the universe .
@xmath58 also is the dilation rate of the mass of test particles in gravitational fields . to satisfy the weak equivalent principle
, @xmath58 should be independent of the species of matter @xcite .
one natural choice of the @xmath58 form is to consider it as a geometrical quantity , a function of @xmath68 .
there is quite a wide range of the choices for @xmath58 which can satisfy the above conditions . in this work , we only present one of the simplest choices , where we describe @xmath58 as a function of the scalar ricci curvature @xmath68 in the einstein frame as @xmath69 where @xmath70 and @xmath71 are constants .
eq.([model ] ) is just one kind of possible phenomenological descriptions since we are lack of the knowledge on the nature of the dilation rate .
once we specify the dilation rate @xmath58 , we can get the @xmath0 gravity in the jordan frame by using the conformal transformation from the einstein frame to the jordan frame .
in this section , we will extend our discussion to the perturbed space - time .
we will first review the cosmological perturbation theory in @xmath0 gravity in the jordan frame .
we will show that perturbations can be consistently connected between the jordan frame and the einstein frame .
we focus on the scalar perturbation .
the perturbed line element in fourier space reads , @xmath72\label{pert_jordan}\quad,\end{aligned}\ ] ] where @xmath73 are scalar fields which indicate the perturbations .
@xmath74 is the scalar harmonic function .
the perturbed form of the modified einstein equations eq.([einsteineq ] ) @xmath75 can be obtained by writing @xmath76 where @xmath77 and expressing @xmath78 inserting the line element eq.([pert_jordan ] ) into the above equations , we can get the perturbation form of the modified einstein equation . from @xmath79
we obtain , @xmath80\nonumber \\ & + & f'(\frac{k}{2}b+3\mathcal{h}\psi-\frac{3}{2}\phi')-\frac{3}{2}\mathcal{h}^2\delta f\nonumber\\ & + & \frac{3}{2}\frac{a''}{a}\delta f-\frac{3}{2}\mathcal{h}\delta f'-\frac{k^2}{2}\delta f\label{je_1}\quad,\end{aligned}\ ] ] where the prime denotes the derivative with respect to the conformal time @xmath81 .
the trace part @xmath82 gives @xmath83\nonumber \\ & + & f'[\frac{2}{3}kb+\psi'+2\mathcal{h}\psi-2\phi']\nonumber\\ & + & \delta f[\mathcal{h}^2+\frac{a''}{a}-\frac{2}{3}k^2]-\delta f''-\delta f'\mathcal{h}+2\psi f''\nonumber\quad,\\\label{je_2}\end{aligned}\ ] ] and the anisotropic part @xmath84 ( @xmath85 ) gives @xmath86+f'(\frac{1}{2}e'+kb)-k^2\delta f.\nonumber\\\label{je_3}\end{aligned}\ ] ] the last part @xmath87 leads to @xmath88+f'[\mathcal{h}b-\frac{1}{2}k\psi]\nonumber \\ & + & \frac{1}{2}k\delta f'-\frac{1}{2}k\mathcal{h}\delta f-\frac{1}{2}b f''.\nonumber\\\label{je_4}\end{aligned}\ ] ] the above perturbation equations are covariant .
this can be easily seen by investigating the infinitesimal coordinate transformation , @xmath89 we use the symbol with hat " to indicate the quantities after coordinate transformation . under the infinitesimal transformation
, we can show that the perturbed quantities in the line element eq .
( [ pert_jordan ] ) behave as @xcite @xmath90 after inserting eq .
( [ transjordan ] ) into eqs .
( [ je_1])-([je_4 ] ) , we find that eqs .
( [ je_1])-([je_4 ] ) keep the same forms under the infinitesimal transformation , which shows that eqs .
( [ je_1])-([je_4 ] ) are covariant . furthermore , if we take @xmath91,@xmath92 , eqs .
( [ je_1])-([je_4 ] ) can go back to the standard cosmological perturbation equations in the einstein gravity as presented in @xcite . after we get the perturbed form of the modified einstein equations , we turn to derive the equation of motion for matter field in the jordan frame .
from @xmath93 we can obtain its perturbed form @xmath94 .
looking at the zero - th component , we have @xmath95 the i - th component gives @xmath96'+4\mathcal{h}(p_{m}+\rho_{m})(b+v_{m } ) \nonumber \\ & = & ( p_{m}+\rho_{m})k\psi+k\delta p_{m}-\frac{2}{3}kp_m\pi_m\quad.\label{eqv}\end{aligned}\ ] ] similarly , it can be shown that the above equations are covariant under the infinitesimal coordinate transformation .
the freedom of choosing coordinates in general relativity will lead to ambiguity in defining a real density perturbation from the coordinate transformation .
the systematic way to get rid of the ambiguity is to fix the gauge , which has been well discussed in @xcite . here
we will argue that the gauge conditions in the jordan frame are not always the same as the conditions in the einstein frame after the conformal transformation .
we only focus our discussion on the most commonly used gauges , the newtonian gauge and the synchronous gauge .
the newtonian gauge is defined by setting @xmath97 and these conditions completely fix the gauge @xmath98 the perturbations in this gauge can be shown as gauge invariant .
for instance , the matter density perturbation can be presented as @xmath99 where @xmath100 is a general perturbation .
we can show that @xmath101 is invariant under the coordinate transformation eq .
( [ transjordan ] ) .
however , in contrast to the newtonian gauge , the synchronous gauge is not completely fixed because the gauge condition @xmath102 only confines the gauge up to two arbitrary constants @xmath103 , @xmath104 in order to fix the synchronous gauge , additional conditions are called for .
usually , @xmath105 is fixed by specifying the initial condition for the curvature perturbation in the early time of the universe and @xmath106 can be fixed by setting the peculiar velocity of dm to be zero , @xmath107 , which is equivalent to say that the gauge is at rest with respect to the cold dm and the observer is comoving with cold dm . after fixing @xmath103 ,
the synchronous gauge can be completely fixed .
as long as the synchronous gauge is completely fixed , it will be equivalent to the newtonian gauge . for instances , the density perturbations in different gauges
are related by @xmath108 where @xmath109,@xmath110 .
we will show later that the gauge condition for the newtonian gauge will be the same both in the jordan frame and the einstein frame .
however we will see that the gauge condition for the synchronous gauge will change when we transform the jordan frame to the einstein frame by using the conformal transformation .
since the newtonian gauge has this advantage , we will work in the newtonian gauge hereafter .
we use the bardeen potentials @xcite @xmath111 , @xmath112 to represent the space time perturbations . since we are only interested in the dm dominated period in the @xmath0 cosmology , we set @xmath113 and @xmath114 .
the perturbed einstein equations can be reduced to @xmath115\nonumber \\ & + & f'(3\mathcal{h}\psi-\frac{3}{2}\phi')-\frac{3}{2}\mathcal{h}^2\delta f\nonumber\\ & + & \frac{3}{2}\frac{a''}{a}\delta f-\frac{3}{2}\mathcal{h}\delta f'-\frac{k^2}{2}\delta f\label{jen_1}\quad , \\
\delta f''&=&-\mathcal{h}\delta f'+f[-\frac{2}{3}k^2\psi-\frac{2}{3}k^2\phi+4\frac{a''}{a}\psi\nonumber \\ & -&2\mathcal{h}^2\psi+2\mathcal{h}\psi'-4\mathcal{h}\phi'-2\phi'']+2\psi f''\nonumber \\ & + & f'[\psi'+2\mathcal{h}\psi-2\phi']\nonumber \\ & + & \delta f[\mathcal{h}^2+\frac{a''}{a}-\frac{2}{3}k^2]\label{jen_2}\quad , \\ \frac{\delta f}{f}&=&-\psi-\phi\label{jen_3 } \quad , \\
-\frac{1}{2}\kappa^2a^2\rho_mv_m&=&f[k\phi'-k\mathcal{h}\psi]-\frac{1}{2}kf'\psi\nonumber \\ & + & \frac{1}{2}k\delta f'-\frac{1}{2}k\mathcal{h}\delta f\quad.\label{jen_4}\end{aligned}\ ] ] @xmath18 is a free scalar field which characterizes the deviation from the einstein gravity .
the perturbation @xmath116 can be obtained by @xmath117 where @xmath118 using eq.([rj ] ) and eq.([jen_1 ] ) to eliminate @xmath119 and @xmath120 in eq.([jen_2 ] ) , we obtain @xmath121 where @xmath122 and @xmath123 eq.([deltf ] ) is consistent with eq.(8.92 ) in @xcite , if we express it in the cosmic time @xmath22 .
> from the equations of motion eqs.([eqm ] ) ( [ eqv ] ) , we obtain @xmath124 combining eqs.([mp_jordan ] , [ v_jordan ] ) , we have the second order differential equation for the evolution of the matter perturbation @xmath125 we will first derive the perturbed form of the equation of motion for matter fields eq.([eqofmotion ] ) under the general conformal transformation .
then we will fix the freedom of the general conformal transformation by setting @xmath126 to obtain the perturbation form of the modified einstein equations .
we will show that the equations obtained under conformal transformation from the jordan frame are consistent with the results derived by doing perturbations directly in the einstein frame . in the background
, the conformal transformation is defined by re - scaling the line element of the jordan frame @xmath127 however , in perturbed spacetime , we need to take account of the perturbation of @xmath128 .
and thus the full line element reads @xmath129\quad .\label{perturbedspacetime}\end{aligned}\ ] ] since we are only interested in the linear perturbation , we can expand the above expression to the linear order @xmath130 where the symbols with
tilde " indicate the quantities in the einstein frame . for the clear discussion , hereafter we will use tilde " to indicate quantities that are different in the einstein frame from their values in the jordan frame .
for the quantities without tilde " , they are the same in both frames .
we will use hat " to indicate the coordinate transformation which should not be confused with the conformal transformation .
the differential operator @xmath131 does not depend on the geometric structure @xmath132 on the spacetime manifold , it will not change under the conformal transformation so that the ordinary derivative operator @xmath133 and the conformal coordinates @xmath134 will be the same in both the einstein frame and the jordan frame . noting @xmath135 and considering eqs.([trans],[transjordan ] )
, we can show that , under the infinitesimal coordinate transformation , the perturbed quantities @xmath136 , @xmath137 behave as @xmath138 where @xmath139 it is clear that in the coordinate transformation @xmath136 , @xmath137 behave in a similar way as in the jordan frame .
this point also holds for other perturbation quantities @xmath140 now we can discuss the gauge conditions in the einstein frame . in the newtonian gauge , the gauge conditions @xmath97 fix the gauge as that in the jordan frame , namely @xmath141 thus the newtonian gauge conditions keep the same under the conformal transformation .
however , in the synchronous gauge , after the conformal transformation , the gauge conditions in the einstein frame read @xmath142 we will find that @xmath107 is a physical choice because it is the solution of the equation of motion in eq.([v_ein ] ) .
it is clear that in the einstein frame the gauge conditions for the synchronous gauge are not exactly the same as those in the jordan frame .
considering the gauge conditions for the newtonian gauge are the same in both the einstein frame and the jordan frame , we will work in newtonian gauge hereafter . in the background , the conformal transformation leads the equation of motion of the matter field eq.([eqofmotion ] ) in the einstein frame to the form @xmath143 where @xmath144 , @xmath145 and @xmath146 @xmath147 is the coupling vector .
we only consider the evolving universe where the coupling vector is time - like .
we focus on the dm dominated phase with @xmath113 .
the conservation equation eq.([efm ] ) can be recasted into @xmath148 where @xmath149 , @xmath58 is the mass dilation rate @xmath150 the perturbed equation of eq.([matter ] ) reads @xmath151 where we have used @xmath152 and @xmath153 . in the einstein frame ,
the perturbation equation for the peculiar velocity has the form @xmath154 where @xmath155 eqs.([permatter ] , [ v_ein ] ) are consistent with the results in @xcite and @xcite by setting the general coupling vector as @xmath156 , where @xmath157 is along the direction of @xmath158 .
eqs.([permatter ] , [ v_ein ] ) can also be consistently obtained from the perturbation equations of eqs.([mp_jordan ] , [ v_jordan ] ) by doing the conformal transformation @xmath159 similarly , combining eqs.([permatter ] , [ v_ein ] ) , we can get the second order differential equation @xmath160 next , we consider the perturbed einstein equations in the einstein frame .
we set @xmath126 to fix the freedom of the general conformal transformations . inserting the conformal transformations eq.([contrans ] ) into eqs.([jen_1])-([jen_4 ] ) , we obtain @xmath161\nonumber\quad , \end{aligned}\ ] ] where @xmath162 , @xmath163 and @xmath164 are defined as @xmath165 the perturbed einstein equations give rise to the perturbation of @xmath166 as @xmath167 where we have used @xmath168 if we directly perturb this @xmath169 expressed in eq.([def ] ) @xmath170 after inserting @xmath171 eliminating @xmath15 by the friedmann equation and further employing @xmath172 we can get the result in consistent with eq.([deltav ] ) .
@xmath116 can be determined by eq.([deltf ] ) . inserting eq.([contrans ] ) , ( [ deltf ] )
turns out to be @xmath173\delta f+\frac{1}{3}\kappa^2\tilde{a}^2f\delta\tilde{\rho}_m\nonumber \\ & -&(2\mathcal{\tilde{h}}+4\tilde{a}\gamma)\delta f'-2\tilde{a}f\gamma\tilde{\psi}'-4f\tilde{a}\psi\gamma'\nonumber\\ & -&12f\psi\tilde{a}'\gamma+6f\tilde{a}\tilde{\phi}'\gamma\quad , \end{aligned}\ ] ] where @xmath174 is given by @xmath175 and @xmath2 can be obtained from eq.([rje ] ) in the einstein frame . after fixing @xmath116 ,
we can obtain the consistent covariant perturbation for @xmath58 by eq.([deltagamma ] ) , since the infinitesimal coordinate transformation for @xmath176 presented in eq.([deltagamma ] ) satisfies the rules of the covariant transformation for a scalar field @xmath177 similarly the perturbation for @xmath178 , @xmath179 is given by its covariant configuration @xmath180 and so does @xmath176 @xmath181
from the above analysis , @xmath0 model in the jordan frame is conformally equivalent to the interaction model in the einstein frame .
this equivalence holds not only in the background dynamics but also in the perturbed spacetime .
therefore , we can construct @xmath0 models by choosing a peculiar form of the mass dilation rate function @xmath58 in the einstein frame .
one of the simplest choice of the @xmath58 is @xmath182 after fixing the mass dilation rate function , we can figure out the evolution of the cosmological model in the einstein frame and then we can transform it into the jordan frame by using the conformal transformation . in the lower panel of fig [ evolution ] , we present some examples for the effective de equation of state @xmath3 for the @xmath0 cosmology in the jordan frame constructed by selecting the above mass dilation rate function in the einstein frame after the conformal transformation .
we can see that the effective de equation of state in the constructed @xmath0 model presents the phenomena in the vicinity of @xmath183 and with the @xmath183 crossing behavior , which is consistent with present cosmological observations .
this shows that with the appropriate choice of the mass dilation function , which refers to the mass , a more fundamental quantity , we can construct the viable @xmath0 model to produce the observationally allowed effective de equation of state .
we have also shown that the linear perturbation equations are exactly equivalent at all scales in both the einstein frame and the jordan frame . in the usual @xmath0 model , it was argued that the scale factor described by the @xmath0 cosmology during the matter phase behaves inconsistently with cosmological observations@xcite .
it is of interest to examine the @xmath0 model constructed from the peculiar mass dilation rate function we have chosen and see whether our @xmath0 model is viable if comparing with observations in the matter dominated phase .
we assume that the quantities defined in the jordan frame are consistent with the actual observational quantities .
we will study the subhorizion approximation in the jordan frame .
following @xcite , from eqs.([jen_1],[jen_3 ] ) , we obtain @xmath184 where @xmath116 is governed by eq.([deltf ] ) @xmath185 and @xmath186 is defined by eq.([mdef ] ) . from eq.([secondme ] ) , the equation for the growth function @xmath187 in the sub - horizon approximation reads @xmath188 where @xmath189 and @xmath190 and the effective de equation of state @xmath3 in the jordan frame for the @xmath0 models we constructed . , width=307,height=268 ]
the @xmath186 can be very large in the early time of the universe , since @xmath191 as @xmath192 . when @xmath193 , the regime is called
general relativistic(gr ) regime " where the matter density perturbation has the standard evolution since the modified gravity does nt show up in this regime where @xmath194 and @xmath195 , @xmath196 .
however , in the late time of the universe , when @xmath197 , the modified gravity effect has prominent influence on the structure formation because the ratio of @xmath198 , which has been altered greatly from the standard einstein model and in this case @xmath199 where @xmath200 is the value of @xmath18 today . this regime is called the scalar - tensor(st ) regime " and obviously matter perturbation does not have the standard evolution as in the einstein gravity in this regime . as shown in the upper panel of fig [ evolution ] , in the transition phase of these two regimes , @xmath201 is strongly scale dependent on the wave number of @xmath202 .
the transition of these two regimes can be characterized by @xmath203 which can occur as early as in the matter dominated epoch , even in small scales , in certain types of @xmath0 models in the expansion history of the universe @xcite . in our model , since the dilation rate function @xmath58 will drop rapidly in the early time of the universe , the transition usually happens very late @xmath204 for interested wave number @xmath202 .
thus in our constructed @xmath0 models , we have the standard matter dominated phase and matter power spectrum , which can avoid the problems for the usual @xmath0 model mentioned in @xcite .
in this work , we have further disclosed the equivalence of the @xmath0 gravity presented in the jordan frame and in the einstein frame from the background dynamics to the perturbation theory . in the jordan frame
we have derived the general analytic solution for the @xmath0 model to have the same expansion history as the @xmath1cdm cosmology .
this analytic solution of the @xmath0 model can provide further insights on how cosmology can be used to test the gravity at the largest scales .
moreover we have presented a systematic and self - consistent way to construct the @xmath0 model in the jordan frame through conformal transformation by using the mass dilation rate function in the einstein frame .
we have also developed linear perturbation theories to further exhibit the equivalence of the @xmath0 gravity in the jordan frame and in the einstein frame .
we have shown that the @xmath0 model constructed from the mass dilation rate can produce observational allowed effective de equation of state and provide the matter phase in consistent with cosmological observations .
it is of great interest to confront our @xmath0 model with latest observations in the future study .
it is interesting to note that our @xmath0 model was constructed from the mass dilation rate , which refers to a fundamental quantity , the mass , rather than the mysterious de equation of state .
this gives the root for understanding the equivalence between the @xmath0 model in the jordan frame and the einstein frame , since we know well that the concept of mass plays a central role in gravitational theories @xcite .
the mass can be classified into the inertial mass and the gravitational mass . in the jordan frame
the @xmath0 gravity can be effectively treated with the modified gravitational constant @xmath205 if we compare it with the einstein gravity .
this is effectively equivalent to rescale the gravitational mass @xmath206 .
however the inertial mass in the jordan frame is conserved so that the equation of motion for a free particle in the jordan frame is described by @xmath207 , which is the same as the description in the standard einstein gravity , where the rest inertial mass of particle is conserved .
any change in the gravitational mass in the jordan frame inevitably changes the gravitational field , which , in turn , will change the inertial frame .
any change in the inertial frame will induce the well - known frame - dragging " effect which was recently confirmed by the gravity probe b mission @xcite .
the inertial frame - dragging " effect can be equivalently considered as the inertial mass - dragging " effect by assuming that the inertial frame is unchanged while the inertial mass of particles defined in this frame is rescaled .
this understanding can help us interpret that after the conformal transformation , in the einstein frame , the equation of motion for particles in the inertial frame is governed by the varying mass equation of motion @xmath208 which is one of the well established physics in general relativity @xcite .
the rest inertial mass in the einstein frame is indeed rescaled as @xmath209 and is no longer conserved .
considering the mach principle , there should be no difference among gravitational theories described in different frames .
thus a viable @xmath0 model in the jordan frame should be consistent with the viable model in the einstein frame in describing cosmology .
the result obtained in the work has solid physics root .
_ acknowledgment : j.h.he would like to thank b. r. granett and l. guzzo for helpful discussions .
j.h.he acknowledges the financial support of miur through prin 2008 and asi through contract euclid - nis i/039/10/0 .
the work of b.wang was partially supported by nnsf of china under grant 10878001 and the national basic research program of china under grant 2010cb833000 . _ jian - hua he , b. wang , jcap 06 ( 2008 ) 010 , arxiv:0801.4233.;jian - hua he , b. wang , p. zhang , phys .
d 80 ( 2009 ) 063530 , arxiv:0906.0677 ; xiao - dong xu , jian - hua he , bin wang , accepted for publication in phys.lett.b , arxiv:1103.2632 .
d. pavon , w. zimdahl , phys .
b 628 ( 2005 ) 206 ; s.campo , r. herrera , d. pavon , phys . rev .
d 78 ( 2008 ) 021302(r ) ; g. olivares , f. atrio - barandela and d. pavon , phys . rev .
d 74 ( 2006 ) 043521 .
jian - hua he , b. wang , e. abdalla , phys .
d 83 ( 2011 ) 063515 , arxiv:1012.3904 . | we review the conformal equivalence in describing the background expansion of the universe by @xmath0 gravity both in the jordan frame and the einstein frame . in the jordan frame
, we present the general analytic expression for @xmath0 models that have the same expansion history as the @xmath1cdm model .
this analytic form can provide further insights on how cosmology can be used to test the @xmath0 gravity at the largest scales .
moreover we present a systematic and self - consistent way to construct the viable @xmath0 model in jordan frame using the mass dilation rate function from the einstein frame through the conformal transformation .
in addition , we extend our study to the linear perturbation theories and we further exhibit the equivalence of the @xmath0 gravity presented in the jordan frame and einstein frame in the perturbed space - time .
we argue that this equivalence has solid physics root . |
squeezing in an ensemble of two - level systems @xcite is a quantum phenomenon with applications ranging from quantum information @xcite to precision metrology @xcite .
since the canonical two - level system is a spin-@xmath0 particle , this is often referred to as `` spin squeezing '' , but physical realizations include a broad range of systems defined by the same lie algebra , such as two - component becs @xcite or polarized light @xcite .
a seminal paper by kitagawa and ueda @xcite established nonlinear dynamics as a natural way to generate spin - squeezing , via two distinct mechanisms : by one - axis twisting ( oat ) on the bloch sphere with hamiltonian @xmath1 , and two - axis counter - twisting ( tact ) with hamiltonian @xmath2 , with @xmath3 being components of the collective spin operator .
the latter scenario is more efficient in generating strong spin squeezing , however no experiment has achieved it directly , but schemes have been proposed to convert oat into effective tact hamiltonians @xcite .
the spin - squeezing hamiltonian quadratic in @xmath3 is part of the more general lipkin - meshkov - glick ( lmg ) model @xcite , an important exactly solvable model introduced to test various many - body approximation methods .
the model is relevant in several fields , including nuclear physics @xcite , quantum criticality and phase transitions @xcite , molecular magnetism @xcite , multi - particle entanglement @xcite and classical bifurcations @xcite .
the broad applicability of the lmg hamiltonian , @xmath4,\end{aligned}\ ] ] arises from the independent variation of the three parameters @xmath5 .
therefore , ultracold atomic systems , with their extreme tunability and control , seem a natural choice for a lmg simulator . but as yet only a limited case with one varied parameter ( keeping @xmath6 fixed ) has been demonstrated with bec @xcite . in this paper , we propose a way to implement general lmg with independent multi - parameter control as well as direct spin squeezing by both tact and oat , with a bec in a ring trap with periodic angular modulation of the external potential and of the inter - particle interactions @xcite .
two counter - propagating rotational modes act as pseudo - spin coupled by the periodic potential ; and the spatially modulated nonlinearity controls and varies the various components of the twisting tensor @xmath7 in the general quadratic hamiltonian @xmath8\end{aligned}\ ] ] that covers oat , tact , and lmg as special cases @xcite .
our study is also motivated by the rapid advances in recent years in toroidal trapping of ultracold atoms by a variety of ways : by intersecting a sheet " beam with a lagueree - gaussian beam @xcite or a beam passed through a ring - shaped mask @xcite , with painted potentials @xcite , and with magnetic waveguides @xcite .
notably , the study of lmg and spin - squeezing provides a distinct new line of study with such traps which have so far been primarily used to examine superfluid flow @xcite and interferometric effects @xcite .
the paper is organized as follows . in section [ derivation ]
we derive the hamiltonian , in sec . [ spinsqueezing ] spin squeezing by oat and tact
is discussed , in sec [ lmgmodel ] we discuss the lmg model and phase transitions in our scheme , in sec .
[ fastestsqueezing ] we show various features of the spin squeezing procedure under the optimum rotation condition , in sec . [ conditions ] we specify the conditions for physical parameters , in sec . [ stateprep ] we discuss various options for initial state preparation and detection , and we conclude with sec . [ conclusion ] .
some mathematical details are in appendices .
consider a bec of @xmath9 atoms in a toroidal trap with a superimposed potential @xmath10 and spatially varying nonlinearity , with the latter having twice the periodicity of the potential around the ring as shown in fig .
[ f - schematic ] . for example , in a laguerre - gaussian implementation this could be achieved by modulating the amplitudes and phases of the component beams .
the spatially varying nonlinearity can be produced by optically induced feshbach resonance @xcite , as demonstrated for linear geometry in a recent experiment @xcite .
the dynamics is driven by the hamiltonian @xcite @xmath11 { \hat{\psi}}(\mathbf{r}),{\nonumber}\\ & & { \rm where}\ \ g(\mathbf{r},t)= \frac{4\pi \hbar^2 a ( \mathbf{r},t)}{m } \label{qf - hamiltonian}\end{aligned}\ ] ] is the nonlinearity parameter , with @xmath12 being the scattering length , @xmath13 the particle mass , and @xmath14 , @xmath15 are , respectively the annihilation and creation bosonic field operators satisfying the commutation relations @xmath16=\delta(\mathbf{r}-\mathbf{r^\prime})$ ] .
using cylindrical co - ordinates @xmath17 , with @xmath18 along ring axis , we assume tight harmonic confinement in radial and axial directions @xmath19 with @xmath20 being mean ring radius . the azimuthal potential is taken to be weakly sinusoidal rotating along the ring with frequency @xmath21 , creating a circular bragg grating , @xmath22 + \hbar u_y \sin\left[2q ( \varphi + \omega t ) \right ] , \end{aligned}\ ] ] with @xmath23 an integer and amplitudes assumed small , @xmath24 .
the nonlinear parameter is given a periodic spatio - temporal dependence , @xmath25,\end{aligned}\ ] ] where @xmath26 , @xmath27 and @xmath28 are constants .
we expand the field operators @xmath29 in terms of the mode functions of the ring trap , @xmath30 being the rms widths of the mode functions in the respective directions . assuming that the tight trapping allows only the ground state to be populated in the axial and radial directions , the hamiltonian can be written in terms of the annihilation and creation operators @xmath31 and @xmath32 as @xmath33.\end{aligned}\ ] ]
the hamiltonian simplifies further if only two counter - propagating modes with @xmath34 are initially populated , and the energy gap manifest in the diagonal terms inhibits transitions to other modes .
we absorb the time dependence of the fields into the definition of the mode operators , @xmath35 , @xmath36 , and we define the pseudo - spin operators @xmath37 @xmath38 the hamiltonian reduces to @xmath39 \label{hamj}\end{aligned}\ ] ] where we have defined effective one - dimensional nonlinear coefficients @xmath40 and dropping a scalar term @xmath41 that creates an irrelevant global phase .
this effective hamiltonian is one of main result of the paper .
additional details of its derivation is provided in appendix [ appendixa ] .
it is worth noting that although the sign of the unmodulated nonlinear term @xmath42 suggests attractive inter - particle interaction , it actually stems from repulsive interaction for positive @xmath26 .
this apparent contradiction stems from the fact that for @xmath43 the counter - rotating modes @xmath44 and @xmath45 are almost equally populated and the condensate forms a standing wave with pronounced interference fringes ( see fig .
[ f - schematic ] ) . therefore , the particles are effectively compressed to half of the volume that would be occupied if they were all circulating in the same direction ( @xmath46 ) with no interference fringes .
thus , states with larger @xmath47 correspond to lower interaction energy .
the terms linear in @xmath37 generate rotation on the bloch sphere with angular velocity vector @xmath48 .
spin squeezing is generated by nonlinear terms the coefficients of which can be written as a _ twisting _
tensor @xcite : @xmath49 since @xmath50 , adding multiples of a unit matrix to @xmath7 only leads to a trivial global phase . for @xmath51 ( no spatial modulation of the nonlinearity ) the position of the bec interference fringes plays no role .
the hamiltonian is invariant with respect to rotations around @xmath52 , and the nonlinear dynamics corresponds to oat . when @xmath53 the nonlinearity is modulated , and the interaction energy is affected by the relative positions of the interference fringes and the maxima of the nonlinearity ( see fig .
[ f - schematic ] ) . in this case
the twisting tensor has eigenvalues @xmath54 and @xmath55 .
the maximum squeezing rate is determined by the difference between the largest and the smallest eigenvalues of the twisting tensor @xcite . for positive @xmath56 and states that are nearly gaussian and optimally positioned on the bloch sphere , on the axis of the middle eigenvalue , the maximum squeezing rate is @xmath57 where the squeezing parameter @xmath58 is defined as the ratio of the minimum variance of the uncertainty ellipse of the state and the variance of the spin coherent state .
appendix [ appendixb ] provides more details about the squeezing . to achieve this maximum rate ,
the state should be centered on the bloch sphere along the axis corresponding to the middle eigenvalue , and kept optimally oriented by continuous rotation of the system at frequency @xmath59 around this axis , where @xmath60 are the eigenvalues in ascending order ; here , if @xmath61 , this frequency is @xmath62 . in the special case
when the middle eigenvalue is exactly between the other two eigenvalues , @xmath7 can be transformed to a diagonal form diag@xmath63 corresponding to tact ( see @xcite for details ) ; here @xmath64 is the difference between the largest and smallest eigenvalues . in our model
it corresponds to the condition @xmath65 .
since the modulation maxima and minima are proportional to @xmath66 , the maximum and minimum are @xmath67 and @xmath68 , so that tact occurs if the maximum value of the modulated nonlinearity is five times larger than the minimum ( see fig .
[ f - schematic ] ) . in this case
, the state naturally remains optimally oriented , no rotation is needed .
special cases of our hamiltonian ( [ hamj ] ) correspond to various instances of the lmg model .
for example , setting @xmath69 leads to lmg model with @xmath70 , @xmath71 , and @xmath72 , and setting @xmath73 leads to lmg model with @xmath74 , @xmath75 , and @xmath76 with the roles of the coordinates in @xmath77 permutated @xmath78 . to illustrate the applicability of our hamiltonian to the lmg model
, we show a possible lmg phase diagram in fig .
[ f - phasediag1 ] . in this example , we assume @xmath79 and @xmath80 , varying @xmath81 and @xmath82 . such a choice may be natural , for instance , if one needs to avoid rotation of the fields .
the zones denoted i
iv correspond to those studied in @xcite , reflecting different qualitative behavior of the energy surfaces and different properties of the density of eigenstates of the hamiltonian .
crossing the boundary between various zones corresponds to various phase transitions ; for example crossing from i to ii by decreasing @xmath83 with @xmath51 corresponds to the classical bifurcation at the transition from rabi to josephson dynamics observed in @xcite . the conditions for fastest squeezing can be achieved with the parameters used for the lmg phase diagram above .
the state is positioned along @xmath84 axis on the bloch sphere and rotation @xmath82 should be applied at the optimum frequency defined earlier .
these conditions are fulfilled by parameters of the red broken line between points a d in fig .
[ f - phasediag1 ] .
we show the corresponding time evolution of the squeezing parameter @xmath58 in fig .
[ f - timevol1 ] .
the best squeezing properties correspond to tact ( line d ) and as can be seen , even a relatively small deviation of the tact condition leads to a strong deterioration of squeezing in the last stages ( line c ) . for the respective parameter values
we also show the density of states @xmath85 as a function of the hamiltonian eigenvalues @xmath86 with qualitatively different shapes in zone ii ( one peak ) and iii ( two peaks that merge in the limit of tact , point d of fig .
[ f - phasediag1 ] ) .
the peaks divide the energy spectrum into regions with qualitatively different features of the hamiltonian eigenstates . as an interesting relationship between squeezing and lmg phase we note that the states undergoing squeezing with the optimum rotation are located at one of the density peaks .
we illustrate this feature in a few figures . in fig .
[ f - eigenenergs1 ] we show the eigenvalues for the hamiltonian from eq .
( [ hamj ] ) as a function of @xmath87 with the remaining parameters fixed at @xmath88 , @xmath89 , @xmath90 , and choosing @xmath91 for fastest squeezing .
we set @xmath92 .
the arrow denoted `` ( example ) '' corresponds to the bloch sphere diagrams in figs .
[ f - qfunctions1 ] and [ f - qfunctions2 ] .
the arrows denoted with ( a)(d ) in fig .
[ f - eigenenergs1 ] correspond to the spectral densities shown in fig .
[ f - timevol1 ] , except that here we use smaller value of @xmath88 for better visibility of the spectral lines . the gap between the two transverse red lines fig .
[ f - eigenenergs1 ] indicate the range of eigenstates with dominant contribution to the state being squeezed , for @xmath88 .
the relative contributions of the eigenstates vary with @xmath87 since the red lines cut across the black lines that trace the eigenenergies as functions of @xmath87 .
the histogram in fig .
[ f - eigenenergs1 ] shows the probabilities of the individual eigenstates for @xmath93 . the spread ( twice the energy uncertainty @xmath94 ) of the histogram corresponds to separation of the two red lines mentioned above , and the correspond gap for @xmath95 equals the width of vertical red bars in the spectral densities plotted in the insets of fig .
[ f - timevol1 ] . in fig .
[ f - qfunctions1 ] we show the @xmath96-functions of the hamiltonian eigenstates in the vicinity of the peak of the energy spectrum , for a fixed value of @xmath93 with all the other parameters the same as in fig . [ f - eigenenergs1 ] , and the fastest squeezing condition yields @xmath97 . as can be seen in fig .
[ f - qfunctions1 ] , the density peak divides the spectrum into qualitatively different regions .
states with energy below the peak are doubly degenerate and their @xmath96-functions form two rings encircling the poles .
states with energy above the peak have @xmath96-functions following a single curve similar to a tennis - ball seam .
the transition between these two regimes at the spectral peak contains pillow - shaped and x - shaped structures .
the highlighted panels in fig .
[ f - qfunctions1 ] show eigenstates involved in the superposition being squeezed .
one can understand the squeezing process under the optimum rotation condition as gradually forming destructive interference of one diagonal pair of corners of the pillow - shaped and x - shaped structures , and constructive interference of the other pair of corners .
snapshots of the time evolution of the superposition state being squeezed is shown in fig .
[ f - qfunctions2 ] as the squeezing is generated , up to a time beyond when maximum squeezing is achieved .
the evolution is confined to the tangential degree of freedom if the radial and axial confinement is sufficiently tight so that occupation of higher modes of these degrees of freedom is energetically forbidden .
this is achieved if @xmath98 where @xmath99 is the atomic de broglie wavelength along the ring . to avoid transitions to other circular modes
, the corresponding couplings should be sufficiently small . in the hamiltonian ( [ hama ] )
the terms with @xmath100 couple modes @xmath34 also to unwanted modes @xmath101
. the probability of such transitions would oscillate with small magnitude near zero provided that the couplings are much smaller than the energy difference of the coupled states , namely @xmath102 this condition puts a lower bound on the time during which the system can complete a @xmath103-flip between the @xmath104 states : @xmath105 for rubidium atoms in visible light , @xmath106s , so the flipping time would take at least @xmath107 @xmath108s .
the nonlinear term in hamiltonian ( [ hama ] ) couples the modes @xmath34 to other modes @xmath109 for which the condition @xmath110 ( homogeneous part @xmath111 ) or @xmath112 ( inhomogeneous part @xmath87 ) is satisfied .
if , say mode @xmath113 is occupied with @xmath9 atoms and the remaining modes are empty , then the energetically closest state satisfying this condition corresponds to @xmath114 atoms in mode @xmath23 and one atom in each of the modes @xmath115 .
the energy difference between these two states is @xmath116 .
thus , to avoid such transitions , the nonlinear coupling should satisfy @xmath117 setting limits on the shortest time during which useful squeezing can be produced @xmath118 much longer than the rotational period of the superposition of two lowest - energy eigenstates of the ring trap .
assuming rubidium in a ring trap with @xmath119 one finds @xmath120 ms .
the process must be performed before losses become significant . with @xmath121 and @xmath122 ,
the peak atomic density @xmath123 @xmath124 and the dominant mechanism of losses are inelastic collisions .
recent experiments with combined magnetic and optical @xcite and purely optical feshbach resonance @xcite report suppressing inelastic loss rates to @xmath125 @xmath126/s , i.e. , the requirements are just on the edge of present experimental possibilities . full demonstration of these effects would require further suppression of @xmath127 that can be expected , e.g. , from dark - state - like interferences by means of multiphoton transitions in combined optical fields @xcite .
more detailed treatment of the problem of squeezing with losses and finite temperature can be found in @xcite .
a spin coherent state can be prepared by rotating a bec with the chosen winding number @xmath23 .
one way to do that is via raman transition to transfer the angular momentum from lg beams to a non - rotating bec @xcite .
the resulting state is an eigenstate of @xmath128 with eigenvalue @xmath129 .
any other spin coherent state can be prepared from this state by application of operators @xmath37 , by varying the parameters @xmath130 and @xmath21 to rotate the state on the bloch sphere . for detection @xmath37 , the procedure can be reversed : first the state is rotated so that the operator of interest corresponds to @xmath128 , then a raman process converts the @xmath131 eigenstate to a non - rotating bec .
finally the proportion of the non - rotating atoms is measured ; for the setup in ref .
@xcite this can be done by releasing the condensate and observing the size of the central hole in the interference pattern produced by the free - falling atoms . alternately , one can measure @xmath128 by coupling the ring resonator to a linear atomic waveguide ( formed , e.g. , by a red - detuned horizontal laser beam ) positioned tangentially near the ring .
atoms circulating in the opposite orientations would leak to the waveguide and propagate in opposite directions towards the waveguide ends where they can be detected .
we have shown that a bec in a toroidal trap with a circular lattice and a spatially modulated nonlinearity can be used to realize a general lmg model and comprehensive spin - squeezing procedures , including direct two - axes twisting ; with independent control over all the relevant parameters .
our model should be useful for creating squeezed states for quantum measurements and testing critical phenomena in a range of physical systems described by the lmg model .
it also provides a different realm of physics to be explored with the rapidly advancing technology of toroidal trapping of ultracold atoms .
we conclude with some comments about our model : unlike most schemes for bosonic pseudospin squeezing @xcite , it can directly generate rotation about an arbitrary bloch axis by varying @xmath130 and @xmath21
. any relevant ratio of the lmg parameters can be achieved by co - ordinate rotation , if @xmath87 can access @xmath132 $ ] or @xmath133 $ ] .
although our examples did not use it , our model allows additional control via temporal modulation ( @xmath134 ) of the nonlinearity .
kkd acknowledges support of the national science foundation under grant no .
phy-1313871 , and of a passhe - fpdc grant and a research grant from kutztown university .
we start from eq .
( [ hama ] ) and assume that only two modes @xmath104 are populated . we can then drop a diagonal term that only gives a constant offset . using the definition of the non - linear coefficients
@xmath135 the hamiltonian takes the form @xmath136 the time dependence of the hamiltonian can be absorbed in the definition of the bosonic operators as @xmath137 expressing the equations of motion of these operators , one finds @xmath138 \nonumber \\ & = & \frac{u_x - iu_y}{2}{\hat{a}}_{-q}e^{i2q\omega t } + \chi_0 \left({\hat{a}}_{q}^{\dag } { \hat{a}}_{q}^{2 } + 2{\hat{a}}_{q } { \hat{a}}_{-q}^{\dag}{\hat{a}}_{-q}\right ) + \frac{\chi_1}{2 } { \hat{a}}_{q}^{\dag } { \hat{a}}_{-q}^{2 } e^{i(4q\omega t-\alpha)}\nonumber \\ & = & \frac{u_x - iu_y}{2}{\hat{b}}e^{iq\omega t } + \chi_0 \left({\hat{a}}^{\dag } { \hat{a}}^{2 } + 2{\hat{a } } { \hat{b}}^{\dag}{\hat{b}}\right ) e^{iq\omega t } + \frac{\chi_1}{2 } { \hat{a}}^{\dag } { \hat{b}}^{2 } e^{iq\omega t - i\alpha}\nonumber \\ & = & i \dot{{\hat{a } } } e^{iq\omega t } - q\omega { \hat{a } } e^{iq\omega t } , \label{amotion } \\
i \dot{{\hat{a}}}_{-q } & = & \frac{1}{\hbar}[{\hat{a}}_{-q},h ] \nonumber \\ & = & \frac{u_x + iu_y}{2}{\hat{a}}_{q}e^{-i2q\omega t } + \chi_0 \left ( 2{\hat{a}}_{q}^{\dag}{\hat{a}}_{q } { \hat{a}}_{-q } + { \hat{a}}_{-q}^{\dag } { \hat{a}}_{-q}^{2 } \right ) + \frac{\chi_1}{2 } { \hat{a}}_{-q}^{\dag } { \hat{a}}_{q}^{2 } e^{-i(4q\omega t-\alpha)}\nonumber \\ & = & \frac{u_x + iu_y}{2}{\hat{a}}e^{-iq\omega t } + \chi_0 \left(2{\hat{a}}^{\dag } { \hat{a } } { \hat{b } } + { \hat{b}}^{\dag}{\hat{b}}^2 \right ) e^{-iq\omega t } + \frac{\chi_1}{2 } { \hat{b}}^{\dag } { \hat{a}}^{2 } e^{-iq\omega t+i\alpha}\nonumber \\ & = & i \dot{{\hat{b } } } e^{-iq\omega t } + q\omega { \hat{b } } e^{-iq\omega t}. \label{amotion2}\end{aligned}\ ] ] thus , the equations of motion for operators @xmath44 and @xmath45 can be written as @xmath139 these equations correspond to the hamiltonian @xmath140 which does not depend on time . using the collective spin operators @xmath37 of eqs .
( [ jx])([jz ] ) , one can express the higher powers of the operators of eq .
( [ pomham2 ] ) as @xmath141 the hamiltonian then can be written as @xmath142 .\end{aligned}\ ] ] dropping the irrelevant last term , one obtains eq .
( [ hamj ] ) .
we define the squeezing parameter @xmath58 , following ref .
@xcite , as the ratio of the smallest variance of the uncertainty ellipse @xmath143 on the bloch sphere to the variance of the spin coherent state @xmath144 , i.e. , @xmath145 the value @xmath143 is determined as the smaller eigenvalue of the 2@xmath1462 projection of the variance matrix to the tangential plane of the bloch sphere with respect to the mean spin vector .
the calculation proceeds as follows .
assume the state is characterized by the first moments @xmath147 and variances @xmath148 .
we calculate the rotated variance matrix as @xmath149 where @xmath150 is matrix composed of @xmath151 and @xmath152 with @xmath153 .
note that @xmath154 rotates the state to the pole of the bloch sphere , in particular @xmath155 the value @xmath143 is then calculated as the smaller eigenvalue of the matrix @xmath156 with @xmath157 where @xmath158 are the elements of @xmath159 defined in ( [ tildev ] ) .
a. e. kozhekin , k. mlmer , and e. polzik , phys .
a * 62 * , 033809 ( 2000 ) ; b. julsgaard , j. sherson , j. i. cirac , j. fiurek , and e. s. polzik , nature * 432 * , 482 ( 2004 ) ; k. hammerer , a. s. srensen , and e. s. polzik , rev .
mod . phys . *
82 * , 1041 ( 2010 ) .
j. esteve , c. gross , a. weller , s. giovanazzi , and m. k. oberthaler , nature * 455 * , 1216 ( 2008 ) ; c. gross , t. zibold , e. nicklas , j. esteve , and m. k. oberthaler , nature * 464 * , 1165 ( 2010 ) ; m. f. riedel , p. bhi , y. li , t. w. hnsch , a. sinatra , and p. treutlein , nature * 464 * , 1170 ( 2010 )
. a. micheli , d. jaksch , j. i. cirac , and p. zoller , phys .
a * 67 * , 013607 ( 2003 ) ; l. pezze , l. a. collins , a. smerzi , g. p. berman , and a. r. bishop , phys . rev .
a * 72 * , 043612 ( 2005 ) ; b. julia - diaz , t. zibold , m. k. oberthaler , m. mele - messeguer , j. martorell , and a. polls , phys .
a * 86 * , 023615 ( 2012 ) ; b. julia - diaz , e. torrontegui , j. martorell , j. g. muga , and a. polls , phys . rev .
a * 86 * , 063623 ( 2012 ) ; m. lapert , g. ferrini , and d. sugny phys .
a * 85 * , 023611 ( 2012 ) .
y. c. liu , z. f. xu , g. r. jin , and l. you , phys .
* 107 * , 013601 ( 2011 ) ; c. shen and l. m. duan , phys .
a * 87 * , 051801 ( 2013 ) ; j. y. zhang , x. f. zhou , g. c. guo , and z. w. zhou , phys .
a * 90 * , 013604 ( 2014 ) .
s. dusuel and j. vidal , phys .
lett . * 93 * , 237204 ( 2004 ) ; f. leyvraz and w. d. heiss , phys .
lett . * 95 * , 050402 ( 2005 ) ; t. caneva , r. fazio , and g. e. santoro , phys .
104426 ( 2008 ) . h. m. kwok , w. q. ning , s. j. gu , and h. q. lin , phys .
e * 78 * , 032103 ( 2008 ) ; g. engelhardt , v. m. bastidas , and t. brandes http://arxiv.org/abs/1405.3514 ( 2014 ) .
s. gupta , k.w .
murch , k.l .
moore , t.p .
purdy , and d.m .
stamper - kurn , phys .
* 95 , * 143201 ( 2005 ) ; m. heimsoth , c. e. creffield , l. d. carr , and f. sols , new j. phys .
* 14 , * 075023 ( 2012 ) ; j. d. pritchard , a. n. dinkelaker , a. s. arnold , p. f. griffin and e. riis , new j. phys . * 14 , * 103047 ( 2012 )
. m. theis , g. thalhammer , k. winkler , m. hellwig , g. ruff , r. grimm , and j. h. denschlag , phys .
* 93 * , 123001 ( 2004 ) ; g. thalhammer , m. theis , k. winkler , r. grimm , and j. h. denschlag phys . rev .
a * 71 * , 033403(2005 ) .
y. li , y. castin , and a. sinatra , phys .
lett . * 100 * , 210401 ( 2008 ) ; b. opanchuk , m. egorov , s. hoffmann , a. i. sidorov , and p. d. drummond , europhys . lett . * 97 * , 50003 ( 2012 ) ; a. sinatra , y. castin , and e. witkowska , europhys . lett . * 102 * , 40001 ( 2013 ) . | we propose a scheme for spin - squeezing in the orbital motion of a bose - einstein condensate ( bec ) in a toroidal trap .
a circular lattice couples two counter - rotating modes and squeezing is generated by the nonlinear interaction spatially modulated at half the lattice period . by varying the amplitude and phase of the modulation ,
various cases of the twisting tensor can be directly realized , leading to different squeezing regimes .
these include one - axis twisting and the two - axis counter - twisting which are often discussed as the most important paradigms for spin squeezing .
our scheme naturally realizes the lipkin - meshkov - glick model with the freedom to vary all its parameters simultaneously . |
in the @xmath14-dominated , cold dark matter ( @xmath14cdm ) cosmology , massive galaxies form hierarchically inside of dark matter haloes ( eg . * ? ? ? * ; * ? ? ? * ) .
the mass evolution of these cdm haloes can place stringent constraints on the cosmological models ( eg . * ? ? ?
* ) , and the most reliable measurements of cdm halo masses are generally derived from modeling gravitationally lensed images . furthermore , gravitational lensing has become a powerful tool for studying properties of both the lensing objects and the more distant sources .
of particular interest are constraints on the mass profiles ( eg . * ? ? ?
* ; * ? ? ?
* ) and the mass - to - light ratios ( eg . * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) of lens galaxies , and the host galaxies of active galactic nuclei ( eg . * ? ? ?
* ; * ? ? ?
while all of these studies are advancing due to increased numbers of known lensing systems @xcite , they are still limited by relatively small sample sizes . in this work ,
we report the discovery of an optical arc which is @xmath15 southwest of the optically unidentified radio source , first j121839.7 + 295325 .
this radio source was discovered in the faint images of the radio sky at twenty - centimeters ( first ) with the very large array ( vla ; * ? ? ?
* ; * ? ? ?
* ) and later detected at 610 mhz @xcite and 74 mhz @xcite .
its relatively bright flux of @xmath16 mjy suggests that it is at @xmath17 ( eg .
* ; * ? ? ?
additionally , j121839.7 + 295325 was _ undetected _ in the pure - parallel observations ( propid : 8090 , pi : s. casertano ) with the wide field and planetary camera 2 ( wfpc2 ) on the hubble space telescope ( hst ) to f606w@xmath18 mag ( @xmath19-band hereafter ; * ? ? ?
* ) , which indicates it is likely at @xmath20 ( eg . * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these two independent constraints yield a coarse redshift estimate of @xmath21 . by correlating the wfpc2 archive with the catalog of first sources , @xcite
note an arc - like feature with @xmath22 mag .
based on the astrometric uncertainties for the hst - wfpc2 and first imaging , they conclude that the likelihood that this arc is the optical identification of the radio source is @xmath23 , following the work of @xcite . since @xcite rule out the possibility that the arc is the optical identification of j121839.7 + 295325 , we will investigate the hypothesis that the optical arc is a gravitationally lensed imaged by the optically unidentified radio source .
this work is organized as follows :
[ mmt ] describes our mmt observations ,
[ model ] outlines the lensing analysis , [ discuss ] discusses the lensing interpretation of these observations , and [ con ]
gives some closing remarks with thoughts toward possible future observations .
unless explicitly stated , all magnitudes are in the ab system @xcite .
we adopt the following cosmological model : @xmath24 , @xmath25 , and @xmath26 km s@xmath27 mpc@xmath27 where @xmath28 @xcite .
while the radio and optical properties provide a broad redshift constraint on the lensing object of @xmath21 , little is known about the optical arc .
therefore , we observed the putative gravitational arc on may 18 , 2007 with the blue - channel spectrograph on the 6.5 m mmt , and obtained 4@xmath292400 s and 1@xmath291200 s exposures . for these observations , our @xmath30 slit tracked through the sky at the parallactic angle , as the object varied in airmass of @xmath31 , which was very closely aligned with the position angle of the arc seen in the left panel of figure [ mod ] . with
the 500 grooves mm@xmath27 grating , the spectral resolution was 3.6 with a dispersion of 1.2 pix@xmath27 .
we observed feige 34 throughout the night to determine the spatial point - spread function ( psf ) and to serve as the spectrophotometric flux standard star .
the two dimensional spectra were reduced by standard algorithms using custom routines in idl .
the ccd read noise , gain , and dark rate were @xmath32 @xmath33 , @xmath34 @xmath33 adu@xmath27 , and @xmath35 @xmath33 pix@xmath27 hr@xmath27 , respectively . to determine an accurate trace , we simply rebinned the two dimensional spectrum in wavelength to boost the relatively poor signal - to - noise ratio .
the one dimensional spectrum was extracted from a given frame using the observed psf as weights in both axes .
finally , the wavelength solution for each frame was determined from an henear comparison lamp , which was observed at the corresponding position on the sky . despite the long integration time
( 3 hours ) on the arc , the total signal - to - noise per pixel is relatively low .
therefore , we rebinned the calibrated , sky - subtracted , and stacked spectrum to @xmath36 for @xmath37 and @xmath38 for @xmath39 . at this resolution ,
a weak spectral break is identifiable at @xmath40 , however we can not conclusively determine whether it is the lyman ( @xmath41 ) or the balmer/4000 break .
therefore , to determine the most likely redshift with its associated uncertainty and relative probability , we computed a spectrophotometric redshift ( eg . * ? ? ? * ) with the redshift code , ` hyperz ` @xcite .
we did not include the @xmath19-band flux in the fit , since the mmt slit did not fully encompass the hst - wfpc2 aperture .
we adopted a grid of synthetic spectra @xcite parameterized by an exponential star formation history with @xmath42-folding times of @xmath43 gyr and ages of @xmath44 , where @xmath45 is the age of the universe .
the final grid of spectral templates is defined by the extinction of @xmath46 for @xmath47 . in figure
[ spec ] , we show the observed spectrum as filled points with the best fit template as a solid line . in the inset , we show the relative probability , which indicates two possible redshift solutions : the balmer/4000 break at @xmath48 or the lyman break at @xmath49 .
if the arc - like object is at @xmath48 , then it is likely an unrelated foreground galaxy to j121839.7 + 295325 with @xmath50 mag .
however , if it is at the higher redshift , then the observed arc morphology in the hst - wfpc2 imaging suggests that j121839.7 + 295325 may be gravitationally lensing a background source . without a flux measurement blueward of the blue channel spectrum
( @xmath51 ) to break this redshift degeneracy ( eg . * ? ? ?
* ) , we can only make statistical statements regarding the redshift of the arc and the physical interpretations of the system . from the inset in figure [ spec ] ,
the integrated probabilities of the two peaks are 13% and 87% for the balmer/4000 and lyman breaks , respectively .
since the lyman break redshift is @xmath52 times more likely , we will further develop the lensing hypothesis .
the best - fit template found by ` hyperz ` has an age of @xmath53 myr , instantaneous burst with @xmath54 mag , which is similar to the lyman break galaxies at @xmath55 @xcite .
therefore the most likely redshift of arc is @xmath0 , which we derive from @xmath56}$ ] and @xmath57 .
we observed j121839.7 + 295325 with the sao wide - field infrared camera ( swirc ) on march 19 , 2008 and obtained @xmath58 s and @xmath59 s dithered images in the @xmath60- and @xmath61-bands , respectively .
these data were reduced by standard algorithms with a similar suite of idl routines used for the blue - channel reductions .
the seeing was relatively constant through the night at @xmath62 and @xmath63 in the @xmath60- and @xmath61-bands , respectively .
the sky image per object frame was created using ` sextractor ` @xcite . in the @xmath64 field - of - view , each object frame contained three stars from the two micron all sky survey ( 2mass ) point source catalog @xcite , which we used to calibrate the astrometry of each frame and to determine the flux zeropoint , with the ( @xmath65 ) magnitude conversions of @xcite .
the astrometrically corrected , sky - subtracted images were averaged together with a 1@xmath66-clip into stacked science frames .
we disregard a small portion of the swirc field - of - view owing to appreciable cosmetic defects .
therefore , the stacked images each contain @xmath67 objects and yield number counts which are consistent with the @xmath60-band galaxy counts of the chandra deep field and hubble deep field , south ( * ? ? ?
* see their figure 2 ) . based on a power - law fit to our observed counts , we estimate the @xmath68 , 50% completeness limits are @xmath69 mag and @xmath70 mag .
the assumed lensing galaxies is not detected in either of these bandpasses , which may suggest that j121839.7 + 295325 is indeed a rather dark object . to estimate our detection limits in the infrared imaging , we determine the limit corresponding to a @xmath68 fluctuation at the position of the radio source .
for a circular aperture with a radius of @xmath71 , these limits are @xmath72 mag and @xmath73 mag and are used below to derive the limits on the mass - to - light ratios .
if we are to interpret this system in the context of gravitational lensing , we must investigate whether lensing can provide a sensible explanation of the optical arc .
the morphology of the arc immediately implies that the lensing galaxy can not be circular : a circular lens that produces such a long arc ( compared with the radius of curvature ) would produce a prominent counter - arc as well .
in contrast , an elongated galaxy can produce a single long arc either as a highly - distorted single image , or as a merged triplet of images . that leaves the question of how to interpret the compact , high surface brightness knot just off the east end of the arc .
since there is no natural way to interpret the arc and knot as lensed images of single source , the knot is either a lensed image of a second source , or an unrelated foreground object .
we attempted to model the knot as a lensed image , although it has little impact on the results since the observed image merely determines the properties of the assumed second source .
we did not make a prior assumption about whether the arc is a single image or merged triplet .
instead , we merely postulated that there are two sources behind an elongated lens galaxy .
we treated the sources as circular gaussians for simplicity , fitting for their positions , sizes , and fluxes .
we treated the lens galaxy as an isothermal ellipsoid , fitting for its position , einstein radius , ellipticity , and position angle . for a given set of model parameters
, we used the software package ` lensmodel ` @xcite to find the images of the two sources , which we compared to the observed arc and knot on a pixel - by - pixel basis to compute the @xmath74 goodness - of - fit .
we used a downhill simplex method to search the parameter space , with an inner loop to optimize the source parameters for a fixed lens galaxy , and an outer loop to optimize the galaxy parameters as well .
in the best - fit model , the lens has an einstein radius of @xmath1 and ellipticity of @xmath75 , the two sources are separated by @xmath76 , and the reduced - chi squared is @xmath77}\!=\!0.94 $ ] . in figure
[ mod ] , we show the hst - wfpc2 image ( left ) , the lens model ( middle ) , and the reconstructed source plane ( right ) .
the mass enclosed within the einstein radius slowly varies with the lens redshift between @xmath78 m@xmath3 ( see the upper panel of figure [ mass ] ) , which is a relatively large mass to have been undetected in the hst - wfpc2 and mmt - swirc imaging .
there are three additional points to make regarding the lens models .
( 1 ) in successful models , the arc is a single highly - distorted image of a source lying just outside a cusp caustic in the source plane .
we did not impose this as an assumption , it emerged naturally from the modeling .
( 2 ) the present model does not explain the variation in surface brightness along the arc , because we used a simple gaussian source simply to reproduce the shape of the arc .
if the arc is indeed singly - imaged , then the surface brightness variation along the arc merely represents the intrinsic variations of the source .
( 3 ) our model predicts a compact , faint counter - image just beyond the center of the lensing galaxy ( see middle panel of figure [ mod ] ) .
there appears to be some flux in roughly the right location in the hst - wfpc2 image , however it is likely associated with a hot pixel .
we did not use the portion of the image as constraints to the lens models .
= 0.75 since j121839.7 + 295325 is undetected in the hst and mmt imaging , we can only determine a lower limit on its mass - to - light ratio .
we must convert the above detection limits ( @xmath79 and @xmath80 ) to the luminosity in a fixed bandpass to determine a mass - to - light ratio , which is independent of redshift . to facilitate direct comparisons with other galaxies , we opt to @xmath81-correct our limiting apparent magnitudes to the rest - frame @xmath82-band absolute magnitudes .
these conversions require assuming a template sed for the undetected lensing galaxy ( eg . * ? ? ?
* ) . to highlight two extreme possibilities
, we show the range of the dynamical mass - to - light ratio consistent with the @xmath19-band ( blue ) , @xmath60-band ( green ) , and @xmath61-band ( red ) imaging as derived from the e ( middle ) and i m ( bottom ) templates @xcite in figure [ mass ] .
the shaded region shows the likely lens redshift , as discussed in [ intro ] .
if a stellar component to j121839.7 + 295325 exists , then we expect it likely has an early - type based on its radio flux ( eg .
. therefore , to be compatible with the observations , j121839.7 + 295325 likely has @xmath83
m@xmath3 l@xmath7 , and could be as dark as @xmath84 m@xmath3 l@xmath7 for the late - type sed .
these limits will slightly decrease / brighten by including dust obscuration , since some fraction of the stellar light may be extincted and increase the maximum luminosity , for fixed flux limits .
however , even with a modest amount of extinction ( @xmath85 mag ) , the intrinsic mass - to - light ratio would still be @xmath6 m@xmath3 l@xmath7 ( for the early - type template ) .
nevertheless , the wfpc2 @xmath19-band observations generally provide the strongest constraint , owing to the relatively shallow swirc observations .
the hst - wfpc2 imaging suggests that j121839.7 + 295325 is gravitationally lensing a background object .
our blue channel spectroscopy supports this hypothesis by placing the optical arc is likely at high redshift ( @xmath86 ) .
given the flux limits from hst - wfpc2 and mmt - swirc ( @xmath87 mag , @xmath72 mag , and @xmath88 mag ) , we find the lensing object may be rather dark , with a @xmath82-band mass - to - light ratio of @xmath83 m@xmath3 l@xmath7 . from the k20 survey , @xcite find a dynamical mass - to - light ratio of @xmath89 m@xmath3 l@xmath7 for their sample of 18 early - type galaxies at @xmath90 .
their upper limit can be extended to @xmath91 m@xmath3 l@xmath7 by including a sample of 28 galaxies from the coma cluster @xcite .
this mass - to - light ratio is roughly consistent with the @xmath92 early - type galaxies of @xcite , which have @xmath93 m@xmath3 l@xmath7 . by comparing to a series of @xmath94 studies , they find the mass - to - light ratio of early - type galaxies evolves as @xmath95 , since the low redshift ellipticals have @xmath96 m@xmath3 l@xmath7 @xcite . therefore ,
if our lower limits derived in [ model ] are correct , then j121839.7 + 295325 may be a relatively dark galaxy , whose total mass is dominated by dark matter . while there are solid statistical arguments that such objects should exist @xcite , there are currently no confirmed examples .
since direct optical observations of these extreme mass - to - light objects will be nearly impossible , they will likely manifest themselves as `` dark lenses '' @xcite .
the most common suggestions of dark lensing have arisen from wide - separation quasar pairs ( wsqps ) , quasars with nearly identical optical spectra separated by @xmath97 , with no identifiable lensing source .
detailed analyses of the spectra of these quasars , from the radio to the @xmath98-rays , have revealed that all known wsqps are simply binary quasars @xcite .
additionally , the cosmic shear survey with hst @xcite has identified a number of gravitational lenses .
in one particular field , @xcite suggested that the conspicuous tangential alignment of galaxies with arc - like morphologies and no obvious over - density of foreground galaxies may be a dark lens .
however , @xcite ruled out this hypothesis by demonstrating the colors of these galaxies are inconsistent with high redshift , lensed objects .
since j121839.7 + 295325 remains unidentified in the optical and near - infrared , it is a good candidate for a dark lens .
however , the bright radio emission of @xmath12 mjy suggests that the lens object likely contains an active galactic nucleus ( eg .
if a baryonic counterpart to this radio source exists , then it likely corresponds to a massive and perhaps extincted galaxy , which would make it similar to the distant red galaxies ( drgs ; * ? ? ?
* ) or the extremely red objects ( eros ; * ? ? ?
if we assume the optical / infrared color of @xmath99 mag typical of an ero , then j121839.7 + 295325 must have @xmath100 mag and a stellar mass of @xmath101 m@xmath3 ( * ? ? ?
* based on their figure 7 ) . in this case , our lens models imply that j121839.7 + 295325would have a dynamical - to - stellar mass ratio of @xmath102 , which is significantly higher than the @xmath103 star - forming galaxies with @xmath104 @xcite .
the hst - wfpc2 and mmt data suggest that j121839.7 + 295325 is an extremely dark galaxy worthy of follow - up observations to confirm or exclude this hypothesis .
in particular , with deeper optical spectroscopy of the arc , our redshift analysis can be verified .
moreover , observations blueward of the blue channel spectroscopy will break the redshift degeneracy of @xmath105 and @xmath106 . owing to the necessary wavelength ( @xmath107 ) and flux limit ( @xmath108 mag ) , these observations could be conducted with the ultraviolet channel of the wide field camera 3 ( wfc3 ) for hst .
since the radio source has not been optically identified , its redshift can only be very coarsely constrained .
the majority of millijansky radio sources have early - type morphologies and colors at @xmath109 ( eg .
* ; * ? ? ?
* ) , consequently its balmer/4000 break will occur at @xmath110 . furthermore , j121839.7 + 295325 is undetected to faint flux - levels , and may require a rather deep ( @xmath111 mag ) imaging campaign . therefore , medium - band or grism observations in the infrared mode with wfc3 would better constrain the lens redshift .
should such observations support our proposed lensing scenario , this system could be among the most distant known gravitational lenses @xcite , a dusty ero with @xmath112 m@xmath3 , or a genuine dark lens , which would be a unique confirmation of the @xmath14cdm paradigm .
we acknowledge support from nasa grant hst ar-10974.01a ( to rer ) , hst ar-08357.01a ( to raw ) awarded by stsci , which is operated by aura for nasa under contract nas 5 - 26555 , and nasa jwst grant nag5 - 12460 ( to raw ) .
we thank warren brown , jacob russell , rolf jansen , alejandra milone , and nimish hathi for helpful discussions and guidance .
we are particularly grateful for the excellent comments and suggestions of the anonymous referee .
becker , r. h. , white , r. l. , & helfand , d. j. 1995 , , 450 , 559 bertin , e. & arnouts , s. 1996 , a&as , 117 , 393 blanton , m. r. , et al .
2001 , , 121 , 2358 bolton , a. s. , burles , s. , koopmans , l. v. e. , treu , t. , & moustakas , l. a. 2006 , , 638 , 703 bolzonella , m. , miralles , j .- m . , & pell , r. 2000 , a&a , 363 , 476 de breuck , c. , et al .
2001 , , 121 , 1241 de breuck , c. , van breugel , w. , stanford , s. a. , rttgering , h. , miley , g. , & stern , d. 2002 , , 123 , 637 bruzual , g. & charlot , s. 2003 , , 344 , 1000 cohen , a. s. , rttgering , h. j. a. , jarvis , m. j. , kassim , n. e. , & lazio , t. j. w. 2004 , , 150 , 417 coleman , g. d. , wu , c .- c . , & weedman , d. w. , 1980 , , 43 , 393 conselice , c. j. , bundy , k. , vivian , u. , eisenhardt , p. , lotz , j. , & newman , j. 2008 , , 383 , 1366 cutri , r. m. , et al .
2003 , 2mass all - sky catalog of point sources ( pasadena : ipac / caltech ) di serego alighieri , s. , et al .
2005 , a&a , 442 , 125 erb d. k. , steidel , c. c. , shapley , a. e. , pettini , m. , reddy , n. a. , & adelberger , k. l. 2006 , , 646 , 107 erben , t. , et al .
2003 , a&a , 410 , 45 frster schreiber , n. m. , et al .
2004 , , 616 , 40 franx , m. , et al .
2003 , , 587 , l79 green , p. j. , et al .
2002 , , 571 , 721 gutirrez , c. m. , et al .
2004 , , 602 , 664 hawkins , m. r. s. , clements , d. , fried , j. w. , heavens , a. f. , veron , p. , minty , e. m. , van der werf , p. 1997
, , 291 , 811 hogg , d. w. , baldry , i. k. , blanton , m. r. , & eisenstein , d. j. 2002 , astro - ph/0210394 kauffmann , g. , white , s. d. m. , & guiderdoni , b. 1993 , , 264 , 201 keeton , c. r. 2001 , astro - ph/0102340 keeton , c. r. , kochanek , c. s. , & falco , e. e. 1998 , , 509 , 561 kochanek , c. s. , keeton , c. r. , & mcleod , b. a. 2001 , , 547 , 50 kochanek , c. s. , falco , e. e. , impey , c. , lehr , j. , mcleod , b. , & rix , h .- w .
2007 , http://www.cfa.harvard.edu/castles/ koopmans , l. v. e. , treu , t. , bolton , a. s. , burles , s. , & moustakas , l. a. 2006 , , 649 , 599 kron , r. g. , koo , d. c. , & windhorst , r. a. 1985 , a&a , 146 , 38 lehr , j. , et al .
2000 , , 536 , 584 miralles , j .- m . , et al . 2002 ,
a&a , 388 , 68 more , a. , mckean , j .-
, muxlow , t. w. b. , porcas , r. w. , fassnacht , c. d. , & koopmans , l. v. e. 2008 , mnras , 384 , 1701 morgan , d. n. , et al .
2000 , , 119 , 1083 mortlock , d. j. , webster , r. l. , & francis , p. j. 1999 , , 309 , 836 muoz , j. a. , falco , e. e. , kochanek , c. s. , lehr , j. , herold , l. k. , fletcher , a. b. , burke , b. f. 1998 , , 492 , l9 oke , j. b. & gunn , j. e. 1983 , , 266 , 713 padmanabhan , n. , et al . 2004 , newa , 9 , 329 papovich , c. , dickinson , m. , & ferguson , h. c. 2001 , , 559 , 620 peng , c. y. , et al .
1999 , , 524 , 572 peng , c. y. , et al .
2006 , , 649 , 616 pirzkal , n. et al .
2001 , a&a , 375 , 351 press , w. h. & schechter , p. 1974 , , 187 , 425 ratnatunga , k. u. , griffiths , r. e. , & ostrander , e. j. 1999 , , 117 , 2010 rengelink , r. b. , tang , y. , de bruyn , a. g. , miley , g. k. , bremer , m. n. , rttgering , h. j. a. , & bremer , m. a. r. 1997 , a&as , 124 , 259 rudnick , g. , et al .
2001 , , 122 , 2205 rusin , d. 2002 , , 572 , 705 rusin , d. , & kochanek , c. s. 2005 , , 623 , 666 russell , j. , ryan jr .
, r. e. , cohen , s h. , windhorst , r. a. , & waddington , i. 2007 , , accepted , astro - ph/0807.2281 de ruiter , h. r. , willis , a. g. , & arp , h. c. 1977 , a&as , 28 , 211 ryan , r. e. , jr . , et al .
2007 , , 668 , 839 saracco , p. , giallongo , e. , cristiani , s. , dodorico , s. , fontana , a. , iovino , a. , poli , f. , & vanzella , e. 2001 , a&a , 375 , 1 spergel , d. n. , et al .
2007 , , 170 , 377 skrutskie , m. f. , et al . 2006 , , 131 , 1163 treu , t. , koopmans , l. v. , bolton , a. s. , burles , s. , moustakas , l. a. 2006 , , 640 , 662 van der marel , r. p. & van dokkum , p. g. 2007 , , 668 , 756 waddington , i. , dunlop , j. s. , peacock , j. a. , & windhorst , r. a. 2001 , , 328 , 882 windhorst , r. a. , miley , g. k. , owen , f. n. , kron , r. g. , & koo , d. c. 1985 , , 289 , 494 windhorst , r. a. 2003 , newa , 47 , 357 white , r. l. , becker , r. h. , helfand , d. j. , & gregg , m. d. 1997 , , 475 , 479 white , s. d. m. & frenk , c. 1991 , , 379 , 52 | we present evidence that the optically unidentified radio source , first j121839.7 + 295325 , may be strongly lensing a background galaxy .
we estimate the redshift of the assumed gravitational arc , discovered in parallel imaging with hst , from mmt - blue channel spectroscopy to be @xmath0 .
we present lens models with an einstein radius of @xmath1 which contains a mass of @xmath2 m@xmath3 , where the uncertainty reflects the range of possible lens redshifts .
the putative lens is not detected to @xmath4 mag and @xmath5 mag in our mmt - swirc imaging . using the flux limits from wfpc2 and swirc
, we estimate that the dynamical mass - to - light ratio of j121839.7 + 295325 is @xmath6 m@xmath3 l@xmath7 for @xmath8 mag , and this lower limit could be as high as @xmath9 m@xmath3 l@xmath7 for @xmath10 mag . since the radio source is optically unidentified ( @xmath11 mag ) and has a radio flux of @xmath12 mjy , it is likely a massive early - type galaxy which hosts a radio - loud agn at @xmath13
. however , the present data can not uniquely determine the mass - to - light ratio of the lensing galaxy , and hence the possibility that this system may be a reasonably dark lens is not ruled out . |
herb wilf , in addition to having done important work on problems related to counting combinatorial objects , has also done pioneering research on algorithms for generating combinatorial objects `` at random '' ( that is , generating an element of a finite combinatorial set so that each element has the same probability of being generated as any other ) ; see @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite and @xcite for fruits of this research .
this survey article describes a recent advance in the area of random generation , with applications to plane partitions , domino tilings , alternating sign matrices , and many other sorts of combinatorial objects .
the algorithm is of the `` random walk '' or `` monte carlo '' variety , but unlike many such algorithms it does not have any initialization bias .
the heart of the algorithm is the method of coupling from the past explored by david wilson and myself in a joint article @xcite . for the sake of readability and motivation
, i will start by focusing on the application of our method to plane partitions .
a beautiful formula of macmahon @xcite says that the number of plane partitions of @xmath0 with at most @xmath1 rows , at most @xmath2 columns , and no part exceeding @xmath3 ( hereafter to be called `` @xmath4-partitions '' ) is given by @xmath5 ( see @xcite and @xcite for definitions of ordinary partitions and plane partitions , and section 2 of @xcite for a fairly simple self - contained proof of macmahon s formula ) .
note that in the case @xmath6 , a plane partition with no part exceeding 1 can be read as the ferrers diagram of an ordinary partition , and macmahon s formula devolves into the assertion that the number of ordinary partitions of @xmath0 with at most @xmath1 parts and no part exceeding @xmath2 is given by the binomial coefficient @xmath7 .
indeed , it is easy to see that such partitions correspond to lattice paths of length @xmath8 joining @xmath9 to @xmath10 , or equivalently , combinations of @xmath8 elements taken @xmath1 at a time . in view of these correspondences
, it is easy to generate a random @xmath11-partition . just as a lattice path in the @xmath1 by @xmath2 rectangle is a 1-complex ( made up of edges in a 2-dimensional grid ) with prescribed boundary ( namely the pair of points @xmath9 and @xmath10 ) , an @xmath4-partition corresponds to a 2-complex ( made up of square 2-cells in a 3-dimensional grid ) whose boundary is a particular non - planar hexagon ( namely the one that goes from @xmath12 to @xmath13 to @xmath14 to @xmath15 to @xmath16 to @xmath17 to @xmath12 in cyclic order ) . in the former case ,
one requires that the lattice path should have ( minimal ) length @xmath8 ; in the latter case , one requires that the surface spanning the hexagon should have ( minimal ) area @xmath18 . in this paper
i will describe an algorithm for generating a random @xmath4-partition with @xmath19 arbitrary .
this algorithm was used to generate figure 1 , which shows a random @xmath20-partition , or rather the spanning surface that it determines , viewed from a point on the ray @xmath21 ; the three different orientations of grid - squares in 3-space are seen in projection as three different orientations of rhombuses in 2-space .
it should be stressed that the size of the plane partition ( that is , the sum of the parts ) was not specified in advance ; it is a random variable with expected value @xmath22 .
note the non - homogeneity of the picture : there is an approximately circular central region in which rhombuses of different orientations are mixed together , surrounded by six regions in which rhombuses of a single orientation predominate . as is shown in @xcite , in a certain strong probabilistic sense the boundary of the central region
does indeed converge to a perfect circle when @xmath0 is large .
this fact was first conjectured on the basis of pictures like figure 1 , and all known proofs depend on steps whose clearest motivation comes from `` knowing the answer in advance '' .
thus we see that there are phenomena pertaining to random @xmath4-partitions that would not have been easy to divine by pure theory , and that an algorithm for generating such plane partitions randomly can be a valuable tool for discovering and investigating such phenomena experimentally . in part 2 of this paper ,
i will show that this problem , along with several others , is a special case of the problem of choosing a random element of a finite distributive lattice . in part 3
, i will describe an algorithm that allows one to solve this problem .
like the well - known folk - algorithm for generating a random element of a 3-element set via independent tosses of an unbiased coin , this algorithm runs in finite expected time , even though it can take arbitrarily long to return an answer .
i conclude in part 4 with some open questions .
the 3-dimensional ferrers diagram associated with an @xmath4-parti- tion is just an order ideal of the poset obtained as the product of chains of cardinalities @xmath1 , @xmath2 and @xmath3 .
writing such chains as * a * , * b * and * c * , respectively , and using @xmath23 to represent the lattice of order ideals of a finite poset ( ordered by inclusion ) , we see that what we are effectively trying to do is generate a random element of the distributive lattice @xmath24 .
( see @xcite for definitions of posets , order ideals , lattices , and distributive lattices , and for the fact that every finite distributive lattice can be represented as @xmath25 for some finite poset @xmath26 . ) here is another way to view this ordering .
consider a hexagon whose internal angles all measure 60 degrees and whose sides in cyclic order have lengths @xmath27 , respectively , and dissect it into @xmath28 unit equilateral triangles .
create a graph @xmath29 whose vertices correspond to these triangles , where two vertices of @xmath29 are adjacent if and only if the corresponding triangles in the dissection share an edge .
( @xmath29 is called a `` honeycomb graph '' for obvious reasons . )
then each @xmath4-partition corresponds to a tiling of the aforementioned hexagon by rhombuses of side length 1 ( each composed of two equilateral triangles joined edge to edge ) , and each of these tilings in turn corresponds to a perfect matching of @xmath29 , that is , to a subset of the edges of @xmath29 in which each vertex of @xmath29 appears just once .
it is shown in @xcite ( which was inspired by earlier work of conway and lagarias @xcite and thurston @xcite ) that the perfect matchings of any bipartite planar graph can be given the structure of a distributive lattice . in the case of the honeycomb - graph @xmath29
described above , one obtains the same distributive lattice structure @xmath24 as before .
if instead one allows graphs @xmath29 that are subgraphs of a square grid , one is able to put a distributive lattice structure on the set of all tilings of a connected subset of the square grid by dominos .
( these special cases are discussed in @xcite . )
more generally ( again see @xcite ) , if @xmath29 is any bipartite planar graph , and @xmath30 is any function from the vertex set of @xmath29 to the non - negative integers , then the set of @xmath31-factors of @xmath29 ( defined as the set of subgraphs of @xmath29 in which vertex @xmath32 has degree @xmath33 for all vertices @xmath32 ) can be given the structure of a distributive lattice ( outside of the trivial case in which no @xmath31-factors exist ) .
another class of examples of distributive lattices in combinatorics comes from constrained lattice paths .
for instance , lattice paths of length @xmath34 that go from @xmath35 to @xmath36 without straying outside of the triangle with vertices @xmath37 , @xmath36 and @xmath36 are a standard incarnation of the `` catalan objects '' .
more generally , we can consider all minimal - length lattice paths that go from one fixed lattice point to another without straying outside some fixed convex region .
assuming that the set of such lattice paths is non - empty , it is easy to show that the set of all such lattice paths is a distributive lattice , with the ordering being defined by inclusion of the associated ferrers diagrams .
one can also extend this approach to the lattice paths enumerated by trinomial coefficients ( with step - vectors @xmath38 , @xmath39 , and @xmath40 ) . a further example is provided by the alternating sign matrices introduced by mills , robbins , and rumsey @xcite .
one can transform each @xmath0 by @xmath0 alternating sign matrix into an @xmath41 by @xmath41 array of numbers in such a fashion that if one takes the component - wise minimum ( or component - wise maximum ) of any two such arrays , one gets another array associated with an @xmath0 by @xmath0 alternating sign matrix . in this fashion ,
the set of alternating sign matrices of fixed size becomes a distributive lattice .
( see @xcite for a more detailed discussion of this . ) a final situation i will mention in which a non - obvious distributive lattice structure exists is the set of independent sets in a bipartite graph @xmath29 .
( this lattice structure has been noticed in the literature several times independently . ) here is the way to see it : if we color the vertices of @xmath29 black and white and write every independent set @xmath42 as @xmath43 where the vertices in @xmath44 and @xmath45 are black and white respectively , then we may define the meet and join of @xmath43 and @xmath46 as @xmath47 and @xmath48 , respectively .
we recall here that every finite distributive lattice admits a unique representation ( up to isomorphism ) as @xmath25 , for @xmath26 some finite poset ( which in fact is easily constructible from the lattice as the sub - poset of join - irreducibles ) . in each of the cases described above ,
the bijection between the combinatorial objects described and order ideals of an associated poset @xmath26 is easily implemented on a computer . thus ,
if we can solve the problem of generating a random element of @xmath25 for an arbitrary finite poset @xmath26 , we will have solved the problem of generating random @xmath31-factors of bipartite planar graphs , the problem of generating random independent sets in a general bipartite graph , and many other problems as well .
in this section i describe a monte carlo approach to sampling from the uniform distribution on the set @xmath25 , using a markov chain whose state are the elements of @xmath25 and whose steady state distribution is the uniform distribution on @xmath25 .
if one merely simulated this markov chain for a large but finite number of steps in the ordinary way , one would have a distribution that was close to the steady - state ( uniform ) distribution , but there would be some residual error ( `` initialization bias '' ) . later in this section
i will explain how one can get rid of this bias by effectively simulating the markov chain for _ infinitely _ many steps , from time minus infinity to time zero ( the method of coupling from the past ) .
it is natural to make @xmath25 into a graph @xmath49 by declaring two order ideals to be adjacent iff their symmetric difference consists of exactly one element of @xmath26 .
it is easy to show that this graph is connected .
moreover , we can associate with each element @xmath50 of @xmath26 a randomization move that preserves the uniform distribution on @xmath25 ( the probability distribution that assigns each order ideal probability @xmath51 ) : given an order ideal @xmath52 , toss a fair coin , and if the coin comes up heads ( resp
. tails ) , replace @xmath52 by @xmath53 ( resp . , @xmath54 )
, unless the resulting subset of @xmath26 is not an order ideal , in which case leave @xmath52 alone . if one performs an infinite sequence of such randomization moves in which each new _ randomization site _
@xmath50 is chosen independently from the uniform distribution on @xmath26 ( or more generally from any probability distribution that assigns positive probability to each element of @xmath26 ) , then this process is simply a stationary markov chain whose state space is the set of order ideals of @xmath26 ; standard ideas from the theory of markov chains guarantee that the probability of any particular order ideal being the current order ideal converges to @xmath51 as time goes to infinity .
this markov chain gives us a way , in principle , of generating a random element of @xmath25 that is as close to unbiased as we like ( i.e. , whose governing distribution is as close to uniform as we like ) .
however , in the absence of estimates of the mixing time of the markov chain , it is not clear for how long a time the chain must be run in order to drive the bias below some predetermined amount deemed acceptable .
moreover , we seek a way of generating samples that has _ no _ bias when truly random bits are used .
fortunately , there is a way around this problem . the markov chain that we have described is monotone in the sense that if we were to run _ two _ instantiations of it in parallel , using the same randomization sites and the same coin - tosses in both runs but starting from different initial order ideals @xmath55 and @xmath56 , then , provided @xmath57 , the order ideals @xmath58 and @xmath59 that result after @xmath0 steps of joint randomization must satisfy the relation @xmath60 . in particular , if we had chosen @xmath55 and @xmath56 to be the empty order ideal @xmath61 and the full order ideal @xmath62 , respectively , and if we find after @xmath0 steps that @xmath58 and @xmath59 are _ equal _ ( call their common value @xmath63 ) , then _ every _ run of the markov chain for @xmath0 steps using those randomization sites and coin - tosses will put us in state @xmath63 , regardless of the initial state @xmath52 ; all such histories are `` squeezed '' between the @xmath64-history and the @xmath65-history , and since these coalesce over the course of the simulation , so must all histories .
it turns out that for most of the examples that arise in practice , this kind of coalescence occurs fairly quickly .
this gives us a way of doing a kind of `` backwards simulation '' of the markov chain that effectively lets us run the chain for a huge number of steps without ( usually ) having to perform anywhere near the number of steps required in a straightforward simulation .
say , for instance , that we want to run the markov chain on the state - space @xmath25 for one million steps , starting from initial state @xmath66 , using random updates from time @xmath67 to time 0 ( we will see shortly why it is convenient to index time in this way ) .
suppose that our markov chain is sufficiently rapidly mixing that over the course of a thousand steps , it s fairly likely for the initial states @xmath64 and @xmath65 to lead to the same final state when evolved in tandem .
if we simulate the markov chain from time @xmath68 to time 0 , using both @xmath64 and @xmath65 as initial states ( let us call this `` phase one '' of our backwards simulation ) , and we indeed find that both histories coalesce at some state @xmath63 at time 0 , then we do not need to run the full simulation from one million steps in the past , for we can already be sure that such a simulation would have given us @xmath63 as our sample .
indeed , in this case we do not even have to choose what the randomization sites and coin - tosses before time @xmath68 are , since they do not enter into the simulation . in the unlikely event that the two histories that were started at time @xmath68 do not coalesce by time 0 ,
then we could go back and do an honest simulation for a million steps , starting from @xmath66 at time @xmath67 ( let us call this `` phase two '' ) . in this way we can simulate the behavior of a million - step random
walk using ( most of the time ) only about two thousand simulation steps .
note however that if one wishes to avoid introducing bias into one s sample , it is crucial that one use the _ same _ randomization sites and coin - tosses from time @xmath68 onward during the long , `` honest '' run ( phase two ) as one did during the short preliminary run ( phase one ) .
if one wants to simulate a _ billion _ steps of random walk , one can modify phase two of the preceding algorithm by running the million - step simulation ( in the rare case where one thousand steps do not suffice ) using @xmath64 and @xmath65 as starting states , rather than @xmath66 . only in the incredibly rare case where this million - step simulation fails to coalesce by time zero
would one need to resort to a billion - step simulation starting from @xmath66 ( `` phase three '' ) , being careful once more to use the already - determined randomization sites and coin - tosses from time @xmath67 to time 0 . note
that the numbers one million and one billion enter into these procedures in the form of bounds on how far back into the past one is willing to go before one `` honestly '' uses @xmath66 as the starting state rather than trying to be clever by starting from @xmath64 and @xmath65 and hoping that they will coalesce by time 0 .
suppose that one removes this upper bound on how far into the past one is willing to go : if coalescence fails , one goes back into the past 1000 times as far as one just did and tries again .
then one can show that with probability 1 the desired coalescence _ will _ eventually occur ( where our notion of `` eventuality '' goes backward in time rather than forward ) , and that the sample returned is effectively a sample `` generated by a run of length infinity '' that is , an absolutely unbiased sample .
this is david wilson s method of coupling from the past .
it can not be overemphasized that the idea of progressing backwards into the past is an indispensable feature of the algorithm . in particular ,
if one were simply to run the markov chain forwards from initial states @xmath64 and @xmath65 in tandem until the histories coalesced and then to output as one s sample the coalescent state , one would in general get a biased sample .
in contrast , samples generated via coupling from the past are entirely free of bias , to the extent that the coin - tosses used are random .
the version of the algorithm discussed above can be improved upon ; for instance , it is much better to go progressively 1 , 2 , 4 , 8 , ... steps into the past rather than @xmath69 , @xmath70 , @xmath71 , ... steps .
moreover , the algorithm can be generalized so as to apply to interesting situations in statistical mechanics where the desired distribution is not uniform , but is the boltzmann distribution for some non - constant energy function on the configuration space .
all of these developments are described more fully in @xcite . here , we content ourselves with showing that the expected running time is finite , in the case where there exists a finite @xmath72 and a positive @xmath73 such that over the course of any time interval of length @xmath72 in the simulation , the conditional probability of coalescence occurring , given initial states @xmath64 and @xmath65 , is at least @xmath73 .
( this holds for all the interesting applications ; in each case , it suffices to find an @xmath72 such that the probability of going from @xmath64 to @xmath65 in @xmath72 steps , conditional upon starting in state @xmath64 , is positive , though the @xmath72 one gets from this approach is much , much larger than the typical time - scale over which coalescence occurs . ) to prove the claim , note that over the course of @xmath74 consecutive steps , divided into @xmath75 blocks of length @xmath72 , the only way in which coalescence could fail to occur is if all @xmath75 blocks are `` non - coalescing '' ; yet independence of the coin - tosses tells us that this is an event of probability at most @xmath76 .
as @xmath0 gets large , this probability shrinks to zero exponentially , implying our claim on the finiteness of expected running time . the algorithm described above is not always efficient ; indeed , in @xcite wilson and i describe a case in which the modified monte carlo algorithm takes exponentially long , even though a direct algorithm for constructing a random order ideal is quite easy to fashion ( see the cautionary note at the end of section 3.2 in @xcite ) .
however , as a practical matter the modified monte carlo algorithm is quite efficient for many sorts of combinatorial objects that arise naturally , such as those mentioned in section 2 .
three final points on implementation deserve note .
first , the algorithm remains valid if instead of choosing randomization sites in accordance with some fixed distribution on @xmath26 one chooses randomization sites according to some other scheme , provided that two conditions are satisfied : the choice of randomization sites should not be affected by the outcomes of the coin flips , and it must be the case that with probability 1 each element of @xmath26 gets chosen as a randomization site infinitely often . under these hypotheses ,
the ( non - stationary ) markov chain on @xmath25 has the uniform distribution as its unique steady - state measure , and the method of coupling from the past will get you there with a finite amount of simulation ( with probability 1 )
. for instance , one may rotate among all the elements of @xmath26 in some fixed order , rather than choosing the randomization sites randomly . in this way one
reduces the amount of information that the algorithm needs to save .
second , if one is using pseudo - random bits given by some trusted pseudo - random number generator ( as i imagine most users of this algorithm will do ) , then , though it is often necessary to _ reuse _ bits , it is not necessary to _ save _ them all .
one can for instance save only seeds that will enable one to re - create the formerly used bits when they are needed again .
third , it should be borne in mind that the output of the monte carlo method is likely to be somewhat correlated to its running - time .
if one were to use the method to generate @xmath77 samples , where @xmath77 was determined on - the - fly rather than chosen in advance , then one might be contaminating one s samples with bias .
the scrupulous investigator may therefore wish to commit to a certain value of @xmath77 ahead of time , based on a preliminary investigation of how long it is likely to take to run the procedure @xmath77 times .
another sort of scheme for randomly generating an order ideal of a finite poset @xmath26 is the recursive approach in which one decides whether or not to include @xmath50 in the random order ideal by tossing a coin whose bias corresponds to the ratio of the cardinalities of two distributive lattices , namely , the lattice of order ideals of @xmath26 containing @xmath50 and the lattice of order ideals of @xmath26 not containing @xmath50 .
this method ( like coupling from the past ) will not always be fast , since the problem of counting antichains ( and hence , equivalently , order ideals ) in a finite partially ordered set is # p - complete ( see @xcite ) .
yet another approach would be one along the lines of recent work of flajolet , zimmermann , and van cutsem ( see @xcite and @xcite ) , who have already created maple software that ( suitably instructed ) can create random plane partitions of the sort shown in figure 1 . _
which way is `` best '' ? _ ( or rather : are there features of a lattice that might dictate when one or another of these approaches will do best ? ) in the case of planar graphs @xmath29 that are not bipartite , it is still possible to generate random perfect matchings of @xmath29 efficiently : the method of kasteleyn @xcite allows one to write the number of perfect matchings as a pfaffian . by calculating the ratio of two such pfaffians
, one can determine the exact proportion of matchings of @xmath29 that contain a given edge , and one can accordingly make an unbiased decision as to whether or not to include this edge in the matching .
applying this recursively , one can generate a random matching of @xmath29 .
surprisingly , wilson @xcite showed that it is possible to randomly generate perfect matchings within a constant multiple of the time needed to compute just one pfaffian ; using sparse linear algebra , this time is @xmath78 arithmetic operations , or @xmath79 bit operations . in the case where @xmath29 is bipartite ,
a specialized version of the pfaffian method ( the permanent - determinant method ) applies in much the same way ; but in this situation one can also apply the markov chain algorithm described in this paper .
_ can the method of this paper be extended to apply in the non - bipartite case ? _
if one uses a biased coin in place of a fair one , one can devise an algorithm in which the probability of a particular order ideal @xmath52 being generated is proportional to @xmath80 , where @xmath81 is the cardinality of @xmath52 and @xmath82 is any positive real number .
that is , we can generate a random element of a distributive lattice so that all the elements of rank @xmath75 in the lattice have individual probability proportional to @xmath83 of being picked .
we might try to choose @xmath82 so as to maximize the collective probability of the elements of rank @xmath75 , but even using this optimal @xmath82 we might have to run the biased procedure many times before we obtain a sample in rank @xmath75 .
note , however , that the resulting sample will be governed by a uniform distribution on the @xmath75th rank of the lattice .
_ is there an efficient procedure for selecting an element in a particular rank of a finite distributive lattice ? _
note that work of flajolet , zimmermann and van cutsem ( @xcite ) provides a solution in some cases .
lastly : _ to what extent can the method of coupling from the past be extended to sampling from more general partially ordered sets , _ such as modular lattices or lattices in general ? for an example of an application of the method to a special kind of non - distributive lattice , see the paragraphs in subsection 3.3 of @xcite that treat permutations .
the article gives the impression that the method of section 3 was used to generate figure 1 .
in fact , a more efficient approach was used , exploiting the special structure of the lattice @xmath25 in question . in particular , the poset @xmath26 of join - irreducibles of the lattice is a cube , and this cube can be divided into `` filaments '' ( where elements @xmath84 belong to the same filament ) . rather than merely attempting to add or delete a particular element of @xmath26 ,
the algorithm attempts to modify @xmath52 by either deleting the largest element of the filament that is in @xmath52 or adjoining the smallest element of the filament that is not in @xmath52 .
putting it differently : the basic steps of the markov chain are `` hexagon - moves '' on the tiling , where one executes a non - trivial hexagon - move by changing the way a small unit hexagon of side - length 1 inside the large region is tiled by three rhombuses ( there are exactly two ways such a hexagon can be tiled ) .
section 3 specifies that one is to choose the successive randomization sites @xmath50 from the uniform distribution on @xmath26 ( or more generally from any probability distribution that assigns positive probability to each element of @xmath26 ) .
however , one has a great deal of leeway here .
for instance , one can use a deterministic sequence of @xmath50 s that cycles through the set of elements of @xmath26 ; as long as the coins used in the algorithm are random , the procedure will still lead to a random order ideal . in the case where @xmath26 is ranked , one particularly natural idea is to do all the @xmath50 s of even rank , followed by all the @xmath50 s of odd rank , and so on , in alternation .
this particular scheme is well - suited to parallel computation , since all the randomizations of the same parity that are considered in any cycle can be done independently of one another .
h. wilf , a unified setting for selection algorithms , ii , _ annals of discrete mathematics _ * 2 * : _ algorithmic aspects of combinatorics _ , north holland ( 1978 ) , 135148 .
h. wilf , the uniform selection of free trees , _
j. algorithms _ * 2 * ( 1981 ) , 204207 . | this survey article describes a method for choosing uniformly at random from any finite set whose objects can be viewed as constituting a distributive lattice .
the method is based on ideas of the author and david wilson for using `` coupling from the past '' to remove initialization bias from monte carlo randomization .
the article describes several applications to specific kinds of combinatorial objects such as tilings , constrained lattice paths , and alternating - sign matrices .
_ this article is dedicated to herbert wilf + in honor of his sixty - fifth birthday . _ |
long gamma - ray bursts ( grbs ) are now considered to be death of massive stars ( so called collapsar scenario ) .
the most convincing observational evidences have been associations with supernovae ( sne ) .
some associations were spectroscopically , others were photometrically . however , there is at least one long grb , grb 060614 ( duration @xmath8 sec , @xcite ) whose sn component was not detectable to very deep limits .
grb 060505 is also a @xmath9 sec burst without a detectable sn @xcite , although its duration @xmath10 4 sec is close to the classical 2 sec threshold which separates the long and short populations of grbs .
various progenitor models have been proposed to explain these events .
some suggested they are short grbs , which are considered to originate in mergers of double compact object binary , with longer duration than typical short grbs @xcite , while others point out similarities of these grbs to other long bursts which originate from collapsars @xcite .
the possibility of new populations which are different from long and short grbs are also suggested for these grbs @xcite .
the sample of long grbs with strong constraints on their sn components is still small , and it is still uncertain what impact do the long grbs without sne have upon our understanding of the grb populations . here we present results of a search for a sn component associated with grb 100418a .
grb 100418a was triggered at 21:10:08 ut april 2010 @xcite .
the duration and the energetics of the prompt emission were @xmath11 second and @xmath12 erg @xcite .
follow - up spectroscopies of the optical transient ( ot ) detected emission - lines and absorption lines with the common redshift of 0.624 @xcite .
@xcite pointed out that the light curve of grb 100418a is similar to previous grbs with sne .
@xcite reported flattening of the optical light curve after 700 ks . throughout this paper
, we assume the fiducial cosmology with the hubble constant of 71 km s@xmath13 mpc@xmath13 , @xmath14 and @xmath15 .
the angular scale is thus 6.8 kpc / arcsec .
we use the ab magnitude system unless otherwise stated .
we observed the field of grb 100418a with focas @xcite on subaru telescope @xcite under a photometric condition ( seeing ) on 2010 may 14 ( ut ) .
we started imaging observations in @xmath16 , and @xmath17 bands at 12:13 ut , and spectroscopic observations followed .
the log of the observations is tabulated in table [ tb : observation ] .
we acquired a single exposure of 120 seconds in each of @xmath16 , and @xmath17 bands with 2@xmath182 binning which yields pixel scale of /pixel .
unbinned @xmath19band images were also taken with /pixel .
spectroscopy was performed with width slit and 3 pixels binned along the spatial direction , and two different settings of grisms and order - cut filter .
one setting covered a wavelength range between 5800 and 10200 with the r300 grism and the o58 order - cut filter , and the other covered between 3900 and 8300 with the b300 grism and no order - cut filter .
the total integration time was 1 hour ( 1200 sec x 3 ) for each of the setting .
we used the atmospheric dispersion corrector , and set the slit position angle to @xmath20 , i.e. , north - south .
the spectral resolution is @xmath21 11 in both settings .
the data were reduced using iraf for the procedures of bias subtraction and flat - fielding .
the spectra were wavelength calibrated , and sky subtracted .
wavelength calibration was performed using night sky emission lines for the red setting and thar arc lines for the blue setting .
and the rms wavelength calibration error is 0.20.3 .
the sensitivity calibration was performed as a function of wavelength by using the spectrum of feige 34 observed with width slit . for imaging data ,
the flux calibration was carried out by observing the standard star pg 1505 - 027 field @xcite at the same night .
( 80mm,80mm)figure1.eps ( 160mm,80mm)figure2.eps the @xmath2band image of /pixel is shown in figure [ fig : r1x1 ] , and the observed spectra are shown in figure [ fig : fullspectra ] .
these spectra are made by summing along the slit in order to include whole galaxy light .
we found strong emission lines such as [ o ] , h@xmath22 , h@xmath23 , [ o ] which are redshifted 0.624 consistently to the previous reports @xcite . [ cols="^,^,^,^ " , ] we use the @xmath24 method described in @xcite , which gives two solutions of possible metallicity ( upper and lower branch ) .
@xmath25 @xcite found [ ne]3869/[o]3727 line flux ratio correlates with metallicity , and thus can be used as a indicator to separate these two solutions .
they showed that galaxies with 12+log(o / h ) @xmath26 8.65 have log [ ne]3869/[o]3727 @xmath27 , while galaxies with 12+log(o / h ) @xmath28 have log [ ne]3869/[o]3727 @xmath29 .
the host galaxy of grb 100418a has log [ ne]3869/[o]3727 = @xmath30 , which suggests that the upper branch solution is the case .
there are several emission line diagnostics proposed to measure metallicity , and their results are not always consistent to each other ( e.g. @xcite ) .
we note that metallicities discussed in @xcite is measuered using the @xcite method for high - metallicity galaxies , and the electron temperature method @xcite for low - metallicity galaxies
. however , the difference between the results of the @xcite method and that of the @xcite method is small in high - metallicity range , and the @xcite method tends to result in higher - metallicity in low - metallicity range compared to other methods @xcite .
hence the difference between the metallicity calibration we use and that used in @xcite would not confuse the separation of the two branches .
we also estimate star formation rate ( sfr ) of the host galaxy from the emission line flux . following h@xmath23sfr relation used in @xcite in which the imf of @xcite is assumed , sfr of the host galaxy is 1.88 @xmath31/yr , or 3.38 @xmath31/yr when re - scaled with the salpeter imf which is consistent to the stellar mass measured in the sed fitting .
the specific sfr @xmath32 gyr@xmath13 is typical of long grb host galaxies with log @xmath33 @xcite .
( 80mm,80mm)figure6.eps ( 80mm,80mm)figure7.eps
we have presented the results of the search for sn component associated with grb 100418a , with focas on subaru telescope .
both imaging and spectroscopic observation show no evidence of sn .
comparison of our @xmath34band magnitude to pre - burst data in sdss catalog puts upper limit on afterglow plus sn luminosity , @xmath35 .
our spectrum shows no sn feature above noise level . assuming that the sn is fainter than the noise level
, we estimate the upper limit on the sn to be @xmath36 ( 44005500 in the restframe ) .
this limit is about 4 magnitude brighter compared to the case of grb 060505 and grb 060614 at @xmath37 , due to higher redshift of grb 100418a ( @xmath38 ) and brightness of the host galaxy .
however , it is still comparable to faintest type ic sne including sn 2002ap which is the faintest broad lined type ic sn ever observed @xcite .
we have also estimated properties of the host galaxy of grb 100418a .
the results of the sed fitting and the emission line diagnostic show that the host galaxy has larger stellar mass and higher - metallicity than majority of long grb host galaxies at @xmath39 ( figure [ fig : hostprop ] ) .
it is notable that among three long grbs with significant limit on their sn component , two bursts occurred in galaxies with 12+log(o / h ) @xmath40 ( grb 060505 and grb 100418a ) .
the other one burst grb 060614 occurred in a very faint galaxy @xcite , and metallicity of the host galaxy is not measured .
@xcite reports possible detection of the sn component of grb 100418a at 28 days after the burst , which is faint and agrees with our upper limit .
the possible sn association suggests that grb 100418a maybe originate from a collapsar .
if their possible detection is a real event , the sn associated with grb 100418a is the faintest sn associated with a grb @xcite , except another possibly detected sn associated with grb 101225a ( @xcite , but see also @xcite ) .
the specific sfr of the host galaxy of grb 100418a is comparable to other long grb host galaxies suggesting collapsar origin of this burst , while short grbs occur in galaxies with lower specific sfr than host galaxies of long bursts @xcite .
however , it should be noted that a short grb ( xrf ) 050416 occurred in a galaxy with similar @xmath41 and sfr to the host galaxy of grb 100418a @xcite . in figure
[ fig : epei ] , we show the spectral peak and the isotropic equivalent gamma - ray energy of grb 100418a @xcite together with those of grb 060505 , 060614 and other grbs at @xmath42 whose host galaxies are metal rich .
grb 100418a and 060614 agree with the @xmath43@xmath44 relation of long grbs by @xcite , while grb 060505 is in upper - left of the relation where most short bursts reside @xcite .
a caveat on this plot is that the @xmath43 of grb 100418a is calculated in @xcite using a relation between photon index in _ swift _ burst alert telescope and @xmath43 @xcite , which is not tested against short grbs .
thus some observational features suggest grb 100418a originates from a collapsar , however none of them are conclusive . the sample of long grbs without bright sne and long grbs in high - metallicity galaxies is still very small , and it is difficult to draw robust conclusions on their nature .
however , localization of a long grb without a bright sn in a high - metallicity galaxy suggests some interesting possibilities .
if long grbs without observable sne originate from collapsars similarly to grbs with hypernove , many of which occur in low - metallicity galaxies .
the possible high - metallicity environments of long grbs without sn features suggest environmental effect on explosion scenario in which a collapsar produce no observable sn ( e.g. , a collapsar with a fallback sn , @xcite ) .
however , it should be noted that metallicity of grb host galaxies maybe systematically different from that of direct environments of grbs @xcite .
@xcite suggested that grbs associated with hypernovae and grbs without sn features belong to a continuous population , predicting the existence of grbs associated with sub - luminous or faint sne with the @xmath45ni mass @xmath46@xmath47 .
the limit on the sn associated with grb 100418a based on the noise spectrum corresponds to 15% of the luminosity of sn 1998bw at the peak of its light curve ( figure [ fig : snlimit ] ) . assuming that the mass of @xmath45ni in sn ejecta is proportional to the luminosity of the sn , the mass of @xmath45ni ejected from grb 100418a is @xmath48 ( the @xmath45ni mass ejected from grb 980425/sn 1998bw is @xmath49 , @xcite ) . in the jet - induced hypernova models by @xcite ,
the ejected amount of @xmath45ni is smaller if the jet formation is slower so that the ram pressure of the jet is lower . if the long grbs without observable sne occur from non - collapsar origin , many of the long grbs that occur in high - metallicity galaxy may be non - collapsar grbs . in the five grbs whose
host galaxies have spectroscopically measured metallicity 12+log(o / h ) @xmath40 at @xmath42 , two are without sn component to significant limit , and the other three do nt have confirmed sn component .
considering the claimed low - metallicity preference of the collapsar model ( @xcite ; see , however , @xcite ) , these observations suggests significant contamination of non - collapsar originate events [ e.g. , grbs originates in merger of double compact object binary , so called short grbs or type i grbs @xcite ] in the sample of @xmath9 sec grbs in high - metallicity galaxies , and possibly also those events in low - metallicity galaxies .
we thank a. de ugarte postigo , g. leloudas and t. sakamoto for helpful discussions .
thanks are also due to our referee whose helpful advice largely improved this paper .
yn is supported by the grant - in - aid for jsps fellows .
amati , l. 2006 , , 372 , 233 amati , l. , della valle , m. , frontera , f. , malesani , d. , guidorzi , c. , montanari , e. , & pian , e. 2007 , , 463 , 913 antonelli , l. a. , et al .
2010 , grb coordinates network , 10620 , 1 baldry , i. k. , & glazebrook , k. 2003 , , 593 , 258 berger , e. 2009 , , 690 , 231 bikmaev , i. , et al .
2010 , grb coordinates network , 10700 , 1 bikmaev , i. , et al .
2010 , grb coordinates network , 10726 , 1 bruzual , g. , & charlot , s. 2003 , , 344 , 1000 caito , l. , et al .
2009 , , 498 , 501 calzetti , d. , armus , l. , bohlin , r. c. , kinney , a. l. , koornneef , j. , & storchi - bergmann , t. 2000 , , 533 , 682 campana , s. , et al .
2011 , , 480 , 69 cucchiara , a. , & fox , d. b. 2010 , grb coordinates network , 10624 , 1 della valle , m. , et al .
2006 , , 444 , 1050 de ugarte postigo , a. , et al . in preparation drout , m. r. , et al .
2011 , , 741 , 97 filgas , r. , klose , s. , & greiner , j. 2010 , grb coordinates network , 10617 , 1 fryer , c. l. , et al .
2007 , , 119 , 1211 fynbo , j. p. u. , et al .
2006 , , 444 , 1047 gal - yam , a. , et al . 2006 , , 444 , 1053 galama , t. j. , et al .
1998 , , 395 , 670 gehrels , n. , et al . 2006 , , 444 , 1044 holland , s. t. , marshall , f. e. , page , m. , de pasquale , m. , & siegel , m. h. 2010 , grb coordinates network , 10661 , 1 iye , m. , et al .
2004 , , 56 , 381 iwamoto , k. , et al .
1998 , , 395 , 672 iwamoto , n. , umeda , h. , tominaga , n. , nomoto , k. , & maeda , k. 2005 , science , 309 , 451 izotov , y. i. , stasiska , g. , meynet , g. , guseva , n. g. , & thuan , t. x. 2006 , , 448 , 955 kashikawa , et al .
2002 , , 54 , 819 kewley , l. j. , & ellison , s. l. 2008 , , 681 , 1183 kobulnicky , h. a. , & kewley , l. j. 2004 , , 617 , 240 levesque , e. m. , & kewley , l. j. 2007 , , 667 , l121 levesque , e. m. , kewley , l. j. , berger , e. , & zahid , h. j. 2010 , , 140 , 1557 lu , y. , huang , y. f. , & zhang , s. n. 2008 , , 684 , 1330 malesani , d. 2010 , grb coordinates network , 10621 , 1 marshall , f. e. , et al .
2011 , , 727 , 132 marshall , f. e. , et al .
2010 , grb coordinates network , 10612 , 1 marshall , f. e. , & holland , s. t. 2010 , grb coordinates network , 10720 , 1 mazzali , p. a. , et al .
2002 , , 572 , l61 mcbreen , s. , et al .
2008 , , 677 , l85 moriya , t. , tominaga , n. , tanaka , m. , nomoto , k. , sauer , d. n. , mazzali , p. a. , maeda , k. , & suzuki , t. 2010 , , 719 , 1445 nagao , t. , maiolino , r. , & marconi , a. 2006 , , 459 , 85 niino , y. 2011 , , 417 , 567 nomoto , k. , mazzali , p. a. , nakamura , t. , iwamoto , k. , danziger , i. j. , & patat , f. 2001 , supernovae and gamma - ray bursts : the greatest explosions since the big bang , 144 ofek , e. o. , cenko , s. b. , gal - yam , a. , frail , d. , kasliwal , m. m. , kulkarni , s. r. , & waxman , e. 2007 , , 662 , 1129 perley , d. a. , et al .
2010 , grb coordinates network , 10727 , 1 richardson , d. 2009 , , 137 , 347 rumyantsev , v. , & pozanenko , a. 2010 , grb coordinates network , 10783 , 1 rumyantsev , v. , & pozanenko , a. 2010 , grb coordinates network , 10883 , 1 sakamoto , t. , et al .
2009 , , 693 , 922 salpeter , e. e. 1955 , , 121 , 161 savaglio , s. , glazebrook , k. , & le borgne , d. 2009 , , 691 , 182 savaglio , s. , et al .
2005 , , 635 , 260 sawicki , m. 2012 , submitted to schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 smith , j. a. , et al .
2002 , , 123 , 2121 soderberg , a. m. , et al .
2007 , , 661 , 982 stetson , p. b. 2000 , , 112 , 925 thne , c. c. , et al .
2011 , , 480 , 72 thne , c. c. , et al .
2008 , , 676 , 1151 tominaga , n. , maeda , k. , umeda , h. , nomoto , k. , tanaka , m. , iwamoto , n. , suzuki , t. , & mazzali , p. a. 2007 , , 657 , l77 tremonti , c. a. , et al .
2004 , , 613 , 898 ukwatta , et al .
2010 , grb coordinates network , 10615 , 1 volnova , a. , msu , s. , ibrahimov , m. , karimov , r. , & pozanenko , a. 2010 , grb coordinates network , 10821 , 1 woosley , s. e. , & weaver , t. a. 1995 , , 101 , 181 woosley , s. e. , & heger , a. 2006 , , 637 , 914 xu , d. , et al . 2009 , , 696 , 971 yabe , k. , ohta , k. , iwata , i. , sawicki , m. , tamura , n. , akiyama , m. , & aoki , k. 2009 , , 693 , 507 yoon , s .- c . , & langer , n. 2005 , , 443 , 643 zhang , b. , et al . 2009 , , 703 , 1696 | we present results of a search for a supernova ( sn ) component associated with grb 100418a at the redshift of 0.624 .
the field of grb 100418a was observed with focas on subaru 8.2 m telescope under a photometric condition ( seeing ) on 2010 may 14 ( ut ) .
the date corresponds to 25.6 days after the burst trigger ( 15.8 days in the restframe ) .
we did imaging observations in @xmath0 , and @xmath1 bands , and two hours of spectrophotometric observations .
we got the resolved host galaxy image which elongated (= 11 kpc ) from north to south .
no point source was detected on the host galaxy .
the time variation of @xmath2band magnitude shows that the afterglow of grb 100418a has faded to @xmath3 without sn like rebrightening , when we compare our measurement to the reports in gcn circulars .
we could not identify any sn feature such as broad emission - lines or bumps in our spectrum . assuming the sn is fainter than the @xmath4 noise spectrum of our observation
, we estimate the upper limit on the sn absolute magnitude @xmath5 in observer frame @xmath6band .
this magnitude is comparable to the faintest type ic sne .
we also estimate host galaxy properties from the spectrum .
the host galaxy of grb 100418a is relatively massive ( log @xmath7 ) compared to typical long grb host galaxies , and has 12+log(o / h ) = 8.75 . |
the paradigm of a quantum spin interacting with its dissipative environment has drawn sustained research interests in a variety of fields including quantum computation @xcite , spin dynamics @xcite , quantum phase transitions @xcite , charge transfer in biological molecules @xcite and impurity effects in magnetic materials @xcite . among the most popular models employed in this
regards is the spin - boson model @xcite that describes a two - level system , i.e. , a spin @xmath2 , coupled linearly to an environment represented by a set of harmonic oscillators .
the coupling between the system and the environment can be characterized by a spectral function @xmath4 .
this model is known to exhibit rich ground state properties . in particular ,
if the bath is characterized by a gapless spectral density @xmath5 , a quantum phase transition is expected to appear , separating a non - degenerate `` delocalized '' phase from a doubly degenerate `` localized '' phase due to the competition between tunneling and environment - induced dissipation . depending on the value of @xmath6 ,
there exist three distinct cases known as sub - ohmic ( @xmath7 ) , ohmic ( @xmath8 ) and super - ohmic ( @xmath9 ) regimes .
recent theoretical studies @xcite show that the transition is of second order in the sub - ohmic regime and kosterlitz - thouless type in the ohmic regime @xcite . in the super - ohmic regime
, however , there is no phase transition . and @xmath10 .
where @xmath6 ( @xmath11 ) and @xmath12 ( @xmath13 ) represent the spectral exponents and coupling strengths , respectively , for the spectral density functions @xmath14 ( @xmath15 ) .
three different phases ( localized , critical and free ) are displayed in the @xmath12-@xmath6 plane with two critical values of the spectral exponents , @xmath16 and @xmath17 . ]
a number of studies have investigated extensions of the standard spin - boson model , for example , to a two - spin system involving a common bath @xcite or two independent baths @xcite , and to a single spin coupled to a bath with simultaneous diagonal and off - diagonal coupling @xcite .
we have recent studied a two - bath spin - boson model @xcite , shown schematically in fig .
[ f0](a ) ] , where @xmath18 and @xmath19 denote the diagonal and off - diagonal coupling , respectively , and the arrow represents a spin .
the bath spectral densities can be described by @xmath20 , where @xmath12 and @xmath13 are the dimensionless coupling strengths , and @xmath6 and @xmath11 denote the spectral exponents characterizing the two baths coupled to the spin diagonally and off - diagonally , respectively .
possible realizations of such two - bath model include impurities in a magnet coupled to two spin - wave modes or two sources of dissipation @xcite , excitonic energy transfer processes in natural and artificial light - harvesting systems @xcite , electromagnetic fluctuations of two linear circuits attached to a superconducting qubit @xcite , two cavity fields coupled to a squid - based charge qubit @xcite , and the process of thermal transport between two reservoirs coupled with a molecular junction @xcite . in the two - bath model with @xmath21
, studies based on the perturbative renormalization group theory predict the presence of two phases , namely , the `` critical phase '' and the free phase , " in the absence of bias and tunneling @xcite .
very recently , the existence of a localized phase " in the two - bath model was discovered numerically in the strong coupling regime @xcite .
the schematic of the phase diagram that emerges is shown in fig .
[ f0](b ) . a continuous quantum phase transition separating the localized phase from the critical phase
was claimed to exist only for the spectral exponent @xmath22 , and a critical value of the spectral exponent , @xmath23 , was estimated from the density matrix renormalization group ( dmrg ) calculations .
when @xmath9 , the impurity behaves as a free spin in the so - called free phase @xcite .
the phase boundary was determined from the response to the external field ( i.e. , the bias or tunneling ) perpendicular to the bath plane .
however , the localized - to - delocalized phase transition will occur under the external field , which renders the phase diagram very complicated .
still unclear is the influence of the external field to the localized - to - critical transition .
moreover , the critical value of the spectral exponent was predicted by a recent mean - field analysis @xcite to @xmath24 , that stands at variance to the aforementioned dmrg result .
it thus remains a challenging task to map out precisely the localized - to - critical phase transition as represented in fig .
[ f0](b ) in the absence of an external field . in the absence of bias and tunneling
, the two - bath model exhibits a high level of symmetry , including the parity symmetry and rotational symmetry @xcite . in the localized phase , spontaneous symmetry breaking takes place due to strong spin - bath coupling .
hence , a symmetry analysis may help distinguish the critical and localized phases .
in addition , a novel quantum phase transition from a doubly degenerate `` localized phase '' to another doubly degenerate `` delocalized phase '' is uncovered with respect to the ratio of the coupling strengths @xmath25 within the two baths @xcite .
the transition is inferred to be of the first order , and the transition point @xmath26 is determined when the spectral exponents of the two baths are identical . since the system at the transition point corresponds to the @xmath27-symmetric spin - boson model , the critical properties of the ground state , i.e. , the spin polarization @xmath28 and generalized susceptibility @xmath29 at @xmath30 and @xmath10 , can also be used to distinguish the localized and critical phases .
the purpose of this paper is to investigate various ground - state phases in the extended spin - boson model involving two identical , independent baths , and determine the critical value of the spectral exponent @xmath16 separating the localized and critical phases in the weak coupling limit of @xmath31 . via the variational approach , the dmrg approach , and the exact diagonalization method , we conduct a comprehensive study on the ground state properties of the two - bath spin - boson model with zero bias and tunneling for the baths described by single mode as well as continuous spectral density function . in this work ,
rotational and parity symmetry breaking is found to occur along the localized - to - critical phase transition , and the phase boundary is obtained with @xmath32 , consistent with the mean - field predictions .
the rest of the paper is organized as follows . in sec .
ii , the two - bath spin - boson model and its symmetry properties are described , along with an introduction to the variational approach .
iii and iv present the numerical results for the localized and critical phases in the two - bath spin - boson model involving a spin coupled to two oscillators or two baths described by a continuous spectral density function , respectively .
conclusions are drawn in the final sec
the two - bath spin - boson model can be described by the hamiltonian below @xmath33 where @xmath34 and @xmath35 is the spin bias and tunneling constant , respectively , @xmath36 is the index of the baths , and @xmath37 ( @xmath38 ) is the diagonal ( off - diagonal ) coupling strength . in order to investigate the quantum phase transition between the critical and localized phases , we focus on the case of @xmath39 as mentioned in the ` introduction ' .
a logarithmic discretization procedure is adopted by dividing the phonon frequency domain @xmath40 $ ] into @xmath41 intervals @xmath42 $ ] ( @xmath43 @xcite .
the coupling strengthes @xmath44 and @xmath37 ( or @xmath38 ) in eq .
( [ ohami ] ) can then be calculated as @xmath45 for convenience , the frequency cut off @xmath46 and the discretization factor @xmath47 are set throughout this paper .
it should be noted that infinite bath modes are considered via the integration of the continuous spectral density @xmath4 , although the number of effective bath modes @xmath41 is finite . since various values of @xmath48 and @xmath49 are possible due to the extended symmetry of the two - bath model with zero bias and tunneling , the spin polarization is introduced as @xmath50 due to hamiltonian invariance under the transformation @xmath51 , the @xmath52 component of the spin polarization is @xmath53 .
hence , @xmath54 can be simplified to be @xmath55 . according to ref.@xcite
, the critical phase is characterized by @xmath56 @xmath57 , which results in @xmath58 .
in addition , the phonon population @xmath59 is introduced to depict the ground state of the two - bath model .
assuming that the wave function of the ground state can be written as @xmath60 where @xmath61 and @xmath62 are the phonon parts of the wave function corresponding to the spin up and down states , respectively , which can be expanded in a series of fock states or coherent states .
the phonon population @xmath59 is thus defined as @xmath63 where @xmath64 represents the trace of the spin freedom in the wave function , @xmath65 and @xmath66 are coordinates in @xmath67 plane corresponding to the off - diagonal and diagonal coupling baths , and @xmath68 is the phonon - component of wave function in the two - dimensional coordinate representation @xmath69 .
the model studied in this paper exhibits a high level of symmetry due to zero bias and tunneling ( @xmath70 ) .
a group theory analysis @xcite shows that the ground state is always doubly degenerate .
we introduce four parity symmetry operators including @xmath71 as the unit operator , @xmath72 and @xmath73 .
the influence of the parity symmetry operations to the ground state @xmath74 is displayed in fig .
[ symmetry](a ) . under the operation @xmath75 ( @xmath76 ) , the sign of the coordinate values corresponding to the displacements of phonons in the diagonal coupling bath ( off - diagonal coupling bath )
will be changed .
the symmetry parameters @xmath77 and @xmath78 of the parity symmetry are defined as @xmath79 when @xmath80 , the results @xmath81 and @xmath82 ( @xmath83 and @xmath84 ) are obtained for the localized phase ( delocalized phase ) in the two - bath spin boson @xcite .
however , @xmath77 and @xmath78 in the case of @xmath10 are still unclear .
a vanishing value of @xmath85 usually indicates breakdown of the parity symmetry .
in contrast , one has @xmath86 when the ground state has perfect parity symmetry along the x or z direction . if @xmath87 , the ground state exhibits partial parity symmetry which may be induced by the numerical errors or the finite number of the degrees of freedom . and rotational symmetry operator @xmath88 to the ground state @xmath74
are displayed in ( a ) and ( b ) , respectively . for convenience
, we use the polygon shape of the ground states to emphasize the influence of symmetry operations . @xmath89 and @xmath90 are new ground states under these symmetry operations , and @xmath91 represent the coordinate values of the ground state in the @xmath18 and @xmath19 directions , respectively . , title="fig : " ] + and rotational symmetry operator @xmath88 to the ground state @xmath74 are displayed in ( a ) and ( b ) , respectively . for convenience , we use the polygon shape of the ground states to emphasize the influence of symmetry operations . @xmath89 and @xmath90 are new ground states under these symmetry operations , and @xmath91 represent the coordinate values of the ground state in the @xmath18 and @xmath19 directions , respectively . , title="fig : " ] in the @xmath27-symmetric spin - boson model with @xmath30 and @xmath10 , the system may exhibit a rotational symmetry , since the hamiltonian is invariant when one simultaneously rotates the spin and the two baths in the @xmath18-@xmath19 plane by an arbitrary angle @xmath92 . according to the abelian @xmath93 symmetry of the two - bath model proposed in ref .
@xcite , the rotational symmetry operator @xmath88 is introduced as @xmath94 where @xmath95 is the generator of the @xmath93 symmetry defined as @xmath96 figure .
[ symmetry](b ) shows the influence of the rotational symmetry operator on the ground state @xmath74 , where the coordinate values of the ground state @xmath91 in the @xmath18 and @xmath19 directions are proportional to the displacement coefficients of the phonons in the off - diagonal coupling and diagonal coupling baths , respectively .
the symmetry parameters @xmath97 and @xmath98 are introduced to quantitatively measure the rotational symmetry as @xmath99|\psi_g\rangle.\end{aligned}\ ] ] it should be noted that only the rotation of the two baths is considered in the definition of @xmath98 .
the system energy @xmath100 is expected to be independent of the rotational angle @xmath92 in the two - bath spin - boson model involving two identical baths .
if the ground state has perfect rotational symmetry , one has @xmath101 for the whole regime of rotational angle @xmath92 .
in contrast , @xmath102 is obtained in the absent of the rotational symmetry , where @xmath103 is a delta function .
when the ground state has partial rotational symmetry , @xmath98 decays with the rotational angle @xmath92 .
furthermore , we investigate the @xmath92-dependent behavior of the parity - symmetry parameters @xmath104 defined as @xmath105 where @xmath106 defined in eq .
( [ order_rotation ] ) is the phonon part of the rotational operator , and @xmath107 is one of the degenerate ground states obtained by rotating the ground state @xmath108 . in the rest of the paper ,
both symmetry parameters , @xmath98 and @xmath104 of the rotational and parity symmetries , respectively , will be comprehensively studied , as they are useful and sensitive to detect the spontaneous symmetry breaking in the localized - to - critical phase transition .
a systematic coherent - state expansion of the ground state wave function , termed as the `` multi-@xmath0 ansatz , '' is introduced as the variational trial ansatz @xcite .
it can be written as @xmath109 |0\rangle_{\textrm{ph } } , \label{vmwave}\end{aligned}\ ] ] where h@xmath110c@xmath110 denotes hermitian conjugate , @xmath111 ( @xmath112 ) stands for the spin up ( down ) state , @xmath113 is the vacuum state of the phonon bath , and @xmath41 and @xmath114 represent the numbers of the bath modes and coherent superposition states , respectively .
in fact , eq . ( [ vmwave ] ) describes a superposition of the spin states @xmath115 that are correlated with the bath modes with displacements @xmath116 and @xmath117 , where @xmath118 and @xmath119 represent the ranks of the coherent superposition state and effective bath mode , respectively .
the displacements ( @xmath120 ) with @xmath121 ( @xmath122 ) correspond to the phonons in the diagonal ( off - diagonal ) coupling bath . using this trial wave function ,
the system energy @xmath123 can be calculated with the hamiltonian expectation @xmath124 and the norm of the wave function @xmath125 as @xmath126 .
the ground state is then obtained by minimizing the energy with respect to the variational parameters @xmath127 and @xmath117 . the variational procedure entails @xmath128 self - consistent equations , @xmath129 where @xmath130 denote the variational parameters .
the `` multi-@xmath0 '' ansatz is much more sophisticated and contains more flexible variational parameters than the silbey - harris ansatz @xcite and nazir s ansatz @xcite , where only @xmath131 and @xmath132 variational parameters are employed , respectively .
for example , if @xmath133 and @xmath134 , our new ansatz has @xmath135 variational parameters , compared to @xmath136 parameters in the silbey - harris ansatz and @xmath137 parameters in nazir s ansatz . for each set of the coefficients ( @xmath138 and @xmath11 ) in the continuous spectral densities @xmath15 and @xmath14 , more than @xmath139 initial states are used in the iteration procedure with variational parameters ( @xmath140 ) uniformly distributed within an interval @xmath141 $ ] .
displacement coefficients @xmath142 of the initial states obey the classical displacements , i.e. , @xmath143 for the diagonal coupling bath and @xmath144 for the off - diagonal coupling bath . in the single - mode case , @xmath120 , @xmath145 and @xmath146
are all initialized randomly . after preparing the initial state , a relaxation iteration technique @xcite
is adopted , and a simulated annealing algorithm @xcite is also employed to improve the energy minimization procedure . the iterative procedure is carried out until the target precision of @xmath147 is reached . theoretically , the number of coherent superposition states @xmath148 is required for the completeness of the environmental wave function in variational method .
however , large values of @xmath114 pose significant challenges in carrying out numerical simulations . to obtain reliable numerical results with large @xmath114 , an approach to improve the variational algorithm
is undertaken based on the parity symmetry .
assuming @xmath108 is the ground state obtained by the variational method with @xmath114 coherent superposition states , an intermediate state @xmath149 can be generated via the parity symmetry operators @xmath150 and @xmath73 , @xmath151 where @xmath152 is the weight coefficient . according to the symmetry analysis @xcite , the symmetry operator @xmath75 or @xmath76 can lead to the other branch of the doubly degenerate ground state with the same energy @xmath153 . hence these four symmetry operators should be equally weighted with @xmath154 . if @xmath155 , @xmath156 and @xmath157 can be derived .
similar to the case of the delocalized davydov @xmath158 variational ansatz in the holstein model @xcite , the energy @xmath159 of the intermediate state is lower than @xmath160 after considering the parity symmetry .
using these eight states as initial states , one can obtain a new ground state @xmath108 by performing the variational procedure with @xmath161 coherent superposition states , which yields a lower ground state energy @xmath162 .
due to numerical errors , however , the state @xmath163 found by the variational algorithm corresponds only to the local minimum in energy in the vicinity of the ground state . to refine the variational results , the rotational symmetry should also be considered in the case of @xmath30 and @xmath164 .
via the rotational operator @xmath88 acting onto the state @xmath165 , a subspace composed of a series of states with respect to the rotational angle @xmath92 is built .
subsequently , the state with the minimum energy in this subspace is regarded as the ground state .
since the generater of the @xmath93 symmetry @xmath95 involves a hopping between the diagonal and off - diagonal coupling baths , the displacement coefficients in the two baths are identical after considering the rotational symmetry , consistent with the argument that the ground state is accompanied by a symmetric distribution of phonon numbers in the diagonal and off - diagonal coupling baths . the phonon population @xmath59 in eq .
( [ density ] ) can be calculated with the multi-@xmath158 variational ansatz in eq .
( [ vmwave ] ) as @xmath166 ^ 2+\left[b_ng_{n}(x , z)\right]^2}{d } , \label{vm_density}\ ] ] where @xmath167 is the norm of the wave function , @xmath145 and @xmath146 denote the weight coefficients of the @xmath118-th coherent superposition state coupled to the spin up and down states , and the phonon functions @xmath168 and @xmath169 represent the phonon component of the wave function @xmath170 in the two - dimensional coordinate representation @xmath69 .
the function @xmath171 denoting a coherent state in the off - diagonal coupling bath can then be deduced , @xmath172 where @xmath173 and @xmath174 are defined as @xmath175 in the same way , the functions @xmath176 and @xmath177 can also be calculated with the displacement coefficients @xmath116 and @xmath117 in eq .
( [ vmwave ] ) as input . in the single - mode case ,
i.e. , @xmath178 , the phonon function can be simplified as @xmath179 where the subscripts @xmath180 and @xmath181 correspond to the diagonal and off - diagonal coupling oscillators , respectively .
the ground state of the model involving two oscillators coupled diagonally and off - diagonally to a spin is investigated in this section .
the corresponding hamiltonian can be written as @xmath182 where @xmath183 and @xmath184 are diagonal and off - diagonal coupling constants , respectively .
it is the simplest version of the two - bath model under current study .
furthermore , we focus on the case of two identical coupling constants @xmath185 as it gives the hamiltonian the rotational symmetry , which may provide some simple insights on the nature of the phase transition between the critical and localized phases . in the exact diagonalization procedure ,
the phonon states @xmath61 and @xmath62 corresponding to the spin up and down states , respectively , are expanded in a series of fock states , @xmath186 and @xmath187 are the coefficients of the fock state @xmath188 for the two oscillators coupled diagonally and off - diagonally to the spin , and @xmath189 is the bosonic truncation number defined as the cutoff value of the phonon occupation number .
we have verified that this value of @xmath190 is sufficiently large for the ground - state energy to converge . solving the schrdinger equation , one can obtain the wave function of the ground state @xmath108 with a series of coefficients , @xmath191 and @xmath187 , and the ground state energy @xmath153 .
thus , the expectation values of @xmath192 and @xmath193 can be calculated as @xmath194 the phonon population @xmath59 can also be obtained by @xmath195 ^ 2 + \left[d_{k_1,k_2}d_{k_1,k_2}(x , z)\right]^2\right),\end{aligned}\ ] ] where @xmath196 and @xmath197 represent the fock states @xmath188 in the coordinate representation @xmath198 , corresponding to the spin up and down states , respectively .
the characteristics of the single - mode spin - boson model involving two oscillators can be investigated intuitively in coordinate representation with the transformation @xmath199 and @xmath200 . in the following discussion ,
we use @xmath65 and @xmath66 as the classical counterparts of the corresponding operators @xmath201 and @xmath202 .
the hamiltonian is then described by @xmath203 where we denote @xmath198 , @xmath204 , @xmath205 , @xmath206 , and @xmath207 is the two - dimensional laplace operator written in polar coordinates .
this hamiltonian describes a spin in a two - dimensional harmonic potential with spin - orbital coupling @xmath208 . when @xmath209@xmath210 , the spatial motion of the particle is too slow compared to the degree of freedom of the spin .
therefore , it is justifiable to introduce the born - oppenheimer approximation .
the spatial motion is thus treated classically , and the spin - part is solved by @xmath211 with the adiabatic eigenstates @xmath212,\nonumber \\
\left|\eta_{-}\right\rangle & = & \left[\begin{array}{c } \sin\left(\frac{\pi}{4}-\frac{\theta}{2}\right)\\ -\cos\left(\frac{\pi}{4}-\frac{\theta}{2}\right ) \end{array}\right ] , \end{aligned}\ ] ] and the corresponding eigenvalues @xmath213 .
here we define @xmath214 .
then , the wave function of the system can be assumed , @xmath215 using the adiabatic eigenstate as the basis , the equations for the spatial part of the wave function are obtained with the stationary schrdinger equation @xmath216 , @xmath217 where @xmath218 , @xmath219 , @xmath220 and @xmath221 are the pauli matrixes , and the non - adiabatic terms are given by the operator @xmath222 which are only in the angular direction , and can be neglected in further analysis . following the standard procedure of variables separation method
, the solution has the following form @xmath223 wherein the radial function @xmath224 is determined by the equation @xmath225r_{\pm}\left(r , m\right ) & = & e_{\pm}r_{\pm}\left(r , m\right ) \nonumber \\ & & \label{eq : e4}\end{aligned}\ ] ] with the effective potential @xmath226 the effective potential contains a harmonic potential , a linear potential and a centrifugal potential , and the angular quantum number is half - integer @xmath227 due to the contribution of the spin @xmath2 part .
thus , the ground state that corresponds to @xmath228 is doubly degenerate . in the strong coupling regime ( @xmath229 )
, we can neglect the centrifugal potential , leading to the expectation of a ring - shaped ground state with the radius @xmath230 . and
@xmath231 is displayed in two - dimensional coordinate representation @xmath232 .
the @xmath65- and @xmath66-coordinate correspond to off - diagonal and diagonal coupling baths , respectively , and the colour represents the phonon population @xmath59 . in ( a ) and ( c ) , the numbers of coherent superposition states @xmath233 and @xmath234 are adopted , respectively , and an intermediate state defined in eq .
( [ intermediate ] ) is shown in ( b ) . ] and @xmath231 is displayed for various numbers of the coherent superpositions @xmath235 and @xmath236 in ( a)-(f ) , respectively .
the @xmath65- and @xmath66-coordinate correspond to off - diagonal and diagonal coupling baths , respectively , and the colour represents the phonon population @xmath59 . ]
we first investigate the ground state of the two - bath model in the case of @xmath231 and @xmath237 via the variational method with @xmath233 and @xmath178 .
according to the aforementioned theoretical arguments , the wave function of the ground state is expected to be ring shaped in the two - dimensional @xmath232 coordinate representation .
however , variational results depict only a quarter of the ring as shown in fig .
[ f6](a ) , where the colour represents the value of the phonon population @xmath59 defined in eq .
( [ vm_density ] ) , and @xmath65- and @xmath66-coordinates correspond to the off - diagonal and diagonal coupling baths , respectively . using the parity symmetry operators onto the ground state , an intermediate state defined in eq .
( [ intermediate ] ) is obtained and shown in fig .
[ f6](b ) .
the energy of this intermediate state , @xmath238 , is found to be slightly below the ground state energy @xmath239 . taking this intermediate state as an initial state , one can seek the ground state via the variational method with @xmath240 .
figure [ f6](c ) shows the phonon population @xmath59 decreasing smoothly with the @xmath66-coordinate , quite different from that of the intermediate state .
it indicates that the parity symmetry in the @xmath66 direction is broken , resulting in doubly degenerate ground states with different values of @xmath241 but the same ground state energy @xmath242 that is much lower than both @xmath243 and @xmath159 . and
@xmath231 is displayed in ( a)-(d ) for various numbers of the coherent superposition states @xmath244 and @xmath245 .
the @xmath65- and @xmath66-coordinate correspond to off - diagonal and diagonal coupling baths , respectively , and the colour reflects the value of the phonon population @xmath59 . ] and @xmath246 , respectively . the phonon frequency @xmath231 is set for both two baths .
the @xmath65- and @xmath66-coordinate correspond to off - diagonal and diagonal coupling baths , respectively , and the colour represents the phonon population @xmath59 . ]
further , the convergence of ground state with respect to the number of the coherent superposition states @xmath114 warrants a careful examination . as shown in fig .
[ f7 ] , the phonon population @xmath59 gradually starts to resemble a ring - like shape as @xmath114 is increased from @xmath247 to @xmath248 and @xmath236 .
in fact , the @xmath18-@xmath19 symmetric spin - boson model exhibits continuous degeneracy by the projecting @xmath249 , where @xmath88 is the rotational symmetry operator defined in eqs .
( [ rotation ] ) and ( [ generator ] ) , and @xmath108 is one branch of the ground state .
moreover , the ground state energy @xmath160 monotonically decreases with @xmath114 and is convergent to an asymptotic value @xmath250 , consistent with the exact diagonalization result @xmath251 with the phonon truncation number @xmath189 .
the ground state of the two - bath model in the case of weaker coupling , @xmath252 and @xmath231 , is investigated next .
the phonon distribution @xmath59 for @xmath244 and @xmath245 is displayed in fig .
[ f8](a)-(d ) .
different from the results shown in fig .
[ f7 ] , the shape of the ground state remains nearly unchanged , indicating that a small value of @xmath114 is sufficient to obtain a reliable numerical result .
the ground state energy @xmath253 is again in an excellent agreement with the exact diagonalization result @xmath254 .
finally , the ground states of two - bath model with @xmath255 and @xmath246 are also plotted in fig .
[ f10](a ) and ( d ) , respectively , to facilitate comparison with those for @xmath252 in fig .
[ f10](b ) and @xmath256 in fig .
[ f10](c ) . in the weak coupling regime @xmath257 , a ground state with a clear rotational symmetry
is found , while it collapses in a corner of the @xmath18-@xmath19 plane in the strong coupling regime @xmath258 .
it supports our conjecture that the rotational symmetry breaks when the coupling strength exceeds a certain value @xmath259 , similar to the picture of the phase transition in the classical xy model . since the radius of the circle in fig .
[ f10](a ) is quite small , any slight shift of the center from the coordinate origin @xmath260 will induce a sharp jump in the spin polarization from @xmath58 to @xmath261 .
it indicates that the spin polarization @xmath54 is unstable in the weak coupling regime , corresponding to the free phase . that the ground state in fig .
[ f10](c ) shows a crescent profile rather than a complete ring may be reasoned from previous analytical results .
the ground state here must be doubly degenerate , and our numerical calculations yield only one branch of the ground state .
upon combining both the degenerate sates , once can readily obtain the complete ring shape of the ground state . and @xmath252 , respectively .
correspondingly , the difference between the exact diagonalization and variational results are displayed in ( b ) and ( d ) .
the @xmath65- and @xmath66-coordinate correspond to off - diagonal and diagonal coupling baths , respectively , and the colour represents the phonon population @xmath59 . ] the ground sates obtained by the exact diagonalization method are shown in figs .
[ f9](a ) and [ f9](c ) for the two cases of @xmath256 and @xmath252 , respectively .
results for both of these cases seem to be identical to those with the variational results shown in figs .
[ f7](f ) and [ f8](d ) . to further verify the consistency ,
the difference between exact diagonalization and variational results are displayed in figs .
[ f9](b ) and ( d ) .
the resulting difference is two orders or magnitude smaller than the phonon population @xmath262 , implying that the wave function obtained by the two methods are nearly the same , thereby lending support to the superior accuracy of our variational results . .
the ground state energy @xmath153 and spin polarization @xmath54 obtained by the variational method ( vm ) and exact diagonalization ( ed ) are listed for three different cases with the diagonal and off - diagonal coupling strengths @xmath263 and @xmath264 , respectively
. the phonon frequency @xmath231 is set , and @xmath114 and @xmath190 denote the coherent - superposition number in the variational method and the bosonic truncated number in the exact diagonalization , respectively . [ cols="<,^,^,^,^",options="header " , ] [ t2 ] the convergence of the variational results with respect to @xmath114 and @xmath41 is carefully tested .
[ iden_2 ] shows the ground - state energy @xmath160 as a function of @xmath114 in the critical phase with @xmath265 and @xmath266 .
a power law decay of the ground state energy with the form @xmath267 is found via numerical fitting , which yields the asymptotic value @xmath268 . in the inset , the spin polarization @xmath269 is also displayed as a function of @xmath114 on a log - log scale .
a perfect power - law behavior of @xmath269 is obtained with the slope @xmath270 and the asymptotic value @xmath271 .
it suggests that the non - zero value of the spin polarization in the critical phase originates in the effects of the finite @xmath114 .
furthermore , the spin polarization @xmath269 in the critical phase is not convergent even for @xmath272 , unlike the case of the localized phase where a small value of @xmath114 is sufficient to obtain reliable results . in a similar manner ,
the influence of @xmath41 to the ground state energy @xmath273 and spin polarization @xmath274 is also depicted in fig .
[ iden_2_mod ] .
both quantities are found to reach asymptotic values when @xmath275 , indicating the sufficiency of @xmath134 .
therefore , in the following discussion on the variational results the number of coherent superposition states and the bath modes are set to @xmath133 and @xmath134 , unless specified otherwise . and
its @xmath65 and @xmath66 components @xmath48 and @xmath241 are plotted as a function of the rotational angle @xmath92 for the states @xmath276 in the case of @xmath265 and @xmath266 .
the dash lines represent the fitting with the trigonometric functions . in the inset
, the shift @xmath277 from the ground state energy is shown . ]
[ t2 ] presents a comparison between the numerical results obtained by the variational method and the dmrg approach @xcite . to ensure reliable results ,
the phonon number used in dmrg algorithm is @xmath278 , much larger than @xmath279 used in the previous work @xcite .
the length of wilson chain is set to @xmath280 and the cutoff dimension of the matrix is @xmath281 . in the three cases of ( @xmath282 ) , ( @xmath283 ) and ( @xmath284 ) ,
only a slight difference of the ground state energy , i.e. , @xmath285 , is found between the variational and dmrg results , further reinforcing the superiority of our variational results .
since the ground state of the critical phase is unstable @xcite , the spin polarization obtained by the variational method is larger than that by the dmrg , as shown in the last two columns in table .
[ t2 ] . in the localized phase , however , a small value of @xmath133 is sufficient to obtain the variational result of the spin polarization @xmath286 , comparable with the dmrg result @xmath287 .
is displayed as a function of @xmath288 for the two cases of @xmath289 ( bottom ) and @xmath290 ( top ) when the coupling strengths are @xmath291 .
the width of the peak @xmath292 is defined as the size of the parity - symmetry regime with @xmath293 . ] figure .
[ iden_3 ] shows the spin polarization @xmath294 and its @xmath65 and @xmath66 components @xmath48 and @xmath241 for the states @xmath276 , where @xmath88 is the rotational symmetry operator defined in eq .
( [ rotation ] ) and @xmath108 is the ground state obtained by the variational method . as the rotational angle @xmath92 increases , the values of @xmath48 and @xmath241 oscillate between @xmath295 and @xmath296 , while the corresponding spin polarization @xmath54 remains almost unchanged .
the obtained curves can be fitted with trigonometric functions , @xmath297 and @xmath298 with @xmath299 .
it indicates that neither @xmath48 nor @xmath241 is a good candidate to characterize the localized - to - critical phase transition , even though they were employed in ref.@xcite . in the inset ,
the shift @xmath300 is plotted .
the sufficiently small value of @xmath301 shows that there are continuous degenerate ground states which have the same energy @xmath153 , consistent with the prediction from the rotational symmetry analysis .
, is displayed in ( a ) as a function of @xmath6 for various values of @xmath12 and in ( b ) as a function of @xmath12 for various values of @xmath6 .
the transition point separating the localized phase from the critical phase is located at the position where the parity index @xmath302 reaches .
, title="fig : " ] + , is displayed in ( a ) as a function of @xmath6 for various values of @xmath12 and in ( b ) as a function of @xmath12 for various values of @xmath6 .
the transition point separating the localized phase from the critical phase is located at the position where the parity index @xmath302 reaches .
, title="fig : " ] apart from the spin polarization , we have also probed the symmetry properties of the ground state .
the symmetry parameter of the parity symmetry , @xmath303 where @xmath77 and @xmath78 are defined in eq .
( [ order_parameter ] ) , is displayed in fig .
[ iden_4 ] for the two cases of @xmath289 and @xmath304 at a sufficiently small coupling strength @xmath291 . in the localized phase with @xmath305 ,
sharp peaks of @xmath104 are found at @xmath306 ( @xmath307 and @xmath247 ) with a small peak width @xmath292 defined as the size of the parity - symmetry regime @xmath308 .
it suggests that the ground state is localized in a corner of the @xmath18-@xmath19 plane . on the other hand
, @xmath104 is always greater than zero in the critical phase with @xmath309 , indicating that the parity symmetry covers the whole subspace .
hence , the parity index @xmath310 reflecting the localization of the ground state can be used to quantitatively distinguish the localized and critical phases .
is displayed against the rotational angle @xmath288 for the two cases of @xmath311 and @xmath312 , corresponding to the critical and localized phases , respectively . ]
shown in fig .
[ iden_5](a ) is the parity index @xmath313 versus the spectral exponent @xmath6 for various values of @xmath12 , in the case with two identical baths , i.e. , @xmath30 and @xmath10 . without any loss of generality
, it is assumed that only the ground state with the parity index @xmath314 belongs to the critical phase , otherwise it belongs to the localized phase . according to this criteria , the transition point @xmath315 between the localized and critical phases is calculated for various values of @xmath12 . with an increase in @xmath12
, @xmath315 increases monotonically , in agreement with the trend shown in fig .
[ f0](b ) . moreover , the parity index @xmath313 is also displayed in fig . [ iden_5](b ) for various values of @xmath6 .
the transition point @xmath316 can be measured as a function of @xmath6 in a similar manner and subsequently the phase boundary in the @xmath18-@xmath19 plane can be obtained . at the angle @xmath317
is displayed in ( a ) as a function of @xmath6 for various values of @xmath12 and in ( b ) as a function of @xmath12 for various value of @xmath6 .
the transition point is located at the position separating the localized phase with @xmath318 from the critical phase with @xmath319 .
, title="fig : " ] + at the angle @xmath317 is displayed in ( a ) as a function of @xmath6 for various values of @xmath12 and in ( b ) as a function of @xmath12 for various value of @xmath6 .
the transition point is located at the position separating the localized phase with @xmath318 from the critical phase with @xmath319 .
, title="fig : " ] the rotational symmetry is studied next in the critical and localized phases , using the two typical cases of @xmath311 and @xmath312 , respectively .
the symmetry parameter @xmath98 is displayed in fig . [ iden_6 ] by rotating the two baths in the @xmath18-@xmath19 plane through an angle @xmath92 . in the localized phase
, @xmath98 quickly depletes to zero , different from that in the critical phase , where it gradually decays to a nonzero value .
compared to the results of the single - mode case in fig .
[ f12 ] , one can find that the localized and critical phases correspond to the strong and intermediate coupling regimes , respectively .
in general , the strong coupling regime has a large coupling strength @xmath12 and a small spectral exponent @xmath6 , while opposite trends ensue in the intermediate coupling regime . for the spectral exponent @xmath9
, however , the system always resides in the weak coupling regime , corresponding to the free phase .
similar to the case of the parity symmetry , the symmetry parameter of the rotational symmetry @xmath320 can also be used to distinguish the localized and critical phases . without loss of generality
, we set a special rotational angle @xmath321 where the phonons in the diagonal and off - diagonal coupling baths are interchanged .
thus , the value of the symmetry parameter is expected to be @xmath322 in the localized phase and @xmath323 in the critical phase .
[ iden_7](a ) shows @xmath324 as a function of @xmath6 for various values of @xmath12 .
the transition point @xmath315 is then located at the position separating the localized phase from the critical phase .
it monotonously increases with @xmath12 , consistent with the results in fig .
[ iden_5](a ) .
moreover , @xmath325 is also plotted in fig .
[ iden_7](b ) as a function of @xmath12 for various values of @xmath6 .
the transition boundary @xmath326 separating the localized and critical phases can then be appropriately calculated . and
the parity index @xmath313 for @xmath327 and @xmath246 are displayed as a function of @xmath6 in a weak coupling case of @xmath328 .
the dash line indicates the transition point @xmath329 . ]
is the transition points determined by the variational calculations at weak coupling @xmath330 ( lower curves ) and @xmath331 ( upper curves ) .
the circles and triangles correspond to the results from the curves of @xmath320 and @xmath313 , respectively .
the dash lines are the linear fitting for the extrapolation of @xmath332 with @xmath333 . ] to accurately estimate the critical value of the spectral exponent @xmath16 shown in fig .
[ f0](b ) , the case of very weak coupling strength @xmath334 is used to investigate @xmath320 and @xmath313 . as shown in fig .
[ iden_9 ] , transition behavior of @xmath313 and @xmath320 is displayed for @xmath327 and @xmath246 .
the near overlap of the two curves of @xmath133 and @xmath246 suggests that @xmath133 is sufficiently large to accurately describe the phase transition . with the increase of @xmath114 , the transition point on the @xmath313 and @xmath320 lines
is found to decrease monotonically tending to an asymptotic value of @xmath335 as marked by the dashed line .
it points to the critical value of @xmath1 in the weak coupling limit of @xmath31 , in perfect agreement with the mean - field prediction of @xmath2 , but stands at variance with the value of @xmath3 by guo _
this discrepancy may be attributed to the fact that the numerically unstable critical phase is beyond the reach of the dmrg algorithm of guo _ et al .
_ , and the external field holds great sway over the phase - transition properties of the two - bath model . in order to get a good estimate of @xmath336 , transition points calculated from the variational approach are in fig . [ iden_95 ] as a function of @xmath337 where @xmath114 is the number of superpositions
two values of coupling strengthes @xmath330 and @xmath331 are used , and we also set @xmath338 , and @xmath30 .
as @xmath337 decreases , the difference between @xmath320 and @xmath313 gradually disappears for both values of @xmath12 . using linear fitting of @xmath339 , the asymptotic values of @xmath340 and @xmath341 for the two cases
are obtained by extrapolation to infinite @xmath114 , which is consistent with the @xmath133 results @xmath342 and @xmath343 within the error bars , further supporting our claim that the number of superpositions @xmath133 is sufficient to obtain reliable results .
the deviation of the critical point from the guo s result @xmath344 is not induced by the effect of the finite value of @xmath114 .
the phase diagram of the extended spin - boson model is displayed in fig .
[ iden_8 ] in the case of @xmath30 and @xmath10 .
the solid triangles and stars represent the phase boundary obtained from the parity index @xmath313 and symmetry parameter of the rotational symmetry @xmath320 , respectively .
the error bars in this phase diagram are estimated via the difference of the transition points measured with the fixed @xmath12 and @xmath6 , respectively .
-@xmath6 plane in the case of @xmath30 and @xmath10 .
the phase boundary separating the critical phase from the localized phase is obtained from the parity index @xmath313 and the symmetry parameter of the rotational symmetry @xmath320 . ] and @xmath117 of two - bath model in the localized phase are plotted as a function of the phonon frequency @xmath44 at @xmath345 and @xmath346 .
the upper and lower panels correspond to the polarized and unpolarized baths , respectively .
the dashed line represents the classical displacements rescaled by a factor @xmath347 for comparison . ]
having studied the symmetry properties in detail , we now turn our attention to the wave function of the ground state for the two - bath model involving the continuous spectral density to understand the nature of the localized and critical phases . to serve this purpose , we chose the case of @xmath348 in the localized phase and the case of @xmath349 in the critical phase as examples .
[ ground_1 ] shows the displacement coefficients @xmath116 and @xmath117 defined in eq .
( [ vmwave ] ) in the localized phase as a function of the phonon frequency @xmath44 .
two different behaviors of the displacement are found in the upper and lower panels for the phonons in the diagonal and off - diagonal coupling baths , corresponding to the polarized bath " and unpolarized bath , " respectively . in the low frequency regimes ,
all the displacement coefficients @xmath116 and @xmath117 converge to a value independent of @xmath118 , i.e. , @xmath350 in the polarized bath and @xmath351 in the unpolarized bath , where @xmath352 is a @xmath210-independent constant . in the low frequency regime , however , @xmath116 and @xmath117 exhibit quite different behaviors , and the relations @xmath353 and @xmath354 are found , indicating that the phonons in the unpolarized bath obey some kind of symmetry constraints . moreover , the quantum fluctuations of the two - bath model in the localized phase are negligible , since the amplitude of the high - frequency displacements @xmath355 in the unpolarized bath is much smaller than that of the classical displacement @xmath356 in the low - frequency limit
. and @xmath117 in the critical phase are plotted at @xmath312 for the polarized and unpolarized baths , respectively .
the dashed line represents the classical displacements rescaled by a factor @xmath357 for comparison . ]
[ ground_2 ] shows the displacement coefficients @xmath116 and @xmath117 in the the critical phase as functions of @xmath44 .
similar to the results in the localized phase shown in fig .
[ ground_1 ] , one bath of the model is in the polarized state , and the other in the unpolarized state .
however , the classical displacement @xmath358 at @xmath359 in the upper panel is compatible with the amplitude of the high - frequency displacements @xmath360 in the lower panel .
it means that the quantum fluctuations play an important role in the critical phase , unlike the case in the localized phase .
( nearby the origin @xmath361 ) and @xmath362 ( in a corner ) , corresponding to the critical and localized phases , respectively .
the @xmath65- and @xmath66-coordinate correspond to off - diagonal and diagonal coupling baths , respectively , and the colour represents the phonon population @xmath59 at the frequency @xmath363 . for convenience , we set the unit of the length in the @xmath18-@xmath19 plane as @xmath364 .
] finally , the ground states of the two - bath model in the localized and critical phases are compared via the phonon population @xmath59 as shown in fig .
[ ground_3 ] . according to the results in figs .
[ ground_1 ] and [ ground_2 ] , the displacement coefficients @xmath116 and @xmath117 in the localized and critical phases are quite different in the low - frequency regime , especially at @xmath363 .
hence , only the bath modes at the frequency @xmath363 are considered , and the unit of the length @xmath364 is set .
the phonon state in the case of @xmath290 and @xmath291 ( critical phase ) is located nearby the origin @xmath361 , but the one in the case of @xmath345 and @xmath346 ( localized phase ) is far away it . in both the cases ,
the distance @xmath365 between the center of the phonon state and the origin is proportional to the displacement @xmath120 in the polarized bath .
moreover , the central angle to the origin is calculated to be @xmath366 for the case of the localized phase , where @xmath367 is the radius of the phonon population @xmath59 and @xmath368 is the distance .
interestingly , the central angle @xmath369 is comparable with the width of the peaks @xmath370 defined in fig .
[ iden_5 ] .
it further supports that the parity index @xmath371 reflects the localized nature of the ground state .
component of the spin polarization @xmath373 for various values of @xmath12 is plotted as a function of the ratio @xmath374 in ( a ) and ( b ) , corresponding to the variational and dmrg results , respectively . in both ( a ) and ( b )
, the dash - dotted line indicates the transition point @xmath375 , and dashed lines represent linear fits .
the value of the spectral exponent @xmath289 is set.,title="fig : " ] + component of the spin polarization @xmath373 for various values of @xmath12 is plotted as a function of the ratio @xmath374 in ( a ) and ( b ) , corresponding to the variational and dmrg results , respectively . in both ( a ) and ( b ) , the dash - dotted line indicates the transition point @xmath375 , and dashed lines represent linear fits . the value of the spectral exponent @xmath289 is set.,title="fig : " ] in the two bath model
, we next investigate the case with @xmath376 to further identify the critical and localized phases . according to ref .
@xcite , there exists a first - order quantum phase transition separating the doubly degenerate `` localized state '' with @xmath377 and @xmath378 from the doubly degenerate `` delocalized state '' with @xmath379 and @xmath380 .
the transition point @xmath381 is expected from the @xmath18-@xmath19 symmetry when the spectral exponents obey @xmath30 . in the following variational calculations , we use @xmath382 and @xmath134 , which have been found to be sufficient in obtaining reliable results of the localized - to - delocalized phase transition @xcite . in fig .
[ multi_1](a ) , the @xmath66 component of the spin polarization @xmath373 obtained by the variational method is plotted with respect to the ratio @xmath374 for various values of the diagonal coupling @xmath12 in the case of @xmath289 . the transition point is determined at @xmath383 , in perfect agreement with the expectation @xmath375 within numerical errors .
furthermore , a linear decay of @xmath373 is found for @xmath384 before showing an abrupt jump to zero at the transition point , thereby verifying the transition to be of the first order . for comparison
, the numerical results obtained by the dmrg algorithm are also displayed in fig .
[ multi_1](b ) . similar behavior of @xmath373 is found , pointing to the validity of the variational method .
is displayed as a function of the coupling strength @xmath12 at @xmath289 .
the circles and triangles correspond to the dmrg and variational results , respectively , and the dashed lines represent exponential fits . ] due to infinite degenerate ground states @xmath276 , the spin polarization @xmath241 at the transition point @xmath375 is variable as shown in fig .
[ iden_3 ] . in the regime with @xmath385 , however , the value of @xmath373 is quite robust , since the rotational symmetry is broken by the anisotropic coupling .
the generalized susceptibility @xmath386 is thus introduced by @xmath387 where @xmath54 is the spin polarization defined in eq .
( [ spin polarization ] ) , and @xmath388 is the reduced coupling strength . as shown in fig .
[ multi_1 ] , the @xmath66 component of the spin polarization @xmath373 can be fitted by a linear behavior .
since the @xmath65 component of the spin polarization is @xmath389 in the localized state @xcite , one obtains @xmath390 .
the generalized susceptibility @xmath391 can then be calculated with the linear fitting procedure for different values of @xmath12 and @xmath6 .
an extended scaling form @xmath392 could lead to a better fitting of the numerical data , and yield @xmath393 in the limit of @xmath394 .
[ multi_2 ] shows the generalized susceptibility @xmath395 at @xmath289 obtained by the variational method ( solid triangles ) and dmrg algorithm ( open circles ) .
both of them decay exponentially with @xmath12 , though the exponents obtained via the fitting with @xmath396 differ for the dmrg ( @xmath397 ) and variational method ( @xmath398 ) .
the two curves @xmath395 follow the same behavior , implying that the system is always in the localized phase at @xmath289 for various values of @xmath12 , and no phase transition occurs when @xmath305 . is displayed as a function of @xmath399 in ( a ) for various values of @xmath12 and in ( b ) for various values of @xmath400 .
the dash - dotted lines indicate the peak positions of the curves for @xmath401 and @xmath402 , pointing to a transition boundary @xmath403 in the phase diagram .
, title="fig : " ] + is displayed as a function of @xmath399 in ( a ) for various values of @xmath12 and in ( b ) for various values of @xmath400 .
the dash - dotted lines indicate the peak positions of the curves for @xmath401 and @xmath402 , pointing to a transition boundary @xmath403 in the phase diagram .
, title="fig : " ] in fig .
[ multi_3](a ) , the generalized susceptibility @xmath29 is displayed as a function of the ratio @xmath399 for various values of @xmath12 . unlike the trend shown in fig .
[ multi_2 ] , @xmath29 increases with the spectral exponent @xmath6 until @xmath404 marked by the dash - dotted line as the position of the peaks when @xmath401 .
it points to the transition boundary separating the localized phase from the critical phase .
moreover , the generalized susceptibility @xmath29 is also displayed in fig .
[ multi_3](b ) for various values of @xmath6 .
the position of the peak similarly marked by the dash - dotted line is found to be at @xmath405 when @xmath406 . from these numerical results ,
one can obtain a relationship @xmath403 , pointing to the transition boundary , i.e. , @xmath407 .
it further supports our contention that the critical value of the spectral exponent is @xmath408 .
the ground states of the spin - boson model with diagonal and off - diagonal coupling to two identical independent baths have been comprehensively studied in this paper by the variational approach , the dmrg algorithm and the exact diagonalization method . adopting a generalized trial wave function , i.e. , multi-@xmath409 ansatz , as the variational wave function
, the spin polarization @xmath54 , ground state energy @xmath153 and wave function @xmath108 are calculated accurately by the variational method , in good agreement with those from the exact diagonalization in the case involving two oscillators and those from the dmrg algorithm in the case involving two baths describing a continuous spectral density function .
three phases ( localized , critical and free ) are identified , corresponding to the strong , intermediate and weak coupling regimes , respectively . via the symmetry parameters @xmath410 and @xmath411 , the nature of the these phases is uncovered .
the breakdown of the parity and rotational symmetries is found to occur along the quantum phase transition between the localized and critical phases .
the phase boundary is determined by the parity index @xmath313 of the parity - symmetry regime with @xmath412 , consistent with that obtained by @xmath320 . moreover
, the critical value of the spectral exponent is estimated as @xmath413 , well in agreement with the mean - field prediction @xmath2 @xcite .
the behavior of the spin polarization @xmath414 and the generalized susceptibility @xmath29 is also investigated for various values of the coupling strengthes @xmath12 and spectral exponents @xmath6 .
both the results point to @xmath415 , further supporting the accuracy of our results in pinning the transition point .
the authors thank bo zheng , yao yao and javier prior for useful discussion . support from the singapore national research foundation through the competitive research programme ( crp ) under project no .
nrf - crp5 - 2009 - 04 is gratefully acknowledged .
this work is also supported in part by and national natural science foundation of china under grant no .
@xmath416 and the u.s .
national science foundation under grant no .
che-@xmath417 .
r. bulla , n .- h .
tong and m. vojta , phys .
lett . * 91 * , 170601 ( 2003 ) ; m. vojta , n .- h .
tong , and r. bulla , phys .
lett . * 94 * , 070604 ( 2005 ) ; f. b. anders , r. bulla and m. vojta , phys .
lett . * 98 * , 210402 ( 2007 ) . | a generalized trial wave function termed as the `` multi-@xmath0 ansatz '' has been developed to study the ground state of the spin - boson model with simultaneous diagonal and off - diagonal coupling in the sub - ohmic regime .
ground state properties including the energy and the spin polarization are investigated , and the results are consistent with those from the exact diagonalization and density matrix renormalization group approaches for the cases involving two oscillators and two baths described by a continuous spectral density function .
breakdown of the rotational and parity symmetries along the continuous quantum phase transition separating the localized phase from the critical phase has been uncovered .
moreover , the phase boundary is determined accurately with the corresponding symmetry parameters of the rotational and parity symmetries .
a critical value of the spectral exponent @xmath1 is predicted in the weak coupling limit , which is in agreement with the mean - field prediction of @xmath2 , but much smaller than the earlier literature estimate of @xmath3 . |
antiferromagnetic spin systems on fully frustrated lattices show many unusual behaviors in magnetic and thermal properties@xcite .
one of the main ingredients is that their unit cell allows for a continuous degree of freedom and that the connectivity ( corner sharing ) allows for an extensive number of these degrees of freedom in the classical ground state .
more subtle phenomena appear when looking at quantum models on these lattices as quantum fluctuations may induce very unsusual ground states . in particular , wether the quantum ground state on fully frustrated lattices may not break the lattice symmetry nor the spin group symmetry is still a highly debated question .
a lot of work has been done to answer this question in the framework of the proposition of anderson@xcite , looking first at resonating valence bond states .
one of the most studied candidates is the quantum @xmath4=1/2 heisenberg antiferromagnet on the kagom lattice . using various methods@xcite
, it has been shown that the low temperature physics should be dominated by short range rvb states which produce a continuum of singlet states between the @xmath5 ground state and the first excited @xmath6 state .
a still controversial question is the presence of a very low temperature peak in the specific heat , much below the one corresponding to the onset of short range correlations , that could be ascribed to a high density of singlet states in the singlet - triplet spin gap .
associated to this low temperature peak is the missing ( or not ) entropy that would characterize an ordered or a disordered ground state .
the experimental relevance of the model is extremely fragile as , in real compounds , many other parameters may drive the physics to very different universality classes , described through inclusion of next nearest neighbor interactions@xcite , disorder@xcite , antisymmetric interactions@xcite , etc ... nevertheless , in many cases , the heisenberg antiferromagnet is a good starting point as it is generally believed that many properties of realistic systems come from deviations from the heisenberg limit .
examples are the layered oxide srcr@xmath7ga@xmath8o@xmath9@xcite ( @xmath10 ) or the organic pseudo-@xmath6 compound m-mpynn.bf@xmath11@xcite . in this paper , the properties of the heisenberg @xmath4=1/2 kagom antiferromagnet are addressed by using a spin green s functions technique .
one of the advantages of the method is that it is well suited for magnetic systems with no long range order as it uses a decoupling scheme based on short ranged spin correlations .
this method was previously introduced by kondo and yamagi@xcite in the context of the one - dimensional heisenberg model as a theory of spin - waves in absence of long - range order .
more recently , it has been used to address different problems@xcite and can be extended to include magnetic phases@xcite . here , the formalism is used to compute thermodynamic and magnetic quantities at all temperatures , like the internal energy , specific heat , entropy , magnetic susceptibility and structure factor .
correlation functions are also computed at all temperatures and for various separation distances . in the next section ,
the approximation is presented and results are discussed in sec . [ results ] .
the kagom lattice.,width=188 ]
the antiferromagnetic heisenberg model is defined by the hamiltonian @xmath12 where we assume @xmath0 and a nearest - neighbor exchange @xmath13 the spin susceptibility @xmath14 is obtained by a fourier transformed green s function @xmath15 where @xmath16\right > .\ ] ] > from the equation of motion of the operator @xmath17 in the heisenberg representation , the green s functions above must satisfy the equation @xmath18
\right > + \ll \left [ a , { \cal h}\right ] ; b\gg _ \omega .\ ] ] the spin susceptibility of the present model can be obtained by @xmath19 the three operator green s functions @xmath20 must also obey their corresponding equations
@xmath21 as higher - order green s functions are generated , one obtains an infinite hierarchy of equations which have to be decoupled in order to obtain @xmath22 . in a frustrated lattice ,
one is constrained to the non - magnetic phase , where @xmath23 and @xmath24 .
here we adopt an extension of the kondo - yamaji decoupling approximation for the kagom lattice , assuming ( only for different sites @xmath25 ) @xmath26 the static correlation functions @xmath27 between spins at the lattice positions @xmath28 and @xmath29 depend on the distance @xmath30 , on the number @xmath31 of bonds between sites @xmath32 and @xmath33 , and on the lattice symmetry . as a shorthand notation
, we shall call the local correlation as @xmath34 , the nearest - neighbor correlation as @xmath35 and the next - nearest - neighbor correlations as @xmath36 and @xmath37 ( the former corresponding to the shortest distance , as illustrated in figure [ rede ] ) .
we also introduce @xmath38 .
the parameter @xmath39 is useful to enforce the condition @xmath40 @xcite .
these correlation functions can be evaluated from the spin susceptibility as @xmath41 thus , @xmath35 , @xmath36 , and @xmath42 and also the parameter @xmath39 appearing in eqs .
[ ap1]-[ap2 ] must be evaluated selfconsistently . applying approximation [ ap1]-[ap2 ] to the kagom lattice yields @xmath43 \chi _ { ij } ( \omega ) = -8jc_1\delta _ { ij}+2c_1j_{ij } \\ + j\tilde{c}_1\sum _ k j_{ik } ' \chi _ { kj}(\omega ) -j\left ( \tilde{c}+2\tilde{c}_1 \right ) \sum _ k
j_{ik } \chi _ { kj}(\omega ) , \label{eq : appr}\end{gathered}\ ] ] where @xmath44 and @xmath45 by working directly with the @xmath46 , the numerical problem is reduced to only two selfconsistent parameters , @xmath47 and @xmath48 , from which one can obtain @xmath49 . with reference to the underlying triangular lattice ,
the approximate equation [ eq : appr ] can be rewritten in matrix form @xmath50 \boldsymbol{\chi } _ { ij}(\omega ) = -2c_1\left ( 4j\delta _ { ij}\boldsymbol{\mathit 1}-\boldsymbol{j}_{ij}\right ) \\ + j\tilde{c}_1\sum _ k
\boldsymbol{j}_{ik } ' \boldsymbol{\chi } _ { kj}(\omega ) -j \left ( \tilde{c}+2\tilde{c}_1 \right ) \sum _ k \boldsymbol{j}_{ik } \boldsymbol{\chi } _ { kj}(\omega ) \ , \label{matrix}\end{gathered}\ ] ] where @xmath51 and @xmath52 are the exchange matrices connecting neighboring triangles . as indicated in fig .
1 , each triangle is surrounded by six neighboring triangles , all of them contributing to both types of exchange matrices .
the elements of the dynamic susceptibility matrix are defined as @xmath53 where @xmath54 , @xmath28 is the lattice position of the @xmath32-th triangle , and the vectors @xmath55 indicate the position of site @xmath56 in the triangular unit cell ( @xmath57 ) .
applying a corresponding fourier transformation to equation [ matrix ] gives @xmath58 where @xmath59 where @xmath60 and @xmath61 are the fourier transform of the exchange matrices @xmath51 and @xmath62 , @xmath63 & { \rm cos}[(x+\sqrt{3}y)/4 ] \\ \\ { \rm cos}[(x-\sqrt{3}y)/4 ] & 0 & { \rm cos}(x/2 ) \\ \\ { \rm cos}[(x+\sqrt{3}y)/4 ] & { \rm cos}(x/2 ) & 0 \\ \end{array } \right ) \ , \ ] ] @xmath64 & { \rm cos}[(3x-\sqrt{3}y)/4 ] \\ \\ { \rm cos}[(3x+\sqrt{3}y)/4 ] & { \rm cos}x+{\rm cos}[(x-\sqrt{3}y)/2 ] & { \rm cos}(\sqrt{3}y/2 ) \\ \\ { \rm cos}[(3x-\sqrt{3}y)/4 ] & { \rm cos}(\sqrt{3}y/2 ) & { \rm cos}x+{\rm cos}[(x+\sqrt{3}y)/2 ] \\
\end{array } \right ) \ .\ ] ] in which @xmath65 and @xmath66 .
it follows that @xmath67 where @xmath68 is a third degree polynomial in @xmath69 with @xmath70-dependent coefficients and the elements of @xmath71 are second degree polynomials in @xmath69 with @xmath70-dependent coeficients .
the dynamic susceptibility can be finally writen as @xmath72 where @xmath73 and @xmath74 the @xmath75 are excitation energies , obtained from the roots @xmath76 of the polynomial @xmath77 .
one of these roots , @xmath78 , is found to be non - dispersive over the whole brillouin zone .
thus , the dispersion in the energy spectrum is due to the other roots ( @xmath79 ) , which are evaluated numerically . by writing the correlation functions of eq .
[ eq : corr ] in matrix form and performing the inverse fourier transformation , we can show that @xmath80 where @xmath81 and @xmath82 is the bose distribution function .
the sum over momenta must be performed numerically .
the dynamic structure factor matrix is related to the dynamic susceptibility by @xmath83 after obtaining the selfconsistent parameters , the static structure factor can be evaluated by @xmath84
we have calculated the selfconsistent values of @xmath35 , @xmath36 , and @xmath37 as a function of temperature .
they are plotted in fig .
[ corr]-b .
the selfconsistent parameter @xmath39 is shown in fig .
[ corr]-a . from the knowledge of these selfconsistent functions
all missing correlation functions can be evaluated . in figure [ cn ]
we show the correlation functions @xmath85 between sites separated by at most 6 lattice bonds .
the distribution of the points is very similar to that obtained in ref . for the pyrochlore lattice , indicating an exponential decay with a correlation length of the order of the lattice parameter @xmath56 .
the internal energy is given by @xmath86 .
the ground - state value @xmath87 is in good agreement with earlier calculations ( see refs . ) . in figure
[ heat ] we present the specific heat @xmath88 .
this curve has a single peak and is very similar to the result of high - temperature expansions @xcite reproducing the high - temperature limit @xmath89 . in the low - temperature limit ,
the time involved in the numerical computation increases with decreasing temperature , because the convergence of the integral in @xmath70 becomes slower , and in addition one needs a higher precision in order to perform the derivative involved in the evaluation of the specific heat .
nevertheless , a careful analysis of the numerical results indicates unambiguously a @xmath3 dependence of the specific heat which extends to very low temperatures .
therefore , there will be no second peak in the present model .
the fact that the specific heat does not vanish exponentially indicates the presence of low energy states even though the low - temperature peak found in ref . is missing .
indeed , it should be emphasized that it is not sure that this peak will persist in the thermodynamic limit@xcite which supports the results obtained in the present work . the change in entropy @xmath90 is also shown in fig .
the total change in entropy is @xmath91 , corresponding to a ground - state entropy @xmath2 , of the same order as found in ref . .
the eigenvalues @xmath92 of the structure factor are shown in fig .
[ lambda ] .
the result is remarkably similar to what is found from the high - temperature expansions of ref .
whithin the present approach one can easily follow the effect of temperature on the structure factor for all wavevectors even at very low @xmath93 .
we observe that its highest eigenvalue is degenerate over the whole brillouin zone .
for the pyroclore , the structure factor is a @xmath94 matrix , which has one additional eigenvalue . in a perturbative expansion @xcite the two highest eigenvalues
are found to lie very close to each other , one of them being completely degenerate while the other one has a weakly lifted degeneracy , which is crucial to reproduce the main features of neutron scattering experiments . as a further work
, it would be interesting to extend the present approach to the pyrochlore antiferromagnet .
this work has been partially supported by brazilian agencies cnpq ( conselho nacional de desenvolvimento cientfico e tecnolgico ) and capes ( coordenao de aperfeioamento de pessoal de nvel superior ) through the french - brazilian cooperation agreement capes - cofecub . | the @xmath0 heisenberg antiferromagnet is studied on the kagom lattice by using a green s function method based on an appropriate decoupling of the equations of motion .
thermodynamic properties as well as spin - spin correlation functions are obtained and characterize this system as a two - dimensional quantum spin liquid .
spin - spin correlation functions decay exponentially with distance down to low temperature and the calculated missing entropy at @xmath1 is found to be @xmath2 . within the present scheme ,
the specific heat exhibits a single peak structure and a @xmath3 dependence at low temperature . |
a _ network _ , in the simplest terms , is a collection of points in some space ( e.g. on the euclidean plane ) , called _ nodes _ or _ vertexes _ , and a collection of node pairs , called _
edges_. the presence of an edge between two nodes indicates that they can directly communicate with each other .
the mathematical name for this network model is _
graph_. a class of networks that has recently attracted particular interest in the wireless communication context , is called _ ad - hoc networks_. it is distinguished by the fact that the network nodes are not subject to any regular ( say periodic ) geometric emplacement in the space but can be rather seen as a snapshot of some _ random point pattern _ , called also _ point process _ ( pp ) in the mathematical formalism typically used in this context .
connectivity , i.e. possibility of indirect , multi - hop , communication between distant nodes , is probably the first issue that has to be addressed when considering ad - hoc networks .
an ubiquitous assumption when studying this problem is that the node randomness is modeled by a spatial _ poisson _ pp .
this latter situation can be characterized by independence and poisson distribution of the number of nodes observed in disjoint subsets of the space .
poisson assumption in the above context is often too simplistic , however analysis or even modeling of networks without this assumption is in most cases very difficult . in this paper
we introduce some class of point processes that might be roughly described as exhibiting less variable point patterns than poisson pp .
we call point processes ( pp s ) of this class _ sub - poisson point processes_. our objective is two - fold .
on one hand we want to argue that this class of pp s allows to extend some classical results regarding network connectivity and on the other hand we want to bring attention to sub - poisson pp s , as they might be useful for modeling of ad - hoc networks , whose nodes are more regularly distributed than a poisson pp .
a random variable is called sub - poisson , if its variance is not larger than its mean ( with the equality holding true for poisson variable ) . intuitively ,
if strictly sub - poisson variables were to describe the number of nodes in different subsets of the space then the resulting point patterns would exhibit less clustering ( or bunching ) than the poisson point pattern having the same number of points per unit of space volume .
this in turn , still intuitively , should have positive impact on the connectivity and perhaps capacity and other network performance metrics , the reason being that the perfectly regular , periodic patterns are commonly considered ( and sometimes can be proved ) as being optimal .
the aim of the present article is to provide rigorous results on the comparison of the connectivity and capacity properties of certain sub - poisson networks to these of the respective poisson networks .
the statistical variability of random variables ( say with the same mean ) can be compared only to some limited extent by looking at their variances , but more fully by convex ordering . under this order one can compare the expected values of _ all _ convex functions of these variables . in multi - dimensions
there is no one single notion of convexity . besides different statistical variability of marginal distributions , two random vectors ( think of number of nodes in different subsets of the space )
can exhibit different dependence properties on their coordinates .
the most evident example here is comparison of the vector composed of several copies of one random variable to a vector composed of independent copies sampled from the same distribution . among several notions of `` convex - like '' ordering of random vectors the so called _ directionally convex _ ( @xmath0 ) order , allowing one to compare expectations of all @xmath0 functions ( see section [ sec : dcx_defns ] below ) of these vectors , takes into account both the dependence structure of random vectors and the variability of their marginals .
it can be naturally extended to random fields by comparison of all finite dimensional distributions , as well as to random ppas and even locally finite random measures by viewing them as non - negative fields of measure - values on all bounded borel subsets of the space ; cf .
@xcite .
using this latter formalism , we say that a pp is _ sub - poisson _ ( or simply _ sub - poisson _ when there is no ambiguity ) if it is @xmath0-smaller than a poisson pp having the same mean measure , i.e. , the mean number of nodes in any given set .
we will also say _ sub - poisson network _ in the case when the nodes of the graph modeling the network are distributed according to some sub - poisson pp .
we shall also see that there are classes of pp whose so - called _ joint intensities _ ( densities of the higher order moment measures , when they exist ; see section [ ss.intensities ] ) are smaller than those of poisson pp
. using this latter ( and weaker ) property we define the class of _ weakly sub - poisson point processes_. another weakening of the comparability assumption , to the expectations of increasing ( or decreasing ) @xmath0 functions allows to define _ increasing @xmath0 _ sub - poisson pp s , which can be seen as being less variable than some poisson pp having possibly smaller ( or larger ) mean measure .
we shall abbreviate increasing @xmath0 by @xmath1 and similarly _ decreasing _
@xmath0 by @xmath2 .
our choice of the @xmath0 order to define sub - poisson pp s has its roots in @xcite , where one shows various results as well as examples indicating that the @xmath0 order implies ordering of several clustering characteristics known in spatial statistics such as ripley s k - function or second moment densities .
namely , a pp that is larger in the @xmath0 order exhibits more clustering ( while having the same mean number of points in any given set ) . full connectivity ( multi - hop communication between any two nodes ) of an ad - hoc network with many nodes is typically hard to maintain , and so , a more modest question of existence of a large enough , connected subset of nodes ( called component ) is studied .
a possible approach to this problem , proposed in @xcite , and based on the mathematical theory of _ percolation _ , consists in studying existence of an _ infinite _ ( called _ giant _ ) _ component _ of the infinite graph modeling a network .
existence of such a component is interpreted as an indication that the connectivity of the modeled ad - hoc network scales well with its number of nodes .
percolation models have been extensively studied both in mathematical and communication literature . the model proposed in @xcite , called now the _ gilbert s model _ , is now considered as the classical continuum model in percolation theory .
it assumes that each node has a given fixed range of communication @xmath3 and direct connection between any two nodes is feasible if they are within the distance @xmath3 from each other , regardless of the positions of other nodes in the network .
the known answer to the percolation problem in this model is given under the assumption that the nodes are distributed according to a homogeneous poisson pp having a _ density _ of @xmath4 nodes on average per unit of volume ( or surface ) .
the result says that there exists a _
non - degenerate critical communication range _ @xmath5 , such that for @xmath6 there is no giant connected component of the network ( the model does not percolate ) , while for @xmath7 there is exactly one such component ( the model percolates ) , both statements holding true almost surely ; i.e. , for almost all poisson realizations of the network . an equivalent statement of the above result says that for a given communication range @xmath3 there exists a non - degenerate critical density of nodes @xmath8 , below which the model does not percolate and above which it does so almost surely . as one of the main results of this paper we will prove an extension of the above result , which says that the _ critical communication range of any homogeneous sub - poisson gilbert s network model is not degenerate_. moreover
, this critical communication range is bounded away from zero and infinity by the constants which depend only on the mean density of nodes and not on the finer structure of the sub - poisson pp of network nodes .
partial results , regarding only non - degeneracy at zero or at infinity can be proved for @xmath9 and @xmath2-sub poisson networks respectively . a more adequate percolation model of a wireless communication network , called the _ sinr _
graph , was studied more recently in @xcite .
it allows one to take into account the interference intrinsically related to wireless communications .
the interference power is modeled by the _ shot - noise _ field generated by the pp of transmitting nodes .
each pair of nodes in the new model is joined by an edge when the signal power to this shot - noise plus some other ( external ) noise power ratio is large enough .
the resulting random graph does not have the independence structure of the gilbert s model and increasing the communication range ( equivalent in this model to increasing the signal power and hence the value of the shot - noise ) is not necessarily beneficial for connectivity .
similar observation holds as regards the increase in the node density .
in fact , the above sinr network model has two essential parameters : the density @xmath4 of poisson pp of nodes and the shot - noise reduction factor @xmath10 .
the percolation domain is characterized in the cartesian product of these two parameters .
the key result of @xcite says that whenever @xmath4 is larger than the critical value corresponding to the percolation of the model with interference perfectly canceled out ( for @xmath11 , which simplifies the model to the gilbert s one ) then there exists a critical value @xmath12 such that the model percolates for @xmath13 and does not percolate for @xmath14 .
as the second main result of this paper , we extend the above result , also to sub - poisson sinr networks .
our proofs rely on the comparison of extremal and additive shot - noise fields generated by @xmath0-ordered pp s studied in @xcite .
while the connection to the sinr model is evident ( additive shot - noise is an element of this model ) the connection to the gilbert s model is perhaps less evident and relies on the fact that this latter model can be represented as a upper level - set of some extremal shot - noise field
. more precisely , from ( * ? ? ?
* propsition 4.1 ) one can conclude that the probability of @xmath15 given locations in the space _ not _ being within the communication range of any of the nodes in the gilbert s model is higher for the network whose nodes are modeled by a pp larger in @xmath0 order . using this property , suitable discretization of the model and the peierls argument ( cf .
@xcite or ( * ? ? ?
* proposition 14.1.4 ) ) one can prove finiteness of the critical transmission range in the sub - poisson gilbert s network .
the strict positivity of this range can be proved by comparing the expected number of paths from the origin to the boundary of an increasing box , again in some suitable discretization of the model .
this latter comparison can be done relying only on weak sub - poisson assumption . _ the remaining part of this paper is organized as follows _ :
in section [ s.preliminaries ] we provide necessary notions and notation .
sub - poisson gilbert s model is studied in section [ s.gilberts ] and sub - poisson sinr model in section [ s.sinr ] . in section [ s.capacity ]
we make some remarks on the impact of @xmath0 ordering on various model characteristics usually called capacities , including the so - called capacity functional , being one of the fundamental characteristics studied in stochastic geometry , as well as on some other capacity quantifiers in the information - theoretic sense . in section [ s.examples ]
we show some examples of sub - poisson and weakly sub - poisson pp , in particular the so - called _
perturbed lattices_. conclusions as well as open questions are presented in section [ s.conclusions ] .
we say that a function @xmath16 is _ directionally convex _ ( @xmath0 ) if for every @xmath17 such that @xmath18 ( i.e. , @xmath19 and @xmath20 with inequalities understood component - wise ) and @xmath21 one has @xmath22 . also , we shall abbreviate _ increasing _ and @xmath0 functions by @xmath1 and _ decreasing _ and @xmath0 by @xmath2 . for two real - valued random vectors
@xmath23 and @xmath24 of the same dimension , @xmath23 is said to be @xmath0 smaller than @xmath24 , ( denoted by @xmath25 ) if @xmath26 for every @xmath0 function @xmath27 for which both expectations are finite . in full analogy one defines @xmath1 and @xmath2 orders of random vectors considering @xmath1 and @xmath2 functions , respectively .
these orders clearly depend only on the distributions of the vectors .
two real valued stochastic processes ( or random fields ) are said @xmath28 or @xmath2 ordered if any finite - dimensional distributions of these processes are ordered .
this definition extends also to locally finite random measures ( in particular pp s ) , by ordering random values of these measures ( in particular numbers of points ) on any finite collection of subsets of the state space .
more precisely , we say for two pp s @xmath29 that @xmath30 if for every finitely many bounded borel subsets @xmath31 we have that , @xmath32 it was shown in @xcite that verifying the above property for all _ mutually disjoint bounded borel subsets _
@xmath33 is a sufficient condition for the respective ordering of the pp s . using the above definition
, we say that _
@xmath34 is sub - poisson if @xmath35 , where @xmath36 is some poisson pp_. noting that for the linear function @xmath37 , both @xmath27 and @xmath38 is convex ( and thus @xmath0 in one dimension ) one can observe that @xmath36 needs to have the same mean measure as @xmath34 ; i.e. , if @xmath35 then @xmath39={{\mathbf e}}[\phi_{poi}(b)]$ ] for every bounded borel subset @xmath40 . in particular ,
_ a stationary pp @xmath34 is sub - poisson if it is @xmath0 smaller than poisson pp @xmath41 of the same intensity _
@xmath42/|b|$ ] , where @xmath43 is the @xmath44dimensional lebesgue measure of the set @xmath40 . from now on @xmath41 or @xmath45 will always denote homogeneous poisson pp of intensity @xmath4 or @xmath46 .
we will also say that pp @xmath34 is homogeneous if its mean measure is equal , up to a constant , to the lebesque measure ; i.e. @xmath39=\lambda |b|$ ] , for some constant @xmath4 all bounded borel sets @xmath40 .
this is a weaker assumption than the stationarity of @xmath34 .
we say that @xmath34 is @xmath1(@xmath2)-sub - poisson if @xmath47 , where @xmath36 is some poisson pp . in this case
the mean measure of @xmath34 is smaller or equal ( larger or equal ) to that of @xmath36 .
the joint intensities @xmath48 of a pp are defined by the following relation for every finitely many disjoint bounded borel sets @xmath31 : @xmath49 = \int\!\!\dots\!\!\int_{b_1\times\dots\times b_k } \rho^{(k)}(x_1,\ldots , x_k ) { \text{\rm d}}x_1 \ldots { \text{\rm d}}x_k,\ ] ] provided they exist ( i.e. , the respective moment measures admit densities ) . a pp @xmath34 with joint intensities of all orders @xmath50 is said to be _ weakly sub - poisson _ if there exists a constant @xmath51 such that for all @xmath52 , @xmath53 due to the @xmath1 property of the function @xmath54 , the ordering @xmath55 implies that for all @xmath52 , @xmath56 \leq { { \mathbf e}}[\prod_{i=1}^k\phi_2(b_i)]$ ] for any disjoint bounded borel sets @xmath57 .
hence , the respective joint intensities @xmath58 of pp s @xmath59 , @xmath60 , provided they exist , obey for all @xmath52 , @xmath61 a.e .. since @xmath62 for poisson pp @xmath63 , we are justified in using the term weakly sub - poisson for pp @xmath34 satisfying ( [ e.weakly-sub-poi ] ) .
given a pp @xmath34 on @xmath64 and a non - negative constant @xmath65 one defines the ( _ spherical _ ) boolean model generated by @xmath34 of ball of radius @xmath66 as the union @xmath67 , where @xmath68 is the ball centered at @xmath69 of radius @xmath66 .
we say that @xmath70 _ percolates _ if there exists an unbounded , connected subset of @xmath70 .
this definition extends to any random subset of @xmath64 .
one defines the _ critical radius _
@xmath71 of the boolean model as @xmath72 . here is the first result of this paper .
[ thm : perc_sub - poisson_pp ] there exist universal constants @xmath73 and @xmath74 depending only on @xmath4 ( and the dimension @xmath75 ) such that if @xmath76 then @xmath77 and if @xmath78 then @xmath79 . thus , for a homogeneous @xmath0-sub - poisson @xmath34 of intensity @xmath4 we have @xmath80 , where the constants depend only on @xmath4 .
the lower bound holds also when @xmath34 is weakly sub - poisson .
the proof is given in the appendix .
the above result can be extended to the so - called @xmath50-percolation models .
note that the connectivity structure of the boolean model @xmath70 corresponds to that of the gilbert s network with nodes in @xmath34 and communication range @xmath81 ; cf .
section [ sss.gilbert ] .
thus the critical communication range @xmath82 in this latter network is twice the critical radius of the corresponding boolean model .
while the finiteness of the critical radius of the boolean model ( and thus communication range of the sub - poisson gilbert s network ) is intuitively a desired property , its positivity at first glance might be seen from the networking point of view as irrelevant , if not a disadvantage .
a deeper inspection of wireless communication mechanisms shows however that sometimes a non - percolation might be also a desired property . the following modification of the gilbert s model , that can be seen as a toy version of the sinr model studied in section [ s.sinr ] , sheds some light on this latter statement . consider a planar ad - hoc network consisting of nodes modeled by a point process @xmath83 on the plane ( in @xmath84 ) and having communication range
this process corresponds to a _ back - bone _ of the network , whose percolation we are looking for .
consider also an auxiliary pp @xmath85 of _ interferers _ also on @xmath84 .
consider the following modification of the gilbert s connection rule : any two given nodes of @xmath83 can directly communicate ( are joined be an edge ) when they are within the communication range @xmath3 from each other , however only when there is no interfering node ( point of @xmath85 ) within the _ sensing range _ @xmath86 of any of these two nodes .
note that any connected component of this network is included in some connected subset of the _ complement _ @xmath87 of the boolean model @xmath88 .
percolation of this _ vacant region _
@xmath89 of the spherical boolean model is thus a necessary condition for the percolation of our modification of the gilbert s network .
now , the vacant region of the planar boolean model can not percolate if the boolean model itself does ( cf ( * ? ? ? * theorem 4.4 ) ) .
thus a non - percolation of some boolean model related to the interferes @xmath85 is a necessary condition for the percolation of the communication network on @xmath83 .
in this section , we shall work only on the plane @xmath84 .
we slightly modify the definition of sinr network introduced in @xcite allowing for _ external interferers_. the parameters of the model are non - negative numbers @xmath90 ( signal power ) , @xmath91 ( environmental noise ) , @xmath10 ( interference reduction factor ) , @xmath92 ( sinr threshold ) and an attenuation function @xmath93 of the distance @xmath94 satisfying @xmath95 , continuous , strictly decreasing on its support , with @xmath96 and @xmath97 .
these assumptions are exactly as in @xcite . given a pp @xmath34 , the ( unit - power ) interference generated by @xmath34 at location @xmath98 is defined as @xmath99 .
more generally this object is also called ( additive ) shot - noise field generated on by @xmath34 withe _ response function _
@xmath100 ; cf ( * ? ? ?
* ch . 2.2 and 2.3 ) .
the sinr from @xmath98 to @xmath101 with interference from @xmath34 is defined as @xmath102 let @xmath83 and @xmath85 be two pp s on the plane .
we do not assume any particular dependence between @xmath83 and @xmath85 .
in particular one may think of @xmath103 .
let @xmath104 and @xmath105 . the sinr network with _ back - bone _ @xmath83 and _ interferers _
@xmath85 is defined as a graph @xmath106 with nodes in @xmath83 and edges joining any two nodes @xmath107 when @xmath108 and @xmath109 .
the snr graph ( i.e , the graph without interference ) is defined as @xmath110 , which is equivalent also to taking @xmath111 .
observe that the snr graph @xmath112 corresponds to the gilbert s network with nodes in @xmath83 of communication range @xmath113 .
percolation in the above graphs is existence of an infinite connected component in the graph - theoretic sense .
firstly , we consider the case when the backbone nodes are distributed according to poisson pp @xmath114 , for some @xmath115 . we shall use @xmath116 and @xmath117 to denote the corresponding sinr and snr graphs respectively .
recall from [ sss.gilbert ] that @xmath118 is the critical intensity for percolation of the poisson gilbert s network of communication range @xmath3 .
the following result guarantees the existence of @xmath119 such that for any homogeneous sub - poisson pp of interferers @xmath85 the sinr network @xmath120 percolates provided @xmath121 percolates .
[ thm : sinr_poisson_perc ] let @xmath122 and @xmath123 for some @xmath124 .
then there exists @xmath125 such that @xmath120 percolates for @xmath13 . the proof is given in the appendix .
recall that we have not assumed the independence of @xmath85 and @xmath126 .
in particular , one can take @xmath127 where @xmath128 are some _ external _ interferers .
if @xmath128 is @xmath1-sub - poisson and independent of @xmath41 then @xmath85 is also @xmath1-sub - poisson ( cf ( * ? ? ?
* proposition 3.2 ) ) .
the result of proposition [ thm : sinr_poisson_perc ] in the special case of @xmath129 was proved in @xcite .
the present extension allows for any @xmath1-sub - poisson pattern of independent external interferers .
the proof of our result can be also modified ( which will not be presented in this version of the paper ) to allow for external interferer s @xmath128 possibly dependent of the backbone . in this full generality
the result says that _ any external pattern of homogeneous @xmath1-sub - poisson interferers added to the sinr network of @xcite can not make the giant component of this network to disappear , provided the interference cancellation factor @xmath10 is appropriately adjusted_. since an @xmath1-sub - poisson point process can be of arbitrarily large intensity , the above observation can be loosely rephrased in the following form : _ it is not the density of interferers that matters for the network connectivity , but their structure ; @xmath1-sub - poisson interferers do not hurt essentially the network connectivity_. we shall now consider the case when the backbone nodes are formed by a sub - poisson pp . in this case , we can give a weaker result , namely that with appropriately increased signal power @xmath90 , the sinr graph will percolate for small interference parameter @xmath119 .
this corresponds to an early version of the result for the poisson sinr network , proved in @xcite , where the percolation of the sinr network is guaranteed for the intensity of nodes possibly larger than the critical one in the corresponding snr network .
[ thm : perc_sinr_sub - poisson ] let @xmath130 for some @xmath131 and @xmath132 for some @xmath124 and also assume that @xmath133 for all @xmath134 .
then there exist @xmath135 such that @xmath106 percolates . the proof is given in the appendix . as in proposition [ thm : sinr_poisson_perc ]
, we have not assumed the independence of @xmath85 and @xmath83 .
in particular one can take @xmath136 , where @xmath128 and @xmath83 are independent and @xmath0-sub - poisson .
now we want to make some remarks on the relation between @xmath0 ordering and some model characteristics usually called capacities , including the so - called stochastic - geometric capacity functional , as well as some other capacity quantifiers in the information - theoretic sense .
our main tool is the following result proved in ( * ? ? ?
* theorem 2.1 ) .
it says that @xmath137 implies the same ordering of the respective shot - noise fields ( with an arbitrary non - negative response function ) @xmath138 for any @xmath139 .
this implies in particular that @xmath140\le { { \mathbf e}}\bigl[\exp\bigl\{s\sum_{i=1}^ni_{\phi_2}(x_i)\bigr\}\bigr]\ ] ] for _ both positive and negative @xmath141_. _ capacity functional _
@xmath142 of a random set @xmath143 is defined as @xmath144 for all bounded borel sets @xmath40 .
a fundamental result of stochastic geometry , called the choquet s theorem ( cf @xcite ) says that the capacity functional defines entirely the distribution of a random closed set .
the complement of it , @xmath145 , is called _ void probability functional_. inequality ( [ e.shot-noise-lt ] ) allows to compare capacity functionals and void probabilities of boolean models generated by @xmath0 ordered pp s . indeed , note that @xmath146 = { { \mathbf e}}\bigl[\exp\bigl\{\sum_{x\in\phi}h(x)\bigr\}\!\bigr],\ ] ] where @xmath147 .
note that the latter expression has a form of the shot - noise variable and thus using the inequality analogous to ( [ e.shot-noise-lt ] ) for shot - noise with the response function @xmath148 ( cf ( * ? ? ?
* theorem 2.1 ) for such a generalization ) we observe that if @xmath137 then @xmath149 in other words , one can say that @xmath0 smaller pp s exhibit smaller voids . in section [ ss.simulation ] we will show some simulation examples which illustrate this statement .
we focus now on capacity quantifiers in the sense of communication theory .
inequality ( [ e.shot-noise-lt ] ) with @xmath150 and @xmath151 and poisson @xmath152 will be used in the proof of proposition [ thm : sinr_poisson_perc ] ( see appendix ) to show that the _ lower level - sets @xmath153 of the interference field generated by @xmath85 percolate through peierls argument _ ( cf .
* proposition 14.1.4 ) ) for sufficiently large @xmath154 .
similarly ( [ e.shot-noise-lt ] ) with @xmath155 can be used to prove that the _ upper level - sets of the interference field @xmath156 generated by @xmath85 percolate through peierls argument for sufficiently large level values @xmath154_. having observed this _
double _ impact of sub - poisson assumption on @xmath85 it is not evident whether the threshold value @xmath157 of the interference reduction factor is larger in sub - poisson network than in the corresponding poisson one .
note that @xmath157 can be related to the information - theoretic capacity ( throughput ) that can be sustained on the links of the sinr graph . in what follows
we will try to explain how sub - poisson assumption can impact some quantifiers of the network capacity .
it is quite natural to consider a network capacity characteristic @xmath158 that depends on some interference field @xmath159 through some @xmath2 function @xmath27 .
we will give a few simple examples of such characteristics in what follows .
then , a _ larger in @xmath2 order interference field ( more variable ! ) @xmath160 leads to larger average capacity @xmath161$]_. in particular in the case of shot - noise interference field @xmath156 , by ( [ e.shot-noise-dcx ] ) we conclude that _ larger in @xmath0 order pp @xmath85 ( clusters more ! ) leads to larger average capacity @xmath161$]_. let us illustrate this somewhat surprising observation by two simple examples .
[ ex.shannon ] let @xmath162 for some constants @xmath163 and random @xmath159 .
clearly this is a decreasing convex function of @xmath159 and larger in convex order @xmath159 gives larger mean capacity @xmath164 $ ] .
the smallest value of @xmath164 $ ] given @xmath165 $ ] is attained for constant @xmath166 $ ] .
assume a random channel fading @xmath167 with convex tail distribution function @xmath168 .
consider @xmath169 for some constant @xmath170 .
assuming independence of @xmath167 and @xmath159 we have @xmath161={{\mathbf e}}[g(t(n+i))]$ ] , which is by our assumption expectation of a convex function of @xmath159 and the same conclusion can be made as in example [ ex.shannon ] .
the above general form of the expression for the outage capacity can be found in many more detailed models based on pp , in particular in the bipolar model of spatial aloha in ( * ? ? ?
* chapter 16 ) .
see also @xcite for a multidimensional version of this observation .
similar conclusion can be made for the ergodic shannon capacity @xmath171 $ ] of this channel .
let us take one step further and consider the interference created by the original pattern of network points ( the back - bone process @xmath83 ) rather than an external ( independent ) interference field .
in this case the interference @xmath172 at receiver @xmath98 usually has to be considered under the so - called palm distribution of @xmath83 _ given _ the location of the emitter ( in @xmath83 ) of this receiver @xmath98 ( again see e.g. the bipolar model of spatial aloha in ( * ? ? ?
* chapter 16 ) for a detailed example ) .
the problem is that @xmath0 ordering of @xmath83 implies only @xmath1 ordering of the respective palm versions of @xmath83 ; cf . @xcite .
the fundamental reason for the required `` extra '' _ increasing _ property of the comparable functions is that a smaller in @xmath0 pp @xmath83 ( having less clustering or even some point `` repulsion '' ) will have under palm probability potentially _ fewer points in vicinity of the conditioned point_. this potentially decreases interference created locally near this point , thus potentially increases our capacity characteristic . more formally : having @xmath1 ordered of @xmath83 under palm probability @xmath173 and the capacity expressed as a @xmath2 function of @xmath174 we can not conclude any inequality for @xmath175 $ ] . the situation is naturally inverted when we are dealing with pp which are @xmath0 larger than poisson pp .
a detailed analysis in @xcite of the outage capacity in the bipolar model generated by some poisson - poisson cluster pp known as neyman - scott pp ( which is @xmath0 larger than poisson pp ) confirms the above observations .
namely , for smaller transmission distance the negative impact of clustering ( locally more interferers ) decreases the outage capacity , while for larger transmission distance the positive impact of interferers being more clustered increases this capacity .
percolation models have been also shown useful to study the transport ( multi - hop ) capacity of ad - hoc networks .
for example , in @xcite , by using a specific multi - hop transmission strategy that involves percolation theory models , it was shown that a poisson sinr network can achieve capacity rate of the order of @xmath176 bits per unit of time and per node , thus closing the gap with respect to the rate @xmath177 shown achievable in poisson networks in @xcite . an interesting and open question ( particularly in view of what was shown above )
is whether these capacity results can be extended to sub - poisson networks .
from ( * ? ? ? * section 5.2 , 5.3 ) , we have a rich source of examples of cox ( doubly stochastic poisson ) pp s comparable in @xmath0 and @xmath2 order .
in particular , we know that the so called lvy - based cox pp ( with poisson - poisson cluster pp as a special case ) is @xmath0 larger than the poisson pp of the same mean intensity . thus , they can be called as _ @xmath0-super - poisson_. however , note that any cox pp , whose ( random ) realizations of the intensity are almost surely bounded by some constant , can by coupled with ( constructed as a subset of ) a poisson pp with intensity equal to this constant .
consequently such cox processes are trivially @xmath1-sub - poisson . in the remaining part of this section we concentrate on the construction of examples of @xmath0-sub - poisson pp s .
consider some lattice ( e.g. the planar hexagonal one , i.e. , the usual `` honeycomb '' model often considered in cellular network context ) .
let us `` perturb '' this `` ideal '' pattern of points as follows
. for each point of this lattice , say @xmath178 , let us generate independently , from some given distribution , a random number , say @xmath179 , of nodes .
moreover , instead of putting these nodes at @xmath180 , let us translate each of these nodes independently from @xmath178 , by vectors , having some given spatial distribution ( say for simplicity , of bounded support ) .
the resulting network , called _ perturbed lattice _ , can be seen as _ replicating and dispersing points from the original lattice_. interesting observations are as follows .
* if the number of replicas @xmath179 are poisson random variables , then the perturbed lattice is a poisson p.p .
* if moreover the node displacement is uniform in the voronoi cell of the original lattice , then the resulting perturbed lattice is homogeneous poisson process @xmath41 with @xmath4 related to @xmath181 $ ] and the original lattice density .
* now , if @xmath179 are convexly smaller than some poisson variable are smaller than the respective expectations for poisson variable of the same mean ; a special case is when @xmath182 is constant ] then the perturbed lattice is @xmath0-sub - poisson .
this is so because the perturbed lattice pp is an independent sum of countably many pp formed by the perturbations of every vertex of the lattice .
this means that @xmath183 where @xmath184 are independent pp . from (
* proposition .
3.2 ) , we have that if @xmath184 are @xmath0 ordered for every @xmath178 , then so is @xmath34 . since each of these pp is formed by @xmath179 i.i.d .
perturbations , it can be shown that these pp s are @xmath0 ordered if their respective @xmath179 s are convexly ordered .
this follows from proving that the following function @xmath185 is convex in its argument @xmath15 : @xmath186 for any @xmath0 function @xmath27 , disjoint bounded borel subsets @xmath187 and any fixed @xmath178 of the original lattice . *
consequently , if @xmath179 s are convexly smaller than some poisson random variable and moreover the node displacement is uniform in the voronoi cell of the original lattice , then the resulting perturbed lattice is @xmath0-smaller than the respective homogeneous poisson process @xmath41 .
the interest in the above perturbed lattice models in the networking context stems from the fact that they provide some spectrum of models that ranges from periodic grids , usually considered in cellular network context , to ad - hoc networks almost exclusively considered under poisson assumptions . in section [ ss.simulation ] we will show some samples of perturbed lattices . on the theoretical side ,
the interest in perturbed lattices stems from their relations to zeros of gaussian analytic functions ( gafs ) ( see @xcite ) .
more precisely @xcite shows that zeros of some gafs have the same distribution as points of some `` non - independently '' perturbed lattice .
another class of pp s , which can be shown as _
weakly sub - poisson _ , are stationary determinantal pp s . for a quick introduction
refer @xcite .
these pp s are defined by their joint intensities satisfying the following relation for all @xmath52 : @xmath188 for some hermitian , non - negative definite , locally square integrable kernel @xmath189 . then by hadamard s inequality for hermitian , non - negative definite matrices @xmath190
, we have that _ stationary determinantal pp are weakly sub - poisson_. these pp are considered as examples of pp whose points `` exhibit repulsion '' . though zeros of gafs are related to determinantal pp s , curiously enough only zeros of i.i.d .
gaussian power series ( i.e , @xmath191 , with @xmath192 i.i.d .
standard complex normal ) are proved to be determinantal pp s ( see @xcite ) .
now , we will show some examples of simulated patterns of perturbed lattices .
we consider hexagonal pattern of original ( unperturbed ) points ( cf .
figure [ f.lattice1 ] , upper - left plot ) .
the replicas are always displaced uniformly in the voronoi cell of the given point of the original point .
consequently , when the numbers of replicas @xmath91 have poisson distribution then the corresponding perturbed lattice is poisson pp .
figure [ f.lattice1 ] , lower - middle plot )
. we always take @xmath181=1 $ ] . in order to generate sub - poisson pp
s we take @xmath91 having binomial distribution @xmath193 .
this family of distributions can be shown to be _ convexly increasing in @xmath15 _ and _ convex upper - bounded by poisson _
@xmath194 distribution .
note on figure [ f.lattice1 ] that in the case @xmath195 , where we have only point displacement ( no replications ) , the resulting perturbed lattice looks much more `` regular '' ( less visible voids , less clusters ) than poisson pp .
this regularity becomes less evident when @xmath15 increases and already for @xmath196 it is difficult to distinguish the perturbed lattice patterns from poisson pp .
however , a more detailed study of gilbert graphs on them ( not presented here ) up to approximately @xmath197 show that they still have significantly different clustering structure than that of poisson pp .
the lower - right plot on figure [ f.lattice1 ] shows an example of perturbed lattice @xmath0 _ larger _ than poisson pp .
it is a doubly stochastic poisson pp ( cox pp ) , where the mean value @xmath198 of the number of replicas @xmath91 is first sampled independently for each original lattice point , with @xmath199 , for some @xmath200 and then , given @xmath198 , @xmath91 has poisson distribution @xmath201 .
this construction is an example of the ising - poisson cluster pp considered in ( * ? ? ?
* section 5.1 ) .
another possibility to generate the number of replicas @xmath91 convexly larger than poisson variable is to `` scale - up '' the poisson variable of a smaller intensity , i.e. , to take @xmath202 where @xmath203 for arbitrary @xmath204 .
in figure [ f.gilbert1 ] we consider a larger simulation window comprising about @xmath205 points of the original hexagonal lattice .
we generate points of the perturbed lattices with @xmath206 for @xmath207 and @xmath208 to approximate poisson pp . for each @xmath15
we show the gilbert graph with the communication radius @xmath3 for which the largest component ( the highlighted one ) starts to significantly out - number all other components .
more precisely for each given simulated pattern of points we find the smallest @xmath209 ( up to the second decimal place ) such that the largest component in the simulation window contains about 60% of the simulated points ( cf .
the bar - plots showing the empirical fraction of the number of points in 10 largest components ) . observing values of @xmath210
we conjecture that _ the critical radius @xmath211 for the percolation of the gilbert graph on the considered perturbed lattices is increasing in @xmath15_. note that for the unperturbed lattice @xmath212 ( the distance between adjacent nodes in our hexagonal lattice ) and for poisson pp of the same intensity @xmath213 is known to be close to @xmath214 .
a more exhaustive numerical study can be found in the extended version of the paper @xcite .
+ hexagonal lattice + @xmath215 + @xmath216 + @xmath217 + @xmath194 + @xmath218
we have extended two results on nontrivial phase transition in percolation of gilbert s and sinr graphs from poisson to @xmath0-sub - poisson pp s .
this means that , regarding existence of this phase transition , poisson pp represents a worst - case scenario within this class of pp s . a natural question in this context is as follows .
consider @xmath137 .
does this ordering imply that the corresponding critical communication ranges in the gilbert s model are ordered as well @xmath219 ?
the answer in the full generality might be negative , as we expect some counterexamples of @xmath0-super - poisson pp s with the critical communication ranges degenerating to @xmath220 or @xmath221 .
however , we conjecture that the critical communication range of a stationary @xmath0-sub - poisson pp s of intensity @xmath4 is smaller than that of a poisson pp of intensity @xmath4 . in this paper we have also discussed the impact of @xmath0 ordering on some local quantifiers of the capacity in networks .
an interesting open question in this context is the relation between @xmath0 ordering and the transport capacity in these networks .
finally , we brought attention to the so called perturbed lattices , which can provide a spectrum of pp s monotone in @xmath0 order that ranges from periodic grids , usually considered in cellular network context , to poisson pp s and to various more clustered pp s including doubly stochastic poisson pp s .
[ s.appendix ]
in this section we prove our main results of this paper .
( _ of proposition _ [ thm : perc_sub - poisson_pp ] ) denote by @xmath222 , the discrete graph formed by the vertexes of @xmath223 , where @xmath224 is the @xmath75-dimensional integer lattice , and edges between @xmath225 such that @xmath226^d\ } \cap \{z_j + [ -\frac{r}{2},\frac{r}{2}]^d\ } \neq \emptyset.$ ] now we define a site percolation on @xmath222 induced by the pp @xmath34 : @xmath227 $ ] where @xmath228^d.$ ] note that when @xmath229 does not percolate , @xmath70 does not percolate and if @xmath230 percolates , so does @xmath70 .
the standard technique to show non - percolation is to show that the expected number of ( self - avoiding ) paths of length @xmath15 starting at the origin in the random sub - graph induced by opened sites of the percolation model tends to zero as @xmath231 .
the probability of a path @xmath232 of length @xmath15 in @xmath233 being supported by open sites is @xmath234 where the first inequality is due to markov s inequality and the last inequality is by weak sub - poisson property of @xmath34 , which in particular is implied by @xmath1-sub - poisson property ( cf .
( [ e.weakly-sub-poi ] ) ) .
since the number of paths starting at the origin , of length @xmath15 in @xmath233 , is bounded by @xmath235 , we have that the expected number of paths of length @xmath15 is at most @xmath236 and this tends to @xmath220 for @xmath66 small enough .
this shows that there exists @xmath237 ( depending also on the dimension ) such that @xmath238 . for the upper bound , we use the peierls argument ( cf .
@xcite ) on the site percolation model induced by @xmath230 on @xmath239 to use this argument , one needs to estimate the probability of the site - percolation model not intersecting a path @xmath232 .
this probability can be expressed as @xmath240 . and
is smaller than @xmath241 by the @xmath2-sub - poisson property of @xmath34 and ( * ? ? ?
* proposition 4.1 ) .
now by choosing @xmath66 large enough this probability can be made as small as we wish and so the peierls argument can be used .
( _ of proposition _ [ thm : sinr_poisson_perc ] ) we follow the proof given in @xcite . assuming @xmath242
, one observes first that the graph @xmath117 also percolates with any slightly larger constant noise @xmath243 , for some @xmath244 .
essential for the original proof of the result is to show that the level - set @xmath153 of the interference field percolates ( contains an infinite connected component ) for sufficiently large @xmath154 .
suppose that it is true .
then taking @xmath245 one has percolation of the level - set @xmath246 .
the main difficulty consists in showing that @xmath117 with noise @xmath243 percolates _ within _ an infinite connected component of @xmath247 .
this was done in @xcite , by mapping both models @xmath117 and the level - set of the interference to a discrete lattice and showing that both discrete approximations not only percolate but actually satisfies a stronger , sufficient condition for percolation , related to the peierls argument ( * ? ? ?
* proposition 14.1.4 ) .
we follow exactly the same steps and the only fact that we have to prove , regarding the interference , is that there exists a constant @xmath248 such that for arbitrary @xmath204 and arbitrary choice of locations @xmath249 one has @xmath250 . to this regard , as in @xcite , using the chernoff bound we dominate this probability by @xmath251 $ ] with arbitrary @xmath150 .
the crucial observation for our extension of the original proof is that @xmath252 $ ] is an @xmath1 function . by the assumption @xmath253 and ( * ? ? ?
* theorem 2.1 ) we have @xmath254\le { { \mathbf e}}\bigl[\exp\bigl\{s\sum_{i=1}^ni_{\phi_\mu}(x_i)\bigr\}\bigr]\ ] ] and we can use the explicit form of the laplace transform of the poisson shot - noise ( the right - hand - side in the above inequality ) to prove , exactly as in @xcite , that for sufficiently small @xmath141 it is not larger than @xmath255 for some constant @xmath256 which depends on @xmath46 but not on @xmath154 .
this completes the proof .
( _ of proposition _ [ thm : perc_sinr_sub - poisson ] ) in this scenario , increased power is equivalent to increased radius in the gilbert s model associated with the sinr model . from this observation
, it follows that by using the discrete mapping and arguments as in the proof of proposition [ thm : perc_sub - poisson_pp ] , we obtain that with increased power the associated gilbert s model percolates .
then , we use the approach from the proof of proposition [ thm : sinr_poisson_perc ] to obtain a @xmath119 such that the sinr network percolates as well . | we consider a class of point processes ( pp ) , which we call _ sub - poisson _ ; these are pp that can be directionally - convexly ( @xmath0 ) dominated by some poisson pp .
the @xmath0 order has already been shown in @xcite useful in comparing various point process characteristics , including ripley s and correlation functions as well as shot - noise fields generated by pp , indicating in particular that smaller in the @xmath0 order processes exhibit more regularity ( less clustering , less voids ) in the repartition of their points . using these results , in this paper we study the impact of the @xmath0 ordering of pp on the properties of two continuum percolation models , which have been proposed in the literature to address macroscopic connectivity properties of large wireless networks .
as the first main result of this paper , we extend the classical result on the existence of phase transition in the percolation of the gilbert s graph ( called also the boolean model ) , generated by a homogeneous poisson pp , to the class of homogeneous sub - poisson pp .
we also extend a recent result of the same nature for the sinr graph , to sub - poisson pp . finally , as examples we show that the so - called perturbed lattices are sub - poisson .
more generally , perturbed lattices provide some spectrum of models that ranges from periodic grids , usually considered in cellular network context , to poisson ad - hoc networks , and to various more clustered pp including some doubly stochastic poisson ones .
percolation , @xmath0 order , gilbert s graph , boolean model , sinr graph , wireless network , poisson point process , perturbed lattice , determinantal point process , connectivity , capacity |
the _ eigenvalues _ of an @xmath4 complex matrix @xmath13 are the roots in @xmath9 of its characteristic polynomial .
we label them @xmath14 so that @xmath15 with growing phases
spectral radius _ is @xmath16 .
we also denote by @xmath17 the _ singular values _ of @xmath13 , defined for all @xmath18 by @xmath19 where @xmath20 is the conjugate - transpose .
the matrix @xmath13 maps the unit sphere to an ellipsoid , the half - lengths of its principal axes being the singular values of @xmath13 .
the _ operator norm _ of @xmath13 is @xmath21 the matrix @xmath13 is singular iff @xmath22 , and if not then @xmath23 . if @xmath13 is normal ( i.e. @xmath24 ) then @xmath25 for every @xmath26 . beyond normal matrices ,
the relationships between the eigenvalues and the singular values are captured by the weyl inequalities ( see lemma [ le : weyl ] ) . let us define the discrete probability measures @xmath27 from now on , we denote `` @xmath28 '' the weak convergence of probability measures with respect to bounded continuous functions .
we use the abbreviations _
_ , _ a.a . _ , and _ a.e . _ for _ almost surely _ , _ lebesgue almost all _ , and _ lebesgue almost everywhere _ respectively .
the notation @xmath29 means _
large enough @xmath30_. let @xmath31 be an infinite table of i.i.d .
complex random variables with finite positive variance @xmath32 .
if one defines the square @xmath4 complex random matrix @xmath33 then the quartercircular law theorem ( universal square version of the marchenko - pastur theorem , see @xcite ) states that a.s .
@xmath34 where @xmath35 is the quartercircular law on the real interval @xmath36 $ ] with lebesgue density @xmath37}(x).\ ] ] additionally , it is shown in @xcite that @xmath38 concerning the eigenvalues , the famous girko circular law theorem states that a.s.@xmath39 where @xmath40 is the uniform law on the disc @xmath41 , known as the circular law .
this statement was established through a long sequence of partial results @xcite , the general case being finally obtained by tao and vu @xcite . from we have a.s.@xmath42 for any fixed @xmath43 , and we get from @xcite and that if additionally @xmath44 and @xmath45 then a.s.@xmath46 the behavior of the ratio operator - norm / spectral - radius suggests that @xmath47 is far from being an asymptotically normal matrix .
following @xcite , if @xmath48 while @xmath45 then a.s .
@xmath49 at speed @xmath50 while @xmath51 remains bounded .
the proof of is partly but crucially based on a polynomial lower bound on the smallest singular value proved in @xcite : for every @xmath52 , there exists @xmath53 such that for any deterministic complex @xmath4 matrix @xmath13 with @xmath54 we have @xmath55 in particular , by the first borel - cantelli lemma , there exists @xmath53 which may depend on @xmath56 such that a.s .
@xmath57 is invertible with @xmath58 for @xmath29 . from now on and unless otherwise stated @xmath31 is an infinite array of nonnegative real random variables with mean @xmath59 and finite positive variance @xmath60 .
let us define the event @xmath61 since @xmath62 we get @xmath63 and thus @xmath64 by the first borel - cantelli lemma , a.s . for @xmath29 , one can define the @xmath4 matrix @xmath3 by @xmath65 the matrix @xmath3 is markov since its entries belong to @xmath66 $ ] and each row sums up to @xmath67 .
we have @xmath68 where @xmath33 and @xmath69 is the @xmath4 diagonal matrix defined by @xmath70 we may define @xmath3 and @xmath69 for all @xmath71 by setting , when @xmath72 , @xmath73 for all @xmath74 and @xmath75 .
the matrix @xmath3 has equally distributed dependent entries . however , the rows of @xmath3 are i.i.d . and follow an exchangeable law on @xmath76 supported by the simplex @xmath77^n : p_1+\cdots+p_n=1\}.\ ] ] from now on , we set @xmath78 .
this is actually no loss of generality since the law of the random matrix @xmath3 is invariant under the linear scaling @xmath79 for any @xmath80 .
since @xmath81 , the uniform law of large numbers of bai and yin ( * ? ? ?
2 ) states that a.s.@xmath82 this suggests that @xmath8 is approximately equal to @xmath83 for @xmath29 .
one can then expect that and hold for @xmath8 .
our work shows that this heuristics is valid .
there is however a complexity gap between and , due to the fact that for nonnormal operators such as @xmath3 , the eigenvalues are less stable than the singular values under perturbations , see e.g. the book @xcite .
our first result below constitutes the analog of the universal marchenko - pastur theorem for @xmath8 , and generalizes the result of the same kind obtained in @xcite in the case where @xmath84 follows an exponential law .
[ th : singular ] we have a.s . @xmath85
our second result provides some estimates on the largest singular values and eigenvalues .
[ th : localization ] we have @xmath86 . moreover , if @xmath45 then a.s .
@xmath87 our third result below is the analogue of for our random markov matrices .
when @xmath84 follows the exponential distribution of unit mean , theorem [ th : circular ] is exactly the circular law theorem for the dirichlet markov ensemble conjectured in @xcite .
note that we provide , probably for the first time , an almost sure circular law theorem for a matrix model with dependent entries under a finite positive variance assumption .
[ th : circular ] if @xmath84 has a bounded density then a.s .
@xmath88 the proof of theorem [ th : circular ] is crucially based on the following estimate on the norm of the resolvent of @xmath8 .
it is the analogue of for our random markov matrices .
[ th : least - singular ] if @xmath84 has a bounded density then for every @xmath89 there exists @xmath53 such that for any @xmath90 with @xmath91 , for @xmath29 , @xmath92 in particular , for some @xmath53 which may depend on @xmath93 , a.s . for @xmath29 ,
the matrix @xmath94 is invertible with @xmath95 .
the proofs of theorems [ th : singular]-[th : localization]-[th : circular]-[th : least - singular ] are given in sections [ se : singular]-[se : localization]-[se : circular]-[se : least - singular ] respectively .
these proofs make heavy use of lemmas given in the appendices [ se : logpot]-[se : spest]-[se : adle ] .
the matrix @xmath3 is the markov kernel associated to the weighted oriented complete graph with @xmath30 vertices with one loop per vertex , for which each edge @xmath96 has weight @xmath97 .
the skeleton of this kernel is an oriented erds - rnyi random graph where each edge exists independently of the others with probability @xmath98 .
if @xmath99 then @xmath3 has a complete skeleton , is aperiodic , and @xmath67 is the sole eigenvalue of unit module @xcite .
the nonoriented version of this graphical construction gives rise to random reversible kernels for which a semicircular theorem is available @xcite .
the bounded density assumption forces @xmath99 .
since @xmath3 is markov , we have that for every integer @xmath100 , @xmath101 where @xmath102 is simply , conditional on @xmath3 , the probability of a loop of length @xmath103 rooted at @xmath104 for a markov chain with transition kernel @xmath3 .
this provides a probabilistic interpretation of the moments of the empirical spectral distribution @xmath105 of @xmath3 .
the random markov matrix @xmath3 is a
_ random environment_. by combining theorem [ th : localization ] with theorem [ th : circular ] and the identity , we get that for every fixed @xmath100 , a.s.@xmath106 the bounded density assumption in theorem [ th : circular ] is only due to the usage of theorem [ th : least - singular ] in the proof .
we believe that theorem [ th : least - singular ] ( and thus [ th : circular ] ) is valid without this assumption , but this is outside the scope of the present work , see remark [ rm : singass ] and figure [ fi : simus ] .
our proof of theorem [ th : circular ] is inspired from the tao and vu proof of based on girko hermitization , and allows actually to go beyond the circular law , see remarks [ rm : clt]-[rm : beyond ] . concerning the extremes , by theorems
[ th : localization]-[th : circular ] , a.s.@xmath107 also , a.s .
the `` spectral gap '' of @xmath3 is a.s . of order @xmath108 ( compare with the results of @xcite ) .
note that in contrast with , we have from theorems [ th : singular]-[th : localization ] , a.s . @xmath109
numerical simulations suggest that if @xmath45 then a.s.@xmath110 unfortunately , our proof of theorem [ th : localization ] is too perturbative to extract this result .
following @xcite , one can also ask if the phase @xmath111 converges in distribution to the uniform law on @xmath112 $ ] as @xmath11 .
another interesting problem concerns the behavior of @xmath113 and the fluctuation as @xmath11 of the extremal singular values and eigenvalues of @xmath8 , in particular the fluctuation of @xmath114 .
classical results on the connectivity of erds - rnyi random graphs @xcite imply that a.s .
for @xmath29 the markov matrix @xmath3 is irreducible .
hence , a.s . for @xmath29 ,
the markov matrix @xmath3 admits a unique invariant probability measure @xmath115 .
if @xmath115 is seen as a row vector in @xmath116 then we have @xmath117 .
let @xmath118 be the uniform law on @xmath119 , which can be viewed as the vector @xmath120 . by denoting @xmath121 the total variation ( or @xmath122 ) distance on @xmath116
, one can ask if a.s.@xmath123 recall that the rows of @xmath3 are i.i.d . and follow an exchangeable law @xmath124 on the simplex @xmath116 . by `` exchangeable '' we mean that if @xmath125 then for every permutation @xmath126 of @xmath119 the random vector @xmath127 follows also the law @xmath124 .
this gives @xmath128 and therefore @xmath129 .
one can ask if the results of theorems [ th : singular]-[th : localization]-[th : circular]-[th : least - singular ] remain essentially valid at least if @xmath3 is a real @xmath4 random matrix with i.i.d .
rows such that for every @xmath130 , @xmath131 these rates in @xmath30 correspond to the dirichlet markov ensemble for which @xmath84 follows an exponential law and where @xmath124 is the dirichlet law @xmath132 on the simplex @xmath116 .
another interesting problem is the spectral analysis of @xmath3 when the law of @xmath84 has heavy tails , e.g. @xmath133 with @xmath134 and where @xmath135 is a uniform random variable on @xmath66 $ ] , see for instance @xcite for the reversible case .
a logical first step consists in the derivation of a heavy tailed version of for @xmath47 .
this program is addressed in a separate paper @xcite . in the same spirit
, one may ask about the behavior of @xmath136 where `` @xmath137 '' denotes the schur - hadamard entrywise product and where @xmath13 is some prescribed profile matrix .
let us start by an elementary observation . the second moment @xmath138 of @xmath139
is given by @xmath140 by using together with the standard law of large numbers , we get that a.s.@xmath141 it follows by the markov inequality that a.s .
the sequence @xmath142 is tight .
however , we will not rely on tightness and the prohorov theorem in order to establish the convergence of @xmath143
. we will use instead a perturbative argument based on the special structure of @xmath3 and on .
namely , since @xmath144 , we get from , for all @xmath26 , @xmath145 additionally , we get from that a.s.@xmath146 this gives that a.s.@xmath147 from , , and , we get that a.s . for @xmath29 , @xmath148 and from and again
we obtain that a.s .
@xmath149 now , from , and by denoting @xmath150 the image probability measure of @xmath35 by @xmath151 , a.s.@xmath152 next , using with lemma [ le : perturb ] provides that a.s .
@xmath153 this implies by the change of variable @xmath154 that a.s.@xmath155 which is the desired result . from and , a.s . for @xmath29 , @xmath156 iff @xmath157 , for all @xmath18 , and thus @xmath158 , and
hence the reasoning can avoid the usage of .
note also that is automatically satisfied when @xmath84 is absolutely continuous since the set of singular matrices has zero lebesgue measure . on the other hand ,
it is also worthwhile to mention that if @xmath45 then one can obtain the desired result without using , by using lemma [ le : perturb ] together with , and this reasoning was already used by aubrun for a slightly different model @xcite .
let us define the @xmath4 deterministic matrix @xmath159 .
the random matrix @xmath160 has i.i.d .
centered entries with finite positive variance @xmath2 and finite fourth moment , and consequently , by , a.s.@xmath161 now , since @xmath162 we have , by lemma [ le : thompson ] , @xmath163 and therefore , a.s .
@xmath164 by combining with and we get that a.s .
@xmath165 in particular , this gives from that a.s .
@xmath166 from theorem [ th : singular ] , since @xmath35 is supported by @xmath36 $ ] , we get by using that a.s.@xmath167 next , since @xmath3 is a markov matrix , it is known that @xmath168 let us briefly recall the proof . if @xmath169 then @xmath170 and thus @xmath67 is an eigenvalue of @xmath3 .
next , let @xmath171 be an eigenvalue of @xmath3 and let @xmath172 be such that @xmath173 and @xmath174
. there exists @xmath26 such that @xmath175 . since @xmath176 and @xmath177 we get @xmath178 , which implies .
let us show now that a.s .
@xmath179 let @xmath180 be as in the proof of theorem [ th : singular ] . from and
, we get a.s .
@xmath181 which gives @xmath182
a.s . on the other hand , from
and we get @xmath183 , which gives .
it remains to establish that a.s.@xmath184 indeed , from we get for every non null @xmath4 complex matrix @xmath13 , @xmath185 with @xmath186 and by using ( [ eq : s2]-[eq : la1]-[eq : s1 ] ) , we obtain that a.s.@xmath187 let us start by observing that from the weyl inequality and , a.s.@xmath188 this shows via the markov inequality that a.s .
the sequence @xmath189 is tight .
however , we will not rely directly on this tightness and the prohorov theorem in order to establish the convergence of @xmath189
. we will use instead the girko hermitization of lemma [ le : girko ] . we know , from the work of dozier and silverstein @xcite , that for all @xmath90 , there exists a probability measure @xmath190 on @xmath191 such that a.s.@xmath192 converges weakly to @xmath190 .
moreover , following e.g. pan and zhou ( * ? ? ?
3 ) , one can check that for all @xmath90 , @xmath193 to prove that a.s .
@xmath189 tends weakly to @xmath40 , we start from the decomposition @xmath194 by using , , and lemma [ le : perturb ] , we obtain that for a.a .
@xmath90 , a.s .
@xmath195 now , arguing as in the proof of theorem [ th : singular ] , it follows that for all @xmath90 , a.s .
@xmath196 suppose for the moment that for a.a .
@xmath90 , a.s .
the function @xmath151 is uniformly integrable for @xmath197 .
let @xmath198 by as in section [ se : logpot ] .
lemma [ le : girko ] implies that there exists @xmath199 such that a.s.@xmath200 where @xmath201 is the logarithmic potential of @xmath202 as defined in section [ se : logpot ] .
now by lemma [ le : unicity ] , we obtain @xmath203 , which is the desired result . it thus remains to show that for a.a .
@xmath90 , a.s .
the function @xmath151 is uniformly integrable for @xmath197 . for every @xmath90 , a.s . by the cauchy - schwarz inequality , for all @xmath204 , and @xmath29 , @xmath205 now the markov inequality and give that for all @xmath90 , a.s . for all @xmath204 @xmath206 where the @xmath207 is uniform in @xmath208 .
consequently , for all @xmath90 , a.s.@xmath209 this means that for all @xmath90 , a.s .
the function @xmath210 is uniformly integrable for @xmath197 .
it remains to show that for all @xmath90 , a.s .
the function @xmath211 is uniformly integrable for @xmath197 .
this is equivalent to show that for all @xmath212 , a.s .
@xmath213 for convenience , we fix @xmath90 and set @xmath214 for all @xmath215 .
now we write @xmath216 from theorem [ th : least - singular ] ( here we need the bounded density assumption ) we get that a.s.@xmath217 and it thus remains to show that a.s.@xmath218 this boils down to show that there exists @xmath219 such a.s . for @xmath29 and @xmath220 , @xmath221 to prove it , we adapt an argument due to tao and vu @xcite .
we fix @xmath220 and we consider the matrix @xmath222 formed by the first @xmath223 rows of @xmath224 by the cauchy interlacing lemma [ le : cauchy ] , we get @xmath225 where @xmath226 for all @xmath227 are the singular values of the rectangular matrix @xmath222 in nonincreasing order .
next , by the tao and vu negative moment lemma [ le : tvneg ] , @xmath228 where @xmath229 is the distance from the @xmath230 row of @xmath222 to @xmath231 , the subspace spanned by the other rows of @xmath222 . in particular , we have @xmath232 let @xmath233 be the @xmath230 row of @xmath47 .
since the @xmath230 row of @xmath3 is @xmath234 , we deduce that @xmath235 where @xmath236 is the canonical basis of @xmath76 .
since @xmath237 is independent of @xmath233 and @xmath238 lemma gives @xmath239 ( note that the exponential bound in lemma [ le : concdist ] kills the polynomial factor due to the union bound over @xmath240 ) .
consequently , by the first borel - cantelli lemma , we obtain that a.s .
for @xmath29 , all @xmath241 , and all @xmath242 , @xmath243 now , the uniform law of large numbers gives that a.s .
@xmath244 we deduce that a.s . for @xmath29 , all @xmath220 , and all @xmath245 , @xmath246 finally ,
from we get @xmath247 and holds with @xmath248 .
[ rm : clt ] the same strategy allows a relatively short proof of . indeed , the a.s
. weak convergence of @xmath249 to @xmath250 follows from the girko hermitization lemma [ le : girko ] and the uniform integrability of @xmath151 for @xmath192 as above using .
this direct strategy does not rely on the _ replacement principle _ of tao and vu .
the replacement principle allows a statement which is more general than involving two sequences of random matrices ( the main result of @xcite on universality ) .
our strategy allows to go beyond the circular law , by letting @xmath251 possibly depend on @xmath252 , provided that and the result of dozier and silverstein @xcite hold .
set @xmath253 . if @xmath254 is large enough for @xmath29 , the limit is no longer the circular law , and can be interpreted by using free probability theory @xcite .
[ rm : beyond ] it is likely that the tao and vu replacement principle @xcite allows a universal statement for our random markov matrices of the form @xmath68 , beyond the circular law , by letting @xmath251 possibly depend on @xmath252 .
this is however beyond the scope of the present work .
note that when @xmath255 , one can get some @xmath256 immediately from ( [ eq : basic2 ] , [ eq : sn - exp - pol ] , [ eq : ulln ] ) .
thus , our problem is actually to deal with @xmath257 . fix @xmath89 and @xmath90 with @xmath258 .
we have @xmath259 for an arbitrary @xmath260 , let us define the event @xmath261 by using the union bound and the chebyshev inequality , we get @xmath262 . now with @xmath263 and @xmath264 we obtain @xmath265 for @xmath29 .
since we have @xmath266 we get by , on the event @xmath267 , for @xmath29 , @xmath268 for every @xmath269 .
now , for every @xmath270 , one may select @xmath53 and set @xmath271 such that @xmath272 for @xmath29 .
thus , on the event @xmath267 , for @xmath29 , @xmath273 consequently , for every @xmath270 there exists @xmath53 such that for @xmath29 , @xmath274 the desired result follows if we show that for some @xmath270 depending on @xmath275 , for @xmath29 , @xmath276 let us prove . at this point , it is very important to realize that can not follow form a perturbative argument based on ( [ eq : sn - exp - pol],[eq : basic1],[eq : ulln ] ) since the operator norm of the perturbation is much larger that the least singular value of the perturbed matrix .
we thus need a more refined argument .
we have @xmath277 with @xmath278 .
let @xmath279 be as in lemma [ le : a ] .
for every @xmath18 , let @xmath280 be the @xmath4 permutation matrix for the transposition @xmath281 .
note that @xmath282 and for every @xmath18 , the matrix @xmath283 is @xmath4 lower triangular . for every column vector @xmath284 of the canonical basis of @xmath76 , @xmath285 now , if @xmath286 and @xmath287 are the rows of the matrices @xmath47 and @xmath288 then @xmath289 define the vector space @xmath290 for every @xmath26 . from lemma [ le : rvdist ] , @xmath291 consequently , by the union bound , for any @xmath292 , @xmath293 the law of @xmath294 does not depend on @xmath104 .
we take @xmath295 .
let @xmath296 be a unit normal vector to @xmath297 .
such a vector is not unique , but we just pick one , and this defines a random variable on the unit sphere @xmath298 .
since @xmath299 and @xmath300 , @xmath301 let @xmath302 be the distribution of @xmath296 on @xmath303 . since @xmath296 and @xmath304 are independent , for any @xmath292 , @xmath305 let us fix @xmath306 .
if @xmath307 are as above then @xmath308 now , since @xmath306 , lemma [ le : a ] provides a constant @xmath309 such that for @xmath29 , @xmath310 but @xmath311 implies @xmath312 for some @xmath313 , and therefore @xmath314 .
we first observe that @xmath315 the real random variable @xmath316 is a sum of independent real random variables and one of them is @xmath317 , which is absolutely continuous with a density bounded above by @xmath318 where @xmath319 is the bound on the density of @xmath84 .
consequently , by a basic property of convolutions of probability measures , the real random variable @xmath316 is also absolutely continuous with a density @xmath320 bounded above by @xmath318 , and therefore , @xmath321}\,\varphi(s)\,ds { \leqslant}bk\sqrt{2n}2u.\ ] ] to summarize , for @xmath29 and every @xmath292 , @xmath322 lemma [ le : a ] shows that the constant @xmath323 may be chosen depending on @xmath93 and not on @xmath324 , and holds with @xmath325 by taking @xmath326 such that @xmath327 for @xmath29 .
[ rm : singass ] our proof of theorem [ th : least - singular ] still works if the entries of @xmath47 are just independent and not necessarily i.i.d . provided that the densities are uniformly bounded and that holds .
the bounded density assumption allows to bound the small ball probability @xmath328 uniformly over @xmath329 .
if this assumption does not hold , then the small ball probability may depend on the additive structure of @xmath329 , but the final result is probably still valid . a possible route , technical and uncertain ,
is to adapt the tao and vu proof of . on the opposite side ,
if @xmath84 has a @xmath330-concave density ( e.g. exponential ) then a finer bound might follow from a noncentered version of the results of adamczak et al @xcite . alternatively
, if @xmath84 has sub - gaussian or sub - exponential moments then one may also try to adapt the proof of rudelson and vershynin @xcite to the noncentered settings .
the derivation of an a.s .
lower bound on @xmath331 is an easy task when @xmath332 and @xmath45 , without assuming that @xmath84 has a bounded density .
let us show for instance that for every @xmath90 , a.s .
@xmath333 this lower bound is meaningful only when @xmath332 . for proving ,
we adopt a perturbative approach .
let us fix @xmath90 . by and we get that a.s .
@xmath334 now we write , with @xmath335 , @xmath336 we observe that gives @xmath337 for the symmetric complex matrix @xmath338 we have for any @xmath90 and @xmath29 , @xmath339 on the other hand , since @xmath45 , by and , a.s . for every @xmath90 ,
@xmath340 putting all together , we have shown that a.s .
for any @xmath90 , @xmath341 combined with , this gives finally .
let @xmath342 be a sequence of complex random matrices where @xmath343 is @xmath4 for every @xmath71 , defined on a common probability space @xmath344 .
for every @xmath345 , the set @xmath346 is at most countable and has thus zero lebesgue measure .
therefore , for * all * @xmath345 and * a.a .
* @xmath90 , we have @xmath347 for all * all * @xmath71 .
note that and theorem [ th : least - singular ] imply respectively that for @xmath348 or @xmath349 , this holds for * all * @xmath90 , * a.s . * on @xmath350 , and @xmath351 .
let @xmath198 be the set of probability measures on @xmath9 which integrate @xmath352 in a neighborhood of infinity
. for every @xmath199 , the _ logarithmic potential _ @xmath201 of @xmath202 on @xmath9 is the function @xmath353 $ ] defined for every @xmath90 by @xmath354 for instance , for the circular law @xmath355 of density @xmath356 , we have , for every @xmath90 , @xmath357 see e.g. @xcite .
let @xmath358 be the set of schwartz - sobolev distributions ( generalized functions ) . since @xmath352 is lebesgue locally integrable on @xmath9 , one can check by using the fubini theorem that @xmath201 is lebesgue locally integrable on @xmath9 . in particular , @xmath359 a.e . and @xmath360 .
since @xmath352 is the fundamental solution of the laplace equation in @xmath9 , we have , in @xmath358 , @xmath361 this means that for every smooth and compactly supported `` test function '' @xmath362 , @xmath363 [ le : unicity ] for every @xmath364 , if @xmath365 a.e . then @xmath366 .
since @xmath365 in @xmath358 , we get @xmath367 in @xmath358
. now gives @xmath366 in @xmath358 , and thus @xmath366 as measures since @xmath202 and @xmath302 are radon measures .
if @xmath13 is an @xmath4 complex matrix and @xmath368 is its characteristic polynomial , @xmath369 for every @xmath370 .
we have also the alternative expression @xmath371 the identity bridges the eigenvalues with the singular values , and is at the heart of the following lemma , which allows to deduce the convergence of @xmath372 from the one of @xmath373 .
the strength of this hermitization lies in the fact that in contrary to the eigenvalues , one can control the singular values with the entries of the matrix .
the price payed here is the introduction of the auxiliary variable @xmath324 and the uniform integrability .
we recall that on a borel measurable space @xmath374 , we say that a borel function @xmath375 is _ uniformly integrable _ for a sequence of probability measures @xmath376 on @xmath377 when @xmath378 we will use this property as follows : if @xmath376 converges weakly to @xmath379 and @xmath380 is continuous and uniformly integrable for @xmath376 then @xmath380 is @xmath379-integrable and @xmath381 .
the idea of using hermitization goes back at least to girko @xcite .
however , the proofs of lemmas [ le : girko ] and [ le : unimaj ] below are inspired from the approach of tao and vu @xcite .
[ le : girko ] let @xmath342 be a sequence of complex random matrices where @xmath343 is @xmath4 for every @xmath71 , defined on a common probability space .
suppose that for a.a .
@xmath90 , there exists a probability measure @xmath190 on @xmath191 such that a.s .
* @xmath382 converges weakly to @xmath190 as @xmath11 * @xmath151 is uniformly integrable for @xmath383 then there exists a probability measure @xmath199 such that * a.s .
@xmath384 converges weakly to @xmath202 as @xmath11 * for a.a .
@xmath90 , @xmath385 moreover , if @xmath342 is deterministic , then the statements hold without the `` a.s . ''
let @xmath324 and @xmath350 be such that ( i - ii ) hold . for every @xmath18 ,
define @xmath386 and set @xmath387 .
note that @xmath388 .
thanks to the weyl inequalities and to the assumptions ( i - ii ) , one can use lemma [ le : unimaj ] below , which gives that @xmath384 is tight , that @xmath389 is uniformly integrable for @xmath384 , and that @xmath390 consequently , a.s .
@xmath199 and @xmath391 a.e . for every adherence value @xmath202 of @xmath384 .
now , since @xmath392 does not depend on @xmath202 , by lemma [ le : unicity ] , a.s .
@xmath393 has a unique adherence value @xmath202 , and since @xmath394 is tight , @xmath384 converges weakly to @xmath202 by the prohorov theorem . finally , by , @xmath202 is deterministic since @xmath392 is deterministic , and ( j - jj ) hold .
the following lemma is in a way the skeleton of the girko hermitization of lemma [ le : girko ] .
it states essentially a propagation of a uniform logarithmic integrability for a couple of triangular arrays , provided that a logarithmic majorization holds between the arrays .
[ le : unimaj ] let @xmath395 and @xmath396 be two triangular arrays in @xmath191 . define the discrete probability measures @xmath397 if the following properties hold * @xmath398 and @xmath399 for @xmath29 , * @xmath400 for every @xmath18 for @xmath29 , * @xmath401 for every @xmath18 for @xmath29 , * @xmath402 converges weakly to some probability measure @xmath302 as @xmath11 , * @xmath151 is uniformly integrable for @xmath402 , then * @xmath394 is tight , * @xmath151 is uniformly integrable for @xmath394 , * we have , as @xmath11 , @xmath403 and in particular , for every adherence value @xmath202 of @xmath394 , @xmath404 * proof of ( jjj)*. from the logarithmic majorizations ( ii - iii ) we get , for @xmath29 , @xmath405 and ( v ) gives @xmath406 and @xmath407 for every @xmath18 and @xmath29 .
now , ( iv - v ) give @xmath408 * proof of ( j)*. from ( ii ) and ( v ) we get @xmath409 respectively .
now the tightness of @xmath394 follows from @xmath410 * proof of ( jj ) .
* we start with the uniform integrability in the neighborhood of infinity .
let us show that for @xmath29 , for any @xmath411 there exists @xmath204 such that @xmath412 if @xmath413 then ( iv ) implies @xmath414 for @xmath29 and follows then from . if otherwise @xmath415 then @xmath416 and one can assume that @xmath417 . let us show that there exists a sequence of integers @xmath418 such that @xmath419 and for @xmath29 , @xmath420 for @xmath421 , let @xmath208 be the infimum over all @xmath422 such that @xmath423 there exists @xmath424 such that @xmath425 , and from ( v ) we get @xmath426 if @xmath427 then holds with @xmath428 and @xmath429 .
otherwise , @xmath430 , and if @xmath431 then @xmath432 , while for any @xmath433 , @xmath434 taking @xmath435 with @xmath436 small enough , we deduce that holds . we have thus shown that holds in all case . now , from ( ii ) and we get for every @xmath437 , @xmath438 in particular , by using ( i ) , we get @xmath439 and @xmath440 since @xmath441 , we deduce that holds with @xmath442 . now , by following the same reasoning , with ( ii ) replaced by ( iii ) , we obtain that for all @xmath443 , there exists @xmath444 such that for @xmath29 , @xmath445 which is the counterpart of needed for ( jj ) .
the logarithmic potential is related to the cauchy - stieltjes transform of @xmath202 via @xmath446 in @xmath358 .
the term `` logarithmic potential '' comes from the fact that @xmath201 is the electrostatic potential of @xmath202 viewed as a distribution of charges in @xmath447 @xcite .
the logarithmic energy @xmath448 is up to a sign the voiculescu free entropy of @xmath202 in free probability theory @xcite .
the circular law @xmath40 minimizes @xmath449 under a second moment constraint @xcite . in the spirit of and beyond matrices , the brown @xcite spectral measure of a nonnormal bounded operator @xmath450 is @xmath451 where @xmath452 is the spectral distribution of the self - adjoint operator @xmath453 . due to the logarithm , the brown spectral measure @xmath454 depends discontinuously on the @xmath455-moments of @xmath450 @xcite . for random matrices ,
this problem is circumvented in the girko hermitization by requiring a uniform integrability , which turns out to be a.s .
satisfied for random matrices such as @xmath83 or @xmath8 .
we gather in this section useful lemmas on deterministic matrices . we provide mainly references for the most classical results , and
sometimes proofs for the less classical ones .
[ le : basic ] if @xmath13 and @xmath319 are @xmath4 complex matrices then @xmath456 and @xmath457 and @xmath458 moreover , if @xmath459 is diagonal , then for every @xmath26 @xmath460 [ le : rvdist ] let @xmath13 be a complex @xmath4 matrix with rows @xmath286 . define the vector space @xmath461 .
we have then @xmath462 the argument behind lemma [ le : rvdist ] is buried in @xcite .
we give a proof below for convenience .
since @xmath463 have same singular values , one can consider the columns @xmath464 of @xmath13 instead of the rows .
for every column vector @xmath172 and @xmath26 , the triangle inequality and the identity @xmath465 give @xmath466 if @xmath467 then necessarily @xmath468 for some @xmath26 and therefore @xmath469 conversely , for every @xmath26 , there exists a vector @xmath470 with @xmath471 such that @xmath472 where we used the fact that @xmath473 .
recall that the singular values @xmath474 of a rectangular @xmath475 complex matrix @xmath13 with @xmath476 are defined by @xmath477 for every @xmath478 .
[ le : tvneg ] if @xmath13 is a full rank @xmath475 complex matrix ( @xmath476 ) with rows @xmath479 , and @xmath461 , then @xmath480 [ le : cauchy ] let @xmath13 be an @xmath4 complex matrix
. if @xmath319 is @xmath475 , obtained from @xmath13 by deleting @xmath481 rows , then for every @xmath478 , @xmath482 lemma [ le : cauchy ] gives @xmath483\subset[s_n(a),s_1(a)]$ ] , i.e.row deletions produce a compression of the singular values interval .
another way to express this phenomenon consists in saying that if we add a row to @xmath319 then the largest singular value increases while the smallest is diminished .
closely related , the following result on finite rank additive perturbations .
if @xmath13 is an @xmath4 complex matrix , let us set @xmath484 if @xmath485 and @xmath486 if @xmath487 .
[ le : thompson ] for any @xmath4 complex matrices @xmath13 and @xmath319 with @xmath488 , we have , for any @xmath489 , @xmath490 even if lemma [ le : thompson ] gives nothing on the extremal singular values @xmath491 where @xmath492 or @xmath493 , it provides however the useful `` bulk '' inequality @xmath494 where @xmath495 and @xmath496 are the cumulative distribution functions of @xmath497 and @xmath498 respectively .
[ le : weyl ] for every @xmath4 complex matrix @xmath13 , we have @xmath499 for all @xmath18 , with equality for @xmath500 . in particular , by viewing @xmath501 as a volume , @xmath502 where @xmath286 are the rows of @xmath13 . moreover , for every increasing function @xmath320 from @xmath503 to @xmath503 such that @xmath504 is convex on @xmath503 and @xmath505 , we have @xmath506 for every @xmath18 . in particular , with @xmath507 for every @xmath80 and @xmath500 , we obtain @xmath508 it is worthwhile to mention that and are optimal in the sense that every sequences of numbers satisfying these inequalities are associated to matrices , see @xcite .
lemma [ le : perturb ] below is used in the proof of theorem [ th : singular ] .
we omit its proof since it follows for instance quite easily from the paul lvy criterion on characteristic functions .
[ le : perturb ] let @xmath395 and @xmath396 be triangular arrays of complex numbers .
let @xmath202 be a probability measure on @xmath9 .
@xmath509 lemma [ le : concdist ] below is used for the rows of random matrices in the proof of theorem [ th : circular ] .
[ le : concdist ] let @xmath510 be i.i.d . random variables on @xmath9 with finite positive variance @xmath511 . for @xmath29 and every deterministic subspace @xmath512 of @xmath513 with @xmath514 , setting @xmath515 ,
@xmath516 the proof of lemma [ le : concdist ] is based on a concentration inequality for convex lipschitz functions and product measures due to talagrand @xcite , see also ( * ? ? ?
the power @xmath517 is used here to fix ideas and is obviously not optimal .
this is more than enough for our purposes ( proof of theorem [ th : circular ] ) .
a careful reading of the proof of theorem [ th : circular ] shows that a polynomial bound on the probability with a large enough power on @xmath30 suffices .
we end up this section by a lemma used in the proof of theorem [ th : least - singular ] .
[ le : a ] for every @xmath518 , let us define the @xmath4 complex matrix @xmath519 then for every @xmath90 we have @xmath520 for @xmath29 while @xmath521 and the convergence is uniform on every compact subset of @xmath9 . note that @xmath522 and @xmath523 for every @xmath524 .
moreover , @xmath525 is invertible if and only if @xmath526 and in that case @xmath527 .
it is a special case of the sherman - morrison formula for the inverse of rank one perturbations .
it is immediate to check that @xmath528 for every @xmath518 .
an elementary explicit computation reveals that the symmetric matrix @xmath529 has rank at most @xmath530 , and thus @xmath525 has at least @xmath531 singular values equal to @xmath67 and in particular @xmath532 . from now , let us fix @xmath90 and set @xmath533 and @xmath534 for convenience .
the matrix @xmath13 is nonsingular for @xmath29 since @xmath535 as @xmath11 .
also , we have @xmath536 for @xmath29 . since @xmath13 is lower triangular with eigenvalues @xmath537 , by , @xmath538 where @xmath539 are two singular values of @xmath13 .
we have also @xmath540 which gives @xmath541 . combined with
, we get that @xmath542 are the solution of @xmath543 .
this gives @xmath544 and the @xmath545 is uniform in @xmath324 on every compact . from this formula
we get that @xmath546 and @xmath547 for @xmath29 , and thus @xmath548 and @xmath549 .
the result of lemma [ le : a ] is more than enough for our purposes .
more precisely , a careful reading of the proof of theorem [ th : least - singular ] shows that a polynomial ( in @xmath30 ) lower bound on @xmath550 for @xmath29 , uniformly on compact sets on @xmath324 , is actually enough .
dc is grateful to o. gudon , m. krishnapur , and t. tao for helpful discussions .
cb and dc are grateful to dipartimento di matematica universit roma tre for kind hospitality .
r. adamczak , o. gudon , a. litvak , a. pajor , and n. tomczak - jaegermann , _ smallest singular value of random matrices with independent columns _ , c. r. math .
paris * 346 * ( 2008 ) , no .
15 - 16 , 853856 .
mr mr2441920 ( 2009h:15009 ) g. akemann and e. kanzieper , _ integrable structure of ginibre s ensemble of real random matrices and a pfaffian integration theorem _ , j. stat .
* 129 * ( 2007 ) , no . 5 - 6 , 11591231 .
mr mr2363393 ( 2009a:82024 ) c .- r .
hwang , _ a brief survey on the spectral radius and the spectral distribution of large random matrices with i.i.d .
entries _ , random matrices and their applications ( brunswick , maine , 1984 ) , contemp . math .
50 , amer . math . soc . ,
providence , ri , 1986 , pp .
145152 .
e. b. saff and v. totik , _ logarithmic potentials with external fields _ , grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences ] , vol .
316 , springer - verlag , berlin , 1997 , appendix b by thomas bloom .
e. seneta , _ non - negative matrices and markov chains _ , springer series in statistics , springer , new york , 2006 , revised reprint of the second ( 1981 ) edition [ springer - verlag , new york ; mr0719544 ] .
mr mr2209438
r. c. thompson , _ the behavior of eigenvalues and singular values under perturbations of restricted rank _
, linear algebra and appl .
* 13 * ( 1976 ) , no . 1/2 , 6978 , collection of articles dedicated to olga taussky todd . ,
and @xmath84 follows the bernoulli law @xmath554 . in both graphics ,
the solid circle has radius @xmath555 .
the left hand side graphic is the superposition of the plot of @xmath556 for @xmath557 i.i.d .
simulations of @xmath8 , made with the gnu octave free software .
the right hand side graphic is the vorono tessellation of @xmath556 for a single simulation of @xmath8 .
since @xmath8 has real entries , its spectrum is symmetric with respect to the real axis . on the left hand side graphic
, it seems that the spectrum is slightly more concentrated on the real axis .
this phenomenon , which disappears as @xmath11 , was already described for random matrices with i.i.d . real gaussian entries by edelman @xcite ,
see also the work of akemann and kanzieper @xcite .
our simulations suggest that theorem [ th : circular ] remains valid beyond the bounded density assumption.,title="fig : " ] , and @xmath84 follows the bernoulli law @xmath554 . in both graphics ,
the solid circle has radius @xmath555 .
the left hand side graphic is the superposition of the plot of @xmath556 for @xmath557 i.i.d .
simulations of @xmath8 , made with the gnu octave free software .
the right hand side graphic is the vorono tessellation of @xmath556 for a single simulation of @xmath8 .
since @xmath8 has real entries , its spectrum is symmetric with respect to the real axis . on the left hand side graphic
, it seems that the spectrum is slightly more concentrated on the real axis .
this phenomenon , which disappears as @xmath11 , was already described for random matrices with i.i.d . real gaussian entries by edelman @xcite ,
see also the work of akemann and kanzieper @xcite .
our simulations suggest that theorem [ th : circular ] remains valid beyond the bounded density assumption.,title="fig : " ] | let @xmath0 be i.i.d .
nonnegative random variables with bounded density , mean @xmath1 , and finite positive variance @xmath2 .
let @xmath3 be the @xmath4 random markov matrix with i.i.d .
rows defined by @xmath5 . in particular , when @xmath6 follows an exponential law , the random matrix @xmath3 belongs to the dirichlet markov ensemble of random stochastic matrices .
let @xmath7 be the eigenvalues of @xmath8 i.e. the roots in @xmath9 of its characteristic polynomial .
our main result states that with probability one , the counting probability measure @xmath10 converges weakly as @xmath11 to the uniform law on the disk @xmath12 .
the bounded density assumption is purely technical and comes from the way we control the operator norm of the resolvent . |
one - way quantum computing @xcite is an important approach for quantum computation based on a series of one - qubit measurements starting from a cluster state of a qubit array .
cluster states are highly - entangled states involving all qubits and are typically generated from an ising - like hamiltonian , starting from an initial product state @xmath0 , where @xmath1 . here ,
@xmath2 and @xmath3 are the two states of the @xmath4-th qubit in an @xmath5-qubit system . in ref .
@xcite , we showed that cluster states in charge qubits@xcite can be created by applying a single gate bias pulse , right after preparing the initial product state ( one - step generation method ) , and are more robust against nonuniformities among qubits than decoherence - free ( df ) states @xcite under a noise environment generated by a quantum point contact ( qpc ) detector , which is a sensitive detector of electric charge distribution@xcite .
however , trap sites are often unavoidable in solid - state qubits owing to their small fabrication size and several experiments show that trap states significantly affect electric transport properties of nanoscale devices@xcite . here , we model the trap site as the island ( discrete energy state ) between double qpcs ( dqpc ) and investigate robustness of cluster states in charge qubits measured by the dqpc detector ( fig .
the charge qubits are based on quantum dots ( qd ) , in which the position of the excess charge in a qubit affects the qpc current electrically , resulting in detection of charged state . owing to this additional external degrees of freedom ,
this setup is considered to be a harsher and more realistic condition for qubits than that of ref .@xcite . here
, we calculate a time - dependent fidelity of four charge qubits with dqpc by solving density matrix ( dm ) equations .
this article is organized as follows . in 2 , formulation of our model is presented .
section 3 is devoted to the numerical calculations for the cluster state and df state .
the conclusion is given in 4 .
the detailed coefficients used in the equations in the text are shown in appendix .
, width=188 ]
here we show the formulation of the density matrix equations for fig .
the hamiltonian for the combined qubits and the qpc for fig .
[ qpc ] is written as @xmath6 .
@xmath7 is a hamiltonian for an array of charge qubits with nearest - neighbor interactions , expressed by @xcite @xmath8 where @xmath9 and @xmath10 are pauli matrices for the @xmath4-th qubit expressed by @xmath11 and @xmath12 .
@xmath13 and @xmath14 refer to the two single - qubit states in which the excess charge is localized in the upper and lower dot , respectively .
@xmath15 is the inter - qd tunnel coupling between two coupled qds .
@xmath16 is the charging energy and corresponds to the energy difference between @xmath2 and @xmath3 controlled by gate voltage .
the coupling constants @xmath17 are derived from the capacitance couplings @xcite .
the hamiltonian for the two serially coupled qpcs @xmath18 is described by@xmath19 \nonumber \\ & + & \sum_{s\!=\!\uparrow,\downarrow } e_d d_{s}^\dagger d_{s } + u d_{\uparrow}^\dagger d_{\uparrow } d_{\downarrow}^\dagger d_{\downarrow}\ , .
\label{eqn : h_qpc}\end{aligned}\ ] ] here , @xmath20
( @xmath21 ) is the annihilation operator of an electron in the @xmath22th ( @xmath23th ) level @xmath24 $ ] of the left ( right ) electrode , @xmath25 is the electron annihilation operator of the island between the qpcs , @xmath26 ( @xmath27 ) is the energy level of electrons in the left ( right ) electrode , and @xmath28 is that of the island . here , we assume only one energy level on the island between the two qpcs , with spin degeneracy .
@xmath29 ( @xmath30 ) is the tunneling strength of electrons between the left ( right ) electrode state @xmath31 ( @xmath32 ) and the island state .
@xmath33 is the on - site coulomb energy of double occupancy in the island .
@xmath34 is the capacitive interaction between the qubits and the qpc , which induces _ dephasing _ between different eigenstates of @xmath10 @xcite .
most importantly , it takes into account the fact that localized charge near the qpc increases the energy of the system electrostatically , thus affecting the tunnel coupling between the left and right electrodes : @xmath35 ( c_{i_l s}^\dagger d_s + d_s^\dagger c_{i_l s } ) \nonumber \\ & + & \sum_{i_r , s } \left [ \sum_{i=3}^4 \delta v_{i_r , is}\sigma_{iz } \right ] ( c_{i_r s}^\dagger d_s + d_s^\dagger c_{i_r s } ) \end{aligned}\ ] ] where @xmath36 ( @xmath37 ) is an effective change of the tunneling strength between the electrodes and qpc island .
hereafter , we neglect the spin dependence of @xmath38 and @xmath39 .
we assume that the tunneling strength of electrons weakly depends on the energy @xmath40 and electrodes are degenerate up to the fermi surface @xmath41 .
then qubit states influence the qpc tunneling rate @xmath42 and @xmath43 by @xmath44 and @xmath45 through @xmath46 and @xmath47 ( @xmath48 and @xmath49 is the density of states of the electrodes ( @xmath50 ) and @xmath51 is a tunneling rate when the island lies in the c " state ) .
the values of @xmath52s are determined by the geometrical structure of the system and depend on the distance between qubits and the qpc as @xmath53 with the measurement strength @xmath54
@xcite .
the dm equations of four qubits and the dqpc detector at zero temperature of fig .
[ qpc ] are derived similarly to ref .
@xcite by @xmath55 -\![\gamma_l^{z_1}\!+\!\gamma_l^{z_2 } ] ) \rho_{z_1z_2}^{a } \nonumber \\ \!&\!-\!&\ !
i\!\sum_{j=1}^{n}\omega_j ( \rho_{g_j(z_1),z_2}^{a}\!-\!\rho_{z_1,g_j(z_2)}^{a } ) \!+\ ! \sqrt{\gamma_r^{z_1}\gamma_r^{z_2 } } ( \rho_{z_1z_2}^{b\uparrow}+\rho_{z_1z_2}^{b\downarrow } ) , \nonumber \\ % \\ % % % ( b)%%% ( b ) & & \frac{d \rho_{z_1z_2}^{b_s}}{dt}\!=\!\left(\!i[j_{z_2}\!-\!j_{z_1}]\ ! -\!\frac{\gamma_l^{z_1 ' } \!+\!\gamma_l^{z_2 ' } \!+\!\gamma_r^{z_1}\!+\!\gamma_r^{z_2}}{2}\right ) \rho_{z_1z_2}^{b_s } \nonumber \\ \!&\!-\!&\ ! i\!\sum_{j=1}^n \omega_j ( \rho_{\!g_j(z_1),z_2}^{b_s}\!\!-\!\!\rho_{\!z_1,g_j(z_2)}^{b_s } ) \!+\ ! \ ! \sqrt{\gamma_l^{z_1}\gamma_l^{z_2 } } \!\rho_{z_1\!z_2}^{a}\ !
! \rho_{z_1\!z_2}^{c } , \nonumber \\ % \\ % % % ( c)%%% c & & \frac{d \rho_{z_1z_2}^{c}}{dt } \!=\!(\!i[j_{z_2}\!-\!j_{z_1 } ] \!-\![\gamma_r^{z_1'}+\gamma_r^{z_2 ' } ] ) \rho_{z_1z_2}^{c } \nonumber \\ \!&\!-\!&\ !
i\!\sum_{j=1}^n \omega_j ( \rho_{g_j(z_1),z_2}^{c}\!-\!\rho_{z_1,g_j ( z_2)}^{c } ) \!+\ !
\sqrt{\gamma_l^{z_1'}\gamma_l^{z_2 ' } } ( \rho_{z_1z_2}^{b\uparrow}+\rho_{z_1z_2}^{b\downarrow } ) , \nonumber\\ \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\ !
\label{eqn : dm}\end{aligned}\ ] ] where @xmath56 indicate qubit states such as @xmath57 and , @xmath58 , @xmath59 and @xmath60 are density matrix elements when no electron ( a " ) , one electron ( b " ) and two electrons ( c " ) exist in the qpc island , respectively .
@xmath61 , @xmath62 , ... , @xmath63 .
@xmath64 is introduced for the sake of notational convenience and determined by the relative positions between qubit states ( see appendix and ref .
@xcite ) . in the present case , there are 768 equations to be solved . time - dependent _ fidelity _ @xmath65 $ ] is calculated by tracing over the elements of the reduced dm obtained from eq .
( [ eqn : dm ] ) .
we use these dm equations to describe four qubits , and compare fidelity of a cluster state @xmath66
in one - way quantum computing , quantum computation is carried out by measuring a series of qubits in cluster states .
cluster states are directly generated if we can prepare an ising - like hamiltonian @xmath67 , starting from the initial state @xmath68 .
a unitary evolution @xmath69 ( we use @xmath70 ) at @xmath71 with an integer @xmath72 transforms @xmath68 into cluster states . in the case of charge qubit in eq .
( [ hcq ] ) , the @xmath9 term needs to be switched off during the creation of the cluster state and then switched on when measurements are carried out . to realize this controllability
, additional gates are required .
however , these additional gates make the qubit system complicated , ending by producing decoherence and crosstalk between qubits themselves and between qubits and the environment .
in addition , for some qubit systems , once the qubit array is made , @xmath15 and @xmath17 are fixed , and only @xmath16 is controllable via the gate voltage bias ( we call these `` simple - design qubits '' ) . in ref .
@xcite , we proposed how to effectively generate cluster states in such simple - design qubits ( one - step generation method " ) . in this method ,
cluster states are obtained by applying a gate bias voltage @xmath16 for the @xmath4-th qubit , expressed as @xmath73 at time @xmath74 . here , @xmath75 with @xmath76 , and we also need a relation @xmath77 with an arbitrary integer @xmath78 and the number of the nearest qubits @xmath79 . as long as these equations are held , we can choose any integers @xmath78 and @xmath79 . in the present one - dimensional qubit array , we have @xmath80 and @xmath81 . here
, we calculate two examples : case @xmath82 is @xmath83 ( @xmath84 ) and case @xmath85 is @xmath86 ( @xmath84 ) . in case @xmath82 , we have @xmath87 , @xmath88 . in case
@xmath85 we have @xmath89 , @xmath90 .
, in case @xmath82 .
@xmath91 , @xmath92 .
the `` weak '' indicates a weak measurement case of @xmath93 and the `` strong '' indicates a strong measurement case of @xmath94 .
, width=207 ]
figure [ fstruct ] shows a time - dependent fidelity when the measurement strength is changed . it can be seen that when the measurement strength is sufficiently large ( @xmath94 ) , the fidelity of cluster states greatly degrades .
this result shows that the cluster state is sensitive to the existence of trap sites .
figures [ fid ] ( a ) and [ fid ] ( b ) show the combined effects of the island structure and the nonuniformities among qubits when parameters @xmath15 , @xmath16 and @xmath95 fluctuate by 10% .
the smaller sizes of qds are preferable for qubits based on qds@xcite .
if we assume that a diameter of each qd is 5 nm@xcite , 10% size fluctuation corresponds to @xmath96 size deviation from 5 nm . in both the strong measurement case [ ( a ) @xmath94 ] and the weak measurement case [ ( b ) @xmath93 ] ,
the difference between the cluster state and the df state is small .
this is in contrast to the case without local island structure in qpc discussed in ref .
@xcite , and shows that the island structure ( trap site ) imposes a larger decoherence environment and would be a major origin of degradation for the entangled states . in order to examine the effect of nonuniformity between qubits in more detail , we calculate a fidelity at time @xmath99 as a function of nonuniformity of qubit parameter @xmath100 ( fig . [ eta ] ) .
we can see that the difference is slight for various types of nonuniformities when @xmath100 is sufficiently small .
we can also see that the fidelity of the larger gate bias case @xmath82 is degraded slightly more than that of the smaller gate bias case @xmath85 .
note that this difference is much smaller in cluster states than in df states , which is a preferable result for one - way computing in solid - state qubits . from figs .
[ fid ] and [ eta ] , we can confirm that the cluster state is robust to nonuniformity , whereas the cluster state is weak against local electronic state ( trap state ) , which is similar to the df state . of a cluster state and a df state when there are nonuniformities in qubit parameter .
@xmath86 and @xmath101 ( case @xmath82 ) .
@xmath91 , @xmath92 , and @xmath94 .
( a ) strong measurement case .
( b ) weak measurement case .
nonuniformity in the qubit parameters is introduced as @xmath102 , @xmath103 , and @xmath104 , with @xmath4 indicating the @xmath4-th qubit . here
@xmath105 for all qubits besides @xmath106 ( i ) , @xmath107 ( ii ) , and @xmath106 ( iii ) .
the fidelities of @xmath108 for ( i ) and ( ii ) mostly overlap . , width=321 ] as a function of nonuniformity @xmath100 .
parameters are the same as those in fig .
[ fid ] . ,
we investigated the effect of local electronic fluctuations ( trap state ) in addition to nonuniformities , as a model of a decoherence mechanism in cluster states .
we found that the island ( trap site ) affects the fidelity of cluster states significantly . because the local electronic state provides _ dynamical _ fluctuation whereas the nonuniformities provide _ static _
fluctuation , we can say that the cluster states are sensitive to _ dynamical _ fluctuations .
experimentally , it is difficult to reduce the number of trap sites .
thus , more detailed understanding of effects of trap states will be required in the future .
the author thanks a. nishiyama , j. koga , s. fujita , f. nori , y. x. liu , and x. hu for useful discussions .
here , we show all coefficients used in eq .
( [ eqn : dm ] ) . for legibility ,
we denote @xmath109 , @xmath110 , @xmath111 , and @xmath112 as @xmath113 , @xmath114 , @xmath115 , @xmath116 , respectively .
four - qubit states are written by @xmath117 , @xmath118 , .. , @xmath119 .
then , @xmath120 are expressed as @xmath121 @xmath122 are given by @xmath123 m. field , c. g. smith , m. pepper , d. a. ritchie , j. e. f. frost , g. a. c. jones , and d. g. hasko : phys .
* 70 * ( 1993 ) 1311 .
y. p. li , d. c. tsui , j. j. heremans , j. a. simmons , and g. w. weimann : appl .
* 57 * ( 1990 ) 774 .
t. sakamoto and k. nakamura : appl .
* 68 * ( 1996 ) 2861 . | we theoretically investigate the robustness of cluster states in charge qubit system based on quantum dot ( qd ) and double quantum point contact ( dqpc ) .
trap state is modeled by an island structure in dqpc and represents a dynamical fluctuation .
we found that the dynamical fluctuations affect the cluster states more than static fluctuation caused by qd size fluctuation . |
heavy meson hadronic matrix elements play an essential role in determining many of the parameters of the standard model , such as the ckm matrix elements .
theoretically , lattice gauge theory provides a means of computing these hadronic matrix elements with better control of systematic errors @xcite .
the heavy quark methods that have been successfully used on the lattice are : sheikholeslami - wohlert ( sw ) action , nonrelativistic qcd , heavy quark effective theory ( hqet ) and fermilab action . here
we focus on the fermilab action because it may best control the systematic errors associated with the large quark mass .
renormalization of lattice operators is necessary in order to obtain physical results from numerical simulations .
non - perturbative renormalization ( npr ) methods are attractive since they avoid possibly large errors associated with lattice perturbation theory .
this is especially important when domain wall fermions are used for the light quarks . among various npr methods ,
we decide to focus on ri / mom scheme npr @xcite , because of the promising results @xcite reported by the rbc collaboration .
we will use npr to represent the ri / mom scheme npr method in the rest of this paper .
our goal is to adopt the ri / mom scheme and to extend its application to heavy quarks .
in particular we will show how to generalize the usual npr `` window restriction '' : @xmath0 to the kinematic region required by the fermilab approach to heavy quarks .
we need to find a new renormalization region , which meets three requirements : i ) it permits evaluation in continuum perturbation theory .
thus , it should involve non - exception external momenta , giving no infrared singularities as light quark and gluon masses are set to zero .
ii ) the fermilab method remains valid , i.e. the region is not too far from the heavy quark mass pole .
iii ) it can be studied using a euclidean , lattice green s function .
let s take a look at a typical feynman diagram with a one loop qcd correction integral , for example , that shown in figure [ fig : feynman ] , with the integral @xmath1 here q is the internal gluon momentum and @xmath2 is the net external quark momentum flowing into @xmath3 internal quark line within the loop .
-0.2 in -7 mm the denominator of the @xmath3 quark line , with minkowski external momenta but euclidean loop momentum is @xmath4 .
since @xmath5 may be small , it does nt guarantee that we can avoid the pole on the @xmath6 axis during integration .
therefore , additional conditions have to be set to avoid singularities and permit the use of continuum perturbation theory .
specifically , we require that the absolute value of each denominator be much larger than @xmath7 .
let s first check our formalism with a massless quark .
we find that @xmath8 , which is consistent with the the original npr method for light quarks .
how about heavy quarks ?
we propose : @xmath9 . in order to keep fermilab discretization errors under control
, we have to constrain the above quantity to be much smaller than @xmath10 as well so that the heavy quark lines remain nearly on - shell .
therefore , our `` window '' of renormalization is @xmath11 .
-7 mm figure [ fig : q_sq_plane ] is a plot in complex @xmath12 plane .
our choice of renormalization points are located within the small region circled .
the distance from this circled region to the mass pole has to be much smaller than @xmath13 and much larger @xmath7 for perturbation theory to be valid .
this is what i mean by `` slightly off - shell condition '' in the rest of the paper .
note that the region we choose is time - like , while the usual npr s choice is space - like . why do we pick a time - like point , instead of a space - like point ?
the constraint comes from the fermilab action , which is an on - shell o(a ) improvement .
if we choose a deep , space - like point , the well - behaved properties of fermilab action would no longer be valid .
recall `` heavy quark '' means the product of @xmath14 and lattice spacing @xmath15 may be larger than or equal to 1 . under this circumstance , it s natural to expect an improved action to break the time - space symmetry .
here is the original fermilab action @xcite : @xmath16 \nonumber\\ & -&\hskip -0.15 in \kappa_s \sum_{n , i } [ \bar{\psi}_n ( r_s-\gamma_i ) \textsl{u}_{n , i } + \bar{\psi}_{n+\hat{i } } ( r_s+\gamma_i ) \textsl{u}_{n , i}^\dag \psi_n ] \nonumber \\ & + & \hskip -0.15 in \frac{i}{2 } c_{b } \kappa_{s } \sum_{n;i , j , k } \varepsilon_{ijk } \bar{\psi}_{n } \sigma_{ij } b_{n;k } \psi_{n } \nonumber \\ & + & \hskip -0.15 in ic_{e } \kappa_{s } \sum_{n;i } \bar{\psi}_{n } \sigma_{0i } e_{n;i } \psi_{n } % \end{equation}\end{aligned}\ ] ] note that those coefficients are the functions of @xmath17 , and are expected to remain finite for both small and large @xmath17 .
what happens when we go slightly off - shell ?
if the external quark masses differ from the physical ones by a small amount @xmath18 , with @xmath19 then the resulting errors will be no larger than the other discretization errors in the fermilab approach .
however , when we go off - shell , we have to include more terms to make the o(a ) improvement complete .
because contact terms and new non - gauge - invariant operators can appear . ) first , we have to add o(a ) off - shell improvement terms in the action .
fortunately , those terms which need to be added in the action can be compensated by improving the quark fields @xcite .
therefore , the action itself remains the same , including the well - behaved coefficients .
the improved quark fields will look similar to those in ref @xcite , but with broken space - time symmetry .
likewise for the composite operators .
the broken symmetry gives us at least double the number coefficients to be determined non - perturbatively , compared with the light quark cases .
we are still exploring how these coefficients may be determined in a more efficient way .
to respect the `` slightly off - shell condition '' , we need to change the time - component fourier transformation to laplace transformation .
take the operator @xmath20 as an example , for a general dirac matrix @xmath21 . the non - amputated green s function in momentum space
is : @xmath22 where @xmath23 .
we calculate the amputated green s function and apply the projector onto @xmath24 to define : @xmath25,\end{aligned}\ ] ] where the s(pa ) is the gauge - averaged propagator . following the usual npr procedure ,
we apply the renormalization condition : @xmath26 where @xmath27 represents tree - level value .
thus , imposing @xmath28 will determine the renormalization factor @xmath29 .
our lattice npr calculations are done in the ri renormalization scheme .
we need continuum ri renormalization calculations to match to physical values given in the continuum @xmath30 scheme .
the previous matching calculations done at @xmath31 , while taking @xmath32 must be extended to our new scheme .
let s take a look at the simplest example : the calculation of @xmath33 and @xmath34 .
our new renormalization conditions become : @xmath35_{m^2-p^2 = \mu^2 } = 1\end{aligned}\ ] ] @xmath36_{m^2-p^2= \mu^2 } = 1\end{aligned}\ ] ] where @xmath37 $ ] .
the calculations are done in the dimensional regularization scheme , where d is 4 - 2@xmath38 . here
are the matching factors for @xmath34 and @xmath33 in the ri and @xmath30 schemes : @xmath39ln \frac{m^2}{\mu^2}\nonumber \\ & + & \hskip -0.15 in ln\frac{\mu'\;^2}{\mu^2 } ( 3 + 2\xi ) \ } + o(\alpha_s^2)\end{aligned}\ ] ] @xmath40 where @xmath41 is defined as @xmath42 .
in this paper , we present a first npr renormalization proposal for relativistic heavy quarks . we propose a new renormalization point : @xmath11 , in a slightly off - shell region .
we must add off - shell improvement terms and determine the corresponding coefficients as well as evaluating laplace transformations in the time . | we discuss the application of the regularization independent ( ri ) scheme of rome / southampton to determine the normalization of heavy quark operators non - perturbatively using the fermilab action . |
experimental information on 3-dimensional imaging " of hadrons , encoded in unintegrated , transverse momentum dependent ( tmd ) parton density and parton decay functions , comes at present from two main sets of experimental data : deep inelastic scattering ( dis ) at high energy , and low-@xmath0 drell - yan ( dy ) and semi - inclusive dis ( polarized and unpolarized ) . in each of these two cases ,
qcd factorization theorems allow one to relate physical , observable cross sections to tmd parton distributions via perturbatively calculable kernels .
these theorems provide the theoretical basis for determining tmd distributions from experimental measurements .
they are also essential to formulate and apply methods of perturbative resummation at all orders in the qcd coupling to a large variety of observables in high - energy hadronic collisions .
examples include processes both at the large hadron collider ( lhc ) and at fixed - target experiments .
this article is based on workshops devoted to these topics held at the university of antwerp in 2014 and provides a concise status report of this field .
the purpose of this article is to give a non - technical introduction to the motivations for experimental and theoretical studies of tmds ; to illustrate this with specific examples of application of tmds to topical issues in high - energy physics phenomenology ; to point to future directions of development .
in particular , we examine implications of two sets of qcd factorization theorems based on tmd parton distribution functions : low-@xmath0 factorization for heavy particle spectra ( including vector bosons , higgs bosons , heavy flavors ) and high - energy factorization .
we focus on production processes in hadronic collisions in two limits : i ) @xmath1 for fixed invariant mass , and ii ) @xmath2 for fixed momentum transfer .
we illustrate this with examples on transverse momentum spectra and angular correlations at the lhc for drell - yan and higgs boson production and associated multi - jets .
we survey computational tools which are being developed to treat the physics of tmds .
we present in particular an application of a new tool , , to tmd parton densities based on fits and parameterizations including qcd evolution .
the paper is organized as follows . in sec .
ii we motivate the use of tmds . in sec .
iii we discuss their role in the physics of large transverse momenta .
iv summarizes experimental prospects and theory developments .
v and vi illustrate the status of fits and parameterizations for tmd parton distributions and of tmd monte carlo tools .
final remarks are given in sec .
transverse momentum dependent parton distributions encode nonperturbative information on hadron structure , including transverse momentum and polarization degrees of freedom , which is essential in the context of qcd factorization theorems for multi - scale , non - inclusive collider observables .
a classic example is given by drell - yan hadroproduction of electroweak gauge bosons .
[ fig : reffigz ] @xcite shows the differential cross section for @xmath3-boson production in @xmath4 collision at the lhc as a function of the @xmath3-boson transverse momentum @xmath0 , in the lepton pair s invariant mass range 60 gev @xmath5 120 gev . in the spectrum of fig .
[ fig : reffigz ] we distinguish the high-@xmath0 region , the peak region , and the low-@xmath0 region . -boson transverse momentum @xmath0 spectrum in @xmath4 collisions at the lhc @xcite . _ _ ] in the high-@xmath0 region
the cross section is expected to be well represented by an evaluation of the partonic @xmath3-boson cross section to finite order in qcd perturbation theory ( leading - order ( lo ) , next - to - leading - order ( nlo ) , and so forth ) , combined with factorization in terms of ordinary ( collinear ) parton distribution functions ( pdfs ) . on the other hand , if this theoretical framework is applied to the region of decreasing @xmath0 it will not be able to describe the approach to the peak region in fig .
[ fig : reffigz ] ( @xmath6 gev@xmath7 ) nor the turn - over region ( @xmath8 gev@xmath7 ) .
rather , the cross section predicted from any finite order of perturbation theory , convoluted with ordinary parton distributions , will diverge as @xmath0 decreases .
the reason for this is that the physical behavior of the @xmath3-boson spectrum near the peak region and below @xcite is controlled by multi - parton qcd radiation , which is not well approximated by truncating the qcd perturbation series to any fixed order but rather requires methods to resum arbitrarily many parton emissions , viz . , scattering amplitudes with an infinite number of real and virtual insertions of soft gluons .
this can be accomplished in a systematic manner via a generalized form of qcd factorization @xcite which now involves quark distribution functions that , unlike the ordinary ones , explicitly depend on transverse momentum and polarization ( tmd pdfs ) .
such tmd pdfs obey evolution equations @xcite which generalize the ordinary renormalization - group evolution equations of collinear pdfs .
these evolution equations , once combined with the tmd factorization of the physical cross section , allow one to resum logarithmically enhanced contributions in the ratio @xmath9 to the perturbation series expansions for the physical observables to all higher orders in the qcd coupling . it is only after this generalized factorization analysis going beyond the collinear factorization is carried through that the physical behavior of the @xmath3 boson spectrum observed in fig .
[ fig : reffigz ] can be predicted .
: gluon density at different mass scales @xmath10 @xcite.__,title="fig : " ] : gluon density at different mass scales @xmath10 @xcite.__,title="fig : " ] : gluon density at different mass scales @xmath10 @xcite.__,title="fig : " ] : gluon density at different mass scales @xmath10 @xcite.__,title="fig : " ] a second example concerns the rise of proton s structure functions at small longitudinal momentum fractions . since in @xmath4 collisions the product of initial - state longitudinal fractions scales like @xmath11 at fixed momentum transfer , where @xmath12 is the squared centre - of - mass energy , as we push forward the high - energy frontier more and more events at small longitudinal fractions contribute to processes probing short - distance physics .
many hard - production cross sections at the lhc receive sizeable contributions from proton s structure functions in this region .
as parton longitudinal momenta become small , the fraction of momentum carried by transverse degrees of freedom becomes increasingly important .
[ fig : reffigstrfun ] shows the proton s gluon density resulting from global fits @xcite to hadronic collision data , performed at lo , nlo , nnlo @xcite of perturbation theory , as a function of the longitudinal momentum fraction @xmath13 for different values of the evolution mass scale @xmath10 . in the low-@xmath13 regime the perturbative higher - order corrections to structure functions are large , and the gluon pdf uncertainty is large .
the strong corrections at low @xmath13 come from multiple radiation of gluons over long intervals in rapidity @xcite , in regions not ordered in the gluon transverse momenta @xmath14 , and are present beyond nnlo to all orders of perturbation theory @xcite .
the theoretical framework to resum these unordered multi - gluon emissions is a generalized form of qcd factorization @xcite in terms of tmd pdfs .
analogously to the drell - yan case discussed earlier , the tmd pdfs obey a suitable set of evolution equations @xcite , appropriate to this kinematic region .
these provide another generalization , valid in the high - energy limit , of the ordinary renormalization - group evolution .
the tmd factorization in this case allows one to resum logarithmically enhanced corrections in the ratio @xmath15 to all higher orders in the qcd coupling . besides the above examples of drell - yan and structure functions , tmd factorization theorems apply to a wide variety of processes at the lhc .
in particular , with extensive measurements of higgs boson production at the lhc run ii , a new set of qcd processes becomes available in which the higgs boson acts as a color - singlet , pointlike source ( in the heavy top limit ) which couples to gluons .
this is to be contrasted with drell - yan and deep - inelastic scattering cases , based on weak and electromagnetic currents providing color - singlet pointlike sources coupled to quarks .
this opens up the possibility of a new program of precision qcd measurements in gluon fusion at high mass scales in the lhc high - luminosity runs @xcite .
analogously to the case of vector bosons in the example of fig .
[ fig : reffigz ] , theoretical predictions for the higgs - boson production differential spectrum over the whole range in transverse momenta accessible at the lhc require generalized qcd factorization , based on initial - state gluon distributions that include polarization and transverse - momentum degrees of freedom . compared to the vector boson case , however , new features arise which are associated with the role of gluon polarizations in gluon - gluon scattering .
gev and longitudinal momentum fraction @xmath16 , and for different values of the nonperturbative parameters discussed in @xcite.the red and green bands around each curve correspond to variations by factor 2 of the resummation scale and rapidity scale in the calculation @xcite.__,title="fig : " ] gev and longitudinal momentum fraction @xmath16 , and for different values of the nonperturbative parameters discussed in @xcite.the red and green bands around each curve correspond to variations by factor 2 of the resummation scale and rapidity scale in the calculation @xcite.__,title="fig : " ] more precisely , in the high - energy limit @xmath17 the higgs boson production from gluon fusion is dominated by a single eikonal gluon polarization @xcite .
the contribution of this polarization depends on the gluon transverse momentum and can be rewritten in terms of the high - energy projection operator defined in @xcite . a complete set of operators for polarization dependent and transverse momentum dependent gluon distributions
is given in @xcite .
in the region of low higgs - boson transverse momenta , @xmath18 , the contributions of polarized gluons to the higgs spectrum have been studied both perturbatively @xcite and nonperturbatively @xcite .
an example is shown in fig .
[ fig : refh - to - f ] @xcite , where the unpolarized and linearly polarized gluon distributions contributing to the higgs boson spectrum at small @xmath0 are plotted as a function of transverse momentum .
the presence of polarized gluon components ( even in unpolarized beams ) characterizes gluon fusion processes and has no analogue in the drell - yan case .
in particular the component in the right hand side plot of fig .
[ fig : refh - to - f ] is a gluon tmd distribution with double spin flip ( see table ii ahead , top right corner ) . from the point of view of perturbative power counting ,
double spin flip effects start to contribute to the higgs @xmath0 spectrum at the nnlo ( but may contribute earlier in more complex , less inclusive observables associated with higgs production ) .
detailed measurements of higgs boson final states will allow the qcd dynamics of polarized gluons and their correlations to be explored experimentally for the first time . for both the drell - yan and higgs cases , in addition to the inclusive spectra an extensive experimental program at the lhc is devoted to the associated production of heavy bosons with jets .
the region in which the boson and leading jet are nearly back - to - back presents features comparable to the discussion given above for the low-@xmath0 part of the inclusive spectra .
for instance , a study of tmd gluon contributions to higgs + jet final states in which the imbalance between the boson and leading - jet transverse momenta is small is reported in fig .
[ fig : refhiggsasymm ] @xcite , showing the boson - jet pair s transverse momentum distribution and azimuthal asymmetries .
asymmetry ( middle ) and @xmath19 asymmetry ( right ) in higgs boson + jet production at small transverse momenta @xmath0 of the higgs + jet pair .
here @xmath20 represents the average of the higgs and jet transverse momenta , and the shaded blue areas represent the range of the asymmetries as @xmath20 varies from 0 to @xmath21 .
_ _ ] the case of associated boson + jet production when the imbalance between the boson and leading - jet transverse momenta is not small , on the other hand , probes the physics of final states with multiple jets .
the role of tmd parton distributions in scenarios with high jet multiplicity is discussed in the next section , and serves to illustrate the connection of tmds with the kinematic region of large transverse momenta .
an extension of the methods discussed above for drell - yan and higgs production applies to the transverse momentum spectra of heavy flavor pairs , e.g. top quarks . unlike the case of color - singlet currents coupled to quarks ( as in drell - yan production ) or gluons ( as in higgs boson production in the heavy top limit ) , heavy - quark pair production constitutes a composite non - pointlike probe , containing color - charged particles in the lowest - order final state and receiving contribution from both quark and gluon tmd channels .
color correlations over long timescales between initial and final states will break factorization in the region of very small transverse - momentum imbalance of the pair @xcite .
studies of this region and the interplay of perturbative and nonperturbative contributions will help understand quantitatively these effects . another area for applications of tmds concerns single spin asymmetries and azimuthal asymmetries in polarized collisions . a classic example is the sivers asymmetry @xcite . fig .
[ fig : sivers ] @xcite shows low - energy measurements @xcite of the sivers transverse single spin asymmetry along with results of the fit @xcite .
for hadron s transverse momenta sufficiently small compared to the virtuality scale @xmath22 of the deep inelastic ( or drell - yan ) process , spin asymmetries obey tmd factorization formulas of the same kind @xcite discussed above for the unpolarized case of low @xmath0 drell - yan .
a combined understanding of current high - energy unpolarized measurements and low - energy spin asymmetry measurements is important for the planning of future polarized collider @xcite and fixed - target @xcite experiments .
we conclude this section by presenting the full leading - twist set of polarization dependent and transverse momentum dependent parton densities in a spin-1/2 hadron .
these are shown in table i and table ii , for the quark @xcite and gluon @xcite cases respectively , including the distributions in unpolarized hadrons ( top rows ) , longitudinally polarized hadrons ( middle rows ) , transversely polarized hadrons ( bottom rows ) .
( see @xcite for slightly different classifications . )
gauge - invariant operator definitions may be given for each of the tmd distributions in terms of nonlocal operator combinations , in which appropriate wilson - line gauge links are associated with quark and gluon fields @xcite .
operator definitions are instrumental in analyzing both factorization and potential sources of factorization breakdown , and in setting up lattice calculations @xcite of parton distributions .
unlike the low-@xmath0 drell - yan factorization theorem @xcite and its extensions for gluon fusion processes , the high - energy factorization theorem @xcite is valid for arbitrarily large momentum transfer .
it is based on the high - energy expansion @xmath2 and can be applied in the ultraviolet region of high @xmath0 .
it allows one , for example , to obtain the structure of logarithmic scaling violations in dis at high energy ( see @xcite ) and to resum logarithmic corrections of higher order in @xmath23 to higgs and top quark production cross sections ( see @xcite ) . in this section
we apply this theorem to discuss the role of tmds in the region of perturbative transverse momenta , in particular in the high-@xmath0 part of the drell - yan spectrum in fig .
[ fig : reffigz ] .
the basic observation is that the lhc kinematics leads to copious production of final states in which a high-@xmath0 vector boson recoils against multiple hard jets .
@xmath24-boson + @xmath25 jets final states using tmd high - energy factorization @xcite .
the motivation for this is twofold : a ) kinematical : it has recently been pointed out @xcite that collinearity approximations , once combined with energy - momentum conservation constraints , give rise to longitudinal momentum shifts and sizeable showering corrections in the monte carlo algorithms used to simulate multi - jet final states at the lhc ; b ) dynamical : it has long been known @xcite that , when the picture of multi - jets from finite - order perturbative matrix elements matched with collinear parton showers is pushed to higher and higher energies , new effects arise in jet multiplicity distributions and angular correlations due to soft but finite - angle multigluon radiation .
both these kinematical and dynamical effects can be taken into account by a tmd treatment of qcd parton shower evolution @xcite . to achieve this , ref .
@xcite uses the exclusive formalism of ccfm evolution equations @xcite implemented in @xcite .
the tmd pdfs to which evolution is applied are determined from fits to the precision dis data @xcite . by evolving these tmd
pdfs up to the scale of @xmath24 + jets and coupling them with appropriate , perturbatively calculated high - energy matrix elements , one obtains predictions for @xmath24-boson + @xmath25 jets observables .
[ fig : ht ] shows the total transverse energy @xmath26 distribution in final states with @xmath27-boson + @xmath25 jets , with @xmath28 , at the lhc . for comparison the experimental measurements @xcite ( jet rapidity @xmath29 , jet transverse momentum @xmath30 gev )
are plotted .
the uncertainty bands on the theoretical predictions are described in @xcite , and largely reflect uncertainties on tmds determinations , estimated according to three different approaches corresponding to the three color bands .
the tmd high - energy factorization predicts azimuthal correlations in the @xmath24 + multi - jet final states .
[ fig : azim ] shows results for the azimuthal correlation between the two leading jets , along with the transverse momentum of the third jet .
current limitations of the approach described above and ongoing improvements are discussed in @xcite , and include in particular the treatment of tmd quark density distributions , and the accuracy of determinations of the gluon density distribution over the whole range of longitudinal momentum fractions @xmath13 relevant to the lhc kinematics .
the results in figs .
[ fig : ht ] and [ fig : azim ] are however encouraging , and sufficiently general , in the context of approaches that aim to go beyond fixed - order perturbation theory and appropriately take account of nonperturbative effects .
as tmds describe nonperturbative transverse momentum dynamics in the hadron , they may provide a suitable framework not only for the factorization of the hard process but also to incorporate effects from soft particle production and multi - parton interactions @xcite .
it is worth noting that while for sufficiently inclusive observables in @xmath24 + jets production calculations based on collinear parton showers matched with finite - order perturbative matrix elements describe measurements at run i very well , this may not necessarily be the case for observables sensitive to the detailed structure of multi - parton emission @xcite .
for example , fig .
[ fig : refatlasm12 ] @xcite shows atlas measurements of the di - jet invariant mass associated with @xmath24 production , compared with several monte carlo calculations .
the comparison with the results from the nlo - matched calculation blackhat + sherpa @xcite suggests that effects beyond nlo + collinear shower may set in for high invariant masses around and above 500 gev . in this region of masses a similar behavior is observed in the comparison of experimental measurements with the alpgen @xcite monte carlo calculation . in fig .
[ fig : refm12 ] we also plot the di - jet invariant mass distribution from the approach @xcite .
-boson + @xmath31 jets , compared with parton - shower monte carlo calculations . _
_ ] -boson + @xmath31 jets final states , computed from the approach @xcite .
the experimental data are from @xcite . _
_ ] for the physics program at run ii it is of much interest to examine the region of very large vector boson transverse momenta of order 1 tev and higher .
[ fig : refhighptz ] @xcite shows cms measurements of the @xmath3-boson @xmath14 in events with @xmath3 + 1 jet and @xmath3 + 2 jets at run i. at the highest @xmath14 one may see dynamics setting in beyond the level of madgraph @xcite and sherpa @xcite multi - leg jet calculations matched with collinear showers , even supplemented with an nnlo k - factor .
-boson transverse momentum measured @xcite in @xmath3 + 1 jet ( left ) and @xmath3 + 2 jets ( right ) events , compared with monte carlo calculations.__,title="fig : " ] -boson transverse momentum measured @xcite in @xmath3 + 1 jet ( left ) and @xmath3 + 2 jets ( right ) events , compared with monte carlo calculations.__,title="fig : " ]
this section gives a brief overview of ongoing theoretical developments and experimental prospects .
_ factorization and resummation for @xmath32_. the factorization @xcite for drell - yan production at low @xmath0 ( along with corresponding extensions to other processes , including semi - inclusive dis and higgs production ) has been reobtained in soft collinear effective theory ( scet ) by different approaches ( @xcite , @xcite , @xcite ) .
the treatment of nonperturbative contributions to the tmd evolution equations @xcite from the region of large transverse distances @xmath33 differs in each of these various approaches and in the classic studies @xcite , and is currently the subject of intense investigations .
such treatment is essential for predictions at @xmath34 gev but its influence may also extend to the peak region .
it is found to be important , and with distinctive features compared to the drell - yan case , in semi - inclusive dis @xcite .
see @xcite for recent discussions of nonperturbative contributions .
the region of small transverse distances @xmath33 , on the other hand , is investigated via perturbative resummations to next - to - next - to - leading accuracy @xcite and computations through two loops @xcite of the perturbative coefficient functions controlling the expansion of the tmds in terms of collinear pdfs .
all these aspects are relevant for the interpretation of the production spectra at low transverse momenta @xmath0 , both in high - energy drell - yan experiments at lhc and tevatron @xcite and in fixed - target experiments @xcite , including polarized drell - yan and semi - inclusive dis @xcite .
_ evolution of tmds and fits to physical cross sections_. the above approaches to low-@xmath0 spectra which make use of tmds currently employ , in practice , either approximate analytic ( or semi - analytic ) solutions of the evolution equations @xcite or perturbative expansions of the tmds in terms of collinear pdfs , or a combination of both .
a different proposal has been put forward in @xcite ( ) , based on global fits to experimental data to obtain tmd parton distributions at different evolution scales , and on using these to make predictions for physical quantities .
this is similar in spirit ( but different in its realization ) to what is done in the case of collinear parton distributions .
theoretical predictions for physical cross sections which obey tmd factorization formulas could then be obtained by applying these formulas , using perturbatively calculable coefficients and appropriately evolved tmds determined from fits to experiment . in this approach , unlike most current implementations of tmd formalisms , the nonperturbative dependence on longitudinal and transverse degrees of freedom is fully coupled , and can be entangled with the dependence on the evolution scale @xcite . for phenomenological applications this can be important when for instance comparing theory with experimental measurements over a wide range in @xmath13 and evolution scales .
_ nonlinear evolution of the gluon tmd and wilson line correlators_. the conventional gauge - invariant operator definition of the gluon tmd @xcite is distinct from the weiszacker - williams operator definition @xcite in terms of wilson lines often used at @xmath35 ( see also @xcite for discussion of the operator definitions ) .
correspondingly , these gluon tmds obey different rapidity evolution equations : in the moderate @xmath36 region one has linear double - logarithmic equations , while in the @xmath35 domain the non - linear single - logarithmic balitsky - kovchegov equation applies @xcite .
the relationship between these two regimes is examined in @xcite , where it is clarified that the non - linear small-@xmath13 evolution transforms into linear rapidity evolution for the conventional gluon tmd .
@xcite consider applications to diffraction , and refs .
@xcite to jets at large rapidity separations . also , the evaluation of the complex combinations of wilson lines entering the gluon tmd at small @xmath13 calls for the development of a dedicated methodology .
essential improvement in the understanding and computation of correlators with wilson lines can be achieved by the eikonal exponentiation methods @xcite , which enable the exact resummation of the diagrams presenting a given correlator as the exponent of series of the so - called web diagrams @xcite .
_ tmds and generalized loop space_. renormalization properties of wilson line correlators control the evolution of tmds @xcite .
in particular , the appearance of light - cone , or rapidity , divergences @xcite in higher - loop corrections to the gauge - invariant correlators calls for a treatment of overlapping divergences , which can be achieved by the introduction of a soft subtraction factor @xcite .
the evolution of the gauge - invariant path - dependent tmds with the light - like cusped wilson lines can also be associated with the geometric evolution in the generalized space @xcite .
the differential shape variations of the underlying contours to the wilson loops are formulated in terms of the frchet derivative @xcite and the equations of motion in the loop space are dual to the energy and rapidity evolution of the tmds having the same structure of the wilson lines @xcite
. _ non - universality and wilson lines .
_ operator definitions of parton distribution functions in terms of quark and gluon fields involve nonlocal operator combinations .
for collinear functions the nonlocality is along the lightcone , for tmds it is along the lightfront involving also transverse separations .
unavoidably , therefore , additional gluonic fields minimally enter in the wilson lines that are needed for an unambiguous gauge invariant description .
the fact that these wilson lines depend on the hard process brings in a calculable non - universality @xcite , which is a generalization of the sign flip between t - odd tmds in going from sidis to dy @xcite .
other examples where these effects appear are jet+jet or photon+jet final states in hadroproduction @xcite as compared to drell - yan or @xmath37 production , and higgs + jet final states as compared to higgs production into colorless final states .
color entanglement can lead to further sources of non - universality affecting both tmd factorization @xcite and collinear factorization @xcite .
a program is ongoing devoted to a careful analysis @xcite of the possible operators that contribute to particular tmd structures followed by the study of their evolution .
_ tmds from exclusive evolution equations_. the gluonic ccfm evolution equation @xcite is being extended along the lines proposed in @xcite to treat the coupled evolution of the flavor - singlet sea quark density and gluon density .
this is important for describing exclusive components of high - multiplicity final states .
in particular , the inclusion of the sea quark density at tmd level is one of the main elements needed to treat drell - yan production across the whole range of central and forward rapidities @xcite measured at the lhc @xcite .
this approach is also being extended to include nonlinear evolution and saturation effects @xcite and to incorporate methods for automated computation of off - shell high - energy matrix elements @xcite .
_ soft particle production and multi - parton interactions_. as tmds encode nonperturbative transverse momentum dynamics in the proton , one may ask whether they are relevant not only for factorization of hard processes but also for the understanding of soft particle production and , in particular , of the multi - parton interactions which are found to be needed at low to moderate transverse momenta for monte carlo simulations to describe experimental data on underlying events , particle multiplicities and spectra .
double parton interactions @xcite including parton s transverse momentum dependence are investigated in @xcite .
the role of parton s transverse momentum in the interpretation of energy flow measurements is discussed in @xcite .
implications for diffraction are considered in @xcite .
tmd effects in multi - parton correlations may be studied in upcoming measurements of charged particle multiplicities and spectra and underlying event at the lhc 13 tev run .
experimental prospects have been discussed for identifying tmd effects based on measurements of benchmark cross sections , both at the lhc and at lower energy experiments . _
drell - yan lepton pair production and drell - yan plus jets_.
as discussed in the previous two sections , both the low-@xmath0 part of the spectrum and the high-@xmath0 part can be sensitive to tmd effects .
multi - differential measurements are especially important as one can access azimuthal correlations in the lepton + jet final states @xcite which constitute distinctive tmd predictions .
comparison of @xmath38 + jet final states at small transverse momentum imbalance @xcite with di - boson @xmath39 final states may shed light on color flow patterns which are eventually responsible for factorization breaking phenomena in hard processes sensitive to very low transverse momentum scales .
_ higgs boson production and higgs boson plus jets_. similar measurements to the drell - yan case , including differential cross sections , are relevant for gluon tmds and qcd studies of polarized gluons and color correlations , once sufficient statistics is reached .
measurements of higgs versus drell - yan at the same invariant mass may be used to reduce the influence of pile - up in the high - luminosity lhc runs @xcite . the boson @xmath0 spectrum , final - state angular distributions and underlying event observables probe different aspects of the higgs coupling to gluons @xcite .
_ heavy flavor production_. measurements of top quark pair production spectra can provide comparable information to the previous two cases but with additional complexity due to the presence of color charges in the final state .
the associated initial - state / final - state color correlations at small @xmath0 could be studied to examine factorization - breaking contributions in the region of very small transverse momenta @xcite , provided sufficient resolution can be reached .
it will also be interesting to investigate kinematic effects of longitudinal momentum reshuffling in parton showers @xcite at top quark scales .
similar studies can be done at lower mass scales with bottom and charm quarks .
_ quarkonium production_. despite the complexity of the bound state , production of @xmath40 and @xmath41 quarkonia is a useful probe of tmd gluon effects at low mass scales .
phenomenological studies are carried out in @xcite .
many features of these processes have been investigated experimentally at the lhc run i @xcite .
measurements of the spectra and especially of the polarization for @xmath42 , @xmath43 and all quarkonium states at run ii will be particularly interesting for studying polarized gluon effects .
color - singlet transitions may have reduced sensitivity to factorization - breaking effects @xcite .
quarkonium measurements are further proposed at fixed - target experiments @xcite and electron - ion collider @xcite .
the polarization dependent and transverse momentum dependent proton s parton densities , in the notation of @xcite , are given in the tables i and ii in sec .
this scheme can be extended to spin-1 targets @xcite .
most existing fits and parameterizations of these distributions may be grouped into categories which broadly correspond to three main areas of tmds applications discussed in sec .
ii : * fits to vector boson @xmath0 experimental data in unpolarized drell - yan production @xcite based on the low-@xmath0 tmd factorization @xcite , in some cases including extension to semi - inclusive dis data @xcite . *
fits to dis structure function data @xcite based on the high - energy tmd factorization @xcite or on other approaches ( e.g. saturation formalism ) to high - energy dis , in some cases including the precision measurements @xcite and tmd pdf uncertainties @xcite .
* fits to spin and azimuthal asymmetries data from low - energy experiments either based on parton model @xcite or including qcd evolution @xcite . for precision phenomenology it will be essential that results of fits and parameterizations are given in a portable form as a determination of tmd pdfs over a given kinematic range , appropriate to the theoretical method and experimental data used
a first step in this direction has been taken in @xcite .
the main point is that if results of fits to experimental data are used to provide tmd pdfs at different evolution scales , theoretical predictions for physical cross sections could then be obtained by using these pdfs in factorization formulas ( or , eventually , in monte carlo event generators implementing these formulas ) . in
@xcite a library has been initiated , , in order to unify and simplify the access of tmds , along with a plotting tool , , for easier comparisons .
commonly used pdf sets are implemented in , with the goal to provide a library of all available tmds . in pdfs
are accessible in an easily callable way within the range of their applicability .
the pdfs currently included range from tmd gluon densities obtained from fits to small-@xmath13 dis data based on high - energy factorization , to tmd gluon densities from fits based on saturation approaches , to tmd quark densities from parton - model fits to low - energy fixed - target data at large @xmath13 and small @xmath44 .
tmd fragmentation functions are not yet implemented , but are foreseen for the future .
an example from is shown in fig .
[ fig : tmdplotter ] , plotting the transverse momentum dependence of valence quark distributions , at fixed values of @xmath13 and renormalization scale @xmath45 , obtained from the fits @xcite . in fig .
[ fig : tmdplotter - glue ] we show results for gluon distributions @xcite .
the nonperturbative parameters of the distribution represented by the red curve are obtained from the fit @xcite to dis experimental data @xcite while the distribution represented by the blue curve is not fitted but based on the model @xcite .
inclusive or semi - inclusive hard cross sections can be calculated by convoluting parton density and decay functions with partonic cross sections . for a detailed description of the exclusive structure of the final states , on the other hand ,
event generators including parton showers and full hadronization are required . in the collinear case ,
cross sections are computed with on - shell initial partons . for many processes ,
higher order calculations exist , and many of these are implemented in monte carlo ( mc ) simulation tools like powheg @xcite , mc@nlo @xcite , amc@nlo @xcite , which combine next - to - leading order partonic calculations with parton showers and hadronization .
these simulations all need a reshuffling of kinematic variables , after the parton shower is generated , in order to satisfy energy - momentum conservation , which can lead to significant kinematic shifts in the longitudinal momentum fraction @xmath13 @xcite .
this is because transverse momentum is generated by the initial - state parton shower , which is not available when the hard scattering is computed . in certain phase
space regions , these longitudinal shifts can affect the accuracy of the calculations significantly . using tmds , this kinematic reshuffling can be avoided from the beginning provided the tmds include transverse momenta generated by perturbative qcd evolution , which in turn can be evaluated according to different approximation schemes such as those in , , @xcite .
if a monte carlo method is used to solve the tmd evolution equation , a further advantage is that the solution of the evolution equation can be directly matched to the simulation of parton showers : the kinematic distributions are the same , whether they come from a solution of the evolution equation or from a simulation of the parton shower @xcite .
while a general purpose monte carlo at tmd level does not yet exist , examples of such algorithms @xcite have been presented for specific cases .
we list a few examples below .
* mc event generators with parton shower and hadronization + @xmath46 @xcite is a full hadron level monte carlo event generator using tmds , originally developed for small @xmath13 processes in @xmath47 , now extended to cover medium and large @xmath13 and @xmath4 processes .
initial state parton showers are treated according to the ccfm formalism , final state parton shower and hadronization is performed by the lund package @xcite .
parton polarizations are included according to the high - energy factorization @xcite .
proton polarizations are not yet included .
+ @xmath46 @xcite . with
the initial and final state parton showers simulated in , one may argue that several elements of tmd physics are effectively included . can be used to mimic spin - dependent cross sections by reshuffling events ( assigning polarization states ) @xcite according to a given cross - section model .
this is especially useful when event topologies are needed ( e.g. , to simulate the interplay of track correlations with detector performance ) , or where no explicit physics model is yet available to be employed in dedicated mc generators .
+ @xmath46 mpythia and mlepto are based on lepto @xcite and @xcite with a modification of the hard process @xcite to treat the azimuthal angle of the scattered ( light ) quark and via momentum conservation of the target remnant according to parameterizations of the sivers function .
while limited to the rather specific case of the sivers effect it can make use of the hadronization embodied in jetset @xcite .
* mc event generators at parton level with fragmentation functions + @xmath46 lxjet ( see @xcite ) is devoted to a calculation of jet cross sections at small @xmath13 in hadron - hadron collisions . it can be also viewed as an event generator as it allows one to generate unweighted events .
it uses high - energy factorization @xcite .
gmc - trans ( see @xcite ) is a mc generator , developed by the hermes collaboration , applying the parton - model expression of the one - hadron semi - inclusive dis cross section using several models / parametrization for various leading - twist tmd pdfs and ffs .
pion and charged - kaon production is simulated , both for proton and neutron targets ( or combinations thereof ) without including nuclear effects . an analytic expression for the semi - inclusive dis cross section
was implemented based on the widely used gaussian ansatz of the transverse - momentum dependences .
@xmath46 tmdgen ( see @xcite ) is an extended version of gmc - trans entirely written in c++ focusing mainly on di - hadron production in semi - inclusive dis .
advances in computation power allowed for other than the gaussian ansatz of the transverse - momentum dependences by employing numeric integration algorithms .
it thus allowed the usage of the spectator model @xcite for various tmd pdfs and ffs .
+ @xmath46 clas ( see @xcite ) uses a similar approach as gmc - trans , though restricted to the unpolarized sector and to longitudinal double - spin asymmetries .
it uses the fully differential single - hadron dis cross section to simulate semi - inclusive dis events
. the transverse momentum dependence can be gaussian , but also light - cone quark - model inspired dependence has been implemented . * semi - analytical calculations of semi - inclusive processes + @xmath46 resbos
@xcite is a package to calculate analytically resummed distributions of inclusive and semi - inclusive observables .
the @xmath0 resummation in resbos and parton showering methods of monte carlo event generators are complementary .
both are based on all - order resummation using sudakov form factors .
resbos allows the user to calculate resummed distributions of the higgs / vector bosons and their decay products up to nnll .
it follows a prescription for matching the resummed contribution onto the fixed - order result and implements a parameterization of non - perturbative effects at small @xmath0 in terms of tmd pdfs .
+ @xmath46 hqt and dyqt @xcite are numerical programs which implement the analytical @xmath0 resummation formalism @xcite to compute , respectively , the @xmath0 spectrum of the standard model higgs and drell - yan lepton pair ( via vector boson production ) in hadronic collisions .
the resummed results are matched to the fixed order calculation valid at high @xmath0 .
the program can be used up to nnll+nlo , with the resummed part evaluated at nnll , the fixed order evaluated at nlo ( higgs / vector bosons plus one or two partons ) and with the normalization fixed to the total nnlo cross section .
+ @xmath46 hres and dyres @xcite are numerical programs which extend the calculations in hqt / dyqt by retaining the full kinematics of the higgs / vector bosons and of its decay products .
the programs implement @xmath0 resummation up to nnll+nnlo and allow the user to apply arbitrary cuts on final states and to plot the corresponding distributions in form of bin histograms .
we studied two sets of examples of multi - scale problems in hadronic collisions which require qcd factorization theorems beyond the collinear approximation and call for the use of tmd parton distributions . in one
set of examples , the transverse momentum scale is small compared to the hard process scale ; in the other , the transverse momentum is of the order of the hard scale but this is much smaller than the total energy of the scattering . in both cases , factorization theorems in terms of tmd parton distributions
are necessary in order both to resum logarithmically - enhanced perturbative corrections to all loops and to properly take into account nonperturbative hadron structure effects .
these multi - scale regimes are relevant to lhc phenomenology .
an example is the low-@xmath0 region of transverse momentum spectra for vector bosons , higgs bosons , heavy flavor pairs at the lhc .
another example is the production of multi - jets associated with heavy bosons and heavy flavors at large jet masses .
further examples include any final state produced by events at small longitudinal momentum fraction @xmath13 , such as final states boosted to high rapidities . besides the lhc , tmd dynamics is central to spin physics in current low - energy experiments and to the planning of future polarized collider and fixed - target experiments .
as the field moves towards the stage of precision studies , appropriate phenomenological tools will be needed .
this includes tools for monte carlo event simulations , which require parton shower evolution algorithms and determinations of tmd parton distributions from experimental data .
first steps toward a new program of portable and accessible tmd pdfs were illustrated with explicit examples in this report .
* acknowledgments*. we are grateful to desy , nikhef and antwerp university for financial support to the 2014 workshops on resummation , evolution , factorization " ( ref 2014 ) .
ioc acknowledges support from the belgian federal science policy office .
the work of mge and as is part of the program of the stichting voor fundamenteel onderzoek der materie ( fom ) , which is financially supported by the nederlandse organisatie voor wetenschappelijk onderzoek ( nwo ) .
the work of pjm is part of the eu ideas programme qwork ( contract no
. 320389 ) .
fh thanks the university of hamburg and desy for hospitality .
the work of fh is supported in part by the dfg sfb 676 programme particles , strings and the early universe . | we provide a concise overview on transverse momentum dependent ( tmd ) parton distribution functions , their application to topical issues in high - energy physics phenomenology , and their theoretical connections with qcd resummation , evolution and factorization theorems .
we illustrate the use of tmds via examples of multi - scale problems in hadronic collisions .
these include transverse momentum @xmath0 spectra of higgs and vector bosons for low @xmath0 , and azimuthal correlations in the production of multiple jets associated with heavy bosons at large jet masses .
we discuss computational tools for tmds , and present an application of a new tool , , to parton density fits and parameterizations .
desy 15 - 111 + nikhef 2015 - 023 + ral - p-2015 - 006 |
the experimental ability to create a superfluid alkali atom fermi gas in optical lattices@xcite opens a new opportunity to study strongly correlated quantum many - particle systems and to emulate high - temperature superconductors .
theoretical description of ultracold atomic bose and fermi gases in optical lattices has attracted much attention.@xcite near the feshbach resonance the atom - atom interaction can be manipulated in a controllable way because the scattering length @xmath0 can be changed from the bcs side ( negative values ) to the bec side ( positive values ) reaching very large values close to resonance . on bec side of the resonance the spin - up and spin - down atoms can form diatomic molecules , and these bosonic molecules can undergo a bec at low enough temperature.@xcite in what follow we focus our attention on the bcs side ( negative scattering length ) where bcs superfluidity is expected analogous to superconductivity . in particular , we consider an equal mixture of atomic fermi gas of two hyperfine states with contact interaction loaded into an optical lattice .
the two hyperfine states are described by pseudospins @xmath1 .
we also assume that the number of atoms in each hyperfine state per site ( the filling factor ) is smaller than unity , and that the lattice potential is sufficiently deep such that the tight - binding approximation is valid .
the system in this case is well described by the single - band attractive hubbard model : @xmath2 here , the fermi operator @xmath3 ( @xmath4 ) creates ( destroys ) a fermion on the lattice site @xmath5 with pseudospin @xmath1 and @xmath6 is the density operator on site @xmath5 with a position vector @xmath7 .
@xmath8 is the chemical potential , and the symbol @xmath9 means sum over nearest - neighbor sites .
@xmath10 is the tunneling strength of the atoms between nearest - neighbor sites , and @xmath11 is the on - site interaction . on the bcs side
the interaction parameter @xmath11 is positive ( the atomic interaction is attractive ) .
we assume @xmath12 and the lattice constant @xmath13 . in the case when the periodic array of microtraps is generated by counter propagating laser beams with different frequencies , the lattice is not stationary anymore .
it is expected that the formation of a bcs superfluidity is possible , but due to the presence of supercurrent ( or quasimomentum @xmath14 , where @xmath15 is the mass of the trapped atoms and @xmath16 is the lattice velocity ) the superflow can break down .
recently , the stability of superfluid fermi gases loaded into an optical lattice in the presence of supercurrent has been studied using the second - order time - dependent perturbation theory@xcite ( this method is a generalization of the earlier approach by belkhir and randeria@xcite ) , and the green s function formalism.@xcite in what follows , we derive the bethe - salpeter ( bs ) equations for the spectrum of the collective modes in the case of a moving optical lattice , and compare the bs formalism with the above - mentioned two methods .
this paper is organized as follows . in sec .
ii , we derive the bs equations for the dispersion of the collective excitations . in sec .
iii , the bs approach is compared with the results obtained by the perturbation theory , and by the green s function formalism .
we consider an imbalance mixture of @xmath17 atomic fermi gas of two hyperfine states @xmath18 with contact interaction loaded into a moving optical lattice .
the total number of @xmath17 atoms is @xmath19 , and they are distributed along @xmath20 sites .
the tight - binding form of the mean - field electron energy is @xmath21 . in the presence of supercurrent with velocity @xmath16
, the cooper s pair has quasimomentum @xmath22 .
the fourier transform of the single - particle green s function is a @xmath23 matrix @xmath24 . in the mean - field approximation
the corresponding matrix elements are as follows : @xmath25 .
\nonumber\end{split}\ ] ] the symbol @xmath26 denotes @xmath27 , @xmath28 , @xmath29 is the boltzmann constant , @xmath30 is the temperature . as can be seen , the one - particle excitations in mean - field approximation are coherent combinations of electronlike @xmath31 and holelike @xmath32 excitations .
the coherent factors @xmath33 and @xmath34 give the probability amplitudes of these states in the actual mixture . here ,
@xmath35 , @xmath36}$ ] , @xmath37}$ ] , and we have used the following notations @xmath38 $ ] , @xmath39 $ ] . the number and
the gap equations in this case are as follows ( @xmath40 is the filling factor ) : @xmath41 the bs equations for the collective mode @xmath42 and corresponding bs amplitudes @xmath43 in the general random phase approximation ( grpa ) are as follows : @xmath44 the first and second terms in the last equation represent the direct and exchange interactions : @xmath45 here , @xmath46 is a bose frequency , and we have introduced the two - particle propagator @xmath47 ( @xmath48 ) : @xmath49 as in the case of a stationary lattice,@xcite the bs equations can be reduced to a set of two equations . at zero temperature ,
the corresponding equations are as follows : @xmath50 g_\textbf{q}^{+}(\textbf{k},\textbf{q})= \frac{u}{2n}\sum_{\textbf{p } } \left[\gamma^\textbf{q}_{\textbf{k},\textbf{q}}\gamma^\textbf{q}_{\textbf{p},\textbf{q}}+ l^\textbf{q}_{\textbf{k},\textbf{q}}l^\textbf{q}_{\textbf{p},\textbf{q}}\right ] g_\textbf{q}^{+}(\textbf{p},\textbf{q})\\&-\frac{u}{2n}\sum_{\textbf{p } } \left[\gamma^\textbf{q}_{\textbf{k},\textbf{q}}\gamma^\textbf{q}_{\textbf{p},\textbf{q}}- l^\textbf{q}_{\textbf{k},\textbf{q}}l^\textbf{q}_{\textbf{p},\textbf{q}}\right ] g_\textbf{q}^{-}(\textbf{p},\textbf{q})-\frac{u}{2n}\sum_{\textbf{p } } \widetilde{\gamma}^\textbf{q}_{\textbf{k},\textbf{q } } \widetilde{\gamma}^\textbf{q}_{\textbf{p},\textbf{q } } \left(g_\textbf{q}^{+}(\textbf{p},\textbf{q } ) -g_\textbf{q}^{-}(\textbf{p},\textbf{q})\right)\\&+\frac{u}{2n}\sum_{\textbf{p } } m^\textbf{q}_{\textbf{k},\textbf{q}}m^\textbf{q}_{\textbf{p},\textbf{q } } \left[g_\textbf{q}^{+}(\textbf{p},\textbf{q})+g_\textbf{q}^{-}(\textbf{p},\textbf{q})\right ] , \label{neweq1sc } \end{split}\ ] ] @xmath51 g_\textbf{q}^{-}(\textbf{k},\textbf{q})= -\frac{u}{2n}\sum_{\textbf{p } } \left[\gamma^\textbf{q}_{\textbf{k},\textbf{q}}\gamma^\textbf{q}_{\textbf{p},\textbf{q}}+ l^\textbf{q}_{\textbf{k},\textbf{q}}l^\textbf{q}_{\textbf{p},\textbf{q}}\right ] g_\textbf{q}^{-}(\textbf{p},\textbf{q})\\&+\frac{u}{2n}\sum_{\textbf{p } } \left[\gamma^\textbf{q}_{\textbf{k},\textbf{q}}\gamma^\textbf{q}_{\textbf{p},\textbf{q}}- l^\textbf{q}_{\textbf{k},\textbf{q}}l^\textbf{q}_{\textbf{p},\textbf{q}}\right ] g_\textbf{q}^{+}(\textbf{p},\textbf{q } ) -\frac{u}{2n}\sum_{\textbf{p } } \widetilde{\gamma}^\textbf{q}_{\textbf{k},\textbf{q}}\widetilde{\gamma}^\textbf{q}_{\textbf{p},\textbf{q } } \left(g_\textbf{q}^{+}(\textbf{p},\textbf{q } ) -g_\textbf{q}^{-}(\textbf{p},\textbf{q})\right)\\&-\frac{u}{2n}\sum_{\textbf{p } } m^\textbf{q}_{\textbf{k},\textbf{q}}m^\textbf{q}_{\textbf{p},\textbf{q } } \left[g_\textbf{q}^{+}(\textbf{p},\textbf{q})+g_\textbf{q}^{-}(\textbf{p},\textbf{q})\right ] .
\label{neweq2sc } \end{split}\ ] ] here @xmath52 @xmath53\sin(q_{\nu}d),\ ] ] and the form factors are defined as follows : @xmath54 @xmath55 the bs equations ( [ neweq1sc])- ( [ neweq2sc ] ) lead to a set of four coupled linear homogeneous equations .
the existence of a non - trivial solution requires that the secular determinant @xmath56 is equal to zero .
here we have introduced symbols @xmath57 and @xmath58 : @xmath59 ^ 2-\varepsilon^2_\textbf{q}(\textbf{k},\textbf{q } ) } , \\ & j^\textbf{q}_{a , b}=\frac{1}{n}\sum_{\textbf{k}}\frac{a^\textbf{q}_{\mathbf{k}% , \mathbf{q}}b^\textbf{q}_{\mathbf{k}% , \mathbf{q}}\left[\omega+\omega_\textbf{q}(\textbf{k},\textbf{q})\right ] } { \left[\omega+\omega_\textbf{q}(\textbf{k},\textbf{q})\right]^2-\varepsilon^2_\textbf{q}(\textbf{k},\textbf{q})}. \label{jsc}\end{split}\ ] ] the quantities @xmath60 and @xmath61 could be one of the four form factors : @xmath62 or @xmath63 .
the secular determinant ( [ bsz ] ) is the main result in our bs approach .
this determinant provides the spectrum of the collective excitations ( in the particle and spin channels ) @xmath64 in a uniform manner . .
the filling factor , the gap and the chemical potential are @xmath65 , @xmath66 and @xmath67 , respectively .
the solid line is obtained by the bethe - salpeter method .
the dotted line represents the spectrum obtained by the green s function method ( see fig .
6a in ref .
[ ] ) , title="fig : " ] [ fig .
there exist another two methods that can be used to calculate the spectrum of the collective excitations of hamiltonian ( [ hubb1 ] ) .
decades ago , belkhir and randeria@xcite calculated the collective modes by linearizing the anderson- rickayzen equations . in the grpa
the anderson- rickayzen equations are reduced to a set of three coupled equations and the collective - mode spectrum is obtained by solving the secular equation @xmath68 , where @xmath69 the quantities @xmath70 and @xmath10 are defined by eq .
( [ jsc ] ) when @xmath71 .
the perturbation approach by ganesh et al .
provides the following @xmath72 secular determinant ( see eqs .
( b8 ) in ref .
[ ] , where there is a negative sign in front of the sum in the definition of @xmath73 ) : @xmath74 since @xmath75 , where @xmath76 one can say that the perturbation method is a generalization of the belkhir and randeria method to the case of moving optical lattice . and filling factor @xmath77 . the superfluid flows along @xmath78 direction , and @xmath79 .
( a ) stationary lattice ( the gap and the chemical potential are @xmath80 and @xmath81 ) .
( b ) moving lattice @xmath82 ( the gap and the chemical potential are @xmath80 and @xmath83 ) .
solid lines are obtained by the bethe - salpeter method while the dotted lines represent the spectrum obtained by the green s function method ( see fig .
11 in ref .
[ ] ) .,title="fig : " ] and filling factor @xmath77 .
the superfluid flows along @xmath78 direction , and @xmath79 .
( a ) stationary lattice ( the gap and the chemical potential are @xmath80 and @xmath81 ) .
( b ) moving lattice @xmath82 ( the gap and the chemical potential are @xmath80 and @xmath83 ) .
solid lines are obtained by the bethe - salpeter method while the dotted lines represent the spectrum obtained by the green s function method ( see fig .
11 in ref .
[ ] ) .,title="fig : " ] [ fig .
2 ] the secular determinant ( [ bsz ] ) can be rearranged as @xmath84 , where @xmath85 means the transpose matrix , and @xmath86 and @xmath87 are @xmath88 and @xmath89 blocks , respectively : @xmath90 thus , the bs secular equation assumes the form @xmath91 .
it should be mentioned that numerical calculations show that the contributions to @xmath92 due to the term @xmath93 are negligible , and therefore , there is an excellent agreement between the dispersions obtained by ganesh et al . in ref .
[ ] and by the bs secular equation not only in the case of a stationary lattice ( see the conclusion section in ref . [ ] ) , but in the case of a moving lattice as well .
the method used by yunomae et al.,@xcite known as the green s function approach , is a generalization of the ct and griffin@xcite approach to the collective excitations in s - wave layered superconductors . according to this approach
the spectrum of the collective excitations of an interacting electron gas ( the model includes a coulomb interaction @xmath94 and a short - range attractive interaction @xmath95 between electrons ) is obtained from the poles of the density and spin response functions in the grpa .
the response functions are obtained by using the kadanoff and baym formalism .
the kadanoff and baym formalism is equivalent to the bs formalism .
the bs equation for the two - particle green s function is @xmath96 , where @xmath97 is the free two - particle propagator , and the kernel @xmath98 is a sum of two parts .
@xmath99 takes into account the direct interaction between the electrons , while @xmath100 describes their exchange interaction .
ct and griffin have ignored the long - range coulomb interaction keeping in @xmath99 only the ladder diagrams involving the short - range interaction .
the exchange interaction involves bubble diagrams with respect to both unscreened coulomb interaction @xmath101 and short - range interaction attractive @xmath95 .
thus , eq . ( 2.33 ) in ref .
[ ] is equivalent to the bs equation @xmath102 , and eq .
( 2.32 ) in ref .
[ ] corresponds to the bs equation @xmath103 . in other words ,
the bs equation @xmath104 is equivalent to both , eqs .
( 2.32 ) and ( 2.33 ) in the paper by ct and griffin .
the kernel of the bs equations in the case when @xmath105 and @xmath95 replaced by a contact interaction@xcite does take into account the direct interaction ( terms which depend on @xmath106 and @xmath107 ) and exchange interaction ( terms which depend on @xmath108 and @xmath15 ) , and therefore , the ct and griffin results in the case of an attractive hubbard model should be similar to the results obtained by the bs approach .
this statement is supported by fig . 1 and fig
the two approaches provide very similar results , but the difference between them increases with increasing the lattice velocity . to illustrate this statement
, we compare the two analytical results for the speed of sound in a weak - coupling regime in a moving 1d optical lattice . by following the calculations that have already been given by belkhir and randeria,@xcite
we reduce our secular determinant ( [ bsz ] ) to the following @xmath23 determinant : @xmath109 where @xmath110 . the solution @xmath111 provides the speed of sound which in a stationary lattice is @xmath112 , where the fermi velocity is @xmath113 .
the same result has been obtained in ref .
[ ] . in the case of a moving lattice , the corresponding determinant is : @xmath114 where @xmath115 and @xmath116q^2\alpha\tan^2(q)}{2\delta^2}. \label{a11}\ ] ] the last determinant provides the following expression for the phononlike dispersion in the long - wavelength limit : @xmath117\tan^2(q)- \delta^2\left[1+\frac{1-\alpha}{\alpha}\tan^2(q)\right ] } \label{1d}\ ] ] here , @xmath118 is the fermi velocity in the presence of a supercurrent .
the speed of sound obtained by the green s function method is:@xcite @xmath119 } \label{1dy}\ ] ] it should be mentioned that in the case when @xmath120 , the term @xmath121\tan^2(q)$ ] is more important than the term @xmath122 in ( [ 1d ] ) and the term @xmath123 in ( [ 1dy ] ) .
we have derived the bs equations for the spectrum of the collective modes in the case when an equal mixture of atomic fermi gas of two hyperfine states with contact interaction is loaded into a moving optical lattice .
it is worth mentioning that there are two advantages of the bs formalism over the green s function approach .
first , within the bs approach we obtain the poles of density and spin response functions in a uniform manner , i.e. one secular equation provides not only the poles of the density response function , but the poles of the spin response function as well .
second , to obtain the poles of the density ( spin ) response function within the green s function method , one has first to take into account the ladder diagrams with respect to the short - range interaction , and after that to sum up the bubble diagrams with respect to both , the short- and long - range interactions .
999 j. k. chin , d. e. miller , y. liu , c. stan , w. setiawan , c. sanner , k. xu , and w. ketterle , nature ( london ) * 443 * , 961 ( 2006 ) .
d. e. miller , j. k. chin , c. a. stan , y. liu , w. setiawan , c. sanner , and w. ketterle , phys .
lett . * 99 * , 070402 ( 2007 ) .
m. greiner , c. a. regal , and d. s. jin , nature ( london ) * 426 * , 537 ( 2003 ) .
s. jochim , m. bartenstein , a. altmeyer , g. hendl , s. riedl , c. chin , j. hecker denschlag , and r. grimm , science * 302 * , 2101 ( 2003 ) . m. w. zwierlein , c. a. stan , c. h. schunck , s. m. f. raupach , s. gupta , z. hadzibabic , and w. ketterle , phys . rev . lett . * 91 * , 250401 ( 2003 ) .
t. bourdel , l. khaykovich , j. cubizolles , j. zhang , f. chevy , m. teichmann , l. tarruell , s. j. j. m. f. kokkelmans , and c. salomon , phys .
lett . * 93 * , 050401 ( 2004 ) .
w. hofstetter , j. i. cirac , p. zoller , e. demler , and m. d. lukin , phys .
* 89 * , 220407 ( 2002 ) .
g. orso and g.v .
shlyapnikov , phys .
lett . * 95 * , 260402 ( 2005 ) .
l. p. pitaevskii , s. stringari , and g. orso , phys .
a * 71 * , 053602 ( 2005 ) .
w. hofstetter phil . mag .
, * 86 * , 1891 ( 2006 ) .
w. yi and l .- m .
duan , phys .
a * 73 * , 063607 ( 2006 ) . t. koponen , j .-
martikainen , j. kinnunen , and p. trm1 , phys . rev . a * 73
* , 033620 ( 2006 ) .
t. koponen , j .-
martikainen , j. kinnunen , l > m. jensen , and p. trm1 , new journal of physics * 8 * , 179 ( 2006 ) .
m. iskin and c. a. r. s de melo , phys .
99 * , 080403 ( 2007 ) . t. k. koponen , t. paananen , j
martikainen , and p. trml , phys .
rev . lett . * 99 * , 120403 ( 2007 ) .
t. paananen , j. phys .
b : at . mol .
. phys . * 42 * , 165304 ( 2009 ) .
ai - xia zhang and ju - kui xue , phys . rev .
a * 80 * , 043617 ( 2009 ) .
t. k. koponen , t. paananen , and p. trml , phys .
lett . * 102 * , 165301 ( 2009 ) .
m. iskin and c. a. r. s de melo , phys .
* 103 * , 165301 ( 2009 ) .
z. g. koinov , physica c * 407 * , 470 ( 2010 ) ; physica status solidi ( b ) , * 247 * , 140 ( 2010 ) .
z. g. koinov , ann .
( berlin ) * 522 * , 693 ( 2010 ) .
r. ganesh , a. paramekanti , and a. a. burkov , phys .
a * 80 * , 043612 ( 2009 ) .
l. belkhir and m. randeria , phys .
b * 49 * , 6829 ( 1994 ) .
y. yunomae , i. danshita , d. yamamoto , n yokoshi , and s tsuchiya , journal of physics : conference series * 150 * , 032128 ( 2009 ) . y. yunomae , d. yamamoto , i. danshita , n. yokoshi , and s. tsuchiya , phys .
a * 80 * , 063627 ( 2009 ) .
r. cot and a. griffin , phys .
b , * 48 * , 10404 ( 1993 ) . | we derive the bethe - salpeter ( bs ) equations for the collective - mode spectrum of superfluid fermi gases of equal mixture of atomic fermi gas of two hyperfine states loaded into a moving optical lattice . in a moving lattice
the superfluid state is unstable due to spontaneous emission of the short - wavelength rotonlike excitations which appear in the spectrum of the collective modes .
it is shown that the spectrum obtained by the bs equations is in an excellent agreement with the collective - mode dispersion calculated by the perturbation approach , while there are some differences between the results obtained by density response function method and by the bs approach .
the difference increases with increasing the lattice velocity , which can be seen in the analytical approximations for the dispersion relation in the long - wavelength limit in a weak - coupling regime . |
it was calabi @xcite who first recognised the rich geometry that can be found on a hypersurface of @xmath6 when the latter is equipped with its natural cross product and @xmath3-structure .
the realization , much later , of metrics with holonomy _ equal _ to @xmath3 allowed this theory to be extended , whilst retaining the key features of the `` euclidean '' theory . the second fundamental form or weingarten map @xmath7 of a hypersurface @xmath8 in a manifold @xmath9 with holonomy @xmath3 can be identified with the intrinsic torsion of the associated @xmath0-structure .
the latter is defined by a 2-form @xmath10 and a @xmath11-form @xmath12 induced on @xmath8 , and @xmath7 is determined by their exterior derivatives .
the symmetry of @xmath7 translates into a constraint on the intrinsic torsion ( equivalently , on @xmath13 and @xmath14 ) that renders the @xmath0-structure what is called _
half flat_. conversely , a @xmath15-manifold @xmath8 with an @xmath0-structure that is half flat can ( at least if it is real analytic ) be embedded in a manifold with holonomy @xmath3 @xcite .
the metric @xmath16 on @xmath9 is found by solving a system of evolution equations that hitchin @xcite interpreted as hamilton s equations relative to a symplectic structure defined ( roughly speaking ) on the space parametrising the pairs @xmath17 .
the simplest instance of this construction occurs when @xmath8 is a so - called _ nearly - khler _ space , in which case @xmath18 is a conical metric over @xmath8 , in accordance with a more general scheme described by br @xcite .
the first explicit metrics known to have holonomy equal to @xmath3 were realized in this way . in this paper , we are concerned with the classification of left - invariant half - flat @xmath0-structures on @xmath5 , regarded as a lie group @xmath19 , up to an obvious notation of equivalence .
one of these structures is the nearly - khler one that can be found on @xmath20 , for any compact simple lie group @xmath21 , by realizing the product as the 3-symmetric space @xmath22 .
indeed , we verify that this nearly - khler structure is unique amongst invariant @xmath0-structures on @xmath23 ( see proposition [ prop : nkunique ] , that has a dynamic counterpart in proposition [ prop : nkunique ] ) .
examples of the resulting evolution equations for @xmath3-metrics have been much studied in the literature @xcite , but one of our aims is to highlight those @xmath3-metrics that arise from half - flat metrics with specific intrinsic torsion , motivated in part by the approach in @xcite .
nearly - khler corresponds to gray - hervella class @xmath24 , and it turns out that a useful generalization in our half - flat context consists of those metrics of class @xmath25 ; see section [ sec : symgrp ] . by careful choices of the coefficients in @xmath26 and @xmath27
, we obtain metrics on @xmath5 of the same class with zero scalar curvature .
another aim is to develop rigorously the algebraic structure of the space of invariant half - flat structures on @xmath5 , and in section [ sec : para ] we show that the moduli space they define is essentially a finite - dimensional symplectic quotient .
this is a description expected from @xcite , and in our treatment relies on elementary matrix theory .
for example , the @xmath28-form @xmath26 can be represented by a @xmath29 matrix @xmath30 , and mapping @xmath10 to the 4-form @xmath31 corresponds to mapping @xmath30 to the transpose of its adjugate .
we shall however choose to use a pair of symmetric @xmath32 matrices @xmath33 to parametrise the pair @xmath34 .
the matrix algebra is put to use in section [ sec : flow ] to simplify and interpret the flow equations for the associated ricci - flat metrics with holonomy @xmath3 .
the significance of the class @xmath35 becomes clearer in the evolutionary setting , as it generates known @xmath3-metrics . in our formulation ,
the equations ( for example in corollary [ cor : flow ] ) have features in common with two quite different systems considered in @xcite and @xcite , but both in connection with painlev equations .
a more thorough analysis of classes of solutions giving rise to @xmath3-metrics is carried out in section [ sec : further ] . some of these exhibit the now familiar phenomenon of metrics that are asymptotically circle bundles over a cone ( `` abc metrics '' ) .
all our @xmath3-metrics are of course of cohomogeneity one , and this allows us to briefly relate our approach to that of @xcite . in the final part of the paper , we present the tip of the iceberg that represents a numerical study of hitchin s evolution equations for @xmath23 .
we recover metrics that behave asymptotically locally conically when @xmath36 belongs to a fixed @xmath28-dimensional subspace .
more precisely , we show empirically that the planar solutions are divided into two classes , only one of which is of type abc .
this can be understood in terms of the normalization condition that asserts that @xmath10 and @xmath27 generate the same volume form , and is a worthwhile topic for further theoretical study . for the generic case
, the flow solutions do not have tractable asymptotic behaviour , but again the geometry of the solution curves ( illustrated in figure [ fig : g2sol3d ] ) is constrained by the normalization condition that defines a cubic surface in space .
this paper grew out of an attempt to reconcile various contributions appearing in the literature .
of particular importance concerning @xmath37-structures are schulte - hengesbach s classifications of half - flat structures ( * ? ? ?
* theorem 1.4 , chapter 5 ) , and hitchin s notion of stable forms @xcite .
in addition , the explicit constructions of @xmath38-metrics appearing in this paper are based on the work of brandhuber et al , cveti et al @xcite , as well as the contributions of dancer and wang @xcite .
throughout the paper @xmath39 will denote the @xmath15-manifold @xmath40 .
as this is a lie group , we can trivialise the tangent bundle .
we describe left - invariant tensors via the identification @xmath41 relative to left multiplication .
we keep in mind that there are lie algebra isomorphisms @xmath42 which at the group level can be phrased in terms of the diagram @xmath43 the cotangent space of @xmath39 , at the identity , consists of two copies of @xmath44
. we shall write @xmath45 and choose bases @xmath46 of @xmath47 and @xmath48 of @xmath49 such that @xmath50 here @xmath51 denotes the exterior differential on @xmath47 and @xmath52 induced by the lie bracket .
we wish to endow @xmath39 with an @xmath0-structure . to this end
it suffices to specify a suitable pair of real forms : a @xmath11-form @xmath27 , whose stabiliser ( up to a @xmath53-covering ) is isomorphic to @xmath54 , and a non - degenerate real @xmath55-form @xmath56 .
these two forms must be compatible in certain ways .
above all , @xmath12 must be a _ primitive _ form relative to @xmath10 , meaning @xmath57 .
so as to obtain a genuine almost hermitian structure we also ask for volume matching and positive definiteness : @xmath58 these forms @xmath27 and @xmath59 are _ stable _ in the sense their orbits under @xmath60 are open in @xmath61 .
the following well known properties ( cf .
@xcite , and @xcite for the study of @xmath11-forms ) of stable forms will be used in the sequel : 1 . there are two types of stable @xmath11-forms on @xmath62 .
these are distinguished by the sign of a suitable quartic invariant , @xmath63 , which is negative precisely when the stabiliser is @xmath64 ( up to @xmath65 ) ; each form of this latter type determines an almost complex structure @xmath66 .
the stable forms @xmath59 and @xmath27 determine `` dual '' stable forms : @xmath59 determines the stable @xmath67-form @xmath68 , and @xmath27 determines the @xmath69-form @xmath70 characterised by the condition that @xmath71 be of type @xmath72 . as
@xmath0-modules @xmath73 decomposes in the following manner : @xmath74\!]\cong{\lambda}^5t^*,\\ { \lambda}^2t^*\cong[\![{\lambda}^{2,0}]\!]{\oplus}[{\lambda}^{1,1}_0]{\oplus}{\mathbb r}\cong{\lambda}^4t^*,\\ { \lambda}^3t^*\cong[\![{\lambda}^{3,0}]\!]{\oplus}[\![{\lambda}^{2,1}_0]\!]{\oplus}[\![{\lambda}^{1,0}]\ ! ] , \end{gathered}\ ] ] using the bracket notation of @xcite . in terms of this decomposition ( see @xcite ) , the exterior derivatives of @xmath75 may now be expressed as @xmath76 where we have used a suggestive notation to indicate the relation between forms and the intrinsic torsion @xmath77 , i.e. , the failure of @xmath78 to reduce to @xmath0 .
obviously , this expression depends on our specific choice of normalisation ( cf . ) .
generally , @xmath77 takes values in the @xmath79-dimensional space @xmath80 our main focus , however , is to study the subclass of _ half - flat @xmath0-structures _ :
these are characterised by the vanishing of @xmath81 , and @xmath82 , i.e. , @xmath83 to appreciate the terminology `` half flat '' , it helps to count dimensions : @xmath84 , @xmath85 , @xmath86 , @xmath87 .
in particular , observe that for half - flat structures @xmath77 is restricted to take its values in @xmath88 dimensions out of @xmath89 possible . in this context
, `` flat '' would mean _
holonomy_. for emphasis ,
we formulate : [ prop : intr_space ] for any invariant half - flat @xmath0-structure @xmath91 on @xmath39 the following holds : if @xmath92 then @xmath93 . if @xmath94 then @xmath95 . in particular , any structure with vanishing @xmath96 component
has @xmath97=0\in h^3(m ) $ ] . in the case
when @xmath92 we shall say the half - flat structure is _
coupled_. the second case above , @xmath98 , is referred to as _ co - coupled_. when the half - flat structure is both coupled and co - coupled , so @xmath99 , it is said to be _ nearly - khler_. [ [ examples - of - type- . ] ] examples of type . + + + + + + + + + + + + + + + + + + as the next two examples illustrate , it is not difficult to construct half - flat structures of type @xmath25 .
[ ex : w1w3 ] in this example we fix a non - zero real number @xmath100 and consider the pair of forms @xmath17 given by : @xmath101 where @xmath102 is defined via the relation @xmath103 clearly , @xmath104 and @xmath105 .
a calculation shows @xmath106 so that @xmath107 the @xmath11-form @xmath108 is given by @xmath109 note that the following normalisation condition is satisfied : @xmath110 in order to verify that the intrinsic torsion is of type @xmath111 , we calculate the exterior derivatives of @xmath112 , @xmath113 , and @xmath114 : @xmath115 finally , note that the associated metric is given by @xmath116 and one finds that the scalar curvature is positive : @xmath117 . [ ex : w1w3s0 ] consider the following pair of stable forms : @xmath118 we find that @xmath119 , and the @xmath11-form @xmath108 is given by @xmath120 the normalisation condition then reads @xmath121 the associated metric takes the form @xmath122 in this case one finds that the scalar curvature is zero .
the author of @xcite uses lie algebra degenerations to study invariant hypo @xmath123-structures on @xmath124-dimensional nilmanifolds . in a similar way
, one could study half - flat structures on the various group contractions of @xmath5 like @xmath125 , where @xmath126 is a compact quotient of the heisenberg group .
( see @xcite for partial studies of such contractions ) .
the invariant half - flat structures on @xmath39 can be described in terms of symmetric matrices . in order to do this ,
we recall the local identifications and set @xmath127 , the space of real @xmath29 matrices , and @xmath128 , the space of real symmetric trace - free @xmath32 matrices .
there is a well known correspondence between @xmath129 and @xmath130 ; a fact which is for example used in the description of the trace - free ricci - tensor @xmath131 on a riemannian @xmath132-manifold .
[ lem : equiv - iso ] there is an equivariant isomorphism @xmath133 which maps a @xmath134 matrix @xmath135 to the matrix @xmath136 by fixing an oriented orthonormal basis @xmath137 of @xmath138 , we make the identifications @xmath139 , @xmath140 via @xmath141 the asserted isomorphism is then given by contraction on the middle two indices , as in the following example : @xmath142 table [ tab : comp - inv - cov ] summarises how invariants and covariants are related under the above isomorphism @xmath143 .
cc ku & sv + k & s + 4(kk^t ) & ( s^2 ) + -2(k^t ) & ( s^2)_0 + -24(k ) & ( s^3 ) + 4(kk^t)k & ( s^2)s + 2kk^tk & 34(s^2)s-(s^3)_0 + 4((kk^t)^2 ) & 3(s)+14 ( s^4 ) + 2(kk^t)^2 & ( s)+14(s^4 ) + -24(k)k & ( s^3)s + 4(kk^t)(k ) & 13(s^3)s-(s^4)_0 + now , let us fix a cohomology class @xmath144 .
we have : [ thm : half - flat - param ] the set @xmath145 of invariant half - flat structures on @xmath146 with @xmath97=c $ ] can be regarded as a subset of the _ commuting variety _ : @xmath147=0\right\}.\ ] ] recall @xmath148 , where @xmath149 so that we have @xmath150 the equation @xmath104 implies that @xmath151 which defines @xmath152 . also note @xmath153 lies in a space isomorphic to @xmath154 .
we may assume that @xmath155 the condition @xmath156 implies @xmath157 lies in the kernel of some @xmath2-equivariant map @xmath158 which must correspond to @xmath159=qp - pq $ ] .
consider the open subset set @xmath160 , @xmath161 , of the commuting variety given by pairs @xmath33
satisfying @xmath162 then @xmath145 is the hypersurface in @xmath160 characterised by the normalisation condition @xmath163 the space @xmath164 has a natural symplectic structure , and @xmath2 acts hamiltonian with moment map @xmath165 given by @xmath166.\ ] ] via ( singular ) symplectic reduction @xcite , we can the simplify the parameter space significantly : [ cor : param ] the set @xmath145 of half - flat structures modulo equivalence relations is a subset of the singular symplectic quotient @xmath167 for later use , we observe that in terms of the matrix framework , the dual @xmath11-form @xmath114 has exterior derivative given as follows : [ lem : hatgamma ] fix a cohomology class @xmath168 . for any element @xmath169 corresponding to an invariant half - flat structure , the associated @xmath55-form @xmath170 corresponds to the matrix @xmath171 , where @xmath172 in particular , if @xmath173 and we set @xmath174 then @xmath175 let @xmath176 : if @xmath33 corresponds to a coupled structure then @xmath177 and @xmath178 for a non - zero constant @xmath179 . if @xmath33 corresponds to a co - coupled structure then @xmath180 for a non - zero constant @xmath181 . obviously , the half - flat pair @xmath33 is of type @xmath111 if and only if the matrices @xmath182 and @xmath183 are proportional , i.e. , we have @xmath184 ; the type does not reduce further provided @xmath185 and @xmath186 . using these conditions
it is easy to show that the structures of example [ ex : w1w3 ] and example [ ex : w1w3s0 ] have the type of intrinsic torsion claimed .
indeed , in the first example , using lemma [ lem : hatgamma ] , we find that @xmath187 whilst the matrices of the second example satisfy @xmath188 [ ex : nk ] in this case , the following conditions should be satisfied : @xmath189 for some @xmath190 .
this is equivalent to solving the equations @xmath191 where @xmath192 .
we find this system of equations can be formulated as @xmath193 keeping in mind that we must have @xmath194 , we obtain only the following solutions @xmath195 : @xmath196 note that these solutions are identical after using a permutation ; the corresponding matrices @xmath36 are of the form @xmath197 respectively .
the above example captures a well known fact about uniqueness of the invariant nearly - khler structure on @xmath5 . in our framework , this can be summarised as follows ( compare with ( * ? ? ?
* proposition 2.5 ) and ( * ? ? ?
* proposition 1.11 , chapter 5 ) ) .
[ prop : nkunique ] modulo equivalence and up to a choice of scaling @xmath198 , there is a unique invariant nearly - khler structure on @xmath39 .
it is given by the class @xmath199 $ ] where @xmath200 as observed in ( * ? ? ?
* proposition 1.8 ) there are no invariant ( integrable ) complex structures on @xmath39 admitting a left - invariant holomorphic @xmath72-form .
indeed , in terms of @xmath201 matrices this assertion is captured by in the notation of lemma [ lem : hatgamma ] , if @xmath202 then @xmath203 .
although we have chosen to focus on the vector space @xmath130 and @xmath204 matrices , we conclude this section with a neat consequence of stability .
consider @xmath205 .
the cayley - hamilton theorem states that @xmath206 where @xmath207 , @xmath208 , and @xmath209 .
consider now the adjugate @xmath210 so that @xmath211 .
table [ tab : comp - inv - cov ] implies that the mapping @xmath212 corresponds to a multiple of @xmath213 .
the following result describes a viable alternative to the square root of a @xmath214 matrix ; it can be proved directly using the singular value decomposition .
any @xmath29 matrix with positive determinant equals @xmath215 for some unique @xmath216 .
let @xmath217 be an interval .
a @xmath3-structure and metric on the @xmath4-manifold @xmath218 can be constructed from a one - parameter family of half - flat structures on @xmath39 by setting @xmath219 where @xmath220 and @xmath221 .
it is well known @xcite the holonomy lies in @xmath3 if and only if @xmath222 . for structures defined via a one - parameter family of half - flat structures ,
this can be phrased equivalently as : the riemannian metric associated with the @xmath3-structure has holonomy in @xmath3 if and only if the family of forms satisfies the equations : @xmath223 differentiation of @xmath224 and @xmath225 gives us : @xmath226 since the one - parameter family consists of half - flat @xmath227-structures , we have @xmath228 ( for each fixed @xmath229 ) , so the conditions @xmath230 reduce to the system . as explained in (
* theorem 8) , the evolution equations can be viewed as the flow of a hamiltonian vector field on @xmath231 .
it is a remarkable fact that this flow does not only preserve the closure of @xmath59 and @xmath27 , but also the compatibility conditions . in order to show that a given @xmath3-metric on @xmath218 has holonomy equal to @xmath3
, one must show there are no non - zero parallel @xmath232-forms on the @xmath4-manifold ( see the treatment by bryant and the second author ( * ? ? ? * theorem 2 ) ) .
for many of the metrics constructed in this paper , the argument is the same , or a variation of , the one applied in ( * ? ? ?
* section 3 ) . in terms of matrices
@xmath176 , we can rephrase the flow equations by [ prop : g2flow - matr ] as a flow , @xmath233 , in @xmath145 , the evolution equations take the form @xmath234 these equations are particularly simple when the cohomology class @xmath235 of @xmath27 satisfies the criterion @xmath173 . in this case , by lemma [ lem : hatgamma ] , we have : [ cor : flow ] for a flow , @xmath233 , in @xmath236 with @xmath173 , the equations take the form : @xmath237 when phrased as above , the preservation of the normalisation essentially amounts to jacobi s formula for the derivative of a determinant .
proposition [ prop : g2flow - matr ] tells us that the @xmath3-metrics on @xmath218 that arise from the flow of invariant half - flat structures , can be interpreted as the lift of suitable paths @xmath238 to paths @xmath239 and moreover these paths lie on level sets of the ( essentially hamiltonian ) functional @xmath240 let @xmath33 be a ( normalised ) solution of the flow equations . then the trajectory @xmath241 lies on the level set @xmath242 inside the space @xmath243 .
[ [ dynamic - examples - of - type- . ] ] dynamic examples of type . + + + + + + + + + + + + + + + + + + + + + + + + + + rephrasing results of @xcite , we now consider the one - parameter family of forms @xmath244 given by @xmath245 in this case , we find that @xmath246 and we shall assume @xmath247 and @xmath248 , so as to ensure @xmath249 . also note that @xmath250 in particular , the normalisation condition reads : @xmath251 in order to solve the flow equations , we also need the @xmath55-form @xmath252 based on the above expressions , the system becomes : @xmath253 these equations can be rewritten as a system of first order odes in @xmath254 and @xmath102 : @xmath255 as we require the normalisation to hold , we can not choose initial conditions @xmath256 freely . after suitable reparametrization
, we find the explicit solution : @xmath257 where @xmath258 , and @xmath259 note that whilst @xmath260 is always non - zero , @xmath261 can be zero . indeed
, this happens if @xmath262 is chosen such that the quadratic equation @xmath263 has a solution @xmath264 for some @xmath265 .
this is the case for any non - zero @xmath266 : if @xmath267 the solution is obtained for @xmath268 and if @xmath269 the solution occurs when @xmath270 introducing @xmath271 , we can express the exterior derivatives of the defining forms via @xmath272 as @xmath273 and @xmath274 , this implies that the constructed one - parameter family of @xmath275-structures consists of members of type @xmath25 .
the associated family of metrics takes the form @xmath276 and has scalar curvature given by @xmath277 zero scalar curvature is obtained for the solution which has @xmath278 .
indeed , in this case the scalar curvature is zero when @xmath279 . finally , let us remark that the associated @xmath3-metric is of the form @xmath280 , or , phrased more explicitly , in terms of the parameter @xmath281 : @xmath282 & \frac{12}{1+as^{-3}}ds\otimes ds\\ & \hskip20pt+\sum_{i=1}^3\left(\frac{s^2(1+as^{-3})}{\sqrt{3}}(e^{2i-1}+e^{2i})\otimes ( e^{2i}+e^{2i-1 } ) + \sqrt{3}s^2(e^{2i-1}-e^{2i})\otimes ( e^{2i}-e^{2i-1})\right ) .
\end{split}\ ] ] if @xmath283 this metric is conical whilst for @xmath284 , the metric is asymptotically conical : when @xmath285 it tends to a cone metric @xmath286 over @xmath39 . in terms of the classification @xcite , the metrics belong to the family ( i ) . in terms of the matrix framework
, the one - parameter families of pairs @xmath33 take the form : @xmath287 in particular , we get another way of verifying the co - coupled condition : @xmath288
[ [ metrics - with - symmetry . ] ] metrics with symmetry .
+ + + + + + + + + + + + + + + + + + + + + + following mainly @xcite , we study examples that relate our framework to certain constructions of @xmath3-metrics appearing in the physics literature .
our starting point in a one - parameter families half - flat pairs @xmath17 of the form : @xmath289 using the normalisation condition , we are able to express the associated one - parameter family of metrics on @xmath39 as follows : @xmath290 and the flow equations read : @xmath291 notice that the @xmath292 action which interchanges the two copies of @xmath293 preserves the metric provided the cohomology class @xmath97 $ ] is of the form @xmath294 , i.e. , @xmath97=(a ,- a ) $ ] . the action interchanges metrics of half - flat structures with @xmath97=(a,0 ) $ ] with those for which @xmath97=(0,-a ) $ ] .
the latter observation is related to the notion of a _ flop _ @xcite .
[ rem : volgrowth ] the quantity @xmath295 can be viewed as the ratio of the volume of @xmath296 relative to a fixed background metric on @xmath1 . as expected , we find that @xmath297 where we have used that @xmath298 , by the normalisation condition . a metric ansatz that has led to the discovery of new complete @xmath3-metrics ( see , for instance , @xcite ) can be expressed in terms of the condition @xmath173 . in this case
, we find @xmath299 where @xmath300 or , alternatively , @xmath301 note that , up to a sign , we have @xmath302
. expressed in terms of the metric function @xmath303 , the flow equations become : @xmath304 the complete metrics constructed by brandhuber et al @xcite arise as a further specialisation of this system .
indeed , if we take @xmath305 and @xmath306 and set @xmath307 , then the system reads @xmath308 which is the same as in ( * ? ? ?
* equation ( 3.1 ) ) , where the authors find the following explicit holonomy @xmath3-metric : @xmath309 asymptotically this is the metric of a circle bundle over a cone , in short an _
abc metric_. in terms of the classification @xcite , it belongs to the family ( ii ) .
[ [ cohomogeneity - one - ricci - flat - metrics . ] ] cohomogeneity one ricci flat metrics .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + any solution of gives us a cohomogeneity one ricci flat metric on @xmath218 .
an important aspect of the cohomogeneity one terminology is to bridge a gap between our framework and the `` lagrangian approach '' appearing in the physics literature ( see , e.g. , ( * ? ? ?
* section 4 ) ) .
for example , consider the metric from the above example , assuming for simplicity that @xmath305 and @xmath310 . by @xcite , we know that the shape operator @xmath311 of the principal orbit @xmath312 satisfies the equation @xmath313 . for the given metric , we find that @xmath314 we also observe that @xmath315 in general , the ricci flat condition can now be expressed as : @xmath316 combined with another equation expressing the einstein condition for mixed directions . if we take the trace of the first equation in , and combine with the second one , we obtain the following conservation law : @xmath317 as explained in @xcite , the above system has a hamiltonian interpretation . it is this interpretation , in its lagrangian guise and phrased with the use of superpotentials , one frequently encounters in the physics literature . in this setting ,
the kinetic and potential energies are given by @xmath318 these definitions agree with those in @xcite up to a multiple of @xmath319 . in @xcite ,
the authors provide a relevant description of the superpotential ; in classical terms this is a solution of a time - independent hamilton - jacobi equation . in the concrete example
, the superpotential @xmath320 can be viewed as a function of @xmath321 .
concretely , we can take @xmath322 in terms of @xmath320 , the flow equations can then be expressed as follows : @xmath323 where @xmath324 ( assuming @xmath325 ) , @xmath326 and @xmath327 finally , we remark that the kinetic and potential terms can be expressed in the form @xmath328 as a further specialisation , let us consider the case when @xmath283 and @xmath329 , @xmath330 ; this is the nearly - khler case
. then the shape operator is proportional to the identity : @xmath331 , and the kinetic and potential terms are @xmath332 respectively .
so the total energy is zero @xmath333 for all @xmath334 .
the superpotential is the fifth oder polynomial @xmath335 [ [ uniqueness - flowing - along - a - line . ] ] uniqueness : flowing along a line . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in the case when @xmath336 , the flow equations turn out to have a unique ( admissible ) solution satisfying for which @xmath36 belongs to a fixed one - dimensional subspace . [
prop : nkunique ] assume @xmath337 is a solution of
. then @xmath36 belongs to a fixed @xmath338-dimensional subspace of @xmath339 if and only if the associated @xmath3-metric is the cone metric over @xmath5 endowed with its nearly - khler structure .
it is easy to see that the solution of which corresponds to the cone metric over @xmath5 ( with its nearly - khler structure ) is represented by @xmath340 so , in this case , @xmath36 indeed belongs to a fixed @xmath338-dimensional subspace of @xmath339 .
conversely , let us assume we are given a solution such that @xmath341 then the system reads : @xmath342 these equations show that there is a purely algebraic constraint to having a solution : @xmath343 where @xmath344 .
uniqueness of the `` nearly - khler cone '' , as a flow solution , now follows by observing that these algebraic equations have the following set of solutions : @xmath345 the solutions with @xmath346 are not `` admissible '' whilst the remaining solutions all result in one - parameter families of pairs equivalent to .
as indicated in the earlier parts of this paper , previous studies of @xmath3-metrics on @xmath218 have focused mainly on metrics with isometry group ( at least ) @xmath347 .
in addition , most of the attention has been centred around solutions in @xmath145 for @xmath348 .
a technique that seems effective if one is specifically looking for complete metrics is to choose the initial values of the flow equations to obtain a singular orbit at that point ( meaning , in our context , one whose stabilizer has positive dimension in @xmath349 ) .
this approach was adopted in @xcite for @xmath350 holonomy . however ,
this final section shifts the focus of our investigation in order to illustrate some more generic behaviour of the flow on the space of invariant half - flat structures on @xmath40 .
we first look for solutions in @xmath351 for which @xmath36 takes the form @xmath352 where @xmath353 are smooth functions on an interval @xmath354 .
a solution of is then uniquely specified by the quadruple @xmath355 we have solved the system for a wide range of initial conditions . a selection of solutions
are shown in figure [ fig : g2sol2d ] . apart from the nearly - khler straight line ,
these solutions are new .
plotting the metric functions , we find that some of the new metrics have one stabilising direction when @xmath356 and no collapsing directions ( they are therefore abc metrics of the sort mentioned in connection with ) .
the others have shrinking directions which cause the volume growth to slow down as shown in figure [ fig : g2sol2dvgrowth ] .
more precisely , in the case @xmath357 , the normalisation forces @xmath358 , written as @xmath359 , to lie on the curve @xmath360 which has two branches separated by the line @xmath361 .
one branch corresponds to positive - definite metrics , including the nearly - khler solution @xmath362 the abc metrics are those for which @xmath363 , and appear on the top left of the nearly - khler line in figure [ fig : g2sol2dqcurvea ] , in green in the coloured version . when @xmath364 , the nearly - khler solution is excluded .
nevertheless , the overall picture remains valid , meaning one branch of the normalisation curve corresponds to positive - definite metrics , and this branch itself has two half pieces , one corresponding to abc curves and one to the other solutions . in the trace - free case , @xmath365
, all solutions degenerate at a point @xmath366 .
the abc solutions are `` half complete '' , meaning that away from the degeneration they are complete in one direction of time .
( see @xcite for other examples of half - complete @xmath3-metrics ) .
the other solutions reach another degeneracy point @xmath367 in finite time .
the singularity at @xmath368 can not be resolved .
in particular , it is not possible to find complete @xmath3-metrics .
one way to circumvent this issue is to consider flow solutions for which @xmath97\neq0 $ ] ; solutions of this form include the metrics discovered by brandhuber et al @xcite .
now , turning to `` less symmetric '' @xmath3-metrics , we consider for solutions in @xmath351 with @xmath36 of the ( generic ) form : @xmath369 where @xmath370 are smooth functions on an interval @xmath354 .
a solution of is then uniquely specified by the sextuple @xmath371 as in the case of planar solutions , we have solved the flow equations for a large number of initial conditions .
in contrast with the planar case , we have not been able to find metrics with one stabilising directions as @xmath372 .
we shall confine our presentation to the class of solutions with the same initial point @xmath373 as the nearly - khler solution , but with varying velocity vector @xmath374 similar to the planar case , the flow lines are governed by the normalization condition , and is replaced by the cubic surface @xmath375 the asymptotic planes corresponding to the vanishing of @xmath376 separate the surface into four hyperboloid - shaped components , and only the one with all factors negative is relevant to our study of positive - definite metrics with holonomy @xmath3 .
the nearly - khler solution @xmath377 ( cf . ) corresponds to its centre point .
families of solutions are shown in figure [ fig : g2sol3d ] which , like those in figure [ fig : g2sol2d ] , were plotted using _ mathematica _ and the command _ ndsolve_. to obtain the curves , it was convenient to further reduce attention to the case in which @xmath378 are all negative .
the corresponding subset of is now a curved triangle @xmath379 with truncated vertices . by issuing a plotting command for @xmath380
, we obtained an abundant sample of mesh points to feed into as initial values .
one can then regard each curve as the continuing trajectory of a particle launched towards a point of @xmath381 , which fits in close to the apex of figure [ fig : g2sol3dqcurvea ] .
all the solutions , apart from the central nearly - khler one , are new .
they tend to have shrinking directions , causing the volume growth to slow down .
the @xmath382 solution curves in figure [ fig : g2sol3dqcurvea ] are plotted for the range @xmath383 since many develop singularities close to @xmath384 ( and close to @xmath385 though positive @xmath386 is not shown ) . in the coloured `` cocktail umbrella '' picture , they are separated into groups distinguished by the value of the function @xmath387 of the initial condition , with the nearly - khler line @xmath388 and its close neighbours in red .
solutions resulting from one of the coordinates being positive can be short - lived in comparison to the others , leading to less coherent plots , and this is why they are absent .
the view looking down the nearly - khler line from a point @xmath389 with @xmath390 is shown in figure [ fig : g2sol3dqcurveb ] .
the @xmath391 symmetry obtained by permuting the coordinates is evident .
the splitting behaviour at the three `` ends '' is to some extent artificial , reflecting as it does the truncation that has resulted from our decision to restrict attention to the negative octant .
the abc two - function solutions of figure [ fig : g2sol2dqcurvea ] in the previous subsection arise when two of @xmath378 coincide and assume a common value greater than @xmath392 .
the projection of these planar curves orthogonal to the nearly - khler line can be seen in figure [ fig : g2sol3dqcurvec ] .
computations confirm that , unlike the generic curves of figure [ fig : g2sol3dqcurveb ] emanating from @xmath393 , these can be extended for all @xmath394 .
in addition to the solutions in @xmath395 , we have investigated solutions in @xmath396 .
regarding the asymptotic behaviour of the associated @xmath3-metrics , the overall picture appears not dissimilar to the one we have described by deforming the nearly - khler velocity . taking account also of the numerical analysis in @xcite , we conjecture that the only solutions that can be extended for @xmath397 or @xmath398 lie in a plane .
both authors thank mark haskins for discussions that helped initiate this research , and in particular for bringing @xcite to their attention .
the first author gratefully acknowledge financial support from the danish council for independent research , natural sciences .
30 v. apostolov , s. salamon , khler reduction of metrics with holonomy @xmath3 .
( 2004 ) , no .
1 , 4361 . m. atiyah , j. maldacena , c. vafa , an m - theory flop as a large n duality .
strings , branes , and m - theory . j. math .
42 ( 2001 ) , no . 7 , 32093220 . c. br , real killing spinors and holonomy .
154 ( 1993 ) , no .
3 , 509521
. l. bedulli , l. vezzoni , the ricci tensor of @xmath399-manifolds . j. geom
57 ( 2007 ) , no .
4 , 11251146 . a. brandhuber , @xmath400 holonomy spaces from invariant three - forms .
nuclear phys .
b 629 ( 2002 ) , no . 1 - 3 , 393416 .
a. brandhuber , j. gomis , s. gubser , s. gukov , gauge theory at large n and new @xmath400 holonomy metrics .
nuclear phys .
b 611 ( 2001 ) , no . 1 - 3 , 179204 . r. bryant , non - embedding and non - extension results in special holonomy
. the many facets of geometry , 346367 , oxford university press , oxford , 2010 .
r. bryant , s. salamon , on the construction of some complete metrics with exceptional holonomy .
duke math . j. 58 ( 1989 ) , no .
3 , 829850 . j .-
butruille , espace de twisteurs dune varit presque hermitienne de dimension 6 .
fourier ( grenoble ) 57 ( 2007 ) , no .
5 , 1451485 . j .-
butruille , homogeneous nearly khler manifolds .
handbook of pseudo - riemannian geometry and supersymmetry , 399423 , irma lect .
phys . , 16 , eur .
, zrich , 2010 .
e. calabi , construction and properties of some 6-dimensional almost complex manifolds .
87 1958 407438 . s. chiossi , a. fino , conformally parallel @xmath401 structures on a class of solvmanifolds .
z. 252 ( 2006 ) , no .
4 , 825848 .
s. chiossi , s. salamon , the intrinsic torsion of su(3 ) and @xmath400 structures .
differential geometry , valencia , 2001 , 115133 , world sci .
publ . , river edge , nj , 2002 . z. chong , m. cveti , g. gibbons , h. l , c. pope , p. wagner , general metrics of @xmath3 holonomy and contraction limits .
nuclear phys .
b 638 ( 2002 ) , no .
3 , 459482 .
d. conti , su(3)-holonomy metrics from nilpotent lie groups .
arxiv:1108.2450 [ math.dg ] . m. cveti , g. gibbons , h. l , c. pope , supersymmetric m3-branes and @xmath3 manifolds . nuclear phys .
b 620 ( 2002 ) , no . 1 - 2 , 328 .
m. cveti , g. gibbons , h. l , c. pope , a @xmath400 unification of the deformed and resolved conifolds .
b 534 ( 2002 ) , no . 1 - 4 , 172180 .
m. cveti , g. gibbons , h. l , c. pope , cohomogeneity one manifolds of @xmath402 and @xmath3 holonomy .
d ( 3 ) 65 ( 2002 ) , no .
10 , 106004 , 29 pp . m. cveti , g. gibbons , h. l , c. pope , orientifolds and slumps in @xmath3 and spin(7 ) metrics .
physics 310 ( 2004 ) , no .
2 , 265301 . a. dancer , mckenzie wang , painlev expansions , cohomogeneity one metrics and exceptional holonomy .
geom . 12 ( 2004 ) , no .
4 , 887926 . a. dancer , mckenzie wang , superpotentials and the cohomogeneity one einstein equations .
( 2005 ) , no .
. j. eschenburg , mckenzie wang , the initial value problem for cohomogeneity one einstein metrics . j. geom
10 ( 2000 ) , no .
1 , 109137 .
e. ferapontov , b. huard , a. zhang , on the central quadric ansatz : integrable models and painlev reductions .
j. phys . a 45 ( 2012 ) , no . 19 , 195204 , 11 pp .
m. fernndez , a. gray , riemannian manifolds with structure group @xmath3 . ann .
pura appl .
( 4 ) 132 ( 1982 ) , 1945 ( 1983 ) . n. hitchin , stable forms and special metrics .
global differential geometry : the mathematical legacy of alfred gray ( bilbao , 2000 ) , 7089 , contemp .
, 288 , amer .
soc . , providence , ri , 2001 .
e. lerman , r. montgomery , r. sjamaar , examples of singular reduction .
symplectic geometry , 127155 , london math .
lecture note ser .
, 192 , cambridge univ .
press , cambridge , 1993 .
w. reichel , ber die trilinearen alternierenden formen in 6 und 7 vernderlichen , dissertation , greifswald , ( 1907 ) .
f. reidegeld , exceptional holonomy and einstein metrics constructed from aloff - wallach spaces .
102 , no . 6 , 11271160 ( 2011 ) .
r. reyes carrin , a generalization of the notion of instanton .
differential geom .
( 1998 ) , no .
1 , 120 . s. salamon , riemannian geometry and holonomy groups .
pitman research notes in mathematics series , 201 .
longman scientific & technical , harlow ; copublished in the united states with john wiley & sons , inc .
, new york , 1989 .
isbn : 0 - 582 - 01767-x
. f. schulte - hengesbach , half - flat structures on lie groups .
phd thesis ( 2010 ) , hamburg .
r. westwick , real trivectors of rank seven .
linear and multilinear algebra 10 ( 1981 ) , no .
3 , 183204 . | we describe left - invariant half - flat @xmath0-structures on @xmath1 using the representation theory of @xmath2 and matrix algebra .
this leads to a systematic study of the associated cohomogeneity one ricci - flat metrics with holonomy @xmath3 obtained on @xmath4-manifolds with equidistant @xmath5 hypersurfaces .
the generic case is analysed numerically .
* half - flat structures on * thomas bruun madsen and simon salamon _ keywords : _ @xmath3- and @xmath0-structures , einstein and ricci - flat manifolds , special and exceptional holonomy , stable forms , superpotential . _
2010 mathematics subject classification : _
primary 53c25 , 53c29 ; secondary 53c44 , 53d20 , 83e15 , 83e30 . |
a bose - einstein condensate ( bec ) in an optical trap @xcite exhibits a rich variety of spin - related phenomena , such as various magnetic phases @xcite and spin domain formation @xcite .
the nuclear spin of @xmath2 and @xmath3 is @xmath4 , and combined with the spin @xmath5 of the outermost electron , the possible hyperfine spins of these atomic species are @xmath6 and @xmath7 .
the @xmath6 bec was first realized at mit with @xmath8 @xcite , for which the antiferromagnetic behavior was observed @xcite . on the other hand , @xmath2 in the @xmath6 hyperfine state was found to be ferromagnetic @xcite .
the @xmath7 @xmath3 condensate was also realized by the mit group @xcite , whose ground state phase at zero magnetic field is predicted to be antiferromagnetic @xcite .
however , the ground state phase and the spin dynamics of the @xmath7 @xmath3 bec have not been studied because of its very short lifetime ( a few milliseconds ) , due to the fact that the energy of the @xmath9 state is higher than that of the @xmath6 state .
the @xmath7 @xmath2 bec also lies energetically higher than its @xmath6 counterpart , but its lifetime is much larger ( @xmath10 ms ) due to a fortuitous coincidence of the singlet and triplet scattering lengths @xcite .
the coherent spin dynamics of the @xmath7 bec have been observed @xcite .
the ground state of the @xmath7 bec of @xmath2 is predicted to be very close to the phase boundary between the antiferromagnetic and cyclic phases @xcite . according to calculations performed by klausen _
et al_. @xcite , the ground state phase of this bec at zero magnetic field is barely in the antiferromagnetic phase .
recent experiments performed by the hamburg group @xcite and by the gakushuin group @xcite appear to support this prediction .
however , due to the experimental uncertainties , the possibility that the ground state phase is cyclic has not been excluded . in the hamburg experiment @xcite , the @xmath11 mixture was shown to be stable , which is the ground state of the antiferromagnetic phase . however
, a magnetic field of 340 mg was applied to the system in this experiment .
since the magnetic field lowers the energy of the @xmath11 states due to the quadratic zeeman effect , the observed stability of the @xmath11 mixture may be due to the presence of the magnetic field . that there is
no spatial separation between the @xmath11 states is a necessary condition for a bec to be antiferromagnetic , but it is not sufficient . in the gakushuin experiment
@xcite , the spin dynamics starting from the @xmath12 state were investigated for various values of the magnetic field .
it was found that the population of the @xmath11 components after 70 ms evolution increased with a decrease of the magnetic field , which might imply that the @xmath7 bec is antiferromagnetic at zero magnetic field . however , the initial @xmath12 state is a highly excited state and the resultant state after 70 ms is far above the ground state .
thus , the experimental observations do not exclude the possibility that the ground state phase at zero magnetic field is cyclic .
in fact , the hamburg group observed that the spin configuration of the cyclic phase has a long lifetime @xcite .
there are two experimental difficulties that hinder the determination of the ground state phase .
one is the short lifetime of @xmath13 100 ms of the upper hyperfine manifold of @xmath2 , which makes the equilibrium spin state hard to reach . in order to realize the ground state phase , we must equilibrate the system at least for a few seconds , as in the @xmath6 case @xcite .
the other difficulty is the small energy scale ( @xmath14 0.1 nk @xcite ) of the spin - singlet - pair term whose sign determines the ground state phase . in order to determine the sign of the spin - singlet pair energy , the magnetic field
must be reduced so that the quadratic zeeman energy is less than the spin - singlet pair energy . however , in such a low magnetic field the system may be affected by stray ac magnetic fields @xcite . to circumvent these difficulties ,
we propose a method to determine the value of the spin - singlet pair energy by spin exchange dynamics .
we show that by controlling the initial population and phase in each magnetic sublevel , we can measure the value of the singlet - pair energy from the initial spin dynamics .
this method therefore has the advantage that equilibrating the system , which takes a long time , is not necessary .
furthermore , the quadratic zeeman energy is allowed to exceed the singlet - pair energy and extreme suppression of stray magnetic fields is not required .
this paper is organized as follows .
section [ s : bogo ] formulates the mean field and bogoliubov theories of a spin-2 bec in the presence of the quadratic zeeman effect .
section [ s : homo ] studies the spin dynamics in a homogeneous system and proposes a method to determine the spin - singlet pair energy from the spin exchange dynamics .
section [ s : trap ] numerically verifies the proposed method for a trapped system with two - body loss .
section [ s : exp ] discusses the initial spin preparation and sec .
[ s : conc ] concludes the paper .
detailed derivations of the analytic solutions are given in the appendix .
from rotational symmetry , the low - energy interaction between two atoms with hyperfine spin @xmath15 can be classified in terms of the total spin @xmath16 , and each scattering channel can be described by the corresponding @xmath17-wave scattering length @xmath18 @xcite . in the case of @xmath7
there are three scattering channels @xmath19 , @xmath20 , @xmath21 , and the interaction is described by @xmath22 , where @xmath23 is the mass of the atom and @xmath24 projects the state of two colliding atoms into the state of total spin @xmath25 .
the interaction hamiltonian can be rewritten @xcite as @xmath26 , where @xmath27 and @xmath28 are the spin operators for atoms 1 and 2 and the interaction coefficients are given by @xmath29 for the linear and quadratic zeeman effects , the energy of the atom depends on the magnetic field @xmath30 as @xmath31 , where @xmath32 and @xmath33 with @xmath34 the bohr magneton and @xmath35 the hyperfine splitting between the states @xmath6 and @xmath7 @xcite .
we assume @xmath36 throughout this paper , which is the case for spin-2 @xmath2 .
the mean - field energy of the entire system is given by @xmath37 denotes the magnetic sublevels , @xmath38 is an external potential , @xmath39 is the total particle - number density , @xmath40 is the spin vector with @xmath41 spin-2 matrix @xmath42 , and @xmath43 is the singlet - pair amplitude . minimizing the action , @xmath44 gives the multicomponent gross - pitaevskii ( gp ) equations for the spin-2 bec as [ gp ] @xmath45 where @xmath46 .
we note that the linear zeeman term only rotates the hyperfine spin about the @xmath47 axis , and hence the spin dynamics are essentially independent of the linear zeeman term .
in fact , setting @xmath48 and noting that the phase factor also arises in the @xmath49 terms as @xmath50 , we can completely eliminate the linear zeeman terms in eqs .
( [ gp1 ] ) and ( [ gp2 ] ) .
this property is due to the rotational symmetry of our system with respect to the @xmath47 axis , which results in the conservation of the projected angular momentum on the @xmath47 axis , @xmath51 . in an ultracold spinor bec
which is isolated from the environment , the total spin angular momentum in the direction of the magnetic field is conserved for a long time ( @xmath52 1 sec ) @xcite .
we therefore minimize the energy of the system in the subspace of a given @xmath53 ; we shall refer to the resulting minimized state as the `` ground state '' . for simplicity ,
we restrict ourselves to the subspace of @xmath54 in this paper , i.e. , we seek the ground state in the subspace of @xmath55 .
we consider the uniform case by setting @xmath56 and @xmath57 , where the density @xmath58 is constant and @xmath59 satisfies @xmath60 .
the energy per atom is given by @xmath61 where @xmath62 , @xmath63 , and @xmath64 for @xmath65 , @xmath66 , @xmath20 .
the linear zeeman term is absent in eq .
( [ ene ] ) because @xmath54 . in order to find the ground state , we compare the energies of the stationary states given in ref . @xcite .
the energy of the antiferromagnetic state , @xmath67 where @xmath68 is the phase of each component , is given by @xmath69 this energy is independent of the phase factor @xmath70 . for the cyclic state ,
@xmath71 the energy of the system takes the form @xmath72 where @xmath73 . when @xmath74 , eq .
( [ en ] ) is minimized by @xmath75 , which corresponds to the antiferromagnetic state @xmath76 . when @xmath77 , eq .
( [ en ] ) becomes minimal with @xmath78 and @xmath79 substituting this into eq .
( [ en ] ) , we obtain @xmath80 the states @xmath76 and @xmath81 satisfy @xmath82 .
even if the local spin @xmath83 is nonzero , the total spin @xmath84 can be zero by the formation of the spatial spin structure @xcite .
for example , a staggered domain structure of the ferromagnetic state gives @xmath55 .
the kinetic energy at the domain wall can be neglected when the region of the wall is much smaller than the region of the domain .
hence , we also consider the case of @xmath85 , bearing in mind that the total spin is maintained as zero , @xmath55 , by some structure formation .
the energy of the ferromagnetic state , @xmath86 or @xmath87 , is given by @xmath88 we also consider a state found in ref .
@xcite , @xmath89 the energy of this state is obtained to be @xmath90 this is minimized by @xmath91 for @xmath92 , and the energy then becomes @xmath93 comparing the energies @xmath94 , @xmath95 , @xmath96 , and @xmath97 , we obtain the phase diagrams shown in fig . [
f : diagram ] . and @xmath36 ( a ) with respect to @xmath98 and @xmath99 and ( b ) with respect to @xmath100 and @xmath101 .
the symbols af , c , f , and m indicate the states @xmath76 , @xmath81 , @xmath102 , and @xmath103 , respectively . in ( b )
, @xmath104 is assumed .
, width=317 ] we have assumed here that one of the states @xmath76 , @xmath81 , @xmath102 , and @xmath103 is the ground state , based on the numerical results of ref .
we have confirmed that this assumption is correct using the monte carlo method . in the case of @xmath2
, @xmath105 is positive and @xmath106 is at most of the same order of magnitude as @xmath105 @xcite .
hence , the ground state of the spin-2 @xmath2 bec is antiferromagnetic or cyclic depending on whether @xmath74 or @xmath77
. therefore , in order to determine the sign of @xmath107 from the ground state , we must suppress the magnetic field so that the condition @xmath108 is met . in the experiment of the hamburg group @xcite ,
a magnetic field of @xmath109 mg was applied , which corresponds to @xmath110 nk .
the mean atomic density was @xmath111 , and if @xmath112 , we have @xmath113 nk .
thus , @xmath114 , which is consistent with the observed stability of the antiferromagnetic state @xmath76 .
in the gakushuin experiment @xcite , @xmath115 mg and @xmath116 , corresponding to @xmath117 nk and @xmath118 nk .
hence , the reported parameters of ref . @xcite belong to the cyclic phase .
the antiferromagnetic ( or stretched ) state , @xmath119 was found in experiments to be stable in the presence of a magnetic field @xcite . to examine this result
, we will investigate the stability of this state using bogoliubov analysis .
substituting eq .
( [ polar ] ) into eqs .
( [ gp ] ) , we find that it evolves as @xmath120 with @xmath121 , and hence eq .
( [ polar ] ) is stationary in the rotating frame @xmath122 .
we can therefore perform bogoliubov analysis with respect to the state ( [ polar ] ) in the rotating frame , which is described by the gp equations ( [ gp ] ) without the linear zeeman terms .
we set the wave function as @xmath123 and substitute it into the gp equations in the rotating frame .
taking the linear terms with respect to @xmath124 , we obtain [ bogo ] @xmath125 these equations can be reduced to the eigenvalue problem by expansion of @xmath124 as @xmath126 and the eigenenergies are obtained as [ eigen ] @xmath127^{1/2 } , \label{mode1 } \\ & & \left [ \varepsilon_k \left ( \varepsilon_k + 8 \tilde c_1 - 2 \tilde c_2 / 5 \right ) \right]^{1/2 } , \label{mode2 } \\ & & \left [ ( \varepsilon_k - 3q ) \left ( \varepsilon_k - 3q + 2 \tilde c_1 - 2 \tilde c_2 / 5 \right ) \right]^{1/2 } , \label{mode3 } \\ & & \left [ ( \varepsilon_k - 4q ) \left ( \varepsilon_k - 4q - 2 \tilde c_2 / 5 \right ) \right]^{1/2 } , \label{mode4}\end{aligned}\ ] ] where @xmath128 . if the eigenenergy is complex , the corresponding mode is dynamically unstable against exponential growth . the eigenvectors @xmath129 and @xmath130 of the first mode ( [ mode1 ] ) are proportional to @xmath131 , which is the same form as @xmath132 , indicating that a collapse occurs if @xmath133 .
the eigenvectors of the second mode ( [ mode2 ] ) are proportional to @xmath134 . since this eigenmode transfers the @xmath11 component to the @xmath135 component , excitation of the mode ( [ mode2 ] ) gives rise to exchange of atoms between the @xmath11 components .
this implies that phase separation between the two components occurs for @xmath136 .
therefore , the fact that no phase separation is observed in experiments @xcite only indicates that @xmath137 ; it is not sufficient to conclude that @xmath138 .
the third mode ( [ mode3 ] ) is two - fold degenerate and the eigenvectors are proportional to @xmath139 and @xmath140 .
therefore , when @xmath141 , the state ( [ polar ] ) is dynamically unstable against the growth of the @xmath142 components .
similarly the last mode ( [ mode4 ] ) is proportional to @xmath143 and describes the growth of the @xmath12 component .
dynamical instability arises in this mode for @xmath77 , consistent with the fact that the ground state has the @xmath12 component for @xmath77 , as shown by eq .
( [ cyclic ] ) . for later use
, we also perform bogoliubov analysis for the stationary state @xmath144 .
the eigenenergies are obtained to be [ eigen2 ] @xmath127^{1/2 } , \label{mode2_1 } \\ & & \left [ ( \varepsilon_k + q ) \left ( \varepsilon_k + q + 6 \tilde c_1 - 2 \tilde c_2 / 5 \right ) \right]^{1/2 } , \label{mode2_2 } \\ & & \left [ ( \varepsilon_k + 4q ) \left ( \varepsilon_k + 4q - 2 \tilde c_2 / 5 \right ) \right]^{1/2}. \label{mode2_3}\end{aligned}\ ] ] the eigenvector of the first mode ( [ mode2_1 ] ) is proportional to @xmath145 , implying that the state @xmath146 collapses if @xmath147 .
we note that this condition for the collapse of the bec is the same as that for @xmath132 .
the second and third modes , eqs .
( [ mode2_2 ] ) and ( [ mode2_3 ] ) , are both two - fold degenerate .
the eigenvectors of the mode ( [ mode2_2 ] ) are proportional to @xmath148 and @xmath140 , and therefore the dynamical instability in this mode increases the @xmath149 components .
similarly , the mode ( [ mode2_3 ] ) has eigenvectors proportional to @xmath150 and @xmath151 .
in contrast to the excitations for @xmath132 , there are always dynamically unstable modes in eqs .
( [ mode2_2 ] ) and ( [ mode2_3 ] ) because of @xmath36 , and therefore the @xmath12 state is always dynamically unstable for @xmath152 in a homogeneous system .
in order to determine the value of @xmath107 from the spin dynamics , we consider the spin dependent interaction that is sensitive only to the spin - singlet term .
the elementary processes in the spin - singlet channel are @xmath153 , @xmath154 , @xmath155 , and @xmath156 . on the other hand ,
the @xmath105 term consists of elementary processes such as @xmath157 .
the spin independent interaction does not flip the spin .
thus , the elementary process appearing only in the @xmath107 term is @xmath155 .
we therefore focus on this process and consider the initial state of the form @xmath158 since the overall phase and spin rotation about the @xmath47 axis do not affect the physics , the relevant phase appears only in the combination of @xmath159 .
we therefore assume that @xmath160 and @xmath161 without loss of generality .
if @xmath162 are exactly zero in the initial state , we find from eq .
( [ gp2 ] ) that @xmath163 are always zero within the mean field approximation .
we therefore assume @xmath164 in the following analysis . in experiments
@xcite , this assumption holds at least for an initial period of @xmath10 ms .
although the @xmath149 components exponentially grow in the presence of the dynamical instability , the assumption @xmath164 is still valid in the early stage of time evolution .
the gp equations then reduce to [ gp02 ] @xmath165 differentiating eq .
( [ gp02_0 ] ) with respect to time and using eq .
( [ gp02_1 ] ) , we can eliminate @xmath166 to obtain @xmath167 where the energy per atom , @xmath168 is a constant of motion .
it is interesting to note that the form of eq .
( [ eqzeta0 ] ) coincides with that describing a 1d bec on a rotating ring if the time derivative is replaced with the spatial derivative @xcite .
the solution is obtained ( see appendix for derivation ) as @xmath169 where @xmath170 , @xmath171 , @xmath172 , @xmath173 , @xmath174 , and @xmath175 are given by eqs .
( [ phi0 ] ) , ( [ a0org ] ) , ( [ eqb0 ] ) , ( [ alphorg ] ) , ( [ beta0org ] ) , and ( [ nuorg ] ) , respectively . the function sn is a jacobian elliptic function @xcite .
if we assume @xmath176 in the initial state , i.e. , @xmath177 the constants in the solution ( [ sol1 ] ) become @xmath178 , @xmath179 , and @xmath180 , \label{a0 } \\
\alpha & = & \sqrt{\frac{8q}{5 \hbar^2 } \left\ { \tilde c_2 \left [ 1 - 2 \zeta_0 ^ 2(0 ) \right ] + 10 q \right\ } } , \label{alpha } \\
\nu & = & \frac{\tilde c_2 ^ 2 \zeta_0 ^ 2(0 ) \left [ 1 - \zeta_0 ^ 2(0 ) \right]}{10 q \left\ { \tilde c_2 \left [ 1 - 2 \zeta_0 ^ 2(0 ) \right ] + 10 q \right\}}. \label{m}\end{aligned}\ ] ] the sign of @xmath171 is opposite to that of @xmath107 because @xmath36 .
the constants @xmath173 and @xmath175 are real and positive if @xmath181 + 10 q < 0 $ ] .
we suppose that the initial state ( [ init ] ) is prepared and the magnetic field satisfies @xmath182 . under this condition , we find that @xmath173 and @xmath175 in eqs .
( [ alpha ] ) and ( [ m ] ) are real and positive .
since the atomic density is typically @xmath183 in experiments @xcite and @xmath184 is of order of the bohr radius at most @xcite , the condition @xmath182 requires @xmath185 mg . at this magnetic field
, the ground state in the subspace of @xmath54 is the stretched state ( [ polar ] ) irrespective of the sign of @xmath107 .
the jacobian elliptic function @xmath186 initially increases for positive @xmath173 and @xmath175 , and hence the @xmath12 component @xmath187 in eq .
( [ sol1 ] ) initially decreases for @xmath188 and increases for @xmath189 .
therefore , we can determine the sign of @xmath107 from the sign of the initial change in the spin population , which is the main result of this paper .
this method is suitable for atomic species with short lifetimes , since only the initial stage of time evolution is needed for the determination of @xmath107 . figure [ f : unifev ] shows time evolution of the spin populations which are obtained by numerically solving eq .
( [ gp02 ] ) .
@xcite and a magnetic field of 250 mg .
the scattering lengths are assumed to be @xmath190 in ( a ) and @xmath191 in ( b ) , where @xmath192 is the bohr radius .
the initial state is chosen to be @xmath193 , @xmath194 , and @xmath195 . the dashed curves are drawn according to eq .
( [ sin ] ) .
, width=317 ] the results confirm the relation between the sign of @xmath107 and the sign of the initial change in the spin populations . in fig .
[ f : unifev ] , the @xmath149 components are assumed to have small initial values ( @xmath195 ) to assess imperfect population transfer in realistic situations .
we find that the dynamical instability in the @xmath149 components have little effect on the initial dynamics of the @xmath12 , @xmath196 components . in fig .
[ f : unifev ] , @xmath197 . for @xmath198
, the jacobian elliptic function can be approximated by the trigonometric function as @xcite @xmath199 and the spin evolution is then approximated by [ sin ] @xmath200 the amplitude of this oscillation is @xmath201 , showing explicitly that whether the spin population initially increases or decreases depends on the sign of @xmath107 .
we can also determine the magnitude of @xmath107 by measuring the amplitude of the oscillations . in fig .
[ f : unifev ] , eq .
( [ sin ] ) is plotted as dashed curves , which are in good agreement with the numerical solutions ( solid curves ) in the early stages of the time evolution . the differences between the numerical and analytic curves in fig .
[ f : unifev ] arise mainly from the growth of the @xmath149 components in the numerical result due to the dynamical instability .
we find that the growth of the @xmath149 components in fig .
[ f : unifev ] ( b ) is larger than that in fig .
[ f : unifev ] ( a ) .
this can qualitatively be understood from the bogoliubov energy ( [ mode2_2 ] ) , which describes the growth of the @xmath149 components from the state @xmath144 .
the imaginary part of eq .
( [ mode2_2 ] ) at @xmath202 , @xmath203^{1/2}$ ] , is larger for @xmath189 than for @xmath188 since @xmath36 and @xmath204 in the present case .
we note that the relative phase @xmath205 in the initial state ( [ initz ] ) is crucial for determining the value of @xmath107 by the above method .
for example , when the initial state is given by eq .
( [ initz ] ) with @xmath206 , or without loss of generality , @xmath207 the time evolution becomes eq .
( [ sol1 ] ) , where @xmath171 , @xmath175 , and @xmath173 are given by eqs .
( [ samea0 ] ) , ( [ samenu ] ) , and ( [ samealph ] ) , respectively ( see appendix for details ) .
these constants satisfy @xmath208 and @xmath209 , and @xmath173 is real . using the relation @xcite @xmath210 and noting that the jacobian elliptic function @xmath211 initially increases
, we find that @xmath187 initially decreases for @xmath189 and increases for @xmath188 .
this behavior is opposite to that in fig .
[ f : unifev ] and therefore control of the relative phase @xmath212 in the initial state is crucial for determining the sign of @xmath107 .
figure [ f : phase ] shows the initial evolution of @xmath213 with @xmath212 varying from @xmath0 to @xmath214 , which indicates that the phase of the oscillation depends on @xmath212 . with the same parameters as in figs . [
f : unifev ] ( a ) and ( b ) . the initial state is given by @xmath193 , @xmath215 , @xmath216 , and @xmath217 , where @xmath218 is an integer ranging from 0 to 4 .
, width=317 ] if @xmath219 , the stretched state ( [ polar ] ) is dynamically unstable against growth of the @xmath12 component , as shown by eq .
( [ mode4 ] ) .
this is because the ground state is in the cyclic phase given by eq .
( [ zeta ] ) if @xmath219 and @xmath137 .
we can therefore conclude that @xmath220 and hence @xmath221 , if we observe the growth of the @xmath12 component from the initial stretched state .
suppose that @xmath222 in the initial state given by eq .
( [ init ] ) and @xmath223 , then @xmath175 is negative and @xmath173 is pure imaginary from eqs .
( [ alpha ] ) and ( [ m ] ) . in this case , the jacobian elliptic function can be rewritten as @xmath224 since @xmath225 and hence @xmath226 , we can make the approximation @xcite @xmath227 where @xmath228 . when @xmath229 , eq .
( [ sdeq ] ) reduces to @xmath230 equation ( [ sol1 ] ) is thus approximated to give @xmath231 the dashed curve in fig .
[ f : cyclic ] shows eq .
( [ tanh ] ) .
the initial state is given by @xmath232 , @xmath233 , @xmath234 .
the dashed curve is drawn according to eq .
( [ tanh ] ) . ,
width=317 ] a comparison of this with the corresponding numerical result ( solid curve ) shows that eq .
( [ tanh ] ) is in good agreement with the numerical result , with a small discrepancy arising mainly from the growth of the @xmath235 components in the numerical result due to the dynamical instability .
the @xmath12 component reaches a maximum at @xmath236 .
we note that a characteristic time scale of the dynamics in fig .
[ f : cyclic ] is @xmath10 ms , which is much larger than the @xmath237 ms seen in fig .
[ f : unifev ] .
this is because the former time scale is set by the dynamical instability , while the latter is set by the time evolution from a non - stationary state .
this method can thus determine the value of @xmath107 , provided that ( i ) @xmath107 is positive , ( ii ) the magnetic field is suppressed so that @xmath238 , and ( iii ) the time scale of the dynamics is much faster than the lifetime of the @xmath7 state .
if @xmath107 is negative , the present method can not determine the value of @xmath107 . in the case of fig .
[ f : cyclic ] ( @xmath219 ) , the @xmath239 population initially increases for @xmath188 , which is opposite to the case of fig . [
f : unifev ] ( @xmath182 ) . nevertheless , the method to determine the sign of @xmath107 in sec .
[ ss : main ] is still applicable , since we can easily find whether we are in a region of @xmath240 or @xmath219 .
if the amplitude of the initial oscillation monotonically decreases without changing sign with an increase in the magnetic field , we can conclude that @xmath240 .
we investigate the spin dynamics in a trap potential to verify the results in sec .
[ s : homo ] in a realistic situation .
we use an axisymmetric trap potential with radial and axial frequencies @xmath241 hz , which was used in the experiment of ref .
the two - body loss rate was found to be roughly independent of the magnetic field and therefore it is considered to be insensitive to the spin - mixing dynamics @xcite .
we therefore take into account the two - body loss by adding the term @xmath242 to the right - hand side of eq .
( [ gp ] ) with @xmath243 .
we prepare the ground state @xmath244 for the @xmath245 state by the imaginary - time propagating method @xcite and transfer it to each spin component as @xmath246 , @xmath247 , and @xmath248 .
we use this state as an initial state , where the initial number of atoms is assumed to be @xmath249 @xcite . the time evolution of each spin population obtained by numerically solving the gp equation ( [ gp ] ) with the two - body loss terms is shown in fig . [ f : trap ] . of each spin component @xmath250 in the trap with @xmath251 hz and @xmath252 hz @xcite .
the scattering lengths are assumed to be @xmath253 in ( a ) and @xmath254 in ( b ) , where @xmath192 is the bohr radius .
the strength of the magnetic field is 250 mg , and the two - body loss coefficient is @xmath255 @xcite . the @xmath245 ground state @xmath244 is prepared , and is initially transferred to each spin component as @xmath246 , @xmath256 , and @xmath257 .
the insets show the column density distributions of the @xmath258 components from top to bottom at @xmath259 ms , each having a @xmath260 field of view .
, width=317 ] we find that the aforementioned dependence of the initial spin evolution on the sign of @xmath107 is valid in the trapped system , i.e. , the quantity @xmath261 first decreases for @xmath188 and increases for @xmath189 .
the growth in @xmath262 due to the dynamical instability and the two - body loss does not affect this dependence .
thus , we have confirmed that the sign of @xmath107 can be determined from the spin dynamics in a realistic situation .
we find that the growth of the @xmath149 components is larger for @xmath263 than for @xmath188 , which is similar to the homogeneous case shown in fig .
[ f : unifev ] . the insets in fig .
[ f : trap ] show the column density distribution of each spin component at @xmath259 ms .
the initial distribution is almost preserved at @xmath259 ms and no spin domains are formed .
as shown in fig . [ f : phase ] , our method to determine the sign of @xmath107 hinges on the relative phase @xmath264 in the initial state ( [ initz ] ) .
therefore we must control the phase experimentally . in the experiments of refs .
@xcite , the rapid - adiabatic - passage technique @xcite was employed to prepare the initial state .
this method can control the phase in the initial state in principle , if the relative phases between the applied rf fields are controlled .
we propose here a new method to prepare the initial state using the raman transition .
we consider the raman transition , for example , between the @xmath7 and @xmath265 @xmath266 states by circularly polarized photons , as illustrated in fig .
[ f : level ] . and @xmath267 @xmath266 states .
the circularly polarized light @xmath268 changes @xmath250 by @xmath269 and hence starting from the @xmath7 , @xmath12 state , the system undergoes transitions only to @xmath267 , @xmath149 and @xmath7 , @xmath11 states with the clebsch - gordan coefficients given by @xmath270 and @xmath271 , respectively . , width=317 ] the initial state is assumed to be the @xmath12 lower state .
hence , the relevant states are the @xmath12 , @xmath196 lower states and the @xmath149 upper states , whose amplitudes are denoted by @xmath272 , @xmath166 , and @xmath273 , respectively .
the equations of motion are given by [ raman1 ] @xmath274 where @xmath275 and @xmath276 are the coupling constants and frequencies of the @xmath268 photons , @xmath277 is the transition energy of the @xmath266 line , and @xmath278 and @xmath279 are the linear and quadratic zeeman energies for the upper hyperfine manifold .
the coupling constants @xmath275 are given by the product of the amplitude of the circularly polarized light and the dipole matrix element between the @xmath17 and @xmath280 states .
using the new variables @xmath281 and @xmath282 , the coefficients on the right - hand side of eq .
( [ raman1 ] ) become time independent . assuming @xmath283 and @xmath284 , and eliminating @xmath273 from eq .
( [ raman1 ] ) , we obtain [ raman2 ] @xmath285 where @xmath286 .
if the amplitudes of the @xmath268 fields are the same , i.e. , @xmath287 is met , the coefficients of the first terms on the right - hand side of eq .
( [ raman2 ] ) become equal , i.e. , @xmath288 . the solutions
are thus obtained as @xmath289 , which is independent of the phase of the circularly polarized light and the applied magnetic field .
this method is thus suitable for preparing the initial state ( [ init ] ) .
we can check whether the prepared state has the correct relative phase @xmath290 .
rotating the spin state ( [ initz ] ) by @xmath291 around the axis on the @xmath292-@xmath293 plane as @xmath294 with @xmath295 , we find that the population of each component becomes @xmath296 where @xmath297 .
let us consider the quantity @xmath298 which can be measured by the spin - dependent non - destructive imaging reported in ref .
we note that if @xmath290 , eq . ( [ fz2 ] ) is independent of @xmath299 and is constant , while if @xmath300 , eq . ( [ fz2 ] ) oscillates , since the phase @xmath301 rotates in the magnetic field .
we can therefore confirm that the desired initial state ( [ init ] ) is obtained if the result of the polarization - dependent phase - contrast imaging @xcite is constant in time .
we have proposed a method to experimentally determine the sign of the singlet - pair energy of a spin-2 bec from spin exchange dynamics , which determines whether the @xmath7 @xmath2 bec at zero magnetic field is antiferromagnetic or cyclic .
we have obtained analytic solutions of the multicomponent gp equations and found that if we prepare the initial state in the magnetic sublevels @xmath12 , @xmath196 with the appropriate relative phase relationship , we can determine the sign of @xmath107 from the sign of the initial change in the spin populations .
for example , in the case of the initial state in eq .
( [ initz ] ) , we can conclude that @xmath107 is positive or negative if the @xmath12 component first decreases or increases .
we can also determine the magnitude of @xmath107 from the amplitude of the oscillation in the spin populations .
since only the initial evolution for a period of @xmath302 ms is needed , we can use this method in the presence of atom loss by inelastic collisions and even in the presence of dynamical instabilities .
we have numerically confirmed that this method is applicable for a trapped system with a realistic two - body loss coefficient @xmath303 .
the required condition for the magnetic field is @xmath240 , so our method works even when the quadratic zeeman energy exceeds the singlet - pair energy . the initial spin state in which the populations and relative phases are controlled as eq .
( [ initz ] ) can be prepared by , e.g. , the raman technique as shown in sec .
[ s : exp ] .
we have shown that the coherent spin dynamics serve as an efficient probe for investigating spinor properties especially when the atomic species have a short lifetime and therefore the equilibrium state can not be reached .
the coherent spin dynamics may also reveal other equilibrium spinor properties as well as nonequilibrium spinor properties that can not be obtained from the static equilibrium .
we would like to thank t. kuwamoto and t. hirano for fruitful discussions .
this work was supported by grant - in - aids for scientific research ( grant no .
17740263 and no .
15340129 ) and by a 21st century coe program at tokyo tech `` nanometer - scale quantum physics , '' from the ministry of education , culture , sports , science and technology of japan .
in this appendix , we solve eq .
( [ gp02 ] ) for the initial condition ( [ initz ] ) with @xmath160 and @xmath304 , and derive the solution ( [ sol1 ] ) . defining a new variable @xmath305 we can rewrite eq .
( [ eqzeta0 ] ) as @xmath306 where @xmath307 with the spin - dependent energy @xmath308 we write the complex variable as @xmath312 with real functions @xmath313 and @xmath314 , where @xmath315 and @xmath316 . substituting this into eq .
( [ eqz0 ] ) and taking the real and imaginary parts , we obtain @xmath317 these equations can be integrated to give @xmath318 where @xmath319 and @xmath320 are constants of integration .
the time derivative of @xmath309 at @xmath321 , @xmath322 , gives @xmath323 and @xmath324 .
the constant @xmath320 is then written as @xmath325 the constant @xmath319 is similarly obtained as @xmath326 we introduce a new variable @xmath327 through @xmath328 where @xmath171 and @xmath172 are constants . substituting eq .
( [ rho ] ) into eq .
( [ rdot2 ] ) , we obtain @xmath329 we determine @xmath172 so that the @xmath330 term in eq
. ( [ rhoeq ] ) vanishes : @xmath331 the right - hand side ( rhs ) of eq .
( [ rhoeq ] ) then becomes quartic with respect to @xmath327 .
we determine @xmath171 so that the rhs of eq .
( [ rhoeq ] ) becomes proportional to @xmath332 ; the result is @xmath333 \nonumber \\\end{aligned}\ ] ] with @xmath334 equation ( [ rhoeq ] ) thus becomes @xmath335 this equation can be integrated to give @xmath336 we rewrite this equation as @xmath337 where @xmath338 the left - hand side of eq .
( [ intrho ] ) is an elliptic integral of the first kind and hence @xmath339 @xcite .
thus , @xmath340 from eqs .
( [ z0 ] ) and ( [ phidot ] ) , the phase @xmath341 of @xmath272 is given by @xmath342 similarly , we can obtain the solution for @xmath343 . from the form of eq .
( [ gp02 ] ) and the initial condition @xmath344 , we find that the relation @xmath345 always holds .
hence , we consider only @xmath346 .
eliminating @xmath347 from eq .
( [ gp02 ] ) , we have the equation of motion for @xmath346 as @xmath348 \zeta_2 - i \frac{2 \tilde c_2}{5 } \hbar \dot{\zeta}_2.\ ] ] the new variable @xmath349 reduces eq .
( [ ddotz2 ] ) to @xmath350 where @xmath351 and @xmath352 .
the initial conditions for @xmath353 are @xmath354 and @xmath355 - \frac{4 i q}{\hbar } \zeta_2(0).\ ] ] equation ( [ z2eq ] ) has the same form as eq .
( [ eqz0 ] ) , in which @xmath356 and @xmath357 are replaced by @xmath358 and @xmath359 , and therefore the derivation from eq .
( [ rphi1 ] ) to eq .
( [ phi0 ] ) can similarly be applied .
we consider the case of @xmath176 in the initial condition given in eq .
( [ initz ] ) . in this case , eq .
( [ zdot0 ] ) is pure imaginary , and eq .
( [ eqb0 ] ) can be solved to give @xmath360 .
equation ( [ a0org ] ) then reduces to @xmath361 , \end{array } \right.\ ] ] where the upper and lower expressions correspond to the plus and minus signs of @xmath362 in eq .
( [ a0org ] ) . using this @xmath171 ,
@xmath175 and @xmath173 become @xmath363 + 10 q \right\}}{\displaystyle \tilde c_2 ^ 2 \zeta_0 ^ 2(0 ) \left [ 1 - \zeta_0 ^ 2(0 ) \right ] } , \\
\displaystyle \frac{\tilde c_2 ^ 2 \zeta_0 ^ 2(0 ) \left [ 1 - \zeta_0 ^ 2(0 ) \right]}{\displaystyle 10 q \left\ { \tilde c_2 \left [ 1 - 2 \zeta_0 ^ 2(0 ) \right ] + 10 q \right\ } } , \end{array } \right .
\\ \label{appalph } \alpha & = & \left\ { \begin{array}{l } \displaystyle \frac{2 |\tilde c_2|}{5 \hbar } \sqrt{\zeta_0 ^ 2(0 ) \left [ 1 - \zeta_0 ^ 2(0 ) \right ] } , \\
\displaystyle \sqrt{\frac{8q}{5 \hbar^2 } \left\
{ \tilde c_2 \left [ 1 - 2 \zeta_0 ^ 2(0 ) \right ] + 10 q \right\}}. \end{array } \right.\end{aligned}\ ] ] it follows from eq .
( [ rho ] ) and @xmath364 that @xmath365 and hence @xmath178 . using the relation @xmath366 we can show that the two expressions in eqs .
( [ appa0 ] ) , ( [ appnu ] ) , and ( [ appalph ] ) are equivalent to each other .
we used the lower expressions in sec .
[ s : homo ] .
the @xmath11 components @xmath367 have the same form as eq .
( [ zeta0 ] ) , in which @xmath171 , @xmath174 , and @xmath172 are replaced by @xmath368 , @xmath369 , and @xmath172 .
these constants are given by @xmath370 , \end{array } \right.\ ] ] @xmath371 , and @xmath372 .
finally , we consider the case of @xmath373 in the initial condition ( [ initz ] ) , i.e. , all the components are real and positive
. we can solve eq .
( [ eqb0 ] ) also in this case to give @xmath364 and hence @xmath178 .
the constants @xmath171 , @xmath175 , and @xmath173 reduce to @xmath374 + 10 q \right\ } - |q| \sqrt{(\tilde c_2 + 10 q)^2 - 40 \tilde c_2 q \zeta_0 ^ 2(0 ) } \right\ } , % % % \nonumber \\ \\ \label{samenu } \nu & = & \frac{q \left\ { \tilde c_2 [ 1 - 2 \zeta_0 ^ 2(0 ) ] + 10 q \right\ } - |q| \sqrt{(\tilde c_2 + 10 q)^2 - 40 \tilde c_2 q \zeta_0 ^ 2(0)}}{q \left\ { \tilde c_2 [ 1 - 2 \zeta_0 ^ 2(0 ) ] + 10 q \right\ } + |q| \sqrt{(\tilde c_2 + 10 q)^2 - 40 \tilde c_2 q \zeta_0 ^ 2(0 ) } } ,
\\ \label{samealph } \alpha & = & \sqrt{\frac{8 \tilde c_2
q a_0}{5 \hbar^2 \nu}},\end{aligned}\ ] ] where we take the lower sign in eq .
( [ a0org ] ) .
since @xmath375 + 10 q \right\}^2 < ( \tilde c_2 + 10 q)^2 - 40 \tilde c_2 q \zeta_0 ^ 2(0)$ ] , we find @xmath376 and @xmath209 . | we propose a method to determine the singlet - pair energy of a spin-2 bose - einstein condensate ( bec ) . by preparing the initial populations in the magnetic sublevels @xmath0 and @xmath1 with appropriate relative phases , we can obtain the coefficient of the spin singlet - pair term from the spin exchange dynamics .
this method is suitable for hyperfine states with short lifetimes , since only the initial change in the population of each magnetic sublevel is needed .
this method therefore enables the determination of the ground state phase of a spin-2 @xmath2 bec at zero magnetic field , which is considered to lie in the immediate vicinity of the boundary between the antiferromagnetic and cyclic phases .
we also show that the initial state in which relative phases are controlled can be prepared by raman processes . |
the international linear collider ( ilc ) @xcite is an electron - positron collider with a centre - of - mass energy up to 1 tev and a luminosity goal of @xmath0 .
the achievement of this goal is challenging , and requires nanometre - sized colliding beams at the interaction point , achieved through emittance reductions in dedicated damping rings and strong focusing before the interaction point .
this , combined with the high bunch charge leads to intense electromagnetic interactions when the beams collide , causing particles of each bunch to be deflected by the space charge of the other .
such deflections strongly enhance the angular divergence of the outgoing beams and significantly dilute their energy distributions , through radiation of so - called beamstrahlung photons .
re - interactions involving these photons also produce additional secondary charged particles .
the production of beamstrahlung , whilst exploitable as a diagnostic , is generally troublesome to the design of the interaction region and subsequent extraction line .
the photons are produced in the direction of the outgoing beam , with considerable power ( in the order of mws ) , in a cone of some opening angle in the forward direction ( typically up to a mrad ) , and must be dumped in a controlled way .
they may irradiate parts of the machine , potentially quench superconducting magnets and produce backgrounds in the interaction region detectors , through backscattering .
hence a detailed knowledge and characterisation of the beamstrahlung production is critical to successful linear collider design .
a further contribution to the photon spectrum arises from the so - called compton photons .
these correspond to quantum interactions between electrons and positrons of the colliding beams , where the accompanying photon radiation has high energy .
the dominant contribution is from the radiative bhabha process , in the particular kinematic domain involving the exchange of a virtual photon only slightly off mass - shell .
the hard part of this process can be described as compton scattering of a beam particle on this quasi - real photon , and mostly factorises from the part corresponding to its emission off the electron or positron on the other side .
the total power corresponding to these additional photons is however only a few tens of watts and they are radiated with angles comparable to those of the beamstrahlung photons . in this report ,
comprehensive calculations of the beamstrahlung and compton photon production for the newest ilc beam parameter sets are presented .
the resulting photon angular distributions are obtained and a set of cone half - opening angles to represent successive total beamstrahlung powers are presented for each ilc parameter set . in section [ seccalcs ] the computational methods and the ilc parameter sets under consideration are discussed . in section [ sectheory ]
approximate analytical expectations of the photon cones are calculated .
they are compared to computational results in section [ secresults ] .
conclusions are presented in section [ secconc ] , including the practical beam stay - clear requirements in terms of cone half - opening angles .
in this section , the current ilc beam parameter sets , the key parameters relevant to beamstrahlung production and the method used to simulate the beamstrahlung and compton photons are described .
the latest ilc beam parameter sets @xcite are similar to the previous working space for the 500 gev machine whilst reducing the magnitude of the beam - beam interaction for the 1 tev case , resulting in smaller mean particle energy losses to beamstrahlung photons , @xmath1 , at that energy .
the parameters for the new working space relevant to this work are shown in table [ tabparasets ] , where e , n@xmath2 , n , @xmath3 , p and @xmath4 are the centre - of - mass energy , the number of bunches per train , the number of particles per bunch , the luminosity , the total power contained in the beamstrahlung photons and the r.m.s bunch sizes in the horizontal , vertical and longitudinal dimensions , respectively
. ._key parameters of the new beam parameter sets used in this work , at 500 gev and 1 tev in the centre - of - mass . _ [ cols="<,^,^,^,^,^,^,^ " , ] [ tabprimanglesextra ] since they use the rigid beam model , these calculations are only valid in the horizontal plane and for cases with small disruption .
comparing tables [ tab500beamcone ] and [ tabtevbeamcone ] indicates that the calculations predict the maximum horizontal angle of photon emissions reasonably well only for some of the parameter sets .
for example , the 500 gev nominal theoretical maximum angle is 0.376 mrad , which agrees well with the 0.35 - 0.39 mrad range computed from the beam - beam simulation for the 100 - 1 w criteria .
this is however not the case for the large y and high luminosity sets , where some differences are seen .
such differences are attributed to extreme colliding particle trajectories in guinea - pig .
these may originate from the input angular distribution of the beams and may also result from some disruption occuring for these parameter sets . in all cases ,
the guinea - pig calculation gives larger values and should be used for conservative results .
similar distributions for the compton photons are presented in figures [ fig500compt ] and [ figtevcompt ] , in the horizontal and vertical planes and for the five parameter sets . while extending to similar angles as for beamstrahlung photons , the corresponding power is orders of magnitude less .
the maximum cone half - opening angles previously computed from beamstrahlung are hence not modified .
in this paper , the angles of photons emitted through the beam - beam interaction at the international linear collider have been computed using both the rigid beam model and the beam - beam simulator guinea - pig .
these calculations , and the subsequent understanding of the beamstrahlung , are crucial to the effective design of the interaction region and extraction line .
the rigid beam theoretical predictions tend to underestimate the photon cone opening angles for some of the beam parameter sets , particularly for cases with large disruption .
guinea - pig is hence used for conservative results .
cones of excluded photon power are defined corresponding to tolerances relevant to different aspects of beam - line design : 100 w for general activation and 10 - 1 w for super - conductive magnet quenching .
two contributions were studied , from beamstrahlung emission and from a compton process involving the exchange of a slightly virtual photon between the incident electrons and positrons . by
far the first contribution dominates for most of the angles .
the results are summarised in table [ tabprimangles ] . for the general design of the extraction line , the 100 w definition is applicable .
in this case , half - opening angles of 0.75 and 0.85 mrad can be used as conservative definitions of the required beam - stay - clear in the horizontal and vertical planes , respectively .
[ secconc ]
we would like to thank daniel schulte , olivier napoly , olivier dadoun , mark briscombe , daniel smallqy and ccile rimbault for helpful advice and assistance , and special thanks to deepa angal - kalinin for checking the manuscript and useful advice .
this work was supported by the commission of the european communities under the 6th framework programme `` structuring the european research area '' , contract number rids-011899 , and by the alliance programme of the british council and ministre des affaires trangres .
99 the ilc bcd , http://www.linearcollider.org/wiki/doku.php?id=bcd:bcd_home the ilc reference design report rdr , http://media.linearcollider.org/report-apr03-part1.pdf and new beam parameters circulated by t.raubenheimer to the rdr leaders http://flc.web.lal.in2p3.fr/mdi/bbsim/bbsim.html http://flc-mdi.lal.in2p3.fr/spip.php?rubrique17 r. appleby and p. bambade , care / elan document 2004 - 21 ( 2004 ) e. keil , cern sl/94 - 78 ( ap ) r. appleby and p. bambade , jinst.1:o10004 ( 2006 ) | the intense beam - beam effect at the interaction point of the international linear collider ( ilc ) causes large disruption of the beams and the production of photons .
these photons , arising dominantly through beamstrahlung emission , are problematic for the machine design as they need to be transported and dumped in a controlled way . in this work ,
we perform simulations of the beam - beam interaction to predict photon production rates and distributions for the different beam parameters considered at ilc .
the results are expressed in terms of a set of cones of excluded power , allowing to define the beam - stay - clear requirements relevant for different cases and contexts .
a comparison is also made with theoretical expectations .
the suggested photon cone half - opening angles are 0.75 and 0.85 mrad in the horizontal and vertical planes , respectively .
these cones cover all machine energies and parameter sets , and include the low power compton photons . |
the rossiter mclaughlin effect ( hereafter the rm effect ) is a phenomenon originally reported as a `` rotational effect '' in eclipsing binary systems by @xcite ( for the beta lyrae system ) and @xcite ( for the algol system ) . in the context of extrasolar planetary science ,
the rm effect is seen as a radial velocity anomaly during a planetary transit caused by the partial occultation of the rotating stellar disk ( see @xcite , @xcite , or @xcite , for theoretical descriptions ) .
the radial velocity anomaly depends on the trajectory of the planet across the disk of the host star , and in particular on the alignment between that trajectory and the rotation field of the star . by monitoring this anomaly throughout a transit
one can determine whether or not the planetary orbital axis is well - aligned with the stellar spin axis .
in the solar system , the orbits of all 8 planets are known to be well - aligned with the solar equator , but this is not necessarily the case for exoplanetary systems , or for hot jupiters in particular .
the key parameter is the sky - projected angle between the stellar spin axis and the planetary orbital axis , @xmath1 , and measurements of this `` misalignment angle '' for various exoplanetary systems will help to place the solar system in a broader context .
specifically , measurements of the rm effect for exoplanetary systems are important because of the implications for theories of migration and hot jupiter formation .
so far , measurements of @xmath1 for two systems have been reported ; and @xcite for hd 209458 , @xcite for hd 189733 . in both of those systems ,
the host star is very bright ( @xmath8 ) , facilitating the measurement .
the observed values of @xmath1 for the two systems are small or consistent with zero , which would imply that the standard migration mechanism ( planet - disk interaction ) does not alter the spin - orbit alignment grossly during the planetary formation epoch .
however , just these few examples are not enough for statistical constraints on other hot jupiter formation theories , including planet - planet interaction @xcite , the `` jumping jupiter '' model @xcite , or the kozai mechanism @xcite , which may lead hot jupiters to have significantly misaligned orbits
. thus further measurements of the rm effect for other transiting systems are valuable . given that most of ongoing transit surveys target relatively faint ( @xmath2 ) host stars , it is also important to extend the reach of this technique to fainter stars .
further observations for new targets would be useful to constrain planet formation theories , and more importantly , have a potential to discover large spin - orbit misalignments , which would be a challenge to some theoretical models . in this paper , we report the measurement of the rm effect and the constraint on @xmath1 for tres-1 ( @xmath9 ) which has a significantly fainter host star than those in previous studies ( @xmath3 ) .
in addition to the fainter host star , this work differs from previous studies of the rm effect in that we have conducted simultaneous spectroscopic and photometric observations .
this new strategy offers several potentially important advantages .
first , the simultaneous photometry eliminates any uncertainty in the results due to the orbital ephemeris and the transit depth .
although this did not turn out to be crucial for the present work , it will be useful for newly discovered targets which still have uncertainty in the times of transits .
second , the transit depth might be expected to vary due to star spots or transient events , and indeed evidence for star spots was reported in hst / acs photometry for this system @xcite .
thus simultaneous monitoring is useful to assess anomalies in the transit depth .
finally , obtaining all of the data on a single night is useful to avoid systematic errors in radial velocities from long - term instrumental instabilities .
moreover , although it need not be simultaneous , the photometry also helps to determine the limb - darkening parameter for the visual band , which can be used in the interpretation of the rm - affected spectra . in this way , we can determine the limb - darkening parameter of the host star directly from the data , instead of assuming a value based on stellar atmosphere models .
we describe our observations in section 2 and report the results in section 3 .
section 4 provides a discussion and summary of this paper .
we observed the planet - hosting k0v star tres-1 on ut 2006 june 21 , the night of a predicted planetary transit , using the subaru 8.2 m telescope at mauna kea and the magnum 2 m telescope at haleakala , both in hawaii .
the event is predicted as the 238th transit from the first discovery , namely @xmath10 ( @xmath11 : integer ) in the ephemeris by @xcite ; @xmath12.\end{aligned}\ ] ] the transit occurred shortly after midnight .
we observed tres-1 during 5 hours bracketing the predicted transit time , through air masses ranging from 1.0 to 1.3 .
the magnum 2 m telescope is located near the haleakala summit on the hawaiian island of maui @xcite .
the magnum photometric observation was conducted in parallel with the subaru spectroscopic observation .
we employed the multi - color imaging photometer ( mip ) using a @xmath13 pixel ccd with a @xmath14 band filter , covering 4750 @xmath15 6180 , and we set 9 dithering positions ( 3@xmath16 positions ) on the ccd .
the mip has a @xmath17 arcmin@xmath18 field of view ( fov ) with a pixel scale of 0.277 arcsec / pixel .
we used 2mass j19041058 + 3638409 as our comparison star for differential photometry .
this star is close enough to fit in the mip field of view , and is known to be photometrically stable at a level sufficient for our study ( e.g. , @xcite ) .
the exposure time was either 40 or 60 seconds according to observing conditions so that the photon counts are close to the saturation level of the ccd , with a readout / setup time of 60 seconds .
we then reduced the images with the standard mip pipeline described in @xcite .
we determined the apparent magnitudes of tres-1 and the comparison star using an aperture radius of 20 pixels .
the typical fwhm of each star ranged from 1.4 to 1.9 arcsec ( from 5 to 7 pixels ) .
we estimated the sky background level with an annulus from 20 to 25 pixels in radius centered on each star , and subtracted the estimated sky contribution from the aperture flux .
then we computed the differential magnitude between tres-1 and the comparison star .
after these steps , we decorrelated the differential magnitude from the dithering positions and eliminated apparent outliers from the light curve , most of which were obtained at the 9th dithering position .
we do not find any clear correlations with other observing parameters . for the analysis of transit photometry
, @xcite studied the time - correlated noise ( the so - called `` red noise '' ) in detail , and introduced a simple and useful method to account at least approximately for the effect of the red noise on parameter estimation . based on these studies , we used the following procedure to determine the appropriate data weights for the magnum photometry ( which are similar to that employed by @xcite ) .
we first fitted the magnum light curve with the analytic formula given in @xcite and found the residuals between the data and the best - fitting model . using only the poisson noise as an estimate of the error in each photometric sample , we found @xmath19 ( @xmath20 : degrees of freedom ) , implying that the true errors are significantly in excess of the poisson noise .
we also found the residuals in the early part of the night ( before @xmath21 [ hjd ] ) to be significantly larger than those from later in the night .
this larger scatter could have been caused by shaking of the telescope by the stronger winds that occurred during the early part of the night .
we thus estimated error - bars separately for the early part of the night and the later part of the night , as described below .
first , we rescaled the error bars to satisfy @xmath22 ( step 1 ) , namely 0.00259 for the early data and
0.00189 for the late data .
the light curve with these rescaled error bars is shown in the upper panel of fig .
[ ourdata ] .
next , in order to assess the size of time - correlated noise for the magnum data , we solved the following equations , @xmath23 where @xmath24 is the standard deviation of each residual and @xmath25 is the standard deviation of the average of the successive @xmath26 points .
@xmath27 is called the white noise , which is uncorrelated noise that averages down as @xmath28 , while @xmath29 is called the red noise , which represents correlated noise that remains constant for specified @xmath26 .
we calculated @xmath27 and @xmath29 for the choice @xmath30 ( corresponding to one hour ) , finding @xmath31 and @xmath32 for the early data . on the other hand
, we found @xmath33 ( suggesting a smaller level of the red noise ) for the late data .
the choice of @xmath30 is fairly arbitrary ; other choices of @xmath26 between 5 and 50 gave similar results .
we then adjusted the error bars for the early night by multiplying @xmath34^{1/2 } \sim 3.6 $ ] ( step 2 ) .
we did not change the error bars for the late night data .
we adopted these rescaled uncertainties for subsequent fitting procedures .
( 70.5mm,70.5mm)figure1.eps lcc time [ hjd ] & value [ m s@xmath35 & error [ m s@xmath35 + 2453907.87018 & 18.7 & 14.0 + 2453907.88139 & 30.5 & 12.5 + 2453907.89262 & 54.6@xmath36 & 12.0 + 2453907.90384 & 24.3 & 10.4 + 2453907.91506 & 26.4 & 11.4 + 2453907.92628 & 30.4 & 10.9 + 2453907.93750 & 22.4 & 14.3 + 2453907.94873 & 2.9 & 11.0 + 2453907.95996 & -7.1 & 12.1 + 2453907.97119 & -22.3 & 13.3 + 2453907.98241 & -40.5 & 13.3 + 2453907.99364 & -39.2 & 13.0 + 2453908.00488 & -9.8 & 12.2 + 2453908.01610 & -30.5 & 13.8 + 2453908.02732 & -17.7 & 13.6 + 2453908.03854 & -24.7 & 12.2 + 2453908.04978 & -27.5 & 11.1 + 2453908.06100 & -38.2 & 13.3 + 2453908.07223 & -23.7 & 11.2 + 2453908.08345 & -23.0 & 9.6 + [ rvsummary ] we used the high dispersion spectrograph ( hds ) on the subaru telescope @xcite .
we employed the standard i2a set - up of the hds , covering 4940 @xmath15 6180 with the iodine absorption cell for measuring radial velocities .
the slit width of @xmath37 yielded a spectral resolution of @xmath38 45000 , and the seeing was between @xmath39 and @xmath40 .
the exposure time for tres-1 was 15 minutes yielding a typical signal - to - noise ratio ( snr ) @xmath38 60 per pixel . in order to estimate systematic errors from short term instrumental variations
, we also obtained spectra of the much brighter ( @xmath41 ) k0v star hd 185144 before and after the series of tres-1 exposures .
this star is known to be stable in velocity @xcite .
we obtained five 30 s exposures of hd 185144 , each having a snr of approximately 200 pixel@xmath42 .
we processed the frames with standard iraf procedures and extracted one - dimensional spectra .
next , we calculated relative radial velocity variations by the algorithm following @xcite .
we used this algorithm because it properly takes into account the fairly large changes of the instrumental profile during the observations .
the hds is known to experience appreciable instrumental variations even within a single night , reported in @xcite and @xcite .
we estimated internal errors of the radial velocities from the scatter of the radial velocity solutions for 2 segments of the spectra .
the typical errors are @xmath43 [ m s@xmath35 , which are reasonable values to be expected from the photon noise limit .
note that we do not find any evidence of star spots or transient events during our photometric observation ( see fig .
[ ourdata ] ) . for this reason
, we have not accounted for possible systematic errors in the velocities due to star spots .
we also reduced the hd 185144 spectra with the same method in order to check for systematic errors due to short - term instrumental instabilities .
the rms of the radial velocity of hd 185144 is less than 5 [ m s@xmath35 , attesting to good instrumental stability .
the resultant radial velocities of tres-1 and their errors are shown in table [ rvsummary ] and the lower panel of fig .
[ ourdata ] .
( 120mm,50mm)figure2a.eps ( 120mm,50mm)figure2b.eps ( 120mm,50mm)figure2c.eps ( 120mm,50mm)figure3.eps l|cc|cc|cc & & & + & & & + parameter & value & uncertainty & value & uncertainty & value & uncertainty + ( all rv samples ) & & & + @xmath44 [ m s@xmath35 & @xmath45 & @xmath46 & @xmath45 & @xmath46 & @xmath47 & fixed + @xmath48 [ km s@xmath35 & @xmath49 & @xmath50 & @xmath51 & @xmath52 & @xmath49 & @xmath50 + @xmath1 [ deg ] & @xmath53 & @xmath54 & @xmath55 & @xmath56 & @xmath57 & @xmath58 + @xmath59 & @xmath60 & @xmath61 & @xmath60 & @xmath61 & @xmath62 & @xmath61 + @xmath63 & @xmath64 & @xmath65 & @xmath64 & @xmath65 & &
+ @xmath66 & @xmath67 & @xmath68 & @xmath67 & @xmath68 & @xmath69 & fixed + @xmath70 [ @xmath71 & @xmath72 & @xmath73 & @xmath72 & @xmath73 & @xmath74 & fixed + @xmath75 [ deg ] & @xmath76 & @xmath50 & @xmath76 & @xmath77 & @xmath78 & fixed + @xmath79 [ m s@xmath35 & @xmath49 & @xmath80 & @xmath81 & @xmath82 & @xmath83 & @xmath84 + @xmath85 [ m s@xmath35 & @xmath86 & @xmath87 & @xmath86 & @xmath87 & & + @xmath88 [ m s@xmath35 & @xmath89 & @xmath90 & @xmath89 & @xmath90 & & + @xmath91 [ hjd ] & @xmath92 & @xmath93 & @xmath92 & @xmath93 & & + @xmath94 [ hjd ] & @xmath95 & @xmath96 & @xmath95 & @xmath96 & & + @xmath97 [ hjd ] & @xmath98 & @xmath99 & @xmath98 & @xmath99 & & + @xmath100 [ hjd ] & @xmath101 & @xmath102 & @xmath101 & @xmath102 & @xmath103 & @xmath102 + ( without the outlier ) & & & + @xmath44 [ m s@xmath35 & @xmath104 & @xmath46 & @xmath104 & @xmath46 & @xmath47 & fixed + @xmath48 [ km s@xmath35 & @xmath49 & @xmath50 & @xmath105 & @xmath52 & @xmath49 & @xmath50 + @xmath1 [ deg ] & @xmath106 & @xmath107 & @xmath108 & @xmath54 & @xmath109 & @xmath110 + @xmath79 [ m s@xmath35 & @xmath111 & @xmath112 & @xmath113 & @xmath114 & @xmath115 & @xmath80 + [ result ]
as described above , we have obtained 20 radial velocity samples and 184 @xmath14 band photometric samples taken simultaneously covering the transit .
in addition , in order to search for an optimal solution of orbital parameters for tres-1 , we incorporate our new data with 12 previously published radial velocity measurements using the keck i telescope ( 7 by @xcite and 5 by @xcite ) and 1149 @xmath116 band photometric measurements spanning 3 transits using the flwo 1.2 m telescope @xcite .
the uncertainties of the flwo data had already been rescaled by the authors such that @xmath22 for each transit ( namely , the step 1 has been done ) . for the step 2 ,
we find @xmath33 for these data , thus we did not modify these error bars further .
we employ the analytic formulas of radial velocity and photometry including the rm effect given in @xcite and @xcite ( hereafter the ots formulae ) in order to model the observed data .
note that based on the previous studies by @xcite and @xcite , we have learned that the ots formulae systematically underestimate the amplitude of radial velocity anomaly by approximately 10% .
this is possibly because the radial velocity anomaly defined by @xcite and that measured by the analysis pipeline are different .
thus we correct @xmath48 in the ots formulas by modifying @xmath117 .
this correction presumably gives more realistic values for @xmath48 and @xmath1 , and has little influence on any of the other parameters .
here we assume circular orbits of the star and the planet about the center of mass ( namely , @xmath118 ) .
we adopt the stellar mass @xmath119 [ @xmath120 @xcite and the orbital period @xmath121 [ days ] and @xmath122 [ hjd ] @xcite . as a result
, our model has 15 free parameters in total .
eight parameters for the tres-1 system include the radial velocity amplitude @xmath44 , the sky - projected stellar rotational velocity @xmath48 , the misalignment angle between the stellar spin and the planetary orbit axes @xmath1 , the linear limb - darkening parameter for @xmath14 band @xmath59 , the same for @xmath116 band @xmath63 , the ratio of star - planet radii @xmath66 , the stellar radius @xmath70 , and the orbital inclination @xmath75 .
here we assume that the limb - darkening parameters for the spectroscopic and photometric models are the same .
lines where the radial velocities are measured . however , in principle the correspondence is not exact because the limb - darkening function may not be identical in the lines as opposed to the continuum , and because the influence of limb - darkening on the rv - measuring algorithm has yet to be investigated in detail .
] we also add three parameters for velocity offsets to the respective radial velocity dataset @xmath123 ( for our template spectrum ) , @xmath124 ( for @xcite ) and @xmath125 ( for @xcite ) , and four parameters for the times of mid - transit @xmath126 . in previous studies of the exoplanetary rm effect , it was possible and desirable to determine both @xmath48 and @xmath1 from the radial velocity data .
in this case , there are two reasons to prefer an external determination of @xmath48 .
first , the signal - to - noise ratio of the anomaly is smaller , because of the faintness of the host star .
this makes it valuable to reduce the number of degrees of freedom in the model .
second , the transit geometry is nearly equatorial , which introduces a very strong degeneracy between @xmath48 and @xmath1 , as explained by gaudi & winn ( 2007 ) .
the alternative we have chosen is to adopt a value for @xmath48 based on previous observations , and use the radial - velocity anomaly to determine @xmath1 .
( we have also investigated our ability to determine both parameters , as described below . )
@xcite reported @xmath127 [ km s@xmath35 for the tres-1 host star from their analysis of the observed spectral line profiles ; this is the most reliable estimate for @xmath48 to date .
we incorporate this information into our model by adding a term @xmath128 ^ 2 $ ] to the @xmath129 fitting statistic .
thus our @xmath129 statistic is @xmath130 ^ 2 + \sum_{j=1}^{n_{f}=1333 } \left [ \frac{f_{j , obs } -f_{j , calc}}{\sigma_{j } } \right]^2 \nonumber\\ & + & \left [ \frac{v \sin i_s - 1.08}{0.30 } \right]^2,\end{aligned}\ ] ] where @xmath131 and @xmath132 represent the values calculated by the ots formulae with the above parameters .
we find optimal parameters by minimizing the @xmath129 statistic of eq .
( 4 ) using the amoeba algorithm @xcite , and estimate confidence levels of the parameters using @xmath133 from the optimal parameter set . to assess the dependence of our results on the _ a priori _ constraint on @xmath48 , we also compute and compare the best - fit values and uncertainties by using another function : @xmath134 ^ 2 + \sum_{j=1}^{n_{f}=1333 } \left [ \frac{f_{j , obs } -f_{j , calc}}{\sigma_{j } } \right]^2,\ ] ] for reference .
in addition , we note that the third radial velocity sample of our data ( @xmath135 [ hjd ] ) may appear to be an outlier , but it lies just about 3@xmath7 from a theoretical radial velocity curve ( e.g. , fig .
[ rvwithapr ] ) . for clarity
, we calculate eq . ( 4 ) and eq .
( 5 ) with and without that sample .
the results for both @xmath129 statistics are presented in table [ result ] .
the minimum @xmath129 is 1308.57 ( 1296.39 ) for eq .
( 4 ) and 1305.18 ( 1298.57 ) for eq . ( 5 ) with 1350 ( 1349 ) degrees of freedom , where the numbers in parentheses refer to the case without the outlier . in fig .
[ rvwithapr ] , we present the radial velocities and the best - fit curve with ( the top figure , marked with `` a '' ) and without ( the middle figure , marked with `` b '' ) the _ a priori _ constraint . in addition , we also compute the best - fit curve without the constraint , but assuming that @xmath136 [ deg ] ( the bottom figure , marked with `` @xmath136 [ deg ] '' ) .
[ vlcontour ] plots ( @xmath137 ) contours calculated with ( the left panel , marked with `` a '' ) and without ( the right panel , marked with `` b '' ) the _ a priori _ constraint .
note that we only show here the results with the possible outlier in fig .
[ rvwithapr ] and fig .
[ vlcontour ] , since the same figures but without the outlier have basically similar appearance and have less information to show . as a result
, we find @xmath138 [ deg ] and @xmath139 [ km s@xmath35 for tres-1 with the _ a priori _ constraint , and our findings except for @xmath1 are in good agreement with previous studies @xcite . on the other hand ,
the result without the _ a priori _ constraint , @xmath140 [ deg ] and @xmath141 [ km s@xmath35 , agrees with the above result within about 1@xmath7 for @xmath1 and @xmath48 , and is also consistent with the previous results for other parameters . in case
we calculate @xmath129 without the constraint but assuming that @xmath136 [ deg ] , we find that the minimum @xmath129 is 1309.85 ( 1298.91 ) with 1351 ( 1350 ) degrees of freedom , and @xmath142 [ km s@xmath35 .
consequently , our results for @xmath1 have fairly large uncertainties .
we find at least that the orbital motion of tres-1b is prograde .
additional radial velocity measurements during transits , and a more precise measurement of @xmath48 , would be desirable to pin down @xmath1 and help to discriminate between different modes of migration .
( 75mm,75mm)figure4.eps ( 60mm,60mm)figure5.eps
we have presented simultaneous spectroscopy and photometry of a transit of tres-1b , exhibiting a clear detection of the rossiter - mclaughlin effect and consequent constraints on the alignment angle between the stellar spin and the planetary orbital axes .
our philosophy has been to use all of the best data available at present . however , it is also interesting to examine how well we are able to determine the system parameters using _ only _ the data gathered on a single night using subaru and magnum .
this is because for future studies of newly discovered transiting exoplanetary systems , higher - precision data from other observatories may not be available .
we repeat the fitting procedure without the keck and flwo data but still assuming the system parameters other than @xmath48 , @xmath1 , @xmath59 , @xmath79 , and @xmath143 to be the values presented in @xcite and @xcite .
we find almost the same values and uncertainties for the parameters above ( the right side of table [ result ] ) as before , indicating that a single night s data would have done almost as well as the full data set .
using transit timing of the tres-1 system , @xcite reported a constraint on the existence of additional planets in the system .
subsequently , @xcite pointed out that @xmath144 are consistent with their ephemeris at about the 2@xmath7 level , but occurred progressively later than expected . here
we present @xmath145 in fig .
our result is also consistent with the published ephemeris within 1.5@xmath7 , occurring only slightly later than expected from @xcite .
in addition to producing transit timing variations , any additional bodies in the tres-1 system could excite the orbital eccentricity of the tres-1b planet , which could be detectable in the radial velocity measurements of the host star .
thus it is worthwhile to put empirical constraints on the orbital eccentricity .
for this purpose , we compute @xmath129 using eq .
( 4 ) for fixed values of ( @xmath146 ) over numbers of grid points to map out the allowed region in ( @xmath147 ) space . in fact
the most appropriate parameters are @xmath148 and @xmath149 since their uncertainties are uncorrelated ( see fig .
[ ewcontour ] ) .
the resulting constraints are @xmath150 and @xmath151 .
our result for @xmath152 is similar to that of @xcite , which was based on the timing of the secondary eclipse .
the value of @xmath152 is consistent with zero within 1@xmath7 , and the value of @xmath153 is consistent with zero within 2@xmath7 .
thus we do not find any strong evidence for a nonzero orbital eccentricity in the tres-1 system .
there is another interesting application of our data .
recently , @xcite studied the detectability of `` hot trojan '' companions near the l4/l5 points of transiting hot jupiters , through any observed difference between the time of vanishing stellar radial velocity variation ( @xmath154 ) and the time of the midpoint of the photometric transit ( @xmath155 ) .
our strategy of simultaneous spectroscopic and photometric transit observations is ideally suited for searching for the hot trojan companions . for the tres-1 system
, @xcite set an upper limit on the mass of the trojan companions @xmath156 at the 3@xmath7 level ( assuming the circular orbit ) , using the radial velocity samples of @xcite .
here we compute @xmath157 using all available out - of - transit radial velocity samples with ( without ) the possible outlier , and both for the circular and the eccentric orbit .
we find @xmath158 [ min ] ( circular ) and @xmath159 [ min ] ( eccentric ) .
accordingly , we set constraints on the mass of the trojan companions , @xmath160 , which is defined as the difference in the mass at l4 ( @xmath161 ) and the mass at l5 ( @xmath162 ) ( namely , @xmath163 ) , through the relation ; @xmath164 where @xmath165 and @xmath166 indicate the uncertainties of @xmath160 and @xmath167 , respectively ( eq . ( 1 ) and eq .
( 2 ) in @xcite ) . note that we adopt @xmath168 [ @xmath169 which is determined by this work .
we find @xmath170 [ @xmath6 ] ( circular ) and @xmath172 [ @xmath6 ] ( eccentric ) . as a result
, we exclude the trojan companions near the l4 point more massive than @xmath173 [ @xmath6 ] if the orbit is circular , and @xmath174 [ @xmath6 ] if the orbit is allowed to be eccentric , both at the 3@xmath7 level .
our constraint under the reasonable assumption of a circular orbit is more stringent than that of @xcite by a factor of 4 , because we have increased the number of radial velocity samples and because our data cover the critical phase to determine @xmath154 .
consequently , we conclude that we do not find any sign of the existence of additional bodies in the tres-1 system at present . in this paper
, we have placed a constraint on the sky - projected angle between the stellar spin axis and the planetary orbital axis for the tres-1 system , namely @xmath0 [ deg ] using all available data and information from previous studies .
although we can not discriminate whether the spin - orbit angle in this system is well - aligned or not at this point , our constraint on @xmath1 clearly indicates the prograde orbital motion of tres-1b . the uncertainty is larger than in previous studies ( @xmath175 [ deg ] for hd 209458 and hd 189733 ) because the host star is significantly fainter in this case .
although further radial velocity measurements during transit would be necessary to pin down @xmath1 more stringently , we have demonstrated for the first time that such measurements are possible for such a faint target .
this is important because most of the newly discovered transiting planets from ongoing transit surveys will have relatively faint host stars .
for example , the new targets that were discovered in 2006 , namely xo-1 @xcite , tres-2 @xcite , hat - p-1 @xcite , wasp-1 and wasp-2 @xcite , are all in this category . combining future measurements of @xmath1 in other transiting systems
, we would be able to determine the distribution of @xmath1 for exoplanetary systems with useful statistical accuracy .
we acknowledge close support of our observations by akito tajitsu , who is a support scientist for subaru hds .
we appreciate the detailed and helpful critique of the manuscript by the referee , scott gaudi .
we also thank kazuhiro yahata , shunsaku horiuchi , takahiro nishimichi , and hiroshi ohmuro for useful discussions .
is supported by a japan society for promotion of science ( jsps ) fellowships for research .
this work is supported in part by a grant - in - aid for scientific research from the jsps ( no.14102004 , 16340053 , 17740106 ) , and mext japan , grant - in - aid for scientific research on priority areas , `` development of extra - solar planetary science , '' and nasa grant nag5 - 13148 .
we wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community . | we report a measurement of the rossiter
mclaughlin effect in the transiting extrasolar planetary system tres-1 , via simultaneous spectroscopic and photometric observations with the subaru and magnum telescopes . by modeling the radial velocity anomaly that was observed during a transit ,
we determine the sky - projected angle between the stellar spin axis and the planetary orbital axis to be @xmath0 [ deg ] .
this is the third case for which @xmath1 has been measured in a transiting exoplanetary system , and the first demonstration that such measurements are possible for relatively faint host stars ( @xmath2 , as compared to @xmath3 for the other systems ) .
we also derive a time of mid - transit , constraints on the eccentricity of the tres-1b orbit ( @xmath4 ) , and upper limits on the mass of the trojan companions ( @xmath514 @xmath6 ) at the 3@xmath7 level . |
supernova explosions feature the interplay of all four known fundamental forces . a complete picture of supernova ( sn ) explosions will therefore require true multi - messenger observations , with data from traditional optical telescopes analyzed alongside coincident data from neutrino and gravitational wave detectors . to provide such multi - messenger data as well as to increase the sensitivity to neutrinos from sne , the icecube collaboration @xcite together with the rotse collaboration @xcite have set up an optical follow - up program that triggers optical observations on multiplets of high - energy muon neutrinos @xcite ( a similar program has been inplemented by the antares collaboration @xcite )
a multiplet is defined as at least two muon neutrinos from the same direction that arrive within a short time window ( e.g. , @xmath2 s ) .
when this happens , an alert is issued to the four rotse - iii telescopes , which immediately observe the corresponding region in the sky .
successful @xmath0-@xmath1 coincident detection would allow one to infer the existence of jets in sne and would probe the expected gamma - ray burst supernova connection @xcite .
absent any other corroborating astrophysical evidence , a standalone neutrino doublet is not a physically interesting occurrence because the background rate of such doublets from atmospheric muon neutrinos in icecube is @xmath3(10/yr ) .
however , given the very small number of sn expected by random chance in the doublet s temporal and directional windows , the significance of a coincident optical observation of a sn rises dramatically : a neutrino doublet and an optical observation in a coincidence time window typically assumed to be @xmath4 day ( which we will show can be narrowed considerably ) is of comparable significance to the detection of a standalone neutrino triplet @xcite .
( neutrino triplets occur by chance only once every few millennia and therefore their detection would be intrinsically significant . ) a gain in sensitivity is thus achieved by effectively lowering the neutrino multiplicity threshold ( from @xmath5 to @xmath6 ) affording a factor of about two increase in the rate of detectable sne @xcite . in other words ,
the ability to narrow the coincidence time window provides a way to reduce the level of accidental coincidences between neutrino doublets and sne .
this useful reduction can be achieved by rejecting coincidences for which the neutrino doublet arrival time is statistically too far outside the narrower time window obtained from a fit to the sn explosion time .
this program thus relies on the ability to match the explosion time as determined from the neutrino multiplet arrival time to that determined from the optical data .
smaller explosion time uncertainties result in better rejection of accidental coincidences , lending more significance to a coincidence detection .
previous studies have assumed that the explosion time can be known with a precision of about one day without , however , supporting this assumption with observational data ( see , e.g. , @xcite ) . in this paper
we present the first study of the determination of the explosion time , @xmath7 , from the sn optical light curve .
we have produced a generic model for the light curve that we test with light curve data from sne with known explosion times .
such a study became attractive in the last few years due to the recent fortuitous discoveries of two nearby type ib / c ( stripped - core ) sne , sn2008d @xcite and sn2006aj @xcite , each with an associated x - ray flash presumably from the shock breakout .
the short x - ray flash provides a time stamp for the explosion that can be compared to the one obtained from fitting the optical light curve data .
furthermore , for obvious reasons the light curve data for these sne begin very early after the x - ray flash , and as such are well - suited to the method described below , because as with an x - ray flash , a neutrino trigger will enable early optical observation of the target sn .
the only other sn that has an explosion time known with even better precision is sn1987a .
it is a low - luminosity type iip sn with a light curve very different from that of sn2008d and sn2006aj .
nevertheless , the physics of the early part of the light curve is similar enough that we find we can successfully extend our analysis to sn1987a as well .
the explosion time of sn1987a is taken to be the time of the mev neutrino burst .
for sn2008d and sn2006aj , we take the x - ray flash as a rough proxy for the explosion time .
the latter is justified by considering both radius and shock velocities for these sne .
the estimated radius at which sn2008d s progenitor system becomes optically transparent to x - rays , @xmath8 @xcite , is relatively small .
for sn2006aj , a larger radius of @xmath9 cm is estimated @xcite , while for sn1987a a photospheric radius of @xmath10 cm is assumed @xcite .
the maximum shock velocity at the shock breakout has been computed as a function of radius , energy and mass in @xcite .
inserting parameters for the sne at hand we obtain @xmath11 c for sn2008d and @xmath12 c for sn1987a . the non - relativistic theory in @xcite yields a maximum shock velocity for sn2006aj that exceeds the speed of light .
the authors in @xcite do a relativistic treatment and estimate 0.85 c. one obtains the minimum time scale @xmath13 s for sn2008d , @xmath14 s for sn2006aj and @xmath15 s sn1987a .
while this crude calculation underestimates by a factor of five the @xmath16 s delay time between explosion and shock breakout predicted by a detailed simulation of sn1987a @xcite , it indicates that for sn2008d and sn2006aj , the shock breakout is not expected to appear much later than @xmath17 s after the explosion . as will be shown in sec .
[ sec : results ] , this theoretical time scale for the shock propagation is much shorter then the resolution of the fit on the time of explosion @xmath7 that we obtain for sn2008d . for sn2006aj , it is comparable to the resolution of the light curve fits .
the remainder of this paper is organized as follows . in sec .
[ sec : data ] we present the light curve data and the model that is used to analyze them . in sec .
[ sec : results ] we present the results of the light curve fits for sn2008d , sn2006aj and sn1987a . in sec .
[ sec : early ] we discuss the importance of early light curve data .
we summarize the implications of our results in sec . [
sec : conclusion ] .
the sn2006aj and sn2008d light curves contain data from times exceptionally soon after their putative explosions , making an accurate estimation of sn explosion times feasible . for sn2006aj
we use the u , b and v band data from the swift uvot @xcite and for sn2008d we use the b , v , r and i band data from flwo @xcite .
additional data from other telescopes is available , but in order to avoid calibration problems arising from different filter and instrument pass bands , we decided to work only with data from a single source . while we have not included it in the fits shown in this paper , we note that it fits the model prediction well . ] . for sn 1987a
, we use the photometric b , v , r , and i band data compiled and analyzed consistently by hamuy _
et al . _
the first data point is 1.14 days after the explosion .
again , to avoid calibration problems , we do not use the earlier discovery data points that exist for the v - band .
we estimate the explosion time by fitting light curves under the assumptions of an initial blackbody emission from the rapidly cooling shock breakout , followed by a phase dominated by the expansion of the luminous shell .
for the latter we test two distinct models .
* shock breakout phase : * to represent the shock breakout phase we use the formulation from waxman et al .
the flux during the shock breakout phase of the sn light curve is approximated by @xmath18 , where @xmath19 is the area and @xmath20 is the intensity .
the intensity is taken as proportional to that produced by a blackbody at a fixed wavelength ( we set @xmath21 nm , but this reference wavelength is not relevant for the results presented here since it appears as multiplicative factor to the fitted temperature ) .
in addition to the explosion time @xmath7 , the other free parameters of the model are the radius and temperature at a fixed reference time .
waxman et al .
@xcite give the sn radius @xmath22 and the shock breakout temperature @xmath23 , where @xmath24 is the elapsed time since the explosion .
inserting these relations in the flux equation yields : @xmath25 with @xmath26 , @xmath27 and @xmath7 free parameters .
* expansion phase : * for the expansion phase we use either a simple expanding photosphere model for the behavior of the light curve or the more complex description from arnett @xcite that uses a time - dependent diffusion equation .
dependence for the late time emission ; the dashed lines the arnett model ( the @xmath28 formulation , which here gives @xmath29 hr , is preferred due to a better fit ; see text in sec .
[ sec : results ] for details ) . ] in the first model , the flux in the pure expansion phase is approximated as @xmath30 with @xmath31 defined above and @xmath32 and @xmath7 free parameters .
this @xmath28 assumption treats the sn photosphere as represented by a blackbody of constant temperature , which expands with constant velocity @xmath33 @xcite .
the area of the photosphere , which is directly proportional to the photon flux , then increases @xmath34 .
this _ ansatz _
works remarkably well for the rising part of type ia sn light curves @xcite .
the model has one free parameter and , when combined with the blackbody emission model , there are a total of four parameters in the fit to the light curve . as an alternative to the expanding photosphere model
, we use the light curve model of arnett @xcite ( also used in @xcite ) , that assumes homologous expansion , radiation pressure dominance , and @xmath35ni present in ejected matter and distributed toward the center of the ejected mass . in this alternative model
there are two free parameters for the rising part of the light curve model , so there are a total of five free parameters in the fit to the light curve . as an example , fig .
[ fig : singleplot ] shows the results of the fits to r - band light curve data of sn2008d .
the full set of light curves for sn2008d , sn2006aj and sn1987a are shown in figs .
[ fig : allsevenplots ] and [ fig : lc1987a ] .
a systematic evaluation of fits to all available bands is the subject of the next section .
we fit our two models to the light curve data in multiple bands for sn2006aj and sn2008d , as shown in fig . [
fig : allsevenplots ] . for each fit
we extract the initial explosion time , @xmath7 , the error on @xmath7 , and the @xmath36 of the fit . for sn2006aj , we find only marginal difference in the accuracy of the fitted @xmath7 if we use the more complex arnett formulation instead of the simpler @xmath28 .
( for the comparison , we restricted the fit to the first six days , since the light curve of sn2006aj evolves faster than other sne , and for later times the @xmath28 approximation does not hold . ) the agreement between both fit models is due to the fact that the earliest part of the light curve is entirely dominated by emission from the shock breakout and hence already strongly constrains @xmath7 .
we obtain an average @xmath7 that is shifted by -0.04 days relative to the x - ray flash , with a statistical error of about 0.005 days .
the light curve data of sn2008d can also be fit by both the @xmath28 and arnett formulations .
however , the early data shown in figs . [ fig : singleplot ] and [ fig : allsevenplots ] is better represented by the @xmath28 model , as determined by the quality of the fit .
we obtain from the fits a @xmath37 for the sum of all four bands .
fitting with the arnett formulation instead of @xmath28 one obtains , with one additional fit parameter per band , a @xmath36 that is significantly worse ( @xmath38 for the sum of all data ) .
we hence proceed with the @xmath28 fit model as our default fit method .
the fit results for all bands using the @xmath28 formulation are shown in fig .
[ fig : summaryplot ] . for sn2008d ,
whose light curves do not start so soon after the explosion time , we find the fitted @xmath7 is consistent with zero for three out of four bands ( 90% cl ) , with an average error of about 0.06 days .
the largest outlier is the v - band , with @xmath39 days .
if the arnett formulation is used instead of the default @xmath28 formulation , the estimated @xmath7 would be shifted by almost 5 hours to late times ( see also fig . [
fig : singleplot ] ) .
in contrast to sn2006aj and sn2008d , sn1987a has its date of birth clearly marked by the observation of a short burst of neutrinos . since its detection
, sn1987a has been studied in great depth , both observationally and theoretically .
we can not expect to have such detailed information for future sne unless they appear in our own galaxy and hence , for the sake of simplicity , we adopt the methodology already used above .
we have fit the light curve data of sn1987a with the model composed of the shock breakout according to eq .
[ eq : phibb ] and the @xmath28 dependence for the expansion phase .
we fit the first eight days of data .
since the photometric data @xcite does not come with estimated uncertainties , we have chosen them to be 0.03 mag to achieve @xmath40 in the fits .
the size of this assumed uncertainty roughly matches the largest scatter of photometric data points observed during a single night .
the fit results for four bands are shown in fig .
[ fig : summaryplot ] . in the figure ,
the larger error bar for the v , r and i - band fits with respect to the b - band fits reflects the fact that the shock breakout feature is not very evident for the redder bands , as can be seen in fig .
[ fig : lc1987a ] .
we have explored whether we can improve the fits by incorporating some key observations for sn1987a into the model , for example that an almost constant bolometric luminosity was observed after the first day after the explosion . in our simple picture , this is achieved by making the photosphere radius expand linearly with time , @xmath41 , while keeping the photosphere temperature dependence as before : @xmath42 .
reinserting this into eq .
[ eq : phibb ] provides a slightly modified model for the shock breakout phase . fitting this shock breakout model results in a systematic shift of @xmath43 days for all bands .
while the b - band result is now consistent with the explosion time obtained from the neutrino burst , the fits of the redder bands appear systematically shifted .
either way , we observe a deviation of the order of 0.3 days for one of the bands indicating the size of the systematic uncertainty involved in the extrapolation .
( using arnett s formulation for the late times is not well justified for sn1987a .
nevertheless , we note that using it does not significantly worsen the fits or change the conclusions . ) summarizing , for sn1987a the light curve data starting 1.14 days after the neutrino burst allows one to fit the explosion time with a fitting error of about 0.2 days and a systematic error of about 0.3 days .
the systematic uncertainty reflects the crudeness of the light curve model employed . nevertheless , relative to simply using @xmath44 days , the fitting technique yields a factor of @xmath45 improvement in the @xmath7 measurement .
the fitted @xmath7 values demonstrate that an estimate of the explosion time with an accuracy of much less than one day can be made using simple analytic light curve models .
the estimates are robust on the scale of a few hours across several independent optical bands .
( the difference between the fitted @xmath7 and the time of the x - ray flash or neutrino burst ) and have a thickness corresponding to the error on the mean .
the @xmath28 formulation is used throughout since it provides comparable or better quality fits relative to that of arnett . ] after each removal .
the solid black line shows the resulting fitted @xmath7 values and the vertical height of the shaded region shows their 1-@xmath46 resolutions , as a function of the time of the earliest used data point in the fit .
the dot - dash line shows the value obtained for @xmath7 simply using the earliest available data point . comparing this curve to the black line from the fitted @xmath7 values
, one sees that if there is a latency of roughly six or more hours after the putative explosion time before the first optical observation is made , the fitting technique provides the more accurate explosion time estimate . ]
using the sn2006aj data , we demonstrate the importance of the early data points by manually removing the earliest data points , one at a time , and re - fitting the data each time . a summary of the result of this exercise is shown in fig .
[ fig : removingpoints ] .
the figure makes evident the importance of the early data points , showing how the accuracy of the fitted @xmath7 depends strongly on these early data , although the accuracy drops most dramatically after about a day .
this is consistent with the observation made for the other sne : with a first data point at @xmath47 days for sn2008d , the explosion date can still be determined to within about 0.2 days , while for sn1987a , with a first data point at 1.1 days after the explosion , the uncertainty is around 0.3 days .
it is also informative to compare the accuracy of the @xmath7 from the fit with that obtained by simply using the first available point on the light curve .
the dot - dash line in fig .
[ fig : removingpoints ] plot shows this @xmath7 estimate , which is simply the difference in time between the x - ray flash and the earliest remaining data point on the light curve .
the figure thus shows that if light curve data is acquired with about a six hour or greater delay from the explosion time , the fitting technique provides a more accurate and precise measure of the explosion time than simply using the earliest point on the light curve .
we have also studied the effectiveness of our method for estimating the @xmath7 when the acquired light curve follows a multi - day cadence , anticipating future survey telescopes such as lsst @xcite .
since we need to have an estimator of the `` true '' @xmath7 to do this , we took the sn2006aj light curve and kept only a subset of its data , choosing those points separated by the lsst cadence of about three days .
this was done for several distinct data subsets . in all cases ,
the fitted resolution on @xmath7 degraded substantially , to about 1.5 days , in agreement with the results shown in fig .
[ fig : removingpoints ] with about 2.5 days of early data removed .
our method thus relies explicitly on the early detection of the light curve . barring future serendipitous discoveries akin to sn2006aj and sn2008d
, we therefore rely on the neutrino - triggered optical follow - up technique ( mentioned earlier ) to provide us with a light curve that extends back suitably close in time to the actual @xmath7 of the explosion .
both the icecube and antares collaborations have started searching for high energy neutrinos from sne by implementing a target - of - opportunity ( too ) program using robotic optical telescopes to identify an optical counterpart to the neutrino signal @xcite . a crucial ingredient therein is the ability to determine the sn explosion time from the optically observed light curve which allows one to establish a temporal coincidence with the neutrino data . using a model for the early part of the light curve , we show for the first time that one can estimate the explosion times with an accuracy much better than the one day generally assumed in the literature ( see e.g. @xcite ) .
we have fitted the light curves of three very different core - collapse sne : sn2008d , sn2006aj and sn1987a . for the expansion phase of the sn we have tested two models , a simple @xmath28 model as well as a more detailed model by arnett .
we found that even with fewer parameters , the @xmath28 dependence generally provids a better fit to the data as well as a better match with the explosion time .
we hence recommend that this be used for a future coincidence search . as shown in fig .
[ fig : summaryplot ] , the estimated @xmath7 and its error , averaged over all available bands , is about @xmath48 days for sn2008d , @xmath49 days for sn2006aj , and @xmath50 days for sn1987a .
for sn2008d , the theoretical uncertainty associated with the use of the time of the x - ray flash as the reference @xmath7 is smaller than the resolutions on @xmath7 from our fits .
the fits in all bands give explosion times that are slightly later than the time of the x - ray flash , indicative of limitations in the rather simple underlying physical model .
for sn2006aj , the explosion date was determined from the fit to the light curve to be @xmath51 s before the x - ray flash . as mentioned earlier ,
this is larger than the estimated time needed for the shock to propagate to the surface of the progenitor .
resolving this discrepancy would require more detailed modeling of the light curve and/or shock propagation . in any case
, the discrepancy for both type ib / c sne investigated is @xmath52 hrs , which can be considered the characteristic size of the systematic uncertainties in @xmath7 .
note that the resolutions on @xmath7 for both sne are longer than what is expected for the onset of gravitational wave or high energy neutrino emission @xcite .
the model does not take into account possible effects due to circumstellar interactions , asymmetries in the ejecta or the differences in the density profiles of the progenitors .
these effects might explain the observed deviations that are difficult to explain with statistical errors alone .
nevertheless , the fitted sn explosion time @xmath7 represents a successful extrapolation of the data to earlier times , and the magnitude of the extrapolation is large compared to the quoted error .
this suggests that the model captures dominant physical properties of the sn during the period shortly after its explosion . for the ongoing programs @xcite
, one can not expect to have similarly detailed multiband , high signal - to - noise observations as we had available for this study , hence we have focused only on one band at a time . if multiband light curves are available , one could do a combined fit and further improve the constraints . in any case , a future optical observation of a sn triggered by a neutrino detector like icecube or antares should start early enough to capture the initial shock breakout . if the initial shock breakout is not observed , and the first observed point on the light curve is more than 1 - 2 days after the actual explosion , the fits give large uncertainties in the explosion time .
this illustrates the cardinal importance of having fast follow - up capabilities in place to perform too observations .
this work shows that the optical data can be fit accurately using the formulation developed above , and that by doing so the statistical significance of the coincidence can not only be quantified but also significantly improved .
the authors would like to thank l. s. finn , m. d. kistler , p. mszros and especially j. f. beacom for helpful and informative discussions .
dfc thanks the deutscher akademischer austausch dienst ( daad ) visiting researcher program and the fulbright scholar program , and acknowledges the support of the national science foundation ( nsf ) .
af and mk acknowledge the support of the deutsche forschungsgemeinschaft ( dfg ) . | core - collapse supernovae are among the prime candidate sources of high energy neutrinos . accordingly , the icecube collaboration has started a program to search for such a signal .
icecube operates an online search for neutrino bursts , forwarding the directions of candidate events to a network of optical telescopes for immediate follow - up observations . if a supernova is identified from the optical observations , in addition to a directional coincidence a temporal @xmath0-@xmath1 coincidence also needs to be established . to achieve this ,
we present a method for estimating the supernova explosion time from its light curve using a simple model .
we test the model with supernova light curve data from sn1987a , sn2006aj and sn2008d and show that the explosion times can be determined with an accuracy of better than a few hours . |
-0.3 cm topological quantities are fundamental to describe low dimensional quantum liquids where standard symmetry breaking do not have a primal importance @xcite .
typical examples are quantum hall liquids @xcite and anisotropic superconductors with time - reversal symmetry breaking @xcite .
recently spin hall conductance for semiconductors are also attracting much current interest @xcite .
the ( spin ) hall conductance has a characteristic geometrical meaning @xcite . in some physical units , they are given by the first chern number of the berry connection @xcite . in practical numerical calculations , we diagonalize hamiltonians on a set of discrete points on the brillouin zone ( bz ) .
it is thus crucial to develop an efficient method of revealing the topological property of infinite systems with continuum bz from corresponding finite systems with discrete bz .
we propose an efficient method for the calculation of the chern numbers on a discretized bz based on a geometrical formulation of topological charges in lattice gauge theory @xcite .
the chern numbers thus obtained are _ manifestly gauge - invariant _ and _ integer - valued _ even for a discretized bz .
one can compute the chern numbers using wave functions in _ any gauge _ or _ without specifying gauge fixing - conditions_. details of the formulation and the basic results were published elsewhere @xcite .
-0.3 cm : let us consider chern numbers in the quantum hall effect as a typical example .
the spin hall conductances is treated similarly .
we take the bz by @xmath0 ( @xmath1 , 2 with integers @xmath2 ) .
since the hamiltonian @xmath3 is periodic in both @xmath4 and @xmath5 directions , the bz is regarded as a two - dimensional torus @xmath6 .
when the fermi energy lies in a gap , the hall conductance is given by @xmath7 , where @xmath8 denotes the chern number of the @xmath9th bloch band , and the sum over @xmath9 is restricted to the bands below the fermi energy @xcite .
the chern number assigned to the @xmath9th band is defined by @xmath10 with @xmath11 , where the berry connection @xmath12 with @xmath13 is defined by a normalized wave function of the @xmath9th bloch band @xmath14 satisfying @xmath15 and @xmath16 @xcite .
the chern number can be nonzero only when the gauge potential can not be defined as a global function over @xmath6 . in this case , one covers @xmath6 by several coordinate patches and then , within each patch , one can take a local gauge ( a phase convention for the wave functions ) such that the gauge potential is a well defined function . in an overlap between two patches , gauge potentials defined on each patch are related by a @xmath17 gauge transformation : @xmath18 and @xmath19 @xmath20 . in the continuum theory , one needs to fix this gauge freedom to perform any explicit calculations . to fix the gauge ,
one first selects an arbitrary state @xmath21 which is globally well defined over the whole bz @xcite .
then the gauge can be specified by @xmath22 ( if @xmath23 ) , where @xmath24 and @xmath25 are gauge independent .
generically , this gauge is only allowed locally because the overlap @xmath26 may vanishes .
it inevitably occurs to have a non vanishing chern number .
then one needs to use several different state @xmath27 ( gauges ) in several regions of patches to cover the whole bz .
: we now switch to the finite systems with discrete bz . a naive replacement of the differential operator to the difference operator breaks the gauge invariance and topological characters of the chern numbers .
here we propose an explicitly gauge invariant and topological definition of the chern number on a lattice .
let us denote lattice points @xmath28 ( @xmath29 , , @xmath30 ) on the discrete bz as @xmath31 , @xmath32 , @xmath33 .
we assume that the state @xmath34 is periodic on the lattice , @xmath35 , where @xmath36 is a vector in the direction @xmath37 with the magnitude @xmath38 .
we first define a @xmath17 link variable from the wave functions @xmath39 where @xmath40 .
it is well defined as long as @xmath41 .
( it is always assumed by an infinitesimal shift of the lattice . )
next we define a lattice field strength by @xmath42 ( @xmath43 ) .
the field strength is defined within the principal branch of the logarithm .
finally , we define the chern number on the lattice as @xmath44 we stress here that @xmath45 is manifestly _ gauge - invariant _ and _ strictly an integer _ for arbitrary lattice spacings . to demonstrate it ,
let us introduce a gauge potential @xmath46 ( @xmath47 which is periodic on the lattice : @xmath48 . by definition
, one finds @xmath49 where @xmath50 is the forward difference operator on the lattice , @xmath51 , and @xmath52 is an _ integer - valued _ field , which is chosen such that @xmath53 takes a value within the principal branch .
now we have @xmath54 it shows that the lattice chern number @xmath45 is an integer . to avoid ambiguities we assume that there are no exceptional configurations in the system under consideration , @xmath55 for all @xmath28 @xcite .
this condition will be referred to as _
admissibility_. under this condition , the chern number is uniquely determined .
since our lattice formulation recovers the continuum one in the limit @xmath56 , we expect the lattice field strength @xmath57 will be small enough for a sufficiently large @xmath58 and the lattice chern number will approach the one in the continuum @xmath59 in the this limit . since both @xmath45 and @xmath8
are integers , we have @xmath60 for meshes of appropriate sizes .
as far as the system is regular , our lattice formulation could reproduce correct chern numbers of the continuum theory even for a coarsely discretized bz , @xmath61 @xcite .
also our method can be extended @xcite to the case of the non - abelian berry connection @xmath62 , which is a matrix - valued one - form associated with a multi - dimensional multiplet @xmath63 @xcite .
: it turns out that the description presented so far proves to be _
unique _ for discretized bz .
namely , under the admissibility , the space of @xmath17 link variables is divided into disconnected sectors and the topological number @xmath45 is uniquely assigned to each sector . the chern number @xmath45 is , moreover , a _
unique _ gauge - invariant topological integer which can be assigned to admissible @xmath17 link variables .
the proof of this statement has been given by lscher @xcite .
a key ingredient of his proof is that any u(1 ) link variable can be decomposed uniquely into @xmath64}(k_\ell)+ia_\mu^{\rm t}(k_\ell ) } \lambda(k_\ell)u^{[w]}_\mu(k_\ell)\lambda(k_\ell+\hat\mu)$ ] , where @xmath65}(k_\ell)$ ] is a gauge field giving rise to the chern number ; @xmath66}(k_\ell)-\delta_2a_1^{[\tilde c_n]}(k_\ell ) = 2\pi i\tilde c_n/(n_1n_2)$ ] , @xmath67 is a periodic transverse field , @xmath68 is a gauge transformation , and @xmath69}_\mu(k_\ell)$ ] is a gauge field giving nontrivial wilson lines but giving zero field strength .
once @xmath65}(k_\ell)$ ] is found , @xmath67 can be determined uniquely by the relation @xmath70 .
readers who are interested in the proof should refer lscher s paper .
the admissibility condition is important for the present lattice formulation and its breakdown is closely related to a singular behavior of the berry connection , that is , it is supplemented with a quantum phase transition . in the present context of the berry connection
, a distribution of the gauge invariant field @xmath57 is completely governed by the @xmath71 dependence of the hamiltonian .
each of the topological ordered states with a nontrivial chern number corresponds to nontrivial topological sector specified by the admissibility and also characterized by the lattice chern number @xmath45 . in the continuum ,
on the other hand , the topological stability of the chern number is assured by the gap - opening condition @xcite .
the topological quantum phase transitions are thus characterized by the gap closing .
namely , nontrivial topological sectors of the continuum , each of which is a topological ordered state , are separated by the gaps . in passing , we would like to mention the kubo formula which is often used in numerical calculations : @xmath72 ^ 2}$ ] .
this formula is equivalent the field strength @xmath73 in the continuum bz , but _ not _ to @xmath74 in the discretized bz . in the latter case
, @xmath75 is no longer topological , although the gauge - invariance remains .
therefore , our topological and gauge - invariant formulation for discrete bz should have strong advantages , especially of describing regions where topological transitions occur .
-0.3 cm to demonstrate the present scheme , we take the hamiltonian for spinless fermions in an external magnetic field : @xmath76 , where the flux per plaquette on the coordinate lattice @xmath77 is @xmath78 .
we consider a model with quantum phase transition , that is , with next nearest neighbor ( nnn ) hopping @xmath79 ( nearest neighbor hopping is @xmath80 ) .
the hamiltonian in the @xmath71-space is given by @xmath81 , where @xmath82 ( @xmath83 , @xmath84 , , @xmath85 ) with @xmath86 and @xmath87 @xcite .
bellow , we will present some results of applying our method to the middle subband of the @xmath88 ( that is , @xmath89 ) system . for simplicity , we set @xmath90 and @xmath91 .
-0.2 cm we show the integer field @xmath52 in figs .
[ f : nfie](a)-(d ) , where we use the global gauge specified by the states @xmath97 .
the black and white circles denote @xmath96 and @xmath98 , respectively , whereas a blank implies @xmath99 .
it is clear that any of them gives the correct chern number @xmath100 . the field @xmath52 is gauge - dependent , but their sum is gauge - invariant .
it also shows a convergence of the gauge dependent field @xmath101 in the limit @xmath102 .
although we have fixed the gauge to calculate the field @xmath101 , calculations of the chern numbers are performed in _
any gauge_. we do _ not _ need specific gauge - fixing to make the gauge connection smooth . an _
arbitrary _ gauge ( e.g. , a phase choice of eigenvectors given by a numerical library ) can be also adopted to compute the chern number .
now let us change a ratio @xmath103 in the nnn model .
[ f : nspec ] , we show the energy spectra of the nnn models .
they are modified hofstadter s butterfly diagrams .
the integers in the figures are the hall conductances when the fermi energy lies in the gap , that is , it is a sum of the chern numbers below the energy gap . due to the sum rule of the chern number @xcite , the chern number of the specific energy band is given by a difference of two integers , above and below the band .
as is known @xcite , the nnn model shows a topological quantum phase transition accompanying a discrete change of the chern numbers .
for example , an increase of @xmath103 causes a rearrangement of the gap connectivity near @xmath104 at @xmath88 .
one can see that the energy gap just above the second band at @xmath88 is rearranged by the small change of @xmath103 .
it is associated with a change of the gap labeling from @xmath105 to + 2 . since the large energy gap just below the middle band is stable and the gap is labeled as + 1 , the chern number of the middle band changes from @xmath106 to + 1 .
this topological change of the chern number is confirmed by a calculation of the gauge dependent field @xmath101 in figs.[f : nf1](b ) and ( d ) .
the field strength @xmath57 just before and the after the quantum transition in also shown in ( a ) and ( c ) .
a small change of the parameter @xmath103 does not change the spectrum so much . on the other hand ,
the field strength @xmath57 is singular at the critical point .
it is consistent with the breakdown of the admissibility as expected .
that is , the quantum phase transition ( of the present lattice model ) occurs at @xmath112 ( @xmath113 ) .
this quantum phase transition is associated with a topological change of the integer value field @xmath101 as shown .
this work was supported in part by grant - in - aid for scientific research from jsps . | it is widely accepted that topological quantities are useful to describe quantum liquids in low dimensions .
the ( spin ) hall conductances are typical examples .
they are expressed by the chern numbers , which are topological invariants given by the berry connections of the ground states .
we present a topological description for the ( spin ) hall conductances on a discretized brillouin zone . at the same time
, it is quite efficient in practical numerical calculations for concrete models .
we demonstrate its validity in a model with quantum phase transitions .
topological changes supplemented with the transition is also described in the present lattice formulation . , and chern number , berry connection , hall conductance , lattice gauge theory |
the strangeness meson photoproduction off the nucleon target is one of the most well - studied experimental and theoretical subjects to reveal the hadron production mechanisms and its internal structures , in terms of the strange degrees of freedom , breaking the flavor su(3 ) symmetry explicitly .
together with the recent high - energy photon beam developments in the experimental facilities , such as lpes2 at spring-8 and clas12 at jefferson laboratory @xcite , higher - mass strange meson - baryon photoproducitons must be an important subject to be addressed theoretically for future studies on those reaction processes . in the previous works @xcite ,
the @xmath0 photoproduction was investigated , the born approximation being used with the regge contributions . in comparison with the preliminary experimental data
@xcite , the theory reproduced the data qualitatively well , but the theoretical cross - section strength was underestimated in the vicinity of the @xmath4 . in the present talk ,
we want to report our recent study to explain this discrepancy observed in the previous work .
based on the theoretical framework as employed in ref .
@xcite , we include the nucleon resonances in the @xmath7-channel baryon - pole contribution . as for the nucleon resonances
@xmath8 , we take into account nine of them , i.e. @xmath9 , @xmath10 , @xmath11 , @xmath12 , @xmath13 , @xmath14 , @xmath15 , @xmath9 , @xmath16 , and @xmath17 , in a full relativistic manner .
we start with explaining the theoretical framework briefly and represent the important numerical results in our study .
we note that the nucleon resonances are carefully taken into account in a full - relativistic manner , in addition to the born diagrams , @xmath18 and @xmath19 meson - exchanges in the @xmath20 channel , and @xmath21 and @xmath22 hyperon exchanges in the @xmath23 channel , which were already employed in ref .
all the effective interaction vertices for the nucleon resonances are taken from ref .
for instance , the invariant amplitudes for the spin-@xmath24 and spin-@xmath25 resonance contributions in the @xmath7 channel can be written as follows : @xmath26 \cr & & \frac{ie_q\mu_{n^*}}{2m_n}\gamma^{\mp}(\not{k_1}+\not{p_1}+m_{n^ * } ) \gamma^{\mp}\sigma^{\mu{i}}k_{1i}u_n(p_1){\varepsilon}_{\mu}(k_1 ) , \cr \mathcal { m}_{n^*}\left(\frac{3}{2}^\pm \right)&= & \frac{g_{{k^*}n^*\lambda}}{s - m_{n^*}^2 } { \varepsilon}_{\nu}^*(k_2){\bar u}_{\lambda}(p_2)({k_2^\beta}g^{{\nu}i}-k_2^ig^{\nu\beta } ) \frac{e_q}{2m_{k^*}}\gamma_i^{\pm}\delta_{\beta\alpha } \cr & & \left[\frac{\mu_{n^*}}{2m_n}\gamma_j\,\mp\,\frac{\bar \mu_{n^ * } } { 4m_n^2}p_{1j } \right]\gamma^{\pm}({k_1^\alpha}g^{{\mu}j}-{k_1^j } g^{\mu\alpha})u_n(p_1){\varepsilon}_{\mu}(k_1),\end{aligned}\ ] ] where @xmath27 stand for the @xmath28 momenta and @xmath29 for the helicity amplitude .
the @xmath30 and @xmath31 denote the strong vector and tensor coupling strengths , respectively .
the polarization vectors for the photon and @xmath32 are assigned as @xmath33 and @xmath34 .
the @xmath35 controls the parity of the relevant resonances in the following way : @xmath36 relevant parameters for the resonances are estimated using the experimental and theoretical information @xcite . in figure
[ fig1 ] , we show the total cross section for the present reaction process , i.e. @xmath37 . the numerical results are drawn separately for the cases including the @xmath2 ( left ) and @xmath3 ( right ) , varying the coupling constants @xmath30 . as shown in the figure , the cross - section enhancement is observed in the vicinity of the threshold region , if we choose the strong coupling strengths as @xmath38 .
we verified that other nucleon resonances are not so effective to interpret the discrepancy . [ cols="^,^ " , ]
in the present work , we have studied the @xmath39 photoproduction theoretically , employing the tree - level born approximation and nucleon - resonance contributions below the @xmath4 . among the nucleon resonances , the @xmath2 and @xmath3 play a dominant role to reproduce the experimental data .
it also turns out that other @xmath8 contributions are not so effective to improve the theoretical results .
we note that the nucleon resonances beyond the @xmath4 may contribute to the threshold enhancement , especially due to the @xmath40 , since the @xmath4 for the present reaction process is about @xmath41 mev .
related works are under progress and appear elsewhere .
the authors are grateful to a. hosaka for fruitful discussions .
the present work is supported by basic science research program through the national research foundation of korea ( nrf ) funded by the ministry of education , science and technology ( grant number : 2009 - 0089525 ) .
the work of s.i.n . was supported by the grant nrf-2010 - 0013279 from national research foundation ( nrf ) of korea .
99 leps2 ( http://www.hadron.jp ) and clas12 ( http://www.jlab.org/hall-b/clas12 ) y. oh and h. kim , phys .
c * 74 * , 015208 ( 2006 ) .
y. oh and h. kim , phys . rev .
c * 73 * , 065202 ( 2006 ) .
l. guo and d. p. weygand [ clas collaboration ] , arxiv : hep - ex/0601010 .
y. oh , c. m. ko and k. nakayama , phys . rev .
c * 77 * , 045204 ( 2008 ) .
k. nakamura [ particle data group ] , j. phys .
g * 37 * , 075021 ( 2010 ) .
s. capstick and w. roberts , phys
. rev .
d * 58 * , 074011 ( 1998 ) .
s. capstick and w. roberts , prog .
* 45 * , s241 ( 2000 ) . | in this presentation , we report our recent studies on the @xmath0 photoproduction off the proton target , using the tree - level born approximation , via the effective lagrangian approach . in addition , we include the nine ( three- or four - star confirmed ) nucleon resonances below the threshold @xmath1 mev , to interpret the discrepancy between the experiment and previous theoretical studies , in the vicinity of the threshold region . from the numerical studies , we observe that the @xmath2 and @xmath3 play an important role for the cross - section enhancement near the @xmath4
. it also turns out that , in order to reproduce the data , we have the vector coupling constants @xmath5 and @xmath6 . |
a logarithmic upturn in the resistivity at low temperature has been observed in graphene with vacancies@xcite .
a fit to the temperature dependence of resistivity with conventional kondo effect yields a large kondo temperature ( with @xmath0 ) which shows a non - monotonic behavior with respect to the gate voltage@xcite .
the vacancies in the graphene sheets are induced by ion irradiation in ultra - high vacuum and the magnetism in sputtered graphite has been experimentally observed@xcite . our goal is to study whether kondo effect alone in graphene can explain the experimental results in ref ..
we start with the anderson impurity model@xcite to study the impurities effect on transport . in the local moment regime
where the impurity occupation for a given spin @xmath1 we use schrieffer wolff transformation@xcite to write down the kondo model from the anderson impurity hamiltonian .
since we are interested in the resistivity due to impurity spin fluctuations we study the kondo model by standard perturbation method .
the perturbative approach fails at @xmath2 with @xmath3 representing kondo temperature but works for @xmath4 .
we compute the scattering rates in this weak coupling regime .
the scattering rates are determined via perturbative calculations of the @xmath5-matrix@xcite .
kondo effect in the pseudogap system has been explored in the context of graphene as well as in that of d - wave superconductor@xcite via various different approach such as nrg or mean field approach@xcite .
the advantage of our approach is the ability to determine the high temperature behavior of the scattering rate and resistivity accurately within perturbation theory .
we assume a dilute concentration of impurities and ignore the spin spin interactions such as the ruderman - kittel - kasuya - yosida ( rkky ) interaction . in graphene these interactions ,
in addition to being oscillatory with distance between impurities , depend on the sublattice on which the impurities are located@xcite . for chemical potential at the dirac point
our results are in agreement with the prediction of the existence of an intermediate coupling fixed point@xcite . near the node
the exchange coupling @xmath6 needs to be larger than a critical value @xmath7 to have the kondo effect .
the dependence of @xmath3 on the chemical potential is qualitatively different for @xmath8 and @xmath9 . for impurities breaking the lattice symmetry ,
a power law in @xmath5 divergence of scattering rate is obtained for @xmath8 while a logarithmic divergence appears for @xmath9 . for impurities preserving the lattice symmetry ,
a power law in @xmath10 divergence of scattering rate is obtained for @xmath8 while a similar logarithmic divergence appears for @xmath9 . for both cases
the scaling of resistivity with single kondo temperature breaks down in the vicinity of the dirac point .
our results for kondo temperature obtained within the @xmath5-matrix formalism is in agreement with the mean field results for the development of the kondo phase@xcite .
the resistivity obtained displays different chemical potential dependences . for impurities breaking the lattice symmetry
the resistivity decreases as chemical potential increases while for impurities preserving the lattice symmetry the resistivity increases as chemical potential increases .
for the same set of physical parameters the dominant source of resistivity is from impurities which breaks the lattice symmetry .
we also explore the region near the empty orbital to mixed valence one in the anderson impurity model to find the resistivity due to charge fluctuations . from the numerical rg@xcite
the kondo effect is suppressed as the critical exchange coupling @xmath11 for chemical potential close to the dirac point .
thus we use unrestricted hatree fock method@xcite on the anderson impurity model to find the resistivity near the empty orbital regime .
the resulting resistivity shows similar dependence on chemical potential as well as dominance from symmetry breaking impurities as the resistivity obtained in the kondo model . near the node
the kondo scale , extracted from the logarithmic temperature dependence region on resistivity , yields a kondo temperature comparable to the observations in the experiment in ref . while away from the dirac point the extracted kondo scale is higher than experiment by one order of magnitude . by combining the charge fluctuation effect for @xmath12 and kondo effect ( spin fluctuations ) for finite @xmath13
we obtain kondo temperature dependence on @xmath13 qualitatively consistent with experimental results@xcite with gate voltage less than @xmath14 .
our conclusion is that the observed experimental results , albeit fitted well by numerical rg for conventional metal kondo model@xcite , can not be solely explained by kondo screening in all range of chemical potential . for chemical potential near the node the charge fluctuations
is responsible for the observed resistivity temperature dependence while away from the node the spin fluctuations take over .
this article is organized as following : we start with the anderson impurity hamiltonian to describe dilute impurities physics in the graphene system . to study the local moment regime we use schrieffer wolff transformation to obtain kondo model from anderson hamiltonian . in section
@xmath15 we evaluate resistivity due to spin fluctuations , with different impurity locations , by perturbation computations on the kondo model . in section
@xmath16 we compute resistivity due to charge fluctuations when impurity occupation is close to zero by using mean field approach on the anderson model . in section @xmath17 we show numerical results of temperature dependence of the resistivity with different symmetry and mechanism . in section
@xmath18 we compare our results with the experiment in ref .. the results are summarized in section @xmath19 .
two appendixes contain derivations for the perturbative results in the kondo model .
we start from graphene hamiltonian in the presence of dilute impurities described by the anderson impurity hamiltonian@xcite @xmath20d_s+h.c.\nonumber\end{aligned}\ ] ] @xmath21 is the nearest hopping in the momentum space with @xmath22 being the nearest neighbor hopping strength .
@xmath13 defines the fermi level measured from the dirac point . @xmath23 and @xmath24 are the particle creation operators on the a and b sublattices .
@xmath25 with @xmath26 , @xmath27 , and @xmath28 being the nearest neighbor lattice vector .
@xmath29 is the lattice constant .
@xmath30 describes the hybridization between the impurity level and graphene electrons with @xmath31 .
@xmath32 describes the coulomb repulsion on the impurity level and @xmath33 is the hamiltonian describing the localized level of @xmath34 electron .
we diagonalize @xmath21 by defining @xmath35 . in this basis
the @xmath21 term becomes @xmath36 with @xmath37 denoting the conduction and valence bands .
the hybridization term @xmath30 in this rotated basis is @xmath38\end{aligned}\ ] ] with @xmath39 .
denote @xmath40 as the energy of the bands evaluated from chemical potential @xmath13 and @xmath41 as combinations of momentum , spin , and @xmath42 the band index .
the anderson impurity hamiltonian describing the impurity in the graphene can be written as @xmath43\end{aligned}\ ] ] to explore the local moment regime where impurity occupation for a given spin @xmath1 we perform schrieffer wolf transformation to project out the charge degree of freedom@xcite .
the exchange hamiltonian or kondo model obtained after this transformation with the additional term @xmath44 describing spin spin interaction at different sites is given by @xmath45 with @xmath46 here @xmath47 and @xmath48 .
the interaction between impurity spins @xmath49 are added for the inclusion of spin spin interaction but is assumed to be small due to small concentration of the impurities in this article . including this term would lead to a time dependent impurity spin via @xmath50 with @xmath51 being the imaginary time@xcite . for impurities preserving the @xmath52 point group symmetry of the triangular sublattice in the graphene system the factor @xmath53 while for impurities breaking the symmetry the factor @xmath54 is a constant . to evaluate the resistivity due to spin fluctuations we use eq.([kondo ] ) as the starting hamiltonian .
we use perturbation expansion on the one particle green function s t - matrix to compute scattering rate and from boltzmann transport to obtain linear response resistivity in both impurity breaking and preserving the lattice symmetry cases .
we use mean field approach on the anderson impurity model shown in eq.([anderson ] ) to obtain resistivity due to charge fluctuations in both symmetry breaking and preserving case .
the following two sections are the computation results for each cases mentioned above .
to study the resistivity due to spin fluctuations we start with the kondo hamiltonian shown in eq.([kondo ] ) .
we calculate transport properties from the @xmath5-matrix which is related to single particle green s function by@xcite @xmath55 where @xmath56 and @xmath57 , @xmath58 are impurity spin state indices .
this expression is related to time dependent green s function by @xmath59 with @xmath60 and @xmath61 being integers ( we put boltzmann constant @xmath62 to simplify the notation ) . based on perturbation in @xmath49 this time dependent green s function can be written as @xmath63_{\alpha',\alpha}\\\nonumber & & \langle t_{\tau}(c_{k ' s ' m}(\tau)\bar{c}_{k s n}(\tau')\bar{c}_{k_1 ' s_1 ' m_1}(\tau_1)c_{k_1 s_1 n_1}(\tau_1)\ldots \bar{c}_{k_n's_n'm_n}(\tau_n)c_{k_n s_n n_n}(\tau_n)\rangle_{h_e}\end{aligned}\ ] ] here @xmath64 and @xmath65 are the time ordering operators and @xmath66 is the s - matrix . the first order in @xmath6 and @xmath67
is given by @xmath68\end{aligned}\ ] ] here @xmath69 with @xmath70 .
for @xmath71 we can simplify above expression by noting that @xmath72 and we get @xmath73 with @xmath74 .
the general second order result of t - matrix , with @xmath75 and @xmath76 , is expressed as @xmath77\}_{\alpha'\alpha}\end{aligned}\ ] ] to focus on the kondo contribution to the scattering rate we may set @xmath78 and we set @xmath79 by assuming dilute impurities . for rkky type of spin spin interactions the interaction strength decays as @xmath80 for symmetry breaking or @xmath81 for symmetry preserving case@xcite .
thus for sufficient dilute impurities we may treat @xmath79 . in this limit eq.([2ndop ] ) is simplified to @xmath82 with @xmath83\end{aligned}\ ] ] for noninteracting spins the third order perturbation , after taking a trace over conduction electron spins and approximating the reduction of three - spin correlation functions to two - spin correlation functions@xcite , is given by @xmath84\end{aligned}\ ] ] from eq.([2nd ] ) and eq.([3rd ] ) we may define a general function @xmath85 which we need to evaluate in computing the @xmath5-matrix @xmath86 by using @xmath87 we may write the continuous form of eq.([expr ] ) as @xmath88 where @xmath89 is the linear spectrum cutoff . in the following
we will separate the discussions into two cases@xcite : one with impurity interactions breaking the @xmath52 lattice symmetry , in which @xmath90 , and we denote @xmath91 in this case .
another with impurity interactions preserving the symmetry , in which @xmath92 , and we denote @xmath93 in this case . for the case of symmetry breaking the cutoff scheme we choose for a linear density of states with a cutoff
@xmath89 is multiplying the argument of right hand side of eq.([exprc ] ) by @xmath94 and extend the integration limit from @xmath95 to @xmath96 .
the resulting @xmath97 , with details shown in appendix .
[ a ] , is @xmath98-\psi(\frac{1}{2}-i\frac{\beta z}{2\pi})\frac{2\lambda^2(z+\mu)}{(z+\mu)^2+\lambda^2}\big\}\end{aligned}\ ] ] here @xmath99 is defined as @xmath100 and @xmath101 is the digamma function .
analytic forms can be obtained in two asymptotic limits by using the asymptotic forms of the digamma function .
for @xmath102 we have @xmath103\end{aligned}\ ] ] where @xmath104 is the euler constant and @xmath105 is the riemann zeta function evaluated at 2 . in the limit @xmath106
we get @xmath107\end{aligned}\ ] ] in eq.([bml ] ) and eq.([bmg ] ) we have assumed @xmath108 . using the boltzmann equation with relaxation time hypothesis@xcite and noticing that the honeycomb symmetry is broken by the impurity we find the scattering rate is related to the @xmath5-matrix by @xmath109\end{aligned}\ ] ] the second line of eq.([reltimesb ] ) represents the scattering process related to different dirac points in the brillouin zone and the third line is the scattering event within the same dirac cone .
we have used the fact that @xmath110=-\pi\sum_{k'}|t_{k , k'}(\epsilon^+)|^2\delta(\epsilon-\epsilon_{k'})$ ] and @xmath111 is independent of angle @xmath112 between momenta @xmath113 and @xmath114 in the symmetry breaking case in the above equation . the scattering rate , @xmath115 with @xmath116 being the relaxation time , to third order is @xmath117\big]\end{aligned}\ ] ] the kondo effect
is reflected in the divergence of the relaxation time in the parquet approximation .
this involves treating the cubic term as the first in an infinite series which is summed to give @xmath118}\end{aligned}\ ] ] plotted as a function of temperature @xmath119 in the unit of @xmath120 .
energy cutoff @xmath121 and @xmath122 . ]
we are mainly interested in the dc response so we shall study the relaxation time when @xmath123 .
fig.[fig1 ] shows the function @xmath124 plotted as a function of temperature for different chemical potential .
we define the kondo temperature as the temperature when the relaxation time diverges when @xmath123 . for the case
@xmath102 the singularities from @xmath125\simeq 0 $ ] can be expressed , by using eq.([bml ] ) , as @xmath126\simeq -1\\\label{tksbs } & & \rightarrow t_k=-\frac{\pi\lambda}{4\gamma+4\ln(2)-4 - 3\zeta(2)}\big(1-\frac{j_c}{j}\big)\end{aligned}\ ] ] where @xmath127 and @xmath128 .
thus for chemical potential @xmath129 we have no kondo effect if @xmath130 .
as one increases the chemical potential we may include the linear order of @xmath13 in eq.([bml ] ) and obtain the expression for kondo temperature as @xmath131 where we have used @xmath132 and @xmath133 .
thus the kondo temperature increases with increasing chemical potential . in the opposite limit where @xmath106 we use eq.([bmg ] ) to obtain the kondo temperature as @xmath134\end{aligned}\ ] ] .
energy cutoff @xmath121 and @xmath122 gives @xmath135 . ] where @xmath136 .
eq.([logtk ] ) can be expressed as @xmath137 with @xmath138 is the electron density of states of graphene .
compared with the kondo temperature in conventional metal @xmath139 there exists critical value of exchange coupling for kondo effect to be realized in this two dimensional pseudo gap system .
fig.[fig2 ] shows the kondo temperature as a function of the chemical potential for various values of @xmath6 . for @xmath140
smaller than @xmath141 , the kondo temperature is smaller than the chemical potential for the range shown . in this regime @xmath3
is given by eq.([logtk ] ) and is exponentially smaller than the energy scale set by the chemical potential .
as @xmath140 approaches one from below , the kondo temperature grows faster than the chemical potential . as @xmath142
the exponential behavior cross over to the linear dependence shown in eq.([tksbm ] ) . given the relaxation time we obtain the linear response conductivity as @xmath143 for small chemical potential or @xmath102 we use eq.([bml ] ) and eq.([tksbm ] ) and approximate @xmath144 .
we get the resistivity at low chemical potential as @xmath145 with @xmath146 for temperature @xmath147 but higher than @xmath3 the same approximation scheme gives @xmath148 thus we see that for @xmath149 the kondo contribution to resistance is not determined by a single scale @xmath3 . for temperature ranged between @xmath147 the scaling of
the resistivity goes like @xmath150 .
this power law behavior indicates that at sufficient low chemical potential the magnetic impurities are not completely quenched while a logarithmic behavior is expected in the conventional metal case . for large @xmath13 or @xmath106 a kondo effect similar to magnetic impurities in the conventional metals
is obtained . for large chemical potential
we approximate @xmath151 . under this approximation
the resistivity @xmath152 is given by @xmath153 use eq.([bmg ] ) for @xmath154 with @xmath155 we get @xmath156\right)^{-1}\\ & \simeq&\frac{3\pi n_i s(s+1)}{4}\left(\frac{j}{2\mu}\right)\frac{h}{e^2}\left(\ln\left[\frac{t_k}{t}\right]\right)^{-1}\end{aligned}\ ] ] for the case of impurities preserving the symmetry of honeycomb lattice the cutoff scheme we choose for a linear density of states with a cutoff @xmath89 is multiplying the argument of right hand side of eq.([exprc ] ) by @xmath157 and extend the integration limit from @xmath95 to @xmath96 .
the resulting @xmath158 , with details shown in appendix [ b ] , is @xmath159\big\}\end{aligned}\ ] ] analytic forms of @xmath158 is obtained by taking the asymptotic behavior of digamma function in the following two limits : @xmath160 and @xmath161 . for @xmath160
we have @xmath162 for the opposite limit @xmath161 we have @xmath163 similar to eq.([bml ] ) and eq.([bmg ] ) we have assumed @xmath108 . using eq.([reltimesb ] ) but with the appropriate relaxation times determined in this section , we compute the resistance . for this case @xmath110=-\pi\sum_{k'}|t_{k , k'}(\epsilon^+)|^2\delta(\epsilon-\epsilon_{k'})$ ] and
@xmath111 is independent of angle @xmath112 between momenta @xmath113 and @xmath114 since @xmath164 .
the scattering rate to third order is @xmath165\big]\end{aligned}\ ] ] the expression for the relaxation time @xmath166 , within the same approach as the previous section , is @xmath167}\end{aligned}\ ] ] plotted as a function of temperature @xmath119 in the unit of @xmath120 .
energy cutoff @xmath121 and @xmath122 . compared with fig.[fig2 ]
@xmath168 shows more temperature variations when @xmath169 , indicating the kondo effect can only be observed on lower temperature compared with symmetry breaking case . ]
the dc conductivity is related to the relaxation time with @xmath123 .
fig.[fig3 ] shows the function @xmath168 plotted as a function of temperature for different chemical potential .
@xmath168 shows small variations with temperature except when temperature is close to zero where exponential growth with decreasing temperature is observed .
for the case @xmath102 the singularities from @xmath170\simeq 0 $ ] can be expressed , by using eq.([bmml ] ) , as @xmath171 in above we have used the leading order correction as @xmath172 since its prefactor is @xmath173 which diverges as we take @xmath174 .
higher order expansion in @xmath173 shows it as a sum of an infinite series in power of @xmath175 with @xmath42 being some integer .
thus the infinite sum gives a factor of @xmath176 . .
energy cutoff @xmath121 and @xmath122 gives @xmath177 for the symmetry preserving case . ] in the opposite limit where @xmath106 we use eq.([bmmg ] ) @xmath178 where @xmath179 , and @xmath180 .
thus for both cases we obtain results similar to mean field results obtained in ref .. fig.[fig4 ] shows the kondo temperature as a function of chemical potential for various exchange coupling strength @xmath6 . for @xmath140
smaller than @xmath181 the kondo temperature is always smaller than the chemical potential for the range shown .
@xmath140 approaches one from below and for @xmath182 the exponential dependence on @xmath13 crosses over to a power law .
given the relaxation time we obtain the linear response conductivity as @xmath183 for @xmath102 , @xmath149 , and @xmath155 we use eq.([bmml ] ) and eq.([kltsp ] ) and again approximate @xmath144 .
the conductivity for @xmath184 has no analytic form as the integral in eq.([cond ] ) involves@xmath185 , which diverges as @xmath186 .
this vanishing resistivity for @xmath186 for impurities preserving the lattice symmetry is due to the fact that the scattering rate goes to zero faster than the chemical potential at the node .
thus the contribution to scattering near the node is dominated by other sources of scattering as compared to exchange scattering of impurities that preserve the lattice symmetry .
for @xmath102 , @xmath187 , and @xmath155 we use eq.([bmml ] ) and eq.([kltsp ] ) and use the same approximation scheme as above we obtain the resistivity as @xmath188 thus even for larger chemical potential the resistivity does not scale solely with kondo temperature .
the behavior for @xmath189 goes like @xmath190 .
for large chemical potential or @xmath191 we approximate @xmath192 . combining with eq.([bmmg ] ) and eq.([khtsp ] )
we get @xmath193\right)^{-1}\end{aligned}\ ] ] thus at large chemical potential the kondo contribution to resistivity is similar in the form as the magnetic impurities in conventional metal .
from @xmath194 , for large coulomb repulsion @xmath195 , it follows that to obtain @xmath196 the impurity level @xmath197 must be close to the fermi surface @xmath13 .
since the density of state in the graphene is proportional to the energy scale away from this fermi surface , or @xmath198 with @xmath199 denoting graphene density of state , the phase space for charge fluctuation is very small and the local moment region is large compared with the case of the magnetic impurities in the conventional metal@xcite .
however it is still likely to have impurity level close to @xmath13 which is not in the local moment region@xcite .
thus it is worthwhile to estimate the resistivity contribution from impurity charge fluctuations .
we use mean field approach on the anderson impurity model shown in eq.([anderson ] ) and rewrite @xmath200 with @xmath201 determined self consistently , to obtain the impurity green s function . from the imaginary part of this green s function we obtain temperature dependence of the linear response resistivity by assuming boltzmann transport . under this mean field approach
we obtain the retarded impurity green s function as@xcite @xmath202 the self energy part @xmath203 is given by @xmath204 in above we have used @xmath205 .
@xmath206 for symmetry breaking case and @xmath207 for symmetry preserving case .
we take the principal part of @xmath203 between @xmath208 with @xmath89 being the linear spectrum cutoff . in the non - magnetic mixed valence regime ,
of which we are interested in , @xmath209 .
the impurity occupation @xmath210 is given by @xmath211\end{aligned}\ ] ] by using eq.([dotgr ] ) and eq.([selfen ] ) we find the relation between @xmath197 and @xmath210 by solving self - consistent conditions numerically .
for impurities breaking the symmetry the self energy @xmath203 obtained from eq.([selfen ] ) is given by @xmath212\end{aligned}\ ] ] since @xmath213 we use eq.([reltimesb ] ) and eq.([sbcond ] ) to obtain the impurity conductivity , denoted as @xmath214 .
the resistivity @xmath215 .
we are mainly interested in the leading order temperature dependence of the resistivity contributed by the charge fluctuation in the anderson impurity model .
thus we use the same approximation @xmath216 in eq.([sbcond ] ) to extract the leading order in temperature dependence .
the resistivity obtained for @xmath217 is @xmath218 here @xmath219 and @xmath220
. analytic result of resistivity for @xmath221 can also be obtained but the expression are cumbersome and we defer a numerical analysis to section v. from eq.([dotocp ] ) we find the non - magnetic region@xcite by demanding @xmath222 when @xmath223 . within this charge fluctuation regime ( @xmath224 )
we study the temperature variation of resistivity at a given @xmath13 and coulomb repulsion @xmath195 . for impurities preserving the honeycomb lattice symmetry
the self energy @xmath203 is given by @xmath225\end{aligned}\ ] ] we again use eq.([reltimesb ] ) and eq.([sbcond ] ) to obtain the impurity conductivity , denoted as @xmath226 .
the resistivity @xmath227 . for temperature dependence
we use the approximation @xmath216 in eq.([sbcond ] ) to extract the leading order . to perform this computation we need to find the @xmath228 dependence in the principal integral of eq.([sesp ] ) .
this is done by fitting numerically the principal value of the integral for large @xmath229 .
this is because the relevant integration region for @xmath228 in the expression of conductivity is @xmath230 which makes @xmath231 in our discussion . from the numerical fit with @xmath232 ( chosen for experimentally accessible range )
we have @xmath233 the conductivity obtained is @xmath234\end{aligned}\ ] ] for @xmath235 the resistivity @xmath236 similar to the case for spin fluctuation in eq.([cond ] ) . for @xmath189
we have @xmath237 here @xmath238 is the parameter from principal integral of the dot self energy .
we find the non - magnetic region from eq.([dotocp ] ) and study the temperature variation of resistivity within this charge fluctuation regime ( @xmath224 ) .
before we proceed to compare the temperature dependences of resistivity due to charge fluctuations , spin fluctuations , and the influence of impurities position , we pause here to discuss the regimes where mean field results are valid in this anderson impurity model .
the order parameter of this unrestricted hatree fock is the d level occupation for a given spin @xmath239 . to make comparison with exact d level occupation done by bethe
ansatz@xcite we need to go back to the case for conventional metal where the mean field results were done by p. w. anderson@xcite .
the d level occupation for a given spin @xmath240 at zero temperature is @xmath241 here @xmath242 with @xmath243 the density of state for conventional metal .
we solve eq.([dot ] ) in the non - magnetic region where @xmath244 and compare the answers with exact results obtained by bethe ansatz .
the comparison for d level occupation for a given spin v.s .
impurity level , with @xmath245 , is shown in fig.[fig5 ] . .
lines denoted ba means bethe ansatz results and lines denoted mf means mean field results .
the two match better in the region where @xmath246 .
this upper bound increases with decreasing @xmath247 . ] from this figure we can see that the mean field results deviate from exact ones in a range @xmath248 for the range of @xmath249 , indicating that mean field is a good approximation when the impurity level @xmath197 is higher than the fermi energy @xmath250 or , in the other words , the impurity is nearby the empty orbital region . for larger coulomb repulsion
@xmath247 the minimum value of @xmath197 of overlapping region is closer to @xmath250 . since for two dimensional system the mean field results are marginal , we expect the mean field result work for @xmath251 , as the s wave scattering in the conventional metal considered above@xcite is a one dimensional problem . since the crossover shifts to lower and lower values of @xmath252 as @xmath195 increases , this is a rough criterion but establishes a basis for the mean field calculations .
in sections iii and iv we have shown the analytic results of temperature dependence of resistivity for @xmath191 and @xmath253 . here
we compute numerically the temperature dependence of resistivity due to spin fluctuations , @xmath254 and @xmath255 for impurities breaking / preserving honeycomb lattice symmetry , and the temperature dependence of resistivity due to charge fluctuations , @xmath256 and @xmath257 .
we use the full form of @xmath258 and @xmath259 and extract the results for @xmath155 with @xmath3 obtained numerically the same way as we obtain the kondo temperature in fig.[fig2 ] and fig.[fig4 ] .
we compare the resistivity for different symmetry with the same sets of parameters .
the resistivity due to symmetry preserving impurities is much smaller than that of symmetry breaking case due to the factor of @xmath260 ( see eq.([relsb ] ) and eq.([relsp ] ) ) .
we examine the resistivity due to impurities spin and charge fluctuations in the symmetry breaking case and make comparison with the experimental results@xcite in the next section .
we use @xmath122 , @xmath121 , @xmath261 , and @xmath262 in all of the numerical results within this section .
we choose different impurity level @xmath197 to explore the resistivity due to spin and charge fluctuations .
the resistivity v.s .
temperature is evaluated numerically between @xmath263 .
let us first study the local moment region .
we choose @xmath264 to ensure the d level occupation @xmath1 .
the chemical potential @xmath13 is chosen between @xmath265 to @xmath266 . from @xmath267
this choice of parameters renders the exchange coupling strength @xmath268 . for both cases these exchange coupling strengths
are less than the critical value @xmath269 and the kondo temperature @xmath3 obtained for both cases are extremely small ( @xmath270 ) . with this choice of parameters @xmath271 and
therefore the analytic expression for @xmath3 corresponds to eq.([logtk ] ) for symmetry breaking case and eq.([khtsp ] ) for symmetry preserving case .
the resistivity v.s .
temperature are plotted in fig.[fig6 ] and fig.[fig7 ] . and various @xmath13 .
the exchange coupling strength @xmath272 is larger than @xmath273 ( for @xmath274 ) and less than @xmath275 ( for @xmath276 ) in the range chosen . in all cases
the kondo temperature is less than @xmath277 so the temperature range chosen is well above the kondo scale . ] and various @xmath13 .
the exchange coupling strength @xmath272 is larger than @xmath278 ( for @xmath274 ) and less than @xmath279 ( for @xmath280 ) in the range chosen . in all cases
the kondo temperature is less than @xmath277 so the temperature range chosen is well above the kondo scale . ] and various @xmath13 .
the d level occupation @xmath239 is greater than @xmath281 ( when @xmath280 ) and lesser than @xmath282 ( when @xmath283 ) in the chosen range . ]
chemical potential @xmath13 ( in unit of @xmath120 ) .
dots are the numerical data of conductivity at various chemical potential and the line is the fitted quadratic curve with @xmath284 .
this quadratic behavior is also seen in fig.3d of ref .. ] and various @xmath13 .
the d level occupation @xmath239 is greater than @xmath285 ( when @xmath286 ) and lesser than @xmath282 ( when @xmath283 ) in the chosen range . ] in fig.[fig6 ] we see the @xmath4 tails of the logarithmic upturns occurring when @xmath2 .
the resistivity goes down as chemical potential increases .
this tendency is quite different from the case of symmetry preserving ones , shown in fig.[fig7 ] .
the dependence of resistivity on chemical potential @xmath13 for symmetry preserving case shows @xmath287 by comparing the resistivity at @xmath288 in fig.[fig7 ] . at
temperature higher than the chemical potential the resistivity goes down with increasing temperature faster than the logarithmic tail for all cases in fig.[fig7 ] .
this is due to the divergence in conductivity when @xmath289 .
the order of magnitude of resistivity at the same temperature for symmetry preserving case is much smaller than the resistivity for symmetry breaking case .
thus we can safely ignore the contributions from symmetry preserving type of impurities when considering the resistivity due to spin fluctuations .
for the case of charge fluctuations we choose @xmath290 to ensure the d level occupation @xmath291 .
the chemical potential @xmath13 is chosen between @xmath265 to @xmath266 .
we compute the resistivity v.s .
temperature for @xmath292 numerically from the mean field results .
the resistivity v.s . temperature are plotted in fig.[fig8 ] and fig.[fig9 ] for symmetry breaking and symmetry preserving cases .
fig.[fig8 ] shows @xmath293 for @xmath294 and tends to a flat region for small temperature , which is very similar to the screening result of kondo effect at @xmath295 .
conductivity at @xmath296 shows quadratic chemical potential dependence , shown in fig.[fig8 - 1 ] , consistent with the gate voltage dependence on conductivity seen in the experiment@xcite .
the experimental fit in ref . for kondo scale , however , is about one order of magnitude smaller compared with the energy scale obtained in logarithmic temperature range in fig.[fig8 ] .
the temperature dependence of resistivity in this charge fluctuation regime is similar to that of kondo model in this case but the physics is not related to spin but charge fluctuation . to facilitate comparing our results with experimental ones in ref .
we refer to the energy scale as a kondo - like temperature in the following discussion .
fig.[fig9 ] also shows @xmath297 for @xmath221 with shorter range of temperature and similarly tends to a flat region for small temperature
. the resistivity increases with increasing chemical potential in fig.[fig9 ] similar to the case of spin fluctuation .
the resistivity for symmetry preserving case is much smaller than that for symmetry breaking impurities and thus ignore the contribution from symmetry preserving impurities . in summary
when both types of impurities are present the resistivity due to impurities preserving lattice symmetry is much smaller than that from impurities breaking the symmetry .
thus we focus our discussions on symmetry breaking cases for spin and charge fluctuations in the next section .
here we make comparisons with the experimental data given in ref .. we start with the perturbative results of kondo model in the case of impurities breaking the honeycomb symmetry . to have large kondo temperature ( @xmath298 in the experiment )
the exchange coupling @xmath272 must be very close to its critical value @xmath269 .
as perturbation breaks down when @xmath299 , we can only analyze the gate voltage dependence of kondo temperature shown in fig.4 of ref .. the strategy is the following : we find the impurity level @xmath197 at a given chemical potential @xmath13 by using the experimental kondo temperature @xmath3 as the kondo temperature obtained by the pole of resistivity , or @xmath300 where @xmath301 . in the experiment
the kondo temperature is obtained as a function of gate voltage .
we assume the gate voltage @xmath302 is connected with chemical potential @xmath13 via capacitive effect , i.e. @xmath303 with @xmath304 representing the electric charges , @xmath305 , and @xmath306 as the capacitance of the graphene . in the experiment of ref .
@xmath307 is regarded as the position of the dirac node .
thus we take @xmath308 by fixing @xmath307 at @xmath309 . using the experimental kondo temperature at a given chemical potential
we compute the corresponding exchange coupling strength @xmath6 and thus determine the relationship between @xmath13 and impurity level @xmath197 .
the results are shown in table.[table1 ] .relation between @xmath13 and @xmath197 obtained by fitting with experimental kondo temperature .
the left two columns show experimental kondo temperature at a given gate voltage . we compute corresponding chemical potential in the third column by using @xmath310 .
the impurity level @xmath197 , shown in the last column , is obtained by evaluating the corresponding exchange coupling stregth @xmath6 . [ cols="^,^,^,^",options="header " , ] as a function of chemical potential @xmath13 .
the relationship is obtained by @xmath311 and @xmath312 which we find by fitting experimental data shown in the lower left figure .
lower left : impurity level @xmath197 as a function of chemical potential .
dots are obtained by the data in fig.4 of ref . shown in table.[table1 ] .
red dashed line is the linear fitting function which gives @xmath313 .
lower right : impurity occupation of a given spin @xmath239 as a function of chemical potential evaluated by mean field approach.,title="fig : " ] + as a function of chemical potential @xmath13 .
the relationship is obtained by @xmath311 and @xmath312 which we find by fitting experimental data shown in the lower left figure .
lower left : impurity level @xmath197 as a function of chemical potential .
dots are obtained by the data in fig.4 of ref . shown in table.[table1 ] .
red dashed line is the linear fitting function which gives @xmath313 .
lower right : impurity occupation of a given spin @xmath239 as a function of chemical potential evaluated by mean field approach.,title="fig : " ] as a function of chemical potential @xmath13 .
the relationship is obtained by @xmath311 and @xmath312 which we find by fitting experimental data shown in the lower left figure .
lower left : impurity level @xmath197 as a function of chemical potential .
dots are obtained by the data in fig.4 of ref . shown in table.[table1 ] .
red dashed line is the linear fitting function which gives @xmath313 .
lower right : impurity occupation of a given spin @xmath239 as a function of chemical potential evaluated by mean field approach.,title="fig : " ] as a function of chemical potential @xmath13 .
the relationship is obtained by @xmath311 and @xmath313 for @xmath314 and @xmath315 ( corresponds to @xmath316 or @xmath317 ) for @xmath318 from table.[table1 ] we find @xmath319 .
the obtained impurity level @xmath197 changes linearly with the chemical potential @xmath13 as shown in lower left of fig.[fig11 ] .
one of the main conclusions of this work is that the observed upturn in resistivity@xcite can be understood in terms of an anderson impurity model only if the impurity level varies with the applied voltage . by using the linear fit in this figure
we obtain the kondo temperature as a function of chemical potential shown in the top of fig.[fig11 ] . between @xmath320 to @xmath266 the kondo temperature grows monotonically from @xmath321 to @xmath322 .
the decrease of @xmath3 with increasing @xmath13 for @xmath323 may indicate the failure of linearity between @xmath13 and @xmath197 for the onset of nonzero chemical potential or the failure of the kondo physics near the node .
the chemical potential dependence shown in top figure of fig.[fig11 ] is roughly consistent with fig.4 in ref . in the intermediate gate voltage . for gate voltage
@xmath302 larger than @xmath14 in ref . , the experimental @xmath3 begins to decrease with increasing gate voltage .
this can be accounted for qualitatively , as shown in fig.[fig11p ] , by assuming that the energy of the impurity level no longer changes with the external gate voltage for @xmath324 due to sufficient charge screening . for small gate voltage ( chemical potential close to the node ) the experimental @xmath3 increases monotonically with increasing gate voltage . in this region
neither constant impurity level nor @xmath312 gives the corresponding experimental dependence on @xmath302 based on our perturbative kondo results .
we also compute the impurity occupation as a function of @xmath13 by using mean field as shown by eq.([dotocp ] ) in symmetry breaking case .
the obtained impurity occupation for a given spin increase from @xmath325 to @xmath326 monotonically between @xmath327 . given that the validity of the mean field is limited to small values of the impurity level occupations ( see section ivc ) , we expect deviations away from the mean field . thus the system is not likely to stay in the local moment region near the node , suggesting a cross over of impurity occupation from local moment to empty orbital regime based on numerical renormalization group results in ref .. thus kondo effect alone would not be able to explain the logarithmic temperature dependence seen in ref .. and chemical potential between @xmath265 and @xmath328 .
lower left : quadratic dependence on @xmath13 for conductivity at zero temperature .
red line is the quadratic fitting which gives @xmath329 and blue dots are conductivity at various chemical potential .
lower right : universality curve after rescaling resistivity and kondo - like temperature scale @xmath330 obtained by the temperature scale when logarithmic behavior is shown . @xmath331 by comparing this universality curve with fig.2b in ref ..
,title="fig : " ] + and chemical potential between @xmath265 and @xmath328 .
lower left : quadratic dependence on @xmath13 for conductivity at zero temperature .
red line is the quadratic fitting which gives @xmath329 and blue dots are conductivity at various chemical potential .
lower right : universality curve after rescaling resistivity and kondo - like temperature scale @xmath330 obtained by the temperature scale when logarithmic behavior is shown . @xmath331 by comparing this universality curve with fig.2b in ref ..
,title="fig : " ] and chemical potential between @xmath265 and @xmath328 .
lower left : quadratic dependence on @xmath13 for conductivity at zero temperature .
red line is the quadratic fitting which gives @xmath329 and blue dots are conductivity at various chemical potential .
lower right : universality curve after rescaling resistivity and kondo - like temperature scale @xmath330 obtained by the temperature scale when logarithmic behavior is shown . @xmath331 by comparing this universality curve with fig.2b in ref .. ,title="fig : " ] the resistivity decreases slowly with temperature . for @xmath332
the logarithmic dependence on temperature begins to appear . ]
let us now investigate whether charge fluctuations can give rise to temperature dependence of resistivity seen in the experiment .
we take @xmath313 and evaluate the resistivity v.s .
temperature from mean field results of impurity green s function for symmetry breaking case .
for chemical potential close to the node , we get reasonable temperature scale ( the logarithmic behavior shows up at @xmath333 ) from charge fluctuations as shown in the top figure of fig.[fig13 ] .
we also have @xmath334 being proportional to conductivity at zero temperature , as seen in the lower left of fig.[fig13 ] which was observed in the ref .. rescaling @xmath335 by @xmath336 and @xmath5 by kondo - like temperature @xmath330 obtained by the temperature at which the resistivity begin to show logarithmic dependence in @xmath5 , we obtain the universal curve shown in the lower right of fig.[fig13 ] . in this range of chemical potential the impurity occupation @xmath337 .
it shows that even for the chemical potential @xmath13 close to the node the one parameter scaling is still possible in this charge fluctuation scheme , while it is shown analytically in eq.([sbr ] ) the one parameter scaling is unlikely for @xmath12 in the kondo case . for chemical potential away from the dirac node we plot the resistivity v.s .
temperature for @xmath338 , @xmath339 and @xmath266 in fig.[fig12 ] .
the overall feature is very similar to the kondo results : near zero temperature the resistivity decreases with @xmath340 while at large temperature @xmath341 . at zero temperature
the conductivity is proportional to @xmath334 as in the case shown in fig.[fig8 - 1 ] .
however the logarithmic behavior shows up at @xmath342 which is about one order of magnitude larger than the experimental results in ref .. thus the charge fluctuation can not explain the experiment for large chemical potential . as a function of chemical potential @xmath13 .
blue dots are the kondo - like temperature , obtained by comparing universal curve with nrg results , for charge fluctuations case .
purple dots are the kondo temperature for spin fluctuation case
. between @xmath343 to @xmath320 both cases give kondo temperature higher than their neighboring chemical potentials .
the `` mixed '' ( brown dots ) case takes kondo scale obtained by charge fluctuations with @xmath344 and spin fluctuations with @xmath345 . by combing these
two the kondo scale obtained grows monotonically with chemical potential . ] by comparing the universal curve obtained by numerical renormalization group@xcite shown in fig.2b in ref . and
the one we have for charge fluctuations in lower right of fig.[fig13 ] we get @xmath346 .
the `` kondo temperature '' for both charge and spin fluctuations as a function of chemical potential is shown in fig.[fig14 ] . from fig.[fig14 ] we observe that charge fluctuations give large kondo scale with increasing chemical potential and a good agreement with experimental results is obtained only if @xmath347 .
away from the node the kondo scale obtained by charge fluctuations grows much faster than the that of the spin fluctuations . the kondo scale obtained from spin fluctuation , on the other hand
, gives large kondo scale for @xmath348 and reaches its minimum when @xmath349 .
the combined picture of the two cases as shown in fig.[fig14 ] , by assuming charge fluctuation for @xmath350 and spin fluctuation for large @xmath13 , can give the overall consistent picture as seen in the experiment for gate voltage less than @xmath14 .
for @xmath351 it shows the cross over from charge fluctuations to spin fluctuations , which is not accounted for in our simple mean field in anderson model nor perturbation in kondo model . for gate voltage
larger than @xmath14 a non - monotonic dependence of @xmath3 on @xmath13 seen in ref .. given our analysis we speculate that the screening due to the finite density of carriers could modify the dependence of the energy of the impurity level on the gate voltage .
a weaker dependence at large gate voltage will lead to a decreasing kondo temperature .
we use anderson impurity model to describe the dilute impurities behavior in graphene .
the goal is to test whether the recent experiment on the resistivity of graphene with vacancies induced by ion irradiation in ultra - high vacuum@xcite can be solely explained by the single impurity kondo effect ( spin fluctuations ) . to study this local moment regime we use schrieffer wolf transformation to freeze the charge degree of freedom and obtain the kondo hamiltonian . in the case of dilute impurities
we may ignore the rkky interactions and treat the problem as single impurity kondo model .
we have computed this kondo contribution to dc resistivity by perturbation in @xmath5-matrix formulation .
analytic expressions are obtained for @xmath102 and @xmath106 by taking asymptotic form of digamma function in the integrand .
the kondo temperature dependence on chemical potential and exchange coupling are obtained .
depending on the location of the impurities the kondo contribution to resistivity is very different . for the type of magnetic impurities which break the @xmath52 symmetry at low chemical potential it shows power law temperature dependence as @xmath150 . for the magnetic impurities preserving the @xmath52 symmetry of the lattice at low chemical potential it shows power law dependence as @xmath352 . at even lower chemical potential
when fermi surface is close to the node , both cases show extra dependence on chemical potential as well as kondo scale . near the node
a critical value of exchange coupling is needed for kondo effect to be realized@xcite .
the critical value @xmath269 is larger for symmetry breaking case . for large chemical potential @xmath13
both cases show logarithmic dependence on temperature scaled by the kondo temperature . with increasing @xmath13 the resistivity at
a given temperature decreases for impurities breaking the honeycomb symmetry while the resistivity increases for the ones preserving the symmetry .
the resistivity obtained with same set of parameters show that the dominant source of resistivity is from the impurities which break the @xmath52 symmetry .
we also have computed the effect of charge fluctuation for impurity occupation @xmath353 by using mean field approach on the anderson impurity model .
the resistivity at a given temperature has similar dependence on the chemical potential as the case for spin fluctuations .
similar to the spin fluctuation case the dominant contribution to resistivity at the same sets of parameters comes from the impurities which break the honeycomb lattice symmetry . by studying the resistivity v.s .
temperature and comparing with experimental results in ref . from both spin and charge fluctuations in the symmetry breaking case
we find that the kondo effect fails to give the correct kondo scale and unable to describe single parameter scaling for chemical potential nearby the node . for @xmath350
the resistivity due to charge fluctuations give reasonable temperature dependence and the resistivity after rescaling also shows single parameter universal behavior .
the same analysis yields large kondo scale for @xmath354 in the charge fluctuation case which is roughly the same chemical potential at which we get non monotonic behavior of kondo temperature in the spin fluctuation ( kondo ) case .
the failure of kondo explanation nearby the node is consistent with the numerical rg results@xcite which find the kondo effect near the node is suppressed for @xmath355 for systems having electronic density of state @xmath356 . by combining the low chemical potential results ( @xmath357 from charge fluctuation with the large @xmath354 results from kondo effect obtain the kondo scale consistent with the experimental results . for chemical potential in between these two cases the system should be in the mixed valence regime . for gate voltage
higher than @xmath14 a weaker dependence of the impurity energy on the applied gate voltage as compared to the dependence at smaller chemical potentials will lead to a decrease in the kondo temperature .
whether this effect or the effect of rkky interactions is responsible for the observed non - monotonic behavior on gate voltage will be the subject of future studies .
the authors wish to acknowledge roland kawakami and shan - wen tsai for useful discussions .
vivek aji and sung - po chao s research is supported by university of california at riverside under the initial complement .
@xmath358 define @xmath359 and use @xmath360 .
as @xmath361 has poles on upper / lower complex plane we may separate @xmath97 into two parts as @xmath362 with @xmath363 and @xmath364 given by @xmath365 we may write @xmath366 with @xmath367 in the third and fourth lines of the above equation we replaced @xmath368 by @xmath369 and @xmath370 .
similar computation gives @xmath371 as @xmath372 combining @xmath373 and @xmath371 we get @xmath374 here @xmath375 denotes the integration path taken from @xmath376 to @xmath377 along imaginary axis .
we may also write @xmath378 in different region as @xmath379 here @xmath380 denotes the integration path taken from @xmath377 to @xmath381 along imaginary axis .
@xmath382 is expressed as @xmath383 the sum of @xmath363 and @xmath364 is then given by @xmath384 rewrite @xmath385 in the expression of @xmath97 along the @xmath375 and @xmath380 paths we get @xmath386\\ & + & \psi(\frac{1}{2}-i\frac{\beta z}{2\pi})\big[\frac{-2\lambda^2(z+\mu)}{(z+\mu)^2+\lambda^2}\big]\big\}\end{aligned}\ ] ] by defining @xmath99 as @xmath100 we may simplify above expression as @xmath387
consider integrals of the form : @xmath388 let @xmath359 we get @xmath389\big\}=\frac{4}{9\pi t^4}(\bar{i}_1+\bar{i}_2)\end{aligned}\ ] ] we take the integration regions into two parts by writing @xmath390 with @xmath391\big]\end{aligned}\ ] ] @xmath392\big]\end{aligned}\ ] ] thus @xmath393 similarly we can write @xmath394 with @xmath395\big\}\end{aligned}\ ] ] and @xmath396\big\}\end{aligned}\ ] ] thus @xmath397\end{aligned}\ ] ] we combine results of @xmath398 and @xmath399 to get @xmath400\big\}\end{aligned}\ ] ] j. h. chen , w. g. cullen , e. d. williams , and m. s. fuhrer , arxiv : 1004.3373 ( 2010 ) .
p. o. lehtinen , a. s. foster , y. ma , a. v. krasheninnikov , and r. m. nieminen , phys .
lett . * 93 * , 187202 ( 2004 ) .
o. v. yazyev , phys .
* 101 * , 037203 ( 2008 ) .
p. esquinazi , d. spemann , r. hohne , a. setzer , k .- h .
han , and t. butz , phys .
. lett . * 91 * , 227201 ( 2003 ) .
m. m. ugeda , i. brihuega , f. guinea , and j. m. gomez - rodriguez , phys .
lett . * 104 * , 096804 ( 2010 ) .
j. j. palacios , j. fernandez - rossier , and l. brey , phys . rev . *
b 77 * , 195428 ( 2008 ) .
b. uchoa , t. g. rappoport , and a. h. castro neto , arxiv:1006.2512 ( 2010 ) .
b. uchoa , v. n. kotov , n. m. r. peres , and a. h. castro neto , phys .
lett . * 101 * , 026805 ( 2008 ) .
j. r. schrieffer and p. a. wolff , phys . rev . * 149 * , 491 ( 1966 )
. k. h. fischer , z. phys . *
b 42 * , 27 ( 1981 ) .
y. nagaoka , phys .
rev . * 138 * , a1112 ( 1965 )
. h. suhl , phys . rev . * 138 * a515 ( 1965 ) .
d. withoff and e. fradkin , phys .
lett . * 64 * , 1835 ( 1990 ) .
k. ingersent , phys . rev . * b 54 * , 11936 ( 1996 ) . c. gonzalez - buxton and k. ingersent , phys
b 54 * , r15614 ( 1996 ) .
a. polkovnikov , phys . rev . *
b 65 * , 064503 ( 2002 ) .
k. sengupta and g. baskaran , phys . rev . *
b 77 * , 045417 ( 2008 ) .
m. vojta and r. bulla , phys .
b 65 * , 014511 ( 2001 ) .
m. vojta , l. fritz , and r. bulla , euro .
. lett . * 90 * , 27006 ( 2010 ) .
l. fritz and m. vojta , phys . rev .
* b 70 * , 214427 ( 2004 ) . l. brey , h. a. fertig , and s. das sarma , phys .
. lett . * 99 * , 116802 ( 2007 ) .
s. saremi , phys . rev . *
b 76 * , 184430 ( 2007 ) .
p. w. anderson , phys . rev . * 124*,41 ( 1961 ) .
t. a. costi , a. c. hewson , and v. zlatic , j. of .
phys : cond .
6 , 2519 ( 1994 ) .
a. c. hewson , _ the kondo problem to heavy fermions _
, cambridge studies in magnetism ( 1993 ) .
p. b. wiegman and a. m. tsvelik , j. phys .
* c. 16 * , 2281(1983 ) ; p. b. wiegman and a. m. tsvelik , adv . in phys.*32 * , 453 ( 1983 ) . | motivated by experiments on ion irradiated graphene , we compute the resistivity of graphene with dilute impurities . in the local moment regime we employ the perturbation theory up to third order in the exchange coupling to determine the behavior at high temperatures within the kondo model .
resistivity due to charge fluctuations is obtained within the mean field approach on the anderson impurity model . due to the linear spectrum of the graphene
the kondo behavior is shown to depend on the gate voltage applied .
the location of the impurity on the graphene sheet is an important variable determining its effect on the kondo scale and resitivity .
our results show that for chemical potential near the node the charge fluctuations is responsible for the observed temperature dependence of resistivity while away from the node the spin fluctuations take over .
quantitative agreement with experimental data is achieved if the energy of the impurity level varies linearly with the chemical potential . |
in many applications of quantum mechanics to communication , the sender and the receiver have to share a maximally entangled quantum state of two particles .
when there is a noiseless quantum communication channel , the sender can send one of two particles in a maximally entangled state to the receiver and sharing of it is easily accomplished .
however , the quantum communication channel is usually noisy , that is , the quantum state of the received particle changes probabilistically from the original state of a particle .
entanglement distillation protocols @xcite and quantum error - correcting codes @xcite are the schemes for sharing a maximally entangled state over a noisy communication channel .
a distillation protocol is said to be _ two - way _ ( resp .
_ one - way _ ) if it involves two - way ( resp . one - way ) classical communication
two - way protocols have larger distillation ability than one - way protocols .
however , few two - way protocols has been proposed so far , namely the recurrence protocol @xcite and the quantum privacy amplification protocol ( qpa protocol ) @xcite .
there may be many two - way protocols better than existing ones , and the discovery of better protocols has been awaited . immediately after the proposal of those schemes , bennett et al
. discovered that one can construct a _
one - way _ entanglement distillation protocol from a quantum code ( * ? ? ?
* section v.c ) , which requires @xmath1 additional qubits where @xmath2 is the number of noisy entangled states to be distilled .
nielsen and chuang ( * ? ? ?
* exercise 12.34 ) observed a construction method of a one - way protocol without extra qubits from a real binary quantum stabilizer code as a generalization of the idea in @xcite . by a conversion method from a quantum code to a distillation protocol
, we can solve problems of distillation protocols from results in quantum codes .
for example , we can construct a good distillation protocol from a good quantum code .
thus such a conversion method deserves further investigation .
it is not known how one can convert a quantum error - correcting code to a _ two - way _ entanglement distillation protocol .
we shall propose a conversion method from an arbitrary quantum stabilizer code to both one - way and two - way entanglement distillation protocols as a generalization of shor , preskill @xcite , nielsen , and chuang @xcite .
benefits of the proposed conversion methods are * we can construct infinitely many two - way protocols .
one can easily construct a two - way protocol better than the recurrence protocol and the qpa protocol from a simple stabilizer code ( see section [ sec42 ] ) .
* it is known that one - way protocols and quantum error - correcting codes without classical communication have the same ability for sharing maximally entangle states over a noisy quantum channel @xcite .
the proposed protocols might be used for further clarification of the relation between distillation protocols and quantum error - correcting codes .
this paper is organized as follows : in section [ sec2 ] , basic notation is introduced . in section [ sec3 ] ,
we present a construction of entanglement distillation protocols from quantum stabilizer codes . in section [ sec4 ] ,
we give examples of converted protocols equivalent to the recurrence protocol and the qpa protocol , and an example better than them . in section [ sec5 ]
, we evaluate the distillable entanglement by the class of one - way protocols converted from stabilizer codes , and show that the converted protocols can distill asymptotically more entanglement from a noisy werner state than the hashing protocol @xcite . in section [ generalfidelity ] , we derive a lower bound on fidelity with a general initial state of protocols .
in this section we fix notation and the problem formulation .
let @xmath3 and @xmath4 be @xmath5-dimensional complex linear spaces with orthonormal bases @xmath6 , , @xmath7 and @xmath8 ,
, @xmath9 , respectively , where @xmath5 is a prime number .
we shall restrict ourselves to @xmath5-ary stabilizer codes because an @xmath10-ary stabilizer code can be constructed as a tensor product of @xmath11-ary stabilizer codes ( * ? ? ?
* , remarks ) , where @xmath11 are prime divisors of @xmath10 , and extension of the proposed conversion method to the @xmath10-ary case is straightforward .
we define the maximally entangled states in @xmath12 by @xmath13 where @xmath14 , @xmath15 , , @xmath16 , and matrices @xmath17 and @xmath18 are defined by @xmath19 with a complex primitive @xmath5-th root @xmath20 of @xmath21 .
the matrices @xmath17 , @xmath18 and their commutation relation were first applied to the quantum mechanics by weyl ( * ? ? ?
* section 4.15 ) .
suppose that charlie prepares @xmath2 pairs of particles in the state @xmath22 , sends the particles corresponding to @xmath3 to alice , and sends the other particles corresponding to @xmath4 to bob .
the quantum channels between alice and charlie and between bob and charlie are noisy in general , and alice and bob share a mixed state @xmath23 , where @xmath24 is the set of density operators on @xmath25 .
the state @xmath26 can be an arbitrary density operator .
the goal of an entanglement distillation protocol is to extract as many pairs of particles with state close to @xmath22 as possible from @xmath2 pairs of particles in the state @xmath26 .
in this section we shall describe how to make an entanglement distillation protocol from a quantum stabilizer code . in the protocol
we extract a state @xmath27 from @xmath28 .
the proposed protocol will be constructed from the nonbinary generalization @xcite of quantum stabilizer codes @xcite .
we assume that the reader is familiar with the formalism of the nonbinary stabilizer code .
let us introduce notation of stabilizer codes .
let @xmath29 @xmath30 , @xmath31 ,
, @xmath32 , @xmath33 , @xmath34 are integers @xmath35 , and @xmath36 a commutative subgroup of @xmath37 . the subgroup @xmath36 is called a stabilizer .
let @xmath38 , , @xmath16 with addition and multiplication taken modulo @xmath5 . for a vector @xmath39 , @xmath31 , , @xmath32 , @xmath40 , let @xmath41 suppose that @xmath42 , , @xmath43 ( and possibly some power of @xmath44 ) @xmath35 is a generating set of the group @xmath36 , where @xmath45 , , @xmath46 are linearly independent over @xmath47 .
let @xmath48 be a complex linear space with the orthonormal basis @xmath49 ,
, @xmath50 , and hereafter we shall identify @xmath48 with @xmath3 and @xmath4 by linear maps @xmath51 and @xmath52 . let @xmath53 be a stabilizer code defined by @xmath36 , that is , a joint eigenspace of @xmath36 in @xmath54 .
there are many joint eigenspaces of @xmath36 and we can distinguish an eigenspace by its eigenvalue of @xmath55 for @xmath56 , , @xmath57 .
hereafter we fix a joint eigenspace @xmath53 of @xmath36 and suppose that @xmath53 belongs to the eigenvalue @xmath58 of @xmath55 for @xmath56 , , @xmath57 .
suppose that we sent @xmath59 , and received @xmath60 .
we can tell which eigenspace of @xmath36 contains the state @xmath60 by measuring an observable whose eigenspaces are the same as those of @xmath55 .
then the measurement outcome always indicates that the measured state @xmath60 belonging to the eigenspace @xmath61 , where @xmath62 is the symplectic inner product defined by @xmath63 for @xmath64 , @xmath31 , , @xmath32 , @xmath65 and @xmath66 , @xmath67 , , @xmath68 , @xmath69 .
we define @xmath70 , @xmath71 , , @xmath32 , @xmath72 .
since the complex conjugate of @xmath20 is @xmath73 , we can see that @xmath74 is a componentwise complex conjugated matrix of @xmath55 .
let @xmath75 be a subgroup of @xmath37 generated by @xmath76 ,
, @xmath77 .
easy computation shows that @xmath75 is again commutative .
so we can consider joint eigenspaces of @xmath75 .
there exists a joint eigenspace @xmath78 of @xmath75 whose eigenvalue of @xmath74 is @xmath79 ( the complex conjugate of @xmath58 ) .
with those notation , our protocol is executed as follows : 1 .
[ step1 ] alice measures an observable corresponding to @xmath80 for each @xmath34 , and let @xmath81 be the eigenvalue of an eigenspace of @xmath75 containing the state after measurement . in what follows we refer to @xmath82 ,
, @xmath83 as a _ measurement outcome_. 2 .
[ step2 ] bob measures an observable corresponding to @xmath84 for each @xmath34 , and let @xmath85 be the eigenvalue of an eigenspace of @xmath36 containing the state after measurement . in what follows we also refer to @xmath86 ,
, @xmath87 as a _ measurement outcome_. 3 .
alice sends @xmath82 , ,
@xmath88 to bob .
[ step4 ] bob perform the error correction process according to @xmath89 ,
, @xmath90 as described below . 5 . [ step5 ]
alice and bob apply the inverse of encoding operators of the quantum stabilizer codes .
[ step6 ] alice and bob discards the last @xmath57 particles .
[ step7 ] if the difference of the measurement outcomes @xmath91 ,
, @xmath92 indicates that the fidelity between the remaining @xmath93 particles and @xmath94 is low , bob discards all of his particles and he tells alice the disposal of particles .
we shall introduce some notation .
for a vector @xmath95 let @xmath96 let @xmath97 [ resp .
@xmath98 @xmath99 be the quantum stabilizer code of @xmath36 ( resp .
@xmath75 ) belonging to the eigenvalue @xmath100 ( resp .
@xmath101 ) of @xmath55 [ resp .
@xmath102 for a vector @xmath103 , , @xmath104 , and @xmath105 [ resp .
@xmath106 be the projection onto @xmath97 [ resp .
@xmath98 .
we have @xmath107 for any @xmath108 . _ proof .
_ let @xmath49 ,
, @xmath109 be an orthonormal basis of @xmath54 consisting of tensor products of @xmath110 , , @xmath111 , and we have @xmath112 for @xmath108 , let @xmath113 ,
, @xmath114 be an orthonormal basis of @xmath97 . for a state @xmath115
we define @xmath116 where @xmath117 is the complex conjugate of @xmath118 . with this notation , @xmath119 ,
, @xmath120 is an orthonormal basis of @xmath121 .
the set @xmath122 , @xmath123 , , @xmath124 is an orthonormal basis of @xmath54 and there exists a unitary matrix on @xmath54 that transforms the basis @xmath49 ,
, @xmath109 to @xmath122 , @xmath123 , , @xmath124 .
let @xmath125 be the componentwise complex conjugate of @xmath126 , that is , @xmath125 transforms @xmath49 ,
, @xmath109 to @xmath127 , @xmath123 , , @xmath124 .
we have @xmath128 @xcite .
therefore @xmath129 since @xmath130 we have @xmath131 \sum_{\vec{x}\in \mathbf{z}_p^{n - k } } \sum_{i=0}^{p^k-1 } \overline{{|\vec{x},i\rangle } } \otimes{|\vec{x},i\rangle } \nonumber\\ & = & \sum_{i=0}^{p^k-1 } \overline{{|\vec{x},i\rangle } } \otimes{|\vec{x},i\rangle } \nonumber\\ & = & \sqrt{p^n}\{p^\star(\vec{x})\otimes p(\vec{x})\ } { |\beta(\vec{0})\rangle}\label{form2}\end{aligned}\ ] ] ' '' '' suppose that we perform the protocol above to the state @xmath132 .
after we get @xmath39 , , @xmath133 as a measurement outcome in step [ step1 ] , the state is @xmath134}.\end{aligned}\ ] ] observe that the vector @xmath135 belongs to @xmath136 , where @xmath137 thus the measurement outcome in step [ step2 ] must be @xmath138 . for the simplicity of presentation
, we assume that the state @xmath139 can be written as @xmath140 where @xmath141 is a probability distribution . a general case will be treated in section [ generalfidelity ] . after performing step [ step1 ] in the proposed protocol to state ( [ restrictedstate ] ) and getting @xmath142 as a measurement outcome , the state is @xmath143 where @xmath144 , @xmath145 and @xmath146 .
suppose that we get @xmath138 as a measurement outcome in step [ step2 ] , and denote @xmath91 ,
, @xmath92 by @xmath147 .
the state @xmath148 belongs to @xmath149 , @xmath150 ,
, @xmath151 , @xmath152 $ ] .
thus the state after step [ step2 ] is @xmath153 where @xmath154 let @xmath155 be the linear subspace of @xmath156 spanned by @xmath45 , , @xmath46 , and
@xmath157 be the orthogonal space of @xmath155 with respect to the symplectic inner product ( [ symplectic ] ) . for vectors
@xmath158 , @xmath159 such that @xmath160 , @xmath161 and @xmath162 has the same effect on states in @xmath163 for any @xmath164 , and we can identify errors @xmath161 and @xmath162 if @xmath160 , which is equivalent to @xmath165 . thus , among errors @xmath161 corresponding to @xmath166 , the most likely error vector @xmath158 is one having maximum @xmath167 in the set @xmath166 .
let @xmath168 be the most likely error vector in @xmath166 .
the set @xmath166 is equal to @xmath169 bob applies @xmath170 to his particles .
this is step [ step4 ] . after applying @xmath170 to bob s particles ,
the joint state of particles of alice and bob is @xmath171 recall that @xmath172 does not change a state in @xmath163 if @xmath173 .
therefore the state ( [ step4state1 ] ) is equal to @xmath174p(\vec{a},\vec{a } ) \rho(\vec{0})p(\vec{a } , \vec{a } ) [ i\otimes \mathsf{xz}(\vec{u}-\vec{e})^ * ] .
\label{step4state2}\end{aligned}\ ] ] we shall explain how to use an encoding operator in step [ step5 ] to extract @xmath94 from the above state .
let @xmath175 be an ancillary state .
consider an encoding operator @xmath176 on @xmath54 sending @xmath177 to @xmath178 for @xmath179 , , @xmath180 , where @xmath181 , , @xmath182 is an orthonormal basis of @xmath163 defined above .
observe that @xmath183 is an encoding operator for @xmath184 sending @xmath177 to @xmath185 for @xmath179 ,
, @xmath180 .
applying @xmath186 to state ( [ step4state1 ] ) yields @xmath187p(\vec{a},\vec{a } ) \rho(\vec{0})\nonumber\\ * & & \mbox { } p(\vec{a } , \vec{a } ) [ i\otimes \mathsf{xz}(\vec{u}-\vec{e})^ * ] ( \overline{u_\mathrm{e}}\otimes u_\mathrm{e } ) \nonumber\\ & = & \sum_{\vec{u } \in \vec{e}+c } \alpha(\vec{u } ) ( \overline{u_\mathrm{e}}^{-1}\otimes u_\mathrm{e}^{-1 } ) p(\vec{a},\vec{a } ) \rho(\vec{0})p(\vec{a } , \vec{a } ) ( \overline{u_\mathrm{e}}\otimes u_\mathrm{e } ) \mbox { [ by eq.\ ( \ref{step4state2 } ) ] } \nonumber\\ * & & \mbox{}+ \sum_{\vec{u } \in \vec{e}+(c^{\perp}\setminus c ) } \alpha(\vec{u } ) ( \overline{u_\mathrm{e}}^{-1}\otimes u_\mathrm{e}^{-1})[i\otimes \mathsf{xz}(\vec{u}-\vec{e})]p(\vec{a},\vec{a } ) \rho(\vec{0})\nonumber\\ * & & \mbox { } p(\vec{a } , \vec{a } ) [ i\otimes \mathsf{xz}(\vec{u}-\vec{e})^ * ] ( \overline{u_\mathrm{e}}\otimes u_\mathrm{e})\nonumber\\ & = & \sum_{\vec{u } \in \vec{e}+c } \alpha(\vec{u } ) ( \overline{u_\mathrm{e}}^{-1}\otimes u_\mathrm{e}^{-1 } ) \left[\frac{1}{p^n } \left\ { \sum_{i=0}^{p^k-1 } \overline{{|\vec{a},i\rangle}}\otimes { |\vec{a},i\rangle}\right\ } \left\ { \sum_{i=0}^{p^k-1 } \overline{{\langle \vec{a},i |}}\otimes { \langle \vec{a},i | } \right\}\right]\nonumber\\ * & & \mbox { } ( \overline{u_\mathrm{e}}\otimes u_\mathrm{e } ) \mbox { [ by eq.\ ( \ref{form2 } ) ] } \nonumber\\ * & & + \sum_{\vec{u } \in \vec{e}+(c^{\perp}\setminus c ) } \alpha(\vec{u } ) ( \overline{u_\mathrm{e}}^{-1}\otimes u_\mathrm{e}^{-1})[i\otimes \mathsf{xz}(\vec{u}-\vec{e})]p(\vec{a},\vec{a } ) \rho(\vec{0})\nonumber\\ * & & \mbox { } p(\vec{a } , \vec{a } ) [ i\otimes \mathsf{xz}(\vec{u}-\vec{e})^ * ] ( \overline{u_\mathrm{e}}\otimes u_\mathrm{e})\nonumber\\ & = & \frac{1}{p^n}\sum_{\vec{u } \in \vec{e}+c } \alpha(\vec{u } ) \left\ { { |\beta(0,0)\rangle}^{\otimes k}\otimes{|\mathrm{a}\rangle}^{\otimes 2}\right\ } \left\ { { \langle \beta(0,0 ) |}^{\otimes k}\otimes { \langle \mathrm{a } |}^{\otimes 2 } \right\ } \mbox { [ by definition of $ u_\mathrm{e}$]}\nonumber\\ * & & + \sum_{\vec{u } \in \vec{e}+(c^{\perp}\setminus c ) } \alpha(\vec{u } ) ( \overline{u_\mathrm{e}}^{-1}\otimes u_\mathrm{e}^{-1})[i\otimes \mathsf{xz}(\vec{u}-\vec{e})]p(\vec{a},\vec{a } ) \rho(\vec{0})\nonumber\\ * & & p(\vec{a } , \vec{a } ) [ i\otimes \mathsf{xz}(\vec{u}-\vec{e})^ * ] ( \overline{u_\mathrm{e}}\otimes u_\mathrm{e } ) \label{step5state}\end{aligned}\ ] ] taking partial trace of the first term over the last @xmath57 qubits yields @xmath94 , which is step [ step6 ] . let @xmath188 be the final state of step [ step5 ] , that is , state ( [ step5state ] ) , and @xmath189 be the state after step [ step6 ] . in step [ step7 ] , bob computes the fidelity between the state @xmath94 and @xmath189 by using knowledge of @xmath147 and @xmath190 .
@xmath191 $ ] is not @xmath21 because @xmath188 is a state after projection .
we have @xmath192 = \mathrm{tr}[\tau_5 ] & = & \mathrm{tr}\left[p(\vec{a},\vec{a } ) \rho(\vec{0})p(\vec{a } , \vec{a})\right ] \sum_{\vec{u}\in \vec{e}+c^\perp } \alpha(\vec{u})\\ & = & { \langle \beta(\vec{0 } ) | } p^\star(\vec{a})\otimes i { |\beta(\vec{0})\rangle } \sum_{\vec{u}\in \vec{e}+c^\perp } \alpha(\vec{u } ) \mbox { [ by eq.~(\ref{form})]}\\ & = & \frac{1}{p^{n - k } } \sum_{\vec{u}\in \vec{e}+c^\perp } \alpha(\vec{u } ) \mbox { [ by eq.~(\ref{form2})]}\end{aligned}\ ] ] if the initial state is @xmath193 such that @xmath194 , we can get @xmath195 as @xmath189 .
therefore we have @xmath196 thus bob estimates that the fidelity between @xmath94 and the normalized state of @xmath189 is at least @xmath197 the value ( [ estimatedfidelity ] ) varies according to @xmath198 , , @xmath92 .
if obtained difference @xmath147 implies low fidelity , bob discards all the particles and tell alice the disposal . note that if we include step [ step7 ] then the whole protocol needs two - way classical communication , but
if we exclude step [ step7 ] then it needs only one - way classical communication .
when alice and bib do not execute step [ step7 ] , the average of fidelity ( [ estimatedfidelity ] ) should be considered instead of respective values of eq .
( [ estimatedfidelity ] ) for each difference @xmath147 of measurement outcomes . the average of eq .
( [ estimatedfidelity ] ) is at least @xmath199 where @xmath200 is the guessed error vector for a given difference @xmath147 of measurement outcomes .
this average fidelity ( [ averagefidelity ] ) will be studied in sections [ sec4 ] and [ generalfidelity ] .
in this section we show how one can construct the well - known recurrence protocol and the qpa protocol from stabilizer codes , and give a two - way protocol constructed from a stabilizer better than the recurrence protocol and the qpa protocol . the recurrence protocol without twirling ( * ? ? ? * step ( a2 ) ) has the same effect on any density operator on @xmath201 as the proposed protocol with @xmath202 , @xmath203 , @xmath204 , the stabilizer @xmath36 generated by @xmath205 , encoding operators @xmath206 @xmath207 for the code belonging to eigenvalue @xmath208 of @xmath205 , @xmath209 @xmath210 for the code belonging to eigenvalue @xmath211 of @xmath205 , and discarding particles in step [ step7 ] if @xmath212 .
this can be seen by a tedious but straightforward computation .
the qpa protocol @xcite has the same effect as the protocol converted from the stabilizer @xmath36 generated by @xmath213 , encoding operators @xmath206 @xmath214 for the code belonging to eigenvalue @xmath208 of @xmath213 , @xmath209 @xmath215 for the code belonging to eigenvalue @xmath211 of @xmath213 , and discarding particles in step [ step7 ] if @xmath212 we shell compare the protocol constructed from the stabilizer generated by @xmath216 , @xmath217 ( @xmath218 ) with the recurrence protocol and the qpa protocol in a similar way to ( * ? ? ?
* fig . 8) .
we discard particles in the protocol unless the measurement outcomes completely agree , i.e. , @xmath219 .
encoding operators for the stabilizer codes belonging to the eigenvalue @xmath220 of @xmath221 and @xmath222 of @xmath223 are described in table [ tab1 ] .
suppose that we have many copies of noisy entangled state @xmath224 and we want to distill the bell state @xmath225 as many as possible by using the hashing protocol and a two - way protocol chosen from the recurrence protocol _ without twirling _ , the qpa protocol , and the protocol constructed from @xmath216 , @xmath217 .
we use the hashing protocol to distill the perfect the bell state @xmath225 after suitable number of iteration of a two - way protocol as described in ( * ? ? ?
* section iii.b.1 ) .
the number of perfect bell state distillable by the three two - way protocols are compared in figure [ fig1 ] .
observe that an example of the proposed protocol has larger distillable entanglement for the range of @xmath226 between @xmath227 to @xmath228 .
in this section , we evaluate the distillable entanglement by one - way protocols constructed from stabilizers .
distillable entanglement is the most important measure of the performance of a class of protocols .
we mean by an @xmath230 $ ] entanglement distillation protocol a protocol always leaving @xmath93 pairs of particles out of given @xmath2 pairs of particles .
let @xmath231 be a class of @xmath230 $ ] entanglement distillation protocol for @xmath232 , @xmath233 , , and @xmath204 , , @xmath2 .
let @xmath234 be a density operator on @xmath235 . the distillable entanglement by the protocol @xmath231 for the sequence of states @xmath236 is the maximum of a real number @xmath237 such that for any @xmath238 and any @xmath239 there exists an @xmath230 $ ] ( @xmath240 ) protocol in @xmath231 such that the protocol extracts a state @xmath241 from @xmath234 such that the fidelity between @xmath242 and a maximally entangled state in @xmath243 is at least @xmath244 .
roughly speaking , the distillable entanglement by @xmath231 is the largest number of maximally entangled pairs in @xmath245 distillable from one pair of particles .
our definition imposes on protocols the restriction that a protocol always produces the same number of pairs of particles .
a general definition without this restriction was given by rains @xcite .
let @xmath246 be a probability distribution , and consider the density operator @xmath247 on @xmath12 .
we shall estimate the distillable entanglement by the proposed protocol for the sequence of states @xmath248 ,
@xmath35 , and show the distillable entanglement is at least as large as the achievable rate of quantum stabilizer codes over the quantum channel @xmath249 on @xmath48 with an error @xmath250 occurs with probability @xmath251 . the achievable rate by quantum stabilizer codes over @xmath249 is the maximum of a real number @xmath237 such that for any @xmath238 and any @xmath239 there exists an @xmath230 $ ] ( @xmath240 ) stabilizer code @xmath53 such that any state @xmath59 can be transmitted over @xmath249 with fidelity at least @xmath244 .
we assume that the decoding of a quantum stabilizer code is implemented as follows : first measure an observable whose eigenspaces are the same as the stabilizer of the code , determine most likely error of the form @xmath252 , and apply the inverse of the guessed error to the codeword . under this assumption ,
the distillable entanglement by the proposed protocol without step [ step7 ] for @xmath248 ,
@xmath35 is at least as large as the achievable rate by quantum stabilizer codes over @xmath249 .
_ let @xmath237 be the achievable rate by quantum stabilizer codes over @xmath249 .
then for any @xmath253 and @xmath254 there exists an @xmath230 $ ] ( @xmath240 ) quantum stabilizer code @xmath53 with stabilizer @xmath36 such that for any state @xmath255 can be transmitted over @xmath249 with fidelity at least @xmath256 .
let @xmath36 be generated by @xmath42 ,
, @xmath43 ( and possibly some power of @xmath44 ) @xmath35 , and @xmath53 belong to the eigenvalue @xmath58 of @xmath55 .
suppose that the decoder guesses the error as @xmath257 when the measurement outcomes indicate that the received state belongs to eigenvalue @xmath258 of @xmath55 for @xmath56 ,
, @xmath57 , where @xmath259 , , @xmath260 .
then the decoder can correct any error @xmath161 if @xmath261 where @xmath155 is a linear subspace of @xmath156 spanned by @xmath45 , , @xmath46 . by lemma [ badcodeword ] ( see appendix [ sec : badcodeword ] ) , there exists a codeword @xmath59 such that if @xmath262 is transmitted and @xmath263 is received with @xmath158 not in the set ( [ correctable ] ) then the fidelity between @xmath262 and the decoded state is at most @xmath264 , because the set ( [ correctable ] ) is equal to the set of correctable errors by @xmath53 in lemma [ badcodeword ] . since @xmath262 can be transmitted through @xmath249 with fidelity at least @xmath256 , the probability of the correctable error ( [ correctable ] ) over @xmath265 is at least @xmath266 .
suppose that we apply the proposed protocol to @xmath267 such that if the difference @xmath147 of measurement outcomes is observed then @xmath268 is applied in step [ step4 ] .
then the average ( [ averagefidelity ] ) of the fidelity is at least @xmath266 , because the errors in the set ( [ correctable ] ) are also correctable by the proposed protocol [ see eq .
( [ step4state2 ] ) ] . for given
@xmath239 set @xmath269 in the above argument , and we can see that the distillable entanglement is at least as large as the achievable rate of quantum stabilizer codes over @xmath249 . ' '' '' the best known lower bound on the achievable rate by quantum stabilizer codes over @xmath249 is given by hamada @xcite , and his lower bound gives the true value for the depolarizing channels .
let us compare the distillable entanglement by the converted protocols and that by the hashing protocol @xcite for the werner state of fidelity @xmath226 , which is given by @xmath270 , @xmath271 and @xmath202 .
the werner state is converted to @xmath272 by applying @xmath273 on bob s particle .
the distillable entanglement of state ( [ convertedwerner ] ) by the hashing protocol is estimated as @xmath274 where @xmath275 is the shannon entropy with base @xmath276 .
the distillable entanglement of state ( [ convertedwerner ] ) by the converted protocols is strictly larger than eq .
( [ binaryhashing ] ) for certain range of @xmath226 , because the achievable rate of the shor - smolin concatenated codes is strictly larger than eq .
( [ binaryhashing ] ) over the depolarizing channel of fidelity @xmath226 @xcite and they can be written as stabilizer codes @xcite .
let us consider the case of @xmath277 , @xmath278 , and @xmath279 for @xmath280 .
the distillable entanglement by the nonbinary generalization @xcite of the hashing protocol is estimated as @xmath281 the achievable rate by the quantum stabilizer codes is strictly greater than eq .
( [ ternaryhashing ] ) for @xmath282 ( * ? ? ? * section vi.c ) , and so is the distillable entanglement by the converted protocols .
in the preceding argument we assumed that the initial state shared by alice and bob was in the form of eq .
( [ restrictedstate ] ) . in this section
we remove this restriction .
let @xmath26 be an arbitrary density operator in @xmath283 .
we shall consider applying the proposed protocol without step [ step7 ] to @xmath26 and calculate the fidelity between the distilled state and @xmath94 .
precisely speaking , we shall calculate the fidelity between @xmath284 and the state after step [ step5 ] , which is equal to that between @xmath94 and the state after step [ step6 ] .
the idea of the following argument is borrowed from section 7.4 of @xcite . since there is no selection of particles in steps [ step1][step6 ] by a measurement , the whole process of steps [ step1][step6 ]
can be written as a completely positive trace - preserving map @xmath285 on the density operators on @xmath283 .
let @xmath286 is a purification of @xmath26 .
since @xmath287 is an orthonormal basis of @xmath283 , we can write @xmath288 as @xmath289 where @xmath290 is a vector in @xmath291 . in step [ step4 ] , the inverse error operator @xmath170 is determined from the difference @xmath147 of measurement outcomes and knowledge of @xmath292 .
when we deal with an arbitrary but known density operator @xmath26 , determine @xmath168 from @xmath147 so that the lower bound ( [ generalbound ] ) below on fidelity becomes large .
once we fix a determination rule of @xmath168 from @xmath147 , we can define @xmath293 the protocol can perfectly distill @xmath94 from @xmath294 .
equation ( [ expansion ] ) can be written as @xmath295 the almost same argument as section 7.4 of @xcite shows that the fidelity between @xmath94 and the state after step [ step6 ] is at least @xmath296
the author would like to thank prof .
tomohiko uyematsu and mr .
toshiyuki morita for helpful discussions .
this research was supported by the japan society for the promotion of science under contract no .
14750278 .
10 m. aschbacher .
, volume 10 of _ cambridge studies in advanced mathematics_. cambridge university press , cambridge , uk , 2nd edition , 2000 . c. h. bennett , g. brassard , s. popescu , b. schumacher , j. a. smolin , and w. k. wootters .
purification of noisy entanglement and faithful teleportation via noisy channels .
, 76(5):722725 , jan .
1996 , quant - ph/9511027 . c. h. bennett , d. p. divincenzo , j. a. smolin , and w. k. wootters .
mixed - state entanglement and quantum error correction .
, 54(5):38243851 , nov . 1996 , quant - ph/9604024 . a. r. calderbank , e. m. rains , p. w. shor , and n. j. a. sloane
. quantum error correction and orthogonal geometry .
, 78(3):405408 , jan .
1997 , quant - ph/9605005 .
a. r. calderbank , e. m. rains , p. w. shor , and n. j. a. sloane . quantum error correction via codes over gf(4 ) .
, 44(4):13691387 , july 1998 , quant - ph/9608006 .
d. deutsch , a. ekert , r. jozsa , c. macchiavello , s. popescu , and a. sanpera . quantum privacy amplification and the security of quantum cryptography over noisy channels .
, 77(13):28182821 , sept .
1996 , quant - ph/9604039 ; erratum : _ phys .
_ , 80(9):2022 , march 1998 .
d. p. divincenzo , p. w. shor , and j. a. smolin .
quantum - channel capacity of very noisy channels .
, 57(2):830839 , feb .
1998 , quant - ph/9706061 ; erratum : _ phys .
a _ , 59(2):1717 , feb .
d. gottesman .
class of quantum error - correcting codes saturating the quantum hamming bound .
, 54(3):18621868 , sept . 1996 , quant - ph/9604038 .
m. hamada .
information rates achievable with algebraic codes on quantum discrete memoryless channels .
version 1 , july 2002 , arxiv : quant - ph/0207113v1 .
m. horodecki and p. horodecki .
reduction criterion of separability and limits for a class of distillation protocols .
, 59(6):42064216 , june 1999 , quant - ph/9708015 .
e. knill .
non - binary unitary error bases and quantum codes . aug .
1996 , quant - ph/9608048 .
m. a. nielsen and i. l. chuang . .
cambridge university press , cambridge , uk , 2000 .
j. preskill .
lecture notes for physics 229 : quantum information and computation , ` http://www.theory.caltech.edu/people/preskill/ph229 ` , 1998 . e. m. rains .
nonbinary quantum codes . , 45(6):18271832 , sept .
1999 , quant - ph/9703048 . e. m. rains .
rigorous treatment of distillable entanglement . ,
60(1):173178 , july 1999 , quant - ph/9809078 .
p. w. shor .
scheme for reducing decoherence in quantum computer memory .
, 52(4):24932496 , oct .
p. w. shor and j. preskill .
simple proof of security of the bb84 quantum key distribution protocol .
, 85(2):441444 , july 2000 , quant - ph/0003004 .
a. m. steane .
error correcting codes in quantum theory .
, 77(5):793797 , july 1996 .
k. g. h. vollbrecht and m. m. wolf .
efficient distillation beyond qubits . aug .
2002 , arxiv : quant - ph/0208152 .
h. weyl . .
methuen , london , 2nd edition , 1931 .
we consider a quantum channel over which an error of the form @xmath297 occurs with the probability @xmath298 for @xmath299 , and we also consider the following decoding method : measure the observable of @xmath54 whose eigenspaces are the same as those of @xmath36 , and apply an operator @xmath300 ( @xmath301 ) determined by the measurement outcome and some deterministic criterion . with this decoding method
, we can correct at most @xmath302 errors among all the @xmath303 errors for an @xmath230 $ ] quantum stabilizer code .
[ badcodeword ] let @xmath53 be an @xmath230 $ ] quantum stabilizer code .
suppose that we have a fixed decoding method as described above .
there exists a codeword @xmath59 such that @xmath304 for all uncorrectable error @xmath297 , where an error @xmath297 is said to be _ correctable _ if a received state @xmath305 is decoded to @xmath262 for all @xmath59 and _ uncorrectable _ otherwise .
_ consider the following map @xmath306 let @xmath307 .
since @xmath36 is commutative , we have @xmath308 .
let @xmath309 be a subspace of @xmath156 such that @xmath310 such a space @xmath309 always exists by the witt theorem ( see sec .
20 of ref .
@xcite ) . since @xmath311
, we have @xmath312 .
the set @xmath313 is a commutative subgroup of @xmath37 , so we can consider a quantum stabilizer code @xmath314 defined by @xmath313 .
we have @xmath315 .
let @xmath316 be a normalized state vector .
we shall construct the desired codeword @xmath262 in lemma [ badcodeword ] from @xmath317 .
let @xmath320 be a set of coset representatives of @xmath309 in @xmath321 , that is , @xmath237 has the same number of elements as @xmath322 , and if @xmath323 and @xmath324 then @xmath318 . we assume @xmath325 .
define @xmath326 which is a normalized state vector in @xmath53 by eq .
( [ orth ] ) .
we want to take @xmath262 in lemma [ badcodeword ] as a multiple of @xmath327 , so let us compute @xmath328 by eq .
( [ orth ] ) we also have @xmath329
. therefore @xmath330 .
define @xmath262 by @xmath331 which is a normalized state vector in @xmath53 .
we shall show that @xmath262 has the desired property .
suppose that an error @xmath332 occurred and we applied @xmath333 as the recovery operator . if @xmath334 , then the error @xmath335 is correctable , otherwise @xmath335 is uncorrectable . if @xmath336 , the decoded state is orthogonal to any transmitted state , so we may assume @xmath337 hereafter .
consider the linear map @xmath340 from @xmath321 to @xmath47 defined by @xmath341 then the kernel of @xmath340 contains @xmath309 because @xmath342 , and @xmath343 implies that @xmath340 is not a zero linear map .
hence we can partition @xmath237 into cosets of @xmath344 in @xmath321 .
each coset of @xmath344 in @xmath321 contains exactly @xmath345 elements of @xmath237 , and each element in a coset has the same value under @xmath340 . therefore @xmath346 | we show how to convert a quantum stabilizer code to a one - way or two - way entanglement distillation protocol .
the proposed conversion method is a generalization of those of shor - preskill and nielsen - chuang .
the recurrence protocol and the quantum privacy amplification protocol are equivalent to the protocols converted from @xmath0 $ ] stabilizer codes .
we also give an example of a two - way protocol converted from a stabilizer better than the recurrence protocol and the quantum privacy amplification protocol .
the distillable entanglement by the class of one - way protocols converted from stabilizer codes for a certain class of states is equal to or greater than the achievable rate of stabilizer codes over the channel corresponding to the distilled state , and they can distill asymptotically more entanglement from a very noisy werner state than the hashing protocol . |
collective phenomena play an important role in heavy - ion collisions , but for a long time it was assumed that , at collision energies much greater than 1 gev / nucleon , only longitudinal and azimuthally symmetric transverse radial flow would survive . during the last few years
the situation has changed qualitatively .
anisotropic , directed as well as elliptic , flow has been observed at the bnl ags @xcite , and strong indications of elliptic flow at the cern sps @xcite have been demonstrated .
the theoretical understanding of the effect and model calculations involving anisotropic flow have progressed significantly ; different anisotropic flow patterns have been linked to such phenomena as quark - gluon plasma formation , softening of the equation of state @xcite , and mean - field effects during the fireball evolution @xcite .
the appropriate tools for flow studies at high energies have been developed @xcite .
it was also noticed that anisotropic flow could be important for other measurements , such as two - particle correlations @xcite .
anisotropic flow has become an essential part of the global picture of heavy - ion collisions at ultrarelativistic energies @xcite ; it is considered one of the key tools to elucidate the dynamics of the collision . in the current paper we present results of the analysis of anisotropic transverse collective flow of identified particles , protons and charged pions , detected in the e877 spectrometer in au + au collisions at a beam momentum of 10.8 and 11.4@xmath0 gev / c .
the data were taken during the 1993 and 1994 ags heavy - ion runs .
using calorimeter data we reconstruct the reaction plane event - by - event , and analyze the particle production with respect to this reaction plane . a similar analysis of charged particle multiplicity and transverse energy flow presented in @xcite displayed a strong directed flow as well as an elliptic flow with the primary axis in the reaction plane , _
i.e. _ perpendicular to the direction of the `` squeeze - out '' effect observed at lower energies @xcite . here
, using the e877 spectrometer data , we apply the same procedure to determine the triple differential distributions @xmath1 of identified protons and charged pions .
the paper is organized as follows . after a description of the apparatus we discuss the procedure of the analysis . in the subsequent sections we present first the results of the fourier analysis of azimuthal particle distributions in different rapidity bins and for different centralities of the collision .
these results are , to a large extent , independent of the uncertainties in the spectrometer efficiency and the detector acceptance .
then the results on azimuthal anisotropies and related quantities ( such as the mean transverse momentum projected into the reaction plane @xmath2 ) are discussed .
different flow scenarios are discussed from the point of view of the observed dependence of the flow signal on the transverse momentum @xmath3 of the particle .
we compare our measurements with the evaluation of nucleon and pion flow derived from the measurements of charged particle and transverse energy flow @xcite , as well as with model predictions ( rqmd versions 1.08 and 2.3 @xcite ) .
the e877 apparatus is shown in fig . 1 .
in the e877 setup , charged particles , emitted in the forward direction and traversing the collimator ( @xmath4 mrad , @xmath5 mrad ) , are analyzed by a high resolution magnetic spectrometer .
the spectrometer identifies particles via simultaneous measurements of momentum and velocity .
the momentum of each particle is measured using two drift chambers ( dc2 and dc3 , position resolution about 300 @xmath6 m ) whose pattern recognition is aided by four multi - wire proportional chambers ( mwpc ) .
the average momentum resolution is @xmath73% limited by multiple scattering .
a time - of - flight hodoscope ( tofu ) located behind the tracking chambers provides the time - of - flight with a typical resolution of 85 ps .
the spectrometer acceptance covers mostly the forward rapidity region with transverse momentum coverage including @xmath8 .
further discussion of the acceptance for different particle species can be found in @xcite .
a clean particle identification is particularly crucial for the study of flow since the extracted signal may be different for different particle species both in magnitude and in sign .
a small admixture in the particle sample of other particle species could therefore heavily bias the final results .
consequently , in the current analysis , strict requirements are applied for particle identification . using the measured momentum and velocity of the particle we calculate the particle mass . in a plot of momentum _ vs. _ the square of the particle mass @xmath9 we select entries in the region of @xmath10 around the @xmath9 peak of the particle under study and reject all entries within a @xmath11 region of another particle species . here
, @xmath12 represents the standard deviation of the ( gaussian ) mass - squared distribution at a given particle momentum @xmath13 .
the determination of the centrality of the collision and of the reaction plane orientation are made using the transverse energy flow measured in the target calorimeter ( tcal ) , and participant calorimeter ( pcal ) .
both calorimeters have @xmath14 azimuthal coverage and , combined together , they provide nearly complete polar angle coverage as viewed from the nucleus - nucleus center of mass system : tcal and pcal cover the pseudorapidity regions @xmath15 and @xmath16 , respectively @xcite . the pseudo - rapidity @xmath17 = - ln tan(@xmath18 )
is defined in terms of the polar angle @xmath19 in the laboratory frame .
the structure of the analysis is very similar to that used in @xcite , and many details can be found there . here
, the distributions @xmath20 of identified protons and charged pions are generated in a coordinate system where the @xmath21- and @xmath22-axes span the reaction plane .
the azimuthal angle in this system is defined as @xmath23 , where @xmath24 is the reaction plane angle , measured for every event using the direction of the transverse energy flow in tcal and pcal , and @xmath25 is the azimuthal angle of an individual particle in the laboratory frame .
the @xmath21-axis is defined in such a way that it points in the direction of the transverse energy flow at forward rapidities .
the transverse momentum components @xmath26 and @xmath27 of identified particles are evaluated in this coordinate system , and the particle distributions with respect to the reaction plane are constructed .
the azimuthal anisotropy of particle production is studied by means of fourier analysis of azimuthal distributions @xcite .
this yields the rapidity , transverse momentum , and centrality dependence of the fourier coefficients @xmath28 ( amplitude of @xmath29-th harmonic ) in the decomposition : e = = ( 1 + 2 v_1 ( ) + 2 v_2 ( 2 ) + 2 v_3 ( 3 ) + ... ) .
[ ed3 ] the e877 spectrometer provides full @xmath30 acceptance in the spectrometer only for a very limited @xmath3 range ( approximately @xmath31 mev / c ) , but the triple differential multiplicity is constructed in full , using the @xmath30 range of the reaction plane angle distribution .
note that the coverage of the calorimeters used for the reaction plane determination does not overlap with the spectrometer coverage , and thus the analysis is largely free from problems related to auto - correlations .
similarly to the analysis presented in @xcite , the reaction plane angle @xmath24 is determined in four non - overlapping pseudorapidity windows .
the reaction plane resolution , i.e. the accuracy with which the reaction plane orientation is determined , is evaluated by studying the correlation between flow angles determined in different windows .
finally , the flow signals are corrected for this resolution .
details of this procedure are described in @xcite . in short ,
the true value of each fourier coefficient @xmath28 is obtained by : v_n = v_n/(n(_1-_r ) , where @xmath32 is the observed signal , and @xmath33 is the mean cosine of the difference of the reconstructed ( @xmath24 ) and true ( @xmath34 ) reaction plane angles , characterizing the reaction plane resolution .
the values of @xmath33 are evaluated as outlined in @xcite . as a result
we obtain the true azimuthal distribution with respect to the reaction plane orientation : e = ( 1 + 2 v_1 ( _ lab-_r ) + 2 v_2 ( 2(_lab-_r))+ ... ) . [ ed3 t ] the e877 spectrometer has a relatively small azimuthal coverage . due to this , the analysis of azimuthal anisotropies is rather sensitive to biases of different kinds , which could simulate an event anisotropy and propagate to the final results .
for instance , during the off - line analysis , it was found that a small fraction of the recorded events ( @xmath35 24% ) have an anomalously high number of hits in the drift chambers . in these events
the hit density was too high to perform reliable tracking and the high occupancy of the spectrometer was found to be correlated with the orientation of the reaction plane .
a bias due to this was avoided by removing these events from the analysis completely and not only from the sample used to generate spectra . the tracking efficiency in the spectrometer ( typically on the order of 90% ) depends slightly on the spectrometer occupancy .
a special correction for this effect was developed and checked by a monte - carlo simulation .
the correction is based on the weighting of each track in accordance with the local track densities in the drift chambers and time - of - flight wall . due to different gain factors and dead towers in the calorimeters
the distribution in the reconstructed reaction plane angle is generally not flat .
special precaution was taken to make the reaction plane angle distribution as flat as possible and to remove possible biases ( see appendix a ) . the fourier coefficients of azimuthal distributions evaluated by the procedure described above ( and corrected for the reaction plane resolution ) , are combined with the measurements of @xmath36 and rapidity spectra @xcite . in this way
the triple differential distributions in @xmath37 , @xmath36 , and @xmath38 ( see equation ( [ ed3 t ] ) ) are determined and analyzed .
anisotropic flow reveals itself already in the dependence on the azimuthal emission angle of the inverse slope parameter of the invariant spectra @xmath39 when plotted as a function of @xmath40 for a given @xmath38 . here
, @xmath41 is the transverse mass .
we extract the inverse slope parameter @xmath42 by fitting the spectra in the region @xmath43 gev / c@xmath44 by a thermal ( boltzmann ) function e m_t ( - ) .
[ efun3 ] the shape of the spectra is not perfectly reproduced by this function , and in order to obtain a good quality description of the spectra in the entire @xmath45 region , the weights of all points were chosen to be equal for the fit and not in accordance with statistical errors . in fig .
[ fspectra ] we show the @xmath45 spectra of protons emitted into the rapidity interval @xmath46 at different angles relative to the reaction plane together with these thermal fits for different centralities
. clearly , the spectra of protons emitted in the direction of flow ( @xmath38=0 ) are significantly flatter than those of protons emitted in the opposite direction ( @xmath47 ) . to visualize the angular dependence of the inverse slope parameters the results from the fit are shown in fig .
[ fteff ] for three rapidity intervals . here
@xmath48 is plotted as a function of the azimuthal emission angle .
the results are presented for four centrality regions , selected in accordance with pcal @xmath49 and corresponding to the values of @xmath50 2313% , 139% , 94% , and @xmath514% ( see fig . 4 in @xcite ) .
the value of @xmath52 is obtained by an integration of @xmath53 from a given value of @xmath49 to the maximal one observed , and the geometric cross section is defined as @xmath54 b , with @xmath55 and @xmath56=1.2 fm .
the results shown in figs . [ fspectra ] and [ fteff ] are not corrected for the reaction plane resolution .
the correction would increase the difference between the maximal and minimal values of @xmath48 by about a factor of 1.5 .
such a correction ( in terms of @xmath57 ) is rather complicated ; for the quantitative description of the flow effects we use a fourier analysis of the azimuthal distributions @xcite where all corrections are implemented . we quantify directed flow by @xmath58 , the amplitude of the first harmonic in the fourier decomposition of the azimuthal particle distribution defined in equation ( [ ed3 ] ) and corrected for the reaction plane resolution .
the coefficient @xmath58 is analyzed as a function of transverse momentum for different rapidity bins and collision centralities .
the results for @xmath59 , corrected for the reaction plane resolution , are presented in figs .
[ fv1prot][fv1pm ] for protons and charged pions .
the error - bars shown in all figures represent statistical errors only .
the systematic uncertainties have mostly two sources : i ) the uncertainty in the determination of the reaction plane resolution leading to a relative error in @xmath58 of the order of 510% , similar for all particle species .
ii ) the uncertainty in the occupancy correction , the accuracy of which we estimate , by inspection of monte - carlo simulations , to be of the order of 2030% .
the correction itself is different for different particle species .
it is negligible for positive pions ( and for negative pions from the 1993 run ) , which are registered in the low occupancy region of the spectrometer .
the correction is maximal , and reaches absolute values of about 0.1 for the data shown , for protons at low @xmath36 and/or high rapidities .
the rapidly increasing uncertainty in the occupancy correction in the spectrometer region close to the beam limits our measurements of proton flow at very low @xmath36 . multiplying the uncertainty in the occupancy correction with its absolute value we end up with a systematic error in the absolute value of @xmath58 of about 0.03 where the correction is maximal . as can be seen from fig .
[ fv1prot ] , proton emission is very strongly correlated with the orientation of the reaction plane .
protons of larger @xmath3 have larger values of @xmath60 with some tendency to saturation in the highest @xmath3 region .
the largest flow signal observed corresponds to a difference in the high @xmath36 proton yields along ( @xmath61 ) and opposite to the flow direction ( @xmath47 ) of almost a factor of 10 ( as can also be seen in fig .
[ fspectra ] ) . both positive ( fig .
[ fv1pp ] ) and negative ( fig .
[ fv1pm ] ) pions exhibit weak flow in the direction opposite to that of protons ( negative values of @xmath58 ) over most of the @xmath36 region studied .
the maximum negative values of @xmath62 are about @xmath63 , significantly less in magnitude than the flow signal observed for protons .
acceptance of negative pions for the magnetic field polarity used during the 1994 run we have combined data from the 1994 run and from the lower statistics 1993 run ( shown in fig . [ fv1pm ] as open points ) , where data exist for both polarities of the magnetic field . ]
independent of any flow scenario @xmath64 must vanish at @xmath65 , due to the continuity of the spectra .
the small non - zero values of @xmath58 for the lowest @xmath36 bin in figs .
[ fv1pp ] and [ fv1pm ] ( for low centralities and rapidities close to beam rapidity ; positive for negative pions and negative for positive pions ) are mostly due to the finite bin size of the data .
the details of the behavior of @xmath58 at very low @xmath36 are presented in fig .
[ fpipm ] where we show , in the same plot , the results for positive and negative pions in the rapidity region @xmath66 for two different centralities and with much finer binning . not only are the sign and magnitude of the flow signal very different for protons and pions but also the functional dependence on @xmath36 varies with particle species .
the @xmath58 values for protons grow almost linearly with @xmath36 over the entire @xmath36 region .
inspecting the transverse momentum dependence of @xmath58 for pions one can distinguish three regions in @xmath36 : i ) the very low @xmath36 region ( below @xmath67 gev / c ) where in fact the flow signals of positive and negative pions are different .
positive pions show @xmath58 values decreasing rapidly and monotonously towards negative values .
conversely , for negative pions , the flow signal becomes at first positive , reaches a peak at about @xmath36=0.01 - 0.02 gev / c , and then decreases towards negative values where the flow signals for both pion charges merge .
this is best seen in fig .
[ fpipm ] .
the merging point appears to depend on centrality , moving to lower @xmath36 for the more central events .
ii ) an intermediate region , approximately at @xmath68 gev / c , where @xmath58 is negative and only weakly dependent on @xmath36 for both pion charges . and
finally iii ) the high @xmath36 region where @xmath58 begins to rise and becomes positive .
there is in fact a very systematic rapidity dependence of the zero - crossing point : it moves to lower @xmath36 values with increasing rapidity occurring at @xmath36 = 0.5 - 0.6 , 0.4 - 0.5 , and 0.3 - 0.4 gev / c for the three rapidity bins at @xmath37=2.8 - 3.2 , 3.2 - 3.6 , and 3.6 - 4.0 , respectively . the centrality dependence of all flow signals is rather pronounced ; in the analyzed centrality region the magnitude of flow for all particles decreases for more central collisions .
directed flow of beam rapidity protons shows a relatively smaller centrality dependence than that of protons at lower rapidities ( see also fig .
[ fmpx ] below ) .
higher order harmonics ( @xmath69 and @xmath70 ) in the particle azimuthal distributions have also been analyzed .
the accuracy in the evaluation of the contribution of higher harmonics deteriorates with increasing order due to the finite reaction plane resolution ( see the discussion in @xcite ) .
this results in larger relative errors .
our results for the proton elliptic flow ( amplitude of the second harmonic in the proton azimuthal distributions ) as a function of transverse momentum are presented in fig .
[ fv2prot ] . a clear positive signal is observed in the high @xmath36 region for rapidities of 2.6 and larger .
there is an indication that the signal moves to higher @xmath36 values with decreasing rapidity .
this combined with the smaller acceptance in @xmath36 at rapidities below 2.6 may be the reason that no significant signal is observed there . in our previous measurements of transverse energy and charged particle flow @xcite a clear signal of elliptic flow was observed at all values of pseudorapidity .
the observed positive values of @xmath69 correspond to an elliptically shaped distribution with the major axis lying in the reaction plane .
this orientation of the elliptic component is perpendicular to what was measured at lower beam energies @xcite . within our spectrometer acceptance
, pions do not exhibit elliptic flow at the level of 23% .
the ( absolute ) accuracy in measuring @xmath70 is about 0.1 .
for all particles , rapidities , and centralities of the collision the observed signals are consistent with zero within this accuracy . weighted with @xmath36 and its probability distribution the coefficient @xmath58 yields @xmath71 , the mean value of the transverse momentum projected onto the reaction plane : = v_1(p_t ) p_t d p_t .
[ empx ] like any other integral quantity , @xmath71 contains less information than @xmath58 .
we nevertheless calculate this quantity in order to compare our results with results from other experiments and model predictions .
this is done by using spectra @xcite measured with the same apparatus .
our results of the value of @xmath71 for protons at different centralities are shown in fig .
[ fmpx ] . due to
the experimental acceptance in @xmath3 , a model - independent evaluation of this quantity is possible only at rapidities @xmath72 .
where it becomes necessary , we extrapolate @xmath73 to high @xmath36 using a thermal parameterization ( as used e.g. in equation ( [ efun3 ] ) ) . the filled points in fig .
[ fmpx ] correspond to @xmath71 calculated in accordance with equation ( [ empx ] ) , using the parameterization of @xmath59 shown in fig .
[ fv1prot ] as the solid ( upper ) curves to extrapolate @xmath58 into the @xmath36 range not measured ( for an analytic expression of this parameterization see equation ( [ ev1mts ] ) below ) .
the small difference between the parameterization and the data in the low @xmath36 region does not contribute visibly to the final result for @xmath71 .
the contribution of the non - measured high @xmath36 part of the spectra to @xmath71 is relatively small at rapidities @xmath74 ( less than 10% ) , where the acceptance in @xmath36 is large . at rapidity
@xmath75 this contribution accounts for about 40% of the value of @xmath71 .
the error - bars shown do not include the uncertainty in the effective slope parameters used for the extrapolation of the @xmath36 spectra or any systematic uncertainty associated with the extrapolation of @xmath58 . in order to assess the systematic error we evaluate the same quantity ( @xmath71 ) by using a different parameterization of the invariant triple differential distribution with @xmath71 as a parameter .
we use the following functional form to parameterize @xmath76 of protons in a given rapidity bin : d^3n / d^3p & & m_t ( - ) , [ efunc1 ] + & & m_t= , _
i.e. _ a thermal distribution with respect to an origin displaced along the @xmath77-axis . with the effective slope parameters taken from
@xcite we use function ( [ efunc1 ] ) to fit the experimental values of @xmath59 ( see fig .
[ fv1prot ] , dashed ( lower ) curves ) .
the extracted values of @xmath71 are shown in fig .
[ fmpx ] as open symbols .
the difference between the results obtained with the two different parameterizations ( filled and open symbols in fig .
[ fmpx ] ) gives an idea of the systematic uncertainty of the results , not including the systematic uncertainty in @xmath58 itself in the range where it is measured ( see above ) .
the latter could be important for the lowest rapidity region of @xmath78 , where the flow signal is very small ; there we estimate an associated systematic uncertainty of @xmath71 of about 20 mev .
the evaluation of @xmath71 values for pions was done by convoluting the experimental values of @xmath59 with the spectra without any extrapolation to the high @xmath36 region .
this is possible due to the relatively large @xmath36 acceptance for pions in the rapidity interval studied .
the extracted values of @xmath71 for pions are about an order of magnitude smaller than those for protons .
the results for the centrality region @xmath79 913% ( centrality 2 ) are presented in fig .
[ fmpx2 ] ( squares and triangles ) together with the corresponding proton results and will be discussed below .
one of the simplest pictures of directed flow would be the motion in the transverse plane of a thermalized source , localized in rapidity .
assuming for the simplest case of no radial expansion , the invariant momentum spectrum for particle emission from a thermalized source is described by : e^ * e^-e^*/t , [ edn ] where @xmath80 is the temperature , and @xmath81 is the particle s energy in the rest frame of the source . in the case of a thermal source moving in @xmath21-direction with velocity @xmath82 ( which we call the directed flow velocity ) , the value of @xmath81 can be obtained by a lorentz transformation : e^*= - _ x _ x p_t ( ) , [ eet ] where @xmath83 , and the energy @xmath84 is evaluated in the system moving longitudinally with the same velocity as the source , @xmath85 . here @xmath86 is the source rapidity . using equations ( [ edn ] ) and ( [ eet ] ) one can evaluate the first fourier coefficient @xmath58 by direct integration to yield v_1 = , [ ev1 ] where @xmath87 , and @xmath88 , @xmath89 are the modified bessel functions . for protons , @xmath90 and @xmath58 does not depend on the , generally unknown , value of @xmath86 . since we expect that @xmath91 it is useful to note that in that case , for relatively small values of @xmath36 , @xmath58 depends almost linearly on @xmath36 : v_1 ( ) .
[ ev1mts ] the solid lines shown in fig . [ fv1prot ] correspond to fits to the data using equation ( [ ev1mts ] ) and inverse slope parameters @xmath48 from @xcite as @xmath80 .
overall , the fit is rather good .
taking into account that , in general , one can not interpret the inverse slope parameters @xmath48 as a source temperature we would like to note that equations ( [ ev1],[ev1mts ] ) are still valid for the case that the invariant spectra at a fixed rapidity have a thermal shape with some effective temperature constant . in fig .
[ fbetax ] , the extracted values of @xmath82 of the proton source are shown for different collision centralities as a function of proton rapidity .
one can see that the transverse source velocity grows with increasing rapidity , possibly peaking around rapidity y=3.1 .
the maximum values are about 10 % of the speed of light .
this transverse velocity is found to decrease with increasing centrality of the collision .
these results have to be put into perspective with the transverse expansion velocities fitted to spectra of various particle species in the same reaction and close to mid - rapidity : in such an analysis transverse expansion velocities of about 50 % of the speed of light are required to describe the data@xcite .
this effect is largest in central collisions .
the transverse expansion velocity decreases away from mid - rapidity . looking more closely at the fits in fig .
[ fv1prot ] one notices a small deviation of the fit using equation ( [ ev1mts ] ) from the data at small transverse momenta ( @xmath92 gev / c ) and at rapidities @xmath93 .
there , protons exhibit a weaker or even opposite flow than expected from the simple model discussed above .
this could be an indication of transverse expansion as pointed out in reference @xcite .
there , it is shown that , depending on the relative magnitude of the sideward flow velocity @xmath94 and the transverse expansion velocity @xmath95 , a reduction or even sign change of @xmath58 at small @xmath36 is possible . for pions
we observe a dependence of @xmath58 on @xmath36 very different from that for protons .
it implies that the physics of pion flow is different from that of a moving source alone ( just as one would expect ) .
a possible explanation of the pion flow signal could be found in a superposition of different effects : absorption in nuclear matter , coulomb interaction with comoving protons , and sideward motion of the source . for pions produced in the center of the collision volume and moving in forward direction with velocities close to speed of light
, one would expect more nuclear matter on the side characterized by the direction of nucleon flow ( positive @xmath21 in our notation ) and , consequently , more absorption .
the effect of pion absorption on nucleons with energy- , or for a fixed rapidity , @xmath36-independent cross section would lead to a negative and constant value of @xmath59 .
such absorption should be comparable for positive and negative pions .
on the other hand , it is known that in the relevant momentum range the pion - nucleon elastic and total cross sections are strongly peaked at the @xmath96 resonance @xcite . in a frame where the nucleon is at rest the cross section peaks at a pion momentum of 0.3 gev / c and falls off rapidly for larger momenta .
if the absorption of pions would occur on nucleons of the same rapidity , this would mean that the negative values of @xmath58 should increase towards zero for @xmath97 0.3 gev / c .
most of the pions we observe are in a rapidity range where there are very few nucleons ( forward of beam rapidity ) .
indeed , analysis of the proton rapidity distribution indicates @xcite that nucleon sources ( fireballs ) are distributed evenly over plus and minus one unit of rapidity in the c.m .
i.e. _ that the most forward nucleon source is at rapidity 2.5 in the laboratory . for pions at rapidities 3.0 , 3.4 , and 3.8
this would imply that absorption effects should decrease for transverse momenta larger than 0.30 , 0.27 , and 0.20 gev / c .
the data indeed show this trend with rapidity at about the @xmath36 values expected ( see figs .
[ fv1pp],[fv1pm ] , and [ fpipm ] ) . at the same time coulomb interaction with
positively charged nuclear matter would be different for pions of different charge , and could qualitatively explain the difference in flow signals at very low @xmath36 shown in fig .
[ fpipm ] .
negative pions are attracted to the protons leading to positive values of @xmath58 ( as argued in the case of absorption on nucleons above ) while positive pions are repelled leading to more negative values of @xmath58 .
the rise of @xmath58 for pions towards large @xmath36 could reflect the sideward motion of the source , the same way it does for protons .
[ fmpx2 ] shows , for the @xmath49-range 200 - 230 gev , the mean directed transverse momenta for protons , positive , and negative pions . due to the acceptance of the spectrometer
the proton and pion results are forward of rapidity 2.2 and 2.8 , respectively .
results were obtained using equation ( [ ev1mts ] ) to complete the integral ( [ empx ] ) beyond the range where @xmath58 is measured .
the error - bars shown include an error in the effective temperature of about 7 % in addition to the statistical errors .
the symmetry of the collision system allows reflection of these results about mid - rapidity , shown in fig .
[ fmpx2 ] by the open symbols , providing values for protons backward of rapidity 0.9 and for pions backward of rapidity 0.3 . to fill the gap at intermediate rapidities we use our measurement of the flow signal in transverse energy @xmath49 and charged particle multiplicity @xmath98 as published in @xcite . in this paper transverse energy and charged particle flow were decomposed into nucleon and pion flow under a few simple assumptions .
both were studied as a function of pseudorapidity . using these results and a simple parameterization of the proton and pion spectra in @xmath37 and @xmath36 from the ags experiment e866 @xcite and our own measurement @xcite which together provide complete rapidity coverage , we evaluate @xmath71 as a function of rapidity .
more specifically , we parameterize the spectra by expressions of the form : & & ( - ) m_t ( - ) , and & & t_b(y)=t_0 ( - ) with values of @xmath99 = 0.89 ( 0.88 ) , @xmath100 = 0.26 ( 0.15 ) gev , and @xmath101 = 1.06 ( 1.70 ) for protons ( charged pions ) . using these parameterizations we calculate the mean rapidity @xmath102 ( weighted with @xmath49 ) and the width of the rapidity interval associated with a bin in pseudorapidity .
then , under the assumption that @xmath103 and @xmath104 are similar , one obtains @xmath71 as a function of rapidity : p_x(y ) p_t(y ) v_1^e_t(y ) .
the resulting mean transverse momenta are presented as stars in fig .
[ fmpx2 ] , together with our spectrometer results , for protons and pions ; the errors shown are a propagation of the statistical errors in the measured flow of @xmath49 and @xmath98 .
the horizontal error - bars shown with the stars represent the widths of the rapidity region actually contributing to the pseudorapidity bins for which @xmath58 was measured .
overall , the agreement between the results of the current analysis of spectrometer data at forward ( and backward ) rapidities and the results derived from the measurements of @xmath49 and @xmath98 flow is good . for the latter method we had estimated the systematic error for proton flow to be about 10% @xcite ; this can also be judged from the symmetry of the data shown in fig .
[ fmpx2 ] with respect to midrapidity ( note that all stars correspond to measured points ) .
the dotted line has been added to aid inspection of the data for the necessary antisymmetry .
the first and the next to last proton point have equal distance from mid - rapidity and should be identical .
they differ by 10 mev / c ( out of 80 mev / c ) . for the spectrometer
results the systematic errors are largest at the lowest rapidities shown ( see fig . [ fmpx ] and discussion ) where they are estimated to be 20 mev / c . within these errors
the points obtained from the two methods merge smoothly and taken together they provide a measurement of @xmath71 for protons and charged pions as a function of rapidity over practically the entire relevant rapidity range .
one can see a proton flow signal rising away from midrapidity to maximal values of about 130 mev / c close to beam and target rapidities .
the pion signal is in comparison very small , about 1/20 or the proton signal , but nevertheless significant .
it is directed opposite to the nucleon flow signal . in fig .
[ fmpx2 ] we compare our results for @xmath71 with the predictions of the rqmd event generator @xcite , versions 1.08 and 2.3 .
version 1.08 , run in cascade mode , has been used for comparison to our earlier results for au + au collisions on pseudorapidity distributions @xcite , proton and pion rapidity @xcite and transverse mass @xcite distributions as well as results for si+a and p+a collisions .
for the au + au system it was found that overall many features of the data are predicted correctly with one striking deviation : generally proton spectra are significantly too steep in the model ; at midrapidity the inverse slope constant is only 2/3 of the experimental value .
predictions from this version of the model have also been compared to our experimental results for flow in @xmath49 and @xmath98 @xcite and generally the model exhibited too little flow as compared to the data . comparing to the proton data in fig .
[ fmpx2 ] , we find that the model also underpredicts the proton flow by nearly a factor of two . for pions , the sign change of the flow relative to protons is properly predicted but in magnitude the model overpredicts this opposite pion flow by a factor of 2 to 3 . meanwhile , a new version of rqmd ( version 2.3 ) has been developed @xcite .
it has been known for some time @xcite that taking into account mean field effects by simulating in the model a skyrme - type nucleon - nucleon potential increases the slopes of proton spectra and at the same time increases the proton flow signal while reducing the opposite pion flow signal in magnitude .
we have used the most recent version of the model in the so - called `` mean field '' mode to compare to the present data as well .
indeed , in this version of the code the proton flow signal is larger by about 50% for the proton @xmath71 at the peak and the data are described rather well ( see fig .
[ fmpx2 ] ) .
simultaneously , the opposite pion flow is reduced , again to a level consistent with the experimental data .
it should be noted that at the same time the slope difference in the proton spectra is reduced ( at midrapidity the slope constant of the model is now 3/4 of what is observed in the data ) but a deviation persists at all rapidities .
[ fv1rqmd ] shows that the agreement of the rqmd predictions ( version 2.3 , mean field ) with the experimental results for @xmath71 of protons as displayed in fig .
[ fmpx2 ] may indeed be accidental .
the functional dependence of the proton flow signal , expressed now in terms of the first fourier coefficient @xmath58 as a function of @xmath36 , is very different from what is seen in the data . in the model ,
the flow signal rises at first with increasing @xmath36 and then becomes rather flat .
in fact , in the rapidity bin around y = 2.9 the model overpredicts @xmath58 but at the same time underpredicts the overall slope of the proton spectrum by 25 % giving overall agreement in @xmath71 .
an interesting question arises whether the rapidly rising and saturating @xmath36 dependence as exhibited by the rqmd predictions ( which is in fact a general trend observed for all rapidity bins ) is due to the rescattering - type mechanism which produces flow in a cascade code in contrast to a hydrodynamic flow mechanism ( reflected _ e.g. _ by the @xmath36 dependence of equations ( [ ev1],[ev1mts ] ) ) .
the elliptic flow signal seen in the anisotropy of @xmath49 and @xmath98 was found to be in good agreement with the rqmd ( version 1.08 ) prediction in our previous publication @xcite . as discussed above , the sign of this elliptic flow signal changes from negative ( preferential emission perpendicular to the reaction plane ) at lower beam energies ( 1 - 2 gev / nucleon range at the bevalac and sis ) to a positive signal ( preferential in - plane emission ) at the present energy .
it was recently noted by sorge @xcite that the @xmath69 anisotropies are sensitive to the pressure at the maximum compression .
positive values of @xmath69 were in fact predicted in a hydrodynamic model by ollitrault for very high energies @xcite .
the observed change in sign between bevalac / sis and ags energies therefore implies that at ags energies already the final in - plane flow overwhelms the initial shadowing effect
. it would be interesting to deduce , from this information , the pressure achieved at maximum baryon density in au + au collisions at the ags .
the measured flow signals of identified particles show that , in semi central collisions , protons ( nucleons ) exhibit strong directed flow .
pions of low @xmath36 exhibit flow in the opposite direction ; at higher @xmath36 pions start to flow in the same direction as protons .
the results on @xmath71 values for proton and pion are complementary in acceptance and match well with nucleon and pion flow values derived from the same experiment by measurement of @xmath49 and @xmath98 flow @xcite . taken together , these results represent a directed flow measurement of protons and pions over nearly the entire rapidity region in au+au collisions at 11@xmath0 gev / c . the flow is found to be maximal around beam and target rapidities with values of @xmath71 for protons of about 130 mev / c and 5% of that for charged pions .
the observed positive values of the second harmonic amplitudes of the proton azimuthal distributions correspond to preferential particle emission in the reaction plane .
this observation agrees with our previous measurements of @xmath49 and @xmath98 elliptic flow @xcite .
the nearly linear dependence of the proton flow signal @xmath58 on @xmath3 can be interpreted in the framework of a transversely moving thermal source .
the corresponding source velocity @xmath94 appears to reach values of 0.1 in the beam rapidity region and for semi
central collisions .
the more complicated @xmath36 dependence of the pion flow signals indicates that the effect there is probably a superposition of several effects such as absorption , coulomb interaction and overall motion of the source ( as for protons ) .
we thank the ags staff , w. mcgahern and dr .
h. brown for excellent support and acknowledge the help of r. hutter in all technical matters . financial support from the us doe ,
the nsf , the canadian nserc , and cnpq brazil is gratefully acknowledged .
one of us ( jpw ) thanks the a. v. humboldt foundation for support , while another ( wcc ) acknowledges the financial support from the gottlieb - daimler and karl - benz - stiftung during his stay in heidelberg , germany , where part of this work was done .
we apply the following procedure to correct for the non - flatness of the reaction plane distribution .
any flattening of the distribution means a correction to the reaction plane angle .
we explicitly introduce this correction defining a new angle as : _
1=_1+_1 . where @xmath105 is written in the form : _
1= _ n ( a_n ( n _ 1 ) + b_n ( n_1 ) ) requiring the vanishing of the @xmath29-th fourier moment of the new distribution , the coefficients @xmath106 and @xmath107 can be evaluated by the original distribution .
b_n & = & ( n_1 ) , + a_n & = & - ( n_1 ) , where the brackets refer to an average over events .
this gives : _
1 = _ 1 + _ n ( -(n_1)(n_1 ) + ( n_1)(n_1 ) ) in practice , we flatten the reaction plane distribution up to the fourth harmonic ( n=4 ) .
note that , due to the small values of @xmath108 and @xmath109 ( typically of the order of a few percent ) , such a flattening of the distribution does not have any effect on the reaction plane resolution .
it can also be shown that the same flattening procedure removes possible trigger biases ( due to imperfect calibration , dead channels or any other asymmetry ) at least up to the second order .
1 . [ fe877 ]
the e877 apparatus .
[ fspectra ] proton @xmath45 distributions in the rapidity interval @xmath46 for different centralities ( indicated by an @xmath49 range in gev ) together with a fit using the function given in equation ( [ efun3 ] ) .
the inverse slope values and their statistical errors are shown in gev .
the open ( filled ) squares correspond to @xmath110 ( @xmath111 ) .
[ fteff ] the dependence of the inverse ( boltzmann ) slope @xmath57 of the proton distributions as a function of azimuthal angle .
the open points are reflections of the filled points about @xmath61 .
[ fv1prot ] the transverse momentum dependence of the first moment ( @xmath58 ) of the proton azimuthal distributions for different particle rapidities and centralities of the collision .
the solid and dashed curves are fits using functions given in eqs .
( [ ev1mts ] ) and ( [ efunc1 ] ) , respectively . for a description
see sections [ sresu ] and [ sdisc ] .
[ fv1pp ] same as fig .
[ fv1prot ] for @xmath112
[ fv1pm ] same as fig .
[ fv1prot ] for @xmath113 .
the open and filled symbols correspond to data from the 1993 and 1994 runs .
[ fpipm ] fourier coefficients @xmath58 for pions in the low @xmath36 region for the rapidity interval @xmath66 and two different centralities .
solid symbols represent data for positive pions .
open symbols are for negative pions ( circles and squares correspond to results from the 1993 and 1994 runs , respectively ) .
[ fv2prot ] the transverse momentum dependence of the second moment ( @xmath69 ) of the proton azimuthal distributions for different particle rapidities and centralities of the collision .
[ fmpx ] the mean projection @xmath71 of the proton transverse momentum onto the reaction plane as a function of rapidity for different centralities of the collision .
results for the second and fourth centrality bin are shifted to the left by 0.025 for clarity of the picture .
solid and open symbols correspond to fits of @xmath59 with functions given in eqs .
( [ ev1mts ] ) and ( [ efunc1 ] ) , respectively . 10 .
[ fbetax ] the transverse velocity @xmath82 of a thermal ( proton ) source for different rapidities and centralities of the collision . for details see text
[ fmpx2 ] mean projection @xmath71 of the proton and pion transverse momentum onto the reaction plane as a function of rapidity for the centrality bin with @xmath49 = 200 - 230 gev .
solid circles , squares , and triangles correspond to measurements in the spectrometer .
open circles , squares , and triangles are reflections of the filled symbols about midrapidity .
the results derived from measurements of @xmath49 and @xmath98 flow @xcite for nucleons and pions are shown by stars ( see text ) .
the dotted line is added to aid inspection of the solid stars for antisymmetry about mid - rapidity .
results from calculations using 2 versions of rqmd ( 1.08 in cascade mode , 2.3 in mean field mode ) are depicted as histograms .
[ fv1rqmd ] comparison of measured values of @xmath58 for protons in two rapidity bins and the centrality bin with @xmath49 = 200 - 230 gev with those predicted by 2 versions of the event generator rqmd .
h. gutbrod , k. h. kampert , b. w. kolb , a. m. poskanzer , and h. g. ritter , phys . lett . *
b216 * , 267 ( 1989 ) ; d. brill _ et al .
lett . * 71 * , 336 ( 1993 ) ; c.pinkenburg and k.d .
hildenbrand , acta phys .
polonica * 27 * , 243 ( 1996 ) ; s. wang _
et al . _ ,
* 76 * , 3911 ( 1996 ) . | results are presented of an analysis of proton and charged pion azimuthal distributions measured with respect to the reaction plane in au + au collisions at a beam momentum of about 11@xmath0 gev / c .
the azimuthal anisotropy is studied as a function of particle rapidity and transverse momentum for different centralities of the collisions .
the triple differential ( in rapidity , transverse momentum , and azimuthal angle ) distributions are reconstructed .
a comparison of the results with a previous analysis of charged particle and transverse energy flow as well as with model predictions are presented . |
an important question in particle physics is the nature of new physics beyond the astoundingly successful standard model .
one compelling scenario for this new physics is the weak scale supersymmetry which provides an answer to a number of puzzles of the standard model such as a solution to the gauge hierarchy problem , the origin of electroweak symmetry breaking as well as providing a candidate for dark matter of the universe .
an added virtue is that it unifies the disparate strengths of the weak , electromagnetic and strong forces at a scale of about @xmath1 gev , opening up another rich landscape of new physics around this scale .
this high scale new physics can only manifest itself at low energies via large radiative correction effects which depend logarithmically on mass as well as through effects such as proton decay which are suppressed by this new high scale .
there is reason to believe that there might be a manifestation of this new physics effect in the domain of neutrinos .
this arises from the consideration of seesaw mechanism@xcite which provides a very natural way to understand the extreme smallness of neutrino masses and which requires the existence of massive right handed neutrinos with at least one having mass close to the gut scale .
it is therefore suggestive that both the seesaw scale and the gut scale are one and the same .
this connection becomes particularly plausible in the context of models based on the grand unifying groups such as so(10)@xcite or @xmath2@xcite that predict the existence of the right handed neutrino as part of the common fermion multiplet along with the fermions of the standard model . in a recent paper
, this concept of unification of gut scale and seesaw scale was taken one step further by making the plausible hypothesis that quark and lepton mixings angles may be same at the seesaw scale due to quark lepton unification whereas the observed large mixings at the weak scale are a consequence of radiative corrections . using the formulae for renormalization group evolution for neutrino mass@xcite , it was shown that this possibility can be realized if the neutrino masses are quasi - degenerate and have same cp@xcite .
models based on @xmath3 group were presented where both the quasidegeneracy as well as mixing unification at the seesaw scale arose from the so - called type ii seesaw formula@xcite .
detailed renormalization group evolution of mixing angles then revealed that one indeed gets the desired bilarge mixing pattern at low energies which are in agreement with the solar and the atmospheric neutrino data .
the third angle also undergoes radiative magnification , but remains small and well within the chooz - palo - verde limit@xcite due to the smallness of the corresponding initial value which is the quark mixing , @xmath4 .
furthermore , our hypothesis has the interesting prediction that the common value for the quasi - degenerate neutrino mass is larger than 0.2 ev , which can be tested in the next round of the neutrinoless double beta decay experiments@xcite and is in the range claimed in ref.@xcite .
it also overlaps with the range accessible to the katrin experiment@xcite and consistent with the bound obtained from wmap@xcite .
threshold corrections at the low - energy susy scale have been found to improve agreement@xcite with the most recent data including kamland and sno@xcite for a particular choice of the susy particle spectrum .
as mentioned , two important ingredients that allow the radiative magnification to work are low - energy supersymmetry(susy ) and large values of @xmath5 which contribute large logarithmic factors for radiative magnification via amplified values of @xmath6-yukawa coupling .
recently , a new unification scenario has been advocated@xcite which is based on the observation that one can maintain gauge - coupling unification and neutralino dark matter without necessarily `` buying '' the entire machinary of weak scale supersymmetry but rather keeping only a subset of susy particles i.e. the gauginos and the higgsino at the weak scale and pushing the rest of the superpartners to high intermediate scales(@xmath7 ) .
this has been called split supersummetry .
this approach discards the naturalness requirement for the higgs mass that motivated the weak scale supersymmetry- thus one has to fine tune the higgs mass to every loop order .
one also has no simple mechanism to understand the electroweak symmetry breaking . despite these disadvantages
, one might consider this as an interesting minimal extension of the standard model that preserves gauge - coupling unification and provides a neutralino dark matter and discuss how neutrinos fit into it .
clearly , there is no inherent obstacle to implementing the seesaw mechanism and quark - lepton unification in these models .
only thing one has to keep in mind is that since the seesaw physics and quark - lepton unification generally tend to add new contributions to the gauge coupling evolution , if we want to retain the simple unification of three standard model gauge couplings , the seesaw scale must be at or above the gut scale . in this paper , we work within a framework of this type and see if we can understand the large neutrino mixings via the radiative magnification mechanism . below the high susy scale(@xmath7 ) in split supersymmetry , apart from the presence of gauginos and higgsinos , the effective theory is governed by nonsusy standard model with one light higgs doublet . as the two important ingredients in the radiative magnification scenario i.e. the long intervals of running in the presence of weak scale susy and the enhancement factor due to @xmath5 are missing in the split supersymmetry below @xmath8 , a natural apprehension emerges about the validity of this mechanism@xcite for large neutrino mixings .
working within a model with quark - lepton unification based on the group @xmath9 and degenerate masses for neutrinos obtained via the type ii seesaw mechanism@xcite and the usual split susy spectrum , we show that this model does indeed lead to bilarge mixing pattern at low energies .
the @xmath10 symmetry guarantees mixing unification between the quark and the lepton sectors at the seesaw scale ( i.e. @xmath11 ) .
the new point which is different from the work of ref.@xcite is that running of neutrino masses is now different .
furthermore , in order to maintain gauge coupling unification , we must have the seesaw scale equal to the gut scale , as already noted .
we carry out this running effect and show that despite the high split susy scale , the radiative magnification mechanism operates successfully to give the bilarge mixings at the weak scale provided that , in addition to the quasi - degenerate majorana neutrino masses with same cp , we have the scalar susy partners with masses starting at least 2 - 3 orders below the seesaw scale . whereas with low - energy susy , the radiative magnification occurs at @xmath12 tev , in split supersymmetry it occurs at the high value of the susy scale which may be any where between @xmath13 gev-@xmath14 gev . below this scale
, the rge running does not affect the neutrino mixings significantly .
what is interesting is that radiative magnification occurs for the scalar susy partner masses close to even the scale @xmath14 gev so that a major part of the running is nonsupersymmetric .
including small threshold effects needed to fit the details like the solar mass - squared difference , the extrapolated low enery values of @xmath15 , @xmath16 , and the mixing angles are in excellent agreement with the current solar , atmospheric , and the reactor neutrino data .
this paper is organized as follows : in sec .
2 we discuss the relevant rges in the context of split supersymmetry . in sec .
3 , we present allowed perturbative region of @xmath5 , radiative magnification of mixing angles at high susy scales and low energy extrapolation of masses and mixings . in sec . 4
, we discuss threshold effects and their estimations .
discussion of the results and conclusions are presented in sec.5 .
in this section while discussing renormalization group equations(rges ) for neutrino masses and mixings we point out some special features of split supersymmetry that contribute to radiative magnification of the mixings at high susy scales . since we will follow the bottom - up approach to determine initial conditions at the seesaw scales for rg - evolutions of neutrino parameters in the top - down approach , we use the stadard rges for the susy , split supersymmetry , and nonsusy cases as applicable in appropriate domains@xcite . in the context of split supersymmetry where the scale(@xmath8 ) of the scalar superpartner masses could be much larger than the weak scale@xcite , the seesaw scale(@xmath17 ) is clearly always larger than the susy scale(@xmath8 ) i , e @xmath18 .
we assume the neutrino mass eigenstates to be quasi - degenerate and possess the same cp .
we also ignore all cp violating phases in the mixing matrix , and adopt the mass ordering to be of type @xmath19 . parametrizing the @xmath20 mixing matrix as u=,[eq1 ] where @xmath21 and @xmath22 )
, the rges for the mass eigen values and mixing angles can be written as @xcite = -2f_m_iu_i^2-m_if_u = b^(m)_im_i,(i=1,2,3).[eq2 ] & = & -f_c_23 ^ 2 ( -s_12u_1d_31+c_12u_2d_32 ) , [ eq3 ] + & = & -f_c_23c_13 ^ 2 ( c_12u_1d_31+s_12u_2d_32 ) , [ eq4 ] + & = & -f_c_12(c_23s_13s_12u_1 d_31-c_23s_13c_12u_2d_32 .
+ & & .+u_1u_2d_21).[eq5 ] here @xmath23 and , for mssm with @xmath24 , f_&=&-h_^2/(16 ^ 2 ^ 2 ) , + f_u&=&(1)(65g_1 ^ 2 + 6g_2 ^ 2- 6h_t^2 ^ 2),[eq6 ] but , for @xmath25 , f_&=&3h_^2/ ( 32 ^ 2 ) , + f_u&=&(3g_2 ^ 2 - 2 - 6h_t^2 - 2 t ) /(16
[ eq7 ] with the definitions of the couplings at @xmath7 , ( m_s)&=&g_2(m_s ) , ( m_s)=g_2(m_s ) , + ( m_s)&=&(3/5)^1/2g_1(m_s ) , ( m _ s ) = ( 3/5)^1/2g_1(m _ s).[eq8 ] the additional term present in eq.(7 ) which is specific to the split susy below @xmath7 and the higgs quartic coupling are , t&=&32(^2+^2)+
12(^2+^2 ) , + ( m_s)&=&4 ^ 2 2.[eq9 ] + the basic mechanism responsible for radiative magnification of mixings of quasi - degenerate neutrinos with same cp in mssm which has been pointed out earlier@xcite is also applicable with split susy but with high value of @xmath7 . since @xmath26 and @xmath27 are scale dependent , the initial difference existing between them at @xmath28 is narrowed down during the course of rg evolution as we approach the susy scale .
this causes @xmath29 large , and hence large magnification to the mixing angle due to radiative effects through eqs.(3)-(5 ) .
another major factor contributing to radiative magnification is due to amplified negative value of @xmath30 in eqs.(3)-(5 ) in the presence of susy by a factor @xmath31 in the large @xmath5 region . an approximate estimation of values of @xmath5 needed for radative magnification in split susy is possible by noting that the magnification has been realized in mssm with weak - scale susy for @xmath32 and in that case the major factor controlling magnification is due to amplification of @xmath6-yukawa coupling@xcite .
expecting similar amplification in split susy gives the appoximate relation , ^2 ( m^0_r / m^0_s)&= & ( h_)^2 ( m_r / m_s ) .
[ eq10 ] where quantities with zero superscript(without superscript ) refer to weak - scale(split ) supersymmetry .
noting that the ratio of the two @xmath6-yukawa coplings in eq.(10 ) is @xmath33 , and using @xmath34@xcite , and @xmath35 gev , the above relation gives the approximate requirements for split supersymmetry case as @xmath36 , and @xmath37 for @xmath38 gev , and @xmath39 gev , respectively .
in fact , it has been noted that split susy permits larger values of @xmath5 than the mssm with weak scale susy@xcite . it would be shown in sec.3 that the allowed perturbative upper limit of @xmath5 increases with increase of the susy scale .
thus even though the supersymmetric part of the running of neutrino parameters is much less than in the weak scale supersymmetry case , larger values of tan beta increases the value of @xmath30 and this in turn leads to the desired enhancement of the radiative magnification effect on the mixing angles .
in addition to the above , here we point out that there is a specific mechanism which operates in the allowed parameter space of split susy that causes mass eigenvalues to approach one another faster and thus , drives radiative magnification even if the susy scale is many orders larger than the weak scale . noting that the @xmath40-function coefficient for the evolution of mass eigenvalue @xmath41 in eq.(2 ) is @xmath42= @xmath43 , near the seesaw scale @xmath44 , for @xmath45 because of high - scale mixing unification constraint , but @xmath46 .
therefore , the evolution of the third mass eigen value would be different from the first two .
further , since supersymmetric @xmath47 is negative for @xmath48 and , depending upon the allowed values of the gauge and the yukawa coupling constants contributing to @xmath49 in eq.(6 ) above the high susy scale , the @xmath40-function coefficient may be positive or negative thus resulting in the decreasing or increasing behavior of the mass - eigen value below the seesaw scale down to @xmath7 .
in particular with split supersymmetry @xmath50 may be positive due to dominance of gauge couplings over the top - yukawa coupling resulting in negative values of @xmath51 for @xmath48 .
but since @xmath30 is negative , the positive value of the first term , @xmath52 , may partly or largely cancel with the second term leading to a large or small positive value of @xmath53 . as a result
, the first two mass eigen values may be expected to increase but the third may decrease , rapidly or slowly , from their input values at @xmath54 .
this increasing behavior of some mass eigen values(@xmath55 ) in the region @xmath56 would be in sharp contrast to that in mssm with weak - scale susy where the allowed high scale values of the gauge and the yukawa couplings and the mixing unification constraint have been found to make all the three @xmath40-function coefficients positive resulting in the decrasing behavior of the three mass eigen values below the seesaw scale although with different rates@xcite .
thus , with split susy the allowed values of the parameter space may cause three mass eigen values to approach sufficiently closer to one another even after much shorter interval of running .
this may occur at high values of @xmath7 which may even be only @xmath57 orders smaller than @xmath54 .
then the functions @xmath29 large near @xmath7 causing magnification of the mixing angles . on the otherhand for @xmath58 ,
the @xmath6-yukawa coupling has negligible contribution and the @xmath40-function coefficients are nearly the same for all the three eigen values and approximately equal to @xmath59 . from eq.(7 ) is clear that in the presence of nonsusy sm , @xmath49 is negative below @xmath7 resulting in positive beta - function coefficients for all mass - eigen values causing them to decrease approximately in the same manner down to @xmath60 .
it is easy to check , using boundary conditions given in eqs.(8)-(9 ) , that @xmath49 is negative at the starting point @xmath7 of the nonsusy theory , with @xmath61 .
although this feature is also common to mssm with the weak - scale susy , in split susy the running interval for nonsupersymmetric theory is much larger .
our numerical solutions to masses and mixing angles reported in sec.3 with split susy are found to corroborate these properties of the rg - evolution .
in this section we discuss the realization of radiative magnification of neutrino mixing angles at high susy scales and low energy extrapolations of mass eigen values and mixing angles for comparison with the available experimental data .
perturbative limit on @xmath5 * + the fact that grand unification with split supersymmetry allows larger values of @xmath5 than those with weak - scale susy has been also noted in ref.@xcite . since the value of @xmath5 is an important ingredient , at first we estimate the maximum allowed values of @xmath5 for every value of @xmath7=@xmath62 gev - @xmath14 gev . for this purpose
we use the standard pdg values of masses and couplings@xcite at @xmath63 and follow the bottom - up approach scanning the values of the three gauge couplings and the third generation yukawa couplings while varying input values of @xmath5 over a wide range at all values of @xmath64 up to the gut - scale using the rges in the appropriate regions@xcite .
the lowest allowed value of @xmath5 is determined when the top - quark yukawa coupling reaches the perturbative limit(@xmath65 ) .
similarly the upper limit is determined by noting that the @xmath6-lepton yukawa coupling attains its perturbative limit at the gut - scale .
these are shown in fig.1 .
it is clear that the upper limit on @xmath5 is larger compared to the weak - scale susy model and this limit also increases quite significantly with increasing value of @xmath7 in split supersymmetry . for examle
wth @xmath7=@xmath66 gev , @xmath67 gev , @xmath39 gev , and @xmath14 gev , , the upper limits are found to be @xmath68 , and @xmath69 , respectively . through this bottom - up approach
we also noted the values of different coupling constants and ckm mixings at different values of the seesaw scales starting from their low - energy values . some of the coupling constants at the seesaw scales are presented in table.1 where the factors like @xmath70 or @xmath71 are included in the values of susy yukawa couplings . with large values of @xmath5 allowed near the upper limit ,
the higgs mass prediction in split susy tend to be independent of this parameter and the two yukawa couplings , @xmath72 and @xmath73 , have negligible effects below @xmath8 .
+ * iii.2.radiative mgnification of mixing angles * + the values of gauge and yukawa couplings , the ckm mixing angles obtained at the seesaw scale from the bottom - up approach and finetuned values of light neutrino masses(@xmath74 ) are used as inputs to obtain solutions from the neutrino rges in the top - down approach .
these input parameters are given in table 1 .
the unknown parameters @xmath75 are determined in such a way that the magnified values of mixing angles are at first obtained at @xmath7 along with the mass eigen - values . these are further extrapolated down to @xmath63 through the corresponding rges to give closest agreement with the experimental data : & = & ( 5 - 8)10 ^ -5ev^2 , = ( 1.2 - 3)10 ^ -3ev^2 , + _ 23&=&0.67 - 0.707 , _ 12= 0.5 - 0.6 , _
130.16 .[eq11 ] as noted in sec.2 the signs and values of the three @xmath40-functions for the mass eigen values at the respective seesaw scales are easily checked from the initial values of couplings given in table 1 using eq.(2 ) .
for example for the case with @xmath76 gev , @xmath38 gev , @xmath77 , the values of coupling constants given in table.1 yield the @xmath40-function coefficients at @xmath17 to be @xmath78 , @xmath79 which , as per the predictions of sec.2 , are responsible for the slow decrease of @xmath80 and rather faster increase of @xmath81 and @xmath82 below the seesaw scale .
these features are clearly displayed in fig.2 .
similarly for @xmath76 gev , @xmath83 gev , @xmath77 , these coefficients are @xmath84 , @xmath85 at the gut - seesaw scale .
it is quite interesting to note that the mutual differences are narrowed down by the increase of the first and the second mass eigenvalues and simultaneous decrease of the third eigen value leading to large @xmath86 and the magnification at @xmath7 .
this may be contrasted with the low - energy susy case where all the three masses showed decreasing behavior although with different rates@xcite .
this increasing behavior of neutrino masses @xmath87 in contrast to the decrasing behavior in low - energy susy is due to different relative values of the gauge and yukawa couplings at higher scales in the large @xmath5-region in split susy some of which are given in table.1 .
below @xmath7 the decrease is almost uniform due to approximate equality of the corresponding @xmath40-function coefficients of the three mass eigenvalues in this region as discussed in sec.2 .
the solutions shown in the fig.4 and fig.5 clearly exhibit radiative magnification to bilarge mixings at the high values of @xmath7=@xmath88 gev , and @xmath7=@xmath39 gev corresponding to the evolutions of mass eigen values as shown in fig.2 , and fig.3 , respectively . + + for the seesaw scale at @xmath89 gev the magnification of mixings takes place even at such high susy - scale as @xmath90 gev .
it is clear that due to larger allowed values of @xmath5 and also due to opposite signs of @xmath40-function coefficients of mass eigen values at high scales , the radiative magnification of neutrino mixings is possible in split susy at high susy scales even if the susy - scale is onle @xmath57 orders smaller than the seesaw scale .
we find that although enhancement due to rg evolution occurs in the @xmath91 sector also , the predicted low energy value remains at @xmath92 which is well within the chooz - palo verde bound@xcite and can be tested in the planned @xmath93 search experiments .
in the previous section we considered the mass squared differences obtained at lower energies purely by rg - evolution and these , as shown in table 1 , are in agreement with gross features of the experimental data on atmospheric neutrinos but falls somewhat on the higher side of solar neutrino data . since threshold effects have been shown to make significant contribution on quasidegenerate neutrinos@xcite in this section we estimate them to spplement our rg - solutions of sec.3 .
in particular we note that the loop facotr at the high - susy - scale threshold(@xmath94 ) , where the heavy superpartners are located , assumes a very simple form .
the threshold loop factor for neutrino mass at the electroweak scale is similar to the sm@xcite in contrast to the weak - scale susy where the superpartners contribute predominantly near @xmath60@xcite , in split susy the dominant threshold corrections to neutrino masses occur near high susy - scale threshold @xmath95 .
since the mixing angles in the radiative magnification scenario are very nearly the same at both the thresholds , it is quite convenient to use the approximation for the pmns matix @xmath96 and evaluate them in the limit @xmath97 and @xmath98 which are very closely cosistent with our solutions@xcite . using the loop factors @xmath99 in the mass basis with ( @xmath100 ) the threshold effects on mass eigen values at the threshold @xmath101
are expressed as@xcite , m_ij(_0)=m_i(_0)_ij+m_i(_0 ) t_ij(_0)+m_j(_0 ) t_ji(_0),(_0= m_z , m_s ) [ eq12 ] transforming the loop factors in terms of those in the flavor basis , @xmath102 , it is straightforward to evaluate the effects in the limiting case and for quasidegenerate neutrinos with slightly different values of the common mass @xmath103 at @xmath104 , @xmath60 leading to the formula@xcite ( m_21 ^ 2)_th(_0 ) = 4m^2(_0)2_12[-t_e(_0 ) + ( t_(_0)+t_(_0))/2 ] , [ eq13 ] + ( m_32 ^ 2)_th(_0 ) = 4m^2(_0)^2_12[-t_e(_0 ) + ( t_(_0)+t_(_0))/2 ] , [ eq14 ] when corrections at both the thresholds are included along with the contribution due to rg - evolution , the analytical formula for the mass squared differences is expressed as , ( m_ij^2 ) ( m_z)= ( m_ij^2)_rg + ( m_ij^2)_th ( m_z)+(m_ij^2)_th ( m_s ) , ( i > j = 1,2,3 ) , [ eq15 ] where the first term in the rhs of eq.(15 ) has been already estimated from rg - evolution and the second and the third terms are estimated through eqs.(13)-(14 ) .
as the rg - evolution effects estimated in sec.3 have already given approximately the correct values of @xmath15 for atmospheric neutrino data and @xmath16 only @xmath105 larger than the kamland and sno data , small and simpler threshold effects might be sufficient to account for these deviations . in split supersymmetry
such contributions can easily arise from the high - scale susy - threshold effects at @xmath106 .
since @xmath107 in split supersymmetry , left - handed charged sleptons and sneutrinos have almost identical masses . denoting the masses of the left - handed sleptons by @xmath108 and
the right - handed charged sleptons mass by @xmath109 we obtain a simple formula for the loop factors at the high susy scale(@xmath106 ) , 16 ^ 2 t_(m_s)&=&(3/8)(g^2_1(m_s)/5+g^2_2(m_s ) ) ( ln(m_^l_^2/m_s^2)-1/2 ) + & & + _ ( 1/4)h_^2(m_s)(1+^2 ) ( ln(m_^r_^2/m_s^2)- 1/2).[eq16 ] + + ignoring the low - energy threshold effect at @xmath60 we find that with @xmath110 , or @xmath109 few times lighter than @xmath7 gives the desired threshold corrections with negative sign . for , example ,
using a simple and plausible choice of approximate degeneracy in the scalar superpartner spectrum around @xmath7 with @xmath111 , we have @xmath112\simeq ( h_\tau^2(m_s)\tan^2\beta ) [ ln(m_{\tilde \tau_r}^2/m_s^2)-1/2]/(128\pi^2)$ ] . then with @xmath113 , numerical values of threshold corrections are obtained as shown in table.1 .
such threshold contributions , when added to the rg - effects , bring the theoretical predictions in concordance with all the available neutrino data including those from kamland and sno .
in this section we discuss the implications of our results briefly and state our conclusions .
our solutions require quasi - degenerate neutrino mass eigen values in the range @xmath114 leading to the effective mass in neutrinoless double beta decay given by @xmath115=@xmath116= @xmath117 ev -@xmath118 ev .
this is accessible to all the experiments being planned to search for the neutrinoless double beta decay .
furthermore , this range of neutrino masses also overlaps with the range accessible to the katrin@xcite tritium beta decay experiment .
the prediction @xmath119=@xmath120 is also accessible to several planned long - baseline neutrino experiments@xcite as well as the planned reactor experiments@xcite .
as discussed in ref.@xcite the range of eigenvalues of neutrino masses is consistent with wmap observations and also with the combined analysis of wmap@xmath1212df grs data@xcite .
+ in summary , we had shown previously that in the mssm with weak - scale susy@xcite , the hypothesis of quark- lepton mixing unification and quasi - degenerate neutrino spectrum at the seesaw scale successfully explains the observed mixing pattern for neutrinos i.e. two large mixings needed for @xmath122 and @xmath123 and small mixing for @xmath124 at low energies . in this paper we have extended this discussion to the case of split supersymmetry where we find that , despite the absence of low - energy susy and the corresponding absence of rg - running with an amplified value of @xmath6-yukawa coupling over a considerable mass interval ( @xmath125-@xmath7 ) , radiative magnification of neutrino mixings does occur .
these deficits seem to be adequately compensated by the larger values of @xmath5 allowed in this case and the positive and negative values of @xmath40-function coefficients for the running of mass eigen values at high scales .
the key tests of the model still remain the common mass of neutrinos above @xmath117 ev and @xmath124 between 0.08 - 0.1 , as in @xcite .
thanks the institute of mathematical sciences for senior associateship .
the work of r.n.m is supported by the nsf grant no .
the work of g.r . is supported by the dae - brns senior scientist scheme of govt . of india .
99 p. minkowski , phys .
b67 * , 421 ( 1977 ) ; m. gell - mann , p. ramond , and r. slansky , _ supergravity _
( p. van nieuwenhuizen et al .
eds . ) , north holland , amsterdam , 1980 , p. 315 ; t. yanagida , in _ proceedings of the workshop on the unified theory and the baryon number in the universe _ ( o. sawada and a. sugamoto , eds . ) , kek , tsukuba , japan , 1979 , p. 95 ; s. l. glashow , _ the future of elementary particle physics _ , in _ proceedings of the 1979 cargse summer institute on quarks and leptons _
( m. lvy et al .
eds . ) , plenum press , new york , 1980 , pp . 687 ; r. n. mohapatra and g. senjanovi , phys .
lett . * 44 * 912 ( 1980 ) .
babu , c.n .
leung and j. pantaleone , phys .
* b319 * , 191 ( 1993 ) ; p. chankowski and z. pluciennik , phys . lett . *
b316 * , 312 ( 1993 ) ; s. antusch , m. drees , j. kersten , m. lindner and m. ratz , phys .
b * 525 * , 130 ( 2002 ) [ arxiv : hep - ph/0110366 ] ; s. antusch , j. kersten , m. lindner , m. ratz and m. a. schmidt , jhep * 0503 * , 024 ( 2005 ) [ arxiv : hep - ph/0501272 ]
. k.r.s .
balaji , a.s .
dighe , r.n .
mohapatra and m.k .
parida , phys .
* 84 * , 5034 ( 2000 ) ; phys .
* b481 * , 33 ( 2000 ) ; k.r.s .
balaji , r.n .
mohapatra , m.k .
parida and e. a. paschos , phys .
rev.*d63 * , 113002 ( 2001 ) .
r. n. mohapatra , m. k. parida and g. rajasekaran , hep - ph/ 0301234 ; phys . rev . *
d69 * , 053007 ( 2004 ) .
klapdor - kleingrothaus et al .
j. * a 12 * , 147 ( 2001 ) ; c. aalseth et al . phys . rev . *
d 65 * , 092007 ( 2002 ) ; h. v. klapdor - kleingrothaus et al .
. lett . * a 16 * , 2409 ( 2001 ) ; hep - ph/0303217 ; h.v .
klapdor - kleingrothaus , a. dietz and i.v .
krivosheina , foundations of physics * 32 * , 1181 ( 2002 ) . for some works on split supersymmetry , see g. f. giudice and a. romanino , hep - ph/0406088 ; n. arkani - hamed and s.dimopoulos , g. f. giudice and a. romanino , hep - ph/0409232 ; k. s. babu , t. enkhbat and b. mukhopadhyaya , arxiv : hep - ph/0501079 ; b. kors and p. nath , nucl .
b * 711 * , 112 ( 2005 ) [ arxiv : hep - th/0411201 ] ; m. drees , hep - ph/0501106 ; a. pierce , phys .
d * 70 * , 075006 ( 2004 ) [ arxiv : hep - ph/0406144 ] ; a. ibarra , hep - ph/0503160 ; r. mahbubani , arxiv : hep - ph/0408096 .
v. barger , m. s. berger , and p. ohmann , phys .
* d47 * , 1093 ( 1993 ) ; c. r. das and m. k.parida , eur .
j. * c 20 * , 121 ( 2001 ) .
d. g. lee and r. n. mohapatra , phys . lett .
* b 329 * , 463 ( 1994 ) .
m. drees , hep - ph/0501106 ; katri huitu , jari laamanen , probir roy , sourov roy , hep - ph/0502052 . .
e. j. chun and s. pokorski , phys .
* d62*,053001 ( 2000 ) ; p.chankowski and p. wasowicz , eur .
j. * c23 * , 249 ( 2002 ) .
s. eidelman _
b592 * , 1 ( 2004 ) .
m. v. diwan _ et al .
_ , phys .
d * 68 * , 012002 ( 2003 ) arxiv : hep - ph/0303081 ; d. ayrea et al .
hep - ex/0210005 ; y. itow et al .
( t2k collaboration ) hep - ex/0106019 ; i. ambats _ et al . _ ( nova collaboration ) , fermilab - proposal-0929 . | radiative corrections to neutrino mixings in seesaw models depend on the nature of new physics between the weak and the gut - seesaw scales and can be taken into account using the renormalization group equations .
this new physics effect becomes particularly important for models with quasi - degenerate neutrino masses where small neutrino mixings at the seesaw scale can get magnified by radiative renormalization effects alone to match observations .
this mechanism of radiative magnification which provides a simple understanding of why lepton mixings are so different from quark mixings was demonstrated by us for the standard supersymmetry scenario where the particle spectrum becomes supersymmetric above the weak scale . in this paper
, we examine this phenomenon in split supersymmetry scenarios and find that the mechanism works also for this scenario provided the susy scale is at least 2 - 3 orders below the gut - seesaw scale and one has larger values of @xmath0 . |
in a recent study @xcite we proposed a frustrated spring - network model to simulate the surface structures of crosslinked polymer gels into which voids are introduced through templating by surfactant micelles . experimentally , templated polyacrylamide gels are created by mixing acrylamide with a crosslinker in the presence of a surfactant @xcite .
the resultant gels have a wide range of pore sizes , making them ideal materials for macromolecular separation @xcite .
the crosslinking process introduces inhomogeneities in the spatial density of the gel that affect its surface configuration .
our results show that the surfaces generated by the spring - network model have a nontrivial , multiaffine scaling behavior at small length scales , with a crossover to scale - independent behavior on large scales . at small length scales we found that the @xmath0th order increment correlation function has a power - law behavior , with @xmath0-dependent exponents that change with the vacancy concentration .
however , experimental surfaces of surfactant - templated hydrogels studied by atomic force microscopy ( afm ) @xcite show @xmath0-independent self - affine scaling at length scales below the crossover length . in ref .
@xcite we discussed the possibility that the multiaffine behavior was due to a finite density of vertical discontinuities due to overhangs in the simulated surfaces .
later work showed analytically and numerically that when vertical discontinuities are introduced into a self - affine surface , it becomes multiaffine @xcite . here
we extend this study to hydrogel surfaces , and we show that smoothing the simulated surfaces by convoluting them with a gaussian changes the scaling behavior from multiaffine to self - affine , in agreement with the experimental afm results .
the smoothing process mimics the effect of the measurement process of an afm tip interacting with the soft substrate .
the model , described in detail in ref .
@xcite , consists of a two - dimensional triangular network of nodes interconnected by massless harmonic springs .
the nodes represent the crosslinkers , and the springs correspond to the polymer chains .
we impose periodic boundary conditions in the horizontal direction , while the top nodes are free to move in two dimensions .
the bottom nodes are fixed , corresponding to bonding to a rigid substrate .
we assume that each spring corresponds to a collapsed polymer chain with equilibrium end - to - end length @xmath1 , which is randomly chosen with probability density @xmath2 thus , @xmath1 is proportional to the square root of an exponentially distributed number of monomers between crosslinkers @xcite .
the constant @xmath3 is proportional to the inverse of the average number of monomers between crosslinkers .
consistent with the picture of a random - coil collapsed polymer chain , we require that the elastic constant of each spring , @xmath4 , be proportional to its equilibrium length , @xmath5 in dimensionless units @xcite . for each statistical realization of the system , the equilibrium spring lengths are chosen independently according to eq .
( 1 ) . in the initial network configuration
, all springs are placed on the bonds of a regular triangular lattice of unit lattice constant , such that all springs are stretched or compressed to unit length , @xmath6 .
this initial configuration is highly stressed .
however , by choosing the value of @xmath3 such that the average force exerted by each spring is zero , we ensure that the initial configuration is not _ globally _ stressed . from the initial configuration , the system is relaxed to one of its many local energy minima by a limited - memory broyden - fletcher - goldfarb - shanno ( l - bfgs ) quasi - newton minimization algorithm @xcite , and by repeating the calculations for several different realizations of the equilibrium spring lengths we verified that the statistical properties of the different relaxed configurations are essentially the same .
spring networks of @xmath7 nodes were generated , and voids were introduced by randomly removing nodes and the springs connected to them .
the hoshen - kopelman algorithm @xcite was applied to eliminate clusters that were not connected to the fixed substrate , thus mimicking the removal of non - bonded material which is washed away during the experimental preparation .
the one - dimensional surfaces are defined as the set of surface heights , @xmath8 , at equally spaced discrete horizontal points , @xmath9 .
the surface height at @xmath9 , @xmath8 , is taken as the largest vertical distance , measured from the bottom , of the intersections of the springs with a vertical line at @xmath9 , where each spring is represented by a straight line segment between the connected nodes .
this definition of the surface is described in more detail in ref .
@xcite .
the @xmath0th order generalized increment correlation function is defined as @xcite @xmath10 for many surfaces , @xmath11 displays a power - law behavior , such that @xmath12 where @xmath13 is the generalized hurst exponent @xcite . for self - affine surfaces ,
the scaling behavior is independent of @xmath0 , @xmath14 , but for multiaffine surfaces , there is an infinite hierarchy of scaling exponents , such that @xmath13 depends continuously on @xmath0 , at least for some range of @xmath0 values and length scales @xcite . very often surfaces present a mixed behavior . for length scales smaller than some crossover ,
@xmath15 , @xmath11 displays a non - trivial scaling behavior , while at length scales larger than this crossover , @xmath16 , @xmath11 is constant , @xmath17 . in our previous work
@xcite it was shown that the simulated gel surfaces are multiaffine , and the scaling exponents , crossover length scales , and saturation values of the increment correlation function are strongly dependent on the void concentration . despite the fact that the experimental surfaces seem to have a similar dependence on the void concentration ,
one major difference between the simulated and experimental surfaces remains : the experimentally measured surfaces are clearly self - affine on short length scales @xcite , while the simulated surfaces are clearly multiaffine @xcite .
recent work @xcite indicates that vertical discontinuities are one important source of multiaffinity , and the surfaces generated by our model certainly have many vertical discontinuities caused by overhangs . even with a weakly interacting afm tip , experimental measurement of a soft surface , such as the polyacrylamide gels studied in @xcite , should have the effect of smoothing out any vertical discontinuities and overhangs , thus showing a self - affine surface , even if the real surface is multiaffine . to mimic the effect of passing a weakly interacting afm tip over the spring - network surfaces
, we have smoothed the simulated surfaces by convolution with a gaussian of @xmath18 in the dimensionless spatial units of the problem .
this process eliminates the overhangs without changing the overall structural properties of the surfaces .
figure 1 shows a detail of the generated surface and the smoothed surface .
figure 2 shows the @xmath0th root of the generalized increment correlation function @xmath19 for systems with @xmath20 , and @xmath21 vacancies , for both the original and smoothed surfaces .
it is evident that when the discontinuities are eliminated by smoothing , the @xmath22 values become independent of @xmath0 , thus indicating that the surfaces are no longer multiaffine but have become self - affine after smoothing .
the @xmath22 exponents are plotted in fig . 3 versus the void concentration for the original spring - network surfaces .
for the smoothed surfaces , @xmath23 for all void concentrations , as expected for a smooth surface .
the multiaffine scaling observed is consistent with that caused by vertical discontinuities @xcite .
th root of @xmath24 from the spring - network surfaces for @xmath25 to @xmath26 in steps of 0.5 . for graphical simplicity
, the @xmath0 labels have been omitted , but @xmath27 , when @xmath28 .
the surfaces are averaged over 8 , 8 , and 10 independent realizations of the equilibrium spring lengths for the networks with 0% , 40% , and 49% vacancies , respectively .
the linear regions of the plot indicate power - law scaling , and for the original simulated surfaces , the different power - law exponents ( slopes in the log - log plots ) indicate @xmath0-dependent multiaffine scaling . after smoothing ,
the power - law scaling at small length scales is self - affine ( parallel lines in the log - log plots ) with @xmath29 for all @xmath0 . ] vs. @xmath30 for the spring - network surfaces .
open symbols indicate the original , multiaffine surfaces , while corresponding filled symbols indicate the smoothed , self - affine surfaces . as expected , @xmath23 for all of the smoothed surfaces .
the observed multiaffine scaling of the original surfaces is consistent with the behavior caused by vertical discontinuities as expected from ref .
the data are averaged over 8 independent realizations of the equilibrium spring lengths for 0% to 47% vacancies , 10 realizations for 49% vacancies , and 14 realizations for 49.5% vacancies . ]
in this paper we have shown that surfaces generated by a frustrated spring - network model can reproduce several aspects of the scaling behavior of surfaces of surfactant - templated polyacrylamide gels , as observed in recent afm experiments @xcite .
both real and simulated surfaces show a nontrivial scaling behavior characterized by a power - law form of the increment correlation function at small and intermediate length scales . at
larger length scales the increment correlation function reaches a saturation value .
this large - scale behavior reflects the average network structure , while the small - scale behavior is due to microscopic density fluctuations .
below some crossover length , the experimental surfaces are self - affine while the original simulated surfaces are multiaffine .
we show that after smoothing by convolution with a narrow gaussian , which eliminates vertical discontinuities , the simulated surfaces become self - affine .
such smoothing mimics the effect of passing a weakly interacting afm tip over the soft hydrogel surfaces .
the large - scale behavior is not affected by the smoothing process .
f. family , t. vicsek , _ dynamics of fractal surfaces _ ( world scientific , singapore , 1991 ) ; h.n .
yang , g.c .
wang , t.m .
lu , _ diffraction from rough surfaces and dynamic growth fronts _ ( world scientific , singapore , 1993 ) . | we show that smoothing of multiaffine surfaces that are generated by simulating a crosslinked polymer gel by a frustrated , triangular network of springs of random equilibrium lengths [ g.m .
buenda , s.j .
mitchell , p.a .
rikvold , phys .
rev .
e 66 ( 2002 ) 046119 ] changes the scaling behavior of the surfaces such that they become self - affine .
the self - affine behavior is consistent with recent atomic force microscopy ( afm ) studies of the surface structure of crosslinked polymer gels into which voids are introduced through templating by surfactant micelles [ m. chakrapani , s.j .
mitchell , d.h .
van winkle , p.a .
rikvold , j. colloid interface sci .
, in press ] .
the smoothing process mimics the effect of the afm tip that tends to flatten the soft gel surfaces .
both the experimental and the simulated surfaces have a non - trivial scaling behavior on small length scales , with a crossover to scale - independent behavior on large scales .
_ pacs : _
61.43.hv ; 89.75.da ; 82.70.gg ; 68.37.ps _ keywords : _ self - affine scaling ; multiaffine scaling ; hydrogels ; atomic force microscopy |
we will explore tropical properties of weighted , or metric trees @xmath2 using the representation theory of the special linear group @xmath1 we direct the reader to the book by fulton and harris @xcite and the book by dolgachev @xcite for an introduction to the representation theory of connected complex reductive groups over @xmath3 . recall that we can choose a borel subgroup @xmath4 , and a maximal torus @xmath5 with @xmath6 and associate to this data a monoid of weights @xmath7 in the characters of @xmath8 which classify irreducible representations of @xmath9 up to isomorphism .
this cone comes with an involution defined by the duality operation on representations @xmath10 the direct sum of all such representations forms a commutative algebra @xmath11 which is the coordinate ring of the quotient of @xmath9 by the unipotent radical of a chosen borel subgroup , @xmath12.$ ] for @xmath13 , this can be taken to be the subgroup of unipotent upper - triangular matrices .
choosing a borel subgroup also fixes a set of positive roots @xmath14 for @xmath15 which define a partial ordering on the weights , we say that @xmath16 if @xmath17 is a member of @xmath18 for @xmath19 the cone @xmath20 is generated over @xmath21 by @xmath22 fundamental weights , @xmath23 the weight @xmath24 is the so - called highest weight of the representation @xmath25 the main result of this paper expresses the @xmath0-dissimilarity vector of an arbitrary tree in terms of the fundamental weights of @xmath13 . in what follows @xmath26 denotes the non - negative members of @xmath27 or @xmath28 let @xmath29 be a trivalent tree with @xmath30 ordered leaves , and let @xmath31 be a function which assigns a weight ( or length ) to each edge of @xmath32 the weight function @xmath33 defines a metric on the leaves @xmath34 of @xmath2 where the distance @xmath35 between the leaves @xmath36 and @xmath37 is the sum of the weights on the edges of the unique path @xmath38 connecting @xmath36 to @xmath37 in @xmath32 we intentionally confuse the path @xmath39 with the set of edges it traverses .
@xmath40 obviously @xmath41 and @xmath42 we call the vector @xmath43 the @xmath44-dissimilarity vector of @xmath45 we may generalize this construction by introducing the convex hull of @xmath0 leaves @xmath46 as the set of all edges which appear in paths connecting some @xmath47 to some @xmath48 @xmath49 the @xmath0-dissimilarity vector @xmath50 is then defined as expected .
@xmath51 the set of @xmath44-dissimilarity vectors of weighted trees @xmath52 is well understood .
a weighted tree can be recovered from its @xmath44-dissimilarity vector , and the set of all @xmath44-dissimilarity vectors is characterized by the following theorem from tropical geometry , see @xcite , @xcite .
the set of @xmath44-dissimilarity vectors coincides with the tropical grassmanian .
@xmath53 the grbner fan of the plcker algebra @xmath54 / i_{m , n}$ ] has support @xmath55 so each vector @xmath56 in this space gives a grbner degeneration of the ideal @xmath57 to the ideal of initial forms @xmath58 .
the set of vectors @xmath56 which for which @xmath58 is monomial free is called the tropical grassmannian @xmath59 the above theorem implies that for a vector @xmath60 to be a @xmath44-dissimilarity vector , it must weight plcker variables @xmath61 in such a way that at least two monomials in each plcker relation @xmath62 have the same weight . for a point @xmath63 to satisfy this requirement ,
the maximum of @xmath64 must be obtained at least twice .
if this is the case , then we may find a tree @xmath65 such that @xmath66 since @xmath44-dissimilarity vectors characterize their respective weighted trees , we should expect that some operation on the @xmath44-dissimilarity vector of a weighted tree @xmath65 , probably tropical in nature , will yield the @xmath0-dissimiliarity vector , and indeed this is the case , see @xcite for the following theorem .
let @xmath67 be the set of length @xmath0 cycles in the set of permutations on @xmath0 letters .
then we have the following formula .
@xmath68 this defines an onto map @xmath69 given that @xmath70 and the tropical grassmannian @xmath71 live in the same space , and coincide for @xmath72 one would hope that these two sets always have a close relationship .
sturmfels and pachter asked if the set of @xmath0-dissimilarity vectors was always contained in the tropical grassmannian @xmath73 @xcite .
cools recently proved this for small @xmath74 @xcite and conjectured that the result holds for all @xmath74 the result was proved in general by giraldo , @xcite .
@xmath75 this means that the entries of @xmath76 always satisfy the tropical plcker equations , and the weighting defined by this vector defines a monomial free initial ideal @xmath77 the purpose of this note is to prove this theorem using tropical properties of the plcker algebra deduced from the related representation theory of @xmath13 .
the plcker algebra @xmath54/ i_{m , n}$ ] is a natural object in the representation theory of @xmath19 it appears as the subring of invariants of the diagonal action of @xmath13 on @xmath78 this is the first fundamental theorem of invariant theory .
@xmath79/ i_{m , n } \cong \c[m_{m\times n}(\c)]^{sl_m(\c)}\\\ ] ] the plcker algebra is exactly the subring generated by the plcker coordinates , @xmath80 .
let @xmath81 @xmath82,$ ] with @xmath83 then the value of the plcker coordinates at @xmath84 is defined by the following .
@xmath85\\\ ] ] we may rewrite this algebra in terms of the category of finite dimensional representations of @xmath13 as follows .
@xmath86^{sl_m(\c ) } = \bigoplus_{\vec{r } \in \mathbb{z}_+^n } [ v(r_1\omega_1^ * ) \otimes \ldots \otimes v(r_n \omega_1^*)]^{sl_m(\c)}\\\ ] ] here @xmath87 is the highest weight of @xmath88 as a representation of @xmath13 and @xmath89 .
with respect to this direct - sum decomposition , the plcker coordinate @xmath80 is a generator of the summand with @xmath90 in the @xmath91th place for all @xmath92 and the trivial representation everywhere else .
multiplication in @xmath93 $ ] has a nice description in terms of this direct sum decomposition as well , it is induced by the cartan multiplication maps in each component , where the tensor product is projected onto its highest weight summand .
@xmath94 we may rewrite each summand in terms of homomorphisms from the category of @xmath13 representations .
@xmath95^{sl_m(\c ) } = hom_{sl_m(\c)}(\c , v(r_1\omega_1^ * ) \otimes \ldots \otimes v(r_n\omega_1^*))\\\ ] ] in this way , the plcker algebra encodes the branching problem of finding copies of the trivial representation of @xmath13 in an irreducible representation of @xmath96 for this reason , we will refer to the plcker algebra as a branching algebra . in general , a branching algebra encodes the branching rules of irreducible representations for some map of reductive groups . in this case
the map is the diagonal map @xmath97 filtrations and associated graded algebras of branching algebras like this one were studied by the author in @xcite , in particular for diagonal embeddings as above , the author described a way to produce filtrations of the branching algebra associated to labelled , rooted trees .
we will review the details of this construction in the next section , but for now we will describe the features that we need .
let @xmath20 be the cone of dominant weights with respect to the standard ordering of weight vectors in the weight lattice for @xmath1 let @xmath98 be a rooted tree with @xmath30 leaves , we consider the orientation induced on the edges of @xmath98 by orienting every edge in the unique path from the root to a leaf in such a way to make the root the unique source .
[ decomp ] let @xmath98 be a rooted tree with @xmath30 leaves .
there is a direct - sum decomposition , @xmath99^{sl_m(\c ) } = \bigoplus \mathcal{w}(\hat{\tree } , \lambda)\\\ ] ] over all @xmath100 such that the root edge is weighted @xmath101 the edge incident to the @xmath36th leaf is weighted @xmath102 and for each internal vertex , the representation associated to the label on the sink appears in the direct sum decomposition of the tensor product of the representations associated to the labels on the sources .
the summand @xmath103 is the vector space of all possible assignments of intertwiners to the internal vertices which realize the weight on a source at a vertex as a summand of the tensor product of the weights on sinks . in figure
[ fig:2 ] we an example of such an object with @xmath13 representations given by young tableaux . recall that @xmath104 proposition [ decomp ] is a formal consequence of properties of semisimple categories with monoidal products .
the tree @xmath98 can be considered as a recipe for inserting parentheses into the tensor product @xmath105 and gives a way to recursively expand the expression into a direct sum .
we can then take the same tree @xmath98 and assign to its edges @xmath106 functionals @xmath107 taking care that @xmath108 is positive on all positive roots .
we apply this functional @xmath109 to each summand , where an element in @xmath103 is given the weight @xmath110 this weights each plcker coordinate @xmath80 with a number dependent on @xmath111 and the tree @xmath112 and so gives a point in @xmath113 after reviewing the construction of this filtration and understanding it with respect to the multiplication operation in the plcker algebra , we will be able to conclude the following .
each @xmath109 defines a point in @xmath114 this will follow from general arguments on filtrations of branching algebras obtained from the associated representation theory , in particular we will give a general way to produce points on the tropical varieties of ideals defining these algebras .
the functionals @xmath109 have a good amount of flexibility , enough to show the following theorem .
there exists for any weighted tree @xmath65 a tree functional @xmath109 such that @xmath115 for all @xmath0 tuples @xmath116 in particular , @xmath0-dissimilarity vectors are points on the tropical grassmannian .
we thank the reviewer for several useful suggestions , including the example [ rev ] .
in this section we will review the construction of filtrations of branching algebras introduced in @xcite .
the basic object we will be working with is the algebra @xmath117^{u_g } = \bigoplus_{\lambda \in c_g } v(\lambda^*),$ ] where @xmath9 is a connected reductive group over @xmath3 , @xmath118 is a maximal unipotent subgroup , @xmath119 are dominant weights , @xmath120 is the irreducible representation with highest weight @xmath121 and @xmath122 is the monoid of dominant weights .
we choose highest weight vectors for each irreducible representation @xmath123 multiplication in @xmath124 is induced by cartan multiplication , see @xcite for an introduction to the algebra @xmath124 .
@xmath125 identify @xmath126 with the dual @xmath127 in the unique way that makes @xmath128 where @xmath129 sends @xmath130 to @xmath131 and @xmath132 is the lowest weight vector of @xmath133 under this identification , cartan multiplication is the dual of the map which sends @xmath134 to @xmath135 .
@xmath136 let @xmath137 be a map of connected reductive groups over @xmath138 we define the branching algebra @xmath139 of @xmath140 as follows .
@xmath141^h = \bigoplus_{(\alpha , \beta ) \in c_h\times c_g } [ v(\alpha^*)\otimes w(\beta^*)]^h\\\ ] ] here @xmath142 acts on @xmath124 through @xmath143 and @xmath140 maps @xmath144 to @xmath145 branching algebras are so - named because the dimension of their multigraded components give the branching multiplicities for irreducible representations of @xmath9 as representations of @xmath146 we will now rewrite the multiplication operation in @xmath139 with respect to the following identity .
@xmath147^h = hom_h(\c , v(\alpha^*)\otimes w(\beta^ * ) ) \cong hom_h(v(\alpha ) , w(\beta^*))\\\ ] ] the isomorphism on the right is given by the following construction , for @xmath148 .
@xmath149 let @xmath150 denote the transformed map . under this isomorphism ,
the multiplication map @xmath151 \otimes [ v(\alpha_2^*)\otimes w(\beta_2^ * ) ] @>c _ * \otimes c _ * > > v(\alpha_1^ * + \alpha_2^ * ) \otimes w(\beta_1^ * + \beta_2^*)\\ \end{cd}\ ] ] becomes @xmath152 this is a straightforward calculation .
now we consider a factorization of @xmath140 in the category of connected reductive groups over @xmath153 @xmath154 we formally get a direct sum decomposition of each multigraded component of the branching algebra @xmath155 @xmath156 this introduces a host of combinatorial representation theory data into the algebra @xmath155 we will see how to multiply two elements , we start by taking the tensor product . @xmath157
the middle representation decomposes as a direct sum of @xmath158 representations , @xmath159 this allows us to represent @xmath160 and @xmath161 as sums of maps .
let @xmath162 and @xmath163 be projections and injections that define the direct sum decomposition with @xmath164 and @xmath165 then we have @xmath166 @xmath167 decomposing the diagram along these sums gives an expansion of the product into components from the direct sum decomposition of @xmath168 , and there is a natural leading term given by the sum of the weights , @xmath169 a general term , @xmath170 is a member of the @xmath171 summand .
the leading term never vanishes , because the defining maps @xmath172 and @xmath173 are the same as the multiplication operation in @xmath174 and @xmath175 respectively .
these algebras are domains because @xmath124 is always a domain .
notice that this analysis depends only on multigraded summands of @xmath176 so the same term decomposition exists for any subalgebra which preserves the multigrading .
we summarize the previous discussion .
[ highest ] for any factorization of a map of connected , reductive groups over @xmath138 @xmath177 there is a direct sum decomposition of @xmath139 into summands @xmath178 with @xmath179 @xmath180 and @xmath181 dominant weights .
this defines a multifiltration of the branching algebra @xmath182 the product of two elements @xmath183 has leading term @xmath184 all lower terms involve @xmath180 which are less than @xmath185 as dominant weights .
we can perform this same construction on a factorization of any length @xmath186 without altering the details , and proposition [ highest ] holds for the resulting multifiltration .
we may use this extra combinatorial data to describe filtrations of @xmath155 to each new summand @xmath187 in the filtration , we attach a number as follows , pick functionals @xmath188 such that @xmath189 has non - negative value on all positive roots of @xmath190 . now apply these functionals to the weights defining the multifiltered summands
, this defines a filtration .
@xmath191 by proposition [ highest ] , the value on a product of elements , computed by summing up the contributions from each element , is always equal to the value on its leading term for any linear functional @xmath192 [ pres ] let @xmath193 \to \mathfrak{a}(\phi)$ ] be a presentation of the branching algebra , and let @xmath194 be a factorization of @xmath195 suppose each @xmath196 is mapped to an element of one of the summands @xmath197 defined by the factorization , and let @xmath198 $ ] be the defining ideal
. then any functional @xmath199 defines a term weighting of @xmath200 which gives a monomial free initial ideal @xmath201 .
pick any expression in the ideal @xmath202 .
@xmath203 we consider the expansion of each monomial term into pure terms , @xmath204 where @xmath205 has the same pure filtration level as the monomial , the existence of this term follows from proposition [ highest ] , which also implies that we must have @xmath206 for every term in this expansion . in general , for pure terms @xmath200 and @xmath207 , we say that @xmath208 if for each component @xmath209 is a positive root . note that not all pure terms are comparable . by definition of the functional @xmath199 if @xmath210 then @xmath211 now suppose some monomial @xmath212 in the expression @xmath213 has the highest filtration weight with respect to @xmath192 we must have @xmath214 so @xmath215 must be canceled by pure terms from the expansion of other monomials .
this implies that some monomial @xmath216 must have a pure term @xmath217 with the same multifiltration level as @xmath218 we must have that @xmath219 as pure terms , by assumption this implies that @xmath216 has the same filtration weight as @xmath212 .
this proposition implies that every @xmath199 defines a point on the tropical variety of the defining ideal @xmath220 it also implies that for any presentation @xmath221 \to \mathfrak{a}(\phi),$ ] and any form in the defining ideal @xmath222 the leading terms of at least two monomials agree , a result independent of a functional @xmath192 the functionals @xmath199 fit into the broader theory of valuations on rings .
roughly these are functions @xmath223 on a ring which satisfy @xmath224 @xmath225 and @xmath226 is @xmath227 or @xmath228 depending on the tropical algebra where @xmath223 takes its values .
generally speaking valuations define `` universal '' tropical points , in that they define a point on the tropical variety of any presentation of a subring of the ring on which they are defined .
we explore these objects in the note @xcite , see also @xcite . for each factorization of @xmath137
@xmath229 @xmath230 we obtain a cone of functionals @xmath231 defined by the conditions on the components of @xmath192 note that the @xmath232 is always an option , indeed this essentially forgets the information in @xmath36-th component of the multifiltration . for each factorization @xmath233 and every @xmath234
there is an operation @xmath235 setting @xmath189 to @xmath227 gives a map of cones @xmath236 which defines @xmath237 as a face of @xmath238 this defines a connected complex of cones @xmath239 over all factorizations of @xmath140 in the category of connected , reductive groups .
the content of the proposition above is that there is a map from this complex into the tropical variety of any presentation of @xmath176 the same holds for any subalgebra of @xmath240 which preserves the multigrading .
in particular , this is true for the subalgebra of invariants , which will be important in the sequel .
@xmath241^h \subset [ r(h)\otimes r(g)]^h = \mathfrak{a}(\phi),\\\ ] ] [ rev ] we can also look at branching deformations for the trivial subgroup of a reductive group @xmath242 this morphism is factored by any flag of subgroups of @xmath15 for instance we can take @xmath243 and look at the flag @xmath244 the branching of @xmath245 over @xmath246 is multiplicity free , so the branching algebra associated to this pair is toric . choosing a functional @xmath247 which is positive on positive roots
then defines a toric deformation of @xmath248 to the monoid of gelfand tsetlin patterns .
any representation @xmath249 of a reductive group @xmath9 defines a morphism @xmath250 for @xmath251 first we note that if @xmath249 is reducible , then the map factors through @xmath252 for some partition of @xmath253 also , the map defined by @xmath249 always defines a factorization of the trivial morphism @xmath254 and can therefore be identified with a cone of filtrations on @xmath255 one can use this technique to define degenerations of a wide range of varieties with symmetry .
let @xmath84 be a commutative ring with the action of a product of reductive groups @xmath256 and @xmath257 be a map of connected reductive groups over @xmath3 .
there is a flat degeneration of @xmath84 defined in @xcite which preserves the action of @xmath256 to the algebra @xmath258^{t_h \times t_g}$ ] , where @xmath259 is a maximal torus of @xmath260 taking @xmath142 invariants for the action of @xmath261 gives a flat degeneration @xmath262^{t_g},$ ] this can then be composed with degenerations of the branching algebra .
this technique was adapted by the author in @xcite to study properties of a quantum analogue of a branching algebra coming from conformal field theory .
a similar sort of universality holds for other types of degenerations defined from the combinatorics of representation theory , for instance toric degenerations of spherical varieties , see @xcite for details . in @xcite
the author also studied the associated graded algebra of a branching filtration . for a factorization , @xmath263 if the functional @xmath111 is strictly positive on the positive roots of @xmath264 then we get a flat deformation over @xmath265,$ ] @xmath266^{t_k}\\\ ] ]
in this section we use the results from the previous section to study @xmath267^{sl_m(\c ) } \subset \mathfrak{a}(\delta_n),$ ] where @xmath268 is the diagonal embedding .
these embeddings have a special class of filtrations classified by rooted trees with @xmath30 leaves .
take such a tree @xmath98 and define a factorization of @xmath269 as follows , let @xmath98 have the orientation induced by the root , as before . for each internal vertex @xmath270 attach the diagonal morphism @xmath271 from one copy of @xmath13 to @xmath272 by well - ordering the non - leaf vertices of @xmath98 in any way such that the first vertex is attached to the root , and two consecutive vertices share an edge
allows us to write this factorization in the style of the previous section .
@xmath273 this results in a direct sum decomposition of @xmath274 into spaces @xmath103 indexed by assignments of dominant weights of @xmath13 to the edges of @xmath98 , along with an assignment of @xmath275linear maps at every vertex intertwining the corresponding tensor products of irreducible representations . from the introduction we know that the plcker algebra is the subalgebra of @xmath274 generated by the unique invariants @xmath276^{sl_m(\c)}$ ] where @xmath0 of the @xmath277 pieces of the tensor product are copies of @xmath278 the first piece , corresponding to the root , is always @xmath138 and the other @xmath279 pieces are the trivial representation @xmath153 each of these spaces is one dimensional , so we should be able to write down the tree diagram of a basis member for a chosen @xmath280 to describe the diagram in general it is simplest to start with a rooted tree @xmath281 with @xmath0 leaves , give this tree an orientation as above .
each leaf of this tree is labeled with @xmath282 and to compute the representation labeling a given edge @xmath283 simply count the number of leaves @xmath284 above @xmath285 with respect to the rooted orientation , and give it the label @xmath286 now dualize the whole picture , so @xmath24 becomes @xmath287 .
the result is shown below in young tableaux .
the root is labeled with @xmath288 which is trivial as an @xmath13 representation . in a general rooted tree
@xmath112 take the convex hull of the root and the non - trivially labeled leaves .
combinatorially , this is the same as some rooted tree with @xmath0-leaves @xmath289 label the edges of @xmath98 accordingly , and label all other edges with the trivial representation .
note that up to scalars the available intertwiners @xmath290 in this diagram are all unique as expected .
the subalgebra @xmath291^{sl_m(\c ) } \subset \mathfrak{a}(\delta_n)$ ] is generated by the @xmath292 elements of this type .
all diagrams are given explicitly in terms of the @xmath22 fundamental weights of @xmath19 and one easily checks that an edge @xmath293 is labeled nontrivially if and only if it is in the combinatorial convex hull of the @xmath0-nontrivially labeled leaves .
now consider the functional @xmath294 defined by @xmath295 for all fundamental weights @xmath296 and note that this functional gives the trivial representation the @xmath227 weight .
pick a non - negative length @xmath297 for edge @xmath298 and consider the functional defined by assigning @xmath299 to the edge @xmath300 for the plcker coordinate @xmath80 we have @xmath301 for any metric tree @xmath65 with @xmath30 leaves there is a rooted tree with @xmath30 leaves @xmath98 , and a functional @xmath111 with @xmath302 to get @xmath112 one may add a root to @xmath29 anywhere .
it is simple to verify that this preserves combinatorial convex hulls .
see figure [ fig:6 ] for an example .
the root is added in the middle of an edge of @xmath2 so in order to preserve the weighting information we must split the weight on this edge among the two new edges created by the addition of the root .
the previous discussion does the rest .
this proposition establishes that we can replicate the @xmath0-dissimilarity vectors of a metric tree @xmath65 with branching filtrations .
the efforts of the previous section confirm that branching filtrations always give tropical points .
together , these facts prove theorem [ main ] .
in this section we will look at dissimilarity maps , plcker coordinates , , and tree weighting functionals in more detail for a specific example .
we will take a look at some elements and relations in the plcker algebra @xmath303.$ ] we choose a rooted tree with @xmath304 leaves , @xmath29 . for simplicity
we give @xmath29 the metric where each edge has length @xmath305 note that the corresponding unrooted tree would have one edge with length @xmath306 and all others with length @xmath307 we will find how @xmath29 weights the plcker relation next we look at the general case of @xmath312 the @xmath44-dissimilarity vectors of a tree are the best understood dissimilarity vectors because of their association with the grassmannian @xmath313 the same is true for @xmath314 branching algebras .
the algebra @xmath274 for @xmath315 is isomorphic to @xmath316^{sl_2(\c)},$ ] indeed we have it follows that the subalgebra of invariants @xmath318 is isomorphic to the @xmath319 plcker algebra . for a rooted tree @xmath112 the functionals @xmath109
are all given by assigning non - negative integers to the edges of @xmath112 as non - negative integers correspond to maps @xmath320 therefore for any metric tree @xmath65 we can construct a branching algebra filtration that weights the plcker monomials the same as @xmath45 in this way , every member of the tropical grassmannian @xmath321 is realizable by a branching filtration . | we prove that @xmath0-dissimilarity vectors of weighted trees are points on the tropical grassmannian , as conjecture by cools in response to a question of sturmfels and pachter .
we accomplish this by relating @xmath0-dissimilarity vectors to the representation theory of @xmath1 |
alfvn waves play important roles in strongly magnetized media .
they propagate along magnetic field lines with the alfvn speed @xmath2 , where @xmath3 is the strength of the mean magnetic field and @xmath4 is density .
alfvn waves moving in opposite directions can interact and result in alfvnic magnetohydrodynamic ( mhd ) turbulence .
alfvnic mhd turbulence in the non - relativistic limit has been studied for many decades and the best available mhd turbulence model is , in spite of all existing controversies ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the one by goldreich & srdihar ( 1995 ; henceforth gs95 ) which was first numerically tested by cho & vishniac ( 2000 ) .
the gs95 model predicts a kolmogorov spectrum ( @xmath5 ) and scale - dependent anisotropy ( @xmath6 ) , where @xmath7 and @xmath8 are wave - numbers along and perpendicular to the local mean magnetic field directions , respectively , and @xmath9 .
when @xmath3 goes to infinity and/or @xmath4 goes to zero , alfvn speed approaches the speed of light and a new regime of turbulence emerges .
more precisely , when the magnetic energy density is so large that the inertia of the charge carriers can be neglected , the medium can be described by relativistic force - free mhd equations @xcite .
cho ( 2005 ) numerically studied three - dimensional mhd turbulence in this extreme relativistic limit and found the following results .
first , the energy spectrum is consistent with a kolmogorov spectrum : @xmath5 .
second , turbulence shows the goldreich - sridhar type anisotropy : @xmath6 .
these scaling relations are in agreement with earlier theoretical predictions by thompson & blaes ( 1998 ) .
the similarity between non - relativistic alfvnic mhd turbulence and relativistic force - free mhd turbulence leads us to the question : to what extent are relativistic and non - relativistic alfvnic turbulence similar ? in this paper , we try to answer this question .
strong imbalanced alfvnic turbulence is an ideal problem for that purpose because interactions between eddies are very complicated in strong imbalanced alfvnic turbulence . in imbalanced alfvnic turbulence ,
the waves traveling in one direction ( dominant waves ) have higher amplitudes than the opposite - traveling waves ( sub - dominant waves ) . by ` strong ' imbalanced turbulence ,
we mean the dominant waves satisfy the condition of critical balance , @xmath10 , at the energy injection scale , where @xmath11 is the strength of the fluctuating magnetic field .
many studies exist for strong imbalanced alfvnic turbulence in the non - relativistic limit @xcite , but no study is available yet for its relativistic counterpart . in this paper , we compare our relativistic simulations with non - relativistic ones .
our study can have many astrophysical implications .
so far , we do not fully understand turbulence processes in extremely relativistic environments , such as black hole / pulsar magnetospheres , or gamma - ray bursts .
if we can verify close similarities between extremely relativistic and newtonian alfvnic turbulence , we can better understand physical processes , e.g. reconnection , particle acceleration , etc .
, in such media .
we describe the numerical methods in section 2 and we present our results in section 3 . we give discussions and summary in section 4 .
we solve the following system of equations in a periodic box of size @xmath12 : @xmath13 where @xmath14 here @xmath15 is the electric field , @xmath16 the poynting flux vector , and we use units such that the speed of light and @xmath17 do not appear in the equations ( see @xcite for details ) .
one can derive this system of equations from @xmath18 where @xmath19 is the dual tensor of the electromagnetic field , @xmath20 the fluid four velocity , and @xmath21 the stress - energy tensor of the electromagnetic field @xmath22 where @xmath23 is the metric tensor and @xmath24 is the electromagnetic field tensor .
we ignore the stress - energy tensor of matter .
we use flat geometry and greek indices run from 1 to 4 .
one can obtain the force - free condition from maxwell s equations and the energy - momentum equation @xmath25 . from equation ( [ eq_cond ] ) , one can derive @xmath26 in our simulations , the mhd condition @xmath27 is enforced all the time .
we solve equations ( 1)-(6 ) using a monotone upstream - centered schemes for conservation laws ( muscl ) type scheme with hll fluxes ( harten , lax , van leer 1983 ; in fact , in force - free mhd these fluxes reduce to lax - friedrichs fluxes ) and monotonized central limiter ( see kurganov et al .
the overall scheme is second - order accurate . after updating the system of equations along the @xmath28 direction , we repeat similar procedures for the @xmath29 and @xmath30 directions with appropriate rotation of indexes .
gammie , mckinney , & tth ( 2003 ) used a similar scheme for general relativistic mhd and del zanna , bucciantini , & londrillo ( 2003 ) used a similar scheme to construct a higher - order scheme for special relativistic mhd .
while the magnetic field consists of the uniform background field and a fluctuating field , @xmath31 , the electric field has only a fluctuating one .
the strength of the uniform background field , @xmath3 , is set to 1 . at @xmath32 ,
no fluctuating fields are present .
we isotropically drive turbulence in the wave - number range @xmath33 .
we adjust the amplitude of forcing to maintain @xmath34 after saturation , where the subscript ` + ' denotes dominant waves .
therefore , we have @xmath35 after saturation . since the energy injection rates for the sub - dominant waves ( @xmath36 ) are equal to or less than those of dominant waves ( @xmath37 ) , where @xmath38 s are forcing vectors , we have @xmath39 and @xmath40 .
simulation parameters are listed in table 1 . to check the stability of our code
, we perform a simulation of relativistic alfvn waves moving in the same direction .
since alfvn waves moving in one direction do not interact each other , their energy spectrum should not change in time .
indeed figure [ fig:1](a ) confirms this : the initial energy spectrum ( the thick solid line ) does not show much change even after t@xmath4163 , which corresponds to @xmath4110 wave crossing times over the box size .
lcl 256-bal & 256@xmath42 & 1 + 256-r0.75 & 256@xmath42 & 0.75 + 256-r0.5 & 256@xmath42 & 0.5 + 256-r0.33 & 256@xmath42 & 0.33 + 512-r0.33 & 512@xmath42 & 0.33
figure [ fig:1](b ) shows time evolution of the energy densities of the dominant waves .
we have @xmath43 at @xmath32 and we drive the medium for @xmath44 . the energy densities of the dominant waves initially rise quickly and reach saturation states .
the values of @xmath45 during saturation in those runs are between 0.5 and 1.0 .
since we drive turbulence isotropically , critical balance is roughly satisfied . in general , the larger the imbalance , the slower the approach to the saturation state .
the largest imbalanced run ( run 256-r0.33 ) shows very slow approach to the saturation state .
figure [ fig:1](c ) shows time evolution of energy densities of the sub - dominant waves . from top to bottom , the degree of imbalance increases .
the top curve corresponds to the balanced turbulence ( run 256-bal ) and the bottom curve to the largest imbalance ( run 256-r0.33 ) . note that , in run 256-r0.33 , @xmath46 goes up very quickly for @xmath47 and then gradually goes down , which may be due to the increase of @xmath48 .
figures [ fig:1](d ) and ( e ) show time evolution of the ratio @xmath49 and @xmath50 , respectively .
figure [ fig:1](d ) clearly shows that the value of @xmath49 increases substantially as the degree of imbalance increases . for @xmath50 , we actually plot @xmath51 , where @xmath52 .
since different theories on imbalanced non - relativistic alfvnic turbulence predict different relations between @xmath49 and @xmath50 , it will be useful to plot the relation for our simulations .
figure [ fig:1](f ) shows the relation between the two ratios . roughly speaking
, the ratio @xmath49 exhibits a power - law dependence on the ratio @xmath53 : @xmath0 with @xmath1 . in figure [ fig:1 ] ,
all simulations are performed on a grid of @xmath54 points .
the top panel of figure [ fig:2 ] , which compares results of runs 512-r0.33 and 512-r0.33 , implies that numerical resolution of @xmath54 would be enough for our current study .
note that two runs have identical numerical set - ups except the numerical resolution ( @xmath54 versus @xmath55 ) .
the values of @xmath45 ( upper curves ) almost coincide , but the value of @xmath46 for @xmath55 is slightly higher than that for @xmath54 ( see lower curves ) .
the bottom panel of figure [ fig:2 ] shows energy spectra for run 512-r0.33 .
although we have only about 1 decade of inertial range , we can clearly observe that the spectral slopes for dominant and sub - dominant waves are different .
the spectrum of the dominant waves ( upper curve ) is slightly steeper than @xmath56 , while that of the sub - dominant ones ( lower curve ) is a bit shallower than @xmath56 . in the presence of a strong mean magnetic field , structure of turbulence
tends to elongate along the direction of the mean field .
therefore elongation of structures , or anisotropy , is an important aspect of mhd turbulence . both relativistic force - free and non - relativistic balanced alfvnic turbulence
. imbalanced non - relativistic alfvnic turbulence is also anisotropic ( e.g. , @xcite ) .
since interactions between eddies are very complicated in imbalanced alfvnic turbulence , it will be interesting to study anisotropy of imbalanced relativistic force - free mhd turbulence .
figure [ fig:3 ] shows the shapes of eddies . in the figure , we plot a contour diagram of the second - order structure function for the magnetic field in a local frame , which is aligned with the local mean magnetic field @xmath57 : @xmath58 where @xmath59 and @xmath60 and @xmath61 are unit vectors parallel and perpendicular to the local mean field @xmath57 , respectively ; see cho & vishniac ( 2000 ) and cho et al .
( 2002 ) for the detailed discussion of the local frame .
the left and middle panels of figure [ fig:3 ] show shapes of dominant and sub - dominant eddies , respectively .
we can clearly see that the dominant eddies ( left panel ) are less anisotropic than the sub - dominant ones ( middle panel ) .
if we plot the relation between perpendicular sizes of eddies ( or , y intercepts of the contours ; @xmath62 ) and the parallel ones ( or , x intercepts ; @xmath63 ) , than we can see that the dominant eddies show anisotropy weaker than @xmath64 and the sub - dominant ones show anisotropy stronger than @xmath64 .
our simulations are consistent with the theory and simulations of the imbalanced non - relativistic mhd turbulence ( beresnyak & lazarian 2008 , 2009 ) .
indeed , the latter results are consistent with our finding of the relation between the ratio of the energy densities of the sub - dominant and dominant waves , their spectral slopes and their anisotropy .
this is suggestive of a close relation between the non - relativistic and relativistic turbulence and implies that the existing theories of non - relativistic turbulence , e.g. theories for magnetic reconnection , particle acceleration , etc .
, can be generalized for the relativistic limit .
this has not yet been done and , naturally , more theoretical / numerical research , especially with high numerical resolutions , for the relativistic case is necessary .
imbalanced turbulence is a generic incarnation of turbulence in the presence of sources and sinks of turbulent energy .
we know from the studies of non - relativistic imbalanced turbulence that its slower decay compared to the balanced one allows the energy transfer over larger distances and its transfer to the balanced one due to parametric instabilities or the reflection of waves from density inhomogeneities can result in local deposition of energy and momentum which provide many astrophysically important consequences .
the properties of imbalanced relativistic turbulence are important for many astrophysical settings including the magnetosphere of pulsars , environments of gamma ray bursts and relativistic jets . 1 .
the magnetic spectrum of dominant waves is steeper than that of sub - dominant waves . 2 .
the dominant waves show anisotropy weaker than and the sub - dominant waves show anisotropy stronger than @xmath64 .
the energy density ratio @xmath49 is roughly proportional to @xmath65 , where @xmath66 s are energy injection rates and @xmath1 .
all these results are consistent with the theory and simulations in beresnyak & lazarian @xcite
. therefore we can conclude that relativistic force - free mhd turbulence is indeed very similar to its non - relativistic counterpart .
our results imply that many results in non - relativistic alfvnic turbulence can be carried over to relativistic force - free mhd turbulence .
for example , theories on magnetic reconnection ( e.g. , @xcite ) , particle acceleration ( e.g. , @xcite ) and thermal diffusion ( e.g. , cho et al .
2003 ) obtained in non - relativistic alfvnic turbulence can also be applicable to relativistic force - free mhd turbulence .
the close similarity between the properties of non - relativistic and relativistic imbalanced turbulence found in this paper elucidates the nature of magnetic turbulence that preserves its properties in both regimes irrespectively of whether turbulence is balanced or imbalanced . from the practical point of numerical studies ,
this allows us to test or double - check theories on non - relativistic alfvnic turbulence using a completely different numerical scheme .
j.c.s work is supported by the national r & d program through the national research foundation of korea ( nrf ) , funded by the ministry of education ( no . 2011 - 0012081 ) .
a.l . is supported by nsf grant ast 1212096 , the center for magnetic self - organization and the vilas associate award .
we thank the international institute of physics ( natal ) for their hospitality .
99 beresnyak , a. 2011 , phys .
lett . , 106 , 075001 beresnyak , a. , & lazarian , a. 2006 , apjl , 640 , l175 beresnyak , a. , & lazarian , a. 2008 , apj , 682 , 1070 beresnyak , a. , & lazarian , a. 2009 , apj , 702 , 1190 | when magnetic energy density is much larger than that of matter , as in pulsar / black hole magnetospheres , the medium becomes force - free and we need relativity to describe it . as in non -
relativistic magnetohydrodynamics ( mhd ) , alfvnic mhd turbulence in the relativistic limit can be described by interactions of counter - traveling wave packets . in this paper
we numerically study strong imbalanced mhd turbulence in such environments . here
, imbalanced turbulence means the waves traveling in one direction ( dominant waves ) have higher amplitudes than the opposite - traveling waves ( sub - dominant waves ) .
we find that ( 1 ) spectrum of the dominant waves is steeper than that of sub - dominant waves , ( 2 ) the anisotropy of the dominant waves is weaker than that of sub - dominant waves , and ( 3 ) the dependence of the ratio of magnetic energy densities of dominant and sub - dominant waves on the ratio of energy injection rates is steeper than quadratic ( i.e. , @xmath0 with @xmath1 ) .
these results are consistent with those obtained for imbalanced non - relativistic alfvnic turbulence .
this corresponds well to the earlier reported similarity of the relativistic and non - relativistic balanced magnetic turbulence . |
the decay @xmath16 ( @xmath17 ) provides unique information about the dynamics of the strong interactions .
it has been a testing ground for such theories as current algebra , pcac , chiral perturbation theory(chpt ) .
the study of this decay has a particular interest in view of new two - loop order ( @xmath18 ) calculations for k@xmath19 decays in chpt @xcite .
the high - order chpt calculations make a definite prediction for the quadratic term in the vector @xmath20 form - factor and link the scalar @xmath21 form - factor linear and quadratic slopes to the @xmath22 corrections . in turn
, @xmath22 is known to be crucial for the @xmath23 measurements .
in fact , the latest measurements do not report any visible non - linearity in the form - factors @xcite .
in this paper we present a high - statistics measurement ( @xmath24 919k events ) of the dalitz plot density in this decay . the description of the experimental setup , trigger and reconstruction procedure can be found in our previous paper @xcite .
the current analysis is based on the high - statistics data collected during run in winter 2001 . in total ,
332 m events were logged on tapes .
this statistics is complemented by about 160 m mc events generated with geant3 @xcite monte carlo program .
the mc generation includes a realistic description of the setup with decay volume entrance windows , tracking chambers windows , chambers gas mixtures , sense wires and cathode structures , erenkov counters mirrors and gas , the shower generation in em calorimeters , etc .
the events with one charged track identified as electron and two or three additional showers in the electromagnetic calorimeter are selected for further processing .
it was observed , that the main background contribution ( about 95% of the background events ) is related to the decay @xmath25 , when the hadronic interaction of the charged pion simulates the electromagnetic shower in the calorimeter . following the method of angular selection used in our analysis of the @xmath26 decay @xcite
, we choose the angle between the beam particle direction and the vector sum of momenta of the final state track and photons as the variable to perform signal - background separation . the expected distribution over this angle is shown in fig.1 together with the real data .
one can observe a clear background peak at small values and a good agreement of the data and monte - carlo .
the cut value was selected in the region where the maximum of the function @xmath27 is reached .
after that , the momenta of the final state particles were refined with 2c @xmath28 kinematic fit . only the convergence of the fit
was required .
the missing energy @xmath29 after 2c fit is shown in fig.2 .
this variable should exhibit a strong peak at @xmath30 in the presence of noticeable background contributions .
the complete absence of any enhancement at small values of the missing energy proofs the quality of the selection procedure .
figure 1 : angle between the beam track and the vector sum of final state particles momenta .
the points with errors are data and the solid histogram is mc .
the shaded area shows the background contribution .
figure 2 : energy of the neutrino compared with mc .
the shaded area shows the surviving background contribution .
the signal monte - carlo events for figures 1 and 2 are weighted with the @xmath31 matrix element where we use @xmath32 ( fixed from our preliminary measurements @xcite ) . finally , 919k events were selected .
we estimate the surviving background contribution to be around 2.1% .
the most general lorentz - invariant form of the matrix element for the @xmath33 decay is @xcite : @xmath34\gamma^{\alpha } + i \frac{2f_{t}}{m_{k } } \sigma_{\alpha \beta}p^{\alpha}_{k}p^{\beta}_{\pi}]v(p_{l})\ ] ] it consists of scalar , vector , and tensor terms .
the @xmath35 form - factors are the functions of @xmath36 . in the standard model ( sm )
, the w - boson exchange leads to the pure vector term .
the scalar and/or tensor terms which are `` induced '' by ew radiative corrections are negligibly small , i.e nonzero scalar or tensor form - factors would indicate the physics beyond the sm .
the term in the vector part , proportional to @xmath37 , is reduced ( using the dirac equation ) to the scalar form - factor . in the same way
, the tensor term is reduced to a mixture of the scalar and vector form - factors .
the redefined vector ( v ) and scalar ( s ) terms , and the corresponding dalitz plot density in the kaon rest frame ( @xmath38 ) are @xcite : @xmath39 with the selected number of events we can not neglect the @xmath40 interference term proportional to the electron mass .
the term proportional to @xmath41 is neglected in our analysis .
for further analysis we assume a general quadratic dependence of @xmath42 on t : @xmath43 the procedure of the extraction of the form - factor parameters starts with the subdivision of the dalitz plot region @xmath44 @xmath45 into @xmath46 bins .
the signal mc was generated with the constant matrix element and the amplitude - induced weights should be calculated during the fit procedure .
one can observe that the dalitz - plot density function @xmath47 in ( 2 ) can be presented in the factorisable form : @xmath48 where @xmath49 are simple bilinear functions of the form - factor parameters and @xmath50 are the kinematic functions which are calculated from the mc - truth information . for each @xmath51 ,
the sums of @xmath50 over events are accumulated in the dalitz plot bins ( i , j ) to which the mc events fall after the reconstruction .
finally , every bin in the dalitz plot gets weights @xmath52 and the density function @xmath53 which enters into the fitting procedure is constructed : @xmath54 this method allows one to avoid the systematic errors due to the `` migration '' of the events over the dalitz plot due to the finite experimental resolution and automatically takes into account the efficiency of the reconstruction and selection procedures . to take into account the finite number of mc events in the particular bin and strong variation of the real data events over the dalitz plot , we minimize a @xmath55 function defined as @xcite : @xmath56 + 2\sum_j ( n_j+m_j+1)\ln\left ( \frac{1+\frac{r_j}{m_j}}{1+\frac{n_j}{m_j+1}}\right),\ ] ] where the sum runs over all populated bins , and @xmath57 , @xmath58 and @xmath59 are the number of data events , expected events and generated monte carlo events respectively . for large @xmath59 eq .
( 8) reduces to the more familiar expression @xmath60\ ] ] the radiative corrections were taken into account by re - weighting every monte - carlo event , using mc - truth information , according to the recent calculations in @xcite .
the minimization is performed by means of the `` minuit '' program @xcite .
the errors are calculated by `` minos '' procedure of `` minuit '' at the level @xmath61 , corresponding to 68% coverage probability for 1 parameter .
a fit of the @xmath17 data with @xmath62 gives @xmath63 .
the total number of bins is 6991 and @xmath64 .
the quality of the fit is illustrated in figures 3 and 4 where the projected variables @xmath65 and @xmath66 are presented .
figure 3 : y distribution .
+ the points with errors are the real data + and the shaded area signal mc .
figure 4 : z distribution .
+ the points with errors are the real data + and the shaded area
signal mc .
the table 1 represents fits with possible nonlinear term in @xmath67 ( eq .
3 ) as well as the fits with tensor and scalar contributions ( eq .
every row of the table 1 represents a particular fit where the parameters shown without errors are fixed . to qualify the statistical significance of the particular fit
, we also show the change in the @xmath68 value obtained with respect to the fit without non - linear or anomalous contributions .
the second row shows a fit where the nonlinearity is allowed in @xmath2
. one can observe @xmath69 ( fig.5 ) correlation that results in a significant @xmath1 errors enhancement and visible shift of @xmath1 parameter . [
cols="^,^,^,^,^",options="header " , ] table 2 . the systematic error contributions .
the @xmath70 decay has been studied using in - flight decays of 25 gev @xmath71 detected by the `` istra+ '' magnetic spectrometer .
the leptoquark - induced amplitudes for @xmath17 decay were considered in @xcite . in this model the relation between tensor and scalar terms is fixed : @xmath77 .
the fit of our data with this constraint gives the following scalar contribution : @xmath78 .
this result can be converted in the upper limit @xmath79 ( 90% c.l . ) . using the expression in @xcite : @xmath80 where @xmath81 is the ratio of leptoquark mass to the square of the yukawa - like coupling , we get @xmath82 99 p.post and k.schilcher , eur . phys .
j. * 25*(2002 ) , 427 .
j.bijnens and p.talavera , nucl.phys . *
b669*(2003 ) , 341 .
s. shimizu et al .
( e246 kek - ps collaboration ) , phys .
lett . * b495*(2000 ) , 33 .
levchenko et al .
( e246 kek - ps collaboration ) , yad.fiz * v65*(2002 ) , 2294 , i.v.ajinenko et al .
( istra+ collaboration ) , phys.lett . *
b574*(2003 ) , 14 .
o.p.yushchenko et al .
( istra+ collaboration ) , phys.lett . *
b581*(2004 ) , 31 .
r brun et al . ,
cern - dd / ee/84 - 1 .
h. steiner et al . , phys.lett . *
b36*(1971 ) , 521 .
chizhov , phys.lett . *
b381*(1996 ) , 359 .
l. rosselet et al . , phys.rev . *
d15*(1977 ) , 574 .
v. cirigliano , m. knecht , h. neufeld , h. rupertsberger , p. talavera , eur .
j. * c23*(2002 ) , 121 .
f. james , m.roos , cern d506,1989 .
kiselev , a.k .
likhoded and v.f .
obraztsov , hep - ph/0204066 ( 2002 ) . | the decay @xmath0 is studied using in - flight decays detected with the `` istra+ '' spectrometer .
about 920k events are collected for the analysis .
the @xmath1 slope parameter of the decay form - factor @xmath2 in the linear approximation ( average slope ) is measured : @xmath3(stat ) @xmath4(syst ) . the quadratic contribution to the form - factor
was estimated to be @xmath5(stat ) @xmath6(syst ) .
the linear slope , which has a meaning of @xmath7 for this fit , is @xmath8(stat ) @xmath4(syst ) .
the limits on possible tensor and scalar couplings are derived : @xmath9(stat ) @xmath10(syst ) , @xmath11(stat ) @xmath12(syst ) . by -by -21.9
cm -10pt 0.5 in 17.0 cm * high statistic measurement of the @xmath13 decay form - factors * + * o.p .
yushchenko , s.a .
akimenko , g.i .
britvich , k.v.datsko , a.p .
filin , a.v .
inyakin , a.s .
konstantinov , v.f .
konstantinov , i.y .
korolkov , v.a .
khmelnikov , v.m .
leontiev , v.p .
novikov , v.f .
obraztsov , v.a .
polyakov , v.i .
romanovsky , v.m .
ronjin , v.i .
shelikhov , n.e .
smirnov , o.g .
tchikilev , v.a.uvarov .
* 0.15 cm * @xmath14 * 0.35 cm * v.n .
bolotov , s.v .
laptev , a.yu . polyarush . * 0.15 cm * @xmath15 * |
in non - relativistic mechanics , the trajectory of a particle is most commonly expressed as a function @xmath0 , which describes how the position of the particle evolves with time . in relativistic mechanics , where space and time are treated in a symmetric way
, the particle s trajectory is regarded as a sequence of spacetime events @xmath1 . in field theory , the field configuration is usually regarded as a function @xmath2 , which describes how the values of the fields vary from point to point in the spacetime
. however , the general relativity suggests @xcite that the spacetime is a dynamical entity , and as such , it should be put with the fields on the same footing .
mathematically , instead of a function @xmath2 , one is therefore motivated to consider the respective graph , i.e. , the collection of points @xmath3 . in this article
, we develop the mathematical formalism for field theories proposed in ( * ? ? ?
3 ) that treats time , space , and fields equally .
all these entities are collectively called _ partial observables _ , and they form a finite - dimensional _
configuration space_. classical field theory studies correlations between the partial observables ( called _ motions _ ) , which have the form of surfaces embedded in the configuration space , and selects the _ physical ( or classical ) motions _
, i.e. , the ones that can be realized in nature .
our dynamical description utilizes a multivector - valued momentum variable , which can be thought of as conjugated to the motion s tangent planes , thus generalizing the canonical momentum conjugated to the velocity vector in classical mechanics .
individual field theories are specified by a choice of the _ hamiltonian _ @xmath4 , which is a function of the configuration space points @xmath5 , and the momentum @xmath6 .
this hamiltonian enters into a variational principle ( sec .
[ sec : varprinc ] ) via the so - called _ hamiltonian constraint _ @xmath7 .
the aim of this article is to establish the hamiltonian constraint formalism for the field theories as a viable , and even superior , alternative to the usual lagrangian formalism .
first , in sec . [ sec : caneq ] , we determine the canonical equations of motion , eqs .
( [ caneom ] ) , that follow from the variational principle .
these equations generalize the hamilton s canonical equations of motion of classical mechanics . in sec .
[ sec : lochj ] , we derive from eqs .
( [ caneom ] ) a local hamilton - jacobi equation , eq .
( [ hjeq ] ) , which generalizes to the field theory the hamilton - jacobi equation of classical mechanics .
( our approach should be compared with refs .
it is worth to emphasize that both , the canonical equations ( [ caneom ] ) , and the hamilton - jacobi equation ( [ hjeq ] ) , contain only partial , and not variational , derivatives . in sec .
[ sec : sym ] , we study transformations of the configuration space , and specify the condition , eq .
( [ symcrit ] ) , under which physical motions are mapped to physical motions .
such transformations are _ symmetries _ of the physical system .
the symmetries imply conservation laws through the noether theorem @xcite , whose hamiltonian version is derived in sec .
[ sec : conslawsfromsym ] .
the corresponding conservation law ( [ conslaw ] ) features a multivector - valued noether current , obtained by contracting the momentum with a symmetry - generating vector field .
two examples are provided to illustrate the universality of the presented formalism .
the first example ( sec . [ sec : exscalar ] ) discusses the theory of a real multicomponent scalar field .
it is shown that the canonical equations reproduce the de donder - weyl equations of motion @xcite , and the local hamilton - jacobi equation reproduces the one invented by weyl @xcite , when the scalar field is regarded as a function defined on the spacetime .
moreover , we examine the symmetries , namely , spacetime translations and rotations , and rotations in the field space , and associate our multivector - valued noether currents with the energy - momentum tensor and the angular - momentum tensor , and the standard vectorial noether currents , respectively . in the second example ( sec . [ sec : exstring ] ) , we treat relativistic particles , strings , or higher - dimensional membranes , depending on the dimensionality of the motions .
the configuration space is identified with the target space of the string theory , the motions are the worldsheets , and the hamiltonian is essentially the simplest and most symmetric function of the momentum variable .
the equations of motion have a simple geometric meaning , namely , they ensure that the mean curvature of the physical motion vanishes .
in fact , this is exactly the condition that defines _ minimal surfaces _ @xcite .
the hamilton - jacobi theory agrees with ref .
we also show that for nearly flat motions , the string theory yields the scalar field theory as a limiting case .
one more remark is in order before we start .
all manipulations are performed in the mathematical language of geometric ( or clifford ) algebra and calculus developed by d. hestenes @xcite ( see also ref .
we will assume that the reader is reasonably familiar with this language .
( a concise introduction into the geometric algebra techniques can be found in the appendices of refs .
@xcite and @xcite . in these articles
, we also provide a more detailed analysis of the subjects treated in the present article . )
let us start with a set of partial observables that constitute a @xmath8-dimensional euclidean configuration space @xmath9 .
( an extension to pseudo - euclidean spaces should be straightforward , but will not be discussed here . )
a point @xmath5 in the configuration space , e.g. , @xmath10 , represents a simultaneous measurement of all partial observables . to establish a _ physical _ theory
, one has to specify the correspondence between the partial observables and physical measuring devices , such as clocks , rulers , or instruments measuring the components of the field . in this article , we take such correspondence for granted , as we will only be concerned with the mathematical aspects of the theory .
let us denote by @xmath11 the dimensionality of motions , i.e. , submanifolds @xmath12 of the configurations space @xmath9 .
with @xmath13 one may study particle mechanics , with @xmath14 one can do the string theory or a field theory in two spacetime dimensions , and so on . we shall not consider systems with gauge invariance , for which the mathematical motion ( the surface in @xmath9 ) has higher dimensionality than the actual physical motion ( the physical trajectory ) .
the tangent space of @xmath12 at a point @xmath5 is spanned by @xmath11 linearly independent vectors @xmath15 , which are conveniently combined into a grade-@xmath11 multivector @xmath16 . the normalized version of this multivector
is called the _ unit pseudoscalar of @xmath12 _ , and it is denoted by @xmath17 . in the terminology used in ref .
* ch . 6 ) , the function @xmath18 represents a @xmath11-dimensional distribution on @xmath9 , with @xmath12 being its integral submanifold .
fundamental for the following formulation of dynamics is the concept of the generalized momentum @xmath6 , which is a grade-@xmath11 multivector defined at each point of @xmath12 ( see fig . [
fig : varprinc ] ) .
it serves as a quantity conjugated to @xmath17 , and in this sense it generalizes the canonical momentum of particle mechanics .
the hamiltonian @xmath19 is a generic function of positions and momenta , which is assumed to be scalar - valued .
( a generalization to the case of multicomponent @xmath4 is straightforward . )
the variational principle that determines the physical ( or classical ) motions of the field theory can now be stated as follows ( cf .
3.3.2 ) ) : a surface @xmath20 with boundary @xmath21 is a physical motion , if the couple @xmath22 extremizes the ( action ) functional @xmath23 = \int_\gamma p(q ) \cdot d\gamma(q)\ ] ] in the class of pairs @xmath24 , for which @xmath25 , and for which @xmath6 , defined along @xmath12 , obeys the hamiltonian constraint @xmath26 the integral in ( [ action ] ) is defined in ( * ? ? ?
* ch . 7 ) ( see also ref .
@xcite ) without resorting to any parametrization of the surface @xmath12 . the inner product between the momentum @xmath6 and the oriented surface element @xmath27 replaces the differential form @xmath28 used in ref .
there , the integral is taken over a submanifold of the bundle of @xmath11-forms over @xmath9 .
since we hesitate to work in spaces that mix points @xmath5 and multivectors @xmath6 , we prefer to integrate over surfaces in @xmath9 , and treat the momentum as a field defined along these surfaces .
we will now derive the equations of motion that follow from the variational principle . for this purpose
, we incorporate the hamiltonian constraint ( [ hamconstraint ] ) into the action ( [ action ] ) by means of a scalar lagrange multiplier @xmath29 .
the augmented action is a functional @xmath30 = \int_\gamma \left [ p(q ) \cdot d\gamma(q ) - \lambda(q ) h(q , p(q ) ) \right ] , \ ] ] where @xmath29 is , in fact , an infinitesimal quantity comparable with @xmath31 , the magnitude of @xmath27 .
the varied action @xmath32 $ ] is an integral over a new surface @xmath33 , with new functions @xmath34 and @xmath35 defined along @xmath33 ( see fig . [
fig : varprinc ] ) .
let @xmath36 be the infinitesimal diffeomorphism mapping the surface @xmath12 to @xmath33 , i.e. , @xmath37 , and let us denote by @xmath38 the variations of the momentum and the lagrange multiplier , respectively .
the infinitesimal variation of the action ( [ actionaugm ] ) , @xmath39 - \mathcal{a}[\gamma , p,\lambda]$ ] , is then given by @xmath40 - \int_\gamma \left [ p(q ) \cdot d\gamma(q ) \!-\ !
\lambda(q ) h(q , p(q ) ) \right ] , \ ] ] where we have employed the integral substitution theorem ( see ( * ? ? ?
* ch . 7 - 5 ) ) to transform the integral over @xmath33 to an integral over @xmath12 . for the infinitesimal diffeomorphism @xmath41 ,
the outermorphism mapping @xmath42 that specifies the transformation rule for multivectors , is given by formula ( [ gcinfsmtr ] ) .
therefore , to the first order in @xmath43 , @xmath44 , and @xmath45 , we find @xmath46 \nonumber\\ & \approx \int_\gamma \left [ - \delta\lambda \ ,
h(q , p ) + \delta p \cdot \big(d\gamma - \lambda \ , \partial_p h(q , p ) \big ) - \lambda \
\delta q \cdot \dot{\partial}_q h(\dot{q},p ) + p \cdot \big((d\gamma \cdot \partial_q ) \wedge \delta q \big ) \right ] , \end{aligned}\ ] ] where @xmath47 is the _ vector derivative _ with respect to a point in @xmath9 ( * ? ? ?
* ch . 2 - 1 ) , and @xmath48 is the _ multivector derivative _ with respect to the momentum multivector @xmath6 ( * ? ? ?
* ch . 2 - 2 ) .
the overdot " notation is used to indicate the scope of the differential operator @xmath47 . without an overdot ,
any differential operator is supposed to act on the functions that stand to its right .
the last term in eq .
( [ actionvar ] ) can be rewritten with a help of the _ fundamental theorem of geometric calculus _
* ch . 7 - 3 ) , @xmath49 where @xmath50 is the oriented surface element on the boundary @xmath51 .
now , the first term on the right - hand side of this equation vanishes , since we assume that @xmath12 and @xmath33 have a common boundary , i.e. , @xmath52 . as concerns the second term , for @xmath13
, @xmath53 is algebraically a scalar , and so the integrand is readily reshuffled , @xmath54 ( mind the priority of the inner product @xmath55 " , and the outer product @xmath56 " before the geometric product , which is denoted by an empty symbol . ) for @xmath57 , we may employ some basic geometric algebra identities to find @xmath58 the two cases have to be treated separately due to the definition of the inner product adopted in ref .
* ch . 1 ) .
after these rearrangements , we arrive at our final expression for the variation of the action , @xmath59 , \ ] ] which holds for @xmath57 , while the case @xmath13 is obtained simply by replacing @xmath60 with @xmath61 .
the requirement that @xmath62 vanish for all @xmath44 , @xmath43 , and @xmath63 yields the following physical motions @xmath20 are obtained by solving the system of differential equations [ caneom ] @xmath64 ( we use the adjective canonical " , because these equations generalize the hamilton s canonical equations of motion of classical mechanics @xcite . ) the first canonical equation ( [ caneom1 ] ) furnishes a relation between the momentum @xmath6 , and the tangent planes of @xmath12 , represented by the oriented surface element @xmath27 .
it asserts that the multivector derivative @xmath65 , which is a grade-@xmath11 multivector , is proportional to @xmath27 , with the proportionality constant equal to @xmath29 .
note that one can always divide @xmath29 and @xmath27 by the magnitude @xmath31 to free eqs .
( [ caneom ] ) from infinitesimal quantities .
the second canonical equation ( [ caneom2 ] ) describes how the momentum multivector @xmath6 changes as it slides along the surface @xmath12 .
it is important to note that @xmath6 is being differentiated , effectively , only in the directions parallel to @xmath12 , as a consequence of the inner product between the surface element @xmath27 , and the vector derivative @xmath47 .
moreover , the overdot " on the left - hand side assures that only the explicit dependence of @xmath4 on @xmath5 is being differentiated , not the dependence through @xmath66 .
the last canonical equation ( [ caneom3 ] ) is simply the hamiltonian constraint ( [ hamconstraint ] ) .
let us remark that had we started with several constraints @xmath67 in the variational principle , we would have introduced the corresponding number of lagrange multipliers @xmath68 , and , consequently , the canonical equations would contain the terms @xmath69 instead of @xmath70 .
one method to deal with the canonical equations is the following .
suppose @xmath66 obeys the hamiltonian constraint @xmath71 in some @xmath8-dimensional region in the configuration space @xmath9 . by differentiation , we obtain @xmath72 and using the first canonical equation ( [ caneom1 ] ) , we find that @xmath73 the right - hand side may be recast , using the identities ( 1.42 ) and ( 1.43 ) from ref .
@xcite , in the form @xmath74 now , we observe that if @xmath75 then eq . ( [ hjeom2 ] ) coincides with the second canonical equation ( [ caneom2 ] ) , which is then automatically fulfilled .
the momentum field that satisfies this condition can be expressed , at least locally , as @xmath76 , where @xmath77 is a multivector of grade @xmath78 ( cf . the relation between closed and exact differential forms ) . the canonical equations ( [ caneom ] ) are then reduced to two equations : @xmath79 and the _ local hamilton - jacobi equation _
@xmath80 if we succeed in finding a solution of eq .
( [ hjeq ] ) , we can plug it into eq .
( [ hjeqgamma ] ) , which then defines a distribution of the tangent planes of a classical motion surface @xmath20 .
this distribution can be integrated to yield the surface itself , provided certain integrability conditions are met ( see ( * ? ? ?
* ch . 6.1 ) ) . if we find a whole family of solution @xmath81 , parametrized by a continuous parameter @xmath82 , then , by differentiating eq .
( [ hjeq ] ) with respect to @xmath82 , and substituting eq .
( [ hjeqgamma ] ) , we obtain the relation @xmath83 now , for @xmath13 , the hamilton - jacobi function @xmath77 is scalar - valued , and we obtain @xmath84 for some constant @xmath85 , meaning that the quantity @xmath86 is conserved along a physical motion . finding @xmath87 such parameters @xmath82 ( recall that the dimension of the configuration space is now @xmath88 ) , the physical motion @xmath20 can be determined from the set of constraints between the partial observables , @xmath89 of course , we assume that the @xmath87 constraints are independent , i.e. , that the gradients @xmath90 are , at every point , linearly independent vectors . when @xmath57 , eq .
( [ hjparamsol ] ) can be rearranged , and integrated using the fundamental theorem of geometric calculus , @xmath91 where @xmath92 is an arbitrary @xmath11-dimensional subset of @xmath20 ( a patch " on @xmath20 ) .
( [ hjconserved ] ) and ( [ hjconteq ] ) express conservation laws for the conserved quantities @xmath93 ( see the noether theorem , eq .
( [ conslaw ] ) , below ) .
a remark is in order before we close this section . in classical particle mechanics ,
one of the solutions of the hamilton - jacobi equation is the action along a classical trajectory , regarded as a function of one of the endpoints . in the field theory
, the classical action may be viewed as a functional of the boundary @xmath94 .
some authors ( e.g. , ( * ? ? ?
3.3.4 ) ) have therefore considered a variational differential equation that describes how the classical action changes under variations of the boundary , using also the name hamilton - jacobi equation " .
note that eq .
( [ hjeq ] ) is substantially different from this kind of approaches , for it contains only partial , not variational , derivatives .
this is why we call it
local hamilton - jacobi equation " .
a local hamilton - jacobi theory is also treated , e.g. , in refs . @xcite and @xcite .
in this section , we will study transformations of the configuration space @xmath9 of partial observables , and identify among them the symmetries of a physical system . a transformation of @xmath9
is expressed mathematically as a diffeomorphism @xmath95 ( see fig . [
fig : transf ] ) . .
] it maps a surface @xmath12 to another surface @xmath96 whose boundary @xmath97 may differ from @xmath51 .
the surface elements on @xmath12 and @xmath33 are related by the induced outermorphism @xmath42 , @xmath98 ( transformations and induced mappings within the framework of geometric calculus are introduced in appendix [ sec : gagc ] , and thoroughly discussed in ( * ? ? ?
* ch.4 - 5 ) . ) note that @xmath41 is an _ active _ transformation , a mapping between the points of the configuration space @xmath9 . in a dual picture
, one could consider the _ passive _ transformations , i.e. , changes of the coordinates on @xmath9 .
since we are working completely without coordinates , all transformations are viewed as active .
the relation between the momentum fields on @xmath12 and @xmath33 is established by demanding that the inner product @xmath99 , and hence the action ( [ action ] ) , be invariant under @xmath41 .
this is achieved by postulating the transformation rule @xmath100 the invariance of the action then implies the following : consider an arbitrary diffeomorphism @xmath95 .
if @xmath20 is a physical motion of a system with hamiltonian @xmath4 , then @xmath101 is a physical motion of a system with hamiltonian @xmath102 , defined by @xmath103 where @xmath104 .
( an explicit proof of this claim on the level of canonical equations of motion is provided in ref .
@xcite . )
we call the transformation @xmath41 a _ symmetry _ , if it maps physical motions to physical motions of the same physical system
. this is the case when @xmath4 and @xmath102 coincide , i.e. , when @xmath105 as an immediate consequence of definition ( [ trham ] ) , we therefore obtain : a transformation @xmath41 is a symmetry of a physical system described by the hamiltonian @xmath4 ( or , in short , a symmetry of @xmath4 ) , if @xmath106 for infinitesimal transformations
@xmath107 , @xmath108 , determined by a vector field @xmath109 , eq .
( [ symcrit ] ) takes the form @xmath110 eq . ( [ symcritinfsm ] ) is obtained from eq .
( [ symcrit ] ) by a straightforward application of the infinitesimal version of the transformation rule ( [ trp ] ) , eq .
( [ gcinfsminvtr ] ) . more rigorously , the infinitesimal transformation arises from a one - parameter group of transformations @xmath111 in the small-@xmath112 limit , when we can approximate @xmath113 conversely , to any vector field @xmath114 corresponds a flow @xmath111 , which can be regarded as a group of transformations parametrized by @xmath112 .
an explicit formula is provided by the _ lie series _
1.3 ) , @xmath115
the symmetries of a physical system are imprinted in its hamiltonian function @xmath19 , and can be explored by analysing eqs .
( [ symcrit ] ) or ( [ symcritinfsm ] ) without any reference to the equations of motion .
however , when the system is assumed to follow a classical trajectory , then the symmetries induce _ conservation laws_. this fact is derived almost instantly in the hamiltonian constraint formalism . substituting canonical equations ( [ caneom1 ] ) and ( [ caneom2 ] ) , respectively , into the first and the second term in eq .
( [ symcritinfsm ] ) , we find ( for @xmath57 ) @xmath116 which can be readily rearranged , @xmath117 and finally combined into one term to yield the equation @xmath118 the derivation for the case @xmath13 is fully analogous .
let us summarize the above considerations in the following hamiltonian version of the celebrated if @xmath119 is an infinitesimal symmetry of @xmath4 , i.e. , if eq .
( [ symcritinfsm ] ) holds , then the solutions of the canonical equations of motion ( [ caneom ] ) satisfy the conservation law @xmath120 the quantities that obey conservation laws play distinguished role in physics .
the noether theorem therefore grants a special status to the @xmath78-vector @xmath121 , and clearly displays the importance of the momentum multivector @xmath6 not only in particle mechanics , but also in the classical field theory .
the integral form of the conservation laws is obtained , analogously to sec .
[ sec : lochj ] , by integrating eq .
( [ conslaw ] ) over an arbitrary connected @xmath11-dimensional subset @xmath92 of a physical motion @xmath20 , and by employing the fundamental theorem of geometric calculus . for @xmath13
, we obtain @xmath122 where @xmath123 , @xmath124 are the endpoints of the curve @xmath92 , whereas for @xmath57 , we find @xmath125 where @xmath50 is the oriented infinitesimal surface element of the boundary @xmath126 .
in this example , we split the configuration space @xmath9 into a @xmath11-dimensional spacetime with the unit pseudoscalar @xmath127 ( we will assume @xmath57 ) , and its @xmath87-dimensional orthogonal complement , the space of fields , with an orthonormal basis @xmath128 , and the unit pseudoscalar @xmath129 .
the points in @xmath9 then have a natural decomposition @xmath130 .
let us assume the following form of the hamiltonian : @xmath131 where @xmath132 is the _ de donder - weyl hamiltonian _
@xcite , which satisfies the conditions @xmath133 geometrically , these conditions mean that @xmath132 depends only on those components of the momentum @xmath11-vector @xmath6 , which are composed of one vector from the @xmath134-space , and @xmath78 vectors from the @xmath135-space .
in order to make contact with the standard theory of fields as functions defined on the spacetime , we represent the motions as @xmath136 where @xmath137 is a spacetime domain ( see fig . [
fig : scfield ] ) .
the surface element of @xmath12 is related to the oriented spacetime element @xmath138 via formula ( [ gcgdgamma ] ) , @xmath139 where the terms with more than one @xmath134 have been neglected .
in fact , they vanish in consequence of the second condition in ( [ dwcond ] ) , and the first canonical equation ( [ caneom1 ] ) , which for the hamiltonian ( [ hamdwform ] ) reads @xmath140 we may in addition assume that the classical momentum satisfies @xmath141 as this condition has no effect on the classical motions . comparing term by term eqs .
( [ sfdgamma ] ) and ( [ caneom1dw ] ) , we find that @xmath142 and @xmath143 the latter equation can be cast as @xmath144 due to the orthogonality of the @xmath135- and @xmath134-spaces .
formula ( [ gcgdgammapartial ] ) can be used to pull " the second canonical equation ( [ caneom2 ] ) down onto the spacetime to yield @xmath145 \cdot \mathsf{p } = ( -1)^d \dot{\partial}_q h_{\rm dw}(\dot{q},\mathsf{p } ) , \ ] ] where we have denoted @xmath146 .
dotting " this equation with a @xmath134-vector @xmath147 , the second term on the left - hand side drops out due to the assumption ( [ sfmomass ] ) , and we arrive at @xmath148 it is now straightforward to show that , choosing an orthonormal basis of the @xmath135-space , the components of eqs .
( [ dw1 ] ) and ( [ dw2 ] ) correctly reproduce the standard equations of motion of the de donder - weyl hamiltonian field theory . for the hamiltonian given by eq .
( [ hamdwform ] ) , the hamilton - jacobi equation ( [ hjeq ] ) reads @xmath149 where @xmath150 is a multivector of grade @xmath78 .
this can be related to the hamilton - jacobi equation derived formerly by weyl @xcite . to this end , let us assume that @xmath77 is a spacetime multivector , and define the vector @xmath151 . taking into account the assumptions ( [ dwcond ] ) , eq . ( [ scfieldhj ] ) is cast as @xmath152 which , when written out in components , is indeed the weyl s hamilton - jacobi equation . from now on ,
we shall be concerned only with a specialized form of the hamiltonian ( [ hamdwform ] ) , @xmath153 eq . ( [ dw1 ] ) in this case reads @xmath154 where @xmath155 denotes the _ reversion _ of @xmath156 ( see the definition ( 1.17 ) in ( * ? ? ?
* ch . 1 - 1 ) ) . writing the field @xmath134 in components , @xmath157
, the latter equation reads @xmath158 the last equality holds in euclidean spaces , where @xmath159 . at this point
it is worth to note that for the hamiltonian @xmath160 the extended action ( [ actionaugm ] ) can be cast , using eqs .
( [ sfdgamma ] ) and ( [ sflambda ] ) , as an integral over the spacetime domain @xmath137 , @xmath161 - |dx| h_{\rm sf } \right\ } \nonumber\\ & = \int_\omega |dx| \left\ { ( i_x \cdot \dot{\partial}_x ) \cdot ( \dot{y } \cdot \mathsf{p } ) - \frac{1}{2 } \sum_{a=1}^n \big ( i_x \cdot ( \mathsf{p } \cdot e_a ) \big)^2 - v(y ) \right\ } .\end{aligned}\ ] ] eliminating the momentum by virtue of eq .
( [ sfmomsubs ] ) , and employing the identity @xmath162 which holds for any spacetime vectors @xmath163 and @xmath164 , we obtain @xmath165 where @xmath166 is the usual lagrangian of an @xmath87-component scalar field @xmath167 .
this observation justifies , a posteriori , the title of this section scalar field theory " .
equation ( [ gcgdgammapartial ] ) can be used to pull " the conservation law ( [ conslaw ] ) down onto the spacetime to recover the standard form of the continuity , and relate the conserved multivectors @xmath121 to the noether currents . for this purpose ,
we define @xmath168 , and calculate @xmath169 \cdot ( \mathsf{p } \cdot \mathsf{v } ) \nonumber\\ & = ( -1)^{d-1 } ( \partial_x \cdot dx ) \cdot \left [ \mathsf{p } \cdot \mathsf{v } + \dot{\partial}_x \wedge \big ( \dot{y } \cdot ( \mathsf{p } \cdot \mathsf{v } ) \big ) \right ] \nonumber\\ & = |dx| ( -1)^d \partial_x \cdot j(x ) , \end{aligned}\ ] ] where we have denoted @xmath170 .\ ] ] this is the standard noether current corresponding to the symmetry generated by the vector field @xmath109 . in view of the conservation law ( [ conslaw ] )
, it satisfies the continuity equation @xmath171 we will now show that the scalar - field hamiltonian @xmath160 enjoys some well known symmetries ( depicted in fig .
[ fig : sfgen ] ) , and exploit the corresponding conserved currents . : spacetime translations @xmath172 , spacetime rotations @xmath173 , and field - space rotations @xmath174 .
the spacetime or the field space are conveniently depicted as two - dimensional planes @xmath175 or @xmath176 , respectively . ] for global spacetime translations @xmath177 where @xmath178 is a constant spacetime vector , the differential mapping is trivial , @xmath179 and so is the adjoint , @xmath180 the vector @xmath178 is at the same time the generator of translations , @xmath181 as can be ascertained by calculating @xmath182 .
the transformation @xmath183 is , according to eq .
( [ symcrit ] ) , a symmetry of the hamiltonian @xmath160 , since @xmath160 does not depend on @xmath135 .
the conserved quantity @xmath184 is related to the noether current @xmath185 via eq .
( [ noethercur ] ) .
explicitly , @xmath186 \nonumber\\ & = - v_x \left [ \mathsf{p } \cdot i_x + ( \mathsf{p } \cdot \dot{y } ) \cdot ( \dot{\partial}_x \cdot i_x ) \right ] + v_x \cdot \dot{\partial}_x \ , i_x \cdot ( \mathsf{p } \cdot \dot{y } ) , \end{aligned}\ ] ] where we have used the fact that @xmath187 .
substituting now for @xmath188 from the hamiltonian constraint @xmath189 , and for @xmath190 from eq .
( [ sfmomsubs ] ) , and using the identity ( [ projidsc ] ) , we arrive at @xmath191 + \sum_{a=1}^n ( v_x \cdot \partial_x \phi_a ) ( \partial_x \phi_a ) \nonumber\\ & = - v_x \mathcal{l}_{\rm sf } + \sum_{a=1}^n ( v_x \cdot \partial_x \phi_a ) \frac{\partial \mathcal{l}_{\rm sf}}{\partial ( \partial_x \phi_a ) } .\end{aligned}\ ] ] this is the standard energy - momentum tensor of a scalar field with lagrangian ( [ lagrsf ] ) . in its natural geometric interpretation
, @xmath192 is an @xmath135-dependent linear mapping of spacetime vectors @xmath178 to spacetime vectors @xmath193 .
a spacetime rotation about a point @xmath194 is defined @xmath195 where @xmath196 is a constant spacetime bivector , and @xmath197 is the corresponding _ rotor_. the associated differential mapping is readily obtained , @xmath198 and the transformation rule for the momentum is found , @xmath199 ( the implementation of rotations using geometric algebra is discussed in ( * ? ? ?
* ch . 3 - 5 ) .
) by expanding the right - hand side of the definition ( [ rotxdef ] ) , and comparing with the lie series , eq .
( [ lieseries ] ) , we find the infinitesimal generator of @xmath200 , @xmath201 in order to show that @xmath200 is a symmetry of @xmath160 , we realize that @xmath202 and @xmath203 , and calculate @xmath204 ^ 2 + v(y ) \nonumber\\ & = p \cdot i_x + \frac{1}{2 } \sum_{a=1}^n \big ( i_x \cdot ( p \cdot e_a ) \big)^2 + v(y ) = h_{\rm sf}(q , p ) .\end{aligned}\ ] ] since @xmath173 is a spacetime vector , it is easy to find an explicit relation between @xmath205 and the corresponding noether current @xmath206 .
we simply replace in eq .
( [ sfenmomtensor ] ) @xmath178 by @xmath207 : @xmath208 this is the angular momentum tensor corresponding to the energy - momentum tensor @xmath209 .
geometrically , @xmath206 is an @xmath135-dependent linear mapping , with a parameter @xmath194 , that maps spacetime bivectors @xmath196 to spacetime vectors @xmath210 ( c.f
. ch . 13.1 in ref .
@xcite ) .
finally , let us consider rotations of the form @xmath211 where @xmath212 is a constant bivector from the field space , i.e. , @xmath213 .
the corresponding differential reads @xmath214 and the momentum transforms as @xmath215 the generator of the field - space rotations is found in the same way as the generator of the spacetime rotations ( [ sfgenrotx ] ) , @xmath216 the hamiltonian @xmath160 transforms under @xmath217 as follows ( note that @xmath218 ) : @xmath219 ^ 2 + v(r_y y { \widetilde{r}}_y ) .\ ] ] if we assume that @xmath220 , which is fulfilled , for example , when the potential @xmath221 depends only on @xmath222 , then the right - hand side of eq .
( [ sfhamroty ] ) is equal to @xmath223 , and hence @xmath217 is a symmetry of @xmath160 .
note that the second term in @xmath160 is invariant under a change of the orthonormal basis of the @xmath134-space , @xmath224 , as can be easily ascertained .
the vector field @xmath174 lies entirely in the @xmath134-space .
therefore , owing to the assumption ( [ sfmomass ] ) , the second term in expression ( [ noethercur ] ) for the noether current drops out , and we obtain @xmath225 a substitution for the momentum from eq .
( [ sfmomsubs ] ) then yields @xmath226 the noether current @xmath227 is an @xmath135-dependent linear mapping of field - space bivectors @xmath212 to spacetime vectors @xmath228 .
probably the simplest nontrivial hamiltonian , which preserves the full symmetry of the configuration space @xmath9 , is @xmath229 where @xmath230 is a scalar constant , and @xmath231 is the magnitude of @xmath6 .
this hamiltonian described the dynamics of a relativistic particle ( for @xmath13 ) , a string ( for @xmath14 ) , or a higher - dimensional membrane ( for @xmath232 ) that propagates in a euclidean spacetime @xmath9 .
the corresponding worldlines ( or worldsheets ) are identified with the motions @xmath12 . the first canonical equation ( [ caneom1 ] ) takes the form @xmath233 which , when substituted into the hamiltonian constraint ( [ caneom3 ] ) , fixes the absolute value of the lagrange multiplier @xmath29 , @xmath234 ) into the second canonical equation of motion ( [ caneom2 ] ) , dividing by @xmath29 , and using eq .
( [ stringeom3 ] ) , we find @xmath235 where @xmath236 is the unit pseudoscalar of the surface @xmath12 .
this equation has a simple geometric interpretation .
it entails vanishing of the _ mean curvature _ of the surface @xmath12 , or , more generally , of its _ spur _ vector ( see ref .
* ch . 4 - 4 ) ) .
( [ stringeom1 ] ) and ( [ stringeom3 ] ) allow us to eliminate @xmath6 and @xmath29 , and rewrite the action ( [ action ] ) in terms of @xmath27 only , @xmath237 where @xmath238 " is the sign of @xmath29 .
this is the euclidean nambu - goto action of the bosonic string theory @xcite .
it is proportional to the volume of the worldsheet @xmath12 , with @xmath239 playing the role of the string tension .
the extremals of the action @xmath240 , i.e. , the solutions of eq .
( [ stringeom ] ) , minimize their volume for a given fixed boundary .
therefore , they are called _ minimal surfaces _ in the mathematical literature @xcite . if assume that the worldsheets are nearly flat , and represent them in the same way as the scalar field , eq .
( [ motionfunc ] ) , then the string action is cast as @xmath241 , \ ] ] where @xmath242 , and the terms of order greater than @xmath243 have been neglected . a comparison with the scalar - field action @xmath244 , eq .
( [ actionsf ] ) , yields @xmath245 and hence we conclude that the string theory for slowly varying worldsheets essentially reduces to a potential - free massless scalar field theory . for the moment , let us focus on the case @xmath13 , which describes a relativistic particle in the euclidean spacetime .
we will present two methods for finding the physical motions @xmath20 .
first , suppose that two points , @xmath247 and @xmath5 , lie on @xmath20 , multiply the equation of motion ( [ stringeom ] ) by @xmath31 , and integrate along @xmath20 from @xmath247 to @xmath5 .
the fundamental theorem of calculus implies that @xmath248 i.e. , @xmath17 is constant along @xmath20 .
the physical motions are therefore straight lines in @xmath9 , @xmath249 where @xmath250 and @xmath109 is an arbitrary constant vector .
the second method makes use of a family of solutions of the hamilton - jacobi equation ( [ stringhj ] ) .
take , for example , @xmath251 the derivative of @xmath77 with respect to @xmath247 yields , according to formula ( [ hjconserved ] ) , the conserved quantities @xmath252 the physical motion are then given by @xmath253 where @xmath109 is an arbitrary constant unit vector . for the hamiltonian @xmath254 , the infinitesimal symmetry condition , eq .
( [ symcritinfsm ] ) , reads @xmath255 this has to be satisfied for all constant @xmath11-vectors @xmath6 .
observe that the left - hand side is equal to @xmath256 \cdot { \widetilde{p } } = \frac{1}{2 } \sum_{j=1}^{n+d } ( \partial_q v \cdot e_j + e_j \cdot \partial_q v ) \cdot \big ( ( e_j \cdot p ) \cdot { \widetilde{p } } \big ) , \ ] ] where the @xmath257 s form an orthonormal basis of the configuration space @xmath9 .
the solution of eq .
( [ strsym ] ) is therefore a vector field @xmath109 , for which @xmath258 holds for all constant vectors @xmath163 .
taking the curl , the right - hand side vanishes , and we find @xmath259 which implies @xmath260 where @xmath261 is a constant bivector .
moreover , note that from eq .
( [ symgenstr ] ) follows that @xmath262 and hence @xmath263 from which we finally obtain an expression for the symmetry generator @xmath109 , @xmath264 where @xmath265 is a constant vector .
the vector field @xmath109 is composed of two terms , the translation generator @xmath266 and the rotation generator @xmath267 the corresponding finite symmetry transformations can be obtained directly from the lie series , eq .
( [ lieseries ] ) , and read ( setting @xmath268 ) @xmath269 and @xmath270 respectively . in view of eqs .
( [ stringeom1 ] ) and ( [ stringeom3 ] ) , the conserved quantities take the form @xmath271 where
@xmath238 " is the sign of @xmath29 .
in this article we studied and developed the formulation of classical field theories proposed in ref .
* ch . 3 ) .
this formulation is based on the hamiltonian constraint @xmath272 between the partial observables @xmath5 , and the generalized momentum @xmath6 , and the fields are viewed as unparametrized submanifolds of the configuration space .
starting from the variational principle of sec .
[ sec : varprinc ] , we derived the canonical equations of motion ( [ caneom ] ) , and , subsequently , deduced the local hamilton - jacobi equation ( [ hjeq ] ) .
these results generalize to the field theory the respective notions from the hamiltonian particle mechanics . in sec .
[ sec : sym ] , we discussed transformations of the configuration space , and identified the symmetry transformations by the condition ( [ symcrit ] ) .
in the ensuing section , taking into account the canonical equations of motion , symmetries were shown to imply conservation laws , eq .
( [ conslaw ] ) , thus establishing a hamiltonian field - theoretical version of the noether theorem .
the simple form of the conserved quantities @xmath121 , where @xmath109 is the vector field that generates the symmetry , clarifies the physical significance of the momentum multivector @xmath6 . with two ensuing examples
, we showed that scalar fields and strings can both be accommodated within our formalism .
one only has to take the appropriate hamiltonian constraint .
in fact , as we also demonstrated , the scalar field theory is a limiting case of the string theory , in a similar way in which the non - relativistic particle mechanics is a limiting case of the relativistic mechanics . to make contact with the standard treatment
, we showed that our hamiltonian constraint approach to the scalar field theory leads to the de donder - weyl formalism .
moreover , we expressed the energy - momentum tensor of the scalar field in terms of the conserved multivector @xmath121 to argue that the latter is a more primitive , and therefore more fundamental , object .
the hamiltonian formalism is especially important when it comes to quantization . in mechanics ,
the momentum is promoted to a differential operator , and the hamilton - jacobi equation is replaced by the schrdinger equation .
it is desirable to have an analogous quantization scheme also for the field theory , which is currently most commonly quantized using lagrangians and feynman path integrals . within the de donder - weyl field theory ,
quantum momentum operators and a schrdinger - like equation have already been proposed @xcite .
we would like to make use of these lessons to develop an analogous formulation of the quantum field theory , based on the more general hamiltonian constraint approach .
the author would like to thank igor kanatchikov for valuable discussions , and the following institutions for financial support : czech technical university in prague , grant sgs13/217/ohk4/3t/14 , czech science foundation ( gar ) , grant ga14 - 07983s , and deutsche forschungsgemeinschaft ( dfg ) , grant kl 256/54 - 1 .
let @xmath95 be a diffeomorphism relating points in the configuration space , and consider the directional derivative @xmath273 @xmath274 is the _ differential _ , a @xmath5-dependent linear mapping of vectors at a point @xmath5 to vectors at @xmath275 .
it can be extended to an _ outermorphism _ acting on the whole geometric algebra by demanding the property @xmath276 for all multivectors @xmath277 and @xmath278 .
the _ adjoint _ of @xmath274 , denoted @xmath279 , is defined via the relation @xmath280 let us specialize to infinitesimal diffeomorphisms @xmath281 where @xmath109 is a vector field on the configuration space @xmath9 .
the action of the differential on an @xmath282-blade @xmath283 is given by @xmath284 for the adjoint outermorphism , we find @xmath285 by the linearity of the above expressions , we therefore conclude that for an arbitrary multivector @xmath277 , @xmath286 up to the first order in @xmath287 . in this approximation
, the inverse of @xmath41 reads @xmath288 , and so we immediately obtain also @xmath289 now , let us briefly consider an example of a mapping between two different manifolds . in the scalar field theory , sec .
[ sec : exscalar ] , we represented the motions @xmath12 by the functions @xmath290 a spacetime blade @xmath291 is then mapped by the associated outermorphism @xmath292 to a blade @xmath293 in the tangent algebra of @xmath12 . here , @xmath294 denotes the vector derivative with respect to a spacetime point , and the ellipsis gathers the terms with two and more @xmath134 s .
formula ( [ gcgdifmap ] ) can be applied to express the oriented surface element of @xmath12 as @xmath295 moreover , from the chain rule for differentiation @xmath296 we find that @xmath297 where @xmath298 , and we have used the identities ( 1.14 ) from ref .
* ch . 3 ) .
the differential operator @xmath299 acts on all functions to its right . whether these include also @xmath300 in the second , or higher , term on the right - hand side of eq .
( [ gcgdgammapartial ] ) has no effect , since @xmath301 , and @xmath302 .
e. noether , _ invariante variationsprobleme _ ,
zu gttingen , math - phys .
klasse , 235 - 257 ( 1918 ) ( see transport theory and stat
. phys . * 1 * 186 - 207 ( 1971 ) for an english translation ) .
g. e. sobczyk , in _ clifford algebras and their applications in mathematical physics : proceedings of second workshop held at montpellier , france , 1989 _ , edited by a. micali , r. boudet , and j. helmstetter , pp 279 - 292 ( 1992 ) [ http://geocalc.clas.asu.edu/pdf-preadobe8/simp_cal.pdf ] . | classical field theory is considered as a theory of unparametrized surfaces embedded in a configuration space , which accommodates , in a symmetric way , spacetime positions and field values .
dynamics is defined via the ( hamiltonian ) constraint between multivector - valued generalized momenta , and points in the configuration space .
starting from a variational principle , we derive the local equations of motion , that is , differential equations that determine classical surfaces and momenta .
a local hamilton - jacobi equation applicable in the field theory then follows readily .
in addition , we discuss the relation between symmetries and conservation laws , and derive a hamiltonian version of the noether theorem , where the noether currents are identified as the classical momentum contracted with the symmetry - generating vector fields .
the general formalism is illustrated by two examples : the scalar field theory , and the string theory . throughout the article
, we employ the mathematical formalism of geometric algebra and calculus , which allows us to perform completely coordinate - free manipulations . |
very long baseline interferometry ( vlbi ) is a powerful technique for obtaining positions of celestial objects with milliarcsecond ( mas ) level accuracy .
the vlbi exploration of radio astrometry ( vera ) telescopes are a japanese vlbi array dedicated to phase referencing vlbi @xcite .
vera consists of four 20 m diameter antennas at mizusawa , ogasawara , iriki and ishigaki - jima ( see figure 1 of @xcite ) . to overcome phase fluctuations and the limited integration time in conventional fast - switching vlbi
, vera has a dual beam system that allows simultaneous observations of the target and reference sources separated by 0.3 to 2.2 degrees .
this advanced capability of vera can be used to measure annual parallaxes and proper motions of masers with a 10 micro arcsecond ( @xmath5as ) level accuracy .
an annual parallax gives a simple geometrical measure of the distance , free from the complex assumptions required in other distance estimators .
studies of asymptotic giant branch ( agb ) stars by @xcite , @xcite , and @xcite are based on the annual parallax derived from vlbi observations , and demonstrated the the ability of vlbi astrometry .
more recently , successful measurements of distances and proper motions with vera have been reported @xcite . in the present paper ,
we report the observations of s crt , concentrating on the distance measurement .
s crt ( iras11501@xmath00719 , irc@xmath010259 , afgl4830s ) is an agb star , with a pulsation period of 155days@xcite .
s crt has a variability type of srb in the general catalog of variable stars @xcite .
the apparent magnitude at @xmath6 band is [email protected] @xcite .
the optical light curve in @xmath7 band is presented in figure [ lightcurves](a ) exhibiting @xmath80.8mag variation .
the distance to s crt has been measured in various ways . @xcite and @xcite determined the distance to be 420pc and 285pc , respectively , using the period - luminosity(pl ) relation of mira variables . @xcite used hipparcos data to measure a parallax of [email protected] , corresponding to a distance of 490pc , but with an large uncertainty .
recently , the new hipparcos catalog @xcite gives the parallax of [email protected] , corresponding to a distance of 787pc .
vera offers the potential to determine the parallax with a better accuracy so that the inconsistencies in these distance measurements can be resolved .
the observations and data reduction are described in section 2 . in section 3 , we present absolute positions and internal motions of maser spots .
the annual parallax of s crt is presented and converted to the absolute distance in this section . finally ,
in section 4 , we examine the properties of s crt in the context of the new distance measurement .
we have conducted a series of monthly vlbi observations from october 2005 to may 2007 with vera .
the duration of each observation was typically 8.5hours , yielding a net integration time of 5 to 6 hours . the typical synthesized beam size ( fwhm )
was 1.5mas @xmath9 0.8mas with a position angle ( p.a . ) of @xmath10 . to obtain the positions of maser spots in s crt ,
the quasar j1147@xmath00724 was simultaneously observed as the position reference .
the j2000 _ a priori _ coordinates of two sources are ( @xmath11 , @xmath12 ) @xmath13 ( 11h 52 m 45.0981s , @xmath007@xmath1
35 48.072 ) for s crt , and ( 11h 47 m 51.554035s , @xmath007@xmath1 24 41.14109 ) for j1147@xmath00724 .
j1147@xmath00724 is classified as a candidate " source in the catalog of the international celestial reference frame ( icrf ) , and its position errors are 270@xmath5as and 290@xmath5as in r.a . and
dec , respectively @xcite .
the separation and p.a . of these sources are 1.23@xmath1 and @xmath14 . from a total of 17 observations
, we used 12 observations in the present study . the observation in 2005
november was not used because the ishigaki - jima antenna did not participate .
bad weather conditions in four observations held between august and october in 2006 resulted in much higher system noise temperatures ( @xmath15500k ) at some of the stations .
the shorter integration times and unsolved phase fluctuations in these four observations degraded the image qualities of s crt in both self - calibration and phase reference analyses .
the observation status is summarized in table [ observations ] , which is organized as follows : column ( 1)the i d number of epoch . the numbers with an asterisk `` * ''
indicates that the observation was used in the estimation of the annual parallax .
column ( 2)date of the observation .
column ( 3)the year and day of the year(doy ) . in the observations in 2007/013 , 2007/096 , and 2007/130 , the phase referenced images are scattered into several components , while the self - calibrated images resulted in a single bright component , clearly indicating that the atmospheric conditions on these days prohibited succesful 22ghz phase referencing .
we did not use these three observations for the parallax estimation . in parallel with the vlbi program ,
we have monitored s crt with single - dish observations at iriki station since 2003 september with the typical interval of one month ( see also @xcite ) .
a data recording rate of 1024mbps was adopted with the vera dir2000 recording system , which yields the total receivable bandwidth of 256mhz with 2-bit digitization .
the 256mhz bandwidth data of left - hand circular polarization were divided into 16 if channels of 16mhz band width , one of which was used to receive the maser emission and the others were used to receive the continuum emission from j1147@xmath00724 .
cross - correlation was carried out with the mitaka fx correlator @xcite at the national astronomical observatory of japan ( naoj ) . in most observations ,
the if channel assigned to the water maser was divided into 512 spectral channels , yielding a frequency spacing of 31.25khz corresponding to a velocity resolution of 0.42kms@xmath2 .
the observations in 2006/069 and 2006/129 provide two times higher frequency resolution than other observations , yielding a velocity resolution of 0.21kms@xmath2 .
for the data of j1147@xmath00724 , the if channels were divided into 64 spectral channels in all observations .
we used the astronomical imaging package software ( aips ) developed in the national radio astronomical observatory in the data reduction .
amplitude calibration was achieved using the system noise temperatures and gains logged during the observations at each station . in the fringe search process of the reference source j1147@xmath00724 , we used the task fring with a typical integration time of 2minutes , with solutions of fringe phases , group delays , and delay rates obtained every 30seconds . using the task tacop ,
these solutions were transfered to the data of s crt in order to calibrate the visibility data .
phase and amplitude solutions obtained from self - calibration of the j1147@xmath00724 were also transferred to the s crt data . since the delay - tracking models used to estimate _ a priori _ delays in the mitaka fx correlator were not accurate enough for astrometry with vera , we have applied better estimates calculated with the calc3/msolv software package @xcite .
@xcite gave a brief description of this package in the report to the international earth rotation service ( iers ) .
calc3 has slight differences in the physical models than those adopted in calc which was developed by the research group at nasa goddard space flight center ( gsfc ) .
comparing the two _ a priori _ values from mitaka fx correlator and the calc3/msolv package , we obtained the difference between the two estimates , then applied them to the visibility data using the tasks tbin and clcal in aips .
we note that the zenith excess path lengths due to the wet atmosphere measured by the global positioning system ( gps ) at each station are considered in the calc3/msolv package .
we estimated the residual atmospheric zenith delay offset using the method described in @xcite .
this correction was applied to one of the four stations of vera where the coherence of the phase - referenced image was maximized .
typically , the offsets were found within @xmath33 cm . the instrumental delay caused by the difference between two signal paths
was estimated using an artificial noise source @xcite .
these delays were then loaded into aips and applied in the same manner as the delay tracking model correction . for each observation , we fitted a two dimensional gaussian model to the brightness distribution to find the position of the maser spot
. then we used these to derive the annual parallax and linear proper motions ( @xmath16 , @xmath17 ) . in the single - beam vlbi imaging of s crt ,
group delays solved with 3c279 were applied , then fringe - fitting was done using the brightest emission showing the lsr velocity of 34.7kms@xmath2 .
thus relative positions of all maser spots in self - calibrated images were determined with respect to this reference spot .
for each velocity channel , position of the spot is defined as the brightness peak of the image . to estimate the relative motion of each spot ,
a linear least - squares analysis was applied to the spot seen at the same velocity channel during at least two continuous observations .
the signal to noise ratio of 10 was adopted as the detection criterion in the self - calibrated images .
the rms noise of the self - calibrated images was typically 90mjybeam@xmath2 .
on the other hand , the rms noise of the phase - referenced images was typically 700mjybeam@xmath2 , about one order of magnitude larger than that of self - calibrated images . due to limited phase coherence
the phase referenced images have limited signal - to - noise ( 30 vs 307 in the self - calibrated images ) .
.observations [ cols="^,<,^,^",options="header " , ]
in figure [ fig - spectr ] , cross - power spectra of s crt from 2005/019(upper ) and 2006/069(lower ) on the mizusawa iriki baseline are presented .
the time variation of the flux densities with the total power spectrum are presented in figure [ lightcurves](b ) with an indication of each lsr velocity .
the 34.7kms@xmath2 spot underwent a radio flare(e.g .
this flare started in february 2006 and reached a maximum of 371jy in march 31(2006/090 ) , then it decreased to 60jy in july . during the flare , this spot did not show structural change .
the cross - power spectrum at the same time showed significantly weaker ( @xmath1830% ) flux density than that of the total power spectrum , indicating that some of the flared emission is resolved on the vera baselines . in figure [ fig - sky ] , we present the positions of maser spot at the lsr velocity of 34.7kms@xmath2 relative to the phase tracking center . throughout the present vlbi observations , this maser spot was bright enough to be detected on all baselines in the phase referencing analyses .
the proper motion was clearly modulated by a parallax .
based on a least - squares fitting analysis , the parallax was determined to be [email protected] which corresponds to a distance of 430@xmath4pc . here , we adopted position errors of each measurement that are obtained as the root sum squares of three error factors , and the details are given in section 4.1 . in the estimation of the parallax , we adopt a very small number of assumptions : the maser spot is moving on a linear trajectory with respect to the star , i.e. , there is no acceleration , and the reference source is fixed on the sky , i.e. , no motions due to core shift or jet features . from the fitting results ,
the linear proper motions of the reference spot ( @xmath16 , @xmath17 ) were obtained to be ( @xmath16 , @xmath17 ) @xmath13 ( @xmath19masyr@xmath2 , @xmath20masyr@xmath2 ) .
this motion is the combination of the proper motion of s crt system and internal motion of the maser spot in the system . considering the offset of the reference spot from the phase tracking center ( @xmath21,@xmath22 ) @xmath13 ( @xmath23mas , @xmath24mas ) , the j2000.0 absolute coordinates of this spot in 2005/292
were obtained to be ( @xmath11,@xmath12 ) @xmath13 ( 11h 52 m 44.96969s , @xmath007@xmath1 35 48.0958 ) .
this is the position as referenced to the position of j1147@xmath00742 and in relative offset to the original phase tracking centre .
the uncertainty of this position is estimated as @xmath8400@xmath5as based on the errors in our phase referencing analysis ( detailed in section 4.1 ) and icrf position of j1147@xmath00724 .
the reference source j1147@xmath00724 exhibited an unresolved structure .
the correlated amplitude as a function of ( @xmath25 ) distance is flat and the upper limit of the source size is 0.8mas ( fwhm ) , which is the minor axis of the synthesized beam in the present observation . in addition , we confirmed that the images of j1147@xmath00724 showed no distinctive change during the observations .
is indicated with the vertical dashed line in the spectra .
the blue - shifted component with respect to the stellar velocity has been brighter than the red - shifted one in the majority of our observations .
, width=302 ] in figure [ fig - map ] , we present the angular distribution and three - dimensional velocity field of maser spots in s crt covering a 60mas @xmath9 60mas region . the maser spot with the lsr velocity of 34.7kms@xmath2
is placed at the map origin . at the distance of 430pc
, 1mas corresponds to 0.43au , and 1masyr@xmath2 corresponds to a velocity of 2.04kms@xmath2 .
the color index in figure [ fig - map ] shows the lsr velocity range from 34.0 to 41.0kms@xmath2 .
the blue- and red - shifted components are separated into the northeast and southwest parts of the area .
the relative motion of each spot ( @xmath26 , @xmath27 ) with respect to the reference spot are used to determine the average motion ( @xmath28 , @xmath29 ) and , hence we obtained ( @xmath28 , @xmath29 ) @xmath13 ( @xmath01.605masyr@xmath2 , @xmath00.252masyr@xmath2 ) .
then , we subtracted ( @xmath28 , @xmath29 ) from ( @xmath26 , @xmath27 ) to obtain the internal motions ( @xmath30 , @xmath31 ) , that are presented with arrows in figure [ fig - map ] .
we successfully detected the internal motions of 26 maser spots .
the typical transverse speed was obtained to be 2.72masyr@xmath2 , corresponding to 5.56kms@xmath2 , by averaging the internal motions of all 26spots .
the parameters of the motions are presented in table [ spot ] in the increasing order of lsr velocity . for the reference spot
, we subtracted the internal motion from the proper motion ( @xmath16 , @xmath17 ) and , thus , the proper motion of s crt system was estimated to be ( @xmath32masyr@xmath2 , @xmath33masyr@xmath2 ) . in the new hipparcos catalog @xcite , the absolute proper motion of s crt was ( @xmath34masyr@xmath2 , @xmath35masyr@xmath2 ) , representing good consistency with our result within the errors .
ccccccccc i d & @xmath36&@xmath37 & @xmath38 & @xmath39 & @xmath30 & @xmath40 & @xmath31 & @xmath41 + & ( kms@xmath2)&(mas ) & ( mas ) & ( jybeam@xmath2)&(masyr@xmath2 ) & & ( masyr@xmath2 ) & + 1 & 34.10 & @xmath01.25 & 0.34 & 2.3 & & & & + 2 & 34.53 & @xmath01.71 & @xmath02.50 & 2.7 & @xmath00.01 & 0.01 & 4.90 & 2.58 + 3 & 34.55 & 0.76 & @xmath05.39 & 4.4 & 4.15 & 0.55 & @xmath01.19 & 0.06 + 4 & 34.68 & 0.00 & 0.00 & 19.5 & 1.61 & & 0.25 & + 5 & 34.68 & @xmath016.97 & @xmath00.34 & 4.6 & @xmath00.55 & 0.02 & 0.69 & 0.09 + 6 & 34.68 & @xmath02.85 & @xmath00.91 & 8.1 & & & & + 7 & 34.72 & 0.73 & @xmath01.42 & 1.8 & 3.50 & 0.29 & @xmath01.09 & 0.45 + 8 & 34.72 & 1.43 & @xmath02.00 & 1.9 & 2.10 & 0.54 & 0.88 & 0.14 + 9 & 34.74 & 1.64 & @xmath010.71 & 5.8 & & & & + 10 & 34.94 & @xmath01.72 & @xmath03.04 & 1.6 & 0.63 & 0.50 & 3.56 & 0.37 + 11 & 34.95 & 1.45 & @xmath01.86 & 35.4 & 2.08 & 0.89 & 0.58 & 0.06 + 12 & 34.97 & 2.51 & @xmath07.51 & 2.3 & & & & + 13 & 34.97 & 1.11 & @xmath027.05 & 2.5 & & & & + 14 & 35.10 & @xmath00.03 & 0.01 & 8.4 & 1.85 & 0.74 & 0.10 & 0.04 + 15 & 35.10 & @xmath02.89 & @xmath00.92 & 4.9 & & & & + 16 & 35.16 & 3.52 & @xmath02.50 & 3.5 & & & & + 17 & 35.16 & 2.13 & @xmath05.20 & 3.2 & & & & + 18 & 35.18 & 2.38 & @xmath07.37 & 2.1 & & & & + 19 & 35.18 & 1.14 & @xmath027.01 & 2.1 & & & & + 20 & 35.52 & 1.30 & @xmath02.21 & 1.5 & 2.05 & 0.90 & 0.68 & 0.36 + 21 & 35.52 & @xmath017.32 & @xmath00.74 & 1.3 & @xmath00.66 & 0.06 & 0.80 & 0.11 + 22 & 35.54 & @xmath05.04 & @xmath01.04 & 0.9 & & & & + 23 & 36.45 & 3.17 & @xmath010.90 & 1.3 & 3.39 & 0.52 & @xmath00.98 & 0.49 + 24 & 37.89 & @xmath014.77 & @xmath022.00 & 0.8 & @xmath04.37 & 0 & @xmath08.49 & 0 + 25 & 38.32 & @xmath027.32 & @xmath026.17 & 1.3 & & & & + 26 & 38.73 & @xmath027.06 & @xmath028.97 & 1.0 & @xmath02.83 & 0.15 & 0.67 & 0.45 + 27 & 38.73 & @xmath019.16 & @xmath021.19 & 0.9 & @xmath04.70 & 0 & 2.46 & 0 + 28 & 38.74 & @xmath029.53 & @xmath023.72 & 0.9 & & & & + 29 & 38.74 & @xmath011.30 & @xmath027.07 & 1.0 & @xmath00.55 & 0.02 & 0.69 & 0.09 + 30 & 38.91 & @xmath04.01 & @xmath026.46 & 1.1 & @xmath01.60 & 0.43 & @xmath00.65 & 0.29 + 31 & 39.15 & @xmath03.67 & @xmath025.96 & 0.8 & @xmath02.69 & 0.30 & @xmath00.30 & 0.18 + 32 & 39.15 & @xmath019.40 & @xmath021.04 & 2.2 & @xmath02.67 & 0.37 & 0.95 & 0.44 + 33 & 39.57 & @xmath019.87 & @xmath020.95 & 3.4 & @xmath03.26 & 0 & 0.14 & 0 + 34 & 39.57 & @xmath07.18 & @xmath022.73 & 2.2 & & & & + 35 & 39.57 & @xmath019.02 & @xmath018.52 & 2.0 & & & & + 36 & 39.59 & 6.15 & @xmath01.49 & 1.8 & & & & + 37 & 39.61 & @xmath04.73 & @xmath028.23 & 0.7 & 1.49 & 0 & 0.23 & 0 + 38 & 39.74 & @xmath05.14 & @xmath023.02 & 1.0 & & & & + 39 & 40.00 & @xmath018.91 & @xmath018.58 & 1.2 & & & & + 40 & 40.01 & @xmath06.22 & @xmath020.02 & 4.7 & & & & + 41 & 40.03 & @xmath04.81 & @xmath028.14 & 1.0 & 1.59 & 0 & 0.18 & 0 + 42 & 40.03 & @xmath02.29 & @xmath026.22 & 2.0 & & & & + 43 & 40.16 & @xmath09.75 & @xmath030.09 & 1.6 & 0.68 & 0.38 & @xmath00.72 & 0.06 + 44 & 40.42 & @xmath020.47 & @xmath020.21 & 1.1 & @xmath01.49 & 0.23 & @xmath03.01 & 0.51 + 45 & 40.58 & @xmath05.73 & @xmath026.12 & 2.6 & 0.54 & 0.05 & @xmath00.26 & 0.17 + 46 & 41.00 & @xmath05.03 & @xmath031.63 & 1.1 & & & & + 47 & 41.00 & @xmath05.70 & @xmath026.20 & 5.4 & 1.32 & 1.27 & @xmath00.81 & 0.42 +
since it is difficult to identify contributions to positional errors in individual vlbi observations @xcite , we applied errors in each observation by evaluating following three error sources : ( 1)a zenith atmospheric delay offset at each station , ( 2)station position errors , and ( 3)image quality .
the best estimates of the zenith atmospheric delay offsets fell within the range of @xmath33 cm .
errors in the _ a priori _ station positions can result in additional phase residuals .
therefore , we assumed the offset of 3 cm even after the atmospheric offset calibration was done . for a given separation of zenith angles between two sources @xmath42 , the difference of signal path lengths @xmath43 between the two directions caused by the zenith atmospheric delay residual @xmath44
can be estimated as follows , @xmath45 where @xmath46 is the mean zenith angle of the observed sources .
we adopted 1.23@xmath1 as the approximation of @xmath42 . during our observations , the elevation angles higher than 30@xmath1 were dominant .
we assumed @xmath47 as the zenith angle for estimating errors of the worst case , and obtained @xmath48 0.12 cm . using a root sum squares of major and minor axes of the synthesized beam(@xmath49 = 1.7mas ) , we estimated a positional error of @xmath50 156@xmath5as from the following expression , @xmath51 where @xmath52 is the wavelength corresponding to the rest frequency of water masers . since the position accuracy of antennas are determined to be @xmath83 mm based on geodetic observations at s and x bands , we estimated the positional error(@xmath53 ) attributed to the baseline errors to be 5@xmath5as using the following expression , @xmath54 where @xmath55 is the separation angle of the source pair .
in addition , we estimated the measurement errors ( @xmath56 ) which depends on the snrs of the phase - referenced images
. finally , we obtained the position errors of each phase - referenced image by adding up quadratically the error factors , @xmath57 , and @xmath58 .
the error bars in figure 3 represent the errors in each observation , and the averaged value of the errors is 167@xmath5as .
figure [ fig - map ] shows the internal motions of the masers pots around s crt found in the present observations . since the maser spots show a bimodal distribution about radial velocities and positions , we extract the kinematic essentials .
the analytic tools based on the diagonalization of the velocity variance - covariance matrix ( vvcm ) @xcite and position variance - covariance matrix are used .
the elements of the vvcm matrix are presented as @xmath59 where the diagonal elements are the velocity dispersion . here
@xmath60 and @xmath61 denote three orthogonal space axes ( e.g. , r.a . ,
dec . , and radial coordinate @xmath62 )
, @xmath63 denotes the @xmath63th maser spot in a collection totalling n(@xmath1326 ) , and the bar indicates averaging over maser spots .
the vector @xmath64 ( figure [ fig - map ] ) , which is a two - dimensional projections of the eigenvector for the largest eigenvalue , is the major principal axis of the vvcm .
this axis of maximum internal velocity dispersion is the outflow axis .
the axis of @xmath64 lies at a p.a .
of @xmath65 , and also make an inclination angle @xmath66 with the plane of the sky , with the southwest lobe directed into the page ( away from the observer ) .
the vvcm was initially evaluated in a coordinate system with axes r.a .
, dec . , and radial coordinate z increasing toward the source , and then it was diagonalized .
we also considered the two - dimensional variance - covariance matrix to quantify the angular distribution of the maser spots .
here , the number of maser spots is 26 , same as the case of vvcm . in the diagonalization of this matrix , one eigenvalue which is larger than the other by a factor of 3.1
was obtained .
the corresponding eigenvector @xmath67 presented in figure [ fig - map ] indicate the spatial elongation of the maser distribution .
the position angle of this axis is found to be @xmath68 .
although maser velocity and position axes show a discrepancy of @xmath69 , the directions of these model - independent axes indicate the bipolar outflow with the major axis along northeast - southwest .
this dynamical property is not revealed by previous observation @xcite .
we estimate the stellar position from the present result by using the kinematic information .
since the motion of each maser spot are powered by the forming star , its position can be estimated as an origin of these motions .
here , we take one of the simplest kinematic models , which assumes a linear motion with a constant velocities from a single origin .
this technique is essentially same as that used in @xcite and @xcite .
based on this modeling , we obtained the origin of the maser motions which most effectively explain the observed positions and velocities . the stellar position from this kinematic analysis ,
the kinematic center was obtained to be ( @xmath37 , @xmath38 ) @xmath13 ( @[email protected] , @[email protected] ) , which is indicated with the cross with error bars in figure [ fig - map ] .
we used the first order moment with respect to the determined stellar position in order to find a typical distance of maser spots .
the obtained distance was [email protected] , and presented with dashed line in figure [ fig - map ] .
at the distance of 430pc , the corresponding linear length is 6.11au .
since we estimated the inclination angle of the outflow axis , this projected length can be converted to the linear radius of 8.96au .
nineteen years ago , the angular distribution of the masers was firstly observed with the nrao very large array in its a configuration @xcite .
the masers were uniformly distributed over a 40mas@xmath7040mas region with the lsr velocities ranging evenly from 33.4 to 44.2kms@xmath2 ( figure 13 of @xcite ) . in their study , the blue-(red-)shifted spots were found in the northeast(southwest ) of the maser region . in addition , masers with intermediate velocities were found in the central region .
although the extent of the distribution is consistent with our result , we did not detect the masers in central region . in figure
[ fig - map ] , the maser spots are mainly found at the peripheral region of the distribution . during the period of 2006.3
2007.0 , we could not find the emission with intermediate velocities(see figure [ fig - spectr](b ) ) , and it can explain the the absence of intermediate velocity maser spots . on the other hand , at the beginning and end of our vlbi observations , the masers with intermediate velocities are found in total power spectrum , and they distributed in the peripheral region . as @xcite did not precisely mention the flux densities and sizes of each spot , it is difficult to specify the reason for our non - detection of the masers in central region .
however , we can speculate as to the reason ; ( 1)our image sensitivities were not high enough to detect them .
the rms noise 38mjybeam@xmath2 of @xcite was lower than that of the present vlbi observations , 90mjybeam@xmath2 .
( 2)the masers in central region have faded before the present vlbi observations .
( 3)the masers were extended and diffuse , therefore they were resolved out with the synthesized beam of vera .
for example , we find the effect of resolve - out ( e.g. @xmath7130% in 2006 march ) from the comparison between cross - power spectrum ( figure [ fig - spectr ] ) and total - power spectrum ( figure [ lightcurves](b ) ) .
we extract the stellar properties of s crt using the distance .
@xcite determined the temperature of photosphere ( @xmath72 ) in s crt to be 3097@xmath3100k by fitting two blackbodies to its infrared spectrum .
using this temperature , we estimated an acceptable temperature range of 2600 3500k . here
we assume that the light variation of 0.8mag ( figure [ lightcurves](a ) ) is same in the infrared and caused only by the temperature variation .
@xcite reported that mira variables show @xmath73 2418 2902k and srvs shows @xmath73 3005 3500k .
the estimated temperature range of s crt is consistent with their finding .
it is worthwhile estimating the stellar radius of s crt , since there is no direct measurement . using the apparent magnitudes in infrared(j , h , and k ) , @xmath72(3097k ) and the distance(430pc )
, we estimated the radius of s crt photosphere ( @xmath74 ) to be @xmath75cm(260@xmath320@xmath76 ) .
this is presented with a shaded - circle in figure [ fig - map ] , and the center of circle is placed on the kinematic center . using this radius , we estimated an acceptable radius range of 213 309@xmath76 , under the assumption that the light variation is caused , now , only by the radius variation .
in the studies of @xcite and @xcite , @xmath74s of dozen of mira variables are presented , and most @xmath74/@xmath76 is larger than 300 . the photospheric radius of s crt is close to the lower limits to those of mira variables in the literature .
since the estimates of @xmath74 and @xmath72 are now available , the luminosity ( @xmath77 ) of s crt is estimated to be 2.29@xmath910@xmath78[w ] , resulting in the ratio @xmath79/@xmath80 of 5970 . a similar analysis of data for
a larger number of mira and srvs would be very useful for a better understanding of the difference and similarities between mira variables and srvs .
we have used vlbi monitoring observations with vera to measure the absolute proper motions of a maser spot in s crt with respect to the quasar j1147@xmath00724 .
we obtained an annual parallax of [email protected] corresponding to a distance of 430@xmath4pc .
nearly two years astrometric observations revealed the internal motions of 26 maser spots for the first time .
taking into account the internal motions of the reference maser spot , the absolute proper motion of s crt system was obtained to be ( @xmath32masyr@xmath2 , @xmath33masyr@xmath2 ) , which is consistent with the result in the new hipparcos catalog .
the three - dimensional velocity field of the maser spots in s crt was also detected , clearly outlining a bipolar outflow with the major axis along northeast - southwest .
the inclination of the axis with respect to the sky plane was @xmath66 .
the angular size of the maser distribution can be converted to the linear maser distribution radius of 8.96au . using the photospheric temperature of 3097k @xcite
, the @xmath6 band apparent magnitude @xmath810.79 , and the distance of 430pc , the linear radius of s crt photosphere ( @xmath74 ) was estimated to be @xmath75cm(260@xmath320@xmath76 ) .
this radius is comparable with the lower limits to those of mira variables in literatures .
antipin , s. v. , pastukhova , e. n. , & samus , n. n. 2005 , informational bulletin on variable stars , 5613 , 1 aringer , b. , hfner , s. , wiedemann , g. , hron , j. , jrgensen , u. g. , kufl , h. u. , & windsteig , w. 1999 , , 342 , 799 benson , p. j. , little - marenin , i. r. , woods , t. c. , attridge , j. m. , blais , k. a. , rudolph , d. b. , rubiera , m. e. , & keefe , h. l. 1990 , , 74 , 911 bloemhof , e. e. 2000 , , 533 , 893 bowers , p. f. , & johnston , k. j. 1994 , , 92 , 189 cutri , r. m. , et al .
2003 , the irsa 2mass all - sky point source catalog , nasa / ipac infrared science archive .
http://irsa.ipac.caltech.edu/applications/gator/ , hachisuka , k. , et al . 2006 , , 645 , 337 haniff , c. a. , scholz , m. , & tuthill , p. g. 1995 , , 276 , 640 hirota , t. , et al .
2007 , , 59 , 897 hirota , t. , et al .
2008 , , 60 , 37 honma , m. , et al .
2007 , , 59 , 889 imai , h. , kameya , o. , sasao , t. , miyoshi , m. , deguchi , s. , horiuchi , s. , & asaki , y. 2000 , , 538 , 751 imai , h. , et al .
2007 , , 59 , in press jike , t. , fukuzaki , y. , shibuya , k. , doi , k. , manabe , s. , jauncey , d.l .
nicolson , g.d . &
mcculloch , p.m. 2005 , polar geoscience , 18 , 26 kawaguchi , n. , sasao , t. , & manabe , s. 2000 , , 4015 , 544 kleinmann , s. g. , sargent , d. g. , & dickinson , d. f. 1978 , , 83 , 1206 kobayashi , h. , et al .
2003 , asp conference series , 306 , 48p kurayama , t. , sasao , t. , & kobayashi , h. 2005 , , 627 , l49 ma , c. , et al .
1998 , , 116 , 516 manabe , s. , yokoyama , k. & sakai , s. 1991 , iers techn .
note , 8 , 61 patel , n. a. , joseph , a. , & ganesan , r. 1992 , journal of astrophysics and astronomy , 13 , 241 perryman , m. a. c. , et al .
1997 , , 323 , l49 petrov , l. , hirota , t. , honma , m. , shibata , k. m. , jike , t. , & kobayashi , h. 2007 , , 133 , 2487 pojmanski , g. 1997 , acta astronomica , 47 , 467 sato , m. , et al .
2007 , , 59 , 743 shibata , k. m. , kameno , s. , inoue , m. , & kobayashi , h. 1998 , iau colloq . 164 : radio emission from galactic and extragalactic compact sources , 144 , 413 shimoikura , t. , kobayashi , h. , omodaka , t. , diamond , p. j. , matveyenko , l. i. , & fujisawa , k. 2005 , , 634 , 459 shintani , m. , et al . 2008 , submitted to .
van leeuwen , f. , feast , m. w. , whitelock , p. a. , & yudin , b. 1997 , , 287 , 955 van leeuwen , f. 2007 , hipparcos , the new reduction of the raw data . by floor van leeuwen , institute of astronomy , cambridge university , cambridge , uk
series : astrophysics and space science library vlemmings , w. h. t. , van langevelde , h. j. , diamond , p. j. , habing , h. j. , & schilizzi , r. t. 2003 , , 407 , 213 whitelock , p. a. , feast , m. w. , & van leeuwen , f. 2008 , , 386 , 313 | we present a distance measurement for the semiregular variable s crateris ( s crt ) based on its annual parallax . with the unique dual beam system of the vlbi exploration for radio astrometry ( vera ) telescopes ,
we measured the absolute proper motion of a water maser spot associated with s crt , referred to the quasar j1147@xmath00724 located at an angular separation of 1.23@xmath1 . in observations spanning nearly two years , we have detected the maser spot at the lsr velocity of 34.7kms@xmath2 , for which we measured the annual parallax of [email protected] corresponding to a distance of 430@xmath4pc .
this measurement has an accuracy one order of magnitude better than the parallax measurements of hipparcos .
the angular distribution and three - dimensional velocity field of maser spots indicate a bipolar outflow with the flow axis along northeast - southwest direction . using the distance and photospheric temperature
, we estimate the stellar radius of s crt and compare it with those of mira variables . |
the time evolution of the intrinsic continuum v - band polarization ( v - pol ) of @xmath0 aqr is shown in figure 1 .
we find evidence of clockwise loop - like structures ( figure 2 ) when comparing the evolution of the polarization across the balmer jump ( bj ) vs v - pol , particularly during polarimetric outburst events ( red in figure 1 ) .
60 cyg also displays this behavior .
we use the non - lte 3d monte carlo code developed by @xcite , hdust , to investigate the origin of the loop - like behavior of the bj vs v - pol during polarimetric outbursts .
interestingly , we find that the clockwise loop structures can be reproduced when the mass - loss from the central star which feeds the disk is turned on ( 6 to 12 o@xmath1clock , figure 3 ) then off ( 12 to 6 o@xmath1clock , figure 3 ) .
we therefore suggest that this diagnostic can provide insight into the time dependence of the density of the innermost disk region .
counter - clockwise loops are also observed in @xmath0 aqr and 60 cyg and will require further modeling to ascertain thier origin .
we note that @xcite found similar loop structures in cmd diagrams of be stars .
in ( * ? ? ? * paper i ) , we detected variations in the pa of disks during polarimetric outbursts , and speculated that these variations could be indicative of warps , non - equatorial blob injections , or spiral density waves in the inner disk .
figure 4 depicts deviations in the mean pa during polarimetric outbursts of @xmath0 aqr .
as shown in figure 5 , we model a density enhancement on one side ( orange ) of the disk and a decrement on the opposing side ( black ) using hdust .
the result of computing the polarization from different viewing angles in a counter - clockwise motion are shown in figure 6 .
this model suggests that asymeteric material being ejected from the star can potentially match observed deviations in the pa . | in @xcite , paper i , we analyzed 15 years of spectroscopic and spectropolarimetric data from the ritter and pine bluff observatories of 2 be stars , 60 cygni and @xmath0 aquarii , when a transition from be to b star occurred .
here we anaylize the intrinsic polarization , where we observe loop - like structures caused by the rise and fall of the polarization balmer jump and continuum v - band polarization being mismatched temporaly with polarimetric outbursts .
we also see polarization angle deviations from the mean , reported in paper i , which may be indicative of warps in the disk , blobs injected at an inclined orbit , or spiral density waves .
we show our ongoing efforts to model time dependent behavior of the disk to constrain the phenomena , using 3d monte carlo radiative transfer codes . |
graphene is recognized as the only two - dimensional crystal that withstands large deformations and high temperatures and is readily integrated into electronic circuits.@xcite the high electron mobility and unique spectral characteristics single out graphene as a promising key element for electronic and optoelectronic applications.@xcite the well - known dirac dispersion of electronic excitations in graphene is responsible for a number of interesting analogies with relativistic quantum electrodynamics,@xcite but prevents the use of graphene in field - effect transistors due to the absence of energy band - gaps.@xcite the strain - engineering of graphene circuits has been recently suggested as a means to bypass this difficulty.@xcite in the dirac - fermion picture of graphene , the strain gradient is equivalent to the presence of a pseudo - magnetic field , which can be manipulated to induce a zero - field quantum hall effect and a topological isolator state.@xcite the variation of strain may also induce substantial variation in the on - site electron energies , however , this effect is reduced by screening.@xcite the experimental realizations of strained graphene confirm its surprisingly high elasticity and pave the way for the development of strain - engineered graphene electronics.@xcite different techniques have been proposed to produce graphene samples with controlled periodical variations of strain.@xcite first transport measurements of graphene superstructures have been already reported.@xcite below we analyze the phase - coherent charge transport in periodically strained graphene samples and propose a way to characterize the quality of graphene superstructures on the basis of their transport properties .
more specifically we use the exact solution of the dirac - kronig - penney model to calculate density of states , conductance , and shot noise in transport through finite size graphene samples with periodic potentials . the scattering off the metal leads is taken into account in all quantities .
the position , the width , and the shape of the conductance minimum as well as the shot noise maximum associated with the periodic superstructure are analyzed in detail . and
the angle @xmath4 specifies the orientation of the honeycomb lattice with respect to the transport direction @xmath5 . ]
this paper is organized as follows . in section [ sec : model ] the effective dirac hamiltonian for deformed graphene
is derived .
the scattering approach to transport is briefly described in section [ sec : scattering ] .
section [ sec : transport ] is devoted to the exact solution of the dirac - kronig - penney model for graphene with one - dimensional modulations of strain in transport direction .
the generalization of this model is discussed at the end of the section [ sec : transport ] .
the results for conductance , shot - noise , and the density of states are qualitatively understood on the basis of the band structure analysis , which is presented in appendix [ app : band ] . in section [ sec:2d ]
we discuss transport in the direction perpendicular to the strain modulation .
we summarize our results in the concluding section [ sec : conclusion ] .
the electromechanical coupling in deformed graphene membranes has been investigated theoretically by many authors@xcite following earlier publications on carbon nanotubes.@xcite the deformations affect the hopping integrals in the tight - binding description of graphene in two distinct ways : by changing the distance between carbon atoms and by tilting the electronic @xmath6-orbitals , which are responsible for conduction . however , for most cases of interest , the in - plane strains play the major role in determining the electronic properties of deformed graphene , while the tilting can be neglected . the effects of in - plane strain are well captured by the tight - binding hamiltonian @xmath7 \left ( a^\dagger_{\boldsymbol{r } + \boldsymbol{\delta}_\alpha } b_{\boldsymbol{r } } + \text{h.c . }
\right),\end{aligned}\ ] ] where the summation runs over the atomic positions , @xmath8 , of a honeycomb lattice , the annihilation operators @xmath9 and @xmath10 refer to the occupation of @xmath6-orbitals at the two non - equivalent positions , @xmath11 and @xmath12 , of the unit cell , and the three vectors @xmath13 shown in fig .
[ fig : theta ] are directed from a @xmath12-atom to its three nearest neighbors .
if the strain varies smoothly on atomic distances , the deviation of the hopping integral from its unperturbed value @xmath14ev in a perfect crystal can be parameterized as @xmath15 where @xmath16 and @xmath17 is the strain tensor of the graphene membrane .
the dimensionless strain tensor describes a change in a metric that can be expressed as @xmath18 where the summation over the spatial indices @xmath19 is assumed .
the length elements @xmath20 and @xmath21 correspond to the metric @xmath22 in the flat space and to the local metric @xmath23 in the membrane , respectively . in subsequent formulas
we let conventionally @xmath24 and @xmath25 .
the graphene crystal withstands very large internal strains so that the values of the strain tensor elements may reach @xmath26%.@xcite even though the strain tensor makes no reference to the crystal structure of graphene , the lattice symmetry is entering the hopping integral in eq .
( [ dt ] ) due to the vectors @xmath13 , @xmath27 .
the slow variation of @xmath17 on the scale of the lattice spacing justifies the effective mass approximation , which is formulated in a generic reference frame shown in fig .
[ fig : theta ] . in this frame
we find @xmath28 while the positions of the two non - equivalent dirac points in the reciprocal space are given by @xmath29 using the fourier ansatz @xmath30 we obtain , to the leading order in spatial gradients , the effective model @xmath31\ , \psi_{\boldsymbol{r}},\ ] ] where @xmath32 , @xmath33 is the pauli matrix in the valley space and the operators are arranged into the four - spinor @xmath34 the vector field @xmath35 is real and its components are related to the strain tensor in a simple way , @xmath36 in the derivation of the effective hamiltonian ( [ ham0 ] ) we neglected the velocity renormalization , which would appear as a correction of the order @xmath37 to the prefactor of the spatial gradients.@xcite unlike the usual vector potential , the vector field @xmath38 preserves the time reversal symmetry of the hamiltonian ( [ ham0 ] )
. it also exhibits the discrete rotational invariance of the honeycomb lattice .
indeed , the hamiltonian ( [ ham0 ] ) remains invariant under the rotation through the angle @xmath39 , while the rotation through the angle @xmath40 is equivalent to an interchange of valleys .
the effect of a constant uniaxial strain has been studied both theoretically@xcite and experimentally@xcite and falls beyond the scope of our consideration .
if the constant uniaxial strain ( in @xmath5 direction ) exceeds a certain critical value , the dirac points merge and a band gap opens .
it is predicted@xcite that , for a crystal expanded uniformly in the zigzag direction ( @xmath41 ) , the critical expansion takes on its minimal value ( @xmath42 ) , while uniaxial strain in armchair direction , @xmath43 , never generates a gap .
these predictions await experimental verification .
the strain also induces an electrostatic potential due to a change of the on - site energies of the tight - binding model .
this effect leads to the appearance of a scalar electrostatic potential , which has to be added to the effective hamiltonian ( [ ham0 ] ) .
in this section we formulate the scattering approach to transport through a deformed graphene sample with metallic leads . we take advantage of the effective single - particle hamiltonian @xmath44 where the fictitious vector potential @xmath38 is related to the strain tensor , @xmath17 , by means of the relation ( [ relation ] ) , and the scalar field @xmath45 describes strain - induced and external electrostatic potentials . in most of the intermediate expressions
we let @xmath46 for simplicity .
the metal leads are modeled by letting @xmath47 for @xmath48 and @xmath3,@xcite in eq .
( [ dirac ] ) .
the vector potential is assumed to be zero in the leads .
the width of the sample in @xmath49 direction is denoted as @xmath50 . the scattering off the metal leads are fully taken into account in the subsequent analysis .
the rectangular sample geometry makes it convenient to employ the fourier transform in @xmath49 , @xmath51 where @xmath52 is the quasiparticle momentum component parallel to the graphene - metal interface .
the momentum takes on the quantized values , @xmath53 , which depend on the boundary conditions in @xmath49 direction , for example , @xmath54 with integer @xmath55 for periodic boundary conditions . in the limit @xmath56 , which we assume below
, the particular type of the boundary conditions is not important .
we also restrict our consideration to small energies , @xmath57 , where the energy @xmath58 is measured with respect to the dirac point . in this approximation
we derive the equation on the transfer matrix in the form@xcite @xmath59{\cal m},\ ] ] where we introduced matrix notation in fourier ( channel ) space , e.g. @xmath60 and @xmath61 .
the transfer matrix has the following structure in @xmath62-space @xmath63 where @xmath64(@xmath65 ) for @xmath66 is the matrix of reflection amplitudes for quasiparticles entering the sample from the left(right ) lead .
the matrices @xmath67 and @xmath68 contain the corresponding transmission amplitudes .
then , the landauer formula for the conductance can be cast in the following form @xmath69^{-1},\ ] ] where the trace is referred to the valley and channel space and the symbol @xmath70 stands for the @xmath71 block of the transfer matrix in @xmath62-space .
the hamiltonian ( [ dirac ] ) with vanishing electrostatic potential , @xmath72 , obeys the chiral symmetry @xmath73 , which is responsible for a non - abelian aharonov - casher gauge invariance at zero energy,@xcite @xmath74 the spatially dependent phases @xmath75 and @xmath76 can be chosen in such a way that the zero - energy spectral equation @xmath77 is reduced to the dirac equation , @xmath78 , with zero vector potential .
this gauge transformation can be applied in the scattering approach in order to demonstrate that the presence of arbitrary vector potential has no effect on charge transport at the dirac point as far as the contribution from edge states can be disregarded.@xcite
one - dimensional modulations of strain were realized experimentally in suspended graphene films using the remarkably large and negative thermal expansion of graphene.@xcite motivated by these experiments we calculate the transport properties of the one - dimensional dirac - kronig - penney model , which has been introduced earlier by several authors.@xcite below we consider a general form of the model , where the variation of both electrostatic as well as vector potentials is included .
using the scattering approach formulated in the section [ sec : scattering ] we provide simple analytical solutions for transport and density of states in this model , which complement previous theoretical studies of dirac fermions in periodic potentials.@xcite a graphene sample with one - dimensional modulation of strain is characterized only by the component @xmath79 of the strain tensor , which depends solely on the @xmath5-coordinate . from eq .
( [ relation ] ) , the components of the pseudo - vector potential take the form @xmath80 in addition , the strain induces a spatial variation of electrostatic potential , @xmath81 .
its relation to the strain tensor is , however , complicated by screening effects . since the potentials depend only on @xmath5 , the transverse momentum @xmath52 is conserved . in this case
the matrices @xmath82 , @xmath83 and @xmath84 in eq . ( [ meq0 ] )
are diagonal in channel space .
since the considered potentials also do not couple the valleys , the scattering problem is reduced to the solution of @xmath85 matrix equation , @xmath86{\cal m}.\ ] ] the @xmath5-component of the vector potential , @xmath87 , enters the equation in a trivial way and can be excluded by the gauge transformation @xmath88 , which does not affect any observable .
the first and the most trivial consequence of this transformation is that the one - dimensional strain modulations in @xmath5 direction have no effect on transport in the zigzag direction ( @xmath89 with integer @xmath55 ) provided the effect of @xmath90 is ignored . for other angles the one - dimensional strain modulations
lead to the appearance of new minima in the gate voltage dependence of the conductivity of the graphene sample .
the numerical solution of eq .
( [ meq ] ) suggests that a periodic vector potential induces much more pronounced minima than a periodic scalar potential of equivalent amplitude .
the weaker effect of @xmath90 can be associated with klein tunnelling , which leads to the suppression of pseudo - gaps in the spectrum of the superstructure . in order to describe the effect of periodic potentials
analytically we introduce the one - dimensional dirac - kronig - penney model , which is characterized by the vector potential @xmath91 given by eq .
( [ afromu ] ) with @xmath92 where the scale @xmath2 stands for half the period of the superlattice , @xmath93 is the heaviside step function , and the dimensionless parameter @xmath94 specifies the amplitude of the strain modulation .
the vector potential introduced by eqs .
( [ afromu],[kronig_u ] ) corresponds to a strain field which is smooth on atomic scale but changes abruptly on distances smaller than the dirac quasiparticle wave length , @xmath95 .
the periodic electrostatic potential is introduced in a similar manner , @xmath96 where @xmath97 is the amplitude of the modulation . ) for finite systems comprising @xmath98 and @xmath99 supercells are compared with the averaged expressions to be derived in eqs .
( [ gav],[fav ] ) , which correspond to the limit @xmath100 and ignore contributions from evanescent modes .
the plots are for graphene with strain modulations , @xmath101 . in a small vicinity of the dirac point , @xmath102 , ( shown at the insets ) the transport is insensitive to strain due to the extended gauge invariance ( [ gauge ] ) and
is dominated by the evanescent modes . ] for the sake of simplicity we consider a sample of length @xmath103 with @xmath104 being an integer . in this case
the solution to eq .
( [ meq ] ) satisfies @xmath105 , where @xmath106 is the transfer matrix corresponding to the size of the supercell , @xmath107 . disregarding the @xmath5-component of the vector potential , @xmath87 , which has been argued above to have no effect on transport
, we write @xmath108 where we take advantage of the definitions @xmath109 the eigenvalues of the transfer matrix , @xmath106 , are conveniently parameterized as @xmath110 where we introduced the real function @xmath111 with @xmath112 the @xmath5-component of the momenta .
the wave number @xmath113 becomes imaginary for some values of @xmath52 and @xmath58 indicating the appearance of pseudo - gaps in the superstructure . in the considered dirac - kronig - penney model
the value of @xmath113 is nothing but the @xmath5-component of the quasi - momentum in the reciprocal space associated with the superstructure . in order to calculate the element @xmath114 of the full transfer - matrix @xmath115 we use the chebyshev identity to calculate the @xmath104-th power of an unimodular matrix@xcite @xmath116 the matrix element @xmath117
is readily determined from eqs .
( [ m0],[chebyshev ] ) with @xmath118 using the fact that both @xmath119 and @xmath120 are real functions of the energy @xmath58 , and the conserved momentum component @xmath52 , we obtain from eqs .
( [ mmatrix],[chebyshev ] ) the exact transmission probabilities for each value of @xmath52 , @xmath121^{-1}.\ ] ] this expression is reduced to the well - known result for a purely ballistic system,@xcite i.e. for @xmath122 , by the substitution @xmath123 and @xmath124 .
the transmission probabilities ( [ exact ] ) determine the energy - dependent transport quantities @xmath125 where @xmath126 is the landauer conductance and @xmath127 is the fano factor for the shot noise . in the limit
@xmath56 the summation over the scattering channels can be replaced by the integration , @xmath128 .
, calculated from eq .
( [ tav ] ) for the dirac - kronig - penney model with strain modulation @xmath129 ( top left ) , @xmath130 ( bottom left ) and electrostatic field modulation @xmath131 ( top right ) , @xmath132 ( bottom right ) . ]
more generally one can define the cumulant generating function for the transport , @xmath133 the dimensionless cumulants @xmath134 determine the so - called full counting statistics of the charge transport .
the conductance and the fano factor are given by @xmath135 and @xmath136 , respectively .
the strength of strain - induced and electrostatic potentials in the dirac - kronig - penney model ( [ kronig_u],[kronig_v ] ) is characterized by the dimensionless parameters @xmath137 respectively . in fig .
[ fig : ncompare ] the conductance and the fano factor in graphene with periodic modulations of strain calculated from eq . are plotted for systems with finite lengths @xmath138 and @xmath139 for @xmath140 .
the energy dependence of conductance and fano - factor ( [ gexact ] ) reveal fast fabry - prot oscillations on the scale @xmath141 . from a physics point of view , these oscillations originate from multiple reflections of propagating modes ( real @xmath113 ) at the metal - graphene interfaces .
the channels with imaginary values of @xmath113 ( evanescent modes or metal - induced states ) also contribute to transport .
even though the individual contribution of each evanescent mode is exponentially small , @xmath142 , their combined effect becomes essential for energies in a vicinity of band - edges .
the role of evanescent modes is especially important at the dirac point due to the absence of propagating modes .
indeed , the conductance and the fano - factor determined by ( [ gexact ] ) take on the values @xmath143 and @xmath144 for @xmath145 irrespective of the vector potential .
this universality is due to the extended gauge invariance ( [ gauge ] ) .
, [ fav ] ) for the one - dimensional dirac - kronig - penney model with different field strengths ( [ dimensionless ] ) .
the effect of the scalar potential is suppressed due to klein tunnelling through potential barriers . ] for @xmath146 , the amplitude of the fabry - prot oscillations in @xmath126 and @xmath127 decreases ( provided @xmath56 ) and the relative contribution of evanescent modes becomes less important . taking the limit @xmath100 is equivalent to ignoring the imaginary values of @xmath113 and averaging over the rapid phase , @xmath147 .
this approximation has been used e.g. in ref . to obtain the full counting statistics of few - layer graphene .
the averaged generating function ( [ full ] ) takes the form @xmath148 where we introduce the mean transmission probability @xmath149 the averaged transport quantities do not depend on the system size @xmath150 and reveal no fabry - prot oscillations . for a ballistic sample , @xmath122 , one finds @xmath151 , where @xmath152 is the @xmath5-component of the momentum . )
( solid ) is compared with an estimate ( [ modecounting ] ) based on the mode - counting argument ( dashed ) for the dirac - kronig - penney model with @xmath153 .
the upper panel shows the corresponding transmission spectrum . ] from the generating function ( [ fullav ] ) we readily find the averaged conductance and noise .
the conductance is given by the landauer formula , @xmath154 while the fano - factor is related to the averaged transmission probabilities in a less evident way @xmath155 the exact and averaged transmission probabilities ( [ exact],[tav ] ) together with the corresponding expressions for the full counting statistics ( [ full],[fav ] ) provide the complete analytical description of the transport properties in the 1d - dirac - kronig - penney model with scalar and vector potentials of arbitrary strength .
the dependence of the transmission coefficient on the transversal momentum component , @xmath52 , is called the transmission spectrum . in fig .
[ fig : tdensity ] we plot the transmission spectra obtained from eq .
( [ tav ] ) for different strengths of the scalar and vector potentials . ) for the dirac - kronig - penney model with different strengths of the periodic potentials . ] the transmission spectra shown in fig .
[ fig : tdensity ] are qualitatively different for weak , @xmath156 , and strong , @xmath157 , strain modulations .
it has to be stressed that the latter regime is well within the experimentally accessible range of parameters . indeed , for an achievable strain modulation of @xmath158% ( @xmath159 ) and the period @xmath160 nm one finds @xmath161 .
neither the modulated strain nor the electrostatic field does open up a full band gap for any values of the parameters @xmath162 or @xmath163 .
nevertheless , very large pseudo - gaps are generated by strain modulations with @xmath164 .
one can see from the lower left panel in fig .
[ fig : tdensity ] that the charge transport in this regime is suppressed for almost all directions of the momenta ( different @xmath52 values ) in a wide energy range .
in contrast , the pseudo - gaps do not open at normal incidence ( @xmath165 ) in the dirac - kronig - penney model with periodic electrostatic potential due to the klein - tunnelling of quasiparticles through electrostatic barriers.@xcite the effects of strain and electrostatic field modulations on the conductance and noise are compared in fig .
[ fig : gfstrength ] .
the most prominent feature of these plots are the dips in conductivity centered around the energies @xmath166 with @xmath167 . the physical mechanism , which is responsible for the appearance of the pseudo - gaps , is equivalent to that of the parametric resonance.@xcite for @xmath164
, the conductance is suppressed in wider energy intervals around @xmath168 . for weak strain or potential amplitude , @xmath156 or @xmath169 ,
significant pseudo - gaps are only located around the energy values @xmath170 with odd values of @xmath55 , since only odd fourier components of the potentials ( [ kronig_u],[kronig_v ] ) exist . for a weak harmonic potential ,
a significant pseudo - gap would only arise around @xmath171 .
the areas of vanishing transmission probability in fig .
[ fig : tdensity ] ( top left ) can be understood on the basis of perturbation theory as presented in appendix [ app : band ] for small amplitudes of periodic potentials .
the pseudo - gap emerging due to a weak periodic strain is described by the functions @xmath172 and @xmath173 given in the first row of tab .
[ tab : gaps ] .
these functions are shown with dashed lines in the upper panel of fig .
[ fig : modecounting ] . inside the pseudo - gap , i.e. for @xmath174 , the transmission coefficient is exponentially small , while it is close to the ballistic value , @xmath175 , outside the pseudo - gap ( one can disregard the change in the resistance of the metal - graphene interface due to the weak periodic potentials ) .
note further , that the pseudo - gap is located near @xmath165 for energies near the dip such that @xmath176 in this region .
we can , therefore , estimate the conductance in the vicinity of the lowest dip ( @xmath171 ) by subtracting the contribution of ballistically propagating modes inside the pseudo - gap from the ballistic result as @xmath177 where @xmath93 is the heaviside step function , @xmath178 , and the limit @xmath179 is assumed .
the generalisation of eq . around the higher resonant energies , @xmath180 ,
is straightforward .
it is shown in fig .
[ fig : modecounting ] for @xmath181 that the result of eq .
( [ modecounting ] ) agrees with the averaged conductance calculated from eq .
( [ gav ] ) . the band - structure analysis ( [ modecounting],[secular ] )
predicts characteristic dips in the conductance at @xmath182 of the depth @xmath183 , where @xmath184 stands for @xmath55-th fourier - component of the vector potential . for the dirac - kronig - penney model one finds @xmath185 hence the value of the conductance dip does not depend on @xmath55 .
we note that the validity of eq .
( [ modecounting ] ) is restricted to weak potentials . for @xmath186 , the pseudo - gaps overlap and
the perturbation theory of appendix [ app : band ] is no longer applicable .
let us now extend the solution of the dirac - kronig - penney model to the density of states . for purely ballistic system with metal leads , both the local and the integrated density has been found in ref . using a green s function approach . below
we take advantage of an alternative route and relate the partial density of states in the channel @xmath52 to the corresponding transmission amplitude , @xmath187 , by the well - known formula@xcite @xmath188.\ ] ] then , the integrated density of states per unit volume is given by @xmath189 where the factor @xmath190 takes into account the spin and valley degeneracy .
the transmission amplitude is readily obtained from eqs .
( [ mmatrix],[chebyshev ] ) as @xmath191 where the wave - number @xmath113 and the real quantity @xmath120 are determined from the expressions ( [ k0 ] ) and ( [ eta ] ) , respectively .
therefore , the partial density of states can be calculated exactly as @xmath192 unlike the conductance or the shot noise ( [ gexact ] ) , the density of states depends on the value of @xmath193 , because the spectrum in a vicinity of the band - edges is dominated by the metal - induced states ( evanescent modes ) .
the metal proximity effect@xcite can be seen already for purely ballistic system near the dirac point .
indeed , for @xmath122 , one finds @xmath194 where @xmath152 . at zero energy the ballistic result ( [ ball ] ) is reduced to @xmath195 .
therefore , the density of states at @xmath196 in eq .
( [ idos ] ) acquires a logarithmic divergency .
this divergency is regularized by the largest available transversal momentum @xmath197 , which is simply equal to the fermi - momentum in the metal lead , @xmath198 .
thus , the density of states in a close vicinity of the dirac point , @xmath145 , is given by @xmath199 in agreement with ref . .
similar logarithmic dependence of the density of states on @xmath200 takes place in the dirac - kronig - penney model for energies inside the pseudo - gaps . in full analogy with the averaged generating function ( [ fullav ] ) we can introduce the averaged density of states , @xmath201 , which corresponds to the limit @xmath202 .
in this limit we disregard the contribution of the metal - induced states by projecting on the real values of the momentum @xmath113 and average the result of eq .
( [ ball ] ) over the rapid phase @xmath203 .
this procedure leads to the simple result @xmath204 hence , in the limit @xmath56 , the mean density is given by @xmath205 in the ballistic limit , @xmath122 , eq . ( [ rhoav ] )
yields the density of states of the clean graphene @xmath206 . in fig .
[ fig : dos ] we plot the averaged density of states calculated from eq .
( [ rhoav ] ) for different strengths of the periodic potentials . , [ kronig_v ] ) with strain modulation @xmath181 ( top left ) , @xmath130 ( bottom left ) and electrostatic field modulation @xmath207 ( top right ) , @xmath132 ( bottom right ) at energy @xmath208 . ]
the results of the previous subsections are readily generalized to a model with arbitrary periodic variation of strain and electrostatic potential in @xmath5 direction .
the analysis of the model is reduced to the calculation of the transfer matrix , @xmath106 , which corresponds to the wave propagation over the distance @xmath107 , the period of the potential . in this generalized model , both the exact and the averaged full counting statistics as well as the density of states
are still given by the expressions ( [ full],[fullav],[nuexact],[nuav ] ) with @xmath209 and the @xmath5-component , @xmath113 , of the quasi - momentum is related to @xmath119 by eq .
( [ k0 ] ) .
thus , for a periodic potential of a general type , the full solution of the problem is reduced to the straightforward numerical evaluation of the functions @xmath119 and @xmath120 .
note , that the exact analytical expressions ( [ lambda],[eta ] ) are restricted to the dirac - kronig - penney model and do not apply generally .
we have used eqs .
( [ num ] ) to calculate the energy - dependent conductance for different amplitudes of the periodic strain in the harmonic potential , @xmath210 . in this case
, the higher pseudo - gaps are found to be suppressed for weak potentials @xmath211 as compared to the dirac - kronig - penney model .
the lowest conductance minima around @xmath212 is essentially the same in both models .
the models become even more similar with increasing potential strength . for @xmath213
the transmission spectra of the single - harmonic model , becomes almost equivalent to those of the dirac - kronig - penney model for all energies .
was set to 1 . ]
we have demonstrated that periodic and @xmath5-dependent scalar and vector potentials modify transport in @xmath5-direction due to the appearance of pseudo - gaps .
let us now argue that such potentials hardly affect transport in @xmath49 direction , i.e. in the direction parallel to the equi - potential lines . in fig .
[ fig : fs ] we plot the fermi surface slightly above the energy value @xmath212 using the exact dispersion relation , @xmath214 , obtained from eqs .
( [ k0],[lambda ] ) for the dirac - kronig - penney model .
we have seen that the transport in @xmath5 direction is dominated by the modes with the momenta @xmath215 . for some values of @xmath52 in this interval one can not find any propagating states , which correspond to the real values of @xmath113 .
this can be seen as the formation of pseudo - gaps in the transmission spectra shown in fig .
[ fig : tdensity ] , which is especially strong for @xmath216 .
in contrast , for each real value of @xmath113 one always finds real values of @xmath52 .
therefore , the transport in @xmath49-direction is not affected by the formation of the pseudo - gaps , and depends only on the details of scattering at the metal - graphene interfaces .
the latter effect is , however , weak and falls beyond the scope of our consideration .
we therefore conclude that one - dimensional superlattices have a major impact only on the transport properties along the direction of their periodicity .
the dirac - kronig - penney model has been applied in order to study phase - coherent charge transport in periodically strained graphene samples . using the exact relations for the band - structure in the considered superlattice
, we calculate the density of states and the full counting statistics for charge transport .
the exact quantities are simplified in the thermodynamic limit by neglecting the contribution of evanescent modes and averaging over the fabry - p ' erot oscillations .
the conductance is found to be suppressed at energies corresponding to the positions of the pseudo - gaps in the modified spectra of the superlattice .
the periodic strain and electrostatic potentials are argued to have the largest effect on transport in the direction perpendicular to the equipotential lines .
the vector potentials are shown to play a greater role in confining the dirac quasi - particles due to the suppression of klein tunnelling .
we thank w .-
hannes for discussions .
s.g . acknowledges support by the epsrc scottish doctoral training centre in condensed matter physics .
w. b. acknowledges support by the dfg through sfb 767 _ controlled nanosystems _ , by the research initiative _ ultraquantum _ , and in part by the project of knowledge innovation program ( pkip ) of chinese academy of sciences , grant no .
in this appendix we perturbatively calculate the band structure of a graphene sample described by the single - valley dirac equation @xmath217 \psi(\boldsymbol r ) = \varepsilon \psi(\boldsymbol r),\ ] ] where @xmath218 is a general 1d - periodic potential that is smooth on the scale of the graphene lattice constant @xmath1 .
to do so , we generalize the standard methods of elementary band structure theory to the dirac case with its non - trivial matrix structure . we find that for purely scalar potentials , this additional structure prevents the creation of a gap at @xmath219 thus making @xmath220 and @xmath221 more relevant for the transport properties ; this fact can be related to the chirality of the quasi - particles in graphene and the phenomenon of klein tunnelling .
let @xmath222 span the 1d - super - lattice ( so that @xmath223 ) and @xmath224 be the corresponding reciprocal vector fixed by the conditions @xmath225 and @xmath226 . making the bloch ansatz @xmath227 for the wave function and writing the periodic potential as a fourier series , @xmath228 , we can write down the central equation @xmath229 where @xmath230 and @xmath231 .
the central equation represents an infinite set of linear equations for the coefficients @xmath232 , which can in general only be solved numerically .
the solution , however , simplifies for sufficiently weak periodic potentials . for @xmath233 close to a degeneracy , @xmath234
, we can neglect all but the two coefficients @xmath235 and @xmath236 in eq . .
the resulting linear system has the simple form @xmath237 here , we introduced the abbreviation @xmath238 and omitted the fourier index @xmath55 of the potentials for brevity .
the solution of eq .
is particularly compact when only one of the coefficients @xmath239 or @xmath240 is different from zero . in this case
we obtain @xmath241 where the pair of quantities @xmath242 take on one of the following values @xmath243 , @xmath244 , @xmath245 , or @xmath246 .
we compare the energy values of the forbidden zone s upper / lower boundary , @xmath247 , as well as the width of the gap in forward direction for the different types of periodic potentials in tab .
[ tab : gaps ] . to calculate @xmath247
we have chosen @xmath233 in eq .
such , that @xmath248 .
typical band gap boundaries are shown in fig .
[ fig : boundaries ] .
.energy value of the forbidden zone s upper / lower boundary , as well as the width of the gap in forward direction for the different types of periodic potentials . in the first row ,
@xmath249 was chosen parallel to @xmath250 , for the remaining rows @xmath251 . [ cols="<,^,^",options="header " , ] t. m. g. mohiuddin , a. lombardo , r. r. nair , a. bonetti , g. savini , r. jalil , n. bonini , d. m. basko , c. galiotis , n. marzari , k. s. novoselov , a. k. geim , and a. c. ferrari , phys .
b , * 79 * , 205433 ( 2009 ) . | motivated by recent proposals on strain - engineering of graphene electronic circuits we calculate conductivity , shot - noise and the density of states in periodically deformed graphene .
we provide the solution to the dirac - kronig - penney model , which describes the phase - coherent transport in clean monolayer samples with an one - dimensional modulation of the strain and the electrostatic potentials .
we compare the exact results to a qualitative band - structure analysis .
we find that periodic strains induce large pseudo - gaps and suppress charge transport in the direction of strain modulation .
the strain - induced minima in the gate - voltage dependence of the conductivity characterize the quality of graphene superstructures .
the effect is especially strong if the variation of inter - atomic distance exceeds the value @xmath0 , where @xmath1 is the lattice spacing of free graphene and @xmath2 is the period of the superlattice .
a similar effect induced by a periodic electrostatic potential is weakened due to klein tunnelling . |
bars are elongated structures frequently present in the central parts of disc galaxies . their formation is ubiquitous in n - body simulations , unless a sufficiently massive and sufficiently centrally concentrated spherical or spheroidal component has been added to stop or at least to delay their formation beyond a reasonable life - time of the disc .
the problems relating to bar formation are far from being solved , but in this review i will leave them aside in order to concentrate on the properties of the gas flow in and around the bar region . such a flow is intimately linked to the main periodic orbits and their structure ( e.g. athanassoula 1992a ; athanassoula 1992b , hereafter a92b ) .
orbital calculations ( for a review see e.g. sellwood & wilkinson 1993 ) have shown that the backbones of bars are the so - called @xmath0 periodic orbits , which are elongated along the bar . if one or two inner lindblad resonances ( hereafter ilrs ) are present , there are also two perpendicular families , called @xmath1 and @xmath2 .
further out we find the 4:1 periodic orbits , which have rectangular - like or diamond - like shapes .
these families , as well as other secondary ones , trap around them a number of regular orbits .
chaotic orbits are also present and their importance depends on the properties of the bar potential .
an example of the gas response to a barred galaxy potential , calculated with the fs2 code ( van albada , van leer & roberts 1982 ; van albada , 1985 ) , is given in fig .
the left panel gives the density response in grey - scale ; lighter shades correspond to higher densities and darker to lower ones .
the imposed bar is at 45 to the horizontal axis , rotates clock - wise and has a semi - major axis of 5 kpc and a semi - minor one of roughly 2.3 kpc .
we note that in most of the bar region the gas has very low density , while most of the gas is concentrated in two narrow strips along the leading edges of the bar . in their inner parts
these density enhancements curve into a nuclear spiral or ring - like structure .
there is also considerable gas concentrated around the extremities of the bar major axis .
the right panel shows the gas flow in the same model and in a frame of reference corotating with the bar .
the length of the vectors is proportional to the velocity at that location and i have also superposed a number of flow lines .
comparison of the two panels , as well as cuts perpendicular to the density maxima ( cf .
a92b ) , show that the density maxima are the loci of shocks .
prendergast ( 1962 , unpublished ) was the first to associate them with the dust lanes observed along the leading edges of bars .
a comprehensive study of the shape of the shock loci and of their dependence on the main parameters of the model can be found in a92b .
the shock loci in strong bars were shown to be straight , while weak bars or ovals have loci which are curved with their concave part towards the bar major axis .
one of the main results of a92b is that , in order for the shock loci to have the shape of the observed dust lanes , the corotation radius , @xmath3 , must be equal to @xmath4 , where @xmath5 is the length of the bar semi - major axis .
a further result of a92b is that , in order for shocks to exist , the curvature of the @xmath0 orbits at apocenter must exceed a certain quantity , or in other words , the orbits that are elongated along the bar should be sufficiently peaked or have loops at their apocenters .
furthermore , in order for the shock loci to be offset towards the leading side of the bar , the @xmath1 orbits must not only exist , but also have a sufficient extent . in cases with no shocks
the flow lines have simple concentric ellipse - like shapes , and there is no net inflow .
however , in cases with shocks the gas flow is considerably more complicated , as can be seen in the right panel of fig . 1 .
at the trailing sides of the bar , where the gas density is very low , the flow is outwards , and stays so until it reaches the shock . at that point
it turns abruptly inwards .
both the outwards and the inwards flow reach high values of the velocity , of the order of , or higher than , 100 km / sec .
the outwards flow is over an area which is large , but has a very low gas density , while the inwards flow is concentrated around the density maxima .
i.e. in a small area with very high gas density
. thus the net inflow , i.e. the density averaged radial velocity , is of the order of less than a km / sec , up to a few km / sec , depending on the model .
care has to be taken when comparing with observations since the low density regions are harder to observe and thus there is a tendency for observations to overestimate the inflow .
a maximum net inflow is found in models with thin , massive , slowly rotating bars , with centrally concentrated axisymmetric components ( a92b ; piner , stone & teuben 1995 , etc ) . since these models have also the strongest shocks ( a92b ) , we come to the conclusion that the most important inflow is in models with strongest shocks , as expected . in cases with ilrs
the gas is brought from regions roughly within a radius equal to the bar semi - major axis to the circumnuclear ring or pseudo - ring and little to the disc within it . on the other hand in cases with no ilr the inflow reaches the center - most area ,
i.e. very near the nucleus .
there are indications , however , that most barred galaxies have one or two ilrs ( athanassoula 1994 ) , so that in general the bar will not push the gas sufficiently inwards to reach the nucleus and eventually feed an agn . nevertheless even in cases with ilrs gas can be pushed to the center - most radii .
one possibility , which will be discussed in section 3 , is that there is a second bar within the main one .
other mechanisms for breaking the ilr barrier rely on the self - gravity of the gas accumulating in the central area .
such mechanisms have been described e.g. by fukunaga & tosa ( 1991 ) , wada & habe ( 1992 , 1995 ) , elmegreen ( 1994 ) and heller & shlosman ( 1994 ) .
the shock loci are regions of high density maxima , and one could have naively thought that this might entail a considerable amount of star formation .
this , however , is not the case , due to the fact that straight shock loci are also the loci of high shear ( a92b ) .
thus a typical molecular cloud will shear out before it has time to collapse , so that no star formation will occur .
this is not necessarily true for shocks whose loci are curved .
the two different types can be seen when comparing e.g. the dust lanes in the bar of ngc 1300 with the corresponding ones in ngc 1566 . with a similar goal , regan ,
vogel & teuben ( 1997 ) stressed the large divergence in gas streamlines before they reach the shock , and argued that this could tear apart molecular clouds and thus inhibit star formation .
it would be interesting to test whether this mechanism works for straight dust lanes as opposed to curved ones , as observations seem to suggest .
the ends of the bar are regions of high density , but not much shear .
thus one expects to find star formation there and this is indeed borne out by observations .
strong gas concentration can also be found in the area covered by the @xmath1 orbits , if the galaxy is sufficiently centrally concentrated ( e.g. early type disc galaxies ) and if the bar is sufficiently strong ( thin and/or sufficiently massive ) . in such cases
star formation can occur in the nuclear pseudo - ring , or in the nuclear disc .
a large number of observations have shown the existence of inner bars in different components , as the co , the visual , or the nir .
their formation could be due either to the existence of a bar unstable stellar inner disc ( i.e. an instability similar to that forming the main bar ) , or to the fact that sufficient gas has been pushed inwards to form a bar unstable gaseous disc ( shlosman , frank & begelman 1989 ) . the main bar and the inner bar do not necessarily turn with the same pattern speed , but this does not mean that they are dynamically independent . indeed the two pattern speeds can be different but coupled .
analytical work has predicted that the location of the ilr of the main bar should coincide roughly with that of the corotation of the inner bar ( tagger et al .
1987 , sygnet et al .
1988 ) , and this has been nicely borne out by the simulations of friedli & martinet ( 1993 ) . as already mentioned in section 2.2 such inner bars can help push the gas to the innermost regions of the galaxy
. the mere existence of an inner bar may nevertheless not be sufficient for this task .
the response of the gas in cases with two imposed bars will of course vary with the angle between the major axes of the two bars .
thus when the two bars have different pattern speeds the response is a function of time .
this is illustrated in fig . 2 , where i display the nuclear parts of the gas response at four different times during a simulation .
the outer ( or main ) bar is always at 45 to the horizontal axis and has a semi - major axis of 5 kpc .
the axisymmetric component has been chosen such that the rotation curve rises very steeply in the inner parts .
the gas response in the model with only the outer bar has strong shocks along the leading sides of the bar that , towards their innermost parts , wind up in the form of a nuclear spiral , as is the case for the model shown in fig .
1 . it is this nuclear region that gets most affected by the inner bar , as expected . in the example shown in fig .
2 the inner ( or secondary ) bar rotates roughly 3.5 times faster than the main bar , so that the corotation of the inner bar is situated roughly at the ilr of the main one .
the gas response in the innermost parts is now bar shaped , but a given end of this gaseous bar does not always link to the same density enhancement of the outer bar .
the gas response adjusts itself so that a given side of the inner gaseous bar links to the nearest density enhancement within the main bar , and in some cases one sees a nuclear pseudo - ring surrounding the inner gaseous bar .
the work described briefly here is part of an extensive study of the response of gas in galaxies with both a primary and a secondary bar , which i have been doing in collaboration with g. d. van albada .
a more comprehensive description of this work will be published elsewhere .
n - body simulations show clearly that stellar bars do not stay flat thin structures , but extend vertically , taking the form of a peanut if seen along the bar minor axis , and a box - like shape if seen edge - on and at an angle to it ( combes & sanders 1981 , combes et al .
1990 , raha et al .
1991 etc ) .
such peanut - like or box - like protuberances are observed in many disc galaxies and are called peanut or boxy bulges .
it is not easy to prove that they are indeed bars seen edge - on by using photometry alone .
thus efforts have been concentrated on kinematics .
kuijken & merrifield ( 1995 ) and bureau & freeman ( 1999 ) have compared the position velocity diagrams ( pvds ) of galaxies with peanut or box - like bulges , to the pvds of galaxies with standard bulges and find importance differences , which , they argue , could well be due to the signatures of the @xmath0 and @xmath1 families of orbits in the peanuts .
bureau & athanassoula ( 1999 ) made a detailed study of the effect of the existence and properties of the families of periodic orbits on the structure of the pvds .
athanassoula & bureau ( 1999 ) have tackled the same problem using hydrodynamical simulations , and they find that the existence of a gap in a pvd , between the signature of the nuclear spiral and that of the outer disc , reliably indicates the presence of a bar .
this gap is due to the fact that most of the main bar region is relatively empty , and this in turn is due to the gas flow described in section 2 .
all arguments seem to converge to the fact that peanut and boxy bulges are misnamed , and are in fact bars seen edge - on . | i briefly review the properties of the gas flow in and around the region of the bar in a disc galaxy and discuss the corresponding inflow and the loci of star formation .
i then review the flow of gas in barred galaxies which have an additional secondary bar .
finally i discuss the signatures of bars in edge - on galaxies .
# 1_#1 _ # 1#1 = # 1 1.25 in .125 in .25 in |
more than thirty years , after the discovery of gamma - ray bursts ( grbs ) we have now a reasonable grb model .
the model is based on the dissipation of an ultra - relativistic outflow . at first the flow
is dissipated by internal shocks ( or another form of internal dissipation ) that produce the prompt @xmath1-rays .
later the interaction of the flow with the circum - burst matter produces an external shock and this blast wave produces the subsequent afterglow .
the role of magnetic fields varies within grb models from crucial to ultimate .
it is generally accepted that the observed afterglow is produced by synchrotron emission @xcite .
synchrotron is also the best - bet model for the prompt @xmath1-rays emission ( note however some criticism of the synchrotron model for the prompt emission @xcite ) . even within the conventional model of a baryonic relativistic outflow
@xcite magnetic fields within the shocked regions are essential to produce this synchrotron emission . in the maximal case
the whole phenomenon is magnetic .
the inner engine that produces the relativistic outflow is driven by the blandford - znaek mechanism @xcite from a magnetized black hole - accretion disk system @xcite , or by a rapidly rotating highly magnetized pulsar @xcite .
the relativistic outflow is a poynting flux @xcite and the dissipation arises due to magnetic field recombination or some other instability .
even with a baryonic outflow magnetic fields may play a dominant role in the internal engine as a possible way to power the relativistic outflow is by reconnection of @xmath0 g magnetic fields within an accretion disk surrounding a compact object @xcite .
we can observe directly only the emitting regions where the dissipation and the emission take place .
the inner engine that accelerates and collimates the relativistic outflow and the outflow itself are hidden .
both could not be observed directly and we have only indirect clues on their nature . our inability to explore directly the inner engine and
the nature of the relativistic outflow ( poynting flux or baryonic ) is quite general .
the same problem arises in modelling relativistic jets seen in many other astronomical objects .
specifically the relativistic jets in agns and galactic micro quasars are the two most similar to grb jets .
grbs might hold the key for the resolution of the question what are these flow made off ?
i begin this short review with a brief discussion of the fireball model focusing on issues most directly relevant to magnetic fields .
i refer the reader to @xcite for a more general and extended review and to @xcite for a rather different perspective on magnetic fields in grbs .
i discuss the implication of synchrotron emission on magnetic fields within the afterglow and the prompt emission .
i examine what can we learn from polarization measurements .
after that i turn to the early afterglow and discuss the role of the prompt optical flash as a diagnostic of the nature of the relativistic outflow .
i conclude with a short discussion of future prospects for progress on these issues .
fig . [ fig:1 ] depicts an overall view of the internal - external shocks model . a compact source
whose size is @xmath2 cm emits a collimated ultra - relativistic wind with a lorentz factor of at least 100 . for a long duration grb
this might take place within a collapsing star , as suggested by the collapsar model @xcite and confirmed by the association of long duration grbs with type ic supernovae @xcite . in this case
the jet has to punch a hole in the stellar envelope , whose radius is @xmath3 cm . this envelope is not depicted in this picture .
[ height=0.3]interexter the kinetic energy of the wind is partially dissipated via internal shocks that take place around @xmath4 cm .
these shocks accelerate the electrons to ultra - relativistic energies ( the typical lorentz factor of an electron is @xmath5 ) . the electrons emit the observed prompt @xmath1-rays via synchrotron radiation .
the needed magnetic field is carried from the inner engine or is generated and amplified by the shocks . at a distance of @xmath6 cm
the loading of the circum - burst material on the outgoing flow becomes effective and the ejecta begins to slow down . at first a two shocks system forms . a forward shock propagating into the circum - burst matter , a reverse shock propagating into the ejecta and a contact discontinuity between the two .
this is a short lived phase during which the early afterglow arises . after the reverse shock crosses the ejecta a rarefaction wave passes and the ejecta cools down adiabatically . at this stage , at distances of @xmath7 cm from the center , the forward shock collects more and more circum - burst material .
it expands adiabatically and approaches the self similar blandford - mckee solution @xcite .
this is the ultra - relativistic analog of the well known sedov - taylor newtonian blast wave solution .
as more material accumulates the shock slows down , with the lorentz factor decreasing like @xmath8 and @xmath9 . sometime during this phase the lorentz factor @xmath10 , drops below @xmath11 the opening angle of the jetted outflow . at this stage
we encounter a jet break in the afterglow light curve . for @xmath12 most of the emitted radiation
is beamed within @xmath13 but for @xmath14 some of the radiation is emitted sideways .
additionally for @xmath14 the outflow expands sideways and if the expansion is rapid enough the additional collected matter causes a faster slowing down .
eventually , around @xmath15 cm the blast wave collects sufficient material to slow it down so that it reaches the newtonian transition .
this takes place a few month to a year after the burst . at this stage
practically all the radiation is in the radio frequencies .
later the blast wave approaches the taylor - sedov solution .
the magnetic field parameters in the different regimes are estimated from the observed emission , that is assumed to be synchrotron ( see the discussion below ) . in table
[ table:1 ] i compare various quantities relating to the magnetic fields in the various regimes .
[ h ] .magnetic parameters within grbs : the size of the region @xmath16 , the magnetic field @xmath17 , the electron s larmour radius , @xmath18 , @xmath19 the maximal correlation length needed for jitter radiation , the skin depth @xmath20 , and the width of the emitting region ( in the observer frame ) @xmath21 . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] an alternative model is based on a poynting flux outflow . in this case
the inner engine is magnetic and the dissipation is in the form of magnetic field recombination .
general arguments suggest that the sizes of the inner engine and of the different emitting regimes should be comparable to those within the baryonic flow model .
the magnetic fields are however , larger as they not only contribute to the synchrotron emission they also carry most of the energy .
[ fig:2 ] depicts a schematic description of this model .
it is also possible of course that the poynting flux energy and the baryonic energy are comparable leading to a mixed model @xcite .
the shock accelerated electrons emit , via synchrotron emission , the observed radiation .
[ fig:3 ] ( from @xcite ) depicts the typical synchrotron spectrum for fast cooling , which is applicable during the grb prompt phase and for slow cooling that is applicable for the afterglow .
it is generally assumed that the energy densities of the relativistic electrons and of the magnetic field can be characterized by equipartition parameters , @xmath22 and @xmath23 that measure the ratios of these energies to the total energy .
it is also assumed that the electrons energy distributed behaves like a power law with an index @xmath24 .
fits to the observed afterglow spectra suggest that @xmath25 , @xmath26 and @xmath27 @xcite .
schematically , the spectrum is composed of four segments separated by three typical frequencies .
the cooling frequency that is the synchrotron frequency of an electron that cools on a hydrodynamic time scale , the synchrotron frequency of the typical electron and the self absorption frequency ( see @xcite and @xcite for details ) .
the blandford - mckee ultra - relativistic self - similar solution combined with the synchrotron radiation model provides an excellent description of the afterglow light curves and spectra .
as one can see from table [ table:1 ] the magnetic fields in this region are relatively large .
the value of 1 g is much larger than the one obtained by a simple shock compression of the intergalactic magnetic field .
even if the burst takes place within a wind that was ejected from the star prior to its explosion the magnetic fields carried out by the wind at @xmath28 cm would be too weak .
an immediate question that arises is what is the origin of this strong magnetic field .
this is particularly puzzling as the afterglow is emitted by shocked circum - burst matter whose magnetic field can not be carried from the inner engine ( as can take place in the internal shocks ) .
the field has to be generated in site by the shock .
[ height=0.35 ] fig1 weibel @xcite suggested already in 1959 , that an instability leading to the growth of the magnetic field would arise whenever two flows cross each other .
kazimura et al . , @xcite have shown that such an instability could lead to growth of the magnetic field and to particle acceleration . as the collisionless shocks that arise in grb afterglows
are composed of two flows that cross each other medvedev and loeb @xcite suggested that the weibel instability is the origin of the strong magnetic fields inferred in grb afterglows .
however , gruzinov @xcite raised the concern that this magnetic field will maintain its equipartition magnitude only over a skin depth @xmath29 which for grb afterglows is around @xmath30 cm - ten orders of magnitude smaller than the width of the emitting region ( see table [ table:1 ] . while i am not aware of a simple analytic explanation of this puzzle recent numerical simulations @xcite suggest that equipartition size magnetic field remains over a region of thousands of skin depth .
moreover , the same instability accelerates particles to relativistic energies @xcite , providing another ingredient of the synchrotron emission . while current simulations can not imitate the realistic conditions they clearly suggest that the required magnetic fields can be amplified in afterglow shocks and hence in other relativistic collisionless shocks .
the energy density within the internal shocks region is much larger than the energy density within the afterglow , and so are the corresponding magnetic fields that can be as large as @xmath31 g ( depending on the exact location of the shocks and on their equipartition parameters ) . here
the field can be generated at the shocks ( by the same wiebel instability ) or it can be carried out from the inner engine .
the toroidal component of the field decays only as @xmath32 and hence a field of @xmath33 g at @xmath30 cm can easily reach @xmath34 g at @xmath33 cm .
note that such a magnetic field is energetically subdominant and it does not require that the flow will be poynting flux .
a poynting flux dominated flow requires at least @xmath0 g at the inner engine and at least @xmath35 g at @xmath33 cm . while synchrotron emission fits nicely the observed afterglow spectra and light curves it is not clear that it fits the observed prompt @xmath1-ray spectrum .
specifically , the low energy part of the synchrotron spectrum behaves like @xmath36 ( see fig .
[ fig:3 ] ) @xcite .
however many grbs show a steeper low energy spectrum .
that is many spectra fall below the `` synchrotron line of death '' @xcite .
the constraint is even more severe , considering the fact that for the required high efficiency the system must be fast cooling and this requires a spectrum less steep than @xmath37 @xcite .
these observations have led some authors @xcite to rule out synchrotron as the source of the prompt @xmath1-ray emission . however ,
the low energy spectral observations of the more sensitive ( in this energy band ) hete @xcite do not show numerous bursts below the `` lines of death '' .
we may have to wait for better data to determine this issue .
the problems that the synchrotron model encountered in fitting the observed grb spectra have led to the suggestions that synchrotron - self - compton @xcite and inverse - compton of an external photon field @xcite are the emission processes .
medvedev @xcite proposed that the prompt emission is produced by jitter radiation - the analogue of synchrotron radiation in a random field @xcite .
jitter radiation arises in cases that the magnetic field s correlation length @xmath38 is smaller than the region over which an electron emits the synchrotron radiation seen by a given observer , @xmath39 ( see fig .
[ fig:4 ] ) .
the spectra of jitter radiation is more complicated than the synchrotron spectra and medvedev suggests that it could have a low energy spectrum that is steeper than @xmath37 or even than @xmath36 .
see however @xcite for remarks on medvedev s results .
as synchrotron emission is intrinsically polarized with a suitable geometry it would lead to polarized afterglow @xcite . in particular the required geometry could occur when we observe a relativistic jet . by now polarized optical emission
has been observed from several afterglows @xcite .
somewhat surprisingly the polarization magnitude and angle do not follow the simple predictions @xcite . while the predicted polarization was at the level of a ten percent
the observed polarization was only around a few percent .
furthermore , the jet structure suggests a specific polarization pattern for a uniform jet @xcite and another one for a universal structured jet @xcite .
none was observed so far . the patchy shell model in which the local geometry of the jet is dominated by hot spots @xcite provides a possible explanation for the lower polarization level and the jumps in the polarization angle @xcite .
granot and taylor @xcite reported recently upper limits on the polarization of grb radio flares ( the early part of the radio afterglow ) .
their best limit , for grb 991216 is around 10% .
they point out that this upper limit already rules out a universal structure jet model with a homogenous magnetic field .
more refines observations will enable us to distinguish between different jet models .
coburn and boggs @xcite announced the discovery of very strong ( @xmath40% ) polarization from the prompt emission of grb 021206 .
it has been argued that this level of polarization shows that the magnetic filed must be uniform @xcite and that this further implies that the flow must be poynting flux dominated @xcite .
however , two independent groups @xcite reanalyzed the same data and found no statistical viable indication of polarization . additionally , synchrotron emission from a relativistic jet @xcite can produce a @xmath41 polarized light even with a random magnetic field ( provided that the field is within the plane of the shock ) .
this is only slightly lower than the maximal emission from a homogenous magnetic field configuration @xcite which is around 60% .
furthermore , a homogenous ( toroidal ) field configuration would arise in any case when the field is dragged from the source @xcite .
thus , homogeneity does not necessarily imply that the flow is poynting flux dominated . we still need a better way to distinguish between poynting flux and baryonic outflow .
one of the most interesting and most vividly debated questions conceding grbs is the nature of the relativistic outflow - poynting flux or a baryonic outflow .
one of the best way to explore this question is by observation and analysis of the early afterglow @xcite .
[ height=0.3]lc1 the initial interaction of the relativistic outflow with the circum - burst material results in two shocks .
a short lived reverse shock that propagates into the outflow and a long lived forward shock that propagates into the circum - burst material .
the later eventually turns into the relativistic blast wave that produces the afterglow .
recently nakar and piran @xcite presented a new analysis of the reverse shock emission ( see also @xcite ) .
they show that a baryonic jet has a unique signature and that a combined analysis of the optical and the radio data could , therefore , enable us to explore the nature of the outflow .
the reverse shock optical emission peaks at @xmath42 when the reverse shock reaches the inner edge of the outflow .
the first unique signature of the reverse shock is a characteristic decay of the optical flux after @xmath42 as @xmath43 @xcite .
this decay arises due to the adiabatic cooling of the shocked material .
the behavior is quite different at the radio .
initially , the typical synchrotron frequency , @xmath44 is above the observed radio frequency @xmath45 , and below the self absorption frequency , @xmath46 .
as the shocked material cools @xmath44 decreases and the radio emission increases as @xmath47 until @xmath48 . at this stage the radio flux increases like @xmath49 until , at @xmath50 @xmath51 and from that moment on the radio emission decreases like @xmath52 just like the optical .
this behavior is schematically depicted in fig . [ fig:5 ] for baryonic outflow there is a unique relation between the time and flux of the optical peak and the time and flux of the radio peak , @xmath42 , @xmath53 , @xmath50 and @xmath54 : @xmath55 , where @xmath56 is the index of the electron energy distribution and @xmath57 is a constant of order unity .
current uncertainties put @xmath57 between 0.5 and 2 . by now the only burst with well determined optical early afterglow is grb 990123 .
its optical flash decayed as @xmath52 . the corresponding factor that relates
the optical and radio peak fluxes and peak times for grb 990123 is 1300 very close to the expected value ( 1000 - 2000 ) .
both results indicate a baryonic outflow in this particular burst .
magnetic fields are clearly a key ingredient of grbs .
they are essential in the emitting regions and it may very well be that they appear in all parts of the grb phenomenon , including the central engine and the relativistic outflow . as nakar & piran have shown @xcite combined measurements of the optical afterglow and the radio flare could enable us to determine the early afterglow parameters for bursts in the nearby future .
swift and several quick response ground telescopes are ideally suited for these optical observations and hopefully radio telescopes will be allocated sufficient time and quickly enough to detect the corresponding radio flares .
this would hopefully shed some light on the central question - the nature of the relativistic outflow - baryonic or poynting flux .
observations of prompt polarization are quite unlikely as one need an extremely bright burst , and such bursts are very rare . on the other hand
the quick notification of swift may allow measurements of many optical polarizations and possibly radio flare polarizations .
these could shed further light on the signature of magnetic fields and on the angular structure of grb jets .
as for the formation of magnetic fields in collisionless shocks , the success of the synchrotron model in explaining the basic properties of grb afterglow demonstrates that nature knows how to generate magnetic fields in collisionless shock , even if we do nt understand yet how .
the progress made in numerical simulations suggest that we are on the right track and indeed it is the weibel instability that is responsible for this field amplification .
this might have some implications for seeding magnetic fields elsewhere in the universe t. j. galama , p. m. vreeswijk , j. van paradijs , c. kouveliotou , t. augusteijn , h. bohnhardt , j. p. brewer , v. doublier , j .- f .
gonzalez , b. leibundgut , c. lidman , o. r. hainaut , f. patat , j. heise , j. in t zand , k. hurley , p. j. groot , r. g. strom , p. a. mazzali , k. iwamoto , k. nomoto , h. umeda , t. nakamura , t. r. young , t. suzuki , t. shigeyama , t. koshut , m. kippen , c. robinson , p.
de wildt , r. a. m. j. wijers , n. tanvir , j. greiner , e. pian , e. palazzi , f. frontera , n. masetti , l. nicastro , m. feroci , e. costa , l. piro , b. a. peterson , c. tinney , b. boyle , r. cannon , r. stathakis , e. sadler , m. c. begam , and p. ianna , _ nature _ , * 395 * , 670672 ( 1998 ) .
j. s. bloom , s. r. kulkarni , s. g. djorgovski , a. c. eichelberger , p. cote , j. p. blakeslee , s. c. odewahn , f. a. harrison , d. a. frail , a. v. filippenko , d. c. leonard , a. g. riess , h. spinrad , d. stern , a. bunker , a. dey , b. grossan , s. perlmutter , r. a. knop , i. m. hook , and m. feroci , _ nature _ , * 401 * , 453456 ( 1999 ) . k. z. stanek , t. matheson , p. m. garnavich , p. martini , p. berlind , n. caldwell , p. challis , w. r. brown , r. schild , k. krisciunas , m. l. calkins , j. c. lee , n. hathi , r. a. jansen , r. windhorst , l. echevarria , d. j. eisenstein , b. pindor , e. w. olszewski , p. harding , s. t. holland , and d. bersier , _ ap . j. lett .
_ , * 591 * , l17l20 ( 2003 ) .
j. hjorth , j. sollerman , p. mller , j. p. u. fynbo , s. e. woosley , c. kouveliotou , n. r. tanvir , j. greiner , m. i. andersen , a. j. castro - tirado , j. m. castro cer ' on , a. s. fruchter , j. gorosabel , p. jakobsson , l. kaper , s. klose , n. masetti , h. pedersen , k. pedersen , e. pian , e. palazzi , j. e. rhoads , e. rol , e. p. j. van den heuvel , p. m. vreeswijk , d. watson , and r. a. m. j. wijers , _ nature _ , * 423 * , 847850 ( 2003 ) . c. barraud , j .- f .
olive , j. p. lestrade , j .- l .
atteia , k. hurley , g. ricker , d. q. lamb , n. kawai , m. boer , j .-
dezalay , g. pizzichini , r. vanderspek , g. crew , j. doty , g. monnelly , j. villasenor , n. butler , a. levine , a. yoshida , y. shirasaki , t. sakamoto , t. tamagawa , k. torii , m. matsuoka , e. e. fenimore , m. galassi , t. tavenner , t. q. donaghy , c. graziani , and j. g. jernigan , _ astron . &
_ , * 400 * , 10211030 ( 2003 ) .
s. covino , d. lazzati , g. ghisellini , p. saracco , s. campana , g. chincarini , s. di serego , a. cimatti , l. vanzi , l. pasquini , f. haardt , g. l. israel , l. stella , and m. vietri , _ astron . &
_ , * 348 * , l1l4 ( 1999 ) . r. a. m. j. wijers , p. m. vreeswijk , t. j. galama , e. rol , j. van paradijs , c. kouveliotou , t. giblin , n. masetti , e. palazzi , e. pian , f. frontera , l. nicastro , r. falomo , p. soffitta , and l. piro , _ ap . j. lett . _ , * 523 * , l33l36 ( 1999 ) .
e. rol , r. a. m. j. wijers , p. m. vreeswijk , l. kaper , t. j. galama , j. van paradijs , c. kouveliotou , n. masetti , e. pian , e. palazzi , f. frontera , and e. p. j. van den heuvel , _ ap . j. _ , * 544 * , 707711 ( 2000 ) .
j. greiner , s. klose , k. reinsch , h. martin schmid , r. sari , d. h. hartmann , c. kouveliotou , a. rau , e. palazzi , c. straubmeier , b. stecklum , s. zharikov , g. tovmassian , o. b " arnbantner , c. ries , e. jehin , a. henden , a. a. kaas , t. grav , j. hjorth , h. pedersen , r. a. m. j. wijers , a. kaufer , h. park , g. williams , and o. reimer , _ nature _ , * 426 * , 157159 ( 2003 ) .
d. bersier , b. mcleod , p. m. garnavich , m. j. holman , t. grav , j. quinn , j. kaluzny , p. m. challis , r. g. bower , d. j. wilman , j. s. heyl , s. t. holland , v. hradecky , s. jha , and k. z. stanek , _ ap
. j. lett .
_ , * 583 * , l63l66 ( 2003 ) . | magnetic fields play a crucial role in the physics of gamma - ray bursts ( grbs ) .
strong observational evidence indicates that the observed afterglow and most likely the prompt emission arise from synchrotron emission .
it is possible that poynting flux plays an important or even dominant role in the relativistic outflow from the inner engine , but like in other astronomical relativistic jets this suggestion is controversial .
finally , it is likely that magnetic fields larger than @xmath0 g occur within grbs inner engines and contribute to the acceleration and collimation of the relativistic jets .
i review here the grb fireball model and discuss the role that magnetic fields play in its various components .
i suggest that the early afterglow , that reflects the initial interaction of the relativistic jet with its surrounding matter is the best available tool to explore the nature of relativistic outflow in astronomical relativistic jets .
address = racah institute for physics , the hebrew university , jerusalem , 91904 israel + theoretical astrophysics , caltech , pasadena , ca 91125 , usa |
high resolution imaging , in particular in biology and materials science , is mostly done using scanning electron microscopy ( sem ) . the ease of use and the widespread availability has made this the number one method for imaging in the aforementioned fields . structuring and manipulation of nano
sized features is traditionally the domain of focused ion beams . here , mainly liquid metal ion sources ( lmis ) using gallium are used
however , other techniques such as various types of gfis,@xcite alloy lmis,@xcite magneto optical trap sources ( motis)@xcite and multicusp plasma sources@xcite are runners up
. good reviews discussing these two techniques can be found in refs . [ ] and [ ] for sem and fib , respectively .
helium ion microscopy presents a technique that unites many of the advantages of sem and fib in a single tool .
more importantly , it also overcomes some of the deficiencies of sem and fib .
in particular , the possibility to image biological and in general insulating samples without coating at highest resolution is important for many scientific questions currently discussed in materials science and biology .
another important charged particle beam technique namely transmission electron microscopy ( tem)depends on very thin samples free of defects from the preparation . the unique nano
sized beam of the him makes it possible to not only mill and pattern smallest features but also do this with minimal damage to the crystal lattice . in the following
we will give an outline of the working principle and signal generation in helium ion microscopy , followed by two sections highlighting specific imaging applications and examples of materials modification .
the initial idea of a scanning ion microscope has been put forward by knoll and ruska already in 1932.@xcite the working principle of a helium ion microscope can be divided into three different stages .
1 . helium ionization and acceleration 2 . beam formation and control 3 .
sample interaction in this introduction we will only briefly touch points 1 and 2 .
instead , we will focus on the physics that is important to understand the application of the technique to imaging and nanofabrication . the first is made possible by using a newly developed gas field ion source ( gfis ) .
gfis have been investigated for a long time@xcite for their use in microscopes.@xcite the idea is based on the initial design of a field ion microscope by e. mller.@xcite however , only recently a stable ion source with a high brightness and small virtual source size has been realized by ward , notte and economou for use in a microscope.@xcite it is based on an emitter whose apex has been shaped into a three sided pyramid ( see fig .
[ fig : him - source ] ) .
work is done currently to understand and optimize the supertip formation process in order to maximise the achievable current.@xcite using scanning field ion microscopy ( sfim ) the apex of the tip can be monitored and shaped using high fields that can ultimately remove weakly bound atoms from the apex . in this way the configuration of the tip apex can be controlled at the atomic level .
although monomers are possible , trimers are more stable .
figure [ fig : him - source](b ) shows a sfim image of the tip .
the trimer in the center and the edge of the first two shelves below the trimer are visible .
the combination of a pyramidally shaped tip and the carefully shaped apex concentrates the electric field , so that subsequent helium ionization predominantly happens at the top most atoms of the pyramid .
using apertures in the beam path allows to select current originating from ionization events on a single apex atom .
typical source parameters are listed in table [ tab : column - param ] .
.typical helium ion source parameters@xcite [ cols="<,^ , < " , ] the second point is technologically demanding and requires a high degree of knowledge on the design and implementation of the involved electrostatic lenses , quadrupoles , octopoles , etc . for more details
the reader is refereed to numerous monographs available on charged particle optics such as [ ] .
the critical source parameters important for the ion optical performance of the column are given in table [ tab : column - param ] .
the energy spread of 1ev is an upper bound .
earlier measurements indicate that the values could in fact be lower by a factor of two to four.@xcite one of the important consequences of the parameters listed in table [ tab : column - param ] is the image side beam convergence angle @xmath0 .
typical values for @xmath0 are well below 1mrad .
this small beam divergence results in a large depth of field @xmath1 here , @xmath2 denotes the minimum feature that can be resolved in the image . assuming identical resolutions the him will have a superior depth of field as compared to low voltage sem with typical @xmath0 values of 8mrad.@xcite once the focused ion beam strikes the sample , the interaction of the accelerated particles with the substrate atoms and electrons will give rise to a large number of different signals .
we will cover the available signals in the following section .
before we do so , we will briefly discuss the processes that occur during ion / sample interaction and that are important for the achievable resolution in charged particle beam imaging .
besides the small beam diameter , the shape and size of the actual interaction volume between the impinging particle and the sample plays an important role to reach ultimate resolution .
figure [ fig : charged - part - beams ] compares monte carlo ( mc ) simulation results for different charged particle beams . from the figure it is evident that the interaction volume relevant for secondary electron ( se ) generation of the focused he beam is smaller than for the other two.@xcite for the case of a ga beam the large cross section of ga with in this case si
leads to substantial scattering in the near surface region relevant for the se signal generation . for a low energy electron beam needed to simultaneously optimize resolution and surface sensitivity in sem electron
electron scattering in the sample widens the beam dramatically in the first few nanometers deteriorating the achievable resolution .
scattering also occurs in him . in the case of a 30kev
he beam scattering occurs with the nuclei of the sample atoms .
however , due to the low mass of helium , scattering is not very efficient in the first few nanometers of sample material .
this results in minimal beam divergence inside the sample .
consequently the collected se originate from a cylinder with a minimal volume .
the somewhat lower energy of the se in him@xcite and the lower characteristic escape depth for se in him for light elements@xcite enhances this difference between sem and him even further .
next we will discuss available particles and correoponding signals .
we will in particular highlight their benefits and drawbacks when used for imaging and what kind of physical quantities can be accessed using them .
the sequence in which they are discussed is determined by their abundance in the tool .
this also corresponds to the ease of use and how widespread the technique is available in the current instrument base .
an initial overview of some of the different contrast mechanisms is given by bell.@xcite secondary electron ( se ) emission generated by ions can be split into two parts . in the logical order we will first discuss se generation followed by the transport of electrons through the material .
the latter is important as the generated electrons still need to reach and subsequently leave the surface into the vacuum so they can be detected . as has been proposed by bethe@xcite
the rate of secondary electron generation @xmath3 ( electrons per incoming ion ) is proportional to the stopping power of the incident particle @xmath4 in ev / .
@xmath5 here , @xmath6 denotes a scaling constant .
it is assumed that at least in the relevant near surface region , atomic collisions do not play an important role and @xmath4 depends only on the electronic stopping power .
the generated se1 are mostly excited by the incoming ions via kinetic emission ( ke).@xcite two types of secondary electrons of type 2 ( se2 ) exist in ion beam imaging .
type 2 electrons can be generated either by recoiling target atoms or from other ses of type 1 .
the second type of se2 generation ( se generated by se ) does not play an important role in him .
this becomes clear when looking at the maximum energy of the se , which is taken to be equal to the energy loss of the impinging he for a head on collision@xcite @xmath7 here , @xmath8 is the electron mass , @xmath9 the ion velocity , and @xmath10 denotes the fermi velocity .
the cross section for such a collision is highest if the ion velocity@xmath11 m/s for a primary energy of 35kev matches the fermi velocity of the electrons in the material . for gold and 35kev
he this yields a maximum se energy of 45ev .
however , this energy is approximately equal to or even below the effective energy needed for se generation by electrons in many materials.@xcite thus the size of the electron
electron collision cascade is restricted . however , kinetic excitation of electrons is also possible by recoil atoms , provided they are fast enough so that their speed still matches the fermi velocity of the target material .
electrons produced by recoiling target atoms are usually called se2 .
ramachandra et al.@xcite calculated the ratio between se2 and se1 as function of energy and material .
the conclusion that can be drawn from their calculation is that for higher primary energies a smaller se2/se1 ratio can be achieved for most materials and consequently a higher resolution is possible .
the other process for electron emission is potential emission ( pe ) via auger neutralization .
however , only for very low energies below 5kev pe becomes dominant@xcite over ke .
once secondary electrons have been generated they still need to escape from the solid .
this process can be described as a diffusion process .
the characteristic length scale of this diffusion process the effective diffusion length of secondary electrons @xmath12is of the order of 1 nm for nearly all materials.@xcite this leads to the fact that only the first few nanometers of the material add to the emitted ses
. measurements of the effective se yield in him show variations between 1 for carbon and values as high as 8 for platinum.@xcite the number and energy distribution of these ion induced secondary electrons differs from what is found in a sem .
a sharper maximum at lower energies is usually found@xcite in him . in fig .
[ fig : se - yield](a ) a comparison of calculated se yields in sem , ga
fib and him is presented .
the calculations done by ohya et al . still overestimate peak position as well as peak width.@xcite these differences between actual measurement results of se yields in a him and calculations are attributed to se generation mechanisms not considered in the calculations .
indications exist that bulk plasmon excitations can play an important role@xcite for se generation in him .
an actual measurement of se yield as a function of se energy and the work function dependence of the distribution maximum is presented in fig .
[ fig : se - yield](b ) . .
( b ) experimentally obtained se yield for three different metals .
the dependence of the peak position on work function is shown as an inset . reproduced with permission from proc .
spie , vol .
8036 , 80360o-180360o-10 ( @xcite ) .
copyright ( 2011 ) spie . ] a consequence of the particular energy distribution of se in him and the small se generation volume is an enhanced surface sensitivity .
this has been shown in a recent comparative study of him and sem imaging performance on carbon coated gold nano
rods.@xcite only at very low acceleration voltages is sem able to visualize the thin carbon layer covering the gold particles .
however , low voltage sem suffers from carbon deposition in the imaged area making detailed studies challenging .
an example of the high sensitivity of him with respect to the material work function is the visualization of different half unit cell surface termination in ti@xmath13sic@xmath14.@xcite depending on the position of the cleavage plane the surface is either terminated by si ( half unit cell ) or ti ( full unit cell ) .
the difference in the chemistry of the top surface layer of atoms results in different se yields for the two terminations .
as a result they can be discriminated in him . a software package called _ ionise _ developed by p. rack and coworkers is available that allows the calculation of the expected se yield for a large number of materials.@xcite although good agreement has been achieved between simulation and experiment , only a limited number of analytical applications based on se yield are known.@xcite sample topography in him is made visible in a similar way as in sem .
the dependence of the se yield on the specimen tilt with respect to the incoming beam can be described the following secant law @xmath15 however , experimental studies showed that the actually measured se yield at the detector deviates from the expected secant behaviour.@xcite the deviations lead to a more linear response curve , which in turn should reduce the strong edge effect known from sem .
nevertheless , a very strong edge effect has been observed in thin layers.@xcite in summary , the achievable high resolution and surface sensitivity in him is based on the fact that the ses originate from a cylinder at the beam penetration point with a diameter of approximately 1nm given by the beam diameter and a length of less than 3nm determined by @xmath12 .
this small probe volume helps to achieve very high image resolution in him .
it should be noted that obtaining such high resolution images has become substantially easier since the introduction of uhv him.@xcite at small fields of views usually high fluences are reached as a consequence of the large pixel density . in an uhv him implantation and sputtering can still negatively affect the sample during imaging .
carbon deposition on the other hand can be excluded.@xcite the removal of hydrocarbons from the sample chamber vacuum prevents the formation of carbon deposits in the imaged area . as a consequence some exceptional imaging results could be achieved ( see fig . [
fig : bio](b ) ) .
what backscattered electrons are to the sem , backscattered helium ( bshe ) atoms and ions are to him .
this rather bold statement is true in several ways , as will be highlighted in the next paragraphs .
two different ways to utilize bshe are available in current him .
first , and most commonly used , is a microchannel plate ( mcp ) detector to acquire qualitative element distribution maps .
second , a silicon drift detector can be used to obtain spectroscopic information from microscopic areas .
the latter allows quantitative element identification based on the same principles as rutherford backscattering spectroscopy ( rbs ) . while se images usually are rich in morphological contrast , bshe images obtained with the mcp are poor in topography and rich in elemental contrast .
in contrast to the se images , the information in these images originates from the bulk of the sample . to measure them the mcp
is inserted below the pole piece in the primary beam path .
while a center hole allows the primary beam to reach the sample , this geometry maximizes the solid angle , and thus the collection efficiency of the detector .
the obtained contrast can be understood by examining the rutherford scattering cross section @xmath16 where @xmath17 is the elementary charge , and @xmath18 and @xmath19 denote the atomic number of impinging and target particle , respectively .
assuming that the target atom is at rest , @xmath20 is the energy of the impinging particle , @xmath21 is an arbitrary element of solid angle , and @xmath22 is the scattering angle . for fixed @xmath18helium for this review and a given energy , a dependence on @xmath23 leads to a strong contrast between different elements .
a more detailed analysis of the scattering cross section shows that a dependence that is related to the structure of the periodic table of elements also exists . this is a result of the change in screening along the rows of the periodic table . in fig .
[ fig : bshe - yield ] , experimentally obtained bshe yields for various elements are presented . .
the row structure of the periodic table can clearly be seen on top of a general increase in bshe yield for heavy elements .
data courtesy of carl zeiss ag . ]
the dependence of the yield on the structure of the periodic table is clearly present . at this point
it is important to realize that a high sensitivity is needed from the detector that is used . under standard imaging conditions in him ,
typically 500 or less ions are used per pixel . assuming a high backscatter yield of 20% and taking into account the detector solid angle
, no more than 50 ions will reach the detector .
a large fraction of these backscattered he particles are also neutral@xcite due a charge transfer process which occurs once they enter the sample .
however , this would represent an ideal case for a heavy element bulk target . in practice these numbers can be substantially smaller and not a single ion should be lost in the detector.@xcite one should realize that contrary to the electrons discussed in the previous section , bshe represent a bulk signal . depending on the atomic number sampling depth , between a few tens of nanometers and a few hundred nanometers can be achieved .
as can be seen from equ .
( [ equ : cross - section ] ) , the backscatter efficiency can be increased by lowering the acceleration voltage . as a side effect , this will also reduce the range of the helium and thus the sampling depth .
in addition to the qualitative element distribution , also quantitative information on the elemental composition can be obtained . for a binary collision
when momentum and energy are conserved the kinetic energy of the backscattered helium @xmath24 depends on the on the ratio of the masses of the impinging ( @xmath25 ) and target ( @xmath26 ) particle . measuring the energy loss at a fixed angle allows the mass of the collision partner , as well as its position relative to the sample surface , to be determined . at high kev or low mev energies this
is known as rutherford backscatter spectroscopy ( rbs ) .
this method is known to deliver high sensitivity and excellent depth - resolution .
the nature of the involved square function makes the method very sensitive to differences between light elements .
a silicon drift detector with a resolution of approximately 4kv can be used for this purpose . however , due to the relatively low primary energy of only 35kev , the measured peaks are not as well defined as in classical rbs .
nevertheless , the nature of him allows such spectra from areas of only several m@xmath27 to be obtained.@xcite the technique has been successfully applied to measure zro@xmath14 layer thicknesses on si substrates with monolayer sensitivity@xcite and for thickness measurements of cobalt / nickel nano
rings.@xcite the achievable resolution is not only limited by the detector . as a result of the already low primary energy ,
the backscattered particles have a high probability to undergo multiple scattering events .
this occurs for the impinging , as well the backscattered helium particle .
consequently , a statistical broadening of the exit energies of the bshe exists .
it has been shown by simulations that even for an ideal combination , such as a 75 nm thick heavy hafnium oxide film on silicon , an intrinsic uncertainty of 14% exists for the depth resolution of hf .
given the current detector resolution this value is further increased to 60%.@xcite compositional characterization will have even larger errors due to a severe peak overlap.@xcite figure [ fig : hfo ] presents an experimentally obtained spectrum from hafnium oxide
. the simnra@xcite calculated fits to obtain elemental composition and layer thickness have been added for different real and ideal detector resolutions .
an increase of detector resolution to 1kv would substantially improve the quality of the obtained spectra and consequently of the fit accuracy .
these curves should be taken to be an indication of what could be possible and not an exact representation of the achievable resolution .
one also needs to keep in mind that the majority of the backscattered particles are neutral,@xcite complicating any post scattering treatment .
the yield of both signals secondary electrons as well as back scattered helium also depends on the crystalline orientation of the sample with respect to the beam .
this allows for two additional contrast mechanisms in him .
the well known channeling contrast@xcite can be exploited to obtain the crystal orientation of samples with a large lateral resolution.@xcite this technique makes use of calculations of the blocked area fraction , which yields results similar to stereographic projections of channeling minima or laue back reflection patterns .
one of the surprising results is that despite the very small wavelength of he ions , scanning transmission ion microscopy ( stim ) is possible .
both dark field and bright field images could be recorded in which thickness fringes and line defects could be identified.@xcite generally speaking , ionoluminescence ( il ) is a phenomenon of light emission due to the optical transitions of an electronic system which has been excited by ion irradiation .
three main stages can be distinguished in the luminescence process:@xcite 1 .
energy absorption and excitation of the electronic system ; 2 .
system relaxation and energy transfer to the emission centers ; 3 .
transition of the system into the ground state by photon emission . during sample bombardment with he@xmath28 ions
, light can be obtained from the excited backscattered neutral he,@xcite excited sputtered atoms and molecular complexes,@xcite or from the sample material itself.@xcite for the case where emission originates from excited species which have left the sample , the spectrum consists of discrete doppler broadened lines .
these lines corresponds to the optical transitions within atomic ( or molecular ) orbitals .
this light is usually observed at a distance up to a few millimeters from the target surface.@xcite in the case of organic samples , ionoluminescence reveals the electronic structure of the organic compounds.@xcite for the investigation of material properties we are mainly interested in the luminescence from the sample itself .
the physics of the emission processes is usually described by considering the sample s electronic band structure , or by using configuration coordinate diagrams.@xcite according to its origin , there are two general types of luminescence : _ extrinsic _ and _ intrinsic_.@xcite in the case of extrinsic luminescence , the light generation is related to the presence of impurity atoms or ions ( so called _ activators _ ) in the sample material . depending on their electronic structure
, activators can be divided in the following groups : 1 .
transition metal ions with d@xmath29 electronic configuration ( e.g. ti@xmath30 , cr@xmath30 , mn@xmath31 ) ; 2 . ions with s@xmath32-configuration ( e.g. tl@xmath28 , pb@xmath31 , sb@xmath30 ) ; 3 .
earth elements ( ree@xmath33 ) ; 4 .
actinides ( e.g. uo@xmath34 , cm@xmath30 ) .
sometimes the presence of a certain impurity ( _ sensitizer _ ) is needed for the luminescence of an activator ( for example ce@xmath30 for activation of tb@xmath30 ions @xcite ) . as opposed to sensitizers , _ quenchers _ suppress light emission from an activator .
for example , fe@xmath31 ions act as quenchers for emission from mn@xmath31 in apatite.@xcite at high activator concentrations self quenching may occur due to resonant absorption processes .
intrinsic luminescence on the other hand , is emission from the sample material itself .
two cases can be distinguished .
first , optical transitions from delocalized states , or in other words the recombination of free electrons from the conduction band with holes in the valence band .
this can include direct and indirect transitions .
the radiative recombination of free excitons also falls in this category .
second , optical transitions from localized states can also be attributed to intrinsic luminescence .
this includes the following processes : recombination of excitons trapped at defect sites ( so called self trapped excitons@xcite ) , emission from excited defects known as color centers ( e.g. nitrogen vacancy centers in diamond@xcite ) , and transitions of charge carriers from delocalized into localized states .
the shape and width of the emission peaks and bands strongly depends on electron phonon interactions and thermal effects .
a strong electron
phonon interaction leads to a stokes shift and peak broadening . as a result
, it is desirable to perform ionoluminescence measurements at cryogenic temperatures , which has not been done to date in him .
since ionoluminescence is in many aspects similar to the cathodoluminescence ( cl ) phenomenon often observed in scanning electron microscopy ( sem ) , databases of cl studies can be employed for the interpretation of il spectra .
ionoluminescence studies are significantly complicated by the fact that an ion beam not only induces light emission , but also directly influences the optical properties of the target due to defect creation .
ion irradiation can lead to target coloration ( e.g. in alkali halides@xcite ) and enhanced emission , but also quench the luminescence ( e.g. semiconductor materials@xcite ) .
however , the use of him to observe il phenomena allows these processes to be followed in
situ with a high lateral resolution .
in this section we will try to give an overview of applications of helium ion microscopy that make use of the special imaging capabilities of the him .
although , the high resolution is the most prominent fact that allows for very accurate critical dimension measurements,@xcite many successful applications of him make use of other distinct characteristics of him .
this overview is by no means complete but will attempt to highlight interesting and eventually surprising imaging applications .
the use of electrons for charge neutralization enables him to obtain high resolution images of insulating , and in particular uncoated , biological samples . in fig .
[ fig : bio](a ) a high resolution him image of a butterfly wing is presented.@xcite the black ground scales of papilio ulysses and other butterflies are imaged without any prior coating , which allows the smallest features to be identified .
the large depth of field in him also makes it ideal for creating anaglyphs using different angles for imaging .
high precision measurements of otherwise not accessible feature heights are possible this way.@xcite however , only very few groups have demonstrated the applicability of him for imaging of cells.@xcite this is astonishing since very high resolution should be possible .
figure [ fig : bio](b ) demonstrates the achievable resolution on such biological , soft and insulating samples .
the imaged protein crystal ( flat bovine liver catalyse ) exhibits a simple rectangular unit cell with a lattice spacing of 8.8 nm @xmath35 6.7 nm .
these values are in excellent agreement with the expected values from literature.@xcite the use of bshe for imaging allows amongst other things subsurface processes such as the formation of buried contacts to be visualized .
this overcomes an existing limitation in how we currently try to follow subsurface diffusion processes .
helium ion microscopy offers a unique , destruction free method to reveal the in
plane shape of a diffusion front . in fig .
[ fig : buried ] results of a him study on the subsurface formation of pd interconnects are presented.@xcite covered pd@xmath14si interconnects .
( b ) simultaneously recorded bshe image showing the pd deposits ( bright , top and bottom ) and the buried pd@xmath14si layer .
the shape of the two pd diffusion fronts is clearly discernible in between the two pd contacts . ] while a large degree of surface detail is present in the se image presented in fig .
[ fig : buried](a ) , no direct indication of the buried pd@xmath14si interconnect is visible .
however , from the simultaneously recorded bshe image shown in fig .
[ fig : buried](b ) the morphology of the pd diffusion front can be seen .
the benefit of this method over cross section approaches is evident . besides the obvious ease with which the shape of curved interfaces can be resolved , sample preparation is also substantially easier .
knowledge of the non straight nature of the diffusion front is important to understand the reason for large variations in device performance .
the shape of the subsurface diffusion front is not accessible using cross section techniques .
because the penetration depth of the ions can be varied by using different primary energies , it is also possible to get an estimate of the thickness of the cover layer.@xcite there are several reviews@xcite introducing various applications of il for material characterization that also reveal possible complications of the technique . here
, we will focus on il imaging with light ions and results obtained with him .
the possibility to do il imaging using a him was demonstrated by boden et al.@xcite the authors tested a wide range of samples such as quantum dots ( qd ) , semiconductors , rare
earth doped nanocrystals , and ce@xmath30 doped garnet .
both bulk samples and nanocrystals doped with ree were found to exhibit il . this is a result of the optical transitions of the @xmath36 electrons in the dopant .
however , the il signal was found to decay with prolonged beam exposure .
this decay has been attributed to defect accumulation which hinders electron transport to the actual emitting rare earth atoms .
several direct bandgap semiconductor materials were investigated in ref .
[ ] , but no il was detected .
however , recently we managed to obtain clear ionoluminescence information from such a material .
figure [ fig : fig_gan ] shows the il image obtained from a thin gan film on sapphire using a fluence of only @xmath37@xmath38 .
beam energy is 35kev .
fov : 45 ] the dark lines in the il image ( fig . [ fig : fig_gan ] ) are dislocations that are initially present in the film . these act as centers for non radiative recombination@xcite and appear dark . the black dot with the grey halo in the center of the image is the result of the long term irradiation of a single pixel prior to recording the image . due to
the high dose applied there all il is quenched rapidly .
this is the result of the creation of various types of defects that provide non radiative paths for the electron de
excitation.@xcite in contrast to bulk semiconductors , the authors of ref .
[ ] could record il images of agglomerates of semiconductor quantum dots .
it is suggested , that crystal defects in qds are generated at a lower rate due to the small size of the particles ( @xmath395 nm ) .
hard nuclear collisions occur deeper in the bulk , while in the first tens of nanometers electronic stopping is predominant .
the observed emission corresponds to the expected band
gap transition .
nevertheless , the signal is quenched with increasing ion fluence . while qd aggregates were detected with relative ease ,
attempts to record il images with higher magnification showing single qds were not successful .
alkali halides are known to exhibit intrinsic il as a result of defect production due to the ion beam radiation.@xcite figure[fig : fig_nacl ] demonstrates beam energy is 35kev .
fov : 25 ] simultaneously acquired images of residues from nacl solution on a metal substrate .
the bright areas in the se image ( fig .
[ fig : fig_nacl](a ) ) are the aluminum substrate .
the darker areas that exhibit a rough appearance are nacl and residue from the drop cast solution used . the il image ( fig .
[ fig : fig_nacl](b ) ) clearly shows the presence of nacl as bright ramified structures .
ogawa et al.@xcite have reported il measurements on sio@xmath40 using him .
interestingly they observe an increase of the il intensity with increasing fluence .
the authors identify several differences between cl and il spectra from the same sample , but the nature of the il signal remains unclear .
the use of il for analysis of biological samples is possible , but not very often used .
the advantage of il imaging over conventional fluorescence imaging is an absence of diffraction limits that restrict the spatial resolution . in that respect ,
application of il in him for bio imaging looks promising .
few groups have used proton beams with energies in the mev range to investigate biological samples such as cells@xcite and skin tissue.@xcite in another study several promising dyes have been identified and can now be used to stain cells specifically for il investigations.@xcite however , only limited work is done using il for imaging of biological samples .
using a classical ion source , a resolution of 150 nm has been demonstrated for ionoluminescence images in stim
mode.@xcite graphene , the prototypical thin layer that currently receives a lot of attention , has successfully been imaged by several groups.@xcite this was done for both suspended and free standing graphene . in fig .
[ fig : graphene - bubble ] a sio@xmath14 supported graphene flake has been imaged using mild imaging conditions . .
using a dose of 1.8@xmath41@xmath38 per image the formation of a he bubble can be observed . as can be seen for the bubble marked in the lower part of the images ,
holes and edges allow the he to escape and the bubble collapses . at a fluence of 2.0@xmath42@xmath38
the bubble is filled and it collapses two frames later after increasing the dose to 2.3@xmath42@xmath38 . with further increase of the dose the cycle starts again .
fov is 20 m .
ion doses in @xmath38 are indicated in the figures . ]
the flakes can be imaged easily with substantial contrast , despite their ultra thin nature . the sequence of images presented in fig .
[ fig : graphene - bubble ] shows the formation and collapse of helium filled bubbles .
graphene has been shown to be impermeable for many gases including helium,@xcite hence the backscattered helium remains trapped below the graphene sheet .
these bubbles fill with he until either the rim of the flake or a hole in the layer is reached
. this will allow the he to escape and the bubble then collapses .
this process can be observed for the central bubble between doses of 2.0@xmath42@xmath38 ( initial filling ) and 2.3@xmath42@xmath38 ( collapse and start of refilling ) .
the filling of the bubbles and subsequent deformation of the graphene sheet results in the stretching of wrinkles and folds that are present in the sheet .
assuming that the bubble has the shape of a spherical cap with a height of 100 nm and a diameter of 4 m we can calculate the pressure under the graphene sheet . using the ideal gas law and a backscatter yield of 0.006 ,
as obtained from srim@xcite , we arrive at a pressure on the order of 10 mbar .
this value agrees reasonably well with the observed pressure range for gas bubbles under graphene on sio@xmath14 obtained elsewhere.@xcite first principles calculations indicate that it should be possible to reveal the graphene lattice using him due to minute changes in the electron emission across the lattice.@xcite however , the realization of such an image might be difficult due to the destructive nature of the helium beam .
it is not clear if sufficient signal can be collected before the defects will alter the local atomic configuration.@xcite defects created in graphene due to the ion
graphene interaction have been investigated theoretically.@xcite recent experimental investigations indicate that already at very low doses severe damage is done to the graphene sheet and many defects are created.@xcite safe imaging doses for suspended graphene seem to be as low as 10@xmath43 to 10@xmath44@xmath38.@xcite for supported graphene , as shown above , the critical dose is certainly lower due to the more destructive nature of recoils from the support material .
an exceptional imaging result has been achieved for carbon nano membranes ( cnm).@xcite such free standing membranes , that are formed by cross linking self assembled monolayers and then removing the substrate , are difficult to image using conventional sem . however , in him these cnms are not only visible , but morphological details such as folds and nanometer sized holes created by slow highly charged ions , are also visible.@xcite given the nature of ultra thin films and the sampling volume of the available secondary imaging particles in him , the ses are the obvious choice for imaging such samples .
an example of the successful application of him to soft layers is the visualization of the phase separation in a mixed poly(3-hexylthiophene)/[6,6]-phenyl - c@xmath45-butric acid methyl ester ( p3ht / pcbm).@xcite such blends are typically used in organic photovoltaic solar cells and represent an important materials class .
although both molecules are essentially carbon , the different number of @xmath46 and @xmath47 bonds in the two polymers results in minute changes of the se yield.@xcite an other example is the clear identification of self assembled monolayers ( sam ) deposited onto sio@xmath14.@xcite here , differences in work function allow the identification of sam molecules and give rise to the observed contrast .
a delicate method of visualizing non continuous ultra thin layers and their different properties utilizes bshe . for a substrate oriented in a channeling condition ,
a very low bshe signal is recorded due to the extended range of he along the low index channeling direction.@xcite however , at places where a thin surface layer with different atomic positions is present , scattering will occur and an increased bshe signal can be observed .
although the relative mass of the elements in the adlayers and the underlying bulk are important , this effect works for any combination of bulk and adlayer elements.@xcite in particular it works for light elements on a heavy substrate.@xcite similar to sem , local electromagnetic fields in the sample will influence the yield , angular distribution and energy of the generated se .
this can be utilized to image dopant distribution@xcite and electronic potential distribution .
a similar application utilizes static capacitive@xcite contrast to reveal conducting features buried below an insulating cover.@xcite
the availability of different gases for gfis such as neon@xcite and the combination with a classic liquid metal gallium focused ion beam ( fib)@xcite makes the technique interesting for various types of materials modifications .
the fine beam produced by the gfis has the potential to engineer structures with a length scale that is well below what is currently possible with lmis based fib techniques .
contrary to e
beam writing , which can achieve similar critical lengths , the removal of material is also possible .
writing structures into resists is usually done using gas injection systems ( gis ) in fib or , if higher resolutions are needed , by using e beam lithography . in particular
the latter suffers from the so called proximity effect.@xcite the deposition parameters necessary to achieve a constant feature size will depend on the distance to the next feature in the proximity of the beam .
this effect is a consequence of the relatively large lateral range of the electrons in the resist material .
structures produced by low mass ion beams are less sensitive to this effect.@xcite the near complete absence of this effect is related to the fact that in the surface near region there is practically no scattering when using ion beams ( see also fig .
[ fig : charged - part - beams ] ) .
furthermore , very little backscattering occurs in the usually relatively light resist materials .
in addition the very low energy se ( < 10ev)@xcite in him are localized close to the beam path , whereas the ses in e beam lithography have a higher energy ( > 100ev ) and therefore travel relatively long distances in the resist material.@xcite as a consequence resist modification by ions happens only in a very small volume along the beam path .
an extensive overview comparing electrons and ions for their useful application in lithography is given by melngailis.@xcite with the availability of he gfis new standards have been set in terms of achievable line width.@xcite this has at least partially been made possible by the above discussed , near complete , absence of the proximity effect when using a highly focused he beam .
however , to date resists initially developed for e beam lithography such as hydrogen sylsesquioxane ( hsq ) and polymethyl methacrylate resist ( pmma ) are used .
new resist materials optimized for the large amount of energy deposited by the ion beam could potentially lead to even better results .
several precursor materials have been explored including mecpptme@xmath13 , teos , and xef@xmath14.@xcite recently , also a tungsten based metal precursors has successfully been used.@xcite with this extremely flexible method a wide range of features can be directly sculpted including 3d structures .
figure [ fig : metal - deposition ] shows examples of structures created using ion beam induced deposition ( ibid ) in the group of p. alkemade.@xcite demagnified copy of a 13th century painting of the chinese poet li bai .
( b ) stacked cross formed by exploiting shadowing of the precursor gas ( fov : 1 m ) .
( adapted with permission from scanning 34 , 90100 ( * ? ? ?
copyright 2012 wiley periodicals , inc . ) ] however , a detailed understanding of the deposition process is needed to unleash the full potential of the method for resist patterning . a model has been proposed that identifies primary ions and secondary electrons as being responsible for the vertical growth of the structures . scattered ions and recoil atoms on the other hand are responsible for the lateral growth of the structures.@xcite simulation software ( _ envision _ ) based on this model is available and
can be used to predict the metal deposition process in him.@xcite the program is based on available concepts and code of an electron beam induced simulation@xcite and the _ _ ionise__@xcite monte carlo program .
experimental confirmation of results showed the applicability of the simulation method.@xcite different from what has been discussed in the previous section , here we will discuss structuring results that do rely on the sputtering of sample atoms . sputtering is a well known phenomenon that has been extensively studied.@xcite however , with only a few exceptions these works usually focus on ions of lower energies of only a few hundred ev to a few kev for the relevant gases helium and neon .
for higher energy ranges the bombardment species is either not a noble gas or generally a heavier element .
the energy range of interest for him is not so well investigated .
surprisingly , many as yet unexplained phenomena , such as very high and fluence dependent sputter yields are observed.@xcite due to the sub nanometer size and the collimated nature of the him ion beam , it is possible to mill features with very small critical dimensions not possible with current fib technology .
we start by discussing results obtained on the most simple milled structure a pore .
such pores are successfully used for biomolecule detection .
the achievable diameter can be as small as 4 nm , which is 60% smaller compared to other one step milling techniques.@xcite most importantly the quality of these pores for biomolecule detection is comparable to state of the art pores fabricated using tem@xcite or fib.@xcite however , these techniques create smallest pores by closing existing bigger pores in a controlled way .
the method relies on the fluidization of the material under irradiation and subsequent deformation which is driven by the surface tension and diffusion of ad atoms created by an ion beam.@xcite although it has not been demonstrated to date , this should also be possible by careful implantation of he into the area surrounding the pore .
in addition to the effects observed elsewhere,@xcite an additional volume increase and subsequent reduction of pore size can occur due to the formation of nano - sized he bubbles.@xcite substantial redistribution of unsputtered material has already been observed when milling bigger pores.@xcite as a result , comparable diameters should be achievable . recently
a gfis using neon@xcite became available .
its combination with a classical liquid metal gallium fib column allows for efficient milling of structures at several length scales.@xcite figure [ fig : triple - milling ] shows a structure milled into gold using all three beams . while initial cut outs are made using the lmis gallium source ( fig .
[ fig : triple - milling](a ) ) finer details are cut using the gfis and neon ( fig . [
fig : triple - milling](b ) ) .
smallest details with a length scale of the order of 10 nm are patterned using helium and the gfis ( fig .
[ fig : triple - milling](c ) ) .
the same gas source combination is also used for imaging of the structures .
another exciting application of helium ion beams for direct write lithography is the successful preparation of devices based on graphene.@xcite the possibility to precisely cut graphene with nanometer precision has been demonstrated earlier by lemme et al.@xcite unfortunately , the unique properties of graphene are very sensitive to damage introduced by the impinging ions.@xcite consequently , only very small fluences of the order of @xmath48@xmath38 can be used to image graphene without damaging it to the extent that the unique properties are lost.@xcite unfortunately , this excludes high resolution imaging of supported as well as free standing graphene .
helium implantation is unavoidable in him .
it is therefore important to understand the associated defect structures , their evolution and response to external influences .
it has to be made clear that the phenomena discussed in this section appear at doses higher than those normally used in him imaging .
however , special applications ( bshe or il imaging ) may require relatively high doses .
obviously , he implantation plays an important role when the him is used for materials modification , as discussed in the previous section .
a dramatic example is shown in ref .
[ ] in the context of the fabrication of domain wall magnetoresistance devices .
although successful fabrication of domain wall magnetoresistance devices was demonstrated at lower doses , at higher doses the implanted he leads to a swelling and surface deformation that in turn results in the destruction of the patterned devices .
initial studies that allow the comparison of experimental damage volumes with srim were made by livengood et al.@xcite one of the results in this publication is summarized in an overview over the different damage regimes that are to be expected for a wide range of he doses ( see fig . [
fig : damage - diagram ] ) .
interestingly , this diagram is valid for both semiconductor materials ( silicon ) and metals ( copper ) .
the validity of these damage regimes has been demonstrated by several groups .
a nice visualization of the bubbles that form at depths comparable to the range of the ions in the material is shown by ref .
the pressure inside these bubbles and blisters is high . for the rather soft and ductile metal gold
, pressures of up to several gpa can be reached in the initial he nano
bubbles.@xcite these bubbles form at open volume defects that are either present in the material or are created by the beam.@xcite the growth of these initial , nano
sized , he bubbles is independent of the primary energy and leads to observable materials modification at fluences in the low 10@xmath49@xmath38 range.@xcite however , with increasing he dose these bubbles start to coalesce , provided they are formed deep enough in the sample to not immediately reach the sample surface .
what follows is a rapid expansion of a blister on the surface .
the pressure inside such a blister of a few hundred mpa@xcite is substantially lower than in the initial nano
sized bubbles .
the effective formation and rapid growth of he induced blisters is at least partly due to the fact that highly compressed he gas is an ideal stopping medium for he .
this results in an amplification of the damage in the vicinity of the bubble and the observed accelerated growth .
in the past half decade helium ion microscopy has proven to be an interesting alternative to its direct and well established competitors : scanning electron microscopy and gallium focused ion beam . however , the authors strongly believe , and hope to have demonstrated above , that him is more than a replacement for sem and fib .
it excels over sem and fib in particular for * high resolution imaging of uncoated biological samples * imaging of insulating samples * a high surface sensitivity * imaging of ultra thin layers * materials modification with unprecedented resolution * direct write lithography * resist patterning at the same time new applications and techniques are constantly being developed and refined . in particular , analytical additions are needed and are currently being developed . while ionoluminescence has its special applications , additions such as secondary ion mass spectroscopy@xcite still need to prove their applicability in real materials science problems . however , although spectroscopy is currently still in its infancy , the extension of the technique to other gases such as neon , and potentially even heavier ones , guarantees an exciting future of the technique in particular for nano fabrication applications .
141ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1116/1.2101792 [ * * , ( ) ] link:\doibase 10.1016/j.ultramic.2004.11.020 [ * * , ( ) ] link:\doibase 10.1021/nl801472n [ * * , ( ) ] link:\doibase 10.1116/1.1526694 [ * * , ( ) ] @noop _ _ , ed .
( , , ) link:\doibase 10.1116/1.2955728 [ * * , ( ) ] link:\doibase 10.1007/bf01342199 [ * * , ( ) ] link:\doibase 10.1063/1.326679 [ * * , ( ) ] link:\doibase 10.1116/1.584328 [ * * , ( ) ] link:\doibase 10.1143/jjap.31.l291 [ * * , ( ) ] http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=432639&tool=pmcentrez&rendertype=abstract [ * * , ( ) ] link:\doibase 10.1116/1.568497 [ * * , ( ) ] link:\doibase 10.1016/0167 - 9317(85)90022-x [ * * , ( ) ] link:\doibase 10.1116/1.584014 [ * * , ( ) ] link:\doibase 10.1007/bf01329651 [ * * , ( ) ] link:\doibase 10.1016/0169 - 4332(95)00351 - 7 [ * * , ( ) ] link:\doibase 10.1116/1.2357967 [ * * , ( ) ] link:\doibase 10.1007/s00339 - 004 - 2869 - 6 [ * * , ( ) ] link:\doibase 10.1016/j.ultramic.2013.03.013 [ * * , ( ) ] link:\doibase 10.1016/j.ultramic.2012.07.026 [ * * , ( ) ] link:\doibase 10.1063/1.4726112 [ * * , ( ) ] link:\doibase 10.1063/1.3702209 [ * * , ( ) ] link:\doibase 10.1016/j.ultramic.2012.11.005 [ * * , ( ) ] link:\doibase 10.1007/978 - 1 - 4615 - 0765 - 9 [ _ _ ] ( , , ) link:\doibase 10.1007/978 - 3 - 540 - 85916 - 1 [ _ _ ] , , vol .
( , , ) link:\doibase 10.1016/0169 - 4332(93)90301-q [ * * , ( ) ] link:\doibase 10.1016/j.phpro.2008.07.088 [ * * , ( ) ] in link:\doibase 10.1063/1.2799423 [ _ _ ] , vol . , ( , ) pp . in link:\doibase 10.1117/12.886347 [ _ _ ] , vol . ,
( , , ) pp .
link:\doibase 10.1016/j.ultramic.2009.01.013 [ * * , ( ) ] link:\doibase 10.1017/s1431927609090138 [ * * , ( ) ] link:\doibase 10.1116/1.3258634 [ * * , ( ) ] link:\doibase 10.1103/physrev.59.913 [ * * , ( ) ] link:\doibase 10.1016/j.nimb.2007.12.058 [ * * , ( ) ] in link:\doibase 10.1007/978 - 1 - 4615 - 2840 - 1_4 [ _ _ ] , , ( , , ) pp .
link:\doibase 10.1002/sia.2107 [ * * , ( ) ] link:\doibase 10.1103/physrevb.24.4412 [ * * , ( ) ] link:\doibase 10.1016/j.nimb.2008.11.003 [ * * , ( ) ] link:\doibase 10.1134/s1027451010050186 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.84.378 [ * * , ( ) ] link:\doibase 10.1016/j.ultramic.2013.07.010 [ * * , ( ) ] , link:\doibase 10.1016/j.scriptamat.2011.03.013 [ * * , ( ) ] link:\doibase 10.1166/jnn.2012.4260 [ * * , ( ) ] link:\doibase 10.1088/0957 - 4484/24/33/335702 [ * * , ( ) ] link:\doibase 10.1017/s1431927611005514 [ * * , ( ) ] link:\doibase 10.3762/bjnano.3.57 [ * * , ( ) ] link:\doibase 10.1116/1.3502667 [ * * , ( ) ] link:\doibase 10.1116/1.3271254 [ * * , ( ) ] in @noop _ _ ( , , ) in link:\doibase 10.1063/1.59188 [ _ _ ] ( , ) pp .
link:\doibase 10.1103/physreva.39.3836 [ * * , ( ) ] link:\doibase 10.1017/s1431927610093682 [ * * , ( ) ] in link:\doibase 10.1007/978 - 3 - 642 - 67112 - 8_5 [ _ _ ] ( , , ) chap . , pp . in link:\doibase 10.1007/978 - 3 - 540 - 87529 - 1_1 [ _ _ ] , , ( , , ) chap . , pp .
link:\doibase 10.1016/0031 - 8914(73)90154 - 7 [ * * , ( ) ] link:\doibase 10.1103/physreva.12.876 [ * * , ( ) ] link:\doibase 10.1016/0029 - 554x(65)90154 - 0 [ * * , ( ) ] link:\doibase 10.1016/0029 - 554x(78)90916 - 3 [ * * , ( ) ] link:\doibase 10.1021/ac50009a001 [ * * , ( ) ] in link:\doibase 10.1007/978 - 1 - 4615 - 5361 - 8_5 [ _ _ ] , ( , ) pp .
link:\doibase 10.1016/s0168 - 583x(97)00182 - 1 [ * * , ( ) ] in @noop _ _ , ( , , ) ed . , pp . link:\doibase 10.1149/1.2408364 [ * * , ( ) ] link:\doibase 10.1306/d4267ab9 - 2b26 - 11d7 - 8648000102c1865d [ * * , ( ) ] link:\doibase 10.1016/0022 - 3697(90)90144 - 5 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2013.02.001 [ * * , ( ) ] in link:\doibase 10.1007/bfb0104335 [ _ _ ] , , vol . , ( , , )
link:\doibase 10.1017/s1431927612013463 [ * * , ( ) ] link:\doibase 10.1088/0957 - 0233/22/2/024004 [ * * , ( ) ] link:\doibase 10.1002/sca.20267 [ * * , ( ) ] in http://www.microscopy.org/mandm/2010/jiang.pdf[__ ] ( , ) link:\doibase 10.1038/srep03514 [ * * , ( ) ] link:\doibase 10.1016/s0022 - 5320(67)80038 - 8 [ * * , ( ) ] link:\doibase 10.1016/j.microrel.2012.06.130 [ * * , ( ) ] link:\doibase 10.1016/j.surfcoat.2006.01.075 [ * * , ( ) ] link:\doibase 10.1016/j.nimb.2011.10.070 [ * * , ( ) ] link:\doibase 10.1143/jjap.37.l398 [ * * , ( ) ] link:\doibase 10.1063/1.1337646 [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/physrevb.14.2583 [ * * , ( ) ] link:\doibase 10.1080/00337577908234484 [ * * , ( ) ] link:\doibase 10.1017/s1431927612005922 [ * * , ( ) ] link:\doibase 10.1016/j.nimb.2009.03.069 [ * * , ( ) ] link:\doibase 10.1016/s0168 - 583x(01)00465 - 7 [ * * , ( ) ] link:\doibase 10.1016/j.nimb.2012.12.052 [ * * , ( ) ] link:\doibase 10.1021/nn900744z [ * * , ( ) ] , in http://meetings.aps.org/link/baps.2010.mar.h21.7[__ ] , vol .
( , , ) link:\doibase 10.1021/nn401992q [ * * , ( ) ] link:\doibase 10.1021/nl801457b [ * * , ( ) ] http://www.worldcat.org/isbn/096542071 http://books.google.com/books?vid=isbn096542071 http://www.amazon.com/gp/search?keywords=096542071&index=books&linkcode=qs http://www.librarything.com/isbn/096542071 http://www.worldcat.org/oclc/191686602[__ ] ( , ) link:\doibase 10.1103/physrevlett.109.265505 [ * * , ( ) ] link:\doibase 10.1088/0957 - 4484/22/17/175306 [ * * , ( ) ] link:\doibase 10.1063/1.4726053 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.153401 [ * * , ( ) ] link:\doibase 10.1016/j.progsurf.2012.05.001 [ * * , ( ) ] link:\doibase 10.1063/1.4792511 [ * * , ( ) ] link:\doibase 10.1021/nl202269n [ * * , ( ) ] link:\doibase 10.1002/app.1990.070391003 [ * * , ( ) ] link:\doibase 10.3762/bjnano.3.58 [ * * , ( ) ] link:\doibase 10.1021/la204437r [ * * , ( ) ] link:\doibase 10.1016/j.susc.2012.02.007 [ * * , ( ) ] link:\doibase 10.1209/0295 - 5075/85/46001
[ * * , ( ) ] link:\doibase 10.1093/oxfordjournals.jmicro.a023780 [ * * , ( ) ] link:\doibase 10.1002/sca.20268 [ * * , ( ) ] link:\doibase 10.1016/j.nima.2010.12.220 [ * * , ( ) ] link:\doibase 10.1017/s1551929512000715 [ * * , ( ) ] link:\doibase
10.1088/0957 - 4484/24/17/175302 [ * * , ( ) ] link:\doibase 10.1116/1.568515 [ * * , ( ) ] link:\doibase 10.1002/sia.740010103 [ * * , ( ) ] link:\doibase 10.1016/0168 - 583x(93)90781-z [ * * , ( ) ] link:\doibase 10.1116/1.3250204 [ * * , ( ) ] in link:\doibase 10.1117/12.862438 [ _ _ ] , vol .
, ( ) p.
link:\doibase 10.1007/978 - 3 - 7091 - 0424 - 8 [ _ _ ] , edited by and ( , , ) pp .
link:\doibase 10.1002/sca.21009 [ * * , ( ) ] link:\doibase 10.1088/0957 - 4484/21/45/455302 [ * * , ( ) ] link:\doibase 10.1017/s1551929511000307 [ * * , ( ) ] in link:\doibase 10.1007/978 - 3 - 7091 - 0424 - 8_11 [ _ _ ] , ( , , ) pp .
link:\doibase 10.1116/1.3237095 [ * * , ( ) ] link:\doibase 10.1116/1.4800983 [ * * , ( ) ] link:\doibase 10.1116/1.3517536 [ * * , ( ) ] link:\doibase 10.1088/0957 - 4484/19/41/415704 [ * * , ( ) ] link:\doibase 10.1016/0168 - 583x(87)90004 - 8 [ * * , ( ) ] link:\doibase 10.1016/0370 - 1573(81)90106-x [ * * , ( ) ] and , eds . , link:\doibase 10.1007/978 - 3 - 540 - 44502 - 9 [ _ _ ] , , vol .
( , , ) link:\doibase 10.1088/0022 - 3727/46/19/195501 [ * * , ( ) ] link:\doibase 10.1088/0957 - 4484/22/28/285310 [ * * , ( ) ] link:\doibase 10.1063/1.1947391 [ * * , ( ) ] link:\doibase 10.1038/35084037 [ * * , ( ) ] link:\doibase 10.3762/bjnano.4.53 [ * * , ( ) ] link:\doibase 10.3762/bjnano.3.67 [ * * , ( ) ] link:\doibase 10.1002/sca.21003 [ * * , ( ) ] link:\doibase 10.1116/1.3511509 [ * * , ( ) ] link:\doibase 10.1016/j.mee.2010.11.041 [ * * , ( ) ] link:\doibase 10.1016/j.carbon.2009.12.057 [ * * , ( ) ] link:\doibase 10.1088/0957 - 4484/23/39/395302 [ * * , ( ) ] link:\doibase 10.1116/1.3237101 [ * * , ( ) ] link:\doibase 10.1016/0001 - 6160(87)90034 - 4 [ * * , ( ) ] link:\doibase 10.1103/physrevb.38.1087 [ * * , ( ) ] link:\doibase 10.1063/1.4739240 [ * * , ( ) ] link:\doibase 10.1002/sca.20262 [ * * , ( ) ] link:\doibase 10.1016/j.nima.2010.12.123 [ * * , ( ) ] | helium ion microcopy ( him ) based on gas field ion sources ( gfis ) represents a new ultra high resolution microscopy and nano fabrication technique .
it is an enabling technology that not only provides imagery of conducting as well as uncoated insulating nano structures but also allows to create these features .
the latter can be achieved using resists or material removal due to sputtering .
the close to free form sculpting of structures over several length scales has been made possible by the extension of the method to other gases such as neon . a brief introduction of the underlying physics as well as a broad review of the applicability of the method is presented in this review . |
dense nano - particle systems have been shown to exhibit collective behavior , as evidenced by aging , and non - equilibrium effects similar to spin glasses.@xcite whether that collective behavior is associated with a `` true '' spin glass phase transition is still controversial . in this paper , we show a dynamic scaling analysis to a spin glass transition for a highly concentrated fe - c nano - particle sample with a narrow particle size distribution .
we also discuss why the scaling analysis may not indicate a phase transition for samples with less interactions or wider size distributions . the sample consisted of ferromagnetic nanoparticles of amorphous @xmath5 ( [email protected] ) , with an average particle size of @xmath7 nm , prepared by the method described by van wonterghem et al . in ref .
the sample was studied in the frozen state and contained 17 vol% of particles .
the ac - susceptibility measurements were performed in a non - commercial squid magnetometer for frequencies in the range @xmath8 17 mhz - 170 hz . fig . 1 shows @xmath9 and @xmath10 at different frequencies . a sample that exhibit a spin glass transition will show critical slowing down , and hence the characteristic relaxation time @xmath11 diverges at the transition temperature according to @xmath12 where @xmath13 is the transition temperature , @xmath14 is related to the relaxation time of the individual particle magnetic moments , and @xmath15 is a critical exponent .
we extracted the freezing temperature @xmath16 , associated with a relaxation time ( @xmath17 ) , from the out - of - phase component of the ac - susceptibility as @xmath18 with @xmath19 .
we also tried other criteria to make sure that the choice of criterion is not significantly influencing the results of the critical slowing down analysis .
the critical slowing down analysis gives @xmath1 , @xmath20 s , and @xmath21 k ( see fig .
we also performed a full scaling of @xmath3 according to , @xmath22 and found data collapse to a single function @xmath23 , for different frequencies , using @xmath2 ( see fig .
3 ) . the value of @xmath15 compares quite well with values found for spin glasses with long range interactions ( rkky ) , while the value of @xmath24 is slightly larger than typical spin glass values , but is consistent with the value of @xmath25 found by jonsson et al .
@xcite for an interacting nanoparticle sample by a static scaling analysis .
the dynamic scaling analysis will only reveal a phase transition if the single particle contribution to @xmath3 is vanishingly small for @xmath4 , i.e. all slow dynamics is due to collective behavior .
two criteria have to be fulfilled for this to be possible ; i ) the interparticle interactions need to be strong and ii ) the particle size distribution needs to be narrow . if we compare the sample used for the scaling analysis with the same sample but much more dilute ( 0.05 vol.% ) we can see that the out - of - phase component is almost zero for @xmath4 ( see inset of fig .
we conclude that the concentrated sample is appropriate to use for scaling analysis .
9 j. l. dormann , et al . , j. magn .
magn . mater . *
187 * , l139 ( 1998 ) .
h. mamiya , i. nakatani , and t. furubayashi , phys .
lett . * 82 * , 4332 ( 1999 ) .
p. jnsson , m. f. hansen , and p. nordblad , phys .
b * 61 * , 1261 ( 2000 ) j. van wonterghem , s , mrup , s.w .
charles , and s wells , j. colloid interface sci . *
121 * , 558 ( 1988 ) . | a highly concentrated ( 17 vol.% ) fe - c nano - particle system , with a narrow size distribution @xmath0 nm , has been investigated using magnetic ac susceptibility measurements covering a wide range of frequencies ( 17 mhz - 170 hz ) .
a dynamic scaling analysis gives evidence for a phase transition to a low temperature spin - glass - like phase .
the critical exponents associated with the transition are @xmath1 and @xmath2 .
the reason why the scaling analysis works for this sample , while it may not work for other samples exhibiting collective behavior as evidenced by aging phenomena , is that the single particle contribution to @xmath3 is vanishingly small for @xmath4 and hence all slow dynamics is due to collective behavior .
this criterion can only be fulfilled for a highly concentrated nano - particle sample with a narrow size distribution . |
the birational automorphism groups of projective varieties are extensively studied . nowadays we know that , for a projective variety of general type @xmath6 over an algebraically closed field of characteristic zero , the number of birational automorphisms of @xmath6 is bounded by @xmath7 , where @xmath8 is a constant which only depends on the dimension @xmath9 of @xmath6 , and @xmath10 is the volume of the canonical divisor @xmath11 ( cf .
furthermore , we know that @xmath12 and @xmath13 from the classical hurwitz theorem and xiao s theorem ( cf . @xcite and @xcite ) .
however , even in low dimensions , it is usually nontrivial to calculate the automorphism groups of explicit varieties of general type ( for example , see @xcite , @xcite , @xcite and @xcite ) .
we focus on automorphisms of minimal smooth complex surfaces of general type with @xmath1 and @xmath2 .
involutions on such surfaces have been studied in @xcite and @xcite .
all the possibilities of the quotient surfaces and the fixed loci of the involutions are listed . in order to find new examples ,
we have tried to classify such surfaces with commuting involutions in @xcite and succeeded in constructing a new family of surfaces in @xcite .
we briefly recall the main results in @xcite . throughout the article
, @xmath0 denotes a minimal smooth surface of general type with @xmath1 and @xmath2 over @xmath14 .
( ( * ? ? ?
* theorem 1.1 , theorem 2.9 and section 4))[thm : classification ] assume that the automorphism group @xmath15 contains a subgroup @xmath16 , which is isomorphic to @xmath3 .
let @xmath17 be the divisorial part of the fixed locus of the involution @xmath18 for @xmath19 .
then the canonical divisor @xmath20 is ample and @xmath21 for @xmath19 .
moreover , there are only three numerical possibilities for the intersection numbers @xmath22 : @xmath23 , @xmath24 and @xmath25 .
the intersection numbers @xmath26 have the following values : @xmath27 , @xmath28 and @xmath29 , respectively . in the above theorem
, we adopt the convention that @xmath30 , @xmath31 in case ( a ) and @xmath32 in case ( b ) . actually , we have completely classified the surfaces in case ( a ) and case ( b ) in @xcite .
but we do not know any example of the surfaces in case ( c ) .
one may ask whether there are noncommutative involutions on @xmath0 . here
we give a negative answer .
[ thm : center ] if @xmath33 is an involution of @xmath0 , then @xmath33 is contained in the center of @xmath15 .
we prove the above theorem in section 3 .
the key step is theorem [ thm : commutative ] which shows that any two involutions on @xmath0 commute .
theorem [ thm : commutative ] also has the following corollary .
[ cor : allinvolutions ] assume that @xmath34 is a pair satisfying the assumption of theorem [ thm : classification ]
. then @xmath35 and @xmath36 are exactly all the involutions of @xmath15 .
the corollary immediately implies that if @xmath15 contains a nontrivial subgroup which is isomorphic to @xmath37 , then @xmath38 or @xmath39 .
we remark that there are surfaces of general type with @xmath1 , @xmath40 and their automorphism groups contain subgroups which are isomorphic to @xmath41 ( cf .
* example 4.24.4 ) ) . as an application
, we calculate the automorphism groups of the surfaces in the case ( a ) of theorem [ thm : classification ] .
these surfaces are those constructed by m. inoue in @xcite and they are the first examples of surfaces of general type with @xmath1 and @xmath2 .
they can be described as finite galois @xmath3-covers of the @xmath42-nodal cubic surface ( see example [ exa : inoue ] , which is from ( * ? ? ?
* example 4.1 ) ) .
[ thm : inoueauto ] let @xmath0 be an inoue surface
. then @xmath43 or @xmath44 .
if @xmath0 is a general inoue surface , then @xmath43 .
inoue surfaces form a @xmath42-dimensional irreducible connected component in the gieseker moduli space of canonical models of surfaces of general type ( cf .
the proof of theorem [ thm : inoueauto ] actually shows that @xmath43 for @xmath0 outside a @xmath5-dimensional irreducible closed subset of this connected component ( see remark [ rem:2dim ] ) .
we also exhibit a @xmath5-dimensional family of inoue surfaces with @xmath44 ( see section 5 ) .
they are finite galois @xmath4-covers of a @xmath45-nodal weak del pezzo surface of degree two , which is the minimal resolution of one node of the @xmath46-nodal del pezzo surface of degree two .
let @xmath6 be a smooth projective surface over the complex number field .
we only consider surfaces with @xmath47 . in this case
, @xmath6 has picard number @xmath48 by noether s formula and hodge decomposition .
also the expotential cohomology sequence gives @xmath49 .
poincar duality implies that the intersection form on @xmath50 is unimodular .
assume that @xmath6 has a nontrivial automorphism @xmath51 .
denote by @xmath52 the fixed locus of @xmath51 .
let @xmath53 be the number of isolated fixed points of @xmath51 and let @xmath54 be the divisorial part of @xmath52 .
then @xmath54 is a disjoint union of irreducible smooth curves .
we denote by @xmath55 the induced linear map on the second singular cohomology group ( note that @xmath56 for @xmath57 ) .
the following proposition follows directly from the topological and holomorphic lefschetz fixed point formulae ( cf .
@xcite , page 567 ; see also ( * ? ? ?
* lemma 4.2 ) ) .
the automorphism @xmath51 is called an involution if it is of order @xmath5 .
[ prop : lefschetz ] if @xmath51 is an involution , then @xmath58 if @xmath51 is of order @xmath59 , then @xmath60 where @xmath61 is the number of isolated fixed points of @xmath51 of type @xmath62 for @xmath63 . throughout this article ,
we denote by @xmath0 a smooth minimal complex surface of general type with @xmath1 and @xmath64 .
then @xmath65 and @xmath0 contains at most one @xmath66-curve ( this follows from poincar duality ; cf .
* lemma 2.5 ) ) .
here an @xmath67-curve ( for @xmath68 ) on a smooth surface stands for an irreducible smooth rational curve with self intersection number @xmath67 . [ lem : oneinvolution](see also the table in @xcite ) let @xmath51 be an involution on @xmath0 . then @xmath69 and @xmath70 . if @xmath71 , then @xmath20 is ample and @xmath54 is irreducible with @xmath72 . for @xmath70 ,
see the proof of ( * ? ? ?
* proposition 3.6 ) . according to (
* lemma 3.2 and proposition 3.3 ( v ) ) , @xmath73 is an odd integer and @xmath74 .
so @xmath69 by proposition [ prop : lefschetz ] .
assume that @xmath71 .
if @xmath20 is not ample , then @xmath0 has a unique @xmath66-curve @xmath75 .
the intersection number matrix of @xmath76 and @xmath75 has determinant @xmath77 .
the determinant equals @xmath78 , for otherwise , the chern classes of @xmath76 and @xmath75 form a basis of @xmath79 and they are @xmath80-invariant , a contradiction to @xmath81 by proposition [ prop : lefschetz ] .
it follows that @xmath82 and @xmath83 .
this is impossible .
so @xmath20 is ample .
the algebraic index theorem gives @xmath84 and thus @xmath85 .
let @xmath86 be the quotient morphism .
we have @xmath87 . if @xmath88 , then @xmath89 and @xmath90 .
so @xmath91 has @xmath92 nodes . if @xmath93 has kodaira dimension @xmath94 , then the minimal resolution @xmath95 of @xmath93 has picard number @xmath96 and it contains @xmath92 disjoint @xmath66-curves . by (
* proposition 4.1 ) , @xmath97 is minimal .
this contradicts @xmath98 .
so @xmath99 and @xmath95 is a rational surface .
this contradicts ( * ? ? ?
* theorem 3.3 ) . hence @xmath100 . in the same manner
we see that @xmath101 ( see also @xcite ) .
so @xmath102 .
because @xmath20 is ample and @xmath103 is a disjoint union of smooth irreducible curves , the algebraic index theorem shows that @xmath103 is irreducible .
we briefly recall some facts from the theory of abelian covers from @xcite .
assume that @xmath104 is a finite abelian cover between projective varieties with @xmath6 normal and @xmath105 smooth .
let @xmath106 be the galois group of @xmath107 and let @xmath108 be the group of characters of @xmath106 . then the action of @xmath106 induces a splitting : @xmath109 , where @xmath110 and @xmath111 . for each nontrivial cyclic subgroup @xmath112 of @xmath106 and each generator @xmath113
, there is a unique effective divisor @xmath114 of @xmath105 associated to the pair @xmath115 .
the cover @xmath107 is determined by @xmath116 and @xmath114 with some specified relations ( cf .
* theorem 2.1 ) ) .
we mainly apply this theory when @xmath117 or @xmath118 .
we set up some notation and conventions .
denote by @xmath119 a group isomorphic to @xmath4 , where @xmath120 are generators of @xmath121 , @xmath122 is of order @xmath5 and @xmath123 is of order @xmath42 .
denote by @xmath124 the group of characters of @xmath121 , where @xmath125 and @xmath126 .
the group @xmath121 contains a unique subgroup @xmath16 which is isomorphic @xmath3 , where @xmath127 and @xmath128 .
denote by @xmath129 the nontrivial character orthogonal to @xmath18 for @xmath19 . when @xmath130 , we simply set @xmath131 and @xmath132 , where @xmath133 is the unique nontrivial character of @xmath134 .
similarly , when @xmath135 , we set @xmath136 for @xmath137 .
for the cyclic group @xmath138 , we set @xmath139 for @xmath140 with @xmath141 .
we adopt similar convention for the cyclic group @xmath142 . in
what follows , the indices @xmath143 should be understood as residue classes modulo @xmath59 . also linear equivalence and numerical equivalence between divisors
are denoted by @xmath144 and @xmath145 , respectively .
@xcite , ( * ? ? ?
* theorem 2.1 and corollary 3.1))[prop : abeliancover ] let @xmath104 be a finite abelian cover between projective varieties . assume that @xmath6 is normal and @xmath105 is smooth . 1 .
if the galois group of @xmath107 is @xmath146 , then @xmath107 is determined by the following data : @xmath147 where @xmath148 are divisors of @xmath105 , @xmath149 is effective and @xmath150 is reduced .
if the galois group of @xmath107 is @xmath121 , then @xmath107 is determined by the following reduced data ( see ( * ? ? ?
* proposition 2.1 ) ) : @xmath151 where @xmath152 and @xmath153 are divisors of @xmath105 , @xmath153 are effective and @xmath154
we first deduce corollary [ cor : allinvolutions ] and theorem [ thm : center ] from the following theorem .
[ thm : commutative ] let @xmath0 be a smooth minimal surface of general type with @xmath155 and @xmath64 .
assume that @xmath15 contains two distinct involutions @xmath33 and @xmath156
. then @xmath157 . on the contrary ,
suppose that @xmath33 is an involution of @xmath15 other than @xmath158 .
theorem [ thm : commutative ] implies @xmath159 .
this group contains seven subgroups of order @xmath42 , say @xmath160 .
each pair @xmath161 must satisfy one of the three numerical possibilities in theorem [ thm : classification ] .
however , this is impossible because any two of these seven subgroups have a common involution . hence @xmath35 and @xmath36 are exactly all the involutions of @xmath15 .
we may assume that @xmath15 contains at least two involutions .
these two involutions generate a subgroup @xmath162 by theorem [ thm : commutative ] .
we still denote by @xmath35 and @xmath36 the involutions of @xmath146 .
let @xmath51 be any automorphism of @xmath0 .
corollary [ cor : allinvolutions ] gives @xmath163 . since @xmath164 , we have @xmath165 and @xmath166 for @xmath19 . from this observation and theorem [ thm : classification ] , we conclude that @xmath167 for @xmath19 and complete the proof .
the remaining of this section is devoted to prove theorem [ thm : commutative ] .
* we assume by contradiction that @xmath168 . * we will deduce a contradiction through a sequence of lemmas and propositions .
we use the same notation as section 2 . recall that @xmath169 , @xmath170 and @xmath171 ( see proposition [ prop : lefschetz ] and lemma [ lem : oneinvolution ] ) .
[ lem : oddorder ] the order of @xmath172 is an odd integer .
assume by contradiction that the order of @xmath173 is @xmath174 and @xmath175 .
let @xmath176 .
then @xmath177 is an involution and @xmath178 .
therefore @xmath179 .
then @xmath180 by theorem [ thm : classification ] .
similarly , @xmath181 and @xmath182 .
so @xmath183 .
let @xmath184 .
note that @xmath185 and @xmath186 are three distinct involutions in @xmath15 and @xmath187 recall that @xmath188 .
now @xmath189 and @xmath190 are not a basis of @xmath79 , for otherwise , implies @xmath191 , which is a contradiction to @xmath192 .
so the intersection number matrix of @xmath193 and @xmath194 has determinant zero . that is @xmath195 , where @xmath196 ( see ) and @xmath197 .
it follows that @xmath198 and the nontrivial linear relation among @xmath199 and @xmath200 is @xmath201 .
this contradicts the fact that the divisor @xmath202 is strictly effective .
hence the order of @xmath173 is an odd integer .
recall that our aim is to deduce a contradiction from the assumption @xmath203 .
according to the previous lemma , from now on , * we may assume that the order @xmath204 of @xmath205 is an odd prime .
* in fact , if @xmath206 for some prime @xmath207 and some integer @xmath208 , then @xmath209 and @xmath210 are involutions and the order of @xmath211 is @xmath212 .
we may replace @xmath185 by @xmath213 and continue our discussion .
the subgroup @xmath214 of @xmath15 is isomorphic to the dihedral group of order @xmath215 .
let @xmath216 denote this subgroup .
since @xmath217 is a prime , all the involutions in @xmath216 are pairwise conjugate and @xmath216 has exactly one nontrivial normal subgroup @xmath218 , which is the commutator subgroup
. any irreducible linear representation of @xmath216 has dimension at most two , and any irreducible @xmath5-dimensional representation of @xmath216 is isomorphic to the matrix representation given by @xmath219 [ lem : nontrivialaction]with the same assumption as above , we have 1 .
the canonical @xmath20 is ample ; 2 . the curves @xmath220 and
@xmath221 generate a pencil @xmath222 of curves with @xmath223 and @xmath224 , and @xmath222 has a simple base point @xmath212 ; 3 .
the group @xmath216 acts faithfully on @xmath222 .
assume that the order of @xmath173 is @xmath225 for @xmath226 .
set @xmath227 .
then @xmath185 and @xmath177 are three distinct involutions and they are pairwise conjugate . therefore @xmath228 and @xmath229 .
since @xmath230 and @xmath231 , @xmath232 we claim that @xmath233 . otherwise , as in the proof of lemma [ lem : oddorder ] , we could deduce a contradiction by calculating the determinant of the intersection number matrix of @xmath193 and @xmath234 and by calculating @xmath235 .
then ( a ) follows from lemma [ lem : oneinvolution ] .
lemma [ lem : oneinvolution ] also gives @xmath236 .
the the algebraic index theorem implies @xmath237 and thus @xmath238 . since @xmath239 , the equality holds and @xmath240 .
similarly , we have @xmath241 let @xmath212 be the unique intersection point of @xmath242 and @xmath243 .
then implies that @xmath193 and @xmath243 pairwise intersect transversely at the point @xmath212 .
recall that @xmath244 and @xmath245 .
let @xmath67 be the smallest positive integer such that @xmath246 .
let @xmath247 be the blowup at @xmath212 , let @xmath248 be the exceptional curve and let @xmath249 be the strict transform of @xmath220 , etc .
then @xmath250 induces a fibration @xmath251 and @xmath252 and @xmath253 are fibers of @xmath254 .
the fibration @xmath254 has @xmath248 as a @xmath67-section .
if @xmath255 , we easily obtain a contradiction by applying the hurwitz formula for @xmath256 .
therefore @xmath257 , @xmath258 and @xmath259 . and
( b ) is proved . for ( c ) , first note that @xmath212 is a fixed point of @xmath216 .
so @xmath216 acts faithfully on the tangent space @xmath260 of @xmath0 to the point @xmath212 . according to the discussion before the lemma , this action is irreducible and the corresponding action of @xmath216 on @xmath261 is faithful . since @xmath223
, @xmath212 is a smooth point of @xmath262 and thus @xmath263 is a @xmath264-dimensional linear subspace of @xmath260 for any @xmath265 . from this , we conclude that @xmath216 acts faithfully on @xmath222 . because @xmath216 acts faithfully on @xmath266 , every automorphism has exactly two invariant curves in @xmath222 . for every involution @xmath267 ,
one of the two @xmath177-invariant curves in @xmath222 is @xmath243 .
denote the other one by @xmath268 .
then @xmath268 contains the seven isolated fixed points of @xmath177 .
denote by @xmath269 one of the two @xmath173-invariant curves in @xmath222 .
then the other one is @xmath270 and @xmath271 .
we shall show that @xmath269 is not @xmath5-connected .
but first we need the following lemma about the action of @xmath216 on the singular cohomology group .
[ lem : quotientpicard ] the automorphism @xmath173 acts trivially on @xmath79 . in particular , the quotient surface @xmath272 has picard number @xmath5 .
we have seen that @xmath185 and thus @xmath216 act trivially on the @xmath5-dimensional linear subspace generated @xmath273 and @xmath274 . because @xmath79 is @xmath59-dimensional and @xmath173 is contained in the kernel of any @xmath264-dimensional representation of @xmath216 , @xmath173 acts trivially on @xmath79 .
hence the invariant subspace of @xmath79 for the @xmath216-action is @xmath5-dimensional and @xmath272 has picard number @xmath5 .
we analysis the members of the pencil @xmath222 , which are not @xmath5-connected .
this will help us to determine the base locus of the linear system @xmath275 in the proof of proposition [ prop : orderthree ] and to find a basis of @xmath276 .
we continue to use the fact that @xmath0 has picard number @xmath59 .
[ lem : non2connected ] assume that @xmath222 contains a curve which is not @xmath5-connected . then 1
. the curves in @xmath222 which are not @xmath5-connected are exactly the @xmath173-invariant curves @xmath269 and @xmath277 ; 2 .
@xmath278 , where @xmath279 and @xmath280 are irreducible curves , and @xmath281 and @xmath282 .
moreover , @xmath279 contains the base point @xmath212 of @xmath222 .
assume that @xmath283 , @xmath284 , @xmath285 and @xmath286 . because @xmath20 is ample and @xmath224
, we may assume @xmath287 and @xmath288 .
then @xmath280 is irreducible .
the algebraic index theorem implies @xmath289 and @xmath290 . in particular , by lemma [ lem : quotientpicard ] , @xmath291 and thus @xmath292 .
hence @xmath293 is one of the @xmath173-invariant curves @xmath269 and @xmath277 . because @xmath294 , the argument above yields @xmath295 , @xmath296 and @xmath282 .
then @xmath297 and @xmath298 .
so the simple base point @xmath212 of @xmath222 belongs to @xmath279 .
it remains to show that @xmath279 is irreducible .
assume by contradiction that @xmath279 is reducible .
because @xmath287 and @xmath20 is ample , @xmath299 , @xmath300 and both @xmath301 and @xmath302 are irreducible .
we may assume @xmath303 and @xmath304 .
then @xmath305 and @xmath306 .
the adjunction formula gives @xmath307 or @xmath308 .
the intersection number matrix of @xmath309 and @xmath302 has determinant @xmath310 .
since the intersection form on @xmath276 is unimodular , we get @xmath307 . since @xmath223 and @xmath311
, the algebraic index theorem implies @xmath312 .
but then @xmath0 contains four disjoint curves @xmath313 and @xmath314 , all with self intersection number @xmath315 , a contradiction to @xmath65 .
the following proposition determines the order of the automorphism @xmath173 .
[ prop : orderthree ] the automorphism @xmath173 is of order @xmath59 .
moreover , @xmath269 is not @xmath5-connected , where @xmath269 is as in lemma [ lem : non2connected ] .
let @xmath262 be any curve in @xmath222 .
the long exact sequence of cohomology groups associated to the exact sequence @xmath316 shows that @xmath317 and the trace of @xmath275 on @xmath262 is complete . thus @xmath275 defines a rational map @xmath318 and @xmath319 is defined on @xmath262 whenever @xmath320 is base point free . in particular , @xmath319 is defined on the smooth curve @xmath321 and @xmath322 is the canonical image of @xmath220 .
the same statement holds by replacing @xmath33 by @xmath156 . because there is a @xmath216-linearization on @xmath323 , the rational map @xmath319 is @xmath216-equivalent .
therefore @xmath322 is contained in the fixed locus of the action of @xmath33 on @xmath324 .
note that an involution on @xmath324 has a line and a point as the fixed locus .
it follows that @xmath33 acts trivially on @xmath324 because @xmath322 is a conic curve or a quartic curve .
similarly , @xmath156 and thus @xmath216 act trivially on @xmath324 .
therefore @xmath318 factors through the quotient morphism @xmath325 .
note that @xmath20 is ample , @xmath262 is nef and @xmath326 .
first assume that @xmath319 is a morphism .
then it is finite and it has degree @xmath327 .
we thus get @xmath328 and @xmath329 .
it follows that the induced morphism @xmath330 is an isomorphism .
so the invariant linear subspace of @xmath79 for the @xmath216-action is isomorphic to @xmath331 , which is @xmath264-dimensional .
this contradicts lemma [ lem : quotientpicard ] and thus @xmath319 is not a morphism .
we now analysis the base locus of @xmath319 .
if @xmath262 is @xmath5-connected , @xmath332 is base point free by ( * ? ? ?
* theorem 3.3 ) and @xmath319 is defined on @xmath262 .
hence the base locus of @xmath275 is contained in the curves of @xmath222 , which are not @xmath5-connected .
according to lemma [ lem : non2connected ] , @xmath222 contains exactly two such curves @xmath278 and @xmath277 .
similar arguments as above show that the trace of @xmath333 ( respectively @xmath334 ) on @xmath279 ( respectively @xmath280 ) is complete .
since @xmath335 and @xmath336 , @xmath337 and @xmath338 are base point free by ( * ? ?
* theorem 3.3 ) . because @xmath339 , @xmath275 has exactly two base points @xmath340 and @xmath341 .
therefore @xmath319 is a finite morphism outside the base locus and @xmath342 .
since @xmath319 factors through @xmath272 , we have @xmath343 and @xmath344 .
it is easy to check that @xmath345 and @xmath346 generate @xmath276 and @xmath347 .
we shall show that @xmath20 is indeed linearly equivalent to @xmath348 , and deduce a contradiction to @xmath155 and complete the proof of theorem [ thm : commutative ] .
for this purpose , we turn to the quotient surface @xmath272 and analysis @xmath349 . [ prop : fixedlocus ] the automorphism @xmath173 has @xmath350 as the divisorial part of the fixed locus and it has five isolated fixed points @xmath351 and @xmath352 , where @xmath353 and @xmath354 are contained in @xmath279 .
each isolated fixed point of @xmath173 is of type @xmath355 .
we have seen that @xmath269 and @xmath277 are @xmath173-invariant and @xmath271 . moreover ,
the curves @xmath356 and @xmath357 are @xmath173-invariant .
also note that a point @xmath358 is a fixed point ( respectively an isolated fixed point ) of @xmath173 if and only if so is the point @xmath359 .
we claim that neither @xmath279 nor @xmath346 is contained in @xmath349 .
otherwise , both @xmath279 and @xmath346 are contained in @xmath349 . since @xmath360 ,
this contradicts the fact that the divisorial part of @xmath173 is a disjoint union of smooth curves .
the claim is proved .
now assume by contradiction that @xmath280 is not contained in @xmath349 .
then nor is @xmath357 and @xmath349 consists of isolated fixed points .
then @xmath349 has five fixed points by proposition [ prop : lefschetz ] and lemma [ lem : quotientpicard ] .
three of these points are @xmath361 and @xmath362 .
denote the other two by @xmath363 and by @xmath364 .
we must have @xmath365 .
otherwise , the nontrivial automorphism @xmath366 has exactly one fixed point @xmath358 , which is a smooth point of @xmath280 since @xmath282 .
this is impossible because @xmath336 .
therefore @xmath365 .
it follows that @xmath367 has exactly two fixed points @xmath212 and @xmath358 , which are smooth points of @xmath279 .
note that @xmath279 has at most two singular points since @xmath335 . because the singular locus of @xmath279 is @xmath173-invariant and @xmath367 has order @xmath59 , we conclude that @xmath279 is indeed smooth
. however , the hurwitz formula shows that @xmath367 has either one or four fixed points , a contradiction .
so @xmath280 and @xmath357 are contained in @xmath349 .
in particular , @xmath280 and @xmath357 are smooth curves .
then @xmath368 consists of five isolated fixed points and each fixed point is of type @xmath369 by proposition [ prop : lefschetz ] and lemma [ lem : quotientpicard ] .
these points must be contained in @xmath370 .
now we are able to describe the quotient map @xmath371 , where @xmath372 and @xmath373 .
the divisorial parts and isolated fixed points of cyclic subgroups of @xmath216 are as follows ( see the discussion before lemma [ lem : quotientpicard ] and proposition [ prop : fixedlocus ] ) : @xmath374 note that @xmath212 is the unique point with the stabilizer @xmath216 . from the action of @xmath216 on the tangent space @xmath260 ( see the proof of lemma [ lem : nontrivialaction ] ( c ) ) , it is easily seen that @xmath375 is a smooth point of @xmath105 .
we conclude that @xmath105 has seven nodes and two @xmath302-singularities @xmath376 and @xmath377 . in particular , @xmath105 is gorenstein .
the ramification formula gives @xmath378 and thus @xmath379 .
let @xmath380 .
then @xmath381 is contained in the smooth locus of @xmath105 .
note that @xmath381 is a smooth elliptic curve and @xmath382 .
so @xmath383 and @xmath384 .
since @xmath385 , @xmath386 by lemma [ lem : quotientpicard ] .
this implies that @xmath387 for @xmath388 .
as the quotient of @xmath0 , @xmath105 has irregularity @xmath389 . therefore @xmath105 is a rational surface .
note that linear equivalence and numerical equivalence between divisors are the same on a smooth rational surface .
since @xmath105 contains only rational double points and @xmath381 is contained in the smooth locus of @xmath105 , we have @xmath390 indeed .
then by , @xmath391 we obtain a contradiction to @xmath155 and complete the proof of theorem [ thm : commutative ] .
as mentioned in the introduction , inoue surfaces are the first examples of surfaces of general type with @xmath1 and @xmath2 ( cf .
@xcite ) . here
we describe them as finite galois @xmath3-covers of the @xmath42-nodal cubic surface , following ( * ? ? ?
* example 4.1 ) . at the end of this section , we prove theorem [ thm : inoueauto ] .
[ exa : inoue ] let @xmath392 be the blowup of the six vertices @xmath393 of a complete quadrilateral on @xmath324 . ] denote by @xmath394 ( respectively @xmath395 ) the exceptional curve of @xmath396 over @xmath397 ( respectively @xmath398 ) and denote by @xmath399 the pullback of a general line by @xmath400 .
then @xmath401 .
the surface @xmath396 has four disjoint @xmath66-curves .
they are the proper transforms of the four sides of the quadrilateral and their divisor classes are @xmath402 let @xmath403 be the morphism contracting there curves
. then @xmath91 is the @xmath42-nodal cubic surface .
let @xmath404 and @xmath405 be the proper transforms of the three diagonals of the quadrilateral , i.e. , @xmath406 for @xmath19 .
note that they are exactly the @xmath315-curves which are disjoint from any @xmath66-curve
. for each @xmath407 @xmath396 has a pencil of rational curves @xmath408 .
observe that @xmath409 for @xmath410 .
we define three effective divisors on @xmath396 @xmath411 we require that @xmath412 ( @xmath19 ) and @xmath413 are smooth @xmath78-curves such that the divisor @xmath414 has only nodes .
it is directly to show that there are divisors @xmath415 and @xmath416 satisfying in proposition [ prop : abeliancover ] .
then there is a smooth finite @xmath146-cover @xmath417 branched on the divisors @xmath418 and @xmath419 .
the ( set theoretic ) inverse image of a @xmath66-curve under @xmath420 is a disjoint union of two @xmath315-curves .
let @xmath421 be the blowdown of these eight @xmath315-curves .
then there is a finite @xmath146-cover @xmath422 such that the following diagram commutes . @xmath423
the surface @xmath0 is a smooth minimal surface of general type with @xmath155 and @xmath64 .
it is called an inoue surface . when the curves @xmath424 and @xmath425 vary , we obtain a @xmath42-dimensional family of inoue surfaces . [ lem : identityinoue ] let @xmath396 be as in example [ exa : inoue ] . 1 .
let @xmath33 be an automorphism on @xmath396 . if the induced map @xmath426 is the identity , then @xmath427 .
2 . let @xmath428 be the involution on @xmath324 such that @xmath429 for @xmath57 .
it induces an involution @xmath430 on @xmath396 .
then @xmath431 consists of the @xmath315-curve @xmath432 and three isolated fixed points @xmath433 , @xmath434 and @xmath435 , where @xmath436 is the unique smooth @xmath430-invariant curve in the pencil @xmath437 .
for ( a ) , the assumption implies that the @xmath315-curves @xmath394 and @xmath395 ( @xmath19 ) are @xmath33-invariant .
so @xmath33 comes from an automorphism on @xmath324 which has @xmath438 as fixed points and thus it is the identity morphism . for ( b ) , note that @xmath439 and @xmath440 . because @xmath441 and the divisorial part of @xmath430 is smooth ,
@xmath431 has @xmath432 as the divisorial part .
then @xmath430 has three isolated fixed points by proposition [ prop : lefschetz ] .
the point @xmath442 is an isolated fixed point of @xmath430 .
note that @xmath430 induces a nontrivial action on the pencil @xmath443 .
so @xmath443 contains exactly two @xmath430-invariant curves @xmath444 and @xmath436 . since @xmath445 and @xmath446 are also @xmath430-invariant , the intersection points @xmath434 and @xmath435 are isolated fixed points of @xmath430 .
let @xmath447 . by theorem [ thm : center ] ,
@xmath448 is @xmath51-invariant for @xmath19 and @xmath51 induces an automorphism @xmath449 on the quotient surface @xmath450 .
so the branch locus @xmath451 ( for @xmath19 ) of @xmath422 is @xmath449-invariant .
assume that @xmath452 , i.e. , @xmath453 .
the automorphism @xmath449 lifts to the minimal resolution @xmath396 of @xmath91 .
denote by @xmath33 the induced automorphism on @xmath396 .
then @xmath454 ( see the diagram ) are @xmath33-invariant because @xmath149 is the inverse image of @xmath451 under the morphism @xmath403 .
these divisors are given by .
it follows that the @xmath315-curves @xmath455 , the @xmath78-curves @xmath456 and the curves @xmath457 are @xmath33-invariant .
note that the chern classes of @xmath455 and the chern classes of @xmath458 generate @xmath79 .
the argument above implies that @xmath459 is the identity morphism .
then @xmath33 is an involution by lemma [ lem : identityinoue ] .
since @xmath460 , the fibration @xmath461 induced by @xmath462 is @xmath33-equivalent for @xmath19 . note that @xmath463 has three singular fibers @xmath464 and @xmath465 .
according to the discussion above , these three fibers are @xmath33-invariant . because any nontrivial automorphism on @xmath466 has at most two fixed points , @xmath33 respects the fibration @xmath463 , i.e. , @xmath467 .
in particular , @xmath445 and @xmath446 are @xmath33-invariant .
note that @xmath468 has three singular fibers @xmath469 and @xmath470 .
if @xmath471 , then all the @xmath66-curves @xmath472 and @xmath473 are @xmath33-invariant since @xmath33 also respects @xmath463
. then @xmath474 is the identity morphism and so is @xmath33 by lemma [ lem : identityinoue ] , a contradiction to our assumption .
so @xmath33 induces a nontrivial action on @xmath475 .
since the singular @xmath476 is @xmath33-invariant , @xmath33 must permute the other two singular fibers of @xmath468 . hence @xmath477 and @xmath478 .
similarly , by considering the action of @xmath33 on @xmath443 , we see that @xmath479 and @xmath480 .
we conclude that @xmath33 is the involution @xmath430 in lemma [ lem : identityinoue ] .
we actually prove that if @xmath481 , then @xmath482 , and in the equation , the curve @xmath425 in @xmath483 is indeed the curve @xmath436 in lemma [ lem : identityinoue ] ( b ) and @xmath484 in @xmath485 . combining with theorem [ thm : center ] and corollary [ cor : allinvolutions ]
, we complete the proof of theorem [ thm : inoueauto ] [ rem:2dim ] when the curves @xmath486 and @xmath487 vary , the inoue surfaces corresponding to the branch divisors with @xmath488 and @xmath484 form a @xmath5-dimensional irreducible closed subset of the total @xmath42-dimensional family of inoue surfaces .
also lemma [ lem : identityinoue ] shows that @xmath489 has three nodes .
moreover , it contains three @xmath66-curves in the smooth locus and these curves are the images of @xmath490 , @xmath491 and @xmath432 under the quotient map from @xmath396 to @xmath489 .
this observation motivates us to construct some special inoue surfaces in the next section .
we construct a @xmath5-dimensional family of inoue surfaces with automorphism groups isomorphic to @xmath4 .
we use the notation in subsection 2.2 .
let @xmath495 be the blowup of these points .
denote by @xmath496 ( respectively @xmath497 , @xmath498 ) the exceptional curve of @xmath105 over @xmath499 ( respectively @xmath500 , @xmath358 ) and by @xmath399 the pullback of a general line by @xmath501 .
then @xmath502 the surface @xmath105 has six disjoint @xmath66-curves and their divisor classes are : @xmath503 let @xmath504 be the proper transform of the line @xmath505 , i.e. , @xmath506 for @xmath19 .
we describe four base - point - free pencils of rational curves on @xmath105 .
they are @xmath507 and @xmath508 ( @xmath19 ) .
the singular members of @xmath509 are @xmath510 and those of @xmath511 ( fixed i ) are @xmath512 . also note that @xmath513 for @xmath19 .
now we define the following effective divisors on @xmath105 : @xmath519 we also define the following divisors : @xmath520 we require that @xmath521 and @xmath522 are smooth curves such that the divisor @xmath523 has only nodes .
these divisors satisfy in proposition [ prop : abeliancover ] .
so there is a finite galois @xmath121-cover @xmath524 and @xmath6 is normal . 1 .
the inverse image @xmath527 ( resp .
@xmath528 ) consists of two points @xmath529 and @xmath530 ( resp . @xmath531 and @xmath532 ) , each of which has stabilizer @xmath533 .
the points @xmath534 and @xmath532 are exactly the singularities of @xmath6 and they are nodes .
3 . the curve @xmath535 is a disjoint union of two irreducible smooth curves @xmath536 and @xmath537 , and @xmath538 has self intersection number @xmath539 and @xmath540 for @xmath63 .
the curve @xmath541 consists of two irreducible smooth curves @xmath542 and @xmath543 , and @xmath544 has self intersection number @xmath539 and @xmath545 for @xmath63 .
the curve @xmath546 is a disjoint union four @xmath315-curves and so is @xmath547 .
the curve @xmath548 is a @xmath315-curve .
* proposition 3.1 ) shows that @xmath6 is smooth outside @xmath527 and @xmath528 .
note that @xmath549 intersects only one irreducible component of @xmath550 ; that is @xmath551 . because @xmath552 , @xmath553 and @xmath554=2 $ ]
, we conclude that @xmath527 consists of two points , each of which has stabilizer @xmath533 .
these two points are nodes of @xmath6 according to ( * ? ? ?
* proposition 3.3 ) .
for the same reason , we have @xmath555 with @xmath556 and @xmath557 is an isomorphism .
we also have @xmath558 .
thus ( a)-(c ) follow from the discussion above .
similar arguments apply to ( d ) and ( e ) . for ( d ) , just note that @xmath559 and @xmath560 are connected irreducible components of @xmath561 . and
( e ) follows from the observation that @xmath562 intersects exactly two irreducible components of @xmath563 and @xmath564 .
now we explain how to obtain the smooth minimal model of @xmath6 . on the minimal resolution @xmath565 of @xmath6 ,
the strict transforms of @xmath566 and @xmath543 are @xmath315-curves .
each of these @xmath315-curves intersects transversely at one point with exactly one of the four @xmath66-curves over the nodes of @xmath6 .
so we can contract the four curves @xmath566 and @xmath543 of @xmath6 to smooth points on another surface .
let @xmath567 be the morphism contracting the disjoint union of the nine @xmath315-curves @xmath568 and the four curves @xmath566 and @xmath543 .
then there is a smooth @xmath121-cover @xmath569 such that the outer square of the following diagram commutes .
@xmath570 we confirm that @xmath0 is the smooth minimal model of @xmath6 by the following proposition . from ( * ? ? ?
* theorem 2.1 ) , we obtain @xmath571
. then @xmath572 by proposition [ prop : abeliancover ] ( b ) and .
let @xmath573 be the double cover branched along the four disjoint @xmath66-curves @xmath574 and @xmath575 .
note that @xmath576 is the unique character of @xmath577 which is trivial on @xmath146 .
so the galois group of @xmath578 is @xmath579 and the cover @xmath580 factors through a @xmath146-cover @xmath581 .
we have @xmath582 and @xmath583 .
the inverse images of @xmath574 and @xmath575 under @xmath578 are @xmath315-curves .
also @xmath584 is a disjoint union of two @xmath66-curves and so is @xmath585 .
let @xmath586 be the morphism contracting three @xmath315-curves @xmath587 and @xmath588 .
then @xmath589 is a weak del pezzo surface of degree three .
let @xmath590 be the morphism contracting the curves @xmath591 and @xmath541 .
we obtain a smooth galois finite @xmath146-cover @xmath592 and a commutative diagram .
the branch locus of @xmath593 is @xmath594 here we denote by @xmath595 , etc .
we claim that 1 .
@xmath596 , @xmath597 and @xmath598 are @xmath315-curves ; 2 .
@xmath599 and @xmath600 are @xmath78-curves , and @xmath601 is a disjoint union of two @xmath78-curves in the same linear system ; 3 .
@xmath602 ( @xmath603 ) is a disjoint union of two @xmath66-curves ; these two @xmath66-curves are disjoint from the @xmath315-curves in ( i ) ; 4 .
@xmath604 , @xmath605 and @xmath606 and are linearly equivalent to @xmath607 .
for example , because the general member of @xmath509 is disjoint from @xmath608 , the curve @xmath609 is a disjoint union of two @xmath78-curves in the same linear system and so is @xmath601 .
in particular , @xmath610 and @xmath611 is well defined in @xmath612 . for the @xmath315-curve @xmath613 on @xmath105 , since @xmath614 and @xmath615 , the curve @xmath616 is a @xmath66-curve , and it intersects with @xmath617 transversely at one point and it is disjoint from @xmath618 and @xmath619 .
so @xmath596 is a @xmath315-curve .
moreover , we have @xmath620 .
finally , the algebraic index theorem yields @xmath621 .
other statements can be proved in the same manner .
we may directly show that @xmath20 is ample , @xmath64 and @xmath155 for the surface @xmath0 in . according to the proof of (
* proposition 4.2 ) , we have @xmath624 it follows that @xmath625 and @xmath64 , where @xmath626 and @xmath627 are base - point - free pencils on @xmath517 induced by @xmath509 and @xmath628 .
the linear system @xmath629 is base point free , and the corresponding morphism contracts exactly the nodal curves @xmath515 and @xmath516 .
hence @xmath630 is ample and so is @xmath20 .
for each @xmath631 , we can calculate @xmath632 by ( * ? ? ? *
theorem 2.1 ) and then easily show that @xmath633 .
it follows that @xmath634 by ( * ? ? ?
* proposition 4.1 ) .
we remark that theorem [ thm : inoueauto ] contributes to the study of the moduli space of the inoue surfaces .
let @xmath635 be the gieseker moduli space of canonical models of surfaces of general type with @xmath636 and @xmath2 ( cf .
let @xmath0 be any inoue surface .
denote by @xmath637 $ ] the corresponding point in @xmath635 and by @xmath638 be the base of the kuranishi family of deformations of @xmath0 .
recall the facts that the tangent space of @xmath638 is @xmath639 , where @xmath640 is the tangent sheaf of @xmath0 , and that the germ @xmath641)$ ] is analytically isomorphic to @xmath642 .
it has been shown in @xcite that the group @xmath146 acts trivially on @xmath639 and @xmath638 is smooth of dimension @xmath42 .
now assume that @xmath0 is a special inoue surface constructed here .
we can use the same method as in the proof of ( * ? ? ? * theorem 5.1 ) to conclude that the invariant subspace of @xmath639 for the @xmath121-action has dimension @xmath5 .
note that @xmath643 and @xmath644 .
combining the result of @xcite , we see that @xmath641)$ ] is analytically isomorphic to @xmath645/(xz - y^2 ) , 0)$ ] .
the author thanks xiaotao sun for all the support during the preparation of the article . the author is greatly indebted to jinxing cai for many discussions and helpful suggestions .
this work is partially supported by the china postdoctoral science foundation ( grant no . : 2013m541062 ) .
i. bauer and f. catanese , _ inoue type manifolds and inoue surfaces : a connected component of the moduli space of surfaces with @xmath646 _ , geometry and arithmetic , 2356 , ems ser .
congr . rep .
soc . , zrich , 2012 .
f. catanese , _ singular bidouble covers and the construction of interesting algebraic surfaces _ , algebraic geometry : hirzebruch 70 ( warsaw , 1998 ) , 97120 , contemp .
* 241 * ( 1999 ) , amer .
provindence , ri , 1999 .
h. cartan , _ quotient dun espace analytique par un groupe dautomorphismes _ , a symposium in honor of s. lefschetz , algebraic geometry and topology , 90102 .
princeton university press , princeton , n. j. 1957 . | let @xmath0 be a smooth minimal complex surface of general type with @xmath1 and @xmath2 .
we prove that any involution on @xmath0 is in the center of the automorphism group of @xmath0 . as an application
, we show that the automorphism group of an inoue surface with @xmath2 is isomorphic to @xmath3 or @xmath4 .
we construct a @xmath5-dimensional family of inoue surfaces with automorphism groups isomorphic to @xmath4 . |
in mutliagent systems , agents form coalitions to perform tasks .
a useful model for analyzing how tasks can be allocated to groups of agents is the _ group activity selection problem _ ( ) , proposed by darmann et al .
@xcite . in s , participants express preferences over pairs of the form ( activity , group size ) .
the activities are then assigned to participants so as to achieve the best performance for the whole system as well as to satisfy individual agents .
the key idea behind this formulation is that ideal group size depends on the task at hand : in a company , an ideal size of the sales team may differ from that of a web developers team .
however , there is one important feature missing from the standard model , namely the _ feasibility _ of resulting groups . in many real - life scenarios ,
smooth communication among members of a group is crucial in order for different individuals to work together , and hence one needs to take into account communication structures among agents . for instance , a group of employers are unable to realize their full potential if no agent knows each other .
nevertheless , the basic framework imposes no restrictions on how agents can split into different groups .
a succinct way to capture such restrictions is to represent communication structures by undirected graphs .
this idea dates back to cooperative games with graph structure , proposed by myerson @xcite . under myerson
s proposal , nodes of the graph correspond to players and edges represent communication links between them .
recently , igarashi et al .
@xcite considered group activity selection on social networks ( ) , where groups need to be connected in their underlying social network in order to achieve certain tasks .
the focus of that work was on core stability and nash stability .
in contrast with similar settings in cooperative games @xcite , many of the computational problems were shown to be np - complete for very simple network structures ; in particular , deciding the existence of core and nash stable outcomes was shown to be np - complete even when the social network is either a path , a star , or has connected components of size at most four . motivated by the work of igarashi et al .
@xcite , we investigate the parameterized complexity of finding stable outcomes in group activity selection problems . a problem is fixed parameter tractable ( fpt ) with respect to a parameter @xmath0 if each instance @xmath1 of this problem can be solved in time @xmath2 where @xmath3 is a computable function .
we show that the problem of deciding the existence of nash stable outcomes for s on acyclic graphs is fixed parameter tractable with respect to the number of activities , thereby solving a problem left open by igarashi et al . @xcite .
for general graphs , we obtain a w[1]-hardness result , implying that this problem is unlikely to admit an fpt algorithm ; in fact , our hardness result holds even for s on cliques ( which correspond to the classic ) .
we then consider a weaker stability concept , namely , _
individual stability_. in the context of hedonic games , it is known that an individually stable solution always exists and can be computed in polynomial time when the communication structure is acyclic @xcite . in contrast , we show that an instance of may have no individually stable outcomes , even if the underlying network is a path .
we then analyze the complexity of computing individually stable solutions , and discover that from an algorithmic point of view individual stability is very similar to nash stability .
= 0.6 mm [ cols="<,<,^,^,^,^",options="header " , ] [ table ] unfortunately , our fpt results do not extend to core stability : we prove that checking the existence of core stable assignments is np - complete even for s on stars with two activities ; for classic , we can prove that this problem is hard if there are at least four activities . on the other hand ,
if there is only one activity , a core stable assignment always exists and can be constructed efficiently , for any network structure .
another parameter we consider is the number of players .
somewhat surprisingly , we show that this parameterization does not give rise to an fpt algorithm for s on general networks .
specifically , for all stability notions we consider , it is w[1]-hard to decide the existence of a stable outcome even when the underlying graph is a clique .
we summarize our complexity results in table [ table ] .
for @xmath4 where @xmath5 , let @xmath6=\{1,2,\ldots , s\}$ ] and @xmath7=\{s , s+1,\ldots , t\}$ ] .
we start by defining a group activity selection problem with a graph structure @xcite .
an instance of the _ group activity selection problem with graph structure ( ) _ is given by a finite set of _ players _ , or _
agents _ , @xmath8 $ ] , a finite set of _ activities _ @xmath9 where @xmath10 and @xmath11 is the _ void activity _ ,
a _ profile _
@xmath12 of complete and transitive preference relations over the set of _ alternatives _
\{(a_{\emptyset},1)\}$ ] , as well as a set of communication links between players @xmath14 . in
what follows , we will write @xmath15 or @xmath16 to indicate that player @xmath17 strictly prefers alternative @xmath18 to alternative @xmath19 ; similarly , we will write @xmath20 or @xmath21 if @xmath17 is indifferent between @xmath18 and @xmath19 . also , given two sets of alternatives @xmath22 and a player @xmath17 , we write @xmath23 to indicate that @xmath17 is indifferent among all alternatives in @xmath24 as well as among all alternatives in @xmath25 , and prefers each alternative in @xmath24 to each alternative in @xmath25 . two non - void activities @xmath26 and @xmath27 are said to be _ equivalent _ if for each player @xmath28 and every @xmath29 $ ] it holds that @xmath30 .
a non - void activity @xmath31 is called _ copyable _ if @xmath32 contains at least @xmath33 activities that are equivalent to @xmath26 ( including @xmath26 itself ) .
player @xmath28 is said to _ approve _ an alternative @xmath34 if @xmath35 .
when describing a player s preferences , by convention we only list the alternatives that she approves as well as @xmath36 .
an outcome of a is a _ feasible assignment _ of activities @xmath37 to players @xmath38 , i.e. , a mapping @xmath39 where for each @xmath40 the set @xmath41 of players assigned to @xmath26 is connected in @xmath42 .
note that we place no constraints on the set of players who can simultaneously engage in the void activity ( i.e. , do nothing ) . for @xmath28 with @xmath43 , we let @xmath44 denote the set of players assigned to the same activity as player @xmath28 ; for @xmath28 with @xmath45 , we set @xmath46 .
a feasible assignment @xmath39 of a is _ individually rational _ ( ir ) if each player weakly prefers her own activity to doing nothing , i.e.@xmath47 for all @xmath28 . a subset @xmath48 of players is called a _ coalition _ , and said to be _ feasible _ if it is connected in the graph @xmath49 . a feasible coalition @xmath50 and an activity @xmath31 _ strongly block _ an assignment @xmath39 if @xmath51 and @xmath52 for all @xmath53
. a feasible assignment @xmath39 of a is called _ core stable _ ( cr ) if it is individually rational , and there is no feasible coalition @xmath50 and activity @xmath31 such that @xmath54 and @xmath26 strongly block @xmath55 . given a feasible assignment @xmath39 of a , a player @xmath28 is said to have * an _ ns - deviation _ to activity @xmath31 if @xmath56 is connected , and @xmath17 strictly prefers to join the group @xmath57 , i.e. , @xmath58 . *
an _ is - deviation _ if it is an ns - deviation , and all players in @xmath57 accept it , i.e. , @xmath59 for all @xmath60 .
a feasible assignment @xmath39 of a is called _ nash stable _ ( ns ) ( respectively , _ individually stable _ ( is ) ) if it is individually rational and no player @xmath61 has an ns - deviation ( respectively , an is - deviation ) to some @xmath31 .
the first stability concept we will consider is nash stability .
the following example , due to igarashi et al .
@xcite , shows that nash stable outcomes may not exist even in very simple instances of .
[ ex : ns : empty ] _ we consider an instance of with two players : one is a loner and another is a stalker .
the loner only approves alternatives of the form @xmath62 , where @xmath63 , and the stalker only approves alternatives of the form @xmath64 , where @xmath63 . clearly , no assignment for this instance is nash stable :
if a loner engages in an activity alone , the stalker would wish to join , and if the loner and the stalker are together , the loner prefers to deviate to @xmath65 . _
igarashi et al .
@xcite show that it is np - complete to decide if a given instance of admits a nash stable outcome , even if the underlying graph @xmath42 is a path or a star ; on the other hand , they demonstrate that for paths and stars the problem of finding a nash stable outcome is fixed - parameter tractable with respect to the number of activities
. we will now show that this fpt result extends to arbitrary acyclic networks .
[ thm : fpt : tree : ns ] the problem of deciding whether an instance of @xmath66 with @xmath67 whose underlying social network @xmath42 is acyclic has a nash stable feasible assignment and finding one if it exists is in fpt with respect to @xmath68 .
we will first present a proof for the case where @xmath42 is a tree ; in the end , we will show how to extend the result to arbitrary forests .
fix an instance @xmath69 of such that @xmath42 is a tree .
we choose an arbitrary node in @xmath38 as the root of this tree , thereby making @xmath42 a rooted tree ; we denote by @xmath70 the set of children of @xmath17 and by @xmath71 the set of descendants of @xmath17 ( including @xmath17 herself ) .
we process the nodes from the leaves to the root . for each @xmath61 , each @xmath72 , each @xmath73 , each @xmath74 \cup \{(a_{\emptyset},1)\}$ ] and each @xmath75 $ ] , we let @xmath76 be _ true _ if there is an assignment @xmath77 where * the set of activities assigned to players in @xmath71 is exactly @xmath78 ; * player @xmath17 is assigned to @xmath26 and is in a coalition with @xmath0 other players ; * exactly @xmath79 players in @xmath71 belong to the same group as @xmath17 ; * the @xmath79 players in @xmath80 weakly prefer @xmath34 to @xmath81 for each @xmath82 , and have no incentive to deviate to the other groups , i.e. , every player in @xmath80 whose children do not belong to @xmath83 likes @xmath84 at least as much as each of the alternatives she can deviate to ; * the players in @xmath85 weakly prefer their alternative under @xmath55 to engaging alone in any of the activities in @xmath86 , have no ns deviation to activities in @xmath87 , and have no incentive to deviate to @xmath17 s coalition , i.e. , if @xmath88 , then every player @xmath89 whose parent belongs to @xmath83 likes @xmath90 at least as much as @xmath91 . otherwise , we let @xmath76 be _
false_. by construction ,
our instance admits a nash stable assignment if and only if @xmath92 is _ true _ for some combination of the arguments @xmath93 , and @xmath84 , where @xmath94 is the root .
if @xmath17 is a leaf , we set @xmath76 to _ true _ if @xmath95 , @xmath96 , and @xmath17 weakly prefers @xmath34 to every alternative @xmath81 such that @xmath82 ; otherwise , we set @xmath76 to _
false_. now , consider the case where @xmath17 is an internal vertex .
we first check whether @xmath17 strictly prefers some alternative @xmath97 such that @xmath82 to @xmath34 ; if so , we set @xmath76 to _ false_. otherwise , we proceed and check for each partition @xmath98 of @xmath99 whether there is an allocation of each activity set @xmath100 to some subtree rooted at @xmath17 s child that gives rise to an assignment with the conditions described above . we do this by using the color - coding technique ; we ` color ' each child of @xmath17 using colors from @xmath98 independently and uniformly at random .
suppose that @xmath55 is an assignment satisfying the properties described above where each activity set @xmath100 is assigned to @xmath101 for some @xmath102 ; we denote by @xmath54 the set of @xmath17 s children whose subtrees are assigned to some @xmath100 .
then , the probability that the subtrees associated with @xmath54 are assigned to these activities by a coloring @xmath103 chosen at random is @xmath104 , since there are @xmath105 possible colorings , and @xmath106 of them coincide with @xmath55 on @xmath54 .
we can then derandomize our algorithm using a family of @xmath0-perfect hash functions @xcite .
now , let us fix a coloring @xmath107 .
for each @xmath108 , we denote by @xmath109 the set of @xmath17 s children of color @xmath110 .
we seek to assign each subtree rooted at the @xmath111-th element of @xmath70 to the activities @xmath112 in such a way that exactly one subtree of each color @xmath108 uses all the activities in @xmath110 .
we will show that there exists an efficient algorithm that finds an assignment compatible with @xmath103 , or determines that no such assignment exists .
we will first determine for each color @xmath108 , each @xmath113 , and each @xmath114 $ ] whether the subtree rooted at @xmath111 can be assigned to the activity set @xmath110 , and exactly @xmath115 players in the subtrees of @xmath116 can be assigned to @xmath26 ; we refer to this subproblem by @xmath117 .
we initialize @xmath117 to _ true _
if * @xmath118 , and @xmath17 and @xmath111 can be _ separated _ from each other , that is , there exists an alternative @xmath119\cup \{(a_{\emptyset},1)\}$ ] such that ( i. ) @xmath120 is _ true _ , ( ii . ) @xmath121 or @xmath17 weakly prefers @xmath34 to @xmath122 , and ( iii . ) @xmath123 or @xmath111 weakly prefers @xmath124 to @xmath91 ; or * @xmath125 , and @xmath126 is _
true_. we set @xmath117 to _ false _ otherwise .
then , we iterate through all the subtrees associated with players in @xmath127 and update @xmath128 : for each @xmath129 and for @xmath130 , we set @xmath128 to _ true _
if * @xmath131 , and , moreover , there exists an @xmath132 $ ] such that both @xmath133 and @xmath134 are _ true _ ; or * both @xmath135 and @xmath128 are _ true _ , and @xmath123 or player @xmath136 weakly prefers @xmath137 to @xmath91 ; we set @xmath128 to _ false _ otherwise . in a similar manner
, we determine whether exactly @xmath115 successors of @xmath17 can be assigned to activity @xmath26 ; we denote this problem by @xmath138 .
we initialize @xmath139 to _ true _ and @xmath138 to _ false _ for each @xmath140 $ ] .
we iterate through all the colors in @xmath98 and update @xmath138 one by one ; for each color @xmath108 and each number @xmath141 , we set @xmath138 to _ true _ if there exists @xmath142 $ ] such that @xmath143 is true for some @xmath113 and @xmath144 is _ true _ , and we set @xmath138 to _ false _ otherwise .
finally , we reject the coloring @xmath103 if @xmath145 is _
, if there exists a nash stable feasible assignment that is compatible with @xmath103 , then the algorithm does not reject the coloring .
we omit the proof for the bound on the running time .
now , if @xmath42 is a forest , we can combine the algorithm described above with the algorithm for graphs with small connected components proposed by igarashi et al .
the running time of the latter algorithm is a product of the time required to find a nash stable assignment for a single connected component and the time required to combine solutions for different components ; in the analysis of igarashi et al . , the former is @xmath146 , where @xmath147 is the maximum component size and the latter is @xmath148 . in our case , each connected component is a tree , so instead of the @xmath146 algorithm for general graphs we can use our fpt algorithm for trees .
this shows that our problem is in fpt for arbitrary forests .
it is unlikely that theorem [ thm : fpt : tree : ns ] can be extended to general graphs or even cliques : our next result shows that the problem of finding a nash stable outcome is w[1]-hard with respect to the number of activities even for ` vanilla ' , i.e. , when the social network imposes no constraints on possible coalitions .
[ thm : w1:clique : ns ] the problem of determining whether an instance of admits a nash stable assignment is _
w[1]_-hard with respect to the number of activities , even if the underlying graph @xmath149 is a clique .
we reduce from clique on regular graphs , which is known to be w[1]-hard ( see e.g. theorem 13.4 in the book of cygan @xcite ) . given a graph @xmath150 and an integer @xmath0 , where @xmath151 , @xmath152 , and each vertex of @xmath153 has degree @xmath154 , we create an instance of whose underlying graph is a clique , as follows .
we define the set of activities as @xmath155 where @xmath156 , @xmath157 .
notice that @xmath158 . for each @xmath159 , we create a vertex player @xmath160 , and for each edge @xmath161 , we create two edge players @xmath162 and @xmath163 .
let @xmath164\}$ ] , and let @xmath165 be a bijection that assigns a distinct number in @xmath24 to each vertex @xmath166 .
note that @xmath167 implies that the intervals @xmath168 $ ] and @xmath169 $ ] are disjoint . similarly , let @xmath170\}$ ] and let @xmath171 be a bijection that assigns a distinct number in @xmath25 to each edge @xmath172 .
for each @xmath166 we construct a set of @xmath173 dummy players @xmath174 ; similarly , for each @xmath172 we construct a set of @xmath175 dummy players @xmath176 .
we let @xmath177 finally , we create five additional players @xmath178 and let @xmath179 intuitively , @xmath180 and @xmath181 will form two instances of the stalker game , with @xmath182 used to stabilize the latter .
let @xmath183 be the set of pairs of distinct players of @xmath38 .
note that @xmath184 depends on @xmath0 , but not on @xmath33 , and the size of our instance of is bounded by @xmath185 .
we will now define the players preferences .
each vertex player @xmath159 and the players in @xmath174 approve each alternative in @xmath186 $ ] .
each edge player @xmath163 approves each alternative in @xmath186 $ ] as well as each alternative in @xmath187 , whereas its dummies only approve the alternatives in @xmath187 .
all of these players are indifferent among all alternatives they approve .
the stabilizer @xmath182 approves each alternative of the form @xmath188 with @xmath189 , @xmath190 $ ] and is indifferent among them ; she also approves @xmath191 , but likes it less than all other approved alternatives .
also , we have @xmath192 we will now argue that the graph @xmath153 contains a clique of size @xmath0 if and only if there exists a nash stable assignment for our instance of .
suppose that @xmath153 contains a clique @xmath54 of size @xmath0 .
we construct an assignment @xmath55 as follows .
we establish a bijection @xmath193 between @xmath54 and @xmath194 , and for each @xmath166 we form a coalition of size @xmath195 that engages in @xmath196 : this coalition consists of @xmath160 , all players in @xmath174 , and all edge players @xmath163 such that @xmath197 .
also , we establish a bijection @xmath198 between the edge set @xmath199 and @xmath78 , and assign the activity @xmath200 to the edge players @xmath162 , @xmath163 as well as to all players in @xmath176 .
finally , we set @xmath201 , and assign the void activity to the remaining players .
we will now argue that the resulting assignment @xmath55 is nash stable . clearly
, no player assigned to an activity in @xmath194 or @xmath202 wishes to deviate .
now , consider a player @xmath203 with @xmath204 ; by construction , @xmath160 only wants to join a coalition if it engages in an activity in @xmath194 and its size is in the interval @xmath205 $ ] , and no such coalition exists .
the same argument applies to players in @xmath174 .
similarly , consider an edge player @xmath163 with @xmath206 .
we have @xmath207 , and therefore @xmath163 does not want to join any of the existing coalitions ; the same argument applies to all dummies of @xmath163 .
further , @xmath208 and @xmath209 do not want to deviate since each activity @xmath189 is assigned to at least three players , and players @xmath210 and @xmath211 do not want to deviate since they are allocated one of their top choices . finally , the stabilizer @xmath182 does not want to deviate , since there is no coalition of size @xmath212 $ ] that engages in an activity @xmath213 .
hence , @xmath55 is nash stable .
conversely , suppose that there exists a nash stable feasible assignment @xmath55 .
notice that @xmath55 can not allocate an activity @xmath189 to @xmath208 or @xmath209 , or leave it unallocated , since no such assignment can be nash stable .
thus , each activity in @xmath194 is allocated to a coalition whose size lies in the interval @xmath214 $ ] for some @xmath159 .
further , nash stability implies that @xmath55 allocates @xmath18 to @xmath210 , @xmath211 and @xmath182 .
now , if some activity @xmath189 is assigned to @xmath215 players , where @xmath216 $ ] for some @xmath159 , the stabilizer @xmath182 would then deviate to that coalition ; hence , for each @xmath213 we have @xmath217 for some @xmath166 .
now , let @xmath218 . by construction , @xmath219 .
we will show that @xmath54 is a clique .
consider a player @xmath220 , and let @xmath221 be the activity assigned to @xmath195 players under @xmath55 .
we have @xmath222 , since otherwise @xmath160 could deviate to @xmath26 ; similarly , all players in @xmath174 are assigned to @xmath26 .
the only other players who approve @xmath223 are edge players @xmath163 .
thus , @xmath224 such players must be assigned to @xmath221 , and the remaining @xmath225 of these players must be assigned to some activity @xmath226 , since otherwise they would deviate to @xmath221 .
now , consider an edge player @xmath163 with @xmath227 for some @xmath226 . by individual rationality
we have @xmath228 and hence @xmath229 consists of @xmath163 , @xmath162 and all dummies of the edge @xmath230 . for each @xmath220 , consider the set of players @xmath231 .
the size of this set is @xmath232 , and we have argued that all players in such sets need to be assigned to activities in @xmath202 .
however , @xmath157 , and each activity in @xmath202 can be assigned to at most two edge players simultaneously .
thus , @xmath233 . as each edge player @xmath234 can belong to at most two sets @xmath235 , it follows that it belongs to exactly two such sets , namely , @xmath236 and @xmath237 .
thus , both @xmath238 and @xmath239 are in @xmath54 , i.e. , @xmath54 is a clique , which is what we wanted to prove . on the positive side , for s on cliques , we can place the problem of finding a nash stable assignment in the complexity class xp with respect to the number of activities .
however , it is not clear if this result can be extended to general s. [ thm : xp : ns ] there exist an algorithm that , given an instance @xmath69 of with @xmath240 , @xmath241 such that @xmath42 is a clique , determines whether it admits a nash stable assignment in time @xmath242 . for every mapping
@xmath243 $ ] , we will check if there is a nash stable assignment such that @xmath244 for each @xmath40 .
there are at most @xmath245 such mappings ; hence , it remains to show that each check will take at most @xmath246 steps .
fix a mapping @xmath243 $ ] .
we construct an instance of the network flow problem as follows .
we introduce a source @xmath215 , a sink @xmath79 , a node @xmath17 for each player @xmath28 , and a node @xmath26 for each activity @xmath40 .
we create an arc with unit capacity from the source @xmath215 to each player , and an arc with capacity @xmath247 from node @xmath40 to the sink @xmath79 .
then , for each @xmath61 we create an arc of unit capacity from player @xmath17 to an activity @xmath40 if and only if @xmath17 weakly prefers @xmath248 to @xmath249 and to all pairs of the form @xmath250 , where @xmath251 .
it can be easily verified that an integral flow of size @xmath252 in this network corresponds to a nash stable assignment where exactly @xmath247 players are engaged in each activity @xmath40 .
it remains to note that one can check in polynomial time whether a given network admits a flow of a given size .
we will next consider a less stringent stability requirement , namely , individual stability .
igarashi and elkind @xcite showed that in the context of hedonic games , acyclicity is sufficient for individually stable outcomes to exist : an individually stable partition of players always exists and can be computed in polynomial time . in contrast , it turns out that for s this is not the case : an individually stable outcome may fail to exist even if the underlying social network is a path ; moreover , this may happen even if there are only three players and their preferences are strict .
[ ex : empty : is ] _ consider a with @xmath253 , @xmath254 , @xmath255 , where players preferences are as follows : @xmath256 _ we will argue that each individually rational feasible assignment @xmath55 admits an is - deviation . indeed ,
if @xmath257 then no player is engaged in @xmath147 and hence player @xmath258 can deviate to @xmath147 . similarly ,
if @xmath259 then no player is engaged in @xmath27 and hence player @xmath260 can deviate to @xmath27 .
there are @xmath261 individually rational feasible assignments where @xmath262 , @xmath263 ; for each of them we can find an is deviation as follows ( we write @xmath264 to indicate that player @xmath17 has an is - deviation to activity @xmath18 ) : 1 .
@xmath265 , @xmath266 , @xmath267 : @xmath268 ; 2 .
@xmath269 , @xmath266 , @xmath267 : @xmath270 ; 3 .
@xmath269 , @xmath266 , @xmath271 : @xmath272 ; 4 .
@xmath273 , @xmath274 , @xmath271 : @xmath275 ; 5 .
@xmath273 , @xmath266 , @xmath276 : @xmath270 ; 6 .
@xmath273 , @xmath266 , @xmath271 : @xmath272 ; 7 .
@xmath273 , @xmath277 , @xmath267 : @xmath270 ; 8 .
@xmath273 , @xmath277 , @xmath271 : @xmath278 ; 9 .
@xmath273 , @xmath277 , @xmath279 : @xmath280 . in example
[ ex : empty : is ] all activities are distinct . on the other hand ,
if all activities are copyable , an individually stable outcome is guaranteed to exist .
moreover , we can adapt the result of igarashi and elkind @xcite for hedonic games and obtain an efficient algorithm for computing an individually stable outcome .
[ thm : is : copyable ] each instance of where each activity @xmath31 is copyable and @xmath49 is acyclic admits an individually stable feasible assignment ; moreover , such as assignment can be found in polynomial time .
the algorithm is similar to the one for hedonic games @xcite .
the basic idea is to create a rooted tree for each connected component and construct an assignment for every subtree in a bottom - up manner . when considering a subtree rooted at player @xmath17 , we start with the assignment obtained by combining the previously constructed assignments for the subtrees rooted at the children of @xmath17 and assigning @xmath17 to the void activity .
we then let player @xmath17 join the most preferred activity among those to which she has an is deviation . after that we keep adding players to @xmath17 s coalition as long as the resulting coalition remains feasible , the player being added is willing to move , and such a deviation is acceptable for all members of @xmath17 s coalition . similarly to the proof in theorem 1 of igarashi and elkind @xcite , one can show that the resulting assignment is individually stable .
in contrast , when activities are not copyable , finding an individually stable feasible assignment is hard , even for very simple social networks . indeed , for every class of simple social networks for which igarashi et al .
@xcite show that finding a nash stable outcome is np - complete , finding an individually stable outcome is np - complete as well .
the proof of the following theorem is similar to the proofs of the respective results of igarashi et al .
@xcite : essentially , we have to replace each instance of the stalker game in these proofs with an instance of the game from example [ ex : empty : is ] .
[ thm : np : is ] given an instance of whose underlying graph is a path , a star , or has connected components whose size is bounded by a constant , it is _
np_-complete to determine whether it has an individually stable feasible assignment .
moreover , the problem of finding individually stable feasible assignments remains hard even if the number of activities is small , and even if the social network is a clique .
the proof if similar to that of theorem [ thm : w1:clique : ns ] and is omitted due to space constraints .
[ thm : w1:clique : is ] the problem of determining whether an instance of admits an individually stable assignment is _
w[1]_-hard with respect to the number of activities , even if the underlying graph @xmath149 is a clique .
again , we reduce from clique on regular graphs . given a graph @xmath150 and an integer @xmath0 where @xmath151 , @xmath152 , and each vertex of @xmath153 has degree @xmath154 , we create an instance of on a clique as follows .
first , we create the same sets
@xmath281 and @xmath78 of non - void activities as in the proof of theorem [ thm : w1:clique : ns ] ; we then create five other activities @xmath282 , @xmath283 , @xmath284 , @xmath285 , and @xmath286 . the set of non - void activities is given by @xmath287 notice that @xmath288 .
also , we create the same set @xmath289 of players as in the proof of theorem [ thm : w1:clique : ns ] ; we create two copies @xmath290 and @xmath291 of the empty is instance together with the stabilizer @xmath292 for the latter ; we let @xmath293 and @xmath183 be the set of pairs of distinct players of @xmath38 . for players in @xmath289 ,
each vertex player @xmath159 and its dummies approve any activity @xmath189 with size within the interval @xmath214 $ ] ; each edge player @xmath163 approves any activity @xmath189 with size within the same interval @xmath214 $ ] as well as any activity @xmath294 with size @xmath295 , whereas its dummies only approve the activities in @xmath78 with size @xmath295 .
the stabilizer @xmath292 approves any activity @xmath189 with size within @xmath296 $ ] and the alternative @xmath297 , and strictly prefers the former alternatives to the latter . for players in @xmath298 , their top alternative is @xmath297 ; however , their preference over the alternatives of @xmath284 , @xmath285 , and @xmath286 is cyclic and given as follows : @xmath299 hence , in an individually stable assignment @xmath55 , the activity @xmath284 should be assigned to all the players in @xmath298 together with its stablizer @xmath292 ; otherwise @xmath55 can not be individually stable as we have seen in example [ ex : empty : is ] .
similarly , the preference for players in @xmath300 is cyclic and satisfies the following relation : for any @xmath189 , @xmath301 notice that in an individually stable assignment @xmath55 , all the activities in @xmath281 must be used outside of @xmath302 , since otherwise @xmath55 can not be individually stable . similarly to the proof for nash stability , one can show that @xmath153 contains a clique of size @xmath0 if and only if there exists an individually stable feasible assignment .
suppose that there exists a clique of size @xmath0 .
we construct an assignment @xmath55 as follows . for each vertex @xmath303
, we choose unused activity @xmath189 and form a coalition of size @xmath195 by assigning @xmath26 to the vertex player @xmath160 , all the dummies of @xmath160 , and the edge players @xmath163 whose endpoint @xmath238 does not belong to @xmath54 . for each edge
@xmath161 such that @xmath304 , we choose unused activity @xmath294 and assign @xmath27 to the edge players @xmath162 , @xmath163 , and all the dummies of @xmath305 .
we assign @xmath284 to all the players in @xmath298 and the player @xmath292 , and assign @xmath282 to all players in @xmath302 .
finally , we assign the void activity to the rest of the players .
we will now show that the resulting assignment @xmath55 is an individually stable .
clearly , no assigned vertex @xmath160 and edge players @xmath163 as well as their assigned dummies wish to deviate .
no unassigned vertex @xmath160 and edge players @xmath163 as well as their unassigned dummies want to deviate , since neither alternative of the form @xmath306 where @xmath189 and @xmath307 $ ] nor alternative of the form @xmath308 where @xmath294 is played among the other players ; no players in @xmath300 wish to deviate since each activity @xmath309 is already assigned to more than @xmath260 players ; clearly , the players in @xmath298 do not want to deviate since they are assigned to their favorite alternative ; lastly , the stabilizer @xmath292 can not deviate since no alternative of the form @xmath306 where @xmath309 and @xmath310 $ ] for some @xmath159 is played .
hence , @xmath55 is individually stable .
conversely , suppose that there exists an individually stable feasible assignment @xmath55 .
notice that @xmath55 can not allocate an activity @xmath189 to a member of @xmath300 , since otherwise no feasible assignment would be individually stable ; hence , all activities in @xmath281 should be assigned to the outside of @xmath302 .
the only individual rational way to do this is to form @xmath0 coalitions of size within the interval @xmath214 $ ] for some @xmath159 and allocate each activity @xmath189 to each of the @xmath0 coalitions .
further , @xmath55 should allocate activity @xmath284 to the players in @xmath298 and the stabilizer @xmath215 , since otherwise @xmath55 would not be individually stable .
if activity @xmath189 is assigned to more than @xmath195 players for some @xmath159 , the stabilizer @xmath292 would then deviate to the coalition ; hence , the @xmath0 coalitions assigned to @xmath26 are of size exactly @xmath195 for some @xmath159 .
now , let @xmath54 be the set of vertex players @xmath160 whose target size @xmath195 is achieved by some coalition .
clearly , @xmath219 .
we will show that @xmath54 is a clique of size @xmath0 .
observe that for each @xmath303 whose corresponding coalition of size @xmath195 is assigned to @xmath189 , @xmath160 should be assigned to @xmath26 , since otherwise the vertex player would have an is deviation to @xmath26 ; likewise , all dummies of @xmath160 must be assigned to @xmath26 ; and @xmath311 edge players @xmath163 must be allocated to @xmath26 , and the other @xmath312 edge players @xmath163 together with @xmath162 and @xmath175 of @xmath305 s copies must be assigned to some activity @xmath27 .
since the activities in @xmath78 can cover at most @xmath313 edge players in an individually rational way , this means that all the activities in @xmath78 should be played among exactly @xmath313 edge players @xmath163 whose left endpoint @xmath160 belongs to @xmath54 .
hence , we can not assign @xmath294 to edge players @xmath163 such that @xmath303 but @xmath314 , since if there exists such a player , this would imply that @xmath315 for some @xmath189 and @xmath314 .
it follows that for each vertex player @xmath220 and edge player @xmath163 such that @xmath227 for some @xmath294 , the other endpoint @xmath238 also belongs to @xmath54 ; thus , each @xmath303 is adjacent to @xmath225 distinct vertices in @xmath54 , thereby concluding that @xmath54 is a clique .
however , just as in the case of nash stability , if we both restrict the structure of the social network and assume that the number of activities is small , we can obtain positive algorithmic results .
the proof of the next theorem is similar to that of theorem [ thm : fpt : tree : ns ] ; again , we omit it due to space constraints .
[ thm : fpt : tree : is ] the problem of deciding whether an instance of @xmath66 with @xmath67 whose underlying social network @xmath42 is acyclic has an individually stable feasible assignment and finding one if it exists is in fpt with respect to @xmath68 .
again , given a tree @xmath49 , we choose an arbitrary node as the root and construct a rooted tree by orienting the edges in @xmath183 towards the leaves ; we denote by @xmath70 the set of children of @xmath17 and by @xmath71 the set of descendants of @xmath17 ( including @xmath17 ) .
we process the nodes from the leaves to the root . for each @xmath28 , each @xmath316 , each @xmath317 , each @xmath74 \cup \{(a_{\emptyset},1)\}$ ] , and each @xmath318 $ ] , we let @xmath76 be _ true _ if there exists a feasible assignment @xmath39 where * the set of activities assigned to players in @xmath71 is exactly @xmath78 ; * player @xmath17 is assigned to @xmath26 and is in a coalition with @xmath0 other players ; * exactly @xmath79 players in @xmath71 belong to the same group as @xmath17 ; * the @xmath79 players in @xmath319 weakly prefer @xmath34 to @xmath81 for each @xmath82 , and has no is - deviation to the activities assigned to players in @xmath320 ; and * the players in @xmath85 weakly prefer their alternative under @xmath55 to engaging alone in any of the activities in @xmath86 , have no is - deviation to activities in @xmath87 . otherwise , we let @xmath76 be _
false_. we let @xmath321 be _ true _ if there exists a feasible assignment @xmath322 with the above properties where @xmath123 or some player in @xmath319 strictly prefers @xmath34 to @xmath91 ; we let @xmath321 be _ false _ otherwise .
we let @xmath323 be _ true _ if there exists a feasible assignment @xmath322 with the above properties where @xmath123 or no player in @xmath324 strictly prefers @xmath91 to their alternative under @xmath55 ; we let @xmath323 be _ false _ otherwise . by construction ,
our instance admits an individually stable assignment if and only if there exists a combination of the arguments @xmath325 , @xmath78 , and @xmath34 such that @xmath326 is _ true _ or @xmath327 is _ true _
, where @xmath94 is the root .
if @xmath17 is a leaf , we set both @xmath76 and @xmath323 to _ true _
if @xmath95 , @xmath96 , and @xmath17 weakly prefers @xmath34 to any other alternative @xmath81 such that @xmath82 ; otherwise , we set both @xmath76 and @xmath323 to _ false_. we set @xmath321 to _ true _ if @xmath76 is true , and @xmath123 or @xmath17 strictly prefers @xmath34 to @xmath91 ; otherwise , we set @xmath321 to _
false_. now consider the case where @xmath17 is an internal vertex .
we first check whether @xmath17 strictly prefers some alternative @xmath97 such that @xmath82 to @xmath34 ; if so , we set @xmath76 , @xmath321 , and @xmath323 to _
false_. otherwise , we proceed and check for each partition @xmath98 of @xmath99 whether there is an allocation of each activity set @xmath100 to some subtree rooted at @xmath17 s child that gives rise to an assignment with the conditions described before .
again , we use the color - coding technique ; we ` color ' each child of @xmath17 using colors from @xmath98 independently and uniformly at random , and check whether or not the coloring gives rise to a desirable outcome .
now , let us fix a coloring @xmath107 .
for each @xmath108 , we denote by @xmath109 the set of @xmath17 s children of color @xmath110 .
we seek to assign each subtree rooted at the @xmath111-th element of @xmath70 to the activities @xmath112 in such a way that exactly one subtree of each color @xmath108 uses all the activities in @xmath110 .
we will show that there exists an efficient algorithm that finds an assignment compatible with @xmath103 , or determines that no such assignment exists . to begin with , for each color @xmath108 , each @xmath113 , and each @xmath328 $ ] , we will run the following three checks : first , whether the subtree rooted at @xmath111 can be assigned to the activity set @xmath110 , and exactly @xmath115 players in the subtrees of @xmath116 can be assigned to @xmath26 ; second , whether the subtree rooted at @xmath111 can be assigned to the activity set @xmath110 , exactly @xmath115 players in the subtrees of @xmath116 can be assigned to @xmath26 , and @xmath123 or some of the @xmath115 players strictly prefers @xmath34 to @xmath91 ; third , whether the subtree rooted at @xmath111 can be assigned to the activity set @xmath110 , exactly @xmath115 players in the subtrees of @xmath116 can be assigned to @xmath26 , and no player in @xmath329 has an incentive to deviate to @xmath17 s coalition ; we refer to these subproblems by @xmath117 , @xmath330 , and @xmath331 , respectively .
we initialize @xmath117 to _ true _
if * @xmath118 , and there exists an alternative @xmath332 \cup \{(a_{\emptyset},1)\}$ ] such that @xmath333 , @xmath17 weakly prefers @xmath34 to @xmath334 , or @xmath335 is true ; or * @xmath336 , and @xmath126 is true .
we set @xmath117 to _ false _ otherwise .
we initialize @xmath331 to _ true _
if * @xmath118 , and there exists an alternative @xmath332\cup \{(a_{\emptyset},1)\}$ ] such that @xmath17 can not deviate to @xmath111 s coalition without hurting the member of the coalition , i.e. , @xmath121 , @xmath17 weakly prefers @xmath34 to @xmath122 , or @xmath335 is true , and @xmath111 has no incentive to deviate , i.e. , @xmath123 or @xmath111 weakly prefers @xmath124 to @xmath91 ; or * @xmath336 , and @xmath337 is true .
we set @xmath331 to _ false _ otherwise .
we initialize @xmath330 to _ true _
if @xmath336 and @xmath338 is true , and initialize @xmath330 to _ false _ otherwise .
then , we iterate through all the subtrees associated with players in @xmath127 and update @xmath128 , @xmath339 , and @xmath331 : for each @xmath129 and for @xmath130 , we set @xmath128 to _ true _ if @xmath131 and there exists @xmath132 $ ] such that both @xmath134 and @xmath133 are true ; or both @xmath128 and @xmath135 are true .
likewise , we set @xmath340 to _ true _ if @xmath341 and there exists @xmath132 $ ] such that both @xmath342 and @xmath343 are true ; or both @xmath340 and @xmath344 are true , and if @xmath88 , then player @xmath136 weakly prefers @xmath137 to @xmath91 .
we set @xmath339 to _ true _
if @xmath131 and there exists @xmath132 $ ] such that both @xmath134 and @xmath133 are true , and at least one of @xmath345 and @xmath346 is true ; or both @xmath330 and @xmath135 are true . finally , we will run the following three checks : first , whether exactly @xmath115 players in @xmath71 can be assigned to @xmath26 ; second , whether exactly @xmath115 players in @xmath71 can be assigned to @xmath26 , and @xmath123 or some of the @xmath115 players strictly prefers @xmath34 to @xmath91 ; third , whether exactly @xmath115 players in @xmath71 can be assigned to @xmath26 , and no player in @xmath329 has an incentive to deviate to @xmath17 s coalition .
we denote these subproblems by @xmath138 , @xmath347 , and @xmath348 , respectively .
we initialize both @xmath139 and @xmath349 to _ true _ , and @xmath138 to _ false _ for each @xmath350 $ ] .
we initialize @xmath351 to _ true _
if @xmath123 or player @xmath17 strictly prefers @xmath34 to @xmath91 , and @xmath351 to _ false _ otherwise ; we set @xmath347 to _ false _ for each @xmath350 $ ] .
we iterate through all the colors in @xmath98 and update @xmath138 , @xmath347 , and @xmath348 : for each color @xmath108 and each number @xmath141 , * we set @xmath138 ( respectively , @xmath348 ) to _ true _ if there exists @xmath352 $ ] such that @xmath143 ( respectively , @xmath353 ) is true for some @xmath113 and @xmath144 ( respectively , @xmath354 ) is true , and we set @xmath138 ( respectively , @xmath348 ) to _ false _ otherwise ; and * we set @xmath347 to _ true _ if there exists @xmath352 $ ] such that @xmath355 is true for some @xmath113 and @xmath144 is true , or @xmath143 is true for some @xmath113 and @xmath356 is true , and we set @xmath347 to _ false _ otherwise . for @xmath357 ( respectively , @xmath358 and @xmath359 )
, we reject the coloring , if @xmath145 ( respectively , @xmath360 and @xmath361 ) is false .
if there exists a feasible assignment that is compatible with @xmath103 and satisfies respective conditions , then the algorithm does not reject the coloring .
we omit the proof for the bound on the running time .
the results for individual stability presented so far indicate that from the complexity perspective it is very similar to nash stability .
however , it is not clear if the xp algorithm presented in theorem [ thm : xp : ns ] extends to individual stability .
the difficulty is that , to determine whether an agent @xmath17 has an is deviation to an activity @xmath26 , it is not sufficient to know how many players engage in @xmath26 : knowing their preferences is important to decide whether @xmath17 s deviation will be vetoed by one of the players currently assigned to @xmath26 .
another important difference concerns copyable activities and games on acyclic graphs : in this setting , individually stable outcomes always exist ( theorem [ thm : is : copyable ] ) , whereas for nash stable outcomes this is not the case , as illustrated by the stalker game .
igarashi et al .
@xcite have demonstrated that the core can be empty even for s on paths with @xmath362 players and @xmath260 activities .
we reproduce their example below .
[ ex : core : empty ] _ consider a with @xmath253 , @xmath363 , @xmath255 , where agents preferences @xmath12 are as follows : @xmath364 it can be shown that this instance admits no core stable assignment @xcite . _ on the positive side , we can show that checking whether a given feasible assignment is core stable is easy , irrespective of the structure of the social network .
the proposition below generalizes theorem 11 of darmann @xcite and theorem 12 of igarashi et al .
@xcite .
[ prop : in - core ] given an instance @xmath365 of and a feasible assignment @xmath55 for that instance , we can decide in polynomial time whether @xmath55 is core stable .
let @xmath366 and let @xmath367 . by scanning the assignment @xmath55 and the players preferences
, we can check whether @xmath55 is individually rational . now , suppose that this is the case .
then , for each @xmath63 and each @xmath368 $ ] we can check if there is a deviation by a connected coalition of size @xmath215 that engages in @xmath26 . to this end
, we consider the set @xmath369 of all players who strictly prefer @xmath188 to their assignment under @xmath55 and verify whether @xmath369 has a connected component of size at least @xmath215 that contains @xmath57 ; if this is the case , @xmath57 ( which is itself connected or empty ) could be extended to a connected coalition of size exactly @xmath215 that strongly blocks @xmath55 .
if no such deviation exists , @xmath55 is core stable . however , core stability turns out to be more computationally challenging that nash stability and individual stability when the number of activities is small : we will now show that core stable assignments are hard to find even if there are only two activities and the underlying graph is a star ( and thus one can not expect an fpt result with respect to the number of activities for this setting ) .
later , we will see that this hardness result can be extended to the case where @xmath42 is a clique , i.e. , to classic , thereby solving a problem left open by the work of darmann @xcite .
[ thm : np : star : core ] it is _
np_-complete to determine whether an instance of admits a core stable assignment even when the underlying graph is a star and the number of non - void activities is @xmath260 .
our problem is in np by proposition [ prop : in - core ] . to establish np - hardness ,
we reduce from the np - complete problem hitting set @xcite .
an instance of hitting set is a family @xmath370 of subsets of a finite set @xmath371 and an integer @xmath0 with @xmath372 .
it is a `` yes''-instance if @xmath373 admits a _ hitting set _
@xmath374 of size at most @xmath0 , i.e. , @xmath375 and @xmath376 for each @xmath377 ; otherwise , it is a `` no''-instance . given an instance of hitting set , we can create three disjoint copies of @xmath378 : for each @xmath166 we create elements @xmath379 , @xmath380 and @xmath381 , set @xmath382 and for each @xmath383 $ ] we let @xmath384 , @xmath385 , @xmath386 and set @xmath387\}$ ] . in
what follows , we will use the fact that @xmath388 admits a hitting set of size @xmath389 if and only if @xmath378 admits a hitting set of size @xmath362 .
consider an instance @xmath390 of hitting set and construct the pair @xmath388 as described above .
for readability , we renumber the elements of @xmath391 as @xmath392 .
we construct an instance of as follows .
we define the set of activities to be @xmath363 .
we introduce a center player @xmath147 , two players @xmath393 and @xmath394 , and a player @xmath239 for each @xmath395 . also , for each @xmath396 $ ] , we let @xmath397 and create a set @xmath398 of dummy players .
we then attach @xmath393 , @xmath394 , each player @xmath399 and each of the dummies to the center .
formally , the graph @xmath49 is given by @xmath400 intuitively , for each @xmath396 $ ] the number @xmath401 is the target coalition size when all players in @xmath402 are engaged in activity @xmath27 , together with @xmath147 and the players in @xmath403 .
the agents preferences over alternatives are defined as follows .
we let @xmath404 $ ] . for each @xmath395 , we let @xmath405 ; also , set @xmath406
. the preferences of each player @xmath395 are given by @xmath407 for each @xmath396 $ ] the dummy players in @xmath403 only approve the alternative @xmath408 .
the preferences of the center player @xmath147 are given by @xmath409 finally , the preferences of players @xmath393 and @xmath394 are given by @xmath410 note that the preferences of @xmath147 , @xmath393 and @xmath394 , when restricted to @xmath411 $ ] , form an instance of with an empty core ( example [ ex : core : empty ] ) .
we will show that @xmath378 admits a hitting set of size at most @xmath0 if and only if there exists a core stable feasible assignment .
let @xmath412 be a hitting set of size at most @xmath0 in @xmath378 .
set @xmath413 ; we have @xmath414 , and , by construction , @xmath415 is a hitting set for @xmath388 .
then , we construct a feasible assignment @xmath55 by assigning activity @xmath26 to @xmath147 , @xmath393 , @xmath394 , and the players @xmath416 , and assigning the void activity to the remaining players . clearly , @xmath55 is individually rational , since @xmath417 . further ,
notice that no connected subset @xmath54 together with activity @xmath26 strongly blocks @xmath55 : every such subset has to contain players @xmath393 and @xmath394 , who are currently enjoying one of their top alternatives .
it remains to show that no connected subset @xmath54 together with activity @xmath27 strongly blocks @xmath55 .
suppose towards a contradiction that such a subset @xmath54 exists ; as @xmath393 and @xmath394 are not interested in deviating , it must be the case that @xmath418 for some @xmath419 $ ] and hence @xmath54 consists of agents who approve @xmath420 , i.e. , @xmath421 for some @xmath396 $ ] .
however , since @xmath415 is a hitting set , there is an agent @xmath422 with @xmath423 , and this agent prefers @xmath424 to @xmath420 , a contradiction .
hence , @xmath55 is core stable .
conversely , suppose that there exists a core stable feasible assignment @xmath55 and let @xmath425 .
we will first argue that @xmath426 . indeed , if @xmath427 , then @xmath428 and agents @xmath147 , @xmath393 and @xmath394 can deviate to @xmath26 . if @xmath429 and the only other agent to engage in @xmath27 is @xmath393 , then @xmath428 and @xmath147 and @xmath394 can deviate to @xmath26 .
if @xmath429 and @xmath430 for some @xmath431 $ ] then @xmath428 , and @xmath147 , @xmath393 and @xmath394 can deviate to @xmath26 .
it follows that @xmath426 .
now , if @xmath432 then @xmath433 and agent @xmath394 can deviate to @xmath27 . similarly , if @xmath434 then @xmath433 , and @xmath147 and @xmath393 can deviate to @xmath27 .
if follows that @xmath435 $ ] and hence @xmath436 ; also , either @xmath437 or @xmath433 .
now , if @xmath415 is not a hitting set for @xmath388 , there exists a set @xmath438 such that all players in @xmath402 are assigned to the void activity . then if @xmath433 the coalition @xmath439 together with the activity @xmath27 strongly blocks @xmath55 , and if @xmath437 the coalition @xmath440 together with the activity @xmath27 strongly blocks @xmath55 . in either case
, we obtain a contradiction with the stability of @xmath55 .
thus , @xmath441 is a hitting set for @xmath388 .
at least one of the three sets @xmath442 , @xmath443 , and @xmath444 contains at most @xmath0 elements ; assume without loss of generality that @xmath445 . by construction
, @xmath446 is a hitting set for @xmath378 .
thus , @xmath378 admits a hitting set of size at most @xmath0 , which is what we wanted to prove .
the hardness result in theorem [ thm : np : star : core ] immediately generalizes to instances of with more than two activities : we can modify the construction in our hardness reduction by introducing additional activities that no player wants to engage in .
also , theorem [ thm : np : star : core ] can be extended from stars to cliques ; however , our proof for cliques relies on having at least four non - void activities .
it remains an interesting open problem whether core stable outcomes of s on cliques can be found efficiently if the number of activities does not exceed @xmath362 .
we conjecture that the answer is ` no ' , i.e. , the problem of computing core stable outcomes remains np - hard for @xmath447 .
[ thm : np : clique : core ] it is _
np_-complete to determine whether an instance of admits a core stable assignment even when the underlying graph is a clique and the number of activities is @xmath448 .
in contrast , checking the existence of core stable assignments in is easy if @xmath449 , irrespective of the structure of the social network .
[ prop : core - single ] every instance of with @xmath450 admits a core stable assignment ; moreover , such an assignment can be computed in polynomial time .
consider an instance @xmath365 of with @xmath450 , and let @xmath367 .
for each @xmath368 $ ] , let @xmath451 be the set of all players who weakly prefer @xmath188 to @xmath249 .
if for each @xmath368 $ ] the largest connected component of @xmath451 with respect to @xmath42 contains fewer than @xmath215 agents , then no outcome in which a non - empty set of agents engages in @xmath26 is individually rational , whereas the assignment @xmath55 given by @xmath452 for all @xmath61 is core stable . otherwise , consider the largest value of @xmath215 such that @xmath451 has a connected component of size at least @xmath215 .
find a connected subset of @xmath451 of size exactly @xmath215 , and assign the agents in this subset to @xmath26 ; assign the remaining agents to @xmath65 . to see that this assignment is core stable , note that for every deviating coalition @xmath54 we would have @xmath453 , and hence such a coalition is either disconnected or some players in @xmath54 prefer @xmath249 to @xmath454 .
it is not clear if the problem of finding core stable assignments in is in fpt with respect to the number of activities when the underlying graph is a path .
however , we can place this problem in xp with respect to this parameter . in fact , we have the following more general result for graphs with few connected coalitions ( see the work of elkind @xcite for insights on the structure of such graphs ) .
[ prop : core - xp ] given an instance @xmath365 of with @xmath240 , @xmath241 , such that the number of non - empty connected subsets of @xmath42 is @xmath455 , we can decide in time @xmath456 whether this instance admits a core stable assignment , and find one such assignment if it exists .
let @xmath457 be the list of all non - empty connected subsets of @xmath42 .
since each assignment of players to activities has to assign a connected ( possibly empty ) subset of players to each activity , we can bound the number of possible assignments by @xmath458 : each of the @xmath68 non - void activities is assigned to a coalition in our list or to no player at all , and the remaining players are assigned the void activity .
we can then generate all these assignments one by one and check if any of them is core stable ; by proposition [ prop : in - core ] , the stability check can be performed in time polynomial in @xmath33 and @xmath68 .
if the social network @xmath42 is a path , it has @xmath459 connected subsets .
thus , we obtain the following corollary .
[ cor : core - xp - path ] the problem of deciding whether a given instance of whose underlying graph is a path admits a core stable assignment is in _ xp _ with respect to the number of activities .
in the previous sections , the parameter that we focused on was the number of activities @xmath68 .
although we expect this parameter to be small in many realistic settings , there are also situations where players can choose from a huge variety of possible activities .
it is then natural to ask if stability - related problems for s are tractable in the number of players @xmath33 is small .
this is the question that we consider in this section .
we first observe that for all stability concepts considered in this paper the problem of finding a stable feasible assignment is in xp with respect to @xmath33 : we can simply guess the activity of each player ( there are at most @xmath460 possible guesses ) and check whether the resulting assignment is feasible and stable .
[ obs : xpplayers ] the problem of deciding whether a given instance of admits a core stable , nash stable , or individually stable assignment is in _ xp _ with respect to the number of players @xmath33 .
the following theorem shows that is unlikely to be fixed - parameter tractable with respect to @xmath33 .
this is somewhat surprising , because an fpt algorithm with respect to @xmath33 could afford to iterate through all possible partitions of the players into coalitions .
interestingly , in the following hardness reduction there is a unique partition of players into coalitions that can lead to a stable assignment ; that is , computational hardness comes solely from assigning known coalitions of players to activities .
[ thm : w1players : clique ] the problem of determining whether an instance of whose underlying graph is a clique admits a core stable ( nash stable , or individually stable ) assignment is _
w[1]_-hard with respect to the number of players @xmath33 . due to space constraints
, we only provide a proof for nash stability .
we describe a parameterized reduction from the w[1]-hard multicolored clique problem .
an instance of this problem is given by an undirected graph @xmath150 , a positive integer @xmath461 , and a vertex coloring @xmath462 .
it is a ` yes'-instance if @xmath153 admits a colorful @xmath463-clique , that is , a size-@xmath463 vertex subset @xmath464 such that the vertices in @xmath465 are pairwise adjacent and have pairwise distinct colors . without loss of generality
, we assume that there are exactly @xmath466 vertices of each color for some @xmath467 , and that there are no edges between vertices of the same color .
let @xmath468 be an instance of multicolored clique with @xmath469 .
for every @xmath470 $ ] , we write @xmath471 to denote the set of vertices of color @xmath17 . for each vertex @xmath160
, we write @xmath472 to denote the set of edges incident to @xmath160 .
we construct our instance as follows .
we have one _ vertex activity _
@xmath473 for each vertex @xmath474 , @xmath470 $ ] , one _ edge activity _
@xmath475 for each edge @xmath476 , two _ color activities _
@xmath477 and @xmath478 for each color @xmath470 $ ] , and one _ colorpair activity _
@xmath479 for each color pair @xmath480 , i \neq j$ ] .
_ we will have one _ color gadget _ for each color @xmath470 $ ] and one _ colorpair gadget _ for each color pair @xmath481 .
for most of the possible assignments , these gadgets will be unstable , unless the following holds : 1 .
for each color @xmath470 $ ] the first two players from the color gadget select a vertex of color @xmath17 ( by being assigned together to the corresponding vertex activity ) .
2 . for each color pair @xmath482 ,
i \neq j$ ] the first two players of the colorpair gadget select an edge connecting one vertex of color @xmath17 and one vertex of color @xmath111 ( by being assigned together to the corresponding edge activity ) .
every selected edge for the color pair @xmath483 must be incident to both vertices selected for color @xmath17 and color @xmath111 .
if the three conditions above hold , then the assignment must encode a colorful @xmath463-clique . _ construction details . _ the color gadget @xmath70 , @xmath484 $ ] consists of the following four players . @xmath485
the colorpair gadget @xmath486 , @xmath487 , i \neq j$ ] , consists of three players .
players @xmath488 and @xmath489 approve all edge activities corresponding to edges between vertices of color @xmath17 and @xmath111 with size two and the colorpair activity @xmath479 with size two and one , and strictly prefer the former alternatives to the latter .
player @xmath490 approves only the colorpair activity @xmath479 with size three . _
_ we will now argue that the graph @xmath153 admits a colorful clique of size @xmath463 if and only if our instance of admits a nash stable assignment .
suppose that there exists a colorful clique @xmath491 of size @xmath463 .
assign the first two players of the color gadget @xmath70 to the activity corresponding to the vertex of color @xmath17 from @xmath491 .
assign the first two players of the colorpair gadget @xmath486 to the activity corresponding to the edge between the vertices of color @xmath17 and @xmath111 in @xmath491 .
assign all other players to the void activity .
clearly , the resulting assignment is feasible .
observe that for a successful nash deviation a player must join an existing non - empty coalition , because no player prefers a size - one activity to the currently assigned one . by construction , the last two players of each color gadget and the last player of the colorpair gadget can not deviate ( no other players engage in an approved activity ) .
consider the first two players of color gadget @xmath70 .
they can not deviate to a vertex activity , because their current activity is their only approved vertex activity that has some players assigned to it .
they can not deviate to an edge activity either , because they would only prefer edge activities corresponding to edges that are not incident to the vertex of color @xmath17 that is part of colorful clique @xmath491 ; these activities , however , have no players assigned to them .
finally , the first two players of a colorpair gadget @xmath486 can not deviate since no player is assigned to any of their more preferred activities .
thus , we have a nash stable feasible assignment .
suppose that there exists a nash stable feasible assignment @xmath55 .
observe that the first two players of each colorpair gadget @xmath486 are assigned to the same edge activity corresponding to some edge @xmath492 .
otherwise , they would need to be assigned to the colorpair activity @xmath479 and be stalked by the third player of the gadget ( similarly to the stalker game from example [ ex : ns : empty ] ) .
we say that these players `` select edge @xmath493 '' . by similar reasoning the first two players of each color gadget @xmath70
are assigned to the same vertex activity corresponding to some vertex @xmath494 .
they can not be assigned to an edge activity , because the third player would necessarily be a player from a colorpair gadget , and these players do not want to engage in activities with two other players .
we say that these players `` select vertex @xmath494 '' .
( note that at this point , the coalition structure of any stable outcome is already fixed : the first two players of each color gadget and of each colorpair gadget must form a coalition of size two , respectively , and all other players must be assigned to the void activity . ) now , assume towards a contradiction that the selected vertices and edges do not form a colorful clique of size @xmath463 .
the size and colorfulness are clear from the construction .
hence , there must be some pair @xmath495 of selected vertices that is not adjacent .
however , this would imply that colorpair gadget @xmath483 selected an edge that is either not incident to vertex @xmath494 or not incident to vertex @xmath496 . without loss of generality
let it be non - adjacent to vertex @xmath494 .
now , there are two cases .
first , if @xmath497 with @xmath498 , then player @xmath499 would deviate to the activity corresponding to @xmath493 .
second , if @xmath497 with @xmath500 , then player @xmath501 would deviate to the activity corresponding to @xmath493 . in both cases we have a contradiction to the assumption that @xmath55 is nash stable . note that although we showed w[1]-hardness for each of the parameters @xmath68 and @xmath33 , parameterizing by the combined parameter @xmath502 immediately gives fixed - parameter tractability , since the input size is trivially upper - bounded by @xmath503 .
we have investigated the parameterized complexity of computing stable outcomes of group activity selection problems on networks , with respect to two natural parameters .
many of our hardness results hold for the classic problem , where there are no constraints on possible coalitions ; however , some of our positive results only hold for acyclic graphs .
interestingly , one of our tractability results holds for s , but it is not clear if it can be extended to s ; thus , while simple networks may decrease complexity , allowing for arbitrary networks may have the opposite effect . in some of our hardness reductions
players have non - strict preferences . in some cases , it is easy to modify our gadgets to rely on strict preferences only ; in other cases ( notably , in individual stability proofs ) our arguments make crucial use of indifferences . confirming our intuition , we showed that finding core stable assignments tends to be computationally more difficult than finding nash stable or individually stable assignments ; we note that this distinction only becomes visible from the parameterized complexity perspective , as all of these problems are np - hard .
our results for games with few players are somewhat preliminary : we have only investigated the complexity of our problems with respect to this parameter for general networks and for cliques , and have not considered simple networks such as stars and paths .
these questions offer an interesting direction for future work .
a. darmann , e. elkind , s. kurz , j. lang , j. schauer , and g. woeginger . group activity selection problem . in
_ proceedings of the 8th international conference on internet and network economics _ , wine 2012 , pages 156169 , 2012 .
a. igarashi and e. elkind .
hedonic games with graph - restricted communication . in _ proceedings of the 15th international conference on autonomous agents and multiagent systems
_ , aamas 2016 , pages 242250 , 2016 .
a. igarashi , d. peters , and e. elkind .
group activity selection on social networks . in _ proceedings of the 31st aaai conference on artificial intelligence _ , aaai 2017 , 2017 . accepted for publication .
long version available as arxiv:1611.04524 [ cs.gt ] . | in group activity selection problem ( ) , players form coalitions to participate in activities and have preferences over pairs of the form ( activity , group size ) . recently ,
igarashi et al .
@xcite have initiated the study of group activity selection problems on social networks ( ): a group of players can engage in the same activity if the members of the group form a connected subset of the underlying communication structure .
igarashi et al .
have primarily focused on nash stable outcomes , and showed that many associated algorithmic questions are computationally hard even for very simple networks . in this paper
we study the parameterized complexity of with respect to the number of activities as well as with respect to the number of players , for several solution concepts such as nash stability , individual stability and core stability .
the first parameter we consider in the number of activities . for this parameter
, we propose an fpt algorithm for nash stability for the case where the social network is acyclic and obtain a w[1]-hardness result for cliques ( i.e. , for classic ) ; similar results hold for individual stability .
in contrast , finding a core stable outcome is hard even if the number of activities is bounded by a small constant , both for classic and when the social network is a star .
another parameter we study is the number of players .
while all solution concepts we consider become polynomial - time computable when this parameter is bounded by a constant , we prove w[1]-hardness results for cliques ( i.e. , for classic ) . |
in the last years there have been several attempts to understand the diverse aspects of cosmology , as the presence of stable vacua and inflationary conditions , in the framework of supergravity and string theory @xcite .
one of the most interesting features emerging from these type of models consists on the study of the consequences of higher dimensional degrees of freedom on the cosmology derived from four - dimensional effective theories @xcite .
furthermore , it has been pointed out that the presence of extra dimensions leads to an interesting connection with the ekpyrotic model @xcite , which generated considerable activity @xcite . the essential ingredient in these models
( see for instance @xcite ) is to consider an effective action with a graviton and a massless scalar field , the dilaton , describing the evolution of the universe , while incorporating some of the ideas of pre - big - bang proposal @xcite in that the evolution of the universe began in the far past .
on the other hand , it is well known that relativistic theories of gravity such as general relativity or string theories are invariant under reparametrization of time .
the quantization of such theories presents a number of problems of principle known as `` the problem of time '' @xcite .
this problem is present in all systems whose classical version is invariant under time reparametrization , leading to its absence at the quantum level .
therefore , the formal question involves how to handle the classical hamiltonian constraint , @xmath0 , in the quantum theory .
also , connected with the problem of time is the `` hilbert space problem '' @xcite referring to the not at all obvious selection of the inner product of states in quantum gravity , and whether there is a need for such a structure at all . the above features , as it is well known , point out the necessity to construct a consistent theory of gravity at quantum level .
one promising attempt is string theory where the structure of the internal space plays a relevant role in the construction of interesting effective models containing cosmological features as the presence of positive valued minima of the scalar potential and slow - roll inflationary conditions .
the usual procedure for that is to consider compactification on generalized manifolds , on which internal fluxes have back - reacted , altering the smooth calabi - yau geometry and stabilizing all the moduli @xcite . in the present work we shall consider an alternative procedure about the role played by the moduli . in particular we shall not consider the presence of fluxes , as in string theory , in order to obtain a moduli - dependent scalar potential in the effective theory .
rather , we are going to promote some of the moduli to time - dependent fields by considering the particular case of a ten - dimensional gravity coupled to a time - dependent dilaton compactified on a 6-dimensional torus with a time - dependent khler modulus . with the propose to track down the role
play by such fields , we are going to ignore the dynamics of the complex structure field ( for instance , by assuming that it is already stabilized by the presence of a string field in higher scales ) .
the work is organized in the following form . in section
ii we present the construction of our effective action by compactification on a time - dependent torus , while in section iii we study its lagrangian and hamiltonian descriptions using as toy model the bianchi type ii cosmological model and we present the general structure for all bianchi class a models .
section iv is devoted on finding the corresponding classical solutions for few different cases involving the presence or absence of matter content . in section
v we present some solutions to the corresponding wheeler - dewitt ( wdw ) equation in the context of standard quantum cosmology , resulting the modified bessel function @xmath1 , solutions that in some sense are similar to this find in previous work @xcite . in order that the unnormalized probability density @xmath2 does not diverge for @xmath3 and at fixed @xmath4 , which are the gravitational variables , we drop the function @xmath5 , remaining only the function @xmath6 .
also in this section we make an analysis for particular case in the constants in the solution found in the asymptotic behavior for large argument in the gravitational variables and it is compare with the bohm s solutions obtained , finding that this corresponds to lowest - order wkb approximation .
the bohm s solutions was obtained in other decomposition of the quantum wdw equation , related with the bohm s formalism , which was applied in the supersymmetric quantum cosmology @xcite . finally our conclusions are presented in section vi .
we start from a ten - dimensional action coupled with a dilaton ( which is the bosonic component common to all superstring theories ) , which after dimensional reduction can be interpreted as a brans - dicke like theory @xcite . in the string frame ,
the effective action depends on two time - dependent scalar fields : the dilation @xmath7 and the khler modulus @xmath8 , where the initial high - dimensional ( effective ) theory is given by @xmath9 + 4\,\hat{g}^{mn}\nabla_{m}\,\phi\,\nabla_{n}\,\phi\bigg ] , \label{action}\ ] ] with a metric described by @xmath10 where @xmath11 are the indices of the 10-dimensional space , and greek and latin indices correspond to the external and internal space , respectively .
the internal metric is given by @xmath12 in order to rewrite it in the einstein frame , we take as usual , the conformal transformation of the four dimensional metric @xmath13 as @xmath14 where @xmath13 is the metric in the string frame and @xmath15 is the metric in einstein frame and the dilaton has been redefined accordingly as @xmath16 .
notice that in this way , the effective dilaton @xmath17 is also a time - dependent field . following this notion , we have that the action reads @xmath18 where we have omitted the upper script @xmath19 in the metric .
now with the expression we proceed to build the lagrangian and the hamiltonian of the theory at the classical regimen employing the anisotropic cosmological bianchi type ii model .
in the last section we have built the reduced effective action , now we will construct the classical hamiltonian from the expression .
we are going to assume that the background of the extended space ( four dimensional ) is a bianchi type ii anisotropic cosmological model . in order to do it ,
let us recall here the canonical formulation in the adm formalism of the diagonal bianchi class a cosmological models .
the metric has the form @xmath20 where @xmath21 is the lapse function , @xmath22 is a @xmath23 diagonal matrix , @xmath24 , @xmath25 and @xmath26 are scalar functions known as misner variables , @xmath27 are one - forms that characterize each cosmological bianchi type model @xcite , and obey the form @xmath28 , with @xmath29 the structure constants of the corresponding model .
the one - forms for the bianchi type ii model are @xmath30 , @xmath31 , @xmath32 .
so , the corresponding metric of the bianchi type ii in misner s parametrization has the form @xmath33 substituting @xmath34 , @xmath35 , @xmath36 , we have the following form @xmath37 the lagrangian density as effective theory for this model , including an intuitive way the matter content for a barotropic perfect fluid as a first approximation , has the structure @xcite @xmath38 where @xmath39 is the parameter that characterize the epochs in the evolution to the universe .
the lagrangian that describes this model is given by @xmath40 .
\label{lagra - i}\ ] ] + computing the momenta associated to the moduli fields @xmath41 and misner variables @xmath42 and using the legendre transformation we obtain the hamiltonian density for this model @xmath43 . \label{ham}\ ] ] we can introduce a new set of variables that involving the misner variables @xmath44 in the gravitational part , @xmath45 with the last set of new variables , the langrangian density can be transformed as @xmath46 and the hamiltonian density in the new variables is @xmath47 -\frac{1}{2}e^{-(\beta_{1 } + \beta_{2 } + \beta_{3})}e^{4\beta_1}. \label{hami}\end{aligned}\ ] ] in the following sections we will analyze each one of the case that involve the terms in our classical hamiltonian .
we will split our analysis in two subsections .
so far , we have built the hamiltonian density of the lagrangian in terms of a new set of variables , but we need to analyze the behavior of each one of the cases , with and without matter , these corresponds to the vacuum case @xmath48 and stiff fluid @xmath49 .
we start our analysis with 1 . _
vacuum case : @xmath50 .
_ + we can see that the classical hamiltonian can be written as + @xmath51\nonumber\\ & -\frac{1}{2}e^{-(\beta_{1 } + \beta_{2 } + \beta_{3})}\,e^{4\beta_{1}}. \label{htwovac}\end{aligned}\ ] ] + by choosing time - gauge condition on the lapse function @xmath52 , the hamilton s equation associated to the hamiltonian are given by + @xmath53 + from the equations we have that the momenta associated to @xmath54 and @xmath55 are given by + @xmath56 + given the above restrictions , we obtain a differential equation for @xmath57 + @xmath58 + where @xmath59 , @xmath60 , and @xmath61 .
the solution to the differential equation is + @xmath62 + where @xmath63 is an integration constant .
+ with this in mind , we see from the , and that + @xmath64 + integrating the last expression , the solution associated to @xmath65 is + @xmath66 + from the expressions , , and we can see that the reparametrization of the misner variables are given by + @xmath67 2 . _
stiff fluid case : @xmath68 . _
+ in this case we have that the hamiltonian is the expression with @xmath68 , so + @xmath69 -\frac{1}{2}e^{-(\beta_{1 } + \beta_{2 } + \beta_{3})}e^{4\beta_1 } , \qquad \rho_1=128\,\pi\,g\ , c_1 \label{hami-1}\end{aligned}\ ] ] + and the hamilton equations are given by the same equations found in the before case , , and , so the solutions become the same , only suffer change in the constant @xmath70 where is necessary to put the constant @xmath71 for the content matter in the stiff scenario .
the volume of the universe for this cosmological model become as @xmath72^{\frac{1}{2 } \frac{\sqrt{\nu^2 + 4\beta}}{\sqrt{\nu^2 + 4\beta } - 4d}},\ ] ] which has a growing trend for any time .
we find in the literature a lot of works on the wheeler - dewitt ( wdw ) equation dealing with different problems , for example in @xcite , they asked the question of what a typical wave function for the universe is . in ref .
@xcite there appears an excellent summary of a paper on quantum cosmology where the problem of how the universe emerged from big bang singularity can no longer be neglected in the gut epoch .
on the other hand , the best candidates for quantum solutions become those that have a damping behavior with respect to the scale factor , represented in our model with the @xmath4 parameter , in the sense that we obtain a good classical solution using the wkb approximation in any scenario in the evolution of our universe @xcite . the wdw equation for this model
is achieved by replacing the momenta @xmath73 , associated to the misner variables @xmath74 and the moduli fields @xmath41 in the hamiltonian .
the factor @xmath75 may be factor ordered with @xmath76 in many ways .
hartle and hawking @xcite have suggested what might be called a semi - general factor ordering which in this case would order @xmath77 as @xmath78 where q is any real constants that measure the ambiguity in the factor ordering in the variable @xmath4 and the corresponding momenta .
we will assume in the following this factor ordering for the wheeler - dewitt equation , which becomes @xmath79\psi = 0 , \label{wdw}\ ] ] where @xmath80 is the three dimensional dlambertian in the @xmath81 coordinates , with signature @xmath82 .
on the other hand , we could interpreted the wdw equation as a time - reparametrization invariance of the wave function @xmath83 . at a glance
, we can see that the wdw equation is static , this can be understood as the problem of time in standard quantum cosmology .
we can avoid this problem by measuring the physical time with respect a kind of time variable anchored within the system , that means that we could understand the wdw equation as a correlation between the physical time and a fictitious time @xcite .
+ when we introduce the ansatz @xmath84 in , we obtain the general set of differential equations ( under the assumed factor ordering ) , @xmath85\ , \theta & = & 0,\label{wdwmod}\\ \frac{48}{11}\,\frac{\partial^2 \phi}{\partial \phi^2 } + \frac{1}{11}\,\frac{\partial^2 \phi}{\partial \sigma^2}-\frac{18}{11}\,\frac{\partial^2 \phi}{\partial\phi \partial \sigma } + \mu^2 \phi&=&0 . \label { phi-1}\end{aligned}\ ] ] the solution to the moduli fields corresponds to the hyperbolic partial differential equation , given by @xmath86 where @xmath87 are integration constants and they are in terms of @xmath88 .
we claim that this solutions is the same for all bianchi class a cosmological models , because the hamiltonian operator in ( [ wdw ] ) can be written in separated way as @xmath89 , where @xmath90 y @xmath91 represents the hamiltonian to gravitational sector and the moduli fields , respectively .
we can rewrite the expression in terms of the misner s variables , this means that we need to transform the dlambertian operator @xmath80 in terms of the new variables . since the dlambertian operator in terms of the misner variables @xmath42 is given by @xmath92 so
, we can transform the dlambertian operator in terms of the new variables as @xmath93 finally , the wdw equation is given by @xmath94 + q\left(\frac{\partial\theta}{\partial\beta_{1 } } + \frac{\partial\theta}{\partial\beta_{2 } } + \frac{\partial\theta}{\partial\beta_{3}}\right ) - 6\left(\frac{\partial^2\theta}{\partial\beta_{1}\partial\beta_{2 } } + \frac{\partial^2\theta}{\partial\beta_{1}\partial\beta_{3 } } + \frac{\partial^2\theta}{\partial\beta_{2}\partial\beta_{3}}\right)\nonumber\\ & & - \biggl[b_{\gamma}e^{(1-\gamma)(\beta_{1 } + \beta_{2 } + \beta_{3 } ) } + 12 e^{4\beta_{1}}-\mu^2\biggr]\theta=0 . \label{newwdw}\end{aligned}\ ] ] with this consideration , the expression is reduced to @xmath95 with @xmath96 .
the last partial differential equation has a solution by taking the following ansatz @xmath97 where @xmath98 are arbitrary constants and @xmath99 is an arbitrary functions depending on the variable @xmath100 . substituting this ansatz into , we find that the @xmath101 function satisfy the ordinary differential equation @xmath102g=0\ ] ] with solution @xmath103 where @xmath104 are the modified bessel functions @xmath105 , @xmath106
, @xmath107 , and @xmath108 its order . in order that the unnormalized probability density @xmath2 does not diverge for latexmath:[$\rm gravitational variables , we drop the function @xmath5 , remain only the function @xmath6 .
so , we have that the solution to the partial differential equation ( pde ) is given by @xmath110 the asymptotic solution for large argument to this bessel function , goes as @xmath111 this solution we will be compared with solution obtained using the bohm s formalism @xcite . employing this formalism
we find that its amplitude of probability @xmath112 is given by @xmath113 for the particular case , where just is depending on the variable @xmath100 we have that its solution is @xmath114 we present the main ideas of this formalism to solve the wdw equation , you can see @xcite .
also we use the hidden symmetry in the potential u for this model @xcite , which seems to be a general property of the bianchi models .
we use the following ansatz for the wave function @xmath116 where @xmath117 is known as the superpotential function , and w is the amplitude of probability which is used in bohmian formalism @xcite , those found in the literature , years ago @xcite .
so is transformed into @xmath118 = 0 , \label{mod}\ ] ] where @xmath119 , @xmath120 , @xmath121 with @xmath122 , @xmath123 is the potential term of the cosmological model under consideration .
eq . can be written as the following set of partial differential equations [ wdwa ] @xmath124 following reference @xcite , first we shall choose to solve eqs . and , whose solutions at the end will have to fulfill eq . , which play the role of a constraint equation .
we were able to solve , by doing the change of coordinates and rewrite in these new coordinates , with this change , the function s is obtained by taking the ansatz .in this section , we shall obtain the solutions to the equations that appear in the decomposition of the wdw equation , , and , using the bianchi type ii cosmological model .
so , the equation @xmath125
^ 2= -(\frac{\partial}{\partial \omega})^2 + ( \frac{\partial}{\partial \beta_+})^2 + ( \frac{\partial}{\partial \beta_-})^2 $ ] can be written in the following way @xmath126 ^ 2 & = & 3 \left [ \left(\frac{\partial}{\partial \beta_1}\right)^2 + \left(\frac{\partial}{\partial \beta_2 } \right)^2 + \left(\frac{\partial}{\partial \beta_3}\right)^2 \right ] - 6\left [ \frac{\partial}{\partial \beta_1 } \frac{\partial}{\partial \beta_2 } + \frac{\partial}{\partial \beta_1 } \frac{\partial}{\partial \beta_3 } + \frac{\partial}{\partial \beta_2 } \frac{\partial}{\partial \beta_3 } \right ] \nonumber\\ & = & 3 \left ( \frac{\partial}{\partial \beta_1}+ \frac{\partial}{\partial \beta_2}+ \frac{\partial}{\partial \beta_3 } \right ) ^2 -12 \left [ \frac{\partial}{\partial \beta_1 } \frac{\partial}{\partial \beta_2 } + \frac{\partial}{\partial \beta_1 } \frac{\partial}{\partial \beta_3 } + \frac{\partial}{\partial \beta_2 } \frac{\partial}{\partial \beta_3 } \right ] .
\label { nab}\end{aligned}\ ] ] the potential term of the bianchi type ii is transformed in the new variables into @xmath127 then for this models is rewritten in the new variables as @xmath128 - 12 e^{4\beta_1 } = 0 .
\label{hanv}\ ] ] now , we can use the separation of variables method to get solutions to the last equation for the @xmath129 function , obtaining for the bianchi type ii model @xmath130 with this result , and using for the solution to in the new coordinates @xmath131 , we have for w function as @xmath132 and re - introducing this result into eq .
we find the constriction on the q factor ordering , @xmath133 and @xmath134 , thus the equation have the form @xmath135 so , the bohm s solutions becomes ( we demanding that @xmath136 does not diverge for @xmath137 , for fixed @xmath4 , the contribution @xmath138 is dropped ) @xmath139 . \label{w2extra1}\ ] ] this class of solution become as the pure bosonic first term and the simple fermionic part in the decomposition of the wave function in the supersymmetric quantum cosmology in the grassmann representation , for this same cosmological model , studied 22 years ago @xcite .
when we compare this solution with the general solution found using separation of variables , we obtain that this corresponds to the asymptotic behavior of this solution , that in some way corresponds to the lowest - order wkb approximation @xcite , which is shown in the figure [ fig ] , where the picks of the plot follow the classical trajectory @xmath140 , which corresponds at zero phase in the probability amplitude in both solutions ; in other words , this wave function has translational symmetry along the last trajectory mentioned .
also is presented a shifting to the wave function in the @xmath141 axe in the @xmath142 direction , how is shown in the graphics to the probability density @xmath143 in both sectors ( full anisotropy and lowest - order wkb approximation ) . with the conditions mentioned above
, we shall calculate @xmath144 as an unnormalized probability density , considering only the anisotropy variables and the @xmath4 parameter fixed . in the full wave function
we have an unnormalized behavior , however in the bohm s sector we have that the maximum is minor to unity , due that only we consider the lowest - order wkb approximation . ) and ( [ w2extra1 ] ) , respectively , considering the anisotropic variables @xmath145 , centering in @xmath146 and 2 , beginning in the up to down in the graphics . the numerical order @xmath147 is considered for the value @xmath134 and the constant @xmath148 appeared in the bohm s formalism . ]
in this work we have explored a compactification of a ten - dimensional gravity theory coupled with a time - dependent dilaton into a time - dependent six - dimensional torus . the effective theory which emerges through this process resembles in the einstein frame
was applied to anisotropic cosmological bianchi type ii model . by incorporating the matter content and by using the analytical procedure of hamilton equation of classical mechanics , in appropriate coordinates
, we found the classical solution for the anisotropic bianchi type ii cosmological model . in the quantum formalism for the standard wheeler - dewitt equation
, we can observe that this anisotropic model is completely integrable without employ numerical methods , similar solutions to partial differential equation have been found in @xcite . in order to have the best candidates for quantum solutions become those that have a damping behavior with respect to the scale factor , represented in our model with the @xmath4 parameter , in this way we dropped in the full solution the modified bessel function @xmath149 and the term @xmath150 in the bohm s solution .
we obtain the relation between these both physical solutions in the asymptotic behavior to the modified bessel function for large argument , and it is compared with the bohm s solutions , , who us indicate that is the lowest - order wkb approximation how is indicate in the literature @xcite . the unnormalized probability density which is shown in the figure [ fig ] , show that the picks of the plot follow the classical trajectory @xmath140 , which corresponds at zero phase in the probability amplitude in both solutions ; in other words , this wave function has translational symmetry along the last trajectory mentioned .
also is presented a shifting to the wave function in the @xmath141 axe in the @xmath142 direction , that in some sense us indicate that the anisotropy of the model in the quantum state finish for large @xmath151 and follow this in the line @xmath140 . for future
work is to consider a more complete compactification process in which all moduli are considered as time - dependent fields as well as time - dependent fluxes and follow the analysis presented in the ref .
@xcite to obtain the other orders to the wkb approximation , and see how these orders contribute to the full solution found by separation of variables .
kuchar , time and interpretations of quantum gravity , _ in _ proceedings of the 4th conference on general relativity and relativistic astrophysics , eds .
g. kunstatter , d. vincent and j. williams ( world scientific , singapore 1992 ) .
m. aguero , j.a.s .
aguilar , c. ortiz , m. sabido and j. socorro , int .
j. of theor .
* 46 * , 2928 ( 2007 ) .
o. obregn , j. socorro and j. bentez , phys .
d * 47 * , 4471 ( 1993 ) , and references therein .
borun d. chrowdhury and samir d. mathur , class .
* 24 * , 2689 ( 2007 ) , _ fractional brane state in early universe_. s. nojiri , o. obregn , s.d.odintsov and k.e .
d * 60 * , 024008 ( 1999 ) . _
( non ) singular kantowski - sachs universe from quantum spherically .
_ s. nojiri , o. obregn , s.d .
odintsov and k.e .
b * 449 * , 173 ( 1999 ) , _ induced wormholes due to quantum effects of spherically reduced matter in large n approximation .
gibbons and l. p. grishchuk nucl .
b * 313 * , 736 ( 1989 ) .
hartle and s.w . hawking ,
d * 28 * , 2960 ( 1983 ) .
s.w . hawking nucl .
b * 239 * , 257 ( 1984 ) .
moniz , _ quantum cosmology -the supersymmetric perspective- vol . 1 & 2 _ , lecture notes in physics * 803 & 804 * , ( springer , berlin ) ( 2010 ) .
j. h. bae , class .
quantum grav .
* 32 * , 075006 ( 2015 ) .
_ mixmaster revisited : wormhole solutions to the bianchi ix wheeler dewitt equation using the euclidean - signature semiclassical method _
d. bohm , phys .
rev . * 85 * , 166 ( 1986 ) .
j. socorro , paulo a. rodrguez , o. nez - soltero , rafael hernndez and abraham espinoza - garca , _ open questions in cosmology _ , intech , chapter 9 , _ quintom potential from quantum anisotropic cosmological models _ ,
pag 219 . edited by gonzalo j. olmo , ( 2012 ) .
r. graham , phys .
lett . * 67 * , 1381 ( 1991 ) .
sukumar , j. phys .
a * 18 * , 2917 ( 1985 ) .
o. obregn and j. socorro , int .
j. of theor .
* 35 * ( 7 ) , 1381 ( 1996 ) . | in this work we construct an effective four - dimensional model by compactifying a ten - dimensional theory of gravity coupled with a real scalar dilaton field on a time - dependent torus without the contributions of fluxes as first approximation .
this approach is applied to anisotropic cosmological bianchi type ii model for which we study the classical coupling of the anisotropic scale factors with the two real scalar moduli produced by the compactification process .
also , we present some solutions to the corresponding wheeler - dewitt ( wdw ) equation in the context of standard quantum cosmology and we claim that these quantum solution are generic in the moduli scalar field for all bianchi class a models . also we gives the relation to these solutions for asymptotic behavior to large argument in the corresponding quantum solution in the gravitational variables and is compared with the bohm s solutions , finding that this corresponds to lowest - order wkb approximation . |
a great deal of attention has been devoted recently to the subject of perfect state transfer ( pst ) in quantum networks .
since an actual quantum processor would require , in fact , the ability to transfer quantum information between spatially separated interacting systems composing a network , protocols have been established for pst in many different and general contexts . among the several interesting theoretical contributions , aiming to advance the understanding and eventual implementation of a quantum processor ,
the construction of effective two - qubit gates from pst between distant nodes in engineered bosonic and fermionic networks was proposed @xcite .
a general formalism of the problem of pst in networks of any topology and coupling configuration was also developed @xcite .
focusing on spin chains , pst has been pursued in networks extending beyond the nearest - neighbor couplings @xcite , and a class of qubit networks allowing pst of any state in a fixed period of time has been devised @xcite .
the problem of the scaling of errors ( arising from network nonidealities ) with the length of the channel connecting the nodes has been discussed @xcite . in this connection , a protocol for arbitrary pst in the presence of random fluctuations in the coupling strengths of a spin chain
was reported @xcite .
the entanglement dynamics in spin chains subject to noise and disorder has also been analyzed in ref .
@xcite . regarding pst in networks of harmonic oscillators ,
a comprehensive analysis of this subject has been presented in ref .
@xcite , and a protocol for high - efficiency transfer of quantum entanglements in translation - invariant quantum chains has been proposed @xcite .
in contrast to the achievements with spin chains , there have been no proposals , until this paper , for arbitrary pst through a network of nonideal harmonic oscillators .
since pst is achieved by appropriate tuning of the intermode coupling and the frequencies of the systems composing the network , the search for the general rules governing these adjustments is of crucial interest . in this regard ,
it is worth noting that the general recipe presented in ref .
@xcite contrasts with the procedure in ref .
@xcite where a degenerate chain of oscillators is considered : all oscillators having the same frequency and interacting with each other with the same coupling strength .
the lack of pst in ref .
@xcite is compensated by a more realistic possibility of implementation of a translation - invariant network , apart from the high efficiency transfer operation reported .
we observe that in the present contribution we also derive general conditions for pst , with the advantage that they are easier to handle mathematically than those outlined in ref .
@xcite . as well as in spin chains ,
significant advances have been made recently in bosonic networks @xcite .
a general treatment of a network of coupled dissipative quantum harmonic oscillators has been presented recently , for an arbitrary topology , i.e. , irrespective of the way the oscillators are coupled together , the strength of their couplings , and their natural frequencies @xcite . regarding the dissipative mechanism ,
two different scenarios are considered in ref .
@xcite : a more realistic in which each oscillator is coupled to its own reservoir and another with all the network oscillators coupled to a common reservoir .
within such a general treatment of dissipation , the emergence of relaxation- and decoherence - free subspaces in networks of weakly and strongly coupled resonators has also been addressed @xcite .
we finally mention the proposition of a quantum memory for the preservation of superposition states against decoherence by their evolution in appropriate topologies of such dissipative bosonic networks @xcite .
building on the achievements mentioned above , in the present study we develop a protocol for quasi - perfect state transfer ( qpst ) in a linear network of dissipative oscillators . more specifically , we envisage the transfer of a state between ideal sender and receiver oscillators through a linear chain of nonideal transmitter oscillators . by analogy with refs .
@xcite , here we again adopt the general and more realistic scenario where each transmitter oscillator is coupled to a distinct reservoir .
therefore , our goal resembles that of arbitrary pst in ref .
@xcite ; however , instead of fluctuations in the couplings of a spin chain , we deal with the fluctuations injected by each of the reservoirs coupled to the transmitter oscillators . anticipating our strategy to achieve qpst despite these sources of nonideality , we adjust the frequency of the sender and the receiver oscillators to be significantly out of resonance with that of the transmitters ; within such an arrangement the state to be transferred occupies the transmitter oscillators only virtually , weakening the undesired effects of their decay mechanisms .
we point out that the recent proposal of a variety of resources for wiring up quantum systems @xcite lends a strongly realistic bias to the possibility of controlling the transfer of information in quantum networks .
following the mastering of the manipulation of the interaction between single atoms and vibrating modes of high - q cavities @xcite and trapped ions @xcite , circuit cavity quantum electrodynamics and photonic crystals seem to enhance the ability to transfer quantum information to a level enabling the implementation of a logic processor @xcite . in this connection , the elaboration of schemes to circumvent the noise injection in the processes of pst in networks of coupled nonideal quantum systems is indispensable , enabling protocols for fault - tolerant information transfer and deepenning our understanding of fundamental quantum phenomena such as entanglement and decoherence .
the plan of the paper is as follows . in section [ sec2 ] , by analogy with the developments presented in ref @xcite ,
we introduce our model in which two ideal oscillators are connected by a transmission line of nonideal oscillators .
the master equation describing the network dynamics is presented , together with its solution . in section [ sec3 ]
, we present a criterion for pst ( qpst ) in ideal ( nonideal ) networks of harmonic oscillators , together with a particular application where the sets of parameters @xmath0 and @xmath1 ensuring pst are derived .
our criterion relies on the definition of matrix @xmath2 , describing the evolution of the system , which serves two purposes : to ensure pst and to compute the exchange time @xmath3 .
the analytical treatment of qpst in small nonideal linear networks of @xmath4 , @xmath5 , @xmath6 , and @xmath7 oscillators are given in section [ sec4 ] . in section [ sec5 ]
we apply numerical procedures to extend our results to large numbers of nonideal transmitters , providing a comprehensive analysis of all the network parameters involved and demonstrating the robustness of our model . finally , we present our concluding remarks in section [ sec6 ] .
before introducing our model for pst , we first revisit the developments in ref .
@xcite , considering a network of @xmath8 interacting dissipative oscillators from the general perspective where each oscillator interacts with each other .
any particular network topology ( or graph ) follows from this general approach , with an appropriate choice of the parameters defining the hamiltonian modelling the network . as stressed above , by topology
is meant @xmath9 which resonators are coupled together , @xmath10 their coupling strengths and @xmath11 their natural frequencies . moreover , in a more realistic approach for most physical systems , it is assumed that each network oscillator interacts with its own reservoir , instead of the special case where all the oscillators interact with a common reservoir .
therefore , assuming from here on that the indexes @xmath12 , @xmath13 , and @xmath14 , labeling the oscillators , run from @xmath15 to @xmath8 , we start from the general hamiltonian @xmath16 , \label{1}\ ] ] where @xmath17 ( @xmath18 ) is the creation ( annihilation ) operator for the @xmath19th bath mode @xmath20 coupled to the @xmath12th network oscillator @xmath21 , whose creation ( annihilation ) operator is @xmath22 ( @xmath23 ) .
the coupling strengths between the oscillators are given by the set @xmath1 , while those between the oscillators and their reservoirs by @xmath24 .
the reservoirs are modeled as a set of @xmath25 modes , and the elements @xmath26 defining the network topology compose the matrix @xmath27 to obtain the master equation of the network we first diagonalize the hamiltonian @xmath28 through a canonical transformation @xmath29 , where the coefficients of the @xmath12th line of matrix @xmath30 define the eigenvectors associated with the eigenvalues @xmath31 of matrix ( [ 2 ] ) . the commutation relations @xmath32
= \delta_{mn}$ ] and @xmath33 = 0 $ ] , following from the orthogonality of matrix @xmath30 , in that @xmath34 , enable us to rewrite hamiltonian ( [ 1 ] ) in terms of decoupled normal - mode oscillators @xmath31 , each of them interacting , however , with all the @xmath8 reservoirs .
the diagonalized @xmath28 helps us to introduce the interaction picture , within which a set of assumptions leads to the master equation describing the network evolution .
we first assume that the couplings between the normal - mode oscillators and the reservoirs are weak enough to allow a second - order perturbation approximation .
moreover , under markovian white noise , the evolved reduced density operator of the network , @xmath35 , is assumed to be factorized from the stationary density operator of the reservoirs .
finally , the reservoir frequencies are assumed to be closely spaced enough to allow a continuum summation , the spectral density @xmath36 and coupling parameter @xmath37 being slowly varying functions .
thus , after tracing out the degrees of freedom of the absolute zero reservoirs , we obtain the generalized lindblad form@xmath38 \right .
\nonumber\\ & + \left .
\frac{\gamma_{mn}}{2}\left ( \left [ a_{n}\rho(t),a_{m}^{\dag } \right ] + \left [ a_{m},\rho(t)a_{n}^{\dag}\right ] \right ) \right\ } \nonumber\\ & \equiv\sum_{m , n}\left\ { \frac{i}{\hbar}\left [ \rho(t),a_{m}^{\dagger } h_{mn}a_{n}\right ] + \mathcal{l}_{mn}\rho(t)\right\ } \mathrm { , } \label{3}\ ] ] where we have defined the effective damping matrix @xmath39 , whose elements are@xmath40 with @xmath41 ^{2}\int_{-\varpi_{n}}^{\infty}\delta\left ( \epsilon\right ) \operatorname*{d}\epsilon$ ] . in eq .
( [ 3]),@xmath42the liouville operators @xmath43 account for both the direct ( @xmath44 ) and indirect ( @xmath45 ) dissipative channels
. through the direct dissipative channels , the oscillators lose excitation to their own reservoirs , at a damping rate @xmath46 , whereas through the indirect channels they lose excitation to all the other reservoirs but not to their own . for markovian white noise reservoirs
the indirect channels disappear , since the spectral densities of the reservoirs are invariant over translation in frequency space , rendering @xmath47 , and , consequently , @xmath48 @xcite .
it is worth stressing that the whole of the working presented in the next subsection , where a solution to the master equation is derived , applies to the case of non - markovian reservoirs , thus including the indirect dissipative channels .
markovian reservoirs are only assumed at the end of subsection c , when the topology of our framework is finally defined as a linear chain of dissipative oscillators . to obtain a solution of eq .
( [ 3 ] ) , we shall employ it to derive the glauber - sudarshan @xmath49 function for the network@xmath50 where we have defined the matrix elements @xmath51 , extending the former @xmath26 , to take into account the dissipative ( @xmath52 ) process .
assuming that the general network described by matrix ( [ 2 ] ) is composed entirely of dissipative oscillators , the matrix @xmath53 assumes the form@xmath54 for the initial state of the network we consider the general pure superpositions of coherent states@xmath55 where @xmath56 is the normalization factor , @xmath57 is the probability amplitude of the product state @xmath58 , and the labels @xmath59 and @xmath60 run from @xmath15 to the integer @xmath61 .
the superscript @xmath59 stands for the @xmath59th state of the superposition while the subscript @xmath12 stands for the coherent state of @xmath42the @xmath12th oscillator .
( we stress that the discrete sum of product states in eq .
( [ rho0 ] ) can be substituted by the continuum sum @xmath62 with no further complication ) . from eqs .
( [ p ] ) and ( [ rho0 ] ) , it is straightforward to show that the network density operator evolves as@xmath63 the excitation of the @xmath12th oscillator , given by @xmath64 follows from the time - dependent matrix elements @xmath65 where the @xmath12th column of matrix @xmath66 defines the @xmath12th eigenvector associated with the eigenvalue @xmath67 of matrix @xmath53 . for the reduced density operator of the @xmath12th oscillator we obtain@xmath68 where the influence of all the other oscillators of the network is present explicitly in the product @xmath69 and implicitly in the states @xmath70 .
a linear ( @xmath71 ) dissipative network of nearest - neighbor interacting harmonic oscillator is built up by coupling the @xmath19th oscillator with the @xmath72th oscillators , leaving the first oscillator ( @xmath73 ) uncoupled from the last one ( @xmath74 ) .
the matrix @xmath75 obtained for this case has the three - diagonal form@xmath76{ccccccc}\omega_{1 } & \lambda_{12 } & 0 & \cdots & 0 & 0 & 0\\ \lambda_{12 } & \omega_{2 } & \lambda_{23 } & \cdots & 0 & 0 & 0\\ 0 & \lambda_{23 } & \omega_{3 } & \ddots & 0 & 0 & 0\\ \vdots & \vdots & \ddots & \ddots & \ddots & \vdots & \vdots\\ 0 & 0 & 0 & \ddots & \omega_{n-2 } & \lambda_{n-2,n-1 } & 0\\ 0 & 0 & 0 & \cdots & \lambda_{n-2,n-1 } & \omega_{n-1 } & \lambda_{n-1,n}\\ 0 & 0 & 0 & \cdots & 0 & \lambda_{n-1,n } & \omega_{n}\end{array } \right ) + \frac{1}{2}\begin{pmatrix } \gamma_{11 } & \gamma_{12 } & \cdots & \gamma_{1n}\\ \gamma_{21 } & \gamma_{22 } & \cdots & \gamma_{2n}\\ \vdots & \vdots & \ddots & \vdots\\ \gamma_{n1 } & \gamma_{n2 } & \cdots & \gamma_{nn}\end{pmatrix } \text{.}\ ] ] as already made clear , we focus on the case of ideal sender and receiver oscillators , here assumed to be the first and the last , respectively , both with the same frequency @xmath77
. all the transmitter oscillators , from the second to the ( @xmath78)th , are assumed from here on to decay at the same rate @xmath79 , and to be tuned out of resonance with the sender and receiver , to frequency @xmath80 . regarding the coupling between the oscillators
, we assume that the sender and the receiver are connected with their transmitter neighbors with the same strength @xmath81 , whereas the transmitters are connected to each other with strength @xmath82 , @xmath83 being a dimensionless parameter allowing the couplings within the transmitting channel to be controlled . finally , assuming markovian white noise reservoirs , to eliminate the indirect dissipative channels , the above matrix simplifies to @xmath84
on the basis of the model developed above , in this section we derive a general criterion for pst , whatever the topology of an ideal ( @xmath85 ) network of harmonic oscillators .
such a criterion for pst is further applied for the particular case of the linear topology modeled by eq .
( [ lin ] ) .
starting from eq .
( [ 4 ] ) written as the matrix product @xmath86 , it is straightforward to conclude that the condition for transferring the state of the first oscillator to the @xmath8th is given by the matrix structure@xmath87 & \genfrac{}{}{0pt}{0}{\genfrac{}{}{0pt}{0}{{}}{0}}{\genfrac{}{}{0pt}{0}{\vdots } { { } } } \\ 1 & 0\hspace{0.2in}\cdots & 0 \end{pmatrix } \text { , } \label{theta}\ ] ] which reversely ensures that the state of the @xmath8th oscillator can equally be transferred to the first one , no matter what happens to the states of the intervening @xmath88 transmitting oscillators , from the second to the last but one . that is why the submatrix formed by removing the first and last rows and columns of @xmath89 is left undefined ;
for any submatrix @xmath90}$ ] , pst is achieved as long as the states of the first and the last oscillators are interchanged .
it is worth noting that the criterion for state exchange fixed by the structure of the matrix in eq .
( [ theta ] ) does not depend on the network state .
evidently , the imposition of this matrix form can be used to compute the exchange time @xmath3 . since the evolution of matrix @xmath2 , given by @xmath91\cdot\mathbf{d}^{-1}\nonumber\\ & = \exp[-(\mathbf{d\cdot}i\mathbf{r}\cdot\mathbf{d}^{-1})t]=\exp [ -i\mathbf{h}t]\text { , } \label{theta1}\ ] ] is governed by the hamiltonian @xmath28 ( whose eigenvalues @xmath92compose the diagonal matrix @xmath93 ) , the sets of parameters @xmath0 and @xmath1 complying with a necessary condition for pst in a particular network topology follow from the commutation relation @xmath94 = 0\text { , } \label{cm}\ ] ] which holds at any time , including the state exchange time @xmath3 .
we observe that not all sets @xmath95 and @xmath96 derived from condition ( [ cm ] ) ensure pst . for *
* a necessary and sufficient condition we must choose , among the sets of parameters @xmath95 and @xmath97 , those ensuring the reduction of matrix ( [ theta1 ] ) to ( [ theta ] ) , our initial premise , at the exchange time .
hence , matrix @xmath2 can be used for two purposes : to ensure pst and to compute the exchange time @xmath3 .
we observed in the introduction that our general condition for pst , given by eq .
( [ cm ] ) , has the advantage of being easier to handle mathematically than that outlined in ref .
in fact , in ref .
@xcite , the condition for pst follows from the diagonalization of a permutation matrix equivalent to the above defined @xmath2 whereas our condition reduces to the computation of a commutation relation .
it must be stressed that in the case where the state transfer is achieved by tunneling , the set of coupling parameters @xmath1 does not play a significant role in the process since the transferred state does not effectively occupy any of the transmitter oscillators .
actually , for the state to be transferred by tunneling , we only have to ensure that the common frequency ( @xmath77 ) of the sender and the receiver oscillators is distinct from those of the transmitter ones ; under this condition , we can verify the relation ( [ cm ] ) independently of the set of coupling strengths @xmath98 , and these may all have the same value @xmath81 .
the extension of the above criterion to state transfer in nonideal networks , the case at hand , follows directly by assuming a finite value of @xmath99 .
the damping rates due to dissipation are thus introduced into eqs .
( [ theta ] ) and ( [ theta1 ] ) , generalizing the relation ( [ cm ] ) to @xmath100 = 0\text{. } \label{cr}\ ] ] however , in this general nonideal case the transferred state exhibits a fidelity that decreases with time and the magnitude of the decay rate @xmath101 , apart from depending on other properties such as the network topology . moreover , there are two further situations , not envisaged in this paper , where the commutation relation ( [ cr ] ) reduces to that of the ideal case ( [ cm ] ) , even when the fidelity decreases with time and the magnitude of @xmath99 : @xmath9 when @xmath102 and @xmath10 @xmath103 . in both cases the sets of parameters @xmath104 and @xmath1 ensuring qpst are exactly those ensuring pst .
it is worth noting again that , confining ourselves to a linear dissipative network , the adopted strategy of considering the ideal sender and receiver oscillators significantly out of resonance with the nonideal transmitters , ensures qpst despite the nonidealities of the transmitter line .
in fact , as anticipated in the introduction and demonstrated below , the virtual occupation of the transmitter line protects the transferred state almost perfectly from the damping mechanisms . considering the hamiltonian @xmath28 in eq .
( [ 2 ] ) and the particular choice @xmath106{ccccc}0 & 0 & \cdots & 0 & 1\\ 0 & 0 & \cdots & 1 & 0\\ \vdots & \vdots & \ddots & \vdots & \vdots\\ 0 & 1 & \cdots & 0 & 0\\ 1 & 0 & \cdots & 0 & 0 \end{array } \right ) \label{theta}\ ] ] ensuring the transfer of the initial state of the @xmath12th oscillator to the @xmath107 $ ] th one , we obtain from relation ( [ cm ] ) the condition for pst:@xmath108 which generalizes the one introduced in ref .
@xcite , given by [ pp]@xmath109 evidently , each choice of the submatrix @xmath90}$ ] in the general form ( [ theta ] ) prompts a different set of parameters @xmath104 and @xmath1 ensuring pst .
another way to compute the exchange time @xmath3 , instead of using matrix @xmath110 , follows from the time profile of the probability of a successful transfer of the desired state or equivalently , the fidelity of the transfer process given by @xmath111 \text{. } \label{f}\ ] ] evidently , in a general situation where pst is sought other than the transfer by tunneling analyzed here the computation of the exchange time @xmath3 from eq .
( [ f ] ) , must take into account the set of parameters @xmath95 and @xmath112 ensuring pst , derived from the commutation relation ( [ cr ] ) .
a further condition is that the ideal case @xmath85 must render @xmath113 .
although a rather demanding task , the maximization of the probability @xmath114 yields , by itself , a necessary and sufficient condition for the derivation of the sets of parameters @xmath95 and @xmath1 ensuring pst . the commutation relation ( [ cr ] ) provides a shortcut for this task . we assume , as usual , that the initial density operator @xmath115 factorizes as @xmath116 , @xmath117 representing the state of the first oscillator , to be transferred to the @xmath8th one , and @xmath118 standing for the initial state of the transmitter and the receiver oscillators . in its turn , the density operator at the exchange time @xmath119 factorizes as @xmath120 , where @xmath121 represents the transferred state of the @xmath8th oscillator , and @xmath122 the final state of the transmitter oscillators plus the sender
evidently , the state of the @xmath8th oscillator at @xmath3 must be , in the ideal case where @xmath85 , exactly that prepared in the first oscillator . with this assumption , by substituting eq .
( [ rho ] ) into @xmath114 , we obtain @xmath123 ^{\ast}\left [ \zeta_{n}^{r^{\prime}}(t_{\mathrm{ex}})-\beta_{1}^{r}\right ] \right\ } \text{. } \label{prob}\ ] ] as an illustrative application of eq .
( [ prob ] ) , we consider the specific case where the state to be transferred to the @xmath8th oscillator is prepared in the first one as the schrdinger cat - like superposition @xmath124 , while all other oscillators are in the vacuum state .
we obtain from eq .
( [ prob ] ) , for the ideal case , the relations [ t]@xmath125 which enable us to determine @xmath3 as well as eq .
( [ cm ] ) , remembering that @xmath92 stands for the eigenvalues of the free hamiltonian @xmath28 .
therefore , the exchange time @xmath3 follows from the computation of the eigenstates and the associated eigenvectors of @xmath28 .
now , focusing on our scheme for the transfer of states by tunneling , we present an analytical treatment of small nonideal linear networks , using the particular case of ( [ theta ] ) to compute the exchange time @xmath3 . as an illustration , consider the case of qpst in a network with @xmath126 , whose dissipative matrix , following from eq .
( [ lin ] ) , is @xmath127 the eigenvalues and eigenvectors of @xmath128 defines the matrix @xmath2 and , consequently , the evolved excitation of the oscillators @xmath129 .
next , we introduce the scaled damping rate @xmath130 , frequency @xmath131 , and time @xmath132 , apart from the detunings @xmath133 and the effective coupling @xmath134 , to stress that we must focus on the regime where @xmath135 and @xmath136 .
these parameters ensure the weak coupling between the transmitter oscillators and their respective reservoirs justifying the master equation derived above apart from enabling the expansion of matrix @xmath2 to second order in @xmath137 , giving @xmath138 where we have defined the functions@xmath139 , \\
\mathfrak{g}_{s}(\tau ) & = \mu\left [ \sin(\varepsilon\mu^{2}\tau ) -\mathfrak{h}_{s}(\tau)\right ] , \\
\mathfrak{h}_{c}(\tau ) & = \exp\left [ -\left ( \eta+i\mu^{-1}\right ) \tau\right ] \cos\left ( \varepsilon\tau\right ) , \\
\mathfrak{h}_{s}(\tau ) & = \exp\left [ -\left ( \eta+i\mu^{-1}\right ) \tau\right ] \sin[\left ( \varepsilon\tau\right ) .\end{aligned}\ ] ] from the condition for transfer of the state of the first oscillator to the last one , fixed by the matrix structure ( [ theta ] ) , and assuming the additional restriction @xmath140 , we thus obtain from eq .
( [ theta4 ] ) the scaled exchange time @xmath141 , \ ] ] which implies the relations [ zeta]@xmath142 + \mathcal{o}\left ( \mu^{2}\right ) \text{,}\label{zeta1}\\ \zeta_{4}^{r}\left ( \tau_{\mathrm{ex}}^{(n=4)}\right ) & \simeq e^{-\eta \mu^{2}\tau_{\mathrm{ex}}^{(4)}}\left [ \beta_{1}^{r}-\left ( \mathfrak{g}_{s}(\tau_{\mathrm{ex}}^{(4)})\beta_{2}^{r}+i\mathfrak{g}_{c}(\tau _ { \mathrm{ex}}^{(4)})\beta_{3}^{r}\right ) \right ] + \mathcal{o}\left ( \mu^{2}\right ) \text { , } \label{zeta2}\ ] ] to ensure that the commutation @xmath143
= 0 $ ] is satisfied . as is evident from the above expressions , the relaxation process represented by the scaled damping rate @xmath144 prohibits a perfect state transfer , attenuating the value of the excitations @xmath145 and @xmath146 .
moreover , the reservoirs also spoil the desired relation @xmath147 by mixing it with the excitations of the nonideal transmitter oscillators . however , and this is the core of our technique , the tunneling mechanism of state transfer prompts the decay function @xmath148 which approaches unity within the ranges of the parameters outlined above , i.e. , @xmath149 as the coupling @xmath82 between the transmitting oscillators is increased .
we note that the increase of @xmath82 decreases the exchange time @xmath150 , and that is the reason for the choice of a strong coupling strength @xmath82 between the transmitter oscillators : to enable the control of the exchange time @xmath151 , decreasing it as the fidelity of the transfer process increases . otherwise , without the dimensionless parameter @xmath83 , increasing of the detuning @xmath152 ( i.e. , the process allowing a significant fidelity of the transferred state ) , would result in an uncontrolled rise in the exchange time @xmath153 . from eq .
( [ zeta ] ) , we conclude that the fidelity of the state transfer mechanism is maximized when the transmitter oscillators are prepared in the vacuum state , apart from weakening the inevitable system - reservoir coupling .
in fact , the amount of excitation of the dissipative transmission channel is directly proportional to the intensity of noise injected into the transferred state .
we finally note that for the case of an ideal transmission channel , i.e. , @xmath154 , we obtain up to first order corrections in @xmath137 carried out in @xmath155 and @xmath156 in the general case of an excited transmission channel the desired relations @xmath157 and @xmath158 . on the basis of the exchange time @xmath159 for the simplest network , composed of two oscillators with coupling strength @xmath81 @xcite , it is useful to assign an effective coupling strength to the network which
, for the case @xmath126 , turns out to be @xmath160 . following the steps described above for the case @xmath126 and adopting the same regime of parameters @xmath162
, we obtain for @xmath163 the result @xmath164 \text{,}\ ] ] giving the expected effective coupling strength @xmath165 between the sender and receiver oscillators . evidently , it is also possible to derive analytical expressions for the scaled exchange time for networks with @xmath166 whenever the diagonalization of the associated matrix @xmath167 generates a characteristic polynomial that factorizes into parts of degree @xmath168 @xmath5 .
we have found that @xmath169 is the limiting case permitting analytical solution , with a characteristic polynomial that factorizes into two parts of degree @xmath5 . for @xmath170 ,
the characteristic polynomial factorizes into two parts of degrees @xmath171 . analyzing the cases @xmath172 and @xmath173 , which factorize into polynomial of degrees @xmath174 and @xmath175 , respectively
, we obtain the results@xmath176 , \\
\tau_{\mathrm{ex}}^{(n=6 ) } & \simeq\frac{\pi}{2\varepsilon^{3}\mu^{4}}\left [ 1+\mathcal{o}(\mu^{2})\right ] \text{.}\ ] ] we finally note that , without the imposition of the restriction @xmath177 , the above expressions for @xmath178 would be approximately rewritten , for small values of @xmath8 , as@xmath179 \mu^{2})\right\ } \text { , } \label{g}\ ] ] where @xmath180 , @xmath181 , and @xmath182 . evidently , for small @xmath8 and @xmath177 the above expression is equivalent to the results derived from @xmath163 to @xmath173 .
the derivation of an analytical expression for @xmath183 in the general case of any @xmath8 is not an easy task . as @xmath8 increases , we find that an involved dependence of the second - order correction @xmath184 on @xmath8 begins to play a significant role .
although the task of identifying this dependence is still a compelling challenge , in the present paper we analyze qpst for large values of @xmath8 numerically , using expression ( [ f ] ) to obtain the fidelity of the transfer process .
now , using the probability of a successful transfer of the desired state , given by eq .
( [ f ] ) , we analyze the qpst for large nonideal linear networks .
consider the transfer of state @xmath185 to the @xmath8th oscillator , with @xmath186 and all other oscillators in the vacuum state . in fig .
1(a , b , c , and d ) we plot the numerical curves for the exchange probability @xmath187 against @xmath188 for the cases @xmath189 and @xmath190 , respectively . considering the above regime of parameters , @xmath191 without requiring the additional restriction @xmath140 we assume , in units of the coupling strength @xmath81 , the fictitious value @xmath192 , giving @xmath193 and @xmath194 , apart from @xmath195 .
we first observe , as expected , that the exchange time @xmath183 increases proportionally to @xmath8 , being around @xmath196 for the case of @xmath172 , in agreement with the analytical result computed in eq .
( [ g ] ) .
moreover , as the state to be transferred occupies the virtual nonideal channel for a time interval proportional to @xmath8 , the fidelity of the transfer process decreases with @xmath8 , as displayed in fig .
1 . in fact , for the choice of decay rate outlined above , we show that the fidelity of the transfer process is about unity for the cases @xmath172 and @xmath197 , beginning to exhibit a significant decrease from @xmath198 . we have not found any sensible decrease of the fidelity for the cases @xmath172 and @xmath197 , even for time intervals many orders of magnitude longer than the exchange time .
the shaded regions in the figures follow from the strong oscillations of the probability @xmath187 , coming from the natural frequencies of the oscillators .
focusing on the case @xmath199 , in fig .
2(a ) we plot the exchange probability @xmath187 against @xmath188 , considering the same parameters as in fig .
1 , except for the excitation @xmath200 of the coherent states assumed to populate the oscillators composing the transmitter channel .
as expected , we verify that the exchange times are exactly those of the associated case in fig .
however , the pattern of the curves changes , giving rise to the expected background oscillations due to the initial excitation of the transmitter oscillators . moreover , additional peaks occurs at the recurrence time @xmath201 , when the superposition state @xmath202 goes back to the sender oscillator .
the magnitudes of these secondary peaks follow from the probability of finding the superposition state in the receiver oscillator at the recurrence time , given by @xmath203 = 1/2 $ ] , where @xmath117 and @xmath204 represent the superposition states standing in the first oscillator at @xmath205 and in the @xmath8th one at @xmath206 . in fig .
2(b ) we assume the same parameters as in fig .
2(a ) to plot the probability of recurrence of the superposition state back to the first oscillator , given by the expression @xmath207 $ ] .
as expected , in the recurrence time the superposition state recurs back to the sender oscillator . in fig .
3(a ) we adopt the same parameters as in fig .
1(b ) , except for the smaller value @xmath208 , to illustrate the expected increase in the exchange time .
an interesting feature of this figure is that the fidelity of the transfer process does not depends on @xmath83 , despite the increase in the exchange time , which is around 7 orders of magnitude greater than in fig .
these features are reinforced in fig .
3(b ) where , again with the same parameters as in fig .
1(b ) , we consider the limiting case @xmath209 , which results in a formidable increase in the exchange time of around 28 orders of magnitude , still preserving the fidelity . with the same parameters as in fig .
3(a ) , except for the smaller value of @xmath210 , in fig .
4(a ) we first see that the smaller detuning pulls the exchange time back , to about the order of magnitude found in fig .
1(b ) , in spite of the quantity @xmath208 .
moreover , we verify the expected continuous decrease of the process due to such a small detuning , which compels the state to populate more effectively the virtual transmitter channel . in fig .
4(b ) we use the same parameters as in fig .
4(a ) , except for the coherent state @xmath211 , to show that a larger excitation of the state to be transferred results in a smaller fidelity of the process . in fact
, the decoherence time of a quantum state varies inversely with its excitation , in accordance with the correspondence principle . finally , to illustrate the advantage of our tunneling - based scheme for state transfer over those where the transfer proceeds non - virtually through all the transmitter oscillators , in fig .
5 we plot the exchange probability @xmath187 against @xmath188 for the case where the sets of parameters @xmath95 and @xmath97 in eq .
( [ pp ] ) are utilized .
assuming , as in fig .
1(a ) , @xmath212 and @xmath195 , we find that , whereas the fidelity of our tunneling scheme is about unity up to at least 10 times the exchange time @xmath213 , that for the case of fig .
5 decays to around zero for @xmath214 .
in this work we have discussed the problem of state transfer in a linear chain of quantum dissipative harmonic oscillators .
assuming the first and last oscillators to be on - resonant with each other and significantly off - resonant with the transmitter channel running from the second to the last - but - one oscillator we take advantage of the tunneling effect to circumvent the decoherence during the transmission process .
we have assumed ideal sender and receiver oscillators connected by nonideal transmitters .
apart from the significant improvement of the fidelity of our quasi - perfect transfer process compared to those where the transfer proceeds through a non - virtual mechanism , we have presented additional contributions to state transferring processes . on the basis of a recently presented general treatment of a network of coupled dissipative quantum harmonic oscillators ,
we have also derived a general criterion for pst , whatever the topology of the network , which is applied here to the particular case of a linear chain .
our criterion , given essentially by the commutation relation between the network hamiltonian @xmath53 and the matrix @xmath215 connecting the initial states of the network to those at the exchange time ( when the transfer process is supposed to be accomplished ) , has the advantage of being easier to handle mathematically than that outlined in ref .
in fact , the matrix @xmath2 , defined in ref .
@xcite , can be used for two purposes : to ensure pst ( or qpst ) and to compute the exchange time @xmath151 , which can also be controlled by manipulating the coupling between the transmitter oscillators . as a matter of fact
, a large detuning between the on - resonant and the off - resonant oscillators ensures not only a high fidelity for the transfer process but also a significant prolongation of the exchange time . in this connection ,
an increase in the coupling between the transmitter oscillators acts to shorten such delays .
the role played by each of the network parameters is analyzed in details in a set of figures presented after the formal development of our scheme .
we thus propose that the tunneling scheme presented here can be useful for transferring quantum states between distant nodes of a quantum circuit without their undergoing a significant coherence decay .
evidently , the control of the network parameters , such as the natural frequencies and coupling strengths of the oscillators , still represents a sensitive issue to be overcome experimentally .
we also point out that recent developments in circuit qed and photonic crystals signal realistic platforms for the experimental implementation of our proposed scheme .
a. blais , r - s huang , a. wallraff , s. m. girvin , and r. j. schoelkopf , phys .
a * 69 * , 062320 ( 2004 ) ; r. j. schoelkopf and s. m. girvin , nature * 451 * , 664 ( 2008 ) ; f. illuminati , nature phys .
* 2 * , 803 ( 2006 ) .
m. a. nielsen and i. l. chuang , quantum computation and quantum information , cambridge university press , cambridge , 2000 ; the physics of quantum information , edited by d. bouwmeester , a. ekert , and a. zeilinger springer , berlin , 2000 . fig .
1 . numerical plots of the exchange probability @xmath187 against @xmath188 for the transfer of the state @xmath216 to the @xmath8th oscillator , with @xmath186 and all other oscillators in the vacuum state . in units of the coupling strength
@xmath81 we set the fictitious value @xmath192 , prompting @xmath217 and @xmath194 , apart from @xmath218 .
curves ( a)-(d ) refer to the cases @xmath189 and @xmath190 , respectively . fig .
numerical plots of the exchange probability @xmath187 against @xmath188 for the transfer of the state @xmath202 to the @xmath8th oscillator , assuming the same parameters as in fig .
1 , except for the excitation @xmath200 of the coherent states populating the oscillators of the transmitter channel . ( b )
probability of recurrence @xmath219 of the initial superposition back to the first oscillator , plotted against @xmath188 .
3 . numerical plots of the exchange probability @xmath187 against @xmath188 for the transfer of the state @xmath216 to the @xmath8th oscillator , assuming the same parameters as in fig .
1(b ) , except for the smaller values of the parameter ( a ) @xmath220 and ( b ) @xmath209 . fig .
( a ) numerical plot of the exchange probability @xmath187 against @xmath188 for the transfer of the state @xmath202 to the @xmath8th oscillator , assuming the same parameters as in fig .
3(a ) , except for the smaller value of the parameter @xmath210 .
( b ) the same as in ( a ) , but with the coherent state @xmath221 , to illustrate that the fidelity of the process decreases as the excitation of the state to be transferred increases . | in this paper we propose a scheme for quasi - perfect state transfer in a network of dissipative harmonic oscillators .
we consider ideal sender and receiver oscillators connected by a chain of nonideal transmitter oscillators coupled by nearest - neighbor resonances . from the algebraic properties of the dynamical quantities describing the evolution of the network state , we derive a criterion , fixing the coupling strengths between all the oscillators , apart from their natural frequencies , enabling perfect state transfer in the particular case of ideal transmitter oscillators .
our criterion provides an easily manipulated formula enabling perfect state transfer in the special case where the network nonidealities are disregarded . by adjusting the common frequency of the sender and the receiver oscillators to be out of resonance with that of the transmitters ,
we demonstrate that the sender s state tunnels to the receiver oscillator by virtually exciting the nonideal transmitter chain .
this virtual process makes negligible the decay rate associated with the transmitter line on the expenses of delaying the time interval for the state transfer process .
apart from our analytical results , numerical computations are presented to illustrate our protocol . |
the term ` animal faith ' in our subtitle is taken from the title of santayana s book skepticism and animal faith"@xcite ; it refers to what gets us through the day and keeps our thought processes going even through our spells of radical doubt .
santayana was concerned to delimit animal faith and to contrast it with other , presumably higher , things .
our agenda here is quite different and is closer to kant s : it is to articulate and transform into explicit principles the animal faith implicit in certain of our concepts that play a key role in science .
once this is done , once these principles become explicit , we ll find that they take on a new life of their own , and are full of surprises .
first , they reveal the simple mathematical structure that unifies quantum and classical .
but then they strangely turn against themselves , revealing their own limitations , and even rather impolitely suggesting that perhaps we should find better ways to get through the day .
this paper , however , is about their more cheerful messages . like
freud s talking cure " , our analysis will start by paying attention to how we talk about everyday things , and then go on to explore subterranean labyrinths in search of hidden meanings .
but it is not a search for what has been repressed , or for what has fallen into the unconsciousness of habit , but for that ancient animal heritage of know how " that we , as thinking people , unconsciously draw upon in formulating our most sophisticated thoughts .
lest this sound too ambitious , let me add that we are confining our analysis to what underlies certain commonplace scientific words , notably : _ information , part , place , event , variable , process , procedure , system , input , output _ and _ cause_. the unconscious beliefs we are searching for are those that belong to the smoothly working hidden machinery behind the easy flow of thoughts in which these words occur .
to put it another way , without certain implicitly held principles these commonplace words , which are indispensable in any discussion of scientific matters , would be quite meaningless , and these are the principles we are trying to capture and articulate .
we shall see that , when these principles are precisely articulated , they become a tightly knit system that has some surprising consequences , including , as mentioned , the two ` core laws ' of quantum mechanics . more generally , the consequences of an analysis of animal faith are of two kinds call them kantian and freudian ; the kantian kind reinforce our animal faith , while the freudian kind force us to question it .
let s begin our analysis with information .
suppose you see something and write down what you see ; this is a _
datum_. _ data _ is what has been seen and noted . to qualify as _ information _ , however , data must have some element of surprise . if you know that a is going to happen , seeing it happen may be gratifying or reassuring , but it is not _
informative_. thus _ information _ is what you have seen , contrasted with what you _ might _ have seen . when you _ describe _ something , you normally supply a number of connected pieces of information , so _ description _ is information broken up into _ connected parts_. these very simple and ordinary observations are beginning to give hints of a pattern , but they come to a halt with the difficult word ` _ parts _ ' .
it s curious how in the long and contentious history of speculations about what the world is made of , almost all the debate has been about what are its _ ingredients _ , and almost none about the more basic problem of what it means for these proposed ingredients to fit together as parts .
taking ordinary material things apart and putting them together is such a natural activity that we blithely extend it to entities of every kind , without ever imagining that this might lead to problems . and
this is certainly the case with information , as when we speak confidently of partial information " , the whole story " , etc .
no doubt we are at ease with such talk because the parts and wholes of information so often coincide with the parts and wholes of the language that conveys it . but
underlying this linguistic idea of part and whole there is a deeper level of meaning , and excavating that deeper level will be the main task of our analysis .
first of all , we must carefully distinguish between two kinds of part- whole relationship : that which we find in space and space - time , and that which we find in material structures like buildings and computers and molecules .
let s call the first relationship _ extension _ and the second _ composition_. the parts of an extension are its _ regions _ , and the identity of a region , i.e. that which makes it different from other regions , is its _ place_. the parts of a composition , on the other hand , are its _ components _ , and the identity of a component is not its place but its _
type_. components are _ interchangeable _ parts ; you can duplicate them , remove them from their places and use them elsewhere , etc . but a region , far from being an interchangeable part , is by its nature unique ; the people of palo alto can leave california , even the buildings of palo alto can leave california , but palo alto itself ca nt leave california since it s a _ place _ in california .
matter , as we currently conceive of it , is composed of elementary particles , the most perfectly interchangeable of parts imaginable , which makes matter a composition par excellence , which is to say , its essential self is nowhere .
there are those today who speak of making matter out of space or space out of matter , but this is nonsense ; you ca nt make extension out of composition nor composition out of extension .
those who claim to have a theory of _ everything _ , even if they are right , have only found half of the holy grail ; the other half is a theory of _ everywhere _ !
the material objects we encounter in everyday life are always both regions and components ; they have both places and types .
thought is a constant dialogue between extension and composition . practically speaking , the important distinction to keep in mind
is between parts that do nt keep their identity when you remove them from their context , like random variables in a joint probability distribution , and those that do , like logic gates in a computer .
only the latter can be used to _ construct _ things . in mathematics ,
the dialogue is between geometry and algebra .
think of analytic geometry , which is just such a dialogue .
in pure euclidean metric geometry , we start out with a homogeneous space whose places are _ points _
, i.e. we can only identify them by _ pointing _ to them as here " or there " or there " etc . in order to keep track of these points ,
we move into algebra .
we do this by giving each point a unique identity as a composition of an x - vector , a y - vector and a z - vector ; this identification scheme is known as a coordinate system .
notice that vectors themselves are not regions or places ; you specify a vector as a certain _ type _ of thing that can be found anywhere .
notice also that we must carefully specify just what it means for one vector to be a part , i.e. a component , of another ; this particular kind of part - whole relation is what defines linear algebra as opposed to other kinds of algebra .
finally , notice that giving coordinates to points does not entirely get rid of _ here _ and _ there _ ; thus the dialogue continues since we must locate the 0 vector _ here _ , and we must then point there " , there " , there " in the directions of x , y and z. nor can we completely get rid of _ this kind _ and _ that kind _ in geometry , since geometric structure involves numerical ratios of distance , so we have _ this _ and _ that _ number , and onward to _ this _ and _ that _ figure , etc . how does the distinction between extension and composition apply to informational structures ?
information accumulates as a progressive _ extension _ of the body of knowledge " .
incoming items of information build on each other , qualify each other , and in general , can only be understood in relationship to their neighbors " .
_ this _ happened , and then _ this _ happened and then _ this _ happened .. " etc . ,
where in each case to know _
what _ happened we must know something about what happened previously . from time to time , though , we _
abstract _ an item of information from its place and give it a kind of autonomy by turning it into a story , a design , a warning , an example , a rule , a law , a procedure etc .
that can be retold or reused in other contexts ; in short , we turn it into a component .
one way to analyze a process is to systematically reconstruct it from components that have been so abstracted ; we ll call this a _ reductive analysis _ and the resulting composition a _
system_. let s reflect a bit on the distinction between process and system , which will play an important role in interpreting our mathematical results .
the word ` process ' , in its most general sense , means something that ` proceeds or moves along ' .
but the word also has the connotation of _ procedure _ as in _
due process_. thus a process of a certain kind refers to that which is allowed to happen , or which can happen , under certain specified conditions .
this is how we shall construe the present technical meaning of the word .
if the possibilities under the specified conditions are assigned probabilities we ll call the process a _ stochastic _ process .
more exactly , a stochastic process is a joint probability distribution on a set of so - called stochastic variables . in chapter 3
, we ll focus on markov processes , which are the simplest and best known stochastic processes , and we ll see that the core quantum laws as they occur in markov processes take the form most familiar to physicists .
a system is a process analyzed into interchangeable parts .
a good system brings order to a complex whole by portraying it as a regular arrangement of a small variety of such parts .
for stochastic systems the paradigm case is a markov chain , which is a chain " of connected copies of a single part called a _
transition matrix _ t@xmath0 .
we can think of t@xmath0 as a representation of a stochastic variable j ( the column variable ) whose probability distribution p(j ) is a function of a free parameter i ( the row variable ) .
connecting means assigning the free parameter i to the stochastic variable j of the prior transition matrix , thereby turning the numbers t@xmath0 into conditional probabilities .
if we assign the free parameter of the first component to an _ unconditioned _ variable , then chaining transmits the definiteness of the first variable down the line to create a joint probability distribution on all the succeeding variables , and the resulting process is a markov process .
the contrast between process and system is roughly that between extension and composition .
a process is extended , its regions being _
events_. like many english nouns ( ` noun ' , for instance ) , the word ` event ' can be taken either in the definite or the indefinite sense ; an event can be something in particular that happened , in which case it has its unique place in space - time , or it can be a particular _ kind _ of happening , as when we speak of the event heads " in probability theory .
one way to describe a process is as an arrangement of events in the second sense , which we can think of as labeling its extended parts by certain of their qualities ; another word for this is _ map - making_. a map is a composition of essences , or predicates , to use the modern term , but the parts of the process that these essences identify need not be removable components , so a map does not in general represent the process itself as a composition . thus we see that there are two ways to analyze a process , or any other extended whole : we can _ map _ it ; this is the way of the naturalist , and also of the explorer , the historian and the astronomer . or we can _ reduce _ it to a heap of autonomous parts , which we then reassemble into a composition that has the same map the naturalist would draw .
nowadays we often hear about the naturalist as the good guy and the reductionist as the bad guy . but building things and taking things apart belong to life as much as exploring and drawing pictures .
when the reductionist goes bad , it s usually because the naturalist hasnt given him a good enough map ; what the reductionist then recreates from his storehouse of interchangeable parts may resemble the naturalist s whole , but the essence of the original is missing . to give to a part the autonomy of a component , we must usually do more than just copy it as it appears in place .
rather , we must de - install " it from its original context , which is to say , we must transform it into a new entity whose features are no longer conditioned by place . to reductively analyze a process is to de - install its partial regions , at least in our imagination , turning them into items that can stand alone , and that can be duplicated in such a way that the duplicates can be reassembled into a duplicate of the original whole . reductive analysis gives the observer a more penetrating gaze , which sees not only what and where things are but where and how they come apart .
the metaphor of de - installing and re - installing components should be a vivid one for anyone who has had to wrestle with the problem of de - installing an app from windows .
in fact , de - installing a computer program is literally a special case of the operation we shall describe here of _ disconnecting _ a component .
we ll go into all this in detail later , but for now let it just serve as a reminder of how much hard work goes into seeing our ordinary world as full of _
things_. this hard work is largely unconscious , and is based on know - how " dating from our dim animal past .
our problem today is that , with the progress of scientific thought , we have wandered into domains where this unconscious skill no longer serves all our needs ; for instance , it falls far short of telling us how to de - install a quark from a nucleus , or an event from a time loop , or for that matter , how to de - install a thought from the stream of consciousness .
thus it has become essential to dredge up and articulate the principles we unconsciously rely on in this work of de - installing things , and consciously learn to do it better , for otherwise we risk disastrous encounters with things we ca nt imagine and therefore ca nt see .
the naturalist , in contrast to the reductionist , sees things in place ; he sees what and where they are in the context of the process as a whole .
this involves seeing extensional separations and boundaries , but it also involves seeing how things _ function _ within the whole . actually , as we ll see , functional structure is best described in terms of the following three - place relation : * separability * : we say that b _ separates _ a and c , or that a and c are _ separable _ at b , if fixing b makes the uncertainty or indefiniteness as to what is the case with a independent of the uncertainty or indefiniteness as to what is the case with c. the most familiar example of this notion is the separability of the future from the past by the present : if we know everything about the present , then getting more information about the past does not give us any more information about the future , and vice versa , or so we suppose .
( remember , we are now examining our presuppositions , not looking for objective truths ) .
note that this says nothing about the determinism of the future by the present all it says is that the past and future , however else they may be related , only communicate " with each other via the present . which brings up another familiar example : the separability of two communicating parties by their line of communication .
if a and c are talking by phone , tapping their phone line ( b ) can tell you everything that s going on between them ; whatever else is going on with the two of them at the time is going on with each of them independently . another term for separability is _
conditional independence_. in the theory of stochastic processes the two terms have essentially the same meaning , though we ll keep them both to refer to slightly different mathematical formulations ( see chapter 2 . ) * independence , unconditional and conditional * : events a and c are called _ independent _ if the probability of a&c is the probability of a times the probability of c. a and c are called _ conditionally independent _ given condition b if they are independent in the probability distribution conditioned by b , i.e. if p(a&c@xmath1b ) = p(a@xmath1b)p(c@xmath1b ) .
if for some random variable x , a and c are independent given condition x = k , where k is any value of x , then we say that x _ separates _ a and c. in the above telephone example , if we regard b as a random variable , then b separates a and c. since every probability distribution is conditioned by something , there is a sense in which all independence is conditional .
the relation of independence , like every notion of being different or separate , is really a three - term relation : a and c are independent in a certain context b. we have just considered the case where that context is the result of placing a certain condition on a probability distribution .
but the more familiar case is that in which the context is the result of _ removing _ a certain condition on a probability distribution .
notice that although there is only one way to impose a condition , there are many ways to remove a condition ; this is one reason why there are so many different ways to represent a given process as a system .
the converse of deconditioned " independence is : * conditional dependence * : we say that a and c are _ conditionally dependent _ given b if p(a&c@xmath1b ) is unequal to p(a@xmath1b)p(c@xmath1b ) .
a very important kind of conditional dependence is that in which the condition b is of the form x = y , where , without b , a&x is independent of y&c .
this is called linking , and we ll return to it shortly .
there is a fundamental theorem that relates linking to separability : * disconnection theorem * : process a&x&c is separable at x if and only if there exists a process in which a&x is independent of y&c which reduces to a&x&c under the condition x = y . this , and other theorems stated in this chapter , are proved in chapter 3 .
speaking more generally , what makes separability so important is that it always coincides with the possibility of breaking a process into two de - installed components of some kind .
we separate a and c by listening in on their phone line b , which is at the same time their connection and their common functional boundary , at least as far as their conversation goes . on the other hand
we _ disconnect _ a and c , turning them into autonomous components , by _ cutting _ their phone line ; this ends their conversation because their phone line is their separation boundary .
when we mark the separation boundaries in a process , we may be functioning as naturalists , but we are also like the butcher with his blue pen marking the chops in a carcass . to reconnect a and c so as to restore the whole process is to equate what is the case with b1 and what is the case with b2 , thus restoring the original situation of a&b&c .
just what it means to equate b1 with b2 turns out to be less obvious than it sounds , and in fact can be understood in two very different ways ; we ll call these two ways _ i - o connection _ and _ linking_. understanding exactly how an i - o connection differs from a link is the key to understanding how quantum processes differ from classical processes such as phone conversations .
let s start with i - o connection .
we ll simplify slightly by supposing that a is talking to c on a one - way line b. if that line is cut , what happens to the signal voltage at the two cut ends b1 and b2 ?
the cut line b1 from a is an _ output_.
this means , ideally , that what comes down the wire depends only on a , so the signal at b1 is unchanged by the cut .
the cut line b2 from c is an _ input _
, however , and our description of the situation does nt specify how inputs behave when they are disconnected .
an open line will normally produce white noise , sometimes mixed with hum and faint radio signals .
however , and this is a crucial point , the designer of a system usually does nt have to take into account disconnected inputs ; he only needs to think about how the system behaves when inputs are driven " by outputs , whether from the user or from other components . to put it another way , the designer can regard inputs as
_ free variables _ ; his components are _ open processes _ , analogous to open formulae in the predicate calculus .
open processes are interconnected by equating inputs to outputs , which means _ assigning _ free variables to non - free or _ bound
_ variables ( assigning variables is a familiar concept for computer programmers . ) a composition created in this way will be called an _
i - o system_. if all the free variables of the components are assigned , we call the system _ closed _ ; otherwise it s an _ open _ system .
we are on familiar ground here ; this is how engineers and programmers are taught to think these days . once again , here are the main ideas : * input * : a free variable in an open process . * i - o connection * : the assignment of a free variable to a bound variable . *
output * : the bound variable to which a free variable is assigned .
disconnecting is just the reverse of connecting .
more exactly : * i - o disconnection * : a and c are disconnected at variable b means we are given two processes a&b1 and b2&c with b2 free such that when b2 is assigned to b1 , the result is the original process a&b&c . * i - o system * : a composition of connected open processes .
every i - o system represents a process , but we must be careful to distinguish this process from the system , just as we must be careful to distinguish a vector from its representation as a sum in a particular basis .
there s something not quite right about all this .
after all , a free variable is only an abstract linguistic entity , whereas our cut wire , lying there on the ground and picking up white noise , remains very much a real material object .
the parts of a telephone system are not indefinite processes " they are just as definite as houses and trees and stones . or so says animal faith .
there is in fact a way to describe the situation that s more in keeping with this bit of animal faith , which is to replace the concept of assigning a free variable by the concept of _ linking _ a so - called _ white _ variable . *
white variable * : a random variable on which there is a uniform probability distribution . in the case of a discrete variable , whiteness means that every value is equally probable , and hence the variable has a finite number of values , since their probabilities must add up to 1 .
a white real variable has a uniform probability density which must integrate to 1 , so it s range is bounded .
* theorem * : if x and y are independent random variables with the same range , and y is white , then the condition x = y on their joint probability distribution does not change the distribution on x. in our telephone example , let s now forget about free variables and think of the components a&b1 and b2&c as closed processes . taken together ,
the two can be regarded as independent regions in the process a&b1&b2&c .
when we connect b1 to b2 , the result is a new process a&b&c .
how does this new process differ from the old ?
the obvious answer is that it is just like the old one except for its behavior being restricted by the condition that b1 equals b2 .
connecting two wire ends means making their signals equal .
it is not immediately clear how this answer , which is couched in the physical language of broken wires , signal voltages etc .
, applies to processes as such .
( remember , a process , as we are now defining it , is simply a joint probability distribution on a set of variables . ) .
however , the phrase .. restricted by the condition .. " gives a strong hint .
to say the probability distribution d1 is just like the probability distribution d2 restricted by the condition c " means that you get d2 from d1 by conditioning all event probabilities in d1 by event c. thus our answer is basically this : to connect any two variables x and y , condition all of the probabilities of the disconnected system by the event x = y . this particular bit of animal faith brought to light will turn out to be the key to making sense out of quantum mechanics , so let s forthwith coin words that nail it down : * links * : place a condition of the form x = y on two variables x and y of a stochastic process , thereby creating a new process in which the unconditional probability p(e ) of any event e is p(e @xmath1 x = y ) in the old process .
this condition is called a link . * link system * : a stochastic process plus a set of links .
we ll go over all of this in considerably more detail in chapters 2 and 3 .
a summary of the relevant concepts , terminology and notation of probability theory will be found in chapter 3 , section 1 .
let s now briefly look at some high spots .
first , if x and y are independent and y is white , then the link between them behaves exactly like an i - o connection , with y acting like the input ; this follows immediately from the above theorem .
we ll show in chapter 3 that we can neatly translate any i - o system into a link system which represents the same process and has essentially the same formal description .
engineers should have no problem adapting to link descriptions .
second , and this is extremely important , we ll prove that this natural mapping of i - o connections to links does not go backward !
though every i - o system is equivalent to a link system , most link systems have no i - o equivalents , and even when one can make an i - o model of a link process , that model is usually exponentially more complicated than some link representation of the same process .
quantum systems are among those having no i - o counterparts , though quantum measurement systems do , as we ll see in section 3.7 .
if technology can learn to really deal with this larger class of non i - o systems , it will turn into a fundamentally new kind of enterprise .
third , a crucial definition , that of a _ link state _ : * link state * : _ given _ any link x = y , the _ state _ of that link is defined as the matrix representing the joint probability distribution on x and y without the link ( actually , it s this matrix normalized by dividing by its trace , but that s a small detail . )
this is the concept that replaces the quantum wave function , and does away with the problem of the collapse of the wave function and other related quantum nuisances , as we discuss later . in the introduction
we briefly mentioned caloric .
heat , in the eighteenth century , was regarded as a fluid called caloric which was somehow related to the mechanical behavior of objects , but subject to its own non - mechanical laws .
a series of experiments during the first half of the nineteenth century , starting with count rumford s crude qualitative observations , and ending with the definitive work of joule and helmholz , convinced physicists that they did nt actually need to postulate a new substance called caloric , since heat is simply energy .
when it was further realized that this energy is the random kinetic energy of molecular motion , thermodynamics became statistical mechanics .
history to some extent repeated itself in the early twentieth century with de broglie s wave mechanics .
his matter waves " were at first thought to be waves in some new kind of fluid , the quantum analogue of caloric , but after born discovered his quantum probability rule , they lost most of their substantiality , and physicists began to call them probability waves " .
quantum mechanics , however , did nt go the way of thermodynamics .
because the so - called probability wave is really a wave of the square root of probability , quantum caloric " did nt seem to be reducible to anything already known . and
, alas , it is still with us , ignored by most working physicists , but hanging around the edge of physics in a kind of limbo where from time to time it is reworked and touted as a wonderful new discovery .
but fortunately this unhappy state of affairs is almost over . here , in a nutshell ,
is how link states reduce quantum caloric " to something already known , namely probability : suppose x and y are independent random variables .
then the probability that x and y will both have some value k is the product of the probabilities that they will have k separately , i.e. p(x = k & y = k ) = p(x = k)p(y = k ) .
let s call p(x = k ) and p(y = k ) the unlinked probabilities of x being k and y being k. now suppose we impose the link condition x = y .
the linked probability of x = k is then the probability , conditioned by x = y , that both x = k and y = k .
the probability that x is k , as a function of k , is proportional to p(x = k)p(y = k ) . in short , linked probabilities
are always quadratic in unlinked probabilities . if the distributions on x and y are identical , which is the quantum situation , then linked probabilities are the squares of unlinked probabilities .
that s essentially all there is to it !
quantum amplitudes are not the intensities of some mysterious new fluid , but are simply unlinked probabilities .
one might ask probabilities of what ? " , but we ll see that we do nt have to answer this question , just as we do nt have to answer the question of what sets of things have numbers x and y in order to understand the meaning and truth of xy = yx
. a generalized form of quantum amplitudes will be found in any statistical situation whatsoever if we subject that situation to link analysis .
quantum amplitudes proper will show up if probabilities can go negative ( we ll return to this in a minute ) and we de - install components in such a way that the link states possess quantum symmetries ; as we ll see in chapter 3 , this is always possible , though not always advisable .
link states are quantum states represented as von neumann density matrices .
however , we can start in the more familiar way with quantum states as vectors over the amplitude field , and when we apply link analysis to the schrodinger equation , amplitudes still turn out to be unlinked probabilities . here , very roughly , is why this is so ; the details are in chapter 3 . in a continuous markov chain , regarded as an i - o system
, the probability distribution on the state variable can be represented as a vector which evolves in time according to the law v = t(v ) , where v is the distribution vector at time t , v the vector at t+dt , and t the differential transition matrix .
the schrodinger equation , expressed in the language of hilbert space , has exactly the same form , the difference being that the components of v are amplitudes rather than probabilities , and t is a unitary matrix rather than a transition matrix .
let s now represent the markov chain as a link system rather than an i - o system .
the changing state vector is now the diagonal of a link matrix s that changes according to the law s = tst@xmath2 .
if we think of s and s as density operators , this is the von neumann generalization of v = t(v ) .
an immediate corollary is the born probability rule in the generalized von neumann form prob(p ) = trace(ps ) .
the rule for turning an i - o system into a link system gives s the form @xmath1 v @xmath3 w @xmath1 , where w is a white " vector , i.e. it represents a uniform probability distribution , so it is only v that varies with time .
the diagonal entries are quadratic in the entries of v and w , but since the constant entries of w are equal they factor out , so the probability vector is just v , the same probability vector that occurs in the i - o system . next ,
let s do the same with the quantum chain .
again we have s = tst@xmath2 , but now t is unitary . and
again we have s = @xmath1v@xmath3w@xmath1 , but now instead of w being white , we have w = v , which makes the diagonal entries into the squares of the entries in v , i.e. state probabilities are the squares of amplitudes " ! the amplitudes are no longer the values of some mysterious wave function but are simply the probabilities in the joint distribution on a pair of unlinked variables .
quantum caloric is nowhere in sight ! that takes care of the main obstacle to a statistical interpretation of quantum mechanics .
the square law is no longer a reason not to regard probability waves as true waves of probability .
it s true , the wave amplitudes are nt the probabilities that we measure ; rather , they are the probabilities that we would measure if the quantum process were disconnected at a separation point .
but the same is true in a markov chain , analyzed as a link process , the only difference being that in the markov chain the output probabilities do nt change when the output is linked to an input . in a quantum chain ,
input " and output " have identical distributions , and since unlinked state variables are independent , linking yields the unlinked probabilities squared .
there is still the problem of amplitudes going negative ( they can also go imaginary , but as mackey observed , the scalar i can be interpreted as a unitary symmetry in real quantum mechanics)@xcite . what could possibly be the meaning of negative probabilities
? how can there be fewer than 0 things having a certain property p ? how can an event e occur less than 0 times ?
these are not questions that can be answered by constructing some ingenious mechanism , since they involve logic itself .
earlier we compared negative probabilities to negative piles of gold , my point being that there is no problem in constructing a useful mathematics that incorporates them .
but that does nt tell us what they really are , and we now doubt very much that their deeper significance can be understood at all within the scope of science as we know it .
this could be enough to drive many people back to quantum caloric .
but before you join the stampede , be aware that the logical mysteries in quantum mechanics do nt come from the link state interpretation .
they come from the bare fact of amplitude cancellation , and practically stare you in the face with epr
. how can it be that a is possible , and also that b is possible , but that ( a or b ) is impossible ?
that s the mystery of the two - slit experiment . here `` or '' is the logician s `` or '' which allows a and b both to be true , a true and b false , a false and b true , but requires ( a or b ) to be false if both are false .
the double slit example we have is mind is : case a : slit a open leading to a detection ; case b : slit b open leading to a detection ; case ( a or b ) : both slits open but there is no detection .
von neumann was well aware of the logical strangeness of quantum phenomena , and tried to cope with it by means of a non - boolean construction that he called quantum logic .
this turned out to be a dead end , but other strange logics dating from the period of quantum logic show more promise .
enough of this for now , however .
we ll return briefly to the logic of negative case - counting in chapter 2 , but it s too big a subject to do justice to here ; suffice to say that the logical anomalies do nt vitiate our simple mathematics .
so where do we stand with quantum mechanics ?
we ve gotten rid of matter waves " but the mystery of quantum interference remains , since to derive the schrodinger equation in link theory does require negative probabilities , even in the process itself .
but now , at least , the mystery of interference has become a _
simple _ mystery , and there is no longer any reason to think of it as belonging to physics in particular .
once one has really grasped how universal the two generalized quantum laws are at the case - counting level , it s as hard to take seriously the arcane models " of quantum mechanics that abound today as it is to take seriously the ancient sheep farmer s spirit of tranquillity " after one has understood that xy = yx .
but we re afraid we are getting somewhat ahead of our story .
we started out to psychoanalyze animal faith , to dredge up some of the unconscious beliefs that underlie our scientific thinking .
wait a minute ! " , you protest ,
i do nt know about you , but i ca nt find any of this fancy mathematics in my unconscious ! "
fair enough , neither of the authors can either .
it s the analyst s job to propose , not to pronounce , and his proposals must always be confirmed by the analysand , who in this case is all of us .
however , what we re looking for in our unconscious is certainly not fancy mathematics .
animal faith in itself , i.e. in place , ca nt even be put into words .
to find " it means to articulate something that does the same job , namely that of supporting our habitual modes of conscious thinking . in our current jargon , to articulate an item of animal faith means to de - install it , i.e. to transform it into a proposition that can stand alone .
however , once articulated and understood , this new creation can then be put to the test by re - installing it in our intuition : does it ring a bell ?
do we think ah , yes , that s what i ve always thought , even if perhaps i never quite said it that way . " ?
if the items we have articulated pass this kind of test , then we can confidently move on to the interesting surprises .
quantum mechanics is one such interesting surprise ; another , which is the topic of chapter 2 , is a very kantian answer to hume s skepticism about the existence of causes .
kant s own answer was that we necessarily see causes everywhere because , to paraphrase the way bertrand russell put it , we wear causal - colored glasses " @xcite . by combining the philosopher von wright s insight into how we operationally define causality with our analysis of process and system , we ll bring this statement up to date : we necessarily see input - output systems everywhere .
since input - output systems are closely related to computer models , it could be said that we necessarily see computers everywhere - we all wear computer - colored glasses !
if all this sounds like the grand march of progress , be aware that the two surprises above are not entirely harmonious .
what makes our kantian analysis of causality possible is the fact that the separability structure of a process only partly determines the form of the link matrix s , and we can show that it is always possible to choose a causal s ( i.e. , an s with an input ) .
however , quantum s s are not causal ! furthermore , and this is crucial , the choice of s for a given disconnection will , in general , affect the separability structure of the disconnected parts .
thus things that come apart neatly with one type of s may exhibit bad non - localities " for another type of s. epr comes apart very neatly with a quantum analysis , but any causal analysis compatible with the data will exhibit non - locality , as bell s theorem shows .
thus , though we are always perfectly free to choose causal s s and thus turn the universe into a giant computer , this computer may have a gosh - awful tangle of long distance wiring that could be eliminated by using other s forms .
kant has had his say now it s freud s turn .
freud s project was not to justify our beliefs but to change our minds .
he was , of course , working with people whose minds were very much in need of changing ( which of course does nt include you and me ) .
still , he may be able to give us a few useful tips .
as long as our assumptions remain unconscious , we have no choice but to believe them , or rather , we have no choice but to act as if we believed them
. the question of belief does nt actually arise until we are able to contemplate and weigh alternatives .
for this to happen with some article of animal faith , we must first of all raise it up from the unconscious depths in such a way that it becomes a proposition , i.e. we must de - install it .
then one of two things may happen .
we may take a clear look at this proposition and recognize its absurdity : wait a minute , my mother is nt really threatening to lock me in my crib if i try to go outdoors ! " , or whatnot . or , we may see that the proposition is obviously true .
if this happens , then another question arises : is this truth objective " , or does it depend on how you look at it ? how do we answer this last question ? how do we find out whether something depends on how we look at it ?
one way is to see whether we can look at it differently .
if we believe that all print is too fuzzy to read , we might try on glasses . of course we may have to experiment to find the right glasses .
kant s error was to confuse our glasses with our eyes , so - to - speak .
the kantian conclusion of chapters 2 and 3 is that we all wear causal glasses , which turn out to be the same as computer - colored glasses .
does this mean that we are forever fated to see everything as a computer , or can we take these glasses off ? as mentioned , causal glasses "
correspond to a particular kind of link matrix .
we can work mathematically with very different kinds of link matrices , and indeed we must do so in order to practice quantum physics .
chapter 3 will make this very clear , and will point us in the direction of an expanded kind of science that works with the full range of acausal state matrices .
but the question still remains : can we , as human beings living our daily lives , ever take off our causal glasses ? for mathematicians , all things are possible
. the experimenter who is willing to push hard enough may occasionally force nature to reveal its unnatural proclivities .
but you and i , when we are just trying to get on with it , must constantly ask questions whose answers begin with because " . that s our human nature . but what will be the effect on how we live and act if we can learn from the coming science how to go beyond why " and because " ? what sort of beings might we someday become ?
all modern discussions of causality begin with hume . it was hume who first clearly pointed out that merely knowing the regular succession of events does not tell us what causes what .
the fact that b always follows a gives us no grounds for concluding that a forces " b to happen .
why speak of causes at all , then ?
one school of thought says we should stop doing so ; here is the young bertrand russell on the subject@xcite : ... the reason why physics has ceased to look for causes is that , in fact , there are no such things .
the law of causality , i believe , like much that passes muster among philosophers , is a relic of a bygone age , surviving , like the monarchy , only because it is erroneously supposed to do no harm . "
the mature bertrand russell , however , found out that like the rest of us he was unable to say very much outside of logic and pure mathematics without bringing in causes .
cause and effect pervades everyday life ; the solution to hume s problem ca nt be confined to the ivory tower of philosophy .
a recent news item reported the discovery of a correlation in mexico city between heavy chili pepper consumption and stomach cancer . for us
chili pepper lovers , this was most unwelcome news : do chilies cause cancer , or might there be some other explanation ? what s at stake here is no mere philosophical abstraction ; it s whether we have to stop eating chilies !
if there is no causal link between a and b , then , even if there is a high correlation between a and b , our wanting to bring about or prevent b is no reason to do anything about a. but if we learn that a causes b , or that it significantly increases the probability of b , then whatever power we may have over a becomes power over b too .
the understanding of cause and effect is essential for practical life , indeed it is essential for our very survival , because it is the knowledge of means and ends .
this obvious fact of life has been curiously neglected by philosophers .
the logician g. h. von wright , however , was an exception ; he actually went so far as to propose that the ends - means relation be used as the defining criterion of causality @xcite .
that is , he would have us say that , by definition , a causes b if and only if a can be used as a means to b , and more generally , that there is a causal connection from a to b if and only if we can change b by changing a. notice that this is an operational definition : it says that the presence or absence of a causal connection between a and b can be detected by performing a certain operation , which is to change a and notice whether there is a concomitant change in b. that is , of course , just how diet scientists would go about seeing whether chilies cause cancer .
does von wright s definition solve hume s problem ?
it certainly gives causality an empirical anchor that can not be neglected in any proposed solution .
however , a very important question still remains : is it possible to find formal patterns in the data itself that correspond to the causal connections revealed by the dispositional relation of means - to - ends ?
this is a difficult and subtle question , and the search for its answer ultimately takes us into territory where the question itself essentially disappears .
first , though , there is a lot of work to be done closer to home .
the people who really live with causality are not philosophers but engineers , and curiously enough , they rarely use the word cause " . that s not because , as russell said , there is no such thing " , but because they have developed a much richer and more discriminating vocabulary for the various causal structures and relationships they deal with .
variable , ` input ' , ` output ' , ` process ' , and ` function ' are standard causal terms .
computer science has added some important newcomers such as , ` assignment ' , ` call ' , and ` subroutine ' .
indeed , the concept of subroutine neatly revives the whole aristotelian foursome of causes : the material cause of a subroutine is the computer it runs in , the formal cause is its source code , the efficient cause is its call , and the final cause is the purpose it serves in the calling program .
it s clear from this brief survey of terms that , for the engineer , causality is an aspect of how something is composed , not of how it is extended .
the engineer s job is to put together certain components into a system that realizes a causal process , i.e. a process having inputs that produce certain desired outputs .
furthermore , the components themselves are causal processes , and putting components together means using the outputs of certain components to causally control the inputs of others . as remarked in chapter 1 , the information engineer does not normally concern himself with the behavior of unconnected inputs . for him , and more generally for low power transfer , the behavior of input b connected to output a is usually the behavior of a in the absence of the connection , so b can be regarded as an indeterminate or _ free _ parameter .
thus , what the engineer designs , in our current jargon , is a functional i - o system ; its components are what we called _
open processes_. the functional dependence of outputs on inputs is called a _
transfer function_. the concept of a transfer function clearly matches von wright s concept of causality as the potential for the relationship of means to ends , so it makes sense to refer to functional i - o systems as _
causal systems_. in a causal system , so defined , the outputs of each component are functions of its inputs , and their composition by assignment of inputs to outputs is functional composition .
however , we now know that , in the case of macroscopic components , the functional dependence of output on input is a large - number effect resting on the statistical behavior of atoms .
a more accurate , and more general , account of the engineer s systems makes the output of a component into a _
function of the input , with the causal transfer function replaced by a transition matrix of conditional probabilities ; these are the i - o systems of chapter 1 .
the composition of transfer functions " is no longer functional composition but matrix multiplication , which reduces to functional composition in the special cases where the transition matrices contain only 0 s and 1 s . having carefully noted the distinction between the mathematician s function and the engineer s transfer function , we ll now relax our language a bit and use the word s cause " and causal " in connection with both .
thus we ll often refer to i - o systems as causal systems and to i - o states as causal states , especially in a context where we are distinguishing these from quantum systems and states , which are essentially a - causal .
a more accurate qualifier would be statistico - causal , but it s too much of a mouthful . since `` classical physics '' is often taken to imply strict determinism , we emphasize that the `` classical '' systems we consider in this paper _ always _ contain stochastic elements .
suppose an input is connected to an output
how do you know which is which ? to find out you disconnect them ; the output is the one that continues to do by itself what they did together when connected - another operational test of causality . but
suppose you ca nt disconnect them , and suppose also that you ca nt manipulate the inputs to the system ; all you can do is watch .
then how do you identify causality ?
is there nothing we can learn about what causes what by just watching ?
hume s problem is still with us .
here s another take on hume s problem : are there patterns in the joint probability distribution of an observed process that mark the causal relations we would see if we were to experiment with its disconnected parts ? to put it another way , do cause and effect belong to the process itself , or are they artifacts that result from taking the process apart ? the theory of link systems throws new light on this question .
recall the disconnection theorem , which says that a process can be disconnected or de - linked " at x if and only if it is separable at x. this shows us a way to transform a process into a link system : first disconnect x into x1 and x2 , where x2 is white " ( corresponding to an input ) , which produces two independent processes , call them a and b. then link a and b by imposing the condition x1=x2 ; this gives us a two - component link system exhibiting our original process .
now do the same for separating variables in a and b etc .
until there are no separation points left .
the result is a system which factors " the original process into prime " components that can not be further subdivided .
recall that separability belongs to the process itself , i.e. it is intrinsic to the probability distribution .
thus the disconnection theorem reveals severe limitations on how we can construe the correlations in a process as causal .
a highly correlated joint probability distribution need not have any separating variables at all , so causality requires something besides mere correlation in the data itself .
we have at least a partial answer to hume .
we ll see in chapter 3 that the earmarks of causality in the data itself can actually be quite distinctive .
given the temporal order of the events in the process , it turns out that there is a unique prime factorization into components with white inputs .
we learned in chapter 1 , and it will be proved in chapter 3 , that such systems are in natural 1 - 1 correspondence with i - o systems representing the same processes . thus from a knowledge of the given alone ,
i.e. of the statistical behavior of the system without our interference , we can completely describe the i - o components we would get if we took the process apart at its ( intrinsic ) separation boundaries .
the cause and effect relation , which we first located in the transfer functions and interconnections of the parts , is already _ there _ in the process itself !
hume , what do you say to that ? not so fast , sir ; you were too glib with your prime factorization .
why should the x2 s be white ? and , if you regard what is given to our understanding to be only the probability of certain events , then on what basis do you bring in temporal order ? if you must call upon such invisible help to reveal causality in the bare order of correlation , you do but reinforce my conviction that causes are phantoms of the imagination . " good points
let s translate them into the modern idiom .
it will help to focus our ideas if we confine ourselves to processes in which the separating boundaries are all in a line ; these are called markov processes .
their definition can be stated in terms of a three - term relation among events called the markov property , which is defined as follows : * markov property * : to say that the triple of events ( a , b , c ) have the markov property means that they satisfy the equation p(c
@xmath1 a&b ) = p(c @xmath1 b ) , which we ll refer to as the _ markov equation_. * markov process * : a process in which the variables are indexed by a time parameter such that , at any time , if a is any event not involving future variables , and b is the state of the present , and c is any event not involving past variables , then a , b and c satisfy the markov equation .
recall that the probability p(x @xmath1 y ) of x conditioned by y is defined as p(x&y)/p(y ) .
bearing this in mind , notice that if we multiply both sides of the markov equation by p(a&b)/p(b ) we get the equation p(a&c @xmath1 b ) = p(a @xmath1 b)p(c @xmath1 b ) .
this relationship among a , b and c is known as _ conditional independence _ ; in words , a is independent of c , given b. recall that in chapter 1 we referred to separability as conditional independence ; this equation defines what we were referring to .
event b separates events a and c means that , given b , the probability of a&c is the product of the probabilities of a and c. it turns out to be more convenient to define separability by a slightly different equation , which is gotten by multiplying both sides of the markov equation by p(a&b)p(b ) .
* separability * : to say that b separates a and c means that p(a&b&c)p(b ) = p(a&b)p(b&c ) . if none of the joint probabilities of a , b and c are zero , then the markov property , conditional independence , and separability are all logically equivalent .
however , for conditional independence to make sense , p(b ) can not be zero , while for the markov equation to make sense p(a&b ) must also not be zero . since separability always makes sense , whatever the probabilities , it s the most convenient of the three to work with .
such fine points need nt concern us now , however ; the really important point is this : conditional independence and separability are symmetrical in a and c , which means that the defining property of a markov process is symmetrical in time ! think about that for a minute .
the markov equation seems to have a temporal arrow , since it s about conditioning of the future by the present and past , but we now see that this is an illusion .
our argument for causality being in the process itself rested on a theorem that assumes a given temporal order of events , and in the usual discussions of markov processes , temporal order is taken for granted .
but hume s challenge was to find causality in the bare order of correlation .
we now see that this bare order does not distinguish past from future , so it would seem that mr .
hume is right : cause and effect really are in the eye of the beholder .
kant , who has been listening to this discussion with growing impatience , can restrain himself no longer .
since his style is notoriously obscure , we will freely translate his remarks , with a little help from the eminent historian of philosophy harald hoffding@xcite .
we gather that by ` process ' you mean something that is accessible to experience .
now experience , in contrast to mere imagination , is a complex composed of elements some of which are due to the faculty of knowledge itself , while others are the result of the way in which this faculty is determined to activity from without . as i have shown , space and time are forms of our perception . for whatever
the nature of our sensations , and however much they may change , the spatial and temporal relations in which their content is presented to us remain the same ; a space or a time does not change , whatever be its filling out .
it follows that the temporal order of the process is given , since it belongs to the form whereby the outer determinant of our experience of the process can manifest itself in experience as a process .
hume is correct that cause and effect are in the ` eye of the beholder ' , but wrong to imagine that we could see otherwise than as we do .
" kant s ambition was to describe once and for all the forms " that belong to the faculty of knowledge " , leaving the outer determinants " of experience in a limbo where they have become known as things in themselves " .
today we are more modest . for one thing
, we no longer make his absolute distinction between the subjective and the objective ; rather we see objectivity as a matter of degree . also , instead of trying to directly analyze the faculty of knowledge , we often rest content with characterizing the contribution of the subject to an experience in terms of a choice of viewpoint " , and of the object as a particular feature of the experience that is invariant under the choice of viewpoint .
the subjective - objective boundary is of course quite mobile , and depends on the range of possible viewpoints we are considering .
even transposed into a modern setting , kant s observations about time direction are telling .
so far we have been taking the notion of process as simply given .
but we must ask : given to what ? if given merely to imagination , then of course processes can be anything we want them to be , and they can go backwards or forwards or any which way in time .
but if we are empirical scientists , then processes are something given to _ experience _ , at least potentially , which means that in thinking about processes going backwards we must consider whether we could actually observe such a process . and
here we must heartily agree with kant that the answer is no , since experience itself has an inexorable temporal arrow , whatever its content .
indeed the process as something given to experience is given one piece at a time , and the order in which these pieces are given is precisely the temporal order that we find in the process itself .
this being the case , what could it possibly mean that time goes backwards ? we ve been describing the process itself as what is given , but here s something else to keep in mind : the time reversal of the _ given _ is the _ taken away _ !
let s be more concrete .
suppose we are observing a process and writing down what we observe .
our observations accumulate in a diary which starts with 100 blank pages and progressively fills with writing . now
each letter that we write reduces the nominal possibilities for what the diary _ could _
contain by a factor of 26 ; this is the shannonian sense in which it constitutes information .
thus the process itself proceeds step - by - step alongside of a process of information gathering , which progressively reduces the vast initial range of possible states of the diary to one state , the chronicle of _
what actually happened_. we decided in chapter 1 that it is not a particular chronology that we will call a process , but rather the conditions and rules that govern it .
let us then imagine an enormous series of repetitions governed by the same conditions and rules , each producing a diary of what actually happened .
we now are given an enormous heap of diaries from whose statistics we abstract the joint probability distribution that we having been calling a stochastic process . the question is whether the stochastic process so abstracted exists as a thing - in itself apart from its potentiality of determining our faculty of knowledge to activity " , namely , the activity that produces this heap of diaries .
the kantian answer is no .
the very _ idea _ of process is inseparable from the idea of a progressive accumulation of observed information in a diary .
such a progression has of course a time arrow , that of the progressive narrowing of a range of possibilities .
thus a markov process backwards in time is nonsense . to be given
a process without a time arrow is inconceivable , since the given - ness of a process is conceptually tied to the given - ness of successive stages of its presentation to a potential observer .
it took a lot of fancy footwork , but time and with it causality seem at last to have returned to their familiar shapes .
finally we are back safe in the bosom of animal faith . or are we ?
kant says it s ok to believe what our animal faith tells us , because we ca nt help it such beliefs belong to the very faculty of knowledge " . darwin s theory of evolution also suggests that it s ok , but for a somewhat different reason , namely that our faculty of knowledge evolved to enable us to survive in an all- too - real world of hostile and dangerous
things - in - themselves " . the two messages , despite their opposite philosophical starting points , are actually not all that far apart .
darwin s message applies an important corrective to kant s , however , which is that in analyzing our faculty of knowledge we must take into account the special conditions that obtained in the environment where that faculty evolved .
thus , the belief that the world is flat is certainly true enough " for creatures who never travel more than a few miles from where they are born , but is badly wrong for long - distance navigators .
homo sapiens is perhaps alone among the animals in being able to modify his animal beliefs in response to drastic changes in the environment , including those of his own making .
furthermore , and in this we humans are probably unique , our faculty of knowledge includes a faculty of abstract reasoning that can extrapolate to conditions far beyond our present environment .
sometimes this precedes a big move : think how we used every branch of modern science to prepare ourselves for exploring and colonizing outer space .
kant lived in the age of enlightenment whose idols were euclid and newton , and rather naively incorporated their systems into his intrinsic " faculty of knowledge , thus confusing a brief phase of human culture with something final and absolute .
one wonders what he would have said about einstein s theory of relativity .
kant thought it was impossible to experience space as non - euclidean , and yet today we have good reason to believe that certain regions of space are quite non - euclidean , and we know at least some of the ways in which this impinges on experience .
the most drastic departures from euclid involve warps not only in space but in time , producing loops where time closes back on itself . for the inhabitants of such a loop ,
if there be such , the future is also the past ! needless to say , time loops pose a serious problem for our neat kantian explanation of time - direction and causality .
if the future is the past , which way does our experience go ? how can there be a diary today about the future ? in brief , _
what happens _ ?
if what happens is what could be experienced by someone involved in the happening , what would such experiences in a time loop be like ? and what would it mean for experience itself to happen ? before getting further into this quagmire , let s briefly check in with the authorities to see whether time loops are anything more than wild science fiction .
we ll start with stephen hawking .
here is a recent item from the new york sunday times : october 1 , 1995 . in a u - turn that has sent shock waves through the universe , professor stephen hawking , britain s leading cosmic physicist ,
has accepted the possibility of time travel .
having ridiculed the concept for years , hawking now says that it is not just a possibility but one on which the government should spend money . "
hawking gives a quick summary of the reasons for his u - turn in the preface to a new book called _ the physics of star trek _ by astronomer lawrence krauss@xcite .
first he addresses the problem of faster - than - light travel , a staple of space opera despite its being forbidden by special relativity , and concludes that fortunately , einstein s general theory of relativity allows the possibility for a way around this difficulty . "
he then turns to time travel : one of the consequences of rapid interstellar travel is that one could also travel backwards in time .
imagine the outcry about the waste of taxpayer s money if it were known that the national science foundation were supporting research on time travel . for that reason , scientists working in this field have to disguise their real interest by using technical terms like ` closed timelike curves ' that are code for time travel .
nevertheless , today s science fiction is often tomorrow s science fact . "
one of the scientists working in this field is the super - string theorist michio kaku . here are two brief excerpts from his recent book _
hyperspace_@xcite : in june 1988 , three physicists ( kip thorne and michale morris at the california institute of technology and ulvi yurtsever at the university of michigan ) made the first serious proposal for a time machine .
they convinced the editors of physical review letters , one of the most distinguished publications in the world , that their work merited serious consideration . " .... if time travel is possible , then the laws of causality crumble .
in fact , all of history as we know it might collapse as well .
" whatever may be the prospects for a time machine , and we personally do nt think they are very good , the conceptual problem posed by time loops in outer space is problem enough in itself . even if we remain comfortably at home in our causal near - euclidean space - time and never venture near a cosmic worm - hole , we still live in the same universe with such aberrations . in the larger picture ,
their time is of a piece with our own , and thus the question still remains : what happens there ?
since this question ca nt be answered by a narrative , what kind of answer should we look for ? here s another brief quote from kaku .
however , the concentrations of matter - energy necessary to bend time backward are so vast that general relativity breaks down and quantum corrections begin to dominate over relativity."@xcite a complementary point is made in an article called the quantum physics of time travel " by david deutch and michael lockwood in the scientific american , march 1994@xcite : quantum mechanics may necessitate the presence of closed time - like curves . these , while hard to find on large scales , may well be plentiful at the submicroscopic scale , where the effects of quantum mechanics predominate . "
these are very interesting points , since , even without time loops , quantum mechanics on the microscopic scale has always given a lot of trouble to those who ask
what happens there ? " .
it seems that we are bracketed by temporal and causal weirdness from both the very small and the very large .
the realm of causality and process , or at any rate due process , appears to be confined to a middle scale , like sentient life itself . but are we so perfectly balanced in this middle that the temporal weirdness at the extremes has no bearing at all on our situation ?
not so long ago we seemed to be bracketed by weirdness of another kind : though pebbles are round and stars and planets are round , the earth we live on is flat .
today we forget how incredibly hard it was for our ancestors to give up their flat earth .
the ancients managed to live precariously on a round earth for almost nine centuries , but by the fourth century ad the earth had become flat again , and flat even for educated people who knew their aristotle , ptolemy and aristarchus .
curiously enough , it was nt flattened by the barbarian invaders , but by learned scholars who were well grounded in alexandrian mechanical technology which they used ( or misused ) to construct ingenious models of the cosmos as a planetarium @xcite . what drove our educated ancestors to such absurdities ?
biblical faith is the usual suspect , but we think animal faith is the real culprit .
animal faith says that up is up and down is down , and never the twain shall meet .
up and down are kantian - darwinian categories , so deeply entrenched in our psyches , our language and our genes that they pervade all thought , as shown by expressions like getting high , feeling let down , being well grounded , etc . etc .
the trouble with a round earth is not that we ca nt imagine its shape but that we ca nt imagine a universe where unsupported things do nt fall , and falling means going down .
the aristotelian and medieval solution to the problem of falling was to have everything fall towards the center of the universe , which was located in the center of the earth .
but after the copernican revolution , the concept of falling once again became quite mysterious .
which way do things fall when they are well away from earth , and why ?
it was newton who solved this mystery by realizing that falling is not something that happens to objects in isolation but is a relationship between two or more objects . with this insight , falling was subsumed into another and very different concept , that of mutual attraction .
past and future are even more ingrained in our genes , psyches , and language than up and down , so quantum superposition and black holes are even greater heresies for animal faith than a round earth wandering around the sun . today it s not falling that is mysterious but _ happening_. does newton s solution offer us any help in making it less so ?
is it conceivable that we could subsume the concept of happening into some more general concept ?
if one is prepared to accept kaku s reading of the current astronomical evidence , this larger concept ca nt have an absolute temporal arrow anymore than newton s concept could have an absolute arrow of up and down .
the past and future have become poles in a relationship . but
a relationship between what and what ? and described how ?
we ll return to this in a moment .
we agreed with kant that _ happening _ is an empirical concept .
we then simplified the meaning of _ empirical _ into being able to keep a diary of observed events .
let s now concentrate on the concept of a _
diary_. to keep a diary , you start out with a stack of white paper and write down your experiences in the order in which they occur . as we noted above , each time you write something down ,
you narrow the set of possibilities for what could end up on the paper , i.e. you take away a subset of _ cases_. each item in your diary corresponds to a certain set of cases taken away , so the act of writing produces a sequence -c1 , -c2 , -c3 , ... of such departing case sets .
this account of writing is incomplete , however , since it only shows cases departing , not cases arriving .
our description of your diary began after you had already acquired a stack of paper . and yet , properly speaking , your first act of diary keeping was to acquire that stack of paper , since it is blank paper that _ supplies _ the set of cases from which the members of c1 , c2 , c3 etc . are taken away when you write .
furthermore , this first act may well be repeated , since you may run out of paper before you are done .
thus a more complete description of your diary writing would go like this : + c1 , -c2 , -c3 , + c4 , -c5 etc .
, where a + c is a set of cases added to those at hand when you get a new sheet of paper , and a -c is a set of cases taken away when paper meets pencil
. now let s address the question : what would it mean for a record of events to be reversed in time ?
what comes to mind is a movie run backward : spilled salt flying from the floor into its shaker , the phoenix rising from its ashes etc .
however , simply running the frames of a movie backwards does not reverse the crucial order of events that creates each frame , namely , light going from object to camera and then leaving its imprint on the blank film .
a backwards movie is only a forward record read backwards ; it s not the record of backwards events .
let s approach our question on the more fundamental level of information and entropy , which is to say , on the level of departing and arriving cases .
once we take this course , its answer is immediate : time reversal turns departing cases into arriving cases and vice versa .
what goes backwards is not the order of movie frames or some other kind of state descriptions " but rather the order of the case sets c1 , c2 , c3 etc . , and this is accompanied by a reversal of their signs .
thus our diary above would go + c5 , -c4 , + c3 , + c2 , -c1 . reversing past and future
turns writing into paper and paper into writing !
time reversal is indeed very simple , but it s also pretty weird .
here s another bit of weirdness .
seen forward in time , you take a piece of paper supplying 100 cases and write something on it that takes away 50 of them .
seen backward , you take a piece of paper supplying 50 cases and take away 100 of them ! what is the meaning of the remaining " minus 50 cases ?
it s certainly nothing familiar .
however , this example shows that negative cases , and thus negative probabilities , ca nt be avoided in describing time - reversal , and if the speculative cosmologists and astrophysicists are right about the circular warping of time , we will somehow have to make our peace with them . in short , the generalized concept of happening we have been looking for , whatever else it may require ,
does require a generalized theory of probability in which we can countenance debts as well as surpluses of cases .
in chapter 1 we mentioned that negative cases play havoc with logic .
let s very briefly see how this happens , since the mathematical illogic " of negative cases is actually a useful tool . a more extended discussion of the following points is given in `` boolean geometry '' , where the `` boolean cube '' is pictured @xcite .
define the logic of a set of cases to be the set of all of its subsets .
consider the logic of a set of three cases .
it has eight members , which can be arranged in a natural way as the vertices of a cube , called the _
boolean cube_. if we stand the boolean cube on the vertex representing the null set 0 , the top vertex becomes the universal set 1 containing all three members , i.e. the principle axis containing 0 and 1 is vertical .
this same construction can be carried out with n cases to give an n - dimensional hypercube , which is also called a boolean cube .
ascending an edge of a boolean cube ( or hypercube ) represents adding a case to the lower vertex , where parallel edges always add the same case .
it follows that going down an edge represents taking away a case .
since writing a diary means successively adding and taking away cases , we can represent the course of writing by an edge path that starts with the vertex representing the initial case set .
suppose now that we stand the boolean cube upside down , putting 1 at the bottom and 0 at the top .
each step in the diary edge path then reverses its sign , so this transformation of the cube represents time - reversal .
it also has a logical meaning , which is negation ; more exactly , a self - reflection of the cube in the vertical direction maps each vertex into its negation . turning the cube upside down is only one of its many geometric symmetries@xcite .
we can classify these according to how far they rotate the 0 - 1 arrow from its normal vertical orientation , the extreme being 180 degrees , as in negation .
the intermediate angles represent partial time reversal .
we can imagine the movable 0 - 1 arrow as a throttle that ranges from full speed ahead in time when it is pointing up to full speed astern when it is pointing down .
if the throttle is exactly halfway between , the diary goes
sideways " in time , so - to - speak ( this can only happen in a hypercube with an even number of dimensions . ) among such sideways transformations are square roots of negation , which we ll come back to in a minute .
it was our aim to come up with a concept of diary that is invariant under time reversal .
the most natural interpretation of its invariant structure is a path on the boolean cube , now understood as a purely geometric structure without an intrinsic up - down direction .
this conception relativizes the logic of the diary , since the operations and , or etc .
depend on the setting of the time throttle .
recall in chapter 1 that we noted the weird logic of interference , where a is possible and b is possible but ( a or b ) is impossible .
relativizing logic explains this puzzling situation : the particle that interferes with itself " is in an eigenstate of an observable for which or has a different meaning than it does in the frame of the classical apparatus . back to the square root of negation .
when we base probability theory on boolean geometry , this transformation gives us imaginary probabilities , just as negation gives us negative probabilities ; link theory turns these into the imaginary and negative amplitudes of quantum mechanics .
let us now go out on a limb and conjecture that the direction sideways in time " is a direction in space .
this could mean that the square root of negation is also behind the i of minkowski space ( see section 3.8 for more on this topic . )
if so , it could be crucial for unifying our theories of space and matter .
let us go even further out on that limb .
first , we ll call on pauli for a bit of support . towards the end of his life pauli
had a mystical vision of the quantum i as the key to unifying physics with psychology @xcite .
it turns out that complex quantum mechanics is mathematically equivalent to a real quantum mechanics ( see section 3.6 ) in which every object contains a particular two - state object , which has been called the janus particle @xcite , whose state is unmeasurable . te s conjecture
is that choosing between the two states of the janus particle corresponds to choosing between the heisenberg and schrodinger representations .
the fact that the absolute phase of amplitude is unobservable then means that quantum matter in itself ( i.e. unmeasured ) is not only symmetrical with respect to time direction but also with respect to _ subject _ and _ object_. the differentiation of the world into particular pieces of matter , particular minds , and particular space warps can thus be understood as three different manifestations of the breaking of the primal symmetry represented by i. as we said before , logic is really out of bounds for this paper .
the last few paragraphs must have seemed pretty cryptic , but we wanted to give a bit of the flavor of things to come .
we ll return to such speculations in the last section of chapter 3 , but until then our business will be with matters closer to home .
time reversal does more than just change the sign of the events in our diary writing
it also reverses their order .
now it s reasonable to formalize the process of registering and recording information as an i - o system . as we saw , the laws of the underlying markov process are symmetrical in time , which means that we could also represent it by an i - o system going backwards in time .
but what sort of system would represent sideways " time ?
here we must turn from i - o systems to link systems .
there turns out to be a whole spectrum of these ranging from full speed ahead to full speed astern , and it is those exactly half - way between that give quantum states ( see section 3.6 ) . note that the state throttle " here is very different from the phase throttle " that rotates the boolean cube , and indeed it need not even involve negative probabilities .
we asked above whether _ happening _ might resemble _ falling _ in being a special case of a more general situation analogous to motion in a gravitational field . as we ll see , the mathematics of the state throttle supports this analogy .
we should speak of rising , however , rather than falling , in the sense that events _ arise _ from the ground " of past conditions . with the state throttle in full reverse ,
events arise backward in time from the ground of future conditions . in the more general case ,
events are repelled " by both past and future , like helium balloons between two massive bodies . at half throttle , when the balloons are exactly balanced in between , we get quantum mechanics .
we ve come a long way from our starting place , which was to respond to hume s challenge to find causality in the bare order of correlation " . in effect
, hume s problem effectively disappeared , giving way to others more puzzling and serious .
the problem with hume s problem is that it assumes we are simply given a world where things happen .
this is indeed true in everyday life , but when we start to think about what happens in time loops or atomic nuclei , we find that it s the concept of happening itself that is the problem .
causality , time direction , and history all seem to be bound up together in a way that makes it hard to say which is the more fundamental .
this tightly knit complex of ideas is a familiar one , and we have made some progress in analyzing it .
but we have also discovered that it does nt transfer very well to either the domain of the very large or of the very small . to make sense of modern physics
seems to require that we give up not only causality and history but even logic .
the next chapter will lay out in a more systematic way the kind of mathematical theory that must take their place .
the main goal of this paper is to show that basic quantum mechanics can best be understood as a branch of the theory of probability . for this to happen
, probability theory must be expanded in two ways : first and foremost , probabilities must be allowed to go negative .
second , and this is more a matter of technique than a change in our assumptions , the composition of transition matrices must be subsumed into a more general method of composition based on the linking of variables .
we ll see that _ in this expanded theory _ , probability amplitude is not some new kind of fluid " or physical field , but simply probability itself . by regarding it as such , we turn some of the most striking basic discoveries of modern physics into pure mathematics , thereby freeing them from their purely physical context to become tools available to every branch of science .
just how much more of physics as we know it will follow in this course remains to be seen .
_ event , sample space , case , variable , range , joint variable , logic , atom , probability measure , probability distribution , random variable , summation theorem , marginal , stochastic process , independence , conditional probability , conditional independence , separability , markov property , markov process .
_ * event*. following feller@xcite we ll take _ event _ to be our basic undefined concept .
we ll use the word in pretty much the ordinary sense . like most english nouns
, it can be either definite or indefinite .
* sample space , case * : a _ sample space _ is a set of mutually exclusive possibilities called _
cases_. we say that an event e _ belongs _ to a sample space l if for every case of l , it makes sense to ask whether e is true or false .
another word for event is _
proposition_. examples : the event _ heads _ belongs to the sample space whose cases are the possibilities heads and tails . the event _ seven _ belongs to the sample space whose cases are the 36 possible throws of a pair of dice .
* variable * : the term will be used in the scientist s sense , meaning something that can vary , rather than in the logician s sense of a place - holder for a term .
a variable x is said to belong to a sample space l if for any value k of x , the event x = k belongs to l. for example , let l be the set of all cases for ten coin tosses , and let x be the number of heads ; then x=1 , x=2 , x=3 ... x=10 , are events in l and x is a variable in l. more abstractly , a variable is simply a function on l ; when we regard x in this way , the proper notation is x(l ) .
the set of all possible values of a variable is called its range . *
joint variable x @xmath4 y * : though variables are often numerical , we ll place no restrictions on what kind of things their values can be . for instance , a variable can range over all of the joint values of several other variables .
the notation x @xmath4 y will refer to a variable that ranges over all joint values of x and y. there is an ambiguity in the term joint value " , which is whether or not it involves the order of things joined .
is x @xmath4 y the same variable as y @xmath4 x ? and is ( x @xmath4 y ) @xmath4 z the same variable as x @xmath4 ( y @xmath4 z )
? actually , there are two concepts of joining , which in everyday language we would render as x and y " and x , then y " .
the first and simplest , which is unstructured joining , is commutative and associative , so we can write x @xmath4 y @xmath4 z @xmath4 ... etc ... without parentheses and with the variables in any order .
the second , which is the one we ll use here , takes joint values to be _ ordered _ n - tuples ; it has the decisive advantage over the first that variables of the same type are in natural 1 - 1 correspondence . there
s actually a third concept called _ list structured _ joining that is very useful in constructing complicated link diagrams , but it belongs to a more advanced and specialized chapter of our theory . * logic , 1 and 0 , generating * : events can be combined with and , or and not , and a set of events closed under these operations will be called a _ logic _ ( this use of the word ` logic ' comes from von neumann ) . we ll abbreviate and by `` & '' and not by `` @xmath5 '' .
a logic is of course a boolean algebra .
null _ element of a logic , abbreviated 0 , is the element a & @xmath5a , which is the same for any a. the _ universal _ element , abbreviated 1 , is the element @xmath50 = ( a or @xmath5a ) .
given any set k of events , the set of all events that result from applying and , or and not to the members of k is a logic , and it will be called the logic _ generated _ by k. given a variable x , the logic generated by all events of the form x = k , where k is a value of x , is called the logic of x , and the members of that logic are called events in x. more generally , the logic of x , y , z ... is the logic generated by all events of the form x = i , y = j , z = k etc . , and
the members of that logic are called events _ in _ x , y , z etc .. * atom*. the atoms of a logic are the cases of that logic .
to put it another way , an _ atom _ of a logic is defined as an event that can not be further refined by any other event of the logic , i.e. a is an atom if , for any e in the logic , either a&e = a or a&e = 0 .
atoms are mutually exclusive possibilities , and the set of all atoms is exhaustive within the logic . in the logic of the variable x ,
the atoms are the events of the form x = k for all k , while in the logic of x , y , z ... the atoms are the events of the form ( x = i)&(y = j)&(z = k ) ... for all i , j , k .... if the number of atoms is finite , then the subsets of atoms uniquely correspond to the events in the logic .
the cases of the sample space are the atoms of its logic , so atom and case are basically synonymous terms . however , we ll generally reserve the word ` case ' for atoms that we count , such as the atoms of the sample space .
* probability*. underlying all our more sophisticated ideas about probability is pascal s classic definition of the probability p(e ) of an event e as the number of cases favorable to e divided by the total number of cases .
if the cases of a sample space can be regarded as equally probable , then this definition defines a so - called _ probability measure _ on the events of that sample space , satisfying the following three axioms : for now we are dealing with _ classical _ probabilities , which are never negative . for these we can always imagine an underlying sample space with equiprobable cases that define p(e ) according to pascal s definition . in dealing with negative probabilities ,
the notion of an underlying case set must be generalized so that it can contain both positive and negative cases . (
the notion of negative set membership has been carefully investigated by mathematicians , and blizard has shown that the zermello - fraenkel axioms can be generalized to allow for it @xcite . ) in this more general form , the case set has a very important role to play in link theory , since it enables us to distinguish among several quite different conceptions of independence ; the details here will be found in future papers .
* probability distribution on x * : a function defined on the range of x which for each k in that range has the value p(x = k ) .
the standard notation for this function is p(x ) , which can be very confusing , since it conflates the scientist s and logician s two very different notions of variable .
however , it s such a well - established tradition that we ll have to live with it ; when you get confused , just remember to replace x by something that looks like a sentence .
* stochastic , or random variable * : a variable on which is given a probability distribution .
the qualifiers ` stochastic ' and ` random ' are only needed when a variable must be distinguished from variables of other kinds , like free variables . since our results have nothing to do with randomness
, stochastic is the preferred qualifier .
* summation theorem and marginals * : the summation theorem says that given a joint probability distribution on variables x and y , we can find the probability distribution on x alone by summing over all values of y ; i.e .. p(x = i ) = @xmath7p((x = i)&(y = j ) ) , or , in our abbreviated notation , p(x ) = @xmath8p(x , y ) .
this is a theorem we ll use repeatedly .
a corollary is that a joint probability distribution uniquely defines the joint probability distribution on any subset of its variables ; such partial joint probability distributions are called _ marginals_. * stochastic process*. define a _ stochastic process _ , or _
process _ for short , as a set of variables on which is given a joint probability distribution .
in the usual definition it is also assumed that these _ stochastic variables _ involve events which can be ordered in time .
this wo nt do for our present purposes , though , since , we need a new concept of process general enough to apply to time loops . we can start out by thinking of process variables as having at least approximate locations in the space - time of current physics .
however , since we aim to eventually construct space - time out of probabilities , this is only a temporary expedient .
* independence*. events e , f , g , ... are called independent if p(e&f&g ... ) = p(e)p(f)p(g ) .... note that for a set of events to be independent , it s not enough that every pair of events in that set be independent ; independence is a joint condition of all the events .
two variables are called independent if events in the first are independent of events in the second .
a set of variables are called independent if events in different variables are independent .
this definition pretty much captures our intuitive concept of independence as lack of correlation , although there is one thing about it that is not so intuitive : suppose that either p(e)=1 or p(e)=0 .
then for any f , p(e&f ) is either p(f ) or 0 ; in either case it is p(e)p(f ) .
such a definite " e is independent of all events , including itself !
this oddity will turn out to be important for the concept of probability space ( see section 3.8 ) . given a finite sample space of equiprobable cases
, there is another way to define the independence of e and f , which is that the sample space can be arranged as a rectangular set in which the cases of e are a vertical stripe and the cases of f are a horizontal stripe . for ordinary probability theory
in which cases always count positively , the two definitions are equivalent .
however , with both negative and positive cases , the two diverge , since p(e ) can be 0 even though its cases set is not
rectangular " with respect to that of f. it turns out that there is even a third definition of independence which applies to imaginary and complex probabilities ( see section 3.6 ) when these are defined in terms of real probabilities .
the above product rule is the definition we ll adopt here , however . since a conditional probability results from dividing by p(f ) ,
it is only well - defined if p(f ) is not 0 .
negative probabilities are constantly producing 0 s in weird places , so it s best to avoid conditional probabilities if there s an alternative ; fortunately , there usually is . *
condition * : an event c that is assumed to be true , thereby reducing the sample spaces to the set of those cases favorable to it .
axioms for a process are often given as conditions on a free " process whose variables are independent .
a very important kind of condition for our present work is the _ link _ , which is an event of the form x = y .
the probability of e in the reduced sample space conditioned by c is p(e @xmath1 c ) in the unconditioned sample space . *
conditional independence * : events e and f are called _ conditionally independent _ given c if they are independent in the sample space conditioned by c. the usual algebraic formulation of this relationship is p(e&f @xmath1 c ) = p(e @xmath1 c)p(f @xmath1 c ) . more generally , a set of events e1 , e2 , e3 ... is conditionally independent given c if p(e1&e2&e3 ...
@xmath1 c ) = p(e1@xmath1c)p(e2@xmath1c)p(e3@xmath1c ) ... as we noted in chapter 2 , these equations becomes meaningless when the probability of the condition is 0 .
the following definition is equivalent to conditional independence when p(b ) is not 0 , but also makes sense when it is 0 : * separability*. events a and c are called separable by event b if p(a&b&c)p(b ) = p(a&b)p(b&c ) .
this definition comes from noting that it can be obtained by multiplying the conditional independence condition , p(a & c @xmath1 b)= p(a @xmath1 b)p(c @xmath1 b ) , on both sides by p(b)@xmath9 .
the equation defining the separability of n events by b is derived by multiplying both sides of the appropriate equation for conditional independence by p(b)@xmath10 .
we say that variable y separates variables x and z if p(x , y , z)p(y ) = p(x , y)p(y , z ) . which is an elliptical way of saying that if for any values i , j , and k of variables x , y and z we have p((x = i)&(y = j)&(z = k))p(y = j ) = p((x = i)&(y = j))p((y = j)&(z = k ) ) .
much of our work will be with markov processes .
feller@xcite begins his chapter on markov processes with the concept of markov chain , which he defines as a series of trials in which the probabilities at each trial depend on the outcome of the previous trial , the rule of dependence being the same for all trials .
he then goes on to say that a markov chain is characterized by a matrix of conditional probabilities , called the _ transition matrix _ , together with an initial state vector " giving the probability distribution at the initial trial .
there are problems with this definition .
part of what motivates it is the need to separate out the _ law _ of the process , given by the transition matrix , from the _ boundary condition _ on the process given by the initial state vector .
and yet there is no reason why the initial vector should not contain 0 s , for which conditional probabilities are meaningless .
thus the markovian law of the process , which makes perfect intuitive sense whatever the initial state , can not properly be described by a matrix of conditional probabilities .
the so - called transition matrix must be defined in some other way .
the remedy for this and related confusions is one we have discussed at length in chapters 1 and 2 ; this remedy is to clearly distinguish between the process itself and its representation as a _
system_. feller began by describing a process and then switched over to describing it as a system , without noticing the change .
section 3.6 will be devoted to markov systems , so here we ll define only the process .
first a more general concept , which we ve already met in chapter 2 : * markov property * : three events a , b and c are said to have the markov property if p(c @xmath1 a&b ) = p(c @xmath1 b ) . multiplying both sides by p(a&b)p(b ) , we see that this is simply separation , i.e. it says that b separates a from c * markov transition matrix * : abstractly , a matrix in which the sum of every column is 1 ( just to confuse matters , such matrices are sometimes called _
stochastic matrices_. ) more concretely , the matrix of conditional probabilities from variable x(t ) to variable x(t+1 ) in a markov process .
* primary variables * : in section 3.1 , a process was defined as a set of random variables on which we are given a joint probability distribution ; these variables will be called the _ primary _ variables of the process .
we will assume that the primary variables are given in a certain linear order , which may or may not be related to the probability distribution . * secondary variables * : a secondary variable is a joint variable in a set of primary variables , called its members .
we will assume that the members always occur in their natural order , so there is one and only one secondary variable for every subset of primary variables .
we ll speak of secondary variables as also belonging to the process .
both primary and secondary variables will be denoted by small letters .
* global variable * : the secondary variable whose membership includes all primary variables . the process itself will be denoted by capitalizing the global variable , e.g. , w is the process whose global variable is w. * independent part , component * : a sub - process u of w is called an independent part of w , or a _ component _ of w , if the primary variables of u are independent of all other primary variables of w. the primary variables of u are also called its _
members_. one often hears that the whole is more than the sum of its parts .
what is generally meant is that the whole is more than the mere heap of its parts , which is true unless the whole happens to be a heap of parts .
a process is indeed the heap of its prime parts , which are in fact components , in the sense that the description of each does not require that we take into account the individual peculiarities of the others .
however , we saw that one of the surprises of negative and imaginary probability theory is that there are several quite different definitions of independence , and thus the members of a heap share , as it were , the nationality " of their heap , which involves the particular kind of independence that its members have from each other .
step 1 : apply the condition x = y to w to obtain a new _ conditioned _ process w with the same variables as w but with a new joint probability distribution w in which the probability p(e ) of an event e is p(e@xmath1x = y ) in w. we can think of a link as an operator l(w , x , y ) on w , x and y that produces a new process w , in which the probability of an event e is its probability in w conditioned by the event x = y , and in which the variable y is dropped .
we ll write w@xmath1(x = y ) for this operation , i.e. w = w@xmath1(x = y ) ; more generally , w@xmath1lmn will mean applying links l , m and n to process w. there is nothing in the definition of link that says x and y must be different variables ; w@xmath1(x = x ) is a perfectly legitimate operator .
what , then , does it do ?
it does nothing to the probabilities , since the condition x = x is already in effect .
it does , however , drop the second variable , namely x. to link a variable to itself simply means to drop it from w. x and y can be either primary or secondary variables . if they are secondary , however , they must contain corresponding members that are also implicitly linked .
thus if x = x1&x2 then y must be of the form y1&y2 , and the condition x = y implies also that x1 = y1 and x2 = y2 .
this means that when y is dropped , we must also drop all of its primary members , and with them all other secondary variables containing any of these members , which are no longer needed , since they are duplicated by secondary variables with corresponding members in x. dropping a variable is a bit more complicated than the name suggests .
the converse of linking could be called de - linking , but the term does nt really fit , since it implies that there are two things to be de - linked , whereas in fact there is only one . to restore the original w
, we must first turn x in two equal variables , and then remove the condition that equates the two .
we ll call this operation : * density matrix * : the joint probabilities on x and y arranged in a matrix with x as the horizontal index and y as the vertical . following _ one _ quantum tradition , we ll use the terms ` state ' and ` density matrix ' interchangeably .
* state of a variable x * : the density matrix of the link x = x .
notice that the off - diagonal elements of this matrix are all 0 , since it s impossible for its two indices to differ , and impossibility has probability 0 !
a more rigorous definition of the state of a variable is the diagonal matrix whose indices range over the range of that variable , and whose diagonal is the probability distribution on that variable .
for instance , if x ranges from 1 to 3 , then its state is the 3 by 3 matrix m@xmath0 such that m@xmath11 is the probability that x=1 , m@xmath12 the probability that x=2 , m@xmath13 the probability that x=3 , and all other entries are 0 .
we ll find that the state of x encompasses all of the classical meanings of the word ` state ' , including markov states and deterministic states like those of a computer .
a markov state is usually represented by a vector ; this traditional vector can now be defined as the diagonal of the state matrix of the time - dependent random variable of the process ( we ll see the details of this later ) . * proposition * : literally , a synonym for event .
( we are borrowing the term from von neumann , who applied it to quantum variables . )
there is a difference in connotation , though .
events are often described by noun phrases , whereas one always represents propositions by sentences .
we can think of propositions as sentences in the predicate calculus whose free variables are assigned to process variables . * predicate * : a propositional function " whose free variables are unassigned . a predicate becomes a proposition when
all of its variables are either assigned or quantified .
a link , for instance , is a proposition resulting from the assignment of the equality predicate to a pair of variables .
* projection * : in linear algebra , a projection p is defined as a linear operator satisfying the idempotent law pp = p. it can be shown that we can always find a basis in which such a p is represented by a diagonal bit matrix .
for now we ll confine ourselves to diagonal projections , i.e. a _ projection _ will be defined as a matrix in which all off - diagonal elements are 0 and the diagonal elements are either 0 or 1 .
von neumann showed that propositions about the measurement of a quantum observable x can be represented by projections that commute with the eigenvalues of that observable .
this led him to an elegant generalization of the born probability rule : if p is a projection representing a certain proposition , and d is a density matrix representing a quantum state , then , given d , the born probability of that proposition is tr(pd ) . * representing predicates by projections * : if p(x ) is a proposition about x , define the _ characteristic function _ of p(x ) to be the bit - valued function b(x ) which is 1 or 0 corresponding to whether p(x ) is true or false for x. the diagonal matrix p having b(x ) as its diagonal is the projection that we ll use to represent the proposition p(x ) .
we ll normally use the same letter p for both the predicate and the projection , and when the predicate is assigned , for the resulting proposition ; it s usually clear from the context which of these things we mean .
if we multiply the projection p by the state matrix d of x , we clearly knock out all cases for which the predicate p is false , so the sum of the remaining entries in d is the probability of proposition p , i.e. p(p ) = tr(pd ) .
this makes born s probability rule start to look a bit less mysterious .
however , what really demystifies born s rule is that it holds not just for the state of a variable but for any link state .
more exactly : proof : first note that the state s of x in w@xmath1x = y is s with all off - diagonal elements set to 0 , normalized by dividing by p(x = y ) .
since p is diagonal , ps is the result of multiplying the rows of s by corresponding diagonal elements of p. thus the diagonal elements of s are simply those of s divided by p(x = y ) = tr(s ) .
since we just saw that p(p ) = tr(ps ) , we conclude that p(p ) = tr(ps ) .
our result differs from von neumann s by the factor tr(s ) .
had we defined a link state as s / tr(s ) , thus making tr(s ) always 1 , we would have gotten von neumann s result exactly .
this was our original definition of state , and it works fines for classical and quantum states .
however , we would like our definition of state to still make sense even as we push our theory into new domains where , because of the possible cancelation of plus and minus cases , the trace of s can be 0 . for this reason , we believe that our present unnormalized state is preferable in the long run .
it follows directly from the definition of independence that the density matrix of a pure state has the form @xmath1v@xmath3u@xmath1 in dirac notation .
if u is white , the state will be called _
causal_. on the other hand , if u = v , the state is quantum . causal and quantum overlap if u and v are both white . , but in no other case .
proof : to say that w is white means that for any i , p(w = i ) = 1/n .
let s be the state of link w@xmath1x = w . since s is pure , p((x = i)&(w = i ) ) = p(x = i)p(w = i ) = ( 1/n)p(x = i ) . but
p((x = i)&(w = i ) ) is s@xmath14 , the ith diagonal term of s. thus tr(s )
= @xmath15 s@xmath14 = ( 1/n)@xmath15 p(x = i ) = 1/n . *
white connection theorem * : roughly : pure causal links do nt affect the component on the output " side .
more exactly , if x is in a component a that is independent of y , and y is white , then the link w@xmath1x = y does nt change the joint probability distribution on a. proof : let z be the join of the variables of a other than x , i.e. a = x&z . then p(x , z , y ) = p(x , z)p(y ) = ( 1/n)p(x , z ) in w. by the above trace theorem , p(x = y ) = 1/n .
thus p(x , z , y@xmath1x = y ) , which , ignoring y , is the joint probability distribution on x and z in w@xmath1x = y , is equal to ( 1/n)p(x , z)/p(x = y ) = p(x , z ) , which is the joint probability distribution on x and z in w. qed .
this theorem is the key to mapping the engineer s input - output systems into link theory .
remember , we asked the question : when an input is wired to an output , how do you know which is which ?
the answer is : cut the wire and see which loose end is unaffected by the cut .
this answer can be amplified a bit .
if the wire is from box a to box b which is otherwise unconnected to a , then cutting it not only leaves the output of a unaffected , but all of a unaffected .
this is essentially the content of the white connection theorem , if we regard the causal link as a connecting wire .
the word causal " means what it says !
* mixed states * : needless to say , pure states play a very important role in our theory . in quantum mechanics ,
impure states are called _ mixed _ states ; this is because they can always be written as linear combinations of pure states .
this turns out to be true quite generally of link states , so we ll borrow the quantum term _
mixed state _ for any state that is not pure .
let us point out , though , that the decomposition into pure states of a mixed classical state , such as one might find in a computer program , is generally quite artificial , since the pure components will seldom be classical and will often involve negative probabilities .
_ disconnection , fundamental theorem of disconnection , link systems , proper systems , predicates as components , relational composition , abstract links and their states , contraction , product theorem , tensors , public variables , systems , transition matrices , causal boxes and systems .
_ * fundamental theorem of disconnection * : let w be the join of three variables x , y and z , i.e. w = x&y&z , , where p(y ) is not 0 for any y. then x can be disconnected from z at y if and only if y separates x and z. ( recall that separation means p(x , y , z)p(y ) = p(x , y)p(y , z ) . ) suppose there exists a disconnection of w that produces a new process w. let y be the new variable of w , i.e. w = w@xmath1(y = y ) . to avoid confusion ,
let s use p for probabilities in w and p for probabilities in w. then in w we have p(x , y , y,z ) = p(x , y)p(y,z ) .
this implies that , in w , x is independent of z for any values of y and y. in particular , x is independent of z if y = k and y=k .
since this is true for any k , it is consistent with the linking condition y = y , which is what turns w into w. this shows that y separates x and z in w. notice that going from disconnection to separation does nt require the assumption that p(y ) is never 0 .
going backwards is a bit harder .
suppose y separates x and z in w. we need to construct a new w containing x , y and z plus a new variable y such that x&y is independent of y&z in w , and such that w = w@xmath1(y = y ) .
separability in w means that p(x , y , z)p(y ) = p(x , y)p(y , z ) .
let w be a process consisting of two independent parts w1 and w2 , where w1 contains x and y , and w2 contains y and z. let p(x , y ) = p(x , y ) .
we construct p(y,z ) as p(z@xmath1 y)/n where n is the number of cases of z for which p(z@xmath1 y ) is non - null .
then p(z@xmath1 y ) = p(z&y)/p(y ) where p(y ) = @xmath16 p(z@xmath1 y ) .
note that this is where the requirement p(y ) @xmath17 0 is needed.w * lemma 1 * : p(x , y , z ) = p(x , y)p(z@xmath1y ) .
proof : because p(y ) is never 0 , we can freely speak of probabilities conditioned by y. thus we can write p(x , y , z ) as p(x , z@xmath1y)p(y ) . we can also write separability as conditional independence , i.e. p(x , z@xmath1y ) = p(x@xmath1y)p(z@xmath1y ) . combining the two we get : p(x , y , z ) = p(x , y)p(y)p(z , y ) = p(x , y)p(z@xmath1y ) .
qed * lemma 2*. the joint probability distribution on x , y , y and z in w is p(x , y)p(z@xmath1y)/n .
proof : by the independence of w1 and w2 , we have p(x , y , y,z ) = p(x , y)p(y,z ) . since p(x , y )
was defined as p(x , y ) this becomes p(x , y)p(y,z ) .
= p(x , y)p(z@xmath1y)/n , where n is the number of rows in the transition matrix p(y@xmath1z ) .
qed now let s link y and y. this involves two steps : first we identify the variables y and y , and then we normalize the resulting distribution by dividing by p(x = y ) .
we see by lemma 2 that the first step turns the joint distribution on w into p(x , y)p(z@xmath1y)/n .
now p(x = y ) is the trace of the state matrix of the link , which we can easily see is 1/n ( the steps showing this are spelled out in the proof of the causal link chain theorem in the next section . ) thus dividing by p(y = y ) yields p(x , y)p(z@xmath1y ) , which by lemma 1 is p(x , y , z ) .
thus p(x , y , z ) = p(x , y , z ) .
the above disconnection theorem is what tells us how to mark a process for dissection , so - to - speak .
first , we find a variable x , primary or secondary , which separates the process into two conditionally independent parts .
we then disconnect the process at x into two unconditionally independent parts , and then repeat for these two parts etc .
until we can repeat no more , at which point we have arrived at a prime factorization .
these prime factors , as such , constitute a mere heap of independent parts .
the disconnection procedure , however , leaves a record of past connections , a kind of wiring diagram , that can be used to reconstitute the original process .
the question arises whether such a prime factorization is unique .
there are actually two questions here : first , whether the factors themselves are unique as processes , to which the answer is definitely no ( except in the trivial case where the process has no factors except itself ) and second , whether the prime boundaries " are unique , or to put it another way , whether every factoring procedure ends up with the same sets of primary variables in its factors .
the second question has no simple yes or no answer .
often it is yes , but there are important cases ( epr , for instance ) where a seemingly arbitrary choice in making the first cut will determine whether or not any further cuts are even possible . proof : divide the linked variables into two sets x and y , where x contains the first variable of each link , and y contains the second .
let x be the join of variables in x , y the join of variables in y. then the link x = y clearly encompasses the links of all the members of x and y , whose order is immaterial as long as they are properly matched .
the linking order can indeed make a difference if the linked pairs have variables in common .
if you link y to z and then x to y , all three of x , y and z will be tied together and called x. however , if you first link x to y , y disappears , so you when then you try to link y to z , there is no process variable to link .
this sort of thing can not happen in a proper system , of course . but suppose you need to link all three of x , y and z together ?
indeed , this is exactly what an engineer does all the time when he branches " an output to go to several inputs .
must the engineer then work with an improper system ?
the answer is no .
there are in fact several ways to branch properly .
one rather interesting way is to extend link theory by treating predicates as abstract " components which , when linked to random variables , become conditions on the process . thus to connect x , y and z
, we introduce the abstract component x=y=z plus the three proper links .
the resulting more general theory merges stochastic composition with the predicate calculus , and may be of mathematical interest .
however , much of what it accomplishes can be done in other ways ; for instance , we can regard x=y=z as a stochastic process with three equal white variables , and the links x = x , y = y and z = z also connect x , y and z. at any rate , abstract components wo nt concern us here . with one exception ,
that is ; links themselves should definitely be regarded as abstract components .
but they are the only ones we need for now . not having to worry about
the order and grouping of links is such a plus that we ll assume that all our systems are proper . under this assumption
, there is a very readable and natural diagrammatic notation for link systems , which uses icons to represent component processes , and arrows between them going from x s to y s to represent links . indeed ,
if we draw and label the icons right , we can often avoid variable names altogether , since the variables will be identified by their places .
unfortunately , the practicalities of word processing , e - mail , etc . make it expedient to have a linear backup notation .
boxes are written thus : a[y1,y2 , .. ; x1 , x2 , .. ] , where the variables before the semicolon are the ingoing arrows , those after , the outgoing .
a link is shown by an equation , so a(y ; x ) ( x = x ) b(x ; z ) links x in a to x in b. if there is no need to reconstruct the particular lost variables in w , we can omit the equation and simply use the same letter in a and b , e.g. a(y ; x ) b(x ; z ) .
needless to say , diagrams are much easier to understand , and it s often best to turn this linear notation into actual diagrams rather than trying to decipher it as it is
. there are more powerful diagrammatic notations than this for dealing with complex structures , but they are nt needed for the simple examples we ll be studying here .
it s worth introducing one new concept from this extended notation , though : * box * : an icon that represents a component of a system , but that may also be a container , implicit or explicit , for a more detailed system . from the outside
a box is to be regarded simply as a process , but it s permissible to look inside it for smaller components ( think of the transistors inside an integrated circuit or the electrons inside an atom . ) in the next two sections we ll show all component icons as boxes , and sometimes refer to them as such
. * transformation * : a two - variable sub - process t of a link system in which both variables are linked to others .
if the variables of t are x and y , in that order , and if x = x is the link to x and y = y is the link to y , then the link state ( density matrix ) of x = x is called the state _ before _
t , while the link state of y = y is called the state _ after _
t. t is called _ causal _ if x is white .
* transformation product theorem*. roughly , to combine two successive transformations , multiply their matrices .
more exactly , if t and u are transformations with variables x,y and y,z , then linking y and y and then ignoring y produces the transformation nut , where n is a normalizer . * corollary * : equivalent cut theorem .
suppose that cut c1 in w produces the process w1 with variables x and y. define w2 as the process that results from multiplying x by diagonal matrix d and y by d@xmath2 .
then the cut w,w2 is equivalent to the cut w,w1 . _
the rule of law , boundary conditions , markov chains , generators , time - reversed markov chains , link chains , the link dynamical rule , differential markov chains , inverse markov chains , double boundary conditions , chains , sharp states , schrodinger s equation , complex amplitudes , the phase particle , the janus particle .
_ the ancient and medieval worlds observed nature as carefully as we do , hoping as we do to discover what remains fixed in the midst of change . but science , as we know it , began when our focus shifted from changeless _
things _ to the changeless laws of _ change itself_. the goal of natural philosophy " then became to find those universal
laws of causality " that determine the future from the present .
the amazing success of newtonian mechanics convinced many people that this goal had actually been attained , and so went the scientific consensus for 200 years . but then came quantum mechanics .
what most shocked people about quantum mechanics was its seeming break with newtonian determinism .
chance , seemingly exiled to the limbo of ancient superstition , had , like napoleon returning from elba , once again forced itself upon the civilized world . could chaos be far behind ?
the ideal goal of scientific theory had to retrench from predicting certainties to predicting probabilities .
our ultimate
law of causality " was an illusion ; the best we could hope for was to find the ultimate rule giving the best _ probability distribution _ on the state variable x(t+1 ) of the world at time t+1 as a function of the state variable x(t ) of the world at time t. which brings us to _ markov chains_. a markov chain is a markov process whose transition matrix is the same at all times t ( see section 3.1 for the definition of a markov transition matrix . )
call this unchanging matrix g the generator of the process . given g plus the state at time 0
, we can calculate p(x(1 ) ) , p(x(2 ) ) .. etc .
, and also p(x1,x2 ) and p(x2,x3 ) .. etc .
it follows from the markov property that we can calculate the whole joint distribution ( this is a well - known theorem . ) a generator , in contrast to an initial state , is the _ law _ of a process .
it s important to realize that not every markov process has such a law .
given the markov property alone , the matrix t of conditional probabilities between successive states can be an arbitrary function of time , in which case the distinction between law and initial condition , between essence " and accident " , is artificial at best . what happens to the generator if we reverse time ?
except in very special cases , it does not exist .
the transition matrices of a time - reversed markov chain are almost always time - dependent .
this makes the chain property a good indicator of time direction
score one more point against hume !
it s far from being a definitive indicator , though , since this reversed time - dependence is always lawful , and the reversed chain can always be embedded in a larger process with the chain property .
we have not yet mentioned systems ; the concept of generator applies only to the process itself .
however , it follows quickly from our results in section 3.5 that we can always represent a markov chain by a link system whose components , except for the first one , look and behave just like g. more exactly : * causal link chain theorem * : given a markov chain c with generator g and initial vector v , there is a link system whose first component is v and whose successive components all have the matrix ( 1/n)g , where n is the * dimension * of v etc .
, i.e. the number of values in the range of the state variables .
let x be the initial chain variable , and x,y be the variables ( indices ) of g. then we must show that the joint distribution on x , y that results from assigning x , the conditioning index in g , to x , is the same as that resulting from linking the joint distribution in x , y having the joint distribution matrix ( 1/n)g to the independent variable x with the link ( x = x ) .
recall that the sum of every column of a transition matrix is 1 , so ( 1/n)g is a joint distribution matrix in which the variable x is white .
thus , by the white connection theorem , linking x and x does not change the distribution on x , and so multiplies the ith column of ( 1/n)g by p(x)/p(x = y ) = np(x ) .
but this is just what you get when you multiply the initial vector by the transition matrix g. * causal transformation * : a transformation t in which the probability distribution on the row index is white , i.e. all columns have the same sum .
the transformations in a causal link chain are of course causal .
a transition matrix as ordinarily defined is a causal transformation for which all elements are non - negative .
we have now mapped markov chains onto link chains in such a way that their generators turn into causal transformations with the same matrices , except for a normalizing factor n. it was important to keep track of such _
normalizers _ in proving our basic theorems , but in the future it will make life easier to ignore them until it comes time to compute actual probabilities .
this is almost always possible because , if we know any multiple of a probability distribution on x , we can divide by its sum to get the probabilities themselves .
when it comes to dealing with negative probabilities , there may be no sensible normalizer at all .
for instance , it can happen that even though all the case counts we need are well defined , the total case count is 0 .
the passage of time in a markov chain can be represented by a time - dependent transition matrix t(t ) , where t(t+1 ) = t(t)g . by the product theorem in section 3.5 , this same transition law " ( modulo a normalizer )
is true in a causal link chain .
the two concepts differ radically , however , when it comes to the meaning of the word ` state ' . in a markov chain ,
the states are what we have called the state vectors of the variables , which directly represent their probability distributions .
these state vectors are indeed still well - defined in link theory , and have the same meaning .
however , the word ` state ' itself has been given a very new meaning , and now refers to the density matrix of a link . since the state vector is the diagonal of the state matrix , probabilities are gotten by the rule p(p ) = tr(ps ) , as we saw in the last section . in markov chains , the law governing change of state , or as we ll call it , the _ dynamical rule _ , is v = t(v ) , where t is the time - dependent transition matrix , v the initial state , and v the final state .
things are very different when we come to link states .
* link dynamical rule*. let t be a causal transformation .
if the matrix of t has an inverse , then s = tst@xmath2 , where s is the state before t , s the state after . if t does not have an inverse ,
then the most that can be said is that ts = st proof .
let x and y be the variables of t , where the links to t are x = x and y = y for external variables x and y. now consider the join of the components other than t , and abstract from this the sub - process with variables x and y. cut the link x = x. by the product theorem , the matrix s in the unlinked variables x and x is ut . similarly , the matrix s in the unlinked variables y and y is tu .
we now have s = ut and s = tu .
thus st = tut = ts .
if t has an invertible matrix , then we can multiply on the right by t@xmath2 giving the stronger rule s = tst@xmath2 .
the classical dynamical rule is the offspring of newtonian determinism , and is based on a certain concept of explanation , which equates our understanding of a phenomenon with our ability to reliably predict it . with the return of chance , reliability becomes a matter of degree .
the best laws are now those that give us the best estimates of probability .
nevertheless , the aim of classical explanation remains the same , namely reducing information about the future to information about the past .
thus the form of an explanatory law is that of a function that yields a probability distribution on the future as a function of a probability distribution on the present .
notice that all this changes with the link dynamical rule .
s is only a function of s if t is invertible .
otherwise , the dynamical rule only specifies a certain relationship between future and past , in which the two enter symmetrically ( note that if t is invertible , we can also write s= tst@xmath2 . ) with this shift come new questions about the very meaning of dynamical explanation .
the traditional scientist asks _ how _ the past dominates the future , but never _
whether_. but link dynamics reveals many other kinds of relational structure besides domination .
we are beginning to see hints of how link theory could enable us to remove our causal colored glasses " .
it greatly simplifies many calculations to write the generator of a differential markov chain in the form g = 1+hdt , where 1 is the identity matrix . ignoring second - order effects
, we then have g@xmath9 = 1 + 2hdt etc .
, so we can calculate t(t ) by integration rather than by repeated matrix multiplication . in differential markov chains , it s customary to speak of h rather than g as the generator of the process
. let @xmath18(t ) be the state vector of the chain at t , and define d@xmath18(t ) as @xmath18(t+dt)-@xmath18(t ) ; then for any t we have d@xmath18 = g(@xmath18 ) - @xmath18 = h(@xmath18)dt .
many people have remarked on the resemblance between the equation d@xmath18 = h(@xmath18)dt and schrodinger s equation .
we ll now see that this resemblance is no accident .
first , note that 1+hdt is always invertible , it s inverse being 1-hdt ( we are ignoring second - order terms , so ( 1+hdt)(1-hdt ) = 1-h@xmath9dt@xmath9 = 1 ) . treating 1-hdt as a generator would give us our original markov chain reversed in time !
but does nt this clash with our earlier observation that the time - reversal of a markov chain is nt a markov chain ?
no , because closer inspection reveals that 1-hdt always contains negative probabilities " .
of course if we allow negative probabilities , then the answer is yes , and we have a useful tool for retrodicting state vectors .
however , be careful here . though the state evolution is reversed by the inverse process
, the forward process itself , which includes the joint distribution on all variables , is not , since the inverse process always has negative joint probabilities .
if we encounter a time - reversed markov chain , there are two things we can do in order to preserve the chain property : we can turn ourselves around in time , in effect putting the boundary condition on the future , or we can allow negative probabilities and work with the inverse chain , which means giving up studying the transitions and contenting ourselves with state vector dynamics .
suppose , though , that we encountered a chain c made up of two chains c1 and c2 , the one forward and the other reversed . then it would seem that we only have the second option
. however , link theory gives us a third , which is to put a separate condition on both the past and future of c , where the initial condition is white for c1 , and the final condition is white for c2 .
this is illustrated in fig .
the generator " g = g1&g2 of c is of course not a causal transformation , so c is not a markov chain .
but g persists throughout the time of the process and thus should qualify as a law .
what is essentially new is that in order to define the process as satisfying this law we must now be given two boundary conditions , one initial and one final , and that without this final condition , the process has no law ! this observation leads us to a basic generalization of the concept of markov chain : we ve been focusing on transformations ; let s now turn to states . as we saw above , the link states of a causal chain are of the form @xmath1v@xmath3w@xmath1 , where w is white . we know from quantum physics
that quantum density matrices are self - adjoint , which means that pure quantum states are of the form @xmath1v@xmath3v@xmath1 .
what kind of link chains have quantum states ?
suppose the initial link state of a quantum chain is @xmath1v@xmath3v@xmath1 . by the dynamical rule
, the second link state will then be g@xmath1v@xmath3v@xmath1g@xmath2 .
for the second state to be quantum , i.e. to be of the form @xmath1v@xmath3v@xmath1 , we must have @xmath19v@xmath1g@xmath2 = ( g@xmath1v@xmath20)@xmath21 = @xmath19v@xmath1g@xmath21 .
( recall that t@xmath21 is the adjoint of t , which for real matrices is just t with rows and columns reversed . )
if this equality is to be true for any initial state vector @xmath1v@xmath20 , i.e. if we are to be able to separate the law of the process from its boundary conditions , then we must have g@xmath2 = g@xmath21 .
but a transformation whose inverse is its adjoint is a unitary transformation .
thus we see that the generator of a quantum chain is unitary .
one question remains ; given the initial vector @xmath1v@xmath20 , what must be the final state vector @xmath19w@xmath1 ? let t be the product gggg ... of all the g s .
then @xmath1w@xmath3w@xmath1= t@xmath1v@xmath3v@xmath1t@xmath2 , so @xmath19w@xmath1 = @xmath19v@xmath1 t@xmath2 . we can write t@xmath2 as g@xmath2g@xmath2g@xmath2g@xmath2 .... which leads to a way of diagramming quantum chains that is very useful in analyzing quantum measurements , as we ll see in the next section . first an important definition : since s is self - adjoint , its non - zero entry must be in its diagonal , which means it is a pure state whose vector components are also sharp .
this leads to the diagram of a quantum system which is prepared in a sharp state given in fig .
3 . as with a differential markov chain , we can write the equation of a quantum chain in the form d@xmath18 = h(@xmath18)dt , where @xmath18 is the state vector .
but where is the i ? you ask .
we still havent arrived at physics yet ; what we have been studying is a more general mathematical structure known as _ real _ quantum mechanics . as mentioned , mackey showed how to get _ complex _ quantum mechanics as a special case of real quantum mechanics by introducing a ( real ) linear operator , which we ll call i , that commutes with every other operator .
we ll take a slightly different route . instead of starting out with i as an operator
, we ll start with a matrix representation of the complex numbers , where a+ib is represented as a 2x2 matrix c@xmath22 , with c@xmath11 = c@xmath12 = a , and c@xmath23 = -c@xmath24 = b. one can quickly verify that these matrices behave like complex numbers under addition and multiplication .
a complex probability measure " is then simply an additive measure on a boolean algebra having these matrices as its values .
let x be a variable on which there is such a complex probability distribution c(x ) .
we can map this onto a real distribution on three variables x , j and k such that for any x , p(j , k ) is the j , kth entry of c(x ) . with this mapping ,
a complex probability acquires a definite meaning in terms of real probabilities : it is a joint distribution on two binary variables j and k satisfying p(j=1 & k=1 ) = p(j=2 & k=2 ) and p(j=2 & k=1 ) = - p(j=1 & k=2 ) .
a complex variable x is then a process in x , j and k such that , for any m , p(j , k@xmath1x = m ) is a complex probability .
suppose e and e are independent events with complex probabilities c and c. what , then , is the probability of e&e ? we would of course like it to be cc. however , if c and c are independent real 2x2 matrices , i.e. independent joint distributions on i , j and i,j , then combining e and e produces a distribution p(j , k , j,k ) = p(j , k)p(j,k ) which does not even make sense as a complex probability .
this presents us with a choice : we must either give up trying to reduce complex probabilities to real probabilities , or else give up our old notion of independence , putting the complex product rule c(e&e ) = cc in its place .
since quantum mechanics uses this complex product rule in defining the inner product of two quantum objects , we ll take the second course , and ask what it means when translated into real probability theory . to multiply c and c means to _ contract _ k and j
, i.e. , to link k to j and then ignore k , or equivalently , since cc = cc , to contract j and k. thus to combine a complex process w with an independent complex process w into a complex process w&w we must always link a binary variable of one with a binary variable of the other . in quantum mechanics ,
an n - dimensional unitary transformation t can be thought of as operation on an object which is an entangled composite of an n - dimensional real object r and a two - dimensional real object j , where the entanglements satisfies the skew - symmetry rules above .
the 2x2 pieces of t can be thought of as operators on j. suppose t operates on an independent r&j. to bring t and t into the same universe of discourse , we must link j and j , thus multiplying the transformations on j times those on j. but it only makes sense to multiply transformations if they operate on the same object , so we conclude that j = j ! in other words
, there is a particular two - state particle that belongs to every quantum object ; this is what we referred to earlier as the janus particle .
that name was based on a conjecture relating its states to the equivalence of the heisenberg and schrodinger representations , so for present purposes lets give it another name : as is well known , only the relative phase of quantum amplitude can be observed , the absolute phase being unobservable .
this is equivalent to saying that the state of j is unobservable .
note that this means that pure " quantum states , regarded as real states , are always mixtures with two - dimensional degeneracy .
an interesting way to construct complex quantum mechanics is to start with real objects ri , add j to each of them , and then confine ourselves of those states of ri&j in which either j is independent of ri ( in the real sense ) , or is entangled in such a way that changing the state of j does not affect the probabilities in measurements of ri .. * janus particle * : a hypothetical two - state particle belonging to all quantum objects , among whose symmetries are the reversal of the heisenberg and schrodinger representations .
let us here go on record with our conjecture that the janus particle is the phase particle , and that resulting subject - object symmetry " is the conjugation symmetry of quantum mechanics . _ the measurement problem in link theory , the paradoxes of negative case counting , standard measurement theory , the projection postulate , the disturbance rule , the three axioms that define laboratory objects , functional objects , experiments , confined quantum objects , quantum time vs. laboratory time , prepared quantum chains , definite present and open future , the projection postulate in link theory , the collapse of the wavefront , the confined quantum object is a laboratory object , the white state , time loops revisited , macroscopic matter .
_ the born rule tells us that , for any proposition p and quantum variable x , the probability of a measurement of x satisfying p is tr(ps ) , where s is the link state of x. the dynamical rule tells how the state s changes with the passage of time .
so does nt that wrap it up ? what more is there to be said about quantum observation ? in the standard hilbert space approach , there is the problem of how to construe our classical " measuring instruments as quantum objects , which led von neumann to his famous construction of the movable boundary between object and observer . but this problem does nt arise in link theory , since quantum and classical can perfectly well coexist in the same joint distribution . indeed , as we have seen , the very distinction between classical and quantum rests in part on arbitrary decisions about how to analyze a process .
so what , then , is left of the infamous quantum measurement problem ? have we in fact solved it ? not quite .
in fact , the measurement problem takes a particularly acute form in link theory , as we ll now see .
4a shows a single complete measurement of a variable x in a pure state @xmath1v@xmath3v@xmath1 ; we simply draw an arrow from x to a recorder , linking x to an input variable of the recorder which is assumed to be white .
so far so good ; there s no measurement problem yet .
if we wait until time t(y ) to measure ( fig .
4b ) , then the state will have been transformed into tst@xmath21 , and an arrow from y to the recorder will again give results in agreement with standard theory .
still no measurement problem . but now let s consider two complete measurements , one before and one after transformation t ( fig .
each of these measurements will give the same probabilities that it did when it was the sole measurement .
however , that s not what the standard theory predicts .
what went wrong ? here
s one thing that went wrong .
consider a _ selection _ from among all the records of just those records for which the first measurement had the value k. within this subset , every event in x will have probability 1 or 0 , so x is independent of all other events , and v , which we ll assume is not sharp , is replaced by a sharp vector v. the result will be in most cases that the state of y is no longer quantum .
in other words , measuring the object throws it out of a quantum state .
but we re in worse trouble than that . in the system of fig .
4c , it may well happen that some of the joint probabilities on x and y are negative , which means our recorder has recorded negative " cases ! so what s wrong with recording negative cases ?
you may ask . remember , recorded cases are not just alternatives ,
they are _ instances _ ; they are actual things , marks on a tape , whatnot .
suppose your closet contains ten positive white shirts and ten negative green shirts .
if you reach into your closet for a shirt , you will come out empty handed , since there are 0 shirts there . on the other hand ,
if you reach in for a white shirt , you ll have ten to choose from . that s quantum logic . to see the problem of repeated measurement more sharply ,
consider a third measurement on z which comes after a second transformation t@xmath21 ( remember , t is unitary , so t@xmath21 = t@xmath2 ) ; this is shown in fig .
5 . as before
, we ll run a long series of recorded experiments . since tt@xmath21 = 1 , x and z
are perfectly correlated in this process .
since the recorder is white and therefore does nt effect the probabilities of x , y and z , this perfect correlation will show up in the total set of records , i.e. x and z are equal in every record . but now consider the subset s@xmath25 of all records for which y = k .
since y separates x and z , we know that x and z are conditionally independent given y , which means that they are independent within s@xmath25 .
thus the records in s@xmath25 will in most cases show x and z as having different values !
s@xmath25 is like your white shirts in the closet
it s a non - null subset of the null set . but
real shirts and real records just do nt behave like that .
obviously we ve got to take another tack . up to now in this paper
we have managed to avoid the logical anomalies that result from negative case counting by carefully steering away from them , the trick here being to think only about the arithmetic of ratios . but this wo nt work forever . sooner or later we have to face the paradox of the white shirts .
this is the paradox of the two - slit experiment seen straight on . with both slits open ,
the case counts cancel at the null in the wave interference pattern " where the detector lies , and no electrons are detected . but with one slit closed , the white shirt " electrons alone are selected , and they make it through .
we ll see that the solution of the measurement problem involves transforming this seeming paradox into a mathematical account of what it means for the future to be _
since the predictions of the standard theory of quantum measurement are in close agreement with experiment , we must represent measurement by link systems that make these same predictions . to see just what is required , here
is a brief summary of the standard theory , which is based on an assumption that von neumann called the _
projection postulate_. * projection postulate*. given a state s and a proposition ( projection ) p , there is an ideal measurement m to test for the truth of p that leaves those things that pass the test in the state psp / p , and those things that fail in the state psp/(1-p ) , where p = tr(ps ) and p is the negation 1-p of p. it follows directly that m leaves the object in the state psp+psp. let q be any observable with eigenvalues q@xmath26 , q@xmath27 , .. q@xmath28 , and let p@xmath26 , p@xmath27 .. p@xmath28 be the projectors " of these eigenvalues , i.e. , the ( ortho ) projections whose ranges are the eigenspaces of q@xmath26 , q@xmath27 .. q@xmath28 .
it is easy to show from the projection postulate that we can measure q by performing a series of tests for , p@xmath26,p@xmath27 ..
. the object will only pass one of these tests , p@xmath29 , which tells us that q has the value q@xmath29 .
the ensemble of objects thus tested will be in the mixture @xmath15p@xmath29sp@xmath29 , which gives us the general rule for the minimum disturbance " of a state by a measurement .
* disturbance rule * : given an object in state s , the least disturbing measurement we can make of quantity q is one which leaves the object in the state @xmath15p@xmath29sp@xmath29 , where the p@xmath29 are the projectors of q. the disturbance rule gives us a way to prepare a beam of objects in any state s , which is to start with a beam in any state and measure it in a way that yields the components of s , and then select out a subset of objects from each component to give the proper weight of that component in s. practically speaking , the standard theory of measurement works very well , at least for simple systems . but the projection postulate has caused a lot of trouble for people concerned with quantum foundations , since , by collapsing the wavefront " , it messes up what is otherwise a beautifully clean and simple unitary dynamics .
since link theory has no wavefront to collapse , it ought to be able to do better , and indeed it does ; in fact , we count its success in this regard as its major accomplishment to date .
we ll see in a while what the projection postulate means in terms of linking .
but first we need a description in our current language of the kinds of things that go on in a physics laboratory .
the first thing to be said about our idealized and simplified laboratory is that there is a clock on the wall .
someone , or something , is always watching that clock and logging every event , so that at the end of every experiment there is a diary of the experimental procedures and the outcomes of measurement .. the second thing to be said is that these experimental procedures can be repeated as often as necessary to obtain meaningful relative frequencies .
let s now characterize the _ objects _ in our lab .
* laboratory object * : a process k in which we distinguish two kinds of _ external _ variables , called _ input _ and _ output _ variables .
each such variable exists for only one clock tick ; for instance x could be the keyboard input to a computer at time t ; we ll write the time at which x occurs as t(x ) .
input variables are those that we control , while output variables are those that we measure , or use to control inputs .
k may have other variables besides its inputs and outputs , and it need not have inputs at all , though it must always have at least one output .
when we speak simply of _ the input _ to k , we ll mean all of its input variables together , or more exactly , the join of all these variables .
similarly for _ the output _ from k. the input ( output ) _ before _ or _ after _ time t@xmath26 is the join of input ( output ) variables x@xmath29 such that t(x@xmath29 ) @xmath19 t@xmath26 , or t(x@xmath29 ) @xmath20 t@xmath26 .
these three axioms broadly capture our common - sense notion of processes that we can observe and manipulate .
notice that they have nothing whatsoever to do with classical mechanics .
bohr characterized the laboratory as the place where newtonian mechanics reigns , but this was a mistake .
newton s laws neither imply nor are implied by the above axioms ; indeed they belong to a wholly different universe of discourse ( that s one reason why the young russell , who admired physics more than common sense , was so eager to jettison causality . ) * functional object * : one in which the output is a function of the input .
the requirement of whiteness on the input imposes the condition on the joint distribution that all probabilities are either 0 or 1/n , where n is dimension ( total number of cases . )
it s convenient to ignore normalization and represent such a joint distribution by the bit array which is the characteristic function of the object s function regarded as a relation .
* experiment * : a link system whose components are laboratory objects , one of which is a functional object known as a _ recorder_. the output of the recorder is a variable at the end of the process whose value is a record of all inputs and outputs to the object on which the experiment is performed .
it is assumed that every input variable in an experiment is linked to a prior output variable , and from this it follows that the process represented by the experiment is itself a laboratory object .
let s now turn to the construction of quantum objects .
we ll show these as enclosed in a dotted box , as in fig . 4 and fig .
our aim now is to design dotted boxes that enclose their quantum weirdness , so - to - speak , i.e. , boxes whose in - going and outgoing arrows satisfy our definition of a laboratory object .
such confined quantum objects are , properly speaking , the true objects of scrutiny in a quantum physics lab .
notice that the dotted boxes in fig .
4 are confined quantum objects ; they satisfy 1 ) and 2 ) because they have no inputs , and 3 ) because their probabilities satisfy born s rule and are therefore non - negative .
the dotted boxes in fig .
5 are not confined objects , however , since their outputs , i.e. the joint distributions on x , y and z , do have negative probabilities .
our challenge now is to redraw fig .
5 as a confined object that gives the right numbers .
there s one thing does nt look right about the box in fig . 5 , which is that to prepare it in state @xmath1v@xmath3v@xmath1 requires placing a boundary condition both before time t(x ) and after time t(y ) .
if we think of these conditions as inputs from the lab , the second condition violates 2 ) , the causality principle .
this is assuming that the time sequences inside and outside the box are the same .
but there is another possibility , and a more likely one , which is that time " in the quantum domain is somewhat autonomous from laboratory time , or to put it more abstractly , that the chains of quantum separability that we compared to markov chains are in themselves not temporal sequences at all . under this assumption
, it is reasonable to define a _ prepared _
quantum chain as one whose beginning and end are both at the initial laboratory time of the experiment ( fig 6a ) .
this construction looks more natural when we draw it with t@xmath21 factored out of w ( fig .
reversing rows and columns of t@xmath21 turns 6b into the equivalent form 6c .
let s then take this as our picture of a chain prepared in a pure state ; notice that the states s and s are identical .
more generally : suppose that when we measure x , the future measurement of y is indeterminate , in the sense that all outcomes are possible .. this indeed is our experience of time , which we express by saying that there does not exist at time t(x ) a record of the outcome of the measurement at t(y ) .
thus it makes no sense to speak at t(x ) of the subset of records for which y has a certain value . but
this implies that x and x are linked at time t(x ) by the matrix tt@xmath21 = 1 , which means that recording the value of x at t(x ) also records the value of x at t(x ) . since we must sum over all values of y to evaluate probabilities at t(x ) , this avoids the paradox of fig .
now what happens when we measure y ?
in our earlier example , this measurement unlinked x and z , making them in fact independent .
however , in the present case , x is already on record as having the same value as x , so it of course can not become independent .
how do we show this in our diagram ? simply by drawing a line between x and x ( fig .
recall that a recording device is a laboratory object having ( white ) inputs whose output at the final time is a 1 - 1 function of the history of its inputs .
before y is measured , this line merely duplicates tt@xmath21 , so drawing it would be redundant .
but after y has acquired a definite value , this line in effect remembers the measurement of x , thus holding the fort against the disruption that would otherwise be produced by fixing y. therefore there will be no records at all where x and x are unequal , and the paradox of a non - null subset of the null set is avoided . * projection postulate for a complete measurement * : to make a complete measurement of the state of a prepared quantum chain at time t(x ) , insert a link between x and x and measure the linked variable x , i.e. _ probe _
x. the choice of whether or not to probe x can be thought of as a double - pole switch .
just what do we mean by the term ` switch ' ? by a _
single - pole switch _ we mean a three - variable component sw[s , x , y ] where s is a binary variable with values on and off , sw[on , x , y ] is the identity matrix , and sw[off , x , y ] is the white matrix in which every element equals 1/n , n being the dimension of x and y. notice that for unlinked x and y , s is white , since the sum of matrix elements in the on condition is the same as that in the off condition . by a _
double - pole switch _ we mean a pair of single - pole switches whose s s are linked .
the crucial question now arises whether this switch input satisfies axioms 1 ) and 2 ) , i.e. whether it is white and causal .
if so , then the dotted box qualifies as a laboratory object . to say that the switch is white means that the probability p of the process is unchanged by drawing the line . by the summation theorem , p is the sum of the joint distribution on x and x. drawing the line eliminates all off - diagonal terms in this sum .
but since tt@xmath21 = 1 , these off - diagonal terms are all 0 anyway , so the line does nt affect p. this takes care of 1 ) ; we ll come to 2 ) shortly . looking at fig .
8b , we see that closing the switch breaks the diagram into two parts separated by x. the second part reduces to a circle , as far as x is concerned , so its presence at x is white .
thus the state s of the link from v to x is unchanged by the measurement of x ; it is still @xmath1v@xmath3v@xmath1 , and the probability distribution on x is still the diagonal of s , in accordance with born s rule .
but what is state s@xmath26 ?
clearly it has the same diagonal as s. but , since tt@xmath21 = 1 , its off - diagonal elements are all 0 .
s@xmath26 is then the mixed state having sharp components weighted by the probability distribution on x. thus we see that this change of state from s to s1 satisfies the standard theory of measurement , and is indeed nothing else but our old friend the collapse of the wavefront ! of course there is now no longer a wavefront to collapse , so we ll have to find another name for it .
whatever we call it , though , it s the crucial event that links the micro to the macro , and it is brought about by the _ passive _ act of recording the state of x. our probe does not disturb what it sees ; its effect comes from seeing alone . that seeing alone is sufficient to collapse the wave function was a point made by hpn in the context of the double - slit experiment performed with low - velocity cannon balls , armor plate and mechanical detectors @xcite .
this point needs a little further discussion here in anticipation of both more abstract _ and _ more physical discussions of the measurement problem which are in preparation .
`` seeing '' the low - velocity cannon ball go through one slit or the other requires illumination , and reflection of at least one photon from the cannon ball itself , together with reflection of enough light from the two slits to distinguish which it went through .
hence there is actual energy transfer from the ball to the observer .
this energy transfer has to be recorded before this particular case can be used to enter as a record in the compilation of single slit cases in contrast with double slit cases .
the act is `` passive '' but not devoid of physical content and ( pace von neumann and wigner ) does _ not _ require a
_ conscious _ observer to `` collapse the wave function '' , as we hope the analysis made in terms of the abstract theory presented above has made clear . more generally , making the `` quantum system '' into a `` laboratory object '' requires conscious experimenters to set up the apparatus and to interpret the results after they are recorded , but this in no way distinguishes it from the protocols of classical experimentation where no question of quantum effects arises .
when such processes are studied in enough detail to observe the progressive ( and predicted ) `` randomization '' of mesoscopic quantum states as has recently been done @xcite with two separating rydberg atoms in @xmath30 states the quantum _ theoretical _ treatment must include many more degrees of freedom than are usually included when discussing quantum measurement and the projection postulate .
we believe that it might prove instructive to see how much of the analysis of such systems depends on how they are `` carved up '' between classical and quantum degrees of freedom and where considerations of experimental accuracy have to enter . closing the record
" switch does not change the observed probability distribution on x , and in this sense it does not disturb the measurement .
however , it does affect the state of y , which would be tst@xmath21 with the switch open but is t s@xmath26t@xmath21 with the switch closed .
from the point of view of the laboratory , closing the switch causes a change in the state of y. inside the dotted box , though , there is nothing resembling cause and effect ; indeed , nothing even happens inside the box , since closing the switch merely duplicates an existing link .
measuring y is just like measuring x , the only difference being that s@xmath26 is mixed ( fig .
thus we see that throwing the switch at y does nt affect the measurement of x , or the whiteness of the x switch , so the switch inputs satisfy 2 ) , the causality axiom .
this shows that the dotted box is a classical object , and the sequence of measured variables is in fact a markov process .
so far we ve only been considering complete measurements .
an incomplete measurement of x is represented by a functional box f(x ) going from x to the recorder , where f is not 1 - 1 .
recall that , in a complete measurement , connecting x to x means that by recording x at t(x ) we have also recorded x at t(x ) .
however , in the case of a partial measurement , it s not x but only f(x ) that gets recorded at t(x ) , so this connection should go from f(x ) to f(x ) rather than from x to x ( fig .
this arrangement satisfies the standard general rule s = @xmath31p@xmath25sp@xmath25 for the change of state due to a partial measurement , where the p@xmath25 are the projectors of the measured observable .
what about preparation ?
as remarked , a preparation is always equivalent to a selection based on a measurement .
but we can also represent a preparation as an input to an unprepared system , i.e. a system in the white state ( fig .
10a ) ; clearly such an input is always white . the white state , whose matrix is 1 , has a unique role in quantum theory , since it is unchanged by measurement ( i.e. , if s=1 , then @xmath31p@xmath25sp@xmath25 = @xmath31p@xmath25 = s ) .
the diagram of an unprepared and unmeasured quantum object is simply a circle ( fig .
10b ) ; we can think of it as representing pure entropy .
perhaps this is the only quantum diagram that makes sense in isolation from the classical world .
our conclusion that the collapse of the wavefront " requires an open future seems to reintroduce an absolute arrow of time .
but this is an illusion , since an unprepared system is symmetrical with respect to past and future , whether it is measured or not .
time s arrow comes not from quantum openness but from classical preparation , which we have conventionally placed at the beginning of the process , but which could just as well occur at the end , or at both the beginning and end , since it is equivalent to a selection of records .
thus our present analysis is consistent with our previous notion of the state throttle " in a time loop .
the topic of measurement , especially partial measurement , deserves a much fuller treatment than we have given it here .
however , even the present abbreviated account calls for a few more remarks concerning macroscopic matter : in our analysis above , we have treated quantum and laboratory objects as essentially different kinds of things . and
yet sooner or later we must face the fact that laboratory objects are made out of atoms , and try to render an account of just what this means . in standard quantum mechanics ,
when you assemble quantum objects you always get other quantum objects , and this indeed does create insoluble problems , since there s no way to bend " quantum - ness into anything else . the present approach does nt have that problem , of course , since it treats quantum - ness as a mode of analysis rather than an intrinsic property of things , and there is no reason why the appropriate mode of analysis should be the same at different scales .
but we can hardly let things go at that , since there is a huge body of empirical knowledge about the relationship between micro and macro that needs to be translated into our present conceptual scheme . in the last century
, the laws of mechanics were thought to form a closed system that completely governs physical change , all other physical facts being relegated to the category of accidents " of the world state .
how , in such a scheme , is one to deal with the undoubtedly lawful character of thermodynamic events like heat flow ?
the nineteenth - century answer , which never quite satisfied anyone , was to impose a secondary layer of
laws among the accidents " , so - to - speak , these being somehow based on the laws of probability . with the advent of quantum mechanics
this idea becomes pretty far - fetched , since the equation governing quantum mechanical change has , except for the factor i , exactly the same form as the diffusion equation , which is one of the alleged laws among the accidents .
let us hope that this classical two - layer concept of law is now well behind us .
how , then , do we describe the relationship of micro to macro without it ?
let s start at the level of process .
remember , in a process as such there are no states until you decide how to carve it up , only correlated variables , so it literally makes no sense to ask whether a process is quantum or causal .
it does , however , make sense to ask whether or not its joint distribution contains negative probabilities .
let s pretend for the moment that it does nt .
what we are then given is a huge number of micro - variables associated with individual atoms plus a much smaller number of macro variables associated with regions large enough to contain many atoms , the latter often closely corresponding to sums or averages of the micro variables .
the macro - variables can be directly and repeatedly measured without disturbing them , and they seem to satisfy pretty well our axioms for laboratory objects , at least for simple inorganic objects .
we implicitly assume that spatial boundaries inside the object correspond to separation boundaries in our present sense
one sometimes speaks of this assumption as non - locality .
applying our fundamental theorem makes it possible to define states at these boundaries , thus turning the macro - behavior of the object into a composition .
this is essentially how an scientist carves something up when he wants to make a computer model of it , and there is no doubt that it works very well .
the reason it works is that macro - behavior has a good deal of autonomy from what we assume about the behavior of the unmeasured micro - variables , and for most purposes we can ignore quantum effects completely , at least under average terrestrial conditions .
it might be supposed that this is simply a consequence of locality plus the complexity of the processes the macro - variables are averaging over , and indeed it would be , if the micro - probabilities were always positive .
however , with negative micro - probabilities the situation is quite different .
consider a chain whose generator is periodic , i.e. , whose nth power is the identity for some n. with only positive probabilities allowed , such a generator would have to be deterministic . but this is not true with negative probabilities , and physicists are well acquainted with periodic unitary transformations ( indeed , any self - adjoint unitary transformation has period two . ) now a probabilistic chain of period
n is of course separable at every variable , and would thus satisfy locality if it were strung through a material object .
however , its first and last variables are perfectly correlated , just as if they were directly linked ! from the point of view of common sense , this is non - locality with a vengeance . clearly we need a new assumption that keeps such non - local virtual links " from playing a significant role in macro - behavior .
physicists already have a name for what is probably the right assumption : decoherence .
since this concept is defined in terms of wave functions , it needs to be reformulated in the language of probability . and
it undoubtedly needs to be generalized to deal with matter at low temperatures , and perhaps with organic matter as well .
though all this could become technically complicated , it looks straightforward enough . there is a much more serious problem which is not solved by decoherence , however , and that is to explain why negative probabilities remain confined to the micro - realm .
what prevents them from breaking out " into the everyday world and playing havoc with logic ?
we avoided this problem in the case of quantum measurements by assuming that making a measurement replaces certain virtual links " due to quantum coherence by actual links .
something of the same sort must apply to the relationship between a macro - variable and the quantum average to which it corresponds .
that is , a macro - variable , by being linked to the world at large , actualizes the virtual link on the quantum level between an average variable and its time - reversed dual . to put it more simply , macro - regions are self - measuring .
notice that measuring an average of many quantum micro - variables extracts only a minute fraction of the information in a pure joint state of these micro - variables , so these macro - measurements " are essentially non - disturbing .
but before we break out the champagne to celebrate the final triumph of logic and common sense , there are a few more questions .
why does nt the time symmetry of the microscopic level carry over to the macro - level , at least to the extent of showing up as independent macroscopic boundary constraints on past and future ? to what extent does our reasoning in terms of states bear on the process itself rather than on how we choose to carve it up ? might it be that hanging on to our causal framework , even for everyday objects , exacts a large price in complexity , thus blinding us to simple and essential forms in nature ? and
finally , why should we assume that we can always find ways to keep the irrationality of negative probability out of our experience ?
this has been a long paper , but we fear it has only scratched the surface of its subject .
we ve concentrated here on the basics rather than on results .
we believe this was the best course , since these basics take some getting used to .
but in the long run , it s results that win the prize .
the history of science is littered with the remains of seductive theories that led nowhere , von neumann s quantum logic being a recent case in point .
the first place to look for results in link theory is obviously physics , about which we ve said relatively little , so here , in conclusion , is a brief prospectus on several new items of physics that we see on the horizon .
first , some general remarks about space .
so far , we ve mostly been drawing diagrams rather than maps ; in the dialogue between composition and extension , composition has had by far the louder voice .
but the world does nt present itself to raw experience as a neat arrangement of interchangeable parts first we have to carve it up . and before we can carve it up , we have to map it , or
mark the carcass " , as we put it earlier .
the places where we mark are the places that separate the information in the extended structure , the boundaries through which correlation flows , and which , when rendered immobile " by fixing the values of their variables , break the whole into independently variable parts .
the question naturally arises as to how the structure of the probabilistic separating relation p(a&b&c)p(b ) = p(a&b)p(b&c ) is related to more familiar geometric structures like metrics and topologies and linear orders .
consider , for instance , a markov process .
we always start out by giving its variables in a certain order , but is this order contained in the probabilities or is it arbitrary ?
we saw that the arrow is arbitrary , so we are really talking about the between - ness relation .
let s define b(x , y , z ) , read
y is between x and z " , to mean p(x , y , z)p(y ) = p(x , y)p(y , z ) .
we know that separation is equivalent to the markov property , which holds for any x , y and z in the given order . thus the question is whether b(x , y , z ) always excludes b(y , x , z ) and b(x , z , y ) ; let s call this the order property . if a process has the order property , then the linear order of its variables is indeed intrinsic to that process itself .
there are some obvious instances of markov processes in which the order property fails .
for instance , a sequence of independent variables is a markov process , but since any one of the variables separates any two others , the variables can be rearranged in any order ; intrinsically , they are a mere heap . at the opposite extreme
is the case where the variables are perfectly correlated , i.e. , where the transition matrices are permutation matrices .
if you fix any variable in such a process , you fix all the others too .
but , as we saw in section 1 , a fixed variable is independent of all others , so , again , any variable separates any other two .
it may seem odd to think of laplace s perfect newtonian clockwork universe as a mere disordered heap of states , but there it is !
the true opposite of the independent variable case is the differential process given by a system where g = 1 + hdt .
if h is not 0 and there are no negative transition probabilities , such a process is always ordered .
hdt does nt have to be truly infinitesimal to guarantee this , only rather small .
if g is constant in time , the probabilities not only define an order but a metric , at least up to a scale constant .
the local order of a continuous process is rather curious .
since it approaches the deterministic case in arbitrarily small neighborhoods , it is , so - to - speak , locally disordered .
that is , it would require arbitrarily many repetitions of the process to statistically determine the order of points that are arbitrarily close together ; this is rather suggestive of the time - energy uncertainty principle .
the same kind of reasoning can be applied to higher - dimensional continuous processes , whose probability distributions in general contain the intrinsic geometry of the space that is used to define them , or at least its topology .
keith bowden@xcite has suggested that if we apply suitable perturbing random fields to classical fields , we can recover their spaces from the resulting joint probability distributions .
it would seem natural in a probability space " for the local structure to be quite different from the global structure , which might have some bearing on super - string theory .
what s the simplest random process of all ?
surely it s a coin toss .
let v be a coin tossing process whose heads / tails variable is x , and define the velocity v of v as the probability of x = heads minus the probability of x = tails ( we re not assuming it s a fair coin . )
notice that if you step one meter to the right or left every second according to the outcome of a toss of v , then v is your average velocity to the right in meters per second .
what s the simplest of all systems with more than one component ?
surely it s a pair of linked coin tosses .
call these components v and v , with binary variables x and x and velocities v and v. prior to linking , v and v are independent . linking x and
x produces the single - variable binary process u = v&v@xmath1x = x , which we can think of as a coin toss too .
what , then , is the velocity u of u in terms of v and v ?
proof : let p and q be the probabilities of heads and tails for v , and similarly let p and q for v. then v = p - q and v = p-q , and from the definition of linking one can quickly verify that u = ( pp-qq)/(pp+qq ) .
thus we must show that ( pp-qq)/(pp+qq ) = ( p - q+p-q)/[1-(p - q)(p-q ) ] .
now in fact these two expressions are not identical as they stand , but only become identical when we bring in the additional fact that probabilities add up to one , i.e. p+q = p+q = 1 .
the easiest way to take these conditions into account is to note that v = ( p - q)/(p+q ) and v = ( p-q)/(p+q ) and substitute these expressions for v and v in ( v+v)/(1+vv ) ; the resulting expression then reduces to ( pp-qq)/(pp+qq ) .
qed . applied to observer and object ,
the boost law implies the lorentz transformation thus we find that the concept of linking , which before led us immediately to the heart of quantum mechanics , has now led us immediately to the heart of relativity ! there is still a lot of work to be done to relate the above theorem to the concept of probability space " based on separability .
one approach here may be to interpret time lines " as binary markov chains from which the left - right variables are abstracted statistically .
1x1 space - time would then be the indefinite process that results from linking these velocity variables in an unspecified collection of such chains .
notice the formal resemblance here to our construction of complex amplitudes , which also resulted from linking an indefinite set of processes via a binary phase variable .
the question arises whether this resemblance is more than just an analogy .
could it be that at some fundamental level , the phase particle and the velocity particle " are one and the same ?
let s briefly consider where this would lead . since in ( complex )
minkowski space boosts are rotations of the complex plane , this identity would make the relativity of amplitude phase into a generalization of the relativity of motion . even more important for the science of the future is that the conjugation symmetry of the phase particle would become the symmetry of v and -v , which is the symmetry that results from reversing object and observer .
this supports our conjecture that the phase particle is indeed the janus particle , and makes more sense of pauli s mystical vision of i as the key to uniting physics with psychology , as described in chapter 2 .
but of course we are now looking rather far ahead .
noyes , h. p. [ 1980
] , `` an operational analysis of the double slit experiment '' , in _ studies in the foundations of quantum mechanics _ , p. suppes ed . , philosophy of science association , east lansing , michigan , p 103 .
a markov process in time can be generated either from a vector given when the process starts and propagate forward in time or from a vector given when the process ends and propagate backward in time ( see text ) . | we shall argue in this paper that a central piece of modern physics does not really belong to physics at all but to elementary probability theory . given a joint probability distribution j on a set of random variables containing x and y ,
define a link between x and y to be the condition x = y on j. define the _ state _ d of a link x = y as the joint probability distribution matrix on x and y without the link .
the two core laws of quantum mechanics are the born probability rule , and the unitary dynamical law whose best known form is the schrodinger s equation .
von neumann formulated these two laws in the language of hilbert space as prob(p ) = trace(pd ) and dt = td respectively , where p is a projection , d and d are ( von neumann ) density matrices , and t is a unitary transformation .
we ll see that if we regard link states as density matrices , the algebraic forms of these two core laws occur as completely general theorems about links .
when we extend probability theory by allowing cases to count negatively , we find that the hilbert space framework of quantum mechanics proper emerges from the assumption that all d s are symmetrical in rows and columns . on the other hand ,
markovian systems emerge when we assume that one of every linked variable pair has a uniform probability distribution . by representing quantum and markovian structure in this way
, we see clearly both how they differ , and also how they can coexist in natural harmony with each other , as they must in quantum measurement , which we ll examine in some detail . looking beyond quantum mechanics ,
we see how both structures have their special places in a much larger continuum of formal systems that we have yet to look for in nature .
6.0 in 8.6 in -0.25truein 0.30truein 0.30truein = 1.5pc slac pub-7890 + august 1998 + * process , system , causality , and quantum mechanics * + _ * a psychoanalysis of animal faith * _ tom etter + 112 blackburn avenue + menlo park , california 940252704 + and + h. pierre noyes + stanford linear accelerator center + stanford university , stanford , ca 94309 + submitted to _ international journal of theoretical physics .
_ * introduction : counting sheep * once upon a time there was a sheep farmer who had ten small barns , in each of which he kept five sheep . when asked how many sheep he had altogether , he replied many " , for people in those days counted on their fingers , and no one had ever thought of counting beyond ten . every morning
he would drive his sheep over the hill and through the woods to their pasture , where they assembled in five fields , ten sheep to a field .
the farmer , who was of a reflective bent , saw here a curious and beautiful law of nature : ten barns each with five sheep , and then five fields each with ten sheep ! "
unfortunately this law did not always hold , and when the wolves howled on the hill at night , it failed quite often .
the farmer had an explanation for this : the howling of the wolves greatly upsets my sheep , and the laws of nature , like the laws of man , are often disobeyed when agitated spirits prevail " .
the farmer realized that to make his law universal he would have to modify it thus : when tranquillity reigns , ten of five turn into five of ten . "
we today who know arithmetic would say that the farmer s law , though true enough in his particular situation , is nt a very good law by scientific standards .
it needs to be
factored " into two laws , the first being the simple and very general law that xy = yx and the second a more complicated and specialized law having to do with sheep and wolves .
the farmer was indeed aware that xy = yx , at least in the case of 5 and 10 , but what he could not see is that the essential condition for xy to be yx has nothing to do with sheep or wolves or tranquillity but is simply that the total number of sheep remain constant .
one reason he could nt see this is that he lacked any conception of the total number of his sheep ; that s because in those days there were no numbers beyond ten , just many " .
there are three morals to this tale .
the first is that it s not enough just to ask whether a law is right or wrong we should also ask whether it gets to the point .
the second is more subtle : if the point escapes us , maybe it s because we lack the raw materials of thought needed to even conceive of it .
the third is not subtle at all : learn to count !
we have learned to count beyond ten sheep and even beyond three dimensions , but we still are under a very stifling conceptual limitation in not being able to count beyond the two types of phenomena that we call _ classical _ and _ quantum_. this paper will set these two among many more .
it will do this by teaching us some new ways to count cases , such as how to keep counting when the count goes below zero !
this will provide us with the raw materials for thought we need to clearly see some crucial points that quantum philosophy has so far missed , notably the significance , or rather the insignificance , of the wave function , and the essentially acausal nature of quantum processes .
quantum mechanics has revealed many puzzling patterns in nature , perhaps the most puzzling being the epr correlation .
the explanations we hear for this phenomenon all too often resemble the farmer s spirit of tranquillity .
we ll see that we can factor " a simple piece of probability theory out of physics that makes sense of things like epr in much the same way that xy = yx makes sense of the farmer s sheep counts .
what s basically new here is that quantum phenomena in general can be represented as simple large number phenomena whose laws belong to the arithmetic of case counting .
but what about non - locality ?
the collapse of the wave front ?
quantum measurement ?
how does the present paper fit in with the more familiar ways of interpreting the formalism of quantum mechanics ?
let me briefly address this question .
it was recognized quite early that quantum mechanics bears the earmarks of a purely statistical theory .
the schrodinger equation looks very much like the equation governing a markov process , and we actually get the schrodinger equation if we multiply the generator of a self - adjoint continuous markov chain by i. now markov processes belong entirely to the theory of probability there s no physics in them at all .
could it be that quantum mechanics does for mechanics what statistical mechanics did for the theory of heat ?
can mechanics , and space and time along with it , be reduced to the statistical behavior of something simpler and more fundamental , or perhaps even to theorems in the bare science of probability itself ?
two obstacles have stood in the way of such a simplification .
the first is that probabilities ca nt be imaginary .
it turns out that we can define imaginary probabilities in terms of negative probabilities , but that would nt seem to be of much help have you ever counted fewer than 0 cases ?
but then again , come to think of it , have you ever seen a pile of gold bricks with fewer than 0 bricks in it ? and yet the mathematics of negative piles of gold bricks has become indispensable for keeping accounts , especially government accounts .
so why not
keep accounts " with negative probabilities ?
this first obstacle does nt look like it ought to stop us in our tracks for long , and if it were the only obstacle , it probably would nt .
the second has proved more obstinate .
we saw that the schrodinger and markov equations have the same form , but we must next ask , do they govern the same quantities ? the answer appears to be no .
the numbers that in the markov equation are probabilities , are the square roots of probabilities in the schrodinger equation .
this little disparity has for sixty years kept alive the notion of the wave function " whose amplitude " squared is probability . and this little disparity , this seeming technicality , has been the logjam that has kept quantum philosophy circling in the same stagnant pool of inadequate ideas for the last sixty years .
the present paper aims to break up that logjam .
let s assume we have overcome obstacle 1 and can now work within an extended theory of probability where cases count negatively as well as positively . as we ll see , most of the math in this extended probability theory is the same as in the all - positive theory .
now suppose we write down the schrodinger equation as a formal markov chain , in which the differential transition matrix operates on a state vector .
the numbers in this vector are the amplitudes of the wave function " , which we would now like to think of as probabilities .
the well - known problem with this construction is that the basic rule for quantum probability , the born rule , says that these amplitudes must be squared to give the probabilities that are actually observed in quantum measurement .
however , and now we are coming to the key idea , this chain of transition matrices we call a markov chain is only one of many ways to represent the joint probability distribution that is the markov process itself .
another is by a chain of joint probability matrices linked by conditioning events of the form x = y .
we define the state s of such a link as the probability matrix on x and y with the link removed . a state
is called quantum if the distributions on x and y are identical , and a chain is called quantum if its states are quantum and its matrices are unitary . now here is the punch - line : in a quantum chain , the unlinked probabilities on x and y behave like quantum amplitudes , while the linked probabilities , which are the diagonal entries of s , are the squares of these amplitudes , and hence behave like the probabilities they are supposed to be .
there is no wave function , only probabilities , positive and negative .
the link method embeds quantum mechanics in the mathematics of markov processes in such a way that quantum amplitudes are represented by unlinked probabilities and quantum measurement probabilities by linked probabilities . as we ll see , the same embedding maps classical markov chains into isomorphic representations of themselves . in linked chains of any kind , linked probabilities
are quadratic in unlinked probabilities , but in classical chains one of the factors is constant and factors out , so probability is in effect linear in amplitude , i.e. the classical state vector " is the usual probability vector .
if this sounds a bit cryptic , do nt worry , it will all be spelled out in detail .
most discussions of the meaning of quantum mechanics these days seem to be about the problem of the collapse of the wave function . " in link theory this problem simply vanishes , since there is no wave function to collapse .
imagine if the eighteenth century caloric were still hanging around as the official theory of heat : we d be chronically plagued by ever more complicated theories explaining the collapse of the caloric field " when you measure an atom s energy .
what a relief to get away from the spell of such nonsense !
this large - number explanation of quantum mechanics raises two basic questions : large numbers of what ? and must we buy it ?
the answer to the first question is implicit in the above discussion , but needs to be said simply : the things we count large numbers of are cases .
simple arithmetic reveals that the core quantum laws , in a generalized form , are features of any probabilistic system whatsoever .
von neumann s formulation of the born probability rule prob(p ) = trace(ps ) holds at every connection between the parts of such a system , and the dynamical rule st = ts governs every part that is connected at two places .
we brought up caloric to draw a parallel between our present situation and the situation in physics when it was discovered that the laws governing heat could be interpreted as statistical laws of atomic motion .
however , there is a big difference . in the case of heat ,
the statistical theory sat on top of the newtonian theory of motion , whereas in our case there is no underlying empirical theory at all .
probability theory is just the arithmetic of case counting , so the generalized quantum laws are like xy = yx in that their truth is assured , the only empirical issue being where and when they apply .
the answer to the second question is no , we do nt have to .
however , the same can be said about the arithmetical explanation of five fields with ten sheep each .
it s logically possible that when true tranquillity reigns , the gods always make sure that every field contains ten sheep ( presumably the age of true tranquillity is long since past ) .
it s also logically possible that the non - local guide wave " explanation of quantum phenomena is the right one . with both sheep and quantum ,
the arithmetical explanation makes so much more sense that it would be most malicious of the gods to reject it just to save our old habits of thought .
we ll see that there is another reason to prefer the arithmetical explanation , which is that , as our discussion of markov processes suggests , it also applies to classical things like computers .
this at last enables us to make sense of quantum measurement , which has always been a great mystery .
quantum and classical now stand revealed as two shapes " made of the same stuff , so there is nothing more mysterious about their both being parts of the same process than there is about round wheels and square windows both being parts of the same car .
the radical path also leads to a good kantian solution of hume s problem , which is that of finding causality in the order of succession , and we ll see that the choice between acausal and causal / classical thinking is to some extent a choice of analytical method , like the choice between polar and rectilinear coordinates . but then comes the big question : what about the other shapes ? the ones other than quantum or classical that we have never before imagined , and therefore never thought to look for in nature ?
we ll briefly touch on the big question , but it calls for a much bigger answer than we can give here , or now . |
the study of ultra - relativistic heavy - ion collisions provides insight into properties of very high density nuclear matter .
since the observed particle distributions to first order only reflect the conditions in the final state of the system , signatures originating from the early stage are needed to conclude whether the system passes through a partonic phase .
this is one necessary step to identify the predicted state of matter called the quark - gluon plasma ; the other one being the proof that the system is thermalized .
one observable which is sensitive to the early stage of the collision is the second harmonic coefficient , @xmath0 , of the fourier expansion of the azimuthal momentum distribution , called elliptic flow @xcite . in non - central collisions
the initial spatial anisotropy is transformed into an anisotropy in momentum - space if sufficient interactions occur among the constituents within the system .
once the system has expanded enough to quench the spatial anisotropy , further development of momentum anisotropy ceases . in the following
we will discuss results on @xmath0 for identified hadrons measured in @xmath1 minimum bias au+au collisions from the star experiment at rhic .
with the high statistics available from rhic s run iv the coverage for identified hadron @xmath0 was extended up to @xmath11 by disentangling the different contributions to the overall @xmath12 , as shown in fig.[fig:1 ] . to suppress non - flow effects
, particles were correlated to the reaction plane determined in the opposite @xmath13-subevent .
the difference of this @xmath0-measurement from the results of the method where the event plane is measured over the full acceptance , is an estimate of one of the contributions to the systematic error ; another factor in the systematic error is based on the estimated level of contamination of protons by kaons .
the systematic errors are shown as bands in fig.[fig:1 ] , while the error bars include statistical uncertainties only . at low
@xmath5 the systematic errors were estimated to be less than 10% @xcite . for identified hadrons ( upper panel ) and ratio ( lower panel ) between the measurements and a polynomial fit through all data points except pions for 200 gev minimum bias au+au collisions . for both figures
the open symbols for @xmath3 and @xmath4 were taken from @xcite.[fig:2 ] ] while the proton @xmath0 saturates at @xmath14 the measurements for @xmath3 show a decline starting from their maximum at about @xmath15 .
the flow of these non - strange hadrons is strikingly similar to that of the strange particles @xmath16 and @xmath17 . the mesons and baryons fall into separate groups and this behavior seems to extend out to rather high transverse momentum .
most surprisingly , even at the highest @xmath5 measured a large @xmath0 is still observed .
this appears to be in contradiction to parton energy loss models @xcite and multi - component calculations @xcite , which both predict a smaller @xmath0 at and above @xmath18 . at these high momenta
the remaining @xmath0 value should be essentially driven by the path - length dependence of jet quenching only @xcite .
to further study the observed grouping into mesons vs. baryons , the top of fig.[fig:2 ] shows the scaled @xmath19 for identified hadrons over a broad range of scaled transverse momentum @xmath20 . here
@xmath21 is the number of constituent quarks ( ncq ) of a given hadron .
figure [ fig:2 ] , bottom displays the ratio between the measurements and a polynomial fit to all data . at low @xmath20
@xmath22 the observed deviations from the fit follow a mass - ordering which is expected from hydrodynamic flow . at higher @xmath5 ,
all @xmath19 measurements are very close to the common ` mean value ' .
while the shown errors include statistical uncertainties only , it is still striking to see that the similarity of the scaled @xmath0 extends out to @xmath20 as high as @xmath23 .
it seems that all mesons fall above the fit while all baryons fall below .
the observation that the @xmath24 is similar for all the different particle species , strongly supports quark coalescence ( see for example @xcite ) as the dominant process of hadronization in the intermediate @xmath5 region . with the available statistics small deviations from exact ncq - scaling are now detectable , which were expected even for simple recombination models @xcite .
compared to @xmath3 , @xmath25 , and @xmath26 @xcite , measurements of particle spectra at rhic have shown that the multi - strange hadrons @xmath6 and @xmath27 are less affected by the hadronic phase @xcite than by the partonic phase . it has been demonstrated @xcite that the @xmath28 and @xmath29 baryons possess a significant amount of @xmath0 , which suggests that this azimuthal anisotropy is generated at an early and partonic stage of the system s evolution . for the @xmath6 meson , earlier measurements disfavored kaon coalescence as the dominant channel for @xmath6 production at rhic energies @xcite . also its lifetime is long compared to the lifetime of the fireball .
hence the @xmath6 is a useful probe of the early stage of the system as well .
the @xmath0 of the @xmath6 is particularly interesting , since the @xmath6-mass of @xmath30 is similar to the mass of the proton @xmath31 . for strange hadrons ( left ) and multi - strange hadrons ( right ) in 200 gev minimum bias au+au collisions .
the dashed lines show a common fit @xcite to the @xmath16 and @xmath32 data .
hydrodynamic model calculations @xcite are shown as shaded areas .
[ fig:3 ] ] the new high statistics measurements of azimuthal anisotropy @xmath0 of multi - strange hadrons show strong elliptic flow for all studied particles : @xmath7 , @xmath8 , and @xmath6 ( see fig.[fig:3 ] , right ) .
the error bars contain statistical uncertainties only .
for the @xmath6-measurement systematic errors obtained by the comparison of two different methods are also shown .
a common fit to the results obtained for @xmath16 and @xmath17 , as depicted in fig.[fig:3 ] , left and motivated by number - of - constituent - quark scaling @xcite , was used to compare the @xmath0-data of multi - strange hadrons to those of the other particles .
a detailed comparison , carried out by calculating @xmath33 for the difference between the measurements and the fit function ( see tab.[table:1 ] ) , suggests that @xmath7 and @xmath29 indeed flow as strongly as the other baryons , while the @xmath6 favors the similarity to other mesons .
note that the fit function does not reproduce the data perfectly due to small deviations from the exact number - of - constituent - quark scaling , as discussed in sec.[ncqoldi ] .
.comparison of @xmath34 measurements for multi - strange hadrons to a common fit @xcite to @xmath34 of @xmath16 and @xmath35 .
[ table:1 ] [ cols="^,^,^ " , ] + since the @xmath6 flows like a meson and not as strongly as baryons , e.g. the proton with its similar mass , it is found that the observed scaling is a meson - baryon effect and not a mass effect .
the observed strong flow of multi - strange hadrons substantiates the development of collectivity among partons at rhic .
we have presented the azimuthal anisotropy parameter @xmath0 for identified particles in 200gev minimum bias au+au collisions .
elliptic flow of @xmath36 and @xmath3 was measured up to @xmath37 , where these particles still show a strong @xmath0 signal .
the observed meson - baryon grouping at intermediate @xmath5 suggests parton coalescence to be the dominant process of hadronization in that region , even though small deviations from ideal number - of - constituent - quark scaling are observed .
this scaling is a meson - baryon effect and not a mass effect . finally , the multi - strange hadrons @xmath6 , @xmath7 , and @xmath38 flow as strongly as the other mesons and baryons , which confirms partonic collectivity at rhic .
the remaining item to address for the discovery of a quark - gluon plasma is the thermalization of the system .
99 a.m. poskanzer and s.a .
voloshin , phys.rev .
* c58 * ( 1998 ) 1671 .
s. adler _ etal .
_ ( phenix collaboration ) , phys.rev.lett .
* 91 * ( 2003 ) 182301 .
j. adams _ etal . _
( star collaboration ) , phys.rev .
* c72 * ( 2005 ) 014904 .
wang , m. gyulassy , phys.rev.lett .
* 68 * ( 1992 ) 1480 .
d. molnr , nucl - th/0503051 .
a. dainese , c. loizides , g. pai , eur.phys.j .
* c38 * ( 2005 ) 461 .
d. molnr and s.a .
voloshin , phys.rev.lett .
* 91 * ( 2003 ) 092301 .
j. adams _ etal . _
( star collaboration ) , phys.rev.lett .
* 92 * ( 2004 ) 112301 .
j. adams _ etal . _
( star collaboration ) , phys.rev.lett .
* 92 * ( 2004 ) 182301 .
j. adams _ etal . _
( star collaboration ) , phys.rev.lett .
* 95 * ( 2005 ) 122301 .
j. adams _ etal . _
( star collaboration ) , phys.lett .
* b612 * ( 2005 ) 181 .
x. dong , s. esumi , p. sorensen , n. xu and z. xu , phys.lett .
* b597 * ( 2004 ) 328 .
p. huovinen , private communication ( 2004 ) . | the azimuthal anisotropy parameter @xmath0 has been measured with high statistics for identified hadrons in @xmath1au+au collisions with the star experiment . at
high transverse momentum @xmath2 a strong @xmath0 for @xmath3 and @xmath4 is observed . in the intermediate @xmath5 region ,
number - of - constituent - quark scaling was tested to high precision .
a detailed comparison of @xmath0 for the multi - strange hadrons @xmath6 , @xmath7 , and @xmath8 with other particle species substantiates the development of collectivity among partons in the early phase of the collisions at rhic . |
with most of our used energy being lost as waste heat , there is a growing need for high - performance thermoelectric materials that can directly and reversibly convert heat into electricity .
the efficiency of a thermoelectric material is determined by the dimensionless figure of merit @xmath2 , where @xmath3 is the seebeck coefficient , @xmath4 is the electrical conductivity , @xmath5 is the temperature , and @xmath6 and @xmath7 are the electronic and phonon thermal conductivities , respectively . to obtain a high @xmath1 value , one must try to increase the power factor ( pf=@xmath8 ) and/or decrease the thermal conductivity
. however , the transport coefficients ( @xmath3 , @xmath4 , @xmath6 and @xmath7 ) are not independently tunable because they all depend strongly on the details of the band structure and the scattering of the charge carriers . as a result
, it is still a challenge to obtain high @xmath1 thermoelectric materials , hindering their wide applications .
several successful concepts have been developed to increase the @xmath1 value , such as the use of low dimensionality,@xcite which could enhance the pf due to the sharper density of states ( dos ) near the fermi energy or reduce the thermal conductivity originating from the increased phonon scattering .
therefore , many efforts have been devoted to search for low - dimensional thermoelectric materials .
due to the weak van der waals interactions between the neighboring layers , the layered materials serve as ideal candidates which can be readily exfoliated to two - dimensional ( 2d ) films.@xcite on the other hand , a high valley degeneracy in the electronic structure also contributes to the enhancement of the thermoelectric performance.@xcite such degeneracy can be engineered by tuning the doping and composition in a bulk material.@xcite if we could combine the above two concepts together , namely , introduce the high valley degeneracy into a low dimensional thermoelectric material , a large @xmath1 value may be achieved .
( a ) top and ( b ) side views of the hfs@xmath0 monolayer .
the black line in ( a ) denotes the primitive cell used in our calculations .
( c ) band structures of the hfs@xmath0 monolayer without ( solid black line ) and with ( dashed red line ) spin - orbit coupling ( soc ) .
the inset is a magnified view of the region marked by the rectangular outline .
( d ) absolute value of the seebeck coefficient and ( e ) power factor ( pf ) as a function of the carrier concentration .
the solid black and dashed red lines represent the results without and with soc , respectively . ] phonon spectra of the hfs@xmath0 monolayer under different external strains of ( a ) 0% ,
( b ) 6% , ( c ) 9% , and ( d ) 10% .
the gap between acoustical and optical phonon bands is highlighted in yellow color . ]
transition metal dichalcogenides ( tmdcs ) , which have the formula m@xmath9 ( m = transition metal ; @xmath10 = s , se , or te ) , represent a large family of layered materials .
tmdcs are presently being intensively researched due to their diverse and attractive properties.@xcite some of them , tis@xmath11 for example,@xcite were found to be good thermoelectric materials . when exfoliated into film ,
the 2d tis@xmath11 was predicted to have much improved thermoelectric performance compared with its bulk counterpart when the thickness fell below 14 layers.@xcite hfs@xmath11 is another typical tmdc .
it was theoretically predicted that the lattice thermal conductivity of the bulk hfs@xmath11 was about 9 wm@xmath12k@xmath12 at 300 k,@xcite which is much smaller than those of the mos@xmath11@xcite and ws@xmath11,@xcite making the hfs@xmath11 system a promising thermoelectric material .
the room - temperature @xmath1 value of the bulk hfs@xmath11 was about 0.06,@xcite , still far away from the requirement of the thermoelectric application.@xcite if using the method of low dimensionality , the thermoelectric performance of the hfs@xmath11 system may be further enhanced . in this work
, we investigate the thermoelectric properties of the 2d hfs@xmath11 monolayer .
it is demonstrated that the hfs@xmath11 monolayer is an excellent n - type thermoelectric material , with much improved thermoelectric performance compared with the bulk .
furthermore , the band valleys in the hfs@xmath11 monolayer can be effectively tuned by the external strain , and the seebeck coefficient and therefore the peak value of the pf are greatly increased . at the strain of 6% , where the degeneracy of the valence band valleys reach the maximum , the @xmath1 value of the p - type doped system is dramatically increased .
our calculations were performed within the framework of the density functional theory ( dft ) , as implemented in the abinit code.@xcite the exchange correlation energy was in the form of perdew - burke - ernzerhof ( pbe)@xcite with generalized gradient approximation ( gga ) .
the brillouin zone was sampled with a @xmath13 monkhorst - pack @xmath14 mesh .
the cutoff energy for the plane - wave expansion was set to be 600 ev .
based on the electronic structure , the electronic transport coefficients are derived by using the semiclassical boltzmann theory@xcite within the relaxation - time approximation , and doping is treated by the rigid - band model.@xcite the electronic thermal conductivity @xmath6 is calculated using the wiedemann - franz law @xmath15 , where @xmath16 is the lorenz number . in this work , we use a lorenz number of @xmath17 w / k@xmath18.@xcite in the calculation of the phonon dispersion , the density functional perturbation theory ( dfpt ) as implemented in the vasp package@xcite was used to calculate the force constant matrices .
a @xmath19 supercell was used and the phonon frequencies were obtained by the phonopy code.@xcite the lattice thermal conductivity was calculated by solving the phonon boltzmann transport equation within the relaxation time approximation , as implemented in the shengbte code.@xcite the second order harmonic and third order anharmonic interatomic force constants ( ifcs ) were calculated by using @xmath19 supercell with @xmath20 monkhorst - pack @xmath14 meshes and @xmath21 supercell with @xmath22 point , respectively .
the interactions up to third - nearest neighbors were considered when calculating the third order ifcs .
the bulk hfs@xmath0 crystallizes in the 1@xmath5-cdi@xmath0 structure with the space group of @xmath231.@xcite due to the weak van der waals interaction between the layers , few - layer hfs@xmath0 could be exfoliated from the bulk structure.@xcite the top and side views of the hfs@xmath0 monolayer are shown in figs .
1(a ) and ( b ) , respectively .
after full relaxation , the lattice parameters are calculated to be @xmath24 , slightly larger than the in - plane parameters of the hfs@xmath0 bulk.@xcite the structural stability is investigated by calculating the phonon spectrum , which is shown in fig .
there is no imaginary frequency in the phonon dispersion , indicating that the hfs@xmath0 monolayer is stable .
the result agrees well with the previous report,@xcite confirming the reliability of our calculations .
we next consider the electronic properties .
the band structures without and with spin - orbit coupling ( soc ) included are shown in fig .
1(c ) , displayed by solid black and short dashed red lines , respectively .
the hfs@xmath0 monolayer is an indirect - band - gap semiconductor , with the valence band maximum ( vbm ) and conduction band minimum ( cbm ) located at the @xmath22 and @xmath25 points , respectively .
the band splitting due to the soc is observed . in particular , at @xmath22 point , the splitting energy @xmath26 of the vbm is 0.13 ev , thus the degeneracy of the two valleys near the vbm changes from 2 without soc to 1 when soc is included .
however , the degeneracy of the conduction band valley near the cbm is unchanged .
the band gap is calculated to be 1.22 ( 1.29 ) ev with ( without ) soc included .
the electronic transport coefficients ( @xmath3 , @xmath27 ) are shown in figs .
1(d ) and ( e ) , respectively .
we can see that for the p - type doping , the seebeck coefficient and therefore the pf ( with relaxation time @xmath28 inserted ) are decreased due to the soc , while for the n - type doping , they keep unchanged .
if we notice the change of the band structure due to the soc , we can deduce that the decrease of the seebeck coefficient may be caused by the decrease of the valley degeneracy .
therefore , a possible way to optimize the pf is to engineer the band valleys .
strain is an effective way to tune the electronic structure of the 2d material . in the following , a biaxial tensile strain , which is defined as @xmath29 ,
is applied to the hfs@xmath0 monolayer . to see how large strain the hfs@xmath0 monolayer can withstand ,
we calculate the phonon spectra when the strain is increased up to 10% .
the results for the strains of 6% , 9% , and 10% are displayed in figs .
2(b ) , ( c ) and ( d ) , respectively .
no imaginary frequencies are observed in the phonon spectra until the strain increases up to 9% .
when the strain reaches 10% , small imaginary frequency appears along the @xmath30 direction , thus the system tends to be unstable .
therefore , the largest strain that the hfs@xmath0 monolayer can withstand is 9% .
the hfs@xmath0 monolayer can withstand a relatively large strain , mainly originating from the special sandwich structure . when the strain is applied , the outer s atom layers move inside , so the thickness of the hfs@xmath0 monolayer is decreased , while the bond length between hf and s atoms is slightly altered .
energy difference between the three valence band valleys ( @xmath31 ) and between the three conduction band valleys ( @xmath32 ) as a function of the applied strain . ]
next , the strain effect on the electronic properties is investigated .
when the doping concentration is not very high , the electronic transport coefficients of the p- and n - type doped systems are determined by the bands near the vbm and cbm , respectively . for the valence bands ,
the three valleys highlighted in different color in fig .
3(a ) are denoted as i , ii , and iii , respectively , while the three conduction band valleys are denoted as iv , v , and vi . when no strain is applied , the energy of band valley ii is the largest among the three valence band valleys and the energy difference between valleys ii and i ( iii ) is denoted by @xmath31 , as shown in fig .
the energy of the valley iv is the smallest among the conduction band valleys and the energy difference between iv and v ( vi ) is denoted by @xmath32 .
when the strain is applied , for the valence band valleys , the valleys i and iii are elevated gradually when the strain is increased , while the valley ii is gradually lowered , resulting in the decrease of @xmath31 , as displayed in fig .
4 . at the strain of 6% ,
the three valleys converge in energy and @xmath31 is decreased to zero . at this strain condition
, the degeneracy of the valence band valley reaches the maximum , as displayed by the inset of fig .
when further increasing the strain , the energy of valleys i and iii becomes larger than that of the valley ii , and the degeneracy is reduced , as demonstrated by the inset of fig .
3(f ) . for the conduction band valleys , as the strain
is increased , the valley iv is gradually elevated , while the valleys v and vi are lowered , therefore , the @xmath32 decreases monotonically with the increase of the strain ( see fig .
the increase of the band valley degeneracy may be beneficial to the thermoelectric performance . in the following ,
the electronic transport coefficients of the hfs@xmath0 monolayer under different strains are calculated by using the semiclassical boltzmann theory .
since the system becomes unstable when the strain reaches 10% , we only consider the condition when the strain is not larger than 9% .
figure 5(a ) shows the calculated absolute values of the seebeck coefficients at 300 k as a function of the doping concentration .
we can see that for the p - type doping , the seebeck coefficient first increases as the increase of the strain , reaches its maximum at the strain of 6% , and then slightly decreases when further increasing the strain . for the n - type doping , however , the absolute value of the seebeck coefficient increases monotonically as a function of the strain , thus it reaches the maximum at the strain of 9% . if we notice the energy difference between the band valleys ( see fig .
4 ) , we can find that the seebeck coefficient increases as the absolute value of the valley energy difference is decreased .
when @xmath31 decreases to zero , the degeneracy of the valence band valleys becomes the largest , and thus the absolute value of the seebeck coefficient of the p - type doping reaches the maximum .
the trend of the electrical conductivity as a function of the strain ( see fig .
5(b ) ) is just opposite to that of the seebeck coefficient . here
the relaxation time @xmath28 is inserted as a parameter .
the electrical conductivity @xmath33 is generally decreased by the applied strain , which is detrimental to the pf . whether the pf will be improved by the strain or not is determined by the balance between the seebeck coefficient and electrical conductivity . in fig .
5(d ) , we plot the pf ( with @xmath28 inserted ) at 300 k as a function of the doping concentration under different strains . for the p - type doping ,
the peak value of the pf first increases as increasing the strain , reaches the maximum at the strain of 6% , and then decreases slightly when the strain is further increased . for the n - type doping , however , the pf increases monotonically with the increase of the strain .
the tread of the pf as a function of the strain is the same as that of the seebeck coefficient , indicating that the negative effect of the strain on the electrical conductivity is overweighed by the increase of the seebeck coefficient .
therefore , the pf of the hfs@xmath0 monolayer can be greatly improved by the valley engineering through the method of strain .
.[table1-relaxation time]effective mass ( @xmath34 ) , carrier mobility ( @xmath35 ) , relaxation time ( @xmath28 ) at 300 k in the zigzag and armchair directions of the unstrained and 6% strained hfs@xmath0 monolayers . [ cols="^,^,^,^,^,^,^ " , ] to evaluate the figure of merit , we have calculated the lattice thermal conductivity @xmath7 , which is 2.92 wm@xmath12k@xmath12 at 300 k for the unstrained hfs@xmath0 monolayer .
the hfs@xmath0 monolayer has much smaller lattice thermal conductivity than those of the mo / w based tmdc monolayers,@xcite the same as the case of the bulk systems.@xcite moreover , when under the external strain of 6% , the room - temperature @xmath7 is further suppressed to be 2.22 wm@xmath12k@xmath12 , due to the phonon softening of transverse and longitudinal acoustic ( ta and la ) modes , as well as the reduced gap between the acoustical and optical phonon bands ( see fig .
the very small lattice thermal conductivity of the hfs@xmath0 monolayer indicating that this system may have much improved thermoelectric performance .
combining all the calculated coefficients together , we plot in fig .
6(b ) the room temperature @xmath1 value as a function of the carrier concentration .
when no strain is applied , a @xmath1 value as high as 4.48 is achieved for the n - type doped system , which is significantly improved compared with the corresponding bulk.@xcite the @xmath1 value of the n - type doped system is much larger than that of the p - type doped one .
the hfs@xmath0 monolayer can be used as an excellent n - type thermoelectric material .
however , in the fabrication of thermoelectric modules , both the p- and n - legs are needed .
interestingly , we find that the @xmath1 value of the p - type doped system is significantly increased by the applied strain , from 0.60 for the unstrained hfs@xmath0 monolayer to 3.67 for the 6% strained one . however , for the n - type doping , in spite of the increased peak value of the power factor ( with relaxation time inserted , see fig .
5(d ) ) , the @xmath1 value is decreased by the external strain , from 4.48 for the unstrained system to 3.08 for the strained one .
this is because that when inserting the thermal conductivity to evaluate the @xmath1 value , the optimal doping concentration gets decreased , moving to the region where the pf of the strained system is slightly smaller than that of the unstrained one .
although slightly decreased , the room - temperature @xmath1 value is still above 3.0 , reaching the requirement of the thermoelectric application.@xcite in table 2 , we summarize the maximum @xmath1 values of the unstrained and 6% strained hfs@xmath0 monolayer , with the optimal doping concentration and the corresponding seebeck coefficient , electrical conductivity , electronic and lattice thermal conductivities included . for the p - type doping , we can see that at the optimal doping concentration , where the maximum @xmath1 value is obtained , both the seebeck coefficient and the electrical conductivity is increased by the strain , while the lattice thermal conductivity is decreased , which lead to the significantly increased @xmath1 value . for the n - type doping , at the optimal doping concentration , although the electronic and lattice thermal conductivities are decreased , the seebeck coefficient and the electrical conductivity are also decreased , resulting in the decreased @xmath1 value for the 6% strained system .
in conclusion , we have investigated the electronic , phonon , and thermoelectric properties of the hfs@xmath0 monolayer .
the band valleys of this 2d material can be effectively engineered by the external strain . at the strain of 6% ,
the three valence band valleys converge in energy and the degeneracy of the valleys reaches the maximum , while the energy difference of the conduction band valleys decreases monotonically as the strain increases up to 9% , the largest strain the hfs@xmath0 monolayer can withstand . although the electrical conductivity is decreased by the strain , the increase of the seebeck coefficient overweighs the decrease of the electrical conductivity and thus the peak value of the power factor ( with relaxation time inserted ) increases monotonically with the decrease of the energy difference among the band valleys . at the strain condition of 6%
, the maximum room - temperature @xmath1 value of 3.67 can be achieved for the p - type doped system , which is five times larger than that of the unstrained one .
our results indicate that the thermoelectric performance of the hfs@xmath0 monolayer can be greatly improved by the band valley engineering through the method of strain . in experiments
, we can introduce strain to a 2d material in various ways .
traditionally , strain can arise from the lattice mismatch between epitaxial thin films and substrates.@xcite in recent years , with the progress in nanotechnology , one could transfer a 2d film to a soft supporting substrate and apply strain to the film by either stretching@xcite or bending@xcite the substrate . by using this method
, extremely large strain could be achieved .
for example , strain as high as 30% has been realized in graphene.@xcite these methods can be readily transferred to the hfs@xmath0 monolayer , which deserves further study in the future experiments .
this work was supported by the national key research and development program under contract no .
2016yfa0300404 , national natural science foundation of china under contracts no . 11404340 , 11274311 , u1232139 , 11674326 and 11574108 , the anhui provincial natural science foundation under contract no .
1408085ma11 , the china postdoctoral science foundations ( grant no . 2014m550352 and 2015t80670 ) .
the calculation was partially performed at the center for computational science , caships .
j. n. coleman , m. lotya , a. oneill , s. d. bergin , p. j. king , u. khan , k. young , a. gaucher , s. de , r. j. smith , i. v. shvets , s. k. arora , g. stanton , h .- y .
kim , k. lee , g. t. kim , g. s. duesberg , t. hallam , j. j. boland , j. j. wang , j. f. donegan , j. c. grunlan , g. moriarty , a. shmeliov , r. j. nicholls , j. m. perkins , e. m. grieveson , k. theuwissen , d. w. mccomb , p. d. nellist and v. nicolosi , _ science _ , 2011 , * 331 * , 568@xmath36571 .
r. j. smith , p. j. king , m. lotya , c. wirtz , u. khan , s. de , a. oneill , g. s. duesberg , j. c. grunlan , g. moriarty , j. chen , j. wang , a. i. minett , v. nicolosi and j. n. coleman , _ adv .
_ , 2011 , * 23 * , 3944@xmath363948
. x. gonze , j .-
m . beuken , r. caracas , f. detraux , m. fuchs , g .-
rignanese , l. sindic , m. verstraete , g. zerah , f. jollet , m. torrent , a. roy , m. mikami , p. ghosez , j .- y .
raty and d. c. allan , _ comput . mater .
_ , 2002 , * 25 * , 478@xmath36492 . x. gonze , g .-
m . rignanese , m. verstraete , j .-
beuken , y. pouillon , r. caracas , f. jollet , m. torrent , g. zerah , m. mikami , p. ghosez , m. veithen , j .- y .
raty , v. olevano , f. bruneval , l. reining , r. godby , g. onida , d. r. hamann and d. c. allan , _ z. kristallogr . _ , 2005 , * 220 * , 558@xmath36562 . | the electronic , phonon , and thermoelectric properties of a two - dimensional hfs@xmath0 monolayer are investigated by using the first - principles calculations combined with the boltzmann transport theory .
the band valleys of the hfs@xmath0 monolayer can be effectively tuned by the applied biaxial strain . the seebeck coefficient and therefore the peak value of the power factor ( with the relaxation time inserted ) increase
when the degeneracy of the band valleys is increased by the strain . when no strain is applied , the hfs@xmath0 monolayer is an excellent n - type thermoelectric material , while the thermoelectric performance of the p - type doped one is poor .
the applied tensile strain of 6% can increase the room - temperature @xmath1 value of the p - type doped system to 3.67 , which is five times larger than that of the unstrained one .
the much more balanced @xmath1 values of the p- and n - type doping are favorable for fabrication of both p- and n - legs of thermoelectric modules .
our results indicate that the thermoelectric performance of the hfs@xmath0 monolayer can be greatly improved by the valley engineering through the method of strain . |
the subject of decoherence of a quantum system has witnessed a renewed interest in the last years .
in addition to being central to our understanding of the transition from quantum to classical physics@xcite , it is of outmost importance in the blooming field of quantum information@xcite , where decoherence or dephasing represent loss of information . a quantum superposition state `` collapses '' due to the dissipative interaction with the environment ( e.g. in the case of trapped atoms
, the coupling to the electro - magnetic vacuum leads to spontaneous scattering of photons ) .
this interaction is irreversible in nature . on the other hand ,
the macroscopic ( ensemble averaged ) response of a quantum system prepared in a superposition state decays due to the dephasing resulting from local variations in the evolution of the system ( e.g. inhomogeneous broadening ) . by stimulating an effective time reversal
, revivals of coherence in dephased systems ( at least partial ones ) have been reported for spin echoes @xcite and photon echoes @xcite , and more recently for a motional wave packet echo using ultra cold atoms in a one - dimensional optical lattice @xcite . in this paper
we discuss the technique of `` echo spectroscopy '' on ultra cold atoms trapped in an optical dipole trap , which utilizes a micro wave ( mw ) analog to the photon echo @xcite .
we show that this technique can be used to study the quantum dynamics of trapped atoms , even when many states are thermally populated .
we show that the stability of eigenstates under perturbations of the hamiltonian plays an important role for the echo signal .
numerical calculations of this quantity for an atom - optics bunimovich stadium are presented .
next , by overcoming the inhomogeneous shift , which is the main limiting factor for the coherence time in dipole traps @xcite , echo spectroscopy opens the way for using dipole trapped atoms in quantum information and possibly for high precision spectroscopy .
finally , our analysis leads to a proposal of a dephasing free dipole trap .
one of the most widely used mw spectroscopy techniques is the so called ramsey method @xcite .
it works on two level atoms after the following principle : the atoms are prepared in one of their two states ( denoted @xmath0 ) .
next , a short mw - pulse close to resonance transfers the atoms into a coherent superposition state of the two levels ( this pulse is usually called a @xmath1 pulse ) , which is left to evolve for a certain time , without mw - irradiation . due to the detuning between the atomic resonance frequency and the mw - frequency
, the time evolution will cause an atomic phase to develop in the rotating frame of the mw field .
a second @xmath1-pulse maps this phase onto populations of the two levels , which can be detected .
if the mw - frequency is scanned , this phase difference is scanned , and sinusoidal oscillations in the populations , known as ramsey fringes , are seen . when ramsey spectroscopy is performed on an ensemble of
inhomogeneously broadened two level atoms , each atom will acquire a slightly different phase in the dark period , causing the macroscopic fringe contrast to decay when increasing the time between the @xmath1-pulses . thus the ramsey fringe contrast is a measure of the dehpasing of the system .
the dephasing can be reversed by stimulating a _
coherence echo_. this is done by adding a population inverting pulse ( a @xmath2-pulse ) between the two @xmath1-pulses .
if the @xmath2-pulse is added exactly in the middle between the two @xmath3-pulses each part of the superposition state generated by the first @xmath1-pulse will have spent an equal amount of time in both levels at the time of the second @xmath4-pulse , and will therefore be exactly in phase with each other .
this means that a coherence echo have appeared at the time of the second @xmath1-pulse even for a system that have dephased completely before the @xmath5-pulse . when measuring the populations the coherence echo is therefore observed by seeing that all the atoms have returned to the initial state after the @xmath1-@xmath5-@xmath6-pulse sequence . if the time ( @xmath7 ) between the first @xmath6 and the @xmath5-pulse is kept constant , and the time between the @xmath5 and the second @xmath1-pulse ( @xmath8 ) is swept , the coherence echo is seen as a decrease in the population of @xmath9 for @xmath10 ( see fig .
[ fi1 ] ) .
the population @xmath11 for @xmath10 is referred in what follows as the `` echo signal '' , where @xmath12 indicates full revival of coherence and @xmath13 indicates complete dephasing .
experimental data shown in this paper is taken using two hyperfine levels of the ground state of @xmath14 atoms trapped in a dipole trap .
these two levels ( @xmath15 , denoted @xmath16 , and @xmath17 , denoted @xmath18 ) are separated by the energy difference @xmath19 where @xmath20 ghz is the hyperfine splitting .
for spectroscopy of trapped atoms we must consider the entire hamiltonian including both the internal and the external degrees of freedom of the atom . since the dipole potential is inversely proportional to the trap laser detuning @xmath21 @xcite
there is a slightly different potential for atoms in different hyperfine states for our experimental parameters . ] and the hamiltonian of our trapped two level atom can be written as : @xmath22 where @xmath23 is the atomic center of mass momentum and @xmath24 [ @xmath25 the external potential for an atom in state @xmath16 [ @xmath9 ] , much smaller than @xmath27 .
the atoms are initially prepared in their internal ground state @xmath0 .
their total wave function can be written as @xmath28 @xmath29 , where @xmath29 represents the motional ( external degree of freedom ) part of their wave function .
if a mw field close to resonance with @xmath30 is applied , transitions between the eigenstates of the hamiltonian corresponding to different internal states , can be driven .
the small size of optical traps ( typically on the order of @xmath31 @xmath32 or less ) as compared to the mw wavelength ( @xmath3310 cm ) , ensures that the momentum of the mw photon can be neglected ( lamb - dicke regime @xcite ) .
even though the internal states of the atoms are described as a two - level system eq.[hami ] describes a multi - level system in two energy manifolds . when the mw - field is added to the hamiltonian in eq.[hami ] and the time - dependent schrdinger equation
is formulated it is seen that the transition matrix elements between eigenstates of eq .
[ hami ] are given by @xmath34 where @xmath35 is the free space matrix element for the internal state transition , and @xmath36 is the overlap between the initial motional eigenstate of @xmath37 and an eigenstate of @xmath38 .
therefore , when a strong and short mw pulse is applied , the motional part of the initial wave function is simply projected into @xmath39 . if @xmath40 then clearly @xmath41 . for a small enough potential difference ( perturbation ) @xmath42 the approximation @xmath43 is valid . in the general case , a projected eigenstate of @xmath37 will not be an eigenstate of @xmath38 , and therefore will evolve in the new potential , causing the overlap @xmath44 to decay ( @xmath45 @xmath46 ) . ) measured as a function of time between @xmath47-pulse and second @xmath48-pulse ( @xmath49 ) for fixed @xmath50 . a dip in @xmath11
is seen , showing a `` coherence echo '' .
@xmath51,width=288 ] if ramsey spectroscopy is performed on atoms initially prepared in an eigenstate of @xmath24 characterized by the quantum number @xmath52 , then the probability to be in the internal state @xmath9 after the @xmath53-@xmath54 pulse sequence can be shown to be : @xmath55 where @xmath56 is the mw - frequency , @xmath57 is the time between the pulses , and @xmath58 is a generalized , state - dependent , detuning ( it reduces to the detuning when only two levels are coupled by the mw - field ) , defined so @xmath59 is real and positive .
scanning @xmath56 for a fixed @xmath60 yields the usual ramsey fringes with a contrast given by @xmath61 . since @xmath62 this contrast serves as a measure of the stability of the quantum evolution under a small perturbation in the hamiltonian @xcite . in our experiments atoms
are not initially in a single motional eigenstate , but a thermal ensemble of atoms incoherently populates many eigenstates . the total population in @xmath9 is then given by an average of @xmath11 over the initial thermal ensemble .
@xmath58 depends on the initial state , and due to this spread in detuning the fringe contrast of the ensemble - averaged @xmath11 decays rapidly even when @xmath63 for all populated states ( note the analogy to the inhomogeneous broadening of two level atoms ) .
hence , the decay of the fringe contrast in ramsey spectroscopy of thermal trapped atoms does not teach us anything about the dynamics in the trap ( @xmath64 ) but simply measures the spread in @xmath58 .
this spread also limits the interrogation time for high precision spectroscopy .
equation [ hami ] indicates that if a special trap can be constructed so that @xmath58 is independent of @xmath52 , then the fast decay of the ramsey fringes will be suppressed .
an example of such a dephasing - free trap is an inverted pyramid of exponentially decaying evanescent waves , in which gravity provides the vertical confinement .
this trap have the special feature that @xmath65 is exactly equal to @xmath66 shifted along the vertical axis .
therefore all eigenenergies are shifted the same amount and @xmath58 is indeed a fixed number .
this makes such a trap an excellent candidate for high precision spectroscopy .
note , that in order to use traps for this kind of measurements , the suppression of the relative spread in @xmath58 , should be accompanied by a reduction in their average value , e.g. by increasing the trap s detuning . since the proposed trap is a `` dark optical trap '' , where atoms are trapped by repulsive light forces mainly in the dark @xcite , hence minimizing the amount of interaction with the trap s light , then in addition to being dephasing - free it has also a low photon scattering rate ( which will now be the limiting factor for the coherence time ) and a small @xmath58 .
apart from using special geometries a possible solution to the above mentioned dephasing is to reverse it using `` echo spectroscopy '' by adding a mw @xmath5-pulse between the two @xmath67 pulses . @xmath11 for atoms initially populating a single eigenstate then becomes : @xmath68 \nonumber \\ & = & \frac{1}{2}\left [ 1-% % tcimacro{\func{re}}% % beginexpansion \mathop{\rm re}% % endexpansion \left ( e^{i\frac{e_{n}^{1}}{\hbar } \tau } \left\langle \varphi _ { n}\left ( t=0\right ) \mid \varphi _ { n}\left ( t=\tau \right ) \right\rangle \right ) % \right ] \label{formel1}\end{aligned}\ ] ] where @xmath69 is a new basis defined by @xmath70 @xmath46 , and @xmath71 @xmath72 . @xmath11 no longer depends on @xmath56 and @xmath27 and therefore not on any @xmath58 but only on the dynamics is often a good approximation @xcite , but the limitations of it is not yet known to the authors . ] .
ensemble averaging over a thermal ensemble therefore does not remove the fingerprints of the dynamics . measured for three different trap laser wavelengths @xmath73 close to the @xmath74 transition wavelength of 795 nm .
lower curve : a `` small perturbation '' ( @xmath73=805 nm ) . a good echo signal ( @xmath75 )
is seen almost independent of time between pulses .
middle curve : a `` medium perturbation '' ( @xmath73=798.25 nm ) .
the echo signal oscillates around a value smaller than 1/2 with partial revivals at @xmath76 and @xmath77 , where @xmath78=3.6 ms is the measured trap oscillation frequency in the transverse direction .
upper curve : a `` strong perturbation '' ( @xmath73=796.25 nm ) .
after a short time the echo signal completely disappears ( @xmath79 ) , but partly revives again at @xmath80 ms , close to @xmath81.,width=288 ] we performed echo spectroscopy on @xmath82 @xmath14 atoms laser cooled to @xmath83 @xmath84 and then trapped in a @xmath85 @xmath84 deep gaussian optical trap with waist size of @xmath86 @xmath32 .
the procedure is detailed in @xcite .
the echo signal @xmath11 was measured as a function of time between pulses for different wavelengths of the trap laser , thereby changing the perturbation strength @xmath87 , while keeping the trap depth constant by adjusting the power of the trap beam .
for a small perturbation a good echo ( @xmath75 ) is seen independent of the time between pulses ( see fig . [ fi2 ] ) . for larger
perturbation damped oscillations to a level smaller than 1/2 are seen .
an even larger perturbation causes a complete decay of the echo coherence ( @xmath79 ) followed by partial revivals , which are clear signatures of the dynamics .
two revivals ( of different amplitude ) are seen when the time between mw pulses is close to @xmath88 and @xmath89 , respectively , where @xmath90 ms is the measured radial oscillation frequency of the trap ( the longitudinal oscillation frequency is @xmath91 @xmath92 , hence the longitudinal motion is essentially fixed for the duration of our experiment ) .
the revival at @xmath88 corresponds to wavepacket oscillations excited by the symmetric change of the trap strength , and the stronger revival at @xmath89 corresponds also to the vertical sloshing of wavepackets induced by the vertical shift of the gravito - optical potential for @xmath93 compared to @xmath0 ( in our experiments this vertical shift is of the order of a few @xmath94 , as compared to a trap cloud size of @xmath86 @xmath32 ! ) .
the lack of complete revival at @xmath89 is mainly due to anharmonicity of the gaussian trap , but also small amount of decoherence due to spontaneous scattering of photons and noise may contribute .
the sensitivity of our technique to map the quantum dynamics of a system is revealed here , since we see the dynamics due to a perturbation ( kick ) , that is about three orders of magnitude smaller than @xmath95 .
) .,width=288 ] in fig .
[ fi2 ] it is seen that for sufficient large detuning the long time asymptotic behavior of the echo signal from thermal atoms is smaller than 1/2 , and eventually becoming @xmath96 .
this can be understood from eq .
[ formel1 ] : if @xmath87 is small there will be , for each initial motional eigenstate @xmath97 of @xmath37 a large matrix element for going into the equivalent motional eigenstate of @xmath38 , since @xmath98 and @xmath38 are almost identical .
if the time between pulses is large enough so the dynamics has dephased completely between them , then the contributions from other states will average to zero .
hence , the long time ensemble average of @xmath11 is simply given by @xmath99 $ ] , where @xmath97 and @xmath100 are corresponding eigenstates of @xmath37 and @xmath38 respectively .
in @xcite a paraxial numerical calculation of @xmath101 , averaged over the @xmath102 thermally - populated states of the trap yielded good quantitative agreement with measured long time echo signal . as a function of perturbation strength @xmath103 for eigenstates with @xmath104 .
three different perturbations are shown for the stadium eigenstate of fig .
3 , namely dilation , stretching and the `` physical perturbation '' ( a dilation followed by a negative stretch ) .
the rectangle state has @xmath105
. for dilation substantial reduction of @xmath106 indeed occurs for @xmath107 , as expected from the intuitive discussion in the text .
, width=288 ] because of the importance of the eigenstate stability to the revival of coherence ( good echo signal ) and dynamics in the trap , we now calculate @xmath108 for perturbations , that we assume approximate our inherent perturbation , i.e. the difference in dipole potential for the two hyperfine states , for different traps .
note , that since a large @xmath108 also implies a large survival probability as defined in @xcite , it also plays an important role in quantum dynamical studies .
atom optic billiards are good candidates for trapping atoms for high precision spectroscopy , since the interaction with the trap light is small because atoms are confined mainly in the dark @xcite .
they have been used to study dynamics in the past , and have experimentally been demonstrated to display the features of ideal billiards , @xcite . for our calculations
we therefore approximate the atom optics billiard with an ideal hard - wall billiard , and use the powerful technique of vergini and saraceno to numerically calculate highly excited eigenstates and eigenenergies @xcite . for an unperturbed bunimovich stadium @xcite ( two semicircles of radius @xmath109 connected by two straight lines of length @xmath110 , for units defined such that @xmath111 ) a plot of the probability density of a highly excited state is shown in fig .
the size of its wave vector is @xmath112 ( @xmath113 -@xmath114 ) hence the wavefunction have on the order of few thousand antinodes . for three states of the stadium with @xmath104 as a function of perturbation strength .
( a ) stretching : a strong dependence on the initial state is seen .
( b ) dilation , for the same three stases as in ( a ) : only a weak dependence on the initial state with similar @xmath115 is seen.,width=432 ] systems respond very different to different types of perturbations @xcite .
perturbations of @xmath66 that can be described as a change of coordinate system such as rotation , translation ( of importance to the inverted pyramid discussed above ) and dilation will give exceptional high @xmath116 .
intuitively this is clear since an eigenstate of @xmath37 under these three perturbations is also an eigenstate of @xmath38 rotated , translated or dilated respectively .
therefore , to cause a significant drop in @xmath117 in case of dilation ( see fig . [ fi3 ] ) . on the other hand perturbations that changes
the symmetry of the billiard can cause much faster changes in @xmath108 .
this is illustrated in fig .
[ fi3 ] where it is seen that dilating a bunimovich stadium causes @xmath118 to drop much slower than changing the lengths of the straight sections ( this perturbation we call `` stretching '' ) .
we are interested in the perturbation induced in an atom - optics billiard by changing the strength of the optical potential .
this is best approximated by moving the billiard s wall a fixed distance perpendicular to itself .
this perturbation , that we name the physical perturbation , can be viewed as a dilation followed by a negative stretch , and it shows features between those of dilation and stretch ( fig . [ fi3 ] ) . ) between an initial state and its ten closest neighboring states with the same parity , as a function of stretching factor .
( b ) eigen @xmath115 s as a function of stretching factor , for the same states as ( a ) .
it is seen that when the state undergo avoided crossing it mixes strongly with the crossed state .
bold lines in ( a ) and ( b ) corresponds to the same states . , width=288 ] in fig [ fi3 ] @xmath108 for a rectangle is also shown in order to illustrate that the criteria stating that the dilation should move the wall of the billiard a distance of the order of the distance between nodes is valid not only for chaotic shapes but also for regular ones . for dilation in 1d systems this criteria is equivalent to the necessary criteria for substantial mixing from perturbation theory , namely that the perturbation energy should be comparable to the level separation . that this criteria also holds for dilating an integrable shape in 2d ( where it corresponds to a perturbation energy of @xmath119 times the level spacing in our case ) is also obvious since it is equivalent to two 1d systems .
but its fulfilment for a chaotic system where variables can not be separated is less obvious unless the above arguments are considered . as a function of dilation factor for eigenstates with @xmath104 in a stadium .
straight lines are observed as excepted.,width=288 ] figure [ fi5 ] shows latexmath:[$|\left\langle n^{\prime } = n|n\right\rangle with @xmath112 .
it is seen that @xmath116 is highly state dependent in case of stretching , but is not in case of dilation . following the explanation given above , it is clear that states with almost the same @xmath121 have almost the same @xmath118 for a given perturbation strength in case of dilation . the sudden drop in @xmath118 in case of stretching is often associated with avoided crossings of energy levels , after which the two states involved still hold the major part of the probability .
this is clearly seen from fig .
[ fi6 ] , where matrix elements to the ten closest states of same parity are shown for stretching . in fig .
[ fi6]b @xmath115 as a function of stretching factor is shown for the same ten states .
as seen , the avoided crossing of the two neighboring states ( drawn in bold lines ) occurs for the same stretch factor as the one causing strong mixing between them ( fig .
[ fi6 ] ) ) . in fig .
[ fi7 ] @xmath115 as a function of the dilation factor is presented , showing straight lines as expected .
matrix elements to close by states are 0 for dilation , in strong contrast to stretching shown in fig .
[ fi6]a @xcite .
we have analyzed echo spectroscopy of trapped ultra - cold atoms . using this technique
, we can combine the high precision of mw - spectroscopy with the high degree of control over experimental parameters provided by laser cooling and trapping techniques , to perform quantum dynamical studies even when many states are thermally populated . based on this analysis we proposed a gravito - optical trap
in which atoms in superposition states does not suffer from dephasing due to inhomogeneous stark shifts .
we analyzed numerically the stability of eigenstates for several perturbations , and verified that the rapid drop in @xmath116 seen for symmetry breaking perturbations is linked to avoided crossings of energy levels .
the symmetry of the effective perturbations , caused by mw transitions in our experimental atom optics billiards was shown to depend on their shape .
for example , rectangle or a circular billiard the physical perturbation corresponds approximately to pure dilation , thus enhancing the applicability of echo spectroscopy for these shapes , as compared to the bunimovich stadium billiard discussed above . finally , the suppression of dephasing achieved with echo spectroscopy yields a dramatic increase in the coherence time for trapped atoms , that may find important applications for precision spectroscopy and quantum information processing .
this work was in part supported by minerva foundation , the united states - israel binational science foundation , and foundation antorchas .
m. f. a. acknowledges help from dansk - israelsk studiefond til minde om josef og regine nachemsohn .
d. cohen , e. j. heller , unification of perturbation theory , random matrix theory , and semiclassical considerations in the study of parametrically dependent eigenstates , phys .
84 ( 13 ) ( 2000 ) 28412844 . | we discuss a recently demonstrated type of microwave spectroscopy of trapped ultra - cold atoms known as `` echo spectroscopy '' [ m.f .
andersen et .
al . , phys .
rev .
lett .
, in press ( 2002 ) ] .
echo spectroscopy can serve as an extremely sensitive experimental tool for investigating quantum dynamics of trapped atoms even when a large number of states are thermally populated .
we show numerical results for the stability of eigenstates of an atom - optics billiard of the bunimovich type , and discuss its behavior under different types of perturbations .
finally , we propose to use special geometrical constructions to make a dephasing free dipole trap . |
compelling evidences from type ia supernovae ( sn ia ) and other cosmological observations show that the expansion of the universe is undergoing an accelerating stage at late times ( see ref . for reviews ) . within the framework of general relativity and assuming homogeneity and isotropy
, this indicates that there should exist an energy source , termed dark energy , which provides a significant negative pressure to cause this acceleration . thus far
the nature of dark energy remains unresolved and is generally regarded as one of the most tantalizing problems in cosmology . while a positive cosmological constant remains the simplest realization of dark energy , current observational data have not ruled out the possibility of a time - evolving dark energy.@xcite as one awaits more information from the future observations , the constraining power of the next - generation observations and new analysis methods are being pursued.@xcite@xmath2@xcite many dark energy models have been proposed and studied . for the cosmological constant , its value has been constrained by observations ( see ref . , for example ) .
the quintessence model , which invokes a time - varying scalar field,@xcite@xmath2@xcite generally allows its energy density and equation of state to evolve with time .
there are various quintessence models with different potential forms ( see , refs .
@xmath2 , , for example ) that have been proposed .
studies of the classification@xcite and the general dynamical behavior@xcite of quintessence have been carried out .
there have been also works on the reconstruction of quintessence potentials@xcite@xmath2@xcite@xmath3@xcite and the investigations on how future observational data can constrain individual models of quintessence@xcite@xmath2@xcite .
the generalized chaplygin gas ( see ref . and references therein ) has been proposed to either unify dark matter and dark energy@xcite or to simply play the role of dark energy@xcite .
the constraint of the generalized chaplygin gas has been obtained.@xcite@xmath2@xcite while these works have helped us study the possible nature of dark energy and obtain constraints of the parameters of an individual dark energy model , it should be desirable to explore possible means to determine whether a particular dark energy model can be ruled out by the observational data .
recently , we introduced a new approach to testing the consistency between a dark energy model and the observational data.@xcite we tested the exponential ( see ref . and references therein ) and the power - law potentials@xcite of quintessence to demonstrate this method , which can be summarized as follows .
for each dark energy model , we look for a _ characteristic _ , @xmath4 , which in general can vary with the redshift but is equivalent to a constant parameter within the domain of the model .
we further define the _ measure of consistency _
, @xmath5 , as the derivative of @xmath4 with respect to the redshift @xmath6 .
the observational data should allow a null value for @xmath5 if the corresponding dark energy model is consistent with them .
if , however , the @xmath5 = 0 line lies outside certain confidence region , then that dark energy model is ruled out by the observational data at the corresponding confidence level . to obtain the constraint on the measure of consistency @xmath5 from the observational data , a parametrization of the relevant physical quantity , such as the equation of state or the luminosity distance , is required .
we have invoked a broadly used form of parametrization of the equation of state,@xcite@xmath3@xcite @xmath7 where @xmath8 and @xmath9 are the pressure and the energy density of dark energy , respectively .
the two parameters in eq .
( [ eq : w0wa - parametrization ] ) and the normalized matter density at present , @xmath10 , define the parameter space , @xmath11 , through which the information from the observational data can be extracted .
a recent work that is close in spirit to ours is that of zunckel and clarkson,@xcite who proposed consistency test of the cosmological constant via a direct parametrization of the luminosity distance . in our terminology
, they use @xmath12 , which is equivalent to the constant @xmath13 in the domain of the cosmological constant , where @xmath14 is the critical density at present .
sahni et al.@xcite also proposed null test of the cosmological constant via the diagnostic @xmath15 $ ] , which is equivalent to the constant @xmath13 in the domain of the cosmological constant , where @xmath16 is the hubble expansion rate and @xmath17 is the hubble constant . in this paper
, we take the more recent data set and apply our method of consistency test to five dark energy models , including the cosmological constant , the generalized chaplygin gas as the dark energy component,@xcite and three quintessence models : exponential , power - law and inverse - exponential potentials@xcite . the data set we use
includes a recently compiled `` constitution set '' of sn ia data,@xcite@xmath2@xcite the cosmic microwave background ( cmb ) measurement from the five - year wilkinson microwave anisotropy probe ( wmap ) observation,@xcite and the baryon acoustic oscillation ( bao ) measurement from the sloan digital sky survey ( sdss)@xcite and the @xmath18df galaxy redshift survey ( @xmath18dfgrs)@xcite . a conventional way to determine
how well a dark energy model can fit the observational data is the _ model - based approach _ , in which one optimizes the parameters of each dark energy model based on the observational data and then statistically assesses the goodness of fit ( see ref . , for example ) . in such approach one has to obtain the best fit for each set of parameters specific to the particular dark energy model , which could be tedious . in particular , in order to optimize the parameters of a quintessence model one has to solve the field equation numerically for each point in the parameter space , which can be computationally intensive and time consuming.@xcite in contrast , in our approach we first constrain the parameters of the chosen parametrization through the observational data , and then test consistency of each dark energy model based on this set of parameters
. this can be more efficient than the model - based approach when one deals with a large number of dark energy models .
it is also more direct and therefore much faster to constrain the parameter space @xmath11 than to optimize the parameters of a quintessence model in the model - based approach .
the potential downside of our method , however , would be that as long as one invokes a specific form of parametrization , one might have simultaneously imposed a prior , or bias , against certain dark energy models .
this issue requires a separate investigation and we are currently pursuing that.@xcite we note that the two methods are different in spirit .
the goodness of fit describes how well a model can fit the observations .
the consistency test , on the other hand , examines whether the condition necessary for a model is excluded by the observations . with in mind
the pros and cons mentioned above , we believe that the two methods , that is , the model - based and ours , should be complimentary to each other in the pursuit of revealing the nature of dark energy .
in this paper we consider a flat friedmann - lemaitre - robertson - walker ( flrw ) universe and assume that it is dominated by pressureless matter and dark energy in the present epoch . the hubble expansion rate , @xmath19 , is given by the friedmann equations as @xmath20 \nonumber \\
& = & h_0 ^ 2 \left [ \omega_{m } ( 1+z)^3 + ( 1-\omega_{m } ) \exp \left ( 3 \int_{0}^{z } \left [ 1 + w(z ' ) \right ] \frac{dz'}{1+z ' } \right ) \right ] , \label{eq : h2(z)}\end{aligned}\ ] ] where the dark energy density @xmath21 \frac{dz'}{1+z ' } \right ) , \label{eq : rho phi ( z)}\ ] ] and @xmath22 for quintessence as a dark energy model , the quintessence field and the potential are related to the equation of state , the hubble expansion rate , and the dark energy density as follows @xcite .
@xmath23\rho_\textsc{de}(z')}}{h(z ' ) } \frac{dz'}{1+z ' } \ , , \label{eq : phi(z)}\ ] ] @xmath24 \rho_\textsc{de}(z ) / 2 \ , .
\label{eq : v(z)}\ ] ] we perform the consistency test of five dark energy models including the cosmological constant , the exponential potential , the power - law potential , the inverse - exponential potential , and the generalized chaplygin gas as the dark energy component . for the cosmological constant ,
the energy density @xmath25 is a constant .
we define the characteristic @xmath26 as the dark energy density @xmath9 , which in general would evolve with the redshift but is equivalent to the constant parameter @xmath25 within the cosmological constant domain , @xmath27 in the same spirit , for the exponential potential , @xmath28 \ , , \ ] ] we identify _
@xmath29 _ as the characteristic constant parameter and accordingly define the characteristic @xmath30 , @xmath31 for the power - law potential , @xmath32 we define the following characteristic corresponding to the index @xmath33 , @xmath34^{-1 } \label{eq : qpower1 } \\ & = & n \quad \textrm{for the power - law potential.}\label{eq : qpower2}\end{aligned}\ ] ] for the inverse - exponential potential , @xmath35 \,,\ ] ] the characteristic is defined as @xmath36^{-2}\label{eq : qinverse - exp1 } \\ & = & m_2 \quad \textrm{for the inverse - exponential potential.}\label{eq : qinverse - exp2}\end{aligned}\ ] ] as the dark energy component , the generalized chaplygin gas has an equation of state govern by @xmath37^{\alpha } \
, , \ ] ] where @xmath38 and @xmath39 .
the corresponding characteristic is defined as @xmath40 we then define the measure of consistency @xmath41 as the derivative of the characteristic @xmath42 with respect to the redshift for each dark energy model , @xmath43 where @xmath44 denotes `` @xmath0 '' , `` exp '' , `` power '' , `` inverseexp '' and `` chaplygin '' , respectively .
@xmath41 can in general evolve with the redshift but should be constant zero in the domain of the corresponding dark energy model .
we use the combined data set from three types of observations including the sn ia observation , the cmb measurement , and the bao measurement .
we use the constitution set of sn ia data compiled by hicken et al.,@xcite@xmath3@xcite@xmath2@xcite which provides the information of the luminosity distance and the redshift .
the luminosity distance - redshift relation is given by @xmath45 we use the cmb shift parameter measured by the five - year wmap observation,@xcite @xmath46 we use the bao measurement from the joint analysis of the sdss and @xmath18dfgrs data,@xcite which gives @xmath47 where @xmath48^{1/3},\ ] ] and @xmath49 is the angular diameter distance , @xmath50 the constraint of the parameter space ( @xmath51 , @xmath52 , @xmath13 ) is obtained by fitting the three parameters to this combined data set .
the best fit of the parameters are found to be @xmath53 the two - dimensional constraint of @xmath54 is shown in fig .
[ fig1 ] . based on a combined data set including the constitution set of sn ia data , the cmb measurement from the five - year wmap , and the bao measurement from the sdss and @xmath18dfgrs .
the dark and light gray areas correspond to the @xmath55 and @xmath1 confidence regions , respectively .
[ fig1 ] ] for the consistency test of each dark energy model , we reconstruct @xmath41 via the constraint of ( @xmath51 , @xmath52 , @xmath13 ) , with the use of the equations in sec .
[ sec : introduction ] and sec .
[ sec : formalism ] .
we perform the test in the redshift region @xmath56 , where the influence of dark energy on the expansion of the universe is most significant .
this region is covered by the current sn ia observations , which is the most sensitive type of observations to probe the behavior of dark energy .
if the @xmath57 line lies outside certain confidence region , the corresponding dark energy model is ruled out at that confidence level . adopting the constraint obtained in sec .
[ sec : ii - b ] , we find that the @xmath58 line lies outside the @xmath1 confidence region while the null lines of the other four models lie inside the @xmath55 confidence region .
the results are shown in fig .
[ fig2 ] . and @xmath1 confidence regions , respectively .
the @xmath58 line lies outside the @xmath1 confidence region for @xmath59 .
the null lines of the measure for the other four models lie inside the @xmath55 confidence regions for @xmath56 .
this indicates that the exponential potential is ruled out at the @xmath1 confidence level while the other four dark energy models are still consistent with the current observational constraints down to the @xmath55 confidence level.[fig2 ] ]
we have preformed consistency test of five dark energy models , including the cosmological constant , the generalized chaplyngin gas , and three quintessence models : exponential , power - law , and inverse - exponential potentials .
this test gives a simple signature if a dark energy model is ruled out by the observational data .
it can be done efficiently via the constraint of a single set of parameters deduced from the observational data , and can test quintessence models without solving the field equation . through our approach and invoking the broadly used parametrization of the equation of state , the exponential potential is found to be ruled out at the @xmath1 confidence level based on the current observational data .
the other four dark energy models remain consistent with the current observations down to the @xmath55 confidence level .
it is worth noticing that in our previous work the power - law potential was ruled out at the @xmath55 confidence level based on a different data set @xcite .
one noticeable change in the new data set is that the cosmological constant is contained in the @xmath55 confidence region , which was not so in the previous one . whether our method can discriminate between the power - law potential and the cosmological constant can in principle be studied with the monte carlo test @xcite .
the flat cosmological constant model was examined by davis et al @xcite . via the model - based approach , in which they used the same data
set except the three - year wmap data instead .
they found the goodness of fit for the flat cosmological constant model to be @xmath60 while we find the model at least consistent with the data at the @xmath55 confidence level .
the two results are not in conflict with each other .
our method of the consistency test can in principle be applied not only to other dark energy models but also to other models explaining the acceleration of the expansion , as long as one looks for the characteristic @xmath4 corresponding to a constant parameter of each model .
one can also choose a different parametrization for better discriminating power between the models in regard .
the discriminating power of the method with different forms of parametrization and the possible bias imposed by the chosen parametrization should be studied via the monte carlo test @xcite .
chen is supported by the taiwan national science council under project no .
nsc 95 - 2119-m-002 - 034 and nsc 96 - 2112-m-002 - 023-my3 , and p. chen by taiwan national science council under project no .
nsc 97 - 2112-m-002 - 026-my3 and by us department of energy under contract no .
de - ac03 - 76sf00515 .
a. albrecht _ et al .
_ , arxiv : astro - ph/0609591 . m. barnard , a. abrahamse , a. albrecht , b. bozek and m. yashar , _ phys .
_ d * 77 * , 103502 ( 2008 ) [ arxiv:0712.2875 [ astro - ph ] ] .
a. abrahamse , a. albrecht , m. barnard and b. bozek , _ phys .
_ d * 77 * , 103503 ( 2008 ) [ arxiv:0712.2879 [ astro - ph ] ] .
b. bozek , a. abrahamse , a. albrecht and m. barnard , _ phys .
_ d * 77 * , 103504 ( 2008 ) [ arxiv:0712.2884 [ astro - ph ] ] .
m. barnard , a. abrahamse , a. j. albrecht , b. bozek and m. yashar , _ phys .
_ d * 78 * , 043528 ( 2008 ) [ arxiv:0804.0413 [ astro - ph ] ] .
y. wang and p. mukherjee , _ phys .
_ d * 76 * , 103533 ( 2007 ) [ arxiv : astro - ph/0703780 ] .
je - an gu , c .- w . chen and p. chen , _ new j. phys . _ * 11 * 073029 , ( 2009 ) [ arxiv:0803.4504 [ astro - ph ] ] c. zunckel and c. clarkson , _ phys . rev
. lett . _ * 101 * , 181301 ( 2008 ) [ arxiv:0807.4304 [ astro - ph ] ] .
e. komatsu _ et al .
_ [ wmap collaboration ] , _ astrophys .
j. suppl . _
* 180 * , 330 ( 2009 ) [ arxiv:0803.0547 [ astro - ph ] ] .
r. r. caldwell , r. dave and p. j. steinhardt , _ phys . rev .
* 80 * , 1582 ( 1998 ) [ arxiv : astro - ph/9708069 ] .
je - an gu and w - y .
p. hwang , _ phys .
_ b * 517 * , 1 ( 2001 ) [ arxiv : astro - ph/0105099 ] .
l. a. boyle , r. r. caldwell and m. kamionkowski , _ phys .
_ b * 545 * , 17 ( 2002 ) [ arxiv : astro - ph/0105318 ] . j. a. frieman , c. t. hill , a. stebbins and i. waga , _ phys . rev .
lett . _ * 75 * , 2077 ( 1995 ) [ arxiv : astro - ph/9505060 ] .
a. j. albrecht , c. p. burgess , f. ravndal and c. skordis , _ phys .
_ d * 65 * , 123507 ( 2002 ) [ arxiv : astro - ph/0107573 ] .
i. zlatev , l. wang and p. j. steinhardt , _ phys .
lett . _ * 82 * , 896 ( 1999 ) [ arxiv : astro - ph/9807002 ] ; p. j. steinhardt , l. wang and i. zlatev , _ phys .
_ d * 59 * , 123504 ( 1999 ) [ arxiv : astro - ph/9812313 ] .
r. r. caldwell and e. v. linder , _ phys .
lett . _ * 95 * , 141301 ( 2005 ) [ arxiv : astro - ph/0505494 ] .
e. v. linder , _ phys .
_ d * 73 * , 063010 ( 2006 ) [ arxiv : astro - ph/0601052 ] .
d. huterer and h. v. peiris , _ phys .
_ d * 75 * , 083503 ( 2007 ) [ arxiv : astro - ph/0610427 ] .
d. huterer and m. s. turner , _ phys .
_ d * 60 * , 081301 ( 1999 ) [ arxiv : astro - ph/9808133 ] .
a. a. starobinsky , _ jetp lett .
_ * 68 * , 757 ( 1998 ) [ _ pisma zh .
fiz . _ * 68 * , 721 ( 1998 ) ] [ arxiv : astro - ph/9810431 ] .
t. nakamura and t. chiba , _ mon . not .
soc . _ * 306 * , 696 ( 1999 ) [ arxiv : astro - ph/9810447 ] .
t. d. saini , s. raychaudhury , v. sahni and a. a. starobinsky , _ phys .
lett . _ * 85 * , 1162 ( 2000 ) [ arxiv : astro - ph/9910231 ] .
b. f. gerke and g. efstathiou , _ mon . not .
soc . _ * 335 * , 33 ( 2002 ) [ arxiv : astro - ph/0201336 ] . v. sahni and a. starobinsky , _ int .
j. mod .
_ d * 15 * , 2105 ( 2006 ) [ arxiv : astro - ph/0610026 ] .
m. sahlen , a. r. liddle and d. parkinson , _ phys .
_ d * 75 * , 023502 ( 2007 ) [ arxiv : astro - ph/0610812 ] .
m. c. bento , o. bertolami and a. a. sen , _ phys .
_ d * 66 * , 043507 ( 2002 ) [ arxiv : gr - qc/0202064 ]
. n. bilic , g. b. tupper and r. d. viollier , _ phys .
_ b * 535 * , 17 ( 2002 ) [ arxiv : astro - ph/0111325 ] .
a. y. kamenshchik , u. moschella and v. pasquier , _ phys .
_ b * 511 * , 265 ( 2001 ) [ arxiv : gr - qc/0103004 ] .
a. a. sen and r. j. scherrer , _ phys .
_ d * 72 * , 063511 ( 2005 ) [ arxiv : astro - ph/0507717 ] .
m. makler , s. quinet de oliveira and i. waga , _ phys .
_ b * 555 * , 1 ( 2003 ) [ arxiv : astro - ph/0209486 ] .
t. m. davis _ et al .
_ , _ astrophys .
j. _ * 666 * , 716 ( 2007 ) [ arxiv : astro - ph/0701510 ] .
m. chevallier and d. polarski , _ int .
j. mod .
_ d * 10 * , 213 ( 2001 ) [ arxiv : gr - qc/0009008 ] .
e. v. linder , _ phys .
lett . _ * 90 * , 091301 ( 2003 ) [ arxiv : astro - ph/0208512 ] .
a. g. riess _ et al .
_ [ supernova search team collaboration ] , _ astrophys .
j. _ * 607 * , 665 ( 2004 ) [ arxiv : astro - ph/0402512 ] .
p. astier _ et al .
_ [ the snls collaboration ] , _ astron . astrophys . _ * 447 * , 31 ( 2006 ) [ arxiv : astro - ph/0510447 ] . a. g. riess _ et al .
_ , _ astrophys .
j. _ * 659 * , 98 ( 2007 ) [ arxiv : astro - ph/0611572 ] .
g. miknaitis _ et al .
_ , _ astrophys
. j. _ * 666 * , 674 ( 2007 ) [ arxiv : astro - ph/0701043 ] . w. m. wood - vasey _ et al .
_ [ essence collaboration ] , _ astrophys .
j. _ * 666 * , 694 ( 2007 ) [ arxiv : astro - ph/0701041 ]
. m. kowalski _ et al .
_ , _ astrophys .
j. _ * 686 * , 749 ( 2008 ) [ arxiv:0804.4142 [ astro - ph ] ] .
m. hicken _ et al .
_ , arxiv:0901.4804 [ astro-ph.co ] .
d. j. eisenstein _ et al .
_ [ sdss collaboration ] , _ astrophys . j. _ * 633 * , 560 ( 2005 ) [ arxiv : astro - ph/0501171 ] .
w. j. percival , s. cole , d. j. eisenstein , r. c. nichol , j. a. peacock , a. c. pope and a. s. szalay , _ mon . not .
_ * 381 * , 1053 ( 2007 ) [ arxiv:0705.3323 [ astro - ph ] ] .
chen , je - an gu and p. chen ( in preparation ) . | recently we proposed a new approach to the testing of dark energy models based on the observational data . in that work we focused particularly on quintessence models for demonstration and invoked a widely used parametrization of the dark energy equation of state . in this paper
we take the more recent sn ia , cmb and bao data , invoke the same parametrization , and apply this method of consistency test to five categories of dark energy models , including the @xmath0cdm model , the generalized chaplygin gas , and three quintessence models : exponential , power - law and inverse - exponential potentials .
we find that the exponential potential of quintessence is ruled out at the @xmath1 confidence level , while the other four models are consistent with data .
this consistency test can be efficiently performed since for all models it requires the constraint of only a single parameter space that by choice can be easily accessed . |
suppose @xmath6 particles are placed on vertices of the @xmath0-path , with no site multiply occupied .
the _ biased exclusion process _
is the markov chain @xmath7 with transitions as follows : * choose uniformly among the @xmath8 edges of the path , * if both vertices of the selected edge are either occupied or unoccupied , do nothing , * if there is exactly one particle on the edge , place it on the right vertex with probability @xmath9 and on the left with probability @xmath10 .
the canonical case is when @xmath0 is even and @xmath11 .
this defines a reversible ergodic markov chain , which has a unique stationary distribution @xmath12 .
it is natural to ask about its _ mixing time _
, @xmath13 we write @xmath14 for @xmath15 . when @xmath16 , proved @xmath17 n^3 \log ( n/{\varepsilon } ) \,,\ ] ] and conjectured that the lower bound is sharp .
recently , lacoin answered this , proving that the process has a _ cutoff _ , i.e. @xmath18 it is worth observing that the eigenfunction lower bound method introduced in wilson turns out to be widely applicable , giving sharp lower bounds for many models . when @xmath19 , the mixing time was first studied by , who proved @xmath20 .
a simpler path coupling proof was given by .
( this proof is repeated here as the upper bound in theorem [ thm : mixbe ] . )
the purpose of this paper is to understand the mixing behavior when the bias may depend on @xmath0 and in particular when @xmath21 as @xmath3 .
we show that in all cases , there is a _ pre - cutoff _ , meaning that there are universal constants @xmath22 so that @xmath23 we find that , depending on the rate at which @xmath24 , the mixing time interpolates between the unbiased and constant bias cases .
below summarizes our results .
we write @xmath25 to mean that there exist constant @xmath26 , not depending on @xmath27 , so that @xmath28 .
[ thm : main ] consider the @xmath27-biased exclusion process on @xmath29 with @xmath6 particles .
we assume that @xmath30 .
if @xmath31 , then @xmath32 if @xmath33 , then @xmath34 if @xmath35 , then @xmath36 we provide more precise estimates on @xmath37 in proposition [ prop : lb1 ] , proposition [ prop : lb2 ] , and theorem [ thm : mixbe ] . in particular , the lower bound in follows from proposition [ prop : lb1 ] , the lower bound in follows from proposition [ prop : lb2 ] , and the lower bound in follows from proposition [ prop : lbu ] .
the upper bounds in and follow from theorem [ thm : mixbe ] , and the upper bound in follows from proposition [ prop : nearun ] .
since the behavior of the individual particles remains diffusive in the @xmath38 regime , it is not surprising that the mixing time has the same order as the unbiased process in this case .
the change of the functional form of the mixing time at @xmath39 is a more unexpected transition .
a path coupling gives useful upper bounds for @xmath40 .
when @xmath41 is small , we use a simple coupling adapted from a coupling for ( unbiased ) random adjacent transpositions given in . in the unbiased case ,
@xmath6 coupled unbiased random walks must hit zero .
the bias introduced when @xmath41 is small does nt overwhelm the diffusive motion , so the same idea works . for lower bounds , when @xmath42 , we use wilson s method ( introduced in ) . thus we need the eigenfunction corresponding to the second eigenvalue , which we explicitly compute .
when @xmath43 , we follow the left - most particle , and show it needs at least order @xmath44 moves to mix .
the organization of the paper is as follows . after giving definitions in section [ sec : defn ] , in section [ sec : spectrallb ]
we compute the eigenfunction needed for wilson s method , and provide the corresponding lower bounds .
in particular , the lower bounds in theorem [ thm : main ] ( i ) and ( ii ) are given in propositions [ prop : lb1 ] and [ prop : lb2 ] , respectively .
we give the two upper bounds in section [ sec : ub ] : the upper bound in is given in proposition [ prop : nearun ] , and the other upper bounds in theorem [ thm : main ] are all immediate from theorem [ thm : mixbe ] .
we conclude with the single particle lower bound needed for theorem [ thm : main ] ( iii ) in section [ sec : lbu ] .
it will sometime be convenient to use a bijection of the state - space @xmath45 of the particle process to the space of nearest - neighbor paths of length @xmath0 which begin at @xmath2 and have exactly @xmath6 _ up _ increments and @xmath46 _ down _ increments . for a particle configuration @xmath47 , let @xmath48 be defined by @xmath49 , and @xmath50 so occupied sites correspond to increments and vacant sites correspond to decrements of the path .
see figure [ fig : exclmov ] for an illustration . .
node @xmath51 of the path is updated in configuration @xmath52 to obtain @xmath53 .
this corresponds to exchanging the particle at vertex @xmath51 with the hole at vertex @xmath54 .
[ fig : exclmov ] ] the dynamics on the path are as follows : pick among the @xmath8 interval vertices of the path .
if the path is a local extremum , refresh it with a local maximum with probability @xmath55 , and a local minimum with probability @xmath56 .
if the chosen vertex is not an extremum , do nothing .
see again figure [ fig : exclmov ] for an illustration of a transition , and figure [ fig : biasedtrans ] for the possible transitions from a particular path . ]
it will be convenient to move back and forth from the particle description and the path description , and we will freely do so .
here we set @xmath57 ; our assumption is always that @xmath58 .
[ prop : eigenfunction ] let @xmath59 . the function @xmath60 , defined for the path @xmath61 as @xmath62 is the second eigenfunction for the biased exclusion process , with eigenvalue @xmath63 we let @xmath64 ; note our convention is @xmath65 . for a path @xmath61 and vertex @xmath66 ,
let @xmath67 be the number of up - edges before @xmath68 .
we have @xmath69 and @xmath70 .
define @xmath71 for @xmath72 .
let @xmath73 be the path obtained by applying an update to @xmath61 at internal vertex @xmath68 .
then @xmath74 = q g^\star_{h}(i-1))+ p g^\star_{h}(i+1 ) \,.\ ] ] consider the case where @xmath68 is a local extremum in @xmath61 . if the path at @xmath68 is refreshed to a local maximum , then @xmath75 , while if the path is refreshed to a local minimum , then @xmath76 . therefore , @xmath77 = q \theta^{i - ( f_h(i-1 ) + 1 ) } + p
\theta^{i - ( f_h(i+1 ) -1 ) } = q g^\star_h(i-1 ) + p g^\star_h(i+1 ) \,.\ ] ] in the case where @xmath78 , the update at @xmath68 must leave the path unchanged . in this case ,
@xmath79 and @xmath80 .
therefore , @xmath81 \,.\ ] ] finally , suppose @xmath82 ; again , the update at @xmath68 does not change the path .
since @xmath83 in this case , @xmath84 for any constant @xmath85 , the function @xmath86 also satisfies @xmath87 = q g_h(i-1 ) + p g_h(i+1 ) \,.\ ] ] define @xmath88 and let @xmath89 define @xmath90 let @xmath91 satisfy @xmath92 that is , @xmath93 is the eigenfunction for the @xmath94 random walk on @xmath95 with absorbing states @xmath2 and @xmath0 .
a direct verification shows that @xmath96 is a solution .
note that @xmath97 \theta^{-1/2 } q \sin(\pi / n ) \\ & \quad + [ \theta^{n - k } - c ] \theta^{-n/2}\theta^{1/2}p \sin(\pi - \pi / n ) \nonumber \\ & = \sqrt{pq } \sin(\pi / n)[1 + \theta^{n/2-k } - c[1 + \theta^{-n/2 } ] ] \nonumber \\ & = 0 \ , .
\nonumber \end{aligned}\ ] ] define @xmath98 let @xmath99 be the configuration obtained after one step of the chain when started from @xmath61 ; as before let @xmath100 be the update given that internal vertex @xmath52 is selected for an update .
@xmath101 & = \sum_{x=1}^{n-1 } { { \mathbf e}}_h [ g_{\tilde{h}}(x ) ] \phi(x)\\ & = \sum_{x=1}^{n-1 } \bigr [ \bigl(1 - \frac{1}{n-1}\bigr ) g_h(x ) + \frac{1}{n-1}{{\mathbf e}}_h [ g_{\tilde{h}^{(x ) } } ] \bigr ] \phi(x ) \\ & = \bigl ( 1 - \frac{1}{n-1 } \bigr ) \phi(h ) + \frac{1}{n-1}\sum_{x=1}^{n-1 } [ q g_h(x-1 ) + p g_h(x+1 ) ] \phi(x ) \end{aligned}\ ] ] the sum on the right equals @xmath102 + [ g_h(0 ) \phi(1 ) q + g_h(n ) \phi(n-1 ) p ] \\ = \lambda \sum_{x=1}^{n-1 } g_h(x ) \phi(x ) = \lambda \phi(h ) \ , , \end{gathered}\ ] ] by . therefore ,
@xmath103 = \bigl ( 1 - \frac{1 - \lambda}{n-1 } \bigr ) \phi(h)\ ] ] note that @xmath104 for @xmath105 , and @xmath106 is increasing in @xmath61 , so @xmath60 is increasing .
an increasing eigenfunction always corresponds to the second eigenvalue , so it must be the one with largest ( non unity ) eigenvalue .
the second largest eigenvalue equals @xmath63 note that @xmath107 , so we have @xmath108 \theta^{-x/2 } \sin(\pi x / n ) \\ & = \sum_{x=1}^{n-1 } \bigl [ \alpha^{h(x ) } - \theta^{(n - x)/2 } \frac{1 + \theta^{n/2-k}}{1 + \theta^{n/2 } } \bigr ] \sin(\pi x / n ) \\ & = \sum_{x=1}^{n-1 } \alpha^{h(x ) } \sin(\pi x / n ) - \xi(n , k,\alpha)\ , . \end{aligned}\ ] ] let @xmath109 since @xmath110 does not depend on @xmath61 , and the eigenfunction @xmath60 must be orthogonal to the constants , it follows that @xmath111 . since @xmath112 , @xmath113 to apply wilson s lower bound
, we need to bound @xmath114 from below , and @xmath115 from above .
define @xmath116 for @xmath117 defined in , @xmath118 using that @xmath119 , we pair together the terms at @xmath52 and @xmath120 in so that @xmath121 the first sum simplifies to @xmath122 and the second to @xmath123 [ lem : efrb ] let @xmath117 be as in , and for a path @xmath61 , let @xmath99 be one step of the exclusion chain started from @xmath61 .
let @xmath124 be the spectral gap .
define @xmath125 if @xmath126 , then @xmath127 \log
n \,.\ ] ] fix @xmath128 . from , @xmath129 if @xmath99 is obtained by a single update to @xmath61 at @xmath52 , the @xmath130 , and @xmath131 thus , if @xmath132 , then @xmath133 letting @xmath134 so that @xmath135 , equations and show that @xmath136 ^ 2 \,.\ ] ] the spectral gap @xmath137 satisfies @xmath138 suppose that @xmath139 .
then from and we have @xmath140\log n \,.\ ] ] if @xmath141 , where @xmath142 , then @xmath143 ^ 2 & \zeta > 0 \\ c_0 \left ( \frac{3\rho^2}{4 } \right)^2 & \zeta = 0 \end{cases } \,.\ ] ] the right - hand side is bounded below for @xmath142 , so we conclude that @xmath127 \log n \,.\ ] ] [ prop : lb1 ] if @xmath144 where @xmath145 , then @xmath146 \bigl(\log n + \log[(1-{\varepsilon})/{\varepsilon } ) ] \bigr ) \,.\ ] ] from , the spectral gap @xmath137 satisfies @xmath147 \,.\ ] ] using lemma [ lem : efrb ] in ( see also theorem 13.5 of for a discussion ) yields @xmath148 \label{eq : wilsonlb } \\ & = \frac{n^3 } { ( \pi^2 + \zeta^2)}[1 + o(1)]\bigl(\log n + \log[(1-{\varepsilon})/{\varepsilon } ] \bigr)\ , , \nonumber \end{aligned}\ ] ] which yields . note that this matches the lower bound in theorem 4 of wilson ( 2004 ) for the symmetric exclusion when @xmath149 .
[ prop : lb2 ] if @xmath150 but @xmath151 , then @xmath152(\log n + \log[(1-{\varepsilon})/{\varepsilon } ] ) \,.\ ] ] this again follows from , and lemma [ lem : efrb ] .
[ prop : nearun ] there exists a constant @xmath153 such that if @xmath154 , then @xmath155 we now define a markov chain @xmath156 so that * @xmath157 and @xmath158 are _ labelled _ @xmath6-particle configurations , * if the labels are erased , @xmath159 and @xmath160 each are biased exclusion processes .
we say a labelled particle is _ coupled _ at time @xmath161 if it occupies the same vertex in both @xmath157 and @xmath158 . we now describe a move of this chain from state @xmath162 :
pick an edge @xmath163 among the @xmath8 edges uniformly at random .
we consider several cases . * _ both @xmath164 and @xmath165 have no particles on @xmath163 .
_ the chain remains at @xmath162 .
* _ one of @xmath166 contains two particles on @xmath163 , and one of @xmath166 contains one particle on @xmath163 . _
suppose , without loss of generality , that @xmath164 contains one particle on @xmath163 .
toss a @xmath56-coin to determine where the particle is placed in @xmath164 .
if the single particle on @xmath163 in @xmath164 is coupled , or has the same label as one of the particles on @xmath163 in @xmath165 , arrange the two particles on @xmath163 in @xmath165 to preserve or facilitate the coupling . otherwise , toss a fair coin to determine the placement of the two particles in @xmath165 . *
_ both @xmath164 and @xmath165 have two particles on @xmath163_. toss a fair coin to determine the placement of the two particles on @xmath163 in @xmath164 .
place the particles in @xmath165 on @xmath163 to preserve or facilitate any couplings ; if no coupling is possible , toss a fair coin to determine the particle placement on @xmath163 .
the distance @xmath167 between particle @xmath68 in @xmath164 and particle @xmath68 in @xmath165 performs a delayed nearest - neighbor walk , with possible bias @xmath27 at each move ( sometimes the bias is to the right , sometimes to the left ) .
the probability it moves is at least @xmath168 .
we can thus couple it to a random walk @xmath169 with constant upward bias @xmath27 so that @xmath170 until @xmath167 hits zero .
consider the biased random walk @xmath169 on @xmath171 with positive bias @xmath27 , holding probability @xmath172 , and @xmath173 ; if @xmath174 then @xmath175 we have @xmath176 where @xmath177 . by the central limit theorem , since @xmath178 , there is a constant @xmath179 such that , for @xmath0 large enough , @xmath180 thus by taking @xmath153 large enough , @xmath181 if we run @xmath182 blocks of @xmath183 moves , then we have @xmath184 setting @xmath185 , @xmath186 if @xmath187 , then @xmath188 , and @xmath189 for @xmath0 large enough . and @xmath53 . [
fig : walkneighbors ] ] we consider configurations @xmath52 and @xmath53 to be adjacent if @xmath53 can be obtained from @xmath52 by taking a particle and moving it to an adjacent unoccupied site . in the path representation ,
moving a particle to the right corresponds to changing a local maximum ( i.e. , an `` up - down '' ) to a local minimum ( i.e. a `` down - up '' ) . moving a particle to the left changes a local minimum to a local maximum .
see figure [ fig : exclmov ] , where @xmath190 .
[ thm : mixbe ] consider the biased exclusion process with bias @xmath191 on the segment of length @xmath0 and with @xmath6 particles .
set @xmath192 . for @xmath193 ,
if @xmath0 is large enough , then @xmath194 \right ] \,.\ ] ] in particular , if @xmath195 , then @xmath196 , so @xmath197 -2 \log \beta + o(\beta ) \bigr ] \,.\ ] ] note that whenever @xmath198 for constants @xmath153 and @xmath199 , the ratio of the upper and lower bounds is bounded . thus there is a pre cut - off for this chain in this regime . for @xmath200 ,
define the distance between two configurations @xmath52 and @xmath53 which differ by a single transition to be @xmath201 where @xmath61 is the height of the midpoint of the diamond that is removed or added .
( see figure [ fig : walkneighbors ] . )
note that @xmath58 and @xmath202 guarantee that @xmath203 , so we can use path coupling see , e.g. , theorem 14.6 of .
we again let @xmath204 denote the path metric on @xmath205 corresponding to @xmath206 .
we couple from a pair of initial configurations @xmath52 and @xmath53 which differ at a single vertex @xmath207 as follows : choose the same vertex in both configurations , and propose a local maximum with probability @xmath208 and a local minimum with probability @xmath56 .
for both @xmath52 and @xmath53 , if the current vertex @xmath207 is a local extremum , refresh it with the proposed extremum ; otherwise , remain at the current state .
let @xmath209 be the state after one step of this coupling .
there are several cases to consider .
the first case is shown in figure [ fig : walkneighbors ] .
let @xmath52 be the upper configuration , and @xmath53 the lower . here
the edge between @xmath210 and @xmath211 is `` up '' , while the edge between @xmath212 and @xmath213 is `` down '' , in both @xmath52 and @xmath53 .
if @xmath207 is selected , the distance decreases by @xmath214 .
if either @xmath211 or @xmath212 is selected , and a local minimum is selected , then the lower configuration @xmath53 is changed , while the upper configuration @xmath52 remains unchanged .
thus the distance increases by @xmath215 in that case .
we conclude that @xmath216 - \rho(x , y ) & = -\frac{1}{n-1}\alpha^{h+n - k } + \frac{2}{n-1}p \alpha^{h+n - k -1 } \nonumber \\ & = \frac{\alpha^{h+n - k}}{n-1 } \left ( \frac{2p}{\alpha } - 1 \right ) = \frac{\alpha^{h+n - k}}{n-1}\left ( 2\sqrt{pq } - 1 \right ) \ , .
\label{eq : aepc1 } \end{aligned}\ ] ] in the case where @xmath52 and @xmath53 at @xmath217 are as in the right panel of figure [ fig : walkneighbors ] , we obtain @xmath216 - \rho(x , y ) & = -\frac{1}{n-1}\alpha^{h+n - k } + \frac{2}{n-1}(1-p ) \alpha^{h+n+1 } \nonumber \\ & = \frac{\alpha^{h+n - k}}{n-1}\left ( 2\alpha ( 1-p ) - 1 \right ) = \frac{\alpha^{h+n - k}}{n-1}\left(2 \sqrt{pq } - 1 \right ) \label{eq : aepc2 } \end{aligned}\ ] ] ( we create an additional disagreement at height @xmath218 if either @xmath211 or @xmath212 is selected and a local maximum is proposed ; the top configuration can accept the proposal , while the bottom one rejects it . ) since @xmath219 , we have @xmath220 , and both and reduce to @xmath221 - \rho(x , y ) = - \frac{\alpha^{h+n - k}}{n-1 } \delta\,.\ ] ] now consider the case on the left of figure [ fig : moreconfigs ] .
we have @xmath216 - \rho(x , y ) & = -\frac{1}{n-1}\alpha^{h+n - k } + \frac{1}{n-1}q \alpha^{h+n - k+1 } + \frac{1}{n-1}p \alpha^{h+n - k-1 } \nonumber \\ & = \frac{\alpha^{h+n - k}}{n-1}\left ( q\alpha + \frac{p}{\alpha } - 1 \right ) \nonumber \\ & = -\frac{\alpha^{h+n - k}}{n-1 } \delta \ , , \end{aligned}\ ] ] which gives again the same expected decrease as .
( in this case , a local max proposed at @xmath211 will be accepted only by the top configuration , and a local min proposed at @xmath212 will be accepted only by the bottom configuration . ) the case on the right of figure [ fig : moreconfigs ] is the same .
thus , holds in all cases .
that is , since @xmath222 , @xmath223 = \rho(x , y)\left(1 - \frac{\delta}{n-1 } \right ) \leq \rho(x , y ) e^{-\frac{\delta}{n-1 } } \,.\ ] ] the diameter of the state - space is the distance from the configuration with @xmath6 `` up '' edges followed by @xmath46 `` down '' edges to the configuration with @xmath46 `` down edges '' followed by @xmath6 `` up '' edges . to move from the former to the latter , first flip the top - most maxima , next the subsequent two maxima , continuing down @xmath224 levels
. at level @xmath225 , there are @xmath225 maxima to flip .
each of the next @xmath226 levels will have @xmath6 maxima to flip .
the number of maxima in the last @xmath224 levels decrease by a unit at each depth .
thus , the distance travelled equals @xmath227 since @xmath228 , corollary 14.7 of gives @xmath229 \right ] \,.\ ] ] note that @xmath196 as @xmath24 , so @xmath230 -2 \log \beta + o(\beta ) \right ] \,.\ ] ] in particular , if @xmath231 , then @xmath232 , which is the same order as the mixing time in the symmetric case . ]
[ prop : lbu ] suppose that @xmath150 . for any @xmath193 and @xmath233 ,
if @xmath0 is large enough , then @xmath234 we use the particle description here .
the stationary distribution is given by @xmath235 where @xmath236 are the locations of the @xmath6 particles in the configuration @xmath52 , and @xmath237 is a normalizing constant . to see this ,
if @xmath238 is obtained from @xmath52 by moving a particle from @xmath225 to @xmath239 , then @xmath240 let @xmath241 be the location of the left - most particle of the configuration @xmath52 , and let @xmath242 be the location of the right - most unoccupied site of the configuration @xmath52 .
let @xmath243 and consider the transformation @xmath244 which takes the particle at @xmath225 and moves it to @xmath206 .
note that @xmath245 is one - to - one on @xmath246 .
we have @xmath247
so @xmath248 letting @xmath249 , we have @xmath250 we consider now starting from a configuration @xmath251 with @xmath252 .
the trajectory of the left - most particle , @xmath253 , can be coupled with a delayed biased nearest - neighbor walk @xmath169 on @xmath171 , with @xmath254 and such that @xmath255 , as long as @xmath256 .
the holding probability for @xmath169 equals @xmath257 . by the gambler s ruin , the chance @xmath258 ever reaches @xmath259 is bounded above by @xmath260 therefore .
@xmath261 by chebyshev s inequality ( recalling @xmath254 ) , @xmath262 taking @xmath263 and @xmath264 shows that @xmath265 as long as @xmath266 . combining with shows that @xmath267
we conclude that as long as @xmath266 , @xmath268 as @xmath3 , whence @xmath269 for sufficiently large @xmath0 .
we thank perla sousi and nayantara bhatnagar for helpful comments on an earlier version of this paper . | we analyze the mixing behavior of the biased exclusion process on a path of length @xmath0 as the bias @xmath1 tends to @xmath2 as @xmath3 .
we show that the sequence of chains has a pre - cutoff , and interpolates between the unbiased exclusion and the process with constant bias . as the bias increases , the mixing time undergoes two phase transitions : one when @xmath1 is of order @xmath4 , and the other when @xmath1 is order @xmath5 . |
filaments support both longitudinal and transverse oscillations ( see , * ? ? ?
the oscillations in the filaments can be used to diagnose the local plasma conditions and magnetic field by applying the principle of mhd seismology ( see , * ? ? ?
* ; * ? ? ?
these oscillations are broadly classified as large amplitude @xcite and small amplitude @xcite oscillations .
large amplitude transverse oscillations , where a filament oscillates as whole , are often associated with the disturbances coming from nearby flares .
in contrast , the large amplitude longitudinal oscillations are always found to be associated with small energetic events like a subflare at one leg of the filament @xcite , a small flare in a nearby active region or jets @xcite . there have been several reports on large amplitude longitudinal oscillations .
developed a 1d self - consistent model which explained both restoring force and damping of large amplitude longitudinal oscillations .
the model is analogous to an oscillating pendulum assuming the radius of curvature of magnetic dip as the length of the pendulum .
used this model to explain the strong damping of large amplitude longitudinal oscillations observed in the filament threads . recently , performed 2d non - linear time dependent mhd simulations where a magnetic field can get distorted in response to the mass loading .
the authors have reported that a simple pendulum model can be used to characterise the large amplitude longitudinal oscillations even in a scenario where the magnetic field responds to the plasma motions .
+ damping of oscillations have been observed in large amplitude transverse oscillations .
used seismology combining the damping time and period of oscillations in a consistent manner to estimate the magnetic field strength and inhomogeneity length scale in the coronal loops .
applied the same technique to the oscillating threads of a prominence . applied this technique to estimate magnetic field and inhomogeneity length scale of a filament as whole .
+ in addition to transverse oscillations , large amplitude longitudinal oscillations are also found to be damped @xcite . unlike transverse oscillations , that are now believed to be damped by the resonant absorption , damping mechanisms in longitudinal oscillations are not well understood .
several damping mechanisms have been proposed , _
e.g. , _ energy leakage @xcite , dissipation @xcite but these mechanisms are under debate .
however , mass accretion due to the condensation of filament material @xcite is shown to be promising in explaining the damping of longitudinal oscillations in a filament .
recently , have reported the evidence of strong damping of longitudinal oscillations in the filament threads that are modelled as a curved magnetic tube .
+ reported that an incoming shock wave can trigger longitudinal and transverse oscillations in the filaments depending on the interaction angle between the filament and the shock wave .
has reported coexistence of two transverse oscillation modes , one in the plane of sky and other along the line of sight , in a filament in response to a moreton wave .
both modes were perpendicular to the filament axis . in this work
, we report a unique observation of the co - existence of large amplitude damped longitudinal oscillations and large amplitude damped transverse oscillations in an active region filament .
the paper is organised as follows . in section 2
, we describe the observations . in section 3
we discuss the methods of data analysis .
we present the results in section 4 , which is followed by the discussion and conclusions in section 5 .
a m1.1 class flare was observed by goes satellite in an active region ar 11692 on march 15@xmath0 2013 .
the flare was associated with a halo coronal mass ejection ( cme ) .
the flare created a global disturbance in the active region .
the cme associated with the flare produced a shock wave . to observe shock fronts , we created running difference images of aia 193 .
the shock fronts are marked with red and green arrows in figure [ shock_fig ] .
it is evident from figure [ shock_fig ] that the shock fronts are moving outward .
a filament was lying to the south - west of the active region as seen in h@xmath1 images from nso / gong ( figure [ context ] ) .
nso / gong provides full - disk observation of the sun in 6563 ( h@xmath1 ) .
it has a pixel resolution of @xmath2 1.07@xmath3and a cadence of 1 sec .
the observational data used in this study are obtained from 06:00:54 ut to 10:22:54 ut .
the flare started at 05:46:00 ut , peaked at 06:58:00 ut and ended at 08:35:00 ut as recorded in goes catalogue .
we note that just after the onset of the flare , oscillations were triggered in the filament .
+ filaments consist of fine threads that are not often seen in gong h@xmath1 images due to its poor spatial resolution .
therefore , we use the images from the extreme ultraviolet ( euv ) passbands of _ atmospheric imaging assembly _ ( aia ) onboard _ solar dynamic observatory _ ( sdo ) .
aia instrument provides almost simultaneous full - disk images of the sun in seven euv band .
it has a spatial resolution of @xmath4 , pixel size of @xmath5 and a cadence of 12 s @xcite .
we find that the filament consists of several threads that are seen more clearly in aia 171 than any other euv passbands of aia .
therefore , we use aia 171 for further analysis . moreover , we find that shortly after the oscillations started , the full disk images of the sun were not available from 06:22:23 ut to 07:36:00 ut .
thus , the first cycle of oscillations is not seen in aia 171 .
+ we overplot the contours of the filament as seen in gong h@xmath1 over the magnetogram obtained from _ helioseismic and magnetic imager _ ( hmi ) onboard sdo ( figure [ hmi ] ) .
we find that the filament lies over the region which separates the positive and the negative polarity .
we note from the movie ( available online ) that at 06:11:59 ut , _
i.e _ before eruption , the filament material as enclosed in two different contours are part of the same filament .
+ by careful inspection of the h@xmath1 images in the movie ( available online ) , it appears that both longitudinal and transverse oscillations are present simultaneously in the filament . in the movie ( available online )
, we mark the filament in aia images with a green arrow .
the transverse oscillations in the filament under study are already reported in . in this paper , we focus on longitudinal oscillations at two parts of the filament adjacent to its two ends .
image of the filament lying at south - west of the active region ar 11692 .
two broad artificial slices along the axis of the filament , used for generating x t maps , are over plotted .
three artificial slits perpendicular to the axis of filament are overplotted in green , red and yellow , which were used in @xcite to investigate the transverse oscillations . ]
_ right _ hmi magnetogram .
overplotted red contours outline the position of the filament . ]
to investigate the longitudinal oscillations in gong h@xmath1 images , we place two artificial broad slices , slit 1 and slit 2 along the axis of the filament as shown in figure [ context ] .
the length ( width ) of the @xmath6 and the @xmath7 slice is 130 pixel ( 15 pixel ) and 60 pixel ( 20 pixel ) respectively .
the filament is oscillating in both transverse and longitudinal directions simultaneously .
we choose broad slices so that the filament material remain inside the slices in spite of the transverse oscillations .
to further characterise the longitudinal oscillations in the filament , first we create a time - distance ( x t ) map with time in x axis and distance along the slice in y axis for both slit 1 and slit 2 .
intensity along each column of the x
t map is fitted with a gaussian curve .
the mean value and one sigma error of the intensity is estimated . finally the x
t map is fitted with a damped sinusoidal function to investigate the oscillation characteristics .
since the filament is oscillating and the amplitude of the oscillations is decreasing with time , it can be fitted with a damped sinusoidal function ( assuming the mass of the filament to be constant ) and represented as @xcite , @xmath8 where @xmath9 is a constant , @xmath10 is the amplitude , @xmath11 is the damping time , @xmath12 is the angular frequency , and @xmath13 is the initial phase of sinusoidal function .
the least square fitting is done using the function _
_ in interactive data language ( idl ) @xcite .
it has been shown that if a filament accretes mass then the bessel function describes the oscillations better than the damped sinusoid @xcite .
therefore , we fit the x
t map using a damped zeroth order bessel function of first kind , represented by @xmath14 where @xmath15 is a constant , @xmath16 is amplitude , @xmath12 is the angular frequency , @xmath17 is the damping time , and @xmath18 is the phase of the bessel function .
we repeat the analysis on aia 171 images but with only one artificial slice .
the filament material at the position of slit 2 is not clearly seen in aia 304 , 171 and 193 due to the presence of overlying post - flare loops ( see movie ) .
the period , damping time and amplitude estimated from the fitting of a damped sinusoidal function represented by equation ( 1 ) are summarized in table [ table1 ] .
the period , damping time and amplitude estimated from the fitting of a damped zeroth order bessel function of first kind represented by equation ( 2 ) is summarized in table [ table2 ] . in the right panel of figure
[ xt ] , we overplot dashed green lines at three time instances .
it is evident that both ends of the filament started oscillating in different phase .
it is also worth noting that the oscillations started at different times in the two ends of the filament .
the period of oscillation is found to be 57@xmath193 min which is similar to the period estimated at the position of slit 1 in gong h@xmath1 .
the damping time is estimated to be 100 min which is larger as compared to the damping time estimated using slit 1 in gong h@xmath1 images .
amplitude of oscillation is found to be 10 mm which is less than the amplitude measured in gong h@xmath1 images .
the main reason of large damping time and smaller amplitude is that , first cycle of the oscillations is missed in aia 171 images , therefore the oscillations start after 100 min as shown in figure [ xt_aia ] ( right panel ) .
we note from figure [ xt ] ( top right panel ) that the estimated damping time of oscillations and amplitude are strongly dependent on first cycle of the oscillations .
if first cycle of oscillations is missed then the damping time will be longer and amplitude will be smaller .
another reason can be that the filament threads apparently merge and separate during the oscillations .
this may be attributed to the fact that multiple threads are stacked along the line of sight , so if they are not moving in phase it may appear that they are merging and separating .
this is probably also related to the presence of transverse oscillations because such effect is not reported in earlier studies @xcite where only longitudinal oscillations were present .
therefore , there is large uncertainty in choosing the thread that is associated with the oscillations ( see right panel of figure [ xt_aia ] ) .
it is worth noting at this point that we did not fit damped bessel function to aia dataset because , as said above , the first cycle of oscillation is missed thus the bessel function will not give the correct estimation of the damping time .
filament material is supported by the magnetic fields against gravity .
we use a simple pendulum model to estimate the oscillation parameters @xcite . assuming restoring force to be gravity , we get the following equation @xmath20 where @xmath21 is the sun s surface gravity
, @xmath12 is the oscillation frequency , @xmath22 is the period of oscillation and @xmath23 is the radius of curvature .
using equation ( 3 ) , the radius of curvature in gong h@xmath1 images is estimated to be @xmath2 80 mm and 85 mm at the position of dip 1 ( where @xmath6 slice is placed ) and dip 2 ( where @xmath7 slice is placed ; see , figure [ context ] ) respectively .
similarly , the radius of curvature at the location of dip 1 ( see left hand panel of figure [ xt_aia ] where artificial slice is placed ) in aia 171 is estimated to be 80 mm .
we calculate a lower limit of the magnetic field strength , assuming that magnetic stress is balanced by the weight of the filament material present in a dip @xcite , using the following expression , @xmath24 where @xmath25 is the magnetic field , @xmath26 is the electron number density in @xmath27 and @xmath22 is the period of oscillation in hours .
here we use the typical value of electron number density @xmath28 @xmath27 as reported in . therefore
, minimum magnetic field strength is estimated to be @xmath2 ( 25 @xmath19 1 ) g both in dip 1 and dip 2 which is consistent with the typical values of the measured magnetic fields in the filaments @xcite . since the period of oscillation is estimated to be the same using aia 171 , the estimated magnetic field strength remains unchanged . in the model given by ,
the damping of longitudinal oscillations in threads of the filaments is explained by the accretion of filament material .
we calculate the mass accretion rate from the best fit parameters of the damped bessel function ( represented by equation ( 2 ) ) , using the following relation @xcite , @xmath29 where , @xmath30 is the initial mass of filament , @xmath12 and @xmath18 are the obtained from fitting of the damped bessel function .
the mass of a filament , @xmath30 can be calculated as @xmath31 , where @xmath32 and @xmath33 are the radius and the length of the thread of the filament .
we take the typical radius of a filament thread to be 100 km as assumed in .
the length of the thread is estimated using x
t map as shown in figure [ xt ] ( top panel ) .
as explained above , the artificial slit is placed along the length of the filament .
thus the extent of the brightness along each column of the x
t map gives an estimate of the length of the filament thread at a particular time instance .
the length of the filament thread is measured at various time instances ( _ i.e , _ along different columns ) from the start to the end of the oscillations .
finally , we estimate the mean value of the length of the filament thread @xmath2 23@xmath19 7 mm at the position of slit 1 . following the above procedure ,
the length of the filament thread at the position of slit 2 is estimated @xmath2 10@xmath194 mm .
it should be borne in mind that the estimation of the length of the filament thread at the position of slit 2 is uncertain because the oscillations are not clearly seen in the x
t maps as shown in the bottom panel of figure [ xt ] . under these assumptions
we estimated the mass accretion rate at the position of slit 1 and 2 to be @xmath2 ( 51 @xmath19 17 @xmath3410@xmath35 kg hr@xmath36 ) and ( 192 @xmath19 72 @xmath3710@xmath35 kg hr@xmath36 ) respectively .
the derived mass accretion rate at the position of slit 1 is comparable to those reported in earlier studies @xcite .
it should be noted that at the position of slit 2 the mass accretion rate is higher while damping time is longer .
one of the reasons for this discrepancy is the uncertainty in the fitting parameters . since the mass accretion rate depends on the phase of the bessel function which is estimated from the best fit parameters .
therefore , if the fitting of a function is not good then the value of phase will have large errors and thus can not be trusted .
.table of parameters of damped sine fitting [ cols="^,^,^,^,^ " , ] in this subsection we investigate the possibility of the existence of both longitudinal and transverse oscillations driven by kink oscillations .
we assume the filament as a cylinder embedded in a uniform plasma with low plasma @xmath38 .
magnetic field is assumed to be constant inside the cylinder and the effects of gravity is ignored .
the model set up is similar to the one described in @xcite . to get the variation of thermodynamic quantities ,
linearized ideal mhd equations are used ( see * ? ? ?
* ) , @xmath39 @xmath40,\ ] ] @xmath41 @xmath42 where @xmath43 is the lagrange displacement vector , @xmath44 , @xmath45 and * b * are perturbed quantities while @xmath46 , @xmath47 and * b@xmath48 * are unperturbed quantities .
neumann boundary conditions are assumed at @xmath49 @xmath50_{r = a}=0,\ ] ] @xmath51_{r = a}=0\ ] ] and dirichlet boundary conditions are assumed at @xmath52 @xmath53 @xmath54 @xmath55 @xmath56 where @xmath57 is the total pressure . for a filament , we define @xmath12 , @xmath58 , @xmath59 as kink , alfvn and tube oscillation frequency respectively and the tube speed , @xmath60 is given by @xmath61 where @xmath62 is the alfvn speed and @xmath63 is the sound speed .
@xmath64 , @xmath47 , @xmath46 are equilibrium magnetic filed , plasma pressure and plasma density respectively and @xmath65 is the magnetic permeability in free space . for kink mode ( m=1 ) , the total pressure perturbation , @xmath66 , in the filament
is assumed to be @xmath67 where @xmath68 is the amplitude of the total perturbed pressure .
since kink mode is not axisymmetric thus cos(@xmath69 ) dependence is assumed . substituting of equation ( 9 ) into equations ( 5 - 8 ) yields , @xmath70 where @xmath71 where @xmath72 .
equation ( 10 ) should also hold true for r(r ) . in this work
we confine ourselves to the inside of the flux tube thus , on solving equation ( 10 ) for r(r ) , we get , @xmath73 in cylindrical coordinates , @xmath74 is the perturbed velocity amplitude of the transverse oscillation and @xmath75 is the perturbed velocity amplitude of the longitudinal oscillation .
inside the cylinder the perturbed radial and axial component of velocity amplitude are defined as ( see * ? ? ? * for details ) , @xmath76 and @xmath77 thus , axial to radial velocity amplitude ratio is given by the following expression , @xmath78 we take the length of slit 2 as length of the filament , which is estimated to be 90 mm .
the radius of the filament is estimated to be @xmath2 1 mm .
it is important to note that here we assume the length of the flux tube ( modelled as a filament ) as the length of the slit covering the filament material because here we compare longitudinal oscillations with the transverse oscillations as reported by , where authors assumed that filament was oscillating as a whole in the transverse direction . to estimate the ratio of the axial to radial velocity amplitude
, we choose three slit positions as shown in green , red and yellow lines in figure [ context ] , that were used by . from observations
, we find that the longitudinal velocity amplitude to the transverse velocity amplitude ratio at these locations are 3.92 , 3.26 and 2.22 respectively . using equation ( 12 ) , we estimate the ratio to be 4 @xmath79 @xmath80 , 2 @xmath79 @xmath80 and 1 @xmath79 @xmath81 respectively assuming typical value of sound speed in chromosphere as 15 km s@xmath36 , alfvn speed in three slit positions as 82 , 88 and 91 km s@xmath36 respectively and perturbation frequency ( assumed to be equal to the kink oscillation frequency ) to be 117 , 124 and 128 km s@xmath36 respectively ( see , * ? ? ?
it should be borne in mind that a linear analysis is presented here .
considering the observed displacement to be large , non - linear effects would become important . therefore to model such events further studies are needed .
nevertheless , from above analysis , we note that it is almost impossible to detect longitudinal oscillation in low-@xmath38 plasma as a result of fast kink mode in our observation
. therefore , large amplitude longitudinal oscillations can not be the kink mode .
in this section , we explore the most likely mechanism for driving the longitudinal and transverse oscillations in a filament channel . reported that shock waves can interact with filaments in two possible geometries . if shock waves come such that the normal vector to the shock wave is perpendicular to the filament axis then a filament oscillates in the transverse direction .
while if the shock wave interacts with a filament such that the normal vector of the shock wave is parallel to the filament axis then a filament oscillates in the longitudinal direction . here we are proposing a more general scenario where a shock wave interacts obliquely with the filament such that the filament material oscillates in both transverse and longitudinal direction as revealed from our observations .
the position of the shock fronts in our observations are clearly seen in figure [ shock_fig ] .
however , it should be borne in mind that from observations it is not straight forward to determine the direction of shock with respect to the filament because the filament is in close proximity of the active region and is associated with overlying loops that expand in response to the shock wave .
this makes it extremely difficult to determine the orientation of the shock wave .
we depict this by a cartoon in figure [ cartoon ] .
we show that when a shock wave hit a filament obliquely , the normal vector to the wave front , @xmath82 ( shown in dark red ) , can be decomposed into perpendicular ( @xmath83 ) and parallel ( @xmath84 ) components as shown in figure [ cartoon ] in blue .
the component parallel to the filament axis triggers the longitudinal oscillations in the filament while the component perpendicular to the filament axis triggers the transverse oscillations in the filament .
thus the shock front interacts with the filament at dip 1 and displaces the filament material in both transverse and longitudinal direction .
similarly in dip 2 the shock front displaces the plasma in both transverse and longitudinal directions , but at a different time . with this schematic picture
we can explain the different start time of longitudinal oscillation at two ends of the filament .
however , with this scenario , one should expect the transverse oscillations to be present at the locations of both dips but transverse oscillations are only observed at the location of dip 1 .
it is worth noting that this is the simple geometry . in principle
many parts of the shock wave can interact with the filament at different orientations and can excite different modes .
further studies are needed to confirm this .
in this work , we find the co - existence of longitudinal and transverse oscillations in an active region filament as observed by nso / gong h@xmath1 .
the oscillations are triggered by the interaction of a shock wave with the filament .
we find that the oscillations in two ends of the filament started with different phase but with almost similar period of oscillations .
we note that the longitudinal oscillations at two slit positions started at different times ( see , figure [ xt ] ) may be because the shock wave front interacts with the two parts of filament at different time instances .
we depict it with a schematic diagram ( see , figure [ cartoon ] ) .
the damping time at the position of slit 1 is estimated to be 46 minute .
this is a bit too strong as compared to the earlier reports @xcite and comparable to what is reported in .
suggested that the mass accretion of the filament material is responsible for damping .
we find mass accretion rate to be 51 @xmath19 17 @xmath3410@xmath35 kg hr@xmath36 at the position of slit 1 and 192 @xmath19 72 @xmath3710@xmath35 kg hr@xmath36 at the position of slit 2 which is comparable with the earlier reports @xcite .
we performed the filament seismology and estimated the magnetic field strength at two dip locations .
we estimated the lower limit of the magnetic field strength @xmath2 ( 25 @xmath19 1 ) g at the location of both the slices which is consistent with typical values of measured magnetic field from observation @xcite .
we also calculate the radius of curvature of two magnetic dips of the filament as 80 mm and 85 mm for dip 1 and dip 2 , respectively .
+ from our observations we exclude the possibility of existence of transverse and longitudinal oscillation simultaneously driven by a fast kink mode .
longitudinal oscillations due to kink wave may be present but their amplitude is far less as compared to the amplitude of the longitudinal oscillations observed in this case . thus we believe that transverse and longitudinal oscillations exist independently in the filament under study .
we would like to thank the referee for his / her valuable comments which has improved the presentation of this article .
34 # 1isbn # 1#1#1#1#1#1#1#1_#1_#1*#1*#1#1#1#1#1#1#1#1#1#1#2#1#1#1#1#1_#1_#1#1#1*#1*#1#1#1_#1_#1#1#1#1#1#1#1#1#1#1#1 # 1http://dx.doi.org/#1 [ ] # 1http://arxiv.org / abs/#1 [ ] # 1http://adsabs.harvard.edu / abs/#1[]#1#1#1#1#1#1#1#1#1#1#1#1 | we report on the co - existence of longitudinal and transverse oscillations in an active filament . on march 15@xmath0 2013 , a m1.1 class flare was observed in the active region ar 11692 .
a cme was found to be associated with the flare .
the cme generated a shock wave that triggered the oscillations in a nearby filament , situated at the south - west of the active region as observed from national solar observatory ( nso)_global oscillation network group_(gong ) h@xmath1 images . in this work
we report the longitudinal oscillations in the two ends of the filament , co - existing with the transverse oscillations .
we propose a scenario in which an incoming shock wave hits the filament obliquely and triggers both longitudinal and transverse oscillations .
using the observed parameters , we estimate the lower limit of the magnetic field strength .
we use simple pendulum model with gravity as the restoring force to estimate the radius of curvature .
we also calculate the mass accretion rate which causes the filament motions to damp quite fast . |
our knowledge of stellar structure and evolution rests heavily on the comparison between theory and observation .
double - lined eclipsing binaries ( hereafter ebs ) have long been at the center of this process , since they allow the mass the most fundamental of all stellar properties as well as the radius to be determined to very high precision ( and accuracy ) , often as good as 1% , independently of the distance and independently of any calibrations .
such high precision enables stringent tests of theory , as described , e.g. , by @xcite . in the last decade
or so it has become clear that stars in the lower main - sequence show significant discrepancies when compared to standard models
. studies of several key systems have shown unambiguously that the radii predicted by the models are systematically up to @xmath810% too small , and the temperatures @xmath85% too high ( see * ? ? ? * and references therein ) .
the deviations are commonly attributed to the high level of chromospheric activity present in these systems .
orbital periods are typically short , and as a result tidal forces tend to synchronize the components rotation with the orbital motion .
the rapid rotation , in turn , induces the activity , which can manifest itself in the form of copious x - ray emission , flaring , h@xmath9 and ca ii h and k emission , spottedness , and other effects .
according to the recent compilation by @xcite , among the eclipsing binaries with at least one component below 0.8 m@xmath10 only five have reliable mass and radius determinations with relative errors that are smaller than 3% .
these are , in order of decreasing mass , uv psc @xcite , gu boo @xcite , yy gem @xcite , cu cnc @xcite , and cm dra @xcite .
a number of other similar systems are known , but are not yet at the same level of precision _ and _ accuracy . for further progress
it is therefore important to either improve the determinations in the latter cases , or to find new ones where those conditions are met .
the eclipsing system reported here is one of these , which in addition presents the largest difference in mass between the components , providing in principle extra leverage for the comparison with models .
i m virginis ( also hd 111487 , 1e 1247.0@xmath110548 , @xmath12 , @xmath13 , j2000.0 ; , @xmath14 ) was detected as an x - ray source with the _ einstein _ observatory by @xcite .
the radial - velocity variability was found by @xcite , and subsequent spectroscopic and photometric studies carried out by @xcite confirmed i m vir to be both double - lined and eclipsing , and to be composed of a g7 star and a late - k or early - m star , thus piquing our interest .
the orbital period was estimated as 1.3085 days .
very little data on this binary have been published since , other than sparse photometry and occasional reports on the chromospheric activity and x - ray flaring @xcite .
we present here extensive spectroscopic and differential photometric measurements that allow us to determine its fundamental properties with very high accuracy .
the observations are presented in [ sec : spectroscopy ] and [ sec : lc_obs ] , followed by a detailed light - curve analysis in
[ sec : lc_analysis ] accounting for the presence of spots .
the absolute dimensions are derived in [ sec : absolute ] with careful consideration of potential sources of systematic error , essential for the results to be useful . the comparison with models of stellar evolution
is discussed in
[ sec : discussion ] , and we summarize our conclusions in [ sec : conclusions ] .
i m vir was observed at the harvard - smithsonian center for astrophysics ( cfa ) with three nearly identical echelle spectrographs on the 1.5-m tillinghast reflector at the f. l. whipple observatory ( mount hopkins , arizona ) , the 1.5-m wyeth reflector at the oak ridge observatory ( harvard , massachusetts ) , and the 4.5-m equivalent multiple mirror telescope ( also on mount hopkins , arizona ) prior to its conversion to a 6.5-m monolithic telescope .
photon - counting intensified reticon detectors ( digital speedometers ; * ? ? ? * ; * ? ? ?
* ) were used in each case to record a single 45 echelle order centered at a wavelength of 5188.5 , featuring the gravity - sensitive lines of the b triplet .
the resolving power provided by this setup is @xmath15 .
nominal signal - to - noise ( s / n ) ratios for the 138 spectra we obtained range from 13 to 58 per resolution element of 8.5 .
the first observation was taken in 1984 january 1 , and monitoring continued until 2009 may 10 .
a handful of the early spectra are the same ones included in the work of @xcite , who discovered the radial - velocity variability , but have been re - reduced and analyzed here with much improved methods , as we now describe .
all of our spectra are double - lined , but the secondary is comparatively quite faint , accounting for only 6% of the light of the primary ( see below ) .
radial velocities for both stars were derived using todcor , a two - dimensional cross - correlation technique introduced by @xcite .
this method uses two templates , one for each component of the binary , which we selected from a large library of synthetic spectra based on model atmospheres by r. l. kurucz ( see * ? ? ?
these templates have been calculated for a wide range of effective temperatures ( @xmath16 ) , surface gravities ( @xmath17 ) , rotational velocities ( @xmath18 when seen in projection ) , and metallicities ( [ m / h ] ) .
following @xcite , the optimum templates were selected by running extensive grids of cross - correlations with todcor , seeking to maximize the average correlation weighted by the strength of each exposure .
because of the faintness of the secondary , only the parameters for the bright star can be determined reliably from our spectra .
initially we assumed @xmath19 , and determined @xmath16 and @xmath18 for a range of metallicities between [ m / h ] @xmath20 and [ m / h ] @xmath21 .
the optimal values were found by interpolation in [ m / h ] .
we then repeated the correlations for a lower value of @xmath22 in order to bracket the estimate from the analysis described later ( @xmath23 ) .
interpolation in @xmath17 to this final value resulted in @xmath24 k and @xmath25 .
the formal metallicity that maximizes the correlation is [ m / h ] = @xmath26 .
while this suggests an overall abundance close to solar , the uncertainty is likely to be at least 0.25 dex , and we consider the estimate primarily as a free parameter included to optimize the match between the synthetic templates and the observed spectra .
we note , furthermore , that because of our narrow spectral window , there is a strong correlation between temperature and metallicity . in this case , however , the @xmath16 value reported above is consistent with various photometric estimates described below in [ sec : absolute ] , in turn lending more credence to the metallicity estimate . for the final velocity determinations we adopted the set of template parameters closest to the values above that
maximizes the average correlation ( @xmath27 k , @xmath28 , @xmath19 , and solar composition ) .
small differences in these template parameters compared to the interpolated values do not affect the velocities significantly . for the secondary template we adopted the same metallicity , and parameters consistent with the results from [ sec : absolute ] : @xmath29 k , @xmath30 , @xmath19 .
in addition to the velocities , we have determined the light ratio between the components at the mean wavelength of our observations , following @xcite , accounting for the difference in line blocking between the primary and the much cooler secondary .
the value we obtain is @xmath31 .
although todcor significantly reduces systematic errors in the radial velocities caused by line blending , residual effects can remain as a result of shifts of the spectral lines in and out of our narrow spectral window as a function of orbital phase .
experience has shown that these effects must be examined on a case - by - case basis .
we investigated them here by means of numerical simulations similar to those described by @xcite ( see also * ? ? ?
* ; * ? ? ?
we generated synthetic composite spectra matching our observations by combining copies of the primary and secondary templates used above , shifted to the appropriate velocities for each of the exposures as predicted by a preliminary orbital solution , and scaled to the observed light ratio .
these synthetic observations were then processed with todcor in exactly the same way as the real spectra , and the resulting velocities were compared with the input shifts .
the differences for i m vir are shown in figure [ fig : cortodcor ] , and are less than 0.5 for the primary but reach values as large as 13 for the secondary .
similarly large differences have been found occasionally for other systems using the same instrumentation ( e.g. , ad boo , gx gem , vz cep ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we have applied these differences as corrections to the raw velocities of i m vir . the impact on the masses
, however , is very small : only 0.26% for the primary and 0.14% for the secondary .
the reason for this is that the large corrections for the secondary are similar and of the same sign at both quadratures , therefore amounting mostly to an overall systematic shift rather than a change in the velocity semi - amplitude .
similar adjustments based on the same simulations were made to the light ratio , and are already included in the value reported above .
lccccccc 45719.9949 & @xmath1167.58 & @xmath32144.06 & 1.54 & 13.01 & @xmath320.30 & @xmath329.61 & 0.1661 + 45743.0126 & @xmath32105.93 & @xmath11116.81 & 0.83 & 6.98 & @xmath321.08 & @xmath323.76 & 0.7555 + 45750.8945 & @xmath32100.91 & @xmath11122.28 & 1.58 & 13.30 & @xmath112.51 & @xmath113.82 & 0.7785 + 45754.0581 & @xmath1174.09 & @xmath32148.80 & 1.54 & 12.97 & @xmath321.10 & @xmath323.56 & 0.1961 + 45754.0603 & @xmath1171.28 & @xmath32149.09 & 1.90 & 16.05 & @xmath324.23 & @xmath323.38 & 0.1977 + the stability of the zero - point of the cfa velocity system was monitored by means of exposures of the dusk and dawn sky , and small run - to - run corrections were applied in the manner described by @xcite .
the final velocities , including these offsets as well as the corrections for systematics , are listed in table [ tab : rvs ] along with their uncertainties .
the median values for the velocity errors are 0.90 for the primary and 7.6 for the secondary .
preliminary single - lined orbital solutions carried out separately for the primary and secondary indicated a significant zero - point difference ( i.e. , a difference in the systemic velocity @xmath33 ) , which is often seen by many investigators in cases where there is a slight mismatch between the spectra of the real stars and the templates used for the cross - correlations ( see , e.g. , * ? ? ?
* ; * ? ? ?
numerous experiments with other templates did not remove the offset , which is about 4 ( the secondary velocities being systematically larger ) . in i
m vir this most likely arises because of the cool temperature of the secondary , and the fact that synthetic templates for such stars become increasingly unrealistic due to missing opacity sources in the models . in order to prevent this offset from affecting the velocity semiamplitudes ,
we have therefore accounted for it by including it as an additional free parameter in our double - lined solution .
effectively this means we allow each component to have its own center - of - mass velocity , without significantly affecting the velocity semiamplitude of the component , and hence without affecting the individual mass determination . on the other hand ,
the offset implies additional uncertainty in the true center - of - mass velocity of the binary , beyond the formal errors listed below . we do not expect a significant template mismatch for the solar - type primary ,
so the systemic velocity is most likely to be closer to the value for that star .
the results are presented in table [ tab : specorbit ] , in which the measurements have been weighted according to their uncertainties in the usual way , and the errors rescaled by iterations to achieve a reduced @xmath34 value near unity , separately for the primary and secondary .
no indication of significant eccentricity was found , as expected for such a short orbital period , and the final orbit was therefore considered circular . the observations and computed curve
are displayed graphically in figure [ fig : specorbit ] , along with the residuals , which are also listed in table [ tab : rvs ] .
lc + @xmath35 ( days ) & 1.30861497 @xmath36 0.00000034 + @xmath37 ( hjd@xmath38 ) & 52402.87420 @xmath36 0.00052 + @xmath39 ( ) & 92.684 @xmath36 0.088 + @xmath40 ( ) & 136.84 @xmath36 0.74 + @xmath41 & 0 + @xmath33 ( ) & + 12.221 @xmath36 0.082 + @xmath42 ( ) & @xmath113.96 @xmath36 0.69 + + @xmath43 ( @xmath44 ) & 0.978 @xmath36 0.012 + @xmath45 ( @xmath44 ) & 0.6621 @xmath36 0.0044 + @xmath46 & 0.6773 @xmath36 0.0037 + @xmath47 ( 10@xmath48 km ) & 1.6678 @xmath36 0.0016 + @xmath49 ( 10@xmath48 km ) & 2.462 @xmath36 0.013 + @xmath50 ( 10@xmath48 km ) & 4.130 @xmath36 0.013 + + @xmath51 ( ) & 0.99 + @xmath52 ( ) & 8.39 + @xmath53 & 138 + time span ( days ) & 9241.7
differential photometric observations of i m vir were conducted in 2006 april and may using the 0.4-m ealing cassegrain reflector of the gettysburg college observatory ( gco , gettysburg , pennsylvania ) .
the camera was a photometrics ( roper scientific ) ch-350 thermoelectrically - cooled unit equipped with a site 003b back - illuminated scientific grade ccd chip and bessell @xmath0 filters .
exposure times were 120 , 60 , 25 , and 20 seconds , respectively .
we obtained 597 , 672 , 630 , and 631 observations in @xmath54 , @xmath55 , @xmath56 , and @xmath57 , with a cadence of about 6 minutes .
the field of view of the gco camera is approximately 18 arcmin , and since i m vir is a relatively bright target , this severely limited our choice of reference stars of comparable magnitude and color to two : bd@xmath1105 3573 ( ` comp ' , @xmath58 , @xmath59 , j2000.0 ; @xmath60 , @xmath61 ) and hd 111427 ( ` check ' , @xmath62 , @xmath63 , j2000.0 ; @xmath64 , @xmath61 ) .
the colors of these two stars are in fact nearly identical to that of the variable , which is @xmath65 ( see [ sec : absolute ] ) .
differential photometry was performed on i m vir and the two reference stars in all our images by means of mira - ap software ( http://www.mirametrics.com/ ) .
we employed standard aperture photometry techniques to derive instrumental magnitudes , setting the radius of the measuring apertures for each night using a standard value of 2.5 times the fwhm of the seeing disk , based on previous curve - of - growth calibrations using the same equipment .
typical errors as represented by the scatter of the comp@xmath11check differences are 0.0132 mag in @xmath54 , 0.0124 mag in @xmath55 , 0.0135 mag in @xmath56 , and 0.0150 mag in @xmath57 .
the primary eclipse is @xmath80.62 mag deep in @xmath55 , while the depth of the secondary is only @xmath80.08 mag .
lccc 53835.58358 & 0.8288 & 0.811 & 0.966 + 53835.58732 & 0.8317 & 0.849 & 0.995 + 53835.59108 & 0.8345 & 0.852 & 1.000 + 53835.59482 & 0.8374 & 0.853 & 1.007 + 53835.59857 & 0.8403 & 0.854 & 1.005 lccc 53835.58481 & 0.8298 & 0.848 & 0.994 + 53835.58855 & 0.8326 & 0.846 & 0.999 + 53835.59230 & 0.8355 & 0.839 & 0.989 + 53835.59605 & 0.8383 & 0.838 & 0.998 + 53835.60705 & 0.8468 & 0.826 & 0.985 lccc 53835.62654 & 0.8616 & 0.584 & 1.027 + 53835.63031 & 0.8645 & 0.660 & 1.011 + 53835.63406 & 0.8674 & 0.685 & 1.002 + 53835.63781 & 0.8703 & 0.704 & 1.003 + 53835.64157 & 0.8731 & 0.708 & 1.005 lccc 53835.58592 & 0.8306 & 0.933 & 1.013 + 53835.58967 & 0.8335 & 0.933 & 1.014 + 53835.59342 & 0.8363 & 0.939 & 1.034 + 53835.59717 & 0.8392 & 0.929 & 1.025 + 53835.60091 & 0.8421 & 0.946 & 1.032 examination of the raw photometry revealed slight trends in the comp@xmath11check differences that changed from night to night and are most likely of instrumental origin . neither star is known to be variable , and we find no periodicities or other discernible patterns . consequently , we de - trended the i m vir measurements by removing the median value of the comp@xmath11check differences calculated over intervals of a few hours .
even after this correction , the sparser second half of the data presents a systematic @xmath80.01 mag difference ( fainter ) in the out - of - eclipse light level compared to the first half , and other occasional changes possibly due to temporal evolution of surface inhomogeneities ( spots ) on the surface of one or both stars .
this would indicate a timescale for evolution of these features of a few weeks .
for these reasons we have chosen to restrict our analysis to the first half of the observations ( spanning 22 days ) , which provide complete coverage of both eclipses and are cleaner overall .
the complete data set is given for all filters in tables [ tab : lcb][tab : lci ] in its original form , i.e. , without the nightly corrections , and we list also the comp@xmath11check differences .
additional photometry of i m vir has been reported by @xcite in the strmgren @xmath66 system .
these data were obtained some twenty years earlier than our own measurements , between 1983 may and 1986 july .
unfortunately the coverage of the eclipses is very incomplete , as shown later , so these data are not useful for determining the geometric parameters of the system .
nevertheless , it is possible to extract an average time of eclipse , which we present below , as well as brightness ratios in the different bands , which will be used in
[ sec : absolute ] to deconvolve the light of the two stars in order to obtain photometric estimates of the temperatures and metallicity .
the ephemeris adopted for the remainder of this paper is the one calculated from the spectroscopy , @xmath67 which , by virtue of the 25-year radial - velocity coverage , is much more accurate than could ever be obtained from the 64-day interval of the photometric observations .
the uncertainties are indicated above in parentheses .
three times of eclipse from the @xmath0 photometry have been measured by fitting the light - curve model described below ( including spot parameters ) to each night with sufficient coverage of a primary or secondary minimum , simultaneously in all four passbands .
the only adjustable parameter allowed in these fits was a time shift .
a similar procedure was followed with the @xcite photometry , with the primary luminosity in each band added as a free parameter and no spots considered .
the resulting eclipse timings for i m vir are collected in table [ tab : minima ] , along with an average time of eclipse from the spectroscopy , and the few additional times found in the literature .
lccccc 46042.99984 & 0.00097 & i & ccd & @xmath110.00561 & 1 + 51274.8750 & & i & ccd & + 0.02690 & 2 + 51885.9610 & & i & vis & @xmath110.01029 & 2 + 52402.87420 & 0.00052 & i & spec & 0.0 & 4 + 53474.62948 & 0.00055 &
i & ccd & @xmath110.00038 & 3 + 53843.65995 & 0.00010 & i & ccd & + 0.00067 & 4 + 53845.62427 & 0.00052 & ii & ccd & + 0.00207 & 4 + 53877.68325 & 0.00038 &
i & ccd & @xmath110.00002 & 4
the differential photometry described in [ sec : lc_obs ] was phase - folded with the adopted ephemeris , and analyzed with the 2003 version of the wd code @xcite . for the reasons indicated earlier ,
only the first half of the data were used ( 22 nights worth ) .
the main light curve parameters adjusted are the orbital inclination ( @xmath68 ) , the pseudo - potentials ( @xmath69 and @xmath70 ) , the secondary effective temperature ( @xmath71 ) , the primary luminosity ( @xmath72 ) , and a phase shift
. the latter would normally be unnecessary given the accuracy of the spectroscopic ephemeris and the short interval of the photometric observations .
however , as we describe below , the light curves of i m vir present distortions we ascribe to spots , which can cause asymmetries in the eclipses that may appear as phase shifts .
since these distortions are modeled here only in a simplified way , it is prudent to include a phase shift as an additional parameter to supplement the spot model .
the primary temperature was held fixed at the spectroscopic value of 5570 k , since the light curves do not provide a strong enough constraint on both temperatures , but only on their ratio .
emergent intensities used in the program were taken from model atmospheres described by @xcite .
square - root limb - darkening coefficients for the johnson - cousins @xmath0 bands were interpolated from the tables by @xcite , and adjusted dynamically according to the current temperatures and surface gravities of the stars at each iteration .
the reflection albedos were fixed at the value 0.5 , appropriate for stars with convective envelopes , and the gravity darkening exponents were set to 0.34 for the primary and 0.40 for the secondary , following @xcite . the orbit was initially considered to be circular , and spin - orbit synchronization was assumed based on the evidence presented in [ sec : absolute ] .
the photometric measurements in the four passbands were fitted simultaneously .
outliers were rejected by 3-@xmath73 clipping based on preliminary solutions .
convergence in the final fit was considered to have been achieved when the corrections to the elements were smaller than the internal errors in three consecutive iterations .
as is common in eclipsing binaries with late - type components , the light curves of i m vir show out - of - eclipse modulations that are attributable to spots .
there is also a slight difference in the light level at the two quadratures .
these variations can in principle be modeled with the wd code in the approximation of circular spots of uniform temperature . in practice , however , such modeling is fraught with difficulties and there is abundant literature discussing problems of indeterminacy and non - uniqueness , particularly when using data of limited quality ( see , e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) .
nevertheless , because we are interested in recovering the geometric parameters of the system as free as possible from systematic errors , we have made an effort to account for these modulations by considering up to two spots .
while the subset of the photometric observations used here spans only 22 days , our spectroscopic coverage extends for more than 25 years .
given the changes in the light curves mentioned in
[ sec : lc_obs ] , the photometry is modeled separately from the radial velocities as any spot model would not apply to both . in a detached system such as this
the photometry does not constrain the mass ratio , so the value of @xmath74 was adopted from the spectroscopy .
the parameters representing the spots in the wd model are the latitude ( @xmath75 ) , longitude ( @xmath76 ) , angular radius ( @xmath77 ) and temperature contrast relative to the photosphere ( @xmath78 ) .
only cool spots have been considered .
the latitude is essentially unconstrained by our data , and the spot size and temperature factor are strongly correlated with each other and can not usually be adjusted simultaneously .
thus , only the longitude and the spot size were adjusted at the same time as the geometric parameters , while the latitude and contrast factor were held fixed and estimated through a grid search .
we explored values over a wide range in @xmath75 from + 80 to @xmath1180 in steps of 20 , a range in the primary @xmath78 from 0.85 to 0.95 , and a range from 0.75 to 0.85 for the secondary @xmath78 , both with a step size of 0.05 .
as indicated above , it is virtually impossible to tell which star has the spot(s ) , so we have considered here only three possible scenarios : two spots on the primary , two on the secondary , or one spot on each star .
lcc + phase shift & + @xmath68 ( deg ) & + @xmath79 & 6.298 @xmath36 0.023 & 7.081 @xmath36 0.038 + @xmath80 & 0.1776 @xmath36 0.0007 & 0.1145 @xmath36 0.0007 + @xmath81 & 0.1796 @xmath36 0.0008 & 0.1151 @xmath36 0.0008 + @xmath82 & 0.1784 @xmath36 0.0007 & 0.1147 @xmath36 0.0007 + @xmath83 & 0.1776 @xmath36 0.0008 & 0.1150 @xmath36 0.0007 + @xmath84 & 0.1785 @xmath36 0.0008 & 0.1146 @xmath36 0.0010 + + @xmath16 ( k ) & 5570 & 4246 @xmath36 16 + @xmath85 @xmath54 band & + @xmath85 @xmath55 band & + @xmath85 @xmath56 band & + @xmath85 @xmath57 band & + albedo & + gravity darkening & 0.34 & 0.40 + + @xmath86 @xmath54 band & 0.625 & 0.904 + @xmath87 @xmath54 band & 0.240 & @xmath110.057 + @xmath86 @xmath55 band & 0.364 & 0.644 + @xmath87 @xmath55 band & 0.450 & 0.190 + @xmath86 @xmath56 band & 0.237 & 0.462 + @xmath87 @xmath56 band & 0.517 & 0.354 + @xmath86 @xmath57 band & 0.137 & 0.269 + @xmath87 @xmath57 band & 0.539 & 0.475 + + latitude ( deg ) & @xmath1160 & @xmath1160 + longitude ( deg ) & 307.3 @xmath36 4.9 & 331.8 @xmath36 4.7 + radius ( deg ) & 26.4 @xmath36 1.2 & 36.0 @xmath36 2.5 + @xmath16 factor & 0.95 & 0.80 + + @xmath88 / @xmath89 & + @xmath90 / @xmath91 & + @xmath92 / @xmath93 & + @xmath94 / @xmath95 & observations of i m vir along with our best - fit model .
the curves are shifted vertically for clarity .
photometric residuals are shown in the bottom panels , in the same order as the top curves . ] .
note the different vertical scales for the primary and secondary eclipses . ]
l c cccc c ccc c cc phase shift & & 0.0007 & 0.0007 & 0.0006 & 0.0004 & & 0.00014 & 0.00010 & 0.00010 & & 0.0006 & 0.0002 + @xmath68 ( deg ) & & 87.38 & 87.22 & 86.91 & 87.83 & & 0.38 & 0.08 & 0.16 & & 87.24 & 0.42 + @xmath71 & & 4272 & 4191 & 4231 & 4304 & & 49 & 56 & 16 & & 4246 & 83 + @xmath96 & & 6.257 & 6.289 & 6.304 & 6.431 & & 0.077 & 0.018 & 0.023 & & 6.298 & 0.082
+ @xmath97 & & 7.091 & 7.080 & 7.003 & 7.220 & & 0.090 & 0.044 & 0.038 & & 7.08 & 0.11 + @xmath98 & & 0.1798 & 0.1788 & 0.1783 & 0.1743 & & 0.0024 & 0.0006 & 0.0008 & & 0.1785 & 0.0026 + @xmath99 & & 0.1144 & 0.1146 & 0.1161 & 0.1119 & & 0.0017 & 0.0009 & 0.0010 & & 0.1146 & 0.0022 + @xmath85 ( @xmath54 ) & & 0.04999 & & & & & & 0.0041 & 0.00021 & & 0.0480 & 0.0041 + @xmath85 ( @xmath55 ) & & & 0.05366 & & & & & 0.0054 & 0.00030 & & 0.0750 & 0.0054 + @xmath85 ( @xmath56 ) & & & & 0.10928 & & & & 0.0065 & 0.00042 & & 0.1096 & 0.0065 + @xmath85 ( @xmath57 ) & & & & & 0.15179 & & & 0.0068 & 0.00050 & & 0.1427 & 0.0068 the best solutions were typically obtained with the spots located in the southern hemisphere , although we do not assign any particular physical significance to this as we only consider the spot modeling as a means of removing a perturbation from the underlying light curve .
the overall best fit has one spot on each component , and a reduced @xmath34 square that is only marginally better than the other two scenarios ( 1.5% lower than the case with two spots on the primary , and 3.4% lower than the two - spot configuration on the secondary ) .
the parameters of this best fit are listed in table [ tab : lc_param ] , and the synthetic curves are shown together with the observations in figure [ fig : lc_plot ] and figure [ fig : eclipses_plot ] . among the quantities derived from this fit ,
the light ratios allow for an important consistency check against the ratio obtained directly from our spectra .
interpolating between @xmath54 and @xmath55 to the mean wavelength of our spectroscopic observations , we obtain @xmath100 , which agrees well with the value of @xmath31 from [ sec : spectroscopy ] .
the effect of the spots on the light curves is illustrated in figure [ fig : spots_lc ] , and is seen to range from @xmath80.025 mag in @xmath54 to @xmath80.015 mag in @xmath57 ( peak - to - peak ) .
a depiction of the location of the spots on each star is shown in figure [ fig : spots_map ] .
the spot on the primary covers 5.2% of its surface , while that of the secondary covers 9.5% . in order to provide more realistic uncertainties for the geometric and radiative parameters than the internal errors reported in table [ tab : lc_param ] , we have considered additional sources of error as follows : ( a ) we ran wd solutions separately for each passband , with the spot parameters fixed at their final values , and examined the differences in the parameters .
these fits are summarized in table [ tab : lc_fit ] .
the dispersion in each parameter about the average of the four bands ( @xmath101 ) was taken as an additional contribution to the overall uncertainty .
( b ) we calculated the range in each parameter from the simultaneous @xmath0 fits in the three spot scenarios , and adopted half of this range as an additional contribution to account for the degeneracy in the spot modeling .
( c ) as a check on the internal errors from wd , we continued the iterations in our adopted fit beyond convergence for another 200 steps , and we examined the scatter of those 200 solutions . for the main parameters ( @xmath68 , @xmath71 , @xmath69 , @xmath70 ) the scatter was only a small fraction of the internal errors , but for the light ratios it was typically a factor of two larger .
we adopted the larger of the two estimates in each case .
these three sources of error ( @xmath101 , @xmath102 , and @xmath103 in table [ tab : lc_fit ] ) were combined quadratically , and are the ones we assign to the final light - curve parameters .
changes in the adopted mass ratio within its error have no additional effect .
the parameters and these errors , which we believe to be realistic , are listed in the last two columns of table [ tab : lc_fit ] .
further solutions were carried out allowing the eccentricity to vary , but in all cases we found the result to be insignificant compared to its error , consistent with the indications from spectroscopy .
third light was also tested for at various stages of the analysis , but was always found to either converge toward negative ( unphysical ) values , or to be consistent with zero .
different limb - darkening laws were tested as well ( linear , logarithmic ) , but the differences with the results in table [ tab : lc_fit ] were minimal ( well within the errors ) .
the final solution indicates the two stars are nearly spherical , the difference between @xmath104 and @xmath105 being only 1.1% for the primary and 0.5% for the secondary .
the primary eclipse is annular ( 48% of the light of that star blocked ) , while the secondary is total ( see figure [ fig : eclipses_plot ] ) , with the totality phase lasting 28 minutes . the second half of our photometry , not used here because of likely changes in the spots and more incomplete phase coverage , is compared with our final model for the geometry in figure [ fig : lc_season2 ] , excluding the spot terms .
the larger scatter is obvious , but the fit is still quite reasonable .
figure [ fig : lc_manfroid ] displays our fits to the sparser strmgren photometry by @xcite .
brightness ratios from these solutions ( 0.0167 , 0.0353 , 0.0599 , and 0.0775 in @xmath66 , respectively ) are used below to separate the light of the two components and derive a photometric estimate of their effective temperatures and metallicity .
photometry of i m vir corresponding to the second half of our data set , which we have not included in the analysis for the reasons described in the text .
the solid curves are the same best - fit model shown in figure [ fig : lc_plot ] ( without the spots ) , adjusted for a slight overall brightness change and a phase shift . ]
photometry on the standard system published by @xcite , but not used in our analysis , compared with our best - fit model . ]
the absolute masses and radii for the components of i m vir follow from our spectroscopic and light curve solutions described in the preceding sections .
the effect that spots may have on the relative radii has been considered explicitly in our light - curve modeling , and included in our error estimates .
similar effects may influence the masses , although we do not expect them to be significant since they would tend to average out over the extended coverage of our spectroscopy .
nevertheless , the spot models in [ sec : lc_analysis ] allow us to quantify the importance of these effects . figure [ fig : spots_rv ] shows the distortions in the radial velocities expected from each of the three spot scenarios considered earlier . as a test
, we perturbed the radial velocities using these curves and repeated the spectroscopic orbital solutions . compared to our adopted fit ( table [ tab : specorbit ] ) ,
the differences were less than 0.6% in the minimum masses .
the differences in @xmath106 , which affects the absolute radii , were less than 0.2% . to be conservative , for the calculation of the final mass and radius uncertainties
we have augmented the errors in table [ tab : specorbit ] by adding in quadrature half of the range in @xmath107 and @xmath106 obtained from the three spot configurations . with this , the absolute masses of i m vir are determined to 1.2% and 0.7% for the primary and secondary , and the absolute radii to 1.5% and 1.9% .
next in importance to the masses and radii are the effective temperatures of the stars .
the temperature _ difference _ ( or ratio ) is accurately determined from the light curves , but the absolute scale is set by the primary value , which is fixed in our analysis . as a check on the spectroscopic value of 5570 k adopted here , we made use of absolute photometry available for i m vir from a variety of sources , along with color / temperature calibrations .
brightness measurements were compiled from the 2mass catalog @xcite , the _
tycho-2 _ catalog @xcite , strmgren photometry from @xcite and @xcite , and our own @xmath108 measurements outside of eclipse , transformed to the standard system using the comparison and check stars .
for the @xcite photometry we considered only the data out of eclipse .
the _ tycho-2 _ and @xcite measurements are quite numerous and can be considered to be little affected by the minima .
the date of the 2mass measurements indicates they were taken during the egress of a secondary eclipse ; corrections of order 0.03 mag ( slightly different in each band ) were applied to @xmath109 based on our light curve solutions , extended to the near - infrared .
the adopted magnitudes for the combined light in all passbands are listed in table [ tab : colors ] . with these we formed 8 different color indices ,
also listed in the table , and then deconvolved the light to obtain the individual indices for the two stars .
the light ratios in @xmath54 and @xmath55 for the deconvolution were adopted from table [ tab : lc_fit ] , and those for the strmgren passbands from our fits in the previous section .
for the other passbands we used model isochrones ( @xcite for the _ tycho-2 _ bands , and @xcite for the @xmath109 bands , after conversion to the 2mass system ) .
two different isochrone ages were tried ( 1 gyr and 5 gyr ) , although the differences were well within the errors .
color / temperature calibrations from @xcite , @xcite , and @xcite were then applied to the primary indices , assuming solar metallicity .
although the color indices are not independent of each other , they do serve to test the internal consistency between the various photometric systems and the three different calibrations .
the interagreement is very good ( typically within 150 k ) .
the weighted average is @xmath110 k when using the 1 gyr isochrones for the light ratios , and 40 k less when adopting 5 gyr .
we regard these to be in very good agreement with the spectroscopic value , considering the uncertainties . changing the metallicity for the calibrations to the rough estimate from [ sec : spectroscopy ] ( [ m / h ] @xmath111 ) reduces the photometric temperatures by less than 10 k. a similar exercise using the indices for the secondary and an additional calibration for cool stars by @xcite gives a more uncertain result of @xmath112 k ( as expected from the faintness of the star ) , but still consistent with the much more accurate value based on the light curve analysis , which is @xmath6 k.
these tests suggest that the adopted temperatures for i m vir are accurate .
they correspond to spectral types of for the primary and for the secondary .
lc + @xmath54 & 10.234 @xmath36 0.030 + @xmath55 & 9.574 @xmath36 0.030 + @xmath113 & 10.209 @xmath36 0.021 + @xmath87 & 9.748 @xmath36 0.017 + @xmath114 & 10.483 @xmath36 0.039 + @xmath115 & 9.768 @xmath36 0.030 + @xmath116 & 8.176 @xmath36 0.020 + @xmath117 & 7.712 @xmath36 0.025 + @xmath118 & 7.634 @xmath36 0.024 + + @xmath119 & 0.661 @xmath36 0.028 + @xmath120 & 0.465 @xmath36 0.032 + @xmath121 & 1.372 @xmath36 0.028 + @xmath122 & 1.832 @xmath36 0.030 + @xmath123 & 1.905 @xmath36 0.031 + @xmath124 & 0.715 @xmath36 0.057 + @xmath125 & 2.099 @xmath36 0.038 + @xmath126 & 0.533 @xmath36 0.031 an additional quantity of great importance for the interpretation of the masses , radii , and temperatures is the chemical composition . beyond our estimate in [ sec : spectroscopy ] of [ m / h ] @xmath127 , a spectroscopic estimate from a composite spectrum of i m vir was reported by @xcite as [ fe / h ] = @xmath128 , based on an effective temperature some 200 k cooler and a higher surface gravity than we derive for the primary , which dominates the light .
it is unclear how accurate this determination is , in view of those differences .
a photometric estimate for the primary may be derived from the out - of - eclipse strmgren measurements of i m vir by @xcite and @xcite , after removing the light contribution of the secondary using the light ratios obtained in [ sec : lc_analysis ] .
the metallicity relation by @xcite gives a rather poorly determined value of [ fe / h ] @xmath129 , in which the uncertainty includes photometric errors as well as the scatter of the calibration .
the secondary is too cool for this calibration and other similar ones based on strmgren indices , but is within range of the near - infrared formula by @xcite , which yields [ fe / h ] @xmath130 .
this again includes all photometric errors and the scatter of the calibration .
it has been noted by @xcite that this latter color / temperature relation appears to underestimate the metallicity of late - type stars of solar composition or greater , by approximately 0.3 dex ; for sub - solar compositions it is not clear that a correction is necessary .
the above estimates suggest a composition of i m vir somewhat below solar , perhaps [ fe / h ] @xmath131 , but the uncertainties are large and this conclusion requires confirmation .
lcc @xmath132 ( @xmath44 ) & 0.981 @xmath36 0.012 & 0.6644 @xmath36 0.0048 + @xmath56 ( @xmath133 ) & 1.061 @xmath36
0.016 & 0.681 @xmath36 0.013 + @xmath16 ( k ) & 5570 @xmath36 100 & 4250 @xmath36 130 + @xmath17 ( cgs ) & 4.379 @xmath36 0.014 & 4.594 @xmath36 0.017 + @xmath134 & @xmath110.012 @xmath36 0.034 & @xmath110.867 @xmath36 0.056 + @xmath135 ( mag ) & @xmath110.12 @xmath36
0.10 & @xmath110.82 @xmath36 0.17 + @xmath136 ( mag ) & 4.88 @xmath36 0.15 & 7.71 @xmath36 0.29 + @xmath18 ( ) & 43 @xmath36 2 & + @xmath137 ( ) & 41.0 @xmath36 0.6 & 26.3 @xmath36 0.5 + @xmath138 ( @xmath133 ) & + distance ( pc ) & the absolute dimensions for the system are summarized in table [ tab : absolute ] , along with derived properties including the luminosities and absolute visual magnitudes . to calculate @xmath136
we have adopted the bolometric corrections from @xcite , with uncertainties that account for the temperature errors as well as a contribution of 0.10 mag added in quadrature , to be conservative .
i m vir does not have an entry in the _ hipparcos _ catalog @xcite .
here we estimate the distance to the system to be @xmath139 pc , ignoring extinction for such a close object .
this is 50% larger than first estimated by @xcite on the basis of the spectral type .
separate distance estimates for each component agree nearly perfectly , showing the internal consistency of the fundamental data on which they are based .
the distance , proper motions from _ tycho-2 _ , and systemic velocity from our spectroscopic orbital solution ( table [ tab : specorbit ] ) lead to space velocities of @xmath140 , @xmath141 , and @xmath142 in the solar frame ( with @xmath143 positive toward the galactic center ) .
these do not appear to associate i m vir with any known moving group in the solar neighborhood . with our accurate radii
we compute projected synchronous rotational velocities of @xmath144 and @xmath145 for the primary and secondary , respectively .
the value for the primary is consistent with our direct measurement of @xmath146 , suggesting that synchronization with the orbital motion has been achieved ( if the spin axis is parallel to the orbital axis ) .
this is expected from the short period .
estimates of 43 for the primary and 23 for the secondary were reported by @xcite , but without errors . at our request
, those spectra were kindly remeasured by f.fekel ( priv .
comm . ) , giving @xmath147 and @xmath148 , based on the calibration by @xcite .
once again the primary agrees well with the synchronous value , and the secondary is probably also consistent , considering the difficulty of the measurement .
an independent value for the primary was reported by @xcite as 36.2 with an uncertainty of 24 , roughly in agreement with ours .
with its solar - type primary and low - mass secondary , i m vir is a particularly interesting system for testing models of stellar evolution .
the accurate masses , radii , and temperatures for both components offer an opportunity to further investigate the discrepancies for low - mass stars mentioned in the introduction .
the leverage afforded by the very different masses is unique among systems with at least one component under 0.8 @xmath44 .
we begin by comparing the measured properties against models from the widely used yonsei - yale series @xcite , which treat convection in the standard mixing - length approximation with a mixing length parameter @xmath149 ( in units of the pressure scale height ) , calibrated against the sun .
evolutionary tracks for solar composition and for the exact masses we measure for each star are shown in the @xmath17@xmath16 diagram of figure [ fig : yaletracks ] , as dashed lines .
these models are seen to be too hot compared to the estimated temperatures . adjusting the metallicity to higher values brings the tracks closer to the observations , but it is not possible to match both stars at the same time , especially considering that the temperature _ difference _ is much better known than the absolute temperatures .
part of the problem has to do with the fact that these models are not intended for low - mass stars such as i m vir b , which require a more sophisticated equation of state , and particularly , non - gray boundary conditions between the interior and the photosphere ( see , e.g. , * ? ? ?
if we therefore focus for the moment only on the solar - type primary , we find an excellent match to an evolutionary track for @xmath150 ( corresponding to [ fe / h ] @xmath151 in these models ) , at the rather old age of about 8 gyr .
the corresponding mass tracks for both stars are drawn with solid lines in figure [ fig : yaletracks ] , and the shaded area indicates the uncertainty in their location that comes from our @xmath81% mass errors .
an isochrone for this age is indicated with the dotted line .
that fits the primary best .
the mass errors are indicated by the shaded areas .
the implied age for the primary is 8 gyr , and the corresponding isochrone is shown as the dotted line .
the track for the secondary is calculated through the end of the main - sequence phase , which is reached at an age of @xmath840 gyr .
the age of the universe is marked with an asterisk on this track .
the observed location of the secondary would thus point to an implausibly old age for the star .
this is a result of the models overestimating the temperature and underestimating the size for this star . ] not surprisingly , the secondary does not fit the evolutionary track at this metallicity , and appears implausibly old .
this is a consequence of it being both too large and too cool compared to the models , which is in the same sense as deviations found for many other late - type dwarfs , and is believed to be due to the effects of chromospheric activity .
previous studies have indicated that better agreement for this class of stars is possible with models such as those of @xcite , which have the required non - gray boundary conditions and use a lower mixing - length parameter of @xmath152 .
the latter tends to diminish ( but does not completely eliminate ) the discrepancies noted with other models in @xmath56 and @xmath16 .
the one property of low - mass stars that appears to be reasonably well reproduced by theory is the luminosity ( see , e.g. , * ? ? ?
* ; * ? ? ?
this provides a means of testing the predictions from the yonsei - yale models for the primary , which point to an unexpected combination of super - solar metallicity and old age , and a composition that is somewhat inconsistent with the indications from [ sec : absolute ] . in the left panels of figure [ fig : testyale ]
we compare the secondary luminosity , radius , and temperature against isochrones from @xcite corresponding to the age of 8 gyr found above , for different compositions including the value [ fe / h ] @xmath151 that best fits the primary ( according to the yonsei - yale models ) .
= 0.0 and @xmath153 . ]
all three of @xmath154 , @xmath56 , and @xmath16 for the secondary are seen to be too large compared to the model favored by the primary , which is represented by the solid lines .
the measured luminosity of the secondary , as well as its effective temperature , suggest a significantly lower abundance , and the measured radius is too large no matter what the metallicity .
changes in age at a fixed metallicity of [ fe / h ] = + 0.15 do not improve the fit , as seen clearly in the right panels of figure [ fig : testyale ] . ) .
_ left panels : _ fixed age of 8 gyr , and different metallicities , as labeled . _
right panels : _ fixed abundance of [ fe / h ] = + 0.15 , and different ages , as labeled . ] there is little doubt that the radius of the secondary is too large compared to models .
furthermore , the age / metallicity inconsistency between the primary and secondary is perhaps an indication that the good match for i m vir a in figure [ fig : yaletracks ] is illusory , and that it too may have anomalies similar to those of the secondary , possibly related also to activity .
the primary is , after all , a rapid rotator ( @xmath155 ) .
standard models such as those of @xcite do not account for the effects of chromospheric activity , although the artificially low value of the mixing length parameter ( @xmath156 ) seems to be a step in the right direction ( see also * ? ? ?
* ; * ? ? ?
as a way of parameterizing the missing physical effects from activity and their impact on the structure of low - mass stars , @xcite explored the use of a correction factor @xmath157 to the theoretical radii , and showed that good fits to empirical data could be achieved by simultaneously correcting the theoretical temperatures by @xmath158 , so as to preserve the bolometric luminosity .
we apply the same procedure here . to account for a possible difference in the activity level
, we have considered separate values of @xmath157 for each star , and examined a wide range of metallicities and ages using the @xcite models to obtain the best simultaneous match to the properties of both components , at a single age and composition . in figure
[ fig : betagrid ] we display the results of our grid search .
a near - perfect fit is achieved for [ fe / h ] @xmath159 , as measured by the @xmath34 represented in the top panel .
interestingly , this value is much more consistent with the rough estimates from photometry and spectroscopy described in
[ sec : absolute ] than with the metal - rich composition suggested by the yonsei - yale models for the primary .
the radius correction factors are @xmath160 and @xmath161 , and the best - fit age is 2.4 gyr ( see middle and bottom panels ) .
the age and metallicity depend almost entirely on the measured stellar luminosities .
the values of @xmath162 , @xmath56 , and @xmath16 are shown in figure [ fig : betamodel ] with the adjusted isochrones from @xcite .
standard models without the corrections to the radii and temperatures are shown as dashed lines in the lower panels , for reference . and
@xmath163 to the theoretical radii to account for the effects of chromospheric activity ( see text ) .
_ top : _ the @xmath34 of the fit as a function of metallicity for the nominal effective temperatures of the stars ( solid line with filled circles ) , and for temperatures perturbed by @xmath36100 k to explore the role of systematics ( dashed lines ) .
_ middle : _ correction factors to the predicted radii .
_ bottom : _ best - fit age as a function of metallicity . ] @xmath159 and an age of 2.4 gyr ( solid lines ) , with the radii adjusted by the correction factors @xmath164 and @xmath163 for the primary and secondary , respectively ( see figure [ fig : betagrid ] and text ) .
the theoretical temperatures are adjusted by factors @xmath158 , and the model luminosity is unchanged .
the dashed lines in the two bottom panels represent the standard models for the same age and metallicity , but without any corrections for activity ( @xmath165 ) .
] the implication of the above fit is that the radius of the primary star is 3.7% larger than predicted by theory , and its temperature is just under 2% ( 100
k ) too cool . for the secondary the models underpredict the radii by 7.5% , and overpredict the temperature by about 3.5% , or 150 k. it appears , therefore , that _ both _ stars are affected by chromospheric activity . while this is not unexpected for the secondary , the fact that the primary also seems to show the same anomalies supports other recent evidence that the phenomenon is not confined to low - mass stars , but reaches stars of solar mass ( 0.981 @xmath44 for i m vir a ) . other examples of solar - type stars where this has been documented include vz cep b ( 1.108 @xmath44 ; * ? ? ?
* ) , cv boo b ( 0.968 @xmath44 ; * ? ? ?
fl lyr b ( 0.958 @xmath44 ; * ? ? ?
* ) , and v1061 cyg b ( 0.932 @xmath44 ; * ? ? ?
there is ample evidence of activity in i m vir , which was in fact discovered through its x - ray emission , as described in
[ sec : introduction ] .
activity is manifested in our light curves in the form of spottedness , and in retrospect , our finding that the best fit to the photometry is achieved with one spot on each star as opposed to spots on any single component is consistent with the results of the stellar evolution modeling in the preceding paragraph , which indicates that both stars are affected .
this is not surprising , given that both have convective envelopes and are rapid rotators .
additional manifestations of activity in this case include a filled - in h@xmath9 line @xcite or the possible detection of h@xmath9 emission from the secondary @xcite , h and k emission @xcite , and flaring in x rays @xcite . with the x - ray count rates and hardness ratio from _ rosat _ , the energy conversion factor prescribed by @xcite , and
our bolometric luminosities and derived distance , we have estimated the ratio of the x - ray to bolometric luminosity for the two components .
we obtain @xmath166 and @xmath167 , respectively , where we have assumed equal x - ray emission from each star since the _ rosat _ observation does not resolve the binary .
these values are consistent with the secondary being completely saturated , and the primary being near saturation .
they are similar to the x - ray luminosities seen in other active binary systems , and are in fact in good agreement with the trend between the radius anomalies and @xmath168 found by @xcite ( see , e.g. , her figure 2 ) .
the differences in radius and temperature between the models and the observations for i m vir are of similar magnitude as those found for other active stars with convective envelopes , but do rely quite strongly on our adopted effective temperatures for the stars because of the way the @xmath157 correction factors have been determined . given the greater difficulty of determining absolute temperatures than other properties such as masses and radii , one may wonder to what extent systematic errors in @xmath16 might affect the result , and also the derived age and metallicity .
we focus here on systematic errors in the primary temperature , since the secondary value is tied to the primary through the light - curve solution .
we explored this question by perturbing the temperatures by @xmath36100 k , with corresponding adjustments to the luminosities , and repeating the grid search for @xmath157 .
the results are illustrated in the top panel of figure [ fig : betagrid ] , where the dashed lines show that the best - fit metallicity changes by @xmath360.10 dex ( marked by the short dotted lines ) .
the changes in the @xmath157 factors are @xmath360.010 for @xmath164 and @xmath360.003 for @xmath163 , and the change in the best - fit age is @xmath360.50 gyr , or @xmath820% .
our extensive spectroscopic and photometric observations of i m vir have enabled us to determine highly accurate values for the absolute masses and radii of both stars to better than 2% , as well as accurate temperatures ( table [ tab : absolute ] ) .
this eclipsing system now joins the ranks of those with the best determined properties ( see * ? ? ?
the primary is of solar type , and the secondary is a late k dwarf .
the very different masses provide increased discriminating power for testing models of stellar evolution .
we find evidence that both components , which have convective envelopes and are rapidly rotating , show discrepancies in their radii and temperatures compared to calculations by @xcite , similar to those reported for other low - mass stars .
the predicted radii are too small by 3.7% and 7.5% for the primary and secondary , and the theoretical temperatures are too high by 100 k and 150 k , respectively .
these effects are ascribed to chromospheric activity , for which there is abundant evidence in this system in the form of x - ray , h and k , and h@xmath9 emission , x - ray flaring , and spottedness seen directly in the light curves . the larger effect observed for
the secondary is consistent with the larger fraction of that star being covered by spots ( a factor of two difference ) .
the fact that the near - solar mass primary is also affected supports recent findings for other binaries indicating that the impact of activity on the structure of stars is not limited to the m dwarfs , as previously thought , but most likely extends to all stars with convective envelopes , reaching up to masses near or larger than that of the sun .
current stellar evolution models do not account for these effects .
some progress in this area has been made to incorporate magnetic activity in the theoretical calculations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and initial results are very encouraging .
a crucial ingredient for the comparison with models is the chemical composition .
the best - fit models for i m vir indicate an age of @xmath169 gyr and a metallicity of [ fe / h ] @xmath170 .
the uncertainties here represent the sensitivity to systematic errors of @xmath36100 k in the effective temperatures , but exclude unquantified systematics in the @xcite models themselves , which are difficult to evaluate . while the metallicity prediction seems to be in good agreement with our rough spectroscopic abundance estimate in
[ sec : spectroscopy ] and that of @xcite , as well as photometric estimates for both stars , a proper detailed spectroscopic analysis of both components is still lacking , and is essential to validate the comparison .
such determinations are notoriously difficult in double - lined binaries , which is why relatively few have them . in i
m vir this task would be challenging for two reasons : the secondary is very faint compared to the primary , and it is of late spectral type .
abundance determinations for late - type stars are still problematic due to shortcomings in the model atmospheres .
however , i m vir offers a unique opportunity because the secondary eclipse is total .
high s / n ratio spectra taken with a sufficiently large telescope during the 28-minute totality phase would be of the primary only , and can be analyzed with standard techniques . the uncertainty in this observing window is estimated to be approximately @xmath365 minutes , based on the uncertainties in the geometric light - curve parameters .
we are grateful to i. ribas and c. jordi for useful discussions on the implementation of the wilson - devinney code .
we also thank f.fekel for communicating his @xmath18 measurements of i m vir , and the anonymous referee for helpful suggestions .
the spectroscopic observations used in this work were obtained with the able help of p. berlind , m. calkins , j. caruso , g. esquerdo , r. davis , e. horine , d. latham , r. mathieu , d. silva , j.stauffer , r. stefanik , s. tokarz , and j. zajac .
we also thank r.davis for maintaining the cfa echelle database .
jcm acknowledges financial support from the spanish ministerio de ciencia e innovacin during the research stay at the harvard - smithsonian center for astrophysics , where most of this work was done , and from the spanish ministerio de educacin y ciencia via grants aya2006 - 15623-c02 - 01 and aya2006 - 15623-c02 - 02 .
gt acknowledges partial support for this work from nsf grant ast-0708229 .
lm and wb were supported by gettysburg college .
additional thanks go to david kraft for preliminary data reductions , and to peter mack and garry hummer for technical support at the gettysburg college observatory .
this research has made use of the simbad database and the vizier catalogue access tool , both operated at cds , strasbourg , france , of nasa s astrophysics data system abstract service , and of data products from the two micron all sky survey ( 2mass ) , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by nasa and the nsf .
we also made use of data provided by the vsop collaboration , through the vsop wiki database operated at eso chile and eso garching .
andersen , j. 1991 , , 3 , 91 baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p. h. 1998 , , 337 , 403 bonfils , x. , delfosse , x. , udry , s. , santos , n. c. , forveille , t. , & sgransan , d. 2005 , , 442 , 635 casagrande , l. , portinari , l. , & flynn , c. 2006 , , 373 , 13 casagrande , l. , flynn , c. , & bessell , m. 2008 , , 389 , 585 chabrier , g. , gallardo , j. , & baraffe , i. 2007 , , 472 , l17 chabrier , g. , & baraffe , i. 1997 , , 327 , 1039 claret , a. 2000a , , 363 , 1081 claret , a. 2000b , , 359 , 289 clausen , j. v. , torres , g. , bruntt , h. , andersen , j. , nordstrm , b. , stefanik r. p. , latham , d. w. , & southworth , j. 2008 , , 487 , 1095 clausen j. v. , bruntt h. , claret a. , larsen a. , andersen j. , nordstrm b. , gimnez a. 2009 , , 502 , 253 cutri , r. m. et al .
2003 , 2mass all - sky catalog of point sources , univ . of massachusetts and infrared processing and analysis center ( ipac / california institute of technology )
dall , t. h. et al .
2007 , , 470 , 1201 dantona f. , ventura p. , & mazzitelli i. 2000 , , 543 , l77 delfosse , x. , forveille , t. , sgransan , d. , beuzit , j .- l . ,
udry , s . ,
perrier , c. , & mayor , m. 2000 , , 364 , 217 demarque , p. , woo , j .- h . ,
kim , y .- c .
, & yi , s. k. 2004 , , 155 , 667 eker , z. 1996 , , 473 , 388 eker , z. 1999 , , 512 , 386 fekel , f. c. 1997 , , 109 , 514 flower , p. j. 1996 , , 469 , 355 girardi , l. , bressan , a. , bertelli , g. , & chiosi , c. 2000 , , 141 , 371 gonzlez hernndez , j. i. , & bonifacio , p. 2009
, , 497 , 497 griffin , r. e. m. , david , m. , & verschueren , w. 2000 , , 147 , 299 helfand , d. j. , & caillault , j .- p . 1982 , , 253 , 760 hg , e. , fabricius , c. , makarov , v. v. , urban , s. , corbin , t. , wycoff , g. , bastian , u. , schwekendiek , p. , & wicenec , a. 2000 , , 355 , l27 holmberg , j. , nordstrm , b. , & andersen , j. 2007 , , 475 , 519 johnson , j. a. , & apps , k. 2009 , , 699 , 933 lacy , c. h. s. , torres , g. , & claret , a. 2008 , , 135 , 1757 latham , d. w. 1985 , in iau coll .
88 , stellar radial velocities , eds .
a. g. d. philip & d. w. latham ( schenectady : l. davis ) , 21 latham , d. w. 1992 , in iau coll .
135 , complementary approaches to double and multiple star research , asp conf .
32 , eds . h. a. mcalister & w. i. hartkopf ( san francisco : asp ) , 110 latham , d. w. , nordstrm , b. , andersen , j. , torres , g. , stefanik , r. p. , thaller , m. , & bester , m. 1996 , , 314 , 864 latham , d. w. , stefanik , r. p. , torres , g. , davis , r. j. , mazeh , t. , carney , b. w. , laird , j. b. , & morse , j. a. 2002 , , 124 , 1144 liu , x. , huang , h. , & zhu , w. 1996 , , 246 , 39 lpez - morales , m. 2007 , , 660 , 732 lpez - morales , m. , & ribas , i. 2005 , , 631 , 1120 manfroid , j. et al .
1991 , , 87 , 481 marschall , l. a. , nations , h. , & witman , k. 1988 , , 20 , 994 marschall , l. a. , stefanik , r. , nations , h. , & karshner , g. 1989 , , 21 , 1083 morale , f. , micela , g. , favata , f. , & sciortino , s. 1996 , , 119 , 403 morales , j. c. , et al .
2009 , , 691 , 1400 mullan d. j. , & macdonald , j. 2001 , , 559 , 353 ogloza , w. , niewiadomski , w. , barnacka , a. , biskup , m. , malek , k. , & sokolowski , m. 2008 , ibvs , 5843 , 1 pandey , j. c. , & singh , k. p. 2008
, , 387 , 1627 perryman , m. a. c. , et al . 1997 , the _ hipparcos _ and _ tycho _ catalogues ( esa sp-1200 ; noordwjik : esa ) popper , d. m. 1996 , , 106 , 133 popper , d. m. 1997 , , 114 , 1195 popper , d. m. 2000 , , 119 , 2391 popper , d. m. , lacy , c. h. , frueh , m. l. , & turner , a. e. 1986 , , 91 , 383 ramrez , i. , & melndez , j. 2005 , , 626 , 446 ribas , i. 2003 , , 398 , 239 ribas , i. , morales , j. c. , jordi , c. , baraffe , i. , chabrier , g. , & gallardo , j. 2008 , , 79 , 562 schmitt , j. h. m. m. , fleming , t. a. , & giampapa , m. s. 1995 , , 450 , 392 siess , l. , dufour , e. , & forestini , m. 2000 , , 358 , 593 silva , d. r. , gioia , i. m. , maccacaro , t. , mereghetti , s. , & stocke , j. t. 1987 , , 93 , 869 strassmeier , k. g. , hall , d. s. , fekel , f. c. , & scheck , m. 1993 , , 100 , 173 strassmeier , k. g. , washuettl , a. , granzer , th . , scheck , m. , & weber , m. 2000 , , 142 , 275 torres , g. 2007 , , 671 , l65 torres , g. , andersen , j. , & gimnez , a. 2009 , , in press ( arxiv:0908.2624v1 ) torres , g. , andersen , j. , nordstrm , b. , & latham , d. w. 2000 , , 119 , 1942 torres , g. , & lacy , c. h. s. 2009 , , 137 , 507 torres g. , lacy c. h. s. , marschall l. a. , sheets h. a. , mader j. a. 2006 , , 640 , 1018 torres , g. , neuhuser , r. , & guenther , e. w. 2002 , , 123 , 1701 torres , g. , & ribas , i. 2002 , , 567 , 1140 torres , g. , stefanik , r. p. , andersen , j. , nordstrm , b. , latham , d. w. , & clausen .
j. v. 1997 , , 114 , 2764 torres , g. , vaz , l. p. r. , & lacy , c. h. s. 2008 , , 136 , 2158 van hamme , w. , & wilson , r. e. 2003 , in gaia spectroscopy : science and technology , asp conf .
298 , ed .
u. munari ( san francisco : asp ) , p. 323
wilson , r. e. 1979 , , 234 , 1054 wilson , r. e. , & devinney , e. j. 1971 , , 166 , 605 yi , s. k. , demarque , p. , kim , y .- c . , lee , y .- w . ,
ree , c. h. , lejeune , t. , & barnes , s. 2001 , , 136 , 417 zucker , s. , & mazeh , t. 1994 , , 420 , 806 | we report extensive spectroscopic and differential photometric @xmath0 observations of the active , detached , 1.309-day double - lined eclipsing binary i m vir , composed of a g7-type primary and a k7 secondary . with these observations
we derive accurate absolute masses and radii of @xmath1 , @xmath2 , @xmath3 , and @xmath4 for the primary and secondary , with relative errors under 2% .
the effective temperatures are @xmath5 k and @xmath6 k. the significant difference in mass makes this a favorable case for comparison with stellar evolution theory .
we find that both stars are larger than the models predict , by 3.7% for the primary and 7.5% for the secondary , as well as cooler than expected , by 100 k and 150 k , respectively .
these discrepancies are in line with previously reported differences in low - mass stars , and are believed to be caused by chromospheric activity , which is not accounted for in current models . the effect is not confined to low - mass stars : the rapidly - rotating primary of i m vir joins the growing list of objects of near - solar mass ( but still with convective envelopes ) that show similar anomalies .
the comparison with the models suggests an age of 2.4 gyr for the system , and a metallicity of [ fe / h ] @xmath7 that is consistent with other indications , but requires confirmation . |
gravitational microlensing surveys were originally suggested by @xcite as a means of detecting dark matter in the galaxy in the form of massive compact objects , commonly abbreviated as machos following @xcite .
various efforts were undertaken to search for such objects , including the optical gravitational lensing experiment ( ogle ; @xcite ) , the macho project ( e.g. , @xcite ) , and several others .
currently ongoing projects , most notably ogle and moa @xcite , are detecting hundreds and dozens of new microlensing events each year , respectively .
the new events found by these groups are made public in real time to maximize scientific gain by enabling follow - up of interesting objects by the astronomical community .
in addition to the possibility of detecting dark matter machos , microlensing surveys also offer the interesting opportunity to study the populations of stellar lenses ( which constitute at least a significant fraction of all events ) and sources .
since most stars are located in multiple systems , many microlensing events show deviations from the standard paczyski ( point lens , point source ) light curve .
for example , some events show the effects of caustic crossings @xcite ; this provides clear evidence of lens binarity and can be used to constrain the combination of the distances to the source and lens , the source - lens relative motion , and the physical extent of the source , as well as revealing the mass ratio of the lens system .
a small additional fraction of microlensing events should show clear signatures of source binarity in multiple peaks and color shifts ; this fraction was estimated to be 25% by @xcite . to date no convincing example of such an event
has been reported in the literature ; the best case is macho-96-blg-4 , but the light curve of this event can be equally well fit by a binary source or binary lens model , with the binary lens providing the most natural interpretation @xcite .
several authors @xcite have proposed explanations for the lack of such clear binary source microlensing events .
essentially , a variety of different effects including blending , unequal luminosity of binary components , and simple coincidence conspire to make most binary source microlensing light curves resemble single source , single lens events . a handful of other events , such as macho-96-lmc-2 @xcite ,
show signatures of binary orbital motion of the source ; in this particular case , it is not clear whether both components of the binary contribute significantly to the total source flux . in any case
the light curves of such events do not match the expectation of having multiple distinct , well - resolved peaks as described above .
this work analyzes the light curve of , which shows two distinct microlensing peaks .
the light curve can be fit by a binary lens or a binary source model ; however , the binary source model is statistically preferred .
the light curve is presented in [ sec : data ] , [ sec : model ] contains a description of the models and fitting procedure , and the results are analyzed in
[ sec : disc ] .
an @xmath0-band light curve of was obtained from the ogle internet archive ( cf .
@xcite ) , and is shown in figure [ fig : lc ] .
the source is located at @xmath1:59:03.04 , @xmath2:11:@xmath3 ; a finding chart is also available from the ogle web site .
the data set consists of 109 observations from three observing seasons spanning a period of 2.2 yr , from august 2001 through 17 october 2003 .
the light curve from the 2001 and 2002 observing seasons ( seasons i and ii hereafter ) is relatively constant ( to within @xmath4% excluding the highest point and the lowest point ) , with a mean @xmath0-band magnitude of 15.58 .
however , due to the high stellar surface number density in the field , some blending seems likely .
the light curve from the 2003 season ( iii ) shows two very distinct smooth peaks , with maximum magnification factors of @xmath5 and @xmath6 . for both peaks , the rise and fall in source intensity
are well covered by observations . in order to simplify the fitting procedure ,
the light curve was converted to units of relative flux by adopting the mean intensity from seasons i and ii as the base flux level .
thus the mean relative flux from seasons i and ii is unity by construction . even for relatively bright sources like the one in consideration , the photometric error estimates reported by the ogle data reduction pipeline ( cf .
@xcite ) are typically somewhat too small to account for the observed scatter in the light curves of sources that do not vary intrinsically ( b. paczyski 2004 , private communication ) . in the light curve from seasons
i and ii , the standard deviation @xmath7 in flux is larger than the mean quoted error @xmath8 by a factor @xmath9 , a ratio which nevertheless is significantly too large to be readily explained by underestimation of uncertainties .
this is probably an indication that the source ( or at least one component of it , if it is a blend ) is variable at the level of at least a few percent . a careful inspection of the season ii light curve indicates that variations are correlated from night to night .
however , no periodic signal is present at a level of significance above 50% in a lomb - scargle normalized periodogram ( cf .
@xcite ) computed for seasons i and ii .
such non - periodic variability substantially complicates the modeling process , causing greatest confusion for model effects that produce variations with similar amplitudes and time scales .
[ sec : disc ] for further discussion . to aid in interpreting fit results ,
the errors were scaled so that @xmath10 per degree of freedom @xmath11 ( dof ) , also known as the reduced @xmath10 and denoted @xmath12 , is unity for a constant fit to the light curve from seasons i and ii . this required scaling all relative flux errors by a factor of 2.49 .
the scaling factor is similar to the ratio @xmath13 quoted in the previous paragraph ; the fact that these two quantities differ could either be attributed to a non - gaussian error distribution ( which seems likely given the apparent systematic variations ) or to the limited number of data points .
since the intensity of the source during seasons i and ii was taken to be constant , microlensing models were applied to the 50 data points from season iii only .
two classes of models were considered : ( 1 ) models with a coplanar binary lens consisting of two point masses , in which the source was taken to be a small ( but finite ) disk of uniform surface brightness ( `` binary lens '' or `` double lens '' models ) ; ( 2 ) models with a single point mass lens and two distinct point sources ( `` binary source '' or `` double source '' models ) .
the source was treated as finite in the former case due to the possibility of caustic crossings ; although there is no direct evidence of such in the light curve , the possibility was considered in order not to restrict the exploration of parameter space ( e.g. , a case in which a brief caustic crossing event occurs in a gap between observations ) . in the latter case ,
such caution was not needed since the smooth , low - magnification light curve precludes any possibility of very small ( close to zero ) impact parameters .
the light curve of suggests that the two binary components ( whether of the lens or of the source ) are separated by an angular distance of at least twice the angular einstein radius @xmath14 . for a typical value of @xmath15 milli - arcseconds ( mas ) and a distance to the binary ( lens or source ) of at least a few kiloparsecs ( kpc ) , which is very likely for an event in the direction of the galactic bulge ,
the implied projected separation of binary components is several astronomical units ( au ) . assuming the masses to be not too different from a solar mass ( @xmath16 ) , the minimum orbital period of the system is several years .
therefore , both classes of models were restricted to the static case ; that is , any relative motion of binary components ( lens or source ) was neglected . for both classes of models ,
two parameters are needed to characterize the vector microlens parallax @xmath17 ( e.g. , @xcite ) .
one parameter can be chosen to be the parallax magnitude @xmath18 , whose inverse gives the projected size @xmath19 of the einstein radius in the observer plane .
a second parameter @xmath20 then gives the lens - source proper motion angle on the sky ( measured counter - clockwise from north ) .
following the method recently proposed by @xcite , all parallax fitting is done in geocentric reference frames defined with respect to the position and velocity of earth at times very close to one of the light curve peaks .
binary lens models are characterized by eight parameters , in addition to the two components of the microlens parallax .
four parameters describe the trajectory of the lens relative to the source : the time @xmath21 of closest approach to the more massive binary component ( corresponding roughly to one of the two event maxima ) , the einstein time scale @xmath22 , the angle @xmath23 between the lens line of centers and the lens trajectory with respect to the source position , and the impact parameter @xmath24 to the more massive binary component .
the remaining four parameters describe the lens and source properties : the binary mass ratio @xmath25 and separation @xmath26 , the fraction @xmath27 of the baseline intensity due to the source , and the ratio of the angular source radius to the angular einstein radius , @xmath28 .
all length scales ( @xmath29 ) are in terms of the einstein radius @xmath30 of a lens with mass equal to the total mass of the binary . for events in which it has been possible to measure or estimate the ratio @xmath31 due to caustic crossings , typical values for microlensing toward the galactic bulge
are found to be on the order of 0.0010.01 @xcite . in the absence of caustic crossings and at low magnification , as in the case of , binary lens model light curves are very insensitive to the precise value of @xmath31 provided it is smaller than a few tens of percent .
hence this parameter is poorly constrained by the available data , and its value is fixed at 0.01 in the subsequent analysis . for a binary lens the model flux
can conveniently be found numerically for a given lens - source configuration by finding the ( multiple ) image locations for a set of points on the source boundary ( e.g. , @xcite ) .
this procedure gives 35 new sets of points defining the image boundaries , depending on whether the source is outside , crossing , or inside a caustic .
since the surface brightness of the source is taken to be constant , the magnification is given by the ratio of the sum of the ( appropriately signed ) areas enclosed by the image boundaries to the area of the source .
the static binary source model is constructed from a linear combination of two point lens , point source models with einstein time scales and vector parallaxes constrained to be the same for each source component .
this model has seven parameters in addition to parallax : two for the fraction of the baseline intensity due to each source ( @xmath32 , where @xmath33 ) , three for the trajectory ( @xmath34 as in the previous case except that the reference point is defined by the closest approach of the lens to the source responsible for the first peak in the light curve ) , source separation @xmath26 , and the einstein time scale @xmath22 . the model flux @xmath35 at time @xmath36 can be expressed as @xmath37\ ] ] where @xmath38 gives the total magnification for a point source , point lens system with separation @xmath39 ( in units of @xmath30 ) .
in this parametrization , the condition @xmath40 must clearly be met in order for the model to be physically meaningful .
the @xmath10minimization was performed using minuit , a function minimization package available as part of the cern program library . for each class of models ,
the fitting algorithm was initialized in several different realizations with different sets of initial parameter values .
all such sets were chosen to produce an initial model light curve that was qualitatively similar to the observed one ( two well - separated smooth peaks with approximately the correct amplitude and width ) .
for the binary lens model , four different initial source trajectories were chosen : two in which the source encountered the more massive lens first ; two in which the source encountered the less massive lens first .
for each of these pairs , one trajectory was chosen to pass in between the two lenses , and one was chosen not to pass in between the two lenses .
the initial values of the other parameters were chosen appropriately to reproduce qualitatively the light curve shape for each of these cases .
for the binary source model , an equivalent variety of initial parameter guesses were used , subject again to the requirement that the model light curve be qualitatively similar to the observations .
after the best fit solutions were obtained for models in which the parallax was fixed at zero , full nine parameter fits ( including parallax ) were performed using the previous solutions as the initial guesses .
the best fit no - parallax model light curves and fit residuals are shown in figure [ fig : fit ] .
tables [ tab : doublelens ] and [ tab : doublesource ] contain the best fit results for the binary lens and binary source models , respectively . for the same number of dof , the no - parallax double source model is preferred over the no - parallax double lens model by a highly significant margin of @xmath41 , and furthermore this model
satisfies the criterion that @xmath42 ( with the caveat that the overall error scaling is rather uncertain ) .
however , the lower panels of figure [ fig : fit ] both show evidence of systematic residuals , indicating that unmodeled complexity is present in both fits .
this is not at all surprising in light of the evidence for low - amplitude source variability mentioned in
[ sec : data ] , and considering that parallax has been neglected . including the parallax effect adds two more free model parameters and results in @xmath43 and @xmath44 for the binary lens and binary source models , respectively .
these large @xmath45 values formally indicate that the parallax effect is detected with a high level of statistical significance .
the inclusion of parallax brings both models into the @xmath46 regime , for the adopted error scaling .
it also reduces the difference in the statistical quality of the two model fits , though the double source model remains superior by @xmath47 for the same number of dof .
it is worthwhile to note that although the overall error scaling is rather arbitrary as discussed in
[ sec : data ] ( and hence the precise value of @xmath10 is far from rigorous ) , @xmath48 is still a meaningful quantity .
the low values of @xmath49 and @xmath50 respectively obtained for the best fit double lens and double source models indicate that the errors have probably been scaled by too large a factor , resulting in values of @xmath10 that are systematically low . in principle
, the errors could be rescaled so that @xmath51 for the best fitting model , but this does not seem warranted , especially in view of the complications introduced by the likely source variability .
regardless , if the errors had been scaled by a smaller factor , @xmath48 values would increase correspondingly .
additionally , there are a number of degeneracies in the problem that are important to consider . in the absence of parallax , that is , when the apparent lens - source relative trajectory is ( modeled as being ) strictly linear , there are two fully degenerate binary lens models and four fully degenerate binary source models .
the degenerate models differ only by the sign of the impact parameter @xmath24 and the orientation angle @xmath23 .
there are only two equivalent binary lens solutions because trajectories on which the two lens components pass on opposite sides of the source produce fundamentally different light curves ( due to proximity to caustics ) from trajectories on which the two lens components pass on the same side of the source . the same does not hold for the binary source models , and hence in this case there exist four degenerate trajectories .
these degeneracies are somewhat alleviated when the parallax effect is included .
for the binary lens model the degeneracy is broken ; the two solutions separate by @xmath52 ( only the preferred solution is given in table [ tab : doublelens ] ) . for the binary source model the degeneracy
is largely broken as well ; only the solution with the lowest @xmath10 is given in table [ tab : doublesource ] .
the three solutions not included in table [ tab : doublesource ] all fall within the range @xmath53 of the preferred solution .
regardless , the differences among these near - degenerate solutions are largely immaterial , since the best fit parameters ( neglecting the sign of the impact parameter @xmath54 and the orientation angle @xmath23 ) lie within or very near the 1-@xmath55 error ranges quoted in tables [ tab : doublelens ] and [ tab : doublesource ] .
for all the solutions , the parallax magnitudes @xmath18 are consistent within 1-@xmath55 with the values from the best fit solutions .
there is a further possible degeneracy discovered recently by @xcite .
this `` jerk - parallax '' degeneracy can produce multiple parallax solutions for a given sign of the impact parameter .
in an effort to find any such solutions , a detailed search of the @xmath17 plane was performed for the double source model ( which is more computationally tractable than the double lens model ) .
the results are shown in figure [ fig : fitcon ] .
no fully degenerate solution is present , although the contours are highly elongated in a direction perpendicular to the projected acceleration vector of earth at the time of the first event maximum , when the geocentric reference frame is defined ( as expected for this degeneracy ) .
numerical searches for other parameter combinations that produce similar trajectories were also performed ( by solving eqs .
1214 from @xcite ) for both classes of models .
no further degenerate solutions were discovered .
an extremely relevant question is whether the parallax effect can be disentangled from the likely variability of the source ( regardless of whether the variability is inherent or due to a hitherto unrecognized systematic effect ) . to address this issue , figure [ fig : modeldiff ] compares the deviations from constant flux in the season ii light curve with the magnitude of the effect of parallax on the model light curves for season iii .
it is apparent that the modulations of the model light curves due to the parallax effect have similar amplitudes and variation time scales to the observed variability of the source , rendering the parallax measurements rather untrustworthy .
further support for the notion of `` contaminated '' parallax measurements comes from examining the implications of the best fit parallax - included models .
there is a simple relationship between the projected einstein radius @xmath19 , the lens mass @xmath56 , and the lens - source relative distance @xmath57 , where @xmath58 and @xmath59 are the distances from the observer to the lens and source ; this relation is given by @xmath60 @xcite .
it is plausible to assume the source to be located in the galactic bulge at a distance of approximately 8 kpc and to assume the lens to have a total mass of approximately one solar mass . substituting the measured values of @xmath18 from the best fit binary lens and binary source models , one obtains @xmath61 kpc and @xmath62 kpc , respectively . in both cases @xmath63 , indicating that @xmath64 .
even accounting for the substantial additional uncertainties introduced by assuming the lens mass and source distance , such a nearby lens seems unlikely . since the apparent variability of the source limits the degree to which the binary lens and binary source models can be statistically distinguished
, a definitive interpretation of will probably require further observations .
at the least , future ogle observations of the source will allow better constraints on its variability . a simple ( though not necessarily straightforward ) test which might clearly confirm the binary source model could be performed through moderate resolution optical spectroscopy to see if the source exhibits double absorption lines or radial velocity variations .
an alternative possibility would be to obtain deep , high resolution imaging of the field to see if the lens can be observed directly .
the light curve of shows two well - separated smooth peaks that can be modeled as either microlensing by a binary lens or microlensing of a binary source .
both models are physically plausible and can possibly be tested by future observations of the source and/or lens .
the binary lens model provides a more familiar explanation of the event since such phenomena have been observed many times in the past , while the binary source model is preferred on statistical grounds .
given the apparent low - amplitude variability of the source ( or at least one component of it ) however , statistical distinctions of the magnitude that separate the two models are somewhat suspect , especially considering the improbably large microlens parallaxes indicated by the best fit models . despite these complications , remains a plausible candidate for microlensing of a well - resolved binary source one of only two such observed to date .
the author wishes to thank b. paczyski and t. sumi for helpful comments and suggestions , m. jaroszyski for providing routines ( originally developed by s. mao ) to calculate binary lens light curves , a. gould for helpful discussion , and the ogle collaboration for making microlensing data publicly available in real time .
this research was supported in part by a ndseg fellowship award to the author .
@xmath18 & 0.0 & @xmath65 & au@xmath66 + @xmath20 & 0.0 & @xmath67 & rad + @xmath25 & @xmath68 & @xmath69 & + @xmath26 & @xmath70 & @xmath71 & @xmath30 + @xmath23 & @xmath72 & @xmath73 & rad + @xmath24 & @xmath74 & @xmath75 & @xmath30 + @xmath21 & @xmath76 & @xmath77 & days + @xmath22 & @xmath78 & @xmath79 & days + @xmath27 & @xmath80 & @xmath81 & + @xmath31 & 0.01 & 0.01 & @xmath14 + & & & + @xmath10 & 57.76 & 32.10 & + dof & 43 & 41 & + @xmath12 & 1.34 & 0.783 & + @xmath18 & 0.0 & @xmath82 & au@xmath66 + @xmath20 & 0.0 & @xmath83 & rad + @xmath84 & @xmath85 & @xmath86 & + @xmath54 & @xmath87 & @xmath85 & @xmath30 + @xmath88 & @xmath89 & @xmath90 & days + @xmath22 & @xmath91 & @xmath92 & days + @xmath93 & @xmath94 & @xmath95 & + @xmath26 & @xmath96 & @xmath97 & @xmath30 + @xmath23 & @xmath98 & @xmath99 & rad + & & & + @xmath10 & 44.45 & 27.95 & + dof & 43 & 41 & + @xmath12 & 1.03 & 0.682 & + | the apparently repeating microlensing event ogle-2003-blg-095 is analyzed .
data were obtained from the ogle internet archive and exist in the public domain .
the source is relatively bright , with an unmagnified ( but possibly blended ) @xmath0-band magnitude of 15.58 , and the signal - to - noise ratio of the data is excellent .
the light curve shows two distinct , smooth peaks characteristic of a double microlensing event .
it can be modeled as either ( 1 ) microlensing by a binary lens or ( 2 ) microlensing of a binary source , with the latter model providing a statistically superior fit .
however due to apparent low - amplitude variability of the source , the interpretation is somewhat ambiguous .
ogle-2003-blg-095 is only the second possible case in the literature for microlensing of a well - resolved binary source . |
this article presents the calculation , using continuum and lattice methods , of boundary terms in 3-dimensional gravity .
the gravity theory is presented in first order palatini form , this being a particular example of the general class of bf models @xcite as this is the most convenient presentation for deriving the discretization .
we find a variety of boundary conditions , and discuss the significance of these for different types of boundaries in space - time .
the bulk theory of three - dimensional gravity is well known to be a topological field theory , however it is also well known that three - dimensional topological field theories can give rise to non- topological boundary degrees of freedom , the classic example being the cs theory giving rise to a wzw model on the boundary @xcite . in the case of three dimensional gravity with cosmological constant , one can utilize a trick that relates the action to the difference of two cs actions , and then use the standard cs - wzw relationship , however the actual boundary conditions are a little more subtle . in three dimensions
this is relevant to the @xmath0 space , or more generally to btz black hole solutions . in this paper
we wish to understand in the context of discretization of quantum gravity the boundary degrees of freedom that correspond to black hole entropy .
this paper is directed towards a longer study of boundary terms in gravity theories , ultimately in @xmath1 dimensions , with the hope of understanding directly in a theory of quantum gravity , the possibile origin of holographic phenomena , and of the microscopic details of black hole entropy , in particular well out of the supersymmetric and extremal limits which have been very well studied in the framework of string theory .
the actual type of discretization that we consider here is maybe at first sight a bit unusual .
the approach is originally due to ponzano and regge @xcite where they considered a simplicial decomposition of a three - manifold and the path - integral is then defined as a summation over the possible sets of lengths of the edges of the dual lattice .
the alternative of course is to fix the size of the simplices and to form the path integral by summation over possible simplicial decompositions .
for the major part of this paper , we will be discussing three dimensional models that have a topological invariance in the bulk and thus the fixed decomposition is somewhat innocuous but again the use of this simplicial decomposition also for the boundary where in general we believe there are physical degrees of freedom needs to be considered more cautiously .
in addition we eventually need to extend our results to the realistic case of four - dimensional gravity where we do not even have topological invariance in the bulk making things more intricate though hopefully still manageable .
we begin however , in the context of euclidean three - dimensional gravity where already we find some interesting results concerning the boundary theories .
we will start off with a discussion of a discretization of the bf theory that corresponds to three - dimensional euclidean gravity in the framework of the ponzano - regge discretization , that is a discretization into tetrahedra with edges labelled by so(3 ) spins , and each tetrahedron then weighted in the path integral ( sum ) by the corresponding @xmath2 symbol . from this discretization
we can then derive a boundary action and will compare this to what we may expect from the corresponding bf theory .
we in fact find that there are two simple types of boundary conditions , one leads to a topological boundary theory and the other to a dynamical boundary theory .
in addition we discuss mixed boundary conditions which are relevant for the boundary at infinity in @xmath0 for example .
we discuss modifications to these boundary actions that arise when one replaces the group @xmath3 with @xmath4 which would correspond to gravity with lorentzian signature .
we also discuss the regularization via quantum groups and find some interesting relationships to work on string theory and @xmath0/cft duality .
finally we make some suggestions for understanding black hole entropy in this context and we discuss briefly the extension of these methods to four - dimensional quantum gravity .
we will now turn to a discretization of the bf representation of three - dimensional gravity and show how it leads to the ponzano - regge action .
the bf action is a generic action for a certain class of topological field theories , @xcite . for three - dimensional gravity
it actually corresponds to the palatini first order action .
we will mostly use the @xmath5 variables which are related to the gravity variables via the dictionary ; @xmath6 is the dreibein and @xmath7 is the curvature of the spin - connection @xmath8 .
the basic action for three - dimensional gravity in this first order formulation is then @xmath9 where @xmath10 is the curvature two form of a potential one - form @xmath11 , and @xmath12 is the dreibein .
these fields transform under the action of an @xmath3 gauge group .
the invariance of the action consists of a gauge transformation in @xmath11 , @xmath13 , coupled with a local gauge rotation @xmath14 , and an additional invariance only acting on the dreibein ( reparametrization ) under which @xmath15 $ ] .
the parameters of these transformations are in @xmath3 and it s lie algebra respectively .
the first transformation expresses the local lorentz invariance , and the second the diffeomorphism invariance .
the theory as formulated is diffeomorphism invariant with no explicit appearance of the metric in the action and thus topological .
the constraint that the metric is torsion free , @xmath16 , in this first order form , arises from the @xmath11 equation of motion .
the total group of local symmetry is @xmath17 @xcite .
we will proceed now to a discrete formulation of three - dimensional gravity .
we will carry out the discretization as a means of studying the continuum theory , however we would like to point out that in @xcite some arguments are given indicating that in three - dimensional gravity the space - time is necessarily discrete .
our study in fact also indicates another possibile method to prove that three - dimensional gravity is discrete .
one of the original motivations leading us to consider a discrete space - time approach to quantum gravity is the following .
we will throughout this paper take the view that black holes in quantum gravity behave like quantum mechanical objects , and that this leads to unitarity in quantum gravity via some type of holographic mechanism @xcite .
if one considers the black hole horizon to be a quantum object capable of storing and retransmitting information , then one would imagine that this horizon follows a null or even time - like path in space - time and that the region inside the global horizon is not something that an outside observer can ever see or discuss .
this is the view of black - hole complementarity developed to reconcile the apparent contradiction that unitary black hole evaporation implies that observers outside the black hole view the physics of the horizon in a very different way to freely falling observers who fall into the horizon of a large black hole @xcite .
as such one may view the formation of a black - hole as the expansion of a planckian bubble in space - time to become macroscopic .
inside such a bubble there is nothing .
thus it seems necessary to think of the microscopic structure of space - time to be a collection of bubbles .
as such there is a discretization of space - time into units of size the planck length .
there are a variety of ways to approach the discretization of the bf theory in three - dimensions , although all constructions give the same final result . for other discussions of the approach that we present here
see @xcite . the simplest approach to discretization is to formally carry out the path integral over the b - field , as it is simply a lagrange mutiplier for the einstein equation .
the result is , @xmath18 = \int { \cal d}a \prod_x \delta(e^{f(x)})\ ] ] where @xmath19 are the co - ordinates on the closed manifold @xmath20 , and the delta function is in the group manifold of @xmath3 .
the delta function can be rewritten using the identity @xmath21 where @xmath22 and the sum is over all representations of the group g. using this identity we can write , @xmath18 = \int { \cal d}a \prod_x \sum_j ( 2j + 1 ) \chi_j(e^{f(x)}).\ ] ] to make this expression tractable we now discretize the manifold @xmath20 by dividing it into tetrahedra . from this tetrahedral decomposition ,
we construct a dual discretization for which the vertices are at the centre of the tetrahedra @xcite , the edges pass between the centres of adjacent tetrahedra , and the faces are then bound by these edges and each dual face will be pierced by precisely one edge of the original tetrahedral decomposition . in figure 1
we show the part of the dual lattice that will live inside one of the original tetrahedra .
we now assign to every face of the dual lattice ( that is every edge of the original lattice ) a representation and to every edge of the dual lattice a group element as shown in figure 2 .
the product of the group elements around a dual face is then the holonomy of that cycle and thus represents a discretization of the curvature . denoting the discretization of @xmath20 as @xmath23
we can finally write , @xmath24 = { \cal n}\int \prod_{e\in\delta } \sum_{j_e } du_e ( 2j_e + 1 ) \chi_{j_e}(\prod_{\partial\tilde{e}}u)\ ] ] where @xmath25 is a normalisation factor .
in this expression , @xmath12 is an edge of the tetrahedral decomposition @xmath23 and @xmath26 is the face dual to the edge @xmath12 . to actually evaluate this expression we notice that the character can be written as a sum of products of the wigner function @xmath27 where @xmath28 is the group element corresponding to an edge of the dual graph , @xmath29 where @xmath30 .
then for each edge of the dual graph there will appear in the integral over the corresponding group elements three wigner functions which can be evaluated immediately using , @xmath31 ( for these and other angular momentum identities that we use below and for the definitions of the various symbols that we use , we recommend that the reader refer to the very complete monograph @xcite ) .
each tetrahedron thus will contribute two @xmath32 symbols for every face , thus eight @xmath32 symbols for every tetrahedron .
half of these are summed over angular momentum projections in pairs , one of the pair coming from each of two tetrahedron with a common face and the orthogonality of the @xmath32 symbols ensures that this term becomes the identity .
the remaining expression is such that summing over the projection quantum number of the angular momentum the four @xmath32 of a given tetrahedron gives a single @xmath2 symbol using the identity , @xmath33 the final result for a closed manifold @xmath34 and simplicial decomposition @xmath23 is ( see @xcite or @xcite for more details ) , @xmath35 this answer is the path sum proposed by ponzano and regge to be a discretization of three dimensional quantum gravity . in fact for a tetrahedron with edge lengths @xmath36 , the corresponding weight is the 6j symbol with angular momenta @xmath37 . in the semi - classical limit @xcite ( large angular momenta and vanishing planck length such that the combination @xmath38 remains constant ) ,
a 6j symbol actually becomes the cosine of the regge action for the tetrahedron as a direct discretization of three dimensional gravity ( the cosine arises as the bf theory path integral sums indiscriminately over positive , negative and degenerate values for @xmath39 ) .
the regge action is the direct discretization of the einstein - hilbert action @xcite .
@xmath40 @xmath41 is the angle between the normals to adjoining faces and @xmath42 is the length of the edge common to the two faces labeled by @xmath43 and @xmath44 .
the extra factor of half between the edge length and corresponding angular momentum is for consistency in this semi - classical limit and we can intuitively justify it by noting that the length of the angular momentum vector for the representation of spin @xmath45 is actually @xmath46 which becomes @xmath47 in the limit of large angular momentum .
the above expression for the discretized path sum is not quite complete .
we have ignored the fact that there could in principle be some normalization factor in front of the sum , and in fact one would hope that there is such a factor simply because the sum itself is divergent .
one can simply choose a normalization factor to subtract the divergence , however it is interesting to see how the divergence arises .
this was already analyzed in the original paper of ponzano and regge , and the reader should look there for the details . in short , one takes the biedenharn - elliot ( be ) identity ( appendix a. ) , which relates a product of three @xmath2 symbols summed over one angular momentum , to a product of two @xmath2 symbols without summation .
geometrically this corresponds to taking three tetrahedra joined together along a common edge , and each with a face in common with two of the others .
removing the common edge ( the sum in the be identity ) leaves one with two tetrahedra sharing one common face . using the orthogonality of the @xmath2 symbols ,
one can change this identity to one that relates a single tetrahedron to four tetrahedra formed by introducing an additional vertex at the centre of the original tetrahedron .
the identity is in appendix a for the interested reader .
the important point is that there is an infinite factor @xmath48 for every vertex of the simplicial decomposition .
therefore we see that an infinite factor of this form must be added to the denominator of the path sum to regularize it , and that there is one such factor for every vertex in the triangulation .
the actual normalization factor is then @xmath49 where @xmath50 is the number of vertices in the discretization . in addition
this discussion has shown us that the path sum is actually invariant under the two transformations derived from the be identity .
these two transformations are known as pachner moves @xcite and these are the discretized version of diffeomorphisms .
we have therefore learnt that the path sum thus defined ( in particular with the regularization discussed ) is diffeomorphism invariant in discretized form just as the @xmath5 theory was before the discretization .
the full path sum is then , @xmath51 where @xmath52 is the number of vertices in @xmath23 . in this form
however it is still not very practical for calculating .
there exists a different regularization that involves a q - deformation of @xmath53 due to turaev and viro @xcite .
the path sum is @xmath54_q \prod_{t\in\delta } ( -)^{\sum_{i=1}^6 j_t^i } \left[\begin{array}{ccc } j_t^1 & j_t^2 & j_t^3\\ j_t^4 & j_t^5 & j_t^6 \end{array}\right]_q\notag\end{aligned}\ ] ] where @xmath55_q & = \frac{q^n - q^{-n}}{q - q^{-1}}\end{aligned}\ ] ] the parameter of the quantum deformation is a root of unity @xmath56 and the sum is regularized as the representations of @xmath57 involve angular momenta only in the range @xmath58
so the path sum now involves all finite sums and @xmath59 has been replaced by @xmath60 which is clearly finite .
the semi - classical limit of the @xmath61 symbol indicates that the q - deformed path sum is related to quantum gravity in three dimensions with a positive cosmological constant .
the limit must be carried out in a way that as the angular momentum become large , correspondingly also @xmath44 must go to infinity .
the limit is @xcite , @xmath62 and we see in particular that the limit which makes contact with the semi - classical physics is the limit in which the cosmological constant goes to zero . note that this can not be derived directly from the action @xmath63 by any simple generalization of the discretization carried out above , as the non - linearity in @xmath39 does not allow us to easily integrate over @xmath39 to get a simple expression involving the curvature @xmath64 .
it would be very interesting to find a derivation of the tv path sum from the discretization of the path integral for @xmath65 . for simple manifolds this sum
can actually be evaluated giving the turaev - viro invariants that are important for the understanding of the topology of three - manifolds .
the restriction on angular momentum in the quantum group representations is the same as that which must be imposed on string states in @xmath0 .
we will take another look at the q - deformed action and limits thereof after we have discussed the boundary discretization and will find that in the context of gravity in @xmath0 there may indeed be a deeper meaning to this regularization .
it is also interesting to consider the relationship between this construction of three - dimensional gravity using a quantum deformation and studies of quantum doubles of groups @xcite . in this article
one has a different type of quantum group that does not have a fixed deformation parameter .
it is used for the discussion of multi - particle states in three - dimensional gravity .
each particle , creates a localized source of curvature , and in general the space is conical at infinity .
it is amusing to notice that for the chern - simons description of three - dimensional gravity , at zero cosmological constant one uses the group @xmath17 @xcite , but at non - zero cosmological constant , one finds instead the group @xmath66 with the level of the chern - simons theory related to the curvature . going to the multi - particle fock space in three - dimensions
means that we are allowing variable localized curvature depending upon the location and mass of the particle sources .
we find that the group @xmath17 is replaced by @xmath67 @xcite but now with no additional parameter , indicating perhaps that all values of curvature are possibile depending on the number and mass of particles present .
this relationship deserves to be studied in more detail as it indicates a possible second quantization that involves also the cosmological constant .
let us consider the general variation of the @xmath5 action for a manifold with boundary , ( other work on this subject can be found in the papers @xcite ) .
@xmath68 we see then that the field equations are not effected by the presence of the boundary provided that the variation of @xmath69 is zero on the boundary
. the path integral in the presence of the boundary will then be a function of the boundary value of the spin connection .
we have another choice , which corresponds to the bf theory with a boundary term @xmath70 the variation of @xmath71 is now @xmath72 and therefore the boundary condition must be that the variation of @xmath39 is zero on the boundary , and the path integral will now be a function of the boundary metric .
the first boundary condition of fixed spin connection on the boundary actually gives rise to a topological field theory on the manifold plus boundary .
the second boundary condition is dirichlet on the metric , and this does not give rise to a topologically invariant boundary action . in a study of asymptotic symmetries in three - dimensional gravity @xcite
, it was shown that with appropriate boundary conditions one can also find a liouville theory on the boundary at infinity of @xmath0 space .
such boundary conditions formulated in terms of the metric and connection are actually mixed boundary conditions , and we will give more details of how these work below . the continuum
boundary action can be easily derived by following a construction similar to that used in @xcite where the wzw - cs relationship was discussed in some detail .
first we consider the boundary condition @xmath73 for which there is no additional boundary term . to examine the boundary theory we will insert the solutions to the bulk equations and then examine the action of the gauge symmetries of the theory in the presence of the boundary . for the wzw - cs relationship
the boundary degrees of freedom arise precisely because the bulk gauge symmetry is only a global symmetry on the boundary thus the breaking of the gauge symmetry by the presence of the boundary gives rise to new degrees of freedom .
the bulk equations of motion are solved by @xmath74 @xmath75 substituting these solutions into the action we find of course that it vanishes identically as @xmath76 .
the gauge variation of the action is identically zero for the local lorentz invariance , but the diffeomorphism transformation has in principle an additional boundary term equal to @xmath77 which we see also vanishes as @xmath76 from the equations of motion ( modulo some topological issues regarding the extension of a flat connection to a boundary of given topology ) .
this boundary action is that of an obviously topological two - dimensional field theory in agreement with the proof by ooguri and sasakura @xcite that with the @xmath73 boundary condition the path sum is a topological invariant not just of the bulk but of the bulk plus boundary theory . for the @xmath78 boundary condition we must add to this result the boundary term @xmath79 .
the boundary condition now seems to indicate that the boundary metric is important in the path sum , in fact the path sum will now be a function of the boundary triangulation . again
the solutions to the bulk equations will be inserted into the action , the bulk again giving zero contribution but the boundary now gives a non zero contribution equal to @xmath80 furthermore the gauge transformations now give rise to non - trivial boundary terms , @xmath81 @xmath82 we can see from these variations that the symmetry of the boundary theory is significantly smaller than that of the bulk theory .
in fact we must have @xmath83 constant for the gauge transformation to vanish and also @xmath84 .
therefore the boundary theory has no diffeomporphism invariance , and is invariant only under global lorentz transformations .
finally we can consider the boundary conditions used in @xcite which are related to three dimensional gravity with cosmological constant . to do this
we make a small deviation into the chern - simons representation of three - dimensional gravity with cosmological constant .
our action is then , @xmath85 we make the change of variables , @xmath86 and we then find that @xmath87 becomes the difference of two chern - simons theories plus an additional boundary term .
@xmath88 - cs[a^- ] ) + \frac{1}{\sqrt{-3\lambda}}\int_{\partial { \cal m}}tr(a^+a^-)\ ] ] we see here that the level of the chern - simons theory is inversely proportional to the square root of the cosmological constant , and also that if we started with a @xmath5 action with no boundary term , then after the change of variables we have a boundary term that is of a mixed form , rather than of the form @xmath89 .
this is due to the fact that using the variables @xmath90 we can consider boundary conditions that would be mixed boundary conditions when expressed in terms of the variables @xmath69 and @xmath39 . indeed ,
if we add @xmath91 to the @xmath5 action then following the construction of @xcite one finds that in the chern - simons variables the action factorizes into two pieces that represent a pair of chiral wzw theories .
the boundary conditions now imply restrictions on a combination of the metric and connection .
it is precisely this setup that was shown to arise for the boundary at infinity of @xmath0 in the work of brown and henneaux and afterwards coussaert , henneaux and van driel @xcite .
the boundary theory is actually a liouville theory .
note that to discuss this case in the discretized framework we really need to use the quantum group representations as it is only then that ones sees a cosmological constant in the semi - classical limit .
the discretized boundary theory will turn out to be very similar to a discretization of liouville theory .
we will show how this relationship arises in more detail once we have set up the formalism for the quantum discrete boundaries .
one may already worry here that we are trying to construct some triangulation of liouville theory in the strongly coupled phase and it is well known that for euclidean surfaces such theories have very non - continuum like phases .
a discussion of these problems and arguments for better behaviour in the lorentzian case are in @xcite . from the bulk calculation of the discretized path sum
, we saw that for every face of the simplicial decomposition , there are two @xmath32 symbols .
indeed if we consider a single tetrahedron as a discretization of the three - dimensional ball then it has a weight , @xmath92 joining now an additional tetrahedron to one of the faces of this tetrahedron , we get another decomposition of the three - ball . on the internal face
there is now a @xmath32 symbol coming from each of the tetrahedra , but we must now sum over the angular momentum projections assigned to the internal faces ( now identified of course ) . using the orthogonality identity for a pair of @xmath32 symbols
@xmath93 we see that the internal @xmath32 symbols vanish and we are left in the path sum with a @xmath2 symbol for every bulk tetrahedron , and a @xmath32 symbol for every boundary face . in general the integral that gave rise to this pair of @xmath32 symbols was along the link of the dual lattice that passes from the centre of one tetrahedra to the centre of an adjacent one piercing one and only one face . in the presence of a boundary only half of this integral is carried out , from the centre of the tetrahedron to the face and this integral gives rise to a @xmath32 symbol for the bulk @xmath2 symbol and additional single @xmath32 symbol for the boundary face .
the new feature that has given rise to the boundary weights for the boundary faces is that now we do not have an entire dual face , but rather the dual face is cut in half by the presence of the boundary as shown in figure 3 .
we therefore need to also consider the group elements that live on the edge of the dual face that is exposed by the boundary . for the bulk path - sum described in the previous section
the connection was integrated away .
now due to the exposed dual faces , we have a boundary dependence on the connection that we may or may not integrate over depending upon the boundary conditions chosen . in figure 4
we have labelled one such edge from @xmath94 to @xmath95 with its weight @xmath96 .
for the boundary conditions that correspond to the action with no boundary term , that is the @xmath97 conditions , we are instructed to keep the connection fixed on the boundary , and thus we must not integrate over the boundary values of @xmath28 .
we thus find a network with trivalent vertices , each vertex is weighted by a @xmath32 symbol , and the vertices are tied together by the matrix elements of the corresponding group elements .
the one and two tetrahedra path sums above easily generalize by gluing faces of tetrahedra together and using the orthogonality condition giving one the general expression for a simplicial decomposition with boundary .
@xmath98 in this expression the normalization factor is the usual one mentioned above and @xmath99 is the dual lattice .
the summation is over the angular momenta assigned to edges in the bulk and the boundary , and over the angular momentum projections assigned to each triangular face of the boundary .
for the situation where the group representations summed over are those of the quantum group this is precisely the bulk plus boundary action derived by ooguri and sasakura @xcite , where they show that the hilbert space of the @xmath100 theory is equivalent to that of a pair of chern - simons theories for which the boundary state is described by wilson lines joined by trivalent vertices with an identical structure to that derived above .
we would also like to note that this path sum ( for all boundary group elements @xmath28 equal to the identity element ) is the same as that derived in @xcite .
in contrast to our present approach , in that paper the boundary action was derived purely on the grounds of topological invariance .
the boundary term @xmath101 required for the @xmath102 boundary conditions when discretized becomes , @xmath103 where @xmath104 refers to the boundary value of @xmath39 . in this expression the dreibein @xmath104 is replaced by its discretized representation that being the length of the corresponding edge of the boundary of the original lattice , and the connection is represented by @xmath28 which is the gauge field assigned to the link of the boundary of the dual lattice that is dual to the edge where @xmath104 resides .
the partition function is in this case a function only of @xmath104 .
we must multiply the path sum derived above for the @xmath97 boundary conditions by this additional term , remove the sum over the boundary values of the spins @xmath105 , and integrate over @xmath28 to derive the final path sum for dirichlet boundary conditions in the metric .
the integral of importance is that over @xmath28 and is @xmath106 inserting this into the path sum gives the final result for fixed metric boundary conditions , @xmath107 where @xmath108 signifies the lattice without boundary components .
the sum is now only over the angular momentum in the interior edges .
the integral over @xmath28 on the boundary has now fixed the angular momentum projections to be associated to edges of the boundary triangulations , rather than with faces as for @xmath109 .
thus we see that the action is quite similar to that for the @xmath97 boundary conditions except that now the boundary values of the angular momenta are fixed corresponding to the fixed boundary metric .
note that in this path sum the factors of @xmath110 are absent for edges that lie in the boundary due to the restriction in the product over edges to @xmath111 .
these factors are important in the angular momentum identities that one uses to prove topological invariance , thus indicating that for this choice of boundary conditions there is no topological invariance on the boundary agreeing with our continuum analysis . in the path integral for a fixed boundary metric
, one would expect that in the quantum gravity there would be a need to sum over all possible boundary configurations that give a discretization of the continuum boundary metric .
a construction of such a type will be seen to be necessary for a calculation of black hole entropy in this discretized setup . in general before fixing boundary conditions we have the expression for @xmath97 without summation over angular momenta and without the integral over the boundary gauge connection .
we need to understand what boundary conditions will allow calculations relevant to black hole physics , and also what representations of the ( quantum ) group one must include in this summation .
the representations and boundary conditions will be discussed in the final section when we consider the construction for lorentzian metrics .
furthermore , we need to know how to implement the boundary conditions that give liouville theory in our path sum construction .
we want to show a point of contact between our calculations and discrete tft s in two dimensions . for @xmath97
everything is topological and there is an easy way to get a two dimensional tft from this theory . in @xmath112
take a thickened wall and remove the bulk tetrahedra using the various pachner moves in the bulk and on the boundary .
the final result will be just two dimensional , but in some sense a double layer as the two faces will both carry their own @xmath32 symbols .
the two dimensional action that one finds by this procedure is , @xmath113 this is indeed a two - dimensional tft , invariant under two - dimensional pachner moves and similar actions have been studied in a collection of works @xcite . for @xmath102 we can not actually remove all the bulk tetrahedra , as the removal process that one uses for the totally topological situation of @xmath97 relies heavily on the topological invariance of the boundary theory and in particular on the elementary shelling operations .
we can however take a limit that is inspired by the bulk boundary correspondence of the ads / cft conjecture @xcite . to do this
we imagine that we take a semi - classical limit of the bulk action leaving the boundary angular momentum fixed .
the relevant limit of the bulk @xmath2 symbols that have a face edge or vertex on the boundary were already studied in the original article of ponzano and regge .
the interesting thing that we find is that the boundary answer depends crucially on the asymptotic properties of the manifold .
this sort of behaviour is maybe not a surprise as it is precisely such a dependence in the ads / cft correspondence that accounts for the simplicity of the near horizon limit in the ads case . for asymptotically flat spaces
however the action is not expected to be similar to the cft as it will live on a null surface rather than on a time - like surface and the asymptotic group of symmetries will be smaller .
the limits of @xmath2 symbols in which only some of the angular momentum are taken to be large are of two basic types .
the first involves removing one vertex to infinity , and thus the three edges connected to that vertex become large , while the three vertices that form the remaining face stay fixed , this face then is a triangle of the boundary configuration .
the result is thus the @xmath32 symbol of the remaining face , where the pairwise differences between the large angular momenta make the @xmath114 quantum numbers in this @xmath32 symbol and we thus find an answer similar to that which we derived from the bf theory , a pair of @xmath32 symbols on the boundary .
the answer is , @xmath115 the other possibility corresponds to holding the length of one edge fixed , this representing a tetrahedra that has only an edge in contact with the boundary . in this case one
still can do one of two things with the remaining angular momenta .
one can take the angular momentum on the unique edge that does not touch our chosen edge to also be fixed , and the other four go to infinity . or one can take all five to be large .
this is where the dependence on the large scale asymptotics of the space have an effect .
if for instance in the euclidean case we are considering a boundary that is a sphere in @xmath112 , then clearly we must take the limit where all five other angular momenta become large . on the other hand , if the boundary is a plane in @xmath112 then one need take only four angular momenta to infinity , the other two corresponding to opposite edges of the tetrahedra remain fixed .
the expressions for these limits contain additional dependence on parameters of the limiting process and can be found in appendix b. the expressions for the path sums in these limits are relatively complicated .
it is interesting to note that the answer for this `` near - boundary '' limit , is basically the two - dimensional double @xmath32 symbol action derived above for purely topological boundary conditions , however , with some additional structure depending upon the asymptotic behaviour of the space - time . in the next section for null surfaces in lorentzian manifolds we will find that the semi - classical limit leads to an hypothesis that simplifies the boundary discretization considerably .
if we replace the @xmath3 of the euclidean construction with @xmath4 then the representation theory becomes somewhat more complicated , and all limits of the corresponding angular momentum coupling coefficients in the various representations have not been fully studied .
however , the original large angular momentum limit of ponzano and regge has also been carried out for the discrete series of representations in the non - compact case @xcite .
the result is basically the same as for the compact group apart from the fact that the angles are now hyperbolic given by the boost required to take the normal to a face into the normal of an adjoining face . in the limit of large angular momentum
we have @xmath116 in this expression , @xmath117 is the angle between the faces @xmath43 and @xmath44 , @xmath118 and @xmath119 is the unit normal to the corresponding face .
thus as one face becomes null , the corresponding normal will also become null , and the angle that this face makes with the three neighbouring faces becomes infinite .
the exponential in the weight for the tetrahedron implies that the corresponding angular momentum must be zero or that two of the sides of the triangle are of equal length and the corresponding angles are of opposite sign . therefore the only configurations that can contribute have equilateral triangles and isosceles triangles where the short edge has length @xmath120 corresponding to zero angular momentum .
the equilateral triangles must have all zero angular momentum labels and thus have all sides of length @xmath120 . so in a path sum involving all discretizations with a given boundary metric the path sum is dominated by discretizations with boundary triangles that have all lengths equal to @xmath120 .
this is modified then by collective structures built from isosceles triangles .
it is clear that the configurations involving isosceles triangles must be collective , as the presence of an isosceles triangle , implies also that neighbouring triangles are isosceles , and so on , until the structure closes again .
an example of such a collective structure is shown in figure 5 .
these structures are reminiscent of macroscopic loop operators in the matrix models of dynamical triangulations @xcite .
we should note here that we have been a bit incautious regarding the order of limits .
we took large @xmath45 and then interpreted the expression for small @xmath45 .
there is indirect evidence that the result is sensible and we will discuss our reasoning below . the precise calculation that one needs to do is to take the null boundary limit of the quantum @xmath2 symbol in a similar way to the original limits studied by ponzano and regge .
thus we get a picture of horizon states in discretized quantum gravity and this is a positive step towards a micrpscopic understanding of black hole entropy . in t
hooft s discussions of horizon states @xcite one finds similarly a special role for the low angular momenta , @xmath121 when the horizon at fixed rindler time is represented as a collection of discretized line segments labelled by angular momenta of @xmath4 . also in the ashtekhar approach to quantum gravity , the entropy calculations indicate that entropy is derived from contributions only from the lowest spin states on the horizon @xcite and similarly in the paper @xcite .
it is interesting to reflect upon the meaning of the boundary action .
if we assume that the semi - classical limit of the clebsch - gordon coefficients for @xmath122 for the discrete representations are an analytic continuation of those for @xmath57 then we should find a negative cosmological constant .
for the situation of @xmath123 gravity in a space of constant negative curvature , one finds as mentioned above that the boundary theory is a liouville theory .
furthermore from recent work on liouville theory @xcite it is known that the representations of the virasoro algebra that arise in the n - point functions , involve the quantum group @xmath124 . in the string theory picture of @xmath0/cft duality @xcite
the mass cut - off on angular momentum representations is also the same as that which arises in the discrete representations of @xmath124 .
beginning as we did from the prtv ( ponzano - regge - turaev - viro ) construction , it appears that we have arrived at almost the same conclusion .
note though that in the prtv construction , after changing to a lorentzian space - time signature it is not necessary that the representations are identical to those used for the euclidean geometries .
maybe one should sum(integrate ) over the continuous representations that arise in the liouville approach for the boundary at infinity . on the other hand , for the null boundary at a global horizon
, it is not so clear how to proceed , however some interesting insight will come from a comparison of our boundary action and string theory on @xmath0 @xcite . if we had the clebsch - gordon coefficients for the discrete series of @xmath122 we could also explicitly calculate the weight of a `` macroscopic loop '' configuration and make a direct comparison with the macroscopic loop wavefunctions calculated for example in @xcite .
the clebsch - gordon coefficients are known for the continuous series @xcite and for these one should be able to directly compute the null boundary limit .
the representation theory of non - compact quantum groups is still very much under development , see @xcite .
also a discussion on the relationship between strings in @xmath0 and quantum groups can be found in @xcite .
there are representations of @xmath122 that are discrete , and agree basically with the discrete ones for the compact group , and give a cut - off in the path sum , see also @xcite for a few more details on these .
the other representations are those that arise from the quantum group representation of the virasoro algebra of the liouville cft at @xmath125 .
these are similar to the continuous representations of @xmath126 .
we can already make some speculative remarks derived from studies of string theory and continuum gravity in @xmath0 @xcite .
one can study the various physical excitations in this space - time both from the perspective of the space - time and that of the string theory . in the space - time picture
, one finds a @xmath127 liouville theory , and indeed if one considers a non - critical string theory with target equal to @xmath0 then again one will find the world - sheet theory also to be liouville with @xmath127 .
the states that arise are classified by quantum group representations @xcite .
however the representations that arise are not those that we are using in the turaev - viro path sum .
this strongly suggests that an extension of the prtv ( ponzano - regge - turaev - viro ) construction to include the representations of @xmath122 that arise in the liouville theory corresponds to extending the quantum gravity path sum , to a string field theory path sum ( albeit with a fixed topology for the target manifold ) .
liouville theory at @xmath128 also appears in the context of @xmath129 compactifications although in this case the theory appears as a consequence of @xmath130 group factors in the internal space @xcite .
it would be interesting to find connections between this structure and the liouville theory that is naturally present for @xmath0 string compactifications .
the liouville theory on the boundary cylinder at infinity for gravity in @xmath0 has a central charge @xmath131 where @xmath132 or @xmath133 and the correlation functions of this theory are constructed from the clebsch - gordon coefficients of the quantum group @xmath122 where @xmath134 the cosmological constant of the @xmath0 space is proportional to @xmath135 . in turn , the cosmological constant that arises in prtv is proportional to @xmath136 , where the deformation for the turaev - viro quantum group is given by , @xmath137 clearly @xmath138 and thus the groups that arise in the two approaches are indeed identical deformations of @xmath139 .
thus the group and deformation parameter agree in a manner which supports the conjecture that the quantum group that arises in the regge calculus is the same as that of the liouville theory on the boundary . however , the representations that arise in the liouville theory have angular momentum in @xmath140 , while those in prtv are identical to those that arose for @xmath57 with angular momentum running from 0 to @xmath141 .
of course once we changed from euclidean to lorentzian discretizations , the question already arose as to which representations one should sum over and now we see that the answer to this question may have deeper significance .
actually one can make the relationship between our discrete boundary action involving the quantum group and the perturbation theory of the liouville theory on the cylinder more concrete in a very geometrical manner by examining the perturbative expansion of the liouville theory on a cylinder ( corresponding to the boundary of @xmath0 ) .
write the path integral with sources and charges for all liouville vertex operators in selected representations .
use bootstrap to argue that all vertices can be reduced to cubic and recall that the cubic vertex for the liouville theory is given precisely by the clebsch - gordon coefficient of the quantum group @xmath122 @xcite .
furthermore the propogator of the perturbative expansion of the liouville theory is the wigner coefficient @xmath142 of the corresponding representations .
such feynman diagrams correspond precisely to the dual lattice with weights as derived in the previous section and as shown in figure 3 .
geometrically all genus zero amplitudes correspond to one of our quantum group boundary terms in structure but with a sum over representations different from those used by turaev - viro .
if one considers the dual lattice to the boundary triangulation , one finds a trivalent graph that lives on the boundary of the manifold , being one of the feynman diagrams discussed above . at
any given time - slicing this will look like a collection of particles with mass given by their spin , and as this gas evolves there are interactions coming from the trivalent graph .
thus one can make a proposal for calculating the entropy using a system of particles making a gas .
the @xmath143 spin chain is a possible starting point for such a calculation .
this model is a chain of spins the solution to which involves the quantum group @xmath122 and which is related to liouville theory for @xmath128 and @xmath144 and also possibly for all @xmath125 @xcite . within this framework
we should be able to formulate the explicit calculation that is necessary to calculate the entropy of the boundary theory and thus black hole entropy .
the feynman diagrams of the boundary action describe the time - evolution of the gas ( figure 6 ) . for a null boundary
the gas will be non - relativistic whereas for a time - like boundary the gas will be relativistic . in the case of a null boundary
these representations will become more restricted and the three point interaction implies that during the evolution of the gas one has both creation and annihilation of particles .
this may even imply some sort of dissipation in the null case .
other works arguing for dissipative behaviour for a theory describing a black hole horizon have appeared in @xcite . from a deeper understanding of this gas
one should be able to directly calculate the entropy and thus the black hole horizon entropy .
another consideration that we have not addressed directly but that has already arisen a few times in our discussions , and also one that is intimately related to the calculation of the entropy is the following . without a boundary , it was clear that the prescription of ponzano and regge to hold fixed the simplicial decomposition was already sufficient due to the topological nature of the theory . now in the presence of a boundary it is possible that one really needs to sum over the boundary triangulations .
the bulk theory is topological and is insensitive to how one describes the sum in detail , however we expect some dynamics on the boundary .
this indicates the possibility of extending the path - sum to dynamical triangulations .
this sounds like trouble as such triangulations give rise to the matrix model of liouville and for @xmath127 these models are badly behaved with very rough surfaces dominating the path sum .
however , discretizations for lorentzian manifolds have been studied in @xcite where the authors have shown that when the simplicial decompositions are restricted by the requirement of a causal structure , the phases of the dynamical triangulations are well behaved involving smoother surfaces than in the euclidean setup .
the haussdorf dimension in particular remaining @xmath145 rather than becoming fractal and equal to 4 as it does in the euclidean case .
indeed , our work also implies that there is another possibly interesting type of dynamical triangulation , where the `` causal '' structure is that implied by the constraint that the surface be not lorentzian , but null
. it would be interesting to study in the context of @xcite null dynamical triangulations , the results of which investigation would certainly shed light on the dynamics of black hole horizons in quantum gravity . for @xmath1 dimensions we now have some intuition for how to approach the discretization
. the simplices will be labelled by @xmath146 representations .
we can write the boundary path sum including boundaries following more or less the same philosophy as above . in this case from the beginning it seems that we probably need to consider dynamical triangulations as otherwise we will end up with a topological bulk theory rather than a theory containing also gravitational dynamics .
various versions of discretizations of four - dimensional lorentzian manifolds have been studied as for example in @xcite .
furthermore the semi - classical limit of the @xmath147 symbols that arise in these bulk path - sums , has been studied in @xcite with results agreeing with the regge discretization once more .
we expect for null boundaries also in @xmath1 dimensions that some restrictions will be placed on the representations arising and that one will probably again find some sort of three - dimensional dynamical triangulation describing the behaviour of the horizon . from the work in @xcite it has been shown that also for three - dimensional lorentzian dynamical triangulations , the branched polymer and crumpled phases , can not be reached leaving hope that such a system will have a nicely behaved continuum phase transition .
it would be interesting to also look at three dimensional null dynamical triangulations to see if the causality restrictions on triangulations introduce some regulator of the geometries .
the way to proceed is we believe clear .
one must determine the representations that are important for the theory that is being investigated , and then one must look at various limits of the @xmath45-symbols .
+ * appendices *
the biedenharn - elliot identity relates the @xmath2 symbols associated to two different ways of combining nine angular momenta .
the sum on the right hand side is replaced by a product on the left .
geometrically this identity is represented by the diagram shown .
@xmath148 using the orthogonality for a pair of @xmath2 symbols this identity can be rearranged as discussed in the text , up to an infinite multiplicative factor .
the regularized version of this identity as first given in ponzano and regge @xcite , is @xmath149 another useful identity for understanding the relationship between bulk and boundary transformations is , @xmath150 the geometrical meaning of the left side is simply a pair of adjoining boundary faces .
the right hand side involves the gluing of two faces of an additional tetrahedron to the original pair of faces resulting in a new pair of boundary triangles .
this results in a @xmath151 pachner transformation in two - dimensions .
using the orthogonality of the @xmath32 symbols one can rewrite this equation to give the algebraic representation of the @xmath152 transformation .
here are the @xmath153 and @xmath154 limits of ponzano - regge .
we will use the following labelling for the tetrahedron ( figure 7 ) . for the @xmath153 limit
, we shift @xmath155 by @xmath10 and take @xmath10 to be much larger than all of @xmath156 . in the figure this limit corresponds to keeping the segments
@xmath157 $ ] and @xmath158 $ ] of fixed length while all other edge lengths go to infinity . we can consider @xmath159 $ ] to be the edge of the tetrahedron that lies in the boundary . for the @xmath154 limit
, we take @xmath161 and @xmath162 , where now @xmath163 are all large , though small with respect to @xmath164 .
this corresponds to keeping only the segment @xmath165 of fixed length and all other edges to infinity .
the one edge of small size is then @xmath165 and in the text this is the edge that lies in the boundary of the manifold , all other edges in this case being internal .
the final answer is , @xmath166 } \mbox{cos}\,(t-\frac{1}{4}\pi),\ ] ] where @xmath167 and @xmath168 is the regge action for the tetrahedron . in this limit
, the dependence on the asymptotic structure of the space enters as the angle that remains in the final expression is the angle between the edges @xmath169 $ ] and @xmath170 $ ] or equivalently between @xmath171 $ ] and @xmath172 $ ] .
these angles enter the expression for the limit through the regge action .
if for instance the boundary is on a sphere of finite volume , then as one takes this limit a tetrahedron with one edge stuck on the sphere boundary , these angles will go to infinity .
if the boundary is planar in flat space then the angles will go to zero and we go back to the @xmath153 result .
+ the author would like to thank g.arcioni , m. blau , g. thompson , g. t hooft and k. krasnov for useful conversations during the various stages of this work .
this work was commenced during the extended workshop in string theory at the abdus salam ictp in trieste in the summer of 1999 .
the author is supported by the pionier programme of the netherlands organisation for scientific research ( nwo ) .
99 m. blau and g. thompson , `` a new class of topological field theories and ray - singer torsion , '' _
_ * b228 * ( 1989 ) , 64 ; `` topological gauge theories of antisymmetric tensor fields '' , _ ann .
* 205 * ( 1991 ) , 130 : g. horowitz , `` exactly soluble diffeomporphism invariant theories '' , _ commun
* 125 * ( 1989 ) 417 .
s. elitzur , g. moore , a. schwimmer and n. seiberg , `` remarks on the canonical quantization of the chern - simons - witten theory . '' _ nucl.phys._*b326 * ( 1989 ) 108 . g. ponzano and t. regge , `` semi - classical limit of racah coefficients , '' in _ spectroscopic and group theoretical methods in physics _
f. bloch , ( north - holland , 1968 ) .
e. witten , `` @xmath173-dimensional gravity as an exactly soluble system '' , _ nucl .
* b311 * ( 1988 ) 46 .
g. t hooft , `` quantization of point particles in @xmath173-dimensional gravity and space - time discreteness '' , _ class .
* 13 * ( 1996 ) 1023 , gr - qc/9601014 .
g. t hooft , `` dimensional reduction in quantum gravity . ''
gr - qc/9310006 .
l. susskind , `` the world as a hologram . ''
hep - th/9409089 .
g. t hooft , _ nucl .
* 256 * ( 1985 ) , 727 ; _ nucl .
_ * b335 * ( 1990 ) 138 ; _ phys .
_ * t36 * ( 1991 ) 247 .
l.susskind , l. thorlacius and j. uglum , _ phys .
_ * d48 * ( 1993 ) 3743 .
l. freidel and k. krasnov , `` spin foam models and the classical action principle . ''
hep - th/9807092 , n.kawamoto , h.b .
nielsen and n sato , `` lattice chern - simons gravity via ponzano - regge model . ''
hep - th/9902165 .
s. chung , m. fukuma and a. shapere , `` structure of topological lattice theories in three dimensions , '' hep - th/9305080 .
m. reisenberger , `` a left - handed action for euclidean general relativity . '' , _
_ * 14 * ( 1997 ) , 1753 .
varshalovich , a.n .
moskalev and v.k .
khersonskii , `` quantum theory of angular momentum '' .
singapore : world scientific , 1988 .
t. regge , _
nuovo cimento _ * 19 * ( 1961 ) 551 .
u. pachner , `` shellings of simplicial balls and p.l .
manifolds with boundary , '' _ discr.math .
_ * 81 * ( 1990 ) , 37 ; `` p.l homeomorphic manifolds are equivalent by elementary shellings , '' _ europ.j.combinatorics _ * 12 * ( 1991 ) , 129 . v.g . turaev and o.ya .
viro , `` state sum invariants of 3-manifolds and quantum 6j - symbols , '' _ topology_*31 * ( 1992 ) , 865 .
s. mizoguchi and t. tada , `` 3-dimensional gravity from the turaev - viro invariant , '' hep - th/9110057 .
f.a . bais and n.m .
muller , `` topological field theory and the quantum double of su(2 ) '' , hep - th/9804130 .
h. ooguri , `` partition functions and topology - changing amplitudes in the three - dimensional lattice gravity of ponzano - regge , '' _
nucl.phys._*b382 * ( 1992 ) 276 . h. ooguri and n.sasakura , `` discrete and continuum approaches to three - dimensional quantum gravity , '' _
mod.phys.lett._*a6 * ( 1991 ) 3591
. m. blau , i. jermyn and g. thompson , `` solving topological field theories on mapping tori , '' _ phys .
lett . _ * b383 * ( 1996 ) 169 , hep - th/9605095 , ic/96/74 .
brown and m. henneaux , `` central charges in the canonical realization of asymptotic symmetries : an example from three - dimensional gravity . '' _
comm.math.phys._*104*(1986 ) , 207 .
o. coussaert , m. henneaux and p. van driel , `` the asymptotic dynamics of three - dimensional einstein gravity with a negative cosmological constant . ''
gr - qc/9506019 .
j. ambjrn and r. loll , `` non - perturbative lorentzian quantum gravity , causality and topology change , '' hep - th/9805108 , j. ambjrn , j. jurkiewicz and r. loll , `` a non - perturbative lorentzian path - integral for gravity '' hep - th/0002050 , j. ambjrn , j. jurkiewicz and r. loll , `` lorentzian and euclidean quantum gravity - analytical and numerical results '' hep - th/0001124 . c. bachas and p.m.s .
petropoulos , _ comm.math.phys .
_ * 152 * ( 1993 ) , 191 .
m. fukuma , s. hosono and h. kawai , `` lattice topological field theory in two - dimensions , '' hep - th/9212154 .
g. carbone , m. carfora and a. marzuoli , `` wigner symbols and combinatorial invariants of three - manifolds with boundary , '' dfnt - t 14/98 , sissa 118/98/fm , `` invariants of spin networks with boundary in quantum gravity and tqft , '' gr - qc/9905076
. j. maldacena , `` the large n limit of superconformal field theories and supergravity '' , _ adv .
* 2 * ( 1998 ) 231 , hep - th/9711200 ; s. gubser , i. klebanov and a. polyakov , `` gauge theory correlators from noncritical string theory '' , _ phys .
lett . _ * b428 * ( 1998 ) 105 , hep - th/9802109 .
e. witten , `` anti de sitter space and holography '' , _ adv .
_ * 2 * ( 1998 ) 2253 hep - th/9802150 .
j.w . barrett and t.j .
foxon , `` semi - classical limits of simplicial quantum gravity , '' gr - qc/9310016 .
s. davids , `` semiclassical limits of extended racah coefficients , '' gr - qc/9807061 .
d. gross and a. migdal , _ phys .
* 64 * ( 1990 ) 127 , t. banks , m. douglas , n. seiberg and s. shenker , _ phys .
_ * b238 * ( 1990 ) 279 , g. moore , n. seiberg and m. staudacher , _ nucl .
_ * b362 * ( 1991 ) 665 .
g. t hooft , `` transplanckian particles and the quantization of time '' , _ class .
_ * 16 * ( 1999 ) , 395 ; gr - qc/9805079 .
a. ashtekar , j. baez , a. corichi and k. krasnov `` quantum geometry and black hole entropy , '' _ phys .
* 80 * ( 1998 ) 904 , gr - qc/9710007 .
v. suneeta , r.k .
kaul and t.r .
govindarajan , `` btz black hole entropy from ponzano - regge gravity , '' gr - qc/9811071 .
h. dorn and h .- j .
otto , `` two and three - point functions in liouville theory , '' hep - th/9403141 .
a.b . zamolodchikov and al.b .
zamolodchikov , `` structure constants and conformal bootstrap in liouville field theory , '' hep - th/9506136 .
b. ponsot and j. teschner , `` liouville bootstrap via harmonic analysis on a non - compact quantum group , '' hep - th/9911110 .
a. giveon , d. kutasov and n. seiberg , `` comments on string theory on @xmath0 '' , _ adv .
* 2 * ( 1998 ) 733 , hep - th/9806194 , and references herein ; j. maldacena and h. ooguri , `` strings in ads(3 ) and sl(2,r ) wzw model '' , hep - th/0001053 ; a. kato and y. satoh , `` modular invariance of string theory on ads(3 ) '' , hep - th/0001063 .
g. arcioni , s. de haro and m. oloughlin , work in progress .
h. steinacker , `` quantum anti - de sitter space and sphere at roots of unity . ''
hep - th/9910037 .
a. jevicki and s. ramgoolam , `` non - commutative gravity from the ads / cft correspondence '' _ jhep _ * 9904 * ( 1999 ) 32 , hep - th/9902059 .
s. randjbar - daemi and d. jatkar , `` type iib string theory on @xmath174 '' , _
phys.lett . _
* b460 * ( 1999 ) 281 , hep - th/9904187 .
l.d . fadeev and o. tirkkonen , `` connections of the liouville model and the @xmath143 spin chain , '' _ nucl .
_ * b453 * ( 1995 ) 647 , hep - th/9506023 .
g. t hooft , `` quantum gravity as a dissipative deterministic system '' , _ class.quant.grav . _
* 16 * ( 1999 ) 3263 ; gr - qc/9903084 .
`` black holes : the membrane paradigm '' eds . k.s .
thorne , r.h .
price and d.a .
macdonald , ( yale university press , london , 1986 ) .
parikh and f. wilczek , `` an action for black hole membranes '' , _ phys.rev .
_ * d58 * ( 1998 ) , 064011 , gr - qc/9712077 .
j.w . barrett and l. crane , `` relativistic spin networks and quantum gravity '' , _ j. math .
* 39 * ( 1998 ) 3296 , gr - qc/9709028 .
j.w . barrett and l. crane , `` a lorentzian signature model for quantum general relativity , '' gr - qc/9904025 .
j.w . barrett and r.m .
williams , `` the asymptotics of an amplitude for the four - simplex '' , gr - qc/9809032 . m. reisenberger and c. rovelli , `` sum over surfaces formulation of loop quantum gravity '' , _ phys.rev . _ * d56 * ( 1997 ) , 3490 . m. reisenberger , `` a lattice world - sheet sum for 4-d euclidean general relativity '' , gr - qc/9711052 . j. barrett and l. crane , `` relativistic spin networks and quantum gravity '' , _ j.math.phys . _
* 39 * ( 1998 ) 3296 . | boundary actions for three - dimensional quantum gravity in the discretized formalism of ponzano - regge are studied with a view towards understanding the boundary degrees of freedom .
these degrees of freedom postulated in the holography hypothesis are supposed to be characteristic of quantum gravity theories .
in particular it is expected that some of these degrees of freedom reside on black hole horizons .
this paper is a study of these ideas in the context of a theory of quantum gravity that requires no additional structure such as supersymmetry or special gravitational backgrounds .
lorentzian as well as euclidean regimes are examined . some surprising relationships
to liouville theory and string theory in @xmath0 are found .
gr - qc/0002092 |
the dependence of the spectrum of cosmological fluctuations in an inflationary universe on hidden assumptions about the physics on length scales much smaller than the planck length has been recently studied in refs .
_ a priori _ this dependence comes about since in typical scalar - field - driven inflationary models the duration of the period of expansion is so long that the physical wavelengths of comoving modes which correspond to the present large - scale structure of the universe are much smaller than the planck length at the beginning of inflation . in weakly coupled scalar field models of inflation ,
the spectrum of fluctuations is calculated by assuming that fluctuation modes start out in the vacuum ( i.e. minimum energy density ) state at the beginning of inflation and subsequently evolve as determined by the equations of motion .
it is hence not unreasonable to expect that modifications of the physics on trans - planckian scales could affect the final spectrum of fluctuations .
while the correct theory of trans - planckian physics is not known , possible effects of the new physics can be modeled by changes in the dispersion relation of the fields corresponding to linear cosmological fluctuations .
such an approach was used earlier @xcite to study the possible dependence of the spectrum of black hole radiation on trans - planckian physics .
indeed , for a class of dispersion relations which deviate so strongly from the usual one such that the evolution of the mode functions is highly non - adiabatic and the effective frequency becomes imaginary , it was found that the spectrum of fluctuations is modified in a significant way compared to what is obtained for a linear dispersion relation @xcite . as shown in @xcite , the final spectrum of fluctuations
is unchanged if the change in the dispersion relation is not too drastic , the wave function evolves adiabatically , and hence the wkb approximation is valid throughout the evolution .
two major deficiencies in the previous work concerned the choice of initial conditions and the fact that complex dispersion relations were required in order to obtain significant deviations from the standard results . in the previous papers , the assumption was made that the state starts out as the one minimizing the energy density .
however , if the dispersion relation differs greatly from the linear one , in particular if it becomes complex , the physical motivation for choosing this state becomes unclear .
this problem does not arise in a bouncing universe . in this case , modes corresponding to present cosmological scales could well have had a physical wavelength smaller than the planck length at the bounce point , but at early ( pre - bounce ) times , the wavelength was larger than the planck length .
in this context , initial conditions can be set up at some very early pre - bounce time during which the mode obeys a linear dispersion relation and the physical meaning of the minimum energy density state is clear .
it is thus interesting to study the spectrum of fluctuations in a bouncing universe which is asymptotically flat such that the choice of initial vacuum becomes unambiguous .
it is also interesting to exhibit a case where complex dispersion relations are not needed .
although complex frequencies are standard in classical physics and in quantum mechanics , they seem to be more problematic in the context of quantum field theory . in this article , we find an example of a bouncing universe in which the spectrum of fluctuations depends sensitively on changes in the physics at length scales smaller than the planck length , without requiring complex frequencies at high wavenumbers , thus improving on the second major deficiency of previous work .
another reason for investigating the dependence of the spectrum of fluctuations on modifications of trans - planckian physics in a bouncing universe is that one could expect ( however , as we show , this is not correct ) that the deviations in the spectrum incurred during the pre - bounce period when the wavelength is blue - shifting in the trans - planckian regime will cancel with the deviation in the post - bounce trans - planckian regime when the mode is being red - shifted .
such a result would then lead to the expectation that the results on the dependence of the spectrum of fluctuations on changes in the trans - planckian physics might be different in the case of black holes and ( non - bounce ) inflationary cosmology . a third reason for analyzing a bouncing universe
is that there has recently been a lot of interest in the field of string cosmology in bouncing universe models . in the einstein frame ,
the pre - big - bang scenario @xcite starts in a collapsing dilaton - dominated phase , and the same is true in the ekpyrotic scenario @xcite .
our arguments show that trans - planckian effects could change the predictions of standard cosmological perturbation theory in these examples .
a model for the evolution of the scale factor in a bouncing inflationary universe is @xmath0 , where @xmath1 denotes the value of @xmath2 at the center of the bounce , and @xmath3 is the hubble expansion ( contraction ) rate during the period of exponential expansion ( contraction ) .
the idea of the calculation is to assume an initial spectrum of fluctuations at some early time @xmath4 when the wavelength of all modes of interest is much larger than the planck length , to propagate the perturbation modes through the bounce and calculate the final spectrum at time @xmath5 . in this model , one of the difficulties encountered in the previous work @xcite , namely the problem of specifying the initial state in the trans - planckian regime ,
is overcome .
however , it is necessary now to set up initial conditions on wavelengths which typically also are much larger than the hubble radius .
the latter problem can be overcome by considering a bouncing universe which is radiation - dominated in the asymptotic past and future , e.g. @xcite @xmath6 , where @xmath7 and @xmath8 are constants . another toy model
in which it is possible to set up well motivated initial conditions on scales much smaller than the hubble radius and at the same time much larger than the planck scale is a model in which the universe is asymptotically flat both at large positive and negative times : @xmath9 where @xmath10 and @xmath11 ( with @xmath12 ) denote the asymptotic size of the universe and the size at the bounce , respectively , and @xmath13 determines the time scale of the bounce , i.e. the time over which the scale factor changes . in the above , we are using conformal time @xmath14 , and employing the convention that the scale factor carries dimension of length while the comoving coordinates ( including conformal time ) are dimensionless . finally , a comment is in order on the issue of backreaction .
it has been shown in refs .
@xcite that the energy density of the particles created by trans - planckian physics can contribute in a significant way to the background energy density .
we do not address this interesting question since this clearly comes as a second step once a case where the spectrum is modified has been explicitly exhibited . in this paper
we concentrate on the first question since it would be useless to treat the backreaction question if no well - motivated example where the spectrum is modified can be found .
however , it should be clear that , once this is done ( as shown in the present article ) , this problem becomes the central question in the study of trans - planckian physics in cosmology .
this article is organized as follows . in section
ii , we review the arguments which show that no change in the final spectrum of fluctuations of a scalar matter field in a bouncing universe is expected independent of the dispersion relation , provided that the wkb approximation remains justified . we give a qualitative reason to expect changes in the amplitude of the final spectrum for dispersion relations for which the adiabatic approximation breaks down .
as mentioned above , our analysis assumes that perturbation theory remains justified and back - reaction can be neglected @xcite . in section iii
we study a concrete model , namely the asymptotically flat bouncing universe given by ( [ model2 ] ) with a dispersion relation modified in the trans - planckian regime according to the prescription of corley and jacobson @xcite .
we show that , as expected , both the overall amplitude and the shape of the spectrum differ from what is obtained for the linear dispersion relation , without ever requiring the effective frequency to become imaginary .
the change in the spectrum is produced by an interesting combination of effects due to the modified dispersion relation and the driving term for parker particle production @xcite .
the equation of motion for a minimally coupled free scalar matter field @xmath15 in a universe with scale factor @xmath16 in momentum space takes on the simple form @xmath17 making use of the re - scaling @xmath18 where @xmath19 denotes the comoving wavenumber linked to the physical wavenumber by the relation @xmath20 . in the above
, a prime denotes the derivative with respect to conformal time .
the previous equations are valid for a spacetime with flat spatial sections .
the advantage of considering a test scalar field and/or gravitational waves is that we do not need to address the origin of the dynamics of the scale factor .
this is in contrast to the case of density perturbations .
however , the friedmann equation reads @xmath21 , where @xmath22 and @xmath23 is the energy density .
this means that a bounce is consistent with the einstein equations only if @xmath24 . for @xmath25
, it would imply @xmath26 for each component .
therefore , it seems that it is inconsistent to use eq .
( [ eom1 ] ) if the scale factor behaves such that there is a bounce .
there are two ways out of the previous argument .
the first one is the following .
if @xmath24 , then eq . ( [ eom1 ] ) takes the form @xmath27 .
in addition the eigenfunctions of the three - dimensional laplacian operator are no longer planes waves . however ,
if we restrict ourselves to modes with wavelengths much smaller than the curvature radius , then they do not feel the curvature of the spacelike section and we can safely work with eq .
( [ eom1 ] ) .
it could be checked that this is indeed the case for the wavenumbers considered in the example presented in the following section .
another possibility is to use the `` nonsingular universe construction '' @xcite . in this case , it is definitely possible to get a spatially flat bounce .
what happens is that the higher derivative terms become important at the bounce and enable the evasion of the previous argument by supplying other terms in the analog of the first friedmann equation . in conclusion , although in this article we consider a scale factor with a bounce , we can nevertheless work with eq .
( [ eom1 ] ) , i.e. we let the nonsingular universe construction enter only via the scale factor . the method introduced in @xcite to study the dependence of the spectrum of fluctuations on trans - planckian physics is to replace the linear dispersion relation @xmath28 by a non standard dispersion relation @xmath29 where the function @xmath30 is _ a priori _ arbitrary . in the context of cosmology
, it has been shown in ref .
@xcite that this amounts to replacing @xmath31 appearing in ( [ eom1 ] ) with @xmath32 defined by @xmath33.\ ] ] for a fixed comoving mode , this implies that the dispersion relation becomes time - dependent .
various forms for @xmath34 have been used , e.g. @xmath35 , \quad n_{\rm eff}^2(n,\eta ) = n^2 + n^2 b_m \biggl[{{\ell_{_{\rm c } } } \over { \lambda(\eta ) } } \biggr]^{2m},\ ] ] the first being a generalization of the relation studied by unruh @xcite whereas the second one is a generalization of the relations used by corley and jacobson @xcite . in the above
, @xmath36 is the physical wavelength , @xmath37 is the cutoff length ( which can be taken to be the planck length ) , @xmath38 and @xmath39 are integers , and @xmath40 is a real number which can be either positive or negative .
thus , the equation of motion to be analyzed is @xmath41\mu \ ,
= \ , 0 \ , .\ ] ] for fixed comoving wavenumber @xmath42 , the evolution of @xmath43 depends crucially on which region @xmath14 lies in .
the first region is the trans - planckian region with @xmath44 let us first consider the standard inflationary scale factor and a monotonic dispersion relation [ like the first one in eq .
( [ disp12 ] ) or the second one with @xmath45 .
this means that , initially , the term @xmath32 dominates in eq .
( [ eom2 ] ) .
the initial conditions are fixed in this region and since the wkb approximation is applicable , we can choose the initial state as the `` minimizing energy state '' @xcite .
then , the ( positive frequency ) solution is given by @xmath46 \
, , \ ] ] where @xmath47 is some initial time .
the second region corresponds to @xmath48 in this region the mode has reached the linear part of the dispersion relation .
the general solution in region 2 is the plane wave @xmath49 with constant coefficients @xmath50 and @xmath51 . for the standard dispersion relation ,
the initial conditions are fixed in this region .
the usual choice of the vacuum state is @xmath52 , @xmath53 . in general @xmath50 and @xmath51
are determined by the matching conditions between regions 1 and 2 . however ,
if the dynamics is adiabatic throughout ( in particular if the @xmath54 term is negligible ) , the wkb approximation holds and the solution is always given by ( [ wkbsol ] ) .
therefore , if we start with @xmath55 and uses this solution , one finds that @xmath51 remains zero at all times .
deep in the region where @xmath56 the solution becomes @xmath57 i.e. the standard vacuum solution times a phase which will disappear when we calculate the modulus .
the phase @xmath58 is given by @xmath59 , where @xmath60 is the time at which @xmath61 .
the situation changes dramatically if we consider non - monotonic dispersion relations .
this is the case if @xmath62 in eq .
( [ disp12 ] ) or for the dispersion relation introduced in ref . @xcite .
two new features can occur .
either the dispersion relation can become complex ( this is in general the case if @xmath62 ) or the term @xmath54 dominates ( this is the case for the dispersion relation introduced in ref .
@xcite or for the corley / jacobson dispersion relation with @xmath62 for a certain range of comoving wavenumbers ) . in this case , the wkb approximation is violated in region 1 and we expect changes in the final spectrum .
unfortunately , in the context of the standard increasing inflationary scale factor , one also looses the ability to fix natural initial conditions . however , this is no longer true if the spacetime is asymptotically flat because then , at infinity , the term @xmath54 goes to zero , see also ref .
therefore , in this case , we can choose well - motivated initial conditions .
a good example of such a situation is provided by a bouncing universe . from the above considerations
, we expect that in this case the final spectrum is modified and the initial conditions can be fixed naturally . as an additional benefit ,
complex frequencies can be avoided . in the next section ,
we consider a toy model where these qualitative arguments can be implemented concretely , at the level of equations .
we will now illustrate the qualitative arguments of the previous section with a concrete quantitative example .
we take the asymptotically flat bouncing universe given by ( [ model2 ] ) .
we consider the type of modified dispersion relation for which in the case of an expanding inflationary universe the deviations in the spectrum were found @xcite , namely the ( generalization of the ) corley / jacobson dispersion relation ( [ disp12 ] ) with @xmath40 negative . in this case , @xmath30 is given by @xmath63 where @xmath64 is the cutoff physical wavenumber .
the dispersion relation is represented in fig .
[ cjdisp ] . and @xmath65 .
in general @xmath66 vanishes at @xmath67 . for @xmath68 ,
the physical frequency becomes imaginary.,width=321 ] the behavior of the solutions of the equation of motion @xmath69 is determined by the competition between the two terms @xmath70 and @xmath54 given by @xmath71^{-2 m } , \quad \frac{a''}{a}= \frac{2(\ell _ 0-\ell _ { \rm b})}{\eta _ 0 ^ 2\ell _ { \rm b } } \frac{1 - 3(\eta /\eta_0)^2}{[1+(\eta /\eta
_ 0)^2]^3},\ ] ] where @xmath72 is the cutoff length .
these two terms are represented in fig .
when @xmath73 , which is always the case when @xmath74 , the wkb approximation is valid and the fundamental solutions can be written as @xmath75 \rightarrow _ { \eta /\eta _ 0\rightarrow \pm \infty}\frac{1}{\sqrt{2n } } e^{\pm in\eta } .\ ] ] in this case , the physical wavenumbers of the modes are such that they correspond to the linear part of the dispersion relation , see fig .
[ cjdisp ] . in this
situation , natural initial conditions can be chosen .
note that these initial conditions are , in a sense , even `` more standard '' than in the previous studies on trans - planckian physics since usually the initial conditions are set in a region where the dispersion relation is not linear ( but , as explained above , since the wkb approximation holds , meaningful initial conditions can nevertheless be considered ) . on the contrary ,
when @xmath76 the wkb approximation is violated and two independent solutions are given by @xmath77 in this case , we are close to the cutoff scale and @xmath78 being close to zero , see fig .
[ cjdisp ] , the @xmath54 term dominates .
this situation is very similar to the one discussed in ref .
@xcite where this part of the dispersion relation has been named `` the tail '' . in fact , it is possible to quantify exactly the accuracy of the wkb approximation . given an equation of the form @xmath79 , the wkb approximation is valid if the quantity @xmath80 , where the quantity @xmath81 is defined by the following expression @xmath82 this standard criterion can be obtained in the following manner .
the wkb solution , @xmath83 , satisfies the equation @xmath84 exactly .
therefore , one has @xmath85 if @xmath80 . in the present context , @xmath86 is of course equal to @xmath87 .
we have plotted the quantity @xmath88 in fig .
[ qwkb ] . when @xmath89 , @xmath90 and the wkb approximation is satisfied , as expected from the previous discussion .
the quantity @xmath91 increases when one approaches the bounce . at @xmath92 ,
where @xmath93 , there is a divergence .
this point is a turning point .
then for @xmath94 , as can be seen in fig .
[ qwkb ] , @xmath91 remains large and the wkb approximation breaks down in agreement with the considerations above .
, the dashed one @xmath95 .
the values used are @xmath96 , @xmath97 , @xmath98 , @xmath99 , @xmath100 and @xmath101.,width=321 ] , @xmath97 , @xmath98 , @xmath99 , @xmath100 and @xmath102.,width=321 ] in a bouncing universe , the modes start in the linear part of the dispersion relation , pass through the non - linear part and then come back in the standard region .
however , things are not so simple because there exists a range of comoving wavenumbers such that the dispersion relation becomes complex .
although this case is _ a priori _ interesting , it clearly requires more speculative considerations , in particular the quantization in the presence of imaginary frequency modes .
since our goal in this paper is to exhibit a case where everything can be done in a standard manner , we will restrict ourselves to modes which never enter the region where the dispersion relation becomes complex . in the following , we determine the range of comoving wavenumbers we are interested in and for which we are going to calculate the power spectrum .
the maximum of the absolute value of @xmath54 is located at @xmath103 and is given by @xmath104 . therefore , if we restrict ourselves to modes such that @xmath105 then one is sure that , with an unmodified dispersion relation , the term @xmath54 can always be neglected and that the initial spectrum is never changed . clearly , this is not a physical restriction but it renders the comparison with the case with a modified dispersion relation easier .
the minimum of the modified part of the dispersion relation is given by ( from now on , we consider the case @xmath99 and @xmath100 since it is more convenient and does not restrict the physical content of the problem in any way ) @xmath106 ,
\quad n_{\rm inf}\equiv 2\pi\frac{\ell _ { \rm b}}{\ell _ { _ { \rm c}}}.\ ] ] therefore , to maintain a real dispersion relation , we should only consider modes such that @xmath107 . of course , for consistency , the parameters of the model must be chosen such that @xmath108 .
this is the case for the most natural choice , i.e. , @xmath109 . in the following , as already mentioned above , the time @xmath110 such that @xmath111 will play a crucial role . for convenience
we will only consider values of @xmath42 such that this time is determined only by the central peak of @xmath54 and not by the two wings . in practice , this amounts to taking @xmath112 since @xmath113 is the maximum height of the wings .
then it is easy to show that @xmath114 $ ] where @xmath115 where @xmath116 is the largest value of @xmath42 such that @xmath117 remains real .
therefore , the range we are interested in is given by @xmath118 in practice , we have @xmath119 and @xmath120 .
this means that @xmath121 , @xmath122 and that the range reduces to @xmath123 having determined the relevant wavenumbers , we can now choose the initial condition and solve the equation of motion . in the first region
where @xmath124 , we only consider positive frequency modes and we have @xmath125,\ ] ] where @xmath126 is an arbitrary initial time . in the second region ,
where @xmath127 , the solution is given by @xmath128 the lower bound of the integral is _ a priori _ arbitrary .
however , it is very convenient to take it equal to zero because in this case the second branch becomes odd whereas the first one ( i.e. the scale factor ) is even .
then , it is easy to show that @xmath129}\frac{\eta } { \eta _ 0 } + \frac{p^2 + 2p-3}{2p^{5/2}}\arctan\biggl(\sqrt{p}\frac{\eta } { \eta _ 0}\biggr ) \biggr\},\ ] ] where @xmath130 . finally , the solution in the third region where @xmath131 can be written as @xmath132+\frac{c_2}{\sqrt{2n_{\rm eff}}}\exp \biggl [ + i\int ^{\eta } _
{ \eta _ { \rm i}}n_{\rm eff}(n,\tau ) { \rm d}\tau\biggr].\ ] ] the goal is now to calculate the coefficients @xmath133 and @xmath134 . using the continuity of the mode function @xmath135 and of its derivative , we find @xmath136}\frac{e^{i[\omega ( \eta _ { \rm j } ) -\omega ( -\eta _ { \rm j})]}}{\sqrt{n_{\rm eff}(n,-\eta _ { \rm j } ) n_{\rm eff}(n,\eta _ { \rm j})}}\biggl\{-[g'+\alpha g](-\eta _ { \rm j } ) [ f'+\bar{\alpha } f](\eta _ { \rm j})+[f'+\alpha f](-\eta _ { \rm j } ) [ g'+\bar{\alpha } g](\eta _ { \rm j})\biggr\ } , \\ \label{coeffc2 } c_2(n ) & = & \frac{-i}{2w[-\eta _ j(n)]}\frac{e^{-i[\omega ( \eta _ { \rm j } ) + \omega ( -\eta _ { \rm j})]}}{\sqrt{n_{\rm eff}(n,-\eta _ { \rm j } ) n_{\rm eff}(n,\eta _ { \rm j})}}\biggl\{-[g'+\alpha g](-\eta _ { \rm j } ) [ f'+\alpha f](\eta _ { \rm j})+[f'+\alpha f](-\eta _ { \rm j } ) [ g'+\alpha g](\eta _ { \rm j})\biggr\ } , \end{aligned}\ ] ] where we have used the short - hand notation @xmath137 and @xmath138 .
we have also utilized the following definitions : @xmath139 is the wronskian , @xmath140 , and the quantity @xmath7 is @xmath141 using the fact that @xmath34 is even , it is easy to see that @xmath142 . using the same property , one could also simplify the factor @xmath143 in the above equation .
the final result is given by eqs .
( [ coeffc1 ] ) , ( [ coeffc2 ] ) where all the functions are explicitly known except the function @xmath144 .
the dependence on the dispersion relation of the final result is completely encoded in this function .
the time @xmath110 is the solution of an algebraic equation that is not possible to solve explicitly .
however , we can find an approximation if the function @xmath145 is taylor expanded around @xmath103 .
for the function @xmath54 , we use a quadratic least square approximation which gives a better result .
we obtain @xmath146.\ ] ] the two functions and their approximations are represented in fig .
[ approx ] .
we see now that @xmath110 is the solution of a quadratic algebraic equation .
straightforward calculations give @xmath147 \biggl[n^2\biggl(\frac{n}{n_{\rm inf}}\biggr)^2 + \frac{6}{\eta _ 0 ^ 2}\biggr]^{-1}.\ ] ] the function @xmath110 can also be found exactly by numerical calculations . , the dot - dashed curve its quadratic approximation .
the dashed curve represents @xmath95 , and its quadratic approximation is the fourth curve .
values chosen are @xmath96 , @xmath97 , @xmath98 , @xmath148 , @xmath100 and @xmath101.,width=321 ] we have done this computation by means of a fortran code . the comparison of the exact result with its approximation given in eq .
( [ approxetaj ] ) is represented in fig .
the function @xmath149 is not monotonic and has a maximum around @xmath150 for the values of the parameters considered in fig .
[ etaj ] . for this value @xmath145
is almost never smaller than @xmath54 .
it is clear that around this value the approximation will be very good whereas for other values of @xmath42 , especially for @xmath151 , the approximation will be less good .
this is due to the fact that when @xmath42 approaches @xmath152 , the asymptotic value of @xmath145 decreases and the curve opens out at the intersection with @xmath54 . as a consequence , the quadratic approximation employed above breaks down .
this is confirmed by the plots in fig .
the approximation can be less good for other values of the parameters . from the above analysis
, it is clear that @xmath153 is a small number .
we can therefore expand everything in terms of the parameter @xmath154 .
this will allow us to get an analytical estimate of the spectrum .
after tedious but straightforward computations , one obtains the following expressions for the coefficients @xmath155 and @xmath156 @xmath157 + { \cal o}(\epsilon ^2 ) , \label{c1c2a } \\
c_2(n)&= & -\epsilon \frac{i\eta _ 0}{2n } e^{-2i\psi _ { \rm i}(n ) } \biggl [ 2n^2\biggl(1-\frac{n^2}{n_{\rm inf}^2}\biggr)^{1/2 } + n_{\rm b}^2\biggl(1-\frac{n^2}{n_{\rm inf}^2}\biggr)^{-1/2 } + n_{\rm b}^2\biggl(1-\frac{n^2}{n_{\rm inf}^2}\biggr)^{-3/2}\biggr ] + { \cal o}(\epsilon
^2 ) , \label{c1c2b}\end{aligned}\ ] ] where @xmath158 .
several comments are in order at this point .
firstly , the expressions of the two coefficients have exactly the expected form .
when the parameter @xmath159 goes to zero , @xmath155 goes to one and @xmath156 goes to zero , i.e. , we recover an unmodified spectrum .
this is in complete agreement with the fact that , when @xmath160 , the effective wavenumber @xmath34 never penetrates the region where the wkb approximation is violated .
this is in this sense that the parameter @xmath159 is the relevant parameter to expand the spectrum in since it is clearly directly linked to the violation of the wkb approximation .
therefore , in the limit @xmath161 , one recovers the usual spectrum .
secondly , it is interesting to notice that the two terms proportional to @xmath159 in @xmath155 and @xmath156 are the same up to a global phase . to compute the phase in eqs .
( [ coeffc1 ] ) and ( [ coeffc2 ] ) , we have written @xmath162 . since we only need to evaluate @xmath163 , one can use the approximate equation for @xmath34 , see eq .
( [ taexpand ] ) , in the integral @xmath164 .
one finds @xmath165 let us remark that the term @xmath166}$ ] in the expression of @xmath156 does not contribute to order @xmath159 whereas the factor @xmath167}$ ] gives a contribution at this order .
this contribution is exactly such that the two terms proportional to @xmath159 in eqs .
( [ c1c2a ] ) , ( [ c1c2b ] ) have the same absolute value .
thirdly , the factor @xmath168 is always positive since we restrict ourselves to the regime where @xmath169 .
therefore , the above expressions of @xmath155 and @xmath156 are always well - defined .
finally , interestingly enough , the two coefficients @xmath155 and @xmath156 have a similar structure to one of the corresponding coefficients found in ref .
@xcite where another example of a modified spectrum has been exhibited .
we are now in a position where we can calculate the spectrum explicitly .
it is defined by the following equation @xmath170 we will evaluate the spectrum in the region where @xmath171 . in this case ,
@xmath172 and @xmath173 .
expanding everything in terms of the small parameter @xmath159 , one obtains @xmath174 \nonumber \\ & & \times \sin \biggl[2\phi ( n ) -2\psi _ { \rm i}(n)+2n(\eta -\eta _
+ ) \biggr ] + { \cal o}(\epsilon ^2),\end{aligned}\ ] ] where @xmath175 is the time at which @xmath176 and the phase @xmath177 is defined by @xmath178 . again the spectrum has the expected form .
the leading order is the vacuum spectrum with spectral index @xmath179 and the next - to - leading order ( proportional to @xmath159 ) has the form of a complicated function of @xmath42 times superimposed oscillations .
we have plotted this complicated function ( without the superimposed oscillations ) in fig .
[ ampli ] . on the wave number @xmath42 for the values @xmath96 , @xmath97 , @xmath98 , @xmath99 , @xmath100 .
the solid curve is the exact result determined numerically , the dotted curve is obtained using the analytical approximations.,width=321 ] we can check that the correction remains small and that , therefore , the approximation employed in this article is consistent . on the other hand , when the wavenumber @xmath42 approaches @xmath180 , the approximation scheme breaks down but , precisely in this limit , the whole problem becomes pointless in agreement with the considerations above .
we also see that the deviation is minimum around @xmath150 ( for these values of the parameters ) because , as already mentioned above , in this case the term @xmath54 almost never dominates .
although the present example is only a toy model , it is quite interesting to see that it bears some close resemblance with the other example of a modified spectrum found in ref .
this suggests that the features previously exhibited may be valid in general .
we have thus reached our main goal : find an example where the initial conditions can be fixed naturally , where the frequency never becomes complex and where the final spectrum is modified .
note that in this example , the change in the spectrum comes about from an interplay between the modified dispersion relation factor @xmath95 and the factor @xmath54 which is responsible for parker particle production @xcite . for an unmodified linear dispersion relation ,
the @xmath54 term is always negligible and there is no particle production .
however , for our modified dispersion relation , for a range of modes the parker particle production term becomes important and leads to non - adiabatic evolution .
the length of time during which the @xmath54 term dominates depends on the specifics of the dispersion relation , and hence the final spectrum will depend on these specifics .
, @xmath97 , @xmath98 , @xmath148 , @xmath100 .
when @xmath42 approaches @xmath180 the correction blows up and the approximation becomes meaningless.,width=321 ]
in this paper we have studied the dependence of the spectrum of a free scalar field in a bouncing universe on trans - planckian effects introduced via a modification of the free field dispersion relation .
we have found that both the amplitude and the slope of the spectrum depend on the dispersion relation , assuming that the dispersion relation leads to non - adiabatic evolution of the mode functions on trans - planckian physical length scales .
such non - adiabatic evolution is possible without requiring the effective frequency to become imaginary .
this result supports our earlier work which shows that the spectrum of cosmological fluctuations in inflationary cosmology can depend sensitively on trans - planckian physics @xcite .
note that results supporting @xcite were recently also obtained by @xcite ( see also ref .
@xcite ) in the context of a mode equation modified from the usual ones by taking into account @xcite effects coming from a short distance cutoff . for a recent paper on the effects of short - distance physics on the consistency relation for scalar and tensor fluctuations in inflationary cosmology see ref . @xcite .
the calculation we presented here is free of two of the possible objections against the earlier work .
the first objection was that the use of the minimum energy density state as initial state is not well defined on trans - planckian scales if the dispersion relation differs dramatically from the linear one ( which it has to in order to get non - adiabatic evolution ) . in our present work ,
the initial conditions are set in the low curvature region and on length scales larger than the cutoff length but smaller than the hubble radius , where the choice of initial vacuum state is well defined .
the second objection concerned the use of dispersion relations for which the effective frequency is imaginary in some time interval .
no imaginary frequencies are used in the present model .
the results obtained in this paper apply also both to bouncing universe backgrounds described by the two other models mentioned in the introduction . in the case of the first model ( where the scale factor is given by an hyperbolic cosine )
, similar results would hold if we would match the _ local minkowski vacuum _
( wkb vacuum ) at some initial time and express the results in terms of the wkb vacuum state at the corresponding post - bounce time .
however , in that case , the initial conditions are not easy to justify since they have to be set on length scales larger than the hubble radius .
obviously , we could consider that model and let the scale factor make a further transition to an asymptotically radiation- dominated or asymptotically flat stage at very large initial and final times . in this case , it is very likely that results similar to the ones we obtained here in model ( [ model2 ] ) would be obtained .
we would like to thank m. lemoine for comments and careful reading of the manuscript .
we acknowledge support from the brown - cnrs university accord which made possible the visit of j. m. to brown during which most of the work on this project was done , and we are grateful to herb fried for his efforts to secure this accord .
s. e. j. acknowledges financial support from cnpq .
r. b. wishes to thank bill unruh for hospitality at the university of british columbia during the time when this work was completed , and for many stimulating discussions .
j. m. thanks the high energy group of brown university for warm hospitality .
the research was supported in part by the u.s .
department of energy under contract de - fg02 - 91er40688 , task a. j. martin and r. brandenberger , proceedings of the ninth marcel grossmann meeting on general relativity , edited by r. t. jantzen , v. gurzadyan and r. ruffini , world scientific , singapore , 2002 , astro - ph/0012031 . | in this paper , we calculate the spectrum of scalar field fluctuations in a bouncing , asymptotically flat universe , and investigate the dependence of the result on changes in the physics on length scales shorter than the planck length which are introduced via modifications of the dispersion relation . in this model , there are no ambiguities concerning the choice of the initial vacuum state .
we study an example in which the final spectrum of fluctuations depends sensitively on the modifications of the dispersion relation without needing to invoke complex frequencies .
changes in the amplitude and in the spectral index are possible , in addition to modulations of the spectrum .
this strengthens the conclusions of previous work in which the spectrum of cosmological perturbations in expanding inflationary cosmologies was studied , and it was found that , for dispersion relations for which the evolution is not adiabatic , the spectrum changes from the standard prediction of scale - invariance . |
over the era of relativistic nuclear collisions , the statistical - thermal models have been widely used in the analysis of particle production . during the past few years
, these models have been successful in describing particle multiplicities in high energy nuclear reactions at the stage where inelastic collisions between hadrons cease ( chemical freeze - out ) , see e.g. @xcite . the easiest to handle , thence
the most used , statistical model calculational framework is based on the grand canonical ( gc ) ensemble . in the gc approach entropy
is maximized using the constraints of conserved _ ensemble averages _ of energy and charges .
this allows the net charges to fluctuate from sample to sample even though the actual charges brought into the physical reactions were exactly same every time . in the large volume and energy limit
, the error in such an assumption due to charge fluctuations is negligible .
on the other hand , when the involved volumes are small , like in e@xmath0e@xmath1 , p@xmath2 , pa reactions or in peripheral aa reactions , the charge fluctuations allowed in gc formalism give a large theoretical error .
so , the gc ensemble turns out to be inadequate for the analysis of experimental results and the statistical ensemble to be used is rather the canonical one , enforcing exact conservation of relevant charges . in this case
entropy is maximized using the constraint of conserved ensemble average of energy , but charges are not allowed to fluctuate .
statistical models based on the canonical ensemble have been able to reproduce particle multiplicities even in elementary collisions , although in those there is clearly no room for kinetic thermalization @xcite .
also , the strangeness enhancement with increasing system size in p a reactions can be explained by a canonical effect @xcite . in this work
, we investigate whether and to what extent the canonical treatment is relevant for actual heavy ion collisions .
the main difficulty of canonical calculations is related to the involved large values of baryon number and electric charge , which make the numerical computation of canonical partition functions quite hard .
we present here an efficient numerical method to carry out such calculations which enabled us to implement exact charge conservation up to baryon number @xmath3 whereas with old methods it was only possible to reach @xmath4 @xcite with very long computation times . after deriving the needed expressions of mean particle numbers , we present a systematical study of the chemical factors appearing in the canonical statistical model .
along with some comparison between grand - canonical and canonical calculations , we show a comparison with experimental results for peripheral to central a a reactions at ags and sps energies . in section [ gamma ]
, we briefly address the role of the so - called relative strangeness chemical equilibrium , usually parametrized with @xmath5 , in canonical statistical - thermal models .
in nuclear physics , the problem of calculating the relativistic canonical partition function has been handled by several different methods , see e.g. @xcite . here , we employ a very general group theoretical method
applicable to any internal symmetry represented by a semi - simple lie algebra
first introduced by cerulus @xcite . by denoting the set of conserved quantum numbers by @xmath6
, the canonical partition function @xmath7 can be obtained from the usual grand - canonical partition function @xmath8 by using a projecting operator onto the conserved quantum numbers . in the case of @xmath9 internal symmetries of type @xmath10 , the projection takes the form : @xmath11 z_{gc}(t , v,\{\lambda_{c_i}\ } ) .
\label{eq : projection}\ ] ] where @xmath12 is a @xmath10 group parameter and a wick - rotated fugacity factor @xmath13 is introduced , for every charge @xmath14 . in order to calculate @xmath8 in the above
integral , we will assume the boltzmann statistics for all hadron species .
the error involved in such an approximation for final particle multiplicities , when resonance decays is taken into account , is some percent in case of pions and much less for all other hadrons .
quantum statistics may be used , but requires somewhat heavier calculations and disregarding it does not essentially affect any forthcoming argument . in heavy ion reactions , the relevant set of conserved charges is @xmath15 , namely baryon number , strangeness and electric charge . denoting the partition function of hadrons carrying none of the previous charges by @xmath16 , and writing @xmath8 by using the one - particle partition function @xmath17 \int { { \rm d}}^3 p \exp [ -\sqrt{p^2+m_i^2}/t]$ ] for each hadron @xmath18 , we find @xmath19\right\}. \label{pro1}\end{aligned}\ ] ] the direct numerical computation of the triple integral above is a formidable task for @xmath20 .
the integral can be turned into a sum over many indices of modified bessel functions @xmath21 , yet its evaluation is very time consuming and becomes impractical with baryon number larger than @xmath22 @xcite .
therefore , we have chosen another approach , which has been applied in elementary collisions @xcite . from ( [ pro1 ] )
one finds the integrand to be violently oscillating for large net quantum numbers , so it is worth trying to eliminate analytically the source of strongest oscillation ; in heavy ion reactions , the baryon number is always the largest of the set @xmath23 , so it is certainly the most beneficial to eliminate the integration over @xmath24 .
in fact , this can be done by taking advantage of a special feature of baryon number , that is no elementary hadron exists with @xmath25 .
we first rewrite equation ( [ pro1 ] ) in the form @xmath26 \exp\left[\sum_{m,\overline{m}}z_i^1 e^{{{\rm i}}(s_i\phi_s+q_i\phi_q)}\right],\label{canpart}\end{aligned}\ ] ] where the two summations run over baryons and mesons , respectively . as @xmath27 for all of the baryons , we can write the baryon summation as : @xmath28 where baryonic and antibaryonic terms have been separated . introducing the notation @xmath29 , the above equation can be rewritten as : @xmath30 substituting this expression in eq .
( [ canpart ] ) and changing the baryon variable according to @xmath31 allows us to perform analytically the integration in @xmath32 , which yields a modified bessel function i@xmath33 : @xmath34 . \nonumber\end{aligned}\ ] ] thus , we are left with a double integration only which can be then performed numerically with no major problem provided that , @xmath24 being very large , the uniform asymptotic expansion of bessel functions for large orders is used to compute i@xmath33 .
it is worth noting , that the form ( [ partt ] ) does not involve a saddle - point approximation used in @xcite .
actually , for very small systems it is not even applicable , because the integrand ( [ pro1 ] ) is rather smoothly distributed in the group parameter space @xmath35 , so the basic assumption of the saddle - point method fails .
the mean particle numbers of primary hadrons ( i.e. those directly emitted from the hadronising source ) @xmath36 are obtained from the equation ( [ partt ] ) @xcite by taking the derivative of the canonical partition function with respect to a fictitious fugacity @xmath37 : @xmath38 where the quantity @xmath39 is called _ chemical factor _ @xcite . in the large volume ( thermodynamical ) limit , this expression becomes the gc one , i.e. @xmath40 . the canonical baryon chemical factor can be defined as that relevant to baryons with vanishing strangeness and electric charge , e.g. neutrons : @xmath41 so that in the gc limit one gets : @xmath42 similarly , one can define chemical factors @xmath43 and @xmath44 for strangeness and electric charge respectively .
it must be emphasized that , unlike in the gc framework where the total fugacity is actually a product of different charge fugacities powered to the number of charges carried by the hadron , the canonical chemical factors can not be factorized and must be calculated , as indicated in equation ( [ cmean ] ) , as a ratio of partition functions for different sets of quantum numbers .
the relative differences between chemical factors and their gc counterparts and quantities related to these differences are defined , in this paper , as _
canonical effects_. in addition to the comparison between full canonical calculations and gc ones , it is very important to assess whether the enforcement of exact strangeness conservation with gc treatment of baryon number and electric charge ( _ strangeness canonical ensemble _ ) is indeed sufficient to produce results close enough to full canonical ones . here
, we only quote the analytical form of strangeness - canonical partition function @xmath45 derived using the method introduced in @xcite : @xmath46 in the equation above , we denote the sum of one particle partition functions carrying the same quantum numbers as @xmath47 by @xmath48 , and similarly for other kinds of multistrange hadrons , while @xmath49 stands for the sum @xmath50 . in this case ,
mean particle numbers read : @xmath51
first of all , we point out that we do not perform fits to experimental results in this paper .
indeed , as our main motivation is to test the statistical canonical formalism in the heavy ion collision regime , we only take some suggestive values for freeze - out temperatures and baryon densities from our recent work on the analysis of particle multiplicities in heavy ion systems in the laboratory momentum range from gsi sis 1.7 @xmath52 gev to cern sps 158 @xmath52 gev @xcite and make comparisons between canonical and grand - canonical calculations for those values .
the peculiar feature of canonical formalism is a nonlinear volume behaviour of mean particle numbers because of the dependence of chemical factors on the volume according to eqs .
( [ cmean],[pro1 ] ) . to show this effect
, we first fix the baryon density @xmath53 so that the dependence of the chemical factors on the volume turns into a dependence on net baryon number .
as the temperature is fixed too , then the grand - canonical baryon chemical potential can be calculated and the difference with the corresponding chemical factor can be studied .
furthermore , net strangeness @xmath54 and electric charge @xmath55 are fixed by the initial conditions of the collision ( namely zero and @xmath56 respectively ) .
the canonical effect is the most significant at comparatively low energies , such as 1.7 gev per nucleon in sis au au collisions . in our recent analysis @xcite
we find the best fit to experimental results with thermal parameters @xmath57 mev and @xmath58 mev .
these values yield , in the pointlike hadron gas framework , a baryon density lower than 0.1 @xmath59 , which is used as a reference value in order to study the canonical effect in heavy ion collisions as a function of the number of participants - i.e. net baryon number - in the collision . in figure [ fig1 ]
the chemical factor @xmath60 is shown in comparison with its gc limit @xmath61 . under these circumstances , the canonical effect related to the baryon chemical factor
is found to decrease from 2.1% at @xmath62 down to 0.7% at @xmath63 , and to further decrease rather slowly towards the gc limit with increasing baryon number or volume . for central au
au reactions , the number of participating nucleons is more than 300 , so we can conclude that the canonical baryon chemical effect is negligible .
however , this is not the case for strangeness . in figure [ fig2 ]
the theoretical production ratio k@xmath0 over number of participating nucleons is shown to point out the canonical strangeness effect .
note that we have been able to compute the full chemical factor even for @xmath64 .
the ratio increases very slowly towards the gc limit , thus strangeness must always be handled canonically when analysing sis results .
the relative difference between full canonical and strangeness - canonical results is 13% for @xmath62 , decreasing to 4% for @xmath63 and to 1% for @xmath65 .
it must be pointed out that the above results are calculated using an isospin symmetric initial configuration whereas @xmath66 for gold nucleus . however
, this variation essentially gives no change on the relative differences above , although the ratio k@xmath67 decreases naturally due to the initial neutron excess .
in au au collisions at ags energies ( 11.6 @xmath52 mev momentum ) , the chemical freeze - out temperature is found to be around @xmath68 mev and the baryon density close to the normal nuclear density @xcite . using these values
, we show the canonical baryon chemical factor in comparison with the gc limit in figure [ fig3 ] keeping a symmetric isospin initial condition .
relative difference between @xmath60 and @xmath61 decreases rather quickly from 23% at @xmath62 to 1% at @xmath65 . a more realistic comparison of results in au
au with pp collisions requires different initial isospin configuration to be into account . by using temperature and
baryon density quoted above , @xmath60 at @xmath69 turns out to be 81.9 and @xmath70 with @xmath71 .
recent experimental results on ags k@xmath72 ratios with increasing centrality @xcite serve as a basis for a further study of strangeness production behaviour at @xmath68 mev . in figure [ fig4 ]
we plot the calculated k@xmath67 with results obtained in peripheral to central au
au , si au and si al collisions .
it can be seen that all the way up from @xmath62 to @xmath73 the full canonical and strangeness - canonical results are essentially the same .
all central reaction results lie on the region where the canonical effects are negligible and the gc formalism applies .
strangeness enhancement in au
au system does not definitely look like being of canonical origin , whereas si
au and si al follow roughly the curve at the normal nuclear density ( not shown in the figure ) .
the above argument also applies for the k@xmath74 enhancement pattern shown in figure [ fig5 ] . whereas the kaon enhancement with increasing centrality seems not to be compatible with a purely canonical effect , inelastic pp and _ central _ p a and a a systems considered in ref .
@xcite show a different behaviour with increasing volume .
therein , it was found that k@xmath75 ratios clearly follow the canonical curves . the best fit temperature for the multiplicities measured by na49 experiment at sps in central pb pb reactions at 158 @xmath52 gev is found to be @xmath76 mev @xcite . in figure [ fig6 ]
we show the strange baryon enhancement by using a baryon density 0.3 @xmath59 .
although the baryon density in our analysis @xcite was found to be @xmath77 @xmath59 this larger value is used in order to probe the largest reasonable canonical effect .
all results in figure [ fig6 ] are normalized to the _ full canonical _ baryon multiplicities per participant at the point @xmath78 .
this choice reveals the slight difference between the results obtained using the @xmath54-canonical approximation and the @xmath79-canonical calculation , and forces both canonical methods to reach for the same gc limit .
an interesting feature here is the fact , that the @xmath54-canonical method leads to an overestimation of the canonical effect .
corresponding curves have also been calculated in ref .
@xcite within the strangeness canonical ensemble with some numerical approximations not used here .
this enhancement picture has been calculated again with the assumption of an initial symmetric isospin configuration .
changing the freeze - out conditions to @xmath80 mev and @xmath81 @xmath82 , the total relative enhancement of @xmath83 from canonical @xmath69 to gc @xmath84 raises to 30% for @xmath85 s , 241% for @xmath86 s and 918% for @xmath87 s .
sps na49 collaboration has also measured the participant dependence of k@xmath88 ratio in pp , c c , si si , s s and pb pb reactions @xcite . in fig .
[ fig7 ] we show that the roughly linear enhancement pattern is far away from the purely canonical shape . in the same figure
, one can see the weak dependence of the k@xmath88 ratio on thermal parameters .
indeed , theoretical curves always lie well above the measured points , which is a strong indication that strangeness is not in complete chemical equilibrium .
usually , this lack of equilibrium is taken into account by introducing a parameter @xmath5 ( see discussion in section [ gamma ] ) , which , in the boltzmann approximation , appears as a linear coefficient in front of the k@xmath75 ratios .
taking our best fit value @xmath89 @xcite brings the theoretical curves asymptotically to the experimental central pb
pb reaction points .
curves with @xmath80 mev ( lowest two ) assume initial isospin symmetry , whilst the @xmath54 canonical curve with @xmath90 mev is corrected for the condition @xmath84 .
it is worth quoting our result for pp to central pb
pb k@xmath91 enhancement using @xmath80 mev and @xmath81 @xmath59 taking into account the initial isospin : @xmath92 with @xmath93 .
experimentally , this ratio is about 2 as it can be inferred by looking at the points in fig .
in statistical model analysis of heavy ion reactions it is found that the assumption of full chemical equilibrium of strangeness is mostly not satisfied ( @xcite and this work ) .
thus , a concept of relative chemical equilibrium was inroduced @xcite and has been applied in many subsequent works .
the relative chemical equilibrium means that the strange quark production is not fully equilibrated , but the distribution of strange quarks among different hadron species occurs according to statistical equilibrium ( _ statistical coalescence _ ) . in order to parametrize the relative chemical equilibrium , another constraint in the derivation of the relevant statistical operator
is needed . in the gc formalism , this constraint is the conservation of the average number of strange and antistrange quarks in addition to usual constraints for energy and charge conservation .
note that this constraint is not owing to any natural symmetry of strong interactions like baryon number , electric charge and strangeness conservation .
hence , as the number of produced strange quarks to be distributed among final hadrons is unknown , it always leads to the introduction of another free parameter to be determined _ a posteriori . _ using the aforementioned constraints , the grand canonical ( gc ) partition function can be written as : @xmath94,\ ] ] where the chemical potentials for all the relevant charges have been grouped under the symbol @xmath95 .
the fugacity factor for the constrained number of strange quarks @xmath96 is usually called @xmath5 @xcite . in the full canonical picture , where charges are not allowed to fluctuate from sample to sample
, the constraint of the average number of strange and antistrange quarks can be turned into the enforcement of a fixed _ exact _ number of @xmath97 quarks .
corresponding @xmath98 symmetry yields another integration variable and another phase factor , @xmath99 into the projection integral ( [ pro1 ] ) @xcite . following the procedure described in section [ model ] the average number of hadrons @xmath100 can be written as : @xmath101 in the grand canonical limit , the corresponding number is @xmath102 as an illustrative example , take the average number of @xmath103 mesons .
it does not carry any relevant charge , but the valence quark content is @xmath104 . in the gc limit @xmath105 .
this leads us to the identification @xmath106 there is one major caveat in those formulae and the use of @xmath5 .
indeed , the canonical formalism for the _ net conserved charges _ in the system is to be applied in any reaction where they are known from the initial state ; if the system is large enough , the fluctuations allowed in gc ensemble give negligible deviation from canonical results and one is allowed to use formulae such as eq .
( [ gcgs ] ) . however , the case of the number of strange quarks @xmath107 is instrinsically different .
in fact , this number is not known from the initial state and may undergo large dynamical fluctuations from event to event , well beyond those ( small ) predicted by the gc formalism and its related fugacity @xmath5 . if this was the case , @xmath5 in eq .
( [ gcgs ] ) must be understood as a sort of _ average _ fugacity rather than a proper fugacity . on the other hand , when the expected mean number of @xmath104 pairs is small
, the canonical formalism should be used ( i.e. eq . ( [ cangs ] ) ) but the probability distribution of generating a given number of @xmath104 pairs should be known in advance . as a reasonable _ ansatz _ , one can assume those pairs are independently produced so that their distribution is a poissonian @xcite .
we have presented a method to calculate particle densities in the canonical framework of the statistical model , which can be effectively used in heavy ion reactions at very large values of baryon number .
this method has allowed us to study the applicability of different approximations in chemical analysis , namely the grand - canonical and the strangeness - canonical ensembles .
it is found that the full @xmath79 canonical formalism is only needed for elementary and very small nuclear systems up to baryon number @xmath22 . for larger systems , the baryon number and the electric charge
can be safely handled in grand - canonical manner .
for central au au and pb pb reactions , exact strangeness conservation is needed at gsi sis energies ( @xmath108 gev a ) , while at ags and sps even strangeness can be handled grand - canonically .
certainly , this also applies to much higher energy rhic and lhc collisions .
the observed strangeness enhancement from peripheral to central nuclear collisions @xcite is generally not reproduced by the canonical curves with fixed temperature and baryon density whilst the ags p p to p a to a a _
central _ reaction measurements are found to follow the canonical results @xcite .
this is an indication of a possible change in reaction systematics going from peripheral to central collisions . the strange hadron enhancement in sps from p p to pb
pb reactions is only qualitatively reproduced by canonical model . whilst @xmath5 was not used in the derivation of our numerical results , the weak dependence of k@xmath75 ratio on the thermal parameters ( shown in fig .
[ fig7 ] ) serves as a further compelling piece of evidence in favour of incomplete strangeness chemical eqilibration in cern sps heavy ion reactions , confirming previous findings @xcite .
+ we would like to acknowledge stimulating discussions with esko suhonen , jean cleymans , krzysztof redlich , ulrich heinz and lszl csernai .
one of us ( a.k . ) acknowledges the support of the bergen computational physics laboratory in the framework of the european community - access to research infrastructure action of the improving human potential programme . | we study the effect of enforcing exact conservation of charges in statistical models of particle production for systems as large as those relevant to relativistic heavy ion collisions . by using a numerical method developed for small systems ,
we have been able to approach the large volume limit keeping the exact canonical treatment of all relevant charges , namely baryon number , strangeness and electric charge .
hence , we hereby give the information needed in a hadron gas model whether the canonical treatment is necessary or not in actual cases .
comparison between calculations and experimental particle multiplicities is shown .
also , a discussion on relative strangeness chemical equilibrium is given . |
inflation @xcite maintains its position as the favourite model for the origin of structure in the universe .
the reasons are two - fold .
firstly , the gaussian adiabatic and nearly scale - invariant density perturbations that the usual models produce currently offer the best framework in which to interpret observational data on structures in the universe . and
secondly , it is a extremely simple paradigm within which to make theoretical predictions ; for example the cmbfast program allows the prediction of microwave background anisotropies to better than one percent accuracy , and as yet calculations in the rival , much more theoretically challenging , topological defects theories @xcite lag some way behind @xcite . however , the inflationary paradigm is quite a broad one , and there exists a wide range of different implementations of the inflationary idea . in this article
i aim to give a flavour of the complexity of models which we might one day find ourselves forced to deal with , if we are to understand structure formation .
i will also stress the importance of these notions for issues such as cosmological parameter estimation from the cosmic microwave background @xcite , which as we will see can not be divorced from the issue of the initial conditions for structure formation .
in particular , i seek to provide at least partial answers and discussion centred around the following key questions : * what do the simplest models of inflation predict ?
* is it possible to test the idea of inflation ? *
if the simplest models are right , what do we learn about the inflationary mechanism ?
* if the simplest models are not right , what happens then ?
inflation is defined as a period of accelerated expansion during the very early stages of the universe s evolution . however to fix the physical significance of inflation in your mind , it is better to rewrite that somewhat , as follows @xmath0 comoving units are the right ones to use , since to a first approximation everything , including linear density perturbations , gets dragged along with the expansion and we are interested in how things evolve relative to that .
the hubble length is the key characteristic scale in the universe , setting the length scales over which causal processes can act .
what this tells us therefore is that inflation is precisely the condition that we see a smaller and smaller region of the comoving universe as inflation proceeds . in that sense , inflation is rather akin to ` zooming in ' on a small region of the initial universe .
it is well known that inflation solves the classic cosmological problems .
for example , the flatness problem is solved because the density parameters of matter , @xmath1 , and a possible cosmological constant , @xmath2 , obey the friedmann equation which can be written as @xmath3 by definition , during inflation the right - hand side tends to zero and so we rapidly approach @xmath4 which is the condition for a spatially - flat universe .
although after inflation is over this tells us that the universe must necessarily be evolving away from flatness , provided sufficient inflation took place we would be forced so close to flatness that all the subsequent evolution , from the end of inflation to the present , would be insufficient to to move us significantly away from flatness see figure [ f : flatness ] .
+ while it is an attractive feature that inflation can solve the flatness and horizon problems , these are ` post - dictions ' and offer little prospect of a further test of the paradigm , far less a way to distinguish between different inflationary models .
consequently , interest has rightly been refocussed on the fact that the condition for inflation is precisely that which permits large - scale perturbations to be generated causally , as it means that scales are ` expanded ' more rapidly than the hubble scale .
this is shown in figure [ f : scales ] .
further , inflation offers a definite mechanism that the perturbations originate as quantum fluctuations which in a given model can be readily calculated @xcite . +
in the simplest scenarios , which have been widely explored ( see ref.@xcite and references therein ) , inflation is driven by the potential energy of a single scalar field @xmath5 .
this potential energy @xmath6 , possibly accompanied by a specific mechanism to end inflation , is the model input , and physical observables will be determined by the value of @xmath7 and its derivatives at the time when those scales cross outside the hubble radius during inflation . in the simplest models
this is the only epoch which matters ; before then the scale is well within the hubble radius and the effect of expansion is negligible , while afterwards the scale is too large to be influenced by causal processes .
however , we will see later that this is true only of this simplest class of models . as far as perturbations are concerned , the simplest models have the following characteristics . 1 . as there is only one type of matter , isocurvature perturbations can not be supported and scalar perturbations must be adiabatic . i ll call their spectrum @xmath8 ( see ref . @xcite for a formal definition ) , where @xmath9 is the comoving wavenumber .
gravitational waves are always present at some level , and i ll write their spectrum as @xmath10 . in many models , especially those of the currently - popular hybrid type ( see ref .
@xcite for a thorough review of inflationary model building ) they in fact turn out to be negligible . 3 .
the two spectra can normally be approximated as power - laws @xmath11 where @xmath12 and @xmath13 are the spectral indices and are readily calculated in a given inflation model .
however , the power - law expansions are only valid if the inflationary potential is flat enough ( see later ) , and if the available observations are of sufficiently poor quality .
high - quality observations make higher accuracy demands on the description of the initial power spectrum .
the relative importance of density perturbations and gravitational waves @xmath14 ( measured in terms of their contribution to large - angle microwave background anisotropies ) is also readily predicted , and equals approximately @xmath15 . in general @xmath12 and @xmath14
are independent .
from reading some parts of the literature , one might form the impression that the cosmological parameters , such as the hubble parameter and the density of the universe , are directly inscribed on the last - scattering surface and hence available to be measured in a model - independent way .
nothing could be further from the truth ; the cosmological parameters govern the _ dynamics _ of perturbations , and a single time - slice in isolation can say nothing whatsoever about their values . to obtain them
, one needs to understand the perturbations at some other epoch , and since this is not accessible to direct observation this initial form must be fixed by assumption .
since there is no unique model supplying these initial conditions , they must be parametrized in some way and those parameters added into the melting pot .
the various parameters can be divided into the cosmological parameters ( whose role is to describe the present state of the universe ) and the inflationary parameters , describing the initial conditions .
a typical set might be cosmological parameters _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ @xmath16 : : : the hubble parameter .
@xmath17 : : : the density of matter in the universe .
@xmath18 : : : a cosmological constant , if present .
@xmath19 : : : the amount of baryonic matter . @xmath20 : : : the amount of hot dark matter , if present .
+ ( it is assumed that there is always some cold dark matter . )
@xmath21 : : : the optical depth to the last scattering surface caused by reionization .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ inflationary parameters _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ @xmath22 : : : the overall amplitude of density perturbations .
@xmath12 : : : the spectral index of the density perturbations .
@xmath14 : : : the effect of gravitational waves on cobe .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ amongst those cosmological parameters , some might be thrown out by assumption ; for example we may restrict ourselves to a flat universe without a hot dark matter component . typically authors have considered fairly wide cosmological parameter sets .
however on the inflationary side most efforts have been very minimal , often with just @xmath23 and @xmath12 considered . while it might be highly desirable to imagine that the initial conditions can be parametrized by so few parameters , and while indeed that may well in the end turn out to be true , it needs to be acknowledged that less kind circumstances might also prevail , as described in the next section .
having decided your parameters , you can do two things .
the first , which has received quite a bit of attention , is to make estimates of how accurately one can expect to measure said parameters , under the assumption that they provide a sufficient set and for some chosen experimental configuration @xcite .
an example will be given next section .
the second thing is to actually go ahead and try and estimate the parameters from some data .
this has begun to be popular @xcite , though so far the only attempt to cover enough parameters that one can argue they are sufficient is a recent attempt by tegmark @xcite to constrain a nine - parameter space based on the simplest inflation models .
current data is not really up to the task of seriously limiting such a wide parameter volume , but hopefully that will change in the medium future .
the simplest inflationary models are the ones most commonly described .
but there are several other possibilities already in the literature , which would impact on the simple picture outlined so far .
* _ there may be only one field , but with a potential energy @xmath6 which is less flat than we might like . _
+ then the power - law approximation might break down .
the simplest manifestation of this is for the spectral index @xmath12 to become scale - dependent , but perhaps amenable to a perturbative approach as outlined later in this section .
more drastic is the case of ` designer ' power spectra , where the potential has sharp features leading to similarly sharp features in the spectra .
this would need to be dealt with on a model - by - model basis .
* _ there may be more than one field . _
+ then the adiabatic perturbations might be accompanied by isocurvature perturbations .
superhorizon adiabatic perturbations are then no longer constant @xcite .
we can not necessarily ignore details of physics while modes are superhorizon , e.g. the ( p)reheating epoch which brings inflation to an end .
* _ quantum tunnelling effects might be important . _
+ then the universe might be open instead of flat @xcite .
i will briefly look at the possible impact of the first two of these .
the better the quality of observations available , the more likely it is that the power - law approximation is inadequate .
it is in fact the beginning of a taylor expansion of the power spectrum of the form @xcite @xmath24 where only the first two terms are kept .
the expansion scale @xmath25 is arbitrary , but a specific choice is in fact favoured in that it can be chosen so that the estimated errors on @xmath12 and on @xmath26 are uncorrelated .
a perturbative approach is therefore suggested where we add further terms to this series ( i.e. adding the coefficients to the parameter estimation process ) until a satisfactory fit is achieved . in a given inflation model
these coefficients are as easily estimated as @xmath12 itself , and indeed some inflation models , especially the so - called ` running mass ' models @xcite , can indeed give an effect large enough to be detectable @xcite .
when more parameters are added , in principle the error on _ all _ parameters is increased .
we computed the effect in ref .
@xcite , assuming a version of the cold dark matter model and observations by a version of the planck satellite including polarized detectors , and the table shows how the uncertainties change as extra derivatives of @xmath12 are added .
in fact the picture is quite encouraging .
the uncertainties on the cosmological parameters hardly increase at all , and even the uncertainty on the spectral index itself suffers little as extra derivatives are incorporated .
it is quite possible therefore that in these models one might get extra information on the inflationary model , through determination of the higher derivatives , while suffering little cost elsewhere .
[ cols= " < , < , < , < , < " , ] if there are multiple scalar fields , then we may find isocurvature perturbations as well as adiabatic ones .
this represents a significant complication , because adiabatic and isocurvature modes can source each other even on superhorizon scales if there are interactions between the field or fluid components .
predicting the perturbations in such models presents no problem of principle , but extra information might need to be provided ( details of the reheating period ending inflation , for example ) , whereas in the single field model causality provides a direct bridge from the inflationary era until the scales re - enter the horizon . in general very complicated post - inflation processing between isocurvature and adiabatic modes might occur @xcite .
one type of model aims to create only isocurvature perturbations @xcite .
the basic idea is that during inflation , the field which later becomes the cdm already exists and experiences perturbations via the usual quantum fluctuation mechanism .
when it later becomes dynamically significant , we have an isocurvature perturbation between cdm and the rest .
alternatively , there may be a mixture of the two .
low - level isocurvature contamination can impact on parameter estimation , since it will contribute either as a detectable component requiring extra parameters for its description , or as an extra undetectable noise contribution .
inflationary models of these types have to be compared to observational data on a model - by - model basis .
this potentially leads to a wide range of predictions .
there is however one saving grace , which is that at least all these models share one characteristic which seems unavoidable , being the existence of a peak structure in the microwave anisotropy spectrum @xcite .
this appears to be a generic prediction , as it follows from a simple argument . because perturbations are established well outside the horizon ,
they are observed entirely in the growing mode .
this necessarily leads to a phase coherence and ultimately to an acoustic peak structure .
the conclusion therefore is that if the observed spectrum fails to feature a series of peaks when measured at high resolution , then inflation can not be the sole source of perturbations originating structure .
note that phrase carefully however to avoid confusion .
if a peak structure is not observed , that does nt completely rule out inflation , because it may partly contribute to the anisotropies , an example being the mixed inflation plus defects scenario where the peak structure vanishes if the defects are sufficiently dominant @xcite .
further , if the peak structure _ is _ observed , that does not ` prove ' inflation .
it would support inflation as it would have passed a further observational hurdle , but other mechanisms may be able to produce similar structures .
the main point i have aimed to stress in this article is that the inflationary paradigm is a broad one , realizable in a number of ways .
the inflationary input is crucial to current views on parameter estimation , not just of the inflationary parameters but the cosmological ons too .
if the simplest inflationary models prove correct , all will be well but we will learn only a limited amount about the inflationary mechanism .
more desirable would be something a little more complicated , allowing us to extract further information about inflation at only a very modest cost to the cosmological parameters .
but we had better hope that the model is not so complicated that it requires so many additional parameters as to have a serious effect on our ability to fit the observational data , or even worse a model which defies parametrization at all .
fortunately , all the current indications are favourable .
i thank ed copeland and ian grivell for collaboration on some of the work described in this article .
a. h. guth , phys .
d * 23 * , 347 91981 ) ; a. d. linde , _ particle physics and cosmology _ ( harwood , chur , switzerland , 1990 ) .
a. r. liddle and d. h. lyth , phys . rep . *
231 * , 1 ( 1993 ) .
a. vilenkin and e. p. s. shellard , _ cosmic strings and other topological defects _
( cambridge university press , 1994 ) .
pen , u. seljak and n. turok , phys .
* 79 * , 1611 91997 ) ; b. allen , r. r. caldwell , s. dodelson , l. knox , e. p. s. shellard and a. stebbins , phys .
lett . * 79 * , 2624 ( 1997 ) .
l. knox , phys .
d * 52 * , 4307 ( 1995 ) ; g. jungman , m. kamionkowski , a. kosowsky and d. n. spergel , phys .
d * 54 * , 1332 ( 1996 ) ; j. r. bond , g. efstathiou and m. tegmark , mon . not
. soc . * 291 * , l33 ( 1997 ) ; m. zaldarriaga , d. spergel and u. seljak , astrophys .
j. * 488 * , 1 ( 1997 ) .
j. e. lidsey , a. r. liddle , e. w. kolb , e. j. copeland , t. barreiro and m. abney , rev .
mod . phys . * 69 * , 373 ( 1997 ) .
d. h. lyth and a. riotto , to appear , phys .
, hep - ph/9807278 . k. ganga , b. ratra and n. sugiyama n. , astrophys . j. * 461 * , l61 ( 1996 ) ; m. white and j. silk , phys . rev . lett . * 77 * , 4704 ( 1996 ) ; j. r. bond and a. h. jaffe , in ` microwave background anisotropies ' , ed f. r. bouchet et al . , editions frontires , astro - ph/9610091 ; c. h. lineweaver , d. barbosa , a. blanchard and j. g. bartlett , astron . astrophys . *
322 * , 365 ( 1997 ) ; s. hancock , g. rocha , a. n. lasenby and c. m. gutierrez , preprint astro - ph/9708254 ( 1998 ) .
m. tegmark , preprint astro - ph/9809201 ( 1998 ) .
a. a. starobinsky and j. yokoyama , in _ fourth workshop on general relativity and gravitation _ , ed k. nakao et al , astro - ph/9502002 91995 ) ; m. sasaki and e. d. stewart , prog .
. phys . * 95 * , 71 ( 1996 ) ; j. garca - bellido and d. wands , phys.rev .
d * 53 * , 5437 ( 1996 ) .
j. r. gott , nature * 295 * , 304 ( 1982 ) ; m. sasaki , t. tanaka , k. yamamoto and j. yokoyama , phys .
b * 317 * , 510 ( 1993 ) ; m. bucher , a. s. goldhaber and n. turok , phys .
d * 52 * , 3314 ( 1995 ) ; s. w. hawking and n. turok , phys . lett .
b * 425 * , 25 ( 1998 ) .
e. j. copeland , i. j. grivell and a. r. liddle , mon . not .
. soc . * 298 * , 1233 ( 1998 ) .
e. d. stewart , phys .
b * 391 * , 34 ( 1997 ) ; e. d. stewart , phys .
d * 56 * , 2019 ( 1997 ) ; l. covi and d. h. lyth , preprint hep - ph/9809562 ( 1998 ) . w. hu , preprint astro - ph/9809142 ( 1998 ) .
a. d. linde , phys .
b * 158 * , 375 ( 1985 ) ; l. kofman , phys .
b * 173 * , 400 ( 1986 ) ; a. d. linde and v. mukhanov , phys .
d * 56 * , 535 ( 1997 ) ; p.
j. e. peebles , preprint astro - ph/9805194 ( 1998 ) .
w. hu and m. white , phys .
lett . * 77 * 1687 ( 1996 ) . c. contaldi , m. hindmarsh and j. magueijo , preprint astro - ph/9809053 ( 1998 ) ; r. a. battye and j. weller , preprint astro - ph/9810203 ( 1998 ) . | various issues concerning the impact of inflationary models on parameter estimation from the cosmic microwave background are reviewed , with particular focus on the range of possible outcomes of inflationary models and on the amount which might be learnt about inflation from the microwave background . |
the use of migdal eliashberg ( me ) theory for the study and analysis of electron - phonon systems is widespread . for conventional superconductors such as lead , where the electron - phonon coupling is higher than can be treated with bcs theory ,
me theory has been extremely successful for understanding the superconducting properties .
recently , a lot of researchers have been interested in an apparent kink in the electronic dispersion of cuprate supercondutors as determined from angle resolved photo - emission spectroscopy ( arpes ) , and their analysis / interpretation typically uses the related engelsberg schrieffer ( es ) result ( i.e. just the lowest order fock diagram ) @xcite .
the es result divides excitations into long - lived ( coherent ) low - energy excitations , and rapidly decaying high - energy excitations , with a kink at the phonon energy .
es based analysis of arpes results suggests a large electron - phonon coupling in the cuprates , with estimates of the dimensionless coupling constant lying between @xmath0 and @xmath1 ( depending on doping ) and very large phonon frequencies of @xmath2 mev @xcite . there has also been a development of a maximum entropy technique for analysing cuprate superconductors which makes use of eliashberg theory @xcite . the analysis in ref . determines a dimensionless electron - phonon coupling of @xmath3 , but this is well above the @xmath4 value where one would normally expect perturbation theory to fail .
the region of intermediate electron - phonon coupling is also relevant to a number of other materials .
for example , electron paramagnetic resonance measurements on the manganites support a strong `` electron - phonon '' coupling leading to jahn
teller polarons @xcite . in light of the current experimental situation , and the importance of the conclusions of refs .
, a full re - analysis of the perturbation theory is of high significance .
the neglect of vertex corrections suggested by migdal leads to a theory where an infinite set of feynman diagrams may be summed . according to an analysis carried out by migdal
, this theory should be valid in the physical regime of electron - phonon problems , where the phonon energy is significantly smaller than the intersite hopping ( `` migdal s theorem '' ) @xcite specifically that the vertex corrections are small when @xmath5 ( @xmath6 is the phonon frequency and @xmath7 the fermi energy ) .
on the basis of migdal s analysis , it is often believed that migdal eliashberg theory is applicable above @xmath4 in the adiabatic limit because of the apparently small size of the vertex corrections , even though in general perturbative approaches break down ( i.e. the functional form of the self - energy becomes incorrect ) when the coupling constant becomes large . in this article , we revisit migdal s analysis by examining the large-@xmath8 limit ( local approximation ) .
the large-@xmath8 limit has been an effective workshop for determining the validity of approximate schemes .
the quantum monte - carlo ( qmc ) solution in the large-@xmath8 limit @xcite has been compared with a number of different diagrammatic approaches @xcite demonstrating that self - consistent second - order perturbation theory worked better than migdal - eliashberg theory at intermediate phonon frequencies and strong coupling ( @xmath9 , @xmath10 ) .
the qmc solution is difficult for very low phonon frequency and low temperature .
qmc self - energies and green functions are generated along the matsubara axis .
as such , it is not easy to make quantitative conclusions about the point of breakdown of me theory as small differences at individual matsubara frequencies may result in relatively large differences in the spectral functions . using an alternative approach , benedetti _ et al .
_ @xcite found the formation of more than one extremum in the path - integral formulation at very low phonon frequencies and intermediate coupling , leading to a breakdown of me theory .
work on the finite - dimensional holstein model also shows deficiencies in me theory . on the basis of a comparison with exact diagonalization results , alexandrov _ et al .
_ @xcite have shown that me theory may break down at intermediate couplings even in the adiabatic limit .
in this article , we investigate the holstein model of electron - phonon interactions @xcite , which employs a number of approximations .
the phonon dispersion is flat corresponding to independently moving ions and phonon anharmonicity is neglected , resulting in a hamiltonian , @xmath11 where @xmath12 @xmath13 create ( annihilate ) electrons at site @xmath14 with spin @xmath15 , @xmath16 is the local ion displacement , @xmath17 the ion momentum , @xmath18 the ion mass ,
@xmath19 the electron hopping parameter and @xmath20 the chemical potential .
a mean - field coulomb pseudopotential may also be included , but in the normal state , it is just absorbed into the chemical potential . in the following , @xmath21 and all energies ( including temperature )
are measured in units of @xmath22 .
an expression for the effective interaction between electrons can be obtained by performing a linear transformation in the phonon variable to remove the electron - phonon term @xcite .
the resulting interaction is retarded , with fourier components : @xmath23 where @xmath24 are the matsubara frequencies for bosons . taking the limit @xmath25 , @xmath26
, while keeping the ratio @xmath27 finite , leads to an attractive hubbard model @xcite with instantaneous attraction of magnitude @xmath28 , where @xmath29 is the half band - width ( note that @xmath30 is related to the bipolaron binding energy for the strong coupling two - electron problem ) . in the opposite limit ( @xmath31 , @xmath32 , keeping @xmath33 finite ) the phonon kinetic - energy vanishes , leaving only a static variable @xmath16 representing the phonon subsystem in the hamiltonian , @xmath34 .
thus the problem looks like that of a single electron in a disordered potential .
the large-@xmath8 limit of this model was exactly solved by millis _
_ @xcite and extended to deal with long range order by ciuchi _ et al _ @xcite .
thus , the phonon frequency may be thought of as a parameter for tuning the level of electronic correlation @xcite . in this paper
, we investigate the validity of the migdal
eliashberg approach within the dynamical mean - field theory formalism .
as we carry out the self - consistency on the real - axis self - energies , we can investigate the behaviour of me theory over a full range of temperatures and phonon frequencies .
we begin by calculating self - energies , spectral functions , the quasi - particle weight and renormalized phonon frequency . to determine the validity of the theory
, we calculate an expression for the lowest - order correction to the vertex function , and evaluate its magnitude .
we find that the theory can also break down in the low frequency regime , contrary to the standard interpretation of migdal s analysis .
finally , we calculate resistivity curves in the regime where the lowest - order corrections are small .
in this article , we use the local approximation or dynamical mean - field theory ( dmft ) as a way of analysing the migdal - eliashberg theory .
the self - energy of correlated - electron systems is momentum independent in limit of large dimensions @xcite , and approximately momentum independent in 3d . in large-@xmath8 ,
lattice models map onto anderson impurity models with a self - consistent hybridisation @xcite .
the self - consistent dmft equations may be obtained by rewriting the action for the model to be considered in terms of an effective single - site action @xcite @xmath35 here @xmath36 plays the rle of the host green function in the equivalent impurity model .
if one assumes that correlations carried via the bath between electrons entering and exiting a single site can be neglected ( true in the case of large coordination number ) , the degrees of freedom associated with all but one site can be integrated out .
this leads to the following self - consistent equation @xcite , @xmath37},\ ] ] where @xmath38 is the site - local ( impurity ) green function , and is itself a functional of @xmath39 .
@xmath40 are the fermionic matsubara frequencies .
@xmath41 $ ] is the reciprocal function of the hilbert transform , defined as @xmath42=\xi$ ] .
the hilbert transform of the non - interacting density of states , @xmath43 , is defined as @xmath44 .
when electrons move on a tight - binding hypercubic lattice , the bare dos takes the form of a gaussian @xcite , @xmath45 and there is no simple expression for the reciprocal function . introducing the modified dyson equation , @xmath46 ( where @xmath47 is the electron self - energy ) , equation ( [ eqn : selfconsistent ] )
may be rewritten as , @xmath48=i\omega _
{ n}+\mu -\sigma ( i\omega _ { n}),\ ] ] and inverted to give an expression for the green function in terms of the self - energy , @xmath49 this also allows the approximation to be interpreted as a course graining of the momentum space @xcite . to complete the scheme , a form for the electronic self - energy
must be calculated in terms of the non - interacting weiss field .
this is normally approximate .
then a self - consistent procedure is followed : compute the green function from equation ( [ eqn : greensfn ] ) , the weiss field from equation ( [ eqn : weissfield ] ) and then re - calculate the self - energy until convergence is reached .
the application of me theory within dmft corresponds to computing the self - energy and phonon propagator from the diagrammatic equations in figure [ fig : feynmandiag](a and b ) .
the self - consistent solution of such a self - energy corresponds to the summation of all feynman diagrams which contain no vertex corrections . in the low ( non - zero )
phonon - frequency limit , migdal s analysis indicates a condition , @xmath50 , for the neglect of corrections to the vertex function @xcite .
the solution of the dyson equation seen in figure [ fig : feynmandiag](a ) results in the full phonon propagator , @xmath51},\ ] ] which is then used for the calculation of the electron self - energy . the phonon polarisation bubble ( dimensionless self - energy ) @xmath52 is given from perturbation theory , @xmath53 and may be analytically continued by introducing the spectral representation , @xmath54 where @xmath55/\pi$ ] and performing the sum over matsubara frequencies to give , @xmath56=-2u\int _
{ 0}^{\omega } dx\rho ( -x)\rho ( \omega -x),\ ] ] at absolute zero , and @xmath57=-u\sinh ( \frac{\omega } { 2t})\int _ { -\infty } ^{\infty } dx\ , \frac{\rho ( -x)}{\cosh ( \frac{x}{2t})}\frac{\rho ( \omega -x)}{\cosh ( \frac{\omega -x}{2t})},\ ] ] at finite @xmath58 . the full spectral function is used so that all diagrams with no vertex corrections are included , consistent with the proper interpretation of migdal s analysis .
the lowest - order skeleton diagram shown in figure [ fig : feynmandiag](b ) , @xmath59 may be analytically continued in the same way as the polarisation bubble to give , @xmath60=u\int _ { 0}^{\omega } dx\rho ( x)\sigma ( \omega -x),\ ] ] with @xmath61/\pi$ ] .
a more complicated expression applies at finite temperature , @xmath62 = u\cosh ( \frac{\omega } { 2t})\int _ { -\infty } ^{\infty } \frac{dx\ , \rho ( \omega -x)\sigma ( x)}{2\cosh ( \frac{\omega -x}{2t})\sinh ( \frac{x}{2t})}.\ ] ] at each stage , the kramers kronig relation is used to compute the real parts of the electron and phonon self - energies , for instance @xmath63=\mathrm{p}\int dx \mathrm{im}[\pi _ { 0}(x)]/\pi(\omega -x)$ ] , where p denotes the principal integral .
after the self - consistent procedure has converged , physical properties are calculated , including the quasi - particle weight ( inverse effective mass ) , @xmath64 the effective phonon frequency @xmath65 from , @xmath66)=0,\ ] ] and the optical conductivity @xcite , @xmath67=\frac{\pi } { \omega } \int _ { -\infty } ^{\infty } d\varepsilon { \mathcal{d}}(\varepsilon ) \int _ { -\infty } ^{\infty } d\nu \ , \rho ( \varepsilon , \nu ) \rho ( \varepsilon , \nu + \omega ) [ f(\nu ) -f(\nu + \omega ) ] , \ ] ] where @xmath68 is the fermi - dirac distribution and @xmath69/\pi$ ] ( taking the limit , @xmath70 , the dc conductivity is recovered . )
although dmft is approximate , the formalism can be expected to work in the non - interacting limit , where course graining will give the exact non - interacting dos for the tight - binding model ( regardless of dimension ) . in the opposite limit ( @xmath71 ) , the neglect of loops through the host
is justified , and the formalism should be exact . such propagation of correlation is also small when the coordination number is high ( e.g. fcc lattices ) .
we have solved the dmft equations using the self - energies in equations ( [ eqn : selfenergyeqn ] ) and ( [ eqn : selfenergyt ] ) to find electron spectral functions .
we show the evolution of these functions with coupling for @xmath72 in figure [ fig : spectralfunctions](a ) and @xmath73 in figure [ fig : spectralfunctions](b ) .
the spectral functions show two features .
there is a low energy peak with a width defined by the phonon frequency , and a high energy shoulder .
this behaviour can be related to the two limits of the holstein model . at low energy scales ( @xmath74 ) , electrons interact via virtual processes and the behaviour is essentially hubbard - like ( correlated ) . at frequencies greater than @xmath75 , the spectral weight is reduced , phonons may be created , and static behaviour emerges ( in the static limit , phonons may always be created as no energy is required to deform the lattice ) .
the central peak narrows with increased coupling , until a critical value is reached . here
, the theory has a brinkman rice - like transition with a diverging effective mass @xcite . in figure
[ fig : selfenergy ] , we plot the imaginary part of the self - energy . for low coupling strengths ( small @xmath30 and energy scales ( @xmath76 ) , @xmath77 $ ]
is small because electrons can not create phonons , consistent with the englesberg
schrieffer analysis ) .
however , as the coupling increases , electrons can be scattered by bipolaron resonances and @xmath78 $ ] rises sharply on either side of the fermi - energy .
the gap in the weight of the self - energy is frequently assumed in the analysis of arpes data , but as we can see here , when the electron - phonon coupling approaches the band width ( i.e. @xmath4 ) , the self - energy has a more complicated form , without the simple picture of coherent and incoherent quasi - particles .
indeed , it has been demonstrated that in low dimensions , vertex corrections are required to achieve a sharp discrimination between coherent and incoherent dressed electrons @xcite .
the onset of this `` transition '' can be studied by examining the inverse quasi - particle mass ( quasi - particle weight ) , shown in figure [ fig : qpweight ] for two values of phonon frequency .
as the coupling increases , @xmath79 becomes smaller and eventually vanishes at a critical value of coupling ( the effective mass diverges ) .
we also examine the phonon spectral function in figure [ fig : phonspec ] for @xmath80 and various couplings . as the coupling
is increased , the phonon modes soften ( figure [ fig : phonfreq ] ) .
the effective frequency does not tend to zero as quickly as the quasi - particle weight .
we will revisit this point later in this article .
it is clearly the case that me theory does not correctly describe the strong coupling limit .
it is known from the exact solution of the static limit @xcite and from approximate `` iterated - perturbation theory '' @xcite and qmc @xcite solutions of the hubbard model , that sub - bands should form at strong coupling .
as our calculations demonstrate , this sub - band formation is not properly reproduced within me theory and this makes it likely that effects of vertex corrections become significantly more important at strong coupling .
since this is at variance with the traditional interpretation of migdal s analysis , we are motivated to re - examine the vertex function in the next section .
the neglect of the lowest - order vertex correction shown in figure [ fig : feynmandiag](c ) and all higher - order corrections is central to migdal eliashberg theory . in this section ,
we compute the lowest - order correction and use this to define the region of validity for me theory . in this sense , we are revisiting migdal s analysis to understand why there is a contradiction between low frequency results computed with me theory and advanced numerical methods .
the ratio of first to zeroth order vertices at finite temperature may be written as @xmath81 as temperature tends to zero , this sum becomes an integral which may easily be evaluated ( @xmath82 and @xmath83 are defined as in figure [ fig : feynmandiag ] ) , @xmath84 in order for vertex corrections to be unimportant , the ratio @xmath85 must be small . as an example
, we show the lowest order correction to the vertex function for ( a ) @xmath86 and @xmath87 ( b ) @xmath88 and @xmath89 in figure [ fig : vertexcorrection ] . for the larger phonon frequencies , the ratio is clearly greater than 10% and the approximation is expected to be significantly changed by the inclusion of vertex corrections . for the weak coupling @xmath30 shown here , the corrections are much less pronounced for small phonon frequencies . in figure
[ fig : diverge ] we show the magnitude of the central peak of the vertex correction ( always the largest part of the function ) for a range of couplings at @xmath90 . between @xmath91 and @xmath92 the vertex correction diverges , and me theory clearly breaks down . as it is necessary to pass through the divergence to reach @xmath93 ,
the theory is not applicable for strong couplings .
figure [ fig : breakdown ] shows the lowest value of @xmath94 for given @xmath95 at which the ratio @xmath96 exceeds 0.1 ( the primed quantities are chosen so that @xmath97 represents an unprimed value @xmath98 , and a primed value of @xmath99 represents an unprimed value @xmath100 ) . below this value
me theory can be expected to give accurate results , while the theory begins to break down above this line . for @xmath101 ( @xmath102 )
the standard migdal criterion works well and correctly predicts the region in which me theory is applicable .
however for frequencies @xmath103 ( @xmath104 ) , the migdal criterion misses the divergence in the vertex correction associated with the divergence in the effective mass
. this breakdown happens at @xmath105 or @xmath4 .
the line of breakdown due to this vertex divergence levels off as the phonon frequency approaches zero and does not scale as @xmath106 ( migdal s criterion ) .
these results imply that there is a breakdown in me theory in the adiabatic limit .
we note that at high phonon frequencies ( @xmath107 ) , our line does not tend to zero .
this is because the contribution @xmath108 does not tend to infinity . instead , the function gets wider as the phonon frequency increases ( see figure [ fig : vertexcorrection ] , panel b ) , and some combination of the magnitude and width of the vertex is probably a better measure of the importance of vertex corrections . at large @xmath6 ,
the breakdown is quite different to that at low phonon frequencies as the second order diagrams ( one of which contains a vertex correction ) develop the same functional form and magnitude .
however the breakdown at @xmath109 ( @xmath110 ) is not too bad an estimate . at this value of @xmath30 , there is only a small mass renormalisation due to vertex corrections
. our results are consistent with that of benedetti and zeyher @xcite , which was obtained with a different technique using a semi - circular density of states ( @xmath111 ) , and predicted breakdown for @xmath112 in the extreme adiabatic limit .
we have included their result in the figure , choosing the bandwidth parameter by matching the bare dos at the fermi energy for the two cases ( the conversion factor is @xmath113 ) which is reasonable when considering low energy excitations at half - filling , and corresponds to a critical @xmath114 for the gaussian dos used here ( see the diamond in figure [ fig : breakdown ] ) .
@xcite also predicted a breakdown in me theory at @xmath105 from their solution of the static limit of the infinite - dimensional holstein model , consistent with the breakdown that we have found .
the solution of the static limit predicts the formation of sub - bands , which is something not reproduced in the me solution .
the lack of subbands in me theory demonstrates that higher order diagrams are essential for the description of the sub - band formation , and therefore for the description of the strong coupling limit .
we suggest that the breakdown of me theory should be understood in the following heuristic manner . as the coupling increases , the host spectral function ( @xmath115 ) narrows and
spectral weight is shifted towards lower energy scales .
there is then an effective bandwidth , @xmath116 , for quasi - particles close to the fermi surface .
the condition for the applicability of me theory then becomes @xmath117 .
as seen in figure [ fig : phonspec ] , there is also a reduction in the effective phonon frequency with increasing interaction strength .
although the renormalisation of the phonon frequency helps to drive against the transition , we note that the band narrowing effect is much stronger , as can be seen by comparing figures [ fig : qpweight ] and [ fig : phonfreq ] .
using the finite temperature form of the self - energy , it is possible to calculate resistivity curves .
these are shown in figure [ fig : resistivityudep ] at various electron - phonon coupling and phonon frequency . at high temperatures ,
the tendency is to linear behavior , with a gradient depending on @xmath30 , but independent of @xmath118 .
the effect of phonon frequency is most dramatically seen at @xmath119 , where a point of inflection is seen , before the curve tends to low temperature @xmath120 behavior , consistent with a weakly renormalized electron gas .
we note that long range order has not been considered in this study , so no superconducting transition is seen .
the negative curvature of the resistivity curve at intermediate temperatures is of interest , since that curvature has recently been interpreted as an onset of the mott limit , but may have a different interpretation as an intermediate coupling phenomenon @xcite .
we have computed spectral functions , resistivity curves , self - energies , quasi - particle weight and effective phonon frequency for the holstein model at half - filling using migdal eliashberg theory , within the dmft framework , and revisited migdal s analysis of the strength of the vertex correction . by analysing the first order vertex correction ,
we have defined the region of applicability of me theory .
we find that me theory breaks down at intermediate coupling in the adiabatic limit , showing that migdal
eliashberg theory should only be trusted at weak coupling within this framework .
the coupling at which breakdown occurs corresponds to a divergence in the effective mass , indicating that both long range order and vertex corrections should be included to correctly describe the strong coupling regime .
the question remains : what was lacking from the analysis of migdal , which estimated the magnitude of the vertex corrections .
this can be understood from the magnitude of the kinetic energy . in the calculations ,
the kinetic energy can be seen to be decreasing due to the localisation at the non - analyticities .
thus in migdal s estimate , the renormalised kinetic energy , and not the bare kinetic energy should have featured .
the renormalised phonon frequency should also have been used , but it tends to zero more slowly than the kinetic energy . at the non - analyticity , where the kinetic energy tends to zero , it can be seen that a vertex divergence is expected from the modified version of migdal s estimate .
our results are timely because of a recent upsurge of interest in the role of electron - phonon coupling in the cuprate superconductors .
there are many experimental results which have been analysed and interpreted using es / me theory , in spite of the large coupling constant that has been predicted from that analysis .
many groups are using migdal s analysis to justify their claims , without a careful consideration of the self - consistent effect of coupling on the kinetic energy of the electrons ( polarons ) .
exact numerical analysis shows that large electron - phonon couplings significantly reduce the kinetic energy of electrons @xcite and thus affect the internal consistency of migdal s analysis .
therefore , the any very large @xmath121 determined from es style analysis of experimental results is not consistent with a theory that neglects vertex corrections ( i.e. a more detailed analysis with higher order effects included is necessary ) . without that analysis ,
the only conclusion that can be reached is that the coupling is large , but no reliable value for that coupling can be determined .
the main conclusion of a breakdown is not expected to change in 1d , 2d or 3d . in three dimensional systems , the role of spatial fluctuations should only be significant very close to the bi - polaron instability .
the role of spatial fluctuations on the spectral properties in 2d has been analysed at weak coupling , and is found to lead to quantitative differences in the coherent excitations @xcite .
spatial fluctuations have a major role in the superconducting state , and it has also been shown that the eliashberg approach to superconductivity is inadequate for optical phonon mediated @xmath8-wave superconductivity @xcite , and incomplete for @xmath122-wave superconductivity @xcite .
we therefore urge researchers to carefully analyse the internal consistency of their theories when dealing with strong coupling materials .
the authors would like to thank f.gebhard and f.essler for useful discussions .
jph thanks the epsrc for partial funding of this work .
is the frequency of the emitted phonon and @xmath123 is the frequency of the incoming electron .
neglect of this diagram is central to migdal s theorem .
thick and thin lines represent the full and bare green functions respectively .
wavy lines represent phonons and straight lines electrons.[fig : feynmandiag],width=491 ] ( panel a ) and @xmath124 ( panel b ) at @xmath125 . a bi - polaronic resonance forms at zero frequency .
no upper and lower sub - bands are formed but spectral weight is shifted away from the fermi - energy .
the central peak narrows with increasing @xmath30 , corresponding to a divergence in the effective mass .
the general form of the curves is similar at all frequencies.[fig : spectralfunctions],title="fig:",height=377 ] ( panel a ) and @xmath124 ( panel b ) at @xmath125 .
a bi - polaronic resonance forms at zero frequency .
no upper and lower sub - bands are formed but spectral weight is shifted away from the fermi - energy .
the central peak narrows with increasing @xmath30 , corresponding to a divergence in the effective mass .
the general form of the curves is similar at all frequencies.[fig : spectralfunctions],title="fig:",height=377 ] ( panel a ) and @xmath124 ( panel b ) at @xmath125 . as the coupling , @xmath126
( see discussion after eqn 2 ) increases , hubbard - like behaviour is seen at low energy scales , and fk - like behaviour at @xmath127 .
note how the gap in the self - energy at weak coupling ( which is central to the engelsberg
schrieffer approach ) fills up as coupling is increased .
this is due to the increasing importance of higher order diagrams at stronger couplings .
this shows that one should be careful about interpreting results from the cuprates ( @xmath128 ) using an es approach ( this type of analysis is typically used on arpes results ) , since the es form for the self - energy ( i.e. gapped at @xmath129 ) can only be seen for very small @xmath121 .
[ fig : selfenergy],title="fig:",height=377 ] ( panel a ) and @xmath124 ( panel b ) at @xmath125 . as the coupling , @xmath126 ( see discussion after eqn 2 ) increases , hubbard - like behaviour is seen at low energy scales , and fk - like behaviour at @xmath127 .
note how the gap in the self - energy at weak coupling ( which is central to the engelsberg schrieffer approach ) fills up as coupling is increased .
this is due to the increasing importance of higher order diagrams at stronger couplings .
this shows that one should be careful about interpreting results from the cuprates ( @xmath128 ) using an es approach ( this type of analysis is typically used on arpes results ) , since the es form for the self - energy ( i.e. gapped at @xmath129 ) can only be seen for very small @xmath121 .
[ fig : selfenergy],title="fig:",height=377 ] , computed using me theory as a function of @xmath30 ( results for phonon frequencies @xmath130 and @xmath131 are shown ) .
@xmath132 is strongly renormalised for intermediate @xmath30.[fig : qpweight],height=377 ] .
the effective phonon frequency ( location of maximum ) and particle lifetime ( inverse width of peak ) are reduced with increasing coupling .
when @xmath133 the curve is skewed in such a way that it is not lorenzian , and the excitation no longer has the properties of a single phonon.[fig : phonspec],height=377 ] ) = 0 $ ] .
the renormalisation of the effective phonon frequency is not as strong as that of the electronic inverse mass ( quasi - particle weight @xmath132 ) shown in figure [ fig : qpweight ] , which goes to zero at the transition .
the inset shows the quasiparticle lifetime given by @xmath134}\protect $ ] .
when the value of @xmath121 tends to 1 , the phonons can no longer be treated as single particle excitations .
[ fig : phonfreq],height=377 ] and @xmath87 ( b ) @xmath88 and @xmath89 .
the central maximum in ( b ) shows a ratio of over @xmath135 and one should expect significant corrections to quantities computed within me theory .
[ fig : vertexcorrection],width=529 ] and @xmath136 , chosen so that a value of 1 represents @xmath98 , etc . also shown is the result of benedetti and zeyher for very low phonon frequency ( filled diamond ) . above the line
, the ratio of the first order vertex correction to the bare vertex , @xmath137 , exceeds 10% and me theory is no longer strictly valid . at low frequencies ( adiabatic limit ) , breakdown occurs at smaller coupling strength than expected as a result of the divergence found in the vertex corrections .
this is in contrast to the interpretation of migdal s analysis which is often used when low frequency experimental data is analysed , and is the main result of this article .
[ fig : breakdown],height=377 ] | we investigate the applicability of migdal eliashberg ( me ) theory by revisiting migdal s analysis within the dynamical mean - field theory framework .
first , we compute spectral functions , the quasi - particle weight , the self energy , renormalised phonon frequency and resistivity curves of the half - filled holstein model .
we demonstrate how me theory has a phase - transition - like instability at intermediate coupling , and how the engelsberg schrieffer ( es ) picture is complicated by low - energy excitations from higher order diagrams ( demonstrating that es theory is a very weak coupling approach ) . through consideration of the lowest - order vertex correction ,
we analyse the applicability of me theory close to this transition .
we find a breakdown of the theory in the intermediate coupling adiabatic limit due to a divergence in the vertex function .
the region of applicability is mapped out , and it is found that me theory is only reliable in the weak coupling adiabatic limit , raising questions about the accuracy of recent analyses of cuprate superconductors which do not include vertex corrections . |
when the first galaxies emerged @xmath4 million years after the big bang , their starlight reionized and heated the intergalactic hydrogen that had existed since cosmological recombination .
much is currently unknown about this process , including what spatial structure it had , when it started and completed , and even which sources drove it .
the ly@xmath0 forest provides one of the only robust constraints on this process , showing that it was at least largely complete by @xmath5 , when the universe was one billion years old @xcite .
this letter argues that there exists another , potentially groundbreaking signature of reionization in the ly@xmath0 forest data . the amount of absorption in the ly@xmath0 forest can be quantified by the effective optical depth , @xmath6 , where @xmath7 $ ] is the transmitted fraction of a quasar s flux
, @xmath8 indicates an average over a segment of the forest of length @xmath9 , and @xmath10 is the optical depth in ly@xmath0 at location @xmath11 along a sightline .
the optical depth , @xmath10 , scales approximately as the h i number density , which after reionization scales as @xmath12 . here
, @xmath13 is temperature , @xmath14 is the gas density in units of the cosmic mean , and @xmath15 is the h i photoionization rate , which scales with the amplitude of the local ionizing radiation background .
observations of high-@xmath16 quasars show a steep increase in the dispersion of @xmath17 among coeval forest segments around @xmath18 @xcite . in the limit of a uniform ionizing background ,
the well - understood fluctuations in @xmath19 fall well short of producing the observed dispersion at @xmath20 , as shown recently by @xcite ( hereafter b2015 ) .
previous studies have attempted to explain this excess with spatial fluctuations in the ionizing background .
the properties of spatial fluctuations in the background depend on the number density of sources and the mean free path of photons , @xmath21 . while @xmath21 is well constrained at @xmath22 , being too large to yield significant background fluctuations for standard source models @xcite , b2015 showed that the excess @xmath17 dispersion at @xmath23 could be matched in a model where @xmath21 decreases by a factor of @xmath24 between @xmath25 and @xmath23 a time scale of just @xmath26 million years .
however , such rapid evolution in @xmath21 is inconsistent with extrapolations based on measurements at lower redshifts @xcite , and would imply that the emissivity of ionizing sources , in turn , increases by an unnatural factor of @xmath24 over the same cosmologically short time interval .
because of these issues , b2015 speculated that the excess dispersion was evidence for large spatial variations in the mean free path .
alternatively , fluctuations in the ionizing background could have been enhanced if the sources of ionizing photons were rarer than the observed population of galaxies
. however , current models require half of the background to arise from bright sources with an extremely low space density of @xmath27@xcite . this scenario is a possibility of current debate ( e.g. * ? ? ? * but see daloisio et al . in prep . ) . in this letter
, we explore a source of dispersion in @xmath17 that has so far been neglected and that , unlike ionizing background fluctuations , has straightforward implications for the reionization process itself .
in addition to @xmath14 and @xmath15 , the ly@xmath0 opacity depends on @xmath28 , mainly because the amount of neutral hydrogen after reionization is proportional to the recombination rate , which scales as @xmath29 .
previous attempts to model ly@xmath0 opacity fluctuations had not included the residual temperature fluctuations that must have been present if reionization were patchy and temporally extended . as ionization fronts propagated supersonically through the igm , the gas behind them was heated to tens of thousands of degrees kelvin by photoionizations of h i and he i. after reionization , the gas cooled mainly through adiabatic expansion and through inverse compton scattering with cosmic microwave background ( cmb ) photons @xcite . since different regions in the igm were reionized at different times , these heating and cooling processes imprinted an inhomogeneous distribution of intergalactic temperatures that persisted after reionization @xcite .
we will show that these residual temperature variations likely account for much of the observed dispersion in @xmath17 at @xmath30 , and may even account for all of it a scenario that would yield new information on the timing , duration , and patchiness of reionization .
the remainder of this letter is organized as follows . in [ sec : methods ] and [ sec : toymodel ] , we describe our simulations and methodology . in [ sec : results ] , we present our main results . in
[ sec : conclusion ] , we offer concluding remarks .
we use comoving units for distances and physical units for number densities .
to model the impact of relic temperature fluctuations from reionization on the distribution of @xmath17 , we ran a suite of 20 cosmological hydrodynamics simulations using a modified version of the code of @xcite .
the simulations were initialized at @xmath31 from a common cosmological initial density field .
we used a matter power spectrum generated by camb @xcite assuming a flat @xmath32cdm model with @xmath33 , @xmath34 , @xmath35 , @xmath36 , @xmath37 , and @xmath38 , consistent with recent measurements @xcite .
our production runs use a cubical box with side length @xmath39 , with @xmath40 dark matter particles and @xmath41 gas cells . in each simulation
, reionization was modeled in a simplistic manner by instantaneously ionizing and heating the gas to a temperature @xmath42 at a redshift of @xmath43 .
subsequently , ionization was maintained with a homogeneous background with spectral index @xmath44 , consistent with recent post - reionization background models @xcite . utilizing the periodic boundary conditions of our simulations
, we trace skewers of length @xmath45 ( following the convention of the b2015 @xmath17 measurements ) at random angles through all of the hydro simulation snapshots .
each skewer is divided into @xmath46 equally spaced velocity bins of size @xmath47 ( where @xmath48 is the cosmological scale factor ) , and ly@xmath0 optical depths are computed using the method of @xcite .
although reionization occurs instantaneously at @xmath43 within each simulation box , we piece together skewer segments from simulations with different @xmath43 to model the effect of an inhomogeneous reionization process , as we describe further in the next section .
the post - reionization temperatures in the simulations are relatively insensitive to the spectrum of the ionizing background , but they are sensitive to the amount of heating that is assumed to occur at the time a gas parcel is reionized .
previous calculations @xcite have bracketed the range of possible reionization temperatures to @xmath49 k. ( we note that previous large - scale reionization simulations do not accurately capture @xmath42 , as they do not resolve the @xmath50 physical kpc ionization fronts . ) thus , we have run two sets of ten simulations
one set with @xmath51 k and the other with @xmath52 k where each set contains instantaneous reionization redshifts of @xmath53 .
this redshift range spans the likely duration of reionization .
simulations with @xmath54 are driven to a common temperature by @xmath55 , so they are well approximated by the @xmath56 simulation .
figure [ fig : physics ] shows the post - reionization thermal and associated hydrodynamic relaxation of intergalactic gas ( and its effect on the ly@xmath0 forest opacity ) in our simulations .
the top - left panel shows the volume - weighted average gas temperature in the ten @xmath52 k simulations .
the bottom - left panel shows this same average , but limited to gas cells with densities of @xmath57 , the deepest voids that dominate transmission in the highly - saturated @xmath58 forest . even at @xmath59 ,
the @xmath60 gas temperatures differ by up to a factor of five between the simulations that were reionized at different times .
the right panels show the h i number densities ( top ) and the transmission ( bottom ) at @xmath61 for the same skewer through our @xmath56 ( blue / solid ) and @xmath62 ( red / dashed ) simulations .
the transmission is nearly zero in the @xmath56 case , owing to colder temperature and hence enhanced h i densities , whereas there is significant transmission in the @xmath62 case .
reionization is a process that is inhomogeneous and temporally extended , unlike in our individual hydro simulations .
modeling the thermal imprint of patchy reionization on the ly@xmath0 forest thus requires an additional ingredient : a model for the redshifts at which points along our skewers are reionized . in this section
, we present a simplified toy model to illustrate how we piece together skewer segments from our hydro simulations , and to provide insight into how the timing , duration and morphology of reionization affect the amplitude of ly@xmath0 opacity fluctuations . for illustrative purposes ,
let us assume that the ly@xmath0 forest is made up of segments of equal length , @xmath63 , where each segment is reionized at a single redshift .
let us further assume that the reionization redshift of each segment is drawn from a uniform probability distribution over the interval @xmath64 $ ] , resulting in a global reionization history in which the mean ionized fraction , @xmath65 , is linear in redshift ( since @xmath65 is the cumulative probability distribution of the reionization redshift ) . assuming that the reionization redshift field is in the hubble flow , each segment of length @xmath63 spans @xmath66 velocity bins of our hydro simulation skewers .
for the first segment , with reionization redshift @xmath67 , we take the initial @xmath68 spectrum velocity bins of a randomly drawn skewer that has @xmath43 closest to @xmath67 . for the next segment , with reionization redshift @xmath69
, we take the next @xmath68 velocity bins _ from the same skewer _ through the simulation that has @xmath43 closest to @xmath69 ( note that all simulations were started from the same initial density field ) .
we repeat this procedure until an entire @xmath70 sightline is filled . in
what follows , we compare these toy models against the b2015 measurements of the cumulative probability distribution function of @xmath17 , @xmath71 .
the measurements considered here span @xmath72 in bins of width @xmath73 .
we construct @xmath71 from 4000 randomly drawn sightlines at the central redshift of each bin . for each redshift , we rescale the nominal post - reionization photoionization rate of our simulations ( @xmath74 ) by a constant factor , such that our model @xmath71 is equal to the observed distribution at either @xmath75 or @xmath76 , depending on which value provides the better visual fit .
we have performed extensive numerical convergence tests using simulations of varying resolution and box size .
we found excellent convergence of @xmath71 for our production runs in both box size and resolution ( especially for our patchy reionization models ) . in our toy model ( colored curves )
compared to the b2015 measurements ( black histograms ) and to an instantaneous reionization model with @xmath77 ( black / dotted curves ) .
* top row : * varying the coherence length , @xmath78 , assuming a uniform @xmath43 probability between @xmath79 . *
bottom row : * varying the redshift interval over which the @xmath43 values are drawn , assuming @xmath80 . , width=340 ] figure [ fig : toymodel_cdfs ] shows the @xmath71 of these toy models for a range of @xmath81 , compared against the b2015 measurements in the two highest redshift bins ( black histograms ) .
the curves with shorter reionization durations , or with smaller @xmath63 , fall closer to the black / dotted curves , which assume that reionization occurs instantaneously ( at @xmath77 , although it does not depend on this choice ) .
figure [ fig : toymodel_cdfs ] affords three insights : ( 1 ) the larger the coherence length , @xmath63 , over which gas shares a similar reionization redshift , the larger the spread in @xmath17 ; ( 2 ) the width of the observed @xmath71 can be fully accounted for only if reionization ended at @xmath82 and was well underway by @xmath2 ; ( 3 ) the maximal width is achieved by a late - ending ( @xmath83 ) and extended reionization model in which large contiguous segments ( @xmath84 ) of the ly@xmath0 forest were reionized at @xmath85 . in the next section ,
we apply these insights to construct more realistic models of reionization that reproduce the observed width of @xmath71 .
we generate physically - motivated reionization redshift fields using simulations based on the excursion - set model of reionization ( esmr ) @xcite , which has been shown to reproduce the ionization structure found in full radiative transfer simulations @xcite . in particular , our esmr uses a realization of the linear cosmological density field and a top hat in fourier space filter to generate a realization of reionization in a cubical box with side @xmath86 , sampled with @xmath87 cells . for details of the algorithm ,
see @xcite .
we use the simplest formulation of the esmr , with two free parameters : ( 1 ) the minimum mass of halos that host galactic sources of ionizing photons , @xmath88 ; ( 2 ) the ionizing efficiency of the sources , @xmath89 .
we tune @xmath89 to obtain a reionization history that is approximately linear in redshift , similar to those in our toy model .
the esmr calculation yields a cube of reionization redshifts .
we construct mock absorption spectra by first tracing @xmath70 skewers through this cube .
we then piece together spectrum segments from our suite of hydro simulations , much like in our toy model , except here we match @xmath43 to the reionization redshifts along the esmr skewers .
this process does not account for how @xmath43 correlates with density , an effect that will underestimate ( making our models conservative ) the width of @xmath71 , which we address shortly .
the left panel of figure [ fig : zreionfield ] shows a 2d slice through our fiducial reionization redshift field in which reionization spans @xmath90 . in the right panels ,
the blue / solid curves show the fraction of transmitted flux along four different sightlines through this field , selected to span a range of @xmath17 .
the red / dashed curves show the corresponding reionization redshifts along the sightlines . on average , regions that reionize at later times
yield more transmission , while regions that reionize at @xmath85 result in dark gaps in the ly@xmath0 forest .
the variation among these mock spectra is similar to the well - known variation seen in observations of the @xmath3 ly@xmath0 forest @xcite . in our interpretation , this variation reflects differences in the reionization redshifts and hence temperatures fluctuations . ] between segments of the ly@xmath0 forest .
the leftmost four panels of figure [ fig : eps_cdfs ] show @xmath71 in three esmr models that take @xmath91 and @xmath92 , with reionization histories shown in the inset of the @xmath93 panel .
the cmb electron scattering optical depths in these models are @xmath94 and @xmath95 , within uncertainties of the latest _ planck _ measurement of @xmath96 @xcite .
these models are compared against the b2015 measurements ( black histograms ) and against the homogeneous reionization reference model with @xmath77 ( dotted curves ) .
the blue / long - dashed and green / short - dashed curves correspond to scenarios in which reionization spans @xmath97 and @xmath98 , respectively .
while these two models produce significantly more width in @xmath71 than the homogeneous reionization model , they fall short of producing the full range of @xmath17 required to match the observations .
however , the fiducial @xmath90 reionization model ( red / solid curves ) generally provides a good match to the measurements .
a significant success of the fiducial model is that the observed redshift evolution of the @xmath71 width is reproduced without any additional tuning of parameters .
( we have checked that this success also holds over @xmath99 , lower redshifts than those shown where the effect is reduced . ) indeed , there are not any parameters that can be tuned in our model to change the post - reionization evolution of the width .
the pink shaded regions indicate the 90% confidence levels of our fiducial model , estimated from bootstrap realizations , showing consistency with all the data aside from a single high - opacity point at @xmath100 and at @xmath23 . the discrepancy at the highest opacities may arise because our method of constructing mock absorption spectra does not capture correlations between temperature and density that should be present , since denser regions are more likely reionized earlier .
such correlations would act to increase the width of @xmath71 , as the denser regions around galaxies are ionized earlier in our models .
one might naively think these correlations are small , because correlations between the density on the much larger scales of the h ii bubbles and the smaller scales of voids in the forest should be weak , but calculations show that they may not be negligible @xcite the right panels of figure [ fig : eps_cdfs ] show the effect of varying @xmath42 and @xmath88 . for @xmath101 ,
the distribution of @xmath17 is somewhat narrower than the case with @xmath91 .
hotter temperatures are likely achieved towards the end of reionization , when h ii bubbles are larger and propagate at quicker speeds .
a model with @xmath102 near the end of reionization , and smaller temperatures earlier on , would likely produce more width than a model with @xmath102 at all times .
however , any conclusions about @xmath42 prior to modeling the density / reionization - redshift correlations are premature .
the blue / long - dashed and green / short - dashed curves in the right panels show the effect of varying @xmath88 . for the former atomic cooling halos ( achs ) curve , @xmath103 is set to the minimum mass required to achieve a halo virial temperature of @xmath104 .
for both curves , we tune @xmath89 to match the reionization history of our fiducial model ( red curve in the inset ) .
we find that the effects of varying @xmath103 are minor .
we have shown that residual temperature inhomogeneities from a patchy and extended reionization process likely account for much of the opacity fluctuations in the @xmath58 ly@xmath0 forest .
inhomogeneities in the ionizing background may also contribute at a significant level , though current models in this vein have required very small mean free paths or extremely rare sources .
we showed that residual temperature fluctuations alone could account for the entire spread of observed @xmath17 .
a significant success of this interpretation is that it is able to reproduce the rapid growth of @xmath17 fluctuations with redshift , despite having very little freedom in its post - reionization evolution . in this scenario ,
the observations favor a late but extended reionization process that is roughly half complete by @xmath2 and that ends at @xmath3 .
unlike ionizing background fluctuations , which do not necessarily signal the end of reionization , temperature fluctuations directly probe the timing , duration , and patchiness of this process .
if most of the opacity variations owe to temperature , it would mean that , on average , the darkest @xmath105 segments of the @xmath106 ly@xmath0 forest were reionized earliest , and the brightest segments last a potentially powerful probe of cosmological reionization . + the authors acknowledge support from nsf grant ast1312724 . h.t .
also acknowledges support from nasa grant atp - nnx14ab57 g .
computations were performed with nsf xsede allocation tg - ast140087 .
we thank paul la plante , jonathan pober , phoebe upton sanderbeck , george becker , adam lidz , and fred davies for helpful discussions . | recent observations of the ly@xmath0 forest show large - scale spatial variations in the intergalactic ly@xmath0 opacity that grow rapidly with redshift at @xmath1 , far in excess of expectations from empirically motivated models .
previous studies have attempted to explain this excess with spatial fluctuations in the ionizing background , but found that this required either extremely rare sources or problematically low values for the mean free path of ionizing photons . here
we report that much or potentially all of the observed excess likely arises from residual spatial variations in temperature that are an inevitable byproduct of a patchy and extended reionization process .
the amplitude of opacity fluctuations generated in this way depends on the timing and duration of reionization .
if the entire excess is due to temperature variations alone , the observed fluctuation amplitude favors a late - ending but extended reionization process that was roughly half complete by @xmath2 and that ended at @xmath3 . in this scenario ,
the highest opacities occur in regions that reionized earliest , since they have had the most time to cool , while the lowest opacities occur in the warmer regions that reionized most recently .
this correspondence potentially opens a new observational window into patchy reionization . |
astronomers commonly seek to estimate the _ space density _ of objects , and a sky survey such as the sloan digital sky survey ( sdss ) @xcite can yield a representative sample useful for this purpose , due to the assumed isotropy of the universe .
figure [ rawdata ] depicts redshift and absolute magnitude measurements for a sample of quasars given in @xcite .
these are a subset of the sdss quasar sample ( data release 3 ) , chosen to be statistically valid for purposes such as exploring the evolution with redshift of the luminosity function , i.e. the space density of quasars as a function of absolute magnitude .
this paper describes a new method for estimating these luminosity functions , and presents results from the analysis of this quasar sample .
for the purposes of the statistical inference problem , imagine the dots in figure [ rawdata ] as observations of bivariate data @xmath7 from some distribution with probability density @xmath2 , i.e. the probability that a randomly chosen quasar falls in a region @xmath8 is @xmath9 .
( equivalently , in a sample of size @xmath10 , one expects that @xmath11 will fall in the region @xmath8 . )
hence , the luminosity function at redshift @xmath12 is , up to a multiplicative constant , the cross - section of the bivariate density at @xmath12 , denoted @xmath13 .
the main challenge is estimation of this bivariate density given truncated data .
only objects with apparent magnitude within some range are observable . when this bound on apparent magnitude is transformed into a bound on absolute magnitude , @xmath14 , @xmath15 is assumed when making this transformation . ]
, the truncation bound takes an irregular shape , varying with redshift .
@xmath16-corrections further complicate this boundary , leading to the dashed region in figure [ rawdata ] .
also , the sample is not assumed to be complete within this region , and the probability of observing an object will vary with position on the sky , along with other factors . incorporating this _ selection function _ into the analysis is a secondary challenge .
nonparametric estimators are advantageous in cases where either there does not exist a commonly agreed upon parametric physical model , or there is a desire to validate a parametric model .
see @xcite for an overview of the potential of nonparametric methods in astronomy and cosmology .
a fully nonparametric approach is not possible here , since some assumptions must be placed on the form of the density in order to infer its shape over the unobservable region . under such conditions , one approach would be to fit a sequence of increasingly complex parametric models in an attempt to obtain a good fit to the data .
a less subjective alternative is a _ semiparametric _ approach which merges a nonparametric method with sufficient structure from a parametric form to obtain useful results .
this work describes a semiparametric approach to estimating the bivariate density , and hence the luminosity functions , under irregular truncation .
this is a long - standing challenge in astronomical data analysis , with a variety of proposed methods .
interesting qualitative and simulations - based comparisons between different approaches can be found in @xcite and @xcite .
a parametric model fit using maximum likelihood is a common choice , since it addresses the truncation bias in a natural manner ; see , for instance , @xcite , @xcite and the parametric models fit and referenced in 6 of @xcite .
these models have the drawback of imposing a tight constraint on the luminosity function in a case where there is not a consensus parametric form .
some proposed methods are nonparametric , but assume that redshift and absolute magnitude are independent , and hence assume that there is no evolution of the luminosity function with redshift .
these include the nonparametric maximum likelihood method described in @xcite and @xcite and adapted for double truncation in @xcite , along with the methods in @xcite , @xcite , the @xmath17 estimator of @xcite and @xcite .
the semiparametric method of @xcite also assumes independence . @xcite
apply a nonparametric technique which assumes independence after having transformed the bivariate data using a parametric form .
any method which assumes independence can be applied over small redshift ranges ( usually called bins ) . @xcite and @xcite describe other binning approaches .
binning forces the difficult choices of bin centers and widths , and independence is still assumed over the width of the bin .
this work was motivated by the goal of developing a statistically rigorous method which ( 1 ) does not assume a strict parametric form for the bivariate density ; ( 2 ) does not assume independence between redshift and absolute magnitude ; ( 3 ) does not require dividing the data into arbitrary bins ; and ( 4 ) naturally incorporates a varying selection function .
this was accomplished by decomposing the bivariate density @xmath2 into @xmath3 where @xmath18 will take an assumed parametric form ; it is intended to model the dependence between the two random variables .
for example , there may be a physical , parametric model for the evolution of the luminosity function which could be incorporated into @xmath18 .
alternatively , one could use @xmath19 as a first - order approximation to the dependence .
the functions @xmath4 and @xmath5 are estimated nonparametrically , with _ bandwidth _ parameters to control the amount of smoothness in the estimate . using the quasar sample of figure [ rawdata ] ,
the estimates obtained here are quite consistent , if not a bit smoother , than those found in @xcite .
this analysis confirms the finding of the flattening of the slope of the luminosity function at higher redshift .
the paper is organized as follows . [ data ] briefly describes the quasar sample used here .
[ ourmodel ] gives an overview of the idea of local maximum likelihood , a nonparametric extension of maximum likelihood , and describes in detail the semiparametric approach taken here .
[ bandselect ] describes how the integrated mean squared error can be approximated using cross - validation ; the bandwidths can then be chosen to minimize this quantity .
[ results ] presents some results from the analysis of the @xcite quasar sample , along with the results from some simulations .
more detailed derivations , along with theory for approximating the distribution of the estimator , can be found in @xcite .
the approach was implemented as a fortran subroutine with r wrapper ] ] .
the full @xcite sample , shown in figure [ rawdata ] , consists of 15,343 quasars . from these ,
any quasar is removed if it has @xmath20 , @xmath21 , @xmath22 , or @xmath23 .
in addition , for quasars of redshift less than 3.0 , only those with apparent magnitude between 15.0 and 19.1 , inclusive , ( after the application of @xmath16-corrections ) are kept ; for quasars of redshift greater than or equal to 3.0 , only those with apparent magnitude between 15.0 and 20.2 are retained .
these boundaries combine to create the irregular shape shown by the dashed line in figure [ rawdata ] .
this truncation removes two groups of quasars from the @xcite sample .
first , there are 62 quasars removed with @xmath22 .
this was done to mitigate the effect of the irregularly - shaped , very narrow region in the lower left corner of figure [ rawdata ] .
second , there are 224 additional quasars with @xmath24 and apparent magnitude larger than 19.1 ; these fall in an extremely poorly sampled region , which can also be noted from figure [ rawdata ] .
hence there are 15,057 quasars remaining after this truncation .
the sample is not assumed to be complete within this region .
associated with each sampled quasar is a value for the _ selection function _ , which can be interpreted as the probability that a quasar at this location , and of these characteristics would be captured by the sample .
details regarding how the selection function was approximated via simulations , along with many other details regarding the sample , can be found in @xcite .
the approach taken here is built upon a nonparametric extension of maximum likelihood called _ local likelihood modeling_. this section begins by describing local likelihood density estimation in the general case .
this is then adapted to the problem at hand , initially for the case assuming the random variables are independent .
the case where dependence is allowed is then described as a simple extension . to contrast the standard _
global _ approach to estimation with the local approach employed here , consider the following .
assume the data @xmath25 are realizations ( observations ) of independent , identically distributed random variables from a distribution with density @xmath26 . with classic maximum likelihood estimation ,
one chooses a single estimate from among a class of candidates for @xmath26 ; let @xmath27 denote this class .
specifically , the maximum likelihood estimator ( @xmath28 ) for @xmath26 is defined as the @xmath29 which maximizes @xmath30 \label{usuallike}\ ] ] or , equivalently , the @xmath29 which maximizes @xmath31 ( the notation @xmath32 simultaneously indicates a random variable with unknown density @xmath26 , and the observed realization of that random variable . ) although written here like a density estimation problem , one could imagine the class @xmath27 being indexed by a parameter @xmath33 ; hence this also captures the usual maximum likelihood estimator for parametric problems .
for example , one could define @xmath27 to consist of all gaussian densities as mean @xmath34 and variance @xmath35 vary . in cases
where each @xmath29 is a density ( e.g. , the aforementioned gaussian case ) , the expression in brackets of equation ( [ usuallike ] ) is always zero , and thus unnecessary .
however , it is often advantageous to let @xmath27 be a wider class of smooth , nonnegative functions ; then the bracketed term forces @xmath28 to be a probability density . with local modeling , instead of seeking the single member of the class @xmath27 to be the estimate of @xmath36 , the goal is to approximate @xmath37 for @xmath38 near @xmath39 , yielding the _ local estimate _ @xmath40 .
typically , @xmath41 can be approximated locally by a polynomial ; in fact , a linear form for @xmath42 usually suffices .
see figure [ explainloclike ] . on the left plot
, the dashed line gives the logarithm of the gaussian density with mean zero and variance one .
local linear estimates @xmath42 are shown for each of @xmath43 .
it is unimportant that @xmath44 is not a good estimate of @xmath37 for @xmath38 far from @xmath39 , since many such local estimates will be found and then smoothed together .
these local estimates were calculated with a simulated data set consisting of 10,000 values . the method for finding these local estimates
is outlined next . in independent work ,
@xcite and @xcite localized the likelihood criterion of equation ( [ shortlike ] ) near @xmath45 by writing @xmath46 where @xmath47 is a kernel function parametrized by @xmath48 . a standard choice would be @xmath49 where @xmath16 is a probability density , but more specific forms will be considered ( and required ) below .
the choice of @xmath50 typically has much more influence on the estimator than does the choice of the kernel function .
the local estimate @xmath40 is found by maximizing @xmath51 over @xmath52 belonging to some simple class , usually degree @xmath53 polynomials expanded around @xmath39 : @xmath54 thus , the model is locally parametric with parameters @xmath55 .
one imagines repeating this procedure at a grid of @xmath39-values , call this grid @xmath56 , and hence obtaining a family of local estimates @xmath57 . as a result , @xmath58 is the family @xmath4 of local estimates which maximizes @xmath59 the final local likelihood estimator @xmath60 is constructed by smoothing together the local estimates : @xmath61 thus making dual use of @xmath50 .
returning to figure [ explainloclike ] , the plot on the right shows @xmath60 , the result of smoothing together 101 local linear estimates ( @xmath56 consists of 101 values between -3 and 3 ) . in this case , @xmath62 .
it is clear that the estimate comes very close to the true density . in what follows , simply assume that @xmath63 is chosen so that @xmath64 for all @xmath38 and hence @xmath65 this is a departure from the original approach of @xcite and @xcite , who instead used @xmath66 .
the criterion @xmath67 appears awkward upon first sight , but it possesses the following property : considering @xmath68 again as random variables with unknown density @xmath26 , then @xmath69 is maximized by choosing the family @xmath4 which sets @xmath70 for all @xmath39 and all @xmath38 .
if that choice were made , the estimate would be @xmath71 .
thus , since @xmath72 , the local estimate @xmath42 will approximate the degree @xmath53 taylor expansion of @xmath73 for @xmath38 around @xmath39 .
the expected value of the standard likelihood criterion is also maximized by setting the density equal to the truth , but this localized version has the advantage of allowing the choice of @xmath50 to adjust the amount of smoothness in the estimator . in [ bandselect ] , an objective method for bandwidth selection is described .
there is an apparent conflict between the choice of @xmath50 and the choice of the number of local models ( the cardinality of @xmath56 ) since small @xmath56 will lead to smooth estimates . in the applications here
, @xmath56 is chosen large , so that the amount of smoothing is completely dictated by @xmath50 .
now return to the bivariate density estimation problem using truncated astronomical data .
the available data are denoted @xmath74 and @xmath75 , the vectors of redshifts and absolute magnitudes , respectively .
let @xmath76 denote the region outside of which the data are truncated and let @xmath77 denote the cross - section of @xmath76 at @xmath12 ; @xmath78 is defined similarly .
let @xmath2 denote the unknown joint density of random variables @xmath12 and @xmath79 .
the approach taken here originates in the following naive method . for the moment assume @xmath12 and @xmath79 are independent so that @xmath80 where @xmath81 is the density for redshift and @xmath82 is the density for absolute magnitude .
clearly , the available data allow estimation of the redshift density for observable quasars , denote this density @xmath83 .
this is related to @xmath81 by @xmath84 where @xmath85 is a normalizing constant which forces @xmath83 to integrate to one .
assuming for the moment that @xmath82 were known , it is possible to turn an estimator for @xmath83 into an estimator for @xmath81 by solving equation ( [ xmarg ] ) for @xmath81 : @xmath86 starting with an initial guess at @xmath82 , we could iterate between assuming @xmath82 is known , and estimating @xmath81 , and vice versa .
this procedure is portrayed in figure [ explainmethod ] . using the quasar data
set described in [ data ] , the upper left plot depicts @xmath87 along the vertical axis , with absolute magnitudes ranging from -29.9 to -25.85 .
an ( arbitrary ) assumption is made regarding the density for absolute magnitude @xmath88 , shown as the solid curve in the upper right plot .
for example , one can find that @xmath89 and thus conclude that the observed sample catches 24% of the quasars at @xmath90 .
( the fact that some quasars are missed within @xmath76 is considered later when the selection function is incorporated into the analysis . )
the lower left plot shows how the proportion of quasars observed varies with redshift , i.e. it is a graph of @xmath91 versus @xmath92 .
the dashed line in the lower right plot is @xmath93 , the estimated redshift density for observable quasars .
the solid curve is @xmath94 , as defined above , found by dividing @xmath93 by the proportion of quasars observed at redshift @xmath12 , and then normalizing to force the estimate to be a density .
figure [ explainmethod ] also illustrates problems with this approach .
first , the sharp corner of @xmath76 at @xmath95 leads to a sharp feature in the estimate @xmath94 .
in other words , smooth @xmath83 does not produce a smooth @xmath94 .
second , consider the behavior of @xmath96 for @xmath12 where @xmath97 is small , for instance @xmath98 : even a small error in the estimate of @xmath97 will lead to a large error in @xmath96 .
the fundamental challenge is that a well - chosen estimator ( i.e. , well - chosen smoothing parameters ) for @xmath83 does not necessarily lead to @xmath94 being a good estimator for @xmath81 .
in addition , it is possible to construct examples where this iterative approach will converge to different estimates starting from different initial values for @xmath82 .
despite the aforementioned problems with the use of @xmath94 , that approach can be improved using the local likelihood methods of [ locliketut ] . in what follows ,
@xmath83 is estimated using local polynomial models which include an additive _ offset _ term .
this offset is chosen so that when subtracted off , what remains is a good estimator for @xmath81 .
the procedure is fundamentally the same as that for constructing @xmath94 : starting with an initial guess as to the value of the density for absolute magnitude ( @xmath82 ) , the relationship between @xmath81 , @xmath82 , and @xmath83 ( shown in equation ( [ xmarg ] ) ) is exploited to construct an estimator for @xmath81 .
( here it is assumed that @xmath2 is normalized so that @xmath99 , but this choice is arbitrary since the estimate can be extended outside of @xmath76 and then renormalized as appropriate . ) to start , rewrite equation ( [ xmarg ] ) as @xmath100 where @xmath85 is the constant required to force @xmath99 .
consider the goal of estimating @xmath101 for @xmath38 near @xmath39 .
ideally , it would be possible to fit a local model @xmath102 to obtain both the local estimate @xmath40 and the needed normalizing constant @xmath85 , but truncation does not allow for direct estimation of @xmath81 .
instead , write a local version of equation ( [ withoff ] ) as @xmath103 and then substitute in the expression for @xmath104 from equation ( [ ideal ] ) into equation ( [ locverswithoff ] ) to get @xmath105 of course , it is possible to estimate @xmath83 with the available data and equation ( [ offsetexp ] ) makes it clear that a good way of doing this would be to fit a local polynomial model with @xmath106 included as an offset .
( recall that , for the moment , @xmath82 is assumed known . ) in other words , instead of maximizing the local likelihood criterion @xmath107 over @xmath108 that are polynomials expanded around @xmath39 ( as in equation ( [ polyexp ] ) ) , maximize over functions of the form @xmath109 write @xmath110 as the local likelihood at @xmath39 when the offset is included .
label the parameters which maximize @xmath110 as @xmath111 . comparing equations ( [ ideal ] ) and ( [ offsetexp ] ) , note that @xmath112 is an estimate of @xmath113 and hence @xmath114 is the local ( near @xmath39 ) estimate of @xmath115 . as before , this is repeated for a grid of values @xmath116 and the result is the family @xmath58 which maximizes @xmath117 and the estimate of @xmath118 is found by smoothing together these local estimates : @xmath119 here , it is stressed that estimates of @xmath118 are smoothed together , instead of estimates of @xmath83 .
this is important because now @xmath50 can be chosen to obtain the optimal amount of smoothing for the best estimate of @xmath118 .
this avoids the problems which were evident at @xmath95 in figure [ explainmethod ] .
a method for choosing @xmath50 is described in [ bandselect ] . also , the constant @xmath85 is present in all of these estimates , but it will turn out in the next step that this is exactly what we need : there is no need to renormalize and get separate estimates of @xmath81 and @xmath85 . in this next step ,
@xmath82 will be estimated holding @xmath118 fixed at its current estimate . to ease notation , define @xmath120 with an estimate of @xmath118 in hand ,
now let @xmath121 denote the density for the observable @xmath79 so that since @xmath122 it follows that @xmath123 now consider local models of the form @xmath124 where @xmath125 and now @xmath126 is the logarithm of the offset ; note that an estimator for this was found above in equation ( [ estoff ] ) : @xmath127d{z}.\ ] ] this leaves @xmath128 as an estimator for @xmath82 .
this is then used to reestimate the offset term used in equation ( [ offsetexp ] ) , and the process repeats .
this is conceptually the same procedure as was used to create @xmath94 above , since the estimate of @xmath129 is found by alternating estimating @xmath81 and @xmath82 .
this section will tie together the ideas of the previous .
the iterative procedure described above is computationally tractable , and has intuitive appeal .
remarkably , it is also possible to pose the estimation problem in another manner which is not as computationally useful , but will lead to analytical results .
define @xmath130 \left [ \sum_{u \in { \cal g } } k^*{\hspace{-.04in}}\left({z},u,\lambda\right ) \exp\!\left ( { \bf a}_u\!\left({z}\right)\right ) \right ] d{m}\:d{z}\right\}\label{critind } \ ] ] as the global criterion .
it is a function of both families of local models , @xmath4 and @xmath5 .
the key is to notice that if @xmath5 is held fixed at its current estimate @xmath131 , maximizing @xmath132 over local models @xmath4 is identical to maximizing @xmath133 with fixed estimate of the offset term . to see this , recall equation ( [ foffset ] ) and note that an estimator for @xmath134 is @xmath135 and from equations ( [ callf ] ) and ( [ loclike ] ) , @xmath136 \right ] d{z}\\ & & \hspace{-1.2 in } = { k'}+ \sum_{j=1}^n \left\ { \sum_{u \in { \cal g } } k^*{\hspace{-.04in}}\left({z}_j , u,\lambda\right ) { \bf a}_u\!\left({z}_j\right ) \right.\\ & & \hspace{-1.0in}\left.- \!\int_{\cal a } \left[\sum_{u \in { \cal g}}k^*{\hspace{-.04in}}\left({z},u,\lambda\right ) \exp\!\left({\bf a}_u\!\left({z}\right)\right)\right ] \left [ \sum_{v \in { \cal g } } k^*\!\left({m},v,\lambda\right ) \exp(\widehat { \bf b}_v\!\left({m}\right))\right ] d{z}\:d{m}\right\ } \\ & & \hspace{-1.2 in } = { k''}+ { \cal l}^*\!\left({\bf f } , \widehat { \bf g } , { \bf { z}},{\bf { m}}\right)\end{aligned}\ ] ] where @xmath137 and @xmath138 are constants which do not depend on @xmath4 , and @xmath139 and @xmath140 are the lower and upper bounds on redshift , respectively .
an analogous statement could be made for finding @xmath5 when @xmath58 is held fixed .
thus , the iterative search method described in [ itermethod ] is equivalent to maximizing this global criterion . until now
, the derivation of the approach has assumed that random variables @xmath12 and @xmath79 are independent .
dependence will be incorporated by including a parametric portion @xmath18 so that the assumption becomes that @xmath141 a restriction placed on @xmath6 is that it must be linear in the real - valued parameters @xmath33 . in the absence of a physically - motivated model ,
a useful first - order approximation is @xmath19 .
the global criterion of equation ( [ critind ] ) is naturally updated to @xmath142 \!\left [ \sum_{u \in { \cal g } } k^*{\hspace{-.04in}}\left({z},u,\lambda\right ) \exp\!\left ( { \bf a}_u\!\left({z}\right)\right ) \right ] \!d{m}\:d{z}\right\}. \label{globcrit}\end{aligned}\ ] ] note that with this form , when @xmath4 and @xmath5 are held constant , maximizing @xmath143 over @xmath33 is equivalent to finding the maximum likelihood estimate of @xmath33 .
note also that the sum over the @xmath10 data pairs has also been updated to allow specification of a weight @xmath144 . in this case , the natural choice for the weight is the inverse of the selection function for that data pair .
the intuition is that a pair with selection function of 0.5 is `` like '' two observations at that location . finally , with a criterion of this form
, this estimator can be fit into a general class of statistical procedures called _ m - estimators_. see the appendix ( [ mest ] ) for an overview of m - estimators .
the described procedure returns an estimate normalized to be a probability density over the observable region @xmath76 .
of course , it could be renormalized to meet the goals of the analysis , but care should be taken if the renormalization involves multiplying by a constant which is itself estimated from the data . in certain cases , namely when there is a small sample , this could result in significantly understated standard errors .
luminosity curves are usually stated in units of @xmath145 , and are obtained by multiplying the bivariate density ( normalized to be a probability density over @xmath76 ) by a redshift - dependent constant ; thus no adjustment of the standard errors is needed in this case .
the choice of the bandwidth @xmath50 ( the smoothing parameter ) is critical . choosing @xmath50 too large results in an oversmoothed , highly biased estimator ; choosing @xmath50 too small leads to a rough , highly variable estimator .
this is the _ bias / variance tradeoff_. fortunately , it is possible to select @xmath50 to balance these two in a meaningful , objective manner .
although this discussion applies in general to the problem of density estimation , here it will be described in terms of estimating the bivariate density @xmath146 over @xmath76 .
let @xmath147 denote a general estimator for @xmath146 which is a function of a smoothing parameter @xmath50 .
then , @xmath148 \:d{z}\:d{m}\nonumber \\ & = & \int_{\cal a } \left [ \mbox{variance}\!\left(\widehat { \phi}_{\lambda}\!\left({z},{m}\right)\right ) + \mbox{bias}^2\!\left(\widehat { \phi}_{\lambda}\!\left({z},{m}\right)\right ) \right ] \:d{z}\:d{m}\label{balance}\end{aligned}\ ] ] is the _ integrated mean squared error _ for @xmath147 .
imse is a natural measure of the error in the estimator , and it is apparent from equation ( [ balance ] ) how it balances the bias and variance of the estimator .
although imse can not be calculated , there is an unbiased estimator .
it holds that @xmath149 where @xmath85 is a constant which does not depend on @xmath50 , so it can be ignored .
let @xmath150 denote the estimate of the density at @xmath151 found using the data set with this @xmath152 data pair removed .
following @xcite , @xmath153 so that the _ least - squares cross - validation score _ ( lscv ) , @xmath154 is an unbiased estimator for imse@xmath155 , and hence minimizing it over @xmath50 approximates minimizing the imse .
see @xcite and @xcite for theoretical results showing the large - sample optimality of choosing smoothing parameters to minimize this criterion .
figure [ explainlscv ] gives an example of bandwidth selection by minimizing lscv .
here , 100 simulated values are taken from the gaussian distribution with mean zero and variance one .
the left plot shows how lscv varies with the choice of bandwidth , and leads to a choice of @xmath156 .
the right plot compares the density estimate using three bandwidths @xmath157 with the true density . with the bandwidth too small , there are nonsmooth features , and the bias is low but the variance is high . with
the bandwidth too large , the estimate is smoothing out the prominent peak in the center . here , the variance of the estimate is low , but the bias is high .
the optimal choice gives an estimate close to the truth , and is found using a bandwidth which balances estimates of the bias and variance .
the weighting due to the selection function needs to be taken into account in the previous discussion . recall that the weight @xmath158 is conceptualized as the number of equivalent observations represented by this data pair .
thus `` leaving out '' observation @xmath159 is achieved by reducing its weight from @xmath158 to @xmath160 in the criterion ( equation [ globcrit ] ) .
but one must imagine repeating this @xmath158 times ( for each equivalent observation which observation @xmath159 represents ) .
let @xmath161 denote the _ effective sample size_. the new relationship is @xmath162 where @xmath150 now indicates the estimator evaluated at @xmath151 when the weight on observation @xmath159 is reduced from @xmath158 to @xmath163 .
direct calculation of the leave - one - out estimates would be computationally intractable .
@xcite describes an approximation based on the second - order taylor expansion of the criterion function .
this approximation proves to be highly accurate and computationally simple .
the method described in [ itermethod ] involves fitting local polynomial models at each of a grid of values @xmath116 , for both the @xmath12 and @xmath79 directions .
these derivations were all performed assuming fixed bandwidth @xmath50 used for each of these models .
this was merely for notational convenience ; there is no reason that different bandwidths could not be chosen for each of these local models .
in fact , given that the variables are on different scales , it would be unreasonable to assume the same bandwidth would be a good choice for each . in the results given in the next section , a stated bandwidth is assumed to be on the scale of the variables after they have been transformed to lie in the unit interval , and bandwidths are given as @xmath164 pairs .
allowing the bandwidth to further vary over the different local models gives the overall model fit much flexibility , and lscv can be minimized as before .
a full search over this high - dimensional space is not feasible in practice , however .
this section describes the results of the application of this method to some real and simulated data sets . in all cases ,
linear models are fit when doing the local likelihood modeling ( @xmath165 ) , and @xmath56 is a grid of 100 evenly spaced values in both the @xmath92 and @xmath166 dimensions .
the parametric portion is set as @xmath19 .
bandwidths @xmath164 are stated as proportions of the range for that variable , e.g. @xmath167 means that the bandwidth for the local models for redshift cover 5% of the range @xmath168 .
this method was applied to the sample of quasars described in [ data ] . as stated above
, the method is capable of incorporating the selection function via differential weighting in the criterion ( equation ( [ globcrit ] ) ) , but the selection function does present some challenges in this case . for quasars with
@xmath169 the selection function drops as low as 0.04 due to difficulty in distinguishing quasars from stars of spectral type a and f. this gives a weight of 25 to these quasars , which would be fine if it were exact , but these weights are calculated based on simulations and @xcite states that the selection function in this region `` is quite sensitive to such uncertain details of the simulation . '' they limit the weight on any observation to 3.0 to account for this .
this limit was also imposed in the analysis here .
figure [ lscvplot ] shows how lscv varies with @xmath170 and @xmath171 .
the criterion is minimized when @xmath172 and @xmath173 .
the grid of values at which lscv is calculated is spaced by 0.01 because , as will be seen below , fluctuations of the bandwidths on this scale lead to very little change in the estimates .
the minimum value is -0.0078262 , but no significance can be attached to this value , since lscv is not an unbiased estimate of imse , but instead of imse plus an unknown additive constant .
figure [ bivest ] shows , using the solid contours , the estimate of the quasar density ( two - dimensional luminosity function ) as a function of @xmath92 and @xmath166 , when @xmath167 and @xmath174 .
this estimate is normalized to integrate to one over the entire dashed ( observable ) region .
recall from [ loclikewithoff ] that this is the form which the algorithm provides .
fortunately , this is the ideal form for the estimate .
the ( effective ) count of quasars in the surveyed region is @xmath175 and the survey covers @xmath176 .
thus , the quasar count in a region @xmath177 of @xmath178 space can be estimated using @xmath179.\ ] ] the estimate of @xmath33 is @xmath180 , with a standard error of 0.03 .
although it is not possible to assign physical significance to this value for @xmath33 , it is clear that the possibility that @xmath181 is ruled out , and hence there is very strong evidence for evolution of the luminosity function with redshift .
this estimate has an apparent irregularity in the shape of the density estimate for @xmath182 .
( note the `` bumps '' in the solid contours for all values of @xmath166 at @xmath182 . )
quasars of this redshift are given larger weight due to interference from stars of spectral type g and k. although it is not possible to be certain , it appears that the weighting may not be sufficiently accurate for the quasars .
the weights may be underestimated leading to a corresponding dip in the density estimate .
the bandwidth ( @xmath170 ) is sufficiently small to pick up this artifact .
in fact , lscv forces the bandwidth to be small enough so it can model this feature .
it is hoped that in future work the uncertainty in the weights can be incorporated into lscv . for comparison ,
another estimate was constructed using @xmath183 for local models centered on redshift values larger than 2.0 , while still using @xmath167 for @xmath184 .
this estimate is shown as the dotted contours .
the increased smoothing removes the artifact .
figure [ margs ] shows the estimated count of quasars with @xmath185 as a function of redshift . as in figure
[ bivest ] , the solid curve is the estimate with the lscv - optimal bandwidths , and the dashed estimate is found using the increased smoothing . figure [ allconds ] shows quasar counts as a function of absolute magnitude at a collection of redshift values .
comparisons are made with the estimates given in @xcite which were found using the bin - based method of @xcite .
the error bars in both figures [ margs ] and [ allconds ] are one standard error , but represent statistical errors only .
the error bars do not account for incorrect specification for the parametric form @xmath6 .
but , if there is bias from the incorrect specification of @xmath6 , the binned estimates must share these biases .
this would be surprising since , while having higher variance , estimates constructed from binning do not make assumptions regarding the evolution of the luminosity function , and hence a well - constructed estimate should be approximately unbiased
. figure [ allconds ] also provides insight into the sensitivity of the estimate to the bandwidth choice .
it would be of great concern if small changes in bandwidth led to significant changes in the estimate . to explore this ,
eight additional estimates were constructed using every possible combination of @xmath186 and @xmath187 .
the results are shown as gray curves in each plot of figure [ allconds ] , but are only visible at @xmath188 and @xmath189 .
the fluctuations are small relative to the size of the error bars . clearly , the estimates are insensitive to these perturbations .
simulations were performed to further explore the behavior of the estimator .
for these , the estimate shown in the dotted contours in figure [ bivest ] is taken to be the true bivariate density ; the truncation region is unchanged .
the idea is to ask the following : if the truth were , in fact , the estimate found here , would this method be able to reach a good estimate of the density under identical conditions ( same sample size and truncation region ) ? hence ,
the new data sets were simulated consisting of 16,589 @xmath178 pairs within the observable region .
the first of these data sets was utilized to find the optimal smoothing parameters ; these were found to be @xmath190 and @xmath191 .
each of the other 19 data sets was analyzed using these values , so that these simulations also provide insight into the adequacy of this approach to bandwidth selection .
figure [ simconds ] shows the results from the simulations by comparing estimates of the cross - sections of the estimates @xmath192 at four different redshifts .
each dashed curve is an estimate from one of the 20 data sets .
the solid curve is the truth .
these results show strong agreement between the estimates and the truth over the regions where data are observed .
there is some bias in the tails , but this is in regions far from any observable data .
in addition , these simulations provide strong evidence that the estimates of the standard errors are accurate : the variability in the estimates is comparable to the size of the error bars .
the semiparametric method described here is a strong alternative to previous approaches to estimating luminosity functions .
the primary advantage is that it allows one to estimate the evolution of the luminosity function with redshift without assuming a strict parametric form for the bivariate density .
instead , one only needs to specify the parametric form for a term which models the dependence between redshift and absolute magnitude .
future work will focus on specifying a physically - motivated form for this parametric portion , but the results from the analysis of a sample of quasars reproduce well those from @xcite while only assuming a simple , first - order approximation to the dependence .
other portions of the bivariate density are modeled nonparametrically , and are functions of smoothing parameters .
using least - squares cross - validation , these smoothing parameters can be chosen in an objective manner , by minimizing a quantity which is a good approximation to the integrated mean squared error .
results from simulations show that , with a data set of this size , the method is indeed capable of recapturing the true luminosity curves under the truncation observed in these cosmological data sets .
the author gratefully acknowledges the comments of the referee , which greatly improved this paper , and the contributions of peter freeman , chris genovese , and larry wasserman of the department of statistics at carnegie mellon university .
the author s work is supported by nsf grants # 0434343 and # 0240019 .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , cambridge university , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
15 natexlab#1#1 boyle , b. , shanks , t. , croom , s. , smith , r. , miller , l. , loaring , n. , and heymans , c. 2000 , mnras , 317 , 1014 choloniewski , j. 1986 , mnras , 223 , 1 efron , b. & petrosian , v. 1999 , j. am .
assoc . , 94 ,
824 efstathiou , g. , ellis , r. , peterson , b. 1988 , mnras , 232 , 431 felten , j. 1976 , apj , 207 , 700 hall , p. 1983 , ann .
stat . , 11 , 1156 hjort , n. & jones , m. 1996 , ann .
stat . , 24 , 1619 jackson , j. 1974 , mnras , 166 , 281 loader , c. 1996 , ann . stat . , 24 , 1602 lynden - bell , d. 1971 , mnras , 155 , 95 maloney , a. , & petrosian , v. 1999 , apj , 518 , 32 nicoll , j. & segal , i. 1983 , a&a , 118 , 180 page , m. and carrera , f. 2000 , mnras , 433 qin , y .- p . , & xie , g .- z . 1999 ,
a&a , 341 , 693 richards , g. , et al .
2006 , apj , 131 , 2766 rudemo , m. 1982 , scan .
j. of stat .
, 9 , 65 sandage , a. , tammann , g. , & yahil , a. 1979 , apj , 232 , 352 schafer , c. ( 2006 ) submitted .
available as cmu dept . of stat .
tech report # 842 http://www.stat.cmu.edu/tr/tr842/tr842.html schmidt , m. 1968 , apj , 151 , 393 stone , c. 1983 , ann .
, 12 , 1285 takeuchi , t. , yoshikawa , k. , & ishi , t. 2000 , apjs , 129 , 1 willmer , c. 1997 aj , 114 , 898 wang , m .- c .
1989 , j. of amer .
84 , 742 wasserman , l. , miller , c. , nicol , r. , genovese , c. jang , w. , connolly , a. , moore , a. , & schneider , j. 2001 , astro - ph/0112050 woodroofe , m. 1985 , ann .
stat . , 13 , 163 york , d. 2000 , aj , 120 , 1579
the procedure described in [ ourmodel ] can be fit into a general class of statistical estimators called _ m - estimators_. in the simplest case , a m - estimator for a parameter is constructed by maximizing a criterion of the form @xmath193\ ] ] where @xmath194 are the observed data , assumed to be realizations of independent , identically distributed random variables and @xmath195 is the parameter to be estimated .
the function @xmath196 is some criterion .
for example , in the case of finding the maximum likelihood estimate of @xmath195 , the function @xmath197 , where @xmath198 is the density corresponding to parameter @xmath195 .
most least squares problems can be stated as m - estimators .
standard theory for m - estimators can be applied to obtain an approximation to the distribution of @xmath199 , which can then be used to find standard errors and form confidence intervals . in the case at hand
, @xmath32 is the pair @xmath200 , @xmath201 , and @xmath202 \left [ \sum_{u \in { \cal g } } k^*{\hspace{-.04in}}\left({z},u,\lambda\right ) \exp\!\left ( { \bf a}_u\!\left({z}\right)\right ) \right ] d{m}\:d{z}. \ ] ] see schafer ( 2006 ) to see the derivations of the approximate distribution for the estimator in this case .
the m - estimator could be generalized to the following : @xmath203\ ] ] where @xmath204 is the weight given to the @xmath152 observation .
this allows for easy incorporation of the selection function into the analysis .
the statistical theory for this _ weighted m - estimator _ is a simple extension of that for the standard m - estimator . | the observational limitations of astronomical surveys lead to significant statistical inference challenges .
one such challenge is the estimation of luminosity functions given redshift @xmath0 and absolute magnitude @xmath1 measurements from an irregularly truncated sample of objects .
this is a bivariate density estimation problem ; we develop here a statistically rigorous method which ( 1 ) does not assume a strict parametric form for the bivariate density ; ( 2 ) does not assume independence between redshift and absolute magnitude ( and hence allows evolution of the luminosity function with redshift ) ; ( 3 ) does not require dividing the data into arbitrary bins ; and ( 4 ) naturally incorporates a varying selection function .
we accomplish this by decomposing the bivariate density @xmath2 via @xmath3 where @xmath4 and @xmath5 are estimated nonparametrically , and @xmath6 takes an assumed parametric form .
there is a simple way of estimating the integrated mean squared error of the estimator ; smoothing parameters are selected to minimize this quantity .
results are presented from the analysis of a sample of quasars . |
eddy currents are induced electric currents in a conducting material that result when either the object moves through a nonuniform magnetic field or is stationary but subject to a time - changing magnetic field .
they were first observed by franois arago@xcite but it was not until michael faraday s discovery of induction@xcite that the mechanism of this phenomenon was understood .
a pioneering experiment to quantify the heat dissipated by the eddy currents , and hence connect electromagnetism to thermodynamics , is von waltenhofen s pendulum@xcite [ figure [ fig : schematic](a)(c ) ] .
the bob of the pendulum is a non ferromagnetic conducting plate which swings between the poles of an electromagnet , generating circulating eddy currents .
the dissipative eddy currents tend to oppose the motion of the bob , and hence damp the pendulum .
the experiment is an example of magnetic braking : when the magnet is turned off , the pendulum oscillates as a classic pendulum , damped only by air resistance and frictional forces in the hinge ; when the magnetic field is switched on , the pendulum comes to a rapid stop .
von waltenhofen s pendulum is a popular demonstration in high school@xcite and introductory undergraduate@xcite classes because it provides a striking visual illustration of the invisible eddy currents .
it is easy to set up , and gives the instructor a chance to discuss real - world applications such as magnetic braking in roller coasters , hybrid cars and the detection of flaws in conductive materials .
moreover , by using different shaped plates as bobs , some of the characteristics of eddy currents can be explored : if the plate is now substituted with one where the slits do not reach the edges [ figure 1(d)(ii ) ] , the damping is only mildly altered .
if a solid plate is replaced by a comb - like plate with slits cut into it [ figure 1(d)(iii)-(iv ) ] the damping effect is greatly attenuated .
students infer from this that the eddy currents must be localized to some region of the plate .
typically , however , the relationship between the shape of the plate and the damping is considered only qualitatively due to the complexity of the calculations required , which require advanced mathematical techniques introduced in higher - level electromagnetism classes . in this paper , therefore , we apply some of these strategies to predict the damping behavior of conducting plates with different geometries .
we investigate four plate shapes , illustrated in figure 1(d ) : ( i ) a square plate , ( ii ) a plate with holes that do not reach the edge , ( iii ) a plate with with two slots , ( iv ) another with four slots . the present work was primarily performed during a project - based graduate electrodynamics class with the objective of giving students a real , tractable example of the utility of their calculations .
it was also an opportunity to address the challenges of connecting theory to experiment , even where the fundamental theory is uncontroversial .
the paper is organized as follows : in section ii , the form of the damping for the present experiment is derived first in terms of a simplified model suitable for undergraduate classes and the effect of these damping terms on the simple pendulum is explored . in section iii , we present a more sophisticated model derived from maxwell s equations in the magneto - quasistatic approximation . within this framework , the rate of power dissipation for each plate is estimated using a conformal mapping technique and a finite element solution . in section iv ,
the motion of the pendulum with each of the four plates is determined experimentally and compared with the predictions of each of the theoretical models in section v. in section vi , the pedagogical context of this work is discussed .
brief conclusions are presented in section vii .
schematic of von waltenhofen s pendulum .
( a ) an aluminium plate swings between the pole pieces of an electromagnet .
as it passes through the fringing field , eddy currents are induced in the plate that dissipate energy and damp the pendulum .
( b ) as the plate passes through the center of the pole pieces , the rate of change in flux through the plate vanishes and so the eddy current momentarily vanishes .
( c ) as the pendulum swings back , the eddy current reverses direction .
( d ) various plate geometries under investigation in the present work ( to scale ) . ]
the damped simple pendulum is , in the presence of dissipative forces , described by the equation , @xmath0 where @xmath1 is the rayleigh dissipation function , @xmath2 is the natural frequency of the pendulum and @xmath3 is a generalized coordinate . in the absence of the magnetic field ,
a stokes - like drag is assumed , @xmath4 arising from friction in the bearing and air resistance . in this section and
the next , the form of the dissipation function in the presence of the magnetic field is derived , first for a simplified model suitable for introductory classes and second a more physically correct model based on the magneto - quasistatic approximation .
the remaining theory section considers the effect of these terms on the solutions of ( [ eq : pendulumeqn ] ) . to develop the simplified model
, we make the gross approximation that the eddy currents are completely localized to the boundary of the plate .
then , the plate may be replaced by a thin closed loop of wire of equivalent shape with area @xmath5 , length @xmath6 , resistivity @xmath7 and with a normal unit vector @xmath8 .
the magnetic flux passing through the loop is , @xmath9 where @xmath10 is the mean value of the magnetic field over the plate when it is located at a particular position @xmath11 .
the emf induced around the loop is , by the faraday - lenz law and using the chain rule , @xmath12 the power dissipated in the loop is precisely the dissipation function desired and readily calculated , @xmath13 which , using @xmath14 , becomes , @xmath15 where the highlighted term represents the `` geometric factor '' arising from the shape of the plate .
the dissipation function ( [ eq : rayleighdissipationfunction ] ) has the form of a position - dependent stokes drag .
it is now necessary to solve the pendulum equation including the new dissipation terms ( [ eq : rayleighdissipationfunction ] ) , @xmath16\dot{q}=0,\ ] ] for which the form of the magnetic field must be known . here ,
the coordinate system is oriented as shown in figure 1 such that the plate normal @xmath8 is parallel to the @xmath17 axis , so only @xmath18 is required ; the origin is chosen to lie in the middle of the pole pieces .
an approximation made in other studies@xcite is that the magnetic field is constant between the pole pieces and zero everywhere else , i.e. that , @xmath19 where @xmath20 is the width of the pole pieces .
if the plate is square of side @xmath6 , the average value of the field over the surface of the plate when it is centered at @xmath11 along the @xmath21-axis may be determined , @xmath22 for the ansatz ( [ eq : bansatz ] ) , the flux through the plate increases ( decreases ) linearly as it enters ( leaves ) the field .
depending on the size of the plate , and the amplitude of the motion , the plate may also experience portions of the motion where there is no change in flux .
if the plate is the same size as the pole pieces @xmath23 , and never fully leaves the field , i.e. @xmath24 , then @xmath25 and the classical damped pendulum is recovered , but with a modified damping constant , @xmath26\dot{x}=0.\label{eq : equationofmotion}\ ] ] in order to gauge the validity of the ( [ eq : bansatz ] ) for the pendulum we calculated the spatial dependence of the magnetic field due to the pole pieces in the magnetostatic formulation , using a scalar potential function defined @xmath27 and subject to laplace s equation @xmath28 .
this was solved using the finite element software ` flexpde ` for square pole pieces of width @xmath29 cm and separation @xmath30 cm .
dirichlet boundary conditions were imposed on the pole pieces . a plot of @xmath18 along the horizontal centerline of the midplane of the pole pieces is displayed in fig .
[ fig : fieldprofile](a ) , showing as expected that the field is constant between the pole pieces with the fringing field decaying rapidly outside . from this calculated profile ,
the flux @xmath31 through a square plate placed at different horizontal positions was also calculated ; the result is displayed in fig . [
fig : fieldprofile](b ) .
notice that the flux depends linearly on position for @xmath32 cm , suggesting that the equation of motion ( [ eq : equationofmotion ] ) is likely to be valid for moderate amplitudes ; however around @xmath33 the flux depends quadratically on @xmath21 , suggesting that ( [ eq : equationofmotion ] ) is likely to be inaccurate for small amplitudes .
( a ) field strength as a function of distance @xmath21 through the center of the pole pieces .
( b ) magnetic flux through a square plate as a function of the position of the plate @xmath21 . ]
since the eddy current is not strictly confined to the edges of the plate , it is necessary to develop a more sophisticated model by solving maxwell s equations @xmath34 together with the continuity equation @xmath35 to determine the true current density in the plate . here ,
the _ magneto - quasistatic _ approximation@xcite has been made , i.e. the maxwell correction to ampere s law is neglected .
this is justified because the timescale of variation of the magnetic field over the surface of the plate is sufficiently long that the eddy current produced does not significantly alter the magnetic field in turn . to perform the analysis ,
consider a thin metal plate of uniform conductivity @xmath36 and thickness @xmath37 and where the shape of the plate is arbitrary .
it is assumed that ohm s law holds for the plate , so @xmath38 in the frame of the laboratory , since the pendulum is rigid , the position of the plate may be described by the single generalized coordinate @xmath3 in eq .
[ eq : pendulumeqn ] . as the pendulum swings
, the plate passes through regions of different magnetic field strength .
viewed from the point of view of the plate , the magnetic field across its surface changes as a function of time .
this may be described by defining a local coordinate system @xmath39 relative to the plate such that the plate lies in the @xmath40 plane , and considering the magnetic field in this frame , @xmath41 we note that the reference frame @xmath39 is _ not _ an inertial frame as the pendulum undergoes periodic acceleration during the course of its motion ; however we neglect these effects as the pendulum velocity @xmath42 . by faraday s law , an electric field and hence an eddy current
is induced in the plate : if @xmath43 the skin depth of the material it may be assumed that @xmath44 is oriented in the @xmath45 plane and does not vary significantly over the thickness of the plate ; hence the problem is quasi - two - dimensional . from equations ( [ eq : maxwell ] ) and ( [ eq : ohm ] ) , it is a standard textbook problem to show that the quantities @xmath44 , @xmath46 , @xmath47 and @xmath48 all obey the diffusion equation , @xmath49 since @xmath50 , the continuity equation implies that @xmath48 must have no divergence and hence can be written using a stream function , _
@xmath51 _ by inserting ( [ eq : jstreamfunction ] ) into ( [ eq : diffusioneq ] ) , together with the given applied magnetic field into maxwell s equations , it may be shown that the stream function obeys poisson s equation @xmath52 where @xmath53 here is the 2d laplacian , @xmath54 is the magnetic field strength and we defined a source term @xmath55 . since @xmath48 must lie tangentially to the edge of the plate , @xmath56 obeys a dirichlet boundary condition @xmath57 on all edges of the plate . from the solution for @xmath56 , the instantaneous rate of power dissipation by the plate
is then calculated , @xmath58 where the integral is to be taken over the plate .
inserting the stream function ( [ eq : jstreamfunction ] ) into ( [ eq : powerdissipation ] ) , integrating by parts and using ( [ eq : poisson ] ) , we obtain @xmath59 a formal solution to ( [ eq : poisson ] ) may be constructed using the green function for the laplacian , @xmath60 inserting this into ( [ eq : powerdissipationintermediate ] ) and rearranging we obtain the dissipation function , @xmath61\zeta(\mathbf{x}')\ da'\\ & = & -d\sigma b_{0}^{2}\dot{q}^{2}\int\int\zeta(\mathbf{x}',q)\zeta(\mathbf{x}'',q)g(\mathbf{x}',\mathbf{x}'')\ da'\ da''.\label{eq : powerdissipation-2}\end{aligned}\ ] ] since the source function @xmath62 depends on the position of the pendulum @xmath11 , this has the form @xmath63 where we define a local dissipation rate @xmath64 . to determine an effective damping coefficient
, the calculations proceed as follows : in subsection [ sub : evaluating - the - fringe ] the form of @xmath62 is determined , then the positional dependence of the drag term is evaluated using a conformal mapping technique in subsection [ sub : conformal - mapping - solution ] , and also by solving ( [ eq : poisson ] ) with ` flexpde ` in subsection [ sub : finite - element - simulations ] ; finally , the effect of the position dependent damping on the motion of the pendulum is considered in the subsection [ sub : effect - on - the ] . before evaluating the power dissipation
, it is first necessary to determine the functional form of the magnetic fringe field that the plate experiences as it moves between the pole pieces and hence the function @xmath65 .
this could also be obtained from a finite element calculation as was done for the simplified model in section ( [ sec : simplified - model ] ) .
we made the approximation that only the fringe fields from the sides of the pole pieces are significant in the damping , and the vertical extent of the pole pieces are neglected ; the formulation above remains sufficiently general that these details could easily be included in a more sophisticated calculation of @xmath65 .
conformal mapping @xmath66 from ( a ) a portion of the strip @xmath67 to ( b ) the depicted region in the @xmath17 plane .
places where the integrand in ( [ eq : scfringe ] ) vanishes are indicated by red circles ; these introduce corners into the map .
contours of constant @xmath68 ( dashed lines ) and @xmath69 ( solid lines ) and their projection are indicated corresponding to equipotentials and field lines respectively .
inset : plots of the field profile @xmath70 and its spatial derivative @xmath71 calculated from the conformal mapping .
, width=288 ] within this approximation , the problem is two dimensional and hence can be solved using conformal mapping , an important technique for solving electrostatics problems . for a full discussion
we refer the reader to @xcite and @xcite .
briefly , however : a map from a domain to its image is called _ conformal _ if angles measured locally around a point in the source domain are preserved under the mapping .
a key property of such maps is that if a function that is a solution to laplace s equation is constructed on some domain , then the projection of that function under the action of the mapping will _ also _ obey laplace s equation .
boundary value electrostatics problems in two dimensions can hence often be solved using this method by constructing a solution to laplace s equation with appropriate boundary conditions in some simple domain and then finding a conformal map to the domain of interest .
the powerful riemann mapping theorem guarantees such a map exists if the domains are simply connected ; mappings between multiply - connected domains may also exist if certain compatibility criteria are met@xcite .
a rich source of such maps are analytic complex functions @xmath66 that map a domain in the complex plane @xmath72 to an image in the plane @xmath73 . the required mapping to find the fringe field is , @xmath74^{1/2}\ d\xi.\label{eq : scfringe}\ ] ]
as shown in figure [ fig : cmfringe ] , this maps the strip defined by @xmath75 in the @xmath72 plane onto the depicted region in the @xmath76 plane ; the resulting shape resembles an overhead view of the edge of two parallel pole pieces .
the integrand in ( [ eq : scfringe ] ) vanishes at @xmath77 and @xmath78 .
the effect of these poles is to introduce corners in the map ; the angle of these is controlled by the strength of the pole , here @xmath79 .
the factor of @xmath80 ) is required to ensure that the boundary of the @xmath81 part of the strip remains horizontal while the @xmath82 half is mapped to the vertical boundaries .
the integral can be performed analytically , @xmath83 where the overall prefactor is chosen to ensure that the @xmath81 half of the strip is mapped onto @xmath84 .
the magnetic field strength can be determined using this map as follows : in the space between the pole pieces @xmath85 and hence the magnetic field can be constructed from a scalar potential @xmath86 subject to laplace s equation @xmath87 . on the boundary
, @xmath46 must be perpendicular to the pole pieces so these must be surfaces of constant @xmath88 . in the @xmath89-plane ,
the system resembles a parallel plate capacitor with the solution @xmath90 where @xmath5 and @xmath91 are arbitrary constants .
since laplace s equation is invariant under the conformal mapping , this solution remains valid when projected into the @xmath17-plane : the lines of constant @xmath92 are equipotentials and lines of constant @xmath93 are the field lines as shown in figure [ fig : cmfringe ] . the magnetic field can be found from @xmath94 ; this can be expressed in complex form in the @xmath89-plane as @xmath95@xcite . evaluating @xmath66 and @xmath96 along the center of the strip @xmath97 yields @xmath98 along the midplane ; the @xmath21 derivative of @xmath99 required for the pendulum motion
is evaluated from these using the chain rule , @xmath100\nonumber \\
b_{y}^{0 } & = & \left(1+e^{2\pi t}\right)^{-1/2}\nonumber \\
\frac{\partial b_{y}^{0}}{\partial x } & = & -\frac{1}{4}\pi\text{sech}(\pi t)^{2}.\label{eq : fringefield}\end{aligned}\ ] ] having tabulated values of * @xmath101 * using the above implicit expressions , a final form for @xmath65 was assembled from two appropriately translated and rescaled copies of @xmath101 as shown in the inset of figure [ fig : cmfringe ] ; this was used in subsequent calculations as an ` interpolatingfunction ` in _ mathematica _ , or exported as a table for use in ` flexpde ` . for the plates and pole pieces considered , it is actually the case that the plates were slightly larger than the width of the pole pieces ; while this could easily be accommodated by a simple rescaling of the solution ( [ eq : fringefield ] ) , it was found in practice not to affect the dissipation enough to affect the quality of the fit .
( a ) map from the unit circle to a paddle with a single slit of variable geometry .
( b ) map from a rectangular domain to the slit .
the position of poles in the canonical domains is indicated with red circles ; these become vertices in the image domain .
( c ) two maps relevant to the experiment : i ) map from a circle to a square ; ii ) map from a rectangular domain to a square with two slitsnote significant crowding in the central spar . ]
conformal mapping can be also used to evaluate the power dissipation of the plate as it passes through the fringing field .
the strategy here is to find a mapping from a simple canonical domain on which the green function is known to one that approximates the shape of the plates ; then the integral ( [ eq : powerdissipation-2 ] ) can be computed . as will be seen
, the choice of canonical domain is important to render the problem tractable .
we therefore examined two choices : first , the mapping from the unit circle ( @xmath89-plane ) to a polygon of interest ( @xmath17-plane ) defined by an ordered set of vertices @xmath102 is given by the schwarz - christoffel formula@xcite , @xmath103 where @xmath104 is the exterior turning angle of the polygon at vertex @xmath105 and the @xmath106 are the points on the unit circle that are mapped to the vertices of the polygon @xmath102 .
the constants @xmath5 and @xmath107 are translation and scaling factors respectively . while the set of parameters @xmath108 can be determined directly from the polygon , the positions of the parameters @xmath109 must be determined numerically@xcite . @xmath5 and @xmath107 are then determined to rotate and scale the result onto the polygon of interest . some representative results for a single slit are shown in fig .
[ fig : conformalmapping-2](a ) . due to the symmetry of the problem , the position of three poles must be determined numerically with the remaining vertices provided by reflection .
the second choice of canonical domain is a rectangular portion of the semi - infinite strip , @xmath110 , @xmath111 .
a mapping onto the shape of interest is given by , @xmath112\right)^{-\alpha_{i}}\ d\xi\label{eq : scstripmap}\ ] ] where similarly @xmath106 are the position of the poles which produce the required corners in the image domain and @xmath104 are the turning angles ; this formula is a more general version of equation ( [ eq : scfringe ] ) used to determine the fringing field .
results for the single slit are shown in fig .
[ fig : conformalmapping-2](b ) .
note that this map is approximate only : the left and right sides of the rectangle are mapped to the bottom of the paddlethe side that the slit is onbut their image under ( [ eq : scstripmap ] ) is not a straight line .
hence , only two poles @xmath109mapped to the interior vertices of the slitneed be determined numerically ; the third parameter is the width of the rectangle @xmath113 . despite the approximate nature of the map ,
the distortion of the lower boundary is very small .
these parameters were found by minimizing an objective function that was constructed from the @xmath114 norm of the difference between the ratios of the sides of the polygon and their desired values ; a notebook to do so is provided as supplementary material .
robust numerical codes including ` scpack ` and the sc toolbox@xcite are available for more complex shapes . comparing the two choices of canonical domain in figure [ fig : conformalmapping-2 ]
, an important difference is apparent .
the poles in the maps from the unit circle tend to lie very close to one another , becoming in some cases indistinguishable in the figure , though they remain numerically distinct . as a consequence of this phenomenon , known as _ _
crowding__@xcite , very small regions adjacent to these crowded poles are mapped to large regions in the image domain .
crowding is problematic , because it makes numerically finding the @xmath106 challenging ; it will also cause problems in evaluating the power dissipation .
notice that using the rectangular source domain eliminates the crowding problem as the poles remain well - separated . in fig .
[ fig : conformalmapping-2](c ) , we show two maps relevant to the experiment . first , the map from the circle to the square is famously @xmath115 an example with two slits that closely matches the experimental plate is shown in fig .
[ fig : conformalmapping-2](c)(ii ) , constructed by introducing additional poles into one side of the one - slit rectangle maps above .
unfortunately , the crowding phenomenon is visible in this map : a very small piece of the rectangle is mapped to the central spar of the plate . as will be seen later
, this will reduce the accuracy of the power dissipation calculation when this spar coincides with the edge of the pole pieces . while it is possible to find maps with crowding localized to different portions of the plate@xcite , it can not be alleviated entirely if more than one slit is present .
hence , we did not pursue maps for plates with more than 2 slits .
turning , then , to the issue of evaluating the power dissipated by the induced current , we initially tried to use the eigenfunction expansion of the green function , @xmath116 which , inserted into the integral in ( [ eq : powerdissipation-2 ] ) allows one to obtain , @xmath117^{2}}{\lambda_{n}}\label{eq : conformalintegral}\end{aligned}\ ] ] the 2d integral in ( [ eq : ci1 ] ) can then be performed in the @xmath89-plane using the normalized eigenfunctions of the unit disk and the mapping @xmath66 , @xmath118 unfortunately , we found that the series ( [ eq : conformalintegral ] ) converges very slowly except for the square map ( [ eq : squaremap ] ) .
the problem is further compounded for the unit circle by the fact that the integrand in ( [ eq : conformal2d ] ) is very sharply peaked and localized to the boundary in many cases due to crowding .
for these reasons , we abandoned this approach and the unit circle maps as viable means to evaluate the dissipation .
the power dissipation integral ( [ eq : powerdissipation-2 ] ) can be re - expressed in the @xmath89-plane as , @xmath119 in order to evaluate this integral , one requires a numerically efficient representation of the green function .
unfortunately , the green s function for a rectangle can not be expressed in closed form , requiring expensive evaluation of elliptic integrals .
recently , however , renewed interest in green function methods has yielded rapidly converging series representations@xcite .
briefly , the strategy is to start from the eigenfunction expansion of @xmath120 ( [ eq : eigenfunctionexpansion ] ) on the rectangle and separate the double summation into two pieces : one that can be summed analytically and a second that remains to be performed numerically . fortuitously
, the analytical part contains the inherent singularity ; the remaining numerical sum is a smooth function requiring few terms to converge to machine precision .
equation ( 7 ) of @xcite presents an expression for the green function in a square ; we derived an equivalent expression for the rectangle @xmath110 , @xmath111 of interest , @xmath121 where @xmath72 and @xmath122 consistent with our earlier definitions . inserting ( [ eq : rapidgreenfunction ] ) , ( [ eq : scstripmap ] ) and @xmath65 derived in the previous section into ( [ eq : dissipation4d ] ) , we evaluated the local power dissipation rate @xmath64 as a function of @xmath11 for a variety of plate geometries .
a notebook to perform these calculations is presented as supplementary material ; we defer consideration of the results until the subsequent section , in which the same function is calculated by finite element simulations .
( a ) dissipation rate as a function of position and ( b ) i - iv corresponding current density profiles obtained numerically for the various shapes from ` flexpde ` . ]
finite element analysis is a commonly used numerical technique for solving partial differential equations in systems too complicated for analytical solution . here
, the pde of interest is , as shown in the previous sections , poisson s equation with a spatially dependent source term ( [ eq : poisson ] ) .
this must be solved to obtain the stream function @xmath123 for each of the paddle shapes of interest and at each point @xmath11 in the motion of the paddle .
suitable input files for ` flexpde ` for each plate are included as supplementary material .
each script contains field variable and parameter definitions , a definition of the pde , a description of the appropriate domain boundary and specification of the boundary conditions , and a list of the outputs desired , i.e. @xmath56 , @xmath48 and the value of the integral ( [ eq : powerdissipation-2 ] ) . a relative error of @xmath124 in the absolute @xmath56 was requested and used by ` flexpde ` to guide adaptive refinement . for each plate ,
the boundary condition on the exterior is simply @xmath57 , which forces the stream function to lie tangentially to the edge of the plate .
the plate with holes requires special treatment : around each of the holes @xmath56 should be constant , i.e. the hole should be an equipotential surface , but the correct value is not _ a priori _ known .
a way to obtain the correct solution is to exploit an electrostatic analogy , modifying poisson s equation by introducing a permittivity @xmath125 , @xmath126 the computational domain is extended so that @xmath56 is also defined inside the holes .
different values of @xmath125 are used for the interior of the plate @xmath127 and the holes @xmath128 . as the ratio @xmath129
, the solution converges to the desired solution where @xmath56 constant inside the holes . for the purposes of calculation ,
@xmath127 is chosen arbitrarily to be @xmath130 and @xmath128 is successively increased ; the value of ( [ eq : powerdissipation-2 ] ) converged to a relative error of @xmath131 for @xmath132 .
plots of the stream function @xmath56 and the corresponding current densities @xmath48 resulting from the calculations are displayed in figure [ fig : flexpde ] , together with a plot of the dissipation rate @xmath64 .
comparison of finite element ( solid lines ) and conformal mapping ( points ) approaches to calculating the local rate of dissipation @xmath64 . ]
it is now possible to compare the conformal mapping and finite element methods for this problem .
to do so , we used both methods to calculate the dissipation rate @xmath64 for the plate geometries displayed in fig .
[ fig : conformalmapping-2](b ) and ( c ) , i.e. those for which we were able to find appropropriate mappings . to perform the integral ( [ eq : dissipation4d ] )
, we used _
mathematica s _ ` nintegrate ` with the ` adaptivemontecarlo ` method and @xmath133 evaluation points .
the calculation for each plate took @xmath134 minutes for ` flexpde ` and @xmath135 minutes for the conformal mapping approach on the same computer ( apple macbook air 11 late 2014 ) .
results for the various plates are shown in fig .
( [ fig : flexpdecomparison ] ) .
the agreement is very good for the single slit geometries , with the position and size of extrema reproduced well by both methods .
for the two slit geometry , however , the conformal mapping method performs less well : the outer maxima are reproduced well , but the central one is not .
this corresponds to the positioning of the plate where the edge of the pole piece coincides the central spar , where as discussed in section [ sub : conformal - mapping - solution ] the mapping is crowded .
clearly from this comparison , the method of choice appears to be finite elements , since the results are less noisy and all plate shapes can be simulated using the method .
having formulated both models , we turn to the effect of the dissipation on the motion of the pendulum . for both the simplified wire model of section [ sec : simplified - model ] and the magneto - quasistatic model of section [ sec : magneto - quasistatic - model ] ,
the equation of motion has the form , @xmath136 for the wire model , @xmath137\ ] ] as read off from eq .
( [ eq : equationofmotion ] ) ; for the magneto - quasistatic model the stokes drag is supplemented by the position - dependent dissipation rate , @xmath138 as is well - known , the equation of motion ( [ eq : generalequationofmotion ] ) with constant @xmath139 can be solved analytically yielding , @xmath140 where @xmath141 and @xmath142 are amplitude and offset parameters , @xmath56 is a phase and @xmath143 .
hence , the motion remains oscillatory for @xmath144 . if the plate is centered with respect to the pole pieces , @xmath145 ; @xmath146 and @xmath56 similarly represent initial conditions that are experimentally controllable . rather than fit to the entire trajectory , it is convenient to look only at the extreme points where the pendulum is stationary and @xmath147 . inserting this into ( [ eq : solutionlinear ] ) and rearranging , we obtain @xmath148 showing that the values of the position of the pendulum at successive stationary points should obey a linear relation on a plot of @xmath149 versus @xmath150 . for the magneto - quasistatic model , analytical solution of ( [ eq : generalequationofmotion ] ) is impossible and hence the equation must be solved numerically .
this was performed in _
mathematica_. because the strategy of looking at successive extrema on log scales remains a valuable fitting strategy even if the equation of motion is solved numerically , we additionally incorporated an extremum - finding routine to identify the stationary points from the calculated trajectory . with these simplified analytical and numerical magneto - quasistatic models ,
we are now ready to compare them with experimental data .
( a ) photograph of the experimental setup , designed as a classroom demonstration for introductory physics .
( b ) detail showing placement of camera and screen for imaging .
( c ) photograph of paddle shapes used .
, width=316 ] in order to test the predictions of the calculations performed in previous sections , we performed an experiment to observe the motion of a pendulum with the different paddles attached and fit the models described in previous sections to the trajectories obtained .
the magnetic brake apparatus used , shown in figure [ fig : app ] , was designed as a classroom demonstration for introductory physics classes .
the apparatus consists of an electromagnet with flat pole pieces of adjustable separation , powered by a 12v car battery . a potential divider and ammeter
were incorporated into the circuit so that the current through the magnet could be changed and measured .
the pole pieces are @xmath151 cm and for the present experiment were separated by @xmath152 cm .
the pendulum hangs from a supporting frame above the pole pieces and is made from a solid rod of length 39.5 cm to which different paddles of interest can be affixed with screws .
four plates were studied , each of side 11.43 cm by 11.43 cm with different arrangements of holes or slits as shown earlier in figure [ fig : schematic](d ) .
students collected videos of the motion of the pendulum using a cellphone camera , and in a later iteration a digital camera was used to achieve higher image quality and frame rate .
although data logging hardware could also be used where available , the video approach was adopted as it is inexpensive , required no additional apparatus beyond the existing demonstration and is easily reproduced by others .
a white piece of paper with a scale bar was attached to the pole piece behind the camera to provide a uniform background and facilitate quantitative analysis of the resulting videos .
the camera was positioned so that the lens was aligned with the rod at rest ; its position was marked with chalk so that it could be easily replaced after each trial .
the plates were released from the same point outside of the magnet , determined by a ruler placed on top of the pole pieces .
the plates were released when the camera started recording and data was collected until the pendulum came to a rest .
two sample frames from one of the movies are shown in figure [ fig : reslice](a ) .
for each plate , 5 videos were collected with @xmath153 , @xmath154 , @xmath155 , @xmath156 and @xmath157 current flowing through the circuit respectively .
selected videos were taken more than once to verify repeatability .
these values obviously depend on the details of the circuit , but were chosen as follows : for the lower bound , a value was sought such that there was a visible difference between the motion of the plate with weakest damping ( the one with 4 slits ) in the on and off states . for the upper bound , we found a current level such that the plate with the strongest damping gave 2 - 3 oscillations before coming to a stop , i.e. we are in the oscillatory regime of @xmath144 in equation ( [ eq : generalequationofmotion ] ) .
( a ) sample frames from a video of the pendulum s motion . from this movie ,
the row of pixels indicated with the dashed line is taken from each frame and composited to produce the resliced image ( b ) . ] having collected the dataset of videos of the pendulum with different plates , the trajectories were extracted from the videos .
this was done in two steps : first , the open - source image processing program ` imagej ` @xcite was used to `` reslice '' the movie , converting it from a stack of images of the @xmath158 plane at different time points @xmath150 to a new stack of images of @xmath159 at different vertical positions @xmath160 . an sample output
is displayed in fig .
[ fig : reslice](b ) , showing the damped motion of the pendulum as a function of time for the row of the movie indicated by the dashed line in fig .
[ fig : reslice](a ) .
the second step in the data extraction was performed by a custom program written in _
, included as supplementary material .
this first subtracted the background , then for each time point located the position of the bob by finding the centroid of the pixels with intensity above a certain value .
the result of this process was to yield for each movie a list of @xmath159 representing the trajectory of the pendulum . using the rulers and
known frame rate of the camera ( 30fps ) , the list was calibrated to physical coordinates ; the program also subtracts the equilibrium position of the pendulum from the @xmath21 position . to facilitate comparison with the models ,
the extremum - finding routine mentioned in the previous section was used to identify a list of stationary points @xmath161 from each trajectory .
a constant was subtracted from the times such that for each trajectory the first extremum was set to @xmath162 .
having obtained a set of experimental trajectories , we now fit both models to the data . beginning with the simplified wire model , the analytical solution in section [ sub : effect - on - the ]
implies that for each individual trajectory the extrema should follow a straight line on semi - log scales with slope @xmath163 from the set of results displayed as points in fig .
[ fig : fit ] , it is seen that this is generally the case , though some deviations are apparent . in particular , the motion starts to decay more rapidly once @xmath164 ; we therefore excluded this data from our fit . for each plate ,
several values of current @xmath165 were used to take a succession of trajectories . although it is obvious that @xmath166 , we did not explicitly measure the magnetic field and therefore absorb this unknown constant of proportionality together with @xmath7 the effective resistivity of the wire into a single fitting parameter @xmath167 .
this fitting parameter could , of course , in principle be calculated or measured separately , but the purpose of fitting here is to establish whether the model is consistent with the results for all plates simultaneously .
the slopes extracted from the ensemble of plates with different @xmath168 ( calculated values are shown in table [ tab : fitsimplified ] ) at different applied currents @xmath165 should therefore universally be described by , @xmath169 estimates for the initial position of each trajectory @xmath141 were obtained by fitting a straight line to the extrema individually on semi - log scales .
we then estimated @xmath170 for each plate from the @xmath171 trajectory for each plate ; the results are shown in table [ tab : fitsimplified ] .
two pairs of plates seem to cluster around similar values ; these were collected at different times indicating that the natural damping of the apparatus , perhaps lubricant in the hinge , changed between them .
.[tab : fitsimplified]geometric parameters and fitted @xmath167 , @xmath170 for each plate using the simplified model . [ cols="^,^,^,^",options="header " , ] results from the universal least - squares one - parameter fit for @xmath167 are shown in fig .
[ fig : fit](a ) . as can be seen ,
the fit is nt bad and follows many of the trajectories quite well .
that said , there are many discrepancies .
we therefore estimated @xmath167 from each individual plate using the data at different applied field . while the details of these are not shown here , the quality of fit was very much better .
the resultant values of @xmath167 are displayed in table [ tab : fitsimplified ] .
in particular , it is not surprising that the plate with 4 holes is such an outlier as the wire model can not account for this geometry .
values for the other plates are the same order of magnitude , which together with fig .
[ fig : fit](a ) demonstrates that the simplified model gives reasonable estimates for the influence of plate geometry even though it is based on quite egregious approximations . in the same spirit
, we assume that the damping term in ( [ eq : generalequationofmotion ] ) for the magneto - quasistatic model can similarly be described by a universal constant @xmath167 , @xmath172 for this model , as well as @xmath141 , @xmath170 , @xmath167 , @xmath173 and @xmath165 , it is also necessary to specify the natural angular frequency @xmath174 ; the value from the @xmath171 trajectory for each plate is used since @xmath175 for the weak damping in the absence of the magnetic field and hence @xmath176 . the results of the one - parameter universal fit for @xmath167 are displayed in [ fig : fit](b ) . the quality of fit is significantly better , especially for large amplitudes and early times , and more consistent between plates than for the simple model , implying that the magneto - quasistatic model yields good predictions of the effect of plate geometry .
of particular importance is the fact that the fit for the 4 hole plate is no worse than that of the other plates , suggesting that these features have been handled correctly by the model .
nonetheless , there remains an unexplained phenomenon in the data , i.e the anomalous increase in damping once the amplitude of the motion drops below @xmath177 .
this is not predicted by either model .
the simplified model predicts a constant rate of damping , while the magneto - quasistatic model predicts a crossover from magnetically dominated damping at large amplitudes to stokes dominated damping at small amplitudes .
we tried adjusting the magneto - quasistatic model in various waysoffsetting the position of the plate with respect to the pole pieces and changing the size of the plate relative to the pole piecesbut found no similar effect for any reasonable parameters for the experiment .
clearly , the stokes form of the non - magnetic damping is most likely at fault , because the source of damping is likely to be primarily the hinge rather than air resistance ; further analysis should be performed to determine the origin of this anomaly .
the scientific results presented in this paper were obtained in part by students in spring 2013 as part of the electromagnetic theory ii class at tufts university , the second course in the e&m sequence in the graduate program at tufts ; two students from the class are co - authors on the paper ( cw and jw ) having spent a considerable time subsequently acquiring additional data and performing calculations . in this section , the pedagogical context of these activities is briefly documented in the hope that it is of use to others .
the idea of incorporating ( simple ) experiments in class was motivated by an observation that key scientific skills expected of graduate studentsconnecting theory and experiment ; article reading ; journal selection and article preparationare not included explicitly in the curriculum .
the project was spread over the second half of the semester as follows : 1 . a prepatory homework where students derived the magneto - quasistatic model [ sec : magneto - quasistatic - model ] .
two @xmath178 hour class periods where students worked in groups .
the students were divided into groups of four and worked on : i ) conformal mapping analysis , ii ) finite element analysis and iii ) experiments .
the instructor provided scaffolding activities to help students learn flexpde and the schwarz - christoffel mapping technique .
these activities continued subsequently outside of class .
3 . a reading activity whereby students had to identify possible journals for publication of this work ;
this was followed by a 15 minute discussion of journal selection in class .
4 . a just - in - time teaching@xcite - like pre - class activity in which students had to identify the structural elements in a typical paper and construct an outline ;
these were passed round anonymously for peer feedback .
a homework activity in which students prepared an introduction for the paper .
students were asked to read each other s work and identify strong features as well as possible modifications as a pre - class activity .
this was followed by a discussion in class .
each group collaboratively produced a draft section for the final paper as part of a homework ; these were combined by tja into a coherent document and extensively rewritten during the second iteration of experiments .
the focus of the class and restricted time available necessitated some design trade - offs .
for example , we chose to use an existing commercial software package , ` flexpde ` , as opposed to implementing a custom finite element program , for several reasons : first , the scale and complexity of problems solvable with a specialist code is much greater than with a simple custom - written program and hence should be applicable to situations encountered in student s research projects .
second , the program contains advanced features such as adaptive refinementallocation of grid points based on estimated errorthat are time - consuming to implement .
third , by removing the focus of the exercise from programmatic implementation , time can be spent on understanding how to properly interpret the output using fundamental principles of numerical analysis such as order , error estimation and convergence .
while the project - based approach is well - grounded in a theory of learning ( constructivism ) , due to the small @xmath179 size of the class , it is difficult to rigorously demonstrate the effectiveness of the activity .
future iterations would strongly benefit from pre and post testing .
even so , the outcomes of the project indicate that students made considerable progress towards learning these challenging scientific skills . moreover , the class was very highly rated in the formal feedback mechanism .
it is hoped that , in future , physics education researchers might attempt a thorough analysis of this and related approaches at the graduate level .
the present work has provided a careful , quantitative analysis of a familiar classroom demonstration , the van waltenhofen pendulum .
a magneto - quasistatic model has been shown to successfully predict the relative damping of a selection of plates with different geometry using conformal mapping and finite elements as tools to carry out the necessary calculations . as a by - product , a number of valuable resources for more general use , including plots of the current distribution during the course of the motion of the pendulum , have been produced .
additionally , a greatly simplified model , suitable for use in introductory physics classes , has been formulated that could be used to make the classroom demonstration more quantitative .
_ the authors wish to thank students from the tufts electromagnetic theory ii class in spring 2013 , the tufts department of physics & astronomy for providing equipment for the project , and badel mbanga and chris burke for helpful discussions .
the authors contributed to the paper as follows : cw , jw and pw obtained the experimental data used in the paper ; conformal mapping solutions were obtained by jw and tja ; finite element simulations were performed by tja based on initial results obtained by students in the class ; the paper was written by tja from sections submitted by each group and revised collectively . _
13ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1098/rstl.1825.0023 [ * * , ( ) ] link:\doibase 10.1098/rstl.1832.0006 [ * * , ( ) ] @noop ( ) @noop ( ) @noop * * , ( ) @noop _ _ ( , ) @noop _ _ ( , ) @noop _ _ , cambridge monographs on applied and computational mathematics ( , ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ , ed .
( , , ) | we quantitatively analyze a familiar classroom demonstration , van waltenhofen s eddy current pendulum , to predict the damping effect for a variety of plate geometries from first principles .
results from conformal mapping , finite element simulations and a simplified model suitable for introductory classes are compared with experiments . |
savchuk and schmdgen @xcite introduce a method to define and classify the integrable representations of certain * -algebras by an inductive construction .
the original goal of this article was to clarify this method and thus make it apply to more situations .
this has led me to reconsider some foundational aspects of the theory of representations of * -algebras by unbounded operators .
this is best explained by formulating an induction theorem that is inspired by @xcite .
let @xmath0 be a discrete group with unit element @xmath1 .
let @xmath2 be a @xmath0-graded unital * -algebra .
that is , @xmath3 , @xmath4 , and @xmath5 .
in particular , the _ unit fibre _ @xmath6 is a unital * -algebra .
many interesting examples of this situation are studied in @xcite .
a _ fell bundle _ over @xmath0 is a family of subspaces @xmath7 of a c*-algebra @xmath8 ( which is not part of the data ) such that @xmath9 and @xmath10 .
the universal choice for @xmath8 is the _ section c*-algebra _ of the fell bundle .
briefly , our main result says the following .
let @xmath11 be a c*-algebra such that `` integrable '' `` representations '' of @xmath6 are `` equivalent '' to `` representations '' of @xmath11 .
under some technical conditions , we construct a fell bundle @xmath12 over @xmath0 such that `` integrable '' `` representations '' of @xmath13 are `` equivalent '' to `` representations '' of the section c*-algebra of this fell bundle . here the words in quotation marks must be interpreted carefully to make this true .
a _ representation _ of a * -algebra @xmath13 on a hilbert @xmath14-module @xmath15 is an algebra homomorphism @xmath16 from @xmath13 to the algebra of @xmath14-module endomorphisms of a dense @xmath14-submodule @xmath17 with @xmath18 for all @xmath19 , @xmath20 .
the representation induces a graph topology on @xmath21 .
we restrict to _ closed _
representations most of the time , that is , we require @xmath21 to be complete in the graph topology .
the difference from usual practice is that we consider representations on _ hilbert modules _ over c*-algebras . a _ representation of a c*-algebra _ @xmath8 on a hilbert module @xmath15 is a nondegenerate * -homomorphism @xmath22 , where @xmath23 denotes the c*-algebra of adjointable operators on @xmath15 .
the notion of `` integrability '' for representations is a choice .
the class of all hilbert space representations of a * -algebra may be quite wild .
hence it is customary to limit the study to some class of `` nice '' or `` integrable '' representations .
for instance , for the universal enveloping algebra of the lie algebra of a lie group @xmath0 , we may call those representations `` integrable '' that come from a unitary representation of @xmath0 .
this example suggests the name `` integrable '' representations . in our theorem , a notion of integrability for representations of @xmath24 on all hilbert modules over all c*-algebras is fixed .
a representation of @xmath13 is called integrable if its restriction to @xmath6 is integrable .
the induction theorem describes the integrable representations of @xmath13 in terms of integrable representations of @xmath6 .
for instance , if @xmath6 is finitely generated and commutative , then we may call a representation @xmath16 on a hilbert module integrable if the closure @xmath25 is a regular , self - adjoint operator for each @xmath26 with @xmath27 ; then our theory reproduces the examples in @xcite .
an `` equivalence '' between the integrable representations of a unital * -algebra @xmath13 and the representations of a c*-algebra @xmath8 is a family of bijections one for each hilbert module @xmath15 over each c*-algebra @xmath14 between the sets of integrable representations of @xmath13 and of representations of @xmath8 on @xmath15 with two extra properties , namely , compatibility with isometric intertwiners and interior tensor products ; these properties require some more definitions .
first , an _ isometric intertwiner _ between two representations is a hilbert module isometry not necessarily adjointable between the underlying hilbert modules that restricts to a left module map between the domains of the representations . for an equivalence between integrable representations of @xmath13 and representations of @xmath8 we require an isometry to intertwine two representations of @xmath8 if and only if it intertwines the corresponding integrable representations of @xmath13 .
secondly , a _ c*-correspondence _ from @xmath28 to @xmath29 is a hilbert @xmath29-module @xmath30 with a representation of @xmath28 .
given such a correspondence and a hilbert @xmath28-module @xmath15 , the interior tensor product @xmath31 is a hilbert @xmath29-module .
a representation of @xmath13 or @xmath8 on @xmath15 induces a representation on @xmath31 .
we require our bijections between integrable representations of @xmath13 and representations of @xmath8 to be compatible with this interior tensor product construction on representations .
we call @xmath8 a
_ c*-hull for the integrable representations of @xmath13 _ , if the integrable representations of @xmath13 are equivalent to the representations of @xmath8 as explained above , that is , through a family of bijections compatible with isometric intertwiners and interior tensor products . the induction theorem builds a c*-hull for the integrable representations of @xmath13 using a c*-hull for the integrable representations of @xmath6 and assuming a further mild technical condition , which we explain below .
many results of the general theory remain true if we only require the equivalence of representations to be compatible with interior tensor products and unitary * -intertwiners , that is , isomorphisms of representations ; we speak of a _
weak c*-hull _ in this case .
the induction theorem , however , fails for weak c*-hulls .
we show this by a counterexample .
some results only need the class of integrable representations to have some properties that are clearly necessary for the existence of a c*-hull or weak c*-hull , but they do not need the ( weak ) c*-hull itself .
this is formalised in our notions of _ admissible _ and _ weakly admissible _ classes of representations .
for example , let @xmath13 be commutative and define integrable representations as above .
let @xmath32 be the space of characters of @xmath13 with the topology of pointwise convergence .
if @xmath32 is locally compact and @xmath13 is countably generated , then @xmath33 is a c*-hull for the integrable representations of @xmath13 as defined above , that is , those representations where each @xmath25 for @xmath20 with @xmath27 is regular and self - adjoint .
if , say , @xmath34 $ ] with @xmath35 , then the c*-hull is @xmath36 . here
the equivalence of representations maps a representation @xmath16 of @xmath37 $ ] to the functional calculus homomorphism for the regular , self - adjoint operator @xmath38 .
if @xmath32 is not locally compact , then the integrable representations of @xmath13 defined above still form an admissible class , but they have no c*-hull . if , say , @xmath13 is the algebra of polynomials in countably many variables , then @xmath39 , which is not locally compact .
the problem of associating c*-algebras to this * -algebra has recently been studied by grundling and neeb @xcite . from our point of view , this amounts to choosing a smaller class of `` integrable '' representations that does admit a c*-hull .
we have now explained the terms in quotation marks in our induction theorem and how we approach the representation theory of * -algebras .
most previous work focused either on representations on hilbert spaces or on single unbounded operators on hilbert modules .
hilbert module representations occur both in the assumptions and in the conclusions of the induction theorem , and hence we can not prove it without considering representations on hilbert modules throughout .
in addition , taking into account hilbert modules makes our c*-hulls unique .
besides the induction theorem , the other main strand of this article are local global principles , which aim at reducing the study of integrability for representations on general hilbert modules to representations on hilbert space .
given a state @xmath40 on the coefficient c*-algebra @xmath14 of a hilbert module @xmath15 , we may complete @xmath15 to a hilbert space using @xmath40 .
thus a representation of @xmath13 on @xmath15 induces hilbert space representations for all states on @xmath14 .
the _ local
global principle _ says that a representation of @xmath13 on @xmath15 is integrable if and only if these induced hilbert space representations are integrable for all states ; the _ strong _ local global principle says the same with all states replaced by all _ pure _ states .
we took these names from @xcite .
results of pierrot @xcite show that the strong local
global principle holds for any class of integrable representations that is defined by certain types of conditions , such as the regularity and self - adjointness of @xmath25 for some @xmath20 with @xmath27 . for instance , this covers the integrable representations of commutative * -algebras and of universal enveloping algebras of lie algebras . in our work ,
affiliation of multipliers or regularity of hilbert module operators play a role as standard ways to characterise which representations of a particular * -algebra @xmath13 are integrable . in all examples that we treat ,
the regularity of some special elements of @xmath13 is part of the definition of its integrable representations .
but some elements of @xmath13 may well act by irregular operators in some integrable representations if @xmath13 is not commutative .
thus affiliation and regularity are important to study the integrable representations in concrete examples , but can not play a foundational role for the representation theory of _ general _ * -algebras .
if @xmath8 is generated in the sense of woronowicz @xcite by some self - adjoint , affiliated multipliers that belong to @xmath13 , then it is a c*-hull and the strong local
global principle holds ( see theorem [ the : woronowicz_local - global ] ) . a counterexample shows that this theorem breaks down if the generating affiliated multipliers are not self - adjoint : both the local global principle and compatibility with isometric intertwiner fail in the counterexample .
so regularity without self - adjointness seems to be too weak for many purposes .
the combination of regularity and self - adjointness is an easier notion than regularity alone .
a closed operator @xmath41 is regular and self - adjoint if and only if @xmath42 is surjective for all @xmath43 , if and only if the cayley transform of @xmath41 is unitary , if and only if @xmath41 has a functional calculus homomorphism on @xmath36 .
now we describe the fell bundle in the induction theorem and , along the way , the further condition besides compatibility with isometric intertwiners that it needs .
our input data is a graded * -algebra @xmath2 and a c*-hull @xmath11 for the * -algebra @xmath6 .
a representation of @xmath13 is integrable if its restriction to @xmath6 is integrable .
we seek a c*-hull for the integrable representations of @xmath13 . as in @xcite
, we induce representations from @xmath6 to @xmath13 , and this requires a positivity condition .
we call representations of @xmath6 that may be induced to @xmath13 _ inducible_. we describe a quotient c*-algebra @xmath44 of @xmath11 that is a c*-hull for the inducible , integrable representations of @xmath13 .
it is the unit fibre of our fell bundle .
if a representation @xmath16 of @xmath13 is integrable , then its restriction to @xmath6 is integrable and inducible .
thus it corresponds to a representation @xmath45 of @xmath44 .
the identity correspondence on @xmath44 corresponds to a particular ( `` universal '' ) inducible , integrable representation of @xmath6 on @xmath44 .
its domain is a dense right ideal @xmath46 .
the operators @xmath47 on @xmath15 for @xmath48 , @xmath49 are adjointable .
their closed linear span is the fibre @xmath50 of our fell bundle at @xmath51 provided @xmath52 is faithful .
the most difficult point is to prove @xmath53 for all @xmath54 ; this easily implies @xmath55 and @xmath56 , so that the subspaces @xmath57 form a fell bundle . to prove @xmath58 , we need compatibility with isometric intertwiners and that induction maps inducible , integrable representations of @xmath6 to _ integrable _ representations of @xmath13 .
two counterexamples show that both assumptions are necessary for the induction theorem
. fell bundles are noncommutative partial dynamical systems .
more precisely , a fell bundle @xmath12 over @xmath0 is equivalent to an action of @xmath0 on @xmath44 by partial morita rieffel equivalences ; this is made precise in @xcite . in the examples in @xcite ,
the group @xmath0 is almost always @xmath59 ; the c*-algebras @xmath11 and hence @xmath44 are commutative ; and the resulting fell bundle comes from a partial action of @xmath0 on the spectrum @xmath60 . in these examples , the section c*-algebra is a partial crossed product .
this may also be viewed as the groupoid c*-algebra of the transformation groupoid for the partial action of @xmath0 on @xmath60 .
we show that the c*-hull @xmath8 for the integrable representations of @xmath13 is a _ twisted _ groupoid c*-algebra of this transformation groupoid whenever @xmath11 is commutative .
we give some criteria when the twist is absent , and examples where the twist occurs .
one way to insert such twists is by rieffel deformation , using a @xmath61-cocycle on the group @xmath0 .
we show that rieffel deformation is compatible with the construction of c*-hulls .
we describe commutative and noncommutative c*-hulls for the polynomial algebra @xmath37 $ ] in [ sec : polynomials1 ] and [ sec : polynomials2 ] ; the noncommutative c*-hulls for @xmath37 $ ] make very good counterexamples .
we classify and study commutative c*-hulls in [ sec : commutative_hulls ] .
many results about them generalise easily to _ locally bounded _ representations .
roughly speaking , these are representations where the vectors on which the representation acts by bounded operators form a core . the only examples of * -algebras for which we treat locally bounded representations in some detail are commutative ones . through the induction theorem , the representation theory of commutative * -algebras is important even for noncommutative algebras because they often admit a grading by some group with commutative unit fibre .
many examples of this are treated in detail in @xcite .
we discuss untwisted and twisted weyl algebras in finitely and infinitely many generators in [ sec : weyl_twisted ] .
the twists involved are rieffel deformations .
since these examples have commutative unit fibres , the resulting c*-hulls are twisted groupoid c*-algebras .
as it turns out , all twists of the relevant groupoids are trivial , so that the twists do not change the representation theory of the weyl algebras up to equivalence .
i am grateful to yuriy savchuk for several discussions , which led me to pursue this project and eliminated mistakes from preliminary versions of this article .
let @xmath13 be a unital * -algebra , @xmath14 a c*-algebra , and @xmath15 a hilbert @xmath14-module .
our convention is that inner products on hilbert spaces and hilbert modules are linear in the second and conjugate - linear in the first variable .
[ def : rep_hilbert_module ] a _ representation _ of @xmath13 on @xmath15 is a pair @xmath62 , where @xmath63 is a dense @xmath14-submodule and @xmath64 is a unital algebra homomorphism to the algebra of @xmath14-module endomorphisms of @xmath21 , such that @xmath65 we call @xmath21 the _ domain _ of the representation .
we may drop @xmath16 from our notation by saying that @xmath21 is an @xmath66-bimodule with the right module structure inherited from @xmath15 , or we may drop @xmath21 because it is the common domain of the partial linear maps @xmath67 on @xmath15 for all @xmath20 .
we equip @xmath21 with the _ graph topology _ , which is generated by the seminorms @xmath68 for @xmath20 .
the representation is _ closed _ if @xmath21 is complete in this topology .
a _ core _ for @xmath62 is an @xmath66-subbimodule of @xmath21 that is dense in @xmath21 in the graph topology .
definition [ def : rep_hilbert_module ] contains the usual definition of a representation of a * -algebra on a hilbert space by unbounded operators for @xmath69 .
this situation has been studied extensively ( see , for instance , @xcite ) . for @xmath70 with the canonical hilbert @xmath14-module structure , we get representations of @xmath13 by _ densely defined unbounded multipliers_. the domain of such a representation is a dense right ideal @xmath71 . this situation is a special case of the `` compatible pairs '' defined by schmdgen @xcite
. given two seminorms @xmath72 , we write @xmath73 if there is a scalar @xmath74 with @xmath75 .
[ lem : graph_norms_directed ] the set of graph norms partially ordered by @xmath76 is directed : for all @xmath77 there are @xmath78 and @xmath79 so that @xmath80 for any representation @xmath62 , any @xmath81 , and @xmath82 .
let @xmath83 .
the following computation implies @xmath84 : @xmath85 [ def : regular ] a densely defined operator @xmath86 on a hilbert module @xmath15 is _ semiregular _ if its adjoint is also densely defined ; it is _ regular _ if it is closed , semiregular and @xmath87 has dense range .
an _ affiliated multiplier _ of a c*-algebra @xmath14 is a regular operator on @xmath14 viewed as a hilbert @xmath14-module .
the closability assumption in @xcite*definition 2.1.(ii ) is redundant by @xcite*lemma 2.1 .
regularity was introduced by baaj and julg @xcite , affiliation by woronowicz @xcite .
[ rem : affiliated ] let @xmath62 be a representation of @xmath13 on @xmath15 and let @xmath20 .
the operator @xmath67 is automatically semiregular because @xmath88 is defined on @xmath21 .
the closure @xmath89 of @xmath67 need not be regular .
the regularity of @xmath25 for _ some _ @xmath20 is often assumed in the definition of _ integrable _ representations . for non - commutative @xmath13
, we should expect that @xmath89 is irregular for some @xmath20 even if @xmath16 is integrable .
for instance , a remark after corollaire 1.27 in @xcite says that this happens for certain symmetric elements in the universal enveloping algebra @xmath90 for a simply connected lie group @xmath0 : they act by irregular operators in certain representations that integrate to unitary representations of @xmath0 .
the usual norm on @xmath15 is the graph norm for @xmath91 .
hence the inclusion map @xmath92 is continuous for the graph topology on @xmath21 and extends continuously to the completion @xmath93 of @xmath21 in the graph topology .
[ pro : closure_rep ] the canonical map @xmath94 is injective and onto @xmath95 thus @xmath62 is closed if and only if @xmath96 .
each @xmath67 extends uniquely to a continuous operator @xmath25 on @xmath93 , and this defines a closed representation @xmath97 of @xmath13 , called the _ closure _ of @xmath62 .
the operator @xmath67 for @xmath20 is semiregular and hence closable by @xcite*lemma 2.1 .
equivalently , the canonical map from the completion of @xmath21 in the graph norm for @xmath98 to @xmath15 is injective .
its image is @xmath99 , the domain of the closure of @xmath67 .
the graph norms for @xmath20 form a directed set that defines the graph topology on @xmath21 .
so the completion of @xmath21 in the graph topology is the projective limit of the graph norms completions for @xmath20 .
since each of these graph norm completions embeds into @xmath15 , the projective limit in question is just an intersection in @xmath15 , giving . in the hilbert space setting , this is @xcite*proposition 2.2.12 .
the operators @xmath100 for @xmath20 are continuous in the graph topology .
thus they extend uniquely to continuous linear operators @xmath101
. these are again @xmath14-linear and the map @xmath102 is linear and multiplicative because extending operators to a completion is additive and functorial .
the set of @xmath103 with @xmath104 for all @xmath20 is closed in the graph topology and contains @xmath105 , which is dense in @xmath106 .
hence this equation holds for all @xmath107 and so @xmath97 is a representation of @xmath13 on @xmath15 .
the graph topology on @xmath93 for @xmath102 extends the graph topology on @xmath21 for @xmath16 and hence is complete .
so @xmath97 is a closed representation .
we shall need a generalisation of that replaces @xmath13 by a sufficiently large subset .
[ def : strong_generating_set ] a subset @xmath108 is called a _
strong generating _ set if it generates @xmath13 as an algebra and the graph norms for @xmath109 generate the graph topology in any representation .
that is , for every element @xmath20 there are @xmath110 in @xmath111 and @xmath112 so that @xmath113 for any representation on a hilbert module and any vector @xmath114 in its domain .
an estimate @xmath115 for any @xmath114 in the domain of any representation of @xmath13 is usually shown by proving that there are @xmath116 with @xmath117 , compare the proof of lemma [ lem : graph_norms_directed ] . [
exa : strong_generating ] let @xmath118 be the set of _ symmetric _ elements . call an element of @xmath13 _ positive _ if it is a sum of squares of elements of @xmath119 . the positive elements and , _ a fortiori _
, the symmetric elements form strong generating sets for @xmath13 .
any element is of the form @xmath120 with @xmath121 , and @xmath122 for @xmath123 .
thus the positive elements generate @xmath13 as an algebra .
the graph norms for positive elements generate the graph topology by the proof of lemma [ lem : graph_norms_directed ] . [
pro : equality_if_closure_equal ] let @xmath108 be a strong generating set .
two closed representations @xmath124 and @xmath125 of @xmath13 on the same hilbert module @xmath15 are equal if and only if @xmath126 for all @xmath109 .
let @xmath127 for @xmath128 .
the completion of @xmath21 for the graph norm of @xmath98 is @xmath99 , compare the proof of proposition [ pro : closure_rep ] .
hence the completion of @xmath21 in the sum of graph norms @xmath129 for @xmath112 is @xmath130 .
these sums of graph norms for @xmath112 form a directed set that generates the graph topology on @xmath21 .
hence @xmath131 for any closed representation , compare the proof of .
so @xmath132 and @xmath133 have the same domain .
moreover , @xmath134 for all @xmath109 .
since @xmath135 generates @xmath13 as an algebra and @xmath136 , this implies @xmath137 .
[ cor : bounded_rep ] let @xmath135 be a strong generating set of @xmath13 and let @xmath62 be a closed representation of @xmath13 with @xmath138 for each @xmath109
. then @xmath139 and @xmath16 is a * -homomorphism to the c*-algebra @xmath23 of adjointable operators on @xmath15 .
equation gives @xmath139 . since @xmath140 and @xmath141 is defined everywhere , it is adjoint to @xmath67 .
so @xmath142 and @xmath16 is a * -homomorphism to @xmath23 .
[ lem : cstar - rep_bounded ] let @xmath13 be a unital c*-algebra . any closed representation of @xmath13 on @xmath15 has domain @xmath139 and is a unital * -homomorphism to @xmath23 .
let @xmath20 .
there are a positive scalar @xmath143 and @xmath78 with @xmath144 ; say , take @xmath145 and @xmath146 .
then @xmath147 for all @xmath81 .
thus the graph topology on @xmath21 is equivalent to the norm topology on @xmath15 .
hence @xmath148 for any closed representation .
[ def : isometric_intertwiner ] let @xmath124 and @xmath125 be representations on hilbert @xmath14-modules @xmath149 and @xmath150 , respectively .
an _ isometric intertwiner _ between them is an isometry of hilbert @xmath14-modules @xmath151 with @xmath152 and @xmath153 for all @xmath154 , @xmath155 ; equivalently , @xmath156 for all @xmath20 , that is , the graph of @xmath157 is contained in the graph of @xmath158 .
we neither ask @xmath159 to be adjointable nor @xmath160 . let @xmath161 be the category with closed representations of @xmath13 on hilbert @xmath14-modules as objects , isometric intertwiners as arrows , and the usual composition .
the unit arrow on @xmath62 is the identity operator on @xmath15 .
[ lem : closure_functorial ] let @xmath124 and @xmath125 be representations on hilbert @xmath14-modules @xmath149 and @xmath150 , respectively , and let @xmath162 be an isometric intertwiner . then @xmath159 is also an intertwiner between the closures of @xmath124 and @xmath125 . since @xmath159 intertwines the representations , it is continuous for the graph topologies on @xmath163 and @xmath164
. hence @xmath159 maps the domain of the closure @xmath165 into the domain of @xmath166 .
this extension is still an intertwiner because it is an intertwiner on a dense subspace .
[ pro : intertwiner_strong_generators ] let @xmath124 and @xmath125 be closed representations of @xmath13 on hilbert @xmath14-modules @xmath149 and @xmath150 , respectively .
let @xmath108 be a strong generating set .
an isometry @xmath167 is an intertwiner from @xmath124 to @xmath125 if and only if @xmath168 for all @xmath109 .
first let @xmath159 satisfy @xmath169 for all @xmath109 .
then @xmath159 maps the domain of @xmath170 into the domain of @xmath171 for each @xmath109 .
now implies @xmath152 . since @xmath172
, we get @xmath173 for all @xmath109 , @xmath155 .
since @xmath135 generates @xmath13 as an algebra and @xmath174 , this implies @xmath175 for all @xmath20 , that is , @xmath159 is an intertwiner .
conversely , assume that @xmath159 is an intertwiner from @xmath124 to @xmath125 .
equivalently , @xmath176 for all @xmath20 . then @xmath177
because @xmath159 is an isometry , and @xmath178 .
thus @xmath179 for all @xmath20 .
now we relate the categories @xmath161 for different c*-algebras @xmath14 .
[ def : cstar - correspondence ] let @xmath28 and @xmath29 be two c*-algebras .
a _ c*-correspondence _ from @xmath28 to @xmath29 is a hilbert @xmath29-module with a representation of @xmath28 by adjointable operators ( representations of c*-algebras are tacitly assumed nondegenerate ) .
isometric intertwiner _ between two correspondences from @xmath28 to @xmath29 is an isometric map on the underlying hilbert @xmath29-modules that intertwines the left @xmath28-actions .
let @xmath180 denote the category of correspondences from @xmath28 to @xmath29 with isometric intertwiners as arrows and the usual composition . by lemma [ lem : cstar - rep_bounded ] ,
our two definitions of @xmath161 for unital * -algebras and c*-algebras coincide if @xmath13 is a unital c*-algebra .
so our notation is not ambiguous .
there is no need to define representations of a non - unital * -algebra @xmath13 because we may adjoin a unit formally .
a representation of @xmath13 extends uniquely to a representation of the unitisation @xmath181 .
thus the nondegenerate representations of @xmath13 are contained in @xmath182 .
to get rid of degenerate representations , we may require nondegeneracy on @xmath13 when defining the integrable representations of @xmath181 , compare example [ exa : nondegeneracy_condition ] .
let @xmath15 be a hilbert @xmath28-module and @xmath30 a correspondence from @xmath28 to @xmath29 .
the interior tensor product @xmath31 is the completion of the algebraic tensor product @xmath183 to a hilbert @xmath29-module , using the inner product @xmath184 see the discussion around @xcite*proposition 4.5 for more details .
we may use the balanced tensor product @xmath185 instead of @xmath183 because the inner product descends to this quotient .
if we want to emphasise the left action @xmath186 that is part of the c*-correspondence @xmath30 , we write @xmath187 for @xmath31 . in addition , let @xmath62 be a closed representation of @xmath13 on @xmath15 .
we are going to build a closed representation @xmath188 of @xmath13 on @xmath189 .
first let @xmath190 be the image of @xmath191 or @xmath192 under the canonical map to @xmath189 .
[ lem : tensor_rep_with_corr ] for @xmath20 , there is a unique linear operator @xmath193 with @xmath194 for all @xmath81 , @xmath195 .
the map @xmath196 is a representation of @xmath13 with domain @xmath197 .
write @xmath198 as images of elements of @xmath199 : @xmath200 with @xmath201 , @xmath202 for @xmath203 , @xmath204
. then @xmath205 an element @xmath206 is determined uniquely by its inner products @xmath207 for all @xmath208 because @xmath197 is dense in @xmath189 .
the right hand side in does not depend on how we decomposed @xmath40 .
hence @xmath209 well defines an operator @xmath210 .
this is a right @xmath29-module map , and @xmath211 is linear and multiplicative because @xmath16 is .
equation says that @xmath212 for all @xmath198 .
thus @xmath213 is a representation .
let @xmath188 be the closure of the representation on @xmath189 defined in lemma [ lem : tensor_rep_with_corr ] .
[ lem : rep_tensor_corr_functor ] let @xmath167 be an isometric intertwiner between two representations @xmath124 and @xmath125 , and let @xmath214 be an isometric intertwiner of c*-correspondences .
then @xmath215 is an isometric intertwiner between @xmath216 and @xmath217 .
the isometry @xmath218 maps the image @xmath219 of @xmath220 to the image @xmath221 of @xmath222 and intertwines the operators @xmath223 on @xmath219 and @xmath224 on @xmath221 for all @xmath20 .
that is , it intertwines the representations defined in lemma [ lem : tensor_rep_with_corr ] .
it also intertwines the closures of @xmath219 and @xmath221 by lemma [ lem : closure_functorial ] .
the lemma gives a bifunctor @xmath225 the corresponding bifunctor @xmath226 for a c*-algebra @xmath8 is the usual composition of c*-correspondences .
this composition is associative up to canonical unitaries @xmath227 for all triples of composable c*-correspondences .
[ lem : tensor_associative ] if @xmath15 carries a representation @xmath62 of a * -algebra @xmath13 , then the unitary in is an intertwiner @xmath228 .
the bilinear map from @xmath229 to @xmath230 is separately continuous with respect to the graph topologies on @xmath21 and @xmath231 and the norm topology on @xmath30 .
since the image of @xmath232 in the hilbert module @xmath233 is dense in the norm topology , the image of @xmath234 in @xmath235 is a core for the representation @xmath236 .
since the image of @xmath192 in @xmath231 is dense in the graph topology , the image of @xmath234 in @xmath237 is a core for the representation @xmath238 .
the unitary in is an intertwiner between these cores , hence an intertwiner between the resulting closed representations by lemma [ lem : closure_functorial ] .
from now on , _ we tacitly assume representations to be closed_. proposition [ pro : closure_rep ] shows that this is no serious loss of generality .
let @xmath13 be a unital * -algebra .
we assume that a class of `` integrable '' ( closed ) representations of @xmath13 on hilbert modules is chosen .
let @xmath239 be the full subcategory with integrable representations on hilbert @xmath14-modules as objects .
being full means that the set of arrows between two integrable representations of @xmath13 is still the set of all isometric intertwiners .
we sometimes write @xmath240 and @xmath241 for the collection of all the categories @xmath242 and @xmath161 for all c*-algebras @xmath14 .
a c*-hull is a c*-algebra @xmath8 with natural isomorphisms @xmath243 for all c*-algebras @xmath14 .
more precisely : [ def : cstar - hull ] a _ c*-hull _ for the integrable representations of @xmath13 is a c*-algebra @xmath8 with a family of bijections @xmath244 from the set of representations of @xmath8 on @xmath15 to the set of integrable representations of @xmath13 on @xmath15 for all hilbert modules @xmath15 over all c*-algebra @xmath14 with the following properties : * _ compatibility with isometric intertwiners _ :
an isometry @xmath245 ( not necessarily adjointable ) is an intertwiner between two representations @xmath246 and @xmath247 of @xmath8 _ if and only if _ it is an intertwiner between @xmath248 and @xmath249 ; * _ compatibility with interior tensor products _
: if @xmath30 is a correspondence from @xmath28 to @xmath29 , @xmath15 is a hilbert @xmath28-module , and @xmath250 is a representation of @xmath8 on @xmath15 , then @xmath251 as representations of @xmath13 on @xmath31 .
the compatibility with isometric intertwiners means that the bijections @xmath252 for all @xmath15 with fixed @xmath14 form an isomorphism of categories @xmath253 which , in addition , does not change the underlying hilbert @xmath14-module .
the compatibility with interior tensor products expresses that these isomorphisms of categories for different @xmath14 are natural with respect to c*-correspondences .
[ def : star - intertwiner ] let @xmath124 and @xmath125 be two representations of @xmath13 on hilbert @xmath14-modules @xmath149 and @xmath150 .
an adjointable operator @xmath254 is an _ intertwiner _ if @xmath255 and @xmath256 for all @xmath20 , @xmath155 .
* -intertwiner _ if both @xmath257 and @xmath258 are intertwiners .
[ def : adjointable_intertwiner ] any adjointable intertwiner between two representations of a c*-algebra @xmath8 is a * -intertwiner .
in contrast , for a general * -algebra , even the adjoint of a unitary intertwiner @xmath259 fails to be an intertwiner if @xmath260 .
[ exa : friedrichs_extension ] by the friedrichs extension , any positive symmetric operator @xmath86 on a hilbert space @xmath261 extends to a positive self - adjoint operator @xmath262 on @xmath261 .
both @xmath86 and @xmath262 give closed representations @xmath16 and @xmath263 of the polynomial algebra @xmath37 $ ] on @xmath261 ( see [ sec : polynomials1 ] for more details ) .
the identity map on @xmath261 is a unitary intertwiner @xmath264 , but not a * -intertwiner unless @xmath265 .
the following proposition characterises when an adjointable isometry @xmath266 between two representations of @xmath13 on hilbert @xmath14-modules is a * -intertwiner .
since @xmath267 if @xmath159 is adjointable , we may as well assume that @xmath159 is the inclusion of a direct summand .
[ pro : isometry_star - intertwiner ] let @xmath149 and @xmath150 be hilbert modules over a c*-algebra @xmath14 and let @xmath124 and @xmath62 be representations of @xmath13 on @xmath149 and @xmath268 , respectively .
the following are equivalent : 1 .
[ enum : isometry_star - intertwiner1 ] the canonical inclusion @xmath269 is a * -intertwiner from @xmath132 to @xmath16 ; 2 .
[ enum : isometry_star - intertwiner2 ] the canonical inclusion @xmath269 is an intertwiner from @xmath132 to @xmath16 and @xmath270 ; 3 . [
enum : isometry_star - intertwiner3 ] there is a representation @xmath125 on @xmath150 such that @xmath271 .
we view @xmath149 and @xmath150 as subspaces of @xmath268 , so we may drop the isometry @xmath159 from our notation . the implication [ enum : isometry_star - intertwiner3]@xmath272[enum : isometry_star - intertwiner1 ] is trivial .
we are going to prove [ enum : isometry_star - intertwiner1]@xmath272[enum : isometry_star - intertwiner2]@xmath272[enum : isometry_star - intertwiner3 ] .
first assume that @xmath159 is a * -intertwiner .
then @xmath159 is an intertwiner .
in particular , @xmath273 . write @xmath81 as @xmath274 with @xmath275 , @xmath276 . since @xmath277 is an intertwiner , @xmath278
. hence @xmath279 .
thus [ enum : isometry_star - intertwiner1 ] implies [ enum : isometry_star - intertwiner2 ] . if [ enum : isometry_star - intertwiner2 ] holds , then @xmath273 is @xmath16-invariant and @xmath280 because @xmath159 is an intertwiner .
we claim that @xmath281 is @xmath16-invariant as well .
let @xmath282 and @xmath283 .
then @xmath284 because @xmath285 is orthogonal to @xmath150 .
since @xmath163 is dense in @xmath149 , this implies @xmath286 , and this implies our claim .
the condition [ enum : isometry_star - intertwiner2 ] implies @xmath287 as a vector space with @xmath288 because @xmath289 .
then @xmath164 is dense in @xmath150 because @xmath21 is dense in @xmath268 .
thus @xmath290 is a representation of @xmath13 on @xmath150 . and @xmath62 is the direct sum of @xmath124 and @xmath290 because @xmath291 .
the representation @xmath62 is tacitly assumed closed , and this implies that @xmath124 and @xmath290 are closed because @xmath292 for @xmath128 is closed in @xmath21 even in the norm topology and the graph topology for @xmath293 is the restriction to @xmath294 of the graph topology for @xmath16 .
thus [ enum : isometry_star - intertwiner2 ] implies [ enum : isometry_star - intertwiner3 ] .
[ def : weak_cstar - hull ] a _ weak c*-hull _ for the integrable representations of @xmath13 is a c*-algebra @xmath8 with a family of bijections @xmath252 between representations of @xmath8 and integrable representations of @xmath13 on hilbert modules that is compatible with unitary * -intertwiners and interior tensor products .
much of the general theory also works for weak c*-hulls .
but the induction theorem [ the : fell_bundle_sections_hull ] fails for weak c*-hulls , as shown by a counterexample in [ sec : counter_induction ] .
[ pro : adjointable_intertwiner ] let a class of integrable representations of @xmath13 have a weak c*-hull @xmath8 .
let @xmath124 and @xmath125 be integrable representations of @xmath13 on hilbert @xmath14-modules @xmath149 and @xmath150 , and let @xmath295 be the corresponding representations of @xmath8 on @xmath296 for @xmath128 .
an adjointable operator @xmath254 is a * -intertwiner from @xmath124 to @xmath125 if and only if it is an intertwiner from @xmath246 to @xmath247 . working with the direct sum representations on @xmath268 and the adjointable operator @xmath297
, we may assume without loss of generality that @xmath298 , @xmath299 , @xmath300 , and @xmath301 .
the adjointable intertwiners for the representation @xmath250 of @xmath8 form a c*-algebra @xmath302 : the commutant of @xmath8 in @xmath23 .
we claim that the * -intertwiners for the representation @xmath16 of @xmath13 also form a c*-algebra @xmath303 .
intertwiners and hence * -intertwiners form an algebra . thus @xmath303 is a * -algebra .
we show that it is closed .
let @xmath304 be a sequence of adjointable intertwiners for @xmath62 that converges in norm to @xmath305 .
let @xmath81 . then @xmath306 because each @xmath307 is an intertwiner . since @xmath308 is norm - convergent for each @xmath20 ,
the sequence @xmath309 is a cauchy sequence for the graph topology on @xmath21 .
since representations are tacitly assumed to be closed , this cauchy sequence converges in @xmath21 , so that @xmath310 .
moreover , @xmath311 for all @xmath20 , @xmath81 , so @xmath257 is again an intertwiner .
thus the algebra of intertwiners is norm - closed .
this implies that @xmath303 is a c*-algebra .
since the family of bijections @xmath312 is compatible with unitary * -intertwiners , a unitary operator on @xmath15 is a * -intertwiner for @xmath13 if and only if it is an intertwiner for @xmath8 .
that is , the unital c*-subalgebras @xmath313 contain the same unitaries .
a unital c*-algebra is the linear span of its unitaries because any self - adjoint element @xmath86 of norm at most @xmath314 may be written as @xmath315 and @xmath316 are unitary .
thus @xmath317 .
this is what we had to prove .
[ cor : integrable_sums ] let @xmath240 have a weak c*-hull @xmath8 .
direct sums and summands of integrable representations remain integrable , and the family of bijections @xmath312 preserves direct sums .
let @xmath318 be representations of @xmath13 on hilbert @xmath14-modules @xmath319 .
let @xmath320 for @xmath128 be the inclusion maps .
first we assume that @xmath318 are integrable .
let @xmath295 be the representation of @xmath8 on @xmath296 corresponding to @xmath293 , and let @xmath16 be the integrable representation of @xmath13 on @xmath268 corresponding to the representation @xmath321 of @xmath8 .
the isometries @xmath322 are intertwiners from @xmath295 to @xmath321 . by proposition
[ pro : adjointable_intertwiner ] , they are * -intertwiners from @xmath293 to @xmath16 .
hence @xmath323 by proposition [ pro : isometry_star - intertwiner ] .
thus @xmath324 is integrable and the family of bijections @xmath312 preserves direct sums .
the same argument works for infinite direct sums .
now we assume instead that @xmath325 is integrable .
let @xmath250 be the representation of @xmath8 corresponding to @xmath325 .
the orthogonal projection onto @xmath149 is a * -intertwiner on the representation @xmath325 by proposition [ pro : isometry_star - intertwiner ] , and hence also on @xmath250 by proposition [ pro : adjointable_intertwiner ] .
thus @xmath326 for some representations @xmath295 of @xmath8 on @xmath296 .
let @xmath327 be the integrable representation of @xmath13 corresponding to @xmath295 .
the isometry @xmath322 is a * -intertwiner from @xmath295 to @xmath321 and hence from @xmath327 to @xmath325 by proposition [ pro : adjointable_intertwiner ] .
this implies @xmath328 , so that @xmath293 is integrable for @xmath128 .
[ def : universal_integrable_rep ] let @xmath8 be a weak c*-hull for @xmath13 .
the _ universal _
integrable representation of @xmath13 is the integrable representation @xmath329 of @xmath13 on @xmath8 that corresponds to the identity representation of @xmath8 .
[ pro : phi_simplifies ] let @xmath8 with a family of bijections @xmath252 between representations of @xmath8 and integrable representations of @xmath13 on hilbert modules be a weak c*-hull for the integrable representations of @xmath13 .
let @xmath329 be the universal integrable representation of @xmath13 .
then @xmath330 for any c*-correspondence @xmath15 from @xmath8 to @xmath14 .
the proof makes this isomorphism more precise .
let @xmath331 be a representation of @xmath8 on a hilbert module @xmath15 .
then @xmath332 , @xmath333 , is a unitary intertwiner between the interior tensor product of the identity representation of @xmath8 with @xmath15 and the representation @xmath250 on @xmath15 .
if @xmath252 is compatible with interior tensor products and unitary * -intertwiners , then @xmath259 must be a * -intertwiner between @xmath334 and @xmath335 .
that is , the domain of @xmath335 is @xmath336 , and the left action is the closure of @xmath337 , where @xmath338 for all @xmath20 , @xmath339 , @xmath340 .
put in a nutshell , the whole isomorphism between integrable representations of @xmath13 and representations of @xmath8 is encoded in the single representation @xmath329 of @xmath13 on @xmath8 .
this is similar to schmdgen s approach in @xcite . in the following ,
we disregard the canonical unitary @xmath259 in the proof of proposition [ pro : phi_simplifies ] and write @xmath341 . a weak c*-hull @xmath8 does not solve the problem of describing the integrable representations of @xmath13
, it only reduces it to the study of the representations of the c*-algebra @xmath8 .
this reduction is useful because it gets rid of unbounded operators .
if @xmath8 is of type i , then any hilbert space representation of @xmath8 is a direct integral of irreducible representations , and irreducible representations may , in principle , be classified .
thus integrable hilbert space representations of @xmath13 are direct integrals of irreducible integrable representations , and the latter may , in principle , be classified .
but if @xmath8 is not of type i , then the integrable hilbert space representations of @xmath13 are exactly as complicated as the hilbert space representations of @xmath8 , and giving the c*-algebra @xmath8 may well be the best one can say about them .
[ pro : unique_hull ] a class of integrable representations has at most one weak c*-hull .
let @xmath342 and @xmath343 be weak c*-hulls for the same class of integrable representations of @xmath13 .
the identity map on @xmath342 , viewed as a representation of @xmath342 on itself , corresponds first to an integrable representation of @xmath13 on @xmath342 and further to a representation of @xmath343 on @xmath342 .
this gives a morphism from @xmath343 to @xmath342 .
similarly , we get a morphism from @xmath342 to @xmath343 .
these morphisms @xmath344 are inverse to each other because the maps they induce on representations of @xmath342 and @xmath343 on arbitrary c*-algebras are inverse to each other .
an isomorphism in the category of morphisms is an isomorphism of c*-algebras in the usual sense by @xcite*proposition 2.10 .
now take any representation @xmath329 of @xmath13 on @xmath8 .
when is this a universal integrable representation in a ( weak ) c*-hull ?
let @xmath14 be a c*-algebra and @xmath15 a hilbert @xmath14-module .
for a representation @xmath345 , let @xmath346 be the induced representation of @xmath13 on @xmath15 as in the proof of proposition [ pro : phi_simplifies ] . a representation of @xmath13
is called _
@xmath8-integrable _ if it is in the image of @xmath252 .
[ pro : cstar - generated_by_unbounded_multipliers ] the c*-algebra @xmath8 is a weak c*-hull for the @xmath8-integrable representations of @xmath13 if and only if 1 .
[ enum : cstar - gen1 ] if two representations @xmath347 on the same hilbert space @xmath261 satisfy @xmath348 as closed representations of @xmath13 , then @xmath349 .
it is a c*-hull if and only if [ enum : cstar - gen1 ] and the following equivalent conditions hold : 1 .
[ enum : cstar - gen2 ] let @xmath350 be a representation of @xmath13 on a hilbert space @xmath261 and let @xmath351 be a subrepresentation on a closed subspace @xmath352 ; that is , @xmath353 and @xmath354 for all @xmath20 .
if both @xmath355 and @xmath16 are @xmath8-integrable , then @xmath356 as vector spaces .
[ enum : cstar - gen2 ] isometric intertwiners between @xmath8-integrable hilbert space representations of @xmath13 are * -intertwiners .
[ enum : cstar - gen2 ] @xmath8-integrable subrepresentations of @xmath8-integrable hilbert space representations of @xmath13 are direct summands
. the conditions [ enum : cstar - gen1][enum : cstar - gen2 ] together are equivalent to 1 .
[ enum : cstar - gen12 ] let @xmath357 be a hilbert space representation and let @xmath350 be the associated representation of @xmath13 on @xmath261 . if @xmath358 is a @xmath8-integrable subrepresentation of @xmath350 on a closed subspace @xmath352 , then the projection onto @xmath359 commutes with @xmath360 .
the condition [ enum : cstar - gen1 ] says that @xmath252 is injective .
it maps onto the @xmath8-integrable representations of @xmath13 by definition .
the map @xmath252 is compatible with interior tensor products by lemma [ lem : tensor_associative ] .
a unitary operator @xmath259 * -intertwines two representations @xmath124 and @xmath125 of @xmath13 if and only if @xmath361 , where @xmath362 denotes the representation with domain @xmath363 and @xmath364 .
similarly , @xmath259 intertwines two representations @xmath246 and @xmath247 of @xmath8 if and only if @xmath365 .
hence [ enum : cstar - gen1 ] implies that a unitary that * -intertwines two @xmath8-integrable hilbert space representations of @xmath13 also intertwines the corresponding representations of @xmath8 .
the converse is clear .
so @xmath8 is a weak c*-hull if and only if [ enum : cstar - gen1 ] holds . the equivalence between [ enum : cstar - gen2 ] , [ enum : cstar - gen2 ] and [ enum : cstar - gen2 ] follows from proposition [ pro : isometry_star - intertwiner ] by writing @xmath366 .
assume that @xmath8 is a c*-hull .
an isometric intertwiner for @xmath13 is also one for @xmath8 .
then it is a * -intertwiner for @xmath13 and its range projection is an intertwiner for @xmath8 by proposition [ pro : adjointable_intertwiner ] .
thus both [ enum : cstar - gen2 ] and [ enum : cstar - gen12 ] follow if @xmath8 is a c*-hull .
conversely , assume [ enum : cstar - gen1 ] and [ enum : cstar - gen2 ] .
we are going to prove that @xmath8 is a c*-hull for the @xmath8-integrable representations of @xmath13 .
we have already seen that @xmath8 is a weak c*-hull .
we must check compatibility with isometric intertwiners .
let @xmath14 be a c*-algebra and let @xmath319 be hilbert @xmath14-modules with representations @xmath367 of @xmath8 .
the corresponding representations @xmath368 of @xmath13 for @xmath128 are the closures of the representations on @xmath369 given by @xmath370 for @xmath20 , @xmath339 , @xmath371 .
hence an isometric intertwiner for @xmath8 is also one for @xmath13 .
conversely , let @xmath167 be a hilbert module isometry with @xmath372 and @xmath373 for all @xmath20 , @xmath155
. we must prove @xmath374 for all @xmath375 .
let @xmath376 be a faithful representation on a hilbert space @xmath377 .
equip @xmath378 with the induced representations @xmath379 of @xmath8 and @xmath380 of @xmath13 for @xmath128 . since the family of bijections @xmath381 is compatible with interior tensor products , it maps @xmath379 to @xmath380 .
the operator @xmath159 induces an isometric intertwiner @xmath382 from @xmath383 to @xmath384 by lemma [ lem : rep_tensor_corr_functor ] .
since @xmath383 and @xmath384 are @xmath8-integrable , we are in the situation of [ enum : cstar - gen2 ] .
so @xmath382 is a * -intertwiner from @xmath383 to @xmath384 .
thus @xmath382 is an intertwiner from @xmath385 to @xmath386 by proposition [ pro : adjointable_intertwiner ] .
since the representation @xmath387 is faithful , so is the induced map @xmath388 .
hence the intertwining condition @xmath389 proved above lifts to @xmath390 .
thus @xmath252 is compatible with isometric intertwiners .
since [ enum : cstar - gen12 ] holds for c*-hulls , we have proved along the way that [ enum : cstar - gen1 ] and [ enum : cstar - gen2 ] imply [ enum : cstar - gen12 ] .
it remains to show , conversely , that [ enum : cstar - gen12 ] implies [ enum : cstar - gen1 ] and [ enum : cstar - gen2 ] .
in the situation of [ enum : cstar - gen2 ] , the projection @xmath391 onto @xmath359 commutes with @xmath8 by [ enum : cstar - gen12 ] .
thus the representation of @xmath8 on @xmath261 is a direct sum of representations on @xmath359 and @xmath392 .
this is inherited by the induced representation of @xmath13 and its domain .
so [ enum : cstar - gen12 ] implies [ enum : cstar - gen2 ] .
in the situation of [ enum : cstar - gen1 ] , form the direct sum representation @xmath321 on @xmath393 and let @xmath394 .
the representation of @xmath13 corresponding to @xmath321 is @xmath395 . since @xmath396 by assumption , the domain of @xmath395 is @xmath397 for some dense subspace @xmath398 , and @xmath399 is a dense subspace in @xmath359 that is invariant for the representation @xmath395 .
the restricted representation on this subspace is @xmath8-integrable because it is unitarily equivalent to @xmath400 .
therefore , the projection onto @xmath359 commutes with the representation of @xmath8 by [ enum : cstar - gen12 ] .
thus @xmath349 .
so [ enum : cstar - gen12 ] implies [ enum : cstar - gen1 ] .
the condition [ enum : cstar - gen1 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] is required in several other theories that associate a c*-algebra to a * -algebra , such as the host algebras of grundling @xcite , the c*-envelopes of dowerk and savchuk @xcite , or the notion of a c*-algebra generated by affiliated multipliers by woronowicz @xcite , see @xcite*theorem 3.3 or the proof of theorem [ the : woronowicz_local - global ] below . the equivalent conditions [ enum : cstar - gen2][enum : cstar - gen2 ] may be easier to check than [ enum : cstar - gen12 ] because they do not involve the c*-hull .
[ pro : envelope_unique ] let @xmath13 be a * -algebra and let @xmath401 be c*-algebras with representations @xmath402 of @xmath13 for @xmath128 .
assume that for each hilbert space @xmath261 , the maps @xmath403 , @xmath404 , are injective and have the same image .
then there is a unique isomorphism @xmath405 intertwining the representations @xmath402 of @xmath13 for @xmath128 .
hence a c*-envelope as defined in @xcite is unique if it exists .
both @xmath342 and @xmath343 are weak c*-hulls for the same class of representations of @xmath13 by proposition [ pro : cstar - generated_by_unbounded_multipliers ] .
proposition [ pro : unique_hull ] gives the isomorphism @xmath406 .
[ rem : envelope_unique ] it is remarkable that the conditions in proposition [ pro : cstar - generated_by_unbounded_multipliers ] and corollary [ pro : envelope_unique ] only involve hilbert space representations because the hilbert space representations of a c*-algebra only determine its bidual @xmath407-algebra , not the c*-algebra itself . for corollary [ pro : envelope_unique ] , this works because the bijection between the representations is of a particular form , induced by representations of @xmath13 .
[ def : integrable_admissible ] let @xmath13 be a * -algebra .
a class of `` integrable '' representations of @xmath13 on hilbert modules over c*-algebras is _ admissible _ if it satisfies the conditions [ enum : admissible0][enum : admissible3 ] below , and _ weakly admissible _ if it satisfies [ enum : admissible0][enum : admissible2 ] . 1 .
[ enum : admissible0 ] if there is a unitary * -intertwiner from an integrable representation to another representation , then the latter is integrable .
[ enum : admissible1 ] if @xmath14 and @xmath408 are c*-algebras , @xmath30 is a correspondence from @xmath14 to @xmath408 , and @xmath62 is an integrable representation of @xmath13 on a hilbert @xmath14-module @xmath15 , then the representation @xmath409 on @xmath410 is integrable .
[ enum : admissible2 ] direct sums and summands of integrable representations are integrable .
[ enum : admissible3 ] any integrable subrepresentation of an integrable representation of @xmath13 on a hilbert space is a direct summand .
[ lem : integrable_admissible ] any class of integrable representations with a weak c*-hull is weakly admissible . if there is a c*-hull , proposition [ pro : cstar - generated_by_unbounded_multipliers ] implies [ enum : admissible3 ] in definition [ def : integrable_admissible ] .
if there is a weak c*-hull , then [ enum : admissible0 ] and [ enum : admissible1 ] in definition [ def : integrable_admissible ] follow from the compatibility with unitary * -intertwiners and interior tensor products in the definition of a c*-hull , and [ enum : admissible2 ] follows from corollary [ cor : integrable_sums ] .
[ pro : representation_hilm_comp ] let @xmath13 be a unital * -algebra and let @xmath15 be a hilbert module over a c*-algebra @xmath14 .
there is a natural bijection between the sets of representations of @xmath13 on @xmath15 and @xmath411 .
it preserves integrability if the class of integrable representations of @xmath13 is weakly admissible or , in particular , if it has a weak c*-hull
. we may view @xmath15 as an imprimitivity bimodule between @xmath411 and the ideal @xmath159 in @xmath14 that is spanned by the inner products @xmath412 for @xmath413 .
let @xmath414 be the inverse imprimitivity bimodule , which is a hilbert module over @xmath411 with @xmath415
. then @xmath416 and @xmath417 .
if @xmath418 is a representation of @xmath13 on @xmath15 , then @xmath419 is a representation of @xmath13 on @xmath420 .
this maps @xmath421 to @xmath422 .
if @xmath423 is a representation of @xmath13 on @xmath411 , then @xmath424 is a representation of @xmath13 on @xmath425 .
this maps @xmath422 to @xmath421 .
we claim that these two maps are inverse to each other . both preserve integrability by [ enum : admissible1 ] in definition [ def : integrable_admissible ] .
the map @xmath426 sends a representation @xmath418 of @xmath13 on @xmath15 to the representation @xmath427 of @xmath13 on @xmath428 by lemma [ lem : tensor_associative ] .
this is the restriction of @xmath16 to @xmath429 .
since @xmath15 is also a hilbert module over @xmath159 , it is nondegenerate as a right @xmath159-module .
therefore , if @xmath430 is an approximate unit in @xmath159 , then @xmath431 for all @xmath340
. then also @xmath432 for all @xmath81 , @xmath20 , so @xmath431 in the graph topology for all @xmath81 .
thus @xmath433 , and we get the identity map on @xmath421 . a similar , easier argument shows that we also get the identity map on @xmath422 .
let @xmath434 $ ] with @xmath35 .
a ( not necessarily closed ) representation of @xmath13 on a hilbert @xmath14-module @xmath15 is determined by a dense @xmath14-submodule @xmath17 and a single symmetric operator @xmath435 , that is , @xmath436 . then @xmath437 .
[ lem : graph_topology_cx ] the graph topology on @xmath21 is generated by the increasing sequence of norms @xmath438 for @xmath439 .
we must show that for any @xmath440 $ ] there are @xmath143 and @xmath439 with @xmath441 .
we choose @xmath442 so that @xmath98 has degree at most @xmath442
. then there is @xmath143 so that @xmath443 for all @xmath444 .
thus the polynomial @xmath445 is positive on @xmath111 .
so the zeros of @xmath446 are complex and come in pairs @xmath447 for @xmath448 with @xmath449 by the fundamental theorem of algebra
. then @xmath450 , where each @xmath451 is a product of either @xmath452 or @xmath453 for @xmath454 , so @xmath455 .
thus @xmath441 .
thus the monomials @xmath456 form a strong generating set for @xmath37 $ ]
. a representation of @xmath37 $ ] is determined by the closed operators @xmath457 for @xmath439 by proposition [ pro : equality_if_closure_equal ] .
in contrast , it is not yet determined by the single closed operator @xmath38 because @xmath458 is not a _
strong _ generating set : [ exa : x_not_strong_generator ] we construct a closed representations of @xmath37 $ ] on a hilbert space with @xmath459 .
we start with the representation @xmath460 of @xmath37 $ ] on a hilbert space @xmath261 associated to some unbounded self - adjoint operator @xmath38 .
since this operator is unbounded , there is a linear functional on @xmath21 that is bounded in the graph norm of @xmath461 , but not in the graph norm of @xmath257 .
let @xmath462 .
this does not change @xmath38 because @xmath40 is unbounded in the graph norm of @xmath257 .
but now @xmath463 on the domain of @xmath464 .
all notions of integrability for representations of @xmath37 $ ] that we shall consider imply @xmath465 . under this assumption
, an _ integrable _ representation of @xmath37 $ ] is determined by the single closed operator @xmath38 .
let @xmath466 .
let @xmath197 be the identity function on @xmath111 , viewed as an unbounded multiplier of @xmath8 .
we define a closed representation @xmath329 of @xmath13 on @xmath36 by @xmath467 [ the : regular_self - adjoint_cstar - hull ] let @xmath62 be a representation of @xmath434 $ ] on a hilbert module @xmath15 over a c*-algebra @xmath14 .
the following are equivalent : 1 .
[ enum : regular_self - adjoint_cstar - hull1 ] @xmath468 for a representation @xmath331 ; 2 .
[ enum : regular_self - adjoint_cstar - hull2 ] @xmath25 is regular and self - adjoint for each @xmath469 ; 3 .
[ enum : regular_self - adjoint_cstar - hull3 ] @xmath457 is regular and self - adjoint for each @xmath439 ; 4 .
[ enum : regular_self - adjoint_cstar - hull4 ] @xmath38 is regular and self - adjoint and @xmath470 for all @xmath439 ; 5 .
[ enum : regular_self - adjoint_cstar - hull5 ] @xmath38 is regular and self - adjoint and @xmath471 .
call representations with these equivalent properties integrable .
the c*-algebra @xmath36 is a c*-hull for the integrable representations of @xmath13 with @xmath329 as the universal integrable representation . if @xmath123 , then @xmath472 is a self - adjoint , affiliated multiplier of @xmath8 . hence @xmath473 is a regular , self - adjoint operator on @xmath474 for any representation @xmath250 of @xmath8 on @xmath15 by @xcite*proposition 9.10 .
thus [ enum : regular_self - adjoint_cstar - hull1 ] implies [ enum : regular_self - adjoint_cstar - hull2 ] .
the implication [ enum : regular_self - adjoint_cstar - hull2]@xmath272[enum : regular_self - adjoint_cstar - hull3 ] is trivial .
the operator @xmath457 is always contained in the @xmath442-fold power @xmath475 .
the latter is symmetric , and a proper suboperator of a symmetric operator can not be self - adjoint .
thus [ enum : regular_self - adjoint_cstar - hull3 ] implies [ enum : regular_self - adjoint_cstar - hull4 ] .
the set @xmath456 is a strong generating set for @xmath37 $ ] by lemma [ lem : graph_topology_cx ] .
equation gives @xmath476 for any ( closed ) representation .
thus [ enum : regular_self - adjoint_cstar - hull4 ] implies [ enum : regular_self - adjoint_cstar - hull5 ] .
assume [ enum : regular_self - adjoint_cstar - hull5 ] and abbreviate @xmath477 .
the functional calculus for @xmath86 is a nondegenerate * -homomorphism @xmath478 ( see @xcite*theorem 10.9 ) .
let @xmath263 be the representation of @xmath13 associated to @xmath250 .
we claim that @xmath479 .
the functional calculus extends to affiliated multipliers and maps the identity function on @xmath111 to the regular , self - adjoint operator @xmath86 .
this means that @xmath480
. then @xmath481 .
this implies @xmath482 because both are self - adjoint .
since the set @xmath456 is a strong generating set for @xmath37 $ ] , the domain of @xmath263 is @xmath483 by condition [ enum : regular_self - adjoint_cstar - hull5 ] and . on this domain ,
@xmath484 and @xmath485 act by the same operator because they have the same closure .
thus @xmath479 and [ enum : regular_self - adjoint_cstar - hull5 ] implies [ enum : regular_self - adjoint_cstar - hull1 ] .
so all five conditions in the theorem are equivalent . to show that @xmath8 is a c*-hull for the class of representations described in [ enum : regular_self - adjoint_cstar - hull1 ] , we check [ enum : cstar - gen12 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] . an integrable representation of @xmath13 on a hilbert space
@xmath261 corresponds to a self - adjoint operator @xmath86 on @xmath261 by [ enum : regular_self - adjoint_cstar - hull5 ] .
an integrable subrepresentation is a closed subspace @xmath359 of @xmath261 with a self - adjoint operator @xmath486 on @xmath359 whose graph is contained in that of @xmath86 .
since @xmath486 is self - adjoint , the subspaces @xmath487 are equal to @xmath359 .
the cayley transform @xmath259 of @xmath86 maps @xmath488 onto @xmath489 .
thus it maps @xmath359 onto itself .
since @xmath490 generates the image of @xmath491 under the functional calculus , the projection onto @xmath359 is @xmath8-invariant .
[ exa : irregular_selfadjoint ] regularity and self - adjointness are independent properties of a symmetric operator .
examples of regular symmetric operators that are not self - adjoint are easy to find , see [ sec : polynomials2 ] .
we are going to construct a representation @xmath16 of @xmath37 $ ] on a hilbert module for which @xmath25 is self - adjoint for each @xmath440 $ ] with @xmath27 , but @xmath38 is not regular .
we follow the example after thorme 1.3 in @xcite , which pierrot attributes to hilsum .
let @xmath261 be the hilbert space @xmath492)$ ] and let @xmath493 and @xmath494 be the operators @xmath495 on @xmath261 with the following cores . for @xmath493 ,
we take @xmath314-periodic smooth functions ; for @xmath494 , we take the restrictions to @xmath496 $ ] of smooth functions on @xmath111 satisfying @xmath497 . both @xmath493 and @xmath494 are essentially self - adjoint .
let @xmath498)$ ] and @xmath499,{\mathcal{h}})$ ] .
let @xmath17 be the dense subspace of all functions @xmath500\times [ 0,1]\to{\mathbb c}$ ] such that @xmath501 is continuous for each @xmath439 , @xmath502 for all @xmath503 $ ] , @xmath504 , @xmath505 , and @xmath506 equivalently , @xmath507 belongs to the domain of @xmath508 for all @xmath439 , @xmath509 and to the domain of @xmath510 for all @xmath439 , @xmath511 ; indeed , this forces @xmath512 to be continuous on @xmath513\times [ 0,1]$ ] and to satisfy the boundary conditions .
these imply by continuity .
let @xmath514 $ ] act on @xmath21 by @xmath495 .
this defines a closed * -representation of @xmath37 $ ] on @xmath15 with @xmath515 .
the closure @xmath38 is the irregular self - adjoint operator described in @xcite .
let @xmath440 $ ] with @xmath27 .
then @xmath516 is ( regular and ) self - adjoint for any _ integrable _ representation @xmath250 of @xmath37 $ ] by theorem [ the : regular_self - adjoint_cstar - hull ] .
therefore , the restriction of @xmath25 to a single fibre of @xmath15 at some @xmath517\setminus\{0\}$ ] is a self - adjoint operator on @xmath492)$ ] because @xmath493 and @xmath494 are self - adjoint and @xmath515 .
the restriction of @xmath518 at @xmath519 is contained in the self - adjoint operators @xmath520 and @xmath521 by continuity .
we claim that @xmath522 .
this claim implies that @xmath518 is contained in @xmath25 , that is , @xmath25 is self - adjoint .
let @xmath523 $ ] have degree @xmath442 .
then the graph norms for @xmath98 and @xmath524 are equivalent in any representation by the proof of lemma [ lem : graph_topology_cx ] .
hence @xmath525 and @xmath526 have the same domain .
the domain of @xmath526 consists of functions @xmath496\to{\mathbb c}$ ] whose @xmath442th derivative lies in @xmath527 and whose derivatives of order strictly less than @xmath442 satisfy the boundary condition for @xmath528 .
hence the domain of @xmath529 consists of those functions @xmath496\to{\mathbb
c}$ ] whose @xmath442th derivative lies in @xmath527 and whose derivatives of order strictly less than @xmath442 vanish at the boundary points @xmath530 and @xmath314 .
this is exactly the domain of the closure of @xmath531 . on this domain
the operators @xmath532 and @xmath533 both act by the differential operator @xmath534 .
the algebra @xmath434 $ ] has many hilbert space representations coming from closed symmetric operators that are not self - adjoint .
there is , however , no larger admissible class of integrable representations : [ pro : symmetric_no_hull ] assume that an admissible class of integrable representations of @xmath434 $ ] contains all representations coming from self - adjoint hilbert space operators .
then any integrable representation of @xmath13 on a hilbert module comes from a regular , self - adjoint operator .
we first prove that there can be no more integrable hilbert space representations than those coming from self - adjoint operators .
let @xmath350 be an integrable representation on a hilbert space @xmath261 .
we may extend the closed symmetric operator @xmath535 on @xmath261 to a self - adjoint operator @xmath536 on a larger hilbert space @xmath537 .
this gives a representation @xmath133 of @xmath13 on @xmath537 as in theorem [ the : regular_self - adjoint_cstar - hull ] , which is integrable by assumption .
the inclusion map @xmath538 is an isometric intertwiner from @xmath16 to @xmath133 .
hence @xmath16 is a direct summand of @xmath133 by [ enum : admissible3 ] in definition [ def : integrable_admissible ] .
thus @xmath457 is self - adjoint for each @xmath439 , and @xmath16 is the representation induced by @xmath86 .
now let @xmath62 be an integrable representation of @xmath13 on a hilbert @xmath14-module @xmath15 .
for any hilbert space representation @xmath539 , the induced representation of @xmath13 on the hilbert space @xmath540 is also integrable by [ enum : admissible1 ] in definition [ def : integrable_admissible ] .
thus @xmath541 is self - adjoint for any hilbert space representation @xmath539 . a closed , densely defined , symmetric operator @xmath41 on a hilbert @xmath14-module @xmath15 is self - adjoint and regular if and only if , for any state @xmath40 on @xmath14 , the closure of @xmath542 on the hilbert spaces @xmath543 is self - adjoint ; here @xmath544 means the gns - representation for @xmath40 .
this is called the local global principle by kaad and lesch ( @xcite*theorem 1.1 ) ; the result was first proved by pierrot ( @xcite*thorme 1.18 ) .
we will take up local global principles more systematically in [ sec : local_global ] .
thus @xmath457 is regular and self - adjoint for every @xmath439 , so @xmath16 is obtained from the regular self - adjoint operator @xmath38 as in theorem [ the : regular_self - adjoint_cstar - hull ] .
[ exa : smaller_integrable_than_regular_self - adjoint ] there are many admissible classes of representations of @xmath37 $ ] that are smaller than the class in theorem [ the : regular_self - adjoint_cstar - hull ] .
there are even many such classes that contain the same hilbert space representations . for instance
, let @xmath545 with the representation of polynomials by pointwise multiplication .
this is a c*-hull for a class of representations of @xmath37 $ ] by theorem [ the : commutative_hull ] below . since the standard topologies on @xmath111 and
@xmath546 have the same borel sets , both c*-hulls @xmath547 and @xmath36 give the same integrable hilbert space representations .
but there are regular , self - adjoint operators on hilbert modules that do not give a @xmath8-integrable representation .
the obvious example is the multiplier @xmath197 of @xmath36 that generates the universal integrable representation of @xmath37 $ ] .
can there be an _
admissible _ class of representations of @xmath37 $ ] that contains some representation on a hilbert space that does not come from a self - adjoint operator ?
we can not rule this out completely .
but such a class would have to be rather strange . by proposition
[ pro : symmetric_no_hull ] , it can not contain all self - adjoint operators .
by example [ exa : friedrichs_extension ] , it can not contain all positive symmetric operators because then there would be isometric intertwiners among integrable representations that are not * -intertwiners .
the following example rules out symmetric operators with one deficiency index @xmath530 : [ exa : deficiency_index_0 ] let @xmath86 be a closed symmetric operator on a hilbert space @xmath261 of deficiency indices @xmath548 for some @xmath549 $ ] .
then @xmath550 is a core for each power @xmath551 by @xcite*proposition 1.6.1 .
thus there is a closed representation @xmath16 of @xmath37 $ ] with domain @xmath552 and @xmath553 for all @xmath554 . by assumption ,
the operator @xmath555 is surjective , but @xmath556 is not .
that is , the cayley transform @xmath557 is a non - unitary isometry .
the operator @xmath86 may be reconstructed from @xmath558 as in @xcite*equation ( 10.11 ) .
here @xmath559 is surjective , so this simplifies to @xmath560 , and @xmath561 for all @xmath562 . thus @xmath563 and @xmath564 because @xmath565 then @xmath566 for all @xmath439 .
thus @xmath558 is an isometric intertwiner from @xmath16 to itself by proposition [ pro : intertwiner_strong_generators ] . if @xmath559 were an intertwiner as well
, then @xmath567 and @xmath568 for all @xmath569 .
so @xmath570 this is impossible because @xmath571 and @xmath555 is surjective .
so the isometry @xmath558 is an intertwiner , but not a * -intertwiner .
this is forbidden for admissible classes of integrable representations .
if @xmath86 has deficiency indices @xmath572 instead , then @xmath573 has deficiency indices @xmath548 and its cayley transform is an isometric intertwiner that is not a * -intertwiner by the argument above .
[ def : local_global ] let @xmath13 be a * -algebra with a weakly admissible class of integrable representations ( definition [ def : integrable_admissible ] ) .
the _ local global principle _ says that a representation @xmath16 of @xmath13 on a hilbert @xmath14-module @xmath15 is integrable if ( and only if ) the representations @xmath574 are integrable for all hilbert space representations @xmath575 .
the _ strong local global principle _ says that a representation @xmath16 of @xmath13 on a hilbert @xmath14-module @xmath15 is integrable if ( and only if ) the representations @xmath574 are integrable for all _ irreducible _ hilbert space representations @xmath575 . roughly speaking , the local
global principle says that the class of integrable representations on hilbert modules is determined by the class of integrable representations on hilbert spaces .
examples where the local
global principle fails are constructed in [ sec : polynomials2 ] and [ sec : commutative_hulls ] .
we do not know an example with the local
global principle for which the strong local global principle fails . an irreducible representation @xmath576 is unitarily equivalent to the gns - representation for a pure state @xmath577 on @xmath14 .
the tensor product @xmath578 is canonically isomorphic to the completion @xmath579 of @xmath15 to a hilbert space for the scalar - valued inner product @xmath580 .
the induced representation @xmath581 of @xmath13 on @xmath582 is the closure of the representation @xmath16 with domain @xmath583 .
any representation @xmath576 is a direct sum of cyclic representations , and these are gns - representations of states . since any weakly admissible class of integrable representations
is closed under direct sums , the local global principle holds if and only if integrability of @xmath574 for all gns - representations @xmath250 of states on @xmath14 implies integrability of @xmath16 .
[ exa : local - global_cx ] define integrable representations of the polynomial algebra @xmath37 $ ] as in theorem [ the : regular_self - adjoint_cstar - hull ] .
thus they correspond to regular , self - adjoint operators on hilbert modules .
the main result in @xcite says that the integrable representations of @xmath37 $ ] satisfy the local global principle .
this is where our notation comes from .
we already used this to prove proposition [ pro : symmetric_no_hull ] . the strong local global principle for integrable representations of @xmath37 $ ] is only conjectured in @xcite .
this conjecture had already been proved by pierrot in @xcite*thorme 1.18 before @xcite was written , based on the following hahn banach type theorem for hilbert submodules : [ the : hahn - banach_submodules ] let @xmath14 be a c*-algebra and let @xmath15 be a hilbert @xmath14-module .
let @xmath584 be a proper , closed hilbert submodule .
there is an irreducible hilbert space representation @xmath576 with @xmath585 .
[ cor : hahn - banach_submodules ] let @xmath15 be a hilbert module over a c*-algebra @xmath14 .
let @xmath586 be two closed hilbert submodules .
if @xmath587 , then there is an irreducible hilbert space representation @xmath575 with @xmath588 as closed subspaces in @xmath589 .
[ cor : local_global_self - adjoint ] let @xmath41 be a closed , semiregular operator on a hilbert @xmath14-module @xmath15 .
the operator @xmath41 is regular if and only if , for each irreducible representation @xmath575 on a hilbert space @xmath261 , the closures of @xmath590 and @xmath591 on @xmath592 are adjoints of each other .
hence @xmath41 is regular and self - adjoint if and only if @xmath593 is a self - adjoint operator on @xmath592 for each irreducible hilbert space representation @xmath576 .
we now apply the above results of pierrot .
first we deduce a criterion for representations to be equal .
then we prove that certain definitions of integrability automatically satisfy the strong local global principle .
[ the : identify_representations ] let @xmath13 be a * -algebra and let @xmath293 for @xmath128 be closed representations of @xmath13 on a hilbert module @xmath15 over a c*-algebra @xmath14 .
the following are equivalent : 1 .
@xmath594 ; 2 .
@xmath595 for each irreducible hilbert space representation @xmath250 of @xmath14 ; 3 .
@xmath126 for each @xmath20 .
the equivalence ( 3)@xmath596(1 ) is proposition [ pro : equality_if_closure_equal ] , and ( 1 ) clearly implies ( 2 ) .
thus we only have to prove that not ( 3 ) implies not ( 2 ) .
assume that there is @xmath20 with @xmath597 .
the graphs @xmath598 and @xmath599 of @xmath170 and @xmath171 are different hilbert submodules of @xmath600 .
corollary [ cor : hahn - banach_submodules ] gives an irreducible representation @xmath250 of @xmath14 with @xmath601 .
this says that @xmath602 because @xmath603 is the graph of @xmath604 .
how do we specify which representations @xmath16 of a * -algebra @xmath13 are integrable ?
there are two basically different ways .
the universal way specifies the universal integrable representation .
that is , it starts with a representation @xmath329 on a c*-algebra @xmath8 that satisfies [ enum : cstar - gen1 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] and takes the class of @xmath8-integrable representations .
the operator way imposes conditions on the operators @xmath67 , such as regularity and self - adjointness of @xmath25 or strong commutation relations .
in good cases , the same class of integrable representations may be specified in both ways .
for instance , theorem [ the : regular_self - adjoint_cstar - hull ] shows that several classes of representations of @xmath37 $ ] are equal .
the first is defined by the universal representation on @xmath36 .
the second asks @xmath25 to be regular and self - adjoint for all @xmath123 . we are going to make the operator way more precise so that _ all _ classes of representations defined in this way satisfy the strong local global principle .
this is a powerful method to prove local global principles .
[ def : restriction_representation ] let @xmath13 be a * -algebra and @xmath605 some weakly admissible class of representations of @xmath13 on hilbert modules over c*-algebras .
a _ natural construction of hilbert submodules _
( of rank @xmath606 ) associates to each representation @xmath16 on a hilbert module @xmath15 that belongs to @xmath605 a hilbert submodule @xmath607 , such that 1
. [ enum : restriction0 ] if @xmath608 is a unitary * -intertwiner between two representations @xmath132 and @xmath133 in @xmath605 , then @xmath609 maps @xmath610 onto @xmath611 ; 2 .
[ enum : restriction1 ] let @xmath28 and @xmath29 be c*-algebras and let @xmath612 be a @xmath613-correspondence ; let @xmath16 be a representation in @xmath605 on a hilbert @xmath28-module @xmath15 ; then the canonical isomorphism @xmath614 maps @xmath615 onto @xmath616 ; 3 .
[ enum : restriction2 ] if @xmath293 for @xmath617 in a set @xmath159 are representations in @xmath605 on hilbert @xmath14-modules @xmath296 over the same c*-algebra @xmath14 , then the canonical isomorphism @xmath618 maps @xmath619 onto @xmath620 . in brief ,
@xmath621 is compatible with unitary * -interwiners , interior tensor products , and direct sums . a smaller class of representations @xmath622 is _ defined by a submodule condition _ if there are two natural constructions of hilbert submodules @xmath623 ,
@xmath128 , of the same rank @xmath442 , such that a representation @xmath16 in @xmath605 belongs to @xmath624 if and only if @xmath625 . a class of representations @xmath626 is _ defined by submodule conditions _ if it is defined by transfinite recursion by repeating the step in the previous paragraph . more precisely , there are a well - ordered set @xmath159 with a greatest element @xmath627 and least element @xmath530 and subclasses @xmath628 for @xmath629 such that 1 .
@xmath630 and @xmath631 ; 2 .
@xmath632 is defined by a submodule condition for each @xmath633 ; 3 .
@xmath634 if @xmath635 and @xmath636 for all @xmath637 .
the following lemma makes this definition meaningful , the following theorem makes it interesting .
[ lem : operator_conditions ] if @xmath638 is weakly admissible and @xmath622 is defined by a submodule condition , then @xmath624 is also weakly admissible . if @xmath639 is a set of weakly admissible subclasses , then @xmath640 is weakly admissible .
any class of representations defined by submodule conditions is weakly admissible . [
the : operator_conditions ] if @xmath626 is defined by submodule conditions , then it satisfies the strong local global principle .
before we prove these two results , we give examples of classes of representations defined by one or more operator conditions , and a few counterexamples . these show that a class of integrable representations defined in the operator way is often but not always defined by submodule conditions .
[ exa : regularity_condition ] the _ regularity condition _ for @xmath123 requires @xmath25 to be regular and self - adjoint .
equivalently , the closures of @xmath641 for both signs are dense in @xmath15 ; this is equivalent to @xmath25 having a unitary cayley transform , which is equivalent to @xmath25 being regular and self - adjoint . sending @xmath16 to the closed image of @xmath642 or @xmath643 is a natural construction of a hilbert submodule .
hence the condition that @xmath25 is regular and self - adjoint is equivalent to the combination of two submodule conditions of rank @xmath314 .
we may rewrite this as one submodule condition of rank @xmath61 .
alternatively , we may proceed as in the definition of regularity for non - self - adjoint operators .
let @xmath644 denote the closure of the graph of an operator @xmath41 .
a closed operator @xmath41 is regular if and only if the direct sum of @xmath644 and @xmath645 is @xmath600 , where @xmath646 .
if @xmath20 is self - adjoint , then regularity and self - adjointness of @xmath25 together are equivalent to the equality of @xmath647 we claim that @xmath648 and @xmath649 are natural constructions of hilbert submodules of rank @xmath61 .
this is trivial for @xmath649 . that @xmath648 is compatible with unitary intertwiners and direct sums
is an easy exercise .
the construction @xmath648 is compatible with interior tensor products because the graph of @xmath650 is @xmath651 .
for instance , [ enum : regular_self - adjoint_cstar - hull2 ] in theorem [ the : regular_self - adjoint_cstar - hull ] defines integrable representations of @xmath37 $ ] by regularity conditions .
we generalise this in theorem [ the : regular_admissible_local - global ] below .
[ exa : alone_regularity ] the class of representations where @xmath25 is regular for some @xmath20 is always weakly admissible by @xcite*proposition 9.10 .
the first example in [ sec : polynomials2 ] shows a class of representations defined by such a condition that does not satisfy the local global principle , in contrast to theorem [ the : operator_conditions ] . hence asking for @xmath25 to be regular for some @xmath20 can not be a submodule condition .
the problem is that the inclusion @xmath652 for a correspondence @xmath612 may be strict .
[ exa : closure_condition ] two operators are equal if and only if they have the same graph . the graph of a closed operator on @xmath15 is a hilbert submodule of @xmath653 .
if two closed operators are constructed from a representation in a way that is compatible with unitary * -intertwiners , direct sums , and interior tensor products , then asking for their equality is a submodule condition .
for instance , in the study of representations of @xmath37 $ ] , we met the condition @xmath465 for all @xmath439 .
this is meaningful because @xmath475 is automatically closed .
more generally , we may require the operators @xmath654 and @xmath655 for some noncommutative polynomial @xmath656 and @xmath77 to have the same closure .
this always gives a submodule condition .
the construction of adjoints , however , is not allowed , compare example [ exa : alone_regularity ] .
[ exa : strong_commutation ] let @xmath121 and suppose that @xmath657 and @xmath658 are self - adjoint , regular operators for all representations in @xmath605 ; we may achieve this by submodule conditions as in example [ exa : regularity_condition ] in previous steps of a recursive definition .
we say that @xmath659 and @xmath536 _ strongly commute _ if their cayley transforms @xmath660 and @xmath661 commute .
equivalently , @xmath660 commutes with @xmath536 , that is , @xmath662 .
the graphs of @xmath536 and @xmath663 are natural constructions of submodules of rank @xmath61 .
therefore , strong commutation of @xmath664 and @xmath665 is a submodule condition .
[ exa : nondegeneracy_condition ] let @xmath666 be an ideal .
a _ nondegeneracy condition _ for @xmath159 asks the closed linear span of @xmath667 to be all of @xmath15 ; here @xmath21 is the domain of @xmath16 .
this is by definition the equality of @xmath668 , the closed linear span of @xmath669 for @xmath670 , @xmath81 , and @xmath671 .
these are easily seen to be natural constructions of hilbert submodules of rank @xmath314 .
so a nondegeneracy condition is a submodule condition .
for instance , let @xmath159 be a non - unital * -algebra and let @xmath672 be its unitisation .
any representation of @xmath159 extends uniquely to a unital representation of @xmath13 .
the class of nondegenerate representations of @xmath159 inside the class of all representations of @xmath13 is defined by a submodule condition .
more generally , let @xmath673 be vector subspaces and ask the closed linear spans of @xmath669 for @xmath674 , @xmath81 to be equal for @xmath675 .
this is a submodule condition as well .
for instance , the condition @xmath676 for @xmath123 is of this form .
it holds if and only if the cayley transform of @xmath25 is an isometry .
often we need a mild generalisation of the above construction , see example [ exa : spectral_condition ] below .
suppose that we have constructed a representation @xmath677 of a unital * -algebra @xmath303 on @xmath15 for any representation @xmath16 in @xmath605 , such that @xmath678 is compatible with unitary * -intertwiners , direct sums , and interior tensor products ; the last property means that @xmath679 as representations on @xmath680
. then we may ask the nondegeneracy condition for an ideal in @xmath303 instead . in particular , @xmath303 may be a weak c*-hull for some class of representations containing @xmath605 .
[ exa : spectral_condition ] let @xmath681 be commuting , symmetric elements and suppose that @xmath682 for @xmath448 are strongly commuting , self - adjoint , regular operators for all representations in @xmath605 ; we may achieve all this by previous submodule conditions as in examples [ exa : regularity_condition ] and [ exa : strong_commutation ] .
a _ closed spectral condition _ asks the joint spectrum of @xmath683 to be contained in a closed subset @xmath684 .
we claim that this is a submodule condition . under our assumptions on @xmath683 , the functional calculus homomorphism @xmath685 exists .
the condition on the joint spectrum is equivalent to @xmath686 , where @xmath687 is a function that vanishes exactly on @xmath197 , that is , @xmath688 if and only if @xmath689 .
the construction of the operator @xmath690 is clearly compatible with unitary * -intertwiners and direct sums .
it is also compatible with interior tensor products because the functional calculus for @xmath691 maps @xmath692 .
hence @xmath686 is a submodule condition as in example [ exa : closure_condition ] .
more generally , let @xmath693 be locally closed , that is , @xmath197 is relatively open in its closure @xmath694 .
suppose that the spectral condition for @xmath694 holds for all representations in @xmath605 , say , by previous recursion steps .
then the functional calculus homomorphism for @xmath683 exists and descends to @xmath695 .
the _ spectral condition _ for @xmath197 asks the restriction of this homomorphism to the ideal @xmath696 to be nondegenerate .
this is a submodule condition by example [ exa : nondegeneracy_condition ] .
[ exa : weyl_regular ] let @xmath697 for some @xmath698 be the unital * -algebra generated by two elements @xmath699 with the relations @xmath700 , @xmath701 , @xmath702 .
this is the algebra of polynomial functions on the quantum group @xmath703 .
the c*-algebra of @xmath703 is a c*-hull for a certain class of integrable representations of @xmath697 that is defined by submodule conditions .
first , we require @xmath704 to be a regular , normal operator ; equivalently , @xmath705 and @xmath706 are regular and self - adjoint , and they strongly commute ; these are submodule conditions by examples [ exa : regularity_condition ] and [ exa : strong_commutation ] .
secondly , we require the spectrum of @xmath704 ( or the joint spectrum of its real and imaginary part ) to be contained in @xmath707 ; this is a submodule condition by example [ exa : spectral_condition ] .
finally , we require the closure condition as in example [ exa : closure_condition ] for @xmath708 for every noncommutative polynomial @xmath656 .
then an integrable representation on @xmath15 is equivalent to a pair @xmath709 consisting of a unitary operator @xmath710 and a regular , normal operator @xmath711 on @xmath15 with spectrum contained in @xmath712 , subject to the relation @xmath713 .
the domain of the representation is the intersection of the domains of @xmath714 , @xmath554 , by the closure conditions , and polynomials in @xmath699 act by the corresponding polynomials in @xmath715 on this subspace .
the crossed product c*-algebra for the automorphism @xmath716 on @xmath717 is a c*-hull for this class of integrable representations .
interesting classes of representations defined by submodule conditions occur in theorems [ the : integrable_ug ] and [ the : commutative_regular_admissible ] .
the examples in @xcite are also defined by submodule conditions , compare proposition [ pro : integrable_induced_nice ] .
[ exa : continuity_condition ] if the algebra @xmath13 carries a topology , then we may restrict attention to representations of @xmath13 that are continuous in some sense .
for instance , if @xmath0 is a topological group and @xmath718 $ ] is the group ring of the underlying discrete group , then representations of @xmath13 are unitary representations of @xmath0 , possibly discontinuous . among them
, we may restrict to the continuous representations ( compare the definition of a host algebra for @xmath0 in @xcite ) . if @xmath0 is an infinite - dimensional lie group , we may restrict further to representations of @xmath719 $ ] that are smooth in the sense that the smooth vectors are dense .
i do not expect continuity or smoothness to be a submodule condition , and i do not know when the classes of continuous or smooth representations satisfy the local global principle or its strong variant .
semiboundedness conditions ask for certain ( regular ) self - adjoint operators to be bounded above , see @xcite .
if we specify the upper bound on the spectrum , this is a spectral condition as in example [ exa : spectral_condition ] .
when we let the upper bound go to @xmath720 , however , then direct sums no longer preserve semiboundedness .
therefore , semiboundedness conditions seem close enough to submodule conditions to be tractable , but the details require further thought .
let @xmath648 and @xmath649 be natural constructions of hilbert submodules of rank @xmath442 that define @xmath624 inside @xmath605 , and let @xmath605 be weakly admissible .
let @xmath293 for @xmath128 be representations on hilbert @xmath14-modules @xmath296 for the same c*-algebra @xmath14 that belong to @xmath605 .
let @xmath721 be unitary . if @xmath722 , then @xmath723 .
thus @xmath133 belongs to @xmath624 if @xmath132 does .
this verifies [ enum : admissible0 ] in definition [ def : integrable_admissible ] using [ enum : admissible0 ] in definition [ def : restriction_representation ] .
similarly , [ enum : restriction1 ] and [ enum : restriction2 ] in definition [ def : restriction_representation ] show that @xmath624 inherits [ enum : admissible1 ] and [ enum : admissible2 ] in definition [ def : integrable_admissible ] from @xmath605 .
thus @xmath624 is again weakly admissible .
it is trivial that weak admissibility is hereditary for intersections . by transfinite induction
, it follows that any class of representations defined by submodule conditions is weakly admissible .
let @xmath648 and @xmath649 be natural constructions of hilbert submodules of rank @xmath442 that define @xmath624 inside @xmath605 , and assume that @xmath605 satisfies the strong local global principle .
let @xmath16 be a representation on a hilbert @xmath14-module @xmath15 that does not belong to @xmath624 .
we must find an irreducible representation @xmath250 of @xmath14 on a hilbert space @xmath261 such that @xmath724 does not belong to @xmath624 . if the representation does not even belong to @xmath605 , this is possible because @xmath605 satisfies the strong local global principle by assumption .
so we may assume that @xmath16 belongs to @xmath605 but not to @xmath624 .
thus @xmath668 and @xmath725 are well defined and different hilbert submodules of @xmath726 .
corollary [ cor : hahn - banach_submodules ] gives an irreducible representation @xmath250 of @xmath14 on a hilbert space @xmath261 such that @xmath727 as closed subspaces of @xmath728 .
identify these with subspaces of @xmath729 .
the condition [ enum : restriction1 ] in definition [ def : restriction_representation ] gives @xmath730 that is , @xmath724 does not belong to @xmath624 .
thus @xmath624 inherits the strong local
global principle from @xmath605 .
the strong local global principle is easily seen to be hereditary for intersections .
hence any class of representations defined by submodule conditions satisfies the strong local global principle by transfinite induction . [
the : regular_admissible_local - global ] let @xmath13 be a * -algebra and let @xmath731 .
let @xmath732 be the class of all representations where the elements of @xmath135 act by regular , self - adjoint operators .
this class is defined by submodule conditions and hence satisfies the strong local global principle .
it is admissible if @xmath135 is a strong generating set for @xmath13 .
asking @xmath25 to be regular and self - adjoint for a single @xmath109 is a submodule condition by example [ exa : regularity_condition ] . in order to ask this simultaneously for a set @xmath135 , let @xmath733 be a well - ordering on @xmath135 , and add an element @xmath627 with @xmath734 for all @xmath109 .
let @xmath735 for @xmath736 be the class of all representations @xmath16 where @xmath737 is regular and self - adjoint for all @xmath738 with @xmath739 .
these subclasses form a recursive definition of @xmath732 by submodule conditions as in definition [ def : restriction_representation ] .
thus @xmath732 is defined by submodule conditions .
then it is weakly admissible and satisfies the strong local
global principle by lemma [ lem : operator_conditions ] and theorem [ the : operator_conditions ] . from now on
, we assume that @xmath135 is a strong generating set . for @xmath732 to be admissible
, we must prove that any isometric intertwiner @xmath740 between two hilbert space representations in @xmath732 is a * -intertwiner . if @xmath109 , then @xmath25 and @xmath741 are regular , self - adjoint operators .
hence they generate integrable representations of @xmath37 $ ] as in theorem [ the : regular_self - adjoint_cstar - hull ] .
the isometry @xmath159 intertwines these representations of @xmath37 $ ] .
hence it is a * -intertwiner by theorem [ the : regular_self - adjoint_cstar - hull ] . in particular , @xmath277 maps @xmath99 to @xmath742 for each @xmath109 . since @xmath135 is a strong generating set , gives @xmath743 and similarly for @xmath16 .
so @xmath744 .
then @xmath745 and @xmath159 is a * -intertwiner by proposition [ pro : isometry_star - intertwiner ] .
[ cor : weak_cstar - hull_full ] let @xmath731 be a strong generating set for a * -algebra @xmath13 and let @xmath8 with universal representation @xmath746 be a weak c*-hull . if the closed multipliers @xmath472 for @xmath109 are self - adjoint and affiliated with @xmath8 , then @xmath8 is a c*-hull .
all @xmath8-integrable representations belong to @xmath732 because the latter is weakly admissible and contains the universal @xmath8-integrable representation .
since @xmath732 is admissible by theorem [ the : regular_admissible_local - global ] , any smaller class of integrable representations inherits the equivalent conditions [ enum : cstar - gen2][enum : cstar - gen2 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] , which characterise c*-hulls among weak c*-hulls . [
the : woronowicz_local - global ] let @xmath13 be a * -algebra , @xmath8 a c*-algebra , @xmath329 a representation of @xmath13 on @xmath8 , and @xmath747 .
assume that @xmath748 are self - adjoint and affiliated with @xmath8 and generate @xmath8 in the sense of woronowicz , see @xcite*definition 3.1 .
then @xmath8 is a c*-hull for the @xmath8-integrable representations of @xmath13 defined by @xmath329 , and these satisfy the strong local global principle . to show that @xmath8 is a c*-hull
, we check the condition [ enum : cstar - gen12 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] .
let @xmath749 be a representation of @xmath8 on a hilbert space @xmath261 and let @xmath62 be the corresponding @xmath8-integrable representation of @xmath13 .
let @xmath750 be a @xmath8-integrable representation on a closed subspace @xmath751 and let @xmath752 be the projection onto @xmath753 .
we must show that @xmath360 is contained in the commutant of @xmath391 .
equivalently , @xmath250 is a morphism to the algebra @xmath754 of all compact operators on @xmath15 that commute with @xmath391 .
let @xmath755 .
since @xmath528 is self - adjoint and regular as an adjointable operator on the hilbert @xmath8-module @xmath8 , it generates an integrable representation of the polynomial algebra @xmath37 $ ] as in theorem [ the : regular_self - adjoint_cstar - hull ] .
these integrable representations form an admissible class .
therefore , a @xmath8-integrable representation of @xmath13 gives an integrable representation of @xmath37 $ ] when we compose with the canonical map @xmath756\to a$ ] , @xmath757 . and
since @xmath16 and @xmath758 are both @xmath8-integrable , @xmath759 is a direct summand in @xmath760 .
equivalently , the unbounded operator @xmath761 is affiliated with @xmath762 .
the extension of @xmath250 to affiliated multipliers maps @xmath763 to @xmath761 , which is affiliated with @xmath762 .
hence @xmath250 is a morphism to @xmath762 because these affiliated multipliers generate @xmath8 .
thus @xmath8 is a c*-hull for the @xmath8-integrable representations by proposition [ pro : cstar - generated_by_unbounded_multipliers ] .
now we check the strong local global principle .
let @xmath62 be a representation of @xmath13 on a hilbert @xmath14-module @xmath15 .
assume that the representation @xmath764 is integrable for each irreducible representation @xmath40 of @xmath14 on a hilbert space @xmath544 in the sense that it comes from a representation of @xmath8 .
we must show that the representation @xmath62 is integrable .
the condition that @xmath761 be self - adjoint and regular is a submodule condition by example [ exa : regularity_condition ] .
hence the class of representations with this property satisfies the strong local global principle by theorem [ the : operator_conditions ] .
therefore , @xmath761 is a regular , self - adjoint operator on @xmath15 for @xmath82 .
let @xmath40 be the direct sum of all irreducible representations of @xmath14 ; this is a faithful representation of @xmath14 on some hilbert space @xmath261 .
the induced representation @xmath765 of @xmath411 on @xmath766 is faithful as well . by assumption ,
the representation @xmath767 of @xmath13 on @xmath377 is integrable , so it comes from a representation @xmath768 of @xmath8 .
the extension of @xmath768 to affiliated multipliers maps @xmath769 to @xmath770 .
since @xmath761 is a regular operator on @xmath15 , it is an affiliated multiplier of @xmath411 , see @xcite or proposition [ pro : representation_hilm_comp ] .
thus @xmath770 is affiliated with the image of @xmath411 on @xmath377 by @xcite*proposition 9.10 .
thus @xmath771 for @xmath82 . since the affiliated multipliers @xmath763 generate @xmath8 in the sense of woronowicz , @xmath768 factors through a morphism @xmath772 .
this is the same as a representation of @xmath8 on @xmath15 .
let @xmath263 be the representation of @xmath13 on @xmath15 associated to @xmath773 . if @xmath250 is an irreducible hilbert space representation of @xmath14 , then @xmath774 by construction of @xmath773 .
hence theorem [ the : identify_representations ] gives @xmath479 . since @xmath263 is integrable by construction , so is @xmath16 .
the first counterexample in [ sec : polynomials2 ] exhibits a _
symmetric _ affiliated multiplier that generates a c*-algebra , such that the local global principle fails and @xmath8 is not a c*-hull . without self - adjointness
, we only get the following much weaker statement : [ lem : woronowicz_weak_hull ] let @xmath13 be a * -algebra , @xmath8 a c*-algebra , @xmath329 a representation of @xmath13 on @xmath8 , and @xmath747 .
assume that @xmath748 are affiliated with @xmath8 and generate @xmath8 in the sense of woronowicz .
then @xmath8 is a weak c*-hull for the @xmath8-integrable representations of @xmath13 . to show that @xmath8 is a weak c*-hull
, we check [ enum : cstar - gen1 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] .
let @xmath367 be representations of @xmath8 on a hilbert space @xmath261 with @xmath775 .
we claim that @xmath776 maps @xmath8 into the multiplier algebra of the diagonally embedded copy @xmath762 of @xmath777 .
this is equivalent to @xmath349 . since @xmath775 , the extension of @xmath321 to affiliated multipliers maps @xmath778 to an operator of the form @xmath779 for @xmath82 ; these are affiliated with @xmath762 . since these affiliated multipliers generate @xmath8
, @xmath321 is a morphism from @xmath8 to @xmath762 .
thus @xmath8 is a weak c*-hull for @xmath13 .
we illustrate our theory by an example .
let @xmath780 be a finite - dimensional lie algebra over @xmath111 and let @xmath781 be its universal enveloping algebra with the usual involution , where elements of @xmath780 are skew - symmetric .
let @xmath0 be a simply connected lie group with lie algebra @xmath780 and let @xmath782 . a representation of @xmath783 on a hilbert module @xmath15 is equivalent to a strongly continuous , unitary representation of @xmath0 on @xmath15 .
given such a representation , let @xmath784 be its subspace of smooth vectors .
this is the domain of a closed representation of @xmath90 .
we call a representation of @xmath90 integrable if it comes from a unitary representation of @xmath0 in this way .
in particular , @xmath0 acts continuously on @xmath783 by left multiplication with unitary multipliers .
let @xmath785 be the right ideal of smooth elements for this @xmath0-action , equipped with the canonical @xmath90-module structure @xmath746 .
the pair @xmath329 is the universal integrable representation , that is , a representation of @xmath90 is integrable if and only if it is of the form @xmath786 for a representation @xmath250 of @xmath783 .
let @xmath787 form a basis of @xmath780 .
[ the : integrable_ug ] a representation @xmath418 of @xmath90 is integrable if and only if it satisfies the following two conditions : 1 .
@xmath788 ; 2 .
the laplacian @xmath789 is regular and self - adjoint , where @xmath790 for some basis @xmath787 of @xmath780 .
the first condition says that the * -representation @xmath16 of @xmath90 comes from a representation of the lie algebra @xmath780 , which is the setting considered in @xcite*thorme 2.12 .
the proof shows that the graph norms for @xmath791 , @xmath792 , generate the graph topology .
hence we may replace the first condition in theorem [ the : integrable_ug ] by @xmath793 compare .
the elements @xmath794 form a strong generating set for @xmath90 .
another strong generating set is @xmath795 .
[ the : cstar_hull_lie_algebra ] the class of integrable representations of @xmath90 has @xmath783 as a c*-hull and is defined by submodule conditions .
so it satisfies the strong local global principle .
the closed multipliers of @xmath783 associated to @xmath796 are regular and affiliated with @xmath783 and generate @xmath783 by @xcite*example 3 in 3 .
hence @xmath783 is a c*-hull and the strong local
global principle holds by theorem [ the : woronowicz_local - global ]
. we may also verify [ enum : cstar - gen12 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] by hand .
let @xmath797 be a hilbert space representation of @xmath0 , coming from a unitary representation of @xmath0 .
let @xmath350 be the associated representation of @xmath90 and let @xmath358 be an integrable subrepresentation on a closed subspace @xmath359 .
we must show that the projection onto @xmath359 commutes with @xmath8 or , equivalently , is @xmath0-equivariant .
theorem [ the : regular_self - adjoint_cstar - hull ] implies that [ enum : cstar - gen12 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] holds for the c*-hull @xmath798 of @xmath37 $ ] . for each @xmath799 , the operator @xmath800 is self - adjoint .
therefore , the projection onto @xmath359 commutes with @xmath801 for all @xmath444 .
the elements of @xmath0 of the form @xmath802 with @xmath803 , @xmath799 generate a neighbourhood of the unit element .
so they generate @xmath0 because @xmath0 is connected .
hence the projection onto @xmath359 is @xmath0-equivariant .
this shows that @xmath783 is a c*-hull for the class of representations of @xmath90 that are induced by a representation of @xmath783 ; these are exactly the integrable representation of @xmath90 .
pierrot s characterisation of the integrable representations in theorem [ the : integrable_ug ] shows that this class of representations is defined by operator conditions . the first condition in theorem [ the : integrable_ug ] holds if and only if @xmath804 for all noncommutative polynomials @xmath656 .
these are submodule conditions by example [ exa : closure_condition ] .
the second condition is a regularity condition as in example [ exa : regularity_condition ] . hence the strong local
global principle for the integrable representations of @xmath90 also follows from theorem [ the : operator_conditions ] .
the results of vassout @xcite get close to proving an analogue of theorem [ the : cstar_hull_lie_algebra ] for a lie groupoid @xmath0 with compact base .
this analogue would replace @xmath780 by the space of smooth sections of the lie algebroid @xmath805 , and @xmath90 by the * -algebra of @xmath0-equivariant differential operators on @xmath0 , a subalgebra of the * -algebra of @xmath0-pseudodifferential operators .
any symmetric , elliptic element of @xmath90 should be a possible replacement for the laplacian in theorem [ the : cstar_hull_lie_algebra ] .
we discuss two classes of `` integrable '' representations of the * -algebra @xmath37 $ ] with @xmath35 that are weakly admissible , but not admissible , and violate the local global principle .
both examples have a weak c*-hull , on which all powers of the generator @xmath257 act by an affiliated multiplier . in the first example
, these affiliated multipliers generate the weak c*-hull , but not in the second . neither theorem [ the : regular_admissible_local - global ] nor theorem [ the : woronowicz_local - global ] apply because the generating affiliated multipliers are not self - adjoint .
the first example shows that a c*-algebra generated by affiliated multipliers in the sense of woronowicz need not be a c*-hull , though it is always a weak c*-hull by lemma [ lem : woronowicz_weak_hull ] .
the second example shows that a weak c*-hull need not be generated by affiliated multipliers .
let @xmath806 be the unilateral shift .
let @xmath807 be the closed symmetric operator on @xmath808 with cayley transform @xmath135 .
thus @xmath807 has deficiency index @xmath809 .
the domain of @xmath807 is @xmath810 , and @xmath811 for all @xmath812 ( see also example [ exa : deficiency_index_0 ] ) .
we may identify @xmath808 with the hardy space @xmath813 .
then @xmath807 becomes the toeplitz operator with the unbounded symbol @xmath814 .
let @xmath815 be the toeplitz c*-algebra , that is , the c*-subalgebra of @xmath816 generated by @xmath135 .
every element in @xmath815 is of the form @xmath817 , where @xmath818 is the toeplitz operator with symbol @xmath819 and @xmath762 is a compact operator .
let @xmath820 be the kernel of the unique * -homomorphism @xmath821 that maps @xmath135 to @xmath314 .
[ pro : toeplitz_generated ] there is a symmetric , affiliated multiplier @xmath807 of @xmath822 with domain @xmath823 and @xmath824 for all @xmath825 .
it generates @xmath822 in the sense of woronowicz .
we claim that the right ideal @xmath826 is dense .
this would fail for @xmath815 because the continuous * -homomorphism @xmath821 , @xmath827 , annihilates this right ideal .
first , @xmath828 is dense in @xmath829 because @xmath830 has dense range on @xmath808 .
so the closure of @xmath831 contains @xmath829 .
secondly , @xmath832 is dense in @xmath833 because the function @xmath834 on @xmath835 vanishes only at @xmath314 .
an affiliated multiplier of @xmath822 is the same as a regular operator on @xmath822 , viewed as a hilbert module over itself .
since @xmath831 is dense in @xmath822 , there is a regular , symmetric operator @xmath836 on @xmath822 that has @xmath135 as its cayley transform , see @xcite*chapter 10 .
the operator @xmath836 has the domain @xmath831 and acts by @xmath837 . rewriting any @xmath825 as @xmath838 , we may replace @xmath839 by @xmath86 here .
thus @xmath840 .
since @xmath841 maps @xmath842 to @xmath843 , it is surjective , and @xmath844 belongs to @xmath822 .
hence @xmath845 is the inverse of @xmath846 .
so @xmath847 has domain @xmath848 and maps @xmath849 .
as expected , @xmath847 contains @xmath807 : we may write @xmath850 , and @xmath847 maps this to @xmath851 .
next we show that @xmath852 is the inverse of @xmath853 .
we compute @xmath854 this implies @xmath855 .
since this is already surjective and @xmath856 is injective , the domain of @xmath852 is exactly @xmath857 , and @xmath858 is the inverse of @xmath859 as asserted .
let @xmath246 and @xmath247 be two hilbert space representations of @xmath822 whose extension to affiliated multipliers maps @xmath807 to the same unbounded operator .
then they also map the cayley transform @xmath135 of @xmath807 to the same partial isometry .
so @xmath860 , which gives @xmath349 .
thus @xmath807 separates the representations of @xmath822 .
since @xmath861 as well , @xcite*theorem 3.3 shows that the affiliated multiplier @xmath807 generates @xmath822 .
the domain of @xmath862 is the right ideal @xmath863 , which is dense in @xmath822 for the same reason as @xmath823 .
even more , the right ideal @xmath864 is dense in @xmath863 in the graph norm of @xmath862 . thus the intersection @xmath865 of this decreasing chain of dense right ideals @xmath866 is still dense in @xmath822 by @xcite*lemma 1.1.2 .
this intersection is the domain of a closed representation @xmath746 of @xmath37 $ ] on @xmath822 with @xmath867 .
we call a representation of @xmath37 $ ] on a hilbert module @xmath15 _ toeplitz integrable _ if it is of the form @xmath868 for some representation @xmath869 .
[ pro : toeplitz_integrable ] the class of toeplitz integrable representations of @xmath37 $ ] is weakly admissible with the weak c*-hull @xmath822 .
it is not admissible , so @xmath822 is not a c*-hull .
the toeplitz integrable representations violate the local global principle .
a representation @xmath62 of @xmath37 $ ] on a hilbert module @xmath15 over a c*-algebra @xmath14 is toeplitz integrable if and only if it has the following properties : 1 .
@xmath465 for all @xmath439 ; 2 .
@xmath870 ; 3 .
@xmath38 is regular .
toeplitz integrable representations on @xmath15 are in bijection with regular , symmetric operators @xmath41 on @xmath15 for which @xmath871 is surjective .
we checked condition [ enum : cstar - gen1 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] in the proof of proposition [ pro : toeplitz_generated ] .
thus @xmath822 is a weak c*-hull for the toeplitz integrable representations , and this class is weakly admissible .
any self - adjoint operator on a hilbert space generates a toeplitz integrable representation of @xmath37 $ ] because @xmath872 ; so does @xmath807 itself .
thus both example [ exa : deficiency_index_0 ] and proposition [ pro : symmetric_no_hull ] show that the class of toeplitz integrable representations is not admissible .
so @xmath822 is not a c*-hull .
the representation @xmath873 of @xmath37 $ ] has the three properties in the proposition .
this is inherited by all toeplitz integrable representations by lemma [ lem : operator_conditions ] for the properties ( 1 ) and ( 2 ) and by @xcite*proposition 9.10 for regularity .
hence these properties are necessary for a representation to be toeplitz integrable .
conversely , let @xmath62 be a representation on @xmath15 with the properties ( 1)(3 ) .
the set @xmath874 is a strong generating set for @xmath37 $ ] by lemma [ lem : graph_topology_cx ] .
since assumption ( 1 ) gives @xmath465 for all @xmath439 , proposition [ pro : equality_if_closure_equal ] shows that @xmath62 is determined by the single closed , symmetric operator @xmath875 .
since @xmath41 is regular , its cayley transform @xmath876 is an adjointable partial isometry such that @xmath877 has dense range ( see @xcite*chapter 10 ) .
even more , @xmath876 is an isometry because @xmath878 .
thus @xmath876 generates a unital representation @xmath250 of @xmath815 .
it remains nondegenerate on the ideal @xmath822 because @xmath877 has dense range .
the extension of @xmath250 to affiliated multipliers maps @xmath807 to @xmath41 , the inverse cayley transform of @xmath876 .
hence the representation @xmath879 of @xmath37 $ ] is the one associated to @xmath41 .
this finishes the proof that toeplitz integrable representations of @xmath37 $ ] are characterised by the properties ( 1)(3 ) and that they are in bijection with regular , symmetric operators with surjective @xmath871 . for a counterexample to the local
global principle , we let @xmath880 be the one - point compactification of @xmath881 and @xmath882 .
let @xmath883 consist of all continuous functions @xmath884 with @xmath885 .
the unilateral shift @xmath135 on @xmath886 restricts to a non - adjointable isometry @xmath876 on this subspace .
let @xmath41 be the inverse cayley transform of @xmath876 .
this is a closed , symmetric operator on @xmath15 that is irregular because its cayley transform is not adjointable .
if @xmath575 is a hilbert space representation , then the induced representation of @xmath37 $ ] is associated to the closed operator @xmath887 .
the operator @xmath888 remains surjective , and @xmath887 is regular because it acts on a hilbert space .
so @xmath889 generates a toeplitz integrable representation for all representations @xmath250 of @xmath14 . since @xmath41 itself does not generate a toeplitz integrable representation , the local
global principle is violated .
the first two conditions for toeplitz integrability are submodule conditions by example [ exa : closure_condition ] and example [ exa : nondegeneracy_condition ] .
regularity without self - adjointness , however , is not a submodule condition ( see example [ exa : alone_regularity ] ) .
the failure of the local
global principle in this case does not contradict theorem [ the : woronowicz_local - global ] because @xmath890 is not self - adjoint .
the identical inclusion @xmath891 is a representation of the weak c*-hull @xmath822 on @xmath829 and thus corresponds to a toeplitz integrable representation of @xmath37 $ ] on @xmath829 .
this is simply the restriction of @xmath873 to the hilbert @xmath822-submodule @xmath892 , with domain @xmath893 and the same action @xmath746 of @xmath37 $ ] . call a representation _ purely toeplitz integrable _ if it is of the form @xmath894 for some representation @xmath895 .
[ pro : comp_integrable_cx ] a representation @xmath62 of @xmath37 $ ] on a hilbert module @xmath15 over a c*-algebra @xmath14 is purely toeplitz integrable if and only if , in addition to the three properties in proposition , it has the following property : 1 .
the closure of @xmath896 is @xmath15 .
the purely toeplitz integrable representations of @xmath37 $ ] form a weakly admissible class that is not admissible , and @xmath829 is a weak c*-hull for it , but not a c*-hull .
this class violates the local global principle .
the closed multiplier @xmath897 of @xmath822 is affiliated with @xmath829 but does not generate @xmath829 .
since @xmath829 has fewer representations than @xmath822 , the condition [ enum : cstar - gen1 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] for @xmath829 follows from the corresponding property for @xmath822 , which we have already checked in the proof of proposition [ pro : toeplitz_generated ] .
hence @xmath829 is a weak c*-hull for the purely toeplitz representations of @xmath37 $ ] .
since @xmath807 gives a purely toeplitz representations of @xmath37 $ ] on @xmath898 , the class of purely toeplitz integrable representations is not admissible by example [ exa : deficiency_index_0 ] .
therefore , its weak c*-hull is not a c*-hull .
the same counterexample as in the proof of proposition [ pro : toeplitz_integrable ] shows that the local
global principle fails for the purely toeplitz representations .
any closed operator on @xmath808 is affiliated with @xmath829 .
in particular , so is @xmath807 . in the identical representation of @xmath829 on the hilbert space @xmath808 ,
the image of @xmath807 is affiliated with @xmath822 by proposition [ pro : toeplitz_generated ] .
but the representation of @xmath829 is not by a morphism to @xmath822 because the inclusion map @xmath899 is degenerate .
hence @xmath807 does not generate @xmath829 in the sense of woronowicz .
the element @xmath900 is the orthogonal projection onto the span of @xmath901 . a representation of @xmath822 maps @xmath902 to an orthogonal projection whose image is the orthogonal complement of the image of @xmath903 .
this is also the orthogonal complement of the image of @xmath904 .
these orthogonal complements form an increasing chain of complementable submodules , and @xmath16 is purely toeplitz if and only if their union is all of @xmath15 .
this proves our characterisation of purely toeplitz representations .
let @xmath13 be a * -algebra .
a _ bounded representation _ of @xmath13 on a hilbert module @xmath15 is a * -homomorphism @xmath905 .
corollary [ cor : bounded_rep ] says that a closed representation is bounded once @xmath25 is globally defined for @xmath98 in a strong generating set of @xmath13 .
finite - dimensional representations are always bounded .
in particular , characters are bounded .
thus commutative * -algebras have many bounded representations .
many other * -algebras , such as the weyl algebra , have no bounded representations . in this section
, we are going to study c*-hulls related to bounded representations .
these are only relevant if @xmath13 has many bounded representations .
any bounded representation @xmath16 of @xmath13 is bounded in some c*-seminorm @xmath906 on @xmath13 , that is , @xmath907 for all @xmath20 .
then @xmath16 extends to the ( hausdorff ) completion @xmath908 of @xmath13 in the seminorm @xmath906 , which is a unital c*-algebra . if @xmath72 are two c*-seminorms on @xmath13 , then @xmath909 is a c*-seminorm as well .
thus the set @xmath910 of c*-seminorms on @xmath13 is directed . for @xmath911 with @xmath912 ,
let @xmath913 be the * -homomorphism induced by the identity map on @xmath13 .
the c*-algebras @xmath908 and the * -homomorphisms @xmath914 for @xmath912 in @xmath910 form a projective system of c*-algebras .
each * -homomorphism @xmath915 is unital and surjective because its image contains @xmath13 , which is unital and dense in @xmath916 . the c*-seminorms in
@xmath910 define a locally convex topology on @xmath13 , where a net converges if and only if it converges in any c*-seminorm .
let @xmath917 with the canonical map @xmath918 be the completion of @xmath13 in this topology .
this is a c*-algebra if and only if there is a _ largest _
c*-seminorm on @xmath13 . in general
, @xmath917 is the projective limit of the diagram of unital c*-algebras @xmath919 described above .
thus @xmath917 is a unital pro - c*-algebra , see @xcite . as a concrete example , we describe @xmath917 for a commutative * -algebra @xmath13 .
[ def : characters ] let @xmath32 be the set of * -homomorphisms @xmath920 , which we briefly call _
characters_. each @xmath20 gives a function @xmath921 , @xmath922 .
we equip @xmath32 with the coarsest topology making these functions continuous .
that is , a net @xmath923 in @xmath32 converges to @xmath924 if and only if @xmath925 for all @xmath20 .
let @xmath926 be the _ compactly generated topology _ associated to this topology , that is , a subset in @xmath32 is closed in @xmath926 if and only if its intersection with any compact subset in @xmath32 is closed .
if @xmath20 , then its _ gelfand transform _ @xmath927 is a continuous function on @xmath32 .
this defines a * -homomorphism @xmath928 .
if the usual topology on @xmath32 is locally compact or metrisable , then it is already compactly generated and hence equal to @xmath926 .
in general , the topology @xmath926 has more closed subsets and hence may have more continuous functions to @xmath929 , so @xmath930 . [
pro : commutative_pro - completion ] let @xmath13 be a commutative * -algebra .
the directed set @xmath910 of c*-seminorms on @xmath13 is isomorphic to the directed set of compact subsets of @xmath32 , where @xmath931 corresponds to the c*-seminorm @xmath932 the c*-completion of @xmath13 in this c*-seminorm is @xmath933 . and
@xmath934 , where the inclusion map @xmath935 is the gelfand transform @xmath936 , @xmath937 .
let @xmath906 be a c*-seminorm on @xmath13 .
let @xmath938 be the subspace of all @xmath906-bounded characters , that is , @xmath939 if and only if @xmath940 for all @xmath20 .
these are precisely the characters that extend to characters on the c*-completion @xmath908 .
conversely , since @xmath13 is dense in @xmath908 , any character on @xmath908 is the unique continuous extension of a @xmath906-bounded character on @xmath13 . and the subspace topology on @xmath941 is equal to the canonical topology on the spectrum of @xmath908 : a net of @xmath906-bounded characters that converges uniformly on @xmath13 also converges uniformly on @xmath908 . thus @xmath942 by the gelfand
naimark theorem , and so @xmath941 is compact for each @xmath943 . if @xmath912 , then @xmath944 and @xmath945 is the restriction map for the subspace @xmath946 .
the pro - c*-algebra @xmath917 is the limit of this diagram of commutative c*-algebras .
since all the maps @xmath944 are injective , @xmath917 is the algebra of continuous functions on @xmath947 , equipped with the inductive limit topology .
that is , a subset of @xmath947 is closed if and only if its intersection with each @xmath948 is closed , where @xmath948 carries the ( compact ) subspace topology from @xmath32 .
any character @xmath949 on @xmath13 is bounded with respect to some c*-seminorm ; for instance , @xmath950 .
thus @xmath951 as a set .
if @xmath931 is compact , then @xmath952 for @xmath20 must be uniformly bounded on @xmath762 , so that @xmath953 is a c*-seminorm on @xmath13 .
thus @xmath954 for some @xmath943 .
hence the inductive limit topology on @xmath947 is @xmath926 .
we return to the general noncommutative case . the class of @xmath906-bounded representations for a fixed @xmath943 is easily seen to be weakly admissible .
the class of bounded representations with variable @xmath906 is not weakly admissible unless @xmath13 has a largest c*-seminorm because it is not closed under direct sums .
we are going to define the larger class of `` locally bounded '' representations to rectify this . roughly speaking , a representation is locally bounded if and only if it comes from a representation of the pro - c*-algebra @xmath917 .
before we define locally bounded representations , we characterise @xmath906-bounded representations by some slightly weaker estimates .
[ pro : bounded_element_fixed_seminorm ] let @xmath13 be a * -algebra and let @xmath906 be a c*-seminorm on @xmath13 .
let @xmath62 be a representation of @xmath13 on a hilbert module @xmath15 over some c*-algebra @xmath14 and let @xmath81 .
the following are equivalent : 1 .
there is @xmath143 with @xmath955 for all @xmath20 ; 2 .
there is @xmath143 with @xmath956 for all @xmath20 ; 3 .
@xmath957 for all @xmath20 .
the set of vectors @xmath114 with these equivalent properties is a norm - closed @xmath66-submodule of @xmath15 .
the representation of @xmath13 on this submodule extends to the c*-completion @xmath908 .
the implications ( 3)@xmath272(2)@xmath272(1 ) are trivial .
conversely , assume ( 1 ) and let @xmath20 .
let @xmath958 be a sequence in @xmath13 that converges in @xmath908 towards the positive square - root of @xmath959 .
then the sequence @xmath960 in @xmath13 converges in the norm @xmath906 to @xmath961 .
if ( 1 ) holds , then @xmath962 since @xmath963 for all @xmath442 , this implies @xmath964 .
thus ( 1 ) implies ( 3 ) .
the set @xmath965 of vectors @xmath81 satisfying ( 2 ) is a vector subspace and closed under left multiplication by elements of @xmath13 and right multiplication by elements of @xmath14 . on this subspace ,
the graph and norm topologies coincide because of ( 3 ) .
the subspace @xmath965 is closed in the norm topology by the principle of uniform boundedness .
the * -representation of @xmath13 on this submodule is globally defined and bounded by the c*-seminorm @xmath906 .
hence it extends to a representation of @xmath908 .
[ def : locally_bounded_representation ] let @xmath62 be a representation of @xmath13 on a hilbert module @xmath15 .
a vector @xmath81 is _ bounded _ if it satisfies the equivalent conditions in proposition [ pro : bounded_element_fixed_seminorm ] for some @xmath943 .
the representation is _ locally bounded _ if the bounded vectors are dense in @xmath21 in the graph topology . by proposition [ pro : bounded_element_fixed_seminorm ] , the @xmath906-bounded vectors in @xmath21 for a fixed @xmath966 form a closed @xmath66-submodule @xmath967 , on which the representation of @xmath13 extends to the c*-completion @xmath908 and hence to a representation of @xmath917 . since @xmath910 is directed and @xmath968 if @xmath912 , the family of sub - bimodules @xmath969 is directed .
the set of bounded vectors is the increasing union @xmath970 since @xmath971 extends to @xmath917 for each @xmath906 , there is a representation @xmath972 of the pro - c*-algebra @xmath917 on @xmath973 .
the representation @xmath62 is locally bounded if and only if @xmath974 is a core for it .
thus @xmath62 is the closure of the `` restriction '' @xmath975 of @xmath972 to @xmath13 .
we do not claim that @xmath972 is closed , and neither do we claim that @xmath975 is locally bounded for any representation of @xmath917 : we need the representation of @xmath917 to be locally bounded as well .
[ def : pro_locally_bounded ] a representation @xmath418 of a pro - c*-algebra @xmath917 is _ locally bounded _ if the vectors @xmath81 for which @xmath976 , @xmath977 , is continuous form a core .
[ pro : locally_bounded_pro - cstar ] composition with @xmath918 induces an equivalence between the categories of locally bounded representations of @xmath13 and @xmath917 , which is compatible with isometric intertwiners and interior tensor products .
a representation of @xmath917 on @xmath15 is locally bounded if and only if the vectors @xmath114 for which the map @xmath978 , @xmath979 , is continuous form a core .
the * -homomorphism @xmath765 induces an isomorphism between the directed sets of c*-seminorms on @xmath13 and @xmath917 .
therefore , a representation @xmath972 of @xmath917 is locally bounded if and only if the vectors for which @xmath980 for all @xmath981 , for some @xmath943 , form a core .
since the image of @xmath765 is dense , this is equivalent to @xmath982 for all @xmath20 .
thus the closure of @xmath975 is locally bounded if and only if @xmath972 is locally bounded .
an isometric intertwiner @xmath983 also intertwines the closures of @xmath984 and @xmath985 by lemma [ lem : closure_functorial ] .
conversely , an isometric intertwiner between two locally bounded representations of @xmath13 must map @xmath906-bounded vectors to @xmath906-bounded vectors for any @xmath943 and thus remains an isometric intertwiner between the canonical extensions of the representations to @xmath917 .
thus the equivalence between the locally bounded representations of @xmath13 and @xmath917 is compatible with isometric intertwiners .
it is also compatible with interior tensor products , that is , the closure of @xmath986 is @xmath987 .
[ pro : locally_bounded_irreducible ] all irreducible , locally bounded hilbert space representations are bounded .
if @xmath16 is irreducible , then the closed @xmath13-submodule @xmath965 for a c*-seminorm @xmath906 is either @xmath988 or @xmath15 .
the latter must happen for some @xmath906 if @xmath16 is locally bounded .
thus local boundedness is not an interesting notion for irreducible representations .
if @xmath13 has no c*-seminorms , then @xmath989 and @xmath13 has no locally bounded representations , so that the following discussion will be empty . even if the map @xmath990 is not injective , there are examples where all `` integrable '' representations of @xmath13 come from @xmath917 .
an important case is the unit fibre for the canonical @xmath59-grading on the virasoro algebra in @xcite*9.3 . in this case , @xmath13 is not commutative , but all irreducible , integrable representations are characters and hence locally bounded .
[ pro : locally_bounded_regular ] if @xmath16 is a locally bounded representation , then @xmath25 is regular and self - adjoint for each @xmath991 .
the map of left multiplication by @xmath992 on @xmath917 is invertible because @xmath993 is symmetric and @xmath917 is a pro - c*-algebra .
therefore , @xmath994 have dense range on @xmath15 .
thus @xmath25 is regular and self - adjoint , see @xcite*chapter 10 .
[ cor : locally_bounded_admissible ] let @xmath13 be a * -algebra .
the class @xmath995 of locally bounded representations of @xmath13 is admissible .
being locally bounded is clearly invariant under unitary * -intertwiners and direct sums .
since a * -intertwiner maps bounded vectors to bounded vectors , it is also invariant under direct summands .
if @xmath340 is bounded , then @xmath996 is bounded for any c*-correspondence @xmath30 .
thus a locally bounded representation on @xmath15 induces one on @xmath997 . since @xmath119 is a strong generating set for @xmath13 by example [ exa : strong_generating ] , the class of representations for which all @xmath123 act by a regular and self - adjoint operator is admissible by theorem [ the : regular_admissible_local - global ] . this class contains the locally bounded representations by proposition [ pro : locally_bounded_regular ] .
hence this subclass is also admissible .
any pro - c*-algebra @xmath917 contains a dense unital c*-subalgebra @xmath998 of bounded elements , see @xcite*proposition 1.11 .
for instance , if @xmath13 is commutative , so that @xmath999 by proposition [ pro : commutative_pro - completion ] , then @xmath1000 consists of the _ bounded _ continuous functions .
let @xmath62 be a locally bounded representation of @xmath13 .
this comes from a locally bounded representation @xmath1001 of @xmath917 by proposition [ pro : locally_bounded_pro - cstar ] .
the closure of the restriction of @xmath972 to @xmath998 is a representation of a unital c*-algebra .
hence it is a unital * -homomorphism @xmath1002 by lemma [ lem : cstar - rep_bounded ] .
[ pro : equality_locally_bounded ] two locally bounded representations @xmath132 and @xmath133 of @xmath13 on a hilbert module @xmath15 are equal if and only if they induce the same representation of @xmath998 . of course , @xmath132 and @xmath133 induce the same representation of @xmath998 if @xmath137 .
conversely , assume that @xmath132 and @xmath133 induce the same representation @xmath250 of @xmath998 .
if @xmath991 , then the cayley transform @xmath1003 of @xmath1004 is a unitary element of @xmath998 .
the cayley transforms of @xmath170 and @xmath171 are both equal to @xmath1005 . hence @xmath1006 .
since this holds for all @xmath123 , proposition [ pro : equality_if_closure_equal ] gives @xmath137 .
the c*-algebra @xmath998 usually has many representations that do not come from locally bounded representations of @xmath13 .
hence it is not a c*-hull .
it is , however , a useful tool to decide when a representation @xmath746 of @xmath13 on a c*-algebra @xmath8 is a weak c*-hull , that is , when @xmath13 separates the hilbert space representations of @xmath8 : [ pro : criterion_cstar - hull_locally_bounded ] let @xmath746 be a locally bounded representation of @xmath13 on a c*-algebra @xmath8 and let @xmath1007 be the associated representation of @xmath998 .
the image of @xmath250 is dense in @xmath1008 in the strict topology if and only if @xmath8 is a weak c*-hull for the class of @xmath8-integrable representations of @xmath13 defined by @xmath746 .
combine proposition [ pro : equality_locally_bounded ] and the following proposition for @xmath1009 .
[ pro : cstar - hull_bounded_criterion ] let @xmath746 be a representation of @xmath13 on a c*-algebra @xmath8 . let @xmath14 be a c*-algebra and @xmath1010 a * -homomorphism .
assume that two representations @xmath367 of @xmath8 on a hilbert space @xmath261 satisfy @xmath1011 if and only if @xmath1012 .
then @xmath8 is a weak c*-hull for a class of integrable representations of @xmath13 if and only if @xmath1013 is dense in @xmath1008 in the strict topology .
we use the criterion for weak c*-hulls in [ enum : cstar - gen1 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] . by assumption , @xmath1014 if and only if @xmath1015 . thus we have to show that @xmath387 has strictly dense range if and only if @xmath1016 implies @xmath1017 for all hilbert space representations @xmath367 . here
@xmath1018 and @xmath1019 involve the unique extensions of @xmath246 and @xmath247 to @xmath1008 , which are strictly continuous .
hence @xmath1020 implies @xmath1017 on @xmath1008 if @xmath1013 is strictly dense in @xmath1008 .
conversely , let @xmath347 be representations with @xmath1021 and @xmath1022 . then there are @xmath375 and @xmath1023 with @xmath1024 .
the linear functional @xmath1025 on @xmath1008 is non - zero and strictly continuous , and it vanishes on @xmath1013 because @xmath1026 .
hence @xmath1013 is not strictly dense in @xmath1008 .
[ rem : criterion_density_weak - cstar - hull ] proposition [ pro : cstar - hull_bounded_criterion ] applies whenever we can somehow produce enough bounded operators from a representation of @xmath13 so that these bounded operators and the original representation have the same unitary * -intertwiners . for instance , it also applies if the elements of a strong generating set for @xmath13 act by regular operators , so that we may take their bounded transforms .
the quotient map @xmath1027 for @xmath1028 in @xmath910 identifies the primitive ideal space @xmath1029 with a closed subspace of @xmath1030 .
let @xmath1031 and equip it with the inductive limit topology .
let @xmath1032 and @xmath1033 .
then the norm @xmath1034 of the image of @xmath98 in @xmath1035 is the same for all @xmath943 with @xmath1036 .
hence the function @xmath1037 on @xmath1038 is well defined .
[ def : compact_in_pro - cstar ] an element @xmath1032 is _ vanishes at @xmath720 _ if for every @xmath1039 there is @xmath943 such that @xmath1040 for @xmath1041 .
an element @xmath1032 is _ compactly supported _ if there is @xmath943 with @xmath1042 for all @xmath1043 .
let @xmath1044 and @xmath1045 be the subsets of elements that vanish at @xmath720 and have compact support , respectively .
it may happen that @xmath1046 . in the following , we are interested in the case where @xmath1044 is dense in @xmath917 .
for instance , @xmath36 is dense in @xmath1047 .
[ lem : compactly_supported_ideal ] the subset @xmath1044 is a c*-subalgebra of @xmath998 .
the subspace @xmath1045 is a two - sided * -ideal in @xmath917 .
it is norm - dense in @xmath1044 .
more generally , if @xmath14 is a c*-algebra and @xmath1048 is a * -homomorphism , then @xmath1049 is dense in @xmath1050 .
the quotient maps @xmath1051 for @xmath1052 are * -homomorphisms .
thus @xmath1044 is a * -subalgebra of @xmath917 .
an element @xmath1053 is bounded if and only if the norms of its images in @xmath908 for @xmath943 are uniformly bounded .
the norm of @xmath98 in @xmath908 is the maximum of @xmath1034 for @xmath1052 .
hence @xmath98 is bounded if and only if the function @xmath1034 on @xmath1038 is bounded .
thus @xmath1044 consists of bounded elements , and it is an ideal in @xmath998 .
we claim that the limit @xmath98 of a norm - convergent sequence @xmath1054 in @xmath1044 again vanishes at @xmath720 .
given @xmath1039 , there is @xmath1055 so that @xmath1056 for all @xmath1057 . since @xmath1058 vanishes at @xmath720
, there is @xmath943 with @xmath1059 for @xmath1060 .
thus @xmath1040 for @xmath1061 .
thus @xmath1044 is a closed ideal in @xmath998 .
the condition @xmath1042 for fixed @xmath1062 defines a closed two - sided * -ideal in @xmath917 .
hence @xmath1045 is a two - sided * -ideal in @xmath917 .
let @xmath1063 and @xmath1039 .
let @xmath1064 be increasing and satisfy @xmath1065 for @xmath1066 and @xmath1067 for @xmath1068 .
then @xmath1069 , and @xmath1070 if @xmath1071 .
hence @xmath1072 for all @xmath1039 .
thus @xmath1045 is dense in @xmath1044 .
similarly , if @xmath1048 is a * -homomorphism , @xmath1073 , and @xmath1074 , then @xmath1075 and @xmath1076 .
[ the : prim_a_locally_compact_gives_hull ] let @xmath13 be a * -algebra and let @xmath917 be its pro - c*-algebra completion . if @xmath1044 is dense in @xmath917 , then @xmath1044 is a c*-hull for the class of locally bounded representations of @xmath13 .
we shall prove a more general theorem that still applies if @xmath1044 is not dense in @xmath917 . then probably there is no c*-hull for the class of all locally bounded representations .
we may , however , find c*-hulls for smaller classes of representations .
we describe such classes of representations by a generalisation of the spectral conditions in example [ exa : spectral_condition ] .
the spectral condition for a locally closed subset in @xmath1077 implicitly uses a subquotient of @xmath1078 .
we are going to describe subquotients of the pro - c*-algebra @xmath917 .
we then associate a class @xmath1079 of locally bounded representations of @xmath13 to a subquotient @xmath1080 .
if @xmath1081 is dense in @xmath1080 , then @xmath1081 is a c*-hull for @xmath1079 .
theorem [ the : prim_a_locally_compact_gives_hull ] is the special case @xmath1082 .
let @xmath1083 be a closed , two - sided * -ideal in the pro - c*-algebra @xmath917 .
being closed , the ideal @xmath1084 is complete in the subspace topology , so it is also a pro - c*-algebra .
thus @xmath1085 , where @xmath1086 is the image of @xmath1084 in the quotient @xmath908 .
the quotient @xmath1087 is complete if @xmath917 is metrisable , that is , its topology is defined by a sequence of c*-seminorms .
it need not be complete in general , however .
therefore , we replace the quotient @xmath1087 by its completion @xmath1088 , which is a pro - c*-algebra as well .
it is the projective limit of the quotients @xmath1089 for all @xmath966 .
subquotient _ of @xmath917 is a closed , two - sided * -ideal @xmath1090 with @xmath1088 as above .
let @xmath1079 consist of all representations @xmath16 of @xmath13 with the following properties : 1 .
@xmath16 is locally bounded , so it comes from a locally bounded representation @xmath263 of @xmath917 ; 2 . the representation @xmath263 annihilates @xmath1084 ; 3 .
the representation @xmath972 of @xmath1088 induced by @xmath263 is nondegenerate on @xmath1080 , that is , @xmath1091 is a core for @xmath972 .
define the c*-algebra @xmath1081 and its dense ideal @xmath1092 by replacing @xmath917 by @xmath1080 in the definitions above .
equivalently , @xmath1093
. we may choose @xmath1094 and @xmath1082 .
then @xmath1095 simply consists of all locally bounded representations of @xmath13 . hence theorem
[ the : prim_a_locally_compact_gives_hull ] is the special case @xmath1082 of the following theorem : [ the : locally_bounded_relative_hull ] if @xmath1081 is dense in @xmath1080 , then @xmath1081 is a c*-hull for @xmath1079 .
first we claim that @xmath1079 is equivalent to the class of nondegenerate , locally bounded representations of the pro - c*-algebra @xmath1080 as in definition [ def : pro_locally_bounded ] . if @xmath1096 , this is proposition [ pro : locally_bounded_pro - cstar ] .
a locally bounded representation @xmath263 of @xmath917 descends to a representation @xmath1097 of the quotient @xmath1087 if and only if it annihilates @xmath1084 ; the induced representation of @xmath1087 remains locally bounded with respect to the family of c*-seminorms from the quotient mappings @xmath1098 .
hence it extends uniquely to a locally bounded representation @xmath1099 of the completion @xmath1088 .
thus locally bounded representations of @xmath13 for which the corresponding representation of @xmath917 annihilates @xmath1084 are equivalent to locally bounded representations of @xmath1088 .
we claim that a nondegenerate , locally bounded representation @xmath250 of @xmath1080 extends uniquely to @xmath1088 .
let @xmath906 be a continuous seminorm on @xmath1088 , also write @xmath906 for its restriction to @xmath1080 .
the @xmath906-bounded vectors for @xmath250 form a nondegenerate @xmath1100-module .
the module structure extends uniquely to the multiplier algebra of @xmath1100 , and @xmath1101 maps to this multiplier algebra because @xmath1102 .
letting @xmath906 vary gives a locally bounded representation of @xmath1088 that remains nondegenerate on @xmath1080 .
conversely , any such representation @xmath1088 is obtained in this way from its restriction to @xmath1080 .
thus representations of @xmath13 that belong to @xmath1079 are equivalent to nondegenerate , locally bounded representations of the pro - c*-algebra @xmath1080 .
the equivalence above is compatible with isometric intertwiners , direct sums and interior tensor products , compare proposition [ pro : locally_bounded_pro - cstar ] .
lemma [ lem : compactly_supported_ideal ] shows that @xmath1103 is dense in @xmath1080 .
this is an ideal in @xmath1088 as an intersection of two ideals .
hence left multiplication defines a representation of @xmath1088 on @xmath1080 with core @xmath1092 , which is locally bounded by construction . through the canonical homomorphisms
@xmath1104 , this becomes a representation of @xmath13 .
this representation clearly belongs to @xmath1079 .
we claim that it is the universal representation for the class @xmath1079 .
so let @xmath16 be any representation in @xmath1079 .
then @xmath16 comes from a nondegenerate , locally bounded representation @xmath972 of @xmath1080 .
we must show that it comes from a nondegenerate representation @xmath250 of @xmath1081 .
let @xmath250 be the restriction of @xmath972 to @xmath1081 .
then @xmath1105 .
we are going to prove that this is a core .
the bilinear map @xmath1106 is separately continuous with respect to the pro - c*-algebra topology on @xmath1080 and the inductive limit topology on @xmath1107 .
we have assumed that it has dense range .
since @xmath1092 is dense in @xmath1080 , the image of @xmath1108 is a core .
thus @xmath1109 is dense in @xmath21 in the graph topology .
the representation @xmath250 is nondegenerate , and the associated representation of @xmath13 is @xmath16 .
this finishes the proof that the canonical representation of @xmath13 on @xmath1081 is universal .
thus @xmath1081 is a weak c*-hull for @xmath1079 .
it is even a c*-hull because the class of locally bounded representations is admissible by corollary [ cor : locally_bounded_admissible ] .
let @xmath13 be a commutative * -algebra .
we are going to describe all _ commutative _ weak c*-hulls for @xmath13 . actually , we describe all locally bounded weak c*-hulls , and these turn out to be the same as the commutative ones .
we characterise when a c*-hull satisfies the ( strong ) local global principle and when the class of all locally bounded representations has a c*-hull .
we compare the class of locally bounded representations with the class of representations defined by requiring all @xmath123 to act by a regular , self - adjoint operator .
[ pro : commutative_representation ] let @xmath13 be a commutative * -algebra and let @xmath1110 be a commutative c*-algebra .
any representation of @xmath13 on @xmath8 has @xmath1111 as a core and is locally bounded .
there is a natural bijection between representations of @xmath13 on @xmath8 , * -homomorphisms @xmath1112 , and continuous maps @xmath1113 .
let @xmath329 be a representation .
since @xmath1114 is dense in @xmath8 , for any @xmath1115 there is @xmath1116 with @xmath1117 .
then there is an open neighbourhood of @xmath257 on which @xmath656 is non - zero .
a compact subset @xmath762 of @xmath197 may be covered by finitely many such open neighbourhoods .
this gives finitely many functions @xmath1118 so that @xmath1119 for all @xmath1120 .
this sum again belongs to the right ideal @xmath1114 , and hence @xmath1114 contains @xmath1111 .
there is an approximate unit @xmath1121 for @xmath1122 that belongs to @xmath1111 . if @xmath339 , then @xmath1123 for all @xmath20 .
that is , @xmath1124 in the graph topology .
since @xmath1125 , @xmath1111 is a core for @xmath329 . given @xmath20 , we define a function @xmath1126 by @xmath1127 for any @xmath1128 with @xmath1129 .
this does not depend on the choice of @xmath446 , and @xmath1130 is continuous in the open subset where @xmath1131 .
thus @xmath1132 .
the map @xmath1112 , @xmath1133 , is a * -homomorphism .
conversely , any * -homomorphism @xmath1112 gives a representation of @xmath13 on @xmath1122 with core @xmath1111 by @xmath1134 for all @xmath20 , @xmath1128 .
the maps that go back and forth between representations on @xmath1122 and * -homomorphisms @xmath1112 are inverse to each other .
a * -homomorphism @xmath1135 gives a continuous map @xmath1136 by mapping @xmath1115 to the character @xmath1137 .
conversely , a continuous map @xmath1138 induces a * -homomorphism @xmath1139 , @xmath1140 , and these two constructions are inverse to each other .
let @xmath1141 be a continuous map
. then @xmath656 maps compact subsets in @xmath197 to compact subsets of @xmath32 .
if @xmath1142 is compact , then any element in @xmath1143 is @xmath1144-bounded for the c*-seminorm on @xmath13 associated to the compact subset @xmath1145 .
thus all elements in @xmath1111 are bounded . since @xmath1111 is a core for the representation of @xmath13 associated to @xmath656
, this representation is locally bounded . [
the : commutative_hull ] let @xmath13 be a commutative * -algebra , let @xmath1110 be a commutative c*-algebra , let @xmath1146 be a continuous map , and let @xmath329 be the corresponding representation of @xmath13 on @xmath8 . call a representation of @xmath13 on a hilbert module @xmath15 _ @xmath197-integrable _ if it is isomorphic to @xmath1147 for a representation @xmath250 of @xmath8 on @xmath15 .
the following are equivalent : 1 .
@xmath1146 is injective ; 2 .
@xmath8 is a weak c*-hull for the @xmath197-integrable representations ; 3 .
@xmath8 is a c*-hull for the @xmath197-integrable representations .
furthermore , any locally bounded weak c*-hull of @xmath13 is commutative
. if @xmath656 is not injective , then there are @xmath1148 in @xmath197 with @xmath1149 .
the evaluation maps at @xmath257 and @xmath1150 are different @xmath314-dimensional representations of @xmath8 that induce the same representation of @xmath13 . hence the condition [ enum : cstar - gen1 ] in proposition
[ pro : cstar - generated_by_unbounded_multipliers ] is violated and so @xmath8 is not a weak c*-hull .
conversely , assume that @xmath656 is injective .
the representation of @xmath13 on @xmath8 associated to @xmath656 is locally bounded by proposition [ pro : commutative_representation ] and hence induces a representation of the unital c*algebra @xmath1151 of bounded elements in @xmath1152 , see proposition [ pro : commutative_pro - completion ] .
explicitly , this representation composes functions with @xmath656 .
since @xmath656 is injective , @xmath1153 separates the points of @xmath197 .
we show that it is strictly dense in @xmath1154 .
if @xmath1142 is compact , then the image of @xmath1155 in the quotient @xmath933 of @xmath1122 separates the points of @xmath762 . since this image is again a c*-algebra , it is equal to @xmath933 by the stone weierstra theorem .
let @xmath1156 .
for any compact subset @xmath1142 , there is @xmath1157 with @xmath1158 . by functional calculus
, we may arrange that @xmath1159 .
the net @xmath1160 indexed by the directed set of compact subsets of @xmath197 is uniformly bounded and converges towards @xmath656 in the topology of uniform convergence on compact subsets .
hence it converges towards @xmath656 in the strict topology ( compare @xcite*lemma a.1 ) .
this finishes the proof that @xmath1161 is strictly dense in @xmath1162 .
proposition [ pro : criterion_cstar - hull_locally_bounded ] shows that @xmath8 is a weak c*-hull for the @xmath8-integrable representations of @xmath13 .
any @xmath197-integrable representation of @xmath13 is locally bounded .
the class @xmath995 of locally bounded representations of @xmath13 is admissible by corollary [ cor : locally_bounded_admissible ] .
hence the smaller class of @xmath197-integrable representations inherits the equivalent conditions [ enum : cstar - gen2][enum : cstar - gen2 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] .
thus @xmath1122 is even a c*-hull .
let @xmath8 with the universal representation @xmath329 be a locally bounded weak c*-hull .
then the image of @xmath1163 in the multiplier algebra of @xmath8 is strictly dense by proposition [ pro : criterion_cstar - hull_locally_bounded ] .
thus @xmath1008 is commutative , and then so is @xmath8 . thus a weak c*-hull is locally bounded if and only if it is commutative .
[ the : commutative_local - global ] let @xmath13 be a commutative * -algebra , let @xmath1110 be a commutative c*-algebra , and let @xmath1146 be an injective continuous map .
let @xmath1164 be the class of @xmath197-integrable representations .
the following statements are equivalent if @xmath32 is metrisable : 1 .
[ enum : commutative_local - global1 ] @xmath1165 is a homeomorphism onto its image ; 2 .
[ enum : commutative_local - global2 ] @xmath1164 is defined by submodule conditions ; 3 .
[ enum : commutative_local - global3 ] @xmath1164 satisfies the strong local global principle ; 4 .
[ enum : commutative_local - global4 ] @xmath1164 satisfies the local global principle ; 5 .
[ enum : commutative_local - global5 ] if @xmath1166 for a sequence @xmath1167 in @xmath197 and @xmath1168 , then already @xmath1169 . the implications [ enum : commutative_local - global1]@xmath272[enum : commutative_local - global2]@xmath272[enum : commutative_local - global3]@xmath272[enum : commutative_local - global4]@xmath272[enum : commutative_local - global5 ] hold without assumptions on @xmath32 .
i do not know whether [ enum : commutative_local - global1][enum : commutative_local - global4 ] are equivalent in general .
the condition [ enum : commutative_local - global5 ] is there to allow to go back from [ enum : commutative_local - global4 ] to [ enum : commutative_local - global1 ] at least for metrisable @xmath32 .
first we check [ enum : commutative_local - global5]@xmath272[enum : commutative_local - global1 ] if @xmath32 is metrisable .
if @xmath656 is not a homeomorphism onto its image , then there is a subset @xmath1170 that is open , such that @xmath1171 is not open in the subspace topology on @xmath1172 .
since @xmath32 is metrisable , there is @xmath1173 and a sequence in @xmath1174 that converges towards @xmath1175 .
this lifts to a sequence @xmath1167 in @xmath1176 such that @xmath1177 .
we can not have @xmath1169 because @xmath1178 never enters the open neighbourhood @xmath1179 of @xmath257 .
the implication [ enum : commutative_local - global2]@xmath272[enum : commutative_local - global3 ] is theorem [ the : operator_conditions ] , and [ enum : commutative_local - global3]@xmath272[enum : commutative_local - global4 ] is trivial .
we are going to verify [ enum : commutative_local - global1]@xmath272[enum : commutative_local - global2 ] and [ enum : commutative_local - global4]@xmath272[enum : commutative_local - global5 ] .
this will finish the proof of the theorem .
assume [ enum : commutative_local - global1 ] .
let @xmath16 be a representation in @xmath995 .
the operators @xmath25 for @xmath123 are regular and self - adjoint by proposition [ pro : locally_bounded_regular ] .
furthermore , their cayley transforms belong to the image of @xmath1180 , which is commutative .
hence the operators @xmath25 for @xmath123 strongly commute with each other .
the class @xmath240 of representations of @xmath13 with the property that all @xmath25 are regular and self - adjoint and strongly commute with each other is defined by submodule conditions by examples [ exa : regularity_condition ] and example [ exa : strong_commutation ] .
let @xmath1181 .
given a representation in @xmath240 , there is a unique representation @xmath1182 that maps the @xmath98th coordinate projection to the cayley transform of @xmath25 .
we map @xmath32 to @xmath1183 by sending @xmath924 to the point @xmath1184 . here
@xmath1185 is the cayley transform of the number @xmath1186 or , equivalently , the value of the cayley transform of the unbounded function @xmath952 at @xmath949 .
this is a homeomorphism onto its image because for a net of characters @xmath1187 and a character @xmath949 on @xmath13 , we have @xmath1188 if and only if @xmath1189 . thus the composite map @xmath1190 is a homeomorphism onto its image as well .
this forces the image to be locally closed because @xmath1183 is compact and @xmath197 locally compact , and a subspace of a locally compact space is locally compact if and only if its underlying subset is locally closed ( see @xcite*i.9.7 , propositions 12 and 13 ) .
let @xmath1191 be the closure of the image of @xmath197 in @xmath1183 .
then @xmath197 is open in @xmath694 .
all representations in @xmath240 carry a unital * -homomorphism @xmath1192 . asking for this to factor through the quotient @xmath1193 of @xmath1194 is a submodule condition as in example [ exa : spectral_condition ] . asking for
the induced * -homomorphism @xmath1195 to remain nondegenerate on @xmath1122 is another submodule condition as in example [ exa : spectral_condition ] .
the class @xmath605 defined by these two more submodule conditions is weakly admissible by lemma [ lem : operator_conditions ] .
the universal @xmath197-integrable representation belongs to @xmath605 ; by weak admissibility , this is inherited by all @xmath197-integrable representations .
conversely , we claim that any representation in @xmath605 is @xmath197-integrable . if @xmath1196 , then the unital * -homomorphism @xmath1192 descends to a nondegenerate * -homomorphism @xmath1197 . by construction ,
the extension of @xmath250 to multipliers maps the cayley transform of @xmath1198 for @xmath123 to the cayley transform of @xmath25 .
let @xmath263 be the @xmath197-integrable representation of @xmath13 associated to @xmath250 .
the regular , self - adjoint operators @xmath1199 and @xmath25 have the same cayley transform for all @xmath991 .
hence @xmath1200 for all @xmath123 .
the subset @xmath119 is a strong generating set for @xmath13 by example [ exa : strong_generating ] .
hence proposition [ pro : equality_if_closure_equal ] gives @xmath1201 .
thus @xmath1164 is the class of representations defined by the submodule conditions above .
this finishes the proof that [ enum : commutative_local - global1]@xmath272[enum : commutative_local - global2 ] .
now we prove [ enum : commutative_local - global4]@xmath272[enum : commutative_local - global5 ] by contradiction .
let @xmath1167 and @xmath257 be as in [ enum : commutative_local - global5 ] .
let @xmath1202 be the one - point compactification of @xmath881 and view the sequence @xmath1203 and @xmath257 as a map @xmath1204 .
this map is not continuous , but composition with @xmath656 gives a continuous map @xmath1205 .
hence proposition [ pro : commutative_representation ] gives a representation @xmath1206 of @xmath13 on @xmath1207 .
this is not @xmath197-integrable because the map @xmath1208 is not continuous .
we claim , however , that the representation @xmath1209 is @xmath197-integrable for any gns - representation @xmath250 on a hilbert space @xmath261 .
a state on @xmath1207 is the same as a radon measure on @xmath1210 .
since @xmath1210 is countable , any radon measure is atomic .
thus the resulting gns - representation is a direct sum of irreducible representations associated to characters .
each character on @xmath1207 gives an @xmath197-integrable representation because @xmath1211 .
hence @xmath1206 is a counterexample to the local
global principle , so [ enum : commutative_local - global4 ] can not hold if [ enum : commutative_local - global5 ] fails .
[ exa : pure - point ] let @xmath434 $ ] , so that @xmath1212 , let @xmath197 be @xmath111 with the discrete topology , and let @xmath1213 be the identity map .
this is a continuous bijection , but not open .
hence the class of @xmath197-integrable representations violates the local global principle by theorem [ the : commutative_local - global ] .
nevertheless , @xmath1122 is a c*-hull for the class of @xmath197-integrable representations of @xmath13 by theorem [ the : commutative_hull ] .
the @xmath197-integrable representations of @xmath13 are integrable as in theorem [ the : regular_self - adjoint_cstar - hull ] , and so they come from a single regular , self - adjoint operator @xmath875 .
the representation of @xmath37 $ ] associated to @xmath41 is @xmath197-integrable if and only if @xmath1214 , where @xmath1215 for @xmath1216 is the @xmath1217-eigenspace of @xmath41 .
another example of a c*-hull for @xmath37 $ ] where @xmath1218 is bijective but not a homeomorphism onto its image is discussed in example [ exa : smaller_integrable_than_regular_self - adjoint ] .
[ the : no_cstar - hull_commutative_not_locally_compact ] there is a c*-hull for @xmath995 if and only if the compactly generated topology @xmath926 on @xmath32 is locally compact , and then the c*-hull is @xmath1219 .
assume first that @xmath1220 is locally compact .
the pro - c*-algebra @xmath917 that acts on locally bounded representations of @xmath13 is @xmath1221 by proposition [ pro : commutative_pro - completion ] .
the primitive ideal space of @xmath933 for a compact subspace @xmath1222 is simply @xmath762 , and @xmath1223 for @xmath1224 and @xmath1225 .
therefore , a function @xmath1226 vanishes at @xmath720 in the sense of definition [ def : compact_in_pro - cstar ] if and only if it vanishes at @xmath720 in the usual sense .
thus theorem [ the : prim_a_locally_compact_gives_hull ] applies in our case and shows that @xmath1227 is a c*-hull for the class of locally bounded representations of @xmath13 .
conversely , let @xmath8 be a ( weak ) c*-hull for the locally bounded representations of @xmath13
. then @xmath8 is commutative by theorem [ the : commutative_hull ] .
let @xmath1183 be the spectrum of @xmath8 .
the representation of @xmath13 on @xmath1228 corresponds to a continuous map @xmath1229 by proposition [ pro : commutative_representation ] .
let @xmath1230 be a commutative c*-algebra .
any representation of @xmath13 on @xmath14 is locally bounded .
so the bijection @xmath1231 is a bijection between the spaces of continuous maps @xmath1136 and @xmath1232 .
more precisely , this bijection is composition with @xmath656 . for the one - point space @xmath197
, this bijection says that @xmath1233 is bijective .
the bijection for all compact @xmath197 means that @xmath656 becomes a homeomorphism if we replace the topologies on @xmath1183 and @xmath32 by the associated compactly generated ones .
the topology on @xmath1183 is already compactly generated because @xmath1183 is locally compact .
hence @xmath656 is a homeomorphism from @xmath1183 to @xmath1220 , and @xmath926 is locally compact .
let @xmath240 be the class of all representations with the property that @xmath25 is regular and self - adjoint for all @xmath123 .
we are going to compare @xmath240 and @xmath995 .
proposition [ pro : locally_bounded_regular ] gives @xmath1234 .
[ the : commutative_regular_admissible ] the class @xmath240 is admissible and defined by submodule conditions .
hence it satisfies the strong local global principle .
the operators @xmath25 for @xmath123 strongly commute for all @xmath1235 .
let @xmath731 be a strong generating set for @xmath13 . if @xmath25 is regular and self - adjoint for all @xmath109 , then already @xmath1235 . the class @xmath732 of representations defined by requiring @xmath25 to be regular and self - adjoint for all @xmath109 for a strong generating set @xmath135 is admissible and defined by submodule conditions by theorem [ the : regular_admissible_local - global ] .
the class @xmath240 is defined by submodule conditions as well by example [ exa : regularity_condition ] .
so is the subclass @xmath605 of all representations in @xmath240 for which the operators @xmath25 for all @xmath123 strongly commute by example [ exa : strong_commutation ] .
hence our three classes of representations satisfy the strong local global principle by theorem [ the : operator_conditions ] . the classes @xmath240 and
@xmath605 have the same hilbert space representations by @xcite*theorem 9.1.2 .
since @xmath135 is a strong generating set , the domain of any representation @xmath16 in @xmath732 is @xmath1236 by .
this contains @xmath1237 .
now @xcite*theorem 9.1.3 shows that @xmath732 and @xmath240 have the same hilbert space representations . since our three classes of representations satisfy the ( strong ) local global principle and have the same hilbert space representations , they are equal . [
the : locally_bounded_versus_regular ] if @xmath13 is countably generated , then @xmath1238 .
proposition [ pro : locally_bounded_regular ] gives @xmath1239 .
conversely , let @xmath62 be a representation on a hilbert module @xmath15 in @xmath240 ; that is , @xmath25 is regular and self - adjoint for each @xmath20 .
let @xmath1240 be a countable generating set for @xmath13 .
we may assume without loss of generality that @xmath1241 for all @xmath1242 and that @xmath1243 is a basis for @xmath13 and hence a strong generating set . let @xmath81 .
we are going to approximate @xmath114 by bounded vectors for @xmath16 . for each @xmath1242
, there is a canonical homomorphism @xmath1244\to a$ ] mapping @xmath1245 .
the closure of @xmath1246 is an integrable representation of @xmath37 $ ] as in condition [ enum : regular_self - adjoint_cstar - hull2 ] in theorem [ the : regular_self - adjoint_cstar - hull ] .
hence it corresponds to a representation @xmath1247 , the functional calculus of @xmath1248 .
the operators @xmath25 for @xmath991 strongly commute by theorem [ the : commutative_regular_admissible ] .
thus the cayley transform of @xmath1249 commutes with @xmath25 and , in particular , maps the domain of @xmath25 to itself .
the same remains true for @xmath1250 for all @xmath1251 because we get them by the ( bounded ) functional calculus for the cayley transform of @xmath1248 .
so @xmath1252 by and @xmath1253 for all @xmath1251 , @xmath20 as operators on @xmath21 .
now we show that @xmath1246 is locally bounded . if @xmath1254 is supported in a compact subset @xmath1255 , then @xmath1256 for all @xmath1257 $ ] ; thus @xmath1258 is bounded for the representation @xmath1246 .
there is an approximate unit @xmath1259 for @xmath36 that lies in @xmath1260 .
then @xmath1261 for all @xmath81 , even in the graph topology for @xmath16 because @xmath1262 for all @xmath20 , @xmath1263 , @xmath81 .
therefore , the bounded vectors of the form @xmath1264 with @xmath1254 , @xmath81 form a core for @xmath1246 .
so @xmath1246 is locally bounded .
we now refine this construction to approximate @xmath114 by bounded vectors for the whole representation @xmath16 .
we construct @xmath295 as above .
fix @xmath1265 and let @xmath1266 . the argument
above gives @xmath1267 with @xmath1268 and @xmath1269 .
thus @xmath1270 in the graph norm for @xmath1271 for @xmath1272 . for @xmath1273
, let @xmath1274 the operators @xmath1275 are norm - contracting , map @xmath21 into itself , and commute with each other and with the unbounded operators @xmath67 for all @xmath20 .
hence @xmath1276 for all @xmath1277 , @xmath1278 . since we assumed @xmath1279 to be a strong generating set , the graph norms for @xmath1271 generate the graph topology .
so the estimate above shows that @xmath1280 with fixed @xmath1281 is a cauchy sequence in @xmath21 in the graph topology .
thus it converges to some @xmath1282 . letting @xmath1283 ,
the above estimate remains true for @xmath1284 and gives @xmath1285 for all @xmath1286 , uniformly in @xmath1287 .
this implies @xmath1288 , so that @xmath1289 in the graph topology .
it remains to show that each @xmath1290 is a bounded vector .
fix @xmath1291 and let @xmath78 .
choose @xmath1292 so that @xmath1293 is supported in @xmath1294 $ ] .
if @xmath1295 , then @xmath1296 because @xmath1297 occurs in the definition of @xmath1298 .
as above , this implies @xmath1299 for all @xmath78 .
thus @xmath1300 is finite for @xmath1301 .
since @xmath1249 is a basis for @xmath13 and @xmath906 is subadditive , we get @xmath1302 for all @xmath154 .
since @xmath1303 is the operator norm of @xmath1304 , it is a c*-seminorm on @xmath13 . by construction , @xmath1305 for all @xmath154 ,
that is , @xmath1290 is bounded .
[ pro : commutative_regular_versus_locally_bounded_necessary ] if @xmath240 has a weak c*-hull , then @xmath1306 .
let @xmath8 with the universal representation @xmath329 be a weak c*-hull for @xmath240 .
first we claim that @xmath8 is commutative .
let @xmath1307 be a faithful representation .
this corresponds to an integrable representation @xmath16 of @xmath13 .
since the equivalence @xmath1308 is compatible with unitary * -intertwiners , the commutant of @xmath1309 is the c*-algebra of * -intertwiners of @xmath16 by proposition [ pro : adjointable_intertwiner ] .
the commutant of this is a commutative von neumann algebra by @xcite*theorem 9.1.7 .
so the bicommutant of @xmath1309 is commutative .
this forces @xmath8 to be commutative .
any representation of @xmath13 on a commutative c*-algebra is locally bounded by theorem [ the : commutative_hull ] .
if the universal representation for @xmath240 is locally bounded , then all representations in @xmath240 are locally bounded , so that @xmath1310 .
thus @xmath240 only has a weak c*-hull if @xmath1310 .
[ exa : repi_infinite_generators ] let @xmath13 be the * -algebra @xmath1311 $ ] of polynomials in countably many symmetric generators .
then @xmath1312 with the product topology .
this is metrisable .
so @xmath926 is the usual product topology .
since this is not locally compact , @xmath995 has no c*-hull , not even a weak one .
since @xmath13 is countably generated , @xmath1306 by theorem [ the : locally_bounded_versus_regular ] .
a commutative ( weak ) c*-hull for some class of representations of @xmath13 is equivalent to an injective , continuous map @xmath1136 for a locally compact space @xmath197 by theorem [ the : commutative_hull ] .
let @xmath0 be a topological group .
host algebra _ for a @xmath0 is defined in @xcite as a c*-algebra @xmath8 with a continuous representation @xmath1217 of @xmath0 by unitary multipliers , such that for each hilbert space @xmath261 , the map that sends a representation @xmath1313 to a unitary representation @xmath1314 of @xmath0 is injective .
we claim that commutative c*-hulls for the polynomial algebra @xmath1311 $ ] are equivalent to host algebras of the topological group @xmath1315 .
let @xmath1316 be the c*-algebra of @xmath0 viewed as a discrete group .
representations of @xmath1316 are equivalent to representations of the discrete group underlying @xmath0 by unitary multipliers .
since any representation of @xmath1316 is bounded , any weakly admissible class of representations of @xmath1316 is admissible by corollary [ cor : locally_bounded_admissible ] .
call a representation of @xmath1316 continuous if the corresponding representation of @xmath0 is continuous .
this class is easily seen to be weakly admissible , hence admissible .
the unital * -homomorphism @xmath1317 associated to the unitary representation @xmath1217 for a host algebra @xmath8 is continuous by assumption .
thus @xmath8-integrable representations of @xmath1316 are continuous .
the injectivity requirement in the definition of a host algebra is exactly the condition [ enum : cstar - gen1 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] , and this is equivalent to @xmath8 being a c*-hull .
thus a host algebra for @xmath0 is the same as a c*-hull or weak c*-hull for a class of continuous representations of @xmath1316 .
it is unclear whether the class of continuous representations of @xmath1316 satisfies the local
global principle . in applications , we would rather study continuous representations of @xmath0 through the lie algebra of @xmath0 instead of through the inseparable c*-algebra @xmath1316 . for @xmath1318 ,
the lie algebra is the abelian lie algebra @xmath1319 , and its universal enveloping algebra is the polynomial algebra @xmath1320 $ ] . call a representation of @xmath13 integrable if it belongs to @xmath1321 .
let @xmath15 be a hilbert module .
we claim that an integrable representation of @xmath13 on @xmath15 is equivalent to a strictly continuous , unitary representation of the group @xmath1319 on @xmath15 . indeed , a unitary representation of @xmath111 is equivalent to a representation of @xmath1322 , and these are equivalent to integrable representations of @xmath37 $ ] as in theorem [ the : regular_self - adjoint_cstar - hull ] . in an integrable representation of
@xmath1311 $ ] , the operators @xmath1323 for @xmath1242 strongly commute .
hence the resulting representations of @xmath36 commute .
equivalently , the resulting continuous representations of @xmath111 commute , so that we may combine them to a representation of the abelian group @xmath1319 .
conversely , a continuous unitary representation of @xmath1319 provides nondegenerate representations of @xmath1324 for all @xmath1325 by restricting the representation to @xmath1326 .
these correspond to a compatible family of representations of the polynomial algebras @xmath1327 $ ] for @xmath1325 .
the intersection of their domains is dense by @xcite*lemma 1.1.2 .
so these representations combine to a representation of @xmath1320 $ ] .
hence an integrable representation of @xmath13 on a hilbert module as in theorem [ the : commutative_regular_admissible ] is equivalent to a continuous representation of @xmath1319 .
let @xmath0 be a discrete group with unit element @xmath1328 .
[ def : grading ] a _ @xmath0-graded * -algebra _ is a unital algebra @xmath13 with a linear direct sum decomposition @xmath2 with @xmath3 , @xmath4 , and @xmath5 for all @xmath1329 .
thus @xmath24 is a unital * -subalgebra .
the articles @xcite study many examples of @xmath0-graded * -algebras .
we fix some notation used throughout this section .
let @xmath15 be a hilbert module over a c*-algebra @xmath14 .
let @xmath62 be a representation of @xmath13 on @xmath15 .
let @xmath1330 for @xmath54 be the restrictions of @xmath16 , so @xmath1331 . since @xmath16 is a * -homomorphism , @xmath1332 for all @xmath1333 , @xmath1334 .
the last condition means that @xmath1335 for all @xmath19 .
in particular , @xmath1336 is a representation of @xmath6 .
[ lem : restriction_closed ] the families of norms @xmath1337 for @xmath20 and for @xmath26 generate equivalent topologies on @xmath21 .
hence the representation @xmath1338 is closed if and only if @xmath16 is closed .
any element of @xmath13 is a sum @xmath1339 with @xmath1333 and only finitely many non - zero terms .
we estimate @xmath1340 , and @xmath1341 by the proof of lemma [ lem : graph_norms_directed ] .
since @xmath1342 , the graph topologies for @xmath1343 and @xmath16 are equivalent .
[ def : induced_integrable ] let a weakly admissible class of integrable representations of @xmath6 on hilbert modules be given .
we call a representation of @xmath13 on a hilbert module _ integrable _ if its restriction to @xmath6 is integrable
. here `` restriction of @xmath16 '' means the representation @xmath1343 with the same domain @xmath21 as @xmath16 .
this is closed by lemma [ lem : restriction_closed ] .
[ pro : integrable_induced_nice ] if integrability for representations of @xmath6 is defined by submodule conditions , then the same holds for @xmath13 .
if the local global principle holds for the integrable representations of @xmath6 , it also holds for the integrable representations of @xmath13 .
if the class of integrable representations of @xmath6 is admissible or weakly admissible , the same holds for @xmath13 . the first two statements and the claim about weak admissibility are trivial because integrability for a representation of @xmath13 only involves its restriction to @xmath6 . lemma [ lem : restriction_closed ] shows that restriction from @xmath13 to @xmath6 does not change the domain .
hence [ enum : cstar - gen2 ] in proposition [ pro : cstar - generated_by_unbounded_multipliers ] is inherited by @xmath13 if it holds for @xmath6 .
that is , admissibility of the integrable representations passes from @xmath6 to @xmath13 .
it is unclear whether @xmath13 also inherits the _ strong _ local global principle from @xmath6 .
this may often be bypassed using theorem [ the : operator_conditions ] .
let @xmath1344 and @xmath1345 for a hilbert @xmath14-module @xmath30 be a representation of @xmath6 .
we try to induce @xmath1346 to a representation of @xmath13 as in @xcite .
thus we consider the algebraic tensor product @xmath1347 and equip it with the obvious right @xmath14-module structure and the unique sesquilinear map that satisfies @xmath1348 for all @xmath1329 , @xmath1349 , @xmath1350 , @xmath1351 .
this map is sesquilinear and descends to the quotient space @xmath1352 .
it is symmetric and @xmath14-linear in the sense that @xmath1353 and @xmath1354 .
let @xmath16 be the action of @xmath13 on @xmath1352 by left multiplication .
this representation is formally a * -homomorphism in the sense that @xmath1355 for all @xmath20 , @xmath1356 .
the only thing that is missing to get a representation of @xmath13 on a hilbert @xmath14-module is positivity of the inner product .
this requires a subtle extra condition .
[ pro : inducible_criteria ] the following are equivalent : 1 .
[ en : inducible_1 ] the sesquilinear map on @xmath1357 defined above is positive semidefinite ; 2 . [ en : inducible_2 ] for all @xmath54 , @xmath439 and all @xmath1358 , @xmath1359 , the element @xmath1360 is positive ; 3 . [ en : inducible_3 ] for all @xmath54 , @xmath439 and all @xmath1358 , @xmath1359 , the matrix @xmath1361 is positive .
the condition [ en : inducible_2 ] for fixed @xmath54 says that the sesquilinear map on @xmath1362 is positive semidefinite .
since the subspaces @xmath1362 for different @xmath51 are orthogonal , this is equivalent to positive semidefiniteness on @xmath1352 . thus [
en : inducible_1]@xmath596[en : inducible_2 ] .
we prove [ en : inducible_2]@xmath596[en : inducible_3 ] .
fix @xmath54 , @xmath439 , @xmath1363 and @xmath1359 .
let @xmath1364 be the matrix in [ en : inducible_3 ] .
by @xcite*lemma 4.1 , @xmath1365 in @xmath1366 if and only if @xmath1367 for all @xmath1368 .
that is , @xmath1369 for all @xmath1370 .
since @xmath1371 is a right @xmath14-module , this condition for all @xmath1372 , @xmath1373 is equivalent to [ en : inducible_2 ] .
[ def : inducible_representation ] a representation @xmath1346 of @xmath6 is _ inducible _ ( to @xmath13 ) if it satisfies the equivalent conditions in proposition [ pro : inducible_criteria ] .
if @xmath6 were a c*-algebra , it would be enough to assume @xmath1374 for all @xmath54 , @xmath1375 , @xmath1376 , which amounts to the condition @xmath1377 in @xmath6 for all @xmath54 , @xmath48 , which is part of the definition of a fell bundle over a group . for
more general * -algebras , the positivity conditions for different @xmath439 in proposition [ pro : inducible_criteria ] may differ , compare @xcite .
let @xmath1378 be the hilbert module completion of @xmath1352 for the inner product above .
the * -algebra @xmath13 acts on @xmath1352 by left multiplication , @xmath1379 for @xmath1380 , @xmath1381 . as in the proof of lemma [ lem : tensor_rep_with_corr ] , this module structure descends to the image of @xmath1352 in @xmath1382 and gives a well defined representation @xmath16 of @xmath13 on @xmath1378 .
its closure is called the _ induced representation _ from @xmath1346 , and its domain is denoted by @xmath1383 .
the decomposition @xmath1384 is orthogonal for the above inner product and @xmath6-invariant .
hence @xmath1385 where @xmath1386 is the closure of the image of @xmath1362 or , equivalently , the hilbert @xmath14-module completion of @xmath1362 with respect to the restriction of the inner product .
each summand @xmath1386 carries a closed representation of @xmath6 with domain @xmath1387 , and @xmath1388 is the direct sum of these representations .
[ lem : induced_repr_inducible ] let @xmath16 be any representation of @xmath13 .
then @xmath1388 is inducible . for @xmath54 , @xmath1363 , @xmath1389 , let @xmath1390 .
then @xmath1391 we assume that the chosen class of integrable representations of @xmath6 has a ( weak ) c*-hull @xmath11 .
we want to construct a fell bundle whose section c*-algebra is a ( weak ) c*-hull for the integrable representations of @xmath13 . at some point , we need @xmath11 to be a full c*-hull ( compatibility with isometric intertwiners ) and one more extra condition .
but we may begin the construction without these assumptions .
first we build the unit fibre @xmath44 of the fell bundle .
it is a ( weak ) c*-hull for the inducible , integrable representations of @xmath6 .
let @xmath1392 be the universal integrable representation of @xmath6 on @xmath11 . given a representation @xmath1346 of @xmath6 , we let @xmath1393 denote the corresponding representation of @xmath11 .
let @xmath1394 for a self - adjoint element @xmath257 in a c*-algebra denote its negative part .
[ def : restrict_to_inducible ] let @xmath44 be the quotient c*-algebra of @xmath11 by the closed two - sided ideal generated by elements of the form @xmath1395 [ pro : inducible_a_b ] an integrable representation @xmath1346 of @xmath6 is inducible if and only if @xmath1393 factors through the quotient map @xmath1396 .
thus @xmath44 is a c*-hull for the inducible , integrable representations of @xmath6 .
fix @xmath54 , @xmath1363 and @xmath1397 and let @xmath1398 be the matrix @xmath1399 .
we claim that @xmath197 is mapped to a positive matrix in @xmath1400 and that @xmath44 is the largest quotient of @xmath11 with this property for all matrices of this form . by @xcite*lemma 4.1 ,
the image of @xmath197 in @xmath1401 is positive if and only if @xmath1402 for all @xmath1403 .
writing @xmath1404 , this condition becomes @xmath1405 .
this is true for all @xmath1406 because @xmath1407 is a right ideal in @xmath11 and a self - adjoint element of a c*-algebra is positive if and only if its negative part vanishes .
conversely , choosing @xmath1408 from an approximate unit , we see that the positivity of @xmath1409 in @xmath44 is necessary for @xmath197 to be positive in @xmath1400 .
let @xmath1410 be an integrable representation of @xmath6 on @xmath30 .
the corresponding representation @xmath1411 descends to @xmath44 if and only if @xmath1412 maps matrices of the form @xmath1413 as above to positive operators on @xmath1414 .
equivalently , @xmath1415 in @xmath14 for all @xmath1416 .
this holds for all @xmath1417 , @xmath1418 , @xmath1419 if and only if the canonical @xmath14-valued sesquilinear form on the algebraic tensor product @xmath1420 is positive definite .
put together , @xmath1393 descends to a representation of @xmath44 if and only if the canonical sesquilinear form on @xmath1420 is positive definite .
we claim that this implies positivity on the larger space @xmath1421 . since @xmath1422 ,
the image of @xmath1423 is a core for @xmath1410 , that is , it is dense in the graph topology .
if @xmath1424 is a finite - dimensional vector subspace , then the inner product on @xmath1425 is continuous in the @xmath6-graph topology on @xmath1371 .
since norm limits of positive elements are again positive , the positivity of the inner product on @xmath1426 implies positivity on @xmath1427 for all finite - dimensional subspaces @xmath1428 and hence on @xmath1429 . summing up , the representation @xmath1393 associated to an integrable representation @xmath1346 of @xmath6 annihilates the terms in if and only if @xmath1346 is inducible .
the quotient map induces a fully faithful embedding @xmath1430 .
the argument above shows that its image consists of those representations of @xmath11 that correspond to inducible , integrable representations of @xmath6 under the correspondence @xmath1431 .
hence @xmath44 is a ( weak ) c*-hull for the class of inducible , integrable representations of @xmath6 .
let @xmath1432 be the image of @xmath1407 in @xmath44 and let @xmath1433 be the induced representation of @xmath6 on this quotient .
this is the universal inducible , integrable representation of @xmath6 on @xmath44 .
[ def : fibre_g ] let @xmath1434 .
this is a well defined hilbert @xmath44-module because the representation @xmath1435 of @xmath6 on @xmath44 is inducible .
let @xmath1436 be the induced representation of @xmath6 on @xmath50 .
it has the image of @xmath1437 as a core , with the representation @xmath1438 for @xmath1439 , @xmath1333 , @xmath1440 .
by definition , the right @xmath44-module structure and the inner product on @xmath50 are the unique extensions of the following pre - hilbert module structure on @xmath1441 : @xmath1442 for all @xmath1333 , @xmath1443 , @xmath1444 , and @xmath1445 for @xmath1446 , @xmath1447 .
this is positive definite by proposition [ pro : inducible_a_b ] . by definition , @xmath50 is the norm completion of this pre - hilbert @xmath44-module , and @xmath1448 is the completion of @xmath1441 in the graph topology for the representation @xmath1449 of @xmath6 .
the hilbert @xmath44-modules @xmath50 are the fibres of our fell bundle . before we continue our construction
, we lighten our notation by _ dropping all @xmath1450-superscripts_. implicitly
, we replace the class of integrable representations of @xmath6 with the c*-hull @xmath11 by the class of _ inducible _ , integrable representations of @xmath6 with the c*-hull @xmath44 .
since the restriction of any representation of @xmath13 to @xmath6 is inducible by lemma [ lem : induced_repr_inducible ] , this does not change the notion of integrability for representations of @xmath13 .
the fell bundle structure on @xmath7 only exists under extra assumptions . before we turn to these , we construct representations of the hilbert @xmath11-modules @xmath1451 from an integrable representation of @xmath13 .
let @xmath48 and @xmath1452 .
the operator @xmath1453 is defined on all of @xmath15 because @xmath1454 maps @xmath15 into the domain @xmath21 of @xmath1455 .
its adjoint contains the densely defined operator @xmath1456 , and the operator @xmath1457 is bounded .
hence @xmath1456 extends to a bounded operator on @xmath15 , which is adjoint to @xmath1453 .
thus @xmath1458 .
define @xmath1459 as above , we check that @xmath1460 for all @xmath1461 , @xmath1462 , where the inner product is the one that defines @xmath1451 .
thus @xmath1463 extends uniquely to a bounded linear map @xmath1464 which still satisfies .
that is , it is a representation of the hilbert module @xmath1451 with respect to @xmath1465 .
if @xmath1466 is faithful ( hence isometric ) , then the induced map @xmath1467 is isometric : @xmath1468 for all @xmath1469 .
next we want to prove that @xmath1470 for all @xmath1329 and for all integrable representations @xmath16 of @xmath13 . this would give @xmath1471 a fell bundle structure , which would lift to @xmath7 itself if @xmath1465 is faithful .
lemma [ lem : be_bg_multiplicative ] gives provided the closed linear span of @xmath1472 is @xmath1473 for all @xmath54 .
but this only holds if we impose two extra assumptions .
first , compatibility of integrability and induction gives @xmath1451 a canonical left @xmath11-module structure .
secondly , compatibility of the weak c*-hull @xmath11 with isometric intertwiners ensures that the representation @xmath1463 is compatible with this left @xmath11-module structure .
we shall need the following assumption : [ def : induced_actions_integrable ] we say that _ integrability is compatible with induction _ if induction of inducible representations preserves integrability ; that is , if @xmath1346 is an inducible , integrable representation of @xmath6 on @xmath15 and @xmath16 is the representation of @xmath13 on @xmath1474 induced by @xmath1346 , then the representation @xmath1475 of @xmath6 is again integrable .
the representation @xmath1388 is automatically inducible by lemma [ lem : induced_repr_inducible ] .
so the compatibility condition in definition [ def : induced_actions_integrable ] stays the same if we replace the integrable representations of @xmath13 by the inducible , integrable representations , as we have implicitly done above when we wrote @xmath11 instead of @xmath44 .
a direct sum of representations is integrable if and only if each summand is integrable by corollary [ cor : integrable_sums ] .
hence integrability is compatible with induction if and only if an inducible , integrable representation @xmath1346 on @xmath30 induces integrable representations of @xmath6 on @xmath1476 for all @xmath54 .
[ pro : inducible_universal ] integrability is compatible with induction if and only if the representations @xmath1477 of @xmath6 on @xmath1451 are integrable for all @xmath54 .
the representations @xmath1477 of @xmath6 on @xmath1451 are integrable for all @xmath54 if and only if their direct sum is integrable . denote this by @xmath1478 .
if integrability is compatible with induction , then @xmath1479 must be integrable because it is the induced representation of the universal integrable ( inducible ) representation @xmath1392 of @xmath6 on @xmath11 .
we claim that induction maps the representation @xmath1480 of @xmath6 associated to a representation @xmath1481 to the representation @xmath1482 .
this claim follows by showing that the image of @xmath1483 is a core for both representations .
this is done as in the proof of proposition [ pro : inducible_a_b ] by first restricting to a finite - dimensional subspace of @xmath13 and using continuity of inner products in the relevant graph topologies as in the proof of proposition [ pro : inducible_a_b ] .
thus all induced representations are integrable once @xmath1478 is integrable .
the ( strong ) local global principle is useful to check that integrability is compatible with induction : [ pro : induced_actions_integrable_hils ] assume that the integrable representations of @xmath6 satisfy the strong local global principle and that induction maps irreducible , inducible , integrable hilbert space representations of @xmath6 to integrable hilbert space representations of @xmath13 .
then integrability is compatible with induction .
the same conclusion holds if the integrable representations of @xmath6 satisfy the local
global principle and induction maps all inducible , integrable hilbert space representations of @xmath6 to integrable hilbert space representations of @xmath13 .
let @xmath44 with the representation @xmath1435 of @xmath6 be the c*-hull for the inducible , integrable representations of @xmath6 . by proposition [ pro : inducible_universal ]
, it suffices to prove that the canonical representation of @xmath6 on @xmath1484 is integrable . by the strong local global principle , this follows if the induced representation @xmath1485 of @xmath6 on @xmath1486 is integrable for each irreducible representation @xmath250 of @xmath44 on a hilbert space @xmath261 .
the representation @xmath250 is equivalent to an irreducible , inducible , integrable representation @xmath16 of @xmath6 on @xmath261 , and @xmath1485 is the induced representation associated to @xmath16 . by assumption , @xmath1485 is integrable .
this finishes the proof in the case of the strong local global principle .
the argument in the other case is the same without the word `` irreducible . ''
[ pro : finit_dim_integrability_compatible_induction ] assume the following .
first , the integrable representations of @xmath6 satisfy the strong local global principle .
secondly , all _ irreducible _ , integrable hilbert space representations of @xmath6 are finite - dimensional .
third , all finite - dimensional inducible representations of @xmath6 are integrable . and fourth
, each @xmath1487 is finitely generated as a right @xmath6-module .
then integrability is compatible with induction .
first , since @xmath44 is a quotient of @xmath11 , its irreducible representations form a subset of the irreducible representations of @xmath11 .
thus the irreducible , inducible , integrable hilbert space representations of @xmath6 are finite - dimensional as well . by proposition
[ pro : induced_actions_integrable_hils ] , it suffices to check that the induced representation of @xmath6 on @xmath1488 is integrable when @xmath261 is a hilbert space with an irreducible , inducible , integrable representation . by our assumptions
, @xmath261 is finite - dimensional and @xmath1487 is finitely generated as an @xmath6-module .
hence @xmath1488 is finite - dimensional .
this representation is a direct summand in a representation of @xmath13 on @xmath1489 and hence inducible by lemma [ lem : induced_repr_inducible ] . by assumption ,
the induced representation of @xmath6 on @xmath1488 is integrable .
if integrability is compatible with induction , the representation @xmath1490 of @xmath6 on @xmath1451 is integrable .
it is inducible as well by lemma [ lem : induced_repr_inducible ] because it is a direct summand in a representation of @xmath13 .
hence there is a unique ( nondegenerate ) representation @xmath1491 of @xmath11 on @xmath1451 such that @xmath1492 is a core for @xmath1490 , and @xmath1493 for all @xmath1494 , @xmath1452 , @xmath1495 .
our next goal is to show that the representations @xmath1496 and @xmath1497 constructed using are compatible in the sense that @xmath1498 this is not automatic .
the following lemma is the most subtle point in the proof of the induction theorem .
[ lem : be_bg_multiplicative ] equation holds if @xmath11 is a c*-hull , not just a weak c*-hull .
then @xmath1499 for all @xmath54 .
let @xmath1500 .
the linear map @xmath1501 , @xmath1502 , for @xmath1503 , @xmath340 , preserves the inner products by .
hence it extends to a well defined isometry @xmath1504 .
the representation @xmath1491 of @xmath11 on @xmath1451 induces a representation @xmath1505 of @xmath11 on @xmath30 by interior tensor product .
this corresponds to the integrable representation @xmath1506 of @xmath6 on @xmath30 .
let @xmath1371 denote its domain .
a core for this representation is the image of @xmath1507 in @xmath30 , equipped with the left action @xmath1508 for @xmath26 , @xmath1509 , @xmath340 .
we shall need a slightly smaller core , namely , the image of @xmath1510 in @xmath1371 with the representation of @xmath6 by left multiplication in the first tensor factor @xmath1487 .
this is a core by an argument as in the proof of lemma [ lem : tensor_associative ] , using that the map @xmath1511 is separately continuous with respect to the graph topology on @xmath1512 and the norm topology on @xmath15 and that @xmath1513 is dense in @xmath1512 in the graph topology .
the isometry @xmath159 maps the image of @xmath1514 into the domain @xmath21 of @xmath1343 because @xmath1515 .
it maps @xmath1516 for @xmath1517 , @xmath1333 , @xmath1518 , @xmath340 to @xmath1519 .
thus @xmath159 is an isometric intertwiner from a core of @xmath1520 to @xmath1343 .
then @xmath159 intertwines @xmath1520 and @xmath1343 by lemma [ lem : closure_functorial ] . the representation @xmath1343 on @xmath15 was assumed to be integrable throughout , and the representation @xmath1520 on @xmath30 is also integrable .
since @xmath11 is a c*-hull and @xmath159 is an isometric intertwiner between two integrable representations of @xmath6 , it also intertwines the corresponding representations of @xmath11 .
this is exactly the meaning of .
this implies @xmath1499 because @xmath1491 is nondegenerate .
[ lem : fell_bundle_rep ] assume @xmath1499 for all @xmath54
. then holds .
we write @xmath1521 to denote that two sets of operators have the same closed linear span . by definition , @xmath1522 , and @xmath1523 because @xmath1407 is dense in @xmath11 .
our assumption @xmath1524 implies @xmath1525 .
we have seen above that @xmath1526 for @xmath1452 , @xmath1527 extends to a bounded linear operator on @xmath15 that is adjoint to @xmath1453 .
therefore , @xmath1528 for all @xmath1529 , @xmath48 ; both sides are globally defined bounded operators because @xmath1530 maps @xmath15 into @xmath21 .
the closed linear spans on the two sides of this equality are @xmath1531 and @xmath1532 , respectively . thus @xmath1533 . as above ,
the operators @xmath1534 for @xmath1535 , @xmath1536 , @xmath48 are bounded and generate @xmath1537 . hence @xmath1538 we used here that @xmath16 is a homomorphism on @xmath13 and that @xmath3 .
[ lem : fell_bundle_structure ] assume that @xmath11 is a c*-hull and that integrability is compatible with induction .
there is a unique fell bundle structure with fibres @xmath7 such that the maps @xmath1539 form a fell bundle representation for any integrable representation @xmath16 of @xmath13 on a hilbert module @xmath15 .
lemmas [ lem : be_bg_multiplicative ] and [ lem : fell_bundle_rep ] show that holds under our assumptions .
hence the multiplication and involution in @xmath23 restrict to a fell bundle structure on the subspaces @xmath1540 for @xmath54 , such that the inclusions @xmath1541 give a fell bundle representation .
the induced representation @xmath1217 of @xmath13 on the hilbert @xmath11-module @xmath1542 gives a faithful representation of @xmath11 because @xmath1543 contains the identity representation .
hence the resulting representations @xmath1544 of @xmath1451 are also faithful , even isometric .
so the fell bundle structure on @xmath1545 lifts to @xmath1451 , so that the maps @xmath1546 form a fell bundle representation .
let @xmath16 be any integrable representation of @xmath13 .
the exterior direct sum @xmath1547 on the hilbert @xmath1548-module @xmath1549 is still integrable .
the resulting maps from @xmath1451 to @xmath1550 simply give block matrices @xmath1551 for @xmath1503 .
the compressions to the direct summands @xmath15 and @xmath1552 therefore restrict to fell bundle representations with respect to the fell bundle structure on @xmath1553 defined above .
since @xmath1217 is faithful , the projection @xmath1554 is a fell bundle isomorphism .
hence the map @xmath1555 is a fell bundle representation .
let @xmath1556 be a fell bundle over a discrete group @xmath0 ( see @xcite ) .
then @xmath1557 is a @xmath0-graded * -algebra using the given multiplications and involutions among the subspaces @xmath1558 .
section c*-algebra _ @xmath1559 of the fell bundle is defined as the completion of @xmath1560 in the maximal c*-seminorm . by construction ,
a representation of @xmath1559 is equivalent to a representation of the fell bundle .
this holds also for representations on hilbert modules .
[ the : fell_bundle_sections_hull ] let @xmath13 be a graded * -algebra for which @xmath6 has a c*-hull .
assume that integrability is compatible with induction as in definition . the section c*-algebra
@xmath8 of the fell bundle @xmath1561 constructed above is a c*-hull for the integrable representations of @xmath13 .
representations of @xmath8 are in natural bijection with fell bundle representations : restricting a representation of @xmath8 to the subspaces @xmath1451 gives a fell bundle representation , and conversely a fell bundle representation gives a representation of the * -algebra @xmath1562 , which extends uniquely to the c*-completion .
lemma [ lem : fell_bundle_structure ] says that any integrable representation @xmath1563 of @xmath13 induces a fell bundle representation @xmath1564 of @xmath7 and thus a representation of @xmath8 . by construction , this family of maps @xmath1565 is compatible with interior tensor products and unitary * -intertwiners .
we are going to show that this is a family of bijections .
first we describe an integrable representation @xmath329 of @xmath13 on @xmath8 . by construction , @xmath1566 is dense in @xmath8 .
this subspace carries a representation of @xmath13 by left multiplication .
we extend this to the right ideal in @xmath8 generated by @xmath1567 to get a representation of @xmath13 on @xmath8 .
let @xmath329 be its closure .
the representations @xmath1491 of @xmath11 on @xmath1451 are defined so that @xmath1568 is another core for the representation @xmath1477 of @xmath6 on @xmath1451 .
therefore , @xmath1569 is a core for the restriction of the representation @xmath329 to @xmath6 .
this core shows that @xmath1570 , where the interior tensor product is with respect to the canonical embedding @xmath1571 .
therefore , the restriction of @xmath329 to @xmath6 is integrable and the corresponding representation @xmath1572 of @xmath11 is simply the inclusion map @xmath1573 .
thus the representation @xmath329 of @xmath13 on @xmath8 is also integrable .
the integrable representation @xmath329 of @xmath13 on @xmath8 yields a representation @xmath1574 of the fell bundle @xmath7 in @xmath1575 . by construction ,
the image of @xmath1576 in @xmath1451 acts by @xmath1577 .
that is , @xmath1451 is represented by the canonical inclusion map @xmath1578 .
the representation of @xmath8 associated to this fell bundle representation is the identity map on @xmath8 .
interior tensor product with @xmath329 gives a family of maps @xmath1579 that is compatible with unitary * -intertwiners and interior tensor products .
since the composite family of maps @xmath1580 is compatible with interior tensor products and maps the identity representation of @xmath8 to itself , the composite map on @xmath1581 is the identity .
let @xmath62 be an integrable representation of @xmath13 on a hilbert @xmath14-module @xmath15 for some c*-algebra @xmath14 .
this yields a representation @xmath1564 of the fell bundle @xmath7 and an associated representation @xmath972 of @xmath8 .
we claim that the integrable representation @xmath1582 is equal to @xmath62 .
since @xmath1569 is a core for @xmath329 , the subspace @xmath1583 is a core for @xmath1584 ( compare the second paragraph in the proof of lemma [ lem : be_bg_multiplicative ] ) .
since @xmath1475 is the representation of @xmath6 associated to @xmath1465 and the graph topologies for @xmath6 and @xmath13 are equivalent by lemma [ lem : restriction_closed ] , the subspace @xmath1583 is a core for @xmath62 as well . on @xmath1585 ,
@xmath1333 acts by mapping @xmath1586 in both representations , where we view @xmath1587 . since @xmath62 and
@xmath1584 have a common core , they are equal .
this finishes the proof that our two families of maps @xmath1588 are inverse to each other .
thus @xmath8 is a weak c*-hull for the integrable representations of @xmath13 .
since @xmath6 is a c*-hull , the integrable representations of @xmath6 are admissible .
so are the integrable representations of @xmath13 by proposition [ pro : integrable_induced_nice ] .
thus @xmath8 is a c*-hull .
[ rem : fell_explicit ] the unit fibre @xmath11 of the fell bundle in theorem [ the : fell_bundle_sections_hull ] is _ not _ the original c*-hull of @xmath6 , but the quotient @xmath44 , which is a c*-hull for the _ inducible _ , integrable representations of @xmath6 .
the fibre @xmath50 is given by definition [ def : fibre_g ] .
we need technical extra assumptions to construct the involution and multiplication in the fell bundle @xmath12
. hence we should not expect simple explicit formulas for them .
one way to construct this fell bundle is by inducing the universal inducible , integrable representation of @xmath13 on @xmath44 to an integrable representation @xmath1589 of @xmath13 on the hilbert @xmath44-module @xmath1590 . by the proof of lemma [ lem : fell_bundle_structure ] ,
the fell bundle @xmath12 is represented faithfully in @xmath1591 .
that is , the multiplication , involution , and norm in our fell bundle are simply the multiplication , involution and norm in the c*-algebra @xmath1591 . the dense image of @xmath1441 in @xmath50 acts on @xmath1590 by @xmath1592 , where @xmath1593 is the representation of the c*-hull @xmath44 associated to the inducible , integrable representation @xmath1594 of @xmath6 on @xmath1595 .
two assumptions limit the generality of the induction theorem [ the : fell_bundle_sections_hull ] .
first , integrability must be compatible with induction .
secondly , @xmath11 should be a c*-hull and not a weak c*-hull .
equivalently , all isometric intertwiners between integrable hilbert space representations of @xmath6 are * -intertwiners .
we show by two simple counterexamples that both assumptions are needed . in particular , there is no version of the induction theorem for weak c*-hulls .
both counterexamples involve the group @xmath1596 .
a @xmath0-graded * -algebra is a _
* -superalgebra _ , that is , a * -algebra with a decomposition @xmath1597 such that @xmath1598 in both examples , @xmath1599 $ ] with @xmath35 . in the first example
, @xmath13 is the crossed product for the action of @xmath1600 on @xmath1599 $ ] through the involution @xmath1601 .
that is , @xmath1602 since @xmath1603 as a right @xmath1604-module , any representation of @xmath1604 is inducible .
let @xmath1605 with the representation of @xmath1604 from the inclusion map @xmath1606 ( see proposition [ pro : commutative_representation ] ) .
this gives a c*-hull for a class of representations of @xmath1604 that is defined by submodule conditions and satisfies the strong local
global principle by theorems [ the : commutative_hull ] and [ the : commutative_local - global ] .
the class of @xmath1607-integrable representations consists of those representations of @xmath37 $ ] that are generated by a regular , self - adjoint , _ strictly positive _ operator . in a representation of @xmath13 , the element @xmath1608 acts by a unitary involution that conjugates @xmath38 to @xmath1609 .
hence @xmath38 can not be strictly positive .
thus the zero - dimensional representation is the only representation of @xmath13 whose restriction to @xmath1604 is @xmath1610-integrable .
the c*-hull for this class is @xmath988 .
theorem [ the : fell_bundle_sections_hull ] does not apply here because induced representations of inducible , integrable representations of @xmath1604 are _ never _ integrable when they are non - zero .
the second example is the * -superalgebra @xmath1611 thus @xmath13 is commutative and @xmath1612 \cong a_0 $ ] with the usual @xmath1604-bimodule structure and the inner product @xmath1613 .
since @xmath1614 is positive for any @xmath523 $ ] , any representation of @xmath1604 is inducible .
let @xmath62 be a representation of @xmath1604 on a hilbert module @xmath15 over a c*-algebra @xmath14 .
the induced representation @xmath1615 lives on the hilbert @xmath14-module completion @xmath149 of @xmath21 for the inner product @xmath1616 .
its domain is @xmath21 , viewed as a dense @xmath14-submodule in @xmath149 , and the representation of @xmath1604 is @xmath16 again .
the operator @xmath1617 on @xmath21 extends to an isometry @xmath1618 because @xmath1619 for all @xmath1620 .
this isometry commutes with @xmath67 for all @xmath20 , so it is an isometric intertwiner @xmath1621 .
now let @xmath1622 with the universal representation @xmath1623 be one of the two noncommutative weak c*-hulls @xmath822 or @xmath829 of @xmath37 $ ] described in [ sec : polynomials2 ] . in a toeplitz integrable representation , @xmath1617 has dense range .
even more , @xmath1624 is dense in @xmath21 in the graph topology .
thus the isometry @xmath1621 is a unitary * -intertwiner for any integrable representation @xmath62 of @xmath1604 . since all representations of @xmath1604 are inducible , the unit fibre of the fell bundle should be @xmath1622 .
the other fibre @xmath342 is @xmath1625 , which we have identified with @xmath1622 .
the unitary @xmath1626 is a * -intertwiner between the representations of @xmath1604 as well .
therefore , integrability is compatible with induction . and
the left @xmath1622-module structure @xmath1627 on @xmath342 in is simply left multiplication .
next we describe the induced representation of @xmath13 on the hilbert @xmath1622-module @xmath1628 the representations of @xmath13 and @xmath1604 on @xmath1629 have the same domain by lemma [ lem : restriction_closed ] , and for @xmath1604 the domain is @xmath1630 .
we claim that @xmath13 acts on this domain by @xmath1631 we have already seen this for @xmath1632 .
left multiplication by @xmath1633 maps @xmath1634 first to @xmath1635 , which is mapped by the isometry @xmath159 to @xmath1636 . and
it maps the element @xmath1637 for @xmath1638 , which corresponds to @xmath1639 in the odd fibre , to @xmath1640 .
this proves the formula for the action of @xmath1633 .
the representation @xmath1641 of @xmath1622 on @xmath1629 is the representation of the weak c*-hull that corresponds to the representation of @xmath1642 described above .
this is @xmath1643 hence @xmath1644 for @xmath1638 acts by the matrix @xmath1645 the map @xmath1646 is the cayley transform of @xmath1647 . for our two weak c*-hulls ,
this is the unilateral shift @xmath1648 by construction .
thus the odd fibre @xmath1649 of our fell bundle should act by @xmath1650 the map @xmath1641 is a * -representation , and gives @xmath1651 for all @xmath1652 .
this is also obvious from our explicit formulas .
but @xmath1653 differ if , say @xmath1654 , @xmath1655 .
in fact , @xmath1656 is not contained in @xmath1657 .
hence there is no fell bundle structure on @xmath1658 for which @xmath1659 would be a fell bundle representation .
we now specialise the induction theorem [ the : fell_bundle_sections_hull ] to the case where the universal integrable representation of the unit fibre @xmath6 is locally bounded . in this case
, we may first construct a pro - c*-algebraic fell bundle whose unit fibre is the pro - c*-algebra completion of @xmath6 .
this is relevant because pro - c*-algebras are much closer to ordinary c*-algebras than general * -algebras .
so a pro - c*-algebraic fell bundle already has important extra properties .
we will see the importance of this in the commutative case , where the pro - c*-algebraic fell bundle gives us a twisted partial group action on the space @xmath1660 of positive characters . as before , let @xmath0 be a group and let @xmath1661 be a @xmath0-graded * -algebra .
we are interested in the locally bounded representations of @xmath6 , and representations of @xmath13 that restrict to locally bounded representations on @xmath6 .
the class @xmath1662 of locally bounded representations of @xmath6 is admissible by corollary [ cor : locally_bounded_admissible ] , so any weak c*-hull for some smaller class of representations will be an ordinary c*-hull .
let @xmath1663 be the pro - c*-algebra completion of the unit fibre @xmath6 , that is , the completion of @xmath6 in the topology defined by the directed set @xmath1664 of all c*-seminorms on @xmath6
. locally bounded representations of @xmath6 are equivalent to locally bounded representations of @xmath1663 by proposition [ pro : locally_bounded_pro - cstar ] . when is a locally bounded representation inducible ?
[ pro : locally_bounded_inducible ] a locally bounded representation @xmath1665 of @xmath6 on a hilbert module @xmath15 is inducible if and only if the operators @xmath1666 are positive for all @xmath1527 , @xmath54 .
the difference to the general criterion for inducibility in proposition [ pro : inducible_criteria ] is that we do not consider matrices .
the subspace @xmath1667 of bounded vectors is a core for @xmath387 . as in the proof of proposition
[ pro : inducible_a_b ] , it suffices to prove the positivity of the inner product for a finite linear combination @xmath1668 with @xmath1417 , @xmath1669 for a fixed @xmath54 .
since there are only finitely many @xmath1290 , there is a c*-seminorm @xmath906 on @xmath6 so that all @xmath1290 are @xmath906-bounded .
thus we may replace @xmath15 by the hilbert submodule @xmath965 of @xmath906-bounded vectors , where the representation of @xmath6 extends to the c*-completion @xmath1670 for @xmath906 .
since we assume @xmath1666 for all @xmath48 , this representation factors through the quotient of @xmath14 by the closed ideal @xmath159 generated by the negative parts @xmath1671 for all @xmath20 .
the @xmath1672-valued inner product @xmath1673 on @xmath1487 is positive definite by construction ; since @xmath1672 is a c*-algebra , we may use the usual notion of positivity here , which does not involve matrices .
then the inner product on the tensor product @xmath1674 is also positive definite .
this is what we had to prove .
a pro - c*-algebra has a functional calculus for self - adjoint elements .
hence we may construct the negative parts @xmath1675 for @xmath48 , @xmath54 .
we let @xmath1676 be the completed quotient of @xmath1663 by the closed two - sided ideal generated by these elements .
this is another pro - c*-algebra , and it is the largest quotient in which @xmath1677 for all @xmath48 , @xmath54 . by proposition [ pro : locally_bounded_inducible ] ,
a locally bounded representation of @xmath6 is inducible if and only if the corresponding locally bounded representation of @xmath1663 factors through @xmath1676 .
[ cor : locally_bounded_inducible_pro - cstar ] there is an equivalence between the inducible , locally bounded representations of @xmath6 and the locally bounded representations of the pro - c*-algebra @xmath1676 , which is compatible with isometric intertwiners and interior tensor products .
proposition [ pro : locally_bounded_inducible ] says that the equivalence in proposition [ pro : locally_bounded_pro - cstar ] maps the subclass in @xmath1662 of inducible , locally bounded representations of @xmath6 onto the subclass @xmath1678 in @xmath1679 .
let @xmath1680 be the directed set of c*-seminorms on @xmath1676 .
this is isomorphic to the subset of @xmath1664 consisting of all c*-seminorms on @xmath6 for which @xmath1681 holds in the c*-completion @xmath1682 for all @xmath48 , @xmath54 .
we would like to complete @xmath13 to a * -algebra @xmath1683 with unit fibre @xmath1676 , where each @xmath1684 is a hilbert bimodule over @xmath1676 .
but such a construction does not work in the following example .
[ exa : locally_bounded_induction_non - compatible ] it can happen that the class of locally bounded representations of @xmath6 is not compatible with induction . let @xmath1685)$ ] be the * -algebra of all @xmath1686-matrix with only finitely many entries in each row and each column , with the usual matrix multiplication and involution .
let @xmath13 be the @xmath1600-graded * -algebra of block @xmath1687-matrices @xmath1688,\ c\in { \mathbb c}[{\mathbb n}],\ d\in \operatorname{end}^*({\mathbb c}[{\mathbb
n}]),\ ] ] with the grading where @xmath1689 are even and @xmath1690 are odd . here @xmath446 and @xmath558 are infinite column and row vectors with only finitely many non - zero entries , respectively .
thus @xmath1691)$ ] with the grading induced by the grading on @xmath1692 $ ] where @xmath1693 is the even part and the span of @xmath1694 for @xmath1695 is the odd part .
the character @xmath1696 is a bounded representation of the unit fibre @xmath1604 .
induction gives the standard representation of @xmath13 on the hilbert space @xmath1697 by matrix - vector multiplication .
this representation is irreducible because already the ideal of finite matrices @xmath1698 in @xmath13 acts irreducibly .
it is not bounded , that is , some elements in @xmath1685)$ ] act by unbounded operators on @xmath898 .
hence it is not locally bounded by proposition [ pro : locally_bounded_irreducible ] . to rule out this problem
, we now _ assume that induction from @xmath6 to @xmath13 and restriction back to @xmath6 maps bounded representations of @xmath6 again to bounded representations of @xmath6 _ , briefly , that boundedness is compatible with induction .
this implies that local boundedness is compatible with induction because a locally bounded representation contains bounded subrepresentations whose union is a core for it .
our assumption is equivalent to the boundedness of the induced representations of @xmath6 on the hilbert @xmath1699-modules @xmath1700 for all @xmath54 and @xmath1701 .
that is , there is another norm @xmath1702 such that @xmath1703 for all @xmath48 , @xmath1704 .
let @xmath1684 be the completion of @xmath1487 in the topology generated by the family of norms @xmath1705 for @xmath1706 .
[ lem : mult_lb_continuous ] the multiplication maps and the involutions in @xmath1707 extend to continuous maps @xmath1708 and @xmath1709 for @xmath1329 . given @xmath1706 ,
let @xmath1710 be such that @xmath1711 for all @xmath123 , @xmath1704 . if @xmath1712 , @xmath123 , then @xmath1713 that is , the multiplication is jointly continuous with respect to the topology defining @xmath1714 and hence extends to a jointly continuous map @xmath1715 .
furthermore , @xmath1716 and hence @xmath1717 for all @xmath991 .
that is , @xmath1718 for all @xmath123 .
thus the involution is continuous as well .
the completion @xmath1719 of @xmath13 is again a * -algebra by lemma [ lem : mult_lb_continuous ] . by construction of @xmath1676 ,
the inner products @xmath1720 are positive for @xmath48 , @xmath54 ; this remains so for @xmath1721 because the subset of positive elements in @xmath1676 is closed .
thus @xmath1714 has the usual properties of a fell bundle over @xmath0 , except that the fibres are only hilbert bimodules over a pro - c*-algebra
. we may interpret @xmath1714 as a partial action of @xmath0 on @xmath1676 by hilbert bimodules as in @xcite .
usually , the norms @xmath1705 and @xmath1722 on @xmath1487 are not equivalent for a fixed @xmath1723 .
this prevents us from completing @xmath1724 to a pro - c*-algebra .
it also means that the integrable representations of @xmath13 are not locally bounded on @xmath13 , but only on @xmath6 .
this happens in interesting examples such as the weyl algebra discussed in
[ sec : weyl_twisted ] .
this phenomenon for fell bundles is related to the known problem that crossed products for group actions on pro - c*-algebras only work well if the action is strongly bounded , that is , the invariant continuous c*-seminorms are cofinal in the set of all continuous c*-seminorms , see @xcite .
[ pro : locally_bounded_fell_representations ] suppose that boundedness for representations of @xmath6 is compatible with induction to @xmath13 .
representations of @xmath13 that restrict to locally bounded representations on @xmath6 are equivalent to representations of the * -algebra @xmath1725 that restrict to locally bounded representations on @xmath1676 ; this equivalence is compatible with isometric intertwiners and interior tensor products .
let @xmath16 be a representation of @xmath13 for which @xmath1343 is a locally bounded representation of @xmath6 .
the representation @xmath1343 is inducible by lemma [ lem : induced_repr_inducible ] .
hence @xmath1343 is the closure of the restriction of a locally bounded representation @xmath1465 of @xmath1676 by corollary [ cor : locally_bounded_inducible_pro - cstar ] .
then the representation @xmath1726 of @xmath1487 for @xmath54 is continuous with respect to the topology defining @xmath1684 and the graph topology on the domain of @xmath1726 because @xmath1727 .
hence it extends uniquely to @xmath1684 , and this gives a representation @xmath972 of @xmath1728 such that @xmath16 is the closure of @xmath975 .
it is easy to see that this equivalence between the locally bounded representations of @xmath13 and the representations of @xmath1729 that are locally bounded on @xmath1676 is compatible with isometric intertwiners and interior tensor products .
we will explore the consequences of this in the case of commutative @xmath6 in [ sec : commutative_ae ] . in that case ,
boundedness is automatically compatible with induction , and the pro - c*-algebraic fell bundle @xmath1676 gives rise to a twisted groupoid with object space @xmath1660 , which explains why the c*-hull produced by the induction theorem [ the : fell_bundle_sections_hull ] is a twisted groupoid c*-algebra when @xmath6 is commutative and the integrable representations of @xmath6 are locally bounded . here
we briefly consider the situation of theorem [ the : prim_a_locally_compact_gives_hull ] where @xmath1730 is dense in @xmath1676 and provides a c*-hull for the class of locally bounded representations .
then we define @xmath1731 that is , @xmath1732 if and only if for all @xmath1039 there is @xmath1706 such that @xmath1733 for all @xmath1734
. since the involutions @xmath1735 and @xmath1736 are both continuous , they are homeomorphisms .
thus @xmath1737 if and only if @xmath1738 . the proof of lemma [ lem : compactly_supported_ideal ] shows that @xmath1739 and @xmath1740 are dense in @xmath1741 .
[ the : fell_lb ] assume that boundedness is compatible with induction from @xmath6 to @xmath13 and that @xmath1730 is dense in @xmath1676 .
then @xmath1742 is a fell bundle over @xmath0 whose section c*-algebra is a c*-hull for the class of all representations of @xmath13 that restrict to a locally bounded representation of @xmath6 .
the assumption that boundedness is compatible with induction allows us to build the pro - c*-algebraic fell bundle @xmath1714 . call a representation of @xmath1743 or @xmath1744 integrable if the restriction to the unit fibre @xmath6 or @xmath1676 is locally bounded , respectively .
these classes of integrable representations are equivalent by proposition [ pro : locally_bounded_fell_representations ] .
since @xmath1730 is dense in @xmath1676 , it is a c*-hull for the locally bounded representations of @xmath1676 by theorem [ the : locally_bounded_relative_hull ] .
equivalently , it is a c*-hull for the inducible , locally bounded representations of @xmath6 .
let @xmath1745 .
representations of @xmath1746 are equivalent to representations of the fell bundle @xmath1747 .
thus the theorem says that the class of all representations of @xmath1746 is equivalent to the class of integrable representations of @xmath1724 .
more precisely , the equivalence maps a representation @xmath250 of @xmath1746 on a hilbert module @xmath15 to the representation @xmath16 of @xmath1724 with the core @xmath1748 and @xmath1749 for all @xmath1750 , @xmath1751 , @xmath340 ; here @xmath1752 is the product in @xmath1724 , which belongs to @xmath1746 if @xmath1751 . in the converse direction
, we may simply restrict a locally bounded representation of @xmath1724 to the * -subalgebra @xmath1746 .
this restriction is nondegenerate because @xmath1753 acts nondegenerately in any integrable representation of @xmath1724 : this is part of the equivalence between representations of @xmath1730 and locally bounded representations of @xmath1676 in theorem [ the : locally_bounded_relative_hull ] .
we claim that the maps from representations of @xmath1746 to integrable representations of @xmath1724 and back are inverse to each other .
let @xmath16 be an integrable representation of @xmath1724 on a hilbert module @xmath15 .
the representations @xmath16 and @xmath1754 have the same domain by lemma [ lem : restriction_closed ] .
since @xmath1754 is locally bounded , @xmath1755 is a core for @xmath1754 . since @xmath1756 for all @xmath54 ,
this subspace is @xmath1757-invariant and thus a core for @xmath16 .
the representation @xmath250 of @xmath1746 is the closure of the restriction of @xmath16 to @xmath1758 . by definition , the representation of @xmath1724 has the core @xmath1759 and acts there by @xmath1760 .
the subspace @xmath1761 is a core for this representation because the map @xmath1762 is continuous in the norm topology on @xmath15 and @xmath21 is dense in @xmath15 .
if @xmath81 , then @xmath1763 and hence @xmath1764 for all @xmath1750 , @xmath1765 , @xmath81 .
this implies @xmath479 , as desired .
now start with a representation @xmath250 of @xmath1746 .
let @xmath16 be the associated integrable representation of @xmath1724 .
it has the core @xmath1748 and acts there by @xmath1766 for all @xmath1750 , @xmath1767 , @xmath340 .
in particular , if @xmath1768 , then @xmath1769 . since @xmath1770 is dense in @xmath1741 for all @xmath54 , the restriction of @xmath250 to @xmath1771 remains nondegenerate .
therefore , the set of @xmath1772 for @xmath1767 , @xmath340 is dense in @xmath15 .
hence @xmath250 is the restriction of @xmath16 to @xmath1773 , as desired .
the proof of theorem [ the : fell_lb ] does not use constructions in section [ sec : graded_to_fell ] and so provides an alternative proof of the induction theorem in case the chosen class of integrable representations of @xmath6 is the class of all locally bounded representations .
in this section , we apply the induction theorem in the case where @xmath6 and the chosen c*-hull @xmath11 are commutative .
this is the only case considered in @xcite .
extra assumptions in @xcite ensure that the c*-hull for the integrable representations of @xmath13 is the crossed product for a partial action of @xmath0 on the space @xmath1774 of positive characters . without these assumptions , we shall get a `` twisted '' crossed product for a partial action .
so let @xmath0 be a discrete group and @xmath1775 a @xmath0-graded * -algebra such that @xmath6 is commutative .
we have already classified the possible commutative c*-hulls for @xmath6 in [ sec : commutative_hulls ] .
in particular , all commutative weak c*-hulls are already c*-hulls by theorem [ the : commutative_hull ] , and they correspond to injective , continuous maps from locally compact spaces to the spectrum @xmath1776 of @xmath6 .
explicitly , let @xmath197 be a locally compact space and let @xmath1777 be an injective , continuous map .
let @xmath1778 and define a representation of @xmath6 on @xmath11 with domain @xmath1111 by @xmath1779 for all @xmath26 , @xmath1780 , @xmath1115 , where @xmath1781 for @xmath1782 .
let @xmath1783 be the closure of this representation of @xmath6 on @xmath11 .
the c*-algebra @xmath11 with the universal representation @xmath1783 is a c*-hull for a class @xmath1784 of representations of @xmath6 by theorem [ the : commutative_hull ] , and any commutative c*-hull is of this form .
let @xmath1164 be the class of representations of @xmath13 that restrict to a representation in @xmath1784 on @xmath6 , as in definition [ def : induced_integrable ] .
if @xmath1784 is compatible with induction as in definition [ def : induced_actions_integrable ] , then theorem [ the : fell_bundle_sections_hull ] gives a fell bundle whose section c*-algebra is a c*-hull for @xmath1164 .
we are going to characterise exactly when this happens and describe the c*-hull for @xmath1164 as a twisted groupoid c*-algebra .
any representation of @xmath6 on a commutative c*-algebra is locally bounded by proposition [ pro : commutative_representation ] . hence the constructions in [ sec : locally_bounded_unit_fibre ] specialise to our commutative case .
actually , we shall make these results more explicit through independent proofs .
first we describe the c*-hull @xmath44 for the inducible representations in @xmath1784 as in proposition [ pro : locally_bounded_inducible ] : [ lem : inducible_ae_commutative ] call a character @xmath1782 _ positive _ if @xmath1785 for all @xmath48 and all @xmath54 .
these form a closed subset @xmath1660 of @xmath1776 , and @xmath1786 .
the positive characters form a closed subset in @xmath1776 by definition of the topology on @xmath1776 .
we have constructed @xmath44 in proposition [ pro : inducible_a_b ] as a quotient of @xmath11 , such that a representation is inducible if and only if it factors through @xmath44 .
thus it corresponds to a certain closed subset of @xmath1776 .
its points are the inducible characters of @xmath6
. let @xmath949 be a character .
any vector in @xmath1787 is of the form @xmath1788 for some @xmath48 , that is , there is no need to take linear combinations .
hence the sesquilinear form on @xmath1787 for all @xmath54 is positive semidefinite if and only if @xmath1789 for all @xmath48 and all @xmath54 , that is , @xmath949 is positive .
thus @xmath44 is the quotient corresponding to those @xmath1790 for which @xmath1791 is positive .
[ the : induce_character ] let @xmath54 and @xmath1792
. then @xmath1793 .
the set @xmath1794 is relatively open in @xmath1660 .
the left @xmath6-module structure on @xmath1787 for @xmath1795 is by a character @xmath1796 that belongs to @xmath1797 .
the map @xmath1798 is a homeomorphism from @xmath1799 onto @xmath1797 , and these maps form a _ partial action _ of @xmath0 on @xmath1660 ,
that is , @xmath1800 and @xmath1801 for all @xmath1802 . by definition
, @xmath1787 is the hausdorff completion of @xmath1487 in the norm coming from the inner product @xmath1803
. let @xmath1804 satisfy @xmath1805 and @xmath1806 .
we must show that @xmath98 and @xmath446 are parallel .
the following computation makes @xcite*footnote 3 explicit : @xmath1807 because @xmath6 is commutative and the terms in parentheses belong to @xmath6 .
hence @xmath1808 since @xmath1805 and @xmath1806 , this implies @xmath1809 the inner product on @xmath1787 annihilates @xmath1810 for all @xmath1811 .
hence @xmath1812 thus all non - zero @xmath1813 are parallel , that is , @xmath1814 .
the space @xmath1815 is non - zero if and only if there is @xmath48 with @xmath1805 .
thus @xmath1816 the latter set is relatively open in @xmath1660 .
let @xmath1795 .
then the dimension of @xmath1815 is @xmath314 .
hence the representation of @xmath6 on it must be by a character , which we denote by @xmath1796 .
this character is an inducible representation by lemma [ lem : induced_repr_inducible ] , and hence positive by lemma [ lem : inducible_ae_commutative ] .
there is @xmath1712 with @xmath1817 . if @xmath26 , then implies @xmath1818 .
thus @xmath1819 for all @xmath26 .
hence @xmath1820 , so that @xmath1821 by .
thus @xmath1798 maps @xmath1799 to @xmath1797 .
equation also implies that the map @xmath1798 is continuous on the open set of characters in @xmath1822 with @xmath1817 .
since these open sets for different @xmath1712 cover @xmath1799 , the map @xmath1798 is continuous on all of @xmath1799 .
let @xmath1823 and let @xmath1824 and @xmath1825 .
then there is @xmath1826 with @xmath1827 , and @xmath1828 with @xmath1829 .
thus @xmath1830 , and so for @xmath1831 describes @xmath1832 .
hence @xmath1833 thus @xmath1834 for all @xmath1802 .
we have shown that the maps @xmath1798 form a partial action of @xmath0 on @xmath1660 , see @xcite .
in particular , @xmath1798 is a homeomorphism from @xmath1799 to @xmath1797 with inverse @xmath1835 . in the examples considered in @xcite ,
the space @xmath1660 is locally compact and the c*-hull for the integrable representations of @xmath13 is the crossed product for the partial action of @xmath0 on @xmath1660 described above . in general , however , certain twists are possible .
the partial action of @xmath0 on @xmath1660 may be encoded in a transformation groupoid @xmath1836 , which has object space @xmath1660 , arrow space @xmath1837 with the disjoint union topology , range and source maps @xmath1838 , @xmath1839 for @xmath54 , @xmath1840 , and multiplication @xmath1841 for all @xmath1329 , @xmath1842 .
the unit arrow on @xmath949 is @xmath1843 , and the inverse of @xmath1844 is @xmath1845 .
this is an tale topological groupoid because @xmath1846 and @xmath876 restrict to homeomorphisms on the open subsets @xmath1799 of the arrow space .
the object space @xmath1660 need not be locally compact .
we are going to construct another topological groupoid @xmath1847 that is a central extension of @xmath1836 by the circle group @xmath1848 .
that is , @xmath1847 comes with a canonical functor to @xmath1836 whose kernel is the group bundle @xmath1849 .
such an extension is also called a _ twisted groupoid _ in @xcite*section 4 , following a definition by kumjian @xcite .
a twisted groupoid with locally compact object space has a twisted groupoid c*-algebra .
for a suitable injective continuous map @xmath1850 , we are going to identify the c*-hull of the @xmath197-integrable representations of @xmath13 with the twisted groupoid c*-algebra of the restriction of @xmath1847 to @xmath1851 .
a point in @xmath1847 is a triple @xmath1852)$ ] , where @xmath1853 , @xmath1840 , and @xmath1854 $ ] is a unit vector in the @xmath314-dimensional hilbert space @xmath1855 .
we represent unit vectors in @xmath1856 by elements @xmath48 with @xmath1857 ; two elements @xmath1804 with @xmath1858 represent the same unit vector @xmath1854=[b]$ ] if and only if @xmath1859 .
we get the same set of equivalence classes if we allow @xmath20 with @xmath1860 and set @xmath1854=[b]$ ] if @xmath1861 : then @xmath1862 and @xmath1863 satisfy @xmath1854=[a_1]$ ] , @xmath1864=[b_1]$ ] , and @xmath1854=[b]$ ] if and only if @xmath1865 by .
the circle group @xmath1848 acts on @xmath1847 by multiplication : @xmath1866 ) { \mathrel{\vcentcolon=}}(g,\chi,[\lambda a])$ ] .
the orbit space projection for this circle action is the coordinate projection @xmath1867 , @xmath1852)\mapsto ( g,\chi)$ ] .
next we equip @xmath1847 with a topology so that this coordinate projection is a locally trivial principal @xmath1848-bundle . for @xmath20 ,
let @xmath1868 .
this is an open subset in @xmath1660 , and @xmath1860 if @xmath1869 because @xmath949 is positive .
the map @xmath1870 , @xmath1871)$ ] , for @xmath48 is a local section for the coordinate projection @xmath1872 .
if @xmath1804 , and @xmath1873 , then @xmath1874 = \left[\frac{\chi(b^ * a ) } { \chi(a^ * a)^{\nicefrac12 } \chi(b^ * b)^{\nicefrac12 } } b\right].\ ] ] since the functions sending @xmath949 to @xmath1875 , @xmath1876 and @xmath1877 are continuous on @xmath6 , the two trivialisations induce the same topology on the restriction of @xmath1847 to @xmath1878 . for any @xmath1879
, there is @xmath48 with @xmath1860 . thus the open subsets
@xmath1880 cover @xmath1799 .
consequently , there is a unique topology on @xmath1847 that makes the local sections @xmath1881 for all @xmath48 continuous , and this topology turns @xmath1847 into a locally trivial @xmath1848-bundle over @xmath1836 .
we define a groupoid with object space @xmath1822 , arrow space @xmath1847 , and @xmath1882){\mathrel{\vcentcolon=}}\theta_g(\chi),\quad s(g,\chi,[a]){\mathrel{\vcentcolon=}}\chi,\quad ( g,[\theta_h(\chi)],[a])\cdot ( h,\chi,[b ] ) { \mathrel{\vcentcolon=}}(g\cdot h,\chi , [ a\cdot b]);\ ] ] we must show that this multiplication is well defined .
we have @xmath1883 and @xmath1884 by , so @xmath1885 ) \in \sigma$ ] . if @xmath1886 and @xmath1887 , then @xmath1888 by computations as in the proof of theorem [ the : induce_character ] .
hence the multiplication is well defined .
it is clearly associative .
the unit arrow on @xmath949 is @xmath1889)$ ] , and @xmath1852)^{-1 } = ( g^{-1},\theta_g(\chi),[a^*])$ ] .
the multiplication , unit map and inversion are continuous and the range and source maps are open surjections ( even locally trivial ) .
so @xmath1847 is a topological groupoid .
the identity map on objects and the coordinate projection @xmath1890 on arrows form a functor , which is a locally trivial , open surjection on arrows .
the kernel of @xmath1872 consists of those @xmath1852)\in\sigma$ ] for which @xmath1891)$ ] is a unit arrow in @xmath1836 .
then @xmath1892 , and @xmath26 is equivalent to @xmath1854 = [ \chi(a)\cdot 1]$ ] because @xmath1893 .
the map @xmath1852)\mapsto ( \chi,\chi(a))$ ] is an isomorphism of topological groupoids from the kernel of @xmath1872 onto the trivial group bundle @xmath1894 .
thus we have an extension of topological groupoids @xmath1895 the three groupoids above are clearly hausdorff . to construct c*-algebras , we need groupoids with a locally compact object space . therefore ,
we replace @xmath1776 by a locally compact space @xmath197 with an injective , continuous map @xmath1896 .
then @xmath1897 is a c*-hull for a class @xmath1784 of representations of @xmath6 . by lemma [ lem : inducible_ae_commutative ] , the c*-hull for the class of @xmath197-integrable , inducible representations of @xmath6 is @xmath1898 with @xmath1899 .
[ cor : commutative_integrability_induction ] let @xmath1896 be an injective , continuous map .
the class @xmath1164 is compatible with induction if and only if @xmath1851 is invariant under the partial maps @xmath1798 in theorem and the resulting partial maps on @xmath1900 are continuous in the topology of @xmath1900 .
we briefly say that the partial action of @xmath0 on @xmath1660 _ restricts _ to @xmath1900 . by proposition
[ pro : inducible_universal ] , it suffices to check that the induced representation of @xmath6 on @xmath1901 is @xmath197-integrable for @xmath54 if and only if the partial map @xmath1902 on @xmath197 factors through @xmath765 and the resulting partial map @xmath1903 on @xmath197 is again continuous .
view the hilbert module @xmath1904 as a continuous field of hilbert spaces over @xmath1900 .
the fibres of this field have dimension at most @xmath314 by theorem [ the : induce_character ] , and the set where the fibre is non - zero is the open subset @xmath1905 .
hence @xmath1906 .
the representation of @xmath6 on @xmath1904 is equivalent to a representation on @xmath1907 by proposition [ pro : representation_hilm_comp ] .
this is equivalent to a continuous map @xmath1908 by proposition [ pro : commutative_representation ] .
this map is @xmath1909 by a fibrewise computation .
hence the induced representation of @xmath6 on @xmath1910 is @xmath197-integrable if and only if @xmath1909 has values in @xmath1911 and the partial maps @xmath1912 on @xmath197 are continuous . from now on
, we assume that the partial action of @xmath0 on @xmath1660 restricts to @xmath1900 . by corollary [ cor : commutative_integrability_induction ] ,
this assumption is necessary and sufficient for @xmath197-integrability to be compatible with induction .
the `` restriction '' of the partial action on @xmath1660 to @xmath1900 is a partial action of @xmath0 on @xmath1900 by partial homeomorphisms .
its transformation groupoid @xmath1913 is constructed like @xmath1836 .
its set of arrows is the subset of @xmath1836 of arrows with range and/or source in @xmath1914 , and the topology on the arrow space is the unique one that makes the inclusion @xmath1915 and the range and source maps @xmath1916 continuous .
there is also a unique topology on the restriction @xmath1917 of @xmath1847 to @xmath1914 so that there is an extension of topological groupoids @xmath1918 since @xmath1900 is locally compact , the groupoids in this extension are locally compact , hausdorff groupoids . since @xmath1913 is tale , it carries a canonical haar system , namely , the family of counting measures .
there is also a unique normalised haar system on @xmath1919 .
these produce a unique haar system on @xmath1917 by @xcite*theorem 5.1 , so that the groupoid c*-algebra @xmath1920 is defined .
the _ twisted groupoid c*-algebra _
@xmath1921 of @xmath1913 with respect to the _ twist _
@xmath1917 is defined in @xcite .
it is related to the groupoid c*-algebra of @xmath1917 in @xcite*corollary 7.2 .
[ the : twisted_groupoid ] let @xmath0 be a discrete group and let @xmath13 be a @xmath0-graded * -algebra with commutative @xmath6 .
let @xmath1922 be an injective , continuous map , such that the partial action of @xmath0 on @xmath1660 in theorem restricts to @xmath1900 as in corollary .
then @xmath1921 is a c*-hull for @xmath1164 .
the c*-algebra @xmath1921 may be defined as the full section c*-algebra of a certain fell line bundle over the tale , locally compact groupoid @xmath1913 .
the fell line bundle involves the space of sections of the hermitian complex line bundle @xmath1923 associated to the principal @xmath1848-bundle @xmath1924 and the multiplication maps @xmath1925 induced by the multiplication of @xmath1917 ( see @xcite ) . by construction ,
the hilbert @xmath44-module @xmath1926 is isomorphic to the continuous sections of this line bundle @xmath1927 over the subset @xmath1928 of @xmath1929 : an element @xmath1930 is mapped to the continuous section that sends @xmath1931 for @xmath1115 with @xmath1932 to @xmath1933 $ ] .
the multiplication in @xmath1917 is defined so that the multiplication maps @xmath1934 are exactly the multiplication maps in the fell line bundle associated to @xmath1917 .
thus the fell bundle @xmath12 constructed in theorem [ the : fell_bundle_sections_hull ] is isomorphic to the fell bundle @xmath1935 , where @xmath1936 is the space of @xmath1937-sections of @xmath1927 over @xmath1938 and the multiplication and involution come from the fell line bundle structure on @xmath1927 over the groupoid @xmath1913 .
the full section c*-algebra of this fell bundle is canonically isomorphic to the section c*-algebra of the corresponding fell bundle over the groupoid @xmath1913 by results of @xcite .
the small issue to check here is that it makes no difference whether we use @xmath1937-sections or compactly supported continuous sections of @xmath1927 over @xmath1938 .
both have the same c*-completion .
this is a special case of general results about fell bundles over tale locally compact groupoids . if @xmath7 is any fell bundle over @xmath0 , then @xmath1562 is a * -algebra , to which we may apply our machinery although all its representations are bounded .
thus any fell bundle over @xmath0 may come up for some choice of the @xmath0-graded * -algebra @xmath13 .
thus the section c*-algebra of a fell bundle @xmath7 with commutative unit fibre is always a twisted groupoid c*-algebras of a twist of an tale groupoid , namely , the transformation groupoid of a certain partial action on the spectrum of the unit fibre associated to the fell bundle .
this result is already known , even for fell bundles over inverse semigroups with commutative unit fibre , see @xcite .
if @xmath1939 as a groupoid , then @xmath1940 .
this is the same as the crossed product for the partial action of @xmath0 on @xmath1900 .
this happens in all the examples in @xcite .
the possible twists have two levels .
first , @xmath1917 may be non - trivial as a principal circle bundle over @xmath1913 .
secondly , if it is trivial as a principal circle bundle , the multiplication may create a non - trivial twist .
the circle bundle @xmath1941 is trivial if and only if its restriction to @xmath1938 is trivial for each @xmath54 . for a circle bundle ,
this means that there is a nowhere vanishing section .
for instance , if there is @xmath48 that generates @xmath1487 as a right @xmath6-module , then @xmath1942 and @xmath1881 is a global trivialisation of @xmath1943 .
the complex line bundles over a space @xmath197 are classified by the second cohomology group @xmath1944 .
if @xmath1927 is a line bundle , then the spaces of @xmath1937-sections of @xmath1945 for @xmath1946 form a fell bundle over @xmath59 , and the direct sum of these spaces of sections is a @xmath59-graded * -algebra such that the given line bundle @xmath1927 appears in the resulting twisted groupoid . if @xmath1947 , the space @xmath197 is at least @xmath61-dimensional .
there are indeed non - trivial complex line bundles over all _ compact _ oriented @xmath61-dimensional manifolds .
the resulting * -algebra , however , has only * -representations by bounded operators if @xmath197 is compact .
examples where unbounded operators appear must involve a non - trivial line bundle over a noncompact space .
these first appear in dimension @xmath1948 .
it is easy to write down a @xmath59-graded * -algebra @xmath13 where @xmath44 is , say , @xmath1949 and @xmath50 involves the bott line bundle over @xmath1950 .
these examples seem artificial , however .
now assume that @xmath1917 is trivial as a principal circle bundle over @xmath1951 , that is , @xmath1952 as a @xmath1848-space .
we may choose this homeomorphism to be the obvious one on the open subset @xmath1953 corresponding to @xmath1954 .
the multiplication must be of the form @xmath1955 for some continuous @xmath1848-valued function @xmath387 with @xmath1956 for all @xmath1957 ; here @xmath387 is defined on the space of all triples @xmath1958 with @xmath1959 and @xmath1960 ; this space is homeomorphic to the space @xmath1961 of pairs of composable arrows in @xmath1913 .
the associativity of the multiplication in @xmath1917 is equivalent to the cocycle condition @xmath1962 for all @xmath1963 , @xmath1964 for which @xmath1965 , @xmath1966 , and @xmath1967 are defined .
a different trivialisation of the circle bundle @xmath1968 modifies @xmath387 by the coboundary @xmath1969 of a continuous function @xmath1970 normalised by @xmath1971 for all @xmath1964 .
thus isomorphism classes of twists of @xmath1913 are in bijection with the groupoid cohomology @xmath1972 , that is , the quotient of the group of continuous maps @xmath1973 satisfying by the group of @xmath61-coboundaries @xmath1974 of continuous @xmath314-cochains @xmath1970 , where @xmath1975 is defined in . in the easiest case , the function @xmath387 above does not depend on @xmath257 .
then @xmath1976 is a normalised @xmath61-cocycle on @xmath0 in the usual sense .
these cocycles appear , for instance , in the classification of projective representations of the group @xmath0 .
this is related to the twists above because the hilbert space representations of the twisted group algebra for a @xmath61-cocycle @xmath1977 are exactly the projective representations @xmath1978 with @xmath1979 for all @xmath1329 .
the group @xmath59 has no nontrivial @xmath61-cocycles .
they do appear , however , for the group @xmath1980 .
a well known example is the noncommutative torus .
its usual gauge action corresponds to a @xmath1980-grading , where @xmath1981 for the canonical generators @xmath1982 has degree @xmath1983 . in this case ,
@xmath1984 , and @xmath1660 has only one point .
the transformation groupoid @xmath1836 is simply @xmath1985 .
this is discrete , so @xmath1847 is always trivial as a principal circle bundle .
thus the only non - trivial aspect of @xmath1847 is a @xmath61-cocycle @xmath1986 .
the cohomology group @xmath1987 is isomorphic to @xmath1848 , and the resulting twisted group algebras of @xmath1980 are exactly the noncommutative tori .
[ pro : commutative_partial_action_no_twist ] if there are subsets @xmath1988 such that @xmath1989 generates @xmath1487 as a right @xmath6-module , @xmath1990 , and @xmath1991 for all @xmath1992 , @xmath54 , @xmath1993 , then the twist @xmath1917 is trivial and so the c*-hull of @xmath13 is @xmath1994 .
we have @xmath1995 because @xmath1989 generates @xmath1487 as a right @xmath6-module ; and @xmath1852)= ( g,\chi,[b])$ ] for all @xmath1996 , @xmath1997 because @xmath1991 for all @xmath1998 .
hence the local sections @xmath1881 of @xmath1943 for @xmath1999 coincide on the intersections of their domains and thus combine to a global trivialisation .
if @xmath1660 itself is locally compact , then we may take @xmath2000 with the inclusion map @xmath765 .
since @xmath1660 is closed in @xmath1776 , this happens if @xmath1776 is locally compact .
[ the : ae_locally_compact_hull ] assume that @xmath1660 is locally compact in the topology @xmath926 .
call a representation of @xmath13 integrable if its restriction to @xmath6 is locally bounded .
let @xmath1847 be the twisted groupoid constructed above .
then @xmath2001 is a c*-hull for the integrable representations of @xmath13 . if @xmath2002 , then integrability is compatible with induction by corollary [ cor : commutative_integrability_induction ] because the construction of the topology @xmath926 is natural and compatible with restriction to open subsets .
theorem [ the : twisted_groupoid ] shows that @xmath2001 is a c*-hull for the class of representations of @xmath13 whose restriction to @xmath6 is @xmath1660-integrable .
the locally bounded representations of @xmath6 are equivalent to the locally bounded representations of the pro - c*-algebra @xmath2003 by propositions [ pro : commutative_pro - completion ] and [ pro : locally_bounded_pro - cstar ] . restrictions of representations of @xmath13 to @xmath6 are automatically inducible by lemma [ lem : induced_repr_inducible ] . by a pro - c*-algebraic variant of lemma [ lem : inducible_ae_commutative ] , the inducible
, locally bounded representations of @xmath6 are equivalent to those representations of @xmath2003 that factor through the quotient @xmath2004 . since @xmath1660 is locally compact
, @xmath2005 is dense in the pro - c*-algebra @xmath2004 .
hence @xmath2005 is a c*-hull for the inducible , locally bounded representations of @xmath6 by theorem [ the : locally_bounded_relative_hull ] .
assume that @xmath6 is countably generated .
then the usual topology on @xmath1776 is metrisable and hence compactly generated , so that @xmath926 is the standard topology on @xmath1776 . a representation of @xmath6 is locally bounded if and only if all symmetric elements of @xmath6 act by regular , self - adjoint operators by theorem [ the : locally_bounded_versus_regular ] .
thus a representation @xmath16 of @xmath13 is integrable as in theorem [ the : ae_locally_compact_hull ] if and only if @xmath25 is regular and self - adjoint for all @xmath26 with @xmath27 .
this class of integrable representations has the c*-hull @xmath2006 if @xmath1660 is locally compact . in particular ,
if @xmath6 is finitely generated , then @xmath1776 is mapped homeomorphically onto a closed subset of @xmath1077 for some @xmath439 by evaluating characters on a finite set of symmetric generators .
thus @xmath1776 is locally compact .
the discussion above gives : [ cor : ae_fin_gen_hull ] assume that @xmath6 is finitely generated . call a representation of @xmath13 integrable if its restriction to @xmath6 is locally bounded .
then @xmath1660 is locally compact and @xmath2006 for the twisted groupoid @xmath1847 constructed above is a c*-hull for the integrable representations of @xmath13 .
moreover , a representation @xmath16 of @xmath13 is integrable if and only if @xmath25 is regular and self - adjoint for all @xmath26 with @xmath27 .
corollary [ cor : ae_fin_gen_hull ] covers all the examples considered in @xcite , except for the enveloping algebra @xmath2007 of the virasoro algebra that is studied in @xcite*9.3 . the * -algebra @xmath2007 is @xmath59-graded .
its unit fibre @xmath2008 is noncommutative .
the first step in the study of its representations in @xcite*9.3 is to replace @xmath2007 by a certain @xmath59-graded quotient @xmath2009 , whose unit fibre @xmath2010 is commutative by construction .
the motivation is that all `` integrable '' representations of @xmath2007 factor through @xmath13 .
the main result in @xcite*9.3 shows that the partial action of @xmath59 on @xmath1660 is free and that the disjoint union @xmath2011 of the three families of characters described in ( 61)(63 ) of @xcite is a fundamental domain , that is , it meets each orbit of the partial action exactly once .
each subset @xmath2012 is closed in @xmath1776 and locally compact and second countable in the subspace topology .
hence so is @xmath1183 .
since @xmath59 acts by partial homeomorphisms and @xmath1183 is a fundamental domain , there is a continuous bijection @xmath2013 each @xmath2014 is an open subset of @xmath1183 , so that @xmath197 is locally compact .
i have not checked whether this continuous bijection is a homeomorphism .
if so , then @xmath1660 would be locally compact and the results in @xcite for the virasoro algebra would be contained in theorem [ the : ae_locally_compact_hull ] after passing to the quotient @xmath2015 . if not , we would use the locally compact space @xmath197 .
the partial action of @xmath59 on @xmath1660 is clearly continuous on @xmath197 as well , so that theorem [ the : twisted_groupoid ] applies .
given a normalised @xmath61-cocycle on @xmath0 , rieffel deformation is a deformation functor that modifies the multiplication on a @xmath0-graded * -algebra by the @xmath61-cocycle .
there is a similar process for fell bundles over @xmath0 , which we may transfer to section c*-algebras .
this is how rieffel deformation is usually considered , but the setting of graded algebras or fell bundles is actually easier .
we now define rieffel deformation more precisely and show that it is compatible with the construction of c*-hulls in theorem [ the : fell_bundle_sections_hull ] .
a _ normalised @xmath61-cocycle _ on a group @xmath0 is a function @xmath2016 with @xmath2017 for all @xmath54 and @xmath2018 for all @xmath2019 .
let @xmath2020 be a @xmath0-graded algebra .
let @xmath2021 be the same @xmath0-graded vector space with the deformed multiplication and involution @xmath2022 where @xmath2023 for all @xmath54 .
we call @xmath2021 the _ rieffel deformation _ of @xmath13 with respect to @xmath2024 .
[ lem : rieffel_def_works ] the deformed multiplication and involution on @xmath2021 give a @xmath0-graded * -algebra with @xmath2025 if @xmath26 or @xmath1704 , and @xmath2026 for all @xmath54 , @xmath1804 .
the multiplication remains associative by the @xmath61-cocycle condition .
the normalisation of @xmath2024 and for @xmath2027 give @xmath2028 for all @xmath54 .
thus @xmath2029 for @xmath1333 .
the normalisation condition and for @xmath2030 and @xmath2031 for @xmath1329 give @xmath2032 hence @xmath2033 .
this implies the condition @xmath2034 for @xmath1333 , @xmath2035 : @xmath2036 thus the deformed multiplication and involution give a * -algebra . the formula @xmath2026 for @xmath1536 , @xmath1804 is trivial , and @xmath2025 if @xmath26 or @xmath1704 follows from the normalisation of @xmath2024 .
the same formulas work if @xmath7 is a fell bundle over @xmath0 .
let @xmath2037 be the same banach space bundle as @xmath1451 with the multiplication and involution @xmath2038 and @xmath2039 for @xmath1329 , @xmath2040 , @xmath2041 .
by lemma [ lem : rieffel_def_works ] , the deformation does not change @xmath2042 for @xmath2043 or @xmath1462 and @xmath2044 and @xmath2045 for @xmath2046 .
hence @xmath2047 as hilbert @xmath11-bimodules , so that the positivity and completeness conditions for a fell bundle are not affected by the deformation .
we call @xmath2048 the _ rieffel deformation _ of the fell bundle @xmath7 with respect to @xmath2024 . for a c*-algebra of the form @xmath2049 for a fell bundle @xmath7 over @xmath0
, we define its rieffel deformation with respect to @xmath2024 as @xmath2050 for the deformed fell bundle . if @xmath0 is an abelian group , then @xmath2051 for a fell bundle over @xmath0 carries a canonical continuous action of @xmath2052 , called the dual action .
conversely , any c*-algebra with a continuous @xmath2052-action @xmath1560 is of the form @xmath2053 , where @xmath7 is the spectral decomposition of the action , @xmath2054 thus rieffel deformation takes a c*-algebra with a continuous @xmath2052-action to another c*-algebra with a continuous @xmath2052-action .
this is how it is usually formulated .
since @xmath2052 is compact , there are no analytic difficulties with oscillatory integrals as in @xcite . [ the : rieffel_deformation ] let @xmath1775 be a @xmath0-graded * -algebra and let @xmath11 be a c*-hull for a class of integrable representations of @xmath6 .
assume that integrability is compatible with induction for @xmath13 .
let @xmath2024 be a normalised @xmath61-cocycle on @xmath0 .
then integrability is also compatible with induction for @xmath2021 , and the c*-hull for the integrable representations of @xmath2021 is the rieffel deformation with respect to @xmath2024 of the c*-hull for the integrable representations of @xmath13 . the compatibility condition in definition [ def : induced_actions_integrable ] is equivalent to the integrability of @xmath2055 for all @xmath1536 , which only involves a single @xmath1487 with its @xmath6-bimodule structure and the @xmath6-valued inner product @xmath2056 for @xmath1804 .
this is not changed by rieffel deformation by lemma [ lem : rieffel_def_works ] .
hence @xmath2021 inherits the compatibility condition from @xmath13 , and theorem [ the : fell_bundle_sections_hull ] applies to both @xmath13 and @xmath2021 .
the hilbert @xmath11-bimodule @xmath1451 depends only on @xmath1487 with the extra structure above and the universal inducible , integrable representation @xmath1392 of @xmath6 by remark [ rem : fell_explicit ] . since none of this
is changed by rieffel deformation , the fell bundle obtained from @xmath2021 has the same fibres @xmath2057 as @xmath1451 .
rieffel deformation changes the multiplication maps @xmath2058 and the involution @xmath1735 for fixed @xmath1329 only by a scalar . inspecting the construction above
, we see that the multiplication maps @xmath2059 and the involution @xmath2060 in the fell bundle are changed by exactly the same scalars .
hence the fell bundle for @xmath2021 is @xmath2061 .
now the assertion follows from theorem [ the : fell_bundle_sections_hull ] .
we illustrate our theory by studying c*-hulls of twisted @xmath442-dimensional weyl algebras for @xmath2062 .
we begin with the case @xmath2063 , where no twists occur .
then we consider the case of finite @xmath442 without twists and with twists .
finally , we consider the case @xmath2064 with and without twists . the ( @xmath314-dimensional ) _ weyl algebra _
@xmath13 is the universal * -algebra with one generator @xmath98 and the relation @xmath2065 .
there is a unique @xmath59-grading @xmath2066 with @xmath2067 .
the * -subalgebra @xmath1604 is isomorphic to the polynomial algebra @xmath2068 $ ] with @xmath2069 , which is commutative .
the other subspaces @xmath2070 for @xmath554 are isomorphic to @xmath1604 as left or right @xmath1604-modules because @xmath2071 and @xmath2072 for all @xmath2073 .
the spectrum @xmath2074 of @xmath1604 is @xmath111 , where the character for @xmath444 evaluates a polynomial in @xmath711 at @xmath86 .
a character is positive if and only if it is positive on @xmath2075 and @xmath2076 for all @xmath2077 .
this happens if and only if @xmath2078 by @xcite*example 10 . since @xmath2079 , the partial automorphism @xmath2080 of @xmath2081 associated to the @xmath1604-bimodule @xmath2082 acts on @xmath1660 by the automorphism @xmath2083 , which corresponds to translation by @xmath2084 . by induction
, we get @xmath2085 . the domain of @xmath2086 is as big as it could possibly be , that is , it contains all @xmath439 with @xmath2087 by ( see also @xcite*example 16 ) . for any @xmath1273 there is a unique @xmath1946 with @xmath2088 .
thus the transformation groupoid @xmath2089 is simply the pair groupoid on @xmath881 .
there can be no twist in this case .
first , the pair groupoid simply has no non - trivial twists . and
secondly , the generators @xmath2090 for @xmath2073 satisfy the positivity condition in proposition [ pro : commutative_partial_action_no_twist ] , which also rules out a twist .
since no proper non - empty subset of @xmath881 is invariant under the partial action @xmath2091 of @xmath59 , a commutative c*-hull for @xmath1604 for which integrability is compatible with induction gives either @xmath2092 or @xmath2093 . in the second case , @xmath13 has no non - zero integrable representations . in the first case , the c*-hull for the integrable representations of @xmath13 is the groupoid c*-algebra @xmath829 of the pair groupoid @xmath2094 .
the universal representation of @xmath13 on @xmath829 is equivalent to a representation @xmath16 of @xmath13 on @xmath808 by proposition [ pro : representation_hilm_comp ] .
the domain of this representation is the space @xmath2095 of rapidly decreasing sequences , with @xmath2096 for @xmath554 , so @xmath2097 , @xmath2098 . by theorem [ the : regular_self - adjoint_cstar - hull ] ,
a representation @xmath16 of @xmath1604 on a hilbert module @xmath15 is integrable if and only @xmath2099 is regular and self - adjoint for each @xmath554 or , equivalently , @xmath2100 is regular and self - adjoint and @xmath2101 for all @xmath554 . by definition ,
a representation of @xmath13 is integrable if and only if its restriction to @xmath1604 is integrable .
the @xmath59-grading on the c*-hull @xmath829 is `` inner '' : it is induced by the @xmath59-grading on @xmath808 where @xmath2102 has degree @xmath1281 .
equivalently , the dual action of @xmath1848 on @xmath829 associated to the @xmath59-grading is the inner action associated to the unitary representation @xmath2103 , where @xmath2104 for all @xmath2105 , @xmath554 .
now let @xmath1325 and let @xmath2106 be an antisymmetric @xmath2107-matrix .
let @xmath2108 .
let @xmath2109 be the * -algebra with generators @xmath2110 and the commutation relations @xmath2111 for @xmath204 , and @xmath2112 for @xmath2113 .
since @xmath2114 , the relations for @xmath2115 and @xmath2116 are equivalent ; so it suffices to require for @xmath2117 .
the * -algebra @xmath2109 is @xmath2118-graded by giving @xmath1271 degree @xmath2119 , where @xmath2120 is the standard basis of @xmath2118 .
we first consider the case @xmath2121 and write @xmath2122 .
this is the _
@xmath2123-dimensional weyl algebra _ , which is the tensor product of @xmath2123 copies of the @xmath314-dimensional weyl algebra , with the induced @xmath2118-grading .
thus the zero fibre @xmath2124 for @xmath2125 is isomorphic to the polynomial algebra @xmath2126 $ ] in the @xmath2123 generators @xmath2127 .
its spectrum is @xmath2128 .
each @xmath2129 for @xmath2130 is isomorphic to @xmath2124 both as a left and a right @xmath2124-module ; the generator is the product of @xmath2131 for @xmath2132 or @xmath2133 for @xmath2134 from @xmath2135 . here
the order of the factors does not matter because @xmath2121 .
we may identify @xmath2129 with the exterior tensor product of the @xmath2136-bimodules @xmath2137 .
hence the space of positive characters on @xmath2138 is @xmath2139 , and the partial action of @xmath2118 on @xmath2139 is the exterior product of the partial actions of @xmath59 on @xmath881 for the @xmath314-dimensional weyl algebras .
that is , @xmath2130 acts on @xmath2139 by translation by @xmath2140 with the maximal possible domain .
thus the transformation groupoid @xmath2141 is isomorphic to the pair groupoid of the discrete set @xmath2139 .
once again , the only @xmath2118-invariant subsets of @xmath2142 are the empty set and @xmath2139 , so that the only commutative c*-hulls of @xmath1604 for which integrability is compatible with induction are @xmath988 and @xmath2143 .
the first case is boring , and the second case leads to the c*-hull @xmath2144 of the @xmath2123-dimensional weyl algebra . as for @xmath2145 , the universal representation of @xmath2138
is equivalent to a representation on @xmath2146 .
this has the domain @xmath2147 , and the representation is determined by @xmath2148 for @xmath2149 and @xmath2135 .
hence @xmath2150 .
a representation of @xmath13 is integrable if and only if its restriction to @xmath1604 is integrable in the sense that it integrates to a representation of @xmath1324 .
this automatically descends to a representation of @xmath2143 by lemma [ lem : induced_repr_inducible ] .
there are several ways to characterise when a representation of @xmath2126 $ ] integrates to a representation of @xmath1324 .
one is that @xmath2151 for @xmath2135 are strongly commuting , regular , self - adjoint operators and @xmath2152 for all @xmath2135 , @xmath554 , compare @xcite*theorem 9.1.13 .
the group @xmath2118 has non - trivial @xmath61-cocycles , and @xmath2109 is , by definition , a rieffel deformation of @xmath2153 for the normalised @xmath61-cocycle @xmath2154 we could also use the cohomologous antisymmetric @xmath61-cocycle @xmath2155 theorem [ the : rieffel_deformation ] says that the c*-hull @xmath2156 of @xmath2109 is the rieffel deformation of the c*-hull @xmath2157 of @xmath2153 for the same @xmath61-cocycle . in the classification of fell bundles with commutative unit fibre ,
the important cohomology is that of the transformation groupoid @xmath1913 , not of @xmath0 itself . here
@xmath1913 is the pair groupoid of @xmath2139 .
the set of composable @xmath442-tuples in the pair groupoid of @xmath197 is @xmath2159 .
the groupoid cohomology with coefficients @xmath2158 is the cohomology of the chain complex with cochains @xmath2160 , the space of all maps @xmath2161 , and with the boundary map @xmath2162 , @xmath2163 here the hat means that the entry @xmath307 is deleted .
pick some point @xmath2164 and let @xmath2165 for all @xmath439 , @xmath2166 , @xmath2167 .
then @xmath2168 for all @xmath2169 , @xmath2170 .
thus the cohomology vanishes in positive degrees .
any twist of the pair groupoid on @xmath2139 is trivial by lemma [ lem : cohomology_pair_groupoid ] .
therefore , the c*-hull @xmath2156 is isomorphic to @xmath2144 , the untwisted groupoid c*-algebra of the pair groupoid .
there is another proof of this isomorphism , which we pursue so far as to describe the isomorphism between the two c*-hulls .
since the @xmath59-grading on the c*-hull @xmath829 is inner or , equivalently , the corresponding action of @xmath1848 is inner , the same holds for the @xmath2118-grading on the c*-hull @xmath2144 and the corresponding @xmath2171-action on @xmath2144 .
explicitly , the @xmath2171-action is induced by the unitary representation of @xmath2171 on @xmath2146 defined by @xmath2172 rieffel deformation of c*-algebras for _ inner _ actions does not change the isomorphism type of the c*-algebra . hence the c*-hull for the integrable representations of @xmath2109 is also isomorphic to @xmath2144 .
we make this more explicit in our fell bundle language .
let @xmath2173 be a strictly continuous homomorphism to the group of unitary multipliers of a c*-algebra @xmath8 and let @xmath2174 for @xmath2175 , @xmath375 be the resulting inner action .
let @xmath2176 be the spectral decomposition of this action , that is , @xmath2177 if and only if @xmath2178 for all @xmath2175 . in particular , @xmath2179 because @xmath2171 is commutative . assume for simplicity that the @xmath61-cocycle @xmath2024 is a bicharacter as above . for fixed @xmath2130
, we view @xmath2180 as an element @xmath2181 of the dual group @xmath2171 .
the map @xmath2182 is a group homomorphism .
the maps @xmath2183 , @xmath2184 , for @xmath554 are banach space isomorphisms that modify the multiplication as follows : @xmath2185 they keep the involution unchanged .
this is exactly what rieffel deformation does .
hence the maps @xmath2186 form an isomorphism between the undeformed and rieffel deformed fell bundles .
this finishes the proof that the rieffel deformed algebra for an inner action is canonically isomorphic to the original algebra .
the universal representation of @xmath2109 on @xmath2144 again corresponds to a representation of @xmath2109 on @xmath2187 .
we may construct it by carrying over the isomorphism @xmath2188 between the c*-hulls .
this is the inverse of the isomorphism above , so it multiplies on the left by the unitary @xmath2189 of degree @xmath530 on elements of degree @xmath1281 .
we do exactly the same on elements of @xmath2109 and so let @xmath2190 for @xmath2130 act on @xmath2146 by the operator @xmath2191 , where @xmath2192 is the universal representation of the untwisted weyl algebra @xmath2153 on @xmath2146 described above .
the same computation as above shows that this defines a * -representation of @xmath2109 .
we compute it explicitly .
first , the action of elements of @xmath2193 on @xmath2146 is not changed .
the domain of a representation of @xmath2109 is equal to the domain of its restriction to @xmath2193 .
hence the domain of our representation is the schwartz space @xmath2147 , as in the untwisted case .
let @xmath2194 .
the generator @xmath1271 has degree @xmath2119 , and @xmath2195 for our first definition of @xmath2024 in .
thus @xmath2196 these operators on @xmath2147 satisfy the defining relations of @xmath2109 .
now we consider the weyl algebra @xmath2197 in infinitely many generators @xmath2198 .
several `` host algebras '' for this * -algebra are constructed by grundling and neeb @xcite and assembled into a `` full host algebra , '' that is , a c*-algebra whose hilbert space representations are equivalent to the `` regular '' representations of @xmath2197 .
the smaller host algebras that are associated in @xcite to a sequence of integers are c*-hulls as well .
but there is no `` full '' c*-hull in our setting .
the _ infinite - dimensional weyl algebra @xmath2197 _ is the universal * -algebra with generators @xmath1271 for @xmath2199 and relations @xmath2200 , @xmath2201 , and @xmath2202 for @xmath2203 .
let @xmath2204 $ ] be the free abelian group on countably many generators .
the weyl algebra @xmath2197 is @xmath2204$]-graded , where @xmath1271 has degree @xmath2205 $ ] , the @xmath765th generator of @xmath2204 $ ] .
the * -algebra @xmath2197 is a tensor product of infinitely many @xmath314-dimensional weyl algebras .
the zero fibre @xmath2206 is the polynomial algebra in the generators @xmath2127 for @xmath2199 .
hence its spectrum is the infinite product @xmath2207 , which is not locally compact .
the tensor product structure of @xmath2197 shows that a character is positive if and only if each component is .
that is , @xmath2208 is a product of countably many copies of the discrete space @xmath881 . since @xmath881 is not compact , this is not locally compact either .
hence to build a commutative c*-hull for @xmath2206 , we must choose some locally compact space @xmath197 with a continuous , injective map @xmath2209 . here
we simplify notation by assuming that already @xmath2210 ; otherwise , the first step in our construction would replace @xmath197 by @xmath2211 .
for @xmath197-integrability to be compatible with induction , we also need @xmath2212 to be invariant under the partial action of @xmath2204 $ ] , and we need the restricted partial action to lift to a continuous partial action on @xmath197 .
the partial action of the group @xmath2204 $ ] on @xmath2213 is the obvious one by translations .
this is a free partial action , and two points @xmath2214 and @xmath2215 in @xmath2213 belong to the same orbit if and only if there is @xmath2216 such that @xmath2217 for all @xmath2218 . briefly ,
such points are called _
tail equivalent_. this partial action is minimal , that is , an open , @xmath2204$]-invariant subset is either empty or the whole space .
hence @xmath2213 has no @xmath2204$]-invariant , locally closed subsets .
thus @xmath2213 has no @xmath2204$]-invariant , locally compact subspaces .
let @xmath762 be any compact subset of @xmath2213 .
then the projection @xmath2219 to the @xmath765th coordinate must map @xmath762 to a compact subset of @xmath881 .
so there is an upper bound @xmath2220 with @xmath2221_{\mathbb n}{\mathrel{\vcentcolon=}}[0,m_j]\cap { \mathbb n}$ ]
. then @xmath2222_{\mathbb n}$ ] , and the right hand side is compact .
the closure of @xmath2223_{\mathbb n}$ ] under tail equivalence is @xmath2224_{\mathbb n}\biggr),\ ] ] the restricted product of copies of @xmath881 with respect to the compact - open subsets @xmath2225_{\mathbb n}$ ] .
there is a unique topology on @xmath2226 where each subset @xmath2227_{\mathbb n}$ ] is open and carries the product topology .
this topology is locally compact , and the partial action of @xmath2204 $ ] on @xmath2226 by translation is continuous . since the map @xmath2228 is not a homeomorphism onto its image and @xmath2229 is metrisable , the local
global principle fails for the @xmath197-integrable representations of @xmath2206 by theorem [ the : commutative_local - global ] . applying induction from @xmath2206 to @xmath2197 to a counterexample for the local
global principle for @xmath2206 produces such a counterexample also for @xmath2197 . explicitly , choose a sequence @xmath2214 such that @xmath2230 for infinitely many @xmath1281
let @xmath2231 , that is , @xmath2232 has only one non - zero entry , which is @xmath2233 in the @xmath1281th place .
this sequence belongs to @xmath2226 and converges to @xmath530 in the product topology on @xmath2229 , but not in the topology of @xmath2226 .
the resulting representation of @xmath2206 on @xmath1207 is not @xmath2226-integrable , but it becomes integrable when we tensor with any hilbert space representation of @xmath1207 , see the proof of theorem [ the : commutative_local - global ] .
now induce this ( inducible ) representation of @xmath2206 to a representation of @xmath2197 on @xmath1207 .
this gives a counterexample for the local
global principle for @xmath2197 .
let @xmath135 be the set of all words in the letters @xmath2234 . if @xmath2235 is a positive character and @xmath2236 for some @xmath2237 $ ] , then @xmath2238 .
hence proposition [ pro : commutative_partial_action_no_twist ] shows that there is no twist in our case , that is , the c*-hull of the @xmath2226-integrable representations of @xmath2197 is the groupoid c*-algebra of the transformation groupoid @xmath2204\ltimes x_{(m_k)}$ ] .
this c*-hull is canonically isomorphic to one of the host algebras for @xmath2197 constructed in @xcite , namely , to the one that is denoted @xmath2239 $ ] in @xcite with @xmath2240 for all @xmath554 .
the compact subset @xmath2241_{\mathbb n}$ ] that we started with is a complete transversal in @xmath2204\ltimes x_{(m_k)}$ ] , that is , the range map in @xmath2204\ltimes x_{(m_k)}$ ] restricted to @xmath2242 is an open surjection onto @xmath2226 .
hence the groupoid @xmath2204\ltimes x_{(m_k)}$ ] is morita equivalent to its restriction to the compact subset @xmath41 .
this restriction is the tail equivalence relation on @xmath41 .
its groupoid c*-algebra is well known : it is the uhf - algebra for @xmath2243 , that is , the infinite tensor product of the matrix algebras @xmath2244 .
the c*-algebra of @xmath2204\ltimes x_{(m_k)}$ ] itself is the c*-stabilisation of this uhf - algebra .
thus the @xmath2226-integrable representations of @xmath2197 are equivalent to the representations of the ( stabilisation of the ) uhf - algebra of type @xmath2243 .
this depends very subtly on the choice of the sequence @xmath2245 .
there is no canonical * -homomorphism between these uhf - algebras if we increase @xmath2245 : for some @xmath2246 , there is not even a non - zero map between their k - theory groups
. instead , there are canonical morphisms , that is , there is a canonical nondegenerate * -homomorphism @xmath2247\ltimes x_{(m_k ' ) } ) \to { \texorpdfstring{\textup c^*}{c*}}({\mathbb z}[n]\ltimes x_{(m_k)})$ ] if @xmath2246 .
they are constructed as follows .
the inclusion map @xmath2248 is continuous with dense range , but not proper .
thus it induces an injective , nondegenerate * -homomorphism @xmath2249 .
therefore , if a representation of @xmath2250 is @xmath2226-integrable , then it is also @xmath2251-integrable .
if a representation of @xmath2197 has @xmath2226-integrable restriction to @xmath2206 , then its restriction to @xmath2206 is also @xmath2251-integrable . when we apply this to the universal representation of @xmath2197 on the @xmath2252-hull @xmath2247\ltimes x_{(m_k)})$ ] , this gives the desired canonical morphism @xmath2247\ltimes x_{(m_k ' ) } ) \to { \texorpdfstring{\textup c^*}{c*}}({\mathbb z}[n]\ltimes x_{(m_k)})$ ] if @xmath2246 .
it is injective because its restriction to the unit fibre @xmath2253 is injective , the abelian group @xmath2204 $ ] is amenable , and reduced crossed products preserve injectivity of morphisms .
now let @xmath2254 be a skew - symmetric matrix .
it corresponds first to a matrix @xmath2255 and then to a @xmath61-cocycle @xmath2024 on @xmath2204 $ ] as in .
the rieffel deformation of @xmath2197 by @xmath2256 is the universal * -algebra @xmath2257 with the same generators @xmath2258 and the relations @xmath2259 , @xmath2260 , and @xmath2261 for @xmath2203 .
we define @xmath2226 for a sequence @xmath2245 and the @xmath2226-integrable representations of @xmath2257 as above . by theorem [ the : rieffel_deformation ] , this has a c*-hull , namely , the rieffel deformation of the c*-hull for the @xmath2226-integrable representations of the undeformed weyl algebra .
the rieffel deformation gives a twist of the groupoid @xmath2204\ltimes x_{(m_k)}$ ] , and the c*-hull is the twisted c*-groupoid algebra of @xmath2204\ltimes x_{(m_k)}$ ] for this twist .
the c*-hull of @xmath2257 is a twisted groupoid algebra of the transformation groupoid @xmath2204\ltimes x_{(m_k)}$ ] , which is isomorphic to the tail equivalence relation @xmath2263 on @xmath2226 .
we are going to prove that any twist @xmath2264 is trivial .
hence the c*-hull of @xmath2257 is isomorphic to the untwisted groupoid c*-algebra of @xmath2265\ltimes x_{(m_k)}$ ] for all @xmath2256 .
the arrow space of @xmath2263 is totally disconnected because @xmath2226 is totally disconnected and @xmath2263 is tale .
hence any locally trivial principal bundle over @xmath2263 is trivial .
thus @xmath2266 as a topological space , and the twist is described by a continuous @xmath61-cocycle @xmath2267 .
we must show that @xmath387 is a coboundary .
let @xmath2268 for @xmath2269 be the equivalence relation on @xmath2226 defined by @xmath2270 if and only if @xmath2271 for all @xmath2272 .
this is an increasing sequence of open subsets @xmath2273 with @xmath2274 , and each @xmath2268 is also an equivalence relation .
the equivalence relation @xmath2268 is isomorphic to the product of the pair groupoid on @xmath2275 and the space @xmath2276 for the shifted sequence @xmath2277 .
thus the cohomology of @xmath2268 with coefficients in @xmath1848 is isomorphic to the cohomology of the pair groupoid on @xmath2275 with values in the abelian group of continuous map @xmath2278 .
the cohomology of a pair groupoid on a discrete set vanishes in positive degrees by lemma [ lem : cohomology_pair_groupoid ] .
therefore , for each @xmath2269 there is @xmath2279 such that @xmath2280 is the coboundary @xmath2281 of some @xmath2282 .
the restriction of @xmath2283 to @xmath2268 and @xmath2284 both have coboundary @xmath2285 . hence @xmath2286 is a @xmath314-cocycle on @xmath2268 .
again by lemma [ lem : cohomology_pair_groupoid ] , there is @xmath2287 with @xmath2288 .
we replace @xmath2283 by @xmath2289 , where @xmath2290 means the coboundary of @xmath949 on the groupoid @xmath2291 .
this still satisfies @xmath2292 , and @xmath2293 .
proceeding like this , we get continuous maps @xmath2294 for all @xmath2269 that satisfy @xmath2295 and @xmath2296 for all @xmath2269 . these combine to a continuous map @xmath2297 with @xmath2298 . | we define a c*-hull for a * -algebra , given a notion of integrability for its representations on hilbert modules .
we establish a local global principle which , in many cases , characterises integrable representations on hilbert modules through the integrable representations on hilbert spaces .
the induction theorem constructs a c*-hull for a certain class of integrable representations of a graded * -algebra , given a c*-hull for its unit fibre . |
the moments of velocity differences over spatial scales of size @xmath3 , the so - called structure functions , provide useful measures of the statistical description of fluid turbulence ( kolmogorov 1941a , b ) .
in particular , the longitudinal structure functions defined as @xmath4^n \big \rangle \label{sn}\ ] ] have been studied extensively . here
, @xmath5 is the velocity vector at position @xmath6 , and @xmath7 is the unit vector along the separation vector @xmath8 .
the special interest in structure functions comes in part from an exact result , known as the 4/5-ths law , @xmath9 valid in the inertial range of scales ( @xmath10 where @xmath11 is the kolmogorov scale characterizing the dissipative scale of motion and @xmath12 is a suitable large scale of turbulence ) . in part
, the interest is spurred by the operational ease with which longitudinal structure functions can be obtained from experimental data if one makes the so - called taylor s hypothesis ( taylor 1935 ) .
the major impetus for measurements , however , is the scaling result of kolmogorov
k41 for brevity that , for high reynolds numbers , the structure functions follow the relation @xmath13 where the scaling exponent @xmath14 . as a result of considerable work ( see , for example ,
anselmet _ et al .
_ 1984 , maurer _ et al .
_ 1994 , arneodo _ et al .
_ 1996 , sreenivasan & antonia 1997 ) , it now appears nearly certain that the scaling exponents deviate from @xmath15 increasingly and nonlinearly as @xmath16 increases .
this is the anomalous scaling .
while some issues remain to be explained satisfactorily ( see , for example , sreenivasan & dhruva 1998 ) , it appears that anomalous scaling is a genuine result worthy of a serious theoretical effort .
consequently , considerable thrusts of research have occurred in this direction ( e.g. , lvov & procaccia 1996 ) .
one obvious concern about high - order moments is that , since they sample the tails of the probability distribution function ( pdf ) of velocity increments and since some of the associated rare events may be related to well - defined flow structures in real space
it is not entirely certain that the results for high - order moments are universal because the mean shear and other non - universal features driving the flow could influence the rare structures of turbulence .
in contrast , low - order moments are less susceptible to such non - universal aspects , because they are determined nearly entirely by the core of the pdf and hence sampled frequently .
thus , it is reasonable to regard anomalous scaling and its universality as more conclusively established if low - order moments also display anomaly .
this is the main subject of this paper .
the lowest non - trivial structure function that has been studied extensively is the second - order , whose scaling exponent has been shown to be @xmath17 . though this is measurably different from the predicted value of @xmath18 , the difference is too small to be conclusive on its own .
it would thus be useful to examine scaling exponents for moments of still lower orders . since moments of order @xmath0 and below do not exist for velocity increments ( @xcite ) , the range of our current interest is limited to @xmath19 , where @xmath16 is necessarily fractional . with decreasing @xmath16 in this range , if the deviation from @xmath15 decreases , we shall at least know that k41 will be exact in the limit of low - order moments , and regard it as a possible reference point for a theory . if , on the other hand , these deviations remain to be non - trivial as @xmath16 vanishes , it may well be that the understanding of the anomaly can be sought more fruitfully in terms of low - order moments , for the simple reason that such a theory can justifiably ignore rare events .
some preliminary measurements were published in sreenivasan _
( 1996 ) and cao _ et al . _
( 1996 ) but the present paper is a more complete account of the data .
perhaps more importantly , the preliminary numbers did not take account of the possible effects of residual anisotropy in both experiments and simulations , an issue whose importance has been recently highlighted ( biferale & procaccia 2004 ) .
we take account of this feature using a recently developed angle - averaging technique ( taylor _ et al .
_ 2003 ) .
the rest of the paper is organized as follows . in sec . 2
, we describe the experimental and numerical data used for the present analysis .
this is followed in sec . 3 by the calculation of the fractional structure functions and their scaling exponents from the datasets described in sec . 2 .
the exponents from all sources of data agree well with each other and deviate measurably from the k41 prediction even as @xmath20 .
section 4 contains a brief discussion of the significance of the results .
hotwire measurements were made in the atmospheric surface layer at a height of 35 m above the ground using a standard meteorological tower at brookhaven national laboratory .
the tower itself presented very little obstacle to the wind because of its low solidity .
the dataset analyzed here is part of a more comprehensive batch obtained at the tower .
the hotwire , 0.7 mm in length and 0.5 @xmath21 m in diameter , was placed facing the wind , about two meters away from the tower .
( for monitoring the wind direction , the tower was equipped with a vane anemometer placed two meters away from the measurement station . ) the calibration was performed in situ using a tsi calibrator and checked later in a windtunnel .
the signals were low - pass filtered at 5 khz and sampled at 10 khz .
the anemometer and signal conditioners were placed nearby at the height of measurement , and the conditioned signal was transmitted to the ground and digitized using a 12-bit a / d converter .
typical data records contained between 10 and 40 million samples , during which time the wind direction and its mean speed were deemed acceptably constant .
more details are given in @xcite , but the essential features for this particular set of data are listed in table [ expt_params ] .
the wind conditions were somewhat unstable .
.some relevant parameters for the atmospheric data . here
, @xmath22 is the mean speed , u@xmath23 is the root - mean - square velocity , @xmath24 is the mean rate of energy dissipation , @xmath11 and @xmath25 are the kolmogorov and taylor microscales , respectively , and @xmath26 , @xmath27 being the kinematic viscosity of air at the measurement temperature . [ cols="^,^,^,^,^,^ " , ] the finite reynolds number of turbulence in numerical work shortens the inertial range over which scaling exponents may be discerned with clarity .
this remains a constraint because the applicable theory concerns the limit of re @xmath28 .
a constraint in experimental data is that some large - scale anisotropy might be present in the range that appears to scale .
the effect could be present even in high - reynolds - number flows because the effects of anisotropic forcing penetrate the scaling range in a subtle but systematic way ; see , for example , @xcite .
in numerical simulations , the statistics are usually calculated with the separation direction @xmath29 oriented parallel to a box - side .
if there is residual anisotropy ( angular - dependence ) in the small - scales , this procedure will bias the results .
since we are concerned here with delicate results , this uncertainty has to be eliminated satisfactorily .
we make use of two recent developments to properly eliminate the effect of anisotropy in the inertial range of the @xmath30 data .
first , we now know that isotropic and anisotropic contributions can be isolated systematically by projecting structure function of a given order over a particular basis function in its so(3 ) group decomposition ( e.g. , arad _ et al .
_ 1998 , 1999 , kurien _ et al .
_ 2000 , biferale & procaccia 2004 ) .
this is a useful step to take even in nominally isotropic turbulence because the effect of forcing might persist in the nominal scaling range .
second , taylor _ et al . _
( 2003 ) developed a method by which the third - order longitudinal structure function was computed in @xmath31 directions of the flow for a given @xmath3 and the results averaged .
the angle - averaged value of the structure function achieved the kolmogorov 4/5 prediction to a remarkable degree , thus providing a convincing scheme for extracting the isotropic component of a mildly anisotropic flow .
the two procedures achieve the same goal : the angle - averaging procedure is in effect a projection on to the isotropic component of the so(3 ) rotation group .
since this method is independent of the order of the structure function , we use it as described below for fractional statistics . the angle - averaged isotropic structure function in the computational domain @xmath32
is given by @xmath33 where the usual average for a particular direction of the unit separation vector @xmath29 is followed by a spherical average of all possible orientations of @xmath29 over the solid angle @xmath34 .
the long - time average is taken in the steady state . in units of the large - eddy turnover time
, @xmath35 is 1.5 from the start of the simulation and @xmath36 is unity . in numerical simulations , since we have the full three - dimensional velocity field , we can in principle integrate over the sphere and project out the isotropic part of @xmath37 .
_ showed that the full spherical average may be approximated to arbitrary precision by first computing the structure function over sufficiently many different directions in the flow , interpolating each of these functions by a simple single - variable cubic spline , and then averaging the interpolated values over all directions .
this angle - averaging of the structure functions is much faster to implement than interpolating the three - dimensional , three - component velocity data over spherical shells in order to perform the integration .
it is also more accurate in the small scales than other spherical averaging schemes .
the angle - averaging procedure was performed for all orders .
for very low - order moments @xmath38 we observed only marginal anisotropy in the inertial range and the statistics computed in the different directions nearly coincided with each other except for large scales @xmath39 . in figure [ fract_localslp ]
we show the logarithmic derivative ( the local slope ) of the structure functions for various fractional orders .
if the local slope is constant over a range of @xmath3 , it provides the scaling exponent @xmath40 .
a rough estimate of the inertial range from figure [ fract_localslp ] is @xmath41 and the error estimate on the value scaling exponent is calculated as the variance over this range . as expected , the smaller the absolute order @xmath42 , the more constant is the local slope , thus indicating that our confidence level improves as the moment order decreases .
the scaling exponents and their uncertainty over the inertial range are given for a range of fractional orders in table [ exponents ] , column 6 .
the exponents calculated for all three datasets display a good degree of agreement .
the scaling exponents computed in this way from the @xmath30 simulation are plotted as a function of order @xmath16 in figure [ fract_exps ] .
for comparison , the k41 exponents are also shown .
the measured exponents deviate from the k41 values for all orders .
this shows that the intermittency , until now thought to be characteristic of only high - order moments , sampling ` fat ' tails of the pdf of velocity increments , persists into the low - order moments sampling the core of the pdf .
figure [ d - fract_exps ] shows the difference @xmath43 from the k41 prediction as a function of @xmath16 . there is a linear approach to zero with a slope of 0.056 .
equivalently , the behavior near zero is @xmath44 instead of @xmath45 .
figure [ relative_exps ] shows the relative ( percent ) departure of the measured scaling exponents from the self - similarity prediction of k41 , @xmath46 , is a smooth function of @xmath16 in the range @xmath47 ( figure [ relative_exps ] ) .
we have also presented the anomalous exponents calculated from the multifractal @xmath48-model ( meneveau & sreenivasan 1988 ) for comparison .
the dependence on @xmath16 in the range @xmath49 is more or less linear and becomes weakly quadratic for @xmath50 .
the exact @xmath51law result of k41 with scaling exponent 1 , takes the relative difference to zero at @xmath52 . a noteworthy point in this curve
is at @xmath53 .
the relative difference from k41 is not defined for @xmath53 but the experimental and model values go smoothly through zero .
we interpret this to mean that anomalous scaling exists in the limit @xmath20 with relative departure from k41 of about @xmath54 ( see the @xmath55-intercept at @xmath53 in figure [ relative_exps ] ) .
the principal result of this paper is that the fractional low - order moments have scaling exponents that are different from @xmath15 . in this sense
, it is reasonable to consider that anomaly exists in very low - order and negative moments .
thus , instead of focusing entirely on high - order moments in search of an explanation for intermittency , it may also be reasonable to attempt to understand anomaly in low - order moments .
this has the advantage that rare events , whose universality may be in some doubt , play far less of a role .
we should remark on a lingering uncertainty . to keep structure functions real - valued for arbitrary orders , we can only consider , in the range @xmath19 , fractional moments of absolute valued velocity increments defined through equation ( [ sn_a ] ) .
the difference between the structure functions defined in ( [ sn ] ) and the corresponding ones defined for absolute - valued velocity increments is by no means clear for large @xmath42 .
our unpublished work appears to indicate that the absolute - valued structure functions have a larger scaling exponent than the classical ones when @xmath16 is large and odd .
it can not therefore be dismissed entirely that the departures from @xmath15 that one may observe for small and fractional @xmath16 may merely suggest the possibility that k41 does not somehow apply to absolute moments . even if this is our only conclusion , it is still new and thus of some interest .
arneodo , a.,baudet , c. , belin , f. , benzi , r. , castaing , b.,chabaud , b.,chavarria , r. , ciliberto , s. , camussi , r. , chilla , f. , dubrulle , b. , gagne , y. , hebral , b. , herweijer , j. , marchand , m. , maurer , j. , muzy , j.f . ,
naert , a. , noullez , a. , peinke , j. , roux , f. , tabeling , p. , van de water , w. & willaime , h. 1996 structure functions in turbulence , in various flow configurations , at reynolds number between 30 and 5000 , using extended self - similarity .
_ europhys .
* 34 * , 411 - 416 .
kolmogorov , a.n .
1941a the local structure of turbulence in incompressible viscous fluid for very large reynolds numbers . _
* 30*. reproduced in _ proc .
a. _ * 434 * 9 - 13 .
kurien , s. & sreenivasan , k.r .
2001 measures of anisotropy and the universal properties of turbulence _
new trends in turbulence _ les houches summer school 2000 proceedings , eds : lesieur , m . , yaglom , a. & david , f. , 55 - 111 .
sreenivasan , k.r . ,
vainshtein , s.i . ,
bhiladvala , r. , san gil , i. , chen , s. & cao , n. 1996 asymmetry of velocity increments in fully developed turbulence and the scaling of low - order moments .
lett . _ * 77 * , 8 - 11 . | it is now believed that the scaling exponents of moments of velocity increments are anomalous , or that the departures from kolmogorov s ( 1941 ) self - similar scaling increase nonlinearly with the increasing order of the moment .
this appears to be true whether one considers velocity increments themselves or their absolute values .
however , moments of order lower than 2 of the absolute values of velocity increments have not been investigated thoroughly for anomaly . here
, we discuss the importance of the scaling of non - integer moments of order between + 2 and @xmath0 , and obtain them from direct numerical simulations at moderate reynolds numbers ( taylor microscale reynolds numbers @xmath1 450 ) and experimental data at high reynolds numbers ( @xmath2 10,000 ) . the relative difference between the measured exponents and kolmogorov s prediction increases as the moment order decreases towards @xmath0 , thus showing that the anomaly that is manifest in high - order moments is present in low - order moments as well .
this conclusion provides a motivation for seeking a theory of anomalous scaling as the order of the moment vanishes .
such a theory does not have to consider rare events which may be affected by non - universal features such as shear and so may be regarded as advantageous to consider and develop . |
an important question in string phenomenology is how and when supersymmetry ( susy ) is broken .
a great deal of effort has been devoted to frameworks in which it is broken non - perturbatively in the supersymmetric effective field theory .
much less effort has been devoted to string theories that are non - supersymmetric by construction . on the face of it ,
the trade off for the second option , is that non - supersymmetric string models do not have the stability properties of supersymmetric ones .
however it can be argued that as long as the susy breaking is spontaneous and parametrically smaller than the string scale , the associated instability is under perturbative control @xcite .
there is then no genuine moral , or even practical , advantage to the former more traditional option , since nature is not supersymmetric . sooner or later , either route to the standard model ( sm ) will lead to runaway potentials for moduli that need to be stabilised .
indeed spontaneous breaking at the string level may even confer advantages in this respect , as discussed in ref.@xcite .
parametric control over susy breaking requires a generic method for passing from a non - superymmetric theory to a supersymmetric counterpart , under certain limiting conditions .
the method that was studied in ref.@xcite is interpolation via compactification to lower dimensions , with susy broken by the scherk - schwarz mechanism @xcite .
the two great advantages of interpolating models are that their compactification volumes can be tuned to make the cosmological constant arbitrarily small , and that some of them exhibit enhanced stability due to a one - loop cosmological constant that is exponentially suppressed with respect to the generic susy breaking scale @xcite .
they can be viewed as natural and phenomenologically interesting extensions of the original observation in refs.@xcite that the @xmath3 tachyon - free non - supersymmetric @xmath4 model interpolates to the heterotic @xmath5 model , via a scherk - schwarz compactification to @xmath6 .
this paper focusses on the properties of @xmath7-dimensional ( @xmath1 ) theories that interpolate between stable , supersymmetric @xmath0 tachyon - free models .
two main results are presented .
first , by deriving and studying the general form of the @xmath0 endpoint theories , in particular how their modular invariance projections are related to the scherk - schwarz deformation , we are able to generalise the latter beyond previous configurations , opening up new possibilities for model building .
second , we present a comprehensive survey of the physical properties that such theories possess ( chiefly their potentials and the possible restoration of susy at small radius ) .
the general details of the interpolations are as shown in figure [ fig : interpolation ] .
beginning with a supersymmetric @xmath0 theory generically referred to as @xmath8 , the theory is compactified to a non - supersymmetric @xmath1 theory @xmath9 by adapting the coordinate dependent compactification ( cdc ) technique first presented in refs.@xcite .
this is the string version of the scherk - schwarz mechanism , which spontaneously breaks susy in the @xmath1 theory with a gravitino mass of @xmath10 where @xmath11 is the largest radius carrying a scherk - schwarz twist .
( we will use `` cdc '' and `` scherk - schwarz '' interchangeably . ) as usual it is the gravitino mass that is the order parameter for susy breaking : it can be continuously dialled to zero at large radius where susy is restored and @xmath8 regained .
supersymmetric theory at infinite radius , @xmath12 and its supersymmetric and non - supersymmetric , interpolated duals , which are defined in the vanishing radius limit . whether or not susy is restored in the @xmath13 model
is determined by the structure of the scherk - schwarz action . _ _ ] one of the main properties that will be addressed is the nature of the theory as the radius of compactification is taken to zero .
this depends upon the precise details of the scherk - schwarz compactification , and indeed we will find that the presence or otherwise of susy at zero radius depend on the choice of basis vectors and structure constants defining the model .
it is possible that the @xmath1 theory interpolates to either a supersymmetric or a non - supersymmetric model ( @xmath14 or @xmath15 respectively ) .
models of the latter kind correspond to a @xmath0 theory in which susy is broken by discrete torsion @xcite .
we begin in [ sec : cdc_overview ] by reviewing the basic formalism for interpolation .
section [ sec : cdc_virasoro_modified_algebra ] then presents the construction of @xmath1 non - supersymmetric models as compactifications of @xmath0 supersymmetric ones .
the modification of the massless spectra in the decompactification and @xmath16 limits ( with the latter corresponding to the decompactification limit of a @xmath0 @xmath2-dual theory ) is analysed , in order to determine the nature of the theories at the small and large radii endpoints .
section [ sec : details of cosmological constant calculation ] discusses the technique for rendering the cosmological constant in an interpolating form , allowing it to be calculated across a regime of small and large radii .
the modification of the projection conditions and massless spectrum by the choice of basis vectors and structure constants is made explicit , and based on these observations , in particular how the cdc correlates with modified gso projections in the @xmath0 endpoint theories , [ sec : most - general ] then derives the general form of deformation within this framework , extending previous constructions .
this more general formulation may prove to be useful for future model building .
the conditions under which susy is preserved or broken at the endpoints of the interpolation are discussed in [ sec : susy restoration and energy scales ] .
particular focus is given to the constraints on the appearance of light gravitino winding modes in the zero radius limit .
it is found that models in which the cdc acts only on the space - time side , are inevitably supersymmetric at zero radius , while models within which the cdc vector is non - trivial on the gauge side as well yield a non - supersymmetric model in the same limit .
this analysis paves the way for a presentation in [ sec : results ] of explicit interpolations ( in terms of their cosmological constants ) in particular models that display various different behaviours : namely we find examples of interpolation between two supersymmetric @xmath0 theories via @xmath1 theories with negative or positive cosmological constant ; interpolation between a non - supersymmetric @xmath0 theory and a supersymmetric one , with or without an intermediate @xmath1 ads minima ; we also find examples of `` metastable '' non - supersymmetric @xmath0 theories ( by which we mean theories that have a positive cosmological constant with an energy barrier ) that can decay to supersymmetric ones .
as mentioned , this paper follows on from a reasonably large body of work on non - supersymmetric strings that is nonetheless much smaller than the work on supersymmetric theories .
following on from the original studies of the ten - dimensional @xmath4 heterotic string @xcite , there were further studies of the one - loop cosmological constants @xcite , their finiteness properties @xcite , their relations to strong / weak coupling duality symmetries @xcite , and string landscape ideas @xcite .
the relationship to finite temperature strings was explored in refs .
further development of the scherk - schwarz mechanism in the string context was made in refs.@xcite .
progress towards phenomenology within this class has been made in refs .
related aspects concerning solutions to the large - volume `` decompactification problem '' were discussed in refs.@xcite .
non - supersymmetric string models have also been explored in a wide variety of other configurations @xcite , including studies of the relations between scales in various schemes @xcite .
some aspects of this study are particularly relevant to the recent work in refs.@xcite .
note that here we will not elaborate on the properties of the non - supersymmetric @xmath1 theory at radii of order the string length .
as we will see , and as found in ref.@xcite , often there is a minimum in the cosmological constant at this point which suggests some kind of enhancement of symmetry at a special radius .
( indeed often it is possible to identify gauge boson winding modes that become massless at the minimum . )
there is therefore the possibility of establishing connections to yet more non - supersymmetric @xmath1 theories .
conversely one can ask if _ every _ non - supersymmetric tachyon - free @xmath1 theory can be interpolated to a supersymmetric higher dimensional theory .
we comment on this and other prospects in the conclusions in [ sec : conclusions ] .
in this section , we revisit the calculation of the cosmological constant in the scherk - schwarzed theories , and in particular derive a formulation for the partition function of interpolating models , that is useful for the later analysis .
the discussion is a natural generalisation of the `` compactification - on - a - circle '' treatment of ref.@xcite , and as we shall see it ultimately leads to an improved and more general construction for this class of theory .
let us begin by briefly summarising the implementation of the scherk - schwarz mechanism described in that work .
as already mentioned , this is incorporated using a coordinate dependent compactification ( cdc ) @xcite of an initially supersymmetric @xmath0 theory , namely the @xmath17 model . for our purposes
it is useful to define it in the fermionic formulation , although any construction method would be applicable . in this formulation ,
the initial theory is defined by assigning boundary conditions to worldsheet fermions .
if they are real , this is encapsulated in a set of 28-dimensional basis vectors , @xmath18 , containing periodic or antiperiodic phases .
the sectors of the theory are given by the set of @xmath19 where @xmath20 so that @xmath21 . the spectrum of the theory at generic radius in any sector is determined by imposing the gso projections governed by the vectors @xmath18 and a set of structure constants @xmath22 , according to the klst set of rules in refs.@xcite ( and equivalently ref.@xcite ) .
the model is then further compactified down to @xmath1 on a @xmath23 orbifold .
in the absence of any cdc the result would simply be an @xmath24 model resulting from an overall @xmath25 compactification .
the cdc is implemented by introducing a deformation , described by another vector @xmath26 , of shifts in the charge lattice that depend on the radii @xmath27 of the @xmath28 ; this will be shown explicitly below . under the cdc ,
the virasoro generators of the theory are modified , yielding an extra effective projection condition governed by @xmath29 on the states constituting the massless spectrum of the @xmath1 theory .
the remaining massless states are then characterized by their charges under the @xmath30 symmetry associated with @xmath29 . to qualify as a scherk - schwarz mechanism
, this @xmath30 symmetry has to include some component of the @xmath31-symmetry in order to distinguish bosons from fermions , thereby projecting out the gravitinos , and breaking spacetime susy .
the effect of the cdc of course disappears in the strict @xmath32 limit where the kaluza - klein ( kk ) spectrum becomes continuous , and the @xmath0 endpoint model @xmath17 is recovered . on the other hand as we shall see the cdc turns into another gso projection vector in the @xmath33 limit , where states either remain massless or become infinitely massive . upon @xmath2-dualising ,
@xmath34 the @xmath35 model becomes the non - compact theory @xmath36 , whose properties depend precisely on the form of @xmath29 .
the theories at the two endpoints can contain a different number of states and charges . because @xmath29 can overlap the gauge degrees of freedom
, @xmath36 will generically have a gauge symmetry that differs from that of @xmath17 , and possibly no susy . as we will see the two are in fact
linked : if @xmath36 is supersymmetric then the gauge group is the same as that of @xmath17 , if it is not , then the gauge group is different .
let us now elaborate on the above description .
the conventions for the fermionic construction are as in refs.@xcite and for the cdc are as outlined in ref.@xcite .
that is the unmodified virasoro operators are defined as @xmath37 where , in terms of the winding and kk numbers , @xmath38 and @xmath39 respectively , the left- and right - moving momenta for a theory compactified on two circles of radii @xmath40 take the unshifted form @xmath41 note that level - matching is preserved by the cdc , but the mass spectrum is modified rather than the number of degrees of freedom contained within the theory , as required for a spontaneous breaking of susy @xcite .
it is clear from eq.([eqn : virasoro_2 ] ) that for zero winding modes ( @xmath58 ) , states for which @xmath59 become massive under the action of the cdc .
conversely all the zero winding states in the ns - ns sector remain unshifted by the cdc since they are chargeless .
as described in ref.@xcite there may or may not be massless gravitinos depending on whether the effective projection @xmath60 is aligned with the other projections : this is in turn dependent on the choice of structure constant , so that ultimately the breaking of susy is associated with breaking by discrete torsion . to evaluate the cosmological constant , at given radii @xmath61 , one must integrate each @xmath62 term ( weighted by its coefficient @xmath63 ) in the total 1-loop partition function over the fundamental domain @xmath64 of the modular group : @xmath65 where @xmath66 is the number of uncompactified spacetime dimensions , and @xmath67 is the reduced string scale . henceforth @xmath68 is set to 1 ; it can be reinserted by dimensional analysis at the end of the calculation if desired .
the integral splits into upper ( @xmath69 ) and lower regions of the fundamental domain . only terms for which @xmath70 can receive contributions from both regions , with the @xmath71 integral yielding zero in the upper region when @xmath72 , enforcing level matching in the infra - red ( but allowing contributions from unphysical proto - graviton modes in the ultra - violet as described in ref.@xcite ) . at general radius the evaluation of the cosmological constant is complicated immensely by the fact that @xmath73 vary with @xmath11 . in order to make the evaluation tractable
, the total partition function , @xmath74 , has to be rearranged into separate bosonic and fermionic factors as follows .
it is convenient to define @xmath75 and @xmath76 .
twisted sectors are not required in this implementation as they do not contribute to the cosmological constant , although we will comment on them later when we come to generalise the construction .
we have@xmath77 where @xmath78\big\}\ , , \ ] ] and @xmath79 in the above the coefficients of the partition function are given by @xmath80\big\}\ , c^\alpha_\beta\ , , \
] ] where @xmath81 are the coefficients of the original theory before cdc , namely @xmath82\ , .\ ] ] it is convenient to use the resummed version of this expression ; certainly for the @xmath83-expansion this is the preferred method as it makes modular invariance explicit .
this removes the @xmath84 prefactor and adds a factor of @xmath85 . the bosonic factor in the partition function @xmath86
depend upon the radii of compactification , the winding and resummed kk numbers and the cdc induced shift in the kk levels , @xmath42 , as follows : @xmath87 the effective shift in the kk number , given by the requisite @xmath88 , arises from the choice of @xmath89 , which gives an overall phase @xmath90 in the partition function ; as we shall see this shift in the kk number ultimately amounts to introducing a new vector @xmath91 in the non - compact @xmath2-dual theory at zero radius , combined with structure constants @xmath92 , @xmath93 .
note that this means in the 4d spectrum one may find states with 1/4-charges @xmath94 , that since they have @xmath95 , become infinitely massive in the zero radius limit . in order to reorder the sum to do it efficiently , a projection in the @xmath96 on @xmath44
is now introduced to select possible values of @xmath42 .
the parameter for this projection will be called @xmath97 .
thus overall one can write,@xmath98 where @xmath99 note that the phases in @xmath100 are precisely what is needed to cancel the contribution coming from the theta functions in @xmath101 , so that overall the spectrum is merely shifted , with the gso projections remaining independent of @xmath29 .
the bosonic contribution to the total partition function is independent of the fermionic sectors within the theory so @xmath102 appears as a pre - factor to the sector sum for any given @xmath42 .
conversely , the fermionic partition function is composed of terms that depend upon the boundary conditions of the fermions within the sectors @xmath103 and is independent of the compactification radii .
the advantage of this reordering is that one can therefore collect 16 representative @xmath104 factors ( @xmath105 ) , @xmath106 which are independent of the radii , and 16 representative @xmath23 factors ( @xmath105 ) , @xmath107 + the latter are radius dependent interpolating functions , analogous to the functions @xmath108 in the simple circular case studied in ref.@xcite .
the total partition function is then compiled by summing over the 16 @xmath109 sectors as @xmath110 an interesting aspect of the above approach is that in the small radius limit , that part of the spectrum with @xmath95 mod(1 ) decouples and can be discarded , leaving the partition function of the non - compact @xmath0 theory at @xmath111 .
indeed , poisson resumming on @xmath112 and @xmath113 gives @xmath114 where the ellipsis indicate terms that are further exponentially suppressed .
thus the total untwisted partition function in the small radius limit can be expressed as @xmath115 note that @xmath116 is simply the expected volume factor of the partition function in the @xmath2-dual @xmath0 theory . in conjunction with the fermionic component of the partition function ,
this then reproduces a @xmath0 model with an additional basis vector @xmath29 , appearing in the sector definitions as @xmath117 , and with eq.([newgso ] ) providing a new gso projection , namely @xmath118 mod ( 1 ) .
( the mod ( 1 ) comes courtesy of the sum over @xmath39 . ) upon inspection therefore , we are finding that eq.([newgso ] ) _ is actually _ the gso projection of an additional vector @xmath91 in the non - compact @xmath0 theory . beginning with the choice of @xmath119
, one can infer that the @xmath0 theory at zero radius for the examples we have been considering has structure constants @xmath92 and @xmath93 , consistent with the modular invariance rules of klst in refs.@xcite .
in fact identifying sectors as @xmath120 with @xmath121 mod(2 ) , the entire partition function at zero radius is that of the @xmath0 theory with the appropriate corresponding gso phases , @xmath122 reversing the line of reasoning above , leads us finally to a generalisation of the construction of interpolating models based on the modular invariance of their endpoint @xmath0 theories : * first , define a @xmath0 theory in terms of a set of vectors @xmath18 , and _ any _ additional @xmath91 vector that obeys the @xmath0 modular invariance rules of ref.@xcite , together with a set of consistent structure constants @xmath123 and @xmath124 .
( the @xmath123 are then fixed by the modular invariance rules in the usual way . ) * in theories that have an additional @xmath125 orbifold action @xmath126 on compactification to @xmath1 , @xmath91 is still constrained by the need to preserve mutually consistent gso projections , with the condition @xmath127 ( as in refs.@xcite and discussed in ref.@xcite ) . *
the partition function is then in the form of eqs.([mini1]),([mini2 ] ) with coefficients as in eq.([eq : be ] ) . the projection obtained by performing
the @xmath128 sum determines the corresponding kk shift to be @xmath129 to prove that these rules are consistent with modular invariance , one may first poisson - resum back to the original expression but retaining @xmath128 , so that entire partition function is @xmath130 note that the sum over @xmath42 provides a projection that equates @xmath131 mod(1 ) .
it is then straightforward to show that the partition function is invariant under @xmath132 provided that @xmath133 and invariant under @xmath134 provided that @xmath135 this overall set of conditions is precisely that of klst @xcite with the original theory enlarged to include the vector @xmath91 .
@xmath136 note that these rules are significantly more general than those of refs.@xcite , in which the choice @xmath137 corresponds to taking @xmath92 and @xmath93 .
now for example the cdc vectors are no longer restricted to obey @xmath138 mod(1 ) , and moreover the kk shifts have additional sector dependence if @xmath139 .
we should add that , as well as being a generalisation , these rules simplify the construction of viable phenomenological models , because the @xmath127 condition can be implemented independently , with consistency then guaranteed with respect to all the other @xmath18 vectors .
one can also conclude that for consistency a theory that is scherk - schwarzed on an orbifold should contain additional sectors that are twisted under the action of both the orbifold and the scherk - schwarz i.e. twisted sectors that have non - zero @xmath140 .
of course @xmath140 for such sectors has no association with any windings , but one finds that those sectors ( which being twisted are supersymmetric ) are required for consistency ( anomaly cancellation for example ) .
let us now move on to the conditions under which the endpoint theories exhibit susy .
we will always consider models in which the theory at infinite radius is supersymmetric ( as would be evidenced by the vanishing of the cosmological constant there ) but we would like to determine whether or not susy is restored at zero radius as well . in this section
we develop arguments to address this question based on the existence or otherwise of massless gravitinos as @xmath33 . as usual the pure neveu - schwarz ( ns - ns ) sector , @xmath141 gives rise to the gravity multiplet , @xmath142 ( the graviton ) , @xmath143 ( the dilaton ) and @xmath144}$ ] ( the two index antisymmetric tensor ) , from the states @xmath145 in the notation of ref.@xcite .
these states are chargeless under @xmath146 and no projection on them can occur , since the cdc vector is always zero in the @xmath1 space - time dimensions @xmath147 . given the inevitable presence of the graviton , the susy properties of the theory are then dictated by the presence or absence of the r - ns gravitinos , namely @xmath148 their scherk - schwarz projections are determined purely by the scherk - schwarz action on the right - moving degrees of freedom the spectrum is found from the expressions for the modified virasoro operators in eq.([eqn : virasoro_2 ] ) . for the non - winding gravitinos , the shifted kk momentum becomes virtually continuous in the @xmath149 limit and the full @xmath0 gravitino state is inevitably recovered there .
the scale at which susy is spontaneously broken by the cdc is set by the gravitino mass @xmath150 . as the compactification is turned on
, the susy of the @xmath0 theory is broken , and then towards the @xmath33 end of the interpolation , new gravitinos may or may not appear in the massless spectrum , perhaps heralding the restoration of susy at small radius as well . to see if they do ,
consider how the cdc modifies the theories that sit at the endpoints of the interpolation .
we denote by @xmath151 the charge of the lightest gravitino state at large radius .
susy is exact even in the presence of @xmath29 , with the state @xmath151 being exactly massless , if both the first and second terms in the modified virasoro operators of eq.([eqn : virasoro_2 ] ) , namely @xmath152 and @xmath153 vanish .
( for convenience we continue for this discussion to use the original more restrictive rules of refs.@xcite ; it would be trivial to extend the discussion to the more general rules of eq.([eq : genrule ] ) . ) with @xmath154 , the first term receives no extra contribution due to the cdc . furthermore , there is no winding contribution to the second term .
therefore gravitinos that have @xmath155 remain massless and indicate the presence of exact susy .
conversely , if the only remaining gravitinos have @xmath156 their mass is @xmath157 and susy is spontaneously broken . without loss of generality
, one can consider susy breaking to amount to a conflict between @xmath26 and a single basis vector , denoted by @xmath158 .
that is , @xmath158 constrains the gravitinos , while the remaining @xmath18 can not project them out of the theory . in order for the above light ( but not massless ) gravitino to be the one that is left un - projected , the projections due to @xmath159 and @xmath158 must disagree , that is the massive @xmath160 state is retained by @xmath158 while the massless @xmath155 state is projected out . again without loss of generality , it is always possible to choose @xmath158 so that the conditions are aligned ; that is @xmath161 .
these modes are preserved ( and have a mass @xmath162 ) while @xmath158 projects the massless @xmath163 modes out of the theory entirely .
now consider the zero radius end of the interpolation , and denote the new would - be massless gravitino state by @xmath164 .
although a different state , it can be related to the infinite radius gravitino @xmath165 by a shift in the charge vector , induced by a potentially non - zero winding number ; @xmath166 as @xmath11 vanish , the spectrum associated with the winding modes becomes continuous , while the kk states become extremely heavy .
as described in the previous section , the requirement that the kk term in eq.([eqn : virasoro_term_2 ] ) vanishes forms an effective projection that constrains the light states at zero radius , selecting the modes for which @xmath167 where we will assume that @xmath168 .
it is clear from the relation between @xmath169 and @xmath170 in eq.([eqn : modified_q_psi ] ) that the projection due to the cdc vector remains unchanged for any gravitino state there , since @xmath171 ; that is @xmath172 this equation together with eqs.([eqn : modified_projection ] ) and ( [ gravcon ] ) , imply that any gravitino of the spontaneously broken theory that becomes light at small radius must be an odd - winding mode . under the shift in @xmath44 given by eq.([eqn : modified_q_psi ] ) , the @xmath158 projection constraining the gravitinos is @xmath173 for the effective projection in eq.(@xmath174 ) to agree with the modified gso condition in eq.([eqn : modified_gso_condition ] ) for @xmath175 , we then require that @xmath176 eq.([cond1 ] ) is a necessary condition for a model with susy spontaneously broken by the scherk - schwarz mechanism to have massless gravitino states in both the infinite and zero radius limits .
let us see what it implies in a specific theory .
consider the basis vector set @xmath177 , together with a cdc vector that is empty in its left - moving elements : @xmath178\nonumber\\ v_1&= & - \mbox{${\scriptstyle \frac{1}{2}}$}[~00~011~011~ | ~1111~11111~111~111~11~111~]\nonumber\\ v_2&= & - \mbox{${\scriptstyle \frac{1}{2}}$}[~00~101~101~ | ~0101~00000~011~111~11~111~]\nonumber\\ v_4&= & - \mbox{${\scriptstyle \frac{1}{2}}$}[~00~101~101~ | ~0101~00000~011~000~00~000~]\nonumber\\ { \bf e } & = & - \mbox{${\scriptstyle \frac{1}{2}}$}[~00~101~101~ | ~0000~00000~000~000~00~000~]\ , . \end{aligned}\ ] ] a suitable and consistent set of structure constants is @xmath179 gravitinos are found in the @xmath180 $ ] sector , with vacuum energies @xmath181 = [ 0 , -1]$ ] .
the charge operator for the non - winding gravitinos in the initial ( infinite radius ) theory takes the same form as the sector vector itself .
they have charges determined by @xmath182 that give the required @xmath183 mod ( 1 ) for spontaneous susy breaking : the positive helicity states with this choice of structure constants are @xmath184\ , , \ ] ] where the @xmath185 signs on the fermions are co - dependent .
it is clear from the vector overlap between @xmath186 and @xmath182 that the latter is playing the role of @xmath158 that constrains the gravitini states .
( the structure constants have been chosen such that @xmath187 yields identical constraints . )
whether or not any of the winding modes of the gravitinos are light at zero radius depends upon them satisfying the modified gso projection condition of eq.([eqn : modified_gso_condition ] ) : @xmath188 as we saw the two projections agree for the odd - winding modes of the @xmath164 states since @xmath189 , and under the cdc , the charge vector for the small radius gravitino is @xmath190\ , .\ ] ] note that non - zero right - moving charges of the small radius gravitino are on the @xmath191 and @xmath192 world - sheet degrees of freedom , and they no longer overlap the susy charges of the large radius theory .
the appearance of gravitino states in the light spectrum in the zero radius limit of this theory reflects a general conclusion . if the left - moving elements of the cdc vector vanish , eq.([cond1 ] ) is automatically satisfied .
any theory with a cdc vector acting purely on the space - time side becomes supersymmetric at zero radius since the projection always preserves the odd - winding modes of the gravitinos .
the non - supersymmetric @xmath1 theory at generic radius is therefore an interpolation between two supersymmetric theories quite generally in these cases , which sit at the zero and infinite radius endpoints .
the supersymmetric nature of the zero radius theory will later be verified by the vanishing of the cosmological constant in the @xmath33 limit ( figure [ fig : plots_order_1_radii_increments_0.02_trivial_e_lm_nb - nf=28_r_only ] ) , as presented in the following section .
note that the necessary cancellation between thousands of terms is highly non - trivial . consider instead a theory composed of the same basis vector set as in eq.([mod : output1 ] ) , but now with a cdc vector containing non - zero left - moving entries : for example @xmath193\ , .\end{aligned}\ ] ] under the cdc , and for convenience of presentation dropping the @xmath185 signs , the charge vector for the odd - winding gravitino modes is modified to @xmath194\ , .\ ] ] as in the previous example the vector contains the same number of non - zero right - moving entries , but lying in different columns , so there is no contribution from eq.([eqn : virasoro_term_1 ] ) to the mass squared on the space - time side .
however the non - zero left - moving elements now result in a non - zero contribution . under the shift , @xmath195
any non - zero shift in @xmath196 will inevitably produce massive gravitinos since in the r - ns sector the charges of massless states must be zero mod ( 1 ) on the left - moving side . we conclude that susy is restored at small as well as large radius if and only if the scherk - schwarz mechanism does not act on the gauge - side .
conversely if susy is broken at zero radius then so is the gauge symmetry .
the dominant terms in the one loop contribution to the cosmological constant are proportional to @xmath198 for the massless states @xcite , so non - supersymmetric models with an equal number of massless bosonic and fermionic states have an exponentially suppressed one - loop cosmological constant , and hence exhibit an increased degree of stability .
unfortunately it seems to be necessary to determine the full massless spectrum in order to deduce whether or not @xmath197 .
there appears to be no principle , or algebraically feasible generic procedure , for choosing the basis vectors @xmath199 , the cdc vector @xmath26 , and the structure constants @xmath22 , that ensures that @xmath197 .
consider a theory containing @xmath200 , @xmath201 and @xmath187 as in the above basis vector set in eqs.([mod : output1 ] ) , a modified @xmath182 , an additional vector @xmath202 , and a cdc vector that acts only on the space - time side : @xmath203\nonumber\\ v_5&= & - \mbox{${\scriptstyle \frac{1}{2}}$}[~00~000~011~ | ~0101~11100~001~000~10~111~]\nonumber\\ { \bf e } & = & \mbox{${\scriptstyle \frac{1}{2}}$}[~00~101~101~ | ~0000~00000~000~000~00~000~ ] \ , .\end{aligned}\ ] ] a suitable and consistent set of structure constants @xmath22 is @xmath204 this model can be investigated using the general method presented in the previous section .
@xmath182 plays the role of @xmath158 , while @xmath202 respects its projections on the gravitinos .
as discussed the interpolation is between two supersymmetric endpoints at both small and large radius .
the cosmological constant takes a non - zero negative value with a minimum at intermediate values , and returns to zero at the two extremes , displayed in figure [ fig : plots_order_1_radii_increments_0.02_trivial_e_lm_nb - nf=28_r_only ] .
$ ] with radius increments of @xmath205 for a model with @xmath206 trivial . @xmath207 . _
_ ] a theory in which @xmath208 can be generated by a performing an alternative modification to the vectors @xmath209 : @xmath210\nonumber\\ v_5&= & - \mbox{${\scriptstyle \frac{1}{2}}$}[~00~000~011~ | ~0101~11100~010~110~00~011~]\nonumber\\ { \bf e } & = & \mbox{${\scriptstyle \frac{1}{2}}$}[~00~101~101~ | ~0000~00000~000~000~00~000~ ] \ , , \end{aligned}\ ] ] with the following structure constants : @xmath211 similarly to the @xmath212 model with a exclusively non - trivial right - moving cdc vector , this model interpolates between two supersymmetric endpoints at both small and large radius , with the cosmological constant now taking a non - zero positive value at intermediate radii , displayed in figure [ fig : plots_order_1_radii_increments_0.02_trivial_e_lm_nb - nf=-228_r_only ] , corresponding to unstable runaway to decompactification at either end of the interpolation .
$ ] with radius increments of @xmath205 for a model with @xmath206 trivial and @xmath213 _ _ ] a theory with bose - fermi degeneracy can be achieved with a theory comprised of the basis vector set in eq.([mod : output1 ] ) , plus a basis vector @xmath202 and cdc vector of the form @xmath214 \ , , \nonumber \\ { \bf e } & = & \mbox{${\scriptstyle \frac{1}{2}}$}[~00~101~101~ | ~1011~00000~000~100~01~111~]\ , , \end{aligned}\ ] ] with @xmath22 given by @xmath215 @xmath216 and @xmath217 are found to be equal despite the fact that the theory is non - supersymmetric ( as can be seen by the absence of any massless gravitini in the spectrum ) . for models in which the cdc vector @xmath26 is non - trivial in both the gauge and the global entries , the cosmological
constant takes a non - zero value at small radius , while it vanishes exponentially quickly for large compactification scales , as displayed in figure [ fig : plots_order_1_radii_increments_0.02_non_trivial_e_lm_nb = nf_r_only ] .
$ ] with radius increments of @xmath205 for a model with @xmath206 non - trivial and @xmath197 . _
an interpolation from susy to non - susy in which @xmath212 , can be achieved by taking the corresponding set of basis vectors in eqs.([mod : susy_to_susy_nb > nf ] ) , but now with a cdc vector of the form @xmath218\ , .\end{aligned}\ ] ] for models in which @xmath212 , the cosmological constant reduces from a constant positive value at small radius reaching a negative minimum at approximately @xmath219 . as the radius increases to @xmath220 , the cosmological constant tends to zero from negative values , consistent with the restoration of susy in the endpoint model , as displayed in figure [ fig : plots_order_1_radii_increments_0.02_non_trivial_e_lm_nb - nf=192_r_only ] .
$ ] with radius increments of @xmath205 for a model with @xmath206 non - trivial , and @xmath221 . _
_ ] finally for a non - susy to susy interpolation with @xmath208 , we take the model in eqs.([mod : susy_to_susy_nb < nf ] ) but now with a cdc vector of the form @xmath222\ , .\end{aligned}\ ] ] the cosmological constant increases from a constant negative minimum at small radius , to a non - susy 6d theory at small radius and a susy 6d theory at infinite radius , as displayed in figure [ fig : plots_order_1_radii_increments_0.02_non_trivial_e_lm_nb - nf=-64_r_only ] .
$ ] with radius increments of @xmath205 for a model with @xmath206 non - trivial and @xmath223 . _
following on from ref.@xcite , the nature of heterotic strings in the context of scherk - schwarz compactification has been investigated , with particular emphasis on their properties under interpolation . from the starting point of supersymmetric @xmath0 theories in the infinite radius limit , scherk - schwarz compactification to @xmath1 yields models that have @xmath224 , each possibility exhibiting different behaviours under interpolation .
the behaviour of their cosmological constants was studied as a function of compactification radius , and it was found that theories can yield maxima or minima in the cosmological constant at intermediate values , as well as barriers with apparent metastability .
the latter feature may have interesting phenomenological and/or cosmological applications .
the nature of the scherk - schwarz action , in particular whether or not it simultaneously acts to break the gauge group , dictates whether or not susy emerges in the @xmath0 theory at zero radius . by studying the relation of the interpolating theory to the @xmath0 theories that emerge at the end - points of the interpolation , in particular their modular invariance constraints ,
a more general class of scherk - schwarz compactifications was also derived .
the aim in this work has been to establish the general features of interpolating models , relating higher , @xmath66-dimensional models to @xmath225-dimensional compactified models .
it is conceivable that very many non - supersymmetric tachyon - free @xmath1 models can be interpolated to higher dimensional supersymmetric ones .
this would imply the existence of a formal relation between the process of interpolation , and the restoration of susy .
looking forward , it may not be possible to show that _ every _ non - supersymmetric theory is related to a supersymmetric counterpart via the process of interpolation .
however , it seems possible that such a relation may always hold for the particular class of theories in which susy is broken by discrete torsion , as in ref.@xcite for example . a goal for future work would be to establish relationships , of the type found in this study , between additional lower dimensional , non - supersymmetric models , ideally of greater phenomenological appeal , and their supersymmetric counterparts .
if it can be shown that non - supersymmetric models generically relate to supersymmetric theories in this way , interpolation could be used as a tool with which to relate many tachyon - free non - supersymmetric string theories to their supersymmetric siblings .
thus it would be possible to locate non - supersymmetric models within the larger network of string theories extending previous work in this direction .
+ * acknowledgements * : we are extremely grateful to keith dienes , emilian dudas and herv partouche for many interesting discussions and comments .
saa would like to thank the cole polytechnique for hospitality extended during this work .
s. abel , k. r. dienes and e. mavroudi , `` towards a nonsupersymmetric string phenomenology , '' phys .
d * 91 * , no .
12 , 126014 ( 2015 ) [ arxiv:1502.03087 [ hep - th ] ] .
s. abel , `` a dynamical mechanism for large volumes with consistent couplings , '' jhep * 1611 * , 085 ( 2016 ) doi:10.1007/jhep11(2016)085 [ arxiv:1609.01311 [ hep - th ] ] .
h. itoyama and t. r. taylor , `` supersymmetry restoration in the compactified o(16 ) @xmath226 o(16)-prime heterotic string theory , '' phys .
b * 186 * , 129 ( 1987 ) .
h. itoyama and t. r. taylor , `` small cosmological constant in string models , '' fermilab - conf-87 - 129-t , c87 - 06 - 25 .
c. kounnas and b. rostand , `` coordinate dependent compactifications and discrete symmetries , '' nucl .
b * 341 * , 641 ( 1990 ) .
s. ferrara , c. kounnas and m. porrati , `` superstring solutions with spontaneously broken four - dimensional supersymmetry , '' nucl .
b * 304 * , 500 ( 1988 ) .
s. ferrara , c. kounnas and m. porrati , `` @xmath227 superstrings with spontaneously broken symmetries , '' phys .
b * 206 * , 25 ( 1988 ) .
s. ferrara , c. kounnas , m. porrati and f. zwirner , `` superstrings with spontaneously broken supersymmetry and their effective theories , '' nucl .
b * 318 * , 75 ( 1989 ) .
l. alvarez - gaume , p. h. ginsparg , g. w. moore and c. vafa , `` an o(16 ) x o(16 ) heterotic string , '' phys .
b * 171 * , 155 ( 1986 ) ; + l. j. dixon and j. a. harvey , `` string theories in ten - dimensions without space - time supersymmetry , '' nucl .
b * 274 * , 93 ( 1986 ) .
k. r. dienes , `` modular invariance , finiteness , and misaligned supersymmetry : new constraints on the numbers of physical string states , '' nucl .
b * 429 * , 533 ( 1994 ) [ hep - th/9402006 ] ; `` how strings make do without supersymmetry : an introduction to misaligned supersymmetry , '' in * syracuse 1994 , proceedings , pascos 94 * 234 - 243 [ hep - th/9409114 ] ; `` space - time properties of ( 1,0 ) string vacua , '' in * los angeles 1995 , future perspectives in string theory * 173 - 177 [ hep - th/9505194 ] . k. r. dienes , m. moshe and r. c. myers , `` string theory , misaligned supersymmetry , and the supertrace constraints , '' phys .
lett . * 74 * , 4767 ( 1995 ) [ hep - th/9503055 ] ; `` supertraces in string theory , '' in * los angeles 1995 , future perspectives in string theory * 178 - 180 [ hep - th/9506001 ] .
r. rohm , `` spontaneous supersymmetry breaking in supersymmetric string theories , '' nucl .
b * 237 * , 553 ( 1984 ) .
v. p. nair , a. d. shapere , a. strominger and f. wilczek , `` compactification of the twisted heterotic string , '' nucl .
b * 287 * , 402 ( 1987 ) ; + p. h. ginsparg and c. vafa , `` toroidal compactification of nonsupersymmetric heterotic strings , '' nucl .
b * 289 * , 414 ( 1987 ) .
g. w. moore , `` atkin - lehner symmetry , '' nucl .
b * 293 * , 139 ( 1987 ) [ erratum - ibid .
b * 299 * , 847 ( 1988 ) ] ; + j. balog and m. p. tuite , `` the failure of atkin - lehner symmetry for lattice compactified strings , '' nucl .
b * 319 * , 387 ( 1989 ) ; + k. r. dienes , `` generalized atkin - lehner symmetry , '' phys . rev .
d * 42 * , 2004 ( 1990 ) .
k. r. dienes , `` new string partition functions with vanishing cosmological constant , '' phys .
* 65 * , 1979 ( 1990 ) .
d. kutasov and n. seiberg , `` number of degrees of freedom , density of states and tachyons in string theory and cft , '' nucl .
b * 358 * , 600 ( 1991 ) .
s. kachru , j. kumar and e. silverstein , `` vacuum energy cancellation in a nonsupersymmetric string , '' phys .
d * 59 * , 106004 ( 1999 ) ; [ hep - th/9807076 ] ; + s. kachru and e. silverstein , `` on vanishing two loop cosmological constants in nonsupersymmetric strings , '' jhep * 9901 * , 004 ( 1999 ) [ hep - th/9810129 ] .
j. a. harvey , `` string duality and nonsupersymmetric strings , '' phys .
d * 59 * , 026002 ( 1999 ) [ hep - th/9807213 ] ; + s. kachru and e. silverstein , `` selfdual nonsupersymmetric type ii string compactifications , '' jhep * 9811 * , 001 ( 1998 ) [ hep - th/9808056 ] ; + r. blumenhagen and l. gorlich , `` orientifolds of nonsupersymmetric asymmetric orbifolds , '' nucl .
b * 551 * , 601 ( 1999 ) [ hep - th/9812158 ] ; + c. angelantonj , i. antoniadis and k. forger , `` nonsupersymmetric type i strings with zero vacuum energy , '' nucl .
b * 555 * ( 1999 ) 116 [ hep - th/9904092 ] ; + m. r. gaberdiel and a. sen , `` nonsupersymmetric d - brane configurations with bose - fermi degenerate open string spectrum , '' jhep * 9911 * , 008 ( 1999 ) [ hep - th/9908060 ] .
g. shiu and s. h. h. tye , `` bose - fermi degeneracy and duality in nonsupersymmetric strings , '' nucl .
b * 542 * , 45 ( 1999 ) [ hep - th/9808095 ] .
r. iengo and c. j. zhu , `` evidence for nonvanishing cosmological constant in nonsusy superstring models , '' jhep * 0004 * , 028 ( 2000 ) [ hep - th/9912074 ] .
e. dhoker and d. h. phong , `` two loop superstrings 4 : the cosmological constant and modular forms , '' nucl .
b * 639 * , 129 ( 2002 ) [ hep - th/0111040 ] ; `` lectures on two loop superstrings , '' conf . proc .
c * 0208124 * , 85 ( 2002 ) [ hep - th/0211111 ] .
a. e. faraggi and m. tsulaia , `` interpolations among nahe - based supersymmetric and nonsupersymmetric string vacua , '' phys .
b * 683 * ( 2010 ) 314 [ arxiv:0911.5125 [ hep - th ] ] . c. angelantonj , m. cardella , s. elitzur and e. rabinovici , `` vacuum stability , string density of states and the riemann zeta function , '' jhep * 1102 * , 024 ( 2011 ) [ arxiv:1012.5091 [ hep - th ] ] .
o. bergman and m. r. gaberdiel , `` a nonsupersymmetric open string theory and s duality , '' nucl .
b * 499 * , 183 ( 1997 ) [ hep - th/9701137 ] ; `` dualities of type 0 strings , '' jhep * 9907 * , 022 ( 1999 ) [ hep - th/9906055 ] ; +
r. blumenhagen and a. kumar , `` a note on orientifolds and dualities of type 0b string theory , '' phys .
b * 464 * , 46 ( 1999 ) [ hep - th/9906234 ] .
j. d. blum and k. r. dienes , `` duality without supersymmetry : the case of the so(16 ) @xmath226 so(16 ) string , '' phys . lett .
b * 414 * , 260 ( 1997 ) [ hep - th/9707148 ] .
j. d. blum and k. r. dienes , `` strong / weak coupling duality relations for nonsupersymmetric string theories , '' nucl .
b * 516 * , 83 ( 1998 ) [ hep - th/9707160 ] .
a. e. faraggi and m. tsulaia , `` on the low energy spectra of the nonsupersymmetric heterotic string theories , '' eur .
j. c * 54 * , 495 ( 2008 ) [ arxiv:0706.1649 [ hep - th ] ] .
k. r. dienes , `` statistics on the heterotic landscape : gauge groups and cosmological constants of four - dimensional heterotic strings , '' phys .
d * 73 * , 106010 ( 2006 ) [ hep - th/0602286 ] .
k. r. dienes , m. lennek and m. sharma , `` strings at finite temperature : wilson lines , free energies , and the thermal landscape , '' phys . rev .
d * 86 * , 066007 ( 2012 ) [ arxiv:1205.5752 [ hep - th ] ] .
e. alvarez and m. a. r. osorio , `` cosmological constant versus free energy for heterotic strings , '' nucl .
b * 304 * , 327 ( 1988 ) [ erratum - ibid .
b * 309 * , 220 ( 1988 ) ] ; `` duality is an exact symmetry of string perturbation theory , '' phys .
d * 40 * , 1150 ( 1989 ) ; + m. a. r. osorio , `` quantum fields versus strings at finite temperature , '' int .
j. mod .
phys . a * 7 * , 4275 ( 1992 ) .
j. j. atick and e. witten , `` the hagedorn transition and the number of degrees of freedom of string theory , '' nucl .
b * 310 * , 291 ( 1988 ) .
m. mcguigan , `` finite temperature string theory and twisted tori , '' phys .
d * 38 * , 552 ( 1988 ) ; + i. antoniadis and c. kounnas , `` superstring phase transition at high temperature , '' phys .
b * 261 * , 369 ( 1991 ) ; + i. antoniadis , j. p. derendinger and c. kounnas , `` non - perturbative temperature instabilities in n = 4 strings , '' nucl .
b * 551 * , 41 ( 1999 ) [ arxiv : hep - th/9902032 ] .
m. j. bowick and l. c. r. wijewardhana , `` superstrings at high temperature , '' phys .
lett . * 54 * , 2485 ( 1985 ) ; + s. h. h. tye , `` the limiting temperature universe and superstring , '' phys .
b * 158 * , 388 ( 1985 ) ; + b. sundborg , `` thermodynamics of superstrings at high - energy densities , '' nucl .
b * 254 * , 583 ( 1985 ) ; + e. alvarez , `` strings at finite temperature , '' nucl .
b * 269 * , 596 ( 1986 ) ; + e. alvarez and m. a. r. osorio , `` superstrings at finite temperature , '' phys . rev .
d * 36 * , 1175 ( 1987 ) ; `` thermal heterotic strings , '' physica * a 158 * , 449 ( 1989 ) [ erratum - ibid .
a * 160 * , 119 ( 1989 ) ] ; `` thermal strings in nontrivial backgrounds , '' phys . lett .
b * 220 * , 121 ( 1989 ) ; + m. axenides , s. d. ellis and c. kounnas , `` universal behavior of d - dimensional superstring models , '' phys . rev .
d * 37 * , 2964 ( 1988 ) ; + y. leblanc , `` cosmological aspects of the heterotic string above the hagedorn temperature , '' phys .
d * 38 * , 3087 ( 1988 ) ; + b. a. campbell , j. r. ellis , s. kalara , d. v. nanopoulos and k. a. olive , `` phase transitions in qcd and string theory , '' phys .
b * 255 * , 420 ( 1991 ) .
e. kiritsis and c. kounnas , `` perturbative and nonperturbative partial supersymmetry breaking : n=4 @xmath228 n=2 @xmath228 n=1 , '' nucl .
b * 503 * , 117 ( 1997 ) [ hep - th/9703059 ] .
e. dudas and j. mourad , `` brane solutions in strings with broken supersymmetry and dilaton tadpoles , '' phys .
b * 486 * , 172 ( 2000 ) [ hep - th/0004165 ] .
c. a. scrucca and m. serone , `` on string models with scherk - schwarz supersymmetry breaking , '' jhep * 0110 * , 017 ( 2001 ) [ hep - th/0107159 ] .
m. borunda , m. serone and m. trapletti , `` on the quantum stability of iib orbifolds and orientifolds with scherk - schwarz susy breaking , '' nucl .
b * 653 * , 85 ( 2003 ) [ hep - th/0210075 ] . c. angelantonj , m. cardella and n. irges , `` an alternative for moduli stabilisation , '' phys .
b * 641 * , 474 ( 2006 ) [ hep - th/0608022 ] .
d. lust , `` compactification of the o(16 ) x o(16 ) heterotic string theory , '' phys .
b * 178 * , 174 ( 1986 ) .
w. lerche , d. lust and a. n. schellekens , `` ten - dimensional heterotic strings from niemeier lattices , '' phys .
b * 181 * , 71 ( 1986 ) .
w. lerche , d. lust and a. n. schellekens , `` chiral four - dimensional heterotic strings from selfdual lattices , '' nucl . phys .
b * 287 * , 477 ( 1987 ) .
a. h. chamseddine , j. p. derendinger and m. quiros , `` nonsupersymmetric four - dimensional strings , '' nucl .
b * 311 * , 140 ( 1988 ) .
a. font and a. hernandez , `` nonsupersymmetric orbifolds , '' nucl .
b * 634 * , 51 ( 2002 ) [ hep - th/0202057 ] .
m. blaszczyk , s. groot nibbelink , o. loukas and s. ramos - sanchez , `` non - supersymmetric heterotic model building , '' jhep * 1410 * , 119 ( 2014 ) [ arxiv:1407.6362 [ hep - th ] ] . c. angelantonj , i. florakis and m. tsulaia , `` universality of gauge thresholds in non - supersymmetric heterotic vacua , '' phys .
b * 736 * ( 2014 ) 365 [ arxiv:1407.8023 [ hep - th ] ] ; + c. angelantonj , i. florakis and m. tsulaia , `` generalised universality of gauge thresholds in heterotic vacua with and without supersymmetry , '' nucl . phys .
b * 900 * , 170 ( 2015 ) doi:10.1016/j.nuclphysb.2015.09.007 [ arxiv:1509.00027 [ hep - th ] ] .
m. blaszczyk , s. groot nibbelink , o. loukas and f. ruehle , `` calabi - yau compactifications of non - supersymmetric heterotic string theory , '' jhep * 1510 * , 166 ( 2015 ) doi:10.1007/jhep10(2015)166 [ arxiv:1507.06147 [ hep - th ] ] .
s. groot nibbelink and e. parr , `` twisted superspace : non - renormalization and fermionic symmetries in certain heterotic - string - inspired non - supersymmetric field theories , '' phys .
d * 94 * , no . 4 , 041704 ( 2016 ) doi:10.1103/physrevd.94.041704 [ arxiv:1605.07470 [ hep - ph ] ] .
a. e. faraggi , c. kounnas and h. partouche , `` large volume susy breaking with a solution to the decompactification problem , '' nucl .
b * 899 * , 328 ( 2015 ) doi:10.1016/j.nuclphysb.2015.08.001 [ arxiv:1410.6147 [ hep - th ] ] .
c. kounnas and h. partouche , `` stringy n = 1 super no - scale models , '' pos planck * 2015 * , 070 ( 2015 ) [ arxiv:1511.02709 [ hep - th ] ] .
h. partouche , `` large volume supersymmetry breaking without decompactification problem , '' arxiv:1601.04564 [ hep - th ] . c. kounnas and h. partouche , `` super no - scale models in string theory , '' nucl . phys .
b * 913 * , 593 ( 2016 ) doi:10.1016/j.nuclphysb.2016.10.001 [ arxiv:1607.01767 [ hep - th ] ] . c. bachas ,
`` a way to break supersymmetry , '' hep - th/9503030 ; + j. g. russo and a. a. tseytlin , `` magnetic flux tube models in superstring theory , '' nucl .
b * 461 * , 131 ( 1996 ) [ hep - th/9508068 ] ; + a. a. tseytlin , `` closed superstrings in magnetic field : instabilities and supersymmetry breaking , '' nucl .
suppl . * 49 * , 338 ( 1996 ) [ hep - th/9510041 ] ; + h. p. nilles and m. spalinski , `` generalized string compactifications with spontaneously broken supersymmetry , '' phys .
b * 392 * , 67 ( 1997 ) [ hep - th/9606145 ] ; + i. shah and s. thomas , `` finite soft terms in string compactifications with broken supersymmetry , '' phys .
b * 409 * , 188 ( 1997 ) [ hep - th/9705182 ] .
a. sagnotti , `` some properties of open string theories , '' in * palaiseau 1995 , susy 95 * 473 - 484 [ hep - th/9509080 ] .
a. sagnotti , `` surprises in open string perturbation theory , '' nucl . phys .
suppl . * 56b * , 332 ( 1997 ) [ hep - th/9702093 ] . c. angelantonj ,
`` nontachyonic open descendants of the 0b string theory , '' phys .
b * 444 * , 309 ( 1998 ) [ hep - th/9810214 ] .
r. blumenhagen , a. font and d. lust , `` tachyon free orientifolds of type 0b strings in various dimensions , '' nucl .
b * 558 * , 159 ( 1999 ) [ hep - th/9904069 ] .
s. sugimoto , `` anomaly cancellations in type i d-9 - anti - d-9 system and the usp(32 ) string theory , '' prog . theor .
phys . * 102 * , 685 ( 1999 ) [ hep - th/9905159 ] .
g. aldazabal , l. e. ibanez and f. quevedo , `` standard - like models with broken supersymmetry from type i string vacua , '' jhep * 0001 * , 031 ( 2000 ) [ hep - th/9909172 ] .
c. angelantonj , `` nonsupersymmetric open string vacua , '' pos trieste * 99 * , 015 ( 1999 ) [ hep - th/9907054 ] .
k. forger , `` on nontachyonic z(n ) x z(m ) orientifolds of type 0b string theory , '' phys .
b * 469 * , 113 ( 1999 ) [ hep - th/9909010 ] .
s. moriyama , `` usp(32 ) string as spontaneously supersymmetry broken theory , '' phys .
b * 522 * , 177 ( 2001 ) [ hep - th/0107203 ] . c. angelantonj and i. antoniadis ,
`` suppressing the cosmological constant in nonsupersymmetric type i strings , '' nucl .
b * 676 * , 129 ( 2004 ) [ hep - th/0307254 ] . c. angelantonj , `` open strings and supersymmetry breaking , '' aip conf .
proc . * 751 * , 3 ( 2005 ) [ hep - th/0411085 ] .
e. dudas and c. timirgaziu , `` nontachyonic scherk - schwarz compactifications , cosmology and moduli stabilization , '' jhep * 0403 * , 060 ( 2004 ) [ hep - th/0401201 ] . b. gato - rivera and a. n. schellekens , `` non - supersymmetric tachyon - free type - ii and type - i closed strings from rcft , '' phys .
b * 656 * , 127 ( 2007 ) [ arxiv:0709.1426 [ hep - th ] ] . i. antoniadis , c. bachas , d. c. lewellen and t. n. tomaras , `` on supersymmetry breaking in superstrings , '' phys .
b * 207 * , 441 ( 1988 ) .
i. antoniadis , `` a possible new dimension at a few tev , '' phys .
b * 246 * , 377 ( 1990 ) .
i. antoniadis , c. munoz and m. quiros , `` dynamical supersymmetry breaking with a large internal dimension , '' nucl .
b * 397 * , 515 ( 1993 ) [ hep - ph/9211309 ] .
i. antoniadis and m. quiros , `` large radii and string unification , '' phys .
b * 392 * , 61 ( 1997 ) [ hep - th/9609209 ] .
k. benakli , `` phenomenology of low quantum gravity scale models , '' phys .
d * 60 * , 104002 ( 1999 ) [ hep - ph/9809582 ] .
c. p. bachas , `` scales of string theory , '' class .
* 17 * , 951 ( 2000 ) [ hep - th/0001093 ] .
e. dudas , `` theory and phenomenology of type i strings and m theory , '' class .
* 17 * , r41 ( 2000 ) [ hep - ph/0006190 ] .
i. florakis , `` gravitational threshold corrections in non - supersymmetric heterotic strings , '' arxiv:1611.10323 [ hep - th ] ; + i. florakis and j. rizos , `` chiral heterotic strings with positive cosmological constant , '' nucl .
b * 913 * , 495 ( 2016 ) doi:10.1016/j.nuclphysb.2016.09.018 [ arxiv:1608.04582 [ hep - th ] ] .
h. kawai , d. c. lewellen and s. h. h. tye , `` construction of fermionic string models in four - dimensions , '' nucl .
b * 288 * , 1 ( 1987 ) .
doi:10.1016/0550 - 3213(87)90208 - 2 h. kawai , d. c. lewellen and s. h. h. tye , `` construction of four - dimensional fermionic string models , '' phys . rev .
lett . * 57 * , 1832 ( 1986 ) erratum : [ phys .
lett . * 58 * , 429 ( 1987 ) ] .
doi:10.1103/physrevlett.57.1832 h. kawai , d. c. lewellen and s. h. h. tye , `` classification of closed fermionic string models , '' phys . rev .
d * 34 * , 3794 ( 1986 ) .
doi:10.1103/physrevd.34.3794 h. kawai , d. c. lewellen , j. a. schwartz and s. h. h. tye , `` the spin structure construction of string models and multiloop modular invariance , '' nucl . phys .
b * 299 * , 431 ( 1988 ) .
doi:10.1016/0550 - 3213(88)90544 - 5 i. antoniadis , c. p. bachas and c. kounnas , `` four - dimensional superstrings , '' nucl .
b * 289 * , 87 ( 1987 ) .
doi:10.1016/0550 - 3213(87)90372 - 5 | the interpolation from supersymmetric to non - supersymmetric heterotic theories is studied , via the scherk - schwarz compactification of supersymmetric @xmath0 theories to @xmath1 .
a general modular - invariant scherk - schwarz deformation is deduced from the properties of the @xmath0 theories at the endpoints , which significantly extends previously known examples . this wider class of non - supersymmetric @xmath1 theories is a promising prospect for the development of viable sm - like phenomenology .
the full one - loop cosmological constant of such theories is studied as a function of compactification radius for a number of cases , and the following interpolating configurations are found : two supersymmetric @xmath0 theories related by a @xmath2-duality transformation , with intermediate @xmath1 maximum or minimum at the string scale ; a non - supersymmetric @xmath0 theory interpolating to a supersymmetric @xmath0 theory , with the @xmath1 theory possibly having an ads minimum ; a `` metastable '' non - supersymmetric @xmath0 theory interpolating via a @xmath1 theory to a supersymmetric @xmath0 theory . |
in this paper we compare different formalisms that can be used to specify and model systems which are distributed , concurrent and mobile , as those that are usually found in the global computing area .
+ global computing is becoming very important because of the great development of networks which are deployed on huge areas , first of all internet , but also other kinds of networks such as networks for wireless communications . in order to build and program these networks
one needs to deal with issues such as reconfigurability , synchronization and transactions at a suitable level of abstraction .
thus powerful formal models and tools are needed .
until now no model has been able to emerge as the standard one for this kind of systems , but there are a lot of approaches with different merits and drawbacks .
+ an important approach is based on process calculi , like milner s ccs and hoare s csp . these two calculi deal with communication and synchronization in a simple way , but they lack the concept of mobility . an important successor of ccs , the @xmath0-calculus @xcite , allows to study a wide range of mobility problems in a simple mathematical framework .
we are mainly interested in the _ fusion calculus _
@xcite , which is an evolution of @xmath0-calculus .
the interesting aspect of this calculus is that it has been obtained by simplifying and making more symmetric the @xmath0-calculus .
+ one of the known limitations of process - calculi when applied to distributed systems is that they lack an intuitive representation because they are equipped with an _ interleaving _ semantics and they use the same constructions for representing both the agents and their configurations .
an approach that solves this kind of problems is based on _ graph transformations _ @xcite . in this case
the structure of the system is explicitly represented by a graph which offers both a clean mathematical semantics and a suggestive representation . in particular
we represent computational entities such as processes or hosts with hyperedges ( namely edges attached to any number of nodes ) and channels between them with shared nodes .
as far as the dynamic aspect is concerned , we use _ synchronized hyperedge replacement _ with _ hoare synchronization _ ( hshr ) @xcite .
this approach uses productions to specify the behaviour of single hyperedges , which are synchronized by exposing actions on nodes .
actions exposed by different hyperedges on the same node must be compatible . in the case of hoare synchronization
all the edges must expose the same action ( in the csp style ) .
this approach has the advantage , w.r.t .
other graphical frameworks such as double pushout @xcite or bigraphs @xcite , of allowing a distributed implementation since productions have a local effect and synchronization can be performed using a distributed algorithm .
+ we use the extension of hshr with _ mobility _
@xcite , that allows edges to expose node references together with actions , and nodes whose references are matched during synchronization are unified .
+ for us hshr is a good step in the direction of _ logic programming _ @xcite .
we consider logic programming as a formalism for modelling concurrent and distributed systems .
this is a non - standard view of logic programming ( see for a presentation of our approach ) which considers goals as processes whose evolution is defined by horn clauses and whose interactions use variables as channels and are managed by the unification engine . in this framework
we are not interested only in refutations , but in any partial computation that rewrites a goal into another .
+ in this paper we analyze the relationships between these three formalisms and we find tight analogies among them , like the same parallel composition operator and the use of unification for name mobility .
however we also emphasize the differences between these models : * the fusion calculus is interleaving and relies on _ milner synchronization _ ( in the ccs style ) ; * hshr is inherently _ concurrent _ since many actions can be performed at the same time on different nodes and uses hoare synchronization ; * logic programming is concurrent , has a wide spectrum of possible controls which are based on the hoare synchronization model , and also is equipped with a more complex data management .
we will show a mapping from fusion calculus to hshr and prove a correspondence theorem .
note that hshr is a good intermediate step between fusion calculus and logic programming since in hshr hyperedges can perform multiple actions at each step , and this allows to build chains of synchronizations .
this additional power is needed to model milner synchronization , which requires synchronous , atomic routing capabilities . to simplify our treatment we consider only reduction semantics .
the interleaving behaviour is imposed with an external condition on the allowed hshr transitions .
+ finally we present the connections between hshr and logic programming . since the logic programming paradigm allows for many computational strategies and
is equipped with powerful data structures , we need to constrain it in order to have a close correspondence with hshr .
we define to this end _ synchronized logic programming _ ( slp ) , which is a transactional version of logic programming .
the idea is that function symbols are pending constraints that must be satisfied before a transaction can commit , as for zero tokens in zero - safe nets @xcite . in the mapping from hshr to slp edges
are translated into predicates , nodes into variables and parallel composition into and composition . +
this translation was already presented in the msc .
thesis of the first author @xcite and in .
fusion calculus was mapped into shr with milner synchronization ( a simpler task ) in where fusion lts was considered instead of fusion reduction semantics .
the paper also contains a mapping of ambient calculus into hshr .
this result can be combined with the one here , thus obtaining a mapping of ambient calculus into slp .
an extensive treatment of all the topics in this paper can also be found in the forthcoming ph.d .
thesis of the first author @xcite .
+ since logic programming is not only a theoretical framework , but also a well developed programming style , the connections between fusion , hshr and logic programming can be used for implementation purposes .
slp has been implemented in through meta - interpretation .
thus we can use translations from fusion and hshr to implement them .
in particular , since implementations of logic programming are not distributed , this can be useful mainly for simulation purposes .
+ in section [ section : background ] we present the required background , in particular we introduce the fusion calculus ( [ subsection : fusion ] ) , the algebraic representation of graphs and the hshr ( [ subsection : shr ] ) , and logic programming ( [ subsection : logicprog ] ) .
section [ section : fusion2shr ] is dedicated to the mapping from fusion calculus to hshr .
section [ section : hshr2lp ] analyzes the relationships between hshr and logic programming , in particular we introduce slp ( [ subsection : slp ] ) , we prove the correspondence between it and hshr ( [ subsection : hshr2lp ] ) and we give some hints on how to implement fusion calculus and hshr using prolog ( [ subsection : meta ] ) . in section [ section :
conclusion ] we present some conclusions and traces for future work .
finally , proofs and technical lemmas are in [ appendix : proofs ] .
[ [ mathematical - notation . ] ] mathematical notation .
+ + + + + + + + + + + + + + + + + + + + + + we use @xmath1 to denote the application of substitution @xmath2 to @xmath3 ( where @xmath3 can be a term or a set / vector of terms ) .
we write substitutions as sets of pairs of the form @xmath4 , denoting that variable @xmath5 is replaced by term @xmath6 .
we also denote with @xmath7 the composition of substitutions @xmath8 and @xmath9 . we denote with @xmath10 the set of elements mapped to @xmath5 by @xmath2 .
we use @xmath11 to denote the operation that computes the number of elements in a set / vector . given a function @xmath12
we denote with @xmath13 its domain , with @xmath14 its image and with @xmath15 the restriction of @xmath12 to the new domain @xmath16 .
we use on functions and substitutions set theoretic operations ( such as @xmath17 ) referring to their representation as sets of pairs .
similarly , we apply them to vectors , referring to the set of the elements in the vector . in particular , @xmath18 is set difference . given a set @xmath16 we denote with @xmath19 the set of strings on @xmath16 . also , given a vector @xmath20 and an integer @xmath21 ,
@xmath22 $ ] is the @xmath21-th element of @xmath23 .
finally , a vector is given by listing its elements inside angle brackets @xmath24 .
the fusion calculus @xcite is a calculus for modelling distributed and mobile systems which is based on the concepts of _ fusion _ and scope .
it is an evolution of the @xmath0-calculus @xcite and the interesting point is that it is obtained by simplifying the calculus .
in fact the two action prefixes for input and output communication are symmetric , whereas in the @xmath0-calculus they are not , and there is just one binding operator called scope , whereas the @xmath0-calculus has two ( restriction and input ) . as shown in , the @xmath0-calculus is syntactically a subcalculus of the fusion calculus ( the key point is that the input of @xmath0-calculus is obtained using input and scope ) . in order to have these properties fusion actions have to be introduced .
an asynchronous version of fusion calculus is described in , , where name fusions are handled explicitly as messages . here
we follow the approach by parrow and victor .
+ we now present in details the syntax and the reduction semantics of fusion calculus . in our work
we deal with a subcalculus of the fusion calculus , which has no match and no mismatch operators , and has only guarded summation and recursion .
all these restrictions are quite standard , apart from the one concerning the match operator , which is needed to have an expansion lemma . to extend our approach to deal with match
we would need to extend shr by allowing production applications to be tagged with a unique identifier .
we leave this extension for future work . in our discussion
we distinguish between _ sequential processes _ ( which have a guarded summation as topmost operator ) and general processes .
+ we assume to have an infinite set @xmath25 of names ranged over by @xmath26 and an infinite set of agent variables ( disjoint w.r.t .
the set of names ) with meta - variable @xmath27 .
names represent communication channels .
we use @xmath28 to denote an equivalence relation on @xmath25 , called fusion , which is represented in the syntax by a finite set of equalities .
function @xmath29 returns all names which are fused , i.e. those contained in an equivalence class of @xmath28 which is not a singleton .
the prefixes are defined by : @xmath30 the agents are defined by : @xmath31 @xmath32 the scope restriction operator is a binder for names , thus @xmath5 is bound in @xmath33 .
similarly @xmath34 is a binder for agent variables
. we will only consider agents which are closed w.r.t .
both names and agent variables and where in @xmath35 each occurrence of @xmath27 in @xmath36 is within a sequential agent ( guarded recursion ) .
we use recursion to define infinite processes instead of other operators ( e.g. replication ) since it simplifies the mapping and since their expressive power is essentially the same .
we use infix @xmath37 for binary sum ( which thus is associative and commutative ) .
+ given an agent @xmath36 , functions @xmath38 , @xmath39 and @xmath40 compute the sets @xmath41 , @xmath42 and @xmath43 of its free , bound and all names respectively .
+ processes are agents considered up to structural axioms defined as follows .
[ defin : congrfusion ] the structural congruence @xmath44 between agents is the least congruence satisfying the @xmath45-conversion law ( both for names and for agent variables ) , the abelian monoid laws for composition ( associativity , commutativity and 0 as identity ) , the scope laws @xmath46 , @xmath47 , the scope extrusion law @xmath48 where @xmath49 and the recursion law @xmath50 .
note that @xmath38 is also well - defined on processes . in order to deal with fusions
we need the following definition . a substitutive effect of a fusion
@xmath28 is any idempotent substitution @xmath51 having @xmath28 as its kernel .
in other words @xmath52 iff @xmath53 and @xmath2 sends all members of each equivalence class of @xmath28 to one representative in the class is a most general unifier of @xmath28 , when it is considered as a set of equations . ] .
the reduction semantics for fusion calculus is the least relation satisfying the following rules .
[ defin : redfusion ] @xmath54 where @xmath55 and @xmath2 is a substitutive effect of @xmath56 such that @xmath57 .
@xmath58 where @xmath2 is a substitutive effect of @xmath28 such that @xmath57 .
@xmath59 synchronized hyperedge replacement ( shr ) @xcite is an approach to ( hyper)graph transformations that defines _ global transitions _ using _ local productions_. productions define how a single ( hyper)edge can be rewritten and the _ conditions _ that this rewriting imposes on adjacent nodes .
thus the global transition is obtained by applying in parallel different productions whose conditions are compatible .
what exactly compatible means depends on which _ synchronization model _ we use . in this work
we will use the hoare synchronization model ( hshr ) , which requires that all the edges connected to a node expose the same action on it . for a general definition of synchronization models see .
+ we use the extension of hshr with _ mobility _
@xcite , that allows edges to expose node references together with actions , and nodes whose references are matched during synchronization are unified .
+ we will give a formal description of hshr as labelled transition system , but first of all we need an algebraic representation for graphs .
an edge is an atomic item with a label and with as many ordered tentacles as the rank @xmath60 of its label @xmath61 .
a set of nodes , together with a set of such edges , forms a graph if each edge is connected , by its tentacles , to its attachment nodes .
we will consider graphs up to isomorphisms that preserve nodes , labels of edges , and connections between edges and nodes .
+ now , we present a definition of graphs as syntactic judgements , where nodes correspond to names and edges to basic terms of the form @xmath62 , where the @xmath63 are arbitrary names and @xmath64 . also , @xmath65 represents the empty graph and @xmath66 is the parallel composition of graphs ( merging nodes with the same name ) .
let @xmath25 be a fixed infinite set of names and @xmath67 a ranked alphabet of labels .
a syntactic judgement ( or simply a judgement ) is of the form @xmath68 where : 1 .
@xmath69 is the ( finite ) set of nodes in the graph .
2 . @xmath70 is a term generated by the grammar + @xmath71 + where @xmath72 is a vector of names and @xmath61 is an edge label with @xmath73 .
we denote with @xmath40 the function that given a graph @xmath70 returns the set @xmath74 of all the names in @xmath70 .
we use the notation @xmath75 to denote the set obtained by adding @xmath5 to @xmath76 , assuming @xmath77 .
similarly , we write @xmath78 to state that the resulting set of names is the disjoint union of @xmath79 and @xmath80 .
[ defin : structural ] the structural congruence @xmath44 on terms @xmath70 obeys the following axioms : @xmath81 @xmath82 @xmath83 the well - formed judgements @xmath68 over @xmath67 and @xmath25 are those where @xmath84 .
axioms ( ag1),(ag2 ) and ( ag3 ) define respectively the associativity , commutativity and identity over @xmath65 for operation @xmath66 .
+ well - formed judgements up to structural axioms are isomorphic to graphs up to isomorphisms . for a formal statement of the correspondence see .
we will now present the steps of a shr computation .
[ defin : shrtrans ] let @xmath85 be a set of actions . for each action @xmath86 ,
let @xmath87 be its arity . + a shr transition is of the form : @xmath88{\lambda,\pi } \phi \vdash g'\ ] ] where @xmath68 and @xmath89 are well - formed judgements for graphs , @xmath90 is a total function and @xmath91 is an idempotent substitution .
function @xmath92 assigns to each node @xmath5 the action @xmath93 and the vector @xmath94 of node references exposed on @xmath5 by the transition .
if @xmath95 then we define @xmath96 and @xmath97 .
we require that @xmath98 , namely the arity of the action must equal the length of the vector .
+ we define : * @xmath99 + set of exposed names ; * @xmath100 + set of fresh names that are exposed ; * @xmath101 + set of fused names .
substitution @xmath0 allows to merge nodes .
since @xmath0 is idempotent , it maps every node into a standard representative of its equivalence class .
we require that @xmath102 , i.e. only references to representatives can be exposed .
furthermore we require @xmath103 , namely nodes are never erased .
nodes in @xmath104 are fresh internal nodes , silently created in the transition .
we require that no isolate , internal nodes are created , namely @xmath105 .
note that the set of names @xmath106 of the resulting graph is fully determined by @xmath76 , @xmath92 , @xmath0 and @xmath107 thus we will have no need to write its definition explicitly in the inference rules .
notice also that we can write a shr transition as : @xmath88{\lambda,\pi } \gamma\pi,\gamma_{\lambda},\gamma_{int } \vdash g'.\ ] ] we usually assume to have an action @xmath108 of arity @xmath109 to denote `` no synchronization '' .
we may not write explicitly @xmath0 if it is the identity , and some actions if they are @xmath110 . furthermore we use @xmath111 to denote the function that assigns @xmath110 to each node in @xmath76 ( note that the dependence on @xmath76 is implicit ) .
we derive shr transitions from basic productions using a set of inference rules .
productions define the behaviour of single edges .
[ defin : production ] a production is a shr transition of the form : @xmath112{\lambda,\pi } \phi \vdash g\ ] ] where all @xmath63 , @xmath113 are distinct .
+ productions are considered as schemas and so they are @xmath45-convertible w.r.t
. names in @xmath114 .
we will now present the set of inference rules for hoare synchronization .
the intuitive idea of hoare synchronization is that all the edges connected to a node must expose the same action on that node .
[ defin : hoarerules ] @xmath115{\lambda , \pi } \phi \vdash g_2 \quad \gamma ' \vdash g'_1 \xrightarrow[]{\lambda',\pi ' } \phi ' \vdash g'_2 \quad ( \gamma \cup \phi ) \cap ( \gamma ' \cup \phi')=\emptyset}{\gamma , \gamma ' \vdash g_1|g'_1 \xrightarrow[]{\lambda \cup \lambda ' ,
\pi \cup \pi ' } \phi , \phi ' \vdash g_2|g'_2}\ ] ] @xmath116{\lambda , \pi } \phi \vdash g_2 \quad \forall x , y \in \gamma .
x\sigma = y\sigma \wedge x \neq y \rightarrow
\operatorname{act}_{\lambda}(x ) = \operatorname{act}_{\lambda}(y)}{\gamma\sigma \vdash g_1\sigma \xrightarrow[]{\lambda ' , \pi ' } \phi ' \vdash g_2\sigma\rho}\ ] ] where @xmath117 is an idempotent substitution and : 1 .
[ merge : rho ] @xmath118 whenever possible 2 .
[ merge : lambda ] @xmath119 3 .
[ merge : pi ] @xmath120 @xmath121{\lambda_{\epsilon},id } \gamma \vdash g\ ] ] @xmath122{\lambda,\pi } \phi \vdash g_2 \quad x \notin
\gamma \quad \vec y \cap ( \gamma \cup \phi \cup \{x\})= \emptyset}{\gamma , x \vdash g_1 \xrightarrow[]{\lambda \cup \{(x , a,\vec y)\},\pi } \phi ' \vdash g_2}\ ] ] a transition is obtained by composing productions , which are first applied on disconnected edges , and then by connecting the edges by merging nodes . in particular rule ( par ) deals with the composition of transitions which have disjoint sets of nodes and rule ( merge ) allows to merge nodes ( note that @xmath2 is a projection into representatives of equivalence classes ) .
the side condition requires that we have the same action on merged nodes .
definition [ merge : rho ] introduces the most general unifier @xmath123 of the union of two sets of equations : the first set identifies ( the representatives of ) the tuples associated to nodes merged by @xmath2 , while the second set of equations is just the kernel of @xmath0 .
thus @xmath123 is the merge resulting from both @xmath0 and @xmath2 .
note that [ merge : lambda ] @xmath92 is updated with these merges and that [ merge : pi ] @xmath124 is @xmath123 restricted to the nodes of the graph which is the source of the transition .
rule ( idle ) guarantees that each edge can always make an explicit idle step .
rule ( new ) allows adding to the source graph an isolated node where arbitrary actions ( with fresh names ) are exposed .
+ we write @xmath125{\lambda,\pi } \phi \vdash g')$ ] if @xmath126{\lambda,\pi } \phi \vdash g'$ ] can be obtained from the productions in @xmath127 using hoare inference rules . + we will now present an example of hshr computation .
we show now how to use hshr to derive a 4 elements ring starting from a one element ring , and how we can then specify a reconfiguration that transforms the ring into the star graph in figure [ figure : star4 ] .
we use the following productions : @xmath128{(x , \epsilon , \langle \rangle ) , ( y , \epsilon , \langle \rangle ) } x , y , z \vdash c(x , z)|c(z , y)\ ] ] @xmath128{(x , r , \langle w \rangle ) , ( y , r , \langle w \rangle ) } x , y , w \vdash s(y , w)\ ] ] that are graphically represented in figure [ figure : prod ] .
notice that @xmath92 is represented by decorating every node @xmath5 in the left hand with @xmath129 and @xmath130 .
the first rule allows to create rings , in fact we can create all rings with computations like : @xmath131 in order to perform the reconfiguration into a star we need rules with nontrivial actions , like the second one .
this allows to do : @xmath132{(x , r , \langle w \rangle),(y , r , \langle w \rangle),(z , r , \langle w \rangle),(v , r , \langle w \rangle)}\\ \rightarrow x , y , z , v , w \vdash s(x , w)|s(y , w)|s(z , w)|s(v , w)\end{gathered}\ ] ] note that if an edge @xmath133 is rewritten into an edge @xmath16 , then all the edges in the ring must use the same production , since they must synchronize via action @xmath134 .
they must agree also on @xmath130 for every @xmath5 , thus all the newly created nodes are merged .
the whole transition is represented in figure [ figure : reconf ] .
it is easy to show that if we can derive a transition @xmath3 , then we can also derive every transition obtainable from @xmath3 by applying an injective renaming .
[ lemma : injren ] let @xmath127 be a set of productions and @xmath2 an injective substitution .
+ @xmath125{\lambda,\pi } \phi \vdash g')$ ] iff : + @xmath135{\lambda',\pi ' } \phi\sigma \vdash g'\sigma)$ ] + where @xmath136 and @xmath137 . by rule induction . in this paper
we are not interested in logic computations as refutations of goals for problem solving or artificial intelligence , but we consider logic programming @xcite as a _ goal rewriting mechanism_. we can consider logic subgoals as concurrent communicating processes that evolve according to the rules defined by the clauses and that use _ unification _ as the fundamental interaction primitive .
a presentation of this kind of use of logic programming can be found in .
+ in order to stress the similarities between logic programming and process calculi we present a semantics of logic programming based on a labelled transition system .
[ defin : plgrammar ] we have for clauses ( @xmath133 ) and goals ( @xmath70 ) the following grammar : @xmath138 @xmath139 where @xmath140 is a logic atom , `` , '' is the and conjunction and @xmath141 is the empty goal .
+ we can assume `` , '' to be associative and commutative and with unit @xmath141 .
the informal semantics of @xmath142 is `` for every assignment of the variables , if @xmath143 are all true , then @xmath140 is true '' .
a logic program is a set of clauses .
derivations in logic programming are called sld - derivations ( from `` linear resolution for definite clauses with selection function '' ) .
we will also consider partial sld - derivations .
[ defin : sld ] let @xmath36 be a logic program . + we define a step of a sld - resolution computation using the following rules : + @xmath144{\theta } ( b_1,\dots , b_k)\rho\theta } \qquad \textrm{atomic goal}\ ] ] where @xmath123 is an injective renaming of variables such that all the variables in the clause variant @xmath145 are fresh .
@xmath146{\theta } f}{p
\vdash g , g ' \xrightarrow[]{\theta } f , g'\theta } \qquad \textrm{conjunctive goal}\ ] ] we will omit @xmath147 if @xmath36 is clear from the context . + a partial sld - derivation of @xmath148 is a sequence ( possibly empty ) of steps of sld - resolution allowed by program @xmath36 with initial goal @xmath70 .
in this section we present a mapping from fusion calculus to hshr . +
this mapping is quite complex since there are many differences between the two formalisms .
first of all we need to bridge the gap between a process calculus and a graph transformation formalism , and this is done by associating edges to sequential processes and by connecting them according to the structure of the system .
moreover we need to map milner synchronization , which is used in fusion calculus , into hoare synchronization . in order to do this
we define some connection structures that we call amoeboids which implement milner synchronization using hoare connectors .
since hoare synchronization involves all the edges attached to a node while milner one involves just pairs of connectors , we use amoeboids to force each node to be shared by exactly two edges ( one if the node is an interface to the outside ) since in that case the behaviour of hoare and milner synchronization is similar .
an amoeboid is essentially a router ( with no path - selection abilities ) that connects an action with the corresponding coaction .
this is possible since in hshr an edge can do many synchronizations on different nodes at the same time
. finally , some restrictions have to be imposed on hshr in order to have an interleaving behaviour as required by fusion calculus .
+ we define the translation on processes in the form @xmath149 where @xmath36 is the parallel composition of sequential processes .
notice that every process can be reduced to the above form by applying the structural axioms : recursive definitions which are not inside a sequential agent have to be unfolded once and scope operators which are not inside a sequential agent must be taken to the outside .
we define the translation also in the case @xmath149 is not closed w.r.t . names ( but it must be closed w.r.t .
process variables ) since this case is needed for defining productions .
+ in the form @xmath149 we assume that the ordering of names in @xmath150 is fixed , dictated by some structural condition on their occurrences in @xmath36 . + for our purposes , it is also convenient to express process @xmath36 in @xmath151 as @xmath152 , where @xmath153 is a linear agent , i.e. every name in it appears once .
we assume that the free names of @xmath153 are fresh , namely @xmath154 , and again structurally ordered .
the corresponding vector is called @xmath155 . + the decomposition @xmath156 highlights the role of amoeboids .
in fact , in the translation , substitution @xmath2 is made concrete by a graph consisting of amoeboids , which implement a router for every name in @xmath41 .
more precisely , we assume the existence of edge labels @xmath157 and @xmath158 of ranks @xmath159 and @xmath160 respectively .
edges labelled by @xmath157 implement routers among @xmath21 nodes , while @xmath158 edges `` close '' restricted names @xmath72 in @xmath161 .
+ finally , linear sequential processes @xmath16 in @xmath153 must also be given a standard form .
in fact , they will be modelled in the hshr translation by edges labelled by @xmath162 , namely by a label encapsulating @xmath16 itself .
however in the derivatives of a recursive process the same sequential process can appear with different names an unbound number of times . to make the number of labels ( and also of productions , as we will see in short ) finite , for every given process
, we choose standard names @xmath163 and order them structurally : + @xmath164 with @xmath165 implying @xmath166 and @xmath167 .
+ we can now define the translation from fusion calculus to hshr .
the translation is parametrized by the nodes in the vectors @xmath20 and @xmath168 we choose to represent the names in @xmath72 and @xmath155 .
we denote with @xmath169}_{x \in s } g_x$ ] the parallel composition of graphs @xmath170 for each @xmath171 .
[ defin : fc2hshr ] @xmath172}_{x \in \vec v } n(x)$ ] + where : + @xmath173 , + @xmath174 , + @xmath175 + and with : + @xmath176 .
+ @xmath177 + @xmath178 with @xmath179 + @xmath180 + @xmath181}_{x \in \operatorname{im}(\sigma ) } m_{k+1}(x,\sigma^{-1}(x))$ ] where @xmath182 . in the above translation
, graph @xmath183 consists of a set of disconnected edges , one for each sequential process of @xmath184 .
the translation produces a graph with three kinds of nodes .
the nodes of the first kind are those in @xmath168 .
each of them is adjacent to exactly two edges , one representing a sequential process of @xmath153 , and the other an amoeboid .
also the nodes in @xmath20 are adjacent to two edges , an amoeboid and an @xmath158 edge
. finally the nodes in @xmath185 are adjacent only to an amoeboid . +
as mentioned above , translation @xmath186 builds an amoeboid for every free name @xmath5 of @xmath187 : it has @xmath188 tentacles , where @xmath189 are the occurrences of @xmath5 in @xmath187 , namely the free names of @xmath153 mapped to it .
notice that the choice of the order within @xmath10 is immaterial , since we will see that amoeboids are commutative w.r.t .
their tentacles .
however , to make the translation deterministic , @xmath10 could be ordered according to some fixed precedence of the names .
let @xmath190 .
the translation of @xmath2 is in figure [ figure : amoeboids ] .
[ ex : processtr ] let us consider the ( closed ) process @xmath191 .
we can write it in the form @xmath192 as : + @xmath193 + furthermore we can decompose @xmath36 into @xmath187 where : + @xmath194 + @xmath195 . + we can now perform the translation .
+ we choose @xmath196 and @xmath197 : @xmath198 now we define the productions used in the hshr system .
+ we have two kinds of productions : auxiliary productions that are applied to amoeboid edges and process productions that are applied to process edges .
+ before showing process productions we need to present the translation from fusion calculus prefixes into hshr transition labels .
[ defin : acttransl ] the translation from fusion calculus prefixes into hshr transition labels is the following : + @xmath199 where + if @xmath200 then @xmath201 , @xmath202 if @xmath203 with @xmath204 , @xmath205 + if @xmath206 then @xmath207 , @xmath202 if @xmath203 and @xmath204 , @xmath205 + if @xmath208 then @xmath209 and @xmath0 is any substitutive effect of @xmath28 .
+ we will write @xmath210 and @xmath211 as @xmath212 and @xmath213 respectively . [ defin : processprod ] we have a process production for each prefix at the top level of a linear standard sequential process ( which has @xmath214 as free names ) .
let @xmath215 be such a process .
its productions can be derived with the following inference rule : @xmath216{{\llbracket \alpha_j \rrbracket } } \gamma,\gamma ' \vdash g|{\llbracket \xi \rrbracket}|{\llbracket \pi \rrbracket}|\operatorname{\big[\!\!\big]}_{x \in \gamma '' } n(x)}\ ] ] if @xmath217 , @xmath214 and @xmath218 are pairwise disjoint with @xmath219 injective renaming from @xmath220 to fresh names , @xmath221 and @xmath222 .
we add some explanations on the derivable productions .
essentially , if @xmath223 is a possible choice , the edge labelled by the process can have a transition labelled by @xmath224 to something related to @xmath225 .
we use @xmath226 instead of @xmath227 ( and then we add the translation of @xmath219 ) to preserve the parity of the number of amoeboid edges on each path ( see definition [ defin : structured ] ) .
the parameter @xmath20 of the translation contains fresh nodes for restricted names that are taken to the top level during the normalization of @xmath227 while @xmath168 contains the free names in the normalization of @xmath227 ( note that some of them may be duplicated w.r.t .
@xmath227 , if this one contains recursion ) .
if @xmath228 is a fusion @xmath28 , according to the semantics of the calculus , a substitutive effect @xmath0 of it should be applied to @xmath227 , and this is obtained by adding the amoeboids @xmath229 in parallel .
furthermore , @xmath68 must be enriched in other two ways : since nodes can never be erased , nodes which are present in the sequential process , i.e. the nodes in @xmath230 , must be added to @xmath76 .
also `` close '' @xmath158 edges must be associated to forgotten nodes ( to forbid further transitions on them and to have them connected to exactly two edges in the result of the transition ) , provided they are not exposed , i.e. to nodes in @xmath231 .
+ note that when translating the rhs @xmath232 of productions we may have names in @xmath233 which occur just once .
since they are renamed by @xmath2 and @xmath219 , they will produce in the translation some chains of @xmath234 connectors of even length , which , as we will see shortly , are behaviourally equivalent to simple nodes . for simplicity , in the examples we will use the equivalent productions where these connectors have been removed and the nodes connected by them have been merged .
[ ex : productions ] let us consider firstly the simple agent @xmath235 .
+ the only production for this agent ( where @xmath236 ) is : + @xmath237{(x_1,out_1,\langle x_2 \rangle ) } x_1,x_2 \vdash n(x_1)$ ] + where we closed node @xmath238 but not node @xmath239 since the second one is exposed on @xmath238 .
+ let us consider a more complex example : + @xmath240 .
+ the process @xmath241 where @xmath242 can be transformed into : + @xmath243 + where @xmath244 . +
its translation ( with @xmath245 and @xmath246 ) is : @xmath247 thus the production is : @xmath248{(x_1,in_1,\langle x_2 \rangle)}\\ x_1,x_2,x_3,x_4,x_5,x_6,y_3,y_4,y_5,y_6,y_7,y_8,y , x'_5,x'_6 \vdash\\ l_{\overline{x_1}x_2.0}(x_3,x_4)| l_{x_1x_2.(\overline{x_3}x_4.0|\operatorname{rec}x .
( ( x ) x_5x.(\overline{x_6}x.0|x)))}(y_3,y_4,y_5,y_6,y_7,y_8)|\\ m_3(x'_5,y_3,y_7)|m_3(x'_6,y_5,y_8)|m_3(y , y_4,y_6)|n(y)|m_2(x_5,x'_5)|m_2(x_6,x'_6)|n(x_1)\end{gathered}\ ] ] where for simplicity we collapsed @xmath249 with @xmath250 and @xmath251 with @xmath252 .
we will now show the productions for amoeboids .
we have auxiliary productions of the form : @xmath253{(x_1,in_n,\vec y_1),(x_2,out_n,\vec y_2 ) } \gamma,\vec y_1,\vec y_2 \vdash m_k(\gamma ) we need such a production for each @xmath189 and @xmath158 and each pair of nodes @xmath238 and @xmath239 in @xmath76 where @xmath76 is a chosen tuple of distinct names with @xmath189 components and @xmath254 and @xmath255 are two vectors of fresh names such that @xmath256 .
+ note that we also have the analogous production where @xmath238 and @xmath239 are swapped .
in particular , the set of productions for a @xmath257 edge is invariant w.r.t .
permutations of the tentacles , modelling the fact that its tentacles are essentially unordered .
+ we have no productions for edges labelled with @xmath158 , which thus forbid any synchronization .
the notion of amoeboid introduced previously is not sufficient for our purposes .
in fact , existing amoeboids can be connected using @xmath234 edges and nodes that are no more used can be closed using @xmath158 edges .
thus we present a more general definition of amoeboid for a set of nodes and we show that , in the situations of interest , these amoeboids behave exactly as the simpler @xmath157 edges .
[ defin : structured ] given a vector of nodes @xmath258 , a structured amoeboid @xmath259 for the set of nodes @xmath16 containing all the nodes in @xmath258 is any connected graph composed by @xmath260 and @xmath158 edges that satisfies the following properties : * its set of nodes is of the form @xmath261 , with @xmath262 ; * nodes in @xmath16 are connected to exactly one edge of the amoeboid ; * nodes in @xmath263 are connected to exactly two edges of the amoeboid ; * the number of edges composing each path connecting two distinct nodes of @xmath16 is odd .
nodes in @xmath16 are called _ external _ , nodes in @xmath263 are called _
internal_. we consider _ equivalent _ all the amoeboids with the same set @xmath16 of external nodes .
the last condition is required since each connector inverts the polarity of the synchronization , and we want amoeboids to invert it .
+ note that @xmath264 is an amoeboid for @xmath16 .
[ lemma : amoeboid ] if @xmath259 is a structured amoeboid for s , the transitions for @xmath259 which are non idle and expose non @xmath265 actions on at most two nodes @xmath266 are of the form : @xmath267{\lambda , id } s , i , i',\vec y_1,\vec y_2 \vdash m(\vec s ) |
|\vec y_1| } m(\vec y_1[i],\vec y_2[i ] ) | \operatorname{\big[\!\!\big]}\tilde m(\emptyset)\ ] ] where @xmath268 and @xmath269 ( non trivial actions may be exposed also on some internal nodes ) and @xmath254 and @xmath255 are two vectors of fresh names such that @xmath256 . here
@xmath270 contains rings of @xmath234 connectors connected only to fresh nodes which thus are disconnected from the rest of the graph .
we call them _ pseudoamoeboids_. furthermore we have at least one transition of this kind for each choice of @xmath238 , @xmath239 , @xmath254 and @xmath271 .
see [ appendix : proofs ] .
thanks to the above result we will refer to structured amoeboids simply as amoeboids . we can now present the results on the correctness and completeness of our translation . [
theorem : hoarecorrectness ] for each closed fusion process @xmath36 and each pair of vectors @xmath20 and @xmath168 satisfying the constraints of definition [ defin : fc2hshr ] , if @xmath272 then there exist @xmath92 , @xmath76 and @xmath70 such that @xmath273{\lambda , id } \gamma \vdash g$ ] .
furthermore @xmath274 is equal to @xmath275 ( for some @xmath276 and @xmath277 ) up to isolated nodes , up to injective renamings , up to equivalence of amoeboids ( @xmath68 can have a structured amoeboid where @xmath275 has a simple one ) and up to pseudoamoeboids .
the proof is by rule induction on the reduction semantics .
+ see [ appendix : proofs ] .
[ theorem : hoarecompleteness ] for each closed fusion process @xmath36 and each pair of vectors @xmath20 and @xmath168 if @xmath273{\lambda,\pi } \gamma \vdash g$ ] with a hshr transition that uses exactly two productions for communication or one production for a fusion action ( plus any number of auxiliary productions ) then @xmath278 and @xmath68 is equal to @xmath279 ( for some @xmath276 and @xmath277 ) up to isolate nodes , up to injective renamings , up to equivalence of amoeboids ( @xmath68 can have a structured amoeboid where @xmath275 has a simple one ) and up to pseudoamoeboids .
see [ appendix : proofs ] .
these two theorems prove that the allowed transitions in the hshr setting correspond to reductions in the fusion calculus setting .
note that in hshr we must consider only transitions where we have either two productions for communication or one production for a fusion action .
this is necessary to model the interleaving behaviour of fusion calculus within the hshr formalism , which is concurrent . on the contrary
, one can consider the fusion equivalent of all the hshr transitions : these correspond to concurrent executions of many fusion reductions .
one can give a semantics for fusion calculus with that behaviour .
anyway in that case the notion of equivalence of amoeboids is no more valid , since different amoeboids allow different degrees of concurrency .
we thus need to constrain them .
the simplest case is to have only simple amoeboids , that is to have no concurrency inside a single channel , but there is no way to force normalization of amoeboids to happen before undesired transitions can occur .
the opposite case ( all the processes can interact in pairs , also on the same channel ) can be realized , but it requires more complex auxiliary productions .
+ note that the differences between the final graph of a transition and the translation of the final process of a fusion calculus reduction are not important , since the two graphs have essentially the same behaviours ( see lemma [ lemma : injren ] for the effect of an injective renaming and lemma [ lemma : amoeboid ] for the characterization of the behaviour of a complex amoeboid ; isolated nodes and pseudoamoeboids are not relevant since different connected components evolve independently ) .
thus the previous results can be extended from transitions to whole computations .
+ note that in the hshr model the behavioural part of the system is represented by productions while the topological part is represented by graphs .
thus we have a convenient separation between the two different aspects .
[ ex : fusion2hoare ] we will now show an example of the translation . let us consider the process : @xmath280 note that it is already in the form @xmath149 .
it can do the following transition : @xmath281 we can write @xmath36 in the form : + @xmath282 + where : + @xmath283 . + a translation of the starting process is : @xmath284 a graphical representation is in figure [ figure : process1 ] .
+ we have the following process productions : @xmath285{(y_1,out_2,\langle y_2,y_3 \rangle)}\\ y_1,y_2,y_3,y_4,y_5 \vdash l_{r(x_1,x_2)}(y_4,y_5)|n(y_1)\end{gathered}\ ] ] @xmath286{(y_1,in_2,\langle y_2,y_3 \rangle)}\\ y_1,y_2,y_3,y_4,y_5 \vdash l_{s(x_1,x_2)}(y_4,y_5)|n(y_1)\end{gathered}\ ] ] in order to apply ( suitable variants of ) these two productions concurrently we have to synchronize their actions .
this can be done since in the actual transition actions are exposed on nodes @xmath287 and @xmath288 respectively , which are connected to the same @xmath289 edge .
thus the synchronization can be performed ( see figure [ figure : process2 ] ) and we obtain as final graph : @xmath290 which is represented in figure [ figure : process3 ] .
+ the amoeboids connect the following tuples of nodes : + @xmath291 , @xmath292 , @xmath293 .
thus , if we connect these sets of nodes with simple amoeboids instead of with complex ones , we have up to injective renamings a translation of @xmath294 as required
. we will show here an example that uses recursion .
let us consider the closed process @xmath295 .
the translation of this process , as shown in example [ ex : processtr ] is : @xmath296 we need the productions for two sequential edges ( for the first step ) : @xmath235 and @xmath297 .
+ the productions are the ones of example [ ex : productions ] ( we write them here in a suitable @xmath45-converted form ) : @xmath298{(u_1,out_1,\langle
z_1 \rangle ) } u_1,z_1 \vdash n(u_1)\hspace{2cm}\end{gathered}\ ] ] @xmath299{u_2,in_1,\langle y_1 \rangle}\\ u_2,y_1,u_3,y_2,u_4,u_5,w_1,w_2,w_3,w_4,w_5,w_6,y',u'_4,u'_5 \vdash\\ l_{\overline{x_1}x_2.0}(u_3,y_2)| l_{x_1x_2.(\overline{x_3}x_4.0|\operatorname{rec}x . ( ( x ) x_5x.(\overline{x_6}x.0|x)))}(w_1,w_2,w_3,w_4,w_5,w_6)|\\ m_3(u'_4,w_1,w_5)|m_3(u'_5,w_3,w_6)|m_3(y',w_2,w_4)|n(y')|m_2(u_4,u'_4)|m_2(u_5,u'_5)|n(u_2)\end{gathered}\ ] ] by using these two productions and a production for @xmath300 ( the other edges stay idle ) we have the following transition : @xmath301{(u_1,out_1,\langle z_1 \rangle)(u_2,in_1,\langle y_1 \rangle)}\\ u , y , z , x , u_1,z_1,u_2,y_1,u_3,y_2,u_4,u_5,w_1,w_2,w_3,w_4,w_5,w_6,y',u'_4,u'_5 \vdash\\ n(u_1)|l_{\overline{x_1}x_2.0}(u_3,y_2)| l_{x_1x_2.(\overline{x_3}x_4.0|\operatorname{rec}x .
( ( x ) x_5x.(\overline{x_6}x.0|x)))}(w_1,w_2,w_3,w_4,w_5,w_6)|\\ m_3(u'_4,w_1,w_5)|m_3(u'_5,w_3,w_6)|m_3(y',w_2,w_4)|n(y')|m_2(u_4,u'_4)|m_2(u_5,u'_5)|n(u_2)|\\ m_6(u , u_1,u_2,u_3,u_4,u_5)|m_2(z_1,y_1)|m_2(z , z_1)|m_3(y , y_1,y_2)|n(u)|n(z)|n(y)\end{gathered}\ ] ] the resulting graph is , up to injective renaming and equivalence of amoeboids , a translation of : + @xmath302 + as required .
we end this section with a simple schema on the correspondence between the two models .
+ l|l|l|l fusion & hshr & fusion & hshr + closed process & graph & reduction & transition + sequential process & edge & name & amoeboid + prefix execution & production & 0 & @xmath303 + as shown in the table , we represent ( closed ) processes by graphs where edges are sequential processes and amoeboids model names .
the inactive process @xmath109 is the empty graph @xmath65 . from a dynamic point of view
, fusion reductions are modelled by hshr transitions obtained composing productions that represent prefix executions .
we will now present a mapping from hshr into a subset of logic programming called synchronized logic programming ( slp ) .
the idea is to compose this mapping with the previous one obtaining a mapping from fusion calculus into logic programming . in this subsection
we present synchronized logic programming .
+ slp has been introduced because logic programming allows for many execution strategies and for complex interactions .
essentially slp is obtained from standard logic programming by adding a mechanism of transactions .
the approach is similar to the zero - safe nets approach @xcite for petri nets .
in particular we consider that function symbols are resources that can be used only inside a transaction .
a transaction can thus end only when the goal contains just predicates and variables . during a transaction , which is called big - step in this setting
, each atom can be rewritten at most once .
if a transaction can not be terminated , then the computation is not allowed .
a computation is thus a sequence of big - steps .
this synchronized flavour of logic programming corresponds to hshr since : * used goals correspond to graphs ( goal - graphs ) ; * clauses in programs correspond to hshr productions ( synchronized clauses ) ; * resulting computations model hshr computations ( synchronized computations ) .
[ defin : goalgraph ] we call goal - graph a goal which has no function symbols ( constants are considered as functions of arity @xmath109 ) . [ defin : syncprograms ] a synchronized program is a finite set of synchronized rules , i.e. definite program clauses such that : * the body of each rule is a goal - graph ; * the head of each rule is @xmath304 where @xmath305 is either a variable or a single function ( of arity at least @xmath160 ) symbol applied to variables .
if it is a variable then it also appears in the body of the clause .
[ ex : synchronized ] + [ cols="^ , < " , ] in the mapping , the transaction mechanism is used to model the synchronization of hshr , where edges can be rewritten only if the synchronization constraints are satisfied . in particular ,
a clause @xmath306 will represent a production where the head predicate @xmath140 is the label of the edge in the left hand side , and the body @xmath143 is the graph in the right hand side .
term @xmath305 in the head represents the action occurring in @xmath63 , if @xmath307 is the edge matched by the production .
intuitively , the first condition of definition [ defin : syncprograms ] says that the result of a local rewriting must be a goal - graph .
the second condition forbids synchronizations with structured actions , which are not allowed in hshr ( this would correspond to allow an action in a production to synchronize with a sequence of actions from a computation of an adjacent subgraph ) .
furthermore it imposes that we can not disconnect from a node without synchronizing on it .
action on the node we disconnect from and make sure that all the other edges can freely do the same action . for example we can rewrite @xmath308 as @xmath309 , which is an allowed synchronized rule .
an explicit translation of action @xmath265 can be used too . ]
now we will define the subset of computations we are interested in . given a synchronized program @xmath36
we write : + @xmath310 iff @xmath311{\theta'}\operatorname{\!\!\mbox{}^*\,\,}g_2 $ ] and all steps performed in the computation expand different atoms of @xmath312 , @xmath313 and both @xmath312 and @xmath314 are goal - graphs .
+ we call @xmath315 a big - step and all the @xmath316 steps in a big - step small - steps . +
a slp computation is : + @xmath317 i.e. a sequence of 0 or more big - steps .
we want to use slp to model hshr systems . as a first step
we need to translate graphs , i.e. syntactic judgements , to goals . in this translation ,
edge labels are mapped into slp predicates .
goals corresponding to graphs will have no function symbols .
however function symbols will be used to represent actions . in the translation
we will lose the context @xmath76 .
we define the translation operator @xmath318 as : @xmath319 @xmath320 @xmath321 sometimes we will omit the @xmath76 part of the syntactic judgement .
we can do this because it does not influence the translation . for simplicity
, we suppose that the set of nodes in the shr model coincides with the set of variables in slp ( otherwise we need a bijective translation function ) .
we do the same for edge labels and names of predicates , and for actions and function symbols .
let @xmath68 and @xmath322 be graphs .
we define the equivalence relation @xmath323 in the following way : @xmath324 iff @xmath325 .
observe that if two judgements are equivalent then they can be written as : + @xmath326 + @xmath327 + where @xmath328 .
the operator @xmath318 defines an isomorphism between judgements ( defined up to @xmath323 ) and goal - graphs .
the proof is straightforward observing that the operator @xmath318 defines a bijection between representatives of syntactic judgements and representatives of goal - graphs and the congruence on the two structures is essentially the same .
we now define the translation from hshr productions to definite clauses .
[ defin : tuostoruletranslation ] we define the translation operator @xmath318 as : @xmath329{\lambda,\pi } g \rrbracket } = l(a_1(x_1\pi,\vec y_1 ) , \dots , a_n(x_n\pi,\vec y_n ) ) \leftarrow { \llbracket g \rrbracket}\ ] ] if @xmath330 for each @xmath331 and if @xmath332 .
if @xmath333 we write simply @xmath334 instead of @xmath335 .
the idea of the translation is that the condition given by an action @xmath336 is represented by using the term @xmath337 as argument in the position that corresponds to @xmath5 .
notice that in this term @xmath93 is a function symbol and @xmath0 is a substitution . during unification ,
@xmath5 will be bound to that term and , when other instances of @xmath5 are met , the corresponding term must contain the same function symbol ( as required by hoare synchronization ) in order to be unifiable .
furthermore the corresponding tuples of transmitted nodes are unified .
since @xmath5 will disappear we need another variable to represent the node that corresponds to @xmath5 .
we use the first argument of @xmath93 to this purpose .
if two nodes are merged by @xmath0 then their successors are the same as required .
+ observe that we do not need to translate all the possible variants of the rules since variants with fresh variables are automatically built when the clauses are applied .
notice also that the clauses we obtain are synchronized clauses .
+ the observable substitution contains information on @xmath92 and @xmath0 .
thus given a transition we can associate to it a substitution @xmath338 .
we have different choices for @xmath338 according to where we map variables .
in fact in hshr nodes are mapped to their representatives according to @xmath0 , while , in slp , @xmath338 can not do the same , since the variables of the clause variant must be all fresh .
the possible choices of fresh names for the variables change by an injective renaming the result of the big - step .
[ defin : associated2 ] let @xmath126{\lambda,\pi } \phi \vdash g'$ ] be a transition .
we say that the substitution @xmath339 associated to this transition is : + @xmath340 + for some injective renaming @xmath123 .
we will now prove the correctness and the completeness of our translation .
[ theorem : tuostocorrectness ] let @xmath127 be a set of productions of a hshr system as defined in definitions [ defin : production ] and [ defin : hoarerules ] .
let @xmath36 be the logic program obtained by translating the productions in @xmath127 according to definition [ defin : tuostoruletranslation ] .
if : + @xmath125{\lambda,\pi } \phi \vdash g')$ ] + then we can have in @xmath36 a big - step of synchronized logic programming : + @xmath341 + for every @xmath123 such that @xmath342 is a fresh variable unless possibly when @xmath343 . in that case
we may have @xmath344 .
furthermore @xmath345 is associated to @xmath346{\lambda,\pi } \phi \vdash g'$ ] and @xmath347 .
finally , used productions translate into the clauses used in the big - step and are applied to the edges that translate into the predicates rewritten by them .
the proof is by rule induction .
+ see [ appendix : proofs ] . [
theorem : tuostocompleteness ] let @xmath127 be a set of productions of a hshr system .
let @xmath36 be the logic program obtained by translating the productions in @xmath127 according to definition [ defin : tuostoruletranslation ] .
if we have in @xmath36 a big - step of logic programming : + @xmath348 + then there exist @xmath123 , @xmath349 , @xmath92 , @xmath0 , @xmath106 and @xmath107 such that @xmath350 is associated to @xmath346{\lambda,\pi } \phi \vdash g'$ ] .
furthermore @xmath351 and @xmath352{\lambda,\pi } \phi \vdash g')$ ] .
see [ appendix : proofs ] .
[ ex : fusion2lp ] we continue here example [ ex : fusion2hoare ] by showing how that fusion computation can be translated into a synchronized logic programming computation .
@xmath281 remember that a translation of the starting process is : @xmath284 we have the following productions : @xmath285{(y_1,out_2,\langle y_2,y_3 \rangle)}\\ y_1,y_2,y_3,y_4,y_5 \vdash l_{r(x_1,x_2)}(y_4,y_5)|n(y_1)\end{gathered}\ ] ] @xmath286{(y_1,in_2,\langle y_2,y_3 \rangle)}\\ y_1,y_2,y_3,y_4,y_5 \vdash l_{s(x_1,x_2)}(y_4,y_5)|n(y_1)\end{gathered}\ ] ] that corresponds to the clauses ( we directly write suitably renamed variants ) : + @xmath353 + @xmath354 + plus the clause obtained from the auxiliary production : @xmath355 we obtain the big - step represented in figure [ figure : bigstep ] .
@xmath356{out_2(u'_1,x_2,y_2)/u_1,x_2/x'_2,y_2/y'_2,u_2/u'_2,x_3/x'_3}\\ l_{q(x_1,x_2,x_3)}(x_1,y_1,z_1),l_{r(x_1,x_2)}(u_2,x_3),n(u'_1 ) , l_{x_1x_2x_3.s(x_4,x_5)}(u_3,z_2,w_1,z_3,w_2),\\ m_4(u , out_2(u'_1,x_2,y_2),u_2,u_3),m_4(x , x_1,x_2,x_3),m_3(y , y_1,y_2),\\ m_4(z , z_1,z_2,z_3),m_3(w , w_1,w_2),n(u),n(x),n(y),n(w),n(z)\\ \xrightarrow[]{u / u'',u'_1/u''_1,x_2/x''_2,y_2/y''_2,u_2/u''_2,in_2(u''_3,z''_2,w''_1)/u_3}\\ l_{q(x_1,x_2,x_3)}(x_1,y_1,z_1),l_{r(x_1,x_2)}(u_2,x_3),n(u'_1),\\ l_{x_1x_2x_3.s(x_4,x_5)}(in_2(u''_3,z''_2,w''_1),z_2,w_1,z_3,w_2),\\ m_4(u , u'_1,u_2,u''_3),m_2(x_2,z''_2),m_2(y_2,w''_1),m_4(x , x_1,x_2,x_3),m_3(y , y_1,y_2),\\ m_4(z , z_1,z_2,z_3),m_3(w , w_1,w_2),n(u),n(x),n(y),n(w),n(z)\\ \xrightarrow[]{u''_3/u'''_3,z_2/z''_2,w_1/w''_1,z_2/z'''_2,w_1/w'''_1,z_3/z'''_3,w_2/w'''_2}\\ l_{q(x_1,x_2,x_3)}(x_1,y_1,z_1),l_{r(x_1,x_2)}(u_2,x_3),n(u'_1),l_{s(x_1,x_2)}(z_3,w_2),n(u''_3),\\ m_4(u , u'_1,u_2,u''_3),m_2(x_2,z_2),m_2(y_2,w_1),m_4(x , x_1,x_2,x_3),m_3(y , y_1,y_2),\\ m_4(z , z_1,z_2,z_3),m_3(w , w_1,w_2),n(u),n(x),n(y),n(w),n(z)\end{gathered}\ ] ] the observable substitution of the big - step is @xmath357 .
this is associated to the wanted hshr transition with @xmath358 and by applying @xmath123 to the final graph of the hshr transition we obtain : @xmath359 that , translated , becomes the final goal of the big - step as required .
we end this section with a simple schema on the correspondence between the two models .
+ l|l|l|l hshr & slp & hshr & slp + graph & goal & transition & big - step + edge & atomic goal & node & variable + parallel comp .
& and comp . &
@xmath303 & @xmath141 + production & clause & action & function s. + essentially the correspondence is given by the homomorphism between graphs and goals , with edges mapped to atomic goals , nodes to variables , parallel composition to and composition and @xmath65 to @xmath141 .
dynamically , hshr transitions are modelled by big - steps , that are transactional applications of clauses which model productions .
finally , hshr actions are modelled by function symbols .
the theorems seen in the previous sections can be used for implementation purposes . as far as synchronized logic programming is concerned , in a simple meta - interpreter
is presented .
the idea is to use prolog ability of dynamically changing the clause database to insert into it a set of clauses and a goal and to compute the possible synchronized computations of given length .
this can be directly used to simulate hshr transitions . in order to simulate fusion calculus processes we have to implement amoeboids using a bounded number of different connectors (
note that @xmath234 , @xmath360 and @xmath158 are enough ) and to implement in the meta - interpreter the condition under which productions can be applied in a single big - step .
this can be easily done .
furthermore this decreases the possible choices of applicable productions and thus improves the efficiency w.r.t . the general case .
in this paper we have analyzed the relationships between three different formalisms , namely fusion calculus , hshr and logic programming .
the correspondence between hshr and the chosen transactional version of logic programming , slp , is complete and quite natural .
thus we can consider hshr as a `` subcalculus '' of ( synchronized ) logic programming .
the mapping between fusion calculus and hshr is instead more involved because it has to deal with many important differences : * process calculi features vs graph transformation features ; * interleaving models vs concurrent models ; * milner synchronization vs hoare synchronization .
hoare synchronization was necessary since our aim was to eventually map fusion calculus to logic programming .
if the aim is just to compare fusion calculus and shr it is possible to use shr with milner synchronization , achieving a much simpler and complete mapping , which considers the lts of fusion calculus instead of reductions ( see ) .
+ we think that the present work can suggest several interesting lines of development , dictated by the comparison of the three formalisms studied in the paper .
first , our implementation of routers in terms of amoeboids is rather general and abstract , and shows that fusion calculus names are a rather high level concept .
they abstract out the behaviour of an underlying network of connections which must be open and reconfigurable .
had we chosen @xmath0-calculus instead ( see a translation of @xmath0-calculus to milner shr in @xcite ) , we would have noticed important differences .
for instance , fusions are also considered in the semantics of _ open _
@xmath0-calculus by davide sangiorgi @xcite , but in that work not all the names can be fused : newly extruded names can not be merged with previously generated names .
this is essential for specifying nonces and session keys for secure protocols .
instead , fusion calculus does not provide equivalent constructs .
looking at our translation , we can conclude that logic programming does not offer this feature , either .
thus logic programming is a suitable counterpart of fusion calculus , but it should be properly extended for matching open @xmath0-calculus and security applications . in a similar line of thought , we observe that we have a scope restriction operator in the fusion calculus , but no restriction is found in our version of hshr .
we think this omission simplifies our development , since no restriction exists in ordinary logic programming , either
. however versions of shr with restriction have been considered @xcite .
also ( synchronized ) logic programming can be smoothly extended with a restriction operator @xcite .
more importantly , fusion calculus is equipped with an observational abstract semantics based on ( hyper ) bisimulation .
we did not consider a similar concept for shr or logic programming , since we considered it outside the scope of the paper .
furthermore our operational correspondence between hshr and slp is very strong and it should respect any reasonable abstract semantics .
the mapping from fusion calculus into hshr deals only with closed terms , thus no observations can be considered .
however a bisimulation semantics of shr has been considered in @xcite , and an observational semantics of logic programming is discussed in @xcite .
another comment concerns concurrency . to prove the equivalence of fusion calculus and of its translation into hshr we had to restrict the possible computations of the latter . on the contrary , if all computations were allowed , the same translation would yield a concurrent semantics of fusion calculus , that we think is worth studying .
for instance in the presence of concurrent computations not all equivalent amoeboids would have the same behaviour , since some of them would allow for more parallelism than others .
finally we would like to emphasize some practical implication of our work .
in fact , logic programming is not only a model of computation , but also a well developed programming paradigm . following the lines of our translation ,
implementations of languages based on fusion calculus and hshr could be designed , allowing to exploit existing ideas , algorithms and tools developed for logic programming .
zero - safe nets : comparing the collective and individual token approaches .
_ 156 , _ 1 - 2 , 4689 . , montanari , u. , and rossi , f. 2001 .
an interactive semantics of logic programming .
_ 1 , _ 6 , 647690 .
\1987 . a model for distributed systems based on graph rewriting .
_ 34 , _ 2 , 411449 . ,
kreowski , h .- j . ,
montanari , u. , and rozenberg , g. , eds . 1999 . .
world scientific . ,
pfender , m. , and schneider , h. j. 1973 .
graph grammars : an algebraic approach . in _ proc . of ieee conference on automata and switching theory_. ieee computer society , 167180 . , montanari , u. , and tuosto , e. 2001 .
a lts semantics of ambients via graph synchronization with mobility . in _ proc . of ictcs01_. lncs ,
springer , 116 .
. explicit fusions . in _ mathematical foundations of computer science_. 373382 . \2004 .
strong bisimulation for the explicit fusion calculus . in _ proc . of fossacs04_. lncs ,
springer , 484498 .
graph transformation models for software architecture styles .
thesis , departamento de computacin , facultad de ciencias exactas y naturales , universidad de buenos aires , argentina .
, inverardi , p. , and montanari , u. 2000 .
reconfiguration of software architecture styles with name mobility . in _ proc . of coordination
2000_. lncs , vol .
springer , 148163 .
\2001 . synchronized hyperedge replacement with name mobility . in _ proc . of concur01_. lncs ,
springer , 121136 .
bigraphs and transitions .
_ 38 , _ 1 , 3849 .
observational equivalence for synchronized graph rewriting . in _ proc . of tacs01_. lncs ,
springer , 145164 .
process synchronization in distributed systems via horn clauses .
thesis , computer science department , university of pisa , pisa , italy .
downloadable from ` http://www.di.unipi.it/~lanese/tesi.ps ` .
synchronization strategies for global computing models . ph.d .
thesis , computer science department , university of pisa , pisa , italy .
forthcoming .
software architectures , global computing and graph transformation via logic programming . in _
proc sbes2002 - 16th brazilian symposium on software engineering_. anais , 1135 .
2004a . a graphical fusion calculus . in _ proc
of the workshop of the cometa project on computational metamodels_. electronic notes in theoretical computer science , vol .
elsevier science , 199215 . 2004b .
synchronization algebras with mobility for graph transformations . in _ proc . of fguc04
foundations of global ubiquitous computing_. electronic notes in theoretical computer science , vol . 138 .
elsevier science , 4360 .
springer . , parrow , j. , and walker , d. 1992 . a calculus of mobile processes .
_ 100 _ , 177 .
algebraic properties of idempotent substitutions . in _ proc . of icalp90_. lncs , vol . 443 .
springer , 386399 .
the fusion calculus : expressiveness and symmetry in mobile processes . in _ proc . of lics
98_. ieee , computer society press .
\1993 . a theory of bisimulation for the pi - calculus . in _ proc .
of concur93_. lncs , vol . 715 .
springer , 127142 .
the fusion calculus : expressiveness and symmetry in mobile processes .
thesis , dept . of computer systems , uppsala university ,
we have here the proofs that are missing in the main part and some lemmas used in these proofs .
lemmas are just before the proofs that use them .
notice that all the auxiliary productions perform two non trivial actions , and because of hoare synchronization and because each node is shared by at most two edges , synchronizing edges form chains .
there are two alternatives : each chain either starts and begins on an external node , or it is a cycle that contains only internal nodes .
exactly one chain must be of the first type .
in fact if we have no chain of that kind we have only trivial actions on nodes in @xmath16 .
also , if we have more than one , we have more than two non trivial actions on nodes in @xmath16 .
thanks to the last condition of definition [ defin : structured ] this chain contains an odd number of connectors , which must be @xmath260 connectors .
one can easily check by induction on the ( odd ) length of the chain that the transition creates an amoeboid @xmath361,\vec y_2[i])$ ] for each component of vectors @xmath362 and @xmath255 of fresh names and that these vectors are exposed on the external nodes together with an @xmath363 and an @xmath364 action , where @xmath365 .
let us now consider the other kind of chains : these chains produce rings of @xmath234 edges connected only to fresh nodes , which thus correspond to isolate subgraphs .
furthermore , they affect only the labels of internal nodes as required .
since we are working up to injective renamings we only have to prove that two names are connected by @xmath367}_{i=1 \dots n } m(x_i , y_i)$ ] iff they are merged by @xmath368 and that all the paths connecting external nodes in the final graph have odd length . by definition
two names are connected by @xmath186 iff they are merged by @xmath2 .
assume that two names @xmath5 and @xmath369 are merged by @xmath368 .
then their images along @xmath2 are merged by @xmath338 .
this means that we have in @xmath28 a chain of equalities from @xmath370 to @xmath371 .
thus we have amoeboids connecting the amoeboids of @xmath370 and of @xmath371 , thus @xmath5 and @xmath369 are connected .
assume now that @xmath5 and @xmath369 are connected .
then there exists an amoeboid connecting the amoeboids for @xmath370 and for @xmath371 .
thus @xmath338 merges @xmath370 and @xmath371 as required .
finally note that all the paths between external nodes are created by connecting existing paths via new amoeboids .
in particular , each new path is composed by @xmath158 old paths and @xmath372 new amoeboids .
thus its length is the sum of @xmath373 odd lengths and thus it is odd . in order for that process to be in the standard form we just need to make @xmath374 sequential by unfolding recursion and taking bound names to the outside if required . for simplicity , we show just the case where @xmath374 is already sequential , the other being essentially equal .
@xmath390{(u'_1,in_{|\vec x'|},\vec x ' ) } \vec p_1,\vec x',\vec v_p \vdash { \llbracket p \rrbracket}_{\vec v_p,\vec w_p}|\operatorname{\big[\!\!\big]}_{x \in \gamma''_p } n(x)\ ] ] @xmath391{(u'_2,out_{|\vec y'|},\vec y ' ) } \vec q_1,\vec y',\vec v_q \vdash { \llbracket q \rrbracket}_{\vec v_q,\vec w_q}|\operatorname{\big[\!\!\big]}_{x \in \gamma''_q } n(x)\ ] ]
the choice of the parameters @xmath392 and @xmath393 is not important since they are fresh names and we work up to injective renamings .
the choice of @xmath394 and @xmath395 is important instead , since they have to correspond to the nodes in the lhs that are still used by the process .
the choice performed in definition [ defin : processprod ] ensures that . note
that @xmath396 contains an amoeboid for a set @xmath397 with @xmath398 .
thus from lemma [ lemma : amoeboid ] the two above productions can synchronize via the amoeboid . using rule ( idle ) for the other edges , we obtain as a result : @xmath399}_{x \in \gamma''_p } n(x)|{\llbracket q \rrbracket}_{\vec v_q,\vec
w_q}|\operatorname{\big[\!\!\big]}_{x \in \gamma''_q } n(x)|{\llbracket \theta \rrbracket}|\operatorname{\big[\!\!\big]}_{x \in \vec v } n(x)| \\
\operatorname{\big[\!\!\big]}_{i=1 \dots note that the fusion @xmath0 of the transition is the identity since each synchronization involves two edges , and at least one of them is an amoeboid which exposes fresh names .
furthermore amoeboids expose each fresh name twice .
thus no names in @xmath76 are merged .
also , in the final graph each node is shared by two edges .
in fact , the only connections whose cardinality is not preserved by productions are the ones between nodes whose references are exposed , namely @xmath400 and @xmath401 , and the corresponding process edges , but these nodes are connected to the new @xmath402 amoeboids created by the auxiliary productions . in particular , @xmath169}_{x \in \gamma''_p } n(x)|\operatorname{\big[\!\!\big]}_{x \in \gamma''_q } n(x)|{\llbracket \theta \rrbracket}$ ] is an amoeboid for @xmath403 .
furthermore thanks to lemma [ lemma : amoeboidconn ] by adding in parallel @xmath169}_{i=1 \dots the rhs of the fusion rule is @xmath406 , which has to be normalized into @xmath407 with @xmath408 and @xmath409 .
furthermore we can choose the names ( since we are reasoning up to injective renamings ) in such a way that @xmath410 ( plus an injective renaming on names in @xmath411 and @xmath393 which corresponds to an equivalence on the resulting amoeboids ) .
in particular , @xmath169}_{x \in \gamma''_p } n(x)|{\llbracket \theta \rrbracket}$ ] is an amoeboid for @xmath418 .
furthermore , note that @xmath186 has the form @xmath169}_{i=1 \dots
n } m_2(x_{i},y_{i})$ ] and thus thanks to lemma [ lemma : amoeboidconn ] by adding it in parallel we obtain an amoeboid for @xmath419 .
the rhs of the fusion rule is @xmath421 , which has to be normalized into @xmath422 with @xmath408 .
furthermore we can choose the names ( since we are reasoning up to injective renamings ) in such a way that @xmath423 ( plus an injective renaming on names in @xmath392 which corresponds to an equivalence on the resulting amoeboids ) .
@xmath59 equivalent agents are converted into the same representative ( up to @xmath45-conversion ) before being translated , thus the translation of @xmath36 and @xmath153 are equal up to injective renamings .
similarly for the translation of @xmath425 and @xmath426 , thus the thesis follows .
[ lemma : det ] given a graph @xmath68 and one or zero productions for each edge in @xmath70 let : + @xmath126{\lambda_1,\pi_1 } \phi_1 \vdash g'_1 $ ] + @xmath126{\lambda_2,\pi_2 } \phi_2 \vdash g'_2 $ ] + be two transitions obtained by applying the chosen production for each edge ( and using the ( idle ) rule if no production is chosen ) .
then there exists an injective renaming @xmath2 such that : the proof is a simple rule induction if one proves that derivations can be done in a standard way , namely by applying to the axioms first rules ( par ) , then rule ( merge ) and finally rules ( new )
. this can be proved by showing that one can exchange the order of rules and that one can substitute two applications of ( merge ) with substitutions @xmath2 and @xmath435 with just one application with substitution @xmath436 .
we have many cases to consider , but they are not so interesting . as examples we will show the detailed proof for commutation of rule ( merge ) and ( par ) and for composition of two different rules ( merge ) .
@xmath437{\lambda_1 \cup \lambda ' , \pi_1 \cup \pi ' } \phi_1 , \phi ' \vdash g_2\sigma\rho|g'_2 } { \infer { \gamma\sigma \vdash g_1\sigma \xrightarrow[]{\lambda_1 , \pi_1 } \phi_1 \vdash g_2\sigma\rho } { \gamma \vdash g_1 \xrightarrow[]{\lambda , \pi } \phi \vdash g_2 } & \gamma ' \vdash g'_1 \xrightarrow[]{\lambda ' , \pi ' } \phi ' \vdash g'_2 } \ ] ] where for readability we have not written explicitly the side conditions ( see definition [ defin : hoarerules ] ) . then we must also have a derivation for the same transition obtained applying rule ( par ) first and then rule ( merge ) : @xmath438{\lambda '' , \pi '' } \phi '' \vdash ( g_2|g'_2)\sigma\rho ' } { \infer{\gamma , \gamma ' \vdash g_1|g'_1 \xrightarrow[]{\lambda \cup \lambda ' , \pi \cup \pi ' } \phi , \phi ' \vdash g_2|g'_2 } { \gamma \vdash g_1 \xrightarrow[]{\lambda , \pi } \phi \vdash g_2 & \gamma ' \vdash g'_1 \xrightarrow[]{\lambda ' , \pi ' } \phi ' \vdash g'_2 } } \ ] ] we have to prove that this derivation is allowed and that the resulting transition is the one derived also by derivation [ eqn : merge - par ] .
the first step is allowed iff @xmath439 . since the first derivation is allowed by hypothesis , then @xmath440 .
thus the only problem is when a name which is after renamed ( by @xmath2 or @xmath123 ) creates a conflict . in that case
thanks to lemma [ lemma : injren ] we can suppose to start with a different name . the final result of the derivation is not changed by that since the name disappears .
thus the first step is legal . for the second step
we need @xmath441 . note that
since @xmath439 and @xmath442 , we have that @xmath2 is the identity on @xmath230 , thus the only @xmath443 such that @xmath444 are in @xmath76 , thus the condition is satisfied since it was satisfied in derivation [ eqn : merge - par ] .
furthermore we have @xmath445 and @xmath446 .
we must now consider @xmath447 .
we have : latexmath:[\[\rho ' = \operatorname{mgu}(\{(\operatorname{n}_{\lambda \cup \lambda'}(x))\sigma= ( \operatorname{n}_{\lambda \cup \lambda'}(y))\sigma|x\sigma= y\sigma \ } \cup \ { x\sigma = y\sigma @xmath456{\lambda '' , \pi '' } \phi '' \vdash g_2\sigma\rho\sigma'\rho ' } { \infer{\gamma\sigma \vdash g_1\sigma \xrightarrow[]{\lambda ' , \pi ' } \phi ' \vdash g_2\sigma\rho } { \gamma \vdash g_1 \xrightarrow[]{\lambda , \pi } \phi \vdash g_2 } } \ ] ] first of all we have to prove that the step is allowed .
the required condition is that @xmath459 .
we have two cases . if @xmath52 then the thesis follows from the analogous condition of the first step of derivation [ eqn : merge - merge ] .
otherwise we can rewrite the condition as @xmath460 .
note that @xmath461 thus we can rewrite the condition as @xmath462 what is the condition for the second step of derivation [ eqn : merge - merge ] .
we add some explanations for that ( long ) sequence of equations .
step @xmath160 is just the definition of @xmath468 .
step @xmath469 is allowed by lemma [ lemma : mgusubst ] .
step @xmath470 is a simple mathematical transformation .
step @xmath471 applies the definition of @xmath123 .
step @xmath472 is lemma [ lemma : mgusubst ] again .
step @xmath473 creates a new @xmath123 on the outside using idempotence , then it brings it inside using lemma [ lemma : mgusubst ] and uses idempotence again to delete it where it is not necessary .
it uses the definition of @xmath474 too .
step @xmath475 is lemma [ lemma : mgusubst ] again .
steps @xmath476 and @xmath477 are trivial mathematics .
step @xmath478 is another application of lemma [ lemma : mgusubst ] .
finally , step @xmath479 is justified since the names in the domain of @xmath480 are neither in the domain of @xmath435 ( since otherwise they would be in @xmath124 ) nor in the domain of @xmath447 ( since @xmath447 is computed after having applied @xmath123 , which is idempotent ) .
thus an allowed @xmath481 is a subset of @xmath482 which can be deleted by idempotence . in order to apply them
we need two sequential process edges to be rewritten .
each production needs to be synchronized with at least another one since each node is shared by exactly two edges . since process edges are connected only through amoeboids we can have a synchronization only if the two actions done by process edges are equal to two actions allowed by amoeboids .
thanks to lemma [ lemma : amoeboid ] , they must be two complementary actions , i.e. an @xmath363 and an @xmath364 . furthermore they have to be done on the same amoeboid , that is on two names merged by the substitution @xmath2 corresponding to the amoeboids .
we must have @xmath487 and @xmath488 ( or swapped ) with @xmath489 .
furthermore @xmath2 merges @xmath287 and @xmath490 thus we have a transition @xmath278 that corresponds to the synchronized execution of the two prefixes . the productions to be applied are thus forced except for the ones inside the amoeboids , but the only difference among the choices ( as shown by lemma [ lemma : amoeboid ] ) amounts to pseudoamoeboids and exchanges between equivalent amoeboids in the result . from lemma [ lemma : det ] we know that the result of a transition is determined up to injective renamings ( and actions on isolated nodes ) by the starting graph and the productions chosen . thus the transition that corresponds to @xmath272 for theorem [ theorem : hoarecorrectness ] is equal up to injective renamings to a transition that differs from @xmath491{\lambda , id } \gamma ' \vdash g$ ] only for pseudoamoeboids and substitutions of equivalent amoeboids . the thesis follows .
we want to build a big - step where the clause unified with @xmath493 is @xmath494 , if any ( some @xmath493 may not be replaced , in that case as notation we use @xmath495 ) . as a notational convention we use @xmath496 to denote the arguments of @xmath493 and @xmath497 to denote the jth argument of @xmath498 if it is a complex term and @xmath499 if it is a variable ( note that we have different names for the same variable , one for each occurrence ) . all these are undefined if @xmath493 is not replaced , @xmath500 and @xmath501 are undefined also if the jth argument of @xmath498 is a variable
. @xmath508,\vec { [ y'_{i , j}]})/x_{i , j } | a_{i , j } \textrm { is defined}\ } \cup \ { [ v_k]/v_k | ( v_k = x'_{i , j } \lor v_k=\vec y'_{i , j}[l ] \lor ( v_k = x_{i , j } \land a_{i , j } \textrm { is undefined } ) ) \land [ v_k ] \neq v_k \}$ ] we will prove a more general result by induction on the number of `` considered '' atoms , that is we consider an increasing chain of derivations , and considered atoms are the ones that , if expanded in the complete derivation , have already been expanded . for simplicity ,
atoms are considered in numeric order , i.e. at step @xmath260 atoms @xmath511 have already been considered , and atom @xmath512 becomes considered .
we will prove that a computation of the form @xmath513{\theta } \operatorname{\!\!\mbox{}^*\,\,}g_1,\dots , g_n$ ] where we substitute only atoms in the starting goal and where the atoms generated by considered atoms do not contain function symbols exists iff : + @xmath514 and that furthermore : * @xmath508',\vec { [ y'_{i , j}]'})/x_{i , j } | a_{i , j } \textrm { is defined } \land i \leq m\ } \cup \ { [ v_k]'/v_k | ( v_k = x'_{i , j } \lor v_k=\vec y'_{i , j}[l ] \lor ( v_k = x_{i , j } \land a_{i , j } \textrm { is undefined } ) ) \land [ v_k ] ' \neq v_k\}$ ] ; * @xmath515 ; base case , @xmath518 ) we have the following transition : + @xmath519{\theta_1 } ( b_1,a_2,\dots , a_n ) \theta_1 $ ] + where @xmath427 is an mgu of @xmath520 .
this transition exists iff @xmath427 exists .
+ we have : + @xmath521 .
+ note that if @xmath522 then the mgu does not exist ( if @xmath500 and @xmath523 are both defined , otherwise they have to be both undefined since if just one of them is defined then a function symbol will remain in the considered part against the hypothesis ) .
thus we have : + @xmath526'=v_k|[v_k ] ' \neq v_k\})$ ] .
+ to have the real mgu we just need to apply @xmath516 to the equations in the first part ( note that @xmath527 is unified with some other variable only if @xmath528 is undefined thus the domain variables of the first part are not changed ) .
+ this proves the first part since this mgu exists and the second one since it has the wanted form . +
the third part follows from the observation that @xmath529 .
inductive case , @xmath530 ) assume that @xmath531{\theta_g } \operatorname{\!\!\mbox{}^*\,\,}g_1,\dots , g_n$ ] is a logic computation where we substitute only atoms in the starting goal , where atoms @xmath532 are considered and where the atoms generated by them do not contain function symbols . * @xmath508,\vec { [ y'_{i , j}]})/x_{i , j } | a_{i , j } \textrm { is defined } \land i \leq m \ }
\cup \ { [ v_k]/v_k | ( v_k = x'_{i , j } \lor v_k=\vec y'_{i , j}[l ] \lor ( v_k = x_{i , j } \land a_{i , j } \textrm { is undefined } ) ) \land [ v_k ] \neq v_k\}$ ] ; + * @xmath533 .
we will now consider the atom @xmath534 .
if it is not substituted then the thesis follows trivially ( note that @xmath534 does not contain variables substituted with a complex term by @xmath338 , since this can happen only if we have @xmath535 with @xmath536 , @xmath537 defined and @xmath538 undefined , and this is forbidden ; thus we have @xmath539 ) .
let us consider the case in which it is substituted .
we will have a small step of the form : + @xmath540{\theta ' } ( \ g_1,\dots , g_{m+1},a_{m+2},\dots , a_n)\theta\theta'$ ] + where @xmath541 ( note that we can assume that @xmath338 is also applied to @xmath542 since we can assume @xmath543 ) .
we have : @xmath544 for each binding we must consider two cases : either the variable @xmath545 appears in already considered atoms ( we call it an _ old variable _ ) or it does not ( we call it a _ new variable _ ) . in the second case
@xmath338 is the identity on that variable . in the first case
if @xmath546 is defined then the mgu exists iff we have @xmath547 where @xmath548 is the function symbol in the binding for @xmath545 in @xmath338 .
note that if @xmath548 is undefined then also @xmath546 must be undefined otherwise the function symbol remains in the final goal .
thus we will have : @xmath549,\vec { [ y'_{p , q } ] } ) = a_{m+1,j}(x'_{m+1,j},\vec y'_{m+1,j})|a_{m+1,j } \textrm { is defined } \land \\
\land x_{p , q}=x_{m+1,j } \land x_{m+1,j } \textrm { is old } \ } \cup \{[x_{m+1,j}]=x'_{m+1,j}|a_{m+1,j } \textrm { is undefined}\ } \cup \\ \cup \ { x_{m+1,j}=a_{m+1,j}(x'_{m+1,j},\vec y'_{m+1,j})|a_{m+1,j } \textrm { is defined } \land x_{k+1,j } \textrm { is new}\}=\\ = \operatorname{mgu}(\ { [ x'_{p , q}]=x'_{m+1,j } , \vec{[y'_{p , q}]}=\vec y'_{m+1,j}|a_{m+1,j } \textrm { is defined } \land \\
\land x_{p , q}=x_{m+1,j } \land x_{m+1,j } \textrm { is old } \ }
\cup \{[x_{m+1,j}]=x'_{m+1,j}|a_{m+1,j } \textrm { is undefined}\ } \cup \\ \cup \ { x_{m+1,j}=a_{m+1,j}(x'_{m+1,j},\vec y'_{m+1,j})|a_{m+1,j } \textrm { is defined } \land x_{m+1,j } \textrm { is new}\}\end{gathered}\ ] ] where we use @xmath550 $ ] to denote the representative of the equivalence class according to @xmath338 .
we can reorder this substitution into : @xmath551=x'_{m+1,j } , \vec{[y'_{p , q}]}=\vec y'_{m+1,j}|a_{k+1,j } \textrm { is defined } \land\\ \land x_{p , q}=x_{m+1,j } \land x_{m+1,j } \textrm { is old } \ } \cup \{[x_{m+1,j}]=x'_{m+1,j}|a_{m+1,j } \textrm { is undefined}\}\end{gathered}\ ] ] note that the second part is a renaming that does not involve variables in the domain of the first part . thus we can define new equivalence classes @xmath550'$ ] according to this substitution @xmath516 and put the substitution in the resolved form : @xmath552',[\vec y'_{m+1,j}]')|a_{m+1,j } \textrm { is defined } \land x_{m+1,j } \textrm { is new}\ } \cup \\ \cup \ { [ v_k]'=v_k|[v_k ] ' \neq v_k\})\end{gathered}\ ] ] thus the mgu exists and the first part of the thesis is proved .
let us consider the substitution @xmath553 .
note that the renaming part of @xmath349 substitutes variables according to the equivalence between variables in the @xmath554 and representatives of the corresponding variables in @xmath534 thus the composed substitution @xmath555 maps each variable to the representative of the equivalence class that is defined by the union of the two sets of equations as required .
furthermore bindings with complex terms ( the first part of the substitution ) have disjoint domains and thus the union of them is made .
bindings coming from @xmath349 have already the wanted representatives in the image , while to bindings in @xmath338 the renaming is applied , mapping variables into the representatives of their equivalence classes .
thus @xmath556 has the wanted form w.r.t .
the equivalence classes determined by all the equivalences on variables . note that this result does not depend on the order of application of clauses and that after having chosen which clauses to apply and to which atoms it is deterministic up to an injective renaming ( which depends on the choice of names for new variables and on the choice of representatives for the equivalence classes ) .
we can have an applicable variant of this clause for each injective renaming @xmath123 that maps each variable to a fresh one .
let us consider the goal : + @xmath565 + it unifies with the clause variant : + @xmath566 + with mgu @xmath339 : + @xmath567 .
+ note that @xmath568 .
we can see that @xmath339 is associated to @xmath569{\lambda,\pi } \phi \vdash g$ ] as required .
observe also that the result of the computation is : + @xmath570 + because @xmath571 .
this proves the thesis . to deal with the missing possibilities for @xmath123 w.r.t .
the theorem statement note that bindings in the last part of @xmath339 can also be resolved also as @xmath572 . in that case
we have no binding for @xmath63 in @xmath339 .
this is equivalent to defining @xmath573 , and this covers the missing cases for @xmath123 .
rule ( par ) ) @xmath574{\lambda , \pi } \phi \vdash g_2 \quad \gamma ' \vdash g'_1 \xrightarrow[]{\lambda',\pi ' } \phi ' \vdash g'_2 \quad ( \gamma \cup \phi ) \cap ( \gamma ' \cup \phi')=\emptyset}{\gamma , \gamma ' \vdash g_1|g'_1 \xrightarrow[]{\lambda \cup \lambda ' , \pi \cup \pi ' } \phi , \phi ' \vdash g_2|g'_2}\ ] ] * @xmath575 + where @xmath339 is associated to @xmath576{\lambda,\pi } \phi \vdash g_2 $ ] and @xmath577 ; * @xmath578 + where @xmath579 is associated to @xmath580{\lambda',\pi ' } \phi ' \vdash g'_2 $ ] and @xmath581 . in both cases by inductive hypothesis we used as clauses the translations of the productions used in the proof of the hshr transition applied to the translations of the edges on which the productions were applied .
+ thanks to lemma [ lemma : injren ] we can assume that the sets of variables used in the two computations are disjoint .
thanks to lemma [ lemma : bigstep ] we know that these big - steps exist iff @xmath582 . since the used variable sets are disjoint the same condition guarantees the existence of a big - step of the form : + @xmath583{\theta_{\rho}\theta'_{\rho ' } } \operatorname{\!\!\mbox{}^*\,\,}t , t'$ ] + where we used as clauses the unions of the clauses used in the two smaller computations , applied to the same predicates .
rule ( merge ) ) @xmath574{\lambda , \pi } \phi \vdash g_2 \quad \forall x , y \in \gamma .
x\sigma = y\sigma \wedge x \neq y \rightarrow \operatorname{act}_{\lambda}(x ) = \operatorname{act}_{\lambda}(y)}{\gamma\sigma \vdash g_1\sigma \xrightarrow[]{\lambda ' , \pi ' } \phi ' \vdash g_2\sigma\rho_g}\ ] ] where @xmath117 is an idempotent substitution and : where we used @xmath591 instead of @xmath123 to avoid confusion with the injective renaming @xmath123 . for inductive hypothesis
we have : + @xmath575 + where @xmath339 is associated to @xmath576{\lambda,\pi } \phi \vdash g_2 $ ] and @xmath577 and we used as clauses the translations of the productions used in the proof of the hshr transition applied to the translations of the edges on which the productions were applied .
* @xmath592 + @xmath593 + * @xmath508,\vec { [ y'_{i , j}]})/x_{i , j } | a_{i , j } \textrm { is defined}\ } \cup \ { [ v_k]/v_k | ( v_k = x'_{i , j } \lor v_k=\vec y'_{i , j}[l ] \lor ( v_k = x_{i , j } \land a_{i , j } \textrm { is undefined } ) ) \land [ v_k ] \neq v_k \}$ ] + * @xmath594 we want to find a big - step that uses the same clauses of the previous one applied to the same atoms , but starting from this new goal .
we will use an overline to denote the components of the new big - step .
thanks to lemma [ lemma : bigstep ] such a big - step exists iff : + @xmath597 + but if @xmath598 this has already been proved , otherwise this is guaranteed by the applicability conditions of the rule .
we have to prove that @xmath599 is associated to : + @xmath600{\lambda ' , \pi ' } \phi ' \vdash g_2\sigma\rho_g$ ] .
+ from lemma [ lemma : bigstep ] we have : + @xmath601',\vec { [ y'_{i , j}]'})/x_{i , j}\sigma | a_{i , j } \textrm { is defined}\ } \cup \ { [ v_k]'/v_k | ( v_k = x'_{i , j } \lor v_k=\vec y'_{i , j}[l ] \lor ( v_k = x_{i , j } \land a_{i , j } \textrm { is undefined } ) ) \land [ v_k ] ' \neq v_k \}$ ] + where @xmath550'$ ] maps each variable to the representative of the equivalence class given by : @xmath602 is never unified with a primed variable unless @xmath500 is undefined . from the hypothesis and from lemma
[ lemma : mgusubst ] we have that @xmath607 .
the first and the third part are equal ( adding primes ) to the equations for @xmath550'$ ] . as far as the second part is concerned , @xmath123 maps variables merged by @xmath0 to the same primed variable , thus these equations become trivial . rule ( new ) ) @xmath574{\lambda,\pi } \phi \vdash g_2 \quad x \notin \gamma \quad \vec y \cap ( \gamma \cup \phi \cup \{x\})= \emptyset}{\gamma , x \vdash g_1 \xrightarrow[]{\lambda \cup \{(x , a,\vec y)\},\pi } \phi ' \vdash g_2}\ ] ] for inductive hypothesis we have in @xmath36 the following big - step of synchronized logic programming : + @xmath612{\theta_{\rho } } \operatorname{\!\!\mbox{}^*\,\,}t$ ] + where @xmath339 is associated to : + @xmath613{\lambda,\pi } \phi \vdash g_2 $ ] + and @xmath614 but because @xmath615 we have that @xmath339 is also associated to : + @xmath616{\lambda \cup \{(x , a,\vec y)\},\pi } \phi ' \vdash g_2 $ ] + the thesis follows .
the translation @xmath617 can be written in the form @xmath492 where the @xmath618 are translations of single edges .
we want to associate a hshr production to each @xmath493 . for edges associated to atoms that are rewritten in @xmath619
we use an instance of the axiom that corresponds to the clause used to rewrite it , otherwise the rule obtained from rule ( idle ) applied to that edge and to nodes that are all distinct .
let @xmath620 be the heads of these rules .
we choose the names in the following way : for each first occurrence of a variable in @xmath492 we use the same name for the node in the corresponding position , we use new names for all other occurrences .
note that there exists a substitution @xmath2 such that @xmath621 and that @xmath2 is idempotent .
note also that for each @xmath21 all names in @xmath622 are distinct . for each @xmath21
we can choose an instance @xmath623 of the associated rule with head @xmath622 such that for each @xmath624 we have @xmath625 .
as notation we use : + @xmath626{\lambda_i , \pi_i } \phi_i \vdash g'_i$ ] + since all the rules have a disjoint set of names we can apply @xmath372 times rule ( par ) in order to have : + @xmath627}_i g_i \xrightarrow[]{\bigcup_i \lambda_i , \bigcup_i \pi_i } \bigcup_i \phi_i \vdash \operatorname{\big[\!\!\big]}_i g'_i$ ] + now we want to apply rule ( merge ) with substitution @xmath2 . we can do it since @xmath2 is idempotent .
we have to verify that @xmath628 .
this happens thanks to lemma [ lemma : bigstep ] .
thus we obtain a rule of the form : + @xmath629{\lambda , \pi } \phi \vdash g'$ ] + thanks to theorem [ theorem : tuostocorrectness ] we can have in @xmath36 the following big - step of synchronized logic programming : + @xmath630 + for every @xmath123 that satisfies the freshness conditions .
furthermore @xmath631 is associated to @xmath632{\lambda,\pi } \phi \vdash g'$ ] and @xmath633 .
finally we used as clauses the translations of the productions used in the proof of the hshr rewriting . note that @xmath634 . since the result of a big - step
is determined up to injective renaming by the starting goal and the used clauses ( see lemma [ lemma : bigstep ] ) then we must have @xmath635 and @xmath636 for some injective renaming @xmath447 .
note that @xmath637 satisfies the freshness conditions ( since variables generated in logic programming are always fresh ) and we also have @xmath638 .
+ thus we have that @xmath338 is associated to @xmath346{\lambda,\pi } \phi \vdash g'$ ] and that @xmath639 .
this proves the thesis . | in this paper we compare three different formalisms that can be used in the area of models for distributed , concurrent and mobile systems .
in particular we analyze the relationships between a process calculus , the _ fusion calculus _ , graph transformations in the _ synchronized hyperedge replacement _ with hoare synchronization ( hshr ) approach and _ logic programming_. we present a translation from fusion calculus into hshr ( whereas fusion calculus uses milner synchronization ) and prove a correspondence between the reduction semantics of fusion calculus and hshr transitions .
we also present a mapping from hshr into a transactional version of logic programming and prove that there is a full correspondence between the two formalisms .
the resulting mapping from fusion calculus to logic programming is interesting since it shows the tight analogies between the two formalisms , in particular for handling name generation and mobility .
the intermediate step in terms of hshr is convenient since graph transformations allow for multiple , remote synchronizations , as required by fusion calculus semantics . to appear in theory and practice of logic programming ( tplp ) .
fusion calculus , graph transformation , synchronized hyperedge replacement , logic programming , mobility |
a node - capacitated undirected network is a quadruple @xmath9 of node set @xmath10 , ( undirected ) edge set @xmath11 , a specified subset @xmath12 of nodes , called _
terminals _ , and a nonnegative integer - valued node capacity @xmath13 on nonterminal nodes . an _
@xmath12-path _ is a path connecting distinct terminals .
a ( node - capacitated ) _ multiflow _ is a pair @xmath14 of a set @xmath15 of @xmath12-paths and a flow - value function @xmath16 satisfying the node - capacity constraint : @xmath17 the total flow - value of a multiflow @xmath18 is defined as @xmath19 .
a multiflow is called _ maximum _ if it has the maximum total flow - value among all possible multiflows .
a multiflow @xmath20 is said to be _ integral _ if @xmath21 is integer - valued , and _ half - integral _ if @xmath22 is integer - valued . in this paper , we address the problem of finding a maximum multiflow in a node - capacitated network .
this multiflow problem appeared in the work by garg , vazirani , and yannakakis @xcite on an approximation algorithm for node - multiway cut .
in fact , the lp - dual of our multiflow problem is a natural lp - relaxation of the _ minimum node - multiway cut problem _ ; see also ( * ? ? ?
* section 19.3 ) .
they showed that this lp - dual always has a half - integral optimal solution .
the half - integrality of the primal problem , i.e. , the existence of a half - integral maximum multiflow , was later shown by pap @xcite .
he also showed that a half - integral maximum multiflow can be found in strongly polynomial time . in these works ,
the polynomial time solvability depends on the use of the ellipsoid method .
thus it is natural to seek a combinatorial polynomial time algorithm . for the case of unit node - capacity ( @xmath23 for all @xmath24 ) ,
babenko @xcite developed a combinatorial @xmath25 time algorithm to find a half - integral maximum multiflow , where @xmath1 is the number of nodes and @xmath2 is the number of edges .
for general node - capacity , babenko and karzanov @xcite developed a combinatorial weakly polynomial time algorithm to find a half - integral maximum multiflow .
their algorithm runs in @xmath26 time , where @xmath27 is the time complexity of solving the max - flow problem in a network with @xmath1 nodes , @xmath2 edges , and the maximum edge - capacity @xmath28 .
the main result of this paper is the first combinatorial _ strongly _ polynomial time algorithm to solve the maximum node - capacitated multiflow problem .
our algorithm uses , as a subroutine , an algorithm of solving the _ maximum submodular flow problem _ ; see ( * ? ? ?
* section 5.5 ( c ) ) .
let @xmath29 denote the time complexity of solving the maximum submodular flow problem on a network with @xmath1 nodes , @xmath2 edges , and the time complexity @xmath7 of computing the exchange capacity of the submodular function describing the problem .
[ thm : main ] there exists an @xmath30-time algorithm to find a half - integral maximum multiflow and a half - integral optimal dual solution in a network of @xmath1 nodes , @xmath2 edges , and @xmath3 terminals .
the current fastest maximum submodular flow algorithm is the push - relabel algorithm due to fujishige and zhang @xcite of the time complexity @xmath31 ; see the survey @xcite on submodular flow algorithms .
thus we can solve the problem in @xmath32 time .
[ [ application-1-node - multiway - cut . ] ] application 1 : node - multiway cut .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + a _ node - multiway cut _ is a subset @xmath33 of nonterminal nodes such that the deletion of @xmath34 makes every pair of distinct terminals unreachable , or equivalently , @xmath34 meets every @xmath12-path .
the _ capacity _ of a node - multiway cut @xmath34 is defined as @xmath35 .
the _ minimum node - multiway cut problem _ asks to find a node - multiway cut with the minimum capacity .
this well - known np - hard problem is naturally formulated as the following @xmath36-integer program : @xmath37 the natural lp - relaxation obtained by relaxing @xmath38 into @xmath39 is nothing but the lp - dual of our multiflow problem . as mentioned above , garg , vazirani , and yannakakis @xcite proved that a half - integral optimal lp solution @xmath40 always exists , and is obtained from any optimal lp solution by a simple rounding procedure ; see ( * ? ? ?
* section 19.3 ) .
then the set of nodes @xmath41 with @xmath42 is a 2-approximation solution of the minimum node - multiway cut problem .
this rounding algorithm needs an optimal lp solution , which is now obtained by our algorithm . to the best of our knowledge ,
this is the first combinatorial strongly polynomial time implementation of the @xmath43-approximation algorithm .
just recently , chakuri and madan @xcite devised a simple method to round any feasible lp solution into a multiway cut of capacity within factor @xmath43 . combining this rounding method with a fast fptas for multiflow ( e.g. , @xcite )
, they obtain a considerably fast @xmath44-approximation algorithm ( with running time dependent of @xmath45 ) .
[ [ application-2-integral - multiflow . ] ] application 2 : integral multiflow .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + our algorithm is also useful in the problem of finding a maximum integral multiflow .
this problem is a capacitated version of openly - disjoint @xmath12-paths packing problem considered by mader @xcite ( that corresponds to the case of @xmath23 @xmath46 ) .
pap @xcite established the strongly polynomial time solvability of the maximum integral multiflow problem .
the first step of his algorithm is to find a maximum half - integral multiflow .
the second step is to construct and solve an instance of the openly - disjoint @xmath12-path packing problem ( on the graph with size polynomial in the numbers of edges and nodes in the original network ) . finally , combining the integer part of the half - integral multiflow with a solution of the packing problem , one obtains a maximum integer multiflow .
the second step can be done by several combinatorial polynomial time algorithms , including @xcite and ( * ? ? ?
* section 73.1a ) . our algorithm can be used in the first step , and makes the whole algorithm fully combinatorial . [
[ outline . ] ] outline .
+ + + + + + + + let us outline our algorithm and the ideas behind it , as well as the structure of the paper .
our algorithm is designed on the basis of the following two ingredients .
one is a combinatorial duality theory for a class of node - capacitated multiflow problems @xcite .
the other is a developing theory of discrete convex functions on certain graph structures @xcite , which aims to extend concepts in _ discrete convex analysis ( dca ) _
( murota @xcite ) to tackle further various combinatorial optimization problems beyond network flows , matroids , and submodular functions .
we will utilize these theories in a self - contained way . in section [ sec : multiflow ] , following @xcite we formulate the dual of our multiflow problem as a facility location problem on a tree .
this formulation gives a fruitful combinatorial interpretation of the lp - dual problem ( [ eqn : lp - dual ] ) , and brings a simple combinatorial algorithm to find a half - integral optimal multiflow from a given dual optimum , under a certain nondegeneracy assumption .
we will deal with a perturbed problem satisfying this nondegeneracy assumption .
our goal is to solve this perturbed problem efficiently .
we will see that the location problem is further formulated as an optimization over a certain discrete structure , and the objective is an _
l - convex function on a euclidean building _ in the sense of @xcite .
this class of discrete convex functions shares many analogous properties with l - convex functions in dca . in particular , as in the case of dca , there is a natural descent algorithm , called the _ steepest descent algorithm _ , to minimize our l - convex function @xmath47 . for each point @xmath48 ,
the steepest descent algorithm chooses a point @xmath49 ( _ steepest direction _ ) from a _ discrete neighborhood _ of @xmath48 with smallest @xmath50 .
if @xmath51 , then @xmath48 is guaranteed to be optimal . otherwise , i.e. , @xmath52 , replace @xmath48 by @xmath49 , and repeat . in section [ sec : algo ]
, we will implement this conceptually simple algorithm .
we will prove that in our case a steepest direction at each point can be found by solving one maximum submodular flow problem .
this part is the heart of our analysis . as a consequence
, we obtain an algorithm in a simple form as follows : * from a dual solution @xmath48 , construct and solve an instance of the maximum submodular flow problem . *
if the minimal minimum cut consists only of the source , then @xmath48 is optimal , and an optimal multiflow is constructed from any maximum submodular flow .
otherwise the minimal minimum cut gives a steepest direction @xmath49 of the neighborhood at @xmath48 .
replace @xmath48 by @xmath49 , and go to 1 .
our maximum submodular flow problem is defined by a disjoint sum of submodular functions on @xmath53-element sets .
this enables us to compute the exchange capacity in constant time .
moreover , the number of iterations is estimated by the _ geodesic descent property _ ( theorem [ thm : bound ] ) of the steepest descent algorithm .
this intriguing property says that a trajectory of the algorithm forms a geodesic to optimal solutions with respect to a certain @xmath54-metric on the domain .
we know in advance the range where an optimum exists , and the diameter of the range is bounded by @xmath55 relative to the above metric . consequently the number of iteration is bounded by @xmath55 .
it should be noted that our algorithm design includes an interesting new technique of reducing bisubmodularity to submodularity .
this technique and related arguments , including basics on submodularity , are summarized in section [ sec : pre ] .
actually step 1 of the above algorithm is essentially the feasibility check of a _
bisubmodular flow _
problem , that is , finding a ( fractional ) bidirected flow with the flow - boundary constrained to a bisubmodular polyhedron .
this seemingly natural class of problems has not been well - studied so far .
it is well - known that ( fractional ) bidirected flows are easily manipulated by ordinary flows in a _ skew - symmetric network
_ obtained by doubling nodes and edges .
we generalize this doubling construction to bisubmodular functions .
we give a condition for a bisubmodular function @xmath56 to be extended to a submodular function @xmath57 on a larger set , so that the bisubmodular polyhedron of @xmath56 is a projection of the base polyhedron of @xmath57 .
we show that a certain bisubmodular function on a 3-element set , which represents the flow - conservation and the node - capacity constraints on a node of degree 3 , has such a submodular extension on a 6-element set .
our bisubmodular flow problem is described by the disjoint sum of these bisubmodular functions , and can be reduced to the submodular flow problem as mentioned above .
[ [ notation . ] ] notation .
+ + + + + + + + + let @xmath58 , @xmath59 , @xmath60 , and @xmath61 denote the sets of reals , nonnegative reals , integers , and nonnegative integers , respectively .
the infinity element @xmath62 is treated as @xmath63 and @xmath64 for @xmath65 .
the set of all functions from a set @xmath10 to a set @xmath66 is denoted by @xmath67 . for a function @xmath68 ,
let @xmath69 . for a function @xmath70 and a subset @xmath71 ,
let @xmath72 denote @xmath73 .
the function value @xmath74 will also be denoted by @xmath75 if no confusion occurs . for a ( directed or undirected ) graph @xmath76 , an edge from @xmath41 to @xmath77
is denoted by @xmath78 . for a subset @xmath34 of nodes , let @xmath79 denote the set of all edges leaving @xmath34 . for an undirected graph @xmath80 with a specified edge - length ,
let @xmath81 denote the shortest path metric on the vertex set with respect to the edge - length . in the following ,
graphs or networks are supposed to have no multiple edges and loops .
[ [ signed - set - and - transversal . ] ] signed set and transversal .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + a _ signed set _
@xmath82 is the product @xmath83 of a set @xmath10 and the sign @xmath84 .
elements @xmath85 and @xmath86 of @xmath82 are simply denoted by @xmath87 and @xmath88 , respectively .
the _ signed extension _ of a set @xmath10 is defined as the signed set @xmath83 and is denoted by @xmath89 .
for @xmath90 , let @xmath91 and @xmath92 . also for @xmath93 , let @xmath94 and @xmath95 .
a subset @xmath34 of @xmath89 is called _ transversal _ if @xmath96 for all @xmath97 , and is called _ co - transversal _ if @xmath98 for all @xmath97
. for @xmath99 let @xmath100 denote the transversal obtained from @xmath34 by deleting all @xmath101 with @xmath102 , and let @xmath103 denote the co - transversal obtained from @xmath34 by adding all @xmath101 with @xmath104 . for @xmath105 , define @xmath106 by @xmath107 if @xmath108 and @xmath109 if @xmath110 .
[ [ skew - symmetric - network . ] ] skew - symmetric network .
+ + + + + + + + + + + + + + + + + + + + + + + a _ skew - symmetric network _
( see e.g. , @xcite ) is a directed network on a signed set such that an edge @xmath111 exists if and only if an edge @xmath112 exists , and the ( lower and upper ) capacities of edges @xmath111 and @xmath112 are the same .
a skew - symmetric network is often useful for dealing with undirected objects . here
we summarize basics on submodular functions and submodular flows ; see @xcite for further details .
submodular function _ on a set @xmath10 is a function @xmath113 defined on @xmath114 satisfying @xmath115 for @xmath116 .
let @xmath113 be a submodular function on @xmath10 with @xmath117 .
the _ base polyhedron _
@xmath118 is defined to be the set of all vectors @xmath119 satisfying @xmath120 for @xmath71 and @xmath121 . for @xmath122 and a pair @xmath123 of distinct elements of @xmath10 , the _ exchange capacity _
@xmath124 at @xmath48 is defined by @xmath125 where @xmath126 is the @xmath41-th unit vector defined by @xmath127 if @xmath128 and @xmath129 otherwise .
we next introduce submodular flows .
let @xmath130 be a directed network with vertex set @xmath10 , edge set @xmath131 , edge - capacity @xmath132 , and terminals @xmath133 .
the set of nonterminal nodes is denoted by @xmath134 .
we are given a submodular function @xmath135 with @xmath136 .
for a function @xmath137 , let @xmath138 denote the function on @xmath10 defined by @xmath139 namely @xmath140 represents the excess of @xmath141 at node @xmath41 .
an _ @xmath142-flow _ , or simply , a _ flow _ is a function @xmath143 satisfying @xmath144 where @xmath145 means the restriction to @xmath82 .
the flow - value of a flow @xmath141 is defined as @xmath146 . an _ @xmath147-cut _
@xmath34 is a subset of nodes containing @xmath148 and not containing @xmath149 .
the _ cut capacity _ of @xmath34 is defined as @xmath150 an @xmath147-flow is called _ maximum _ if it has the maximum flow - value , and an @xmath147-cut is called _ minimum _ if it has the minimum capacity .
the maximum flow - value of an @xmath147-flow is equal to the minimum cut - capacity of an @xmath147-cut @xmath34 .
if @xmath151 and @xmath113 are both integer - valued , then there exists an integer - valued maximum flow . for any maximum flow @xmath141 ,
the set of nodes reachable from @xmath148 in the residual network @xmath152 of @xmath141 is the unique minimal minimum @xmath147-cut . here
the _ residual network _
@xmath152 of @xmath141 is a directed network on @xmath10 constructed as follows : for each edge @xmath78 in @xmath130 with @xmath153 , add an edge @xmath78 to @xmath152 .
for each edge @xmath78 in @xmath130 with @xmath154 , add an edge @xmath155 to @xmath152 .
for each pair of distinct nodes @xmath156 in @xmath82 with @xmath157 , add an edge @xmath78 to @xmath152 .
there are several combinatorial polynomial time algorithms for computing an integral maximum submodular flow , under the assumption that an oracle of computing the exchange capacity is available .
they are designed by extending existing max - flow algorithms ; see the survey @xcite for further detail .
we note one basic property for the case where the network is skew - symmetric .
[ lem : minimal_minimum ] suppose that @xmath10 is a signed set with @xmath158 , and @xmath130 is skew - symmetric . if @xmath113 satisfies @xmath159 for all @xmath160 , then the unique minimal minimum @xmath147-cut is a transversal . by the assumption for @xmath113
, it suffices to show @xmath161 .
suppose that some edge @xmath162 appears in @xmath163 . in this case , it holds @xmath164 and @xmath165 .
this means that @xmath166 and @xmath167 .
hence edge @xmath168 ( of the same capacity ) does not appear in @xmath163 .
consequently the cut - capacity does not increase .
let @xmath169 .
let @xmath170 denote the set of ordered pairs of disjoint subsets of @xmath10 .
for a function @xmath171 on @xmath172 , the polyhedron @xmath173 in @xmath174 is defined to be the set of all vectors @xmath175 satisfying @xmath176 if @xmath171 is a bisubmodular function on @xmath172 is called _ bisubmodular _ if it satisfies @xmath177 ] , then @xmath173 is known as a bisubmodular polyhedron .
we study the case where @xmath173 is a projection of the base polyhedron of some submodular function .
a representative of such polyhedra is the polyhedron of all flow - boundaries of a bidirected network ; see @xcite .
let @xmath178 be the signed extension of @xmath10 .
for a function @xmath171 on @xmath172 , a _
normal submodular extension _ of @xmath171 is a submodular function @xmath113 on @xmath179 with @xmath180 satisfying @xmath181 define a map ( projection ) @xmath182 by @xmath183 [ lem : projection ] let @xmath171 be a function on @xmath172 and @xmath113 a normal submodular extension of @xmath171
. then it holds @xmath184 .
note that @xmath185 .
we first show that @xmath186 .
take an arbitrary @xmath48 in @xmath118 .
for @xmath187 , let @xmath188 .
note that @xmath189 and @xmath190 .
then we have @xmath191 hence we have @xmath192 also we have @xmath193 hence @xmath194 belongs to @xmath173 .
next we show the converse .
take an arbitrary @xmath195 .
define a vector @xmath48 in @xmath196 by @xmath197 and @xmath198 for @xmath97 .
obviously @xmath199 .
it suffices to show that @xmath48 belongs to @xmath118 . since @xmath200
, we have @xmath201 . for @xmath99
, we have @xmath202 thus @xmath122 , and hence @xmath203 . if @xmath204 has a normal submodular extension , then @xmath171 is necessarily a bisubmodular function .
not all bisubmodular functions admit submodular extensions ( y. iwamasa 2015 ) . we consider a special bisubmodular function on @xmath205-element set @xmath206 , which plays a key role in section [ sec : multiflow ] . for @xmath207 ,
let @xmath208 be the function on @xmath209 defined by @xmath210 [ lem : flow_conservation ] the polyhedron @xmath211 is the set of nonnegative vectors @xmath212 satisfying @xmath213 observe that these inequalities appears in ( [ eqn : d(h ) ] ) .
so it suffices to show that inequalities ( [ eqn : d(h ) ] ) are implied by the above inequalities .
this is a routine verification .
for example , @xmath214 is obtained by adding @xmath215 and @xmath216 .
also @xmath217 is implied by @xmath218 and @xmath219 , and @xmath220 is implied by @xmath221 and @xmath222 .
the polyhedron @xmath211 is a simplex of vertices @xmath223 , @xmath224 , @xmath225 and @xmath226 .
we will see in section [ sec : multiflow ] that @xmath211 represents the flow - conservation law and the node - capacity constraint on a node of degree @xmath205 . this bisubmodular function
@xmath208 has a normal submodular extension .
the following example was found by yuni iwamasa via a computer calculation .
notice that each @xmath227 is one of the following six types : type 1 : : : @xmath228 and @xmath229 . type 2 : : : @xmath230 and @xmath231 for some @xmath232 .
type 3 : : : @xmath233 or @xmath234 .
type 4 : : : @xmath235 and @xmath236 .
type 5 : : : @xmath230 , @xmath237 , and @xmath238 for some @xmath232 .
type 6 : : : @xmath239 for some @xmath232 .
define @xmath240 by @xmath241 [ lem : delta * ] @xmath242 is a normal submodular extension of @xmath208 .
it suffices to consider the case of @xmath243 ; we denote @xmath244 and @xmath245 by @xmath246 and @xmath247 , respectively .
first we show ( [ eqn : extension1 ] ) . for @xmath248 ,
. then @xmath249 and @xmath250 .
if @xmath251 , then @xmath252 , and @xmath34 is of type 1 ; hence @xmath253 . if @xmath254 and @xmath252
. then @xmath34 is of type 5 , and hence @xmath255 . if @xmath256 is empty , then @xmath233 , and @xmath34 is type 3 ; hence @xmath257 . if @xmath258 and @xmath259 , then @xmath254 , and @xmath34 is of type 2 ; hence @xmath257 .
second we show ( [ eqn : extension2 ] ) . it suffices to show that @xmath260 holds for any transversal @xmath34 , and that @xmath261 holds for any @xmath34 that is not transversal and not co - transversal .
the former property follows from @xmath262 , @xmath263 , @xmath264 , @xmath265 , and @xmath266 .
the latter property follows from @xmath267 , @xmath268 , and @xmath269 .
finally we show the submodularity of @xmath247 .
take @xmath270 .
we can assume that @xmath271 and @xmath272 .
case 1 : @xmath34 is of type 6 . in this case , @xmath273 is the whole set , and is of type 3 .
therefore it suffices to consider the case where @xmath274 is of type 1 and @xmath256 is not of type 1 .
necessarily @xmath256 is of type 4 or 6 .
thus submodular inequality @xmath275 holds .
case 2 : @xmath34 is of type 2 with @xmath276 . if @xmath256 is also of type 2 , then @xmath277 is of type 3 , @xmath278 is also of type 3 . if @xmath256 is of type 3 with @xmath279 , then both @xmath274 and @xmath273 are of type 3 .
if @xmath256 is of type 3 with @xmath280 , then @xmath256 can not have @xmath281 , and thus both @xmath274 and @xmath273 are of type 3 . in these cases ,
submodularity @xmath282 holds .
thus we may assume that @xmath256 is of type 1,4 , or 5 .
observe that neither @xmath274 nor @xmath273 is of type 1 .
we may assume that @xmath256 contains @xmath281 and does not contain @xmath283 .
otherwise @xmath274 or @xmath273 is of type 3 , and submodularity holds .
necessarily @xmath256 is of type 1 .
then @xmath274 is of type 5 , and @xmath273 is of type 4 or 6 ; submodularity @xmath284 holds .
case 3 : @xmath34 is of type 3 and @xmath256 is not of type 2 .
if @xmath256 is also of type 3 , then both @xmath274 and @xmath273 is of type 3 ; submodularity holds .
since one of @xmath274 and @xmath273 is of type 3 , it suffices to consider the case where @xmath274 or @xmath273 is of type 1 .
we show that @xmath256 is also type 1 .
if @xmath274 is of type 1 , then @xmath285 , and necessarily @xmath256 is of type 1 .
if @xmath273 is of type 1 , then @xmath286 , and @xmath256 consists of at least two elements in @xmath287 ( type 1 ) .
case 4 : @xmath34 is of type 4 or 5 , and @xmath256 is type 1 , 4 , or 5 .
suppose that @xmath34 is of type 4 .
then @xmath273 is not type 1 .
we may consider the case where @xmath274 is of type 1 and @xmath256 is not of type 1 .
then @xmath256 contains @xmath288 .
thus @xmath256 is of type 4 , and @xmath273 necessarily contains @xmath289 ( type 3 ) ; submodularity ( @xmath275 ) holds .
suppose that @xmath34 is of type 5 with @xmath276 .
then @xmath274 is not type 1 .
we may consider the case where @xmath273 is of type 1 .
if @xmath256 does not contain @xmath281 , then @xmath274 is of type 3 ; submodularity ( @xmath275 ) holds .
thus @xmath290 and @xmath291 or @xmath292 ; @xmath256 is of type 1 .
the intersection @xmath274 is of type 5 ; thus @xmath293 holds .
in this section , we introduce a combinatorial duality theory , developed by @xcite , for a class of node - capacitated multiflow problems .
we consider the following multiflow problem . now
assume that network @xmath130 also has a nonnegative edge - cost @xmath294 ; so the network is a 5-tuple @xmath295 . for a multiflow @xmath20 , the total flow - values on node @xmath41 and edge @xmath296 are denoted by @xmath297 and @xmath298 , respectively . the _ cost _ @xmath299 is defined by @xmath300 next we define the value of a multiflow . a _
tree - embedding _
@xmath301 is a pair of a tree @xmath302 and a family @xmath303 of vertices of @xmath302 indexed by terminal set @xmath12 .
the _ @xmath304-value _
@xmath305 of a multiflow @xmath18 is defined by @xmath306 where @xmath307 denote the ends of an @xmath12-path @xmath308 , and @xmath81 denotes the shortest path metric of @xmath80 with respect to unit edge - length .
we are now ready to define our multiflow problem .
an instance of the problem is a pair of a network @xmath309 and a tree - embedding @xmath310 , and the task is to find a multiflow @xmath56 that maximizes @xmath311 .
this somewhat artificial formulation turns out to be useful , and actually generalizes the original problem .
indeed , take @xmath302 as a star with @xmath312 leaves @xmath313 @xmath314 , let @xmath315 , and let @xmath316 for each edge @xmath296 .
then @xmath311 is twice the total flow - value of @xmath56 . in section [ subsec
: duality ] , we present a combinatorial duality theorem and an optimality criterion . we introduce a nondegeneracy concept of the problem , and give an algorithm to find a half - integral optimal multiflow from a dual optimum under the nondegeneracy assumption .
we also explain how to reduce the original problem to a nondegenerate problem . in section [ subsec : dc ]
, we see that our dual objective can be viewed as an l - convex function on a certain graph structure , and present the steepest descent algorithm ( sda ) to minimize l - convex functions and its iteration bound .
let a pair of @xmath317 and @xmath301 be an instance of the problem .
the vertex set of @xmath302 is also denoted by @xmath302 ( instead of @xmath318 ) .
let @xmath319 denote the edge - subdivision of @xmath302 , where @xmath320 , the edge - length of @xmath319 is defined as @xmath321 uniformly , and the shortest path metric @xmath322 is also denoted by @xmath323 .
a pair @xmath324 of a tree - valued function @xmath325 and a nonnegative half - integer - valued function @xmath326 is called a _ potential _ if it satisfies the following conditions : * for each node @xmath41 , @xmath327 is an integer if and only if @xmath328 belongs to @xmath302 .
* for each edge @xmath78 , it holds @xmath329 . * for each terminal @xmath148 , it holds @xmath330 .
then the following min - max formula and optimality criterion hold : [ thm : duality ] suppose that @xmath331 is even - integer - valued .
the maximum of @xmath311 over all multiflows @xmath56 is equal to the minimum of @xmath332 over all potentials @xmath324 .
[ lem : optimality ] suppose that @xmath331 is even - integer - valued .
a multiflow @xmath18 and a potential @xmath324 are both optimal if and only if they satisfy the following conditions : * for each path @xmath308 in @xmath15 with @xmath333 , it holds @xmath334 * for each edge @xmath78 with @xmath335 , it holds @xmath336 * for each nonterminal node @xmath41 with @xmath337 , it holds @xmath338
. we will use the if part ( and the weak duality in theorem [ thm : duality ] ) only , which is proved for completeness .
( if part ) . for any multiflow @xmath20 and any potential @xmath324 ,
the difference @xmath339 is equal to @xmath340 where we use @xmath341 thus , if @xmath56 and @xmath324 satisfy conditions ( o1 ) , ( o2 ) , and ( o3 ) , then the equality holds in ( [ eqn : weak_duality ] ) , and both @xmath56 and @xmath324 are optimal .
[ [ nondegenerate - case . ] ] nondegenerate case .
+ + + + + + + + + + + + + + + + + + + an instance @xmath342 is said to be _ nondegenerate _ if the edge - costs @xmath331 are positive even - valued and the degree of each node in @xmath302 is at most @xmath205 .
suppose that @xmath342 is nondegenerate .
we further assume , for notational simplicity , that tree @xmath302 has no vertex of degree one ( by attaching paths of infinite length ) .
let @xmath343 and @xmath344 denote the sets of vertices of @xmath302 with degree 2 and 3 , respectively .
for a vertex @xmath345 in @xmath302 , the neighbors of @xmath345 in @xmath302 are denoted by @xmath346 if @xmath347 and @xmath348 if @xmath349 .
let @xmath319 denote the edge - subdivision of @xmath302 . for vertex @xmath350
, the neighbors of @xmath345 in @xmath319 are denoted by @xmath351 if @xmath347 or @xmath352 , and @xmath353 if @xmath349 .
let @xmath354 denote the connected component of @xmath355 containing @xmath356 .
we are going to characterize the flow support of an optimal multiflow .
let @xmath324 be a potential .
motivated by ( o2 ) , define the edge subset @xmath357 by @xmath358 for a nonterminal node @xmath41 , let @xmath359 denote the set of edges @xmath360 with @xmath361 .
since each edge cost @xmath362 is positive , it holds @xmath363 for @xmath360 .
thus @xmath359 for @xmath364 ( or @xmath365 ) partition the set @xmath366 of all edges in @xmath357 incident to @xmath41 .
a _ @xmath324-admissible support _ is a function @xmath367 satisfying the following conditions : * for each nonterminal node @xmath41 with @xmath368 , it holds @xmath369 * for each nonterminal node @xmath41 with @xmath370 , it holds @xmath371 * for each nonterminal node @xmath41 with @xmath337 , it holds @xmath372 * for each edge @xmath296 , @xmath373 is a half - integer , and for each nonterminal node @xmath41 , @xmath374 is an integer . recall the notational convention @xmath375 .
notice that a @xmath324-admissible support is viewed as an edge - weight whose degree vector belongs to a bisubmodular polyhedron described by @xmath376 .
it is not difficult to see from lemma [ lem : optimality ] that for any half - integral optimal multiflow @xmath56 , the flow - support @xmath377 of @xmath56 , defined by @xmath378 , is a @xmath324-admissible support .
the converse also holds .
[ lem : mn ] let @xmath324 be a potential . if a @xmath324-admissible support @xmath377 exists and is given , then @xmath324 is optimal and a half - integral optimal multiflow is obtained in @xmath379 time .
thus our problem is to find a potential @xmath324 so that a @xmath324-admissible support exists .
an algorithm for lemma [ lem : mn ] is the following .
algorithm 1 : : : construction of an optimal multiflow from a @xmath324-admissible support . input : : : a potential @xmath324 and a @xmath324-admissible support @xmath377 .
output : : : a half - integral optimal multiflow @xmath20 .
step 0 : : : @xmath380 .
step 1 : : : choose a terminal @xmath148 and an edge @xmath381 with @xmath382 .
if such a terminal does not exist , then @xmath18 is a half - integral optimal multiflow ; stop .
otherwise let @xmath383 , @xmath384 , @xmath385 , @xmath386 , and go to step 2 .
step 2 : : : if @xmath387 is a terminal , then add path @xmath388 to @xmath15 with flow - value @xmath389 , let @xmath390 for each edge @xmath296 in @xmath308 , and go to step 1 .
otherwise go to step 3 .
step 3 : : : if @xmath391 and @xmath392 for @xmath393 , then choose an edge @xmath394 from @xmath395 with @xmath396 and @xmath397 , and let @xmath398 . + if @xmath399 and @xmath392 for @xmath400 , then choose an edge @xmath401 from @xmath395 with @xmath396 , @xmath397 , and @xmath402 for @xmath403 .
let @xmath404 let @xmath405 and go to step 2 .
this algorithm is essentially the proof of ( * ? ? ?
* lemma 4.5 ) .
let us sketch the correctness of the algorithm ; we show that the resulting multiflow @xmath56 satisfies the conditions ( o1),(o2 ) , and ( o3 ) in lemma [ lem : optimality ] with @xmath324 .
the condition ( o2 ) follows from @xmath406 for @xmath407 . in step 3
, we can always choose a required edge by ( a1 ) and ( a2 ) .
also @xmath377 still satisfies the conditions ( a1 ) , ( a2 ) , and ( a4 ) , thanks to the way of the update .
each produced path @xmath408 satisfies @xmath409 since @xmath410 and @xmath411 for @xmath412 .
also each @xmath413 is positive ( since @xmath331 is positive ) .
since @xmath302 is a tree , we have @xmath414 ; see e.g. , ( * ? ? ?
* lemma 3.9 ) .
thus each produced path satisfies ( o1 ) . by the same argument
, every edge @xmath296 with @xmath415 extends to an @xmath12-path consisting of edges @xmath416 with @xmath417 satisfying ( o1 ) .
this means that if no terminal @xmath148 is chosen in step 1 , then @xmath418 . by ( a3 ) , the resulting multiflow @xmath56 satisfies ( o3 ) .
notice that @xmath419 is a half - integer by ( a4 ) .
hence @xmath56 is half - integral and optimal .
once an @xmath12-path @xmath308 is obtained , @xmath377 becomes zero on some edge , or @xmath420 becomes zero on some node @xmath41 ; they remain zero in subsequent iterations .
thus the algorithm terminates after @xmath421 paths are obtained , where each path is found in @xmath422 time by keeping @xmath423 @xmath424 as lists .
we estimate the range in which an optimal potential exists .
let @xmath425 denote the minimal subtree in @xmath302 containing @xmath426 , and let @xmath427 denote the diameter of @xmath425 , i.e. , @xmath428 .
[ lem : exists ] there is an optimal potential @xmath324 with @xmath429 and @xmath430 for @xmath97 .
let @xmath324 be a potential .
suppose that there is a nonterminal node @xmath431 with @xmath432 .
then there is @xmath433 such that the path between @xmath434 and @xmath435 is outside of @xmath425 , take such @xmath436 with @xmath437 maximum .
we can assume that @xmath438 contains @xmath434 .
let @xmath34 be the set of nodes @xmath77 with @xmath439 .
suppose that @xmath440 .
then @xmath441 for all @xmath442 . for each @xmath442 , replace @xmath443 by @xmath444 . for an edge @xmath78 with @xmath445 and @xmath446 , both @xmath447 and @xmath448 decrease by @xmath321 , and thus ( p2 ) remains to hold .
for other edge @xmath78 , quantity @xmath449 is nonincreasing or remains nonpositive ( if @xmath450 ) .
the feasibility ( p2 ) still holds ( since @xmath362 is nonnegative ) .
thus the resulting @xmath324 is a potential , and the objective value decreases .
suppose that @xmath451 . for each @xmath442
, replace @xmath452 by @xmath453 . for each edge @xmath78 , distance @xmath454 does not increase .
thus the feasibility ( p2 ) holds , and the objective value does not change . by repeating this procedure ,
we can make @xmath324 so that @xmath455 for @xmath97 , without increasing the objective value .
suppose that @xmath456 for some @xmath41 ; necessarily @xmath457 .
for each edge
@xmath78 connecting @xmath41 , @xmath458 ( since @xmath459 ) and @xmath449 is an integer ) .
thus we can replace @xmath460 by @xmath461 to decrease the objective value . repeating this procedure
, @xmath324 satisfies @xmath462 , as required .
[ [ reduction - to - a - nondegenerate - instance . ] ] reduction to a nondegenerate instance .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + here we explain how to reduce our original problem to a nondegenerate problem .
an instance of the original problem is viewed as a pair of network @xmath463 and a tree - embedding @xmath464 such that @xmath465 for all edges @xmath296 and @xmath302 is a star with center @xmath466 and leaves @xmath313 @xmath314 .
we are going to construct a nondegenerate instance .
define edge cost @xmath467 by @xmath468 for each edge @xmath469 .
let @xmath470 .
next we define a tree - embedding @xmath471 .
let @xmath472 be any ( finite ) trivalent tree with @xmath312 leaves @xmath473 @xmath314 and diameter @xmath474 .
for each @xmath433 , consider an infinite path @xmath475 having a vertex @xmath476 of degree one .
identify @xmath473 and @xmath476 , i.e. , glue @xmath475 and @xmath472 at @xmath473 .
the resulting infinite tree is denoted by @xmath477 .
define @xmath434 as the vertex in @xmath475 having distance @xmath478 from @xmath479 .
see figure [ fig : perturbation ] for the construction of @xmath477 .
] now we obtain a nondegenerate instance @xmath480 .
let @xmath481 and @xmath20 be an optimal potential and an optimal multiflow , respectively , for this perturbed instance @xmath480 .
we show that @xmath56 is a maximum multiflow , i.e. , optimal for the original instance @xmath342 .
we are going to construct an optimal potential @xmath324 for @xmath342 from @xmath481 .
let @xmath482 denote the subgraph of @xmath483 induced by vertices @xmath484 with @xmath485 .
namely @xmath482 is ( the intersection of @xmath302 and ) the ball with center @xmath486 and radius @xmath487 .
then it holds @xmath488 for @xmath489 and @xmath433 , let @xmath490 denote the subgraph of @xmath477 consisting of edges @xmath111 in @xmath491 such that @xmath492 .
let @xmath493 be the union of @xmath490 over @xmath433 .
we say that an edge @xmath78 in @xmath11 _ hits _
@xmath493 if @xmath494 and the path between @xmath482 and @xmath495 meets an edge of @xmath493 . for an edge @xmath78 , it holds @xmath496 .
therefore the path between @xmath482 and @xmath495 consists of at most two edges .
since @xmath497 subgraphs @xmath498 are edge - disjoint , there is an index @xmath3 such that every edge in @xmath11 does not hit @xmath493 .
fix such an index @xmath3 . for each @xmath433 , choose the edge @xmath499 of @xmath490 nearest to @xmath434 ( furthest from @xmath473 ) . delete all @xmath499 from @xmath477 .
there are @xmath500 connected components @xmath501 @xmath314 , where @xmath502 is the connected component containing @xmath503 , and @xmath504 is the connected component containing @xmath434 . for each @xmath97 , define @xmath505 by @xmath506 where @xmath507 denotes the vertex in @xmath508 obtained by subdividing edge @xmath509 in @xmath302 .
we show that @xmath324 is a potential for @xmath510 and satisfies ( o1 ) , ( o2 ) , and ( o3 ) with @xmath56 .
we first show the feasibility ( p2 ) @xmath511 for @xmath512 . since @xmath513 implies @xmath514 , we may consider the three cases : ( i ) @xmath515 , @xmath516 , ( ii ) @xmath515 , @xmath517 for @xmath518 , and ( iii ) @xmath515 , @xmath519 for @xmath518 . for ( i )
, @xmath495 necessarily contains two @xmath520 ; otherwise @xmath78 hits @xmath493 at @xmath499 .
this implies @xmath521 .
( ii ) and ( iii ) can not occur since , otherwise , @xmath78 hits @xmath493 at @xmath499 .
thus ( p2 ) holds , and hence @xmath324 is a potential ; ( p1 ) and ( p3 ) are clearly satisfied .
next we show ( o1),(o2 ) , and ( o3 ) for @xmath56 and @xmath324 . to show ( o2 ) , take an edge @xmath78 with @xmath335 .
by ( o2 ) for @xmath56 and @xmath481 in @xmath480 , it holds @xmath522 .
thus the balls @xmath482 and @xmath495 are disjoint and have distance @xmath43 .
suppose that @xmath482 has two edges @xmath520 .
then @xmath482 must meet @xmath490 for every @xmath433 .
necessarily @xmath495 can not be contained by @xmath502 .
also @xmath495 can not have @xmath523 for any @xmath524 ; otherwise @xmath78 hits @xmath493 .
thus @xmath495 is contained in @xmath525 for @xmath526 , and @xmath527 and @xmath528 hold .
if @xmath482 has ( only one ) @xmath529 and @xmath495 has ( only one ) @xmath530 , then @xmath148 and @xmath531 must be different , and necessarily @xmath532 and @xmath533 .
if @xmath482 has only @xmath499 and @xmath495 does not have any of @xmath523 , then necessarily @xmath495 is contained in @xmath504 or @xmath502 ; hence @xmath534 and @xmath535 or @xmath536 .
if both @xmath482 and @xmath495 do not have any of @xmath499 , then both @xmath482 and @xmath495 are contained in @xmath504 for some @xmath537 , and @xmath538 and @xmath539 . in all the cases , it holds @xmath540 , implying ( o2 ) .
consider the condition ( o1 ) .
take a path @xmath541 with @xmath333 .
there is an index @xmath542 such that @xmath543 contains @xmath499 ; otherwise @xmath493 is hit by some edge .
moreover such an index @xmath542 is unique .
otherwise , the balls @xmath544 and @xmath545 with @xmath546 contain @xmath499 .
then @xmath547 .
however , by ( o1 ) and ( o2 ) for @xmath56 and @xmath548 , we have @xmath549 , and @xmath550 ; a contradiction .
similarly there is a unique index @xmath551 such that @xmath552 contains @xmath523 .
if @xmath553 , then @xmath554 must be @xmath555 .
if @xmath556 , say @xmath546 , then @xmath554 must be @xmath557 or @xmath558 .
thus we obtain ( o1 ) . since @xmath559 implies @xmath560 ,
we obtain ( o3 ) .
hence @xmath324 is an optimal potential , and @xmath56 is a maximum multiflow . here we briefly introduce a class of discrete convex functions ( l - convex functions ) on a certain graph structure and the steepest descent algorithm to minimize them .
we then explain that our problem falls into the minimization of an l - convex function .
the full discussion will be given in @xcite .
first we equip the space of all potentials with a graph structure .
let @xmath561 denote the set of half - integers .
let @xmath562 denote the set of pairs @xmath563 such that @xmath564 if and only if @xmath565 .
two points @xmath324 and @xmath566 are adjacent if and only if @xmath567 and @xmath568 are adjacent in @xmath319 and @xmath569 .
fix an arbitrary vertex @xmath570 of @xmath302 .
let @xmath571 ( resp .
@xmath572 ) denote the subset of @xmath562 consisting of pairs @xmath324 with @xmath564 , @xmath565 , and @xmath573 even ( resp .
orient each edge of @xmath562 by @xmath574 if @xmath575 or @xmath576 .
namely @xmath571 is the set of sinks and @xmath572 is the set of sources .
this orientation is acyclic , and induces a partial order @xmath577 on @xmath562 . ]
see figure [ fig : building ] , where nodes in @xmath571 and @xmath572 are colored black and white , respectively .
next we define midpoint operations on @xmath562 .
let @xmath578 denote the edge - subdivision of @xmath319 with edge - length @xmath579 , let @xmath580 denote the set of quarter - integers , and let @xmath581 denote the set of pairs @xmath582 such that @xmath583 if and only if @xmath584 . for two points @xmath585 in @xmath562 , there exists a unique _ midpoint _
@xmath586 such that @xmath587 , @xmath588 , and @xmath589 .
this @xmath49 is denoted by @xmath590 ; accordingly @xmath484 is denoted by @xmath591 . for @xmath592 , there uniquely exists a pair @xmath593 of vertices in @xmath594 with the property that @xmath595 and @xmath596 .
we denote @xmath48 and @xmath49 by @xmath597 and @xmath598 , respectively .
we are ready to define l - convex functions . for a natural number @xmath1 , consider the product @xmath599 ; a point @xmath48 in @xmath599 is represented by a pair @xmath324 of @xmath600 and @xmath601 .
a function @xmath602 is _ l - convex _ if it satisfies the following analogue of the _ discrete midpoint convexity _
* section 7.2 ) : @xmath603 where @xmath604 \rfloor$ ] and @xmath605 for @xmath606 . for @xmath607 , let @xmath608 ( resp .
@xmath609 ) denote the set of points @xmath49 with @xmath610 ( resp .
@xmath611 ) for @xmath606 .
the set @xmath612 is called the _ neighborhood _ of @xmath48 .
the _ steepest descent algorithm _ is given as follows : algorithm 2 : : : steepest descent algorithm ( sda ) input : : : an l - convex function @xmath602 , and a point @xmath613 in @xmath614 .
output : : : a minimizer of @xmath47 .
step 0 : : : let @xmath615 .
step 1 : : : find a minimizer @xmath49 of @xmath47 over the neighborhood @xmath616 of @xmath617 .
step 2 : : : if @xmath618 , then output @xmath617 and stop ; @xmath617 is a minimizer .
step 3 : : : otherwise , let @xmath619 , @xmath620 , and go to step 1 .
the fact that the output is a minimizer easily follows from ( [ eqn : midpoint ] ) ; see ( * ? ? ? * theorem 2.5 ) .
we discuss the number of iterations of this algorithm . for @xmath621 ,
an _ @xmath54-path _ between @xmath48 and @xmath49 is a sequence @xmath622 such that for each @xmath3 and @xmath41 , the @xmath41-th components @xmath623 and @xmath624 belong to a common 4-cycle of @xmath562 , where @xmath2 is called the _ length _ of @xmath308 .
the _ @xmath54-distance _ between @xmath48 and @xmath49 , denoted by @xmath625 , is defined as the minimum length of an @xmath54-path between @xmath48 and @xmath49 . for distinct @xmath626 in a 4-cycle
, it holds @xmath627 . from this
we observe that @xmath628 notice that a sequence @xmath629 generated by sda is an @xmath54-path .
let @xmath630 denote the set of all minimizers of @xmath47 .
the length @xmath2 , which is the number of the iterations , is at least @xmath631 .
this lower bound is almost tight .
[ thm : bound ] let @xmath632 be a sequence of points generated by sda applied to an l - convex function @xmath47 and an initial point @xmath48 .
then @xmath633 . if @xmath634 or @xmath635 , then @xmath636
a similar bound for original l - convex functions in dca was established in @xcite . for a similar but different class of l - convex functions ( called _ alternating l - convex functions _ ) , the same bound was proved by @xcite . by using this result , we give a shorter proof of theorem [ thm : bound ] in appendix .
we now return to our problem .
let a pair of @xmath309 and @xmath301 be an instance .
let @xmath637 .
then the set of potentials @xmath638 is naturally regarded as a subset of @xmath599 .
define a function @xmath639 by @xmath640 if @xmath641 is a potential and @xmath642 otherwise .
[ prop : g_is_l - convex ] suppose that @xmath331 is even - valued .
then @xmath643 is l - convex . in appendix
, we give a shorter proof of this proposition for our case where @xmath331 is positive .
in this section , we prove the main result ( theorem [ thm : main ] ) by presenting an algorithm ( _ dual descent algorithm _ ) to solve a nondegenerate instance of our multiflow problem .
we first show that the optimality check of a potential @xmath324 , or finding a @xmath324-admissible support , is reduced to the submodular flow feasibility problem on a certain skew - symmetric network , called the _
double covering network_. this extends the earlier result on the minimum - cost edge - capacitated multiflow problem by karzanov @xcite , in which the optimality is checked by the classical circulation problem .
partial adaptations of this idea to the node - capacitated setting have been given in @xcite ; but the full adaptation using submodular flow is new . checking the feasibility of a submodular flow
is reduced to the maximum submodular flow problem .
we prove that the minimal minimum cut naturally gives a steepest direction at each potential .
then we obtain a simple descent algorithm mentioned in introduction .
let a pair of @xmath309 and @xmath301 be a nondegenerate instance .
we can assume that there is no edge such that both of its ends are terminals .
let @xmath324 a potential for @xmath342 .
a nonterminal node @xmath41 is said to be _ flat _ if @xmath368 , and _ singular _ if @xmath370 .
a nonterminal node @xmath41 is said to be _
zero _ if @xmath560 and _ positive _ if @xmath337 .
we are going to construct a skew - symmetric network @xmath644 associated with @xmath324 .
we first define a signed node set @xmath645 and edge set @xmath646 indexed by each node @xmath41 in @xmath130 , together with lower edge - capacity @xmath647 and upper edge - capacity @xmath648 . for each terminal @xmath148 , let @xmath649 , and let @xmath650 with @xmath651 and @xmath652 . for each flat node @xmath41 ,
let @xmath653 , and let @xmath654 with @xmath655 .
if @xmath41 is positive , then @xmath656 .
otherwise @xmath657 .
for each positive singular node @xmath41 , let @xmath658 and let @xmath646 consist of @xmath659 , @xmath660 , @xmath661 @xmath662 with @xmath663 , @xmath664 , and @xmath665 @xmath662 .
for each zero singular node @xmath41 , let @xmath666 and let @xmath667 .
define a submodular function @xmath668 on @xmath645 by @xmath669 , where @xmath670 is obtained from @xmath34 by replacing each @xmath671 ( resp .
@xmath672 ) with @xmath673 ( resp .
@xmath674 ) . for an edge @xmath675 with @xmath676 and @xmath677 , add edges @xmath678 and @xmath679 with @xmath680 and @xmath681 . in the case where @xmath77 is a terminal @xmath148 ,
let @xmath682 and @xmath683 .
double covering network _
@xmath684 is defined by the disjoint union of all these edges and nodes .
see figure [ fig : doublecovering ] , where @xmath41 , @xmath77 , @xmath685 , and @xmath686 are positive singular , zero singular , zero flat , and positive flat nodes , respectively . a node @xmath345 in @xmath644
is said to be _ unusual _ if @xmath345 belongs to @xmath645 for a zero singular node @xmath41 .
other nodes @xmath345 are said to be _
usual_. a _ circulation _
@xmath141 in @xmath644 is a function on edge set @xmath131 such that @xmath687 for a circulation @xmath141 , define @xmath688 by @xmath689 suppose that an integral circulation @xmath141 in @xmath644 exists .
then @xmath690 is a @xmath324-admissible support .
the half - integrality of @xmath690 ( in ( a4 ) ) is clear .
take a nonterminal node @xmath41 .
let @xmath691 .
suppose that @xmath41 is flat . then @xmath692 . since @xmath693 and @xmath694 .
we obtain ( a1 ) .
in addition , if @xmath41 is positive , then @xmath695 must hold , and we obtain ( a3 ) . since @xmath696 and @xmath697 are half - integers with @xmath698 , we have ( a4 ) .
suppose that @xmath41 is positive singular .
since @xmath699 , we obtain @xmath700 .
moreover it holds @xmath701 , implying ( a3 ) and ( a4 ) .
if @xmath702 , then @xmath703 ; this contradicts @xmath704 .
therefore @xmath696 , @xmath697 , and @xmath705 satisfy ( a2 ) .
suppose that @xmath41 is zero singular .
notice that @xmath706 , and @xmath707 .
thus @xmath708 ; see ( [ eqn : def_phi ] ) for @xmath709 .
since @xmath710 , because of lemmas [ lem : projection ] and [ lem : delta * ] , the vector @xmath711 satisfies ( a2 ) . since @xmath712
, we have @xmath713 and obtain ( a4 ) by @xmath714 to check the existence of a circulation in @xmath644 , we construct an instance of the maximum submodular flow problem .
add a super source @xmath715 and super sink @xmath716 .
for each edge @xmath717 in @xmath644 having nonzero lower capacity @xmath718 , replace @xmath719 by two edges @xmath720 and @xmath721 with ( upper ) capacity @xmath722 ( and lower capacity @xmath723 ) .
those edges are @xmath724 for positive singular nodes
@xmath41 and @xmath725 for positive flat nodes @xmath41 .
the resulting ( skew - symmetric ) network is denoted by @xmath726 , where modified edge sets are denoted by @xmath727 @xmath728 and the ( upper ) edge - capacity is denoted by @xmath729 . consider the maximum @xmath730-submodular flow problem on @xmath731 , where submodular function @xmath113 on @xmath82 is given as @xmath732 [ lem : a ] if @xmath733 is a minimum @xmath730-cut in @xmath726 , then a circulation @xmath141 in @xmath644 exists , and is obtained from any maximum flow @xmath734 by the following procedure : * for each edge @xmath735 in @xmath644 having nonzero lower capacity , let @xmath736 , and for other edge @xmath296 in @xmath644 , let @xmath737 . indeed , in this case , any max - flow saturates all edges leaving @xmath715 and all edges entering @xmath716 , and consequently the resulting @xmath141 satisfies @xmath738 for all edges @xmath296 having nonzero lower - capacity , and is a circulation of @xmath644 .
next we show that the minimal minimum cut gives rise to a steepest descent direction of @xmath643 at @xmath324 .
an @xmath730-cut @xmath34 is said to be _ normal _ if it satisfies the following conditions : * @xmath34 does not meet @xmath739 for any terminal @xmath433 .
* @xmath34 is a transversal .
* for each positive node @xmath41 , @xmath740 is equal to @xmath741 , @xmath742 , or @xmath743 , and @xmath744 is equal to @xmath741 , @xmath745 , or @xmath746 .
* for each zero node @xmath41 , @xmath747 is equal to @xmath741 , @xmath742 , @xmath743 , or @xmath748 .
then the following holds ; the proof is given in the end of this subsection .
[ lem : normal_mincut ] a normal minimum @xmath730-cut exists , and is obtained from the unique minimal minimum cut @xmath34 by applying the following procedure : * for each singular node @xmath41 , if @xmath749 , then replace @xmath34 by @xmath750 .
let @xmath751 be the subset of @xmath82 consisting of @xmath752 for @xmath24 and @xmath400 such that @xmath753 , or @xmath754 and @xmath755 .
let @xmath756 be the node subset defined by replacing @xmath572 with @xmath571 in the definition of @xmath751 .
a normal cut @xmath34 is said to be _
@xmath757-normal _ if @xmath758 , and is said to be _ of @xmath759-normal _ if @xmath760 . for an @xmath757-normal or @xmath759-normal cut @xmath34 ,
define @xmath761 by @xmath762 for each nonterminal node @xmath24 and @xmath763 for each terminal @xmath433 .
see figure [ fig : neighborhood ] for the correspondence between @xmath764 and @xmath765 , where nodes in @xmath765 are colored black .
[ prop : steepest ] let @xmath34 be a normal minimum @xmath730-cut in @xmath726 , and let @xmath766 and @xmath767 . then @xmath768 is a minimizer of @xmath643 over @xmath769 , and @xmath770 is a minimizer of @xmath643 over @xmath771 . the proof is given in the end .
we are now ready to describe our algorithm .
algorithm 3 : : : dual descent algorithm ( for a nondegenerate instance ) input : : : a nondegenerate instance @xmath309 , @xmath301 .
output : : : a half - integral optimal multiflow @xmath56 .
step 0 : : : choose any vertex @xmath345 in @xmath425 , and let @xmath772 be a potential defined by @xmath773 for @xmath24 and @xmath774 for @xmath433 .
step 1 : : : construct network @xmath726 . obtain an integral maximum @xmath730-flow @xmath734 and the minimal minimum @xmath730-cut @xmath34 in @xmath726 by a submodular flow algorithm .
make @xmath34 normal by procedure ( b ) .
step 2 : : : if @xmath775 , then @xmath324 is optimal , obtain a feasible flow @xmath141 in @xmath644 from @xmath734 by procedure ( a ) . obtain a @xmath324-admissible support @xmath690 by ( [ eqn : zeta_phi ] ) , and obtain a half - integral optimal multiflow @xmath56 from @xmath690 by algorithm 1 ; stop .
step 3 : : : choose @xmath776 with @xmath777 .
let @xmath778 , and go to step 1 .
[ thm : bound2 ] the dual descent algorithm runs in @xmath779 time . by proposition [ prop : steepest ] , @xmath780 is a steepest direction of @xmath643 at @xmath324 .
hence the dual descent algorithm is viewed as the steepest descent algorithm applied to @xmath781 that is l - convex ( proposition [ prop : g_is_l - convex ] ) . by lemma [ lem :
exists ] and theorem [ thm : bound ] , the number of iteration is bounded by @xmath427 .
our submodular function @xmath113 is a disjoint sum of submodular functions @xmath668 for zero singular nodes @xmath41 ( see ( [ eqn : our_submo ] ) ) . so the exchange capacity @xmath782 for a pair of @xmath783 and @xmath345 can take positive if @xmath784 belong to @xmath645 for some zero singular node @xmath41 , and is equal to the exchange capacity for submodular function @xmath668 on a 6-element set @xmath645 .
hence this can be computed in constant time .
the number of nodes of @xmath726 is at most @xmath785 , and the number of edges is at most @xmath786 .
thus step 2 is done in @xmath787 time .
[ [ proof - of - the - main - result - theoremthmmain . ] ] proof of the main result ( theorem [ thm : main ] ) .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + as in section [ sec : multiflow ] , construct nondegenerate instance @xmath480 . then @xmath788 .
apply the dual descent algorithm for @xmath480 . by theorem [ thm : bound2 ]
, we obtain an optimal potential @xmath481 and an optimal multiflow @xmath56 in @xmath789 time . as we have shown in section [ sec : multiflow ] , @xmath56 is also a maximum multiflow , and an optimal potential @xmath324 for the original instance is obtained from @xmath548 ( in @xmath790 time ) . then @xmath791 is a half - integral optimal solution of lp - dual ( [ eqn : lp - dual ] ) .
indeed , @xmath792 is equal to the maximum flow - value .
the feasibility of @xmath791 follows from @xmath793 for every @xmath12-path @xmath308 .
[ [ proof - of - lemmalemnormal_mincut . ] ] proof of lemma [ lem : normal_mincut ] .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath34 be the unique minimal minimum @xmath730-cut ( of finite cut capacity ) .
by lemma [ lem : minimal_minimum ] , @xmath34 is a transversal , implying ( c1 ) .
also @xmath34 can not meet @xmath739 for any terminal @xmath148 .
otherwise , @xmath34 contains @xmath794 ( and does not contain @xmath795 ) .
however the deletion of @xmath794 from @xmath34 does not increase the cut capacity , contradicting the minimality .
consider a zero flat node @xmath41 .
suppose that @xmath796 . if @xmath797 @xmath798 , then the deletion of @xmath672 from @xmath34 decreases the capacity , contradicting the minimality .
thus we have ( c3 ) .
consider a zero singular node @xmath41 .
suppose that @xmath799 .
if @xmath800 , then @xmath747 is of type 1 , 3 , or 5 , the change @xmath801 preserves the type at @xmath645 , and the cut capacity ; contradicting the minimality .
similarly , if @xmath802 , then the change @xmath803 preserves the cut capacity ; a contradiction .
thus @xmath804 .
so suppose that @xmath805 or @xmath806 . if @xmath807 , then necessarily @xmath808 .
suppose that @xmath805 .
if @xmath809 ( type 1 ) , then the change @xmath810 ( procedure ( b ) ) does not increase the cut capacity ( and keeps @xmath34 being a transversal ) . in particular , the resulting @xmath747 is one of the patterns in ( c3 ) .
consider a positive singular node @xmath41 .
first we show that @xmath811 .
suppose to the contrary that @xmath812 .
the the change @xmath813 preserves the capacity ; a contradiction to the minimality .
also @xmath814 is impossible , and @xmath815 implies @xmath816 .
thus the pattern of @xmath817 is one given in ( c2 ) .
next consider @xmath740 .
if @xmath34 contains two nodes in @xmath818 , and necessarily @xmath819 ; the change @xmath810 ( procedure ( b ) ) keeps @xmath34 being a transversal , and does not increase the cut capacity .
if @xmath34 contains @xmath820 , then @xmath34 contains at least two nodes in @xmath742 , reduced to the case above .
thus , after procedure ( b ) , the resulting cut is a normal minimum cut , as required . [ [ proof - of - propositionpropsteepest . ] ] proof of proposition [ prop : steepest ] .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in @xmath562 , vertices in @xmath571 are colored black , and vertices in @xmath572 are colored white .
other vertices have no color . for @xmath821 and @xmath822 with @xmath823
, let @xmath824 for the unique neighbor @xmath783 of @xmath567 with @xmath825 .
in addition , if @xmath826 , then let @xmath827 for the ( unique ) neighbor @xmath783 of @xmath567 with @xmath828 .
[ lem : parity ] for @xmath829 with @xmath830 , @xmath831 is even if and only if one of following pairs has the same color : @xmath832 for @xmath833 , observe that @xmath834 .
hence @xmath835 is even if and only if @xmath324 and @xmath566 have the same color .
also observe that @xmath831 does not change when @xmath324 is replaced by @xmath836 .
the claim follows from these facts .
[ lem : noedge ] there is no edge between @xmath756 and @xmath751 . in particular ,
any normal cut @xmath34 is decomposed into @xmath757-normal cut @xmath766 and @xmath759-normal cut @xmath767 such that @xmath837 edge @xmath838 ( or @xmath839 ) in @xmath726 belongs to @xmath645 for node @xmath41 of colored @xmath840 , and hence belongs to @xmath756 or @xmath751 .
so consider an edge @xmath841 for distinct @xmath842 .
then @xmath843 ( since @xmath844 ) and @xmath845 is even ( since @xmath846 is even ) . by the previous lemma ,
if both @xmath847 and @xmath452 are colored , then they have the same color .
suppose that @xmath840 has no color . if @xmath443 has a color , then @xmath848 has the same color .
if @xmath443 has no color , then @xmath849 and @xmath850 have the same color .
consequently @xmath851 or @xmath852 for all cases .
let @xmath853 ( resp .
@xmath854 ) denote the subset of @xmath769 ( resp .
@xmath771 ) consisting @xmath324 with @xmath855 for @xmath606 .
proposition [ prop : steepest ] follows from the above lemma [ lem : noedge ] and the following .
* the map @xmath856 is a bijection between @xmath853 ( resp .
@xmath854 ) and the set of all @xmath757-normal cuts ( resp .
@xmath759-normal cuts ) .
* for an @xmath757-normal or @xmath759-normal cut @xmath34 , it holds @xmath857 ( 1 ) .
we claim that @xmath858 and @xmath859 are in one - to - one correspondence by ( [ eqn : x ] ) .
suppose that @xmath860 , which implies @xmath861 . then @xmath862 and @xmath863 .
thus the claim is true .
suppose that @xmath864 , which implies @xmath865 .
the claim can be seen from figure [ fig : neighborhood ] .
suppose that @xmath866 ; necessarily @xmath41 is positive .
we can assume that @xmath867 ( and @xmath868 )
. then @xmath869 .
since @xmath870 and @xmath871 , it holds that @xmath872 .
thus we have the claim , implying the statement ( 1 ) .
let @xmath873 .
we first show that @xmath874 if and only if @xmath875 is finite .
suppose that @xmath874 .
pick @xmath796 and edge @xmath876 .
we show @xmath877 .
recall that @xmath878 and @xmath844 .
hence @xmath879 .
since @xmath796 , the change @xmath880 increases @xmath881 by one .
necessarily the change @xmath882 must decrease @xmath883 by one .
this means that @xmath884 , @xmath885 or @xmath886 must hold . for all cases
, @xmath34 contains @xmath887 .
suppose that @xmath875 is finite .
we show that @xmath874 . here
@xmath888 is clear .
we need to show that @xmath889 for each edge @xmath512 .
suppose that @xmath890 .
namely @xmath360 and @xmath891 . in this case
, the argument is the same as the above . indeed
, suppose that the change @xmath892 increases @xmath893 by one .
then @xmath34 contains @xmath894 with @xmath895 , and hence contains @xmath896 with @xmath897 .
the change @xmath898 decreases @xmath899 by one , implying @xmath900 .
thus it suffices to show that @xmath901 can not occur .
otherwise , @xmath902 , @xmath893 is odd , and @xmath903 , @xmath904 .
if both @xmath840 and @xmath905 have colors , then these colors are different ( by lemma [ lem : parity ] ) , and @xmath906 or @xmath907 must hold ; a contradiction .
suppose that @xmath847 has no color
. then @xmath908 must hold .
if @xmath909 has a color , then this color must equal the color of @xmath910 ( since @xmath911 and @xmath912 have different colors ) , and hence @xmath913 ; a contradiction .
if @xmath909 has no color , then @xmath914 must hold ; but this is impossible since colors of @xmath911 and @xmath915 are different .
finally we show the equation ( [ eqn : g_ne(p , r)^x ] ) .
the left hand side is equal to @xmath916 and the right hand side is equal to @xmath917 one can verify from the network construction ( see figure [ fig : doublecovering ] ) , definitions of @xmath918 ( see ( [ eqn : x ] ) ) and @xmath919 ( see ( [ eqn : delta * ] ) ) that @xmath920 is equal to @xmath921 if @xmath41 is positive , @xmath922 if @xmath41 is zero flat , and @xmath923 if @xmath41 is zero singular .
thus we obtain the equation ( [ eqn : g_ne(p , r)^x ] ) .
we thank satoru iwata for information on submodular flow , satoru fujishige for helpful comments , and yuni iwamasa for computation of @xmath242 . the work was partially supported by jsps kakenhi grant numbers 25280004 , 26330023 , 26280004 . 1 k. ando , s. fujishige , and t. naitoh , balanced bisubmodular systems and bidirected flows , _ journal of the operations research society of japan _ * 40 * ( 1997 ) , 437447 .
m. a. babenko , a fast algorithm for the path 2-packing problem , _ theory of computing systems _ * 46 * ( 2010 ) , 5979 . m. a. babenko and a. v. karzanov , a scaling algorithm for the maximum node - capacitated multiflow problem , in : _ proceedings of 16th annual european symposium on algorithms _ ( esa08 ) , _ lncs _ * 5193 * ( 2008 ) , 124135
. m. a. babenko and a. v. karzanov , min - cost multiflows in node - capacitated undirected networks , _ journal of combinatorial optimization _ * 24 * ( 2012 ) , 202228 . c. chekuri and v. madan , simple and fast rounding algorithms for directed and node - weighted multiway cut , in : _ proceeding of acm - siam symposium on discrete algorithms ( soda16 ) _ , to appear , arxiv:1507.04674 .
m. chudnovsky , w. h. cunningham , and j. geelen , an algorithm for packing non - zero @xmath131-paths in group - labelled graphs , _ combinatorica _
* 28 * ( 2008 ) , 145161 .
p. m. dearing , r. l. francis , and t. j. lowe , convex location problems on tree networks , _ operations research _ * 24 * ( 1976 ) , 628642 .
a. frank , _ connections in combinatorial optimization _ , oxford university press , oxford , 2011 .
s. fujishige , _ submodular functions and optimization , 2nd edition _ , elsevier , amsterdam , 2005 .
s. fujishige and s. iwata , algorithms for submodular flows , _ ieice transactions on information and systems _ * 83 * ( 2000 ) , 322329 .
s. fujishige and x. zhang , new algorithms for the intersection problem of submodular systems , _ japan journal of industrial and applied mathematics _ * 9 * ( 1992 ) , 369382 .
n. garg , v. v. vazirani , and m. yannakakis , multiway cuts in node weighted graphs , _ journal of algorithms _ * 50 * ( 2004 ) 4961 .
n. garg and j. knemann , faster and simpler algorithms for multicommodity flow and other fractional packing problems , _ siam journal on computing _ * 37 * ( 2007 ) , 630652 . a. v. goldberg , and a. v. karzanov , path problems in skew - symmetric graphs , _ combinatorica _ * 16 * ( 1996 ) , 353382 .
h. hirai , half - integrality of node - capacitated multiflows and tree - shaped facility locations on trees , _ mathematical programming ,
series a _ * 137 * ( 2013 ) , 503530 .
h. hirai , discrete convexity and polynomial solvability in minimum 0-extension problems , _ mathematical programming , series a _ , to appear .
( the conference version in _ soda13 _ ) h. hirai , l - extendable functions and a proximity scaling algorithm for minimum cost multiflow problem , _ discrete optimization _ * 18 * ( 2015 ) , 137 . h. hirai , l - convexity on graph structures , in preparation
karzanov , a minimum cost maximum multiflow problem , in : _ combinatorial methods for flow problems _ , institute for system studies , moscow , 1979 , pp . 96103 ( in russian ) .
a. v. karzanov , minimum cost multiflows in undirected networks , _ mathematical programming , series a _ * 66 * ( 1994 ) , 313324 .
w. mader , ber die maximalzahl kreuzungsfreier @xmath924-wege , _ archiv der mathematik _ * 31 * ( 1978/79 ) , 387402 .
k. murota , _ discrete convex analysis _ ,
siam , philadelphia , 2003 .
k. murota and a. shioura , exact bounds for steepest descent algorithms of l - convex function minimization , _ operations research letters _ * 42 * ( 2014 ) , 361366 .
g. pap , some new results on node - capacitated packing of a - paths , in : _ proceedings of the 39th annual acm symposium on theory of computing ( stoc07 ) _ , acm , new york , 2007 , pp . 599604 .
g. pap , strongly polynomial time solvability of integral and half - integral node - capacitate multiflow problems , _ egres technical report _ ,
tr-2008 - 12 , ( 2008 ) .
a. schrijver , _ combinatorial optimization polyhedra and efficiency _ , springer - verlag , berlin , 2003 . v. v. vazirani , _ approximation algorithms _ , springer - verlag , berlin , 2001 .
for @xmath925 , the first and second components of @xmath48 are denoted by @xmath926 and @xmath927 , respectively , i.e. , @xmath928 .
first consider the case where @xmath302 is a path ( with no ends ) .
then @xmath562 is isomorphic to the product of two zigzagly - oriented paths . in this case , @xmath929 is identified with the product @xmath930 of @xmath931 paths , and l - convex functions on @xmath929 coincide with _ alternating l - convex functions _ in the sense of @xcite .
also @xmath932 is equal to @xmath323 in @xcite
. then theorem [ thm : bound ] was shown in ( * ? ? ?
* theorem 2.6 ) .
in particular , the following holds : we show that this proposition holds for a general tree @xmath302 .
pick an arbitrary @xmath940 with @xmath941 .
since @xmath942 and @xmath943 are equal or adjacent in @xmath302 or @xmath319 , there is a path @xmath944 in @xmath319 containing @xmath945 , and @xmath946 .
then @xmath947 , and @xmath934 are points in @xmath948 , and the restriction @xmath949 of @xmath47 to @xmath950 is ( alternating ) l - convex . thus proposition [ prop : d_inf = d_inf-1 ] is applicable to @xmath951 , and hence @xmath952 .
by @xmath953 , the equation ( [ eqn : d_inf = d_inf-1 ] ) holds .
theorem [ thm : bound ] is proved as follows . by the description of the steepest descent algorithm
, it holds that @xmath954 or @xmath955 for all @xmath956 . thus @xmath957 holds . since @xmath958
, we obtain @xmath633 .
in addition , if @xmath937 or @xmath938 , then @xmath959 holds .
let @xmath971 and @xmath972 .
then it holds that @xmath973 .
hence we can assume that @xmath974 .
by @xmath975 and @xmath976 , it must hold @xmath977 then there are two edge - disjoint paths in @xmath578 , one containing @xmath978 and the other containing @xmath979 .
also they are contained by a single path . from this ,
we see @xmath980 also @xmath981 and @xmath982
. hence we obtain @xmath983 .
notice that both @xmath984 and @xmath985 are integers .
if @xmath986 , then @xmath987 , and hence @xmath988 and @xmath970 .
so it suffices to show that @xmath989 and @xmath990 can not occur .
suppose not .
then both @xmath984 and @xmath985 are odd .
if both @xmath991 and @xmath992 ( or @xmath993 and @xmath994 ) are colored , then they have different colors ( lemma [ lem : parity ] ) , @xmath995 or @xmath996 holds , and @xmath990 is impossible .
hence we can assume that @xmath997 , and @xmath998 or @xmath999 .
then @xmath1000 , @xmath1001 , and @xmath1002 . hence @xmath1003 ( even )
however @xmath998 or @xmath999 implies that @xmath1004 or @xmath1005 is odd ; a contradiction .
we are ready to prove proposition [ prop : g_is_l - convex ] .
it is not difficult to see that the function on @xmath562 defined by @xmath1006 if @xmath1007 and @xmath62 otherwise is l - convex .
so it suffices to show that the function on @xmath1008 defined by @xmath1009 if @xmath968 and @xmath62 otherwise is l - convex .
we show that @xmath968 and @xmath1010 imply @xmath1011 and @xmath1012 . by lemma [ lem : h_is_convex ]
, we have @xmath1013 . by lemma [ lem : h(x , y)<=0 ] , we have the desired results . | in this paper , we develop an @xmath0-time algorithm to find a half - integral node - capacitated multiflow of the maximum total flow - value in a network with @xmath1 nodes , @xmath2 edges , and @xmath3 terminals , where @xmath4 denotes the time complexity of solving the maximum submodular flow problem in a network with @xmath5 edges , @xmath6 edges , and the complexity @xmath7 of computing the exchange capacity of the submodular function describing the problem . by using fujishige - zhang algorithm for submodular flow , we can find a maximum half - integral multiflow in @xmath8 time .
this is the first combinatorial strongly polynomial time algorithm for this problem .
our algorithm is designed on the basis of a developing theory of discrete convex functions on certain graph structures .
applications include ellipsoid - free " combinatorial implementations of a 2-approximation algorithm for the minimum node - multiway cut problem by garg , vazirani , and yannakakis .
keywords : node - capacitated multiflow , discrete convex analysis , submodular flow , node - multiway cut |
the well - founded semantics was introduced in @xcite to provide 3-valued interpretations to logic programs with negation . since its introduction , the well - founded semantics has become one of the most widely studied and most commonly accepted approaches to negation in logic programming @xcite .
it was implemented in several top - down reasoning systems , most prominent of which is xsb @xcite .
the well - founded semantics is closely related to the stable - model semantics @xcite , another major approach to logic programs with negation .
the well - founded semantics approximates the stable - model semantics @xcite .
moreover , computing the well - founded model of propositional programs is polynomial @xcite while computing stable models is np - hard @xcite .
consequently , evaluating the well - founded semantics can be used as an effective preprocessing technique in algorithms to compute stable models @xcite .
in addition , as demonstrated by smodels @xcite , at present the most advanced and most efficient system to compute stable models of datalog@xmath3 programs , the well - founded semantics can be used as a powerful lookahead mechanism . despite the importance of the well - founded semantics , the question of how fast it can be computed
has not attracted significant attention .
van gelder @xcite described the so called _ alternating - fixpoint _ algorithm .
van gelder s algorithm runs in time @xmath0 , where @xmath4 is the set of atoms occurring in a logic program @xmath2 , @xmath5 denotes the cardinality of @xmath4 , and @xmath1 is the size of @xmath2 ( the total number of atom occurrences in @xmath2 ) . improving on
this algorithm turned out to be difficult .
the first progress was obtained in @xcite .
the algorithm described there , when restricted to programs whose rules contain at most two positive occurrences of atoms in their bodies , runs in time @xmath6 , where @xmath7 is the number of rules in @xmath2 . for programs
whose rules have no more than one positive atom in the body a better estimate of @xmath8 was obtained . for some classes of programs
this is an asymptotically better estimate than the @xmath0 estimate that holds for the algorithm by van gelder .
a different approach to computing the well - founded model was proposed in @xcite .
it is based on the notion of a program transformation @xcite .
the authors describe there several transformations that can be implemented in linear time and that simplify a program while ( essentially ) preserving the well - founded semantics .
these transformations are : the positive reduction , success , negative reduction , and failure ( psnf transformations , for short ) .
they allow one to compute in linear time the kripke - kleene semantics @xcite of the program . to compute
the well - founded semantics one also needs to detect the so - called _ positive loops_. the complexity of this task dominates the asymptotic complexity of the well - founded semantics computation .
no improved algorithms for the positive - loop detection are offered in @xcite so the worst - case asymptotic complexity of the algorithm presented there remains the same as that of the alternating - fixpoint method . however , due to the use of psnf transformations , that simplify the program , the algorithm based on program transformations may in practice run faster .
in contrast to the approach studied in @xcite , we focus here on the positive - loop detection task .
the alternating - fixpoint algorithm works by successively improving lower approximations @xmath9 and @xmath10 to the sets of atoms that are true and false ( under the well - founded semantics ) , respectively .
the algorithm starts with @xmath11 . using this estimate
, it computes the first estimate for @xmath10 .
next , using this estimate , in turn , it computes a better estimate for @xmath9 .
the algorithm continues until further improvements are not possible .
it returns the final sets @xmath9 and @xmath10 as the well - founded semantics .
a dual version of the alternating - fixpoint algorithm , starting with @xmath12 and then alternatingly computing approximations to @xmath9 and @xmath10 , is also possible .
the most time - consuming part of this algorithm is in computing estimates to the set of atoms that are false ( in this part , in particular , positive loops are detected ) . in the van gelder algorithm ,
the best possible approximation ( given the current estimate for @xmath9 ) is always computed by using a bottom - up approach . in this paper
we focus on the problem of detecting positive loops and computing new false atoms .
we restrict our attention to the class of programs that have at most one positive atom in the body .
we denote this class of programs by @xmath13 .
we show that for programs from @xmath13 , false atoms can be computed by means of a top - down approach by finding atoms that do not have a proof .
moreover , we show that it is not necessary to find _ all _ atoms that can be established to be false at a given stage .
finding a proper subset ( as long as it is not empty ) is also sufficient and results in a correct algorithm .
we apply these techniques to design a version of an alternating - fixpoint algorithm computing the well - founded semantics of programs from the class @xmath13 .
we demonstrate that the resulting algorithm is asymptotically better than the original alternating - fixpoint algorithm by van gelder .
specifically , we show that our algorithm runs in time @xmath14 .
thus , for programs with @xmath15 , the asymptotic estimate of the running time of our algorithm is better than that of algorithms by van gelder @xcite and berman et al .
@xcite .
as mentioned above , our approach is restricted to the class @xmath13 .
applicability of our method can , however , be slightly extended .
let us denote by @xmath16 the class of these logic programs that , after simplifying by means of psnf transformations ( or , equivalently , with respect to the kripke - kleene semantics ) fall into the class @xmath13 .
since psnf transformations ( the kripke - kleene semantics ) can be computed in linear time , the asymptotic estimate of the running time of our method extends to all programs in the class @xmath16 .
the paper is organized as follows . in the next section we provide a brief review of the key notions and terminology . in section [ algs ]
we describe several modifications to the original van gelder algorithm , we show their correctness and estimate their running time .
the ultimate effect of our considerations there is a general template for an algorithm to compute the well - founded semantics .
any algorithm computing some ( not necessarily all ) atoms that can be established as false given a current estimate to the well - founded can be used with it .
one such algorithm , for programs from the class @xmath13 , is described and analyzed in section [ cluster ] .
it constitutes the main contribution of the paper and yields a new , currently asymptotically most efficient algorithm for computing the well - founded semantics for programs in @xmath13 .
the last section contains conclusions .
we start by reviewing basic concepts and notation related to logic programs and the well - founded semantics , as well as some simple auxiliary results . in the paper
we consider the propositional case only .
let @xmath2 be a normal logic program .
by @xmath4 we denote the set of atoms occurring in @xmath2 .
let @xmath17 ( throughout the paper we often drop a reference to @xmath2 from our notation , whenever there is no danger of ambiguity ) . by @xmath18
we denote the program obtained from @xmath2 by removing all rules whose bodies contain negated literals of the form @xmath19 , where @xmath20 .
further , by @xmath21 we denote the program obtained from @xmath2 by removing from the bodies of its rules _ all _ negative literals . clearly , the program @xmath22 coincides with the _ gelfond - lifschitz _ reduct of @xmath2 with respect to @xmath23 ( throughout the paper , we write @xmath24 for @xmath22 , to simplify notation ) .
the _ gelfond - lifschitz _ operator on the algebra of all subsets of @xmath25 , @xmath26 ( following our convention , we omit the reference to @xmath2 from the notation ) , is defined by @xmath27 where @xmath28 stands for a least model of a horn program @xmath29 .
we now present characterizations of the well - founded semantics .
we phrase them in the language of operators and their fixpoints .
all operators considered here are defined on the algebra of subsets of @xmath4 .
we denote a least fixpoint ( if it exists ) of an operator @xmath30 by @xmath31 .
it is well known that @xmath26 is antimonotone .
consequently , @xmath32 is monotone and has a least fixpoint .
the set of atoms that are true with respect to the well - founded semantics of a program @xmath2 , denoted by @xmath33 , is precisely the least fixpoint of the operator @xmath34 , that is , @xmath35 @xcite .
the set of atoms that are false with respect to the well - founded semantics of a program @xmath2 , denoted by @xmath36 , is given by @xmath37 ( throughout the paper , @xmath38 denotes the complement of a set @xmath39 with respect to @xmath4 ) .
one can define a dual operator to @xmath34 by @xmath40 it is easy to see that @xmath41 is monotone and that its least fixpoint is @xmath36 .
thus , @xmath42 and @xmath43 .
we close this section by discussing ways to compute @xmath44 for a given finite propositional logic program @xmath2 and a set of atoms @xmath45 .
a straightforward approach is to compute the gelfond - lifschitz reduct @xmath24 and then to compute its least model .
the resulting algorithm is asymptotically optimal as it runs in time linear in the size of the program .
however , in this paper we will use a different approach , more appropriate for the computation of the well - founded semantics .
let @xmath2 be a logic program with negation .
we define @xmath46 . for every set @xmath47
, we define @xmath48 . if we interpret literals of @xmath49 as new _ atoms _ , then for every set @xmath17 , the program @xmath50 can be viewed as a horn program .
thus , it has a least model .
it is easy to see that @xmath51 here , @xmath2 appearing at the left - hand side of the equation stands for the original logic program , while @xmath2 appearing at the right - hand side of the equation stands for the same program but interpreted as a horn program .
thus , using the algorithm of dowling and gallier @xcite , the gelfond - lifschitz reduct can be computed in time @xmath52 ( since @xmath17 , @xmath53 ) .
the departure point for our discussion of algorithms to compute the well - founded semantics is the _ alternating - fixpoint _ algorithm of van gelder @xcite . using the terminology introduced in the previous section
it can be formulated as follows .
= = = = = = = = = = + * algorithm 1 ( van gelder ) + @xmath54 ; + + @xmath55 ; ( * or equivalently : @xmath56 ; * ) + @xmath57 ; ( * or equivalently : @xmath58 ; * ) + no change in @xmath10 ; + @xmath9 and @xmath10 . *
let @xmath59 and @xmath60 be the values of the set @xmath10 just before and just after an iteration of the * repeat * loop in algorithm 1 . clearly , @xmath61 thus , after iteration @xmath62 of the * repeat * loop , @xmath63 .
consequently , it follows from our earlier remarks that when algorithm 1 terminates , the set @xmath10 that is returned satisfies @xmath64 .
since there is no change in @xmath10 in the last iteration , when the algorithm terminates , we have @xmath65 .
that is , algorithm 1 is correct .
we will now modify algorithm 1 .
the basis for algorithm 1 is the operator @xmath41 .
this operator is not _
progressive_. that is , @xmath23 is not necessarily a subset of @xmath66
. we will now introduce a related progressive operator , say @xmath67 , and show that it can be used to replace @xmath41 .
let @xmath2 be a logic program and let @xmath9 and @xmath10 be two subsets of @xmath4 . by @xmath68
we denote the program obtained from @xmath2 by removing 1 .
all rules whose heads are in @xmath10 2 .
all rules whose bodies contain a positive occurrence of an atom from @xmath10 3 .
all rules whose bodies contain a negated literal of the form @xmath19 , where @xmath69 .
clearly , @xmath70 .
we define an operator @xmath71 as follows : @xmath72 where @xmath73 and @xmath74 abbreviates @xmath75 .
the following result gathers key properties of the operator @xmath67 .
[ l1a ] let @xmath2 be a normal logic program .
then : 1 .
[ l11 ] @xmath67 is monotone 2 .
[ l12 ] for every @xmath76 , @xmath77 3 .
[ l13 ] for every @xmath78 , @xmath79 4 .
[ l15 ] @xmath80 5 .
[ l14 ] for every @xmath76 , @xmath81 , where @xmath73 .
proof : ( [ l11 ] ) assume that @xmath82 .
set @xmath83 , @xmath84 .
clearly , @xmath85 and , by antimonotonicity of @xmath26 , @xmath86 . by the definition of @xmath68 ,
consequently , @xmath88 and , so , @xmath89 .
+ ( [ l12 ] ) let @xmath73 .
clearly , @xmath70 .
thus , @xmath90 .
+ ( [ l13 ] ) we have , @xmath91 .
it follows that removing from @xmath92 rules with heads in @xmath93 and those that contain an atom from @xmath93 in their bodies does not change the least model .
that is , @xmath94 since , @xmath95 , @xmath96 .
let @xmath97 .
then , by ( [ l11 ] ) , @xmath98
. thus , we have @xmath99 ( [ l15 ] ) the least fixpoint of @xmath67 is given by @xmath100 . by ( [ l13 ] ) , @xmath101 .
on the other hand , by ( [ l11 ] ) and ( [ l12 ] ) , @xmath102 .
thus , @xmath103 .
it follows that @xmath104 .
( [ l14 ] ) let @xmath73
. since @xmath68 has no rules with head in @xmath10 , @xmath105 and , consequently , @xmath106 .
thus , the assertion follows .
@xmath107 theorem [ l1a ] allows us to prove the correctness of the following modification of algorithm 1 .
= = = = = = = = = = + * algorithm 2 + @xmath54 ; + + @xmath55 ; + @xmath108 ; + @xmath109 ; + no change in @xmath10 ; + @xmath9 and @xmath10 . * by theorem [ l1a ] , each iteration of the * repeat * loop computes @xmath71 as the new value for the set @xmath10 . more formally ,
the set @xmath10 just after iteration @xmath62 , satisfies @xmath110 .
thus , when the algorithm terminates , the set @xmath10 that is returned is the least fixpoint of @xmath67 .
consequently , by theorem [ l1a]([l15 ] ) , algorithm 2 is correct
. we will now modify algorithm 2 to obtain a general template for an alternating - fixpoint algorithm to compute the well - founded semantics .
the key idea is to observe that it is enough to compute a subset of @xmath111 in each iteration and the algorithm remains correct .
let us assume that for some operator @xmath112 defined for pairs @xmath113 , where @xmath114 and @xmath29 is a horn program such that @xmath115 ( the complement is , as always , evaluated with respect to @xmath4 ) , we have : ( w1 ) : : @xmath116 ( w2 ) : : @xmath117 if and only if @xmath118 .
let @xmath119 . by the definition of @xmath68 ,
thus , we define @xmath121 , where @xmath122 .
it is clear that for every @xmath76 , @xmath123 , the latter inclusion follows from theorem [ l1a]([l14 ] ) and ( w1 ) .
consequently , for every @xmath62 , @xmath124 it follows that @xmath125 .
it also follows that there is the first @xmath62 such that @xmath126 .
let us denote this set @xmath127 by @xmath128
. then @xmath129 . in the same time
, by condition ( w2 ) , @xmath130 . since @xmath36 is the least fixpoint of @xmath67 , @xmath131 .
it follows that a modification of algorithm 2 in which line @xmath132 is replaced by @xmath133 correctly computes the well - founded semantics of a program @xmath2 .
thus , we obtain the following algorithm for computing the well - founded semantics .
= = = = = = = = = = + * algorithm 3 + @xmath54 ; + + @xmath55 ; + @xmath134 ; + @xmath109 ; + no change in @xmath10 ; + @xmath9 and @xmath10 .
* we will now refine algorithm 3 .
specifically , we will show that the sets @xmath9 and @xmath10 can be computed incrementally .
let @xmath135 be a horn program .
we define the _ residual _ program of @xmath135 , @xmath136 , to be the horn program obtained from @xmath135 by removing all rules of @xmath135 with the head in @xmath137 and by removing from the bodies of the remaining rules those elements that are in @xmath137 .
we have the following technical result . [ l0 ]
let @xmath135 be a horn program and let @xmath23 be a set of atoms such that @xmath138 .
then @xmath139 .
@xmath107 lemma [ l0 ] implies that ( we treat here negated literals as new atoms and @xmath2 as horn program over the extended alphabet ) @xmath140 thus , if the set @xmath10 is expanded by new elements from @xmath111 , then the new set @xmath9 can be computed by increasing the old set @xmath9 by @xmath141 .
important thing to note is that the increment @xmath142 can be computed on the basis of the residual program and the increment @xmath111 .
similarly , we have @xmath143 thus , computing @xmath68 can also be done incrementally on the basis of the program considered in the previous iteration by taking into account most recently computed increments @xmath111 and @xmath142 .
this discussion implies that algorithm 3 can be equivalently restated as follows : = = = = = = = = = = + * algorithm 3 + 1@xmath144 ; + 2@xmath145 ; ( * @xmath135 will be treated as a horn program * ) + 3@xmath146 ; + 4 + 5@xmath147 ; + 6@xmath148 ; + 7@xmath149 ; + 8@xmath150 ; + 9@xmath151 ; + 10@xmath152 ; + 11 no change in @xmath10 ; + 12 @xmath9 and @xmath10 . * we will now estimate the running time of algorithm 3 .
clearly line 1 requires constant time .
setting up appropriate data structures for programs @xmath135 and @xmath29 ( lines 2 and 3 ) takes @xmath153 steps . in each iteration , @xmath142 is computed and the current program @xmath135 is replaced by the program @xmath154 ( lines 5 and 6 ) . by modifying the algorithm from @xcite and assuming that @xmath135 is already stored in the memory ( it is avaliable either as the result of the initialization in the case of the first iteration or as a result of the computation in the previous iteration ) , both tasks can be accomplished in @xmath155 steps . here
@xmath156 denotes the old version of @xmath135 and @xmath157 denotes the new version of @xmath135 .
consequently , the total time needed for lines 5 and 6 over all iterations is given by @xmath158 ( where @xmath159 is the program @xmath135 , when the algorithm terminates ) .
the time needed for all lines 7 is proportional to the number of iterations and is @xmath160 . given a logic program @xmath29 and sets of atoms @xmath142 and @xmath111 , it takes @xmath161 steps to compute the program @xmath162 in line 8 .
we assume here that @xmath29 is already in the memory as a result of the initialization in the case of the first iteration , or as the result of the computation in the previous iteration , otherwise .
it follows that the total time over all iterations needed to execute line 8 is @xmath163 .
thus , we obtain that the running time of algorithm 3 is given by @xmath164 , where @xmath165 is the total time needed to compute @xmath166 over all iterations of the algorithm . in the standard ( van gelder s )
implementation of algorithm 3 , we compute the whole set @xmath167 as @xmath166 .
in addition , computation is performed in a bottom - up fashion .
that is , we first compute the least model of @xmath168 and then its complement with respect to @xmath169 .
such approach requires @xmath170 steps per iteration to execute line 9 and leads to @xmath0 running - time estimate for the alternating - fixpoint algorithm .
in this section we will focus on the class of programs , @xmath13 , that is , programs whose rules have no more than one positive atom in their bodies .
we assume that we have a procedure @xmath171 that , given a horn program @xmath172 , returns a subset of the set @xmath173 .
we also assume that @xmath171 returns the empty set _ if and only if _ @xmath174 .
for every pair @xmath113 , where @xmath76 and @xmath29 is a horn program such that @xmath115 , we define @xmath175 it is easy to see that this operator @xmath176 satisfies conditions ( w1 ) and ( w2 ) . consequently , it can be used in algorithm 3 . clearly , the procedure @xmath112 and its computational properties are determined by the procedure @xmath171 . in the remainder of the paper
, we will describe a particular implementation of the procedure @xmath171 and estimate its running time .
we will use this estimate to obtain a bound on the running time of the resulting version of algorithm 3 . a straightforward way to compute the least model of @xmath29 and so , to find @xmath177 ,
is `` bottom - up '' .
that is , we start with atoms which are heads of rules with the empty bodies and use the rules of @xmath29 to compute all atoms in @xmath28 by iterating the van emden - kowalski operator .
an efficient implementation of the process is provided by the dowling - gallier algorithm @xcite .
the approach we follow here in the procedure @xmath171 is `` top - down '' and gives us , in general , only a part of the set @xmath177 . more precisely , for an atom @xmath178 we proceed `` backwards '' attempting to construct a proof or to demonstrate that no proof exists . in the process , we either go back to an atom that is the head of a rule with empty body or we show that no proof exists . in the former case , @xmath179 . in the latter
one , none of the atoms considered while searching for a proof of @xmath178 are in @xmath28 ( because @xmath180 and each rule has at most one antecedent ) .
the problem is that we may find an atom @xmath178 that does not have a proof only _
after _ we look at all other atoms first .
thus , in the worst case , finding one new false atom may require time that is proportional to the size of @xmath29 . to improve the time performance , we look for proofs simultaneously for all atoms and grow the proofs `` backwards '' in a carefully controlled way .
namely , we never let one search to get too much ahead of the other searches . this controlled way of looking for proofs is the key idea of our approach and leads to a better performance
. we will now provide an informal description of the procedure @xmath171 followed later by a formal specification and an example . in the procedure , we make use of a _ new _ atom , say @xmath181 , different from all atoms occurring in @xmath29 .
further , we denote by @xmath182 the atom in the head of a rule @xmath183 and by @xmath184 the atom which is either the unique positive atom in the body of @xmath185 , if such an atom exists , or @xmath181 otherwise .
we call an atom @xmath186 _ accessible _ if there are rules @xmath187 in @xmath29 such that @xmath188 , for @xmath189 , @xmath190 and @xmath191 . clearly , the least model @xmath28 of @xmath29 is precisely the set of all accessible atoms . in each step of the algorithm ,
the set of atoms from @xmath192 is partitioned into _ potentially false sets _ or _ pf - sets _ , for short .
we say that a set @xmath193 is a _ pf - set _
if for each pair of _ distinct _ atoms @xmath194 there are rules @xmath187 in @xmath29 such that @xmath195 , for @xmath189 , @xmath196 and @xmath191 .
it is clear that if @xmath197 is a pf - set then either all its elements are accessible ( belong to the least model of @xmath29 ) or none of them does ( they are all false ) . clearly , singleton sets consisting of individual atoms in @xmath198 are pf - sets . in the algorithm , with each pf - set we maintain its cardinality .
current information about the state of all top - down searches and about the dependencies among atoms , that were discovered so far , is maintained in a directed graph @xmath199 .
the vertex set of this graph , say @xmath200 , consists of @xmath201 and of a family of pf - sets forming a partition of the set @xmath192 .
the edges of @xmath199 are specified by a _ partial _
function @xmath202 .
we write @xmath203 if @xmath204 is undefined for @xmath197 .
thus , the set of edges of @xmath199 is given by @xmath205 .
since @xmath204 is a partial function , it is easy to see that the connected components of the graph @xmath199 are unicyclic graphs or trees rooted in those vertices @xmath197 for which @xmath206 is undefined . throughout the algorithm we always have @xmath207 .
thus , the connected component of @xmath199 containing @xmath201 is always a tree and @xmath201 is its root .
if @xmath208 and @xmath197 are two different pf - sets , the existence of the edge @xmath209 in @xmath199 means that we have already discovered a rule in the original program whose head is in @xmath197 and whose tail is in @xmath208 .
thus , if vertices in @xmath208 are accessible , then so are the vertices in @xmath197 . a pf - set that is the root of a tree forming a component of @xmath199 is called an _ active _ pf - set .
if @xmath197 is an active pf - set then no rule @xmath185 with @xmath210 and @xmath211 has been detected so far .
thus , @xmath197 is a candidate for a set of atoms which does not intersect the least model of @xmath29 .
let us note that even though @xmath201 is a root of a tree in @xmath199 it is never active as it is not a pf - set in the first place .
we let active pf - sets grow by gluing them with other pf - sets .
however , we allow to grow only these active pf - sets whose cardinalities are the least . in each iteration of the algorithm
the value of the variable @xmath212 is a lower bound for the cardinalities of active pf - sets . to grow an active pf - set @xmath197
, we look for rules with heads in @xmath197 and with tails in pf - sets _ other _ than @xmath197 ( not necessarily active ) or in @xmath201 .
the dependencies between pf - sets discovered in this way are represented as new directed edges in @xmath199 .
pf - sets that appear in the same cycle are glued together ( in the procedure _ cycle _ ) .
since @xmath201 is not an active pf - set , it never becomes an element of a cycle in @xmath199 . if , when attempting to grow a pf - set @xmath197 we discover a rule with head in @xmath197 and with the tail in a vertex of the tree of @xmath199 rooted in @xmath201 , then @xmath197 is from now on ignored ( all its vertices belong to the least model of @xmath29 ) .
indeed , @xmath197 gets connected to a tree of @xmath199 rooted in @xmath201 .
consequently , it can not become a member of a cycle in @xmath199 in the future and is never again considered by the procedure _
cycle_. the main loop ( lines 6 - 23 ) of the algorithm @xmath171 below starts by incrementing @xmath212 followed by a call to the procedure @xmath213 .
this procedure scans the graph @xmath199 and identifies all its cycles .
it then modifies @xmath199 by considering each cycle and by gluing its pf - sets into a single pf - set . to this end
, it modifies the vertex set @xmath214 of @xmath199 and the function @xmath204 defining the edges of @xmath199 .
each such new pf - set becomes the root of its tree in @xmath199 and so , it becomes active .
the procedure _ cycle _ computes the cardinality of each new active pf - set .
finally , it creates a list @xmath215 so that it consists of active pf - sets of cardinality @xmath212 .
if no such set is found ( @xmath215 is empty ) , we move on to the next iteration of the main loop and increment @xmath212 by 1 .
we give a more detailed description of the procedure @xmath216 later in the paper when we analyze the time complexity of our method .
for each active pf - set @xmath217 we consider the tail of each rule with head in @xmath197 ( lines 9 - 22 ) .
if there is a rule @xmath185 with @xmath210 and @xmath211 then it is detected ( line 15 ) .
the value @xmath206 is set to this element in @xmath214 that contains @xmath184 ( it may be that this set is @xmath201 ) .
we also set the variable @xmath218 to * true * ( line 16 ) .
the pf - set @xmath197 stops to be active .
we move on to the next active pf - set on @xmath215 .
if such a rule @xmath185 does not exist then @xmath219 and @xmath197 is a set of cardinality @xmath212 consisting of atoms which are not in the least model of @xmath29 .
this set is returned by the procedure @xmath171 ( line 21 ) .
hence , for an active pf - set considered in the loop 6 - 23 , either we find a pf - set @xmath220 ( and we have to consider the next pf - set on @xmath215 ) or @xmath197 is returned as a set of atoms which are not in the least model of @xmath29 ( and the procedure @xmath171 terminates ) .
thus , the procedure @xmath171 is completed if either a nonempty set @xmath197 of atoms which are not in the least model of @xmath29 is found or , after some passes of the loop 6 - 23 , the graph @xmath199 has no active pf - sets . in the latter case
@xmath199 is a tree with the root in @xmath221 .
thus , @xmath222 and @xmath223 is returned ( line 24 ) . in the procedure @xmath171 , as formally described below , an input program @xmath29 is represented by lists @xmath224 , @xmath186 , of all atoms @xmath225 such that @xmath225 is the body of some rule with the head @xmath178 . if there is a rule with the head @xmath178 and empty body , we insert @xmath181 into the list @xmath224 .
we also use an operation @xmath226 on lists and elements .
let @xmath227 be a list and @xmath208 be an element , either belonging to @xmath227 or having a special value * undefined*. then @xmath228 the value * undefined * should not be mixed with * nil * which indicates the end of a list .
finally , we use a procedure @xmath229 which , for an atom @xmath208 and a collection @xmath214 of disjoint sets , one of which contains @xmath208 , finds the name of the set in @xmath214 containing @xmath208 ( it follows from our assumptions that such a set is unique ) .
elements of @xmath214 are maintained as linked lists .
each element on such a list has a pointer to the head of the list .
the head serves as the identifier for the list . when the procedure @xmath229 is called , it returns the head of the list to which @xmath208 belongs .
= = = = = = = = = = + 1**procedure @xmath230 ; + 2@xmath231 ; + 3 @xmath232 * do * @xmath233 ; + 4 @xmath234 * do * \{@xmath235 ; @xmath236 } ; + 5@xmath237 ; + 6 @xmath238 * do * + 7@xmath239 ; + 8@xmath240 ; + 9 all @xmath241 * do * + 10@xmath242 ; + 11@xmath243 ; + 12 @xmath244 * nil and * * not * @xmath218 * do * + 13@xmath245 ; + 14 @xmath246 * nil and * * not * @xmath218 * do * + 15@xmath247*if * @xmath248 + 16 @xmath249 ; @xmath250 + 17 @xmath245 + 18 ( 14)@xmath251 ; + 19 @xmath218 * then * @xmath243 + 20 ( 12)@xmath251 ; + 21 @xmath218 * then return * @xmath197 ( * the procedure terminates * ) + 22 ( 9)@xmath251 + 23 ( 6)@xmath251 ; + 24**return @xmath223 + 25 @xmath171 ; + * * *
* we will now illustrate the operation of the algorithm .
let us consider the following horn logic program @xmath29 : @xmath252 this program is represented as a graph , @xmath253 , in fig .
the vertices of this graph correspond to the atoms of the program .
in addition , @xmath253 has an auxiliary vertex @xmath254 .
an edge @xmath255 , where @xmath256 , represents the clause @xmath257 from @xmath29 .
an edge @xmath258 , where @xmath259 , represents the clause @xmath260 .
when illustrating the algorithm , we assume that atoms from @xmath192 ( atoms @xmath261 in our example ) appear on the lists @xmath262 , @xmath263 , in the alphabetical order .
we also assume that whenever @xmath181 belongs to a list @xmath262 , it appears as the first atom on the list . in the algorithm _
false _ , the current state of knowledge about the possibility of proving an atom from @xmath29 is represented by the graph @xmath264 .
initially , @xmath199 consists of isolated vertices .
indeed , line 3 of the algorithm sets @xmath265 to undefined , for every vertex @xmath266 of @xmath214 ( see fig .
[ rys1a ] ( left ) ) .
all of the vertices of @xmath199 , except for @xmath201 are active pf - sets .
the procedure @xmath216 ( line 8) , called with @xmath267 , puts all of them on the list @xmath215 .
the algorithm considers next ( line 9 ) all elements on the list @xmath215 , that is , all vertices of @xmath199 that are active pf - sets and have cardinality equal to _ size_. during the first iteration of the loop 6 - 23 , @xmath215 consists of all vertices of @xmath199 , except for @xmath201 ( that is , singleton sets @xmath268 , where @xmath269 ) . for each vertex @xmath197 of @xmath199 on @xmath215 ,
the algorithm looks for a _ back rule _ for @xmath197 , that is , a rule in @xmath29 with the head in @xmath197 and the tail in a pf - set other than @xmath197 or in @xmath201 . in our graphical representation of @xmath29 by means of the graph @xmath253 , a back rule for @xmath197 corresponds to an edge ( referred to as a _
back _ edge ) in @xmath253 with the head in @xmath197 and the tail in a vertex of @xmath199 other than @xmath197 ( possibly in @xmath201 ) . to find a back rule ( edge ) for @xmath197 , all atoms @xmath270 of @xmath29 ( equivalently ,
all vertices @xmath270 of @xmath253 ) that belong to @xmath197 are considered ( the loop 12 - 20 ) . for each such atom @xmath270 ,
the algorithm searches for the first atom on the list @xmath271 that does not belong to @xmath197 .
let us recall that @xmath271 is the list of atoms that are the tails of rules with the head @xmath270 or , in the terms of the graph @xmath253 , that are the tails of edges with the head @xmath270 .
if such an atom is found , together with @xmath270 it determines a back rule ( edge ) @xmath185 for @xmath197 .
the algorithm sets @xmath206 to be equal to the pf - set containing the tail of @xmath185 ( line 16 ) .
that is , an edge from @xmath206 to @xmath197 is added to @xmath199 .
the algorithm moves then on to the next element of the list @xmath215 . in our example
, in the first iteration of the loop 6 - 23 , a back rule is found for every element on @xmath215 , that is , for every vertex of @xmath199 other than @xmath201 .
for instance , for the vertex @xmath272 , the algorithm considers atoms on the list @xmath273 ( let us recall that atoms on lists @xmath262 are arranged alphabetically with the exception of the special atom @xmath181 which , if present on a list , is always its first element ) .
the first atom on the list , @xmath274 does not belong to @xmath272 .
thus , it defines , together with @xmath275 a back rule for @xmath272 , @xmath276 .
the resulting graph @xmath199 is shown in fig . [ rys1a ] on the right .
let us note that when scanning the list @xmath277 in subsequent iterations the algorithm resumes the scan with the first atom that has not been looked at yet ( cf .
the definition of the operation _ next _ ) .
thus , the next time @xmath275 is considered as an element of an active pf - set for which a back rule is searched for , the scan of @xmath277 will start with @xmath278 .
the same holds true for all lists @xmath262 , @xmath279 .
consequently , each atom on each of these lists is considered just once .
such an approach still guarantees that finding back rules works correctly ( that is , that they are found by the algorithm whenever they exist ) .
indeed , when an atom on a list @xmath262 is considered , it either defines a back rule with the head @xmath266 ( and , thus , can not define any new back rule with the head @xmath266 in the future ) or it is in the same active pf - set as @xmath266 ( and , thus , it neither defines a back rule now nor it will define it in the future , as it will remain in the same pf - set as @xmath266 till the algorithm terminates ) .
the second iteration of the loop 6 - 23 starts with the procedure _ cycle _ contracting each cycle in the graph @xmath199 to a single vertex . the resulting graph is shown in fig .
[ rys2a ] on the left .
the procedure _ cycle _ then creates a new list @xmath215 .
it consists of all active pf - sets of cardinality 2 . in our case
, @xmath215 contains @xmath280 and @xmath281 ( @xmath282 is also active but has cardinality 3 )
. continuing with the second iteration , the algorithm next considers each vertex on @xmath215 ( the loop 9 - 22 ) and looks for back rules . in this iteration ,
a back rule is found for each of the nodes on @xmath215 and the modified graph @xmath199 is given in fig .
[ rys2a ] on the right . in the third iteration , the procedure _
cycle _ contracts the only cycle in @xmath199 to a single active pf - set of cardinality 4 ( figure [ rys3a ] , left side ) .
it also creates a new list @xmath215 .
this time it consists of active pf - sets of cardinality 3 . there is just one such set - @xmath282 .
subsequently , the algorithm _ false _ looks for a back rule for @xmath282 .
it starts by considering edges ending in @xmath275 ( line 11 ; we assume that @xmath197 is represented by the list @xmath283 ) .
it scans the list @xmath277 starting at the first atom that has not been inspected so far , that is , @xmath278 .
however , since @xmath278 belongs to the same pf - set as @xmath275 , @xmath278 does not specify a back rule .
since there are no more atoms on the list @xmath277 , we move on to the next iteration of the loop 12 - 20 and consider atom @xmath274 .
we have @xmath284 . since @xmath278 was already considered ( and yielded a back rule for @xmath285 ) in the first iteration , we consider @xmath286 . since @xmath287 , it defines a back rule for @xmath282 , @xmath288 . the resulting graph @xmath199 is shown in figure [ rys3a ] ( on the right ) .
it has no cycles .
so , the only thing done by the procedure _ cycle _ in the iteration 4 is that it puts on @xmath215 active pf - sets of cardinality 4 .
there is just one such set in @xmath199 , @xmath289 . the algorithm _ false _
looks for a back edge for @xmath289 and does not find any .
the variable _ success _ remains * false*. the algorithm returns @xmath289 and terminates ( line 21 ) .
let us note that this set is a proper subset of the set @xmath173 .
the following theorem formally establishes two key properties of the procedure @xmath171 .
[ t1 ] 1 .
the procedure @xmath171 returns a set @xmath197 such that @xmath290 .
@xmath171 returns the empty set if and only if @xmath291 .
proof : ( 1 ) the statement is trivially true if _ false _ returns the empty set .
thus assume that the returned set @xmath292 .
it means that the value of the variable @xmath218 is * false * after all passes of the loop 12 - 20 for some active pf - set @xmath197 in the list @xmath215 .
thus every rule in @xmath29 with the head in @xmath197 has been considered .
suppose there is a rule @xmath185 in @xmath29 with @xmath293 and @xmath294 .
this rule was considered by the procedure @xmath171 when @xmath295 was a member of some active pf - set , say @xmath296 .
since larger pf - sets are obtained by gluing smaller ones , @xmath297 . while @xmath185 was being considered , the value of @xmath298 in the loop 14 - 18 was @xmath225 and the value of @xmath197 was @xmath296 .
consequently , @xmath299 in line 15 because @xmath297 and @xmath300 so @xmath301 .
hence the value of @xmath218 was set to * true * and @xmath302 was defined to be , say , @xmath303 in line 16 .
the pf - set @xmath296 stopped to be active .
recall that @xmath197 is active when the procedure stops .
hence @xmath296 had to be glued with other pf - sets to obtain @xmath197 .
this is , however , impossible because if @xmath296 were glued with some other pf - sets to form a larger pf - set @xmath266 then @xmath304 .
notice that @xmath305 .
we have got a contradiction with @xmath300 . hence , there are no rules @xmath185 in @xmath29 with @xmath210 and @xmath306 .
thus no atom in @xmath197 is accessible so @xmath290 .
+ ( 2 ) suppose _ false _ returns the empty set and consider the last pass of the loop 6 - 23 , for @xmath307 .
if the list @xmath215 is empty then no vertex of @xmath199 is an active pf - set .
hence , @xmath199 is a tree with the root @xmath221 .
thus all atoms in @xmath198 are accessible and consequently @xmath308 .
if the list @xmath215 is nonempty then it contains one pf - set @xmath309 .
the empty set is returned by the procedure @xmath171 so the value of the variable @xmath218 in line 16 is * true * for @xmath309 .
it means that for some rule @xmath185 in @xmath29 with @xmath310 , @xmath311 so @xmath312 .
hence , @xmath270 is accessible and , consequently , all atoms in @xmath198 are accessible .
that is , we have @xmath291 .
the converse of the implication proved above follows immediately from the first part of the theorem .
@xmath107 we shall now consider the procedure @xmath216 a little bit more carefully . the procedure can be informally written in the following form .
= = = = + * procedure * @xmath240 + 1.initialize @xmath215 to empty .
+ 2.find all cycles @xmath313 in the graph @xmath199 .
put @xmath314 .
+ 3.for every cycle @xmath315 , @xmath316 , do ( i)-(iv ) .
+ ( i)set @xmath317 ; + ( ii)compute @xmath318 ( sum up the cardinalities of all vertices in @xmath319 ) ; + ( iii)update the set @xmath214 : set @xmath320 ; ( * @xmath321 becomes + an active pf - set * ) + ( iv)update the function @xmath204 : for every @xmath322 , if @xmath323 ( for + some @xmath324 ) then @xmath325 ; + 4.for every vertex of @xmath199 that is an active pf - set , if @xmath326 , insert @xmath197 + into the list @xmath215 .
since @xmath199 is a directed graph whose connected components are either unicyclic graphs or trees , step 2 of the procedure @xmath216 can be implemented in @xmath327 time . since pf - sets are represented as linked lists , with each node on the list pointing to the head of the list , step ( i ) can be implemented to take @xmath328 steps .
the time needed for step ( ii ) is , clearly , @xmath329 .
each execution of step ( iii ) takes also @xmath330 . finally , the running time of each execution of step ( iv ) is @xmath331 , where @xmath332 is the size of the connected component of the graph @xmath199 containing @xmath319 .
thus , an iteration of the loop 3 for a cycle @xmath316 takes @xmath333 . clearly , @xmath334 .
moreover , @xmath335 and @xmath336 ( they are all disjoint subsets of @xmath192 ) .
thus , the total time needed for the loop 3 is @xmath337 .
it is easy to see that the time needed for the loop 4 is also @xmath337 .
consequently , the running time of the procedure @xmath216 is @xmath337 .
we are now in a position to estimate the running time of the procedure @xmath171 .
[ l1 ] if the procedure @xmath230 returns a nonempty set @xmath197 , then the running time of @xmath171 is @xmath338 .
if @xmath230 returns the empty set then its running time is @xmath339 .
proof : let @xmath340 and @xmath341 .
as we have already observed the procedure @xmath216 runs in time @xmath342 .
it is not hard to see that , since we represent all sets occurring in the procedure @xmath171 as linked lists , with each node on a list pointing to the head of the list , the operations : @xmath343 and @xmath226 require a constant time .
first assume that the output @xmath197 of the procedure @xmath171 is nonempty .
let us estimate the number of passes of the * while * and * for * loops in the procedure .
clearly , the loop 6 - 23 is executed @xmath344 times .
hence the total running time of all calls of the procedure @xmath216 is @xmath345 .
the number of passes of the loop 9 - 22 is not larger than @xmath346 , where @xmath347 denotes the list @xmath215 in an iteration @xmath62 of the loop . since @xmath347 is a list of disjoint pf - sets of cardinality @xmath62 , @xmath348 , for each @xmath349 .
hence the number of passes of the loop 9 - 22 can be very roughly estimated by @xmath350 .
the loop 12 - 20 is executed at most @xmath351 times .
this inequality follows from the fact that the sets @xmath197 in the lists @xmath347 are disjoint subsets of atoms so @xmath352 .
the estimation of the number of passes of the loop 14 - 18 is a little bit more complicated .
first notice that in each execution of the loop we check a rule of the program @xmath29 and rules are checked only one time .
the rules @xmath185 checked in the loop have either both the head and the tail in some pf - set @xmath232 or @xmath210 and @xmath184 is in some other pf - set @xmath353 . in the latter case @xmath206 is defined in line 16 .
the number of executions of line 16 is not larger than the number of passes of the loop 9 - 22 so it is bounded by @xmath350 .
when the procedure returns the output , the pf - sets have cardinalities not larger than @xmath344 .
hence the number of rules with both the head and the tail in the same pf - set that has been checked before the procedure stops is not larger than @xmath354 thus the number of passes of the loop 14 - 18 in the whole procedure @xmath171 is less than @xmath355 .
it follows that if the output @xmath197 of @xmath171 is nonempty then the running time of @xmath171 is @xmath338 .
now consider the case when the procedure @xmath171 returns the empty set .
clearly the number of passes of the loop 6 - 23 is @xmath356 so it takes @xmath357 time for all executions of the procedure @xmath216 . since the rules are checked in the loop 14 - 18 only one time , the number of passes of this loop is not larger than the number @xmath165 of rules in @xmath29 .
obviously @xmath358 so the running time of @xmath171 in this case is @xmath339 .
@xmath107 by lemma [ l1 ] and considerations in section 3 we get an estimation of the running time of algorithm 3 .
[ t2 ] if @xmath2 is a program whose rules have at most one positive atom in the body then algorithm 3 can be implemented so that its running time is @xmath359 . @xmath107
the method for computing the well - founded semantics described in this paper is a refinement of the basic alternating - fixpoint algorithm .
the key idea is to use a top - down search when identifying atoms that are false .
our method is designed to work with programs whose rules have at most one positive atom in their bodies ( class @xmath360 ) .
its running time is @xmath14 ( where @xmath2 is an input program ) .
thus , our algorithm is an improvement over other known methods to compute the well - founded semantics for programs in the class @xmath13 .
our algorithm runs in linear time for the class of programs @xmath361 for which @xmath362 .
however , it is not a linear - time algorithm in general .
it is an open question whether a linear - time algorithm for computing the well - founded semantics for programs in the class @xmath13 exists .
our results extend to the class @xmath16
. however , the extension is straightforward and the class @xmath16 is still rather narrow .
moreover , it is not specified syntactically ( it is described by means of the kripke - kleene semantics ) .
the question arises whether our top - down approach to positive - loop detection can be generalized to any class of programs significantly extending the class @xmath13 and possessing a simple syntactic description . finally , let us note that the general problem of computing the well - founded semantics still remains a challenge .
no significant improvement over the alternating - fixpoint algorithm of van gelder has been obtained for the class of arbitrary finite propositional logic programs .
this research was supported by the nsf grants cda-9502645 and iri-9619233 .
rao , p. , ramskrishnan , i.v .
, sagonas , k. , swift , t. , warren , d. s. , & freire , j. ( 1997 ) . : a system for efficiently computing well - founded semantics . .
springer - verlag .
ecture notes in computer science , 1265 . | the well - founded semantics is one of the most widely studied and used semantics of logic programs with negation . in the case of finite propositional programs , it can be computed in polynomial time , more specifically , in @xmath0 steps , where @xmath1 denotes the total number of occurrences of atoms in a logic program @xmath2 .
this bound is achieved by an algorithm introduced by van gelder and known as the alternating - fixpoint algorithm .
improving on the alternating - fixpoint algorithm turned out to be difficult . in this paper
we study extensions and modifications of the alternating - fixpoint approach .
we then restrict our attention to the class of programs whose rules have no more than one positive occurrence of an atom in their bodies . for programs in that class
we propose a new implementation of the alternating - fixpoint method in which false atoms are computed in a top - down fashion .
we show that our algorithm is faster than other known algorithms and that for a wide class of programs it is linear and so , asymptotically optimal .
[ firstpage ] |
decoherence in two - state quantum systems is a significant obstacle to the realization of proposed quantum information technologies .
coupling between quantum bit ( `` qubit '' ) states and unobserved environmental degrees of freedom leads to decoherence effects which limit the practical implementation of proposed quantum algorithms @xcite .
photon modes , including polarization and spatial modes , provide an easily accessible system in which simple quantum circuits can be investigated @xcite . here , we examine the process of decoherence by subjecting single photons to a controllable birefringent `` environment '' and observing the evolution of the polarization state . in section [ one - photon ]
, we investigate the evolution of a single photon state under the action of a unitary coupling between polarization and frequency modes .
such non - dissipative `` phase errors '' give rise to decoherence effects , whereby the photon evolves from a definite polarization state to an unpolarized state .
we then introduce and examine an optical implementation of so - called bang - bang " quantum control of decoherence by rapidly exchanging the eigenstates of the coupling operation @xcite .
the term quantum control " is justified since such an operation will be shown to preserve a coherent polarization state in some special cases , and to reduce decoherence in more general cases . in section [ dfs ] we will introduce a particular two - photon polarization - entangled state that , due to its symmetry properties , is immune to collective decoherence of the type mentioned above .
that is , this state is a decoherence - free subspace ( dfs ) of the hilbert space of photon polarization @xcite .
photon pairs entangled in both polarization and frequency degrees of freedom , such as hyper - entangled photons produced in down - conversion sources ( see @xcite ) , further complicate this particular decoherence mechanism .
in particular , energy conservation imposes frequency correlations which affect the coherence properties of these two - photon states . in the experimental case
, it will be shown that this frequency correlation can be effectively suppressed and the dfs recovered by a simple and physically intuitive modification of the apparatus .
in this section , we will show that non - dissipative ( unitary ) coupling between photon frequency and polarization in a birefringent `` environment '' followed by a trace over frequency leads to decoherence.7.5.8 ) . ]
these results are relevant to the study of decoherence in quantum systems which arises when coupling to environmental states , followed by a trace over those degrees of freedom , leads to a loss of phase information between qubit basis states . in the following argument ,
the term `` environmental '' will be used to describe the coupling of information - carrying states ( photon polarization modes ) to degrees of freedom which are not utilized for information representation ( photon frequency modes ) .
it will be shown that even non - dissipative coupling between these modes leads to a loss of information in the qubit states .
a single photon characterized by its frequency spectrum and polarization can be represented by the state ket @xmath0 where @xmath1 , and @xmath2 is the complex amplitude corresponding to the frequency @xmath3 , normalized so that @xmath4 for simplicity , we assume that the frequency spectrum is independent of polarization , so that we need not index @xmath5 by polarization mode . in physical terms , this means that polarization and frequency are not entangled .
since @xmath6 represents a pure state , the density operator can be written as @xmath7 ( see @xcite , @xmath83.4 ) , so that we have the initial state @xmath9
the subscript indicates that @xmath10 includes frequency degrees of freedom .
the usual @xmath11 density matrix representing the polarization state of the photon is given by @xmath12 .
we will now seek the dependence of the density operator @xmath13 on some spatially extended `` environment , '' by introducing the operator @xmath14 which we require to be both linear and unitary .
furthermore , we demand that @xmath14 have eigenkets @xmath15 with ( frequency - dependent ) eigenvalues @xmath16 where , as before , @xmath17 indexes the polarization mode .
since @xmath14 exhibits frequency- and polarization - dependent eigenvalues , it introduces a coupling between qubit states ( polarization modes ) and non - qubit states ( frequency modes ) . as shown in appendix
[ eigenvalues ] , @xmath14 introduces purephase errors " on polarization modes @xmath15 , so that @xmath18 where the phase factor @xmath19 is a real - valued function of @xmath3 and @xmath20 .
now we can calculate the spatial dependence of the density operator under the influence of @xmath14 and following a frequency - insensitive measurement , effected by a partial trace over frequency degrees of freedom : @xmath21}. \label{rho1noqc}\end{aligned}\ ] ] since this process realizes the entangling of qubit states with environmental states ( frequency modes which are not involved in information representation or manipulation ) , we expect to observe decoherence effects in the off - diagonal elements of the qubit state ( polarization ) density matrix @xcite . in order to investigate the coherence properties of a photon under the action of such an operator
, we observe that in a completely mixed ( or decohered " ) state , the density matrix is simply a multiple of the identity matrix .
the diagonal elements of @xmath13 are indexed by @xmath22 .
for these values of @xmath23 and @xmath24 , the argument of the exponential in eq .
[ rho1noqc ] vanishes and by the normalization condition ( eq . [ norm1 ] ) , we have @xmath25 .
in other words , the diagonal elements of @xmath26 are unaffected by @xmath14
. the off - diagonal elements of the density matrix are indexed by the values @xmath27 and @xmath28 . to compute these explicitly
, we must define the functions @xmath19 . in the case of a single photon in a birefringent , linearly dispersive
crystal of thickness @xmath20 ( e.g. , a quartz crystal ) , @xmath29 where @xmath30 is the index of refraction corresponding to polarization @xmath24 . with @xmath3 . by assuming @xmath15 to be an eigenstate of @xmath31 ( eq . [ phi ] ) , we require the optic axis of the birefringent element to be aligned with one of the polarizations @xmath15 . in the case where the @xmath15 represent circular ( elliptical ) polarizations
, this means that the environment is optically active as opposed to ( as well as ) birefringent . ]
making the substitution @xmath32 and writing @xmath33 , the off - diagonal elements of the density matrix are given by @xmath34 where @xmath35 is the fourier transform of @xmath36 ( up to a constant , depending on convention ) .
evidently , @xmath37 implies that @xmath38 , so that eqs .
[ rho12 ] and [ rho21 ] preserve the required hermiticity of @xmath39 . in terms of @xmath40 and @xmath41 ,
the density matrix at some later position @xmath20 becomes a simple modification of the density matrix at @xmath42 : @xmath43 eq .
[ rhonoqc ] allows us to calculate the density matrix representing the polarization of an optical qubit subjected to a non - dissipative frequency - polarization coupling , given the initial state density matrix , @xmath44 , ( written in the basis of eigenmodes of the coupling ) and the functional form of the frequency - amplitude function @xmath5 .. ] in general , @xmath45 is peaked at some central value @xmath46 with a finite width @xmath47 . under these conditions , the fourier transform @xmath41 ( and the off - diagonal elements of @xmath39 ) will fall significantly on the time scale @xmath48 .
in other words , there is no coherent phase relationship between polarization basis states @xmath49 and @xmath50 after the photon has traveled a distance @xmath51 through the environment @xmath14 . to summarize
, we have developed a prescription for calculating the polarization state density matrix representing a photon in a birefringent environment . from this expression
, we see that a photon with definite polarization phase information before entering such an environment will , in general , lose phase information between polarization basis states after traveling a suitably long distance . the physical mechanism which destroys such phase information is the entanglement between polarization modes and frequency modes , which is an instance of the entanglement between qubit states and environmental degrees of freedom which are not useful for information manipulation . because we trace over frequency ( environmental ) degrees of freedom in this case , the phase coherence between the qubit basis states is destroyed . in the previous section
, we showed that a qubit consisting of photon polarization modes @xmath15 may lose its capacity for information representation under the action of a unitary operator , @xmath14 , which entangles frequency and polarization .
in other words , decoherence occurs even for non - dissipative `` phase errors '' in which the qubit basis states are eigenstates of the environmental coupling , and there is no chance for a bit flip error @xcite .
may appear as bit - flip errors in another basis , e.g. @xmath52 . ] in @xcite , the authors describe this result in the general case of qubit states coupled to environmental degrees of freedom and introduce a scheme for reducing such decoherence effects via rapid , periodic exchanging of the eigenstates of the environmental coupling . in our case , this corresponds to an interchange of the polarization basis states @xmath49 and @xmath50 , which can be accomplished by an appropriate reflection or rotation operation .
these exchanges are rapid in the sense that the period of flipping must be short compared to the decoherence time @xmath53 and the time scales of any other relevant decoherence mechanisms ( e.g. scattering , dissipation ) .
so called `` bang - bang '' control reduces the degree of decoherence by averaging out the time - dependent coupling between qubit states and environmental states . in this section ,
the previous results are expanded to include an implementation of `` bang - bang '' quantum control by periodic exchanging of the environmental eigenstates @xmath15 .
the mathematical formulation is considerably simplified if we move to the discrete regime in which the `` environment '' acts in a stepwise fashion , and the exchange of eigenstates also occurs in discrete steps . by including a rotation operator acting simultaneously with @xmath14
is considerably complicated , since this rotation operator will not , in general , commute with @xmath54 for different @xmath20 values . ] to this end , we define @xmath55 to be a reflection ( or rotation ) operator which exchanges the eigenstates of the frequency - polarization coupling : @xmath56 here , the plus ( minus ) sign refers to a reflection ( rotation ) .
the choice of sign will not affect the results of a measurement in the experimental case , and for concreteness , we take the sign to be positive .
we define @xmath31 , a step - wise " operator analogous to @xmath14 in sect .
[ review ] and also define @xmath57 to be the thickness of the birefringent element ( in our experiments , usually a quartz crystal ) ; as before , @xmath30 is the refractive index corresponding to polarization state @xmath24 , and @xmath40 .
this gives a relative phase shift of @xmath58 , so that we have , which is common to eqs .
[ scriptu1 ] and [ scriptu2 ] .
such a global phase shift is never observable and can not affect the coherence properties of the polarization states . ]
@xmath59 eqs .
[ r1]-[r2 ] and [ scriptu1]-[scriptu2 ] have the important consequence that @xmath60 where it is understood that such operator identities are only meaningful when the operators are applied to kets or bras ( see appendix [ proof1 ] for a proof of these identities ) . in this discrete regime
, we consider `` rapid '' exchanging of eigenstates to correspond to the alternating action of the operators @xmath31 and @xmath61 .
regarding @xmath31 and @xmath61 as optical elements ( for example , a quartz crystal and a half - wave plate ) , we consider a cavity scheme in which a single photon passes through this two - element unit @xmath62 times .
symbolically , after @xmath62 passes through the system , we have ( compare eq .
[ rho1noqc1 ] ) , @xmath63 application of identities [ identity1 ] and [ identity2 ] leads to a simplification , depending on whether @xmath62 is odd ( @xmath64 ) or even ( @xmath65 ) .
@xmath66 note that @xmath67 is related to the continuous case of sect .
[ review ] by @xmath68 these relations lead to a simple analog of eq .
[ rho1noqc ] ( where we use @xmath69 , given by eq .
[ rho0 ] , as the input state in both cases ) : @xmath70 finally , writing out @xmath71 in matrix form , substituting @xmath72 for the fourier transform of @xmath73 , and denoting by @xmath74 the characteristic time parameter , we have @xmath75 the subscript @xmath76 ( for `` quantum control '' ) indicates the case in which @xmath61 is included . note that @xmath77 is dependent only on the parity of @xmath62 , so that under the influence of the quantum control procedure , any number of cycles through this particular birefringent environment causes _ at most _ a decoherence equal to that of the first pass .
in fact , eq . [ rhoqc1 ] shows that an even number of cycles causes no decoherence whatsoever . by making the identification @xmath78 ( the distance traveled through the crystal after n passes ) in eq .
[ rhonoqc ] , we can directly compare the results of sect .
[ review ] with those of sect .
[ bang - bang ] : @xmath79 letting @xmath80 be the coherence time such that @xmath41 is significantly reduced for @xmath81 , then @xmath71 represents a partially mixed state for @xmath82 .
are small so that @xmath83 . if the initial state , @xmath44 , has diagonal elements of @xmath84 , then @xmath39 represents a completely random ensemble for @xmath85 .
] on the other hand , @xmath86 never loses all phase information in the off - diagonal elements since @xmath65 recovers the initial state for all @xmath87 . in this sense , @xmath61 acts as a quantum control element , suppressing decoherence and maintaining state purity indefinitely .
we may ask what happens if the strength of the environmental coupling varies with @xmath20 , so that @xmath31 varies on a time scale comparable to the period of the exchange operator , @xmath61 . for simplicity , we model this `` slow - flipping '' case by introducing two distinct operators which differ only in their eigenvalues ( not their eigenstates ) . in more concrete terms , we consider two coupling operators , @xmath88 and @xmath89 , characterized by phase errors @xmath90 and @xmath91 so that , for @xmath92 : @xmath93 the exchange operator @xmath55 acts in alternation between each of these two operators , so that the strength of the coupling changes faster than the eigenstates of the coupling are exchanged .
after n passes through the system , we have @xmath94 intuitively , we expect to see a partial reduction in the degree of decoherence when the quantum control procedure is implemented by introducing the exchange operator . to see these results quantitatively , we make the following substitution which reduces this problem ( mathematically ) to the previously solved one ( see appendix [ proof2 ] for a derivation of these relations ) : @xmath95 so that , @xmath96 @xmath97 with these relations , each operator is mathematically analogous to the operator @xmath31 defined by eq .
[ phi ] , with @xmath98 redefined as appropriate given relations [ relation1]-[relation4 ] and the details of the particular environment . in the case of birefringent crystals ,
the elements corresponding to @xmath88 and @xmath89 differ only in their thickness ( not in their orientation with respect to the polarization basis states @xmath15 ) . defining @xmath99 to be the thickness of a crystal represented by @xmath100
, we have @xmath101 we see that before introducing the exchange operator , the birefringent environment acts effectively as a single birefringent crystal of thickness @xmath102 ( eq . [ relation2 ] ) . after implementing `` bang - bang '' control by introducing the exchange operator @xmath61
, the environment looks effectively like a single crystal of thickness @xmath103 ( eq . [ relation4 ] ) . in the absence of quantum control ,
the photon is subjected to the _ sum _ of the individual phase errors ; in the presence of a quantum control element , the photon is subjected to the _ difference _ of the individual phase errors . in this case ,
where the strength of the environmental coupling changes on roughly the same time scale as the exchanging of eigenstates , the quantum control procedure reduces but does not eliminate decoherence effects .
it will be convenient in our experimental analysis to measure the _ visibility _ , traditionally defined as @xmath104 where @xmath105 ( @xmath106 ) denotes the maximum ( minimum ) number of photons passing through orthogonal settings of a polarizer . denoting these polarization analyzer settings by @xmath107 and @xmath108 ,
we define visibility at the single quantum level by calculating probabilities from the density operator , @xmath13 , is intended to suggest that these polarization modes may , in general , represent any elliptical orthonormal basis states . ]
: @xmath109 note that we no longer include the denominator corresponding to the total number of photons in eq .
[ v_old ] , since the denominator here is the trace of @xmath39 in the @xmath110 basis , which is unity .
@xmath111 quantifies the similarity between the state represented by @xmath26 and the state @xmath107 in the following sense : if @xmath26 represents an unpolarized state , it is a multiple of the identity matrix , and @xmath112 ; at the opposite extreme , if the photon is in a pure state with polarization @xmath113 , @xmath114 .
if we choose to measure with respect to the state @xmath115 , we find that ( see appendix [ vis_proof ] for a proof ) @xmath116 where we have dropped the subscript @xmath113 for this choice of measurement , and @xmath117 indicates the real part . in this form , it is clear that visibility @xmath118 depends on the off - diagonal elements of @xmath13 , and so corresponds with our previous notion of coherence . if we write @xmath39 in the @xmath119 basis , measuring the visibility with respect to @xmath120 and @xmath121 gives us information about the off - diagonal elements of @xmath39 . to see the results of the previous sections experimentally ,
we seek a systematic and controllable implementation of the operators @xmath31 and @xmath61 in the cavity scheme of sect .
[ bang - bang ] .
for the first case , in which there is a single operation which produces a phase error , we use a michelson polarization interferometer as a tunable decoherence mechanism in the @xmath119 basis . a polarizing beam - splitter ( pbs )
separates the path of @xmath122- and @xmath123-polarized light .
the unbalanced arms of the interferometer produce a controllable , @xmath3-dependent phase shift between @xmath124 and @xmath125 ( see fig .
[ michelson ] ) .
[ .35]schematic diagram of experimental apparatus for producing controllable decoherence in the h / v basis .
pbs indicates a polarizing beam splitter ; @xmath61 indicates the ( removable ) exchange element , a quarter - wave plate in this case . the input state is @xmath126.,title="fig : " ] a photon in either path is reflected back to the pbs and retraces its path to the input . the difference in path lengths between arms of the interferometer
will be denoted by @xmath20 .
the unbalanced interferometer is the optical realization of the operator @xmath127 of sect .
[ review ] and @xmath31 of sect .
[ bang - bang ] .
an @xmath128 reflector at the input recycles the light so that a photon passes through the unbalanced arms of the interferometer more than once , in general . to denote the _ probability _ of reflection at a surface .
] a partially reflective out - coupling mirror directs a photon out of the interferometer with a probability of @xmath129 at each pass . at each air - glass interface . here
, one surface of the out - coupling mirror is broad - band anti - reflection ( bbar ) coated so that we have effectively only a single air - glass interface . ]
we send in light from a greatly attenuated diode laser at @xmath130 nm , pulsed at @xmath131 khz . in our experiment
, we observe photons passing through the interferometer @xmath132 times . the total distance traveled by a photon passing through our interferometer @xmath133 times
is @xmath134 m , so that any photon spends at most @xmath135 s in the system . eg&g spcm - aq silicon avalanche photodiodes ( dark count @xmath136 , efficiency @xmath137 ) are used as photon counters . using electronic timing techniques
, we record photon counts only in a coincidence window of 50 ns between the pulse generator and detection event .
these timing techniques , together with spectral filters at the detectors , reduce background coincidence counts to @xmath138 s@xmath139 .
we graphically display photon counts vs. arrival time so that photons leaving the system after n cycles are recorded in the nth peak . in a typical experiment ,
our photon counters register no more than @xmath140 coincidence counts per second , and we conclude that @xmath141 photons per second leave the system ( @xmath142 per pulse ) .. these photon numbers are comparable to those used in current demonstrations of free - space quantum cryptographic key distributions @xcite . ] at the output , we count for a few seconds ( less than a minute in most cases ) .
although we can not determine the unknown polarization state of a single photon , we can accurately find the statistical distribution of the ensemble over the observation time , from which useful information about the density matrix can be inferred . by placing a @xmath143 waveplate with optic axis at @xmath144 at the recycling reflector , we realize an optical flipping " operator .
the @xmath143-reflector-@xmath143 combination takes @xmath124 to @xmath125 and _ vice versa_. this is the optical realization of the exchange operator , @xmath61 in sect .
[ bang - bang ] . by recording photon counts in coincidence with the diode pulse generator and using a polarizer at the detector , we can measure the visibility at @xmath145 .
the frequency spectrum of laser light is restricted using an interference filter with bandwidth @xmath146 nm and central wavelength @xmath147 nm . for our calculations
, we use a gaussian amplitude function with frequency spread @xmath47 corresponding to @xmath148 nm , central frequency @xmath149 corresponding to @xmath150 nm , and normalization factor @xmath151 : ) component at @xmath46.[monofootnote ] ] @xmath152.\ ] ] the corresponding coherence length of the photons is given by @xmath153 m , so we expect substantial decoherence for a path difference of @xmath154 m at cycle @xmath155 . [ cols="<,^ " , ] we see that the singlet state does indeed maintain full coherence in the presence of a unitary frequency - polarization coupling . since polarization degrees of freedom are analyzed and frequency degrees of freedom are unobserved , such a result demonstrates the robustness of a decoherence - free subspace of two qubits against non - dissipative errors arising from coupling with unutilized ( environmental ) degrees of freedom .
in summary , we have developed a prescription for calculating density matrices representing one- and two - photon states in the presence of unitary frequency - polarization coupling . with these methods ,
we have investigated decoherence in quantum systems arising from coupling to unobserved ( environmental ) degrees of freedom . by inducing frequency - dependent phase errors in photon polarization states and tracing over frequency degrees of freedom ,
we observe information loss as a particular form of decoherence . in the one photon case
, we applied the results of @xcite and established the feasibility and utility of `` bang - bang '' quantum control of decoherence in an optical system . by periodically exchanging the eigenstates of the `` environmental '' coupling , photon coherence ( polarization information )
is preserved .
we observed this effect in the case in which the _ strength _ of the coupling is varied ( the `` slow - flipping '' case ) ; the case in which the _ basis _ of the coupling is varied ; and the case in which both the basis _ and _ the strength of the coupling are varied simultaneously .
additionally , we demonstrated the utility of bang - bang control in the presence of a non - unitary ( dissipative ) coupling operation . in all cases ,
we see a qualitative reduction in decoherence for the case in which the quantum control procedure is implemented . in the two - photon case , we applied general results concerning the existence of decoherence - free subspaces in @xcite to the case of polarization entangled photon pairs produced in a down - conversion scheme .
we found that energy conservation in the down - conversion process gives rise to frequency entanglement which alters the assumptions underlying the prediction of a dfs .
in particular , frequency anti - correlation breaks the permutation symmetry of the environmental coupling between opposite photon paths by anti - correlating their respective phase errors .
we found that modifying the experimental apparatus to restore the symmetry of the phase errors allows us to produce a truly collective decoherence process , and consequently we recovered the singlet state as a dfs . in conclusion , we have investigated two procedures for reducing the decoherence of an optical qubit subjected to non - dissipative phase errors and non - unitary dissipative errors .
the first procedure , `` bang - bang '' quantum control , preserves polarization information in a single - photon by rapid , periodic control operations which effectively average out decoherence effects .
the second procedure requires entanglement between two photons so that the resulting state exhibits symmetry properties which are immune to collective decoherence .
these cases together demonstrate the feasibility of both active and passive protection of coherence properties in an optical qubit .
unitarity restricts the possible eigenvalues , @xmath16 , of @xmath156 : @xmath157 since @xmath16 is complex with modulus 1 , we may write @xmath158 where @xmath159 is a real - valued function of @xmath3 and @xmath20 , and the subscript refers to polarization mode @xmath24 . since @xmath31 is a linear operator ,
we need only show that eqs .
[ identity1 ] and [ identity2 ] hold for the basis states , @xmath160 and @xmath161 . from the definitions of @xmath31 and @xmath61 ( eqs .
[ r1]-[scriptu2 ] ) , we have @xmath162 similarly , @xmath163 thus we have , @xmath164 where the upper ( lower ) sign occurs when @xmath61 represents a polarization reflection ( rotation ) .
it follows immediately that @xmath165 since equations [ identity1 ] and [ identity2 ] hold for the basis states they hold for all linear combinations of basis states , which is to say , for all states .
all four relations follow easily from the definitions of @xmath61 and @xmath100 ( eqs . [ r1]-[r2 ] and [ scriptu11]-[scriptu22 ] ) .
equation [ relation1 ] is trivial , since by definition , both operators @xmath88 and @xmath89 act as the identity on @xmath49 . for the other three
, we apply the definitions in succession .
@xmath166 again , if @xmath61 represents a reflection ( rotation ) we use the upper ( lower ) sign . from these results
, we see that @xmath167 can be treated as a single operator @xmath168 with an associated phase error of @xmath169 . factoring out the global phase shift of @xmath91 in the last two relations ( indicated by the arrow in eqs .
[ a7 ] and [ a8 ] ) , we see that @xmath170 can be treated as a single operator @xmath171 with an associated phase error of @xmath172 .
we begin by defining @xmath113 in eq .
[ v_lambda ] to be an equal superposition of the basis states @xmath15 : @xmath173 then we have @xmath174 where @xmath117 indicates the real part .
we begin by modeling two - photon density matrices including frequency anti - correlation effects . substituting the appropriate phase functions ( eq . [ phase_before_modification ] ) into eq .
[ rho2 ] , we have an explicit expression for @xmath175 : @xmath176 we will now calculate @xmath39 assuming the initial state matrix elements , @xmath177 , are given by the density matrix representation of the four bell states , latexmath:[\[|\phi^\pm\rangle=\frac{1}{\sqrt{2}}(|hh\rangle\pm|vv\rangle),\ way , we see the effect of @xmath127 on each of these states . for brevity ,
let @xmath179 note that the dependence on @xmath180 ( the variable of integration ) drops out of the @xmath181 and @xmath182 terms of @xmath183 so that there is no decoherence due to the frequency distribution of the down - converted photons .
however , @xmath180 does not drop out of the @xmath184 and @xmath185 terms of @xmath186 . due to the finite width of the frequency term @xmath187 , for large @xmath57 ,
the integral in these terms approaches zero in the same manner as the fourier integrals of sect .
[ review ] .
note that the results developed above apply only to decoherence in the @xmath119 basis .
the states @xmath188 look different in other bases , so they are subject to decoherence in general . in order to model decoherence in other bases , and also to model non - collective decoherence , we must include rotation operators in eq .
[ rho2_nodfs ] .
we define an operator @xmath189 which represents a rotation by @xmath190 in arm l and a rotation by @xmath191 in path r. it can be shown that in the present matrix notation , @xmath192 it follows that @xmath193
. we can now generalize eq .
[ rho_formal ] by writing the reduced density operator representing polarization degrees of freedom after an input state @xmath194 has passed through crystals of thicknesses @xmath195 and @xmath196 at angles @xmath190 and @xmath191 in the photon paths labeled l and r : @xmath197 eq . [ rho_general ] can be solved numerically given @xmath19 ( @xmath17 ) , the eigenvalues which define @xmath198 . with the same techniques , we can model density matrices including both frequency and path anti - correlation effects . substituting the modified phase functions ( [ phase_after_modification ] ) into eq .
[ rho2_modified ] , we have an expression which includes both frequency and path anti - correlation effects : @xmath199 again , using the four bell states as input states , we see the effect of @xmath200 on each of these states .
@xmath201 note that in this case , the dependence on @xmath180 ( the variable of integration ) drops out of the @xmath185 and @xmath184 terms of @xmath186 . on the other hand , @xmath180 does not drop out of the @xmath181 and @xmath182 terms of @xmath183 .
again , the results developed above apply only to decoherence in the @xmath119 basis .
the analytical form of @xmath39 is helpful in developing a physical picture of the decoherence process . in order to calculate the corresponding result for decoherence in an arbitrary basis
, we simply set @xmath202 in eq .
[ rho_general ] and solve numerically .
this corresponds to the path anti - correlation procedure described earlier .
these results will be presented elsewhere @xcite , however , the theoretical predictions are unchanged for the state @xmath203 , since it looks the same in every basis .
experimental and theoretical work was performed in the quantum information laboratory of p. g. kwiat at lanl .
we wish to thank a. g. white for his willingness to assist in all stages of this work . at dartmouth ,
ajb gratefully acknowledges w. lawrence and m. mycek for offering support , advice and feedback at various stages . | we investigate coherence in one- and two - photon optical systems , both theoretically and experimentally . in the first case
, we develop the density operator representing a single photon state subjected to a non - dissipative coupling between observed ( polarization ) and unobserved ( frequency ) degrees of freedom .
we show that an implementation of `` bang - bang '' quantum control protects photon polarization information from certain types of decoherence . in the second case
, we investigate the existence of a decoherence - free " subspace of the hilbert space of two - photon polarization states under the action of a similar coupling .
the density operator representation is developed analytically and solutions are obtained numerically . |
over the past six decades , much effort has been devoted to determining the critical behavior of statistical properties of lattice models , which would allow a deeper understanding of order - disorder phenomena in magnetic solids .
following onsager s pioneering exact solution for the square lattice ising model @xcite , exact solutions were also obtained for other regular two - dimensional lattice structures @xcite . in particular ,
exact results have been attained for the triangular , honeycomb , kagome , and bathroom - tile lattices @xcite , @xcite , @xcite , as well as for two - dimensional models , such as the union jack ( centered square ) @xcite and the square kagome @xcite lattices .
geometrical frustration is mainly based on the triangle and tetrahedron structures , but it was also found in the ising model on a pentagonal penrose lattice proposed by waldor _ _ et al.__@xcite and solved exactly using the transfer matrix approach .
more recently urumov @xcite considered the ising model on the cairo pentagonal lattice using the decoration transformation @xcite .
this model has been mapped onto the union jack lattice @xcite and its critical temperature and spontaneous magnetization properties have been discussed .
this model is interesting from a mathematical point of view .
a few years ago , real materials with a cairo pentagonal lattice structure were found ; for example , the @xmath0 lattice in @xmath1 ( described as a pentagonal heisenberg model ) was discussed by ressourche _
et al_. @xcite .
this material shows magnetic frustration .
also , theoretical calculations of the phonon structure of antiferromagnetic @xmath1 ( space group _ pbnm _ no .
55 , @xmath2 ) were studied using lattice dynamics and these results were confirmed experimentally by polarized raman spectroscopy from 10 to 300 k @xcite .
more recently , some additional experimental studies were performed @xcite .
ralko @xcite also discussed the hard - core extended boson hubbard model on the cairo pentagonal lattice , using the numerical quantum monte carlo study of stochastic series expansion and cluster mean - field theory .
the purpose of this paper is to present a general exact solution of the pentagonal ising model and as a special case we obtain the urumov solution using a standard decoration transformation approach @xcite .
furthermore , we present a more simplified solution through the direct decoration transformation @xcite instead of the standard one @xcite . the generalized version of the latter @xcite is widely used to solve some two - dimensional decorated ising@xcite and ising - heisenberg @xcite models .
this paper is organized as follows . in sec .
ii we consider the detailed description of the ising model on the cairo pentagonal lattice . in sec .
iii we discuss its phase diagram at zero temperature .
section iv is devoted to the pentagonal ising model mapping , using the direct decoration transformation @xcite for the isotropic free - fermion vertex model @xcite , presenting the most relevant results and discussion . in sec .
v we obtain the finite - temperature phase diagrams , critical temperature , magnetization , entropy , and specific heat .
section vi summarizes our discussion .
the highly anisotropic heisenberg model on the cairo pentagonal lattice considered by ressourche _ et al_. @xcite could be reduced to the ising model on a cairo pentagonal lattice .
therefore , let us consider the ising model on a planar lattice where the tiling is achieved with non regular pentagons ; the lattice may be viewed as an assembly of checkerboard ordering with the elementary cell ( see fig .
[ fig : transf - pent ] ) rotated by @xmath3 in the neighboring square plaquettes , as shown in fig .
[ fig : pent - latt ] ( for more details see ref .
@xcite ) .
the hamiltonian of the cairo pentagonal ising model , ( represented schematically in fig . [ fig : pent - latt ] ) , discussed previously by urumov @xcite , is expressed by @xmath4 where the first summation is the contribution of the interaction between the nearest neighbor with spin @xmath5 ( @xmath5 interacting with coordination number 3 ) and @xmath6 corresponds to the interaction between @xmath5 and @xmath7 . while the second summation is the contribution of the nearest - neighbor interaction @xmath8 between spin @xmath9 and spin @xmath10 ( @xmath10 s interacting with coordination number 4 ) , conveniently we assume @xmath11 and @xmath12 .
in this section we discuss the phase diagram at zero temperature of the hamiltonian given in eq . .
in order to discuss the phase diagram at zero temperature , we define the magnetization @xmath13 for the pentagonal lattice that is used throughout the paper , given by @xmath14 with @xmath15 and @xmath16 .
the energy per plaquette of three ground states that appear for the pentagonal lattice ising model is expressed in terms of an elementary cell ( see fig . [
fig : transf - pent ] ) .
it is worth highlighting that the elementary cell should not be confused with the unit cell of the pentagonal lattice .
\(i ) the ferromagnetic ( fm ) state or saturated state has a total magnetization @xmath17 and ground - state energy per plaquette @xmath18 .
thus the fm state can be represented as @xmath19 this state is limited by @xmath20 for @xmath21 and@xmath22 for @xmath23 , as displayed in fig .
[ fig : ph - dgm ] .
\(ii ) the ferrimagnetic state ( fim ) has a total magnetization @xmath24 and ground - state energy per plaquette @xmath25 , which corresponds to the configuration displayed in fig .
[ fig : ph - dgm ] .
analogous to the previous case , we describe the state by @xmath26 this state is limited by @xmath27 for @xmath21 and @xmath28 for @xmath23 , as illustrated in fig .
[ fig : ph - dgm ] .
\(iii ) the frustrated state ( fru ) is given as a combination of states @xmath29 with its rotated elementary cell and spin inversion on the elementary cell with ground - state energy per plaquette @xmath30 , which can be expressed by @xmath31 this state is limited by @xmath32 ( see fig .
[ fig : ph - dgm ] ) . unit cells @xmath33
( @xmath34 ) with magnetization @xmath35 and unit cell @xmath36 with @xmath37 . ]
( a ) in the top ( left ) panel , we illustrate schematically an arbitrary lattice configuration , which is included in a frustrated state ; the dashed square is shown explicitly in the panel below .
( b ) in the top ( right ) panel , we assemble a lattice with two different unit cells @xmath38(@xmath39 and @xmath40(@xmath41 ) . in the panel
below , the sublattice ( the dashed square ) is explicitly shown ; the contributions of unit cells @xmath38 and @xmath40 are in different concentrations . ] we define @xmath42 as the magnetization for each frustrated - state configuration in the range between @xmath41 and @xmath43 ( we denote by @xmath13 the average of total magnetization ) . combining the state of the elementary cell displayed in eq . and its rotation in @xmath3 of the elementary cell , it is possible to generate the geometrically frustrated state . in particular , when we combine half states @xmath44 and the remaining states with @xmath45 , we obtain an antiferromagnetic state , with null total magnetization .
other intermediate states with magnetization @xmath46 also could be obtained by combining the elementary cell state with different relative amounts of the state given by eq . .
more specifically , the unit cell magnetization could be classified as displayed in fig .
[ fig : sublatt ] , i.e , @xmath33 ( @xmath34 ) and @xmath47 with magnetization @xmath48 and @xmath49 , respectively .
certainly , this is not the only way to classify the unit cell by its magnetization ; any other classifications of the unit cell lead us to the same kind of lattice configuration . in fig .
[ fig : latticeab ] we show two particular situations of such a lattice configuration , formed by the unit cells of type @xmath33 and @xmath40 . in the bulk limit
we have a lattice with total magnetization @xmath43 . in fig .
[ fig : latticeab](b ) the lattice is composed of unit cells @xmath50 and @xmath36 with a different concentration , generating a different total magnetization of the lattice .
thus we obtain the total magnetization of each particular configuration .
however , the average magnetization of all these configurations with equal energy will be null @xmath51 of the frustrated state in the interval @xmath52 .
phase diagram at zero temperature of the pentagonal ising model . ]
through the direct decoration transformation @xcite , it is shown that the pentagonal ising model is equivalent to a union jack lattice , which is in turn mapped onto the isotropic free - fermion eight - vertex model @xcite . from this equivalence
we can obtain the thermodynamics of the pentagonal ising model . in order to study the spin-1/2 ising model on a pentagonal lattice
, we introduce the notation for each elementary square plaquette depicted in fig .
[ fig : transf - pent ] ( this should not be confused with the unit cell ) . in this case
we assume @xmath53=@xmath54 , @xmath55=@xmath54 and @xmath56 .
schematic representation of the elementary plaquette and its mapping to the union jack lattice . ] the hamiltonian for an elementary plaquette as displayed in fig .
[ fig : transf - pent ] is given by @xmath57,\label{pa}\ ] ] where we are assuming @xmath58 and @xmath59 and @xmath60 is defined by @xmath61 , with @xmath62 the boltzmann constant and @xmath63 the absolute temperature .
instead of applying the standard decoration transformation @xcite as applied by urumov @xcite , we apply the direct decoration transformation @xcite to transform the plaquette into the union jack lattice @xcite . in this case
the hamiltonian associated with union jack lattice can be expressed by @xmath64 the boltzmann weights of the hamiltonian may be written @xmath65}},\label{eq : w1}\ ] ] whereas the boltzmann weights for the transformed plaquette is given by the relation @xmath66 similar to the previous notation , here we consider @xmath67 , @xmath68 and @xmath69 , where @xmath70 is taken as a constant shift energy , @xmath71 is the interaction parameter between the internal spin @xmath72 and each spin @xmath73 , and @xmath74 is the coupling term between spins @xmath73 . using the direct decoration transformation proposed in ref .
@xcite , we need to impose the condition , @xmath75 between eqs . and for arbitrary @xmath73 . therefore , we obtain only four non equivalent configurations @xmath76@xmath77 ,
@xmath78 , @xmath79 , and @xmath80 ; any other permutation or spin inversion falls onto one of these configurations .
thus the boltzmann weight for each configuration is given by @xmath81 the above equations satisfy the isotropic free - fermion condition @xcite fig .
[ fig:8-vertex ] . hence the free - fermion condition may be rewritten in terms of @xmath82 as @xmath83 therefore , the boltzmann factor of an effective union jack lattice can be expressed in terms of the pentagonal ising model coupling parameters @xmath84 for simplicity we used the notation @xmath85 and @xmath86 . due to the step by step decoration transformation performed by urumov @xcite @xmath6
was restricted only to @xmath21 or @xmath87 ( ferromagnetic ) ; otherwise , if we consider @xmath23 , we get a imaginary parameter in the intermediate transformation proposed by urumov .
however , using the direct decoration transformation , we do not have such a restriction ( fig .
[ fig : transf - pent ] ) , but only @xmath88 , which means that @xmath6 exchange coupling could be ferromagnetic or antiferromagnetic .
hence the pentagonal ising model is completely equivalent to the ising model on the union jack lattice @xcite with the isotropic nearest - neighbor interactions defined by @xmath89 and non crossing diagonal interactions between the second nearest neighbor given by @xmath90 .
in contrast , the union jack lattice was mapped onto the isotropic free - fermion eight - vertex model @xcite by choy and baxter @xcite .
therefore , we relate the boltzmann factor given by eqs .
( [ eq : rot - w1])-([eq : rot - w5 ] ) and the boltzmann factor of the union jack lattice given by eq . ( 4 ) of ref . @xcite .
these relations are given by @xmath91 the schematic representation of the eight - vertex model is given in fig .
[ fig:8-vertex ] .
( color online ) eight - vertex model diagrams ] these boltzmann weights will be used in the following section to study the critical temperature and spontaneous magnetization of the cairo pentagonal ising model .
in this section we discuss thermodynamical properties , such as the entropy , specific heat , and magnetization , as a function of temperature , as well as the critical temperature behavior .
the thermodynamics of the pentagonal ising model can be expressed following the results given by fan and wu @xcite .
the exact result for the free energy of the pentagonal ising model is then given by @xmath92\mathrm{d}\phi,\label{eq : mr}\ ] ] where @xmath93^{2}\nonumber \\ & + \xi_{1}\xi_{2}\left(4\xi_{2}\xi_{3}-\left(\xi_{1}+\xi_{2}\right)^{2}\right).\label{eq : q1}\end{aligned}\ ] ] once the free energy is known , we can obtain straightforwardly the critical temperature , magnetization , entropy , and specific heat . in order to study the spontaneous magnetization , following the result obtained by choy and baxter @xcite and using eq .
( [ eq : ome ] ) , the magnetization @xmath15 for spins with coordination number 4 is described by @xmath94{1-k^{2}},\label{eq : mag2}\ ] ] where @xmath95 equation ( [ eq : k11 ] ) is expressed in terms of the pentagonal ising model boltzmann factor .
it is important to note that this relation is valid for arbitrary spins @xmath96 and @xmath97 , as shown in fig .
[ fig : transf - pent ] . using the results obtained in ref .
@xcite , the critical point of the union jack lattice is obtained from the condition @xmath98 for @xmath99 or @xmath100 for @xmath101 ; in terms of @xmath102 , this means that the critical points occur at @xmath103 .
equivalently , using the pentagonal ising model boltzmann factor @xmath82 , we have @xmath104 this condition must satisfy the critical point . rewriting eq .
( [ eq : k22 ] ) in terms of @xmath105 and @xmath106 , the critical points must satisfy the relation @xmath107 where @xmath108 and @xmath109 denote @xmath105 and @xmath106 evaluated at the critical temperature @xmath110 .
the same expression could be obtained from eq.(8 ) of the ref .
@xcite ; however , due to the standard decoration transformation @xcite used by urumov@xcite for this model , we have to eliminate the intermediate parameter @xmath111 .
once the intermediate parameter @xmath111 is eliminated from eq .
( 8) of ref .
@xcite , it becomes identical to our eq .
for the case of @xmath21 ( ferromagnetic coupling ) , which was previously studied by urumov @xcite . in eq .
( [ eq : kk ] ) we provide a closed expression for the critical point of the pentagonal ising model .
the curves where the critical points occur are illustrated in fig .
[ fig : crit - points ] . critical points curve fo the pentagonal ising model as a function of the parameters @xmath112 and @xmath113 . ] the finite - temperature properties of the system arer investigated by considering the effect of parameters @xmath8 and @xmath6 on the critical behavior . in fig .
[ fig : crit - points ] we display the critical point regions in units of critical temperature , where the phase transitions of the fim region to a disordered phase ( dp ) and a dp to a fm phases are illustrated .
it is important to highlight that fig .
[ fig : crit - points ] becomes similar to fig .
[ fig : ph - dgm ] when @xmath114 .
phase diagrams as a function of temperature @xmath110 and the parameter @xmath115 for two different values of @xmath6 . in ( b )
the dashed lines show the limiting values of @xmath116 and @xmath114.,title="fig : " ] phase diagrams as a function of temperature @xmath110 and the parameter @xmath115 for two different values of @xmath6 . in ( b )
the dashed lines show the limiting values of @xmath116 and @xmath114.,title="fig : " ] phase diagram in the @xmath117 plane for the pentagonal ising model . ]
an alternative phase diagram is depicted in fig .
[ fig : phaset ] , where the phase diagrams are illustrated in the @xmath118 plane ( since @xmath110 is invariant unde ther @xmath119 exchange ) for fixed parameter @xmath6 . in fig .
[ fig : phaset](a ) the second - order phase transition line is shown in the @xmath118 plane when the parameter @xmath6 is fixed at @xmath120 ; in this case there is a dp .
concretely , for @xmath27 we show two regions : the fim phase and the dp ; for @xmath20 we have a dp and a fm phase . in fig .
[ fig : phaset](b ) we show the behavior of the critical temperature when @xmath121 , displayed as a solid line . for low values of @xmath122 , ( dashed lines ) , which is valid within the limit @xmath116 and @xmath114 .
this result is in agreement with the phase diagram at zero temperature ( see , for instance , fig .
[ fig : ph - dgm ] ) ; more specifically , for @xmath123 , the phase transition occurs at @xmath124 .
we now comment on the finite - temperature phase diagrams displayed in fig .
[ fig : grafico1 ] , in which the critical temperature @xmath125 is shown as a function of the parameter @xmath126 . using the equation of the critical points ( [ eq :
kk]),we obtain the plot illustrated in fig .
[ fig : grafico1 ] . from
that we can analyze three limiting cases .
( i ) : : for @xmath127 we obtain @xmath128 ( @xmath129 ) and the pentagonal lattice is reduced to a ferromagnetic ( ferrimagnetic , with total magnetization @xmath24 ) decorated square lattice .
( ii ) : : for @xmath130 , from our calculation , for @xmath20 and @xmath21 , we find the solution @xmath131 ; in this case the pentagonal lattice is reduced to the ferromagnetic square lattice . in the case
@xmath27 and @xmath23 , in the limit under consideration ( @xmath130 ) , the pentagonal lattice falls into the bottom righ- hand corner , which is a dp state ( see fig .
[ fig : grafico1 ] ) .
( iii ) : : for @xmath132 we obtain @xmath133 for @xmath27 and @xmath21 and the pentagonal lattice is reduced to a ferrimagnetic ( with total magnetization @xmath24 ) square lattice . meanwhile ,
for @xmath132 for @xmath20 and @xmath23 the pentagonal lattice falls into a disordered state , which corresponds to the top left corner fig .
[ fig : grafico1 ] . when @xmath134 , according to a pentagonal lattice mapping onto an effective square lattice ( see fig .
[ fig : transf - pent ] ) , within the limit @xmath130 , the effective lattice is reduced to a ferromagnetic square lattice . in the limit @xmath132 the pentagonal ising model reduces to an antiferromagnetic square lattice .
the internal energy of the pentagonal lattice ising model defined by @xmath135 can be obtained straightforwardly from eq . .
[ intrn - energy ] we display the internal energy as a function of coupling parameter @xmath8 in the low - temperature limit in order to observe the low - lying energy contribution close to the critical temperature assuming @xmath120 .
the dashed line corresponds to the dp internal energy at @xmath136 ( the ground - state energy ) , given by @xmath137 , while the dash - dotted line indicates the internal energy @xmath138 from which the phase transition at zero temperature occurs at @xmath139 . in the case
@xmath114 and assuming @xmath140 with @xmath141 , we have @xmath142 and then it is possible to write @xmath143 .
further , by substituting into eq .
( [ eq : kk ] ) , after some algebraic manipulations we obtain @xmath144 , where @xmath145 was defined previously .
thus a second - order phase transition occurs at @xmath146 .
it is worth highlighting that the lowest critical temperature occurs at @xmath147 .
therefore , the contribution of the low - lying energy is absorbed by the second - order phase transition .
certainly there is no second - order phase transition for @xmath148 @xmath149 .
the solid blue ( thick ) line and the dark - blue ( thin ) curve represent the internal energy in the disordered phase , while by the red ( thick ) line and the orange ( thin ) line curve represent the ferromagnetic region for two critical temperatures @xmath150 and @xmath151 .
( color online ) internal energy @xmath152 as a function of @xmath8 for a fixed value @xmath120 .
the dashed line corresponds to the dp ground energy and the dash - dotted line corresponds to the fm ground energy . ] in fig .
[ fig : int - enrg - ut ] we plot the internal energy as a function of temperature for several values of @xmath8 around the second- order phase transition , assuming a fixed value for @xmath120 .
the black dotted line represents the internal energy @xmath153 evaluated at the critical temperature @xmath110 given by eq . at low temperature . below the critical temperature ,
the internal energy is almost constant ( ferromagnetic phase ) , which means that it is mainly given by the zero - temperature ground - state energy .
although at a critical temperature there is a sudden change of curvature , this change becomes dramatic for lower critical temperature ; for higher critical temperature this change of curvature becomes smoother .
the lowest critical temperature occurs at @xmath147 for @xmath139 ; therefore , for lower values of the coupling parameter @xmath8 there is no second - order phase transition . for a sufficiently high temperature
the internal energy leads to an asymptotic limit , whereas the internal energy increases almost proportionally to the temperature .
( color online ) internal energy @xmath152 as a function of @xmath63 and a fixed value @xmath120 .
the dotted curve corresponds to @xmath153 . ]
the total magnetization @xmath13 is given by eq .
for the pentagonal ising model , which will be discussed in order to show the spontaneous magnetization .
it ought to be pointed out that the calculation of the magnetization of internal spin @xmath5 , @xmath154 , may be obtained following the results obtained by choy and baxter @xcite , which could be expressed by the relation @xmath155 where after some algebraic manipulations the coefficients become @xmath156 defining @xmath157 and @xmath158 in analogy to the boltzmann factors @xmath82 , we have @xmath159 in order to obtain the three - spin correlation function @xmath160@xmath161 we use a checkerboard ising model equivalent to that used by choy and baxter @xcite . here
we use the relation obtained in ref .
@xcite , but in our case we rewrite this relation in terms of the pentagonal ising model boltzmann factor . hence , using some algebraic manipulations we have @xmath162 where @xmath163 therefore , the magnetization @xmath164 can be expressed as a function of magnetization @xmath165 , which is given by @xmath166m_{0}.\label{eq : mag3}\ ] ] it is important to highlight that eq .
is expressed in terms of the original parameters of the pentagonal ising model instead of parameters of the effective hamiltonian such as those obtained by urumov @xcite .
using eq . , we are easily able to manipulate the parameters of the pentagonal ising model in order to discuss the spontaneous magnetization . at the critical point
we need to substitute the expression @xmath167 , where @xmath108 is defined in eq . . from eqs
and we can obtain a closed expression for the total magnetization of the ising model on a pentagonal lattice , using the relation given by eq . .
total magnetization of the pentagonal ising model for three different values of @xmath63 as a function of the parameter @xmath8 and fixed parameter @xmath6 .
( a ) @xmath120 and ( b ) @xmath121.,title="fig : " ] total magnetization of the pentagonal ising model for three different values of @xmath63 as a function of the parameter @xmath8 and fixed parameter @xmath6 .
( a ) @xmath120 and ( b ) @xmath121.,title="fig : " ] we now discuss the behavior of the total magnetization of the pentagonal ising model as a function of the parameter @xmath8 for the low - temperature limit . in fig .
[ fig : t vs j ] we plot the magnetization at low temperature as a function of @xmath8 , where we display two types of plateaus for the fim state and the fm state .
this is in agreement with the phase diagram displayed in fig .
[ fig : ph - dgm ] and [ fig : crit - points ] , whereas the intermediate state corresponds to the fru phase and the dp , respectively .
hence , in fig . [ fig : t vs j](a ) , for @xmath120 , when @xmath136 there are three well defined regions : fm phase with @xmath17 , the fim phase with @xmath24 , and the intermediate fru phase ; also , by increasing the temperature ( for example , from @xmath168 to @xmath169)the disordered phase increases(the @xmath115 increases ) .
meanwhile , in fig .
[ fig : t vs j](b ) , for @xmath121 and at zero temperature , we have a direct phase transition between the fm state and the fim state .
however , for a nearly zero temperature @xmath168 , a small intermediate region arises that corresponds to the disordered phase region . for higher temperature such as @xmath170 ,
the dp region is even larger .
temperature dependence of the total magnetization of the pentagonal ising model for two different values of @xmath112 and a fixed value @xmath120 . ]
another way to analyze the total magnetization is by exploring the temperature dependence @xmath171 of the total magnetization . in fig .
[ fig : m vc t ] we plot the magnetization as a function of temperature for two values of @xmath172 and @xmath120 , where the magnetization of the saturated fm state and the fim state is illustrated . for the case of
@xmath173 we have the fm region with @xmath17 the total magnetization at zero temperature .
the total magnetization vanishes at @xmath174 as temperature increases ; therefore , a second - order phase transition occurs at the critical temperature . for @xmath175
, the total magnetization corresponds to the fim region , with @xmath24 at zero temperature , and it vanishes at @xmath174 as displayed in fig .
[ fig : m vc t ] .
thus a second - order phase transition occurs again at the critical temperature , which is in agreement with the critical point curve displayed in fig .
[ fig : crit - points ] . the entropy can be easily obtained as a negative temperature derivative from the free energy @xmath176 , while the specific heat can be written as a temperature derivative from the entropy @xmath177 . in what follows
we consider only the case @xmath20 since for @xmath27 we have the same behavior . in fig .
[ fig : entro ] we display the entropy as a function of temperature for several values of @xmath8 with @xmath120 fixed .
figure [ fig : entro ] ( a ) shows the low - temperature behavior of entropy , where the residual entropy appears at @xmath178 for @xmath179 .
this means that we are in a geometrically frustrated region , which is in agreement with the illustration of the phase diagram in fig .
[ fig : ph - dgm ] .
the residual entropy is proportional to @xmath180 ; this number comes from the two configurations given in eq .. for @xmath181 the residual entropy has a different nontrivial value as displayed in fig .
[ fig : entro](b ) .
this is due to the degeneration of the phase boundary between the fru and fm ( or fim ) regions at @xmath136 ( for detail see fig .
[ fig : ph - dgm ] ) .
this result was derived in a way similar to that discussed by wannier @xcite for the case of a two - dimensional triangular lattice . while in fig .
[ fig : entro](c ) there is no residual entropy for @xmath182 , the standard temperature dependence of entropy appears with a strong change of curvature located at critical points where a second - order phase transition occurs .
it ought to be highlighted that in the low - temperature limit ( below the critical temperature ) the entropy ( in the fm state ) for @xmath183 and with @xmath149 can be obtained from eq . .
more explicitly , by fixing @xmath140 , where @xmath184 , it is possible to write @xmath185 , with @xmath186 and @xmath187 .
substituting eqs .
( [ eq : a1 ] ) and ( [ eq : q1 ] ) into eq .
( [ eq : mr ] ) , the integration of eq .
( [ eq : mr ] ) results in @xmath188 finally , the entropy @xmath160@xmath176 can be written as @xmath189 where @xmath190 is the energy gap . in figs . [
fig : entro](c ) and [ fig : entro ] ( d ) the low - temperature limit is well fitted by the above limiting expression .
( color online ) entropy as a function of temperature for @xmath120 and ( a ) and ( b ) @xmath191 and ( c ) and ( d ) @xmath183 ] ( color online ) entropy as a function of temperature for @xmath120 and ( a ) and ( b ) @xmath191 and ( c ) and ( d ) @xmath183 ] ( color online ) entropy as a function of temperature for @xmath120 and ( a ) and ( b ) @xmath191 and ( c ) and ( d ) @xmath183 ] ( color online ) entropy as a function of temperature for @xmath120 and ( a ) and ( b ) @xmath191 and ( c ) and ( d ) @xmath183 ] . an additional plot of entropy @xmath193 against @xmath8 in the low - temperature limit is displayed in fig .
[ fig : s - j - fxd - t ] , where the residual entropy is illustrated by the dashed black line at zero temperature . for @xmath148
there is residual entropy @xmath194 , while for @xmath139 the residual entropy becomes @xmath195 ; for higher values of @xmath183 there is no residual entropy .
thereafter , we observed the entropy in the low - temperature limit , where we can show the effects of residual entropy .
the low - lying energy contribution for the entropy between the dp and the fm phase is absorbed by the second - order phase transition as a consequence of the entropy falling dramatically to zero entropy for @xmath183 ; for higher temperature the entropy change curvature becomes softer .
( color online ) entropy @xmath193 as a function of the parameter @xmath8 for a fixed low temperature and @xmath120 . ]
finally , we conclude our analysis of thermodynamics by exploring the temperature dependence of the specific heat .
some typical thermal variations of the specific heat of the pentagonal ising model are plotted in fig .
[ fig : calor ] for several values of @xmath8 and @xmath120 fixed . in figs .
[ fig : calor](a ) and ( b ) we present the temperature dependence of the specific heat in the dp states and show that there is no phase transition at finite temperature because we are observing the frustrated region . in figs .
[ fig : calor](c ) and ( d ) the specific heats are logarithmically divergent at the critical temperature , which is associated with a continuous phase transition between the spontaneously ordered and disordered phases .
clearly , this means that we are facing a fm region , which can be verified in the phase diagram illustrated in fig .
[ fig : ph - dgm ] .
( color online ) specific heat as a function of temperature for @xmath120 and ( a ) and ( b ) @xmath191 and ( c ) and ( d ) @xmath183 . ]
( color online ) specific heat as a function of temperature for @xmath120 and ( a ) and ( b ) @xmath191 and ( c ) and ( d ) @xmath183 . ]
( color online ) specific heat as a function of temperature for @xmath120 and ( a ) and ( b ) @xmath191 and ( c ) and ( d ) @xmath183 . ]
( color online ) specific heat as a function of temperature for @xmath120 and ( a ) and ( b ) @xmath191 and ( c ) and ( d ) @xmath183 . ]
plotted in fig .
[ fig : calor](c ) when @xmath196 and @xmath149 is the specific heat versus temperature ; a surprisingly almost null specific heat is displayed , until the critical temperature is achieved .
when the absolute value of the exchange interaction is only slightly above 1 , the critical temperature ( an order - disorder phase transition ) can be obtained easily from eq . ; this critical temperature is approximately given by the expression @xmath197 in the low - temperature limit . in fig .
[ fig : cvsj - fxd - t ] we display the specific heat as a function of @xmath8 in the low - temperature limit , where we show the specific heat behavior around the second - order phase transition .
as discussed previously , the low - lying energy contribution is absorbed by a second - order phase transition in the case of @xmath183 , while for @xmath148 there is no second - order phase transition ; then we can observe the low - lying energy contribution as a small anomalous broad peak .
( color online ) specific heat as a function of @xmath8 in the low - temperature limit and @xmath120 . ] similarly , the low - temperature asymptotic limit for specific heat can be derived from eq . so that @xmath198 the energy gap is large enough even when @xmath199 because the order - disorder transition occurs for @xmath200 . in fig .
[ fig : lw - spc - ht ] we show the magnification of fig .
[ fig : calor](c ) in low - temperature limit , which is well fitted by eq .. the solid line corresponds to the exact specific heat and the dashed line represents the low - temperature asymptotic limit of the specific heat .
for @xmath199 the corresponding low - temperature approximation is valid for @xmath201 , while for @xmath202 and @xmath203 clearly the low - temperature curve accompanies quite well the exact solution .
( color online ) low - temperature limit specific heat against temperature for the same values as in fig .
[ fig : calor](c ) .
the solid line corresponds to the exact specific heat and the dashed line is the low - temperature limit of the specific heat . ]
using the direct decoration transformation @xcite , we have solved the pentagonal ising model with a more general coupling parameter and compared it with urumov s @xcite solutions .
we have found a frustrated phase of the pentagonal ising model .
in addition , we have obtained a simplified solution for the free energy , as well as a closed expression for the critical temperature .
although this model has already been solved by urumov through the standard decoration transformation @xcite in the nonfrustrated region ( @xmath21 ) , such a result contains unnecessary intermediate parameters that can be avoided so that a closed expression similar to eq . can be obtained .
we have studied the ground - state phase diagram , which exhibits a ferromagnetic state , a ferrimagnetic state , and a frustrated state at @xmath204 . following the exact solution for the pentagonal ising model
, we have discussed the finite - temperature phase diagram , as shown in figs .
[ fig : phaset ] and [ fig : grafico1 ] , identifying the phase transition between the fim state and the dp state and also between the dp state and the fm state . the analysis of the limits in fig .
[ fig : grafico1 ] allows one to find three relevant phases . for @xmath205
the pentagonal lattice reduces to a ferromagnetic ( ferrimagnetic ) decorated square lattice . for @xmath206 and @xmath20
the pentagonal lattice reduces to the ferromagnetic square lattice . finally , for @xmath207 and @xmath27 the pentagonal ising model reduces to a ferrimagnetic square lattice .
the total magnetization as a function of the parameter @xmath8 and for a fixed value of @xmath6 for the ferromagnetic state @xmath208 and the ferrimagnetic state @xmath209 is shown in fig .
[ fig : t vs j ] . for a fixed value of @xmath149
there is a residual entropy @xmath210 for @xmath179 and @xmath181 a nontrivial residual entropy @xmath211 is found , as shown in fig .
[ fig : entro ] .
because of the frustrated state , the entropy below the critical temperature shows a strong change of curvature for @xmath196 .
the specific heat capacity was also investigated at fixed @xmath149 and @xmath212 [ see figs .
[ fig : calor](a ) and [ fig : calor](b ) ] . for @xmath149 and @xmath213
we have unusual behavior due to frustration of the entropy and the heat capacity at temperatures below the critical value , as shown in figs .
[ fig : calor](c ) and [ fig : calor](d ) .
m.r . acknowledges fapemig for financial support .
o.r . and s.m.d.s .
thank cnpq and fapemig for partial financial support . | through the direct decoration transformation approach , we obtain a general solution for the pentagonal ising model , showing its equivalence to the isotropic free - fermion eight - vertex model .
we study the ground - state phase diagram , in which one ferromagnetic ( fm ) state , one ferrimagnetic ( fim ) state , and one frustrated state are found . using the exact solution of the pentagonal ising model
, we discuss the finite - temperature phase diagrams and find a phase transition between the fim state and the disordered state as well as a phase transition between the disordered state and the fm state .
we also discuss some additional remarkable properties of the model , such as the magnetization , entropy , and specific heat , at finite temperature and at its low - temperature asymptotic limit .
because of the influence of the second - order phase transition between the frustrated and ferromagnetic phases , we obtain surprisingly low values of the entropy and the specific heat until the critical temperature is reached . , |
in relativistic plasmas , such as a quark - gluon plasma ( qgp ) possibly formed in relativistic heavy ion collisions or a qed plasma in a supernova explosion , and in relativistic , degenerate matter , as e.g. strange quark stars , medium effects are of great importance .
for example , the originally massless or light particles of the system acquire an effective mass caused by the interactions with the other particles of the system . also screening effects due to the presence of ( color ) charges in the system ( debye screening ) prevent or at least reduce infrared singularities coming from the long range interactions in qed or qcd .
these effects can be calculated from the self energy of the particles .
evaluated perturbatively to lowest order at high temperature @xmath1 or density ( corresponding to a large chemical potential @xmath2 ) the effective mass and the debye mass are of the order @xmath3 or @xmath4 , where @xmath5 is the coupling constant .
these masses can be large compared to the bare masses , which are negligible at high temperature . the dispersion relations of the quasiparticles thus deviate from the ones of bare particles leading to important modifications of the equation of state @xcite .
therefore medium effects are an essential feature of relativistic systems at finite temperature or density .
so far these medium effects have mainly been studied by considering the self energies in the high temperature approximation @xcite or , equivalently , in the hard thermal loop ( htl ) limit @xcite . in this approximation
the momenta of the internal particles of the self energy are assumed to be hard , i.e. of order of the temperature or chemical potential , whereas the external momenta are soft , i.e. of the order @xmath3 ( @xmath4 ) .
these momentum scales can be distinguished only in the weak coupling limit @xmath6 , on which the htl approximation is based .
the reason for these restrictions is twofold .
first this approximation allows for analytic expressions of the self energies . secondly , which is more important , the self energies calculated in this way are gauge independent guaranteeing gauge independent results for the effective masses , the dispersion relations , and the debye screening . as a matter of fact
, the htl approximation builds the basis for a consistent extension of the naive perturbation theory , which suffers from infrared singularities and gauge dependent results for observables .
resumming the htl self energies ( and vertices ) leads to effective green s functions , which have to be used in the case of soft external momenta ( braaten - pisarski method @xcite ) . in this way consistent results for physical quantities can be found , which are gauge independent and complete to leading order in the coupling constant . at the same time the infrared behavior
is greatly improved due to the use of resummed propagators , although there might be still infrared divergences owing to the absence of a static magnetic screening in the htl self energies of the gauge bosons . for applications of this powerful method to interesting problems of the qgp
we refer to ref.@xcite . here
we want to compute the self energies and dispersion relations of photons , gluons and fermions at finite temperature but zero chemical potential within the one - loop approximation abolishing the htl assumption and compare the results to the ones of the htl approximation .
this will give us a hint for the validity of the weak coupling assumption , on which the htl approximation and thus the braaten - pisarski method relies .
up to now there are only explicit expressions for the gluon self energy beyond the htl limit , where the next term of the high temperature expansion has been evaluated @xcite . for a complete one - loop calculation , which can be done only numerically ,
there exist only integral expressions @xcite , results for certain momentum limits @xcite , or analytic expressions for zero temperature but non - zero chemical potential @xcite .
furthermore , it is instructive to have a look at quantities which vanish at leading order of the one - loop high-@xmath1 expansion . as examples we consider here the imaginary parts of the self energies above the light cone and the gauge dependence of the fermion self energy
first we consider the photon self energy or polarization tensor of qed within the one - loop approximation . in the vacuum it reads ( @xmath7 , @xmath8 ) @xmath9 \label{e1}\ ] ] where @xmath10 denotes the electron propagator and @xmath11 . in order to compute @xmath12 at finite temperature
we apply the imaginary time formalism and , for temperatures @xmath13 , neglect the electron mass , i.e. @xmath14 where @xmath15 with @xmath16 for integers @xmath17 . to perform the sum over the matsubara frequencies
@xmath18 it is convenient to use the saclay representation of the propagators @xcite : @xmath19 with @xmath20 e^{-k \tau } - n_f(k ) e^{k \tau } \right\ } , \label{e3}\ ] ] where @xmath21 $ ] denotes the fermi distribution . defining @xmath22 and @xmath23 , ( @xmath24 ) , the longitudinal and transverse projections , respectively , of the self energy read _
l^qed(p ) & = & - 4 e^2 t
_ k_0 ( k_0 q_0 + * k * ) _ f(q ) _
f(k ) + _ t^qed(p ) & = & - 4 e^2 t
_ k_0 ( k_0 q_0 - ( * p*)(*p * ) ) _
f(k ) . [ e4 ] in the htl approximation , i.e. assuming @xmath25 , analytic expressions can be obtained _
l^htl(p ) & = & -3 m_^2 , + _
t^htl(p ) & = & 32 m_^2 , [ e5 ] where @xmath26 denotes the effective , thermal photon mass .
the derivation of ( [ e5 ] ) can be found , e.g. , in the appendix of ref.@xcite .
now we want to evaluate the complete one - loop self energies starting from ( [ e4 ] ) , which we will write as _
l^qed(p ) & = & - 4 e^2 , + _
t^qed(p ) & = & - 4 e^2 , [ e6 ] where @xmath27 \left ( \frac{1}{p_0+k+q } - \frac{1}{p_0-k - q } \right ) \right . \nonumber \\ & & \quad + \left .
[ n_f(q ) - n_f(k ) ] \left ( \frac{1}{p_0+k - q } - \frac{1}{p_0-k+q } \right ) \right\ } \label{e7}\end{aligned}\ ] ] and @xmath28 \left ( \frac{1}{p_0+k+q } - \frac{1}{p_0-k - q } \right ) \right . \nonumber \\ & & \qquad -\left .
[ n_f(q ) - n_f(k ) ] \left ( \frac{1}{p_0+k - q } - \frac{1}{p_0-k+q } \right ) \right\}. \label{e8}\end{aligned}\ ] ] next we neglect the vacuum parts , i.e. the terms which do not contain any distribution functions in ( [ e7 ] ) and ( [ e8 ] ) , as we are interested only in the temperature dependent medium effects .
the angle integration over @xmath29 can be performed analytically .
( for this purpose we remove the @xmath30-dependence in @xmath31 by shifting @xmath32 in the corresponding terms . )
the remaining integral over @xmath33 has to be done numerically .
the final results for the real part of the self energy are shown in fig . 1 and fig . 2 , where they are also compared to the htl results ( [ e5 ] ) .
for the longitudinal component , we observe a surprisingly good agreement between the two approaches even when the htl assumption @xmath34 is not fulfilled .
the transverse component , however , is more sensitive to the violation of the htl assumption . for small @xmath35 and @xmath36 the one - loop result approaches the htl limit as expected . obviously , near the light cone
the results of both approaches coincide even outside the range of validity of the high-@xmath1 assumption .
indeed , the behavior of , e.g. , the longitudinal self energy , for @xmath37 and _ all _ values of @xmath35 is readily found to be @xmath38 $ ] , which is just the htl result ( [ e5 ] ) extrapolated to arbitrary , nearly light - like momenta .
next we consider the dispersion relations , which are given by the poles of the effective photon propagator following from resumming the self energy .
for the longitudinal mode ( plasmon ) , which is absent in the vacuum , and the transverse mode , respectively , they are determined by ^-1 & = & p^2-_l^qed(p)=0 + d_t^ * qed(p)^-1 & = & p^2-_t^qed(p)=0 .
[ e9 ] as in the vicinity of the light cone the deviations of both approaches are small , the same is expected for the dispersion relations for not too large couplings .
indeed , even in the case of extrapolating the coupling constant to @xmath39 , the difference is less than 5% ( fig .
we note that in a good approximation the one - loop dispersion relations follow > from the htl approximations by a suitable rescaling of @xmath40 .
as expected , both branches meet for @xmath41 in either approximation .
here we compute the complete one - loop gluon self energy which , in contrast to the abelian boson polarization , is gauge dependent . in feynman gauge , the pure gauge contribution of the gluon loop , the gluon tadpole and the ghost loop diagram is @xcite @xmath42 \delta_b(k ) \delta_b(q ) \ , , \label{e30}\ ] ] where @xmath43 with @xmath44 . the saclay representation of the bosonic propagator reads @xmath45 with @xmath46 e^{-k \tau } + n_b(k ) e^{k \tau } \right\ } \label{e13}\ ] ] and the bose distribution @xmath47 $ ] . as for the photon
, we consider the longitudinal and transverse components of @xmath48 .
expressed by the functions @xmath49 , @xmath50 defined in analogy to the fermionic functions in ( [ e7],[e8 ] ) with @xmath51 replaced by @xmath52 , they read _
l^qcd & = & 3 g^2 , + _ t^qcd & = & 3 g^2 . [ e31 ] under the htl assumption , terms @xmath53 drop out and the integrals over the bosonic functions yield up to a factor of two the same expressions as for the fermionic case , so the self energies obviously approximate to just the abelian expressions ( [ e5 ] ) with @xmath40 now replaced by the thermal gluon mass @xmath54 . in comparison to the abelian case for the same values of @xmath55 ,
the difference to the htl result is evidently larger ( fig . 4 and fig . 5 ) .
again we find that for soft momenta as well as for @xmath37 the htl limit is approached and that the transverse part is more sensitive to the overstraining of the htl assumption .
in general , beyond the leading contribution of the high-@xmath1 expansion , the ( one - loop ) self energies have a non - vanishing imaginary part above the light cone , for @xmath56 . for the gluon self energies it is shown in figs . 6 and 7 . while the general requirement @xmath57 , corresponding to a positive damping rate , is fulfilled for the one - loop photon self energy , it is violated by , e.g. , the transverse gluon self energy ( in feynman gauge ) .
this inconsistency indicates that extending the high-@xmath1 expansion beyond leading order requires higher - loop contributions to obtain a consistent propagator .
the positive imaginary parts above the light cone thus provide a measure of the magnitude of these sub - leading terms which may be of a similar order as the imaginary htl contributions for @xmath58 .
although self energies containing gluon propagators depend on the choice of the gauge , the dispersion relations are , in principle , observable and thus gauge independent . while the gauge dependent terms of the gluon self energy are ( somewhat surprisingly ) found to be of sub - leading order in the high-@xmath1 expansion and the htl - dispersion relations are thus gauge invariant , this is not the case in the one - loop approximation . postponing the question of the gauge effects to the next section , here in feynman gauge the influence of the sub - leading terms neglected in the htl - limit is considered . for the dispersion relations of the longitudinal and transverse gluons we find a decreasing gluon energy @xmath59 with increasing @xmath5 .
this qualitatively resembles to the next - to - leading order behavior of the gluon mass @xcite .
the deviations from the htl result are larger than in the case of photons for the same extrapolated coupling strength ( fig .
here we consider the electron or quark self energy within the one - loop approximation . at zero temperature
it is given by @xmath60 where the casimir invariant of the fundamental representation is @xmath61 for quarks ( for electrons , @xmath62 ) .
we will first choose the feynman gauge for the gauge boson propagator , @xmath63 combining ( [ e10 ] ) and ( [ e11 ] ) , the finite temperature self energy is given by @xmath64 for evaluating this expression we utilize a general representation of the fermion self energy in the rest frame of the heat bath @xcite : @xmath65 with the scalar functions @xmath66 - p_0 tr[\gamma_0 \sigma ] \right ) \nonumber \\
b(p ) & = & \frac{1}{4 p^2 } \left ( p^2 tr[\gamma_0 \sigma ] - p_0 tr[p{\!\!\!\!/ \,}\sigma ] \right ) . \label{e16}\end{aligned}\ ] ] evaluating the traces @xmath67 \ , , \nonumber \\
t_2(p ) & \equiv & tr[p{\!\!\!\!/ \,}\sigma(p ) ] \ , , \label{e17}\end{aligned}\ ] ] we are able to determine the functions @xmath68 and @xmath69 .
performing the sum over @xmath18 we get @xmath70 where @xmath71 \left ( \frac{1}{p_0+k+q } + \frac{1}{p_0-k - q } \right ) \right .
\nonumber \\ & & \qquad + \left .
[ n_b(q ) + n_f(k ) ] \left ( \frac{1}{p_0+k - q } + \frac{1}{p_0-k+q } \right ) \right\}. \label{e19}\end{aligned}\ ] ] again we neglect the vacuum contribution . after integrating analytically over the angle and numerically over @xmath33 we end up with the final result for @xmath72 . in fig . 9
the real part of @xmath72 is compared to the htl result @xmath73 which contains the effective , thermal fermion mass @xmath74 as for the boson self energies we observe the one - loop results approaching the htl predictions for @xmath34 and @xmath37 ( fig .
similar results can be found for @xmath75 , which follows from @xmath76 , \label{e22}\end{aligned}\ ] ] where @xmath77 \left ( \frac{1}{p_0+k+q } - \frac{1}{p_0-k - q } \right ) \right . \nonumber \\ & & \qquad + \left . [ n_b(q ) + n_f(k ) ] \left ( \frac{1}{p_0+k - q } - \frac{1}{p_0-k+q } \right ) \right\}. \label{e23}\end{aligned}\ ] ] the comparison of the real part of @xmath75 with the htl result @xmath78 is shown in fig . 10 .
in contrast to the htl approximation , both one - loop fermion self energy functions develop an imaginary part . as shown in fig .
11 and fig . 12
, these may be as large as the real parts .
we note that the relevant combinations @xmath79 ( see below ) have non - positive imaginary parts .
next we consider the dispersion relations following from the resummed fermion propagator .
it is convenient to decompose this propagator according to its helicity eigenstates @xcite : @xmath80 with @xmath81 the dispersion relations following from the roots of ( [ e26 ] ) are depicted in fig . 13
. the lower branch ( plasmino ) corresponding to @xmath82 and displaying a minimum at finite momentum , which has a negative helicity to chirality ratio @xcite , is absent in the vacuum . in fig .
13 the complete one - loop dispersion relations ( in feynman gauge ) are compared to the htl ones .
the differences increase with increasing coupling constant .
however , even for @xmath83 the maximum deviation is of the order of 10% .
finally we investigate the gauge dependence of the one - loop fermion self energy . for this purpose
we adopt the gauge boson propagator in a general covariant gauge @xmath84 ^ 2 } \label{e27}\ ] ] with the gauge fixing parameter @xmath85 ( see e.g. @xcite ) . for treating the double pole in the second term
we introduce a regularization parameter @xmath86 @xcite and write @xmath87 ^ 2 } = \left .
\frac{\partial}{\partial \lambda } \frac{1}{k^2-\lambda } \right|_{\lambda=0 } = \left .
\frac{\partial}{\partial \lambda } \delta_b(k_\lambda ) \right|_{\lambda=0}. \label{e28}\ ] ] here @xmath88 indicates the four momentum @xmath89 with @xmath90 . substituting the gauge dependent part of the boson propagator ( [ e27 ] ) into the self energy ( [ e14 ] ) we obtain @xmath91 after some manipulations similar to the one for computing @xmath72 and @xmath75 we find the gauge dependent parts @xmath92 and @xmath93 of the complete one - loop expressions for @xmath72 and @xmath75 as a function of @xmath36 and @xmath35 ( fig .
14 and fig .
the fact that by ward - identities and power counting arguments @xcite they are of the same order as typical corrections of htl by sub - leading terms is numerically found to hold also for hard momenta @xmath94 .
as expected , the gauge dependent contributions vanish for small @xmath36 and @xmath35 , where the one - loop self energy approaches the htl result .
in this work we have calculated the complete one - loop self energies for photons , gluons , and fermions ( electrons or quarks ) and compared them to their htl limits . in the case of the gauge independent one - loop photon self energy we found a good agreement between these two approaches . especially for the dispersion relations both approximations agree to a surprisingly high accuracy .
as far as the real parts of the gluon and fermion self energies are concerned , the difference between the both approaches is more distinct .
for momenta near the light cone , however , both approaches coincide even for hard momenta violating the htl assumption @xmath34 .
this means that for the htl approximation being valid we need not to require that both @xmath36 and @xmath35 are soft , but assuming a soft @xmath95 is sufficient .
this also justifies the use of the htl resummed quark propagator for soft four - momenta in the calculation of the production rate of hard photons from the qgp @xcite .
for the imaginary parts there are somewhat larger differences between the complete one - loop and the htl calculation .
in particular the one - loop fermion self energy shows a significant imaginary part above the light cone which may be of the same order as the real part .
( the same is expected for the ( complete ) gluon self energy beyond leading htl order . ) finally we have studied the gauge dependence of the one - loop fermion self energy in the class of covariant gauges .
the results found above indicate that the htl approximation , based on the weak coupling limit assumption @xmath6 or equivalently on the high temperature assumption @xmath96 , is qualitatively correct for a large momentum regime even for large values of @xmath97 or temperatures of the order @xmath35 .
this gives us some confidence in the usefulness of the effective perturbation theory ( braaten - pisarski method ) , which is based on the htl resummation , for realistic values of the coupling constant @xmath98 2.5 at temperatures @xmath99 300 mev that are accessible in relativistic heavy ion collisions .
furthermore , the use of htl effective masses for describing the equation of state of a gluon gas @xcite or strange quark matter @xcite appears to be justified from the point of view that the deviation of the htl masses from the complete one - loop effective masses can be neglected in those models .
kalashnikov and v.v .
klimov , sov .
( 1980 ) 699 and 33 ( 1981 ) 847 ; k. kajantie and j.i .
kapusta , ann .
160 ( 1985 ) 477 ; u. heinz , k. kajantie , and t.toimela , ann .
176 ( 1987 ) 218 . | the complete one - loop self energies ( real and imaginary parts ) for photons , gluons , electrons and quarks at finite temperature are calculated numerically and compared to the results of the hard thermal loop ( htl ) approximation used for the resummation technique of braaten and pisarski . in this way
some light is shed on the validity of the weak coupling limit assumption ( @xmath0 ) or equivalently the high temperature assumption , on which the htl approximation is based .
furthermore , the gauge dependence of the fermion self energy beyond the htl approximation is considered .
finally the dispersion relations following from the real part of the self energies are compared to the htl results . |
the long - standing quest for a solid state realization of the kitaev honeycomb model@xcite has triggered much of the experimental and theoretical interest in 4d and 5d compounds with two- and three - dimensional tri - coordinated lattices , in which the interplay of the strong spin - orbit coupling ( soc ) and electronic correlations leads to the dominance of the strongly anisotropic kitaev - like interactions.@xcite a lot of experimental effort has been focused on iridium oxides belonging to the a@xmath0iro@xmath1 family@xcite and , more recently , to @xmath2rucl@xmath1.@xcite the kitaev honeycomb model belongs to the class of the compass models .
it is intrinsically frustrated due to the bond - depended nature of the interactions . in the quantum case
, this frustration leads to the appearance of the non - trivial quantum spin liquid ( qsl ) phase with fractionalized excitations , dubbed kitaev qsl.@xcite kitaev qsl is not a unique example of non - trivial ground states of the compass models , @xcite however , it is probably the only one which allows an exact analytic solution
. in honeycomb iridates and ruthenates , the magnetic degree of freedom described by an effective magnetic moment @xmath5 , arises in the presence of strong soc from electrons occupying @xmath6-manifold of states of ir@xmath7 and ru@xmath8 ions . in a@xmath0iro@xmath1 compounds ,
edge - shared iro@xmath9 octahedra provide 90@xmath10 paths for the dominant nearest neighbor kitaev coupling between iridium magnetic moments .
a similar situation takes place in the isostructural layered honeycomb material @xmath2rucl@xmath1 and three - dimensional harmonic honeycomb iridates , @xmath11li@xmath0iro@xmath1 and @xmath12li@xmath0iro@xmath1.@xcite it is believed that the sign of the kitaev interaction may be either antiferro- ( af ) or ferromagnetic ( fm ) depending on the compound.@xcite isotropic heisenberg couplings are also present in these compounds due to the octachedra edge sharing geometry and direct overlap of @xmath13 or @xmath14orbitals which , due to their extended nature , often reach beyond nearest neighbors .
further anisotropies , such as the isotropic off - diagonal @xmath3 interactions , can also be present , mainly as a result of crystal field distortions .
@xcite the competition between all these couplings leads to a rich variety of experimentally observed magnetic structures.@xcite here we discuss in detail the models and the mechanisms which lead to the stabilization of magnetic ordering in two compounds : na@xmath0iro@xmath1 and @xmath2rucl@xmath1 .
several experiments have shown that the low - temperature phase of na@xmath0iro@xmath1 has collinear zigzag long - range magnetic order.@xcite in addition , recent diffuse magnetic x - ray scattering data have determined the spin orientation in this zigzag state and showed that it is along the 44.3@xmath15 direction with respect to the @xmath16 axis , which corresponds to approximately half way in between the cubic @xmath17 and @xmath18 axes.@xcite both of these findings are in disagreement with the original kitaev - heisenberg model,@xcite which predicts the zigzag phase only for the antiferromagnetic nearest neighbor kitaev interaction with the magnetic moments along the cubic axes , while the kitaev interaction in na@xmath0iro@xmath1 is ferromagnetic.@xcite this shows that one needs to extend the nearest neighbor model by including some additional interactions in order to explain these experimental observations .
@xmath2rucl@xmath1 also shows collinear antiferromagnetic zigzag ground state.@xcite recent x - ray and neutron scattering diffraction data@xcite indicate that the best fit to the collinear structure is obtained for the antiferromagnetic nearest neighbor kitaev interaction and when the spin direction points 35@xmath15 out of the @xmath19-plane , i.e. along one of the cubic directions .
this suggests that the microscopic origin of the zigzag ground state in @xmath2rucl@xmath1 might be quite different from the one in na@xmath0iro@xmath1 , and that it can be described by the nearest neighbor kitaev - heisenberg model.@xcite in both cases , the available experimental data provides an important check of the validity of any model proposed to describe the magnetic properties of na@xmath0iro@xmath1 and @xmath2rucl@xmath1 , as it should correctly predict not only the type of the magnetic order but also its orientation in space . in this work
we consider two models , the nearest neighbor kitaev - heisenberg model@xcite and its more complicated counterpart , dubbed @xmath4 model,@xcite and study how the preferred directions of the mean field order parameter are selected in these models .
the formal procedure which we will be using here is based on the derivation of the fluctuational part of the free energy by integrating out the leading thermal fluctuations , and by determining which orientations of the order parameter correspond to the free energy minimum .
this approach is based on the hubbard - stratonovich transformation and was outlined in refs.@xcite to which we refer the reader for technical details . in both models , the thermal fluctuations select the cubic axes as the preferred directions for spins , which describes the experimental situation in @xmath2rucl@xmath1 but not in na@xmath0iro@xmath1 .
we have also checked that in both models the quantum fluctuations ( taken into account either using the quantum version of hubbard - stratonovich approach or within the semiclassical spin - wave approach ) lift the accidental degeneracy of the classical solution and also select the cubic axes as the preferred directions for spins .
we did not present these calculations here as they bring no new results compared with more simple analysis of thermal fluctuations . the important point which we stress in our paper
is that the selection of correct `` diagonal '' direction of the spins observed in na@xmath0iro@xmath1 might happen already on the mean - field level by inclusion of small off - diagonal positive interaction @xmath3 as soon as it is larger than the energy gain of order @xmath20 due to the quantum fluctuations .
this paper is organized as follows . in sec .
ii we study the order by disorder mechanism of the selection of the direction of the order parameter in the nearest neighbor kitaev - heisenberg model on the honeycomb lattice . in sec .
iii we extend our consideration to the @xmath4 model . in sec .
iv , we discuss the role of the off - diagonal @xmath3-term and study the selection of the direction of the magnetic order in na@xmath0iro@xmath1 and in @xmath2rucl@xmath1 .
we summarize our conclusions in sec .
v. appendix a discusses in detail the degeneracy of the classical manifold of the kitaev - heisenberg model .
appendices b and c contain some technical details . and
@xmath21 are two primitive translations .
the bond vectors are @xmath22 , @xmath23 and @xmath24 .
, title="fig : " ] and @xmath21 are two primitive translations .
the bond vectors are @xmath22 , @xmath23 and @xmath24 .
, title="fig : " ] and @xmath21 are two primitive translations .
the bond vectors are @xmath22 , @xmath23 and @xmath24 .
, title="fig : " ] and @xmath21 are two primitive translations .
the bond vectors are @xmath22 , @xmath23 and @xmath24 .
, title="fig : " ] interaction in the four - sublattice transformation : @xmath3 picks up a negative sign on the solid bonds but keeps the sign from unrotated reference frame on the dashed bonds . ]
the kitaev - heisenberg model on the honeycomb lattice reads @xcite @xmath25 where @xmath26 is the interaction between @xmath27-component of the pseudospin @xmath28 , on sublattices @xmath29 .
hereafter , we call these pseudospins simply spins .
@xmath30 and @xmath31 correspond to the heisenberg and kitaev interactions , which in the extended model can be both af and fm .
@xmath32 denote the spin components in the global reference frame .
the classical phase diagram of the model ( 1 ) contains four magnetic phases:@xcite the ferromagnetic phase ( fig .
1 ( a ) ) , the nel antiferromagnet ( fig .
1(b ) ) , the stripy antiferromagnet ( fig . 1 ( c ) ) and the zigzag antiferromagnet ( fig . 1 ( d ) ) .
the latter two magnetic states have a four sublattice structure .
all these phases have macroscopic classical degeneracy .
while the classical degeneracy of the simple fm state and of the af nel state comes straightforwardly from the infinite number of degenerate collinear states , the macroscopic degeneracy of the af stripy and zigzag phases is more complex , and the degenerate ground state manifold consists of six collinear states and a set of non - collinear multi-@xmath33 states . in appendix
a we discuss this question in detail and show that using the four - sublattice klein transformation for the stripy and the zigzag af states,@xcite the nature of the classical degeneracy of all four magnetically ordered states can be understood in a similar way .
importantly , in all cases , the classical degeneracy is accidental and is removed by the order by disorder mechanism which selects a set of collinear states , each with a particular direction of the order parameter .
following chaloupka _
et al_,@xcite we introduce four auxiliary sublattices a , b , c and d ( see fig.2 ) , fix the direction of the spins on the sublattice a and rotate the spins on the subllatices b , c , and d such that the component of spin corresponding to the bond direction ( @xmath17 for b , @xmath18 for c and @xmath34 for d ) stays the same but two other spin components change sign .
this results in the transformed hamiltonian with the same form as ( [ kh ] ) but with transformed couplings .
here we consider the kitaev - heisenberg model in the full parameter space .
for the parameters of the model for which either stripy or zigzag are the ground states , we perform four - sublattice transformation and treat the model ( [ kh ] ) in the rotated basis , in which the stripy order maps to the fm and the zigzag order maps to the simple two - sublattice afm nel state . next , using a hubbard - stratonovich transformation of the partition function,@xcite we discuss how the preferred directions of the order parameter in all these phases are selected by thermal below the ordering temperature .
the partition function of the system of classical spins is given by the integral over the boltzmann weights of the configurations @xmath35[d\mathbf{s}_{1,j}\right]\delta(\mathbf{s}_{0,i}^{2}-1)\delta(\mathbf{s}_{1,j}^{2}-1)\nonumber\\ & & \exp\left[-\beta\sum_{\langle ij \rangle_\alpha}\sum_{\gamma}j^{\alpha\gamma}s_{0,i}^{\gamma } s_{1,j}^{\gamma}\right],\end{aligned}\ ] ] where @xmath36 and @xmath37 are classical spins on sublattices @xmath38 and @xmath39 , and @xmath40 is the inverse temperature . similarly in the case of a quantum system the partition function is a trace of the boltzman weights over the spin operators , @xmath41 $ ] .
+ it is more convenient to perform the hubbard - stratonovich transformation by representing the hamiltonian matrix in the basis of the eigenfunctions of the exchange matrix , which can be easily obtained in the momentum space . to this end , we first introduce a six - component vector @xmath42 , with the components given by the fourier transforms of the @xmath43 components of the spins on @xmath44 and @xmath45sublattices , correspondingly .
this allows us to write the hamiltonian matrix in the momentum space as @xmath46where the @xmath47 exchange matrix @xmath48 is defined as @xmath49 with matrix elements given by @xmath50 e^{\imath{\bm q}\cdot ( { \bm d } _ { \alpha } -{\bm d } _ { z } ) } = j_{{\bm q}}+k_{\bm q}^{\gamma}.\end{aligned}\ ] ] here we drop the overall phase factor @xmath51 and denote @xmath52 , @xmath53 , where @xmath54 and @xmath55 are the lattice vectors .
the matrix @xmath48 is then diagonalized by a unitary transformation , @xmath56 , leading to the following form of the hamiltonian @xmath57where the normal amplitudes of spin - like variables are defined as @xmath58 note that , depending on the form of the interaction matrix , this transformation in general will mix the spin operators on different sites of the unit cell as well as different components of the spin .
however , in the case of the kitaev - heisenberg model , while the two sublattices of the honeycomb lattice are mixed , the @xmath17 , @xmath18 , and @xmath34 components stay separate .
the partition function ( [ z - clas ] ) then looks like : @xmath59[d\mathbf{s}_{1,j+{\bm d}_z}\right]\delta(\mathbf{s}_{0,j}^{2}-1)\delta(\mathbf{s}_{1,j+{\bm d}_z}^{2}-1)\nonumber\\ & & \exp\left[-\beta \sum_{\bm q,\nu } \kappa_{{\bm q},\nu } { \tilde s}^*_{{\bm q},\nu } { \tilde s}_{{\bm q},\nu}\right ] .\label{z - norm}\end{aligned}\ ] ] following the steps outlined in refs.@xcite , we can separate the mean - field and the fluctuational contributions to the partition function , @xmath60 . in the gaussian approximation ,
the fluctuation part of the partition function , @xmath61\exp\left [ -\beta{\mathcal{s}}_{\mathrm{fluct}}\right],\end{aligned}\ ] ] where @xmath62 can be obtained by integration over the fluctuation amplitudes @xmath63 . the explicit expression for the matrix elements of the fluctuation matrix @xmath64 computed for an orientation of the mean - field order parameter along arbitrary direction @xmath65
are given in appendix b. now , the fluctuation contribution to the free energy can be written as @xmath66 while the mean - field part of the free energy has the full rotational symmetry , its fluctuational part , @xmath67 , is sensitive to the direction of the mean - field order parameter .
thus , by finding the minima of the fluctuational part of the free energy , we can pin the spontaneous magnetization along some preferred direction of the lattice .
fig.3 ( a ) shows the angular dependence of fluctuational free energy @xmath68 computed for representative parameters @xmath69 mev and @xmath70 mev , at which the ground state order is the af zigzag .
the magnitude of @xmath68 is presented as a color - coded plot on the unit sphere , where the minima and maxima of the free energy are shown by the deep blue and red colors , correspondingly .
we see that the minima of @xmath68 are achieved when the magnetization is directed along one of the cubic axes .
this finding shows that the contribution of the fluctuations to the free energy removes the degeneracy of the ground state found on the mean field level .
the states which are selected by the thermal fluctuations are the collinear states with the order parameter pointing along one of the cubic axes , thus confirming previous results of the monte carlo simulations@xcite and spin wave analysis by chaloupka et al.@xcite mev and @xmath70 mev and ( b ) @xmath4 model computed with @xmath71 mev , @xmath72 mev , @xmath73 mev , @xmath74 mev , and @xmath75 mev .
, title="fig : " ] mev and @xmath70 mev and ( b ) @xmath4 model computed with @xmath71 mev , @xmath72 mev , @xmath73 mev , @xmath74 mev , and @xmath75 mev .
, title="fig : " ] we discuss the relevance of our findings for the nearest neighbor kitaev - heisenberg model for @xmath2rucl@xmath1 in sec .
iv . however in the next section , we will first consider the selection of the direction of the order parameter in the extensions of the kitaev - heisenberg model relevant for na@xmath0iro@xmath1 .
despite extensive efforts , no consensus concerning the minimal model for na@xmath0iro@xmath1 has been reached yet .
the most natural extension of the kitaev - heisenberg model with ferromagnetic kitaev interaction which captures the zigzag magnetic order can be obtained by inclusion of farther neighbor couplings . in na@xmath0iro@xmath1
, these couplings might not be negligible due to the extended nature of the @xmath76-orbitals of the ir ions . in the early works suggesting this possible extension,@xcite second and third neighbor couplings were taken into account phenomenologically and
only the isotropic part of these interactions was included .
the importance of additional nearest neighbor @xmath77-symmetric anisotropic terms ( @xmath3-terms)@xcite or of the spatial anisotropy of the nearest neighbor kitaev interactions,@xcite were also discussed in the literature as a possible source for the stabilization of the zigzag phase .
here we consider the @xmath4 model,@xcite which still has the same symmetry as the original kitaev - heisenberg model but contains kitaev interactions between both nearest and second nearest neighbors .
the model reads @xmath78 where @xmath79 , @xmath80 , @xmath81 , @xmath82 , and @xmath83 ; @xmath84 , @xmath85 and @xmath86 denote nearest neighbor , second nearest neighbor and third nearest neighbor , respectively . @xmath32 and @xmath87 denote the three types of nearest neighbor and second nearest neighbor bonds of the honeycomb lattice , respectively .
it is important to note that the second neighbor kitaev interactions do not change the space group symmetries of the original kitaev - heisenberg model . for realistic sets of the parameters describing na@xmath0iro@xmath1 , the second neighbor kitaev interaction , @xmath88 , computed from the microscopic approach based on the ab - initio calculation by foevtsova _
et al_,@xcite appeared to be the largest interaction after the nearest neighbor kitaev interaction , @xmath89 , and turn out to be antiferromagnetic .
the mechanism behind the large magnitude of @xmath88 in na@xmath0iro@xmath1 is physically very clear : it originates from the large diffusive na ions that reside in the middle of the exchange pathways , and the constructive interference of a large number of pathways .
moreover , the @xmath89-@xmath88 model , that only includes kitaev interactions,@xcite already stabilizes the zigzag phase for the proper signs of @xmath89 and @xmath88 .
however , as we have discussed in ref.@xcite , the @xmath89-@xmath88 model is still not sufficient to comply with all available experimental data .
the classical degeneracy of the zigzag state obtained within the @xmath4 model with fm @xmath89 is different from the one of the zigzag state realized in the extended kitaev - heisenberg model with afm @xmath89 interaction . to see what difference the sign of @xmath89 makes ,
let us consider the zigzag pattern in fig.[fig : orders ] ( d ) . with afm @xmath89 , the pattern , that minimizes the classical energy in the zigzag state with ferromagnegnetic @xmath18 and @xmath34 bonds , has the spins pointing along the @xmath90axis to take advantage of the kitaev interaction on the afm @xmath17-bonds .
on the other hand the same pattern with fm @xmath89 takes advantage of the kitaev interaction on the fm @xmath91 and @xmath92 bonds by putting spins in the @xmath93plane .
thus the degenerate ground state manifold for a given zigzag pattern with fm @xmath89 is one of @xmath94 , @xmath93 , or @xmath95 planes .
furthermore , when the klein duality 4-sublattice transformation@xcite is applied to the @xmath4 zigzags , these states do not turn into nel afm state , and instead turn into non - collinear states , that are more difficult to work with than the original zigzag states . working with
the zigzag states directly increases the magnetic unit cell to 4 sites , labeled in fig [ fig : orders](d ) .
the hamiltonian matrix in the momentum space can be again written in the form of eq .
( [ hammoment ] ) , however this time due to the larger unit cell the exchange matrix @xmath48 is @xmath96 , instead of @xmath47 .
its matrix elements are given in appendix c. the fluctuations matrix @xmath97 is calculated as before according to equation ( [ amatrix ] ) , with the constraint matrix @xmath98 of equation ( [ constraint ] ) now containing 4 identical blocks instead of 2 .
the fluctuation matrix again contains the information on the direction of the spins and transmits this information to the free energy corrections that we plot in fig .
[ fig : nogamma](b ) .
since the spins are confined to a plane for a given zigzag state we have only the angle of the direction of spins in that plane .
the color of the band at a given angle then gives the size of the fluctuational correction to the free energy , with violet being lowest and red highest energy states .
we see that again the kitaev anisotropies prefer to align the magnetization along the cubic axes .
note , however , that unlike the extended kh model , where there were 6 equivalent states , here there are 4 directions for each of the three zigzag patterns , giving a total of 12 states .
model obtained with the luttinger - tisza method is shown on the first brillouin zone .
we use @xmath99 mev , @xmath72 mev , @xmath100 mev , @xmath101 mev , @xmath75 mev , and ( a ) @xmath102 mev , ( b ) @xmath103 mev , ( c ) @xmath104 mev , and ( d ) @xmath105 mev .
, title="fig : " ] model obtained with the luttinger - tisza method is shown on the first brillouin zone .
we use @xmath99 mev , @xmath72 mev , @xmath100 mev , @xmath101 mev , @xmath75 mev , and ( a ) @xmath102 mev , ( b ) @xmath103 mev , ( c ) @xmath104 mev , and ( d ) @xmath105 mev .
, title="fig : " ] model obtained with the luttinger - tisza method is shown on the first brillouin zone .
we use @xmath99 mev , @xmath72 mev , @xmath100 mev , @xmath101 mev , @xmath75 mev , and ( a ) @xmath102 mev , ( b ) @xmath103 mev , ( c ) @xmath104 mev , and ( d ) @xmath105 mev .
, title="fig : " ] model obtained with the luttinger - tisza method is shown on the first brillouin zone .
we use @xmath99 mev , @xmath72 mev , @xmath100 mev , @xmath101 mev , @xmath75 mev , and ( a ) @xmath102 mev , ( b ) @xmath103 mev , ( c ) @xmath104 mev , and ( d ) @xmath105 mev .
, title="fig : " ]
the discussion above has clearly shown , that neither the original kitaev model nor the extended @xmath4 model can correctly explain the experimental data in na@xmath0iro@xmath1 .
since the easy axes directions are determined solely by the anisotropy terms , only the inclusion of other types of the anisotropies can improve the situation . here
we consider the off - diagonal symmetric @xmath3-terms .
the role of these terms in the nearest - neighbor kitaev model has been studied in refs.@xcite .
these studies have shown that the small @xmath3-terms do not immediately destabilize the zigzag phase , but lead to a deviation of the magnetic moments from the cubic axes .
the origin of @xmath3-terms can be easily seen from the most general form of the bilinear exchange coupling matrix , which on the bond @xmath106 has the form given by @xmath107 while the kitaev term comes from the anisotropy of the diagonal matrix elements of @xmath108 , e.g. @xmath109 , the symmetric and antisymmetric combinations of off - diagonal elements represent other types of possible bond - anisotropies . in the absence of the trigonal distortion , the inversion symmetry prohibits the existence of antisymmetric interactions but some of the symmetric combinations are allowed , i.e. on a given @xmath27-bond , the interaction between the other two spin components , @xmath110 , where @xmath111 , is non - zero .
our previous results@xcite suggest that in na@xmath0iro@xmath1 the magnitude of the strength of @xmath3 on the nearest neighbor bonds is about 2 - 3 mev and vanishes for the second neighbors . here
we consider the @xmath112 model with the previous choice of heisenberg and kitaev interactions and treat @xmath3 as a free parameter .
a straightforward classical minimization in momentum space using luttinger - tisza approach@xcite shows that up to very large values of @xmath113 mev the minima of the classical energy are located at the @xmath114 points corresponding to the zigzag states .
this is clearly seen in fig .
4(a ) where we plot the lowest eigenvalues obtained for @xmath115 mev . at larger values of @xmath3 , the minima shift along the lines connecting @xmath114 points and the center of the brillouin zone ( see fig . 4 ( b ) for @xmath116 mev ) , indicating the transition to incommensurate order .
the incommensurability of the luttinger - tisza solution increases further with larger values of @xmath3 , which is shown in figs .
4 ( c ) and ( d ) .
the exact value of @xmath3 at which the transition occurs is difficult to determine due to the transition being so smooth , note , however , that the transition occurs at values of @xmath3 far beyond those predicted from our microscopic calculations at ambient pressure.@xcite after we have demonstrated that adding small @xmath3 interactions to the @xmath4 model does not destabilize the zigzag order , let us now show that in the presence of @xmath3 the mean - field degeneracy is already lifted and the preferred directions of the order parameter are selected .
this is clearly seen in fig . 5 ( a ) and ( b ) , where the mean field energy of the zigzag order is computed for @xmath115 mev and @xmath117 mev , respectively .
by inspection , we can see that non - zero @xmath3 selects the face diagonals as easy axes for magnetic ordering , and the sign of @xmath3 determines which of the two face diagonals are preferred . for concreteness ,
let us consider the zigzag with afm @xmath92bonds .
as we discussed above the case for @xmath118 , the easy @xmath119-plane is selected at the mean - field level of the @xmath4 model .
then , the inclusion of positive @xmath3 interaction on @xmath17 and @xmath18 bonds gives zero contribution to the energy since on these bonds it involves the spin component perpendicular to the easy plane , but it gives maximal lowering of the energy on the @xmath34-bonds if the spins point along @xmath120 $ ] and @xmath121 $ ] , @xmath122 $ ] and @xmath122 $ ] directions correspondingly for positive and negative values of @xmath3 .
the estimate for the smallest @xmath3 , at which the selection of face diagonals takes place , can be done by comparing the mean - field energy gain due to @xmath3 with the energy gain due to fluctuations at @xmath118 , which at @xmath123 is equal to the zero point energy and is a function of @xmath89 and @xmath88 . at finite temperature ,
the contribution to the energy from the gaussian fluctuations at each @xmath124 can be computed by our method , and this energy will give the lower bound for the magnitude of @xmath3 needed to change the orientation of magnetic order from the cubic to the face diagonal . model with the contribution of ( a ) @xmath115 mev and ( b ) @xmath117 mev .
, title="fig : " ] model with the contribution of ( a ) @xmath115 mev and ( b ) @xmath117 mev .
, title="fig : " ] the microscopic calculations for @xmath2rucl@xmath1 emphasized the importance of the off - diagonal nearest neighbor @xmath3 interactions.@xcite the effect of adding @xmath3 interaction to the nearest neighbor kitaev - heisenberg model is easiest to understand in the rotated reference frame of the four - sublattice klein transformation.@xcite the kitaev and heisenberg interactions do not change their form and only change the value of the coupling constants under this transformation .
on the other hand , @xmath3-interaction picks up a bond dependent sign as shown in fig .
2 . in fact , @xmath3 changes the sign on half of the bonds , i.e. there are just as many negative bonds as there are positive bonds for each kitaev type of bonds .
since the klein transformed version of the zigzag state is the afm nel state , all the bonds are afm and involve the same pair of spins .
thus the contribution of the @xmath3 interaction to the mean - field energy cancels out , and the set of states remains degenerate .
this means that as long as we remain in the small window where @xmath3 does not destabilize the zigzag order found by rau et al.,@xcite we can perform our order - by - disorder approach to see what state is chosen .
figs.6 ( a)-(c ) show the fluctuation free energy computed for the @xmath125 model for @xmath126 mev and @xmath127 mev , suggested by banerjee _
et al._,@xcite and @xmath128 mev , 0.8 mev and 0.9 mev , respectively . in fig
. 6 ( a ) , @xmath128 mev , the minima of the fluctuational free energy are still along cubic directions . for larger @xmath3-interaction ,
the system prefers the states with at least two nonzero spin components and , therefore , the transition towards [ 111 ] preferred directions of the order parameter takes place .
this is shown in fig . 6 ( b ) and ( c ) ,
in which the fluctuational energy is plotted for @xmath129 mev and 0.9 mev .
while in fig .
6 ( b ) only very shallow minima are seen along [ 111 ] directions , in fig .
6 ( c ) both the pronounced minima along the cubic body diagonals and maxima along the cubic axes are very clearly seen .
remember that the computation is done in the rotated reference frame .
therefore , only the states with the orientation of the order parameter along the cubic axes will give the collinear states in the unrotated reference frame .
the states with order parameter pointing along [ 111 ] directions in the rotated reference frame correspond to non - collinear states in the unrotated reference frame . since recent experiments by cao et
al.@xcite have established that spins point along a cubic axis , by calculating the fluctuational corrections as a function of @xmath3 , we can find an upper bound on its value , such that the kitaev fluctuations dominate and keep the cubic axes as the preferred directions . from our calculations
it follows that for @xmath126 mev and @xmath127 mev the upper bound for @xmath3 is about 0.8 mev .
finally , for this set of parameters the transition to the 120@xmath10- afm order occurs around @xmath130 mev .
note that this estimate is far smaller than the @xmath3 values resulting from _ ab initio _ calculations.@xcite .
in this paper we explored how the direction of the magnetic moments in the zigzag ground state order is chosen in na@xmath0iro@xmath1 and @xmath2rucl@xmath1 . in both compounds
, the kitaev interaction plays an important role . for the case of fm nearest neighbor kitaev interaction , like in na@xmath0iro@xmath1 , farther neighbor interactions are essential for stabilizing the zigzag ground state . for the afm nearest neighbor kitaev interaction , which was widely suggested to be
the dominant interaction in @xmath2rucl@xmath1,@xcite the zigzag order can be stabilized already within the nearest neighbor model .
we proposed that the @xmath112 model can explain all the experimental finding in na@xmath0iro@xmath1 . in this model
the selection of the experimentally observed face diagonal direction of the order parameter happens already on the mean - field level due to the small bond - dependent anisotropic term @xmath3 . in @xmath2rucl@xmath1 ,
if the the nearest neighbor kitaev interaction is afm , the original kitaev - heisenberg model@xcite is sufficient to explain both the collinear zigzag ground state and the cubic directions of the order parameter .
we studied the effect of the @xmath3-term and showed that while on the mean - field level it does nt affect the ground state degeneracy , it favors non - collinear 3-*q * states , instead of the experimentally observed zigzag state with spins along cubic axes , once the gaussian fluctuations are included .
thus , it appears to be an upper bound for @xmath3-term , which can be estimated for a given set of nearest neighbor parameters .
after the completion of our paper , we became aware of an independent study by winter _
et al_. @xcite of the magnetic interactions in the kitaev materials na@xmath0iro@xmath1 and @xmath2rucl@xmath1 . in this work , the authors treated all interactions up to third neighbours on equal footing by combining exact diagonalization and ab - initio techniques .
one of the main findings of this work is that the third neighbor heisenberg interaction is important in all kitaev materials .
let us briefly compare the results of ref .
@xcite with our findings .
the conclusions of the authors of ref .
@xcite about the ordering in na@xmath0iro@xmath1 are in agreement with our findings , despite the fact that their estimates for @xmath88 suggest significantly smaller values than the ones that we obtained by including only the dominant superexchange processes between the second neighbors .
the agreement holds because the second neighbor kitaev interaction @xmath88 and the third neighbor interaction @xmath131 favor the same type of afm zigzag ground state . for @xmath2rucl@xmath1 , the authors of ref.@xcite suggest ( _ i _ ) that there may be possible variations of in - plane interactions due to lattice distortions , and ( _ ii _ ) that the nearest neighbor kitaev interaction may be fm and
the third neighbor coupling @xmath131 may be large and afm .
the fm sign of the nearest neighbor kitaev interaction was also suggested by yadav _
et al _ in ref.@xcite .
if this is indeed the case , the physics of @xmath2rucl@xmath1 is similar to that of na@xmath0iro@xmath1 .
this , however , still needs to be verified by a detailed comparison with the experimental data . _
_ we acknowledge insightful discussions with c. batista , g. jackeli , m.garst , g. khaliullin and i. rousochatzakis .
n.p . and y.s .
acknowledge the support from nsf grant dmr-1511768 .
n.p . acknowledges the hospitality of kitp and partial support by the national science foundation under grant no .
nsf phy11 - 25915 .
in this appendix we provide detailed discussion of the classical degeneracy of the extended kitaev - heisenberg model at parameters for which either the stripy or the zigzag af phases are realized as the ground state and the manifold of classically degenerate states is rather complex . to be specific ,
let us first consider the stripy phase .
it contains six inequivalent collinear stripy states with fm bonds along either kitaev @xmath90 , @xmath91 or @xmath92bonds .
it also contains infinite number of non - collinear ( coplanar and non - coplanar ) states .
the spin order in the @xmath90 , @xmath91 or @xmath92 stripy states can be described either with a help of four magnetic sublattices or by a simple spiral characterized by a single-@xmath33 wave vector : @xmath132 , @xmath133 and @xmath134 .
one of the stripy states with fm @xmath34-bonds is shown in fig.1 ( c ) . in each of these stripy states the spins
are aligned along one of the cubic directions which is locked to the spatial orientation of a stripy pattern by the kitaev interaction , i.e. the direction of the order parameter is defined by the wave vector @xmath135 or @xmath136 determining the breaking of the translation symmetry . the structure of the manifold of the non - collinear states , which looks rather complex in the original model , can be easily understood with the help of the four - sublattice transformation ( see fig.2 ) based on the klein duality.@xcite in the new rotated basis , the stripy phase is mapped to the fm order with a unique ordering vector @xmath137 .
classically , all states with arbitrary direction of the fm order have the same energy .
fm states with order parameter along the cubic axes give the six stripy phases in the unrotated spin basis discussed above .
arbitrary directions of the fm order parameter lead to a set of non - coplanar states in which each component of spin , @xmath138 , @xmath139 , and @xmath140 , transforms with its own @xmath141 , @xmath142 and @xmath136 wavevector , which coincide with the @xmath143 vectors describing the spatial orientation of the stripes in the respective collinear states . using these three ordering vectors
, we can write the non - coplanar phase of the unrotated spins as @xmath144 where @xmath145 and @xmath146 are the polar and azimuthal angles of the fm order parameter . @xmath147 denote the spins on the sublattice @xmath38 and the spins on the sublattice @xmath39 are obtained from @xmath147 by a constant phase shift coming from the spin rotation on that bond as prescribed by the four sublattice transformation . as in fig.1 ( c ) , the sublattices 0 and 1 are connected by the @xmath34 bond , the order of the spins on the subllatice 1 is given by @xmath148 in the zigzag phase , the structure of the classical states manifold is very similar to the stripy phase .
the four - sublattice transformation maps the zigzag phase onto the nel af phase .
the generic state is again described by the three-@xmath143 spiral state .
the only difference is that the spins on sublattice 1 have an overall phase factor of @xmath149 , @xmath150 .
the matrix elements @xmath97 can be written as @xmath151 where a repeated index implies a summation over .
the first term in ( [ amatrix ] ) is the contribution from the interaction term and the second term is from the constraint term.@xcite for convenience , the constraint matrix @xmath152 can be first written in the original basis , in which the interaction term is not diagonal , and then transformed to the eigenbasis of the hamiltonian with a help of the unitary transformation @xmath153 . in the original basis
the constraint matrix @xmath152 consists of two blocks , one for each sublattice .
the a - sublattice block has elements @xmath154 with @xmath155 and the b - sublattice block has the elements with @xmath156 .
the two blocks are identical , so @xmath152 takes the following form : @xmath157 with matrix elements given by @xmath158{l}c_{{\bm q},11}= -\frac{2}{3}\bigl [ \beta_c(1-s_{\theta}^{2}c_{\phi}^{2})+3\beta r s_{\theta}^{2}c_{\phi}^{2}\bigr ] , \\
c_{{\bm q},22}= -\frac{2}{3}\bigl [ \beta_c(1-s_{\theta}^{2}s_{\phi}^{2})+3\beta r s_{\theta}^{2}s_{\phi}^{2}\bigr ] , \\
c_{{\bm q},33}= -\frac{2}{3}\bigl [ \beta_c s^2_{\theta}+3\beta r c^2_{\theta}\bigr ] , \\ c_{{\bm q},12}=c_{{\bm q},21}=-\frac{2}{3 } ( 3\beta r-\beta_c ) s^2_{\theta } c_{\phi}s_{\phi } , \\ c_{{\bm q},13}=c_{{\bm q},31}=-\frac{2}{3 } ( 3\beta r-\beta_c ) s_{\theta}s_{\theta } c_{\phi } , \\ c_{{\bm
q},23}=c_{{\bm q},32}=-\frac{2}{3 } ( 3\beta r-\beta_c ) s_{\theta}s_{\phi } c_{\phi } , \end{array } \label{cmatrix}\ ] ] where , to shorten notations , we denote @xmath159 and @xmath160 .
for shortness we define @xmath162 , @xmath163 , and @xmath164 .
the diagonal matrix elements for @xmath165 and 10 are equal to @xmath166 , all other diagonal elements are equal to @xmath167 .
the non - zero off - diagonal elements @xmath161 for @xmath168 are @xmath169 y. singh and p. gegenwart , phys .
b * 82 * , 064412 ( 2010 )
. y. singh , s. manni , j. reuther , t. berlijn , r. thomale , w. ku , s. trebst , and p. gegenwart , phys .
lett . * 108 * , 127203 ( 2012 ) .
x. liu , t. berlijn , w .- g .
yin , w. ku , a. tsvelik , young - june kim , h. gretarsson , y. singh , p. gegenwart , and j. p. hill , phys .
b * 83 * , 220403 ( 2011 )
. f. ye , s. chi , h. cao , b. c. chakoumakos , j. a. fernandez - baca , r. custelcean , t. f. qi , o. b. korneta , and g. cao , phys .
b * 85 * , 180403 ( 2012 ) . s. k. choi , r. coldea , a. n. kolmogorov , t. lancaster , i. i. mazin , s. j. blundell , p. g. radaelli , yogesh singh , p. gegenwart , k. r. choi , s .- w .
cheong , p. j. baker , c. stock , and j. taylor , phys .
rev . lett . * 108 * , 127204 ( 2012 ) .
h. gretarsson , j. p. clancy , yogesh singh , p. gegenwart , j. p. hill , jungho kim , m. h. upton , a. h. said , d. casa , t. gog , and young - june kim , phys .
b * 87 * , 220407(r ) s. h. chun , j .- w .
kim , j. kim , h. zheng , c. c. stoumpos , c. d. malliakas , j. f. mitchell , kavita mehlawat , yogesh singh , y. choi , t. gog , a. al - zein , m. moretti sala , m. krisch , j. chaloupka , g. jackeli , g. khaliullin , b. j. kim , nature physics * 11 * , 462 ( 2015 ) .
a. biffin , r. d. johnson , s. choi , f. freund , s. manni , a. bombardi , p. manuel , p. gegenwart , and r. coldea , phys .
b * 90 * , 205116 ( 2014 ) .
a. biffin , r.d .
johnson , i. kimchi , r. morris , a. bombardi , j.g .
analytis , a. vishwanath , and r. coldea , phys .
lett . * 113 * , 197201 ( 2014 ) .
t. takayama , a. kato , r. dinnebier , j. nuss , h. kono , l.s.i .
veiga , g. fabbris , d. haskel , and h. takagi , phys .
lett . * 114 * , 077202 ( 2015 ) .
i. pollini , phys .
b * 53 * , 12769 ( 1996 ) .
k. w. plumb , j. p. clancy , l. j. sandilands , v. v. shankar , y. f. hu , k. s. burch , h .- y .
kee , and y .- j .
kim , phys .
b * 90 * , 041112 ( 2014 ) .
j. a. sears , m. songvilay , k. w. plumb , j. p. clancy , y. qiu , y. zhao , d. parshall , and young - june kim , phys .
b * 91 * , 144420 ( 2015 ) .
m. majumder , m. schmidt , h. rosner , a. a. tsirlin , h. yasuoka , and m. baenitz , phys .
b * 91 * , 180401 ( 2015 ) .
r. d. johnson , s. c. williams , a. a. haghighirad , j. singleton , v. zapf , p. manuel , i. i. mazin , y. li , h. o. jeschke , r. valenti , and r. coldea phys .
b * 92 * , 235119 ( 2015 ) . a. banerjee , c. bridges , j .- q .
yan , a. a. aczel , l. li , m. b. stone , g. e. granroth , m. d. lumsden , y. yiu , j. knolle et al . , nature materials * 15 * , 733 - 740 , ( 2016 ) .
cao , a. banerjee , j .- q .
yan , c.a .
bridges , m.d .
lumsden , d.g .
mandrus , d.a .
tennant , b.c .
chakoumakos , s.e .
nagler , arxiv:1602.08112 .
j. chaloupka , g. jackeli , g. khaliullin , phys .
lett . * 110 * , 097204 ( 2013 ) .
v. m. katukuri , s. nishimoto , v. yushankhai , a. stoyanova , h. kandpal , s. choi , r. coldea , i. rousochatzakis , l. hozoi , j. van den brink , new j. phys .
* 16 * , 013056 ( 2014 ) .
y. sizyuk , c. price , p. wlfle , and n. b. perkins , phys .
b * 90 * , 155126 ( 2014 ) .
j. reuther , r. thomale and s. rachel , phys .
b * 90 * , 100405 ( r ) ( 2014 ) .
i. rousochatzakis , j. reuther , r. thomale , s. rachel , and n. b. perkins , phys .
x * 5 * , 5 041035 ( 2015 ) .
k. foyevtsova , h. o. jeschke , i. i. mazin , d. i. khomskii , and r. valenti phys .
b * 88 * , 035107 ( 2013 ) . in ref .
@xcite we obtained large @xmath88 interaction by considering only dominant super - exchange processes between second neighbors .
the authors of recent study@xcite claim that the second neighbor kitaev interaction might be suppressed due to the interference of the various second and third order hopping processes , which we did not include in our derivation .
however , as it is discussed in the text , the combined effect of @xmath88 and @xmath131 interactions leads to the same physics . | the magnetic orders in na@xmath0iro@xmath1 and @xmath2rucl@xmath1 , honeycomb systems with strong spin - orbit coupling and correlations , have been recently described by models with the dominant kitaev interactions . in this work
we discuss how the orientation of the magnetic order parameter is selected in this class of models .
we show that while the order - by - disorder mechanism in the models with solely kitaev anisotropies always select cubic axes as easy axes for magnetic ordering , the additional effect of other small bond - dependent anisotropies , such as , e.g. , @xmath3-terms , lead to a deviation of the order parameter from the cubic directions .
we show that both the zigzag ground state and the face - diagonal orientation of the magnetic moments in na@xmath0iro@xmath1 can be obtained within the @xmath4 model in the presence of perturbatively small @xmath3-terms .
we also show that the zigzag phase found in the nearest neighbor kitaev - heisenberg model , relevant for @xmath2rucl@xmath1 , has some stability against the @xmath3-term . |
recent years have witnessed a growing technological , experimental , and theoretical interest in scaling standard machinery down to micro- and nano - scales as needed for the development of `` lab on a chip '' devices . for applications in ,
e.g. , drug - delivery systems or micromechanics one of the most challenging problems at this stage is to develop ways to enable small - scale objects to perform autonomous , controlled motion @xcite .
although the research in this area is still in its early stages , several such proposals have already been tested experimentally ( see , e.g. , refs . ) .
a review of the recent progress in this field can be found in ref . .
whitesides and co - workers proposed a design of self - propelling devices based on an asymmetric decoration of the surface of small objects by catalytic , active sites promoting a chemical reaction in the surrounding liquid medium @xcite .
this asymmetric distribution can provide motility through a variety of mechanisms , such as surface tension gradients and/or cyclic adsorption and desorption @xcite .
the use of an asymmetric surface distribution of a catalyst has been further proposed for an autonomous diffusiophoretic motion emerging as a result of self - created density gradients @xcite .
an experimental realization of such systems , using platinum coated polystyrene spheres , has been recently reported @xcite .
the issue of designing optimal distributions of a catalyst for spherical and rod - like particles has also been approached @xcite . for most of the applications in biological systems or in lab on a chip-type devices
one has to deal with a complicated internal structure of the systems , such as networks of narrow channels or pores and various impenetrable impurities . in some situations
, one may even encounter a quasi two- or one - dimensional behavior , like in the cases of bacteria motion on planar nutrient substrates @xcite and of the motion of small particles within a biological membrane @xcite or in a monolayer adsorbed on a three - dimensional ( _ 3d _ ) liquid subphase @xcite .
thus spatial confinement is a relevant feature so that the assumption of the presence of an unconfined _ 3d _ bulk reactive solvent @xcite may break down .
intuitively , one expects that spatial confinement does influence the resulting motion of self - propelling objects like the ones discussed above .
but _ a priori _ it is not clear if such effects are significant . starting from the model used in ref .
, here we study the effects of spatial confinement on the phoretic motion of a particle that generates number density gradients of the product molecules emerging from the chemical reactions in the depleting newtonian liquid solvent .
we shall focus on the simple case in which the particle and the _
3d _ solution of solvent and product molecules are bounded by a spherical shell because ( i ) this simple geometry allows for an exact solution and ( ii ) it has been experimentally shown for a variety of phoretic systems that confinement effects are dominated by the smallest confining length scale @xcite .
thus this geometry has a paradigmatic character .
similarly to the earlier studies @xcite , our work is based on adopting the standard theory of phoresis for the present case , in which the gradients are self - generated rather than being produced and maintained by external sources . in doing so ,
one is bound to make a number of assumptions that are either already present in the classical theory of phoresis , or arise as a result of mapping the description of such `` active '' surface particles onto the framework of a theory developed to describe the case of inert particles immersed in a pre - defined , externally controlled concentration gradient . since in the literature these assumptions are often overlooked or not spelled out explicitly , we consider it as necessary to provide also a critical discussion of the significant assumptions involved by this mapping , as well as of some of those assumptions implicitly contained in the standard theory of phoresis . accordingly , the outline of this paper is as follows . in section [ model ]
we define the model and discuss some general aspects of systems with self - generated motion , with particular emphasis on the assumptions involved in adopting the standard theory of phoresis .
section [ sec_diff_phor_velocity ] is devoted to the computation of the diffusiophoretic velocity .
this includes the calculation of the diffusiophoretic slip velocity and the determination of the phoretic hydrodynamic flows and density profiles of the product molecules around a self - propelling particle .
the results are discussed in section [ discussion ] , and we conclude in section [ summary ] with a brief summary of our results and general conclusions .
important details of our calculations are presented in the appendices [ app_a ] , [ app_b ] , and [ app_c ] .
the system we consider is shown in fig .
it consists of an impermeable , spherical particle of radius @xmath0 with a point - like catalytic site ( black dot in fig .
[ fig1 ] ) on its surface , which promotes the chemical conversion of a surrounding solvent into product molecules of diameter @xmath1 ( small hatched circles in fig . [ fig1 ] ) .
the particle and the surrounding solution ( solvent plus the solute , i.e. , the product molecules ) of viscosity @xmath2 are enclosed in a concentric , impermeable , spherical shell of radius @xmath3 ( @xmath4 .
an impermeable , spherical particle of radius @xmath0 with a point - like catalytic site at @xmath5 ( depicted as a black dot ) on its surface , enclosed by a concentric , impermeable , spherical wall of radius @xmath3 ( @xmath4 . the product molecules of diameter @xmath1 are shown as small horizontally hatched circles . ] in general , such a chemical conversion of the solvent gives rise to several types of product molecules .
here we shall focus on the particular case in which the chemical conversion of a solvent molecule ( @xmath6 ) leads to two molecules only ( @xmath7 and @xmath8 ) , one very similar in size and properties with the solvent itself ( @xmath9 ) , the other one ( @xmath8 ) different [ @xmath10 ; in the following this latter is denoted as `` product molecule '' and plays the role of a solute in the solvent .
in other words , we consider a situation in which the net result of the chemical conversion can be approximated as the generation of a solute , solely , and the reaction does not lead to a solvent depletion near the catalytic site which otherwise would reduce the current of solvent towards the catalytic site acting as a solvent sink .
for example , this is approximately the case for the pt catalyzed decomposition of hydrogen peroxide ( h@xmath11o@xmath11 ) in aqueous solution into water ( h@xmath11o ) and oxygen ( o@xmath11 ) molecules , as discussed in refs . .
in these experimental studies the oxygen plays the role of the product molecule the properties of which differ significantly from those of the solvent . here
the solvent is actually a binary liquid mixture of h@xmath11o and h@xmath11o@xmath11 for which h@xmath11o is passive and does not participate in the chemical conversion .
we note that actually it is rather difficult to assess whether or not h@xmath11o and h@xmath11o@xmath11 can be treated as being the same in such systems , and probably the answer would be on a case by case basis .
this is not only because the h@xmath11o - particle and h@xmath11o@xmath11-particle interactions may be quite different for different materials ( the simplest example is exactly the catalytic decomposition of h@xmath11o@xmath11 on pt ) , but also because of the sensitivity of phoresis to the details of the _ solvent mediated _ interactions and the weakly acidic nature of h@xmath11o@xmath11 may play an important role here . in providing this example as a possible realization of a catalyzed reaction @xmath12 we had to rely on : ( i ) the statement in ref .
[ second entry , bottom of left column on page 13427 therein ] , which is based on some previously published results by phibbs and gigure @xcite , that hydrogen peroxide and water in contact with au have almost identical interfacial tensions ( this translates into similar particle - solvent interactions ) , and thus that there is no additional solvent - density gradient to be considered , and ( ii ) on the implicit statement in the same ref .
[ second entry , eq .
( 8) therein ] that the effective interaction of the o@xmath11 molecule with the au surface of the rod does not depend on the composition of the solvent h@xmath11o - h@xmath11o@xmath11 mixture .
we thus assume that the reaction at the catalytic site , ( located at @xmath13 , where @xmath14 is the unit vector of the @xmath15-axis ) , acts effectively only as a point - like source of product molecules @xcite of diameter @xmath1 which are diffusing in the solvent with diffusion coefficient @xmath16 . in passing we note that this is in contrast to the situation considered in ref . , where the catalytic site acts both as a source for the product and as a sink for the solvent , i.e. , the generation of a product molecule is accompanied by the annihilation of a solvent one so that @xmath17 which implies that @xmath8 can only be a different configuration of @xmath6 , because it must have the same chemical constituents . in the situation considered here ,
the amplitude of the production rate of @xmath8 is denoted as @xmath18 .
we shall neglect any interaction between the product molecules .
thus the number density of product molecules is considered to be so low that among themselves they behave like an ideal gas .
there is an interaction potential between the product molecules and the moving particle ( see appendix [ app_a ] ) , which includes the impermeability condition at the surface of particle .
the interactions between the product molecules and the solvent are accounted for in an effective way via the stokes - einstein expression @xmath19 for the diffusion coefficient @xmath16 of the product molecules @xcite , where @xmath20 is the boltzmann constant and @xmath21 is the temperature .
the presence of a source of solute ( product molecules ) on the surface of the particle creates a non - uniform and time dependent distribution of solute in the solution ( see fig .
[ fig1 ] ) , i.e. , a non - uniform composition of the solution . because the understanding of the way in which such a non - uniform composition gives rise to phoretic motion is not straightforward , we discuss here in some detail how the general model introduced in section [ general ] can be put into the context of the standard theory of phoresis @xcite .
a recent clear exposition of the basic concepts and a discussion of different phoretic transport scenarios can be also found in ref . .
if the production rate of the source is not very large and the diffusion coefficient of the product molecules through the solvent is not too small , the solute density varies smoothly in space and slowly in time .
this justifies the assumption of local equilibrium and the definition of a position- and time - dependent chemical potential of the solution ( the spatial gradients of which describe the diffusion of the solute away from the source ) . in a quasi steady - state of the solution , corresponding to slow time - variations of the composition ,
the net solute current is approximately zero and the solute density gradients are balanced by pressure gradients [ see also eq .
( b3 ) in ref . ] .
because in the absence of external body forces the mechanical equilibrium for a liquid solution is generally established much faster than the chemical one ( see , e.g. , ref . ) , it is physically plausible to assume that _ far from the boundaries _
( i.e. , far from the wall and far from the particle surface ) the pressure adjusts instantaneously to accommodate the spatially varying solute density profile @xcite and keeps varying slowly with time ( on time scales much larger than the diffusion time @xmath22 ) because the overall density of the solution increases in time . in other words , the pressure is determined by the solute density profile and is obtained from the equation of state of the solvent - solute binary mixture considered as a thermodynamic system in local equilibrium described by the corresponding free energy density .
clearly , the situation is different in the region of contact between the solution and the particle surface ( and also of contact with the wall ) where the interactions of the particle with the product molecules ( solute ) and with the solvent molecules are relevant .
this interaction disturbs the distribution of solute molecules , net currents of solute and solvent result , and the pressure field becomes a quantity determined by the incompressibility requirement for the hydrodynamic flow , and not by an equation of state .
we shall discuss this important point below . at a given time
the spatial variations of the number density @xmath23 of the solute and of the number density @xmath24 of the solvent are characterized as follows : @xmath24 is constant throughout the solution apart from the close vicinity of the particle surface and of the surface of the confining sphere .
there @xmath24 vanishes and reaches its constant bulk value via density oscillations which are induced by local packing effects caused by the finite diameter @xmath25 of the solvent particles .
away from any kind of surface phase transition the approach of the bulk value occurs exponentially on the scale of the bulk correlation length @xmath26 of the solvent .
( the presence of long - ranged dispersion forces causes power - law decays ; however , their amplitude is small and is neglected in the present context . ) away from bulk phase transitions @xmath26 is comparable with the range of the interaction potential between the solvent molecules , which in turn is proportional to @xmath25 with a prefactor of order unity . within this picture @xmath24
does not vary along the surface of the particle . the solute number density @xmath23 is characterized by two important features . on the scale of the system size @xmath27
, @xmath23 varies due to the diffusion process . on the much smaller length scale of the solute diameter @xmath1
, @xmath23 also varies near the particle surface and near the surface of the confinement as discussed above for the solvent . since , however , according to our earlier assumption the solute particles can be considered to form an ideal gas , near the walls @xmath23 varies proportional to @xmath28 , where @xmath29 is the _ effective _ substrate potential ( see appendix [ app_a ] ) in the sense that it describes the interaction of the solute molecule with the substrate in the presence of the solvent , and @xmath30
. typically the range @xmath31 of @xmath29 is proportional to the solute diameter @xmath1 . accordingly
the solute molecules interact directly with the particle only if they are within a thin surface film of thickness @xmath31 , which is assumed to not be deformed by the motion of the particle @xcite .
the standard theory of phoresis assumes @xmath32 so that within this surface film of thickness @xmath31 the solvent can be considered as a continuum with constant @xmath24 .
we note that this assumption is rather convenient from a computational point of view .
however , this assumption is at odds with the basic underlying model according to which the solute particle is created from a solvent particle via a catalytic reaction . under these circumstances
one expects @xmath33 ( see the above example of the reaction @xmath34 h@xmath11o@xmath11 @xmath35 @xmath34 h@xmath11o + o@xmath11 ) . in the absence of a more detailed theory of phoresis , which treats the sizes of the solvent and solute particles on the same footing , we proceed along the lines of the standard theory .
nonetheless we point out that there is an urgent need for improvement
. it might be that due to the fact that continuum hydrodynamics remains quantitatively reliable down to surprisingly small length scales ( here @xmath25 ) , it is conceivable that the aforementioned continuum description yields reliable results , too .
but this remains unproven .
according to the standard theory a hydrodynamic description applies within the aforementioned surface film .
within this picture the solute molecules and their effective interaction with the particle are replaced by a corresponding distribution of `` point forces '' acting on the solvent in the film . within the limitations of such an approach the hydrodynamic description of the solution naturally splits into that of an `` inner '' region formed by the surface film and that of the `` outer '' region formed by the exterior space beyond the surface film .
following refs .
, the ensuing asymmetric , non - uniform solute number density @xmath36 around the particle will give rise , _ within the surface film only _
, to a pressure gradient along the surface of the particle .
this is because the surface film is very thin on the scale of the system size @xmath0 so that the equilibration of the composition profile of the solution within the surface film can be assumed to be fast along the direction normal to the surface of the self - propelling particle compared with the diffusional relaxation time of the composition gradient along its surface , which typically involves a length scale of the order of the particle size @xmath0 . therefore , within the surface film the solute density in the direction normal to the surface of the particle is given by a boltzmann distribution corresponding to the local equilibrium configuration in the presence of the effective interaction potential between the particle and the solute molecules , with a prefactor which depends on the position along the surface of the particle ( for details see appendix [ app_a ] , which builds on ref . ) .
mechanical equilibrium of the solvent within the surface film along the direction normal to the surface of the particle ( no flow along this direction ) requires the pressure gradient along the normal to be equal to the body force densities due to the effective particle - solute interactions .
therefore the pressure within the surface film differs from the `` outer '' pressure field by an `` osmotic pressure '' term , i.e. , a term proportional to the extra solute density @xmath23 in excess to the constant solvent density @xmath24 .
since this osmotic pressure varies along the surface of the particle , there is a gradient of pressure along the surface of the particle .
this lateral pressure gradient is not balanced by any body force and thus generates shear stress within the surface film .
therefore it induces hydrodynamic flow of the solution along the surface of the particle and entails motion of the particle with a velocity @xmath37 . because the system has azimuthal symmetry , the motion is along the @xmath15-axis , i.e. , @xmath38 .
the hydrodynamic flow of the solution , the diffusive transport of the solute , and the coupling between the two giving rise to phoretic motion of the particle are analyzed in detail in the next section ( see also the appendices [ app_a]-[app_c ] ) .
our calculation of the diffusiophoretic velocity proceeds along the lines of ref . and is based on dividing the problem into an inner one within the film around the particle surface discussed in subsec .
[ sec_standard_phoresis ] , and an outer one beyond the range of the effective interaction between the solutes and the particle . in view of the arguments presented in sec .
[ model ] , we base our analysis on the following assumptions:(i ) the chemical reaction leads to a change in the solute density only .
the thickness of the surface film , defined by the range of the effective interaction between the product molecules ( solute ) and the particle , is much smaller than the particle radius @xmath0 .
the spatial variations of the solute number density along the surface of the particle occur over length scales of the order of @xmath0 , which allows one to use the approximation of a locally planar interface .
( ii ) the number density of product molecules is sufficiently low , so that the solute can be viewed as an ideal gas and the solvent - solute mixture behaves like an ideal dilute solution . in this case , the pressure gradient is simply proportional to the gradient of the number density of the product particles ( see appendix [ app_a ] ) .
( iii ) temporal variations of the number density of the product molecules due to their creation occur on time scales much longer than those needed for the fluid flow ( as seen from the moving particle ) to relax to a steady state corresponding to the number density profile at that moment .
( iv ) the flow field of the solution within the surface film can be described by the laws of hydrodynamics .
additionally , we assume that both the reynolds number @xmath39 , where @xmath40 is the _ mass density _ of the solvent , and the peclet number @xmath41 are small , such that one can approximate the hydrodynamic description with the creeping flow ( stokes ) equations and disregard the convection of the solute compared to its diffusive transport . here
we have assumed that the magnitude of the hydrodynamic flow @xmath42 is similar to that of the phoretic velocity @xmath43 ; this assumption is supported _ a posteriori _ by the fact that for our system the phoretic velocity is basically the average of the slip - velocity over the surface of the particle [ see , c.f . , eqs .
( [ slip_vel ] ) and ( [ velocity ] ) ] . for a @xmath2 m size particle moving through water ( density @xmath44 , viscosity @xmath45 ) with a velocity of the order of @xmath2 m/s , which is typical for phoresis
, one has @xmath46 . for the diffusion at room temperature ( @xmath47 j ) of o@xmath11 ( @xmath48 m ) in h@xmath11o@xmath11 ( @xmath49 ) ,
the stokes - einstein relation leads to an estimate @xmath50 for the diffusion coefficient ( in agreement with ref . ) , and thus @xmath51 . therefore the latter assumptions are justified .
note that if one uses @xmath52 rather than @xmath0 as a characteristic length scale , the above results imply that the @xmath53 and @xmath54 numbers remain both very small as long as @xmath55 . as discussed in sec .
[ model ] , the pressure gradient along the surface of the particle , induced by its interaction with the non - uniform distribution @xmath36 of product molecules , leads to flow of the solution relative to the particle .
as shown in appendix [ app_a ] , the hydrodynamic flow within the surface film translates into a ( phoretic ) slip - velocity , @xmath56 at the outer edge @xmath57 of the surface film as a boundary condition for the hydrodynamic flow in the outer region . in this equation
@xmath58 denotes the projection of the gradient operator onto the tangential planes of the surface of the particle , @xmath59 where @xmath60 and @xmath61 are the polar and azimuthal unit vectors , respectively , while @xmath62 is an effective `` mobility '' , and @xmath63 is a characteristic length scale .
the latter is determined by the effective interaction potential @xmath29 between the particle and the product molecules @xcite [ which determines their distribution within the surface film along the direction @xmath64 normal to the particle surface within the local coordinate system ( fig .
[ fig3 ] ) or the radial direction here ] : @xmath65 in the case of a purely steric repulsive interaction , one has @xcite
@xmath66 so that @xmath67 . for a single catalytically active site as shown in fig . [ fig1 ]
@xmath23 and thus @xmath68 do not depend on the azimuthal angle @xmath69 .
we focus here on the quasi - static approximation ( iii ) described above and the limit of low re numbers . in the laboratory frame ,
in which the particle moves with a yet unknown velocity @xmath70 , the hydrodynamic flow @xmath71 at time @xmath72 in the domain beyond the surface film around the particle obeys the steady - state force - free stokes equations @xmath73 where @xmath74 is the pressure tensor , @xmath75 is the hydrostatic pressure , and @xmath76 is the shear stress tensor , i.e. , @xmath77 , subject to boundary conditions ( bc ) of no - slip at the wall at @xmath52 and prescribed slip velocity on the surface @xmath78 , i.e. , @xmath79 note that the parametric dependences of the flow field @xmath42 on @xmath72 and @xmath80 stem from the boundary conditions . note that because for the outer problem the variations of the flow field and those of the number density of the product molecules are over length scales that are much larger than @xmath31 , one can replace everywhere in the calculations the ( unknown ) value of @xmath78 by @xmath0 ; we shall use this approximation in the following .
the solution of the stokes equations with boundary conditions on spherical surfaces is based on expressing both the flow field @xmath81 and the pressure field @xmath75 as series of solid harmonics @xcite @xmath82 ( @xmath83 ) , which are the eigenfunctions of the _
laplace operator . in appendix
[ app_b ] we provide a brief outline of the general method for solving eq .
( [ st_eq ] ) in spherical coordinates and derive the solution obeying the bcs given by eq .
[ bc_flow ] . equation ( [ bc_flow ] ) at @xmath52 corresponds to a no - slip condition imposed on a wall fixed with respect to the laboratory frame .
this deserves further discussion .
( [ bc_flow ] ) the bc is that @xmath81 at the surface of the moving particle , i.e. , @xmath84 where @xmath85 is the position of the particle center , takes the value @xmath86 .
as long as @xmath87 , concentricity holds and one obtains the first of the two bcs given by eq .
( [ bc_flow ] ) .
if , however , the velocity @xmath88 is not small , we readjust the center of @xmath52 in order to impose concentricity using the following `` protocol '' : the particle is allowed to move for a short time @xmath89 with the instantaneous velocity @xmath37 while the spherical shell at @xmath52 is fixed , after which the spherical shell is displaced by @xmath90 to its new fixed position in such a way that it does not perturb significantly the density and flow fields ; this procedure is then repeated .
it is not clear to which extent this latter assumption , which allows us to obtain an analytical solution , can be realized experimentally .
but it is expected that the results we derive for the present geometrical setup are relevant also for more general geometries @xcite , such as a particle moving along a channel in _
3d _ , for which the constraint that the confinement moves with the particle is not needed .
we note that in the general case of a non - spherical particle , or of a particle with non - uniform surface properties ( e.g. , a spatially varying effective mobility @xmath91 ) the particle can also rotate and the bc at the particle surface , eq . ( [ bc_flow ] ) , should include a term accounting for a rigid - body rotation with angular velocity @xmath92 . however ,
this angular velocity turns out to be identically zero in the case of a spherical particle with an effective mobility @xmath91 which is constant on its surface @xcite ; therefore we completely disregard it here .
the velocity @xmath70 of the particle is obtained by requiring that the hydrodynamic force @xmath93 where @xmath94 is the radial unit vector and @xmath95 the surface area element on the spherical surface @xmath96 ( note that we replaced @xmath78 by @xmath0 , as discussed in the previous subsection ) , exerted by the fluid on the composite domain particle plus surface film vanishes [ see the vector identities in eq .
( [ vec_ident ] ) , ( i ) - ( iii ) below this equation , and @xmath97 , with summation over @xmath98 : @xmath99 = 0 \,.\ ] ] if , as discussed above , a rotational motion with angular velocity @xmath100 would have to be considered , too , this will be determined by the additional requirement that the motion is not only force free but also torque free @xcite .
this is again due to the fact that there are no net forces acting on the object composed of the particle and its surface film .
the above argument for determining the velocity @xmath43 has been discussed in detail by anderson ( see ref . and references therein ) , but it is often overlooked and replaced by the incorrect argument of a balance between a drag force - i.e. , the integral of the non - uniform osmotic pressure proportional to the density of solute ( see appendix [ app_a ] ) - exerted on the particle and a stokes - like viscous friction from the solvent ( see , e.g. , refs . ) .
it is important to realize that the occurrence of composition gradients in the solution does not give rise , by itself , to an osmotic pressure ( see also ref . ) , in contrast to such an assumption made in ref . .
such gradients will simply lead to diffusion of the product molecules , while the pressure in the solvent will adjust to accommodate the spatially varying chemical potential corresponding to the quasi - stationary density profile ( mildly time dependent due to the overall increase of the solute number density , in the case of the confined system ) , reflecting mechanical equilibrium @xcite . as a matter of fact , the origin of this osmotic pressure resides in the interaction between the particle and the product molecules , i.e.
, it requires the explicit consideration of the effective interaction potential between the particle and the product molecule ( for a detailed illuminating discussion of this point see refs . ) . an intuitive argument that the use of such a `` stokes - formula '' , which stems from a standard `` drag balanced by viscous friction '' type of reasoning , is incorrect can be formulated as follows . at distances
far from the particle the flow field should look like that produced by the superposition of a point force @xmath101 at the origin , i.e. , the center of the particle , which is the integrated ( over the particle surface ) effective `` product molecules on particle '' interaction ( the forces @xmath102 in fig .
[ fig3 ] in appendix [ app_a ] ) , and a distribution of effective point forces oriented radially ( of the particle acting on the product molecules ) in a small shell region around the surface of the particle ( which is the aforementioned surface film ) , which upon integration gives exactly @xmath103 .
note that for repulsive effective interactions and for an axisymmetric distribution of product molecules with an increasing density towards the source at @xmath104 , as in fig .
[ fig1 ] , @xmath105 is oriented into the positive @xmath15-direction ; for attractive interactions , @xmath101 would be oriented into the negative @xmath15-direction . this can be seen as follows . for repulsive interactions ,
each of the forces @xmath102 is oriented radially inwards ; since the density of the product molecules is increasing towards the source located on the lower hemisphere , the magnitude and the projection onto the @xmath15-direction of the force due to the domain @xmath106 located at any @xmath107 is smaller than the one in the corresponding domain @xmath106 located at @xmath108 .
thus the contribution into the positive @xmath15-direction from the lower hemisphere will dominate .
for attractive interactions , the argument is simply reversed .
this is in accordance with the following two statements regarding the characteristics of the far - field ( i.e. , on length scales over which the particle plus surface film are seen as point - like ) hydrodynamic flow : ( i ) in the absence of external body forces such as gravity or centrifugal forces the motion of the particle plus its surface film is net force free .
( ii ) the generated flow corresponds , within a first approximation , to that produced by a `` force dipole '' [ which is the superposition of a distribution of `` force dipoles '' ( @xmath102 acting on the center and @xmath109 on the product molecules in @xmath110 ) as in fig . [ fig3 ] in appendix [ app_a ] ] , or a higher order `` force multipole '' , e.g. , quadrupole ( if it happens that the net force dipole vanishes , too ) @xcite at the origin , i.e. , the position of the center of the particle , rather than to the one due to a point force , which would be the case for an object uniformly dragged against the viscous stokes friction ( see , e.g. , ref . ) . for an unbounded system ,
these forces translate into a radial decay of the phoretic flow field proportional to @xmath111 ( force dipole ) or @xmath112 ( force quadrupole ) , in contrast to the decay proportional to @xmath113 corresponding to a point force ; for a bounded system , the differences between the flow fields can not be any longer summarized by such a simple criterion as different power laws for the radial decay , but they remain significant nevertheless .
these features of the hydrodynamic flow are discussed in more detail in the appendix [ app_b ] ( see also , c.f . , fig .
[ fig4 ] ) . using the expansion of the velocity and pressure fields in terms of the solid harmonics @xmath114
( see appendix [ app_b ] ) , the hydrodynamic force on the particle , defined by eq .
( [ hydro_force ] ) and expressed as on the left - hand side of eq .
( [ zero_force ] ) , reduces to @xmath115\,,\ ] ] where @xmath116 is the coefficient of @xmath117 in the expansion of the pressure @xcite [ eqs .
( [ press_expansion ] ) and ( [ solid_harmonics ] ) ] , and @xmath118 is the legendre polynomial of order one .
all other terms vanish since the corresponding integrals are exactly equal to zero .
thus the requirement of a vanishing @xmath119 implies @xmath120 , which leads to ( see appendix [ app_b ] ) @xmath121 where @xmath122 the structure of the expression on the right - hand side ( rhs ) of eq .
( [ velocity ] ) clarifies the meaning of the factor @xmath123 , which varies between zero , at @xmath124 , and one , at @xmath125 ( see fig .
[ fig2 ] ) . without this factor
, one has on the rhs the phoretic velocity in the unbounded space due to a source which generates a composition profile @xmath126 such that @xmath127 at all times @xmath72 , as derived in ref . [ with @xmath128 entering into @xmath91 ; see the text around eq .
( [ lambda_def ] ) ] .
thus @xmath123 is a `` hydrodynamic wall - correction '' factor which quantifies and summarizes the effects solely due to the confinement induced changes in the characteristics of the solvent flow .
note that @xmath129 means that the hydrodynamic effects due to confinement will tend to decrease the velocity from the value corresponding to an unbounded system .
according to eq .
( [ velocity ] ) , knowledge of the density @xmath130 of product molecules completely determines the velocity @xmath88 of the particle as a function of @xmath131 for confined systems and thus allows one to quantify the effects of confinement on the resulting phoretic motion . within the assumptions that the diffusion of product molecules is fast compared with the convection by the solvent flow , i.e. , in the limit of small peclet numbers , and that the product distribution @xmath36 is undisturbed by the flow , i.e. , neglecting any so - called polarization effects of the surface film @xcite , the time evolution of the number density @xmath36 of product molecules around the moving particle is governed , in the co - moving frame , by the diffusion equation @xmath132 this equation is to be solved subject to the initial condition ( ic ) of zero density of product molecules and to the boundary conditions of zero normal current at the confining wall ( assuming that the latter is co - moved , without perturbing the solution , such that at all times it remains concentric with the particle ) and at the outer edge @xmath133 of the surface film
actually , the latter condition should be imposed at the surface of the particle , but we shall consider only the case of a thin surface film so that due to its small volume the transport of solute to the surface film is negligible compared to the transport in the outer region and the zero normal current condition at @xmath78 is a good approximation .
hence , the ic and bcs are : @xmath134 we remark that the linearity of eqs .
( [ diff_eq ] ) and ( [ velocity ] ) with respect to @xmath23 imply that by superposition the solution of the present problem can be extended to the case of an arbitrary spatial arrangement of several catalytic sites on the particle surface .
this will allow one to find an optimal decoration for providing stability against particle rotations ( caused by thermal noise ) , which can otherwise spoil the unidirectional motion ( see ref . for such examples of designed optimal surface distributions ) .
equation ( [ diff_eq ] ) subject to the ic and bc conditions given by eq .
( [ bc_ic ] ) is solved by using the laplace transform .
similarly to the approximation employed in sec .
[ subsec_phor_flow ] , from this point further in the calculations we replace everywhere @xmath78 by @xmath0 because the difference @xmath31 ( the thickness of the surface film ) between the two is negligible on the macroscopic length scales characterizing the transport in the outer region .
the solution is obtained as a series in products of legendre polynomials and modified spherical bessel functions of the first and third kind , respectively ; the details are provided in appendix [ app_c ] .
focusing on the particular case in which after its start the activity of the catalytic site is time independent , i.e. , @xmath18 is constant : @xmath135 , where @xmath136 is the heaviside step function and @xmath137 is the average production time for the creation of a product molecule .
the laplace transform of the integral in eq .
( [ velocity ] ) is computed by using the laplace transformed density @xmath138 and the fact that @xmath139 . formally inverting the laplace transform ,
one obtains the diffusiophoretic velocity as a function of time and confinement : @xmath140\right)|_s\,,\ ] ] where @xmath141 \,,\hspace*{.2in}\end{aligned}\ ] ] @xmath142 and @xmath143 are modified spherical bessel functions of the first and third kind , respectively @xcite , and the dimensionless coefficients @xmath144 and @xmath145 are fixed by imposing the boundary conditions [ see eq .
( [ alpa_beta_3d_eqs ] ) and appendix [ app_c ] with @xmath146 so that @xmath147 , @xmath148 , and @xmath149 $ ] are dimensionless ] .
the complicated structure of the function @xmath150 makes it rather laborious to carry out the inverse laplace transform , so that the full time dependence of the velocity @xmath151 can not be derived easily .
however , the asymptotic ( @xmath152 ) value of the velocity @xmath153 can be straightforwardly derived by using the inversion formula for the laplace transform @xcite by noticing that @xmath154 has a simple pole at @xmath155 , which determines the asymptotic value of the velocity as the residue of @xmath154 at @xmath155 @xcite .
the existence of such a constant asymptotic value , which at first glance seems to be in conflict with the fact that the density @xmath156 does not reach a steady state ( since we consider a closed system with a source continuously producing non - interacting point - like particles ) , is due to the fact that the phoretic motion is determined solely by the gradient of the number density of the product molecules along the surface of the particle . at long times ,
the total density of product molecules is large and any redistribution ( which would lead to a change of the density gradient ) proceeds on very slow time scales .
( ultimately the density of product molecules becomes so high that they no longer behave as an ideal gas and the latter assumption breaks down . )
this yields @xmath157 = v_0 \chi(\eta)\,,\ ] ] where @xmath158 is the asymptotic velocity in the case of an unbounded system and [ see eq .
( [ def_chi1 ] ) ] @xmath159 thus @xmath160 is a `` diffusion wall - correction '' factor , and @xmath161 quantifies the combined wall effects , which in the present case factorize into a hydrodynamic and a diffusion contribution .
these two contributions oppose each other such that the confinement in hydrodynamics decreases the particle velocity ( @xmath162 ) whereas the confinement of the diffusion enhances it ( @xmath163 ( see fig .
[ fig2 ] ) ; the latter dominates so that @xmath164 and there is an overall enhancement of the velocity ( see fig .
[ fig2 ] ) .
( solid line ) as a function of the confinement @xmath80 ( see fig .
[ fig1 ] ) and the hydrodynamic and diffusion `` wall - correction '' factors @xmath129 ( dotted line ) and @xmath165 ( dashed line ) . ] assuming the stokes - einstein relation for the diffusion coefficient of the product molecules in the solvent ( see sec .
[ intro ] ) and using the expression eq .
( [ b_def ] ) for the mobility @xmath91 , the velocity @xmath166 [ eq .
( [ vasymp ] ) ] can be rewritten as @xmath167 surprisingly , this expression has no explicit dependence on the viscosity of the solvent or on the temperature .
in fact , the dependence on viscosity drops out upon using the stokes - einstein relation . in turn ,
the dependence on @xmath21 is hidden in the temperature dependence of the ( effective ) interaction between the particle and the product molecules which determines the thickness of the surface film , i.e. , the length scale @xmath168 [ see eq .
( [ lambda_def ] ) ] . in the case of only steric repulsion between the particle and the product molecules , for a particle of radius @xmath169 [ @xmath170 , a product molecule with diameter @xmath171 , @xmath172 and a reaction rate ( following ref . ) of @xmath173
, one finds the velocity @xmath174 pm / s [ 1 nm / s ] ; consequently @xmath175 is of the same order of magnitude .
note that these values are much smaller than the estimate @xmath176 of ref .
, which is based on a significantly larger value @xmath177 than the one @xmath178 corresponding to the reasonable estimate @xmath179 nm for the diameter of the product molecule
. however , this estimate corresponds to the case of a single catalytic site .
one can argue that several such single reaction sites distributed closely around @xmath180 may lead to a significant increase of the velocity ( intuitively , by a factor equal to the surface density of such reaction sites ) .
thus the resulting velocities may eventually be closer to the experimentally reported values @xcite
@xmath181 for objects with a @xmath2m@xmath182 area covered by catalyst if the surface density of reaction sites is of the order of nm@xmath183
. moreover , if the effective interaction between the particle and the product molecules is attractive , the ratio @xmath184 can be very large @xcite .
another possibility leading to a significant increase is that of a classical hydrodynamic slip boundary condition with a large slip - length at the surface of the particle for the flow within the surface film , assuming that ( see appendix [ app_a ] ) classical hydrodynamics provides an accurate description for the solvent flow even at such small scales @xcite .
( according to the text following eq .
( [ g_def ] ) , for the derivation of the effective mobility @xmath91 in the phoretic slip - velocity [ eqs .
( [ slip_vel ] ) , ( [ b_def ] ) , and ( [ slip_vel_hatx ] ) ] we have used a no - slip boundary condition for the flow in the surface film at the particle surface .
this condition can be generalized to a slip condition characterized by a slip length . )
such hydrodynamic slip may be , e.g. , due to roughness of the particle surface , and its effect on the phoretic motion can be intuitively understood as follows .
the slip over the surface of the particle facilitates the solvent flow within the surface film as compared to the case of zero slip ( a `` lubrication '' effect ) , and thus for the same pressure gradient along the surface as in the case of zero slip the flow velocity in the surface film will be increased for nonzero slip - lengths .
the phoretic slip @xmath185 , which is the flow velocity at the outer boundary of the surface film , is thus also increased compared to its value in the case studied here so far that there is a no - slip condition at the surface of the particle ; this leads to an increase ( by up to orders of magnitude @xcite ) of the phoretic velocity of the particle .
the velocity @xmath175 of the particle can be positive or negative , depending on the details of the interaction potential .
this means that the particle can travel either following the gradient of solute density , or against it . according to eqs .
( [ vasymp ] ) and ( [ v0_simple ] ) as well as the definition of the effective mobility @xmath91 , the sign of the phoretic velocity is opposite to the sign of the parameter @xmath91 ( or , equivalently , that of the parameter @xmath186 ) , which in turn depends , inter alia , on the attractive or repulsive character of the effective interactions between the particle and the solute molecules [ see eq .
( [ lambda_def ] ) ] .
equations ( [ slip_vel ] ) and ( [ surf_grad ] ) show that the sign of the phoretic slip velocity , i.e. , of the polar component @xmath187 is also opposite to the one of @xmath91 , because for our system @xmath188 is positive ( see fig .
[ fig1 ] ) . in the case of hard core interactions only one
has @xmath189 and thus @xmath190 ; this implies @xmath191 , i.e. , the particle moves in the direction of @xmath14 and thus _ away _ from the catalytic site , while the slip velocity points in the same direction as @xmath60 , which has a negative @xmath15-component , so that the flow around the particle is also oriented towards the catalytic site @xcite . the confinement does not change the direction of @xmath88 , which for an unbounded system is , as expected , in agreement with ref . . a simple intuitive explanation for the direction of the motion of the particle follows from the observation that there is no net force acting on the composite consisting of the particle plus the surface film , so that there is conservation of momentum . because the solvent flow around the particle is in the negative @xmath15-direction [ see fig . [ fig4](a ) ] , the particle should move in the positive @xmath15-direction .
note that while this simple argument is clear in the unbounded case , for the confined system it breaks down because the formation of vortices [ see fig .
[ fig4](c ) ] implies that there are spatial regions where the flow is in the positive @xmath15-direction .
thus a quantitative analysis is required . if @xmath192 , on the contrary , @xmath193 , which means that the particle will move _ towards _ the source of the product molecules .
this is in contrast to the results presented in ref . , which predict motion _ always away _ from the source . as discussed before the reason for this discrepancy is the ignorance in ref . of the mediating role of the solvent and its assumption of a mapping of the non - uniform density of the solute onto a non - uniform `` osmotic pressure '' acting on the particle ( i.e. , assuming that more product molecules impinge on the particle from the side with higher density ) , rather than mapping onto gradients in the solution pressure along the surface of the particle within its surface film .
this opposing directionality in the case @xmath192 thus provides a simple criterion for an experimental discrimination between the predictions of the two proposals .
spatial confinement leads to an overall enhancement of the phoretic motion , because @xmath194 for all finite @xmath80 . as shown in fig .
[ fig2 ] , even at moderate values of @xmath80 the effect of the confinement is important , e.g. , at @xmath195 there is a ca .
@xmath196 increase in the velocity compared with the unbounded case , and at @xmath197 the increase reaches ca .
@xmath198 . for @xmath124
the velocity stays finite .
this is , of course , an artifact which stems from the assumption that the product molecules are point - like . actually , there is a lower cut - off @xmath199 , where @xmath200 is of the order of the hard core diameter @xmath1 of the product molecules , below which this assumption breaks down and eq .
( [ vasymp ] ) is no longer valid .
however , for @xmath201 , which is a reasonable assumption , one has @xmath202 , and thus the steep increase of the velocity near @xmath203 is physically relevant .
this is a clear counterexample for the statement in the conclusions of ref .
that `` the mobility of such swimmers will be hindered by channel boundaries '' .
note that the velocity remains non - zero and finite for all values of @xmath80 , thus the two opposing effects described by the factors @xmath204 and @xmath205 are of similar magnitude . this can be understood qualitatively from the behavior of the velocity and density fields in the unbounded system : the hydrodynamic flow velocity decays as @xmath112 , where @xmath206 is the distance from the center of the particle @xcite , and the gradients in the number density of product particles ( for phoresis the gradients are relevant , not the density in itself ) are are also varying as @xcite @xmath112 . accordingly
the confinement becomes relevant for both fields at similar length scales and with a similar power - law behavior .
therefore one can expect that the confinement has an equally strong influence on the hydrodynamics and the diffusion .
in summary , we have discussed the effect on the phoretic velocity of a self - propelled particle due to a confining wall for the solvent and for the reaction product emerging from a catalytic site on the surface of the dissolved particle .
the analysis is based on considering the present model in the context of the standard theory of phoresis .
we have critically analyzed the assumptions involved in such an approach .
some of them are already present in the classical theory , others arise as a result of mapping the description of such `` active '' surface particles onto the framework of a theory developed to describe the case of inert particles immersed in a pre - defined , externally controlled concentration gradient .
for example , within the standard theory of phoresis the perturbation of the steady - state , externally controlled concentration gradient is induced by the interaction with the surface of the particle _ upon _ immersion .
while this is an acceptable concept in that context , it is not clear how one can justify this if the concentration gradient develops as a function of time with the particle already present , as it is the case for a self - propelled object as the one we have discussed .
it is beyond the scope of the present work trying to improve the general theory , but we consider a clear understanding of its limitations and of its possible shortcomings as a crucial step for both avoiding confusions such as those involving the application of a stokes - force argument and for future theoretical developments .
the main conclusion from this part of our work is that the development of a microscopic model for the dynamics in the surface film , eventually along the lines of ref . which treats the solvent and the solute molecules on equal footing , seems to be very important . within the confines of the standard theory of phoresis
, we have shown that the presence of a confining wall for the solvent and for the reaction product emerging from a catalytic site on the surface of a dissolved particle leads to a significant increase of the velocity of the self - propelled particle .
this results from two competing effects : an increase of the solute density gradients along the surface of the particle and a simultaneous increase of the hydrodynamic viscous friction .
the former one dominates . if only steric repulsion between the particle and the product molecules is present , the absolute value of the velocity is expected to remain , in general , rather small .
a direct experimental realization of the co - moving geometry considered here seems to be difficult .
but the results which we have derived are expected to be applicable ( at least qualitatively ) for more general geometries ( see ref . ) , such as the motion of spherical particles along cylindrical tubes . in this sense ,
an experimental test of our results may be possible .
further extensions may focus on new phenomena emerging from more complicated geometries , as well as from relaxing the assumption of no interaction between the product molecules . by taking into account the actual finite size of the product molecules and by considering generic chemical reactions ,
in which the reaction products are indeed different from the solvent molecules , it is expected that the production of such solute particles is accompanied by a non - uniform depletion of the solvent around the particle .
thus in general the motion of the particle will be determined by the gradients in both the solvent and the solute ( see also ref . ) .
moreover , such a depletion zone would also lead to a decrease of the production rate of the catalytic site and consequently , to a reduction of the density gradients .
accordingly one might expect that there are optimal values for the reaction rates for which the phoretic velocity is maximal .
the interaction between the product molecules may play a significant role , and this deserves further discussion .
if the density of the product particles is not low ( which is reasonable to expect at least near the catalytic reaction sites , for fast reactions , and for slow diffusion of product particles ) and the ( solvent - mediated ) solute - solute interactions have attractive or repulsive components longer ranged than the hard - core interaction discussed above , several other effects have to be carefully considered , especially if they are as important as the particle - solute interactions . in this case , the distribution of the solute particles in the direction normal to the surface of the big particle is no longer given by eq .
( [ boltzmann ] ) , but it is determined both by the particle - solute interactions and by the contributions of the solute - solute interactions , the latter depending on the whole distribution of solute particles around the big particle . thus it becomes a complicated non - local problem which can not be easily addressed analytically . within a mean - field approximation , one may still assume that eq .
( [ boltzmann ] ) holds if the potential is modified to include an averaged , effective solute - solute interaction .
it is evident that even such a simplistic approach will lead to a different expression for the phoretic - slip velocity , and thus one can reasonably expect qualitative differences compared to the predictions of the classic theory of phoresis .
moreover , the solute - solute interactions will lead to a dynamics which differs from simple diffusion and is determined , roughly speaking , by the density- and interaction - dependent collective diffusion coefficient . for example , attractive solute - solute interactions will lead to a tendency of `` clustering '' and thus will hinder the relaxation of the solute gradients , which intuitively would lead to an increase in the velocity of the particle , while repulsive interactions will help to dissipate the particle gradients and , intuitively , would lead to a decrease in the velocity . however , such intuitive arguments for changes in the velocity have to be carefully considered because , e.g. , it is not clear if in the presence of significant solute - solute interactions the assumption of an ultrathin `` unpolarized '' surface film still holds .
the authors thank prof .
u. seifert for very useful remarks concerning the role of the solvent and the phoretic hydrodynamic flow .
m.n.p . gratefully acknowledges very fruitful discussions with a. gambassi , l. harnau , and m. tasinkevych .
g.o . acknowledges partial financial support by agence nationale de la recherche ( anr ) under the grant `` dyoptri - dynamique et optimisation des processus de transport intermittents '' .
m.n.p . and
g.o acknowledge the hospitality of the max - planck institute fr metallforschung ( mpi - mf ) in stuttgart , as well as partial financial support by the mpi - mf .
the following considerations are connected to the corresponding ones in ref . and put them into the context of the microscopic model discussed in sect . [ model ] . due to the azimuthal symmetry of the system , the flow field @xmath207 of the solvent
has non - zero components only along the radial and polar directions . according to the considerations in sect .
[ model ] , the dissolved product molecules ( solute ) of diameter @xmath1 are exposed to an effective interaction with the particle within a surface film of thickness @xmath208 . accordingly , the solute molecules do not interact with the particle beyond the radial distance @xmath209 ( measured from the center o of the particle ) . since the particle radius @xmath0 is much larger than @xmath1 so that @xmath210 , this surface film of thickness @xmath31 can be approximated to be locally planar ( see fig .
[ fig3 ] ) .
schematic expanded view of a domain @xmath211 of the contact region between the particle and the solution , far from the reaction site , approximated locally by a planar geometry .
the effective interaction between the reaction products , shown as horizontally hatched circles , and the particle is described by an effective potential @xmath212 ; @xmath64 is the normal distance from the surface of the particle , also schematically depicted .
the arrows at the product molecules indicate the forces @xmath213 acting on them , while the arrow at the particle center shows the force @xmath214 on it , due to the same interaction with the product molecules , which is equal in magnitude and opposite to the sum of @xmath213 over the product molecules positions in @xmath211 . here
the case is shown that the product molecules in @xmath211 exert a net attractive force on the particle .
the `` outer edge '' @xmath78 of the surface film beyond which the effective interactions between the particle and the product molecules are negligible is indicated by the horizontal dashed line .
the solvent , within which the product molecules move , is taken to be a homogeneous background ( not shown ) .
actually , @xmath215 ( of the order of the product molecules diameter @xmath1 ) is considered to be smaller than indicated here . ] in a small domain @xmath211 centered at @xmath216 ( fig .
[ fig3 ] ) , the local coordinate system ( ^ ) is chosen such that @xmath217 is along the polar direction @xmath60 while @xmath64 is along the radial ( normal ) direction @xmath94 . within the planar film approximation , which is reliable because the film thickness @xmath31 is taken to be much smaller than the particle radius @xmath0 ( @xmath218 ) , and for slow and smooth variations of the solute density along the surface of the particle ( i.e.
, the solute density is assumed to vary over length scales of the order of the particle radius @xmath0 ) , the flow of the solvent relative to the particle surface is @xmath219 ( with the meaning that the ratio @xmath220 is of the order @xmath221 ) @xcite .
this can be intuitively understood starting from the incompressibility condition @xmath222 , where @xmath223 indicates that the derivatives are taken with respect to the local coordinates .
the component @xmath224 is of the order of the slip velocity @xmath225 , which is expected to be proportional to the gradient along the surface , i.e. , along the @xmath217 direction , of the solute density @xmath23 evaluated at @xmath78 . noting that @xmath226 , one obtains @xmath227 . on the other hand
, the component @xmath228 has to rise steeply within the surface film of thickness @xmath31 from the value zero at the surface of the particle to the value corresponding to the radial component of the outer flow , which is expected to be also of the order of @xmath225 .
therefore @xmath229 are expected to not depend on the surface film thickness @xmath31 and to be finite ( due to assumed smooth , slow variations over length scales of the order of the particle radius ) , it follows that @xmath230 so that @xmath231 implies that the leading contribution @xmath232 must vanish separately because it can not be canceled by subleading contributions @xmath233 .
therefore one has @xmath234 ( up to terms of relative order @xmath235 ) .
this implies that @xmath228 does not depend on @xmath64 ; because @xmath236 on the surface of the particle ( independently of @xmath217 ) , due to the surface being impermeable , it follows that @xmath237 ( in the sense of correction terms @xmath238 ) everywhere in @xmath239 . as indicated in fig .
[ fig3 ] , the effective interaction of the product molecules ( solute ) with the particle , characterized by the effective interaction potential @xmath212 ( in the limit of low solute density many - body effects can be ignored so that @xmath29 is independent of @xmath217 ) , gives rise to a force @xmath240 acting on the solute molecules . assuming that the solvent can be approximated as a continuum on the length scale of the solute
, the no - slip condition at the surface of the _ solute _ molecules implies that a body force density @xmath241 is transmitted to the solvent .
( note that , as discussed also in the main text , the approximation of the solvent as a continuum breaks down if the solute molecules have a size similar to that of the solvent molecules , as in the case of the experiments discussed in refs .
; in such a case this transmission of the body force has to be considered as an assumption to be verified _ a posteriori_. ) evidently , such forces are present only within the surface film , where the effective interaction potential is non - zero . within these assumptions ,
the stokes equations for the flow @xmath219 within the surface film take on the form : @xmath242 the equation corresponding to the @xmath217 component of eq .
( [ stokes_complete ] ) can be further simplified by noting ( with an argument similar to the one used in the paragraph above ) that the partial derivative @xmath243 is much smaller than @xmath244 ( by a factor of the order @xmath245 ) .
thus for the @xmath217 component eq .
( [ stokes_complete ] ) leads to [ inner_stokes ] @xmath246 whereas the @xmath64 component yields @xmath247 we further assume that within the very thin surface film the relaxation of the solute density profile along the direction normal to the surface of the self - propelled particle towards a steady - state ( zero net current , i.e. , the diffusion current due to gradients in @xmath23 along the @xmath64 direction is balanced by the convective current generated by the force - field @xmath248 ) is fast compared to the diffusional relaxation time of its gradient along the surface @xcite . in this case , _ within _ the surface film the number density of the solute , assumed to behave like an ideal gas , is given by the boltzmann distribution corresponding to the effective interaction @xmath249 : @xmath250 ( the prefactor @xmath251 reflects the fact that @xmath252 becomes negligibly small ; we emphasize that eq .
( [ boltzmann ] ) applies _ only within _ the surface film . )
we note that this assumption can break down , especially in the immediate vicinity of the reaction site , if the reaction rate is high .
in such a case one has to explicitly consider the coupled equations of mass transport for the solvent and solute within the surface film ( see ref . ) ; as already mentioned before , we consider here only situations in which eq .
( [ boltzmann ] ) holds . combining eqs .
( [ y_inner_stokes ] ) and ( [ boltzmann ] ) , integrating with respect to @xmath64 , and using that the inner and outer solutions for the hydrodynamic flow should match smoothly at @xmath78 so that at @xmath78 the pressure reaches its `` outer '' solution value @xmath253 , one obtains the following expression for the pressure field within the surface film : @xmath254\,.\ ] ] this is the so - called `` osmotic equilibrium '' condition ( because on the right hand side the term after the plus sign has the form known as `` osmotic pressure '' ) along the surface normal , which varies along the surface of the particle due to the gradient in the solute density @xcite .
note that eq .
( [ inner_pressure ] ) also applies _ only within _ the surface film . by : ( i ) combining eqs .
( [ x_inner_stokes ] ) , ( [ boltzmann ] ) , and ( [ inner_pressure ] ) , ( ii ) noting that @xmath255 can be neglected because @xmath256 ( this is because the outer solution satisfies the force - free ( @xmath257 ) version of eq . ( [ stokes_complete ] ) ; by evaluating it at @xmath78 and along the surface of the particle and by noting that the slip velocity varies over the macroscopic length scale @xmath0 , the above conclusion follows ) while with @xmath258 and @xmath259 one finds from eq .
( [ x_inner_stokes ] ) that @xmath260 , ( iii ) introducing @xmath261 ( iv ) integrating once , and ( v ) using @xmath262 ( which holds because the inhomogeneity causing the flow vanishes at the outer edge of the surface film ) , one obtains @xmath263 integrating by parts , using the boundary condition @xmath264 of no - slip at the particle surface , and noting that @xmath265 for @xmath266 ( because @xmath267 is bounded with respect to @xmath268 , so that @xmath269 is finite for all @xmath268 ) and @xmath270 [ by definition of @xmath271 , one obtains the phoretic slip - velocity @xmath272 as @xmath273 since @xmath267 decays rapidly for @xmath274 the integral in eq .
( [ slip_vel_hatx ] ) can be extended to infinity without causing a significant error : @xmath275 \hspace*{.1in}\nonumber\\ & = & - \mathbf{e}_\theta \ , b \
, \dfrac{d \rho(\hat x = r_+ \theta , r_+)}{d \hat x}\,.\end{aligned}\ ] ] the derivation of the expression for the phoretic slip - velocity is concluded by noting that @xmath276 corresponds to the gradient of the solute number density along the particle surface [ see eqs .
( [ slip_vel])-([lambda_def ] ) ] .
a general solution for the three - dimensional steady - state force - free stokes equations , [ stokes_no_force ] @xmath277 @xmath278 in spherical coordinates has been obtained by lamb @xcite . here
we briefly present the derivation , adapted to the present case which has an additional azimuthal symmetry , and we also compute the hydrodynamic force in the case of a flow subject to the boundary conditions given by eq .
( [ bc_flow ] ) .
the general solution is written as @xmath279 , where @xmath280 is the solution of the homogeneous equations , i.e. , @xmath281 ( so that the components of @xmath280 are harmonic functions , i.e. , they obey the laplace equation ) and @xmath282 , while @xmath283 is a particular solution of eqs .
( [ st_flow ] ) and ( [ div_flow ] ) .
the calculation of @xmath42 proceeds by separately determining the components @xmath280 and @xmath283 defined above . the construction of the component @xmath284 [ a particular solution of the inhomogeneous eq .
( [ st_flow ] ) ] starts from the observation that the pressure field is also a harmonic function , i.e. , @xmath285 [ this follows from taking the divergence of eq .
( [ st_flow ] ) ] .
therefore , it can be expanded in terms of the solid harmonics @xcite , which are the eigenfunctions of the laplace operator in _
3d _ : @xmath286 where @xmath287 with @xmath288 and @xmath289 is the legendre polynomial of degree @xmath290 . in this representation ,
a particular solution @xmath283 is given by @xcite @xmath291\,\ ] ] where @xmath292 this can be checked by inserting eqs .
( [ part_sol ] ) and ( [ coeff_a_b ] ) into eqs .
( [ stokes_no_force]a , b ) and by noting that @xmath293 and @xmath294 .
note that the pole in @xmath295 at @xmath296 is cancelled in eq .
( [ part_sol ] ) .
the construction of the solution @xmath280 starts from the identity @xmath297 expressing the vorticity @xmath298 as @xmath299 [ this is possible because @xmath300 and denoting the product @xmath301 by @xmath302 , and noting that : ( i ) @xmath303 ; ( ii ) @xmath304 ; ( iii ) @xmath305 ; ( iv ) @xmath306 , eq .
( [ vec_ident ] ) can be re - written as @xmath307 due to @xmath308 one has @xmath309 so that @xmath310 is harmonic , and due to @xmath311 also @xmath302 is a harmonic function . by inserting the series expansions in solid harmonics
[ compare eq .
( [ press_expansion ] ) ] of @xmath280 , @xmath310 , and @xmath302 into eq .
( [ hom_sol_form ] ) and by using @xmath294 , one finds that the solution @xmath280 is given by @xmath312\,,\ ] ] where @xmath313 and @xmath314 are the corresponding expansion coefficients of @xmath310 and @xmath302 as in eq .
( [ press_expansion ] ) .
the sum of eqs .
( [ part_sol ] ) and ( [ hom_sol ] ) provides the general solution @xmath315 expressed as a series expansion in spherical harmonics @xcite with the coefficients @xmath316 determined by boundary conditions _ on spherical surfaces_. note that for a problem without azimuthal symmetry the derivation of the solution proceeds analogously and the only change is that the legendre polynomials @xmath317 ( @xmath318 ) are replaced everywhere by a linear combination ( with respect to the index @xmath319 ) of the spherical harmonics @xmath320 .
the general solution derived in the previous subsection allows one to determine the hydrodynamic flow obeying eqs .
( [ stokes_no_force]a , b ) in a spherical domain with prescribed velocity on the boundaries of the domain ( or at infinity ) .
this is carried out in the usual manner by expanding the prescribed surface fields in terms of legendre polynomials and then equating them with the series representation of the general solution evaluated at the boundary to determine the expansion coefficients @xmath321 .
however , the prescribed boundary conditions for velocity fields can be used to obtain an equivalent set of boundary conditions , which exploits the simplicity of the derivatives of the solid harmonics with respect to the radial coordinate and thus significantly simplifies the algebra @xcite ; here we follow this approach , and discuss the most general case , i.e. , without azimuthal symmetry .
let @xmath322 denote the prescribed velocity field on a spherical boundary @xmath323 , i.e. , @xmath324 ; in our case @xmath325 for @xmath326 and @xmath327 for @xmath328 . * ( i ) * by multiplying @xmath329 with @xmath94 , one obtains the radial component @xmath330 of the flow field @xmath331 where @xmath332 denotes the radial component of the velocity @xmath322 prescribed at the spherical boundary . *
( ii ) * the radial component of the vorticity is given by @xmath333\ , , \label{vort_radial}\ ] ] where @xmath334 and @xmath335 are the polar and azimuthal components of @xmath42 , respectively .
on the spherical boundary @xmath323 , within the right hand side of eq .
( [ vort_radial ] ) @xmath335 and @xmath334 can be replaced by @xmath336 and @xmath337 , respectively .
thus the radial component of the vorticity obeys @xmath338|_{r = c } & = & \mathbf{r } \cdot ( \nabla \times \mathbf{u } ) \nonumber\\ & = & \dfrac{1}{\sin \theta}\ , [ \partial_\theta ( \sin\theta \ , u_\phi ) - \partial_\phi u_\theta ] \,.\hspace*{.3in}\end{aligned}\ ] ] * ( iii ) * since @xmath339 , one has @xmath340\ , .
\label{div_modify}\end{aligned}\ ] ] since @xmath341 , @xmath342 , and @xmath343 , eq . ( [ div_modify ] ) renders on @xmath323 the boundary condition @xmath344|_{r = c } = - r ( \nabla \mathbf{u})\,.\ ] ] determining @xmath330 , @xmath345 , and @xmath346 from the solution @xmath347 [ eqs . ( [ part_sol ] ) and ( [ hom_sol ] ) ] , evaluating the expressions at @xmath348 , and equating them to the rhs of eqs .
( [ v_radial_bc ] ) , ( [ vort_radial_bc ] ) , and ( [ div_bc ] ) , respectively , after expanding the latter in terms of spherical harmonics , yields the equations of condition for the coefficients @xmath349 .
applying this procedure for all boundaries , combining all the resultant conditions , and solving for the coefficients @xmath321 yields the velocity @xmath42 in terms of @xmath350 .
we now apply the general results discussed in the previous subsections to the particular confined system shown in fig .
[ fig1 ] , which exhibits azimuthal symmetry and obeys the boundary conditions in eq .
( [ bc_flow ] ) on the spherical surfaces @xmath351 and @xmath3 , respectively . using the expansion of @xmath352 in terms of solid harmonics [ eqs .
( [ part_sol ] ) and ( [ hom_sol ] ) ] and exploiting the azimuthal symmetry ( i.e. , the solution does not depend on @xmath69 ) , splitting this series into two , corresponding to @xmath318 and @xmath353 , respectively , changing in the latter the index of summation according to @xmath354 , and then renaming @xmath355 by @xmath290 , one obtains @xcite [ lhs_expansions ] @xmath356 p_\ell(\cos\theta)\,,\end{aligned}\ ] ] @xmath357 p_\ell(\cos\theta)\,,\ ] ] @xmath358 p_\ell(\cos\theta)\,.\nonumber\\\end{aligned}\ ] ] equations ( [ rad_vel_leg_pol ] ) , ( [ rad_vort_leg_pol ] ) , and ( [ rad_div_leg_pol ] ) relate to eqs .
( [ v_radial_bc ] ) , ( [ vort_radial_bc ] ) , and ( [ div_bc ] ) , respectively .
the rhs of eqs .
( [ rad_vel_leg_pol ] ) and ( [ rad_div_leg_pol ] ) do not depend on the coefficients @xmath359 because the contributions @xmath360 in eq .
( [ hom_sol ] ) have no radial component .
similarly , the rhs of eq .
( [ rad_vort_leg_pol ] ) does not depend on @xmath361 and @xmath362 because @xmath363 [ eq . ( [ part_sol ] ) ] and @xmath364 [ eq . ( [ hom_sol ] ) ] , respectively . by noting that @xmath365 , @xmath366 [ because the particle velocity @xmath38 depends only on @xmath72 and @xmath80 , see eq .
( [ velocity ] ) ] , and @xmath367 [ because the slip velocity is given by a gradient , see eq .
( [ slip_vel ] ) ] , one obtains that on the spherical surface @xmath351 one has [ see eq .
( [ v_radial_bc ] ) ] @xmath368 [ because @xmath38 and @xmath369 , see eq .
( [ slip_vel_hatx ] ) ] , @xmath370 = \mathbf{r } \cdot [ \nabla \times ( \mathbf{v}+\mathbf{v}_s ) ] = 0 $ ] [ see eq .
( [ vort_radial_bc ] ) ] , and [ see eq .
( [ div_bc ] ) ] @xmath371 where [ due to @xmath372 @xmath373 @xmath23 denotes @xmath374 , and @xmath375 ( the last equality follows upon integrating by parts . ) by evaluating the right hand sides of eqs .
( [ rad_vel_leg_pol ] ) , ( [ rad_vort_leg_pol ] ) , and ( [ rad_div_leg_pol ] ) at @xmath0 and @xmath52 and equating them on the left hand sides with @xmath376 and with 0 , 0 , 0 , respectively , the coefficients @xmath321 are obtained for any @xmath318 as the solution of the system of equations given by [ system_coef ] @xmath377 @xmath378 @xmath379 @xmath380 @xmath381 @xmath382 note that eqs .
( [ rad_curl_r ] ) and ( [ rad_curl_r1 ] ) imply that @xmath383 for all @xmath290 .
therefore , for any given @xmath290 one is left with a linear system of four equations for four unknowns : @xmath384 , which ( in the generic case ) admits a unique solution .
taking @xmath385 in eq .
( [ system_coef ] ) , one obtains a system of four linear equations for the unknowns @xmath386 , which depend parametrically on @xmath387 , and @xmath2 [ see eq .
( [ coeff_a_b ] ) ] .
the absence of a body force acting on the particle implies @xmath388 [ see eqs .
( [ hydro_force ] ) - ( [ force_p2 ] ) ] .
this leads to : @xmath389\,.\ ] ] by noting that @xmath390 one obtains the expression eq .
( [ velocity ] ) in the main text . because @xmath188 is positive for the system depicted in fig .
[ fig1 ] the second equality in eq .
( [ tilde_k1 ] ) implies that the sign of @xmath391 , and therefore that of @xmath43 [ see eq .
( [ v_k1 ] ) ] , is opposite to the one of the effective mobility @xmath91 . before concluding this appendix
, we note that the general solution for the hydrodynamic flow [ eqs .
( [ part_sol ] ) , ( [ hom_sol ] ) , and ( [ system_coef ] ) ] allows one to gain insight into the qualitative differences between the phoretic motion mechanism and the one of an `` osmotic pressure '' propeller @xcite .
normalized velocity fields @xmath392 in the laboratory frame corresponding to the diffusiophoretic motion due to steric repulsion only ( @xmath190 ) between the solute and the particle for the case of ( a ) an unbounded system and ( c ) a confined system ( @xmath393 ) . the panels ( b ) and
( d ) show the corresponding results in the case that the motion occurs as a result of a drag provided by the `` osmotic pressure '' integrated along the surface of the particle . the dashed circles in ( a ) and ( b ) indicate the position at which the confining wall is placed in ( c ) and ( d ) .
in all cases ( a ) - ( d ) the particle moves upwards , i.e. , its velocity @xmath88 is in the positive @xmath15-direction ( as shown by the vector on the particle ) .
here we consider only time scales on which the movement of the particle does not change the flow fields . ] in figs .
[ fig4](a ) and ( c ) we show the normalized velocity fields @xmath392 corresponding to diffusiophoretic motion due to steric repulsion only . in this case @xmath190
( [ lambda_def ] ) ] so that the particle velocity @xmath38 is oriented along the positive @xmath15-direction , i.e. , away from the source .
figure [ fig4](a ) corresponds to an unbounded system whereas fig .
[ fig4](c ) depicts the corresponding confined system with @xmath393 ; both cases refer to the same ( chosen ) constant surface gradient @xmath394 of the solute number density .
( being a constant , this gradient scales out and is absorbed in the velocity scale . therefore the dimensionless ratio @xmath392 is independent of it . ) since @xmath91 enters the problem only via the multiplication of @xmath395 [ see eq .
( [ tilde_k1 ] ) ] , also the absolute value of @xmath91 drops out from the normalized flow field .
the viscosity @xmath2 drops out , too .
for the unbounded case , the solution [ see also eq .
( 35 ) in ref . ]
is determined from eqs .
( [ rad_vel_r ] ) , ( [ rad_curl_r ] ) , and ( [ rad_div_r ] ) and the requirement that the flow field vanishes at infinity .
this implies that in eq .
( [ lhs_expansions ] ) all coefficients multiplying terms @xmath396 with @xmath318 are zero . in the confined case ,
the flow field is approximated by keeping terms up to @xmath397 [ eqs .
( [ system_coef ] ) ] in the series expansion of eqs .
( [ part_sol ] ) and ( [ hom_sol ] ) ; this provides a reasonably good approximation , except near the boundary at @xmath52 and for polar angles close to zero or @xmath398 where apparently there are deviations from the no - slip boundary condition .
however , one should keep in mind that in this normalized representation of the flow @xmath392 small numerical deviations from @xmath399 at @xmath52 are sharply overemphasized . for comparison , in figs . [ fig4](b ) and ( d ) we present the corresponding results for the case in which the motion of the particle occurs as a result of a drag provided by an `` osmotic pressure '' integrated over the surface of the particle . in this case
the particle velocity is oriented along the positive @xmath15-direction , too , because of the assumption that the pressure is proportional to the solute number density and thus is larger at the lower hemisphere . in this case
( [ system_coef ] ) is solved under the assumption of a no - slip boundary condition at @xmath0 , i.e. , all the coefficients @xmath400 are set to zero ; this allows one to obtain a solution in closed form in both the unbounded and the confined cases .
note that in this case @xmath401 is non - zero .
it is fixed by the given drag [ integrated `` osmotic pressure '' , eq .
( [ force_p2 ] ) ] , and thus eq . ( [ system_coef ] ) , evaluated at @xmath402 , leads to a relation between the velocity @xmath43 and the drag force .
( in the unbounded case , this relation is the well - known stokes formula for the viscous friction on a sphere . ) in eq .
( [ rad_vel_r ] ) @xmath43 sets the velocity scale , which drops out ( as well as the drag force , because @xmath43 is proportional to it ) of the dimensionless ratio @xmath392 .
qualitative differences between these two mechanisms are particularly clearly visible in the case of the unbounded systems .
for example , the flow along the @xmath15-axis is in the negative direction ( i.e. , opposite to the motion of the particle ) in the phoretic case , while it is in the positive direction for the `` osmotic '' propeller ; the flow lines for the phoretic motion are closing back and then `` slide '' along the surface of the particle . since closed - form
analytical expressions are available in these unbounded cases @xcite , the qualitative differences noticed above can be correlated with the different characteristics of the flows . in the case of phoretic motion
the flow field decays as @xmath112 at large distances @xmath206 from the particle , which is a `` multipole '' far - field corresponding to the action of a force - quadrupole disturbance , in contrast to the @xmath113 decay in the osmotic case , which is a `` monopole '' far - field corresponding to the action of a point - force ( `` stokeslet '' ) disturbance
. the faster decay of the velocity field in the first case means that a boundary at @xmath52 will perturb the phoretic flow much less .
this explains why for the phoretic motion the viscous friction increases with the confinement weaker than what would be expected from naively using a stokes - friction concept .
the flow fields remain qualitatively different even in the presence of the confinement ( in particular with respect to the opposite direction of flow along the @xmath15-axis ) , although the differences are not as pronounced as in the unbounded case .
a further qualitative difference between the two confined flows is that in ( d ) a vortex structure is formed fully detached from the particle whereas in ( c ) the vortex involves flow along the surface of the particle . due to the presence of the boundary at @xmath52 , all the terms @xmath396 in the expansions of eq .
( [ lhs_expansions ] ) are present in the solution and an analysis of the flow in terms of fundamental solutions , as for the unbounded case , is no longer possible .
note that since @xmath308 and @xmath403 [ see the text following eqs .
( [ vec_ident ] ) and ( [ rad_div_r1 ] ) ] one has @xmath404 so that with eqs .
( [ part_sol ] ) and ( [ coeff_a_b ] ) one obtains @xmath405 which is in general nonzero [ e.g. , for osmotic " flow @xmath406 ( eq .
( [ force_p2 ] ) ) ] . thus in agreement with fig .
[ fig4 ] the flow field can contain vortices .
equation ( [ diff_eq ] ) , subject of the ic and bc conditions given in eq .
( [ bc_ic ] ) is solved by using the laplace transform and the inversion theorem @xcite : @xmath407 = \int\limits_0^\infty dt \,e^{-\zeta t } f(t ) \label{lt}\ ] ] and @xmath408:= \frac{1}{2\pi i } \int\limits_{\gamma - i \infty}^{\gamma + i \infty } d\zeta \,e^{\zeta t}\bar f(\zeta ) , \label{ilt}\ ] ] respectively , where @xmath409 ( the laplace transformed quantities are indicated by an overbar ) is assumed to be well - defined for @xmath410 and @xmath411 is sufficiently large such that all singularities of @xmath409 lie to the left of the integration path . by taking the laplace transform of eqs .
( [ diff_eq ] ) and ( [ bc_ic ] ) and by using the ic condition of zero number density of the product molecules , one finds that the laplace transformed density is given by @xmath412 where @xmath413 and @xmath414 is the green s function for the helmholtz operator @xmath415 , @xmath416 , satisfying the bcs of vanishing normal derivative at @xmath417 .
decomposing @xmath414 as @xmath418 , where the singular part @xmath419 is the free space green s function for the helmholtz operator ( which is known in any spatial dimension @xmath420 , see , e.g. , ref . ) , the initial problem is reduced to that of finding the solution @xmath421 of the homogeneous helmholtz equation subject to the boundary conditions @xmath422 in _
3d _ the singular part of the green s function is given by @xcite @xmath423 while the regular part ( i.e. , the general solution of the homogeneous helmholtz equation ) can be written as @xmath424 p_\ell(\cos\theta),\ ] ] where @xmath425 and @xmath426 are the modified spherical bessel functions of the first and third kind , respectively @xcite , @xmath427 is the legendre polynomial of degree @xmath290 , while the coefficients @xmath428 and @xmath429 will be fixed to fulfill the boundary conditions . using one of the addition theorems for the bessel functions @xcite and noting that the angle between @xmath430 and @xmath431 is @xmath108 ( see fig . [ fig1 ]
) , @xmath419 can be re - written as @xmath432 where @xmath433 , @xmath434 , and @xmath435 . by combining eqs .
( [ g_3d ] , [ gs_3d_expand ] ) and the boundary conditions [ eq .
( [ bc_for_g ] ) ] , by noting that @xmath436 ( i.e. , the source of particles is _ on _ the surface of the particle ) , and by re - writing the coefficients as @xmath437 and @xmath438 , the dimensionless coefficients @xmath439 and @xmath440 are determined by the solution of the following closed system of two linear equations with two unknowns ( @xmath441 and @xmath442 ) : where @xmath444 and @xmath445/[d i_{\ell+1/2}(z)/dz]$ ] . the solution @xmath446 and @xmath447 acquires a dependence on @xmath80 via eq .
( [ bc_at_r1 ] ) : @xmath448 and @xmath449 , with @xmath450 .
this determines the green s function @xmath414 and thus the laplace transformed density @xmath451 [ eq .
( [ lt_dens ] ) ] from which , in principle , the density @xmath452 is obtained via the inverse laplace transformation .
note that although the final result is in the form of a series , the quantities in which we are interested , in particular the phoretic velocity @xmath151 [ eq . [ velocity ] ] , involve integrals over the polar angle @xmath453 of the product between this series and a specific @xmath290 legendre polynomial @xmath289 [ @xmath385 in the case of the velocity , @xmath454 in the case of the total number of product molecules ( see below ) ] and thus only one of the terms from the series will contribute .
we make the following three remarks : + * * ( * * i ) since for @xmath455 one has @xmath456 , in the limit @xmath125 one finds @xmath457 and eq .
( [ alpa_3d ] ) reduces to the corresponding bc in ref . ,
i.e. , as expected the solution for the unbounded case is recovered . + * * ( * * ii ) equations ( [ bc_at_r ] ) and ( [ bc_at_r1 ] ) will not coincide in the limit @xmath124 because even in this limit of extreme confinement a source singularity is present at @xmath458 and is picked up by @xmath459 [ eq . ( [ gs_3d_expand ] ) ] . + * * ( * * iii ) the laplace transform @xmath460 of the total number of product particles in the system at time @xmath72 is @xmath461 \bar g(q)\nonumber\\ & & = - 4 \pi \bar a(p ) c_0(q ) i_{1/2}(q r ) \nonumber\\ & & \times
\int\limits_{r}^{r_1 } dr r^2 [ \bar \alpha_0 i_{1/2}(q r ) + ( \bar \beta_0 - 1 ) k_{1/2}(q r ) ] \\ & & = d \bar a(p ) /p \rightarrow n(t ) = \int\limits_{0}^{t } dt ' a(t')\ , , \nonumber\end{aligned}\ ] ] i.e. , as expected @xmath462 is given by the time integral of the production rate , providing a welcome consistency check .
paxton , a. sen , and t.e .
mallouk , chem.eur . j. * 11 * , 6462 ( 2005 ) .
paxton , s. sundararajan , t.e .
mallouk , and a. sen , angew .
, int . ed . *
45 * , 5420 ( 2006 ) ; w.f .
paxton , k.c .
kistler , c.c .
olmeda , a. sen , s.k.st .
angelo , y. cao , t.e .
mallouk , p.e .
lammert , and v.h .
crespi , j. am .
soc . * 126 * , 13424 ( 2004 ) .
ismagilov , a. schwartz , n. bowden , and g.m .
whitesides , angew .
, int . ed .
* 41 * , 652 ( 2002 ) .
catchmark , s. subramanian , and a. sen , small * 1 * , 1 ( 2005 ) .
r. golestanian , t.b .
liverpool , and a. ajdari , phys .
lett . * 94 * , 220801 ( 2005 ) . j.r .
howse , r.a.l .
jones , a.j .
ryan , t. gough , r. vafabakhsh , and r. golestanian , phys .
* 99 * , 048102 ( 2007 ) .
g. rckner and r. kapral , phys .
lett . * 98 * , 150603 ( 2007 ) .
r. golestanian , t.b .
liverpool , and a. ajdari , new j. phys .
* 9 * , 126 ( 2007 ) .
f. peruani and l.g .
morelli , phys .
* 99 * , 010602 ( 2007 ) .
saffman and m. delbrck , proc .
usa * 72 * , 3111 ( 1975 ) ; p.g .
saffman , j. fluid mech . * 73 * , 593 ( 1976 ) . c. barentin , c. ybert , j .-
di meglio , and j .- f .
joanny , j. fluid mech . * 397 * , 331 ( 1999 ) ; c. barentin , p. muller , c. ybert , j .- f .
joanny , and j .-
m . di meglio , eur . phys .
j. e * 2 * , 153 ( 2000 ) .
anderson , ann .
fluid mech .
* 21 * , 61 ( 1989 ) .
m.k . phibbs and p.a .
gigure , can . j. chem . *
29 * , 173 ( 1951 ) .
a. einstein , `` on the movement of small particles suspended in a stationary liquid demanded by the molecular - kinetic theory of heat '' in _
investigations on the theory of the brownian motion _ , ed .
r. frth , transl . by a. d. cowper ( dover , new york , 1956 ) .
f. juelicher and j. prost , arxiv:0812.2924v1 ( 2008 ) .
s.r . de groot and p. mazur , _ non - equilibrium thermodynamics _
( north - holland , amsterdam , 1962 ) , ch .
v.2 , p. 44 .
a. ajdari and l. bocquet , phys .
lett . * 96 * , 186102 ( 2006 ) . j. happel and h. brenner , _
low reynolds number hydrodynamics _ ( noordhoff international , leyden , 1973 ) , ch . 3 - 2 , pp .
h. lamb , _ hydrodynamics _
( dover , new york , 1945 ) , p. 594 . a.l .
zydney , j. colloid interface sci . *
169 * , 476 ( 1995 ) .
f.a . morrison jr . , j. colloid interface sci . *
34 * , 210 ( 1970 ) .
crdova - figueroa and j.f .
brady , phys .
lett . * 100 * , 158303 ( 2008 ) .
n. bala saidulu and k. l. sebastian , j. chem . phys . *
128 * , 074708 ( 2008 ) .
a.t . chwang and t.y .- t .
wu , j. fluid mech . *
67 * , 787 ( 1975 ) .
m. abramowitz and i.a .
stegun , _ handbook of mathematical functions with formulas , graphs , and mathematical tables _ ( dover , new york , 1965 ) pp .
374 , 443 .
carslaw and j.c .
jaeger , _ conduction of heat in solids _ ( university press , oxford , 1959 ) , pp .
377 , 381 .
p. tierno , r. golestanian , i. pagonabarraga , and f. sagus , phys .
* 101 * , 218304 ( 2008 ) .
s. hassani , _ mathematical physics : a modern introduction to its foundations _
( springer , new york , 1998 ) p. 631 | we study theoretically the effects of spatial confinement on the phoretic motion of a dissolved particle driven by composition gradients generated by chemical reactions of its solvent , which are active only on certain parts of the particle surface .
we show that the presence of confining walls increases in a similar way both the composition gradients and the viscous friction , and the overall result of these competing effects is an increase in the phoretic velocity of the particle . for the case of steric repulsion only between the particle and the product molecules of the chemical reactions ,
the absolute value of the velocity remains nonetheless rather small . |
the appearance of filaments on the solar disk is associated with the distribution of large - scale magnetic fields and magnetic fields of active regions ( ar ) .
the filaments are observed virtually at all latitudes in the solar atmosphere .
systematic @xmath0 observations have been carried out approximately since 1915 at the kodaikanal observatory ( india ) , since 1919 at the meudon observatory ( france ) ( spectroscopic observations ) , and since 1959 at the kislovodsk mountain astronomical station of the main ( pulkovo ) astronomical observatory of russian academy of sciences .
the filaments observed in the @xmath0 line in the solar chromospheres are one of the basic indices of solar activity ( see the review by @xcite ) .
solar filaments arise in the vicinity of the magnetic field polarity inversion line ( @xcite , @xcite , @xcite ) . as a rule ,
they are divided into several types ( @xcite , @xcite , @xcite ) : ar filaments , quiescent filaments , and polar filaments ( @xcite ) .
the former are usually short ( about 10 mm ) , have small lifetimes ( hours to days ) , and are located quite low ( @xmath7 mm ) .
the filaments outside the active regions may be hundreds of megameters long , have lifetimes up to a few rotations , and are located at a height of about 100 mm ( @xcite ) . while the basic properties of solar filaments have long been known , filament channels and filaments have been classified more recently in terms of their chirality ( @xcite ) .
they classified filament channels as either dextra or inistral depending on the direction of the axial component of the field as seen by an observer standing on the positive - polarity side of the channel . with respect to the polarity inversion line
, the filaments are also divided into several groups .
e. g. , they can be located within one or several bipolar active regions ( @xcite ) or between diffuse magnetic fields of different polarity ( @xcite ) . despite a century of observations and investigation of the solar filaments the physical mechanisms of their formation
are not yet fully understood ( see the review by @xcite ) .
however , the role of solar filaments in the occurrence of solar flares , coronal mass ejections , and geomagnetic disturbances is extremely high ( @xcite ) .
the objective of this work is to study long - term variations in the filament characteristics by analyzing digitized @xmath0 synoptic maps for the period 1919 - 2014 .
we have mainly focused on the change of the filament tilt angles relative to the heliographic coordinates .
we have used two types of the source data .
the first type is meudon synoptic maps . in the period 1919 - 1989 ,
the meudon observatory published synoptic maps with plotted long - lived eruptive filaments , centers of activity , and faculae ( @xcite ) .
the data are partly processed and are available in tabular form on the web site ( * @xmath8 * ) .
the tables contain information on the coordinates , time , and other characteristics of the filaments for two time intervals ( 1919 - 1956 and 1957 - 2000 ) with the addition of data from other observatories .
later , the procedure has been changed , and since 1990 the maps have been digitized and plotted .
they are provided in digital format ( @xcite ) . in this work ,
we have used the original synoptic maps .
thus , the maps for the period 1919 - 1989 ( rotations 876 - 1850 ) have been scanned from atlas ( @xcite ) .
the maps in the new format for the synoptic rotations 1824 - 1850 ( 1989 - 1991 ) and rotations 1931 - 2008 ( 1998 - 2003 ) are available on the web site ( @xmath9 ) .
the second group of the source data comprises the maps of kislovodsk mountain station for the period 1979 - 2014 .
they are based on daily chromospheric observations with the use of an @xmath0 filter with the bandwidth of @xmath10 .
the data are regularly processed , and daily maps of solar activity are provided .
then , the positions of the filaments are marked on synoptic charts .
unlike the meudon synoptic maps , the kislovodsk ones give the polarity inversion lines of the large - scale magnetic field ( @xcite ) . during 1979 - 1995 ,
the @xmath0 synoptic charts were published in the solnechnye dannye ( solar data ) bulletin , and thereafter , it has been carried on in the electronic version of the bulletin on the web site @xmath11 . at present , the scanned maps are available for the synoptic rotations 1677 - 2155 .
besides the synoptic chart , there are tables that contain characteristics of solar filaments since 1959 till the present ( * @xmath12 * ) . for the purposes like , for example
, studies of some characteristics with the use of other types of observational data or determination of the filament geometric parameters , the filament images need to be vectorized .
the kislovodsk mountain astronomical station provides data on solar filaments both in the form of daily images and @xmath0 synoptic charts . since 2004 , the procedures of identification of filaments , determination of their characteristics , and plotting them on synoptic charts are automated ( @xcite ) .
other algorithms for automatic filament detection have been developed and implemented into various software and data products by e.g. , ( @xcite ; @xcite ; @xcite ; @xcite ) .
figure [ fig1 ] shows an example of the meudon and kislovodsk synoptic charts .
the original maps represent solar filaments schematically showing only their median line and provisional width .
special software was developed to digitize the filaments .
the procedure is as follows .
first , the synoptic chart boundaries and coordinate grid are applied to the scanned image .
then , the filaments are vectorized . in this process
, the operator draws the median line along the filament and chooses an appropriate width in semi - automatic mode .
the following parameters were defined for each filament : mean coordinates of the filament and coordinates of its tips , length , latitudinal and longitudinal dimensions , tilt angle and polarity of magnetic field .
the total number of digitized filaments is 81013 from meudon data for the period 1919 - 2003 and 17787 from kislovodsk data .
the database comprises both the filaments associated with active region magnetic dipoles and quiescent filaments arising between diffuse magnetic configurations
. figure [ fig2 ] represents schematically the positions and tilt angles of the filaments of various types . for further analysis
, we constructed a combined series comprising the data from meudon synoptic charts for the period 1919 - 1985 ( rotations 876 - 1769 ) and from kislovodsk synoptic charts for the period after 1985 ( rotations 1770 - 2155 ) .
altogether , the combined series included the data for the period 1919 - 2014 on 74572 filaments .
figure [ fig3]a illustrates variations in the number of filaments on synoptic charts .
one can clearly identify an 11-year mode , which reflects variation of solar activity . on the other hand ,
the average number of filaments has been increasing during 1919 - 2005 .
this may be due to the changes in the methods and quality of observations
. a more adequate characteristic can be the total length of filaments on the synoptic chart represented in figure [ fig3]b .
according to the meudon data , the total length of filaments on the synoptic maps before 1988 has been more stable .
this is probably due to the fact that the new methods adopted in 1988 allowed one seeing weak filaments and filament channels .
the kislovodsk and meudon maps for cycles 21 and 22 appear more similar as concerned with the length of the filaments rather than their number , probably , because the filaments on meudon maps are plotted as a chain of individual feature lines , while on kislovodsk maps they are represented by solid lines .
figure [ fig4 ] represents the distribution of the lengths of filaments of the 1919 - 2014 composite series .
the distribution of the relative number of filaments versus the logarithm of their length is close to the lognormal distribution with a peak at about 70 mm .
filaments can be efficiently used as an activity index when analyzing long - term variations of solar activity .
solar filaments are present virtually everywhere from the equator to the highest latitudes .
figure [ fig5 ] shows the latitude - time diagram of the centers of digitized solar filaments .
there are two groups of filaments : filaments associated with the polar drift of the large - scale magnetic field and those associated with the distribution of sunspots .
polar filaments appear on the polarity inversion lines of the large - scale magnetic field in high latitudes . at the minimum and the rise phases of sunspot activity cycle , the filament centers drift towards the poles , ceasing their motion in the epoch of the cycle maximum .
there have been several branches of the poleward drift in some solar cycles , e.g. , cycles 18 , 19 , and 20 in the northern hemisphere and in cycles 18 and 21 in the southern hemisphere . in this case , one can see the so - called `` triple reversal '' of the large - scale magnetic field polarity ( @xcite ; @xcite ) .
the drift velocity depends on the height of the activity cycle ( @xcite , see their table 3 for details ) .
the other group of filaments is associated with sunspot activity .
these filaments appear in the mid - latitudes and equatorial zones following the butterfly diagram of sunspot distribution .
the number of the filaments is maximum at latitudes of about @xmath13 ( figure [ fig5 ] ) which is somewhat higher than for the maximum of sunspot distribution .
contrary to the latitudinal distribution of sunspots , a lot of filaments appear in immediate proximity to the equator .
likewise the number filament , their average length is getting higher in low and middle latitudes .
although the total length of filaments decreases to high latitudes , their longitudinal extent increases with latitude from the average of about @xmath14 near the equator to about @xmath15 near the pole .
the longitudinal extent of filaments in the southern hemisphere in the period 1919 - 2014 was somewhat longer than on the northern one .
in the context of different aspects of the solar dynamo , an important characteristic is the angle ( @xmath1 ) of the north - south tilt of the active region magnetic dipoles .
it is believed that the tilt angle for sunspots and bipolar groups determines the efficiency of conversion of toroidal fields into the poloidal magnetic field ( @xcite ) .
solar filaments usually lie along the magnetic polarity inversion lines .
measurements of the tilt angles may help us understand the mechanism of generation of the solar magnetic field . to find the tilt angle , we draw a straight line through each filament approximating it by the least square method , and measured its inclination .
the filaments whose eastern tips are closer to the poles than the western ones were given positive tilts , and negative if vice versa .
figure [ fig6 ] represents the relative distribution of the tilt angle @xmath1 of solar filaments calculated with respect to their length as a weighting factor .
the distribution is strongly asymmetric about @xmath16 .
although most filaments have the tilt close to @xmath17 , the mean tilt angle is positive .
the total number of filaments with negative tilts is 24316 , which is about a half of the number of the filaments with positive tilt angles .
16700 filaments of the latter group arise in the sunspot formation zone , i.e. , at latitudes below @xmath18 .
this is quite remarkable . for the filaments which are located between the isolated bipolar groups of plages and active regions
, we would expect their tilt angles to be negative .
for the filaments associated with sunspots , the tilt angles must be close to normal to the magnetic axis of the dipoles ( see the schemes in ( @xcite , @xcite ) .
it is likely that the filaments plotted on synoptic maps mainly originate at the boundaries of large - scale diffuse magnetic configurations arising in the process of the decay of active regions ( @xcite ) . the mean tilt angles changes with the phase of the cycle ( figure [ fig7 ] ) being the highest in the epoch of solar maximum .
we may note that the smoothed tilt angle values also could vary with the secular ( gleissberg ) cycle attaining their maximum value in the middle of 20th century with the only exception of the present cycle 24 .
figure [ fig7 ] also shows the cycle - mean tilt angles with activity cycles
. it also shows the cycle - mean tilt angles of filaments averaged over the period from @xmath19 to @xmath20 for different cycles .
the maximum tilt angle has been recorded in cycle 19 . in the prolonged minimum solar activity of cycles 23/24
the number of observed filaments was really small ( see figure 5 ) .
this may be a possible reason why the average tilt turned to negative values then .
this issue is worth further investigation .
there is a positive correlation between the amplitude of the activity cycle and the tilts of filaments .
the correlation coefficient of linear regression for cycles 15 - 23 is @xmath21 .
the tilt angles of filaments vary with their length increasing from about @xmath22 to @xmath23 for relatively shorter filaments .
the maximum positive tilt angles correspond to the length of @xmath24 mm .
for further longer filaments the tilt angles are lower while there are fewer of them and for long filaments our definition of tilt by approximation of a straight line may not be that robust .
the latitudinal distributions of the relative number of filaments with positive @xmath25 and negative @xmath26 tilts differ from each other ( figure [ fig8 ] ) .
the number of filaments with positive tilt angles is maximum at the latitudes of about @xmath27 , while the distribution of filaments with negative tilts number peaks at the vicinity of the equator and at latitudes about @xmath28 .
the tilt angles of filaments also depend on latitude .
figure [ fig9 ] illustrates the latitudinal distribution of the tilt angles separately for the filaments with negative and positive tilts .
the absolute value of the tilt angle is maximum in the low - latitude zone and decreases with absolute latitude . since the total number of filaments with positive tilt is larger than the number of those with negative tilts , the distribution demonstrates on average mainly positive tilts with exception for the polar regions ( figure [ fig9]c ) .
the increase of the tilt angles in absolute value for the filaments with @xmath26 ( figure [ fig9]b ) can probably be explained by the joy law for sunspot dipoles .
the tilt angles of sunspots increase with latitude but noting that filaments are located on the polarity inversion lines , one may expect them to be normal to the magnetic axis of the dipoles ( cf .
other works on filaments : @xcite ; @xcite ) . for the filaments with positive tilts ,
the tilt angles have maximum of about @xmath13 at latitudes of about @xmath29 and decrease to the polar regions .
figure [ fig10 ] represents the time - latitude distribution of the tilt angles .
the red areas correspond to positive tilts and the blue ones to the negative tilts . in the mid - latitude zone ,
the tilt is mainly positive . at high latitudes ,
it is mainly negative , especially for the filaments associated with the polar drift ( figure [ fig5 ] ) .
in this work , we have analyzed the solar filaments recorded on the meudon and kislovodsk @xmath0 synoptic charts for about 100 year long period .
we have found that different types of filaments as mentined in introduction ( figure [ fig2 ] ) have different characteristic tilt angles @xmath1 .
variations in the tilt angles in different latitudinal zones for low- and mid - latitude filaments ( in the range @xmath30 ) and polar filaments ( @xmath3 ) in different latitudinal zones during cycles 15 - 23 are shown in figure [ fig11 ] .
the number of polar filaments is 10309 or about @xmath31 of the total number .
the number of filaments with @xmath26 at mid latitudes ( @xmath32 ) is 18303 or @xmath33 of the total number .
low - latitude filaments have mainly positive tilt angles @xmath34 .
their maximum was attained in the middle of the 20th century .
the tilt angles of polar filaments are usually smaller and were negative in cycles 15 - 20 .
most of the filaments are located in the mid - latitude zone .
these are both the filament associated with active regions and those associated with diffuse magnetic fields .
it may be suggested ( @xcite ) that the filaments of different types have the tilt angles of different sign .
figure [ fig12 ] illustrates the time variation of filaments with negative and positive tilt angles .
the mean tilt angle of filaments with @xmath25 decreased during cycles 17 - 22 .
filaments with positive tilts @xmath25 have a long - term trend with @xmath35 variability ( figure [ fig12]a ) . the filaments with @xmath26 in the mid - latitude region are mainly associated with active regions ( see figure [ fig1 ] and the scheme in ( @xcite ) . figure [ fig12]b represents the mean tilt angles for even and odd cycles .
the tilt angles for even cycles are smaller than in the subsequent odd cycles which corresponds to the @xmath36 alternating variability in odd / even cycles according to gnevyshev - ohl rule ( @xcite ) . in the vicinity of the equator , the tilt angles are highest in absolute value ( figure [ fig9]a ) .
this suggests that the magnetic fields of opposite polarity alternate with longitude in accord with the concept of sector structure of the low - latitude large - scale magnetic field . at high latitudes ,
the tilt angles of filaments are small in accord with the concept of the zonal structure of the high - latitude field .
we used the data from @xmath0 synoptic charts obtained at kislovodsk to reconstruct the maps of the magnetic field polarities for the period from 1887 to the present day ( figure [ fig1]b ) ( @xcite ; @xcite ; @xcite ) .
we also compared the positions of filaments with the polarity of the large - scale magnetic field .
figure [ fig13 ] shows variations of tilt angles for the filaments lying on the polarity separation line of the large - scale magnetic field ( @xmath37 ) where `` + '' corresponds to the case when positive magnetic field is near to the north pole and `` - '' when the negative field is near to the south pole ) , and for those lying on the polarity separation ( @xmath38 ) .
this is in agreement with the results of ( @xcite ) .
the filaments lying on ( @xmath37 ) polarity separation have the maximum tilt angles in odd cycles , and the filaments lying on ( @xmath38 ) polarity separation , in the even cycles
. this may reflect the presence of the meridional component of the global magnetic field i.e. the mean poloidal field .
our paper deals with statistics of quantitative characteristics of solar filaments in the century - long period 1919 - 2014 .
we especially focus on analysis of the tilt angles of filaments .
we have shown that the tilts of the filaments of different kinds have different average values of tilts and have different regularities of long - term variability .
the mean tilt angles decrease in the epoch of minimum of the 11-year sunspot activity cycle and have probably positive correlation with the amplitude of the sunspot cycle ( figure [ fig7 ] ) . over the secular gleissberg cycle , the tilt angles in the mid - latitude zone @xmath18 were maximum in the middle of the 20th century in cycles 18 - 19 .
we also found latitudinal migration of tilt filaments with the solar cycle both equatorwards in low latitudes , and polewards in high latitudes .
for the filaments which are associated with solar active regions and have negative tilts , the tilt angles are larger in odd cycles than in the preceding even cycles ( figure [ fig12]b ) . for filaments lying within one the same system of the large - scale magnetic field polarity separation lines [ either ( + /-
) , or ( -/+ ) ] , there is 22 year periodicity in variation of the tilts @xmath1 ( figure [ fig13 ] ) .
this may be connected with variation of latitudinal extent of the filaments which belong to different polarity systems ( @xcite , @xcite ) , and the above described dependence of the tilt angles of filaments with their latitude .
the statistical properties of solar filaments can be used as additional coherent measure of manifestation of the solar cycle which covers all latitudes .
the data series on filaments is systematically calibrated and is available for almost a century long period .
this opens new perspectives for studies of the solar cycle and underlying hydro - magnetic dynamo mechanisms using this universal and long term systematically measured index of manifestations of the solar activity .
the authors would like to thank the anonymous referee for useful and helpful criticism and suggestions for improving the paper .
would like to acknowledge support from russian science foundation no .
15 - 12 - 20001 , k.k . would like to acknowledge partial support from the russian foundation for basic research under grant no .
13 - 02 - 01183 , and v.v .
also partial support under grants rfbr no . 15 - 02 - 03900 and 15 - 02 - 10069 . in degrees .
the data averaged over each synoptic rotation are shown as a weak grey line , and the same data smoothed by running average for 13 rotations are shown as a thick red line .
the mean tilt angles of all filaments in cycles 15 - 24 from @xmath19 to @xmath20 are shown by blue color ; their student s 90% confidence intervals are of order one degree .
cycle numbers are given below .
, width=384 ] confidence intervals . for filaments in equatorial zone error bars
are indistinguishable ( less than @xmath41 of the mean values ) .
, title="fig:",width=384 ] confidence intervals . for filaments in equatorial zone error bars
are indistinguishable ( less than @xmath41 of the mean values ) .
, title="fig:",width=384 ] ( a ) and @xmath26 ( b ) .
the student s @xmath42 confidence intervals error bars are indistinguishable ( about @xmath43 of the mean values ) . ,
title="fig:",width=384 ] ( a ) and @xmath26 ( b ) .
the student s @xmath42 confidence intervals error bars are indistinguishable ( about @xmath43 of the mean values ) . , title="fig:",width=384 ] | the spatial and temporal distributions of solar filaments were analyzed using data from the meudon observatory for the period 1919 - 2003 and the kislovodsk mountain astronomical station for the period 1979 - 2014 .
we scanned @xmath0 solar synoptic charts on which the filaments were isolated and digitized .
the data on each filament comprise its location , length , area , and other geometrical characteristics .
the temporal distributions of the number and total length of the filaments have been obtained .
we also found latitudinal migration of filament locations with the solar cycle , and analyzed the longitudinal distribution and asymmetry of filaments in the northern and southern hemispheres , and other properties of their distribution .
the tilt angles of filaments with respect the solar equator ( @xmath1 ) were analyzed .
on average , the eastern tips of filaments are closer to the poles than the western ones ( @xmath2 ) . on the other hand , the filaments in the polar regions ( @xmath3 , where @xmath4 is the latitude ) usually have negative tilts ( @xmath5 ) .
the tilt angles vary with the phases of the 11 year sunspot cycle and are at their highest values in the epoch of the activity maximum . in the century - long modulation of the solar activity ( gleissberg cycle ) ,
the mean tilt angles of filaments in the mid - latitude zone ( @xmath6 ) were maximum in the middle of the 20th century in solar sunspot cycles 18 - 19 .
we hereby propose using the statistical properties of solar filaments as an additional coherent measure of manifestation of the solar cycle which covers all latitudes and for which almost a century long systematically calibrated data series is available . |
bose - einstein condensates ( becs ) in optical lattices ( ols ) are presently attracting a great deal of interest @xcite due both to their flexibility in parameter design and to the possibility they offer to observe interesting phenomena such as superfluid to mott - insulator transition @xcite , bloch oscillations @xcite , landau - zener tunneling @xcite , generation of coherent atomic pulses ( atom laser ) @xcite , atom interferometry @xcite , etc . in this respect , ols allow to control important properties of bec by means of their periodic structure , this allowing , for example , the existence and stability of localized nonlinear excitations with chemical potentials inside band - gaps ( so called gap - solitons ( gss ) ) even in the presence of repulsive interactions ( positive scattering lengths ) .
this fact , that would be obviously impossible in absence of the ol @xcite , has been experimentally demonstrated in @xcite .
modulations of the ol can be used to accelerate , decelerate or to scatter gss as well as to control their velocities @xcite .
uniform accelerations of the ol combined with periodic modulations of the scattering length , either in space or in time , were shown to be effective tools to induce long lived bloch oscillations in the nonlinear regime @xcite , as well as band - gap tunneling phenomena such as the landau - zener tunneling @xcite .
periodic time dependent ol accelerations were also used to achieve the dynamical localization of nonlinear matter waves @xcite and the rabi - oscillations of gs states across a band - gaps @xcite which survive on a long time scale in the presence of nonlinearity .
moreover , the combination of the above phenomena permits the stirring of gss both in the reciprocal and in the direct lattice space as recently demonstrated in @xcite . besides the one - dimensional contexts in which these effects have been investigated
, ols also play an important role for stabilizing gss against collapse or decay in higher dimensions @xcite .
all what said above refers to the case of perfect ols , e.g. ols without distortions or defects which compromise the periodicity structure . in this case
the interplay between periodicity and nonlinearity is the only source for the localization of matter waves in the system .
ol defects , however , introduce additional bound states in the band - gaps ( so called impurity modes ) @xcite providing an alternate source of localization in the system .
opposite to gss , impurity modes exist both in absence and in presence of nonlinearity and can interfere in the scattering process of gss by ol defects .
scattering properties of solitons with single ( non periodic ) potential wells have been extensively investigated during the past years . for the case of the nonlinear schrdinger equation ( nlse ) scattering of solitons by extended defects
were numerically investigated in @xcite where the occurrence of a series of reflection , transmission , and trapping regions as a function of the defect strength was reported .
a similar problem has been recently investigated for solitons of the gross - pitaevskii equation ( gpe ) with rectangular potential wells and for attractive interatomic interactions @xcite .
similar studies were also done for point defects of the discrete nlse equation @xcite which corresponds , under suitable conditions , to a tight binding model of bec in a deep ol @xcite .
the resonant transmission , reflection and trapping of discrete breathers by point - defects was investigated in @xcite .
in contrast to these studies , however , the scattering of continuous gpe gss by ol defects have been scarcely investigated . in this context , we mention the numerical study performed in @xcite where the scattering properties of gss were suggested to be useful to construct quantum switches and quantum memories . to our knowledge , however , the mechanism underlying resonant transmissions of gpe gss by ol defects and the possibility of multiple defects resonant scattering have not yet been discussed . in the present paper
we provide an extensive numerical investigation of the scattering properties of gss by ol defects and identify the physical mechanism underlying the phenomenon of the resonant transmission .
in particular , we show that the presence of repeated reflections transmission and trapping regions observed for increasing strengths of an ol defect is associated to the impurity modes inside the defect potential with energies ( chemical potentials ) and numbers of atoms matching corresponding quantities of the incoming gs . as the ol defect strength
is increased , new impurity modes enter from the gap edges , moving toward the center of the gap ( bottom of the potential ) .
this implies that for incoming gss with chemical potentials very close to band edges , the number of resonances observed in the scattering coincides with the number of bound states which can exist in the defect potential for the given defect strength .
this fact is demonstrated both by studying stationary states inside the defect potential and by direct numerical integrations of the gpe . an excellent agreement between the two approaches is found , this confirming the correctness of our interpretation .
the dependence of the resonant transmission on the incoming velocity of the gs is also investigated both for fundamental gss in the semi - infinite gap and for gss in the first gap zone . as a result
we show that the transmission resonant peaks become wider as the incoming velocity is increased , with very sharp peaks at small velocities . the multiple resonant transmission through a series of ( two and three ) ol defects is also demonstrated .
we show in this case that for equally spaced identical ol defects the widths of the full transmission resonances in general decreases as the number of defects is increased .
we demonstrate , however , that the resonant transmission through a series of defects can be achieved if defects are designed so to compensate off - resonance detunings introduced by velocity changes .
this fact gives rise to the possibility of using arrays of ol defects as very precise filters for matter wave dynamics .
the paper is organized as follows . in sec .
ii we present the model equation and discuss the basic properties of gss of the gpe . in the sec .
iii we consider the interaction of small amplitude gss with a localized gaussian impurity in the ol .
the problem is investigated by means of direct numerical integrations of the gpe for both attractive and repulsive interactions as well as for attractive and repulsive defects . in section
iv we use a stationary defect mode analysis to show that the repeated resonant transmission , reflection and trapping regions occurs in correspondence of resonances with impurity modes inside the defect potential and investigate their dependence on gs incoming velocities . in section v the multiple resonant transmission across a series of equidistant ( equal and unequal ) ol defects , is investigated . in sect .
vi the main results of the paper are shortly resumed .
let us consider a cigar - shaped bec described by the following normalized one - dimensional gpe @xcite @xmath0 where @xmath1 denotes an external potential of the form : @xmath2 , with @xmath3 a perfect ol of period @xmath4 : @xmath5 , and @xmath6 a defect potential consisting of a sum of @xmath7 single wells potentials localized on a distance of several lattice period around the ol sites @xmath8 , @xmath9 . in the following we assume @xmath10 and @xmath11 to have the form @xmath12
\label{defect}\end{aligned}\ ] ] where without loss of generality the period of the ol is taken @xmath13 .
we remark that results of this paper will not qualitatively dependent on type of defect ( we used gaussian defects just for numerical convenience ) and similar results can be obtained for other shapes of the defects , like square well defects , for example ) . in the following
we will mainly restrict to the cases of few ol defects : @xmath14 only .
we also remark that in eq.([gp ] ) the normalization has been made by measuring the energy in units of recoil energy @xmath15 , where @xmath16 and @xmath17 is the lattice constant , the space coordinate and time in units of @xmath18 and @xmath19 , respectively .
the dimensionless macroscopic wave function is also normalized as @xmath20 , where @xmath21 is the s - wave scattering length .
it is well known that in the absence of the defect potential @xmath22 , eq.([gp ] ) posses families of exact gs solutions with energy located in the band - gaps of the linear eigenvalue problem @xmath23\varphi_{\alpha k}=0,\ ] ] where @xmath24 are orthonormal set of bloch functions with @xmath25 denoting the band index and @xmath26 the crystal - momentum inside the first brillouin zone ( bz ) : @xmath27 $ ] .
it is also known that small - amplitude gss with chemical potentials @xmath28 very close to band edges are of the form @xmath29 with the envelope function @xmath30 obeying the following nlse @xmath31 where @xmath32 and @xmath33 are slow temporal and spatial variables , @xmath34 denotes the soliton effective mass and @xmath35 the effective nonlinearity @xcite .
the condition for the existence of such solitons is @xmath36 and coincides with the condition for the modulational instability of the bloch wavefunctions at the edges of the bz @xcite .
examples of small amplitude gss with chemical potential inside the band - gap structure are depicted in fig . [ fig_e(k ) ] . in the presence of very diluted ol defects
, gss will continue to exist and away from defects they practically coincide with gss of the undistorted ol .
an attractive ( local potential well ) or repulsive ( local potential barrier ) ol defect will be seen by the gss differently , depending on the sign of their effective mass .
thus , for example , in the case of repulsive interactions and a negative effective mass , a gs approaching a repulsive defect will see it as a trapping potential ( rather than as a potential barrier ) , thus besides being totally or partially transmitted / reflected , it can also be trapped at the defect site , a fact which would be impossible in absence of the ol . in all the numerical simulations presented in this paper
we have used stationary gss of eqs .
( [ gp])([defect ] ) , exactly determined by the shooting method @xcite or by self - consistent calculations @xcite , and have put them in action by means of phase imprinting ( e.g. we multiply the state by the phase factor @xmath37 , with @xmath38 being the gs velocity ) .
this provides initial condition for the gpe numerical time integration of the form : @xmath39 . for possible experimental implementations of our results ,
an alternate method to use to put the gs in action could be the acceleration of the ol for a short time interval to move the stationary state away from the bz edges ( center ) so that it can acquire a small bloch velocity @xmath40 ( for details on how this can be done see @xcite ) .
we consider first the scattering of a gs by a single localized defect ( @xmath41 in eq .
( [ defect ] ) ) . in order not to perturb the soliton initially
, the distance between the soliton center and the ol defect is taken much larger than the width of the gs ( @xmath42 ) . in the following
we compute the trapping , transmission and reflection coefficients defined as @xmath43 , @xmath44 and @xmath45 , with @xmath46 , @xmath47 and @xmath48 denoting the numbers of atoms trapped , transmitted and in the initial state , respectively . the final time
@xmath49 depends on the initial velocity of the gs and in the numerical experiment is determined as the time necessary for the coefficients @xmath50 to become stationary after the scattering process has occurred .
the trapping region @xmath51 $ ] has the size of the initial soliton and in all our calculations we fix @xmath52 . for the scattering of a gs in the semi - infinite gap to occur
, the existence criterion for a gs near the bottom edge of the first band ( where the effective mass is positive ) implies that the nonlinear coefficient must be negative sign@xmath53 . by applying a small initial velocity to the gs in the defect direction , depending on the amplitude @xmath54 and width @xmath55 of the defect
, three possible scenarios can occur : i ) complete reflection , @xmath56 ; ii ) complete transmission , @xmath57 ; iii ) partial trapping , @xmath58 .
the regions of the parameter space @xmath59 where these different regimes occur , as obtained from direct numerical integration of eq.([gp ] ) , are reported in fig .
[ eta(d)_w-0125_v005 ] .
the dependence of @xmath60 on the defect strength @xmath54 for two different incoming gs velocities and for a fixed value of the defect width @xmath61 is depicted in fig .
we see that by changing the strength of the defect it is possible to achieve complete reflections ( @xmath62 ) or transmission ( @xmath63 ) as well as partial trappings ( @xmath58 ) .
the profiles of the initial , reflected and transmitted gs are depicted in the top panels of fig.[fig3 ] for defect strengths corresponding to points labeled in the middle panel by letters a , b , c. by comparing the middle and the bottom panels of fig.[fig3 ] it is clear that the sharp peaks at small incoming velocities ( @xmath64 ) for which @xmath63 ( see points a , c , e , f , g in the middle panel ) become wider as the velocity is increased while the regions for which @xmath62 are a bit reduced ( also see fig.[fig3_transm_0.475 ] for the case of first band - gap gss ) .
the first four impurity modes corresponding to the transmission peaks @xmath65 in the middle panel of fig.[fig3 ] have been depicted in fig .
[ modes_fig3 ] .
notice the alternating odd - even symmetry of these modes ( modes d and f being odd and modes e and g being even with respect to the center @xmath66 of the defect potential ) as usual for eigenstates of one - dimensional trapping potentials .
we remark that the existence of four resonant transmission peaks ( and reflection regions ) seen in fig.[fig3 ] for @xmath67 correlates with the existence of four impurity modes for the given defect strength region ( see stationary defect mode analysis below ) .
similar results as those of the previous section can be found for gss inside the first band - gap , with the only difference that now there are two possibilities for the existence of small - amplitude solitons : i ) in the vicinity of the top edge of the first band where the effective mass is negative and therefore gss can exist only for repulsive interactions @xmath68 ; ii ) and in the vicinity of the bottom edge of the second band where effective mass is positive and gs exist only for attractive interactions @xmath69 .
the small - amplitude gs near the first band - gap edges b , c , in fig.[fig_e(k ) ] are shown by the corresponding profiles @xmath70 and @xmath71 depicted in the figure . as remarked
before , the sign of the effective mass determines the type of the interaction of the gs has with the defect potential and for negative gs effective mass ( repulsive interatomic interactions ) the defect will be seen as a defect trapping potential ( supporting therefore bound states ) if the defect strength @xmath54 is positive rather than negative ( as seen for the case of a positive effective mass ) . except for this ,
results go in parallel with those of the previous section and have been collected in figs.[fig3_transm_0.475 ] , [ fig3_transm_0.475_v ] for the case of an initial gs close the top edge of the first band ( negative effective mass ) . in particular , from the bottom panel of fig .
[ fig3_transm_0.475 ] and fig .
[ fig3_transm_0.475_v ] we clearly see that for a fixed defect width the region of the resonant transmission ( corresponding to red color ) becomes wider as the incoming gs velocity is increased , while the full reflected regions are achieved mainly for small velocities , as one could have expected .
we remark that also in this case the number of resonant transmission reflection and trapping regions are found to correlate with the number of impurity modes present in the defect potential for the given range of the defect strength ( first four modes corresponding to the maxima of the trapping coefficient c in fig.[fig3_transm_0.475 ] are depicted in fig.[fig3_transm_0.475_modes ] , [ modes_fig4_0475 ] ) .
as remarked before , the presence of a repulsive ( attractive ) ol defect in the gpe affects the existing gs states for @xmath72 and introduces additional localized states inside the gap in presence of repulsive ( attractive ) nonlinearities .
numerical calculations show that the band structure is only slightly affected by an ol impurity , the main effect being the introduction of bound states spatially localized at the impurity sites and with chemical potentials inside the band - gap .
note that repulsive ( attractive ) ol defects in the presence of an attractive ( repulsive ) nonlinearity can not introduce additional bound states because the corresponding impurity potentials correspond to barriers rather than potential wells , due to the positive ( negative ) effective mass . recalling that the opposite signs of nonlinearity and effective mass is a necessary condition for the gs existence , one has that the effective impurity potential acts as a trapping potential when @xmath54 and @xmath73 have equal signs .
away from the ol impurity localized states are practically the same as for @xmath72 case . at the defect site , however , gss levels get slightly shifted by the impurity potential and additional impurity modes enter the gap .
a gs moving through the impurity will have , in general , a mismatch in energy and in number of atoms with the impurity modes , this giving a partial reflection / transmission of the incoming matter wave .
a total transparency of the ol defect is expected for incoming gs energies and number of atoms exactly matching those of an impurity modes inside the defect potential ( notice that the energy of impurity modes depend on @xmath54 and on the number of atoms ) . by increasing the strength of the impurity ,
the depth of the defect potential increases and more impurity modes enter the gap .
as @xmath74 is increased , the energies of these modes enter the gap from the top ( bottom ) of a band for impurity strength and the nonlinearity both positive ( negative ) .
this implies that the transparency ( complete transmission ) of the impurity occurs in correspondence of each impurity mode entering the gap and matching the energy of the incoming gs given by @xmath75 where the first term @xmath76 represents the energy of the stationary gs state at @xmath77 ( eg . with @xmath78 ) , while the second one is the contribution due to the kinetic energy ( here @xmath79 is the bloch velocity and @xmath80 the effective mass ) . for small velocities
the energy ( [ energyshift ] ) practically coincides with @xmath76 but in general the kinetic term should be accounted in the matching in energy with the impurity levels ( see below ) .
notice that eq .
( [ energyshift ] ) is only valid near stationary points @xmath81 ( bottoms or tops of a band ) where @xmath82 and @xmath83 can be written as @xmath84 in the range of initial velocities @xmath85 $ ] we have considered , the energy curves @xmath86 in vicinity of @xmath87 and @xmath88 are very well approximated by eq .
( [ energyshiftapprox ] ) with @xmath89 , @xmath90 , and @xmath91 and @xmath92 for bottom and top edges of the band , respectively . by knowing @xmath93 ( e.g. @xmath26 ) and @xmath94 one can compute the energy shift due to the nonzero velocity to be accounted in the matching between the gs and the impurity levels ( see lower panels of figs .
[ modes_fig4_-0125 ] , [ modes_fig4_0475 ] ) .
notice that the kinetic energy has the sign of the effective mass so that @xmath83 is pushed forward the corresponding band edge for finite @xmath93 , meaning that inside the impurity potential the gs matching condition with an impurity mode can be achieved for a lower values of @xmath74 . from this
we expect the resonance transmission peaks to be shifted away from the @xmath95 resonance toward lower values of @xmath74 as @xmath93 is increased . to confirm these prediction with gpe calculations we have solved the stationary problem @xmath96u-\sigma u^3=0\end{aligned}\ ] ] searching for bound states @xmath97 localized at the ol defect using both shooting method and self - consistent calculations . using these approaches we found that the values of @xmath54 for which the gaussian defect becomes transparent to the gs dynamics , correspond to the values for which a localized mode inside the defect potential has energy and number of atoms matching the corresponding quantities of the incoming gs .
since the discussion is very similar ( a part implications due to signs of the effective masses ) for gs of the semi - infinite and for the ones inside finite - gaps , we refer to the case of an initial gs inside the semi - infinite gap , with a positive effective mass with chemical potential close to the bottom of the first band . in the top panel of fig.[modes_fig4_-0125 ]
we have shown the dependence of the number of atoms @xmath98 in a defect mode of a given symmetry on @xmath54 , for an energy @xmath99 corresponding to a gs with zero incoming velocity ( point a of figs .
[ fig_e(k ) ] , [ fig3 ] ) .
the horizontal dotted line refers to the numbers of atoms in the initial gs . in the bottom panel of fig .
[ modes_fig4_-0125 ] we show the energy mismatch @xmath100 between the energy of a defect modes @xmath83 and the one of an incoming gs given by eq .
[ energyshiftapprox ] for different incoming velocities .
notice that the intersection points @xmath101 of the dotted line @xmath102 with the curves @xmath103 are in coincidence with the zeros of the function @xmath100 displayed in the bottom panel for @xmath64 and correspond to the maxima of the transmission coefficient @xmath104 in fig.[fig3 ] ( see curves @xmath64 in the middle panel ) .
[ modes_fig4_-0125 ] ( see also [ modes_fig4_0475 ] for repulsive case ) also explains the decay of the reflection coefficient and the rising of the trapping coefficient coinciding up to the maxima observed ( just before the resonance ) for increasing values of @xmath74 away from the resonance points @xmath101 .
to the right of these points in fig.[modes_fig4_-0125 ] , the number of particles in the defect mode is higher then the number of particles in the incoming soliton @xmath105 . in this case
the gs can not `` use '' a defect mode to pass the ol impurity and will be fully reflected . on the left of the resonance points ,
the number of atoms in the gs is higher than the one in the defect mode so that the gs can be captured by the ol impurity by releasing the excess number of atoms into the reflection and transmission channels .
it is clear that the peak of the capture coefficient occurs just before the resonant transmission ( e.g. for @xmath106 ) of the gs through the defect ( achieved when the condition @xmath107 is exactly fulfilled ) .
notice from fig.[modes_fig3 ] that at the resonant transmission , the profiles of the stationary impurity modes obtained by solving eq.([matheiu_def ] ) exactly coincide with the solution of the gpe during the passage trough the defect . in fig.[-0125_odd_even_eta(d ) ] we have also depicted , similarly to fig.[eta(d)_w-0125_v005 ] , the level curves @xmath102 as a function of the amplitude @xmath54 and the width @xmath55 of the defect . from the bottom panel of fig .
[ modes_fig4_-0125 ] it is also clear that for @xmath108 the resonances shift in the direction of lower values of @xmath74 ( as expected from our analysis ) .
this well correlates with the gpe calculations reported in the bottom panel of fig.[fig3 ] .
similar results are obtained for a gs near the bottom of the first band gap ( point b in fig . [ fig_e(k ) ] ) for the case of repulsive interactions .
this is shown in fig.[modes_fig4_0475 ] .
from which we see that the shift of the resonance due to the finite bloch incoming velocity is always in the direction of lower values of @xmath54 and are in good agreement with the gpe numerical results reported in the bottom panel of fig .
[ fig3_transm_0.475 ] .
in particular , for the velocity @xmath109 and the resonance near @xmath110 we obtain from fig .
[ modes_fig4_0475 ] that the resonance peak is at @xmath111 while from the gpe result in fig .
[ fig3_transm_0.475 ] we obtain the value @xmath112 .
it is also worth to note from the bottom panels of figs .
[ modes_fig4_-0125 ] , [ modes_fig4_0475 ] , that for a fixed energy mismatch the widths of the curves increase as the incoming velocity is increased , a fact which correlates with the broadening of the transmission peaks observed in the gpe calculations ( compare bottom panels in fig . [ fig3 ] and in fig [ fig3_transm_0.475 ] , respectively ) . from this
we conclude that the above impurity mode analysis fully confirms the interpretation of the transmission peaks as resonances between incoming gss and impurity modes .
an interesting question to ask is whether the soliton resonant transmission could survive multiple impurity scatterings . in fig .
[ fig10 ] we show the transmission , reflection and trapping regions of an attractive gs of the gpe with two identical gaussian ol impurities , placed at @xmath66 and @xmath113 , respectively .
the tcr curves depend also on the distance between the two impurities and this could be varied so to find optimal values for double resonant transmission to occur .
the phenomenon , however , is more sensitive to variations of the incoming gs velocity as one can see by comparing fig.[fig10 ] with the case of single impurity in fig .
[ fig3 ] .
notice that while the transmission resonance peaks at small velocities are practically unaffected by the presence of the second impurity , they become more narrow at larger velocities for the resonant scattering on two impurities .
the shrinking of the resonances at larger velocities can be understood by the fact that for higher incoming velocities the velocity of the gs after the first impurity is slightly reduced and the variation introduces a detuning from resonance in the scattering with the second impurity which in turn reduces the double resonance width . from this
one can expect that in the presence of more lattice defects the multiple resonance transmission peaks become very narrow and only solitons with very precise initial matching velocities will be able to pass .
in general the gs may be able to pass only a finite number of impurities before remaining trapped at one impurity or becoming scattered back and forth between them .
this is shown in fig.[fig11 ] where contour plots of the gs space - time dynamics in presence of two identical gaussian defects are reported for different strengths @xmath54 and for the same parameters as for the bottom panel of fig.[fig10 ] . in the left top and bottom panels we see the occurrence of total reflection and transmission for a resonant value of @xmath54 while in the corresponding right panels we show the back and forth dynamics of the soliton between two impurities ( top panel ) and the trapping by the second impurity ( top panel ) for a non resonant values of @xmath54 .
notice the small change of the soliton velocity after the passage through the first defect . to overcome the detuning from resonance induced by the change of velocity one could design the strength of the second impurity so to match the value of the intermediate velocity and still achieving @xmath63 . in this way one can realize a double ( or multiple ) filtering of the soliton motion so that only gss having very precise initial velocity can overcome the defect series . this possibility is illustrated in fig .
[ fig12 ] where the resonant transmission of a gs though a series of three gaussian defects of the ol , with defect strengths designed so to compensate the detunings introduced by the velocity changes , is shown .
notice from the left panel of fig .
[ fig12 ] that in the case of equal defects with the initial velocity matching the resonance transmission peak of the first defect , the gs can not be transmitted through all the three defects but it is stopped at the second defect .
the dynamics of the soliton in presence of multiple defects may be quite complicated and is beyond the aim of this work ( an investigation in the parameter space of the possible scenarios for the gs time evolution will be discussed elsewhere ) .
in this paper we have investigated the scattering properties of matter wave gap - solitons with optical lattice defects in the framework of the mean - field gross - pitaevskii equation .
we have shown that the occurrence of repeated reflection , transmission and trapping regions are in correspondence of the defects strengths for which the number of atoms and energies of additional bound states created by the optical lattice defect , match the ones of the incoming gap - soliton .
this has been demonstrated by a study of the stationary defect modes energies ( chemical potentials ) and number of atoms as a function of the defect strength .
a very good agreement between the predicted values of the resonant transmission peaks by means of impurity modes and the ones found by direct time integrations of the gross - pitaevskii equation , is found .
the behavior of the reflection and trapping curves have also been explained by impurity mode analysis .
these investigations have been performed both for attractive and repulsive interactions and for localized states both in the semi - infinite gap ( attractive case ) and in the first gap zone ( attractive and repulsive cases ) .
the dependence of the resonant transmission on the initial gap - soliton velocity has been also investigated .
we have shown that the positions of resonant transmission peaks shift toward lower values @xmath74 as the initial gap - soliton velocity is increased , while the widths of resonances shrinks to zero at very small gap - soliton velocities .
the possibility of multiple resonant transmission through an arrays of defects was also demonstrated . in particular , we have shown that for an optical lattice with two equally spaced identical gaussian defects the widths of the full transmission resonances for larger incoming velocities , decreases as the number of defects is increased .
the sharpening of the transmission peaks has been explained in terms of the detuning from resonance introduced by the small velocity change after the passage of an optical lattice defect . finally , the resonant transmission of a gap - soliton though a series of gaussian defects with unequal strengths designed so to compensate detunings introduced by the velocity changes , was demonstrated .
these results give the possibility to construct very precise filters for the matter wave gap - soliton dynamics by means of properly designed arrays of ol defects .
v.a.b . acknowledges support from the fct grant , ptdc / fis/64647/2006 .
ms acknowledges the miur ( prin-2008 initiative ) for partial financial support .
see review papers o. morsch and m. oberthaler , rev .
phys . * 78 * , 179 ( 2006 ) and references therein ; i. bloch , j. phys .
b * 38 * , s629 .
( 2005 ) ; d. jaksch , and p. zoller , ann .
phys . n.y . * 315 * , 52 ( 2005 ) ; v.a .
brazhnyi and v. v. konotop , mod .
. lett . * 18 * 627 ( 2004 ) .
o. morsch , j. h. mller , m. cristiani , d. ciampini , and e. arimondo , phys .
lett . * 87 * , 140402 ( 2001 ) ; i. carusotto , l. pitaevskii , s. stringari , g. modugno , and m. inguscio , phys . rev . lett . * 95 * , 093202 ( 2005 ) . m. jona - lasinio , o. morsch , m. cristiani , n. malossi , j. h. mller , e. courtade , m. anderlini , and e. arimondo , phys .
lett . * 91 * , 230406 ( 2003 ) ; s. wimberger , r. mannella , o. morsch , e. arimondo , a. r. kolovsky , and a. buchleitner , phys .
a * 72 * , 063610 ( 2005 ) .
baizakov , v.v .
konotop , and m. salerno , j. phys .
b * 35 * 5105 ( 2002 ) ; e.a .
ostrovskaya , yu.s .
kivshar , phys .
lett . * 90 * , 160407 ( 2003 ) ; j. yang and z. musslimani , opt . lett . * 23 * , 2094 ( 2003 ) ; b.b .
baizakov , b.a .
malomed , and m. salerno , europhys .
lett . * 63 * , 642 ( 2003 ) . | the physical mechanism underlying scattering properties of matter wave gap - solitons by linear optical lattice defects is investigated .
the occurrence of repeated reflection , transmission and trapping regions for increasing strengths of an optical lattice defect are shown to be due to impurity modes inside the defect potential with chemical potentials and numbers of atoms matching corresponding quantities of an incoming gap - soliton . for gap - solitons with chemical potentials very close to band edges ,
the number of resonances observed in the scattering coincides with the number of bound states which can exist in the defect potential for the given defect strength .
the dependence of the positions and widths of the transmission resonant on the incoming gap - soliton velocities are investigated by means of a defect mode analysis and effective mass theory .
the comparisons with direct integrations of the gross - pitaevskii equation provide a very good agreement confirming the correctness of our interpretation . the possibility of multiple resonant transmission through arrays of optical lattice defects is also demonstrated .
in particular , we show that it is possible to design the strength of the defects so to balance the velocity detunings and to allow the resonant transmission through a larger number of defects . the possibility of using these results for very precise gap - soliton dynamical filters is suggested . |
it was recognized by shlyakhter @xcite that isotopic ratios of fission products and secondary neutron absorption reactions that occurred with the oklo natural reactor phenomenon some 2 billion years ago could be used to test whether the fine - structure constant @xmath0 varies with time .
specifically , if a neutron absorber ( e.g. , @xmath3sm ) has a low - energy neutron absorption resonance , and the excited compound state can decay by gamma emission , the resonance energy will change in a predictable way if @xmath0 varies . the overlap of the resonance with the reactor neutron spectrum will subsequently change thereby altering the ratios of nearby isotopes .
in addition to @xcite , two other detailed analyses of the oklo phenomenon have been completed @xcite .
the purpose of this note is to assess the validity of the assumption in these analyses that the neutron spectrum can be described by a maxwell - boltzmann ( mb ) distribution , and to determine the effects on the limit of the time variation of @xmath0 for different neutron spectra .
the possibility of a non - mb spectrum was discussed in @xcite ( footnote 3 ) , but was not considered to be important . however , for the high - accuracy results presented in @xcite , the effects due to deviations from a mb spectrum become very important .
the fact that the oklo phenomenon occurred was indicated when uranium taken from this mine was found to be deficient in @xmath4u ; subsequent tests confirmed this discovery . from core samples and knowledge of reactor dynamics , the age of the oklo phenomenon was determined to be about 2 billion years . the conditions necessary for the multiple reactor zones to exist can be used to place constraints upon many parameters such as concentrations of various elements , temperature , moderator type and concentration , _ etc .
_ while the reactor was operating .
for example , it can be shown that only hydrogen or deuterium could have served as moderators at oklo and hydrogen present in water or as hydration of the uranium oxide appears to be the only plausible choice . as a first approximation to the reactor geometry
, it seems reasonable to assume that the oklo phenomenon was a homogeneous reactor of infinite extent .
this simplifies the estimation of the multiplication factor @xmath5 and the neutron flux @xmath6 which we will see depends on the amount of hydrogen and neutron absorbing impurities in the ore . with the present isotopic ratio of @xmath4u/@xmath7u , the maximum @xmath5 that can be obtained in a homogeneous mixture of water and natural uranium is @xmath8 , which occurs where there are 2.43 molecules of water per atom of uranium @xcite ( see table 18.1 , p. 612 ) .
the fractional isotopic abundance of @xmath4u in natural u at present is 0.711% . extrapolating this result
implies a @xmath4u relative isotopic concentration of about 1.2% to achieve @xmath9 .
two billion years ago , the relative concentration of @xmath4u was about 3.7% ( due to the different lifetimes of @xmath4u and @xmath7u ) .
this implies that two billion years ago , @xmath10 for a system with 2.4 h@xmath11o molecules / u atom .
if such a large fractional concentration of water was present , the oklo reactor would have been highly divergent . of the possible effects that can reduce @xmath5 , the two most likely are a low water ( or hydrogen ) concentration in the u deposit , and the presence of impurities .
it is interesting to note that the reactor would have been somewhat self - stabilizing as a runaway to high temperatures would drive water out of the ore deposit
. precise knowledge of the nature and concentration of neutron - absorbing impurities at the time the oklo reactor was running is lacking .
however , we can make a reasonable estimate of the possible maximum absorption by impurities .
the forgoing arguments can made rigorous by estimating @xmath5 as a function of the hydrogen atomic number density fraction , @xmath12 and the atomic fraction of absorbing impurities @xmath13 , relative to the number density of u atoms . with @xmath4u isotopic fraction @xmath14 , for thermal neutrons : @xmath15+\rho_h\sigma_a^h+\sum_i \rho_i\sigma_a^i}\nonumber\\ & \approx & 2.47 e^{-.7031 ( 1/f_h)^{.58}}\nonumber\\ & \times&{580x\over 690x+2.7(1-x)+0.33 f_h + \sum_i f_i\sigma_a^i}\end{aligned}\ ] ] where @xmath16 is the number of neutrons released per @xmath4u fission , @xmath17 is the resonance escape probability obtained from eqs .
( 10.29 ) and ( 13.3 ) of @xcite , @xmath18 b is the @xmath4u fission cross section , @xmath19 b is the total @xmath4u total neutron absorption cross section , @xmath20 b is the @xmath7u total neutron absorption cross section , @xmath21 b is the hydrogen absorption cross section
. the total impurity absorption can be parameterized by @xmath22 furthermore , the effects of oxygen have been neglected . for @xmath23 , @xmath5 as a function of @xmath24 , and as a function of impurity concentration parameterized by @xmath25 ,
is shown in fig .
1 . with no impurities ,
@xmath9 is achieved when @xmath26 . for a higher hydrogen concentration
, the reactor would have run away and the resulting high temperatures would have driven water out of the ore ; the system was likely self - stabilizing to @xmath9 . as the impurity concentration , parameterized by @xmath25 in fig .
1 , increases , @xmath24 required for @xmath9 increases also ; at @xmath27 , @xmath28 to achieve @xmath9 . for @xmath29
b , it is not possible to achieve @xmath30 .
it is also known that the reactor operated until @xmath31 which implies a more restrictive limit of @xmath32 b , for which @xmath33 .
we note that criticality calculations at these high @xmath34u concentrations are subject to a number of corrections ( see @xcite , pp .
463 - 468 ) so the figures derived here are to be considered as providing guidelines for the operation of the natural reactor which indicates that @xmath24 is about 3 , with @xmath35 . a ubiquitous substance that offers a significant thermal neutron absorption cross section is sodium chloride .
the thermal absorption cross section for sodium is @xmath36b , but for chlorine @xmath37 b which limits @xmath38 .
the relative atomic concentrations of other impurities can be similarly estimated .
it is well - known that when neutrons are moderated in the presence of a @xmath40 absorber ( such as u ) , the low - energy part of the spectrum becomes suppressed relative to a high - energy @xmath41 component of the flux .
the relevant parameter that describes this suppression is ( @xcite , pp . 335- 340 ) @xmath42 where @xmath43 is net absorption cross section of the homogeneous mixture for a neutron rms velocity corresponding to the moderator temperature , @xmath44 is the net scattering cross section of the moderating atoms , and @xmath45 is the atomic mass of the moderating atoms . in the present case , for neutron energies about 1 ev , the atomic mass @xmath46 for h which is responsible for most of the neutron moderation . when @xmath47 , the @xmath41 tail contains the bulk of the neutrons .
this is illustrated in fig .
2 where the flux as a function of @xmath48 as obtained from a monte carlo calculation following the procedure in @xcite .
for the parameters derived in the last section , that is @xmath49 , @xmath50 , and @xmath35 , @xmath51 .
however , for neutron energies less than 1 ev , molecular effects become important ; it is no longer a good approximation to assume that the neutron moderation occurs by scattering from free atoms .
the chemical bonding effect can be estimated by setting @xmath52 , so in the low neutron energy region ( @xmath53 ev ) , @xmath47 at 300 k , and scales with temperature as @xmath54 the elements @xmath3sm and @xmath55gd have low energy neutron resonances and are particularly interesting to test for possible time variation of @xmath0 .
the formalism developed in @xcite for the effective cross section of a substance with a low - energy neutron resonance can be understood as follows .
the interaction of a neutron with a nucleus can be parameterized through the optical model , where the hamiltonian is @xmath56 where the absorption rate for neutrons with energy @xmath57 is @xmath58 .
the absorption cross section can be expressed as @xmath59 where @xmath60 is the neutron momentum , @xmath61 is a parameter that depends upon the spins of the neutron and interacting nuclei , @xmath62 is the resonance energy and @xmath63 are the rates for neutron and gamma ray absorption respectively .
the parameters @xmath61 , @xmath62 , and @xmath63 for elements of interest here can be found in @xcite .
the absorption cross section at a particular energy can also be described as the product of the rate of absorption and the times @xmath64 spent in the material , normalized to the propagation length @xmath65 , @xmath66 far away from @xmath62 , @xmath67 is nearly constant and @xmath68 , where @xmath69 is the velocity at which the cross section was measured .
nuclei whose cross sections can be described this way are know as `` @xmath40 '' absorbers .
this is a more efficient representation for thermal neutron absorption of many nuclei such as u. the net absorption cross section is given by the integral of the cross section time the neutron flux @xmath70 , @xmath71 in assessing the effects of a varying low - energy resonance , because all @xmath40 cross sections follow the same universal function which have exactly the same form up to a constant , it is useful to define an effective relative cross section for the nuclei with a varying @xmath67 as @xmath72 where @xmath73 is the energy - dependent cross section , @xmath74 is the moderator temperature , @xmath6 and @xmath75 are the neutron flux and energy distributions , respectively , and @xmath69 is the velocity at which @xmath40 absorption cross sections are measured . in this analysis and the others that have been performed so far , the finite temperature and motion of the absorbing nuclei is ignored .
this is a small correction to the resonance width because @xmath76 for the nuclei being considered , so the absorber velocity is much lower than the neutron velocity .
assuming @xmath77 at 300 k , as a function of @xmath74 , @xmath78 . a procedure similar to that described in @xcite ( numerical integration )
was used to calculate the change in @xmath39 as a function of change in resonance energy , but replacing the mb spectrum with spectra derived from fig .
2 . the same resonance parameters as employed in @xcite were used , and the numerical integration was tested with a mb spectrum for which the same results as fig . 1 of @xcite were obtained .
it is possible to determine the value of the cross section for @xmath3sm , @xmath79 , from detailed isotopic analyses of core samples from oklo and similarly for @xmath55gd to determine @xmath80 .
such an analysis was performed in @xcite and indicates that @xmath81 this result can be used subsequently to determine the ancient value of @xmath62 which can be compared to the present - day measured value .
results of the numerical integration for @xmath3sm are shown in fig .
2 as a function of moderator temperature with @xmath77 at 300 k. in @xcite , the mean temperature of the oklo reactor was estimated to be about 600 k. from the curve for 600 k in fig . 3 ( expanded around the @xmath82 kb in the lower plot ) , we find that the change in resonance energy is @xmath83 \times 10^{-3}\ { \rm ev}\ ] ] which indicates more than a @xmath2 deviation from zero .
note that thetwo solution " problem does not really exist in this case , but the upper magnitude range is expanded . in the calculation of the moderated neutron spectrum , effects of chemical bonds ( which suppress the low energy spectrum ) have been approximately included .
as the atoms become more tightly bound , @xmath84
so ( @xmath85 ) , and in this limit the neutron flux spectrum becomes pure @xmath41 ; the result of a calculation for this case is shown in fig .
4 . in this case ,
for which two solutions exist , @xmath86 this result is roughly consistent with the @xmath77 result ; reality likely is somewhere between the @xmath872 and @xmath88 results .
however , the result for @xmath77 clearly indicates a non - zero change in @xmath62 at a level of more than @xmath2 .
results for a calculation of @xmath55gd are shown in fig .
the interpretation of the @xmath55gd data is complicated by the presence of impurities , and an extensive analysis is given in @xcite .
the oklo data imply @xmath89 kb , which , from fig .
5 , indicates , for the left - hand solution , @xmath90 this is consistent with the @xmath3sm result because impurities shift @xmath62 slightly , and the simple @xmath41 spectrum overestimates @xmath62 .
a shift in resonance energy can be related to a shift in @xmath0 by @xmath91 a reasonable estimate of @xmath92 mev was presented in @xcite , and is roughly the same for @xmath3sm and @xmath55gd . taking the result eq .
( [ 149result ] ) determines @xmath93\times 10^{-9}\ ] ] which is non - zero by more than @xmath2 .
the central value plus twice the upper uncertainty can be used to set a 95% confidence limit on the magnitude of the integrated rate of change over the last two billion years @xmath94 by assuming higher temperatures , this limit decreases slightly . however
, a significant modification of this result requires temperatures beyond physically reasonable values as assumed in @xcite .
it is expected that the neutron absorption resonance parameters ( e.g. , the neutron and gamma widths ) , other than @xmath62 , do not change significantly with @xmath0 . in estimating @xmath95 it was assumed that @xmath62 represents a fixed energy above the ground states of nuclei with @xmath45 and @xmath96 , and @xmath95 is approximately the difference in the coulomb energies of the ground states .
this result is in agreement with a more sophisticated analysis @xcite .
if a @xmath40 absorber is present in the moderator of a neutron chain reactor , the neutron energy spectrum can deviate significantly from the a maxwell - boltzmann distribution .
the oklo phenomenon was rich in uranium which is a strong @xmath40 absorber . by considering realistic deviations from a maxwell - boltzmann low - energy neutron spectrum
, we have shown that a recent analysis of the oklo natural reactor phenomenon implies an non - zero ( @xmath97 ) variation of @xmath0 over the last 2 billion years .
the result is of opposite sign and an order of magnitude smaller than a recent astrophysical determination of the change in @xmath0 @xcite , which is non - zero , but over a different time scale .
it has been suggested that the time variation of @xmath0 is not monotonic @xcite .
our results are probably accurate to within 20% ; a full mcnp model of the reactor , assuming reasonable estimates of the temperature and impurities are possible , would be useful in light of the importance and interest in the possibility of a varying @xmath0 . finally , these results might be interpreted more efficiently in terms of @xmath98 for which the sensitivity is about two orders of magnitude higher than the sensitivity to @xmath0 variation @xcite . | in previous analyses of the oklo ( gabon ) natural reactor to test for a possible time variation of the fine - structure constant @xmath0 , a maxwell - boltzmann low energy neutron spectrum was assumed .
we present here an analysis where a more realistic spectrum is employed and show that the most recent isotopic analysis of samples implies a decrease in @xmath0 , over the last two billion years since the reactor was operating , of @xmath1 ( @xmath2 confidence ) .
issues regarding the interpretation of the shifts of the low energy neutron absorption resonances are discussed . |
among active galactic nuclei ( agns ) , blazars show intense and variable emission above 100 mev @xcite , and the variability time scale can be as short as a few days , or last a few weeks .
they emit across several decades of energy , from the radio to the tev energy band and their spectral energy distributions ( seds ) are typically double humped with a first peak occurring in the ir / optical band in the flat - spectrum radio quasars ( fsrqs ) and low - energy peaked bl lacs ( lbls ) , and at uv / x - rays in the high - energy peaked bl lacs ( hbls ) .
this peak is commonly interpreted as synchrotron radiation from high - energy electrons in a relativistic jet .
the second sed component is commonly interpreted as inverse compton ( ic ) scattering of soft seed photons by relativistic electrons , and peaks in the mev gev and in the tev energy bands in the fsrqs / lbls and in the hbls , respectively .
a recent review of the blazar emission mechanisms and energetics is given in @xcite .
the fsrq ( pks 2251@xmath2158 ; @xmath3 ) is certainly one of the most active extragalatic sources at high energy . in the era , it was detected in 1992 during an intense flaring episode @xcite when its flux @xmath4 was observed to vary within the range @xmath5photons@xmath6s@xmath7 . in 1995 , a 2-week campaign detected a flux @xmath8 of its historical maximum @xcite .
figure [ fig:3c454:egret ] shows the light curve for @xmath9mev as observed by egret in the period 19911995 .
mev in the period 1991 - 1995 .
the downward arrow represents a @xmath10 upper - limit .
data from @xcite.,width=302 ] in 2005 , underwent a major flaring activity in almost all energy bands ( see @xcite ) . in the optical
, it reached @xmath11mag @xcite and it was detected by at a flux level of @xmath12photons@xmath6s@xmath7 in the 3200 kev energy band @xcite . since the detection of the exceptional 2005 outburst , several monitoring campaigns were carried out to follow the source multifrequency behavior @xcite . during the last of these campaigns , 3c 454.3 underwent a new optical brightening in mid july 2007 , which triggered observations at all frequencies , including the one . in the following ,
we briefly describe the satellite , and we discuss the various campaigns triggered on .
the agile satellite @xcite is a mission of the italian space agency ( asi ) devoted to high - energy astrophysics , operative since april 2007 .
the scientific instrument combines four active detectors yielding broad - band coverage from hard x - rays to : a silicon tracker ( st ; * ? ? ?
* mev30 gev ) , a co - aligned coded - mask hard x - ray imager , super agile ( sa ; * ? ? ?
* kev ) , a non - imaging csi mini
calorimeter ( mcal ; * ? ? ?
* mev ) , and a segmented anti - coincidence system ( acs ; * ? ? ?
the gamma - ray imaging detector ( grid ) reaches a 5-@xmath13 sensitivity of about @xmath14@xmath6s@xmath7 above 100 mev for an integration time of 2 days , sources positioned 30 degrees off - axis , an earth occultation fraction of 20% and near the galactic plane ( high background ) .
the hard x - ray monitor on board agile , super- ( sa ) , reaches a sensitivity of about 15 mcrab at 5-@xmath13 in the energy range 1860 kev in a one - day observation .
at the epoch of the optical flare ( 2007 july 1921 ) the closest pointing position allowed by the solar panel constraints ranged within 35@xmath1540@xmath15 from the source .
because of the remarkably large ( @xmath16sr ) field of view ( fov ) of the grid and the successful detection of the vela pulsar at @xmath17 off - axis obtained during the science verification phase , agile could reliably study despite its large off - axis position .
the observations were performed between 2007 july 24 14:30 ut and 2007 july 30 11:40 ut , for a total pointing duration of @xmath18220 ks .
the source detection significance is 13.8@xmath13 as derived from a maximum likelihood analysis @xcite .
figure [ fig:3c454:lcjul07 ] , panel ( a ) , shows the agile / grid light curve for @xmath9mev during the flare , while panel ( b ) shows the comparison with the historical egret flux values .
we note that is detected at a 4-@xmath13 level during almost the whole period on a 1-day timescale .
the average flux above 100 mev for the whole period is @xmath19 .
the average flux above 100 mev for the whole period was the highest ever detected from this source , as shown in figure [ fig:3c454:lcjul07 ] , panel ( b ) .
the source was not detected ( above 5-@xmath13 ) by the super - agile iterative removal of sources ( iros ) applied to the @xmath20 image , in the 2060 kev energy range .
assuming a power law spectral shape with photon index @xmath21 , we find a 3-@xmath13 upper limit of @xmath22 on the average flux from . in november
2007 began pointing at high off - axis angle ( about @xmath23 ) .
nevertheless , in a few days was detected at more than 5-@xmath13 , exhibiting variable activity on a day time - scale . immediately after the source detection ,
a multiwavelength campaign started @xcite .
data were collected during two different periods , the first ranging between 2007 - 11 - 10 12:17 ut and 2007 - 11 - 25 10:57 ut and the second between 2007 - 11 - 28 12:05 ut and 2007 - 12 - 01 11:39 ut , for a total of about 592ks .
the three - day gap between them was due to a pre - planned grid calibration activity .
data were collected during a dedicated too on revolutions 623 ( between 2007 - 11 - 20 03:35 ut and 2007 - 11 - 22 08:46 ut ) and 624 ( between 2007 - 11 - 22 20:45 ut and 2007 - 11 - 24 15:50 ut ) , for a total of about 300ks , while /xrt data were obtained during several too pointings for a total of about 10ks .
data ( radio to optical ) as well as /uvot data were published in @xcite , while data were acquired following a too request . in both cases ,
optical data were acquired continuously during the whole campaign .
mev ) ; pentagons represent /ibis data ( 20200 kev ) ; triangles represent /xrt data ( 0.310 kev ) ; solid and open squares represent @xmath1-band webt @xcite and rem data , respectively . adapted from @xcite.,width=340 ]
figure [ fig:3c454:lcnov07 ] shows the multi wavelength light curves during the november 2007 observing campaign .
we investigated the expected @xmath24-optical flux correlation by means of the discrete correlation function ( dcf ) .
the dcf peak occurred at @xmath25 , and its value is @xmath26 .
this indicates a moderate correlation , with no significant time delay between the and optical flux variations .
we fit the seds for two major flaring episodes by means of a one - zone leptonic model , considering the contributions from synchrotron self - compton ( ssc ) , and from external seed photons originating both from the accretion disk and from the broad - line region ( blr ) .
we note that during both flaring episodes , the external compton scattering of direct disk radiation ( ecd ) contribution can account for the soft and hard x - ray portion of the spectrum , which show a moderate , if any , time variability .
however , we note that the ecd component alone can not account for the hardness of the spectrum .
we therefore argue that in the energy band a dominant contribution from the external compton scattering of the blr clouds ( ecc ) seems to provide a better fit of the data during the -ray flaring states
. in december 2007 began a multi wavelength campaign involving _ spitzer _ , _ swift _ , _ suzaku _ , the webt consortium , the rem and mitsume telescopes , offering a broad band coverage that allowed a truly simultaneous sampling of the synchrotron and ic emissions @xcite .
figure [ fig:3c454:seddec07 ] shows the sed during 2007 december 13 , characterized by the broadest multi wavelength coverage . during this epoch
the source was in a state characterized by a flux for @xmath9 mev of the order of @xmath27 .
we found that a model accounting for ecd and ecc components does not fully describe the seds in the three epochs .
an additional contribution , possibly from the hot corona with @xmath28 surrounding the jet , is required to account simultaneously for the synchrotron and the inverse compton emissions during those epochs . moreover , a detailed analysis of the correlation between the flux variations in the and optical energy band yields a possible @xmath29 day delay of the emission with respect to the optical one .
starting from may 2008 , conducted several multi wavelength observing campaigns on .
a detailed discussion of these results is presented in @xcite .
figure [ fig:3c454:18months ] shows the light curves at different energies over the whole period july 2007 january 2009 .
the different panels show , from bottom to top , the /grid light curve at @xmath30-day resolution for e@xmath31100 mev in units of @xmath32 , the /bat light curve in the energy range 15150 kev at @xmath33-week resolution , the /xrt light curve in the energy range 210 kev , the /uvot light curve in the uv @xmath34 filter , the /uvot light curve in the optical @xmath35 filter , the gasp - webt light curve in the optical @xmath1 filter , and the vlbi radio core at 15 ghz ( filled circles ) and the umrao 14.5 ghz ( open squares ) light curves , respectively .
we investigated the correlation between the flux and the optical flux density in the @xmath1 band by means of the dcf method , computing the dcf on four distinct periods : july 2007 ( mid 2007 ) , november december 2007 ( fall 2007 ) , may august 2008 ( mid 2008 ) , and october 2007 - january 2009 ( fall 2008 ) .
the period `` fall 2007 '' offers a good opportunity to test the correlation , since the @xmath24-ray flux , and even more the optical flux , exhibited strong variability , and the period of common monitoring lasted for more than a month .
the corresponding dcf shows a maximum @xmath36 for a null time lag .
however , the shape of the peak is asymmetric , and a centroid calculation yields a time lag of @xmath37days .
this result is in agreement with what was found by @xcite and by @xcite .
since july 2007 , has been playing the same role for as 3c 279 had for egret , and during the period july 2007 - january 2009 we acquired data not only in the energy band , but across 14 decades in energy .
this allowed us to construct simultaneous seds , sampling high , intermediate , and low emission states , involving both ground and space based observatories .
we found that the role of the external compton on the disk and the broad - line region radiation ( and possibly also on hot corona photons ) is crucial to account for the hard spectrum states .
moreover , thanks to the extremely dense optical coverage provided by the gasp - webt , we were able to study the correlations between the and optical fluxes .
we found a @xmath38 day possible lag of the high energy photons with respect to the optical ones .
the simultaneous presence of two satellites , and fermi , the extremely prompt response of wide - band satellites as , and the long - term monitoring provided from the radio to the optical by the gasp - webt consortium will assure the chance to investigate and study the physical properties of and of several more blazars both at high and low emission states .
i am grateful to l. maraschi and to all the como workshop soc and loc members for having organized such an excellent and fruitful meeting .
the results presented here were obtained in collaboration with the agile agn working group ( a. bulgarelli , a.w .
chen , f. dammando , i. donnarumma , a. giuliani , f. longo , l. pacciani , g. pucella , v. vittorini , and m. tavani ) , the webt - gasp team ( in particular m. villata and c.m .
raiteri ) , and the team ( in particular p. romano , h. krimm , n. gehrels , the duty scientists , and science planners ) .
the agile mission is funded by the italian space agency ( asi ) with scientific and programmatic participation by the italian institute of astrophysics ( inaf ) and the italian institute of nuclear physics ( infn ) .
s.v . acknowledges financial support by the contracts asi i/088/06/0 and asi i/089/06/0 . | during the period july 2007 - january 2009 , the agile satellite , together with several other space- and ground - based observatories monitored the activity of the flat - spectrum radio quasar 3c 454.3 , yielding the longest multiwavelength coverage of this quasar so far .
the source underwent an unprecedented period of very high activity above 100 mev , a few times reaching flux levels on a day time scale higher than @xmath0 , in conjunction with an extremely variable behavior in the optical @xmath1-band , even of the order of several tenth of magnitude in few hours , as shown by the gasp - webt light curves .
we present the results of this long term multiwavelength monitoring campaign , with particular emphasis on the study of possible lags between the different wavebands , and the results of the modeling of simultaneous spectral energy distributions at different levels of activity . |
[ cols="<,^ , < " , ] trim conditions for the aircraft at 20 m/s are : @xmath0 , @xmath1 and @xmath2 and have been used as initial conditions for the controller . again a prediction window of 5 seconds along with 50 coincidence points were used . the upper limit on the thrust constraint continually changes based on the true airspeed of the aircraft according to the equations given in . the minimum thrust level set always to @xmath3 . if however an engine failure
is detected the upper limit on thrust is set to the filter estimate plus @xmath4 uncertainty . +
a ukf filter was designed to perform fdi with the following process noise and noise covariance matrices : @xmath5 where @xmath6 is the filter update rate 0.01 secs .
the initial state vector and covariance matrix are : @xmath7^\intercal , \quad \mathbf{p}(0 ) = \begin{bmatrix } ( 0.5)^2 & 0 & 0 & 0\\ 0 & ( 0.5)^2 & 0 & 0\\ 0 & 0 & ( 0.0850)^2 & 0\\ 0 & 0 & 0 & ( 6)^2 \end{bmatrix}.\ ] ] finally the fault detection logic is based on that given in @xcite . the following scenarios were set up to test the active ftc system on the uav model : 1 .
no fault case 2 .
engine failure - @xmath8 power loss 20 seconds into flight , 3 .
engine failure - @xmath9 power loss 30 seconds into flight .
the aircraft was required to follow the flight trajectory given in figure [ fig : chap6_activeftc_reftraj ] ( note : wind effects have been taken into account using the dryden wind model in matlab ) .
figure [ fig : chap6_activeftc_faultflag ] presents the fault detection logic for each scenario .
there was no fault present in scenario 1 and this is reflected in the results of the fault detection logic as the value of the fault flag remained zero throughout the duration of the flight .
the results for scenarios 2 and 3 show that the fault flag is triggered ( ie .
the value of the flag switches to 1 ) within seconds of the fault occurring .
the control inputs produced by the controller for each scenario are given in figures [ fig : chap6_activeftc_controls_nf ] , [ fig : chap6_activeftc_controls_50 t ] and [ fig : chap6_activeftc_controls_30 t ] .
the results clearly indicate that the controller is able to successfully reconfigure based on fdi data and as the loss of power increases the demand on the elevator increases .
loss of power - controls ] loss of power - controls ] the true airspeed response of the aircraft is given in figure [ fig : chap6_activeftc_controls_vt ] .
the demanded speed was @xmath10 and the results show that for a @xmath8 loss in power the true airspeed demand is unachievable during climb to altitude .
however during straight and level flight the velocity demand is gradually achieved and is finally reached once the aircraft begins the descent phase . during the descent
the true airspeed demand is similar to the no fault case . in the case of @xmath9 power loss
the aircraft is unable to meet the true airspeed demand during straight and level flight , however halfway through the descent phase the aircraft picks up speed and is able to maintain the reference . in any given scenario stall speed is never reached . .
stall speed did not occur in any scenario . ]
the climb rate ( or vertical speed ) response is shown in figures [ fig : chap6_activeftc_controls_vd ] and [ fig : chap6_activeftc_controls_vd_subs ] and shows that during power loss the vertical speed oscillates between the upper and lower constraints values .
this is to be expected as the elevator is working harder to regulate the speed , and hence bounces between the two limits . ] ] finally the trajectories flown by the aircraft in each scenario are shown in figure [ fig : chap6_activeftc_controls_height ] .
the interesting points to note here are that in the highest loss of power case ( scenario 3 ) the aircraft did not have enough power to reach the highest flight altitude so instead cruised at an altitude it was capable of flying . once the straight and level phase of the flight was over the controller was able to fly the aircraft back onto the demanded trajectory .
the @xmath8 loss of power case shows that aircraft still had enough power to fly back onto the demanded path halfway through the straight and level at altitude flight phase .
in this paper we demonstrated the successful application of an nmpc based active ftc design on an actual uav which is currently in operation .
the controller was implemented using actual aircraft data and shows great promise for active fault tolerant flight control .
an appendix , if needed , appears before the acknowledgements .
an acknowledgements section , if used , immediately precedes the references .
sponsorship and financial support acknowledgements should be included here .
9 kale , m. m. , chipperfield , a.j . , `` robust and stabilized mpc formulations for fault tolerant and reconfigurable flight control , '' _ proceedings of the 2004 ieee international symposium on intelligent control _ , 2004 , pp . 222227 ,
doi : 10.1109/isic.2004.1387686 vos , david w. , motazed , b. , `` application of fault - tolerant controls to uavs , '' _ aerospace / defense sensing and controls _ , 1996 , pp .
khan , r. , williams , p. , riseborough p. , rao , a. and hill , r. , _ designing a nonlinear model predictive controller for fault tolerant flight control_. arxiv e - prints 2016 ; https://arxiv.org/abs/1609.01529 , [ last accessed : 02 - 10 - 2016 ] .
khan , r. , williams , p. , riseborough p. , rao , a. and hill , r. , _ fault detection and identification - a filter investigation_. arxiv e - prints 2016 ; https://arxiv.org/abs/1609.02407 , [ last accessed : 02 - 10 - 2016 ] .
boskovic , j.d .
, sai - ming l. and mehra , r.k .
, `` on - line failure detection and identification ( fdi ) and adaptive reconfigurable control ( arc ) , '' _ proceedings of the 2001 american control conference _ , vol . 4 , 2001 , pp . 26252626 ,
doi : 10.1109/acc.2001.946269 gopinathan , m. , boskovic , j.d .
, mehra , r.k . and rago , c. , `` a multiple model predictive scheme for fault - tolerant flight control design , '' _ proceedings of the 37th ieee conference on decision and control _ , vol . 2 , 1998 , pp . 13761381 ,
doi : 10.1109/cdc.1998.758477 maciejowski , j.m .
, `` fault - tolerant aspects of mpc , '' _ iee two - day workshop on model predictive control : techniques and applications - day 2 ( ref
1999/096 ) _ , 1999 , pp . 1/11/4 .
bateman , f. , noura , h. , and ouladsine , m. , `` actuators fault diagnosis and tolerant control for an unmanned aerial vehicle , '' _ ieee international conference on control applications .
cca 2007 _ , 2007 , pp .
bateman , f. , noura , h. , and ouladsine , m. , `` a fault tolerant control strategy for an unmanned aerial vehicle based on a sequential quadratic programming algorithm , '' _
47th ieee conference on decision and control .
cdc 2008 _ , 2008 , pp .
copeland , r.p . and rattan , k.s .
, `` a fuzzy logic supervisor for reconfigurable flight control systems , '' _ proceedings of the 1994 ieee national aerospace and electronics conference ( naecon ) _
, 1994 , pp .
579586 . chen , j. , patton , r.j . and
chen , z. , `` an lmi approach to fault - tolerant control of uncertain systems , '' _ proceedings of the 1998 intelligent systems and semiotics ( isas ) and intelligent control ( isic ) . held jointly with ieee international symposium on computational intelligence in robotics and automation ( cira ) _ , 1998 , pp .
wu , s. , grimble , m.j . , wei , w. , `` qft based robust / fault tolerant flight control design for a remote pilotless vehicle , '' _ proceedings of the 1999 ieee international conference on control applications _ , vol . 1 , 1999 ,
beainy , f. , mai , a. and commuri , s. , `` unmanned aerial vehicles operational requirements and fault - tolerant robust control in level flight , '' _ 17th mediterranean conference on control and automation .
med09 _ , 2009 , pp .
krger , t. , schnetter , p. , placzek , r. and vrsmann , p. , `` fault - tolerant nonlinear adaptive flight control using sliding mode online learning , '' _ neural networks _ , vol .
32 , 2012 , pp .
khan , r. , williams , p. , riseborough p. , rao , a. and hill , r. , _ active fault tolerant flight control system design_. arxiv e - prints 2016 ; https://arxiv.org/abs/1610.02139 , [ last accessed : 11 - 10 - 2016 ] .
williams , p. , `` uav model , ''
_ private communication _ , full details of the uav model can not be disclosed due to a confidentiality condition , 2015 mccormick , b.w .
, `` aerodynamics , aeronautics and flight mechanics , '' _ john wiley & sons inc , usa , ny _ , 1995 . | fault tolerance is achieved through multiply redundant hardware systems in large civil aircraft .
this means of achieving fault tolerance is infeasible for small compact unmanned aerial vehicles . in this paper
we apply a fault tolerant control system which exploits analytical redundancy rather than hardware redundancy to an actual uav model currently in operation via model - in - the - loop simulation .
the fault tolerant control system comprises a nonlinear model predictive controller integrated with an unscented kalman filter for fault detection and identification .
the results show that our fault tolerant control system design is able to identify engine failure within seconds of fault occurrence and distribute control authority to the healthy actuators to maintain safe flight . |
the new family of the charmonium or charmonium - like states include @xmath5 , @xmath6 , @xmath7 , @xmath8 , @xmath9 , @xmath10 , @xmath11 , @xmath12 , @xmath13 , @xmath14 and @xmath15 etc @xcite .
many states sit on the the threshold of two charmed mesons , which inspired some of them ( especially those charged ones ) to be candidates of heavy moleculues @xcite . in the heavy quark limit
, the s - wave and p - wave heavy mesons can be categorized into three doublets : @xmath16 , @xmath17 , @xmath18 .
we collect their masses from pdg in table [ hs ] .
the bottom mesons in the @xmath19 doublet are still missing experimentally .
thus , we will adopt the theoretical predictions of the bottom meson masses in the @xmath19 doublet when we study the heavy flavor molecular system composed of the bottom and anti - bottom mesons . in the framework of the meson exchange model
, we have investigated the possible loosely bound molecular states composed of a pair of heavy mesons in refs .
@xcite . in this work
, we will investigate the possible heavy molecular system constructed by the charmed and anti - charmed mesons , where one meson is in the @xmath20 doublet and the other one is in the @xmath19 doublet . in the following ,
we denote the heavy flavor molecular system as the @xmath21 system for the convenience .
the @xmath22 system can be categorized into four subsystems : @xmath23 $ ] , @xmath24 $ ] , @xmath25 $ ] and @xmath26 $ ] .
they correspond to different quantum number combinations @xmath27 , @xmath28 , @xmath29 and @xmath30 , respectively . since charmed mesons belong to the fundamental representation of flavor @xmath31 , the system constructed by the charmed meson and anti - charmed meson forms an octet and a singlet : @xmath32 as illustrated in table [ wavefunction ] .
the parameter @xmath33 in the flavor wave functions corresponds to the charge parity @xmath34 respectively for the neutral systems as pointed out in refs .
@xcite .
this paper is organized as follows .
we review the formalism in section [ sec2 ] and present the results in section [ sec3 ] .
the last section is a short summary .
the potential model is an effective approach to study the two - body bound state problem . for the @xmath22 system
, the scattering between the charmed and anti - charmed meson occurs via exchanging the light pseudoscalar , scalar and vector mesons , which play the role of providing long - distant , intermediate - distance and short - distance forces . at the hadron level , there exist two types of diagrams in the scattering of the charmed and anti - charmed mesons , i.e. , the cross and direct diagrams as shown in table [ diagrams ] .
the exchanged light mesons relevant to the four subsystems @xmath23 $ ] , @xmath24 $ ] , @xmath35 $ ] and @xmath36 $ ] are also presented in table [ diagrams ] .
when writing out the scattering amplitude , the monopole form factor is introduced at every interaction vertex to compensate the off - shell effect of the exchanged light meson @xmath37 where the phenomenological cutoff parameter @xmath38 is about 1 gev . @xmath39 and
@xmath40 denote the four - momentum and the mass of the exchanged meson .
.the flavor wave functions of the @xmath23 $ ] , @xmath24 $ ] , @xmath35 $ ] and @xmath36 $ ] systems .
we use @xmath69^{a}_{b}$ ] to name the different states , where the superscript @xmath70 denotes the charge of the state and subscript @xmath71 is introduced to distinguish the different states in a subsystem .
[ wavefunction ] [ cols="^,^",options="header " , ] | we study the possible heavy molecular states composed of a pair of charm mesons in the h and s doublets . since the p - wave charm - strange mesons @xmath0 and
@xmath1 are extremely narrow , the future experimental observation of the possible heavy molecular states composed of @xmath2 and @xmath3 may be feasible if they really exist .
especially the possible @xmath4 states may be searched for via the initial state radiation technique . |
communication is the exchange of physical systems aimed at establishing correlations between the communicating parties . in most circumstances we are interested in the total amount of correlations , _
i.e. _ the mutual information , between the sender and receiver @xcite .
information theory can thus be used to make statements governing any communication process @xcite .
one of the most fundamental of them claims that the gain of information achieved by communicating parties can not exceed the amount of information that is actually transmitted @xcite . beside agreeing fully with everyday experience ,
this adheres well with the expectation that in a communication protocol no form of correlation can be established without transmitting that kind of correlation .
remarkably , quantum physics defies our intuition in this case too . as shown for the first time by cubitt _
et al . _
@xcite , quantum entanglement @xcite , which is a purely non - classical type of correlation enabling tasks such as quantum teleportation @xcite , secure cryptography @xcite , improved communication complexity @xcite , and quantum dense coding @xcite , can be established between distant parties by exchanging a carrier system that , at all times , is not entangled with them . with the carrier system @xmath0 , which is then sent to bob s site .
if all the particles are quantum , it is possible to establish entanglement between their respective laboratories even though there was no initial entanglement between them and no entanglement is communicated .
this is accomplished as follows : in step ( 1 ) , the fully separable initial state of the three systems is prepared . in step
( 2 ) alice applies a suitable operation on @xmath1 and @xmath0 , which keeps the latter separable from the rest of the systems but creates entanglement between @xmath1 and joint system made out of @xmath2 and @xmath0 together . in step ( 3 ) , the unentangled carrier @xmath0 is transmitted to bob .
as shown in panel ( 4 ) , this establishes entanglement between the laboratories of alice and bob . ]
this result is particularly relevant for designing efficient methods to distribute entanglement across networks of non - interacting quantum systems @xcite .
two protocols have been identified to achieve this goal : the transfer of pre - available entanglement to chosen nodes of the network @xcite , and the quantum communication scenario discussed above based on the exchange of a carrier quantum system . in the latter case , all the quantum algorithms relying on entanglement
can actually be implemented without communicating any entanglement . in such a scenario , the gain in entanglement between communicating sites is bounded from above @xcite by the amount of communicated nonclassical correlations as measured by quantum discord @xcite .
discord can be present also in separable states giving rise to the possibility of entanglement distribution via separable states . here
we demonstrate this concept experimentally by implementing the protocol proposed in ref .
@xcite on a quantum optical platform based on the use of single - photon polarisation qubits . using the polarisation degree of freedom of a three - photon quantum register
, we show that an initially separable , yet discorded , state of two photons can be transformed into an entangled one using only local operations that never generate entanglement between themselves and the communicated carrier photon .
our experiment proves the tantalising possibility to establish quantum channels among the nodes of a quantum network without the need of previously available entangled resources .
a typical communication scenario is depicted in fig .
[ fig_comm ] .
each of two distant parties , alice and bob , holds a quantum system , which we label @xmath1 and @xmath2 respectively . under the transmission of a carrier quantum system @xmath0 ,
encoding communication from alice to bob , the information gain between their respective laboratories satisfies the inequality @xcite @xmath3 where @xmath4 is the information between alice and bob after bob has received system @xmath0 , @xmath5 is the information shared before the communication protocol , and @xmath6 is the actual information carried by @xmath0 . by introducing the quantum mutual information @xmath7 between two general systems @xmath8 and @xmath9 we can identify @xmath10 , @xmath11 , and @xmath12 . as expected , the gain @xmath13 can not exceed the information carried by system @xmath0 , which is the communicated information . a similar relation for quantum entanglement @xmath14 can not be stated : the entanglement gain is not bounded by the communicated entanglement but rather by the communicated quantum discord , a more general type of quantum correlations @xcite . in particular , we can write @xmath15 where , using a notation similar to the one introduced before and indicating with @xmath16 the relative entropy of entanglement between systems @xmath8 and @xmath9 @xcite , @xmath17 and @xmath18 refer to situations before and after the communication process , respectively . with these choices , in eq . the communicated discord @xmath19 is quantified by the relative entropy of discord @xcite , which is also known as the one - way quantum deficit @xcite .
the fact that discord quantifies a more general type of correlation opens up the possibility to create entanglement between remote parties via separable states alone .
indeed , the bound in eq .
( [ ent_gain ] ) is achieved in the protocol presented in ref .
@xcite , albeit in general the communicated discord does not provide a tight bound on the entanglement that is correspondingly gained @xcite . to the best of our knowledges such
a general tight bound is currently unknown . in the protocol realised here , the two - level particles @xmath1 and @xmath2
are prepared in a separable state @xmath20 that is a mixture of the four bell states @xmath21 ( each occurring with probability @xmath22 ) and @xmath23 ( with probability of occurrence @xmath24 ) .
such a state is separable if and only if the highest probability in the mixture does not exceed @xmath25 @xcite .
the two - level carrier system @xmath0 is initially with alice and is uncorrelated from the other systems , so that the overall initial state is taken as @xmath26 . here
@xmath27 with @xmath28 the identity matrix , @xmath29 the pauli @xmath30 matrix , and @xmath31 $ ] .
alice now generates the state @xmath32 by applying a controlled - phase gate @xmath33 on her systems .
the carrier qubit should remain separable from the other systems , _
i.e. _ we require @xmath34 , while system @xmath1 should become entangled with the subsystem composed of @xmath2 and @xmath0 ( that is , we should have @xmath35 ) .
finally , system @xmath0 is transmitted to bob and in this way the laboratories of alice and bob share entanglement .
the description so far leaves room for the choice of the initial state , which should be taken as one that guarantees a sizeable degree of entanglement in the @xmath36 bipartition , keeping @xmath0 separable , after the protocol .
a possible instance is given by the @xmath37 separable state @xmath38 which is a mixture of two - qubit states formed by the eigenstates @xmath39 of pauli operators @xmath40 , with eigenvalue @xmath41 .
as a measure of entanglement we use the negativity @xmath42 @xcite , which is defined as the most negative eigenvalue of the matrix obtained from @xmath43 under partial transposition of system @xmath1 @xcite . within the class of initial states @xmath44
on which one applies controlled - phase gate , the state built using eq .
and @xmath45 gives @xmath46 , which is the highest amount of distributed entanglement via separable states .
we choose to focus on the negativity because its presence in state @xmath43 guarantees that ( i ) the entanglement established between the sending and receiving laboratories can be localised into entanglement between systems @xmath1 and @xmath2 only using local operations performed at bob s site @xcite , and ( ii ) such localised entanglement is distillable @xcite . therefore , by repeating this protocol a sufficient number of times and performing entanglement distillation , one can in principle obtain maximally entangled pairs between alice and bob , although no entanglement has been shared between them .
-barium borate crystal ( bbo ) pumped by a frequency - doubled femtosecond ti : sapphire laser at @xmath47 nm .
one photon serves as a trigger , while the other three are initialised with polarising beamsplitters ( pbs ) , half - wave ( hwp ) and quarter - wave plates ( qwp ) .
the photons representing systems a and c are subjected to a probabilistic controlled - phase gate based on two - photon interference at a partially polarising beamsplitter ( ppbs ) @xcite .
projective measurements are made with a combination of hwp , qwp and pbs , and detected by single - photon avalanche photodiodes ( apd ) connected to a coincidence logic . ]
the circuit diagram in fig .
[ fig_setup]a shows the conceptual implementation of the described protocol , while the experimental setup is shown in fig .
[ fig_setup]b . using four single photons ( one as a trigger , the other three as the qubits @xmath1 , @xmath2 , and @xmath0 ) , we prepare the discorded state @xmath20 and the mixed state @xmath48 by summing up individual pure - state terms , with measurement acquisition times corresponding to the weights in eq .
( similar technique has been used _
e.g. _ in ref .
this approach guarantees that the initial state is separable .
systems @xmath1 and @xmath0 are subjected to a photonic controlled - phase gate @xcite prior to performing a full three - qubit quantum state tomography on the output state @xmath49 @xcite .
the total integration time was @xmath50 hours , during which we counted @xmath51@xmath52 four - fold coincidence events .
the experimentally obtained density matrix , whose fidelity with the ideal state is @xmath53 , is shown in fig . [ fig_tomography ] . to estimate the uncertainty
, we generate @xmath54 samples with poisson - distributed random noise added to the raw data .
the corresponding reconstructed density matrices are used to evaluate an average fidelity of @xmath55 , which is extremely close to the experimental value .
, obtained via three - qubit state tomography .
( b ) ideal density matrix @xmath49 . ]
the next step is to use the density matrix to calculate the entanglement in the relevant bipartitions : for the bipartitions @xmath56 and @xmath57 , which are separable according to the theoretical protocol , the smallest eigenvalues of the partially transposed density matrices should be exactly zero . in practice , due to experimental imperfections and the statistics of limited photon count rates ,
it is likely that such eigenvalues for the experimentally reconstructed states could have a ( small ) negative nominal value , which would make such instances compatible with entangled bipartitions . to conclusively prove that the bipartitions @xmath56 and @xmath57 are separable , we add increasing amounts of white noise to the initial state , which thus becomes @xmath58)$ ] .
white noise contributes no correlations , so this raises the value of the smallest eigenvalues of partially transposed density matrices , as shown in fig .
[ fig_evalues]a .
a similar method was previously studied to assess the generation of bound - entangled states @xcite . for no added noise , the smallest eigenvalues of the partially transposed experimental states suggest that the protocol is successful : @xmath59 is clearly negative and both @xmath60 and @xmath61 are non - negative . however , as shown by the shaded bars in fig .
[ fig_evalues]a , out of the @xmath54 states simulated for the error estimation only 17.4% have the required features at the relevant bipartitions . fig .
[ fig_evalues]b shows that the proportion of such states rises rapidly with the addition of only a small amount of white noise : 96.5% of the sampled matrices fulfil the requirements for a successful protocol already at @xmath62 . note that this is a conservative estimate : as the histograms in fig .
[ fig_evalues]c make clear , the data ( indicated by the solid coloured bars ) consistently lie in the upper part of the estimated error range .
this skew can be understood by considering the fidelity of the experimental state at @xmath62 with the ideal state , @xmath49 .
as the nominal value of the fidelity is very close to unity ( @xmath63 ) , simulating a poisson distribution with limited count statistics is likely to lead to lower fidelities , pushing the estimates to below the measured values , see fig .
[ fig_evalues]d .
the minimum eigenvalues of the partially transposed experimental states at maximum amount of added noise are @xmath64 , @xmath65 , and @xmath66 . by repeating the limited - count analysis in an otherwise ideal case , in the appendix 3 we show that the skew and the clear splitting between the positive values seen in fig .
[ fig_evalues]a are not artefacts of imperfect state preparation or gate operation . moreover , in order to exclude the possibility of entanglement distribution via bound entangled states @xcite , in appendix 2 we also provide the explicit decomposition of the experimental states in terms of convex sums of separable states of the @xmath57 bipartition .
we have experimentally demonstrated that distillable entanglement can be established between parties who exchange only unentangled systems . the success of our protocol is confirmed by the thorough data analysis performed in our study , which proves unambiguously the entangled nature of the @xmath67 bipartition and the separability of the other two . needless to say ,
an equally interesting albeit weaker statement on entanglement distribution via bound - entangled states would be possible by having @xmath57 a bipartition with positive partial transposition yet not separable .
our experiment demonstrates that entanglement - based quantum information protocols can be executed without communication of entanglement and might pave the way to bypassing the detrimental effects of noisy mechanisms affecting a quantum channel .
in fact , the transmission of separable states would result in more robust protocols than those based on communicated quantum entanglement .
we would like to thank k. kristinsson , p. raynal , and k. yczkowski for discussions .
mz and tp are supported by the national research foundation and ministry of education in singapore , and by the start - up grant of nanyang technological university .
mp thanks the uk epsrc for financial support through a career acceleration fellowship and a grant under the new directions for epsrc research leaders " initiative ; af and ma for support by australian research council discovery early career awards , de130100240 and de120101899 , respectively ; agw acknowledges support from a uq vice - chancellor s senior research fellowship . _ note added_. during the completion of this work we became aware of an independent demonstration of the phenomenon discussed here based on the use of continuous variable systems @xcite . [ [ section ] ]
here we provide a detailed analysis of the techniques used to reveal the separable nature , across the @xmath57 bipartition , of the states that have been produced experimentally .
we begin with the ideal theoretical case and later will apply some of the techniques discussed here to the experimental density matrices .
consider the initial state of system formed by qubits @xmath1 and @xmath2 @xmath68 where @xmath69 ( @xmath70 ) denotes horizontal ( vertical ) polarisation state , @xmath71 ( @xmath72 ) denotes @xmath73 ( @xmath74 ) linear - polarisation state , and @xmath75 ( @xmath76 ) denotes right ( left ) circular polarisation state @xmath77 these are the embodiment of the logical states @xmath78 , @xmath79 , and @xmath80 introduced in the main manuscript .
the carrier qubit is initially in the state : @xmath81 after applying a controlled - phase gate on qubits @xmath1 and @xmath0 , the overall system s state becomes @xmath82 where the density matrix is written in the standard basis and the dots represent zeros .
this state is entangled across the cut @xmath67 as revealed by the partial transposition criterion : the smallest eigenvalue of the corresponding partially transposed density matrix equals @xmath83 .
the state is also separable across the @xmath57 cut , as shown by the following explicit decomposition into product states for this bipartition where @xmath84 @xmath85 represents the standard bell basis and @xmath86 .
in the experiment , additionally to producing states close to the ideal state @xmath49 , we also prepared the set of states with increasing amount of admixed white noise .
all of such states are entangled in the cut @xmath67 , as demonstrated in the main text by the existence of a negative eigenvalue in the spectrum of the matrix obtained after partial transposition of @xmath1 .
figure 4a of the main text also shows that the other two cuts are associated with states having positive eigenvalues after partial transpositions . as this criterion is not a necessary and sufficient one for three - qubit states
, this does not exclude the possibility of having bound entanglement in the one of the cuts that is positive under partial transposition .
therefore , in order to exclude the possibility of performing entanglement distribution via bound entangled states , we explicitly show separability across the bipartition @xmath57 by constructing a separable decomposition of the corresponding states , in analogy with eq . . to this end , we use the following algorithm : 1 .
we generate a set of random product states for the chosen cut and use them to complement the set of product state vectors that enter the decomposition given in eq . .
we write a separable state as @xmath87 where @xmath88 is a label for the chosen separable states @xmath89 discussed at step 1 ) , and @xmath90 are the associated probabilities of occurrence ( @xmath91 ) .
we equate this expression with the experimental state and numerically solve for @xmath90 s . only about @xmath92 product states are sufficient to find explicitly the separable decompositions of all experimental density matrices reported in the main manuscript and thus wash out any possibility for bound entanglement .
the protocol is thus faithfully based on the use of separable states .
our error analysis showed that our data points , fig .
4 , lie slightly outside the most likely range obtained via the standard method for single - photon experiments : maximum - likelihood estimation of states with simulated poissonian noise added to the obtained photon counts .
in addition , we observe that the minimum eigenvalues for the unentangled bipartitions split , with @xmath56 having a more negative bias than @xmath57 . in fig . 5
, we present results of a similar analysis starting with ideal states and an ideal gate operation .
these numerical simulations show the same effects , highlighting that they are not an artefact of non - ideal state preparation or gate imperfections , but solely due to statistics based on limited photon counts . for different cuts of the density matrix .
the dashed ( dash - dotted ) black lines show the theoretical values for the unentangled , @xmath57 and @xmath56 , ( entangled @xmath67 ) bipartitions , assuming an infinite number of counts .
the solid blue , green and red lines show the eigenvalues obtained from reconstructed density matrices starting with the ideal theory states but simulated with equivalent finite count statistics to those in the experiment .
the shaded regions show the one - standard - deviation range obtained from adding poissonian counting statistics to the ideal states .
we observe a similar offset from theoretical states simulated with infinite counts , as in the experimental data ( see fig.4 in main text ) , including the splitting of the eigenvalues for the unentangled cuts .
each data point is calculated from @xmath93 simulated density matrices at @xmath94 separate noise values.,scaledwidth=48.0% ] t. m. cover and j. a. thomas , _ elements of information theory _
( john wiley and sons , ny , 1991 ) .
m. a. nielsen and i. chuang , _ quantum computation and quantum information _
( cambridge university press , cambridge , 2000 ) . m. pawowski , t. paterek , d. kaszlikowski , v. scarani , a. winter , and m. zukowski , nature * 461 * , 1101 ( 2009 ) .
r. horodecki , p. horodecki , m. horodecki , and k. horodecki , rev .
phys . * 81 * , 865 ( 2009 ) . c. h. bennett , g. brassard , c. crepeau , r. jozsa , a. peres , and w. k. wootters , phys .
lett . * 70 * , 1895 ( 1993 ) .
a. k. ekert , phys .
lett . * 67 * , 661 ( 1991 ) .
h. buhrman , r. cleve , s. massar , and r. de wolf , rev .
phys . * 82 * , 665 ( 2010 ) . c. h. bennett and s. j. wiesner , phys .
lett . * 69 * , 2881 ( 1992 ) . h. j. kimble ,
nature * 453 * , 1023 ( 2008 ) .
b. kraus and j. i. cirac , phys .
* 92 * , 013602 ( 2004 ) .
m. paternostro , w. son , and m. s. kim , phys .
lett . * 92 * , 197901 ( 2004 ) .
a. streltsov , h. kampermann , and d. bru , phys .
* 108 * , 250501 ( 2012 ) .
l. henderson and v. vedral , j. phys .
theor . * 34 * , 6899 ( 2001 ) .
h. ollivier and w. h. zurek , phys .
lett . * 88 * , 017901 ( 2002 ) .
a. kay , phys .
lett . * 109 * , 080503 ( 2012 ) .
v. vedral , m. b. plenio , m. a. rippin , and p. l. knight , phys .
lett . * 78 * , 2275 ( 1997 ) .
k. modi , t. paterek , w. son , v. vedral , and m. williamson , phys .
lett . * 104 * , 080501 ( 2010 ) .
m. horodecki , p. horodecki , r. horodecki , j. oppenheim , a. sen(de ) , u. sen , and b. synak - radtke , phys .
a * 71 * , 062307 ( 2005 ) .
r. horodecki and m. horodecki , phys .
a * 54 * , 1838 ( 1996 ) .
g. vidal and r. f. werner , phys .
a * 65 * , 032314 ( 2002 ) .
a. peres , phys . rev .
lett . * 77 * , 1413 ( 1996 ) .
m. horodecki , p. horodecki , and r. horodecki , phys .
* 78 * , 574 ( 1997 ) .
n. k. langford , t. j.weinhold , r. prevedel , k. j. resch , a. gilchrist , j. l. obrien , g. j. pryde , and a. g. white , phys .
. lett . * 95 * , 21 ( 2005 ) .
j. lavoie , r. kaltenbaek , m. piani , and k. resch , phys .
* 105 * , 130301 ( 2010 ) .
d. f. v. james , p. g. kwiat , w. j. munro , and a. g. white , phys .
a * 64 * , 52312 ( 2001 ) .
m. horodecki , p. horodecki , and r. horodecki , phys .
80 * , 5239 ( 1998 ) . c. e. vollmer , d. schulze , t. eberle , v. handchen , j. fiurasek , and r. schnabel , arxiv:1303.1082 ( 2013 ) | information gain in communication is bounded by the information encoded in the physical systems exchanged between sender and receiver .
surprisingly , this does not hold for quantum entanglement , which can increase even though the communicated system carries no entanglement at all . here
we demonstrate this phenomenon in a four - photon experiment where two parties sharing initially separable ( unentangled ) state get entangled by exchanging a photon that is _ at all times _ not entangled with either of them .
our result validates a long - standing assert in quantum information and has important practical implications in quantum networking , where entanglement must be reliably distributed across many nodes at low resource - cost . |
classical and statistical thermodynamics deal with the equilibrium states of matter @xcite . driving the system from one equilibrium state to another
, however , involves the passage of the system through a sequence of instantaneous states that do not satisfy the conditions for thermodynamic equilibrium , and hence , constitute a non - equilibrium process @xcite .
the description of these processes fall outside the realm of classical and statistical thermodynamics , unless the sequence of non - equilibrium states do not depart appreciably from a sequence of equilibrium states .
such idealized process can be thought of as an infinite sequence of infinitesimally small changes in the driving control parameter , after each of which the system is given sufficient time to equilibrate .
this so - called quasistatic process is an excellent representation of real process when the equilibration times of the system are sufficiently short .
however , when the equilibration kinetics is very slow , virtually any change will involve intrinsically non - equilibrium states whose fundamental understanding must unavoidably be done from the perspective of a non - equilibrium theory @xcite .
these concepts become particularly relevant for the description of the slow dynamics of metastable glass - forming liquids in the vicinity of the glass transition @xcite .
it is well known that the decay time of the slowest relaxation processes ( the so - called @xmath4-relaxation time @xmath16 ) increases without bound as the temperature @xmath17 is lowered below the glass transition temperature @xmath18 .
it is then natural to think that the _ equilibration _ time of the system must also increase accordingly . to be more precise ,
let us imagine that a glass - forming liquid , initially at an arbitrary temperature @xmath19 , is suddenly cooled at time @xmath20 to a final temperature @xmath17 , after which it is allowed to evolve spontaneously toward its thermodynamic equilibrium state .
imagine that we then monitor its @xmath4-relaxation time @xmath21 as a function of the evolution or `` waiting '' time @xmath3 elapsed after the quench .
we say that the system has equilibrated when @xmath21 reaches the plateau that defines its final equilibrium value @xmath22 , which must only depend on the final temperature @xmath17 .
the beginning of this plateau occurs at a certain value of the waiting time @xmath3 , that we refer to as the _ equilibration _ time @xmath23 ; this equilibration time must also depend on the final temperature @xmath17 .
there are strong indications , from recent computer simulation experiments @xcite , that in the metastable regime these two characteristic times , @xmath22 and @xmath23 , are related to each other as @xmath24^\eta$ ] , with an exponent @xmath25 ( more specifically , @xmath26 @xcite ) .
this implies that in order to measure the actual equilibrium value @xmath22 we have to wait , before starting the measurement of @xmath22 , for an equilibration time @xmath23 that will increase faster than @xmath22 itself .
this poses an obvious practical problem for the measurement of @xmath22 when the temperature @xmath17 approaches the glass transition temperature @xmath18 , since sooner or later we shall be unable to wait this required equilibration time .
this situation then implies that it is impossible to discard a scenario in which the _ equilibrium _ @xmath4-relaxation time @xmath22 diverges at a singular temperature @xmath27 , since the equilibration time @xmath23 needed to observe this divergence will also diverge at that temperature , i.e. , it will be impossible to equilibrate the system at a final temperature near or below @xmath27 within experimental waiting times . of course
, a measurement carried out at a finite @xmath3 , will always report a result for @xmath16 , but this result will correspond to @xmath21 , the non - equilibrium value of the @xmath4-relaxation time registered at that waiting time @xmath3 .
thus , the analysis of these experimental measurements can not be based on the postulate that the system has reached equilibrium ; instead , one needs to interpret these experiments in the framework of a quantitative theory of slowly - relaxing non - equilibrium processes . until recently
, however , no _ quantitative , first - principles _ theory had been developed and applied to describe the slow _ non - equilibrium _ relaxation of _ structural _ glass - forming atomic or colloidal liquids . about a decade ago latz @xcite attempted to extend the conventional mode coupling theory ( mct ) of the ideal glass transition @xcite , to describe the aging of suddenly quenched glass forming liquids .
a major aspect of his work involved the generalization to non - equilibrium conditions of the conventional equilibrium projection operator approach @xcite to derive the corresponding memory function equations in which the mode coupling approximations could be introduced .
similarly , de gregorio et al .
@xcite discussed time - translational invariance and the fluctuation - dissipation theorem in the context of the description of slow dynamics in system out of equilibrium but close to dynamical arrest .
they also proposed extensions of approximations long known within mct .
unfortunately , in neither of these theoretical efforts , quantitative predictions were presented that could be contrasted with experimental or simulated results in specific model systems of _ structural _ glass - formers . in an independent but similarly - aimed effort ,
on the other hand , the self - consistent generalized langevin equation ( scgle ) theory of colloid dynamics @xcite and of dynamic arrest @xcite has recently been extended to describe the ( non - equilibrium ) spatially non - uniform and temporally non - stationary evolution of glass - forming colloidal liquids .
such an extension was introduced and described in detail in ref .
@xcite , and will be referred to as the _ non - equilibrium _ self - consistent generalized langevin equation ( ne - scgle ) theory .
as one can imagine , the number and variety of the phenomena that could be studied with this new theory may be enormous , and to start its systematic application we must focus on simple classes of physically relevant conditions .
thus , as a first simple illustrative application , this theory was applied in ref .
@xcite to a model colloidal liquid with hard - sphere plus short - ranged attractive interactions , suddenly quenched to an attractive glass state .
the aim of the present work is to start a systematic exploration of the scenario predicted by this theory when applied to the simplest irreversible processes , in the simplest and best - defined model system . in the present case
we refer to the irreversible isochoric evolution of a glass - forming liquid of particles interacting through purely repulsive soft - sphere interactions , initially at a fluid - like state , whose temperature is suddenly quenched to a final value @xmath28 , at which the expected equilibrium state is that of a hard - sphere liquid at volume fraction @xmath12 .
such process mimics the spontaneous search for the equilibrium state of this hard - sphere liquid , driven to non - equilibrium conditions by some perturbation ( shear , for example @xcite ) which ceases at a time @xmath20 .
one possibility is that the system will recover its equilibrium state within an equilibration time @xmath6 that depends on the fixed volume fraction @xmath12 .
the other possibility is that the system ages forever in the process of becoming a glass .
the application of the ne - scgle theory to these irreversible processes results in a well - defined scenario of the spontaneous non - equilibrium response of the system , whose main features are explained and illustrated in this paper . in the following section we provide a brief summary of the non - equilibrium self - consistent generalized langevin equation theory , appropriately written to describe the equilibration of a monocomponent glass - forming liquid constrained to remain spatially uniform .
section [ sectioniii ] defines the specific model to which this theory will be applied , discusses the strategy of solution of the resulting equations , and illustrates the main features of the results .
section [ sectioniv ] presents the scenario predicted by the ne - scgle theory for the first possibility mentioned above , namely , that the system is able to reach its thermodynamic equilibrium state . in this case
we find that the equilibrium @xmath4-relaxation time @xmath29 , and the equilibration time @xmath6 needed to reach it , will remain finite for volume fractions smaller than a critical value @xmath9 , but that both characteristic times will diverge as @xmath12 approaches this dynamic - arrest volume fraction @xmath30 , and will remain infinite for @xmath31 .
although it is intrinsically impossible to witness the actual predicted divergence , the theory makes distinct predictions regarding the transient non - equilibrium evolution occurring within experimentally - reasonable waiting times @xmath3 . in sect .
[ sectionv ] we analyze the complementary regime , @xmath32 , in which the system , rather than reaching equilibrium within finite waiting times , is predicted to age forever . in this regime
we find that the long - time asymptotic limit of @xmath0 will no longer be the expected equilibrium static structure factor @xmath33 , but another , non - equilibrium but well - defined , static structure factor , that we denote as @xmath34 , and which depends on the protocol of the quench .
furthermore , contrary to the kinetics of the equilibration process , in which @xmath0 approaches @xmath33 in an exponential - like fashion , this time the decay of @xmath0 to its asymptotic value @xmath34 follows a much slower power law .
in section [ sectionvi ] we put together the two regimes just described , in an integrated picture , which outlines the predicted scenario for the crossover from equilibration to aging . there
we find that the discontinuous and singular behavior underlying the previous scenario is intrinsically unobservable , due to the finiteness of the experimental measurements , which constraints the observations to finite time windows .
this practical but fundamental limitation converts the discontinuous dynamic arrest transition into a blurred crossover , strongly dependent on the protocol of the experiment and of the measurements .
the main purpose of the present paper is to explain in sufficient detail the methodological aspects of the application of the theory , so as to serve as a reliable reference for the eventual application of this non - equilibrium theory to the same system but with different non - equilibrium processes ( e.g. , different quench protocols ) , or in general to different systems and processes .
thus , we shall not report here the results of the systematic quantitative comparison of the scenario explained here with available specific simulations or experiments , which are being reported separately .
thus , the final section of the paper briefly refers to the main features of those comparisons , and discusses possible directions for further work .
let us mention that the referred non - equilibrium self - consistent generalized langevin equation ( ne - scgle ) theory derives from a non - equilibrium extension of onsager s theory of thermal fluctuations @xcite , and it consists of the time evolution equations for the mean value @xmath35 and for the covariance @xmath36 of the fluctuations @xmath37 of the local concentration profile @xmath38 of a colloidal liquid .
these two equations are coupled , through a local mobility function @xmath39 , with the two - time correlation function @xmath40 .
a set of well - defined approximations on the memory function of @xmath41 , detailed in ref .
@xcite , results in the referred ne - scgle theory . as discussed in ref .
@xcite , for given interparticle interactions and applied external fields , the ne - scgle self - consistent theory is in principle able to describe the evolution of a strongly correlated liquid from an initial state with arbitrary mean and covariance @xmath42 and @xmath43 , towards its equilibrium state characterized by the equilibrium local concentration profile @xmath44 and equilibrium covariance @xmath45 .
these equations are in principle quite general , and contain well known theories as particular limits . for example , ignoring certain memory function effects , the evolution equation for the mean profile @xmath35 becomes the fundamental equation of dynamic density functional theory @xcite , whereas the conventional " equilibrium scgle theory @xcite ( analogous in most senses to mct @xcite ) is recovered when full equilibration is assumed and spatial heterogeneities are suppressed . the ne - scgle theory , however , provides a much more general theoretical framework , which in principle describes the spatially heterogeneous and temporally non - stationary evolution of a liquid toward its ordinary stable thermodynamic equilibrium state .
this state , however , will become unreachable if well - defined dynamic arrest conditions arise along the equilibration pathway , in which case the system evolves towards a distinct and predictable dynamically arrested state through an evolution process that involves aging as an essential feature . to start the systematic application of this general theory to more specific phenomena
we must focus on a simple class of physical conditions .
thus , let us consider the irreversible evolution of the structure and dynamics of a system _ constrained _ to suffer a programmed process of spatially _ homogeneous _ compression or expansion ( and/or of cooling or heating ) . under these conditions , rather than solving the time - evolution equation for @xmath46
, we assume that the system is constrained to remain _ spatially uniform _ ,
@xmath47 , according to a _ prescribed _ time - dependence @xmath48 of the uniform bulk concentration and/or to a prescribed uniform time - dependent temperature @xmath49 . among the many possible programmed protocols ( @xmath48 , @xmath49 ) that one could devise to drive or to prepare the system , in this paper we restrict ourselves to one of the simplest and most fundamental protocols , which corresponds to the limit in which the system , initially at an equilibrium state determined by initial values of the control parameters , @xmath50 , must adjust itself in response to a sudden and instantaneous change of these control parameters to new values @xmath51 , according to the program " @xmath52 and @xmath53 , with @xmath54 being heavyside s step function
furthermore , just like in the first illustrative example described in ref .
@xcite , here we shall also restrict ourselves to the description of an even simpler subclass of irreversible processes , namely , the isochoric cooling or heating of the system , in which its number density is constrained to remain constant , i.e. , @xmath55 , while the temperature @xmath49 changes abruptly from its initial constant value @xmath19 to a final constant value @xmath56 at @xmath20 . under conditions of spatial uniformity ,
@xmath41 can be written as @xmath57 f(k,\tau;t),\ ] ] with @xmath58 , and where @xmath59 is the _
t_-evolving non - equilibrium intermediate scattering function ( ne - isf ) .
similarly , the covariance @xmath60 can be written as @xmath61 s(k;t ) \label{ftsigma}\ ] ] with @xmath62 being the time - evolving static structure factor . under these conditions
, the ne - scgle theory determines that the time - evolution equation for the covariance ( eq .
( 2.11 ) of ref .
@xcite ) may be written as an equation for @xmath0 which , for @xmath63 , reads @xmath64 .
\label{relsigmadif2pp}\ ] ] in this equation the function @xmath65 is the fourier transform ( ft ) of the functional derivative @xmath66 \equiv \left [ { \delta \beta\mu [ { \bf r};n]}/{\delta n({\bf r}')}\right]$ ] , evaluated at @xmath67 and @xmath68 . as discussed in refs .
@xcite , this thermodynamic object embodies the information , assumed known , of the chemical equation of state , i.e. , of the functional dependence of the electrochemical potential @xmath69 $ ] on the number density profile @xmath70 .
the solution of this equation , for arbitrary initial condition @xmath71 , can be written as @xmath72^{-1}\left(1-e^{-\alpha ( k ) u(t)}\right ) , \label{solsigmadkt}\ ] ] with @xmath73 and with @xmath74 in the equations above , the time - evolving mobility @xmath75 is defined as @xmath76 , with @xmath77 being the short - time self - diffusion coefficient and @xmath78 the long - time self - diffusion coefficient at evolution time @xmath3 . as explained in refs . @xcite and @xcite , the equation @xmath79^{-1 } \label{bdt}\ ] ] relates @xmath75 with the @xmath3-evolving , @xmath2-dependent friction coefficient @xmath80 given approximately by @xmath81 ^ 2 \\
\times f(k,\tau ; t)f_s(k,\tau ; t ) .
\end{split } \label{dzdtquench}\ ] ] thus , the presence of @xmath75 in eq .
( [ udt ] ) couples the formal solution for @xmath0 in eq .
( [ solsigmadkt ] ) with the solution of the non - equilibrium version of the scgle equations for the collective and self ne - isfs @xmath82 and @xmath83 .
these equations are written , in terms of the laplace transforms ( lt ) @xmath84 and @xmath85 , as @xmath86 and @xmath87 with @xmath88 being a phenomenological interpolating function " @xcite , given by @xmath89 , \label{lambdadk}\ ] ] with @xmath90 , where @xmath91 is the position of the main peak of @xmath92 ( in practice , however , @xmath93 ) @xcite ) .
the simultaneous solution of ecs .
( [ relsigmadif2pp])-([fluctsquench ] ) above , constitute the ne - scgle description of the spontaneous evolution of the structure and dynamics of an _ instantaneously _ and _ homogeneously _ quenched liquid .
of course , one important aspect of this analysis refers to the possibility that along the process the system happens to reach the condition of dynamic arrest . for the discussion of this important aspect it is useful to consider the long-@xmath2 ( or small @xmath94 )
asymptotic stationary solutions of eqs .
( [ fluctquench])-([dzdtquench ] ) , the so - called non - ergodicity parameters , which are given by @xcite @xmath95 and @xmath96 where the @xmath3-dependent squared localization length @xmath97 is the solution of @xmath98 ^{2}\lambda^2 ( k;t)}{\left[\lambda ( k;t)s(k;t ) + k^2\gamma(t)\right]\left[\lambda ( k;t ) + k^2\gamma(t)\right]}. \label{nep5pp}\ ] ] notice also that these equations are the non - equilibrium extension of the corresponding results of the equilibrium scgle theory ( referred to as the bifurcation equations " in the context of mct @xcite ) , and their derivation from eqs .
( [ dzdtquench])- ( [ fluctsquench ] ) follows the same arguments as in the equilibrium case @xcite .
the solution @xmath99 of eq .
( [ nep5pp ] ) and the mobility @xmath75 constitute two complementary dynamic order parameters , in the sense that if @xmath99 is finite ( or @xmath100 ) , then the system must be considered dynamically arrested at that waiting time @xmath3 , whereas if @xmath101 is infinite , then the particles retain a finite mobility , @xmath102 , and the instantaneous state of the system is ergodic or fluid - like .
we recall that the first relevant application of eq .
( [ nep5pp ] ) is the determination of the _ equilibrium _ dynamic arrest diagram in control - parameter space ( which , in the present case , is the density - temperature plane @xmath103 ) .
this diagram determines the region of fluid - like states , for which the solution @xmath104 ( of eq .
( [ nep5pp ] ) , with @xmath105 ) is infinite .
the complementary region contains the dynamically - arrested states , for which @xmath104 is finite .
the borderline between these two regions is the dynamic arrest transition line . due to the complementarity of the dynamic order parameters @xmath101 and @xmath75 ,
this curve is also the borderline between the region where the mobility @xmath75 will reach its equilibrium value , @xmath106 , and the region of arrested states , where @xmath107 .
thus , since @xmath108 , where @xmath109 is the equilibrium long - time self - diffusion coefficient at the point @xmath103 , this line is also the iso - diffusivity curve corresponding to @xmath110 .
let us now discuss some general features of the solution of the ne - scgle equations just presented .
this discussion has a general character , but for the sake of clarity we shall illustrate the main concepts in the context of one specific application .
thus , consider a mono - component fluid of soft spheres of diameter @xmath111 , whose particles interact through the truncated lennard - jones ( tlj ) pair potential that vanishes for @xmath112 , but which for @xmath113 is given , in units of the thermal energy @xmath114 , by @xmath115 . \label{truncatedlj}\ ] ] the state space of this system is spanned by the volume fraction @xmath116 and the reduced temperature @xmath117 . in order to apply eqs .
( [ relsigmadif2pp])-([fluctsquench ] ) to this model system , we first need to determine its thermodynamic property @xmath118 .
as indicated above , this is the fourier transform of the functional derivative @xmath66 \equiv \left [ { \delta \beta\mu [ { \bf r};n]}/{\delta n({\bf r}')}\right]$ ] , which can also be written as @xmath66 = \delta ( \textbf{r}-\textbf{r}')/\overline{n}- c(\mid\textbf{r}-\textbf{r}'\mid;n , t)$ ] , with @xmath119 being the ordinary direct correlation function @xcite .
this is an intrinsically thermodynamic property , related with the _ equilibrium _ static structure factor @xmath120 by the ornstein - zernike ( oz ) equation , which in fourier space reads @xmath121 .
the oz equation is the basis for the construction of the approximate integral equations of the _ equilibrium _ statistical thermodynamics of liquids @xcite .
in fact , we shall employ one such approximation to determine @xmath122 for our soft - sphere system .
this approximation , explained in detail in the appendix of ref .
@xcite and denoted as py / vw , is based on the percus - yevick approximation @xcite within the verlet - weis correction @xcite for the hard sphere system , complemented by the treatment of soft - core potentials introduced by verlet and weis themselves @xcite .
let us emphasize that for the present purpose , approximations such as these must be regarded solely as a practical and approximate mean to determine the thermodynamic property @xmath122 , which is essentially the local curvature of the free energy surface at the state point @xmath123 @xcite .
this property directly determines the _ equilibrium _ structure factor @xmath120 through the equilibrium relationship @xmath121 , and in practice we actually use this relationship to determine @xmath122 .
the main message of eq .
( [ relsigmadif2pp ] ) , however , is that the experimentally observable , non - equilibrium , static structure factor @xmath0 is not determined by any ornstein - zernike equilibrium condition , but by eq .
( [ relsigmadif2pp ] ) itself , with the thermodynamic property @xmath122 driving the non - equilibrium evolution in the manner indicated by its explicit appearance in this equation . in
what follows , we are interested in studying the scenario revealed by the solution @xmath0 of eq .
( [ relsigmadif2pp ] ) , for the process of isochoric equilibration ( or lack of equilibration ) of the static structure of a system subjected to a temperature control protocol @xmath53 , corresponding to a an instantaneous temperature quench to a final temperature @xmath56 denoted simply as @xmath17 .
thus , the system is assumed to be prepared at an initial equilibrium homogeneous state characterized by a bulk particle number density @xmath124 and temperature @xmath19 , at which its initial static structure factor is @xmath71 . upon suddenly changing the temperature of this system to the new value @xmath17
, one normally expects that the system will reach full thermodynamic equilibrium , i.e. , that the long - time asymptotic limit of @xmath0 will be the equilibrium static structure factor @xmath125 . according to eq .
( [ relsigmadif2pp ] ) , reaching this value is also a sufficient condition for @xmath0 to reach a _
stationary _ state . according to the same equation , however , this is _ not a necessary _ condition for the stationarity of @xmath0 , which could also be attained if @xmath126 , even in the absence of thermodynamic equilibrium ( i.e. , even if @xmath127 ) .
if the long - time stationary state attained is the thermodynamic equilibrium state , we say that the system is ergodic at the point @xmath123 .
the second condition , in contrast , corresponds to dynamically arrested states , in which the long - time asymptotic limit of @xmath0 might differ from the expected thermodynamic equilibrium value @xmath125 . clearly , these are two mutually exclusive and fundamentally different classes of possible stationary states which can only be distinguished if we know the long - time limit of @xmath75 .
this is , however , not a thermodynamic property , and hence , the discrimination of the ergodic or non - ergodic nature of the state point @xmath123 must be based on a dynamic or transport theory that allows the determination of @xmath75 .
one such theory is precisely the scgle theory : to decide if the long - time stationary state corresponding to the point @xmath123 will be an ergodic or an arrested state one can use the equilibrium static structure factor @xmath120 in eq .
( [ nep5pp ] ) to calculate @xmath128 .
if the solution is infinite , we say that the asymptotic stationary state is ergodic , and hence , that at the point @xmath123 the system will be able to reach its thermodynamic equilibrium state without impediment , so that @xmath129 . on the other hand ,
if the solution for @xmath128 turns out to be finite , this means that the system will become dynamically arrested , and that the long - time limit of @xmath0 at the point @xmath123 will not necessarily be its thermodynamic equilibrium value @xmath125 .
instead , we shall have that @xmath130 , with a truly non - equilibrium structure factor @xmath34 , different from @xmath120 , and obtained as an alternative stationary solution of eq .
( [ relsigmadif2pp ] ) . in this manner , by calculating @xmath128 at all state points @xmath123 one can scan the state space to determine the region of dynamically arrested states of the system . ) in the @xmath131 state space of the truncated lennard - jones fluid , eq .
( [ truncatedlj ] ) , with @xmath132 .
the vertical downward arrows represent two fundamentally different classes of irreversible isochoric processes : in the first case the fixed volume fraction @xmath12 is smaller than the dynamic arrest volume fraction @xmath133 , whereas in the second @xmath12 is larger than @xmath9 . ]
we have employed in this manner the py / vw approximation for the equilibrium static structure factor @xmath120 of the tlj soft - sphere model , to determine the region of its fluid - like ergodic states and the region of its dynamically arrested states .
the resulting dynamic arrest transition line is represented by the solid curve in fig .
[ fig1 ] for the tlj fluid with @xmath132 , whose @xmath134 limit coincides with the dynamic arrest volume fraction @xmath9 of the hard sphere liquid , predicted to occur at @xmath135 @xcite . as indicated in the figure ( and as explained at the end of the previous section ) , this transition line is also the iso - diffusivity curve corresponding to @xmath110 . for concreteness ,
let us consider the case in which the system was initially prepared to be in the _ equilibrium _ state corresponding to a point @xmath136 located in the fluid like region .
we then have two fundamentally different possibilities , also illustrated in fig .
[ fig1 ] : either the final point @xmath137 lies in the ergodic region of the dynamic arrest diagram , or else , it lies in the region of dynamically arrested states . the first case is achieved , for example , if the volume fraction of the isochoric irreversible process is smaller than the dynamic arrest volume fraction @xmath135 of the hard sphere liquid .
this isochoric quench @xmath138 will then eventually lead to the full equilibration of the system . in the second case , in which the fixed volume fraction @xmath12 must be larger than @xmath135 ( and the final temperature sufficiently low ) the solution of eqs .
( [ relsigmadif2pp ] ) , ( [ bdt])-([fluctsquench ] ) will describe the irreversible aging of the glass - forming liquid quenched to a point inside the dynamically arrested region .
in either case , solving ec .
( [ relsigmadif2pp ] ) for @xmath0 starts with the formal solution in eq .
( [ solsigmadkt ] ) , written as @xmath139 this expression interpolates @xmath140 between its initial value @xmath141 and its expected long - time equilibrium value @xmath142^{-1}$ ] . clearly , the solution @xmath0 in eq .
( [ solsigmadkt ] ) can be written as @xmath143 with @xmath144 defined in eq .
( [ udt ] ) .
the inverse function @xmath145 is such that @xmath146 and @xmath147 .
the differential form of eq .
( [ udt ] ) can be written as @xmath148 .
upon integrating this equation , we have that @xmath149 , which can also be written , after the change of the integration variable @xmath150 , to @xmath151 , as @xmath152 with the function @xmath153 defined as @xmath154 .
these general observations greatly simplify the mathematical analysis and the numerical method of solution of the full ne - scgle theory under the particular conditions considered here . to see this ,
let us consider a sequence @xmath155 of snapshots of the static structure factor , generated by the simple expression in eq .
( [ solsdktexp ] ) when the parameter @xmath156 attains a sequence of equally - spaced values @xmath157 , say @xmath158 ( with a prescribed @xmath159 and with @xmath160 ) .
the fact that @xmath0 can be written as @xmath161 implies that this sequence will be identical to the sequence @xmath162 generated by the exact solution in eq .
( [ solsigmadkt ] ) , evaluated at a different sequence @xmath163 ( @xmath164 , i.e. , at a sequence of values of the time @xmath3 , given by @xmath165du'$ ] . in other words , the @xmath166th member of the sequence of static structure factors can be labeled either with the label @xmath157 , as @xmath155 , or with the label @xmath163 , as @xmath162 .
for sufficiently small @xmath159 , the discretized form of the previous relationship between @xmath163 and @xmath157 can be written as @xmath167 thus , in practice what we do is to solve the self - consistent system of equations ( [ bdt])-([lambdadk ] ) with @xmath0 replaced by each snapshot @xmath168 of the sequence of static structure factors .
this yields , among all the other dynamic properties , the sequence of values @xmath169 of the function @xmath153 . this sequence can then be used in the recurrence relation in eq .
( [ recrel0 ] ) to obtain the desired time sequence @xmath163 , which allows us to ascribe a well - defined time label to the sequence @xmath162 of static structure factors and to the sequence @xmath170 of the instantaneous mobility @xmath75 .
of course , since the solution of equations ( [ bdt])-([lambdadk ] ) yields all the dynamic properties , we also have in store the corresponding sequence of snapshots of dynamic properties such as @xmath171 , @xmath172 , the @xmath4-relaxation time @xmath173 , etc .
we have applied the protocol just described , which solves the full ne - scgle theory ( eqs .
( [ solsigmadkt])-([lambdadk ] ) ) , to the description of the isochoric irreversible evolution of the structure and the dynamics of the tlj soft sphere liquid , after the instantaneous quench starting from an equilibrium fluid state . to continue the analysis
, however , it is convenient to discuss separately the two mutually exclusive possibilities illustrated by the two vertical arrows in fig .
[ fig1 ] . in this section
we shall concentrate on the conceptually simplest case of the full equilibration of the system , and in the following section we shall discus the process of dynamic arrest .
let us thus illustrate the isochoric quench in which both , the initial and the final points , lie in the fluid - like region . for concreteness , we consider a cooling process , @xmath174 , such that @xmath175 , with @xmath176 , @xmath177 , and with the final temperature corresponding to the deepest quench , @xmath178 .
the initial and final equilibrium static structure factors , @xmath179 and @xmath180 , are presented in fig .
[ fig2 ] . to visualize the transient non - equilibrium relaxation of @xmath0
, we generate a sequence of snapshots @xmath155 using eq .
( [ solsdktexp ] ) with @xmath181 ( @xmath160 ) and with @xmath182 \ ( \approx 1/4\alpha ( k)$ ] , for @xmath183 , the position of the main peak of @xmath184 ) . from now on we shall use @xmath185 $ ] as the time unit and @xmath111 as the unit length . in fig .
[ fig2 ] we include four representative intermediate snapshots of this sequence , corresponding to @xmath186 and 7 .
let us emphasize that although these snapshots of the transient structure factor are linear combinations of two equilibrium static structure factors ( namely , @xmath187 and @xmath184 ) , they themselves represent fully non - equilibrium structures .
corresponding to the quench @xmath138 at fixed volume fraction @xmath188 and @xmath176 , and @xmath178 .
the darker thick ( black ) solid line is the initial structure factor @xmath189 .
the lighter thick ( blue ) solid line is the asymptotic limit @xmath190 .
the sequence of thinner ( green ) solid lines represent @xmath140 for @xmath191 and @xmath192 , with @xmath193 ( we use @xmath185 $ ] as the time unit ) .
these six structure factors also correspond , according to eq .
( [ tdu00 ] ) , to @xmath0 for @xmath194 , and @xmath195 .
inset : inverse relaxation constant @xmath196 . ] fig .
[ fig2 ] exhibits the fact that within the resolution @xmath193 employed to visualize @xmath140 , this non - equilibrium structure relaxes very quickly to its long - time equilibrium limit @xmath184 at most wave - vectors , except in two regions : in the vicinity of @xmath197 , as appreciated in the figure , and in the long - wavelength limit , @xmath198 , not apparent in the main figure , but illustrated and discussed below .
thus , except in these two wave - vector domains , the non - equilibrium snapshots of @xmath140 shown in the figure are already indistinguishable from @xmath184 .
the fact that for large wave - vectors , @xmath199 , the structure @xmath140 approaches very fast its final equilibrium value @xmath184 is understood by the fact that @xmath200 increases with @xmath201 while @xmath184 decreases from its maximum value towards its unit value at large @xmath202 . to the left of @xmath197 , on the other hand , although @xmath203 decreases with @xmath201 , there is a dramatic drop of the static structure factor from its large value at the main peak towards the very small value of @xmath204 of a strongly incompressible liquid . in support of this proposed scenario , in the inset of fig .
[ fig2 ] we plot the inverse relaxation constant @xmath196 as a function of @xmath202 , which clearly exhibits a dominant peak at @xmath183 , and a divergence at @xmath205 .
this explains the quick thermalization of @xmath140 in both , the large wave - vector domain and in the moderately small wave - vector regime @xmath206 . in the really small wave - vector limit @xmath198 , however , the @xmath207 divergence of @xmath208 dominates , and prevents the thermalization of @xmath140 within finite values of @xmath156 .
the crossover from this long - wavelength perfect slowdown , to the faster moderately - small wave - vector regime @xmath209 , is revealed by zooming in at the small-@xmath202 behavior of the snapshots of @xmath140 , as illustrated in fig . [ fig3](a ) .
in contrast with this rather trivial long - wavelength slowing down , the slow relaxation at and around @xmath210 has its origin in the large value attained by @xmath211 , i.e. , in the large strength of the interparticle correlations of spatial extent similar to the mean distance between the particles .
thus , this slowing down of the main peak of the @xmath156-evolving static structure factor is a non - equilibrium manifestation of the so - called cage effect . in fig .
[ fig3](b ) we present a zoom of the snapshots of @xmath140 of fig . [ fig2 ] , exhibiting in more detail the slower relaxation of the structure at these wave - vectors . in fig .
[ fig2 ] corresponding to ( a ) the long - wavelength limit @xmath212 and ( b ) to the neighborhood of the position @xmath197 of the main peak of @xmath184.,title="fig:"]1.5 cm in fig .
[ fig2 ] corresponding to ( a ) the long - wavelength limit @xmath212 and ( b ) to the neighborhood of the position @xmath197 of the main peak of @xmath184.,title="fig : " ] let us notice that the simple expression for @xmath140 in eq .
( [ solsdktexp ] ) , which interpolates this function of @xmath156 between @xmath187 and @xmath184 , may be written as @xmath213 this means that if we plot the static structure factor @xmath140 as @xmath214 vs. the @xmath202-dependent variable @xmath215 $ ] , the results for all the wave - vectors @xmath202 must collapse onto a master curve independent of @xmath202 and of the initial and final values @xmath187 and @xmath184 .
in fact , such a master curve will be essentially a simple exponential function .
this simplicity , however , will be partially lost if we plot @xmath214 directly as a function of the parameter @xmath156 , since such exponential function , @xmath216 , will decay with @xmath156 at a different rate @xmath217 for different values of the wave - vector @xmath202 , as illustrated in fig . [
fig3p](a ) .
in fact , if we define a @xmath202-dependent _ equilibration _ value @xmath208 by the condition @xmath218 , we have that @xmath219 .
thus , except for the arbitrary factor of 5 , the inset of fig .
[ fig2](a ) exhibits the wave - vector dependence of @xmath208 .
there we see that @xmath208 attains its largest value at the wave - vector @xmath197 , corresponding to the position of the main peak of @xmath184 .
this slowest mode imposes the pace of the overall equilibration process , thus characterized by the @xmath202-independent _ equilibration _ value @xmath220 .
the previous discussion illustrates the properties of the ordered sequence @xmath155 of snapshots of the function @xmath140 for equally - spaced values @xmath157 of the parameter @xmath156 .
this sequence of snapshots , however , do not fully reveal the most important features of the real relaxation scenario implied by the solution ( [ solsigmadkt ] ) of eq .
( [ relsigmadif2pp ] ) , which provides @xmath0 as a function of the actual evolution time @xmath3 .
nevertheless , since @xmath221 , these features are fully revealed by simply relabeling the referred sequence @xmath155 using the ( not equally - spaced ) sequence of labels @xmath163 given by the recurrence relation in eq .
( [ recrel0 ] ) .
this results in the sequence @xmath162 of snapshots that describes the actual time evolution of @xmath0 . in order to carry out this program , however , we must first determine the sequence @xmath169 needed in the referred recurrence relation . as indicated before , from any sequence of snapshots @xmath155 , with @xmath222 ( @xmath160 ) , we may generate a sequence @xmath169 of values of the time - dependent mobility @xmath153 by solving the self - consistent system of equations ( [ bdt])-([lambdadk ] ) with @xmath0 replaced by @xmath155 for each snapshot .
the resulting sequence @xmath169 is a discrete representation of the function @xmath153 , shown in fig .
[ fig3p](a ) , whose resolution in the parameter @xmath156 may be improved arbitrarily by taking @xmath159 as small as needed .
the first feature to notice in the result of this procedure is the fact that @xmath153 decays monotonically from its initial value @xmath223 to its final value @xmath224 .
this implies that the system will always remain fluid - like and will have no impediment to reach its expected equilibrium state .
we also find that the function @xmath153 attains its asymptotic value @xmath225 for @xmath226 for the quench illustrated in the figure ) .
$ ] corresponding to three different wave - vectors , @xmath202 = 6.1 , 7.26 @xmath227 , and 8.1 , plotted as a function of the parameter @xmath156 ( dashed lines ) .
the circles correspond to @xmath228 , where @xmath229=e^{-5}$ ] .
the dotted and the solid dark lines represent , respectively , the functions @xmath230 and @xmath231 ( ecs . ( [ deltabdu ] ) and ( [ deltabinvdu ] ) ) , and the solid circle corresponds to the estimated equilibration value @xmath232 . the inset compares the function
@xmath145 calculated according to its exact definition in eq .
( [ tdu00 ] ) ( solid line ) and according to the approximation in eq .
( [ tduapprox ] ) .
( b ) same information as in ( a ) , but now plotted as a function of the actual _ evolution _ time @xmath233 . in the main figure we plot @xmath234 and in the inset we plot @xmath235.,title="fig : " ] $ ] corresponding to three different wave - vectors , @xmath202 = 6.1 , 7.26 @xmath227 , and 8.1 , plotted as a function of the parameter @xmath156 ( dashed lines ) .
the circles correspond to @xmath228 , where @xmath229=e^{-5}$ ] .
the dotted and the solid dark lines represent , respectively , the functions @xmath230 and @xmath231 ( ecs .
( [ deltabdu ] ) and ( [ deltabinvdu ] ) ) , and the solid circle corresponds to the estimated equilibration value @xmath232 .
the inset compares the function @xmath145 calculated according to its exact definition in eq .
( [ tdu00 ] ) ( solid line ) and according to the approximation in eq .
( [ tduapprox ] ) .
( b ) same information as in ( a ) , but now plotted as a function of the actual _ evolution _ time @xmath233 . in the main figure we plot @xmath234 and in the inset we plot @xmath235.,title="fig : " ] in fig .
[ fig3p](a ) we also present the results for @xmath153 plotted as @xmath236 we see that this plot does not exhibit any simple relationship between the decay of @xmath230 and the decay of @xmath214 . in the same figure , however , the results for @xmath153 are plotted as @xmath237 plotted in this manner we observe a more apparent correlation between the decay of both , @xmath238 and @xmath231 , with the parameter @xmath156 .
this feature remains , of course , when these properties are expressed as functions of the actual evolution time @xmath3 , as we now see .
once we have determined the function @xmath153 , using the expression for @xmath145 in eq .
( [ tdu00 ] ) , or its discretized version in the recursion relation of eq .
( [ recrel0 ] ) , we can determine the desired real - time evolution of @xmath0 and @xmath75 . in this manner
we determine that in our illustrative example the sequence @xmath239 and @xmath240 corresponds to the sequence @xmath241 and 0.533 . in the inset of fig . [
fig3p](b ) we present the resulting time - evolution of @xmath0 for the same three wave - vectors as in fig . [ fig3p](a ) .
this inset emphasizes the fact that @xmath0 evolves monotonically from its initial value @xmath242 to its final value @xmath184 , sometimes increasing and sometimes decreasing , depending on the wave - vector considered . in order to exhibit a less detail - dependent scenario , in the main frame of fig .
[ fig3p](b ) we present the same information , but formatted as @xmath234 , which is the re - labeled version ( @xmath243 ) of @xmath214 in eq .
( [ solsdktexp2 ] ) , namely , as @xmath244 we similarly relabel the definitions of @xmath230 and @xmath231 in eqs .
( [ deltabdu ] ) and ( [ deltabinvdu ] ) to define the functions @xmath245 and @xmath246 , which are also plotted in fig . [ fig3p](b ) . the comparison of this figure with fig .
[ fig3p](a ) indicates that , except for the stretched metric of @xmath3 , the overall scenario described by the @xmath156-dependence illustrated in fig .
[ fig3p](a ) is preserved in the @xmath3-dependence illustrated in fig . [
fig3p](b ) .
let us notice in particular that the existence of the equilibration value @xmath247 of the parameter @xmath156 , beyond which @xmath248 , allows us to define an _
equilibration _ time , @xmath249 , as the time that corresponds to @xmath247 through eq .
( [ tdu00 ] ) , @xmath250 for our specific illustrative example , this yields @xmath251 .
the fact that @xmath248 for @xmath252 implies , according to eq .
( [ tdu00 ] ) , that for @xmath253 the function @xmath145 will be linear in @xmath156 , i.e. , @xmath254 with @xmath255du'$ ] . in the inset of fig .
[ fig3p](a ) we compare this asymptotic expression , applied to our illustrative case ( for which @xmath256 and @xmath257 ) , with the actual @xmath145 calculated from eq .
( [ tdu00 ] ) .
since for each snapshot of the static structure factor @xmath0 the solution of eqs .
( [ bdt])-([lambdadk ] ) determines a snapshot of each of the dynamic properties of the system , the process of equilibration may also be observed , for example , in the @xmath3-evolution of the collective and self intermediate scattering functions , @xmath59 and @xmath1 . in fig .
[ fig5](a ) we illustrate this irreversible time - evolution with the snapshots of the self - isf @xmath258 , corresponding to the same set of evolution times @xmath163 as the snapshots of @xmath0 in fig .
we see that the function @xmath1 starts from its initial value @xmath259 , and quickly evolves with waiting time @xmath3 towards the vicinity of its final equilibrium value @xmath260 . at @xmath183 , corresponding to the equilibration process in fig .
[ fig2 ] for evolution times @xmath194 and @xmath195 .
( b ) non - equilibrium evolution of the dimensionless @xmath4-relaxation time , displayed as @xmath261 itself ( inset ) and formatted as in eq .
[ taualphastar ] ( main figure ) .
the solid circle represents the equilibration point ( @xmath262 ) .
, title="fig : " ] at @xmath183 , corresponding to the equilibration process in fig .
[ fig2 ] for evolution times @xmath194 and @xmath195 .
( b ) non - equilibrium evolution of the dimensionless @xmath4-relaxation time , displayed as @xmath261 itself ( inset ) and formatted as in eq .
[ taualphastar ] ( main figure ) .
the solid circle represents the equilibration point ( @xmath262 ) .
, title="fig : " ] the equilibration process of @xmath263 can be best summarized in terms of the dependence of the @xmath4-relaxation time @xmath261 as a function of the evolution time @xmath3 .
the @xmath4-relaxation time may be defined by the condition @xmath264 the dependence of @xmath261 on the evolution time @xmath3 can be extracted from a sequence of snapshots of @xmath263 , such as those in fig . [
fig5](a ) .
the results are illustrated in fig .
[ fig5](b ) , in which the solid line corresponds to @xmath265 .
the solid circle indicates the crossover from the @xmath3-regime where @xmath265 is still in the process of equilibration , to the regime where it has reached its final equilibrium value @xmath266 . in the inset of the figure we plot @xmath265 itself and in the main figure we plot the same information , but formatted as @xmath267 with @xmath268 . in the same figure we also exhibit similar results corresponding to two additional wave - vectors , different from @xmath197 ( dashed lines ) .
these results show that the equilibration time of @xmath261 for these three wave - vectors is largely independent of @xmath202 , and can be well approximated by the equilibration time @xmath249 defined in eq .
( [ equiltime1 ] ) , in contrast with the notorious wave - vector dependence of the predicted evolution of the static structure factor illustrated in fig . [ fig3p](b ) .
let us finally mention another theoretical prediction regarding the kinetics of the equilibration process .
this refers to the similarity of the equilibration kinetics exhibited by the time - dependent mobility @xmath75 , the @xmath4-relaxation time @xmath261 at all wave - vectors , and the static structure factor @xmath269 at the wave - vector @xmath197 , when plotted in terms of the reduced properties @xmath245 , @xmath270 , and @xmath271 .
this similarity is exhibited in fig .
[ fig6 ] for the illustrative quench at fixed @xmath188 , and means that indeed the evolution of @xmath269 ( which is slower than the evolution of @xmath0 for other wave - vectors ) sets the overall relaxation rate exhibited by the dynamic properties @xmath245 and @xmath270 .
thus , from this point of view , we may use either of these characteristic dynamic properties to describe the predicted kinetics of the equilibration process .
-relaxation time @xmath265 ( solid line ) mobility @xmath75 ( dash line ) and static structure factor @xmath269 ( dotted line ) as a function of evolution time @xmath3 , and expressed as @xmath272 , @xmath246 , and @xmath273 . ]
up to this point we have illustrated the main features of the isochoric quench @xmath138 at fixed volume fraction @xmath188 , using for concreteness the values @xmath176 and @xmath274 .
we are now ready to analyze how the scenario just described depends on the initial temperature @xmath275 and on the volume fraction @xmath12 at which we perform the quench .
let us start by considering the dependence on @xmath275 . rather than attempting a comprehensive illustration of this dependence in terms of the evolution of the static structure factor @xmath0 and of the various dynamic properties
, we use the dimensionless mobility @xmath75 as a representative property bearing the essential information about the equilibration process .
this @xmath202-independent property determines the mapping from the parameter @xmath156 to the real time @xmath3 , through the definition of the functions @xmath144 and @xmath145 in eqs .
( [ udt ] ) and ( [ tdu00 ] ) . and
( b ) @xmath4-relaxation time @xmath5 , as a function of evolution time @xmath3 , for the isochoric quench @xmath276 at fixed volume fraction @xmath188 , for three different initial temperatures , @xmath275= 1.0 , 0.1 , and 0.01 . the insets exhibit a power law fit of the transient before saturation.,title="fig : " ] and ( b ) @xmath4-relaxation time @xmath5 , as a function of evolution time @xmath3 , for the isochoric quench @xmath276 at fixed volume fraction @xmath188 , for three different initial temperatures , @xmath275= 1.0 , 0.1 , and 0.01 .
the insets exhibit a power law fit of the transient before saturation.,title="fig : " ] thus , in fig .
[ fig7](a ) we present plots of @xmath277 as a function of @xmath3 for three representative values of the initial temperature @xmath275 , namely , @xmath275= 1.0 , 0.1 , and 0.01 , keeping the same final temperature @xmath274 and the same volume fraction @xmath12=0.56 .
this figure reveals two remarkable features . in the first place
, the equilibration time @xmath249 seems to be rather insensitive to the temperature @xmath275 of the initial state . in other words ,
the system will reach the final equilibrium state in about the same time , @xmath278 , no matter if the initial temperature is @xmath275= 1.0 , 0.1 , or 0.01 . to emphasize this feature
we have highlighted the common equilibration point of the three curves .
the second remarkable feature is that during the transient stage of the equilibration process , the evolution of @xmath277 as a function of @xmath3 follows approximately a power law , @xmath279 , with the exponent @xmath280 and the amplitude @xmath281 depending on the initial temperature @xmath275 . in the inset of the figure we exhibit the power law fit of the transient , indicating the resulting value of the exponent @xmath280 and amplitude @xmath281 .
exactly the same trend is also reflected in the evolution of the intermediate scattering function , as observed in the results for the @xmath4-relaxation time shown in fig .
[ fig7](b ) .
this information is important , since many times it is this dynamic parameter what is monitored in simulations and in some experiments .
let us now discuss the dependence of the equilibration process on the volume fraction @xmath12 .
once again we first use the time evolution of @xmath75 to illustrate this dependence . in fig .
[ fig8](a ) we plot @xmath277 as a function of @xmath3 for a set of values of the volume fraction @xmath12 , corresponding to the metastable regime of the hard - sphere liquid .
according to these results , the inverse mobility @xmath282 reaches its equilibrium value @xmath283 after a @xmath12-dependent equilibration time @xmath6 . to emphasize this prediction ,
the solid circles in the figure highlight the points @xmath284 .
these highlighted points , as indicated in the figure , align themselves to a good approximation along the dashed line of the figure , corresponding to the approximate relationship @xmath285^{1.05}$ ] this relationship between the equilibration time @xmath6 and @xmath286 is one of the most remarkable predictions of the present theory , bearing profound physical implications . to see this
let us recall that the dimensionless mobility @xmath286 is just the scaled long - time self - diffusion coefficient @xmath287 of the fully equilibrated system at the final point @xmath288 , which for the present isochoric quench down to zero temperature , @xmath289 , is the dimensionless equilibrium long - time self - diffusion coefficient of the hard sphere liquid , @xmath290 .
this property can be calculated using the equilibrium version of the present theory @xcite and , as discussed below ( see fig . [ fig9](b ) ) , such calculation leads to the prediction that @xmath291 vanishes at @xmath135 , according to the power law @xmath292 . as a consequence ,
if @xmath293 , we must expect that as @xmath294 the equilibration time will diverge according to @xmath295 .
this predicted divergence of the equilibration time constitutes a strong and interesting proposal , which requires , of course , a critical assessment and validation .
we shall return to this discussion later on in the paper , but at this point , let us carry out a similar analysis , now using the @xmath4-relaxation time @xmath296 ( whenever we omit the wave - vector @xmath202 as argument of @xmath297 is because a specific value for @xmath202 is being assumed fixed , most frequently @xmath298 ) .
thus , in fig .
[ fig8](b ) we plot @xmath296 as a function of @xmath3 for the same set of values of the volume fraction @xmath12 as in fig .
[ fig8](a ) . here
again we find that @xmath299 reaches its equilibrium value @xmath300 after the same evolution time @xmath6 as in the case of @xmath301 .
also here the solid circles highlight the points @xmath302 , and the dashed line of the figure imply that @xmath303^{0.95}(\phi)$ ] .
this implies that the time @xmath6 required to equilibrate the system will grow at least about as fast as the equilibrium value @xmath304 of the @xmath4-relaxation time , and that both properties increase strongly with @xmath12 . and ( b ) @xmath4-relaxation time @xmath5 , as a function of evolution time @xmath3 for various volume fractions corresponding to the metastable regime of the hard - sphere liquid .
the arrows indicate the direction of increasing @xmath12 .
the dashed lines are the power laws @xmath285^{1.05}$ ] and @xmath305^{1.05}$].,title="fig : " ] and ( b ) @xmath4-relaxation time @xmath5 , as a function of evolution time @xmath3 for various volume fractions corresponding to the metastable regime of the hard - sphere liquid .
the arrows indicate the direction of increasing @xmath12 .
the dashed lines are the power laws @xmath285^{1.05}$ ] and @xmath305^{1.05}$].,title="fig : " ] although one can discuss additional features of the class of irreversible process corresponding to the full isochoric equilibration of the system after its sudden cooling , it is now important to contrast the scenario just described , with that of the second class of irreversible processes .
this involves the dynamic arrest of the system , and is the subject of the following section .
let us recall at this point that the ne - scgle description of the spontaneous evolution of the structure and dynamics of an _ instantaneously _ and _ homogeneously _ quenched liquid is provided by the simultaneous solution of ecs .
( [ relsigmadif2pp])-([fluctsquench ] ) . as discussed in subsection [ subsectioniii.2 ] , there exist two fundamentally different classes of irreversible isochoric processes , represented by the vertical downward arrows in fig .
[ fig1 ] . in the previous section we described the resulting scenario for the most familiar of them , namely , the full isochoric equilibration of the system . in this section
we present the ne - scgle description of the second class of irreversible isochoric processes , in which the system starts in an ergodic state and ends in the region where it is expected to become dynamically arrested .
thus , let us continue considering the tlj model system introduced in sect .
[ sectioniii ] ( eq .
[ truncatedlj ] , with @xmath132 ) , subjected to the sudden isochoric cooling at fixed volume fraction @xmath306 , larger than @xmath135 , from the point @xmath307 in the ergodic region , to the point @xmath308 in the region of dynamically arrested states . by construction , the solution @xmath309 of eq .
( [ nep5pp ] ) , obtained using @xmath310 as the structural input , is @xmath311 . in this sense ,
the present class of process is identical to the first one , discussed in the previous section .
the main difference lies , of course , in the fact that in the present case the solution of eq .
( [ nep5pp ] ) for the squared localization length @xmath312 , obtained using the equilibrium static structure factor @xmath313 of the final point as input , will now have a _ finite _ value . to see the consequences of this difference ,
let us go back to subsection [ subsectioniii.3 ] , and consider the function @xmath140 in eq .
( [ solsdktexp ] ) , with @xmath314 . for each value of @xmath156
we may use @xmath140 in the bifurcation equation ( [ nep5pp ] ) for @xmath101 , now denoted as @xmath315 . throughout the previous section it was implicitly assumed that @xmath316 for @xmath314 , an assumption based on the fact that the system started and ended in a fluid - like state . in the present case , however , although the system starts with the condition that @xmath317 , we know that the final point @xmath308 corresponds to an arrested state , so that @xmath318 has a finite value .
this means that somewhere between @xmath319 and @xmath320 the function @xmath315 changed from infinity to a finite value , and this then implies the existence of a finite value @xmath321 of @xmath156 , such that @xmath315 remains infinite only within the interval @xmath322 .
thus , in the present case the simultaneous solution of ecs .
( [ relsigmadif2pp])-([fluctsquench ] ) starts in practice with the precise determination of @xmath321 . to determine the critical value @xmath321 ,
let us consider again the sequence @xmath155 of snapshots of the static structure factor , generated by the expression in eq .
( [ solsdktexp ] ) with @xmath158 ( @xmath160 ) . since we have assumed that initially the system is fluid - like , the value of @xmath321 can not be @xmath323 .
thus , let us employ each snapshot of the sequence @xmath155 , with @xmath324 , as the static input of eq .
( [ nep5pp ] ) , thus determining the sequence @xmath325 of values of @xmath315 , which starts with @xmath326 .
if @xmath327 turns out to be finite , then one may take a smaller @xmath156-step @xmath159 , until this does not happen .
for a sufficiently small @xmath159 , there will be an integer @xmath328 such that @xmath329 for @xmath330 and @xmath331 is finite for @xmath332 , i.e. , such that @xmath333 .
this process can be refined by decreasing @xmath159 , so that one can determine @xmath321 with arbitrary precision for the given initial and final conditions @xmath136 and @xmath137 .
for example , one can readily perform this procedure for the quench indicated by the right arrow of fig .
[ fig1 ] ( from the point @xmath307 to the final point @xmath308 at fixed @xmath306 ) , with the result @xmath321= 0.0128 .
once one has determined @xmath321 with the desired precision , one can construct a new sequence @xmath334 of @xmath335 equally - spaced values of @xmath156 , defined as @xmath336 with @xmath337 , along with the corresponding sequence @xmath338 of snapshots of the static structure factor ( using eq .
( [ solsdktexp ] ) ) . since the sequence @xmath338 is identical to the sequence @xmath339 ( with @xmath340 such that @xmath341 ) , to each member of this sequence , the self - consistent eqs .
( [ dzdtquench])-([fluctsquench ] ) assigns a snapshot of the full dynamics of the system .
in particular , the use of eq .
( [ bdt ] ) generates a sequence of values @xmath342 of the mobility @xmath153 , with arbitrary resolution ( set by the number @xmath343 of @xmath156-steps ) . to illustrate these concepts , in fig .
[ fig9](a ) we present the results for @xmath153 corresponding to the specific quench under discussion .
notice that , as expected , @xmath344 as @xmath156 approaches @xmath321 from below .
a simple ansatz to model this limiting behavior is @xmath345 in the inset of fig .
[ fig9](a ) we plot @xmath153 vs. @xmath346 to determine the value of the exponent @xmath347 and the pre - factor @xmath348 , with the result @xmath347= 2.2 and @xmath348= 9.5 .
we performed similar calculations varying the initial temperature @xmath275 , and found the value @xmath347= 2.2 of the exponent is independent of @xmath275 , so that the dependence of @xmath153 on the initial temperature is carried only in the pre - factor @xmath348 .
for example , we found that @xmath349 , and 490 , for @xmath275= 0.1 , 0.05 , and 0.01 , respectively .
plotted as a function of the parameter @xmath156 and of the difference @xmath346 ( inset ) for the sudden isochoric cooling from the point @xmath350 to the point @xmath351 .
( b ) scaled long - time self - diffusion coefficient @xmath352 of the hard - sphere liquid as a function of volume fraction @xmath12 and of the difference @xmath353 ( inset).,title="fig : " ] plotted as a function of the parameter @xmath156 and of the difference @xmath346 ( inset ) for the sudden isochoric cooling from the point @xmath350 to the point @xmath351 .
( b ) scaled long - time self - diffusion coefficient @xmath352 of the hard - sphere liquid as a function of volume fraction @xmath12 and of the difference @xmath353 ( inset).,title="fig : " ] another remarkable feature of the @xmath156-dependence of @xmath153 illustrated in the inset of fig .
[ fig9](a ) is its similarity with the volume fraction dependence of the scaled long - time self - diffusion coefficient of the _ fully equilibrated _ hard - sphere system , @xmath354 .
this property can be calculated using the equilibrium version of the scgle theory @xcite , and the results are exhibited in fig . [ fig9](b ) . as discussed before @xcite
, the theoretical prediction is that @xmath291 vanishes at the dynamic arrest volume fraction @xmath135 .
the results of fig .
[ fig9](b ) show that in the vicinity of @xmath9 , the function @xmath291 follows the power law @xmath292 , i.e. , it vanishes at @xmath9 with the same exponent as @xmath153 vanishes at @xmath355 . at this point
let us notice that the sequence @xmath342 must be identical to the sequence @xmath357 of values of @xmath75 at the times @xmath358du'$ ] .
the sequence of times @xmath340 can be determined by means of the approximate recurrence relationship in eq .
( [ recrel0 ] ) , i.e. , @xmath359 with @xmath360 , and it allows us to transform the sequence @xmath342 into the discrete representation @xmath357 of the function @xmath75 .
the results for @xmath75 are plotted in fig .
[ fig10 ] , to exhibit the fundamentally different behavior of the functions @xmath153 and @xmath75 .
while the former has a well - defined zero at a finite value of its argument , namely , at @xmath355 , the function @xmath75 decays to zero in a much slower fashion .
in fact , as we now discuss , one of the main predictions of the ne - scgle theory is that @xmath75 will remain finite for any finite time @xmath3 , and only at @xmath361 the mobility will reach its asymptotic value of zero .
thus , the system in principle will always remain fluid - like , and the dynamic arrest condition @xmath100 will only be reached after an infinite waiting time . as a function of evolution time @xmath3 for the quench processes from the point @xmath350 to the point @xmath351 .
the inset exhibits the long-@xmath3 asymptotic decay @xmath362 described by eq .
( [ bdtasympt ] ) . ]
let us actually demonstrate that the value of @xmath3 corresponding to @xmath321 is @xmath363 , and that the mobility @xmath75 decays as a power law with @xmath3 . to discuss the first issue , let us recall eq .
( [ tdu00 ] ) , which writes the function @xmath144 as @xmath364 where the function @xmath153 is , of course , @xmath154 .
according to this result , and to eq .
( [ bstardu ] ) , we can write @xmath365 for @xmath156 in some vicinity @xmath366 of @xmath321 .
this implies that , if the exponent @xmath347 is larger than unity , then @xmath145 will diverge as @xmath156 approaches @xmath367 according to @xmath368 as a consequence , the dynamic arrest time @xmath369 will be infinite , which is what we set out to demonstrate .
let us now discuss the possibility that @xmath75 decays as a power law with @xmath3 . for this
, let us invert the function @xmath145 in the previous equation , and write it as @xmath370 since , according to eq .
( [ udt ] ) , @xmath371 , the time derivative of this asymptotic expression will yield the asymptotic form for @xmath75 , namely , @xmath372 with @xmath373^{-\frac{1}{(\beta-1 ) } } , \label{bmincero}\ ] ] and @xmath374 the latter result implies that if one of the exponents ( @xmath347 or @xmath375 ) is larger than unity , then the other is also larger than unity .
it also implies that if one of them is larger than 2 , then the other is smaller than 2 , and viceversa . in the inset of fig .
[ fig10 ] we compare the actual ne - scgle results for @xmath75 in the main figure , with the approximate asymptotic expression in eq .
( [ bdtasympt ] ) with a fitted exponent @xmath375 , with the result that @xmath376 .
this value coincides with the expected result @xmath377 with @xmath378 .
as indicated above , we performed similar calculations varying the initial temperature @xmath275 , and found that the scenario just described is indeed independent of @xmath275 .
thus , in the asymptotic expression in eq .
( [ bdtasympt ] ) only the pre - factor @xmath379 depends on @xmath275 , and the approximate expression in eq .
( [ bmincero ] ) provides an indicative estimate of its actual value .
the properties of the non - equilibrium mobility function @xmath75 that we have just described reveals the main feature of the time evolution of the static structure factor @xmath0 when the system is driven to a point @xmath131 in the region of dynamically arrested states .
we refer to the fact that under such conditions , the long - time asymptotic limit of @xmath0 will no longer be the expected equilibrium static structure factor @xmath380 , but another , well - defined non - equilibrium static structure factor @xmath34 , given by @xmath381 this non - equilibrium static structure factor not only depends on the final point @xmath131 , but also on the protocol of the quench ( in the present instantaneous isochoric quench , this means on the initial temperature @xmath275 ) . to see the emergence of this scenario , let us consider the sequence @xmath338 of snapshots of the static structure factor generated with eq .
( [ solsdktexp ] ) , for the finite sequence @xmath334 of @xmath335 equally - spaced values of @xmath156 defined as @xmath382 with @xmath337 . according to eq .
( [ tdup ] ) , and to its asymptotic version in eq .
( [ tduasympt ] ) , in the present case the finite range @xmath383 maps onto the infinite physically relevant range @xmath384 of the evolution time @xmath3 ( in contrast with the equilibration processes studied in the previous section , in which the infinite range @xmath385 maps onto the infinite range @xmath384 ) .
since the sequence @xmath338 is identical to the sequence @xmath339 , with @xmath386 , then the sequence of snapshots @xmath339 describing the full evolution of @xmath0 will be generated by a sequence of snapshots of @xmath140 with @xmath156 only in the range @xmath383 .
in other words , in the present case none of the snapshots of @xmath140 with @xmath387 will map onto any physically observable snapshot of @xmath0 , and this applies in particular to the snapshot @xmath388 , corresponding to the expected equilibrium static structure factor @xmath33 . in this manner ,
the long time limit of @xmath0 , normally being the ordinary equilibrium value @xmath389^{-1})$ ] , is now replaced by a non - equilibrium dynamically arrested static structure factor @xmath34 given , according to eq .
( [ solsdktexp ] ) , by the expression in eq .
( [ nesadk ] ) . besides the remarkable prediction of the existence of this well - defined non - equilibrium asymptotic limit of @xmath0
, the second relevant feature refers to the kinetics of @xmath0 as it approaches @xmath34 . to exhibit this feature ,
let us subtract eq .
( [ nesadk ] ) from eq .
( [ solsigmadkt ] ) .
this leads to @xmath390}-1\right ] , \label{sdklong}\ ] ] with @xmath391 at long times , when @xmath392 $ ] is small , this equation reads @xmath393 . \label{sdklonglong}\ ] ] from eq .
( [ udtasympt ] ) , however , we have that @xmath394 , so that the previous long - time expression for @xmath0 can be written as @xmath395 with @xmath396 ( thin solid ( green ) lines ) corresponding to the isochoric quench at fixed volume fraction @xmath306 , from @xmath176 to @xmath178 .
the dashed ( black ) line is the initial structure factor @xmath187 .
the dotted ( blue ) line is @xmath184 .
the thick solid ( red ) line is the dynamically arrested asymptotic solution @xmath397 of eq .
( [ relsigmadif2pp ] ) , given by eq .
( [ nesadk ] ) . in the inset ,
the solid line is the maximum of @xmath0 as a function of the evolution time @xmath3 , and the dashed line is the maximum of @xmath398^{-1}(1-e^{-\alpha ( k)b^{(i ) } t})]$ ] . ]
thus , we conclude that , contrary to the kinetics of the equilibration process , in which @xmath0 approaches @xmath33 in an exponential - like fashion , this time the decay of @xmath0 to its stationary value @xmath34 follows a power law . at very short times , however , @xmath399 , and hence , @xmath400 .
thus , according to eqs .
( [ solsigmadkt ] ) and ( [ udt ] ) , we have that the very initial evolution of @xmath0 might seem to approach its expected equilibrium value @xmath401^{-1}$ ] in an apparently exponential " manner , with a relaxation time @xmath402 .
this apparent initial exponential evolution , however , crosses over very soon to the much slower long - time evolution of @xmath0 described by the asymptotic expression in eq .
( [ sdklonglong ] ) .
[ fig11 ] illustrates with a sequence of snapshots the predicted non - equilibrium evolution of @xmath0 after the isochoric quench at @xmath306 from @xmath176 to @xmath274 .
there we highlight the initial static structure factor @xmath403 and the dynamically arrested long - time asymptotic limit @xmath34 of the non - equilibrium evolution of @xmath0 . for reference , we also plot the expected , but inaccessible , equilibrium static structure factor @xmath404 corresponding to the final temperature @xmath274 . regarding the kinetics of the non - equilibrium evolution , in the inset we plot the evolution of the maximum of @xmath0 as a function of @xmath3 to illustrate the fact that @xmath0 approaches @xmath34 much more slowly , in fact as the power law @xmath405\propto t^{-0.83}$ ] .
for reference , we also plot the maximum of the function @xmath398^{-1}(1-e^{-\alpha ( k)b^{(i ) } t})]$ ] which , according to eq .
( [ solsdktexp ] ) , would describe the evolution of @xmath0 if @xmath75 remained constant , @xmath406 .
let us now discuss how the scenario just described manifests itself in the non - equilibrium evolution of the dynamics .
we first recall that for each snapshot of the static structure factor @xmath0 , the solution of eqs .
( [ bdt])-([lambdadk ] ) determines a snapshot at waiting time @xmath3 of each of the dynamic properties of the system .
thus , the process of dynamic arrest may also be observed , for example , in terms of the @xmath3-evolution of the self intermediate scattering function @xmath1 or of the @xmath4-relaxation time @xmath261 . in fig .
[ fig12](a ) we present a sequence of snapshots of the isf @xmath1 ( thin solid lines ) , evaluated at the fixed wave - vector @xmath407 , plotted as a function of correlation time @xmath2 , for a sequence of waiting times @xmath3 after the sudden temperature quench from @xmath176 to @xmath178 at fixed volume fraction @xmath306 . in the figure we highlight with the dashed line the initial isf @xmath408 .
the ( arrested ) non - equilibrium asymptotic limit @xmath409 is indicated by the solid line , whereas the dotted line denotes the inaccessible equilibrium isf @xmath410 , i.e. , the solution of eqs .
( [ dzdtquench])-([lambdadk ] ) in which the final _ equilibrium _ static structure factor @xmath411 ( also inaccessible ) is employed as static input .
we observe that at @xmath20 , @xmath1 shows no trace of dynamic arrest , but as the waiting time @xmath3 increases , its relaxation time increases as well . in the figure we had to stop at a finite waiting time , but the theory predicts that the isf @xmath1 will always decay to zero for any finite waiting time @xmath3 , and continues to evolve forever , yielding always a finite , ever - increasing , @xmath4-relaxation time @xmath412 . the relaxation of @xmath1 is characterized by a fast initial decay ( @xmath347-relaxation ) to an increasingly better defined plateau , whose height @xmath413 is not determined by the expected equilibrium isf @xmath414 , but by the non - equilibrium asymptotic limit @xmath415 . in other words , @xmath413 is the true " non - equilibrium non - ergodicity parameter @xmath416 . at @xmath407 ( thin solid ( green ) lines )
plotted as a function of correlation time @xmath2 for a sequence of values of the waiting time @xmath3 ( @xmath417 and 1600 ) after the sudden temperature quench at fixed volume fraction @xmath306 , from @xmath176 to @xmath178 .
the dashed line is the initial isf @xmath408 , the thick solid ( red ) line is the non - equilibrium asymptotic limit @xmath418 , and the dotted ( blue ) line is the expected ( but inaccessible ) equilibrium isf corresponding to the hard - sphere system at @xmath306 .
( b ) same sequence of snapshots of @xmath1 plotted as a function of the time @xmath3 scaled with the @xmath4-relaxation time @xmath5 .
the dot - dashed line here is the stretched exponential @xmath419}$ ] , title="fig : " ] at @xmath407 ( thin solid ( green ) lines ) plotted as a function of correlation time @xmath2 for a sequence of values of the waiting time @xmath3 ( @xmath417 and 1600 ) after the sudden temperature quench at fixed volume fraction @xmath306 , from @xmath176 to @xmath178 .
the dashed line is the initial isf @xmath408 , the thick solid ( red ) line is the non - equilibrium asymptotic limit @xmath418 , and the dotted ( blue ) line is the expected ( but inaccessible ) equilibrium isf corresponding to the hard - sphere system at @xmath306 .
( b ) same sequence of snapshots of @xmath1 plotted as a function of the time @xmath3 scaled with the @xmath4-relaxation time @xmath5 .
the dot - dashed line here is the stretched exponential @xmath419}$ ] , title="fig : " ] from this sequence of snapshots of @xmath263 we can extract the @xmath3-evolution of the @xmath4-relaxation time @xmath412 defined in eq .
( [ taualphadef ] ) .
the results allows us to notice one of the main features of the predicted long-@xmath2 decay of @xmath1 , namely , the long - time collapse of the curves representing @xmath1 , corresponding to different evolution times @xmath3 ( like those in fig . [ fig12](a ) ) , onto the same stretched - exponential curve upon scaling the correlation time @xmath2 with the corresponding @xmath261 . in other words , at long times @xmath1 scales as @xmath420 where @xmath421 is the height of the plateau of @xmath415 and @xmath422 ( so that @xmath423 ) , and with @xmath347 being a fitting parameter .
this scaling is illustrated in fig .
[ fig12](b ) with the sequence of results for @xmath1 in fig .
[ fig12](a ) now plotted in this scaled manner , which are then well represented by the stretched - exponential function above , with @xmath424 , @xmath425 and @xmath426 .
the non - equilibrium evolution of the dynamics can be summarized by plotting @xmath412 as a function of waiting time @xmath3 .
this is done here in fig .
[ fig13 ] , where we plot @xmath427 as a function of @xmath3 . the thick dark solid line in fig .
[ fig13](a ) and ( b ) derive from the sequence of snapshots of @xmath263 in fig .
[ fig12](a ) , corresponding to the quench at @xmath306 with initial temperature @xmath176 .
as indicated in these figures , at long waiting times we find that @xmath261 increases with @xmath3 according to a power law that is numerically indistinguishable from @xmath428 with @xmath429 .
in other words the present theory predicts , taking into account eq .
( [ bdtasympt ] ) , that at long waiting times , @xmath261 diverges with @xmath3 with the same power law as @xmath277 . besides these results , fig .
[ fig13](a ) also presents theoretical results for two additional quench programs that differ only in the initial temperature , namely , @xmath430 and @xmath431 .
these three initial temperatures lie above the dynamic arrest transition temperature @xmath432 corresponding to the isochore @xmath306 , which is @xmath433 .
the first feature to notice is that the detailed waiting time dependence of @xmath16 at short times may be strongly quench - dependent , but the asymptotic power law @xmath428 with the exponent @xmath429 is independent of the initial temperature @xmath275 . complementing this information , fig .
[ fig13](b ) describes the dependence of the evolution of @xmath434 on the value of the volume fraction @xmath12 at which these isochoric processes occur , assuming that each of them start and end at the same initial and final temperatures , @xmath176 and @xmath274 .
the main feature to notice in these results is that the long - time asymptotic growth of @xmath435 with waiting time @xmath3 is also characterized by the power law @xmath436 .
-relaxation time @xmath437 ( defined in the text ) after the sudden temperature quench at fixed volume fraction @xmath12 , from an initial temperature @xmath275 to a final temperature @xmath178 . in ( a )
we present the results for the initial temperatures @xmath176 , 0.05 , and 0.001 at the same volume fraction @xmath306 . in ( b )
we fix the initial temperature @xmath176 and present results for @xmath306 and other volume fractions .
the dashed lines indicate the asymptotic power - law @xmath438 that fits the results in the indicated regimes.,title="fig : " ] -relaxation time @xmath437 ( defined in the text ) after the sudden temperature quench at fixed volume fraction @xmath12 , from an initial temperature @xmath275 to a final temperature @xmath178 . in ( a ) we present the results for the initial temperatures @xmath176 , 0.05 , and 0.001 at the same volume fraction @xmath306 . in ( b )
we fix the initial temperature @xmath176 and present results for @xmath306 and other volume fractions .
the dashed lines indicate the asymptotic power - law @xmath438 that fits the results in the indicated regimes.,title="fig : " ]
of course , one could continue describing the predictions of the ne - scgle theory regarding the detailed evolution of each relevant structural and dynamic property of the glass - forming system along the process of equilibration or aging . at this point , however , we would like to unite the main results of the previous two sections in a single integrated scenario that provides a more vivid physical picture of the predictions of the present theory . with this intention , in fig .
[ fig14](a ) we have put together the results for the isochoric evolution of @xmath435 previously presented in figs . [ fig8](b ) and [ fig13](b ) , corresponding to the quench at fixed volume fraction @xmath12 , from an initial temperature @xmath176 to a final temperature @xmath178 , for volume fractions @xmath12 smaller and larger than @xmath135 .
-relaxation time @xmath435 for a sequence of fixed volume fractions ( the arrow indicates increasing @xmath12 ) .
( b ) @xmath12-dependence of the @xmath4-relaxation time @xmath435 for the sequence of fixed waiting times @xmath439 ( from bottom to top ) .
the dashed line is @xmath440 , which is the _ equilibrium _
@xmath4-relaxation time of the hard - sphere system , predicted by the equilibrium scgle theory .
the solid circles highlight the crossover points @xmath441 for each waiting time @xmath3 shown .
the inset of ( b ) shows the evolution of the crossover volume fraction @xmath15 predicted by the ne - scgle theory ( solid line ) and determined in the simulations of ref .
@xcite ( solid squares ) .
, title="fig : " ] -relaxation time @xmath435 for a sequence of fixed volume fractions ( the arrow indicates increasing @xmath12 ) .
( b ) @xmath12-dependence of the @xmath4-relaxation time @xmath435 for the sequence of fixed waiting times @xmath439 ( from bottom to top ) .
the dashed line is @xmath440 , which is the _ equilibrium _
@xmath4-relaxation time of the hard - sphere system , predicted by the equilibrium scgle theory .
the solid circles highlight the crossover points @xmath441 for each waiting time @xmath3 shown .
the inset of ( b ) shows the evolution of the crossover volume fraction @xmath15 predicted by the ne - scgle theory ( solid line ) and determined in the simulations of ref .
@xcite ( solid squares ) .
, title="fig : " ] displaying together these results allows us to have a richer and more comprehensive scenario of the transition from equilibration processes to aging processes in the soft - sphere glass - forming liquid , discussed separately in the previous two sections . according to the ne - scgle theory , the dynamic arrest transition is in principle a discontinuous transition , involving the abrupt passage from one pattern of evolution ( equilibration ) to the other ( aging ) when the control parameter @xmath12 crosses the singular value @xmath135 .
the discontinuous nature of this kinetic transition is rooted in the abrupt transition contained in the _ equilibrium _ version of the scgle theory , which actually predicts the existence and location of the dynamic arrest transition line ( see fig .
[ fig13 ] ) .
the zero - temperature limit of this transition line corresponds to the critical volume fraction @xmath135 .
thus , the evolution of the system after the temperature quench from an initial temperature @xmath176 to a final temperature @xmath274 is dramatically different if the volume fraction of the isochoric process is smaller or larger than this critical volume fraction .
however , in order to actually witness this dramatic difference , we would have to perform observations at volume fractions infinitesimally closer to @xmath9 , and within an evolution time window much larger ( in fact , infinite ) than in any real experiment or simulation .
in fact , what we would like to illustrate now is that the experimental observation of the consequences of this theoretically - predicted singularity will be blurred by this unavoidable finiteness of the time window of any experimental observation . to see this ,
let us display the same information presented in fig .
[ fig14](a ) , which plots @xmath435 as a function of @xmath3 for a sequence of volume fractions @xmath12 , in a complementary format .
this is done in fig .
[ fig14](b ) , which plots @xmath435 as a function of @xmath12 for a sequence of waiting times @xmath3 .
the main feature to notice in each of the curves corresponding to a fixed waiting time @xmath3 , is that one can distinguish two regimes in volume fraction , namely , the low-@xmath12 ( equilibrated ) regime and the high-@xmath12 ( non - equilibrated ) regime , separated in a continuous fashion , and not as an abrupt transition , by a crossover volume fraction @xmath15 . focusing , for example , on the results corresponding to @xmath442 , we notice that @xmath443 . in fig .
[ fig14](b ) we have highlighted the crossover points @xmath444 , @xmath445 , corresponding to each waiting time @xmath3 considered .
we observe that the resulting crossover volume fraction @xmath15 first increases rather fast with @xmath3 , but then slows down considerably , reaching a theoretical maximum crossover volume fraction , @xmath446 , given by @xmath135 , as indicated in the inset of the figure .
the scenario illustrated by fig .
[ fig14](b ) has additional physical implications .
although it is impossible to witness the infinite - time implications of the theoretically - predicted singular dynamic arrest transition , it is important to stress that its finite - time consequences , such as those illustrated in this figure , can be predicted , and could be corroborated by performing measurements at _ intentionally _ finite , accessible waiting times .
it is thus important to test if these predictions make sense by comparing them with available experimental or simulation data .
although this comparison falls out of the scope of the present paper , we can say that the picture that emerges from the predicted dependence of @xmath435 on waiting time and volume fraction just discussed , is fully consistent with the most relevant qualitative features observed in a simulation experiment consisting precisely of the equilibration of a hard - sphere liquid , initially prepared in a non - equilibrium state @xcite .
in fact , such simulation experiment was originally inspired by the very initial version of the theoretical scenario of the present non - equilibrium theory .
to have an idea of the level of agreement , in the inset of fig .
[ fig14](b ) we have included the simulation data for the evolution of the crossover volume fraction @xmath15 with waiting time reported in ref @xcite . in a separate communication @xcite we shall analyze in detail other aspects of the comparison between the predictions of the ne - scgle theory and available simulation results , which indicates a general agreement and exhibits some well - defined limitations of the present non - equilibrium theory .
in summary , in this work we have started the systematic exploration of the predicted ne - scgle scenario of the irreversible isochoric evolution of a soft - sphere glass - forming liquid whose temperature is suddenly quenched from its initial value @xmath19 to a final value @xmath28 . as we explained here
, the response falls in two mutually exclusive possibilities : either the system will reach its new equilibrium state within an equilibration time @xmath6 that depends on the fixed volume fraction @xmath12 , or the system ages forever in the process of becoming a glass . in the first case the equilibrium @xmath4-relaxation time @xmath29 , and the equilibration time @xmath6 needed to reach thermodynamic equilibrium , are predicted to remain finite for volume fractions smaller than a critical value @xmath30 , but as @xmath12 approaches this hard - sphere dynamic - arrest volume fraction , both characteristic times will diverge and will remain infinite for @xmath31 .
although it is intrinsically impossible to witness the actual predicted divergence , the theory makes distinct predictions regarding the transient non - equilibrium evolution occurring within experimentally - reasonable waiting times @xmath3 , which could , thus , be compared with realizable experiments or simulations .
this applies even more to the predictions regarding the complementary regime , @xmath32 , in which the system , rather than ever reaching equilibrium , is predicted to age forever . as discussed in the previous section , under these circumstances the long - time asymptotic limit of @xmath0
will no longer be the expected equilibrium static structure factor @xmath33 , but the non - equilibrium , but well - defined , dynamically arrested static structure factor @xmath34 .
furthermore , @xmath0 is predicted to approach @xmath34 in a much slower fashion ( a power law ) , in contrast with the exponential - like manner in which @xmath0 approaches @xmath33 when the system equilibrates .
putting together the two regimes just described , we have presented the scenario predicted to emerge for the crossover from equilibration to aging .
as discussed in the previous section , the discontinuous and singular behavior is intrinsically unobservable in practice , due to the finiteness of the time windows of experimental measurements .
this forces the discontinuous dynamic arrest transition to appear as a blurred crossover , which may depend on the protocol of the experiment and of the measurements .
testing these predictions by comparing them with available experimental or simulation data is an issue that we shall leave for future studies , since the main purpose here was to provide the details of the methodologies needed to solve the equations that define the ne - scgle , and to illustrate its use with the application to the specific system and processes considered here .
as indicated at the end of the previous section , we can say that the picture that emerges from the predicted dependence of @xmath435 on waiting time and volume fraction is consistent with the most relevant qualitative features observed in the simulation experiment of the equilibration of the hard - sphere liquid @xcite . in a separate paper
we shall establish a more direct contact with those simulation results , and with other simulation or experimental data . in the meanwhile
, it will also be interesting to interrogate the ne - scgle theory on the variations of the scenario just described , when the system and conditions employed here are modified .
for example , one may be interested in understanding how this scenario might change when protocol of the quench is modified .
other questions may refer to the dependence of this scenario on the particular class of model system and interactions ( involving here only soft repulsions ) , particularly when attractive forces are incorporated .
the answer to these questions will surely use the methods and experience developed in the presented work , and will be the subject of future research . | the recently - developed non - equilibrium extension of the self - consistent generalized langevin equation theory of irreversible relaxation [ phys .
rev .
e ( 2010 ) * 82 * , 061503 ; ibid .
061504 ] is applied to the description of the irreversible process of equilibration and aging of a glass - forming soft - sphere liquid that follows a sudden temperature quench , within the constraint that the local mean particle density remains uniform and constant . for these particular conditions ,
this theory describes the non - equilibrium evolution of the static structure factor @xmath0 and of the dynamic properties , such as the self - intermediate scattering function @xmath1 , where @xmath2 is the correlation _ delay _
time and @xmath3 is the _ evolution _ or _ waiting _ time after the quench .
specific predictions are presented , for the deepest quench ( to zero temperature ) .
the predicted evolution of the @xmath4-relaxation time @xmath5 as a function of @xmath3 allows us to define the _ equilibration _ time @xmath6 , as the time after which @xmath5 has attained its equilibrium value @xmath7 .
it is predicted that both , @xmath6 and @xmath7 , diverge as @xmath8 , where @xmath9 is the hard - sphere dynamic - arrest volume fraction @xmath10 , thus suggesting that the measurement of _ equilibrium _ properties at and above @xmath9 is experimentally impossible .
the theory also predicts that for fixed finite waiting times @xmath3 , the plot of @xmath11 as a function of @xmath12 exhibits two regimes , corresponding to samples that have fully equilibrated within this waiting time @xmath13 , and to samples for which equilibration is not yet complete @xmath14 .
the crossover volume fraction @xmath15 increases with @xmath3 but saturates to the value @xmath9 . _ _ |
a gapped state of quantum condensed matter is called topological phase when it supports stable gapless boundary modes , such as an edge or a surface state . the integer quantum hall effect
( qhe ) , which exists in @xmath3 spatial dimensions and under a strong magnetic field , is the best known example of such a phase .
the recent discovery of the quantum spin hall effect ( qshe ) in @xmath3 and the @xmath0 topological insulator in @xmath4 @xcite shows topological phases can exist even in @xmath5 spatial dimensions , and can be protected by some discrete symmetries such as time - reversal symmetry ( trs , t ) , particle - hole symmetry ( phs , c ) , and chiral ( or sublattice ) symmetry ( sls , s ) . for non - interacting fermions , an exhaustive classification of topological insulators ( tis ) and superconductors ( tscs )
is proposed in refs .
@xcite : tis / tscs are classified in terms of spatial dimensions @xmath6 and the @xmath7 symmetry classes ( two `` complex '' and eight `` real '' classes ) ( table [ charge ] ) .
the ten symmetry classes are in one - to - one correspondence to the riemannian symmetric spaces ( without exceptional series ) and , as pointed out in @xcite , they are equivalent to k - theory classifying spaces @xcite . for example , the iqhe , qshe , and @xmath0 ti are a topologically non - trivial state belonging to class a ( @xmath3 ) , aii ( @xmath3 ) , and aii ( @xmath4 ) , respectively .
the complete classification of non - interacting tis and tscs opens up a number of further questions , most interesting among which are interaction effects : do non - interacting topological phases continue to exist in the presence of interactions ?
can interactions give rise to novel topological phases other than non - interacting tis / tscs ? what is a topological field theory underlying tis / tscs , which can potentially describe tis / tscs beyond non - interacting examples ?
on the other hand , the ten - fold classification of tis / tscs reminds us of d - branes , which are fundamental objects in string theory , and are also classified by k - theory @xcite ( table [ dbrane ] ) via the open string tachyon condensation @xcite .
it is then natural to speculate a possible connection between tis / tscs and of d - branes . in this paper , we propose a systematic construction of tis / tscs in terms of two d - branes ( d@xmath8- and d@xmath9-branes ) , possibly with an orientifold plane ( o - plane ) . besides the appealing mathematical similarity between tis / tscs and d - branes , realizing tis / tscs in string theory has a number of merits , since string theory and d - branes are believed to be rich enough to reproduce many types of field theories and interactions in a fully consistent and uv complete way . indeed
, our string theory realizations of tis / tscs give rise to massive fermion spectra , which are in one - to - one correspondence with the ten - fold classification of tis / tscs , and come quite naturally with gauge interactions .
these systems , while interacting , are all topologically stable , as protected by the k - theory charge of d - branes .
we thus make a first step toward understanding interacting tis / tscs @xcite .
we are also separately preparing a regular paper with more details and expanded results in @xcite . in d@xmath8-d@xmath9-systems ,
massive fermions arise as an open string excitation between the two d - branes .
the distance between the branes corresponds to the mass of fermions .
open strings ending on the same d - branes give rise to a gauge field , which we call @xmath10 ( d@xmath8 ) and @xmath11 ( d@xmath9 ) with gauge group @xmath12 and @xmath13 , respectively , and couple to the fermions .
these two gauge fields play different roles in our construction : the gauge field @xmath10 `` measures '' k - theory charge of the d@xmath9-brane , and in that sense it can be interpreted as an `` external '' gauge field . in this picture , the d@xmath9-brane charge is identified with the topological ( k - theory ) charge of tis / tscs . on the other hand , @xmath11 is an internal degree of freedom on the d@xmath9-brane . for example , in the integer / fractional qhe , the external gauge field is the electromagnetic u(1 ) gauge field , which measures the hall conductivity , while the internal gauge field is the chern - simons ( cs ) gauge field describing the dynamics of the droplet itself .
the massive fermions can be integrated out , yielding the description of the topological phase in terms of the gauge fields .
the resulting effective field theory comes with terms of topological nature , such as the cs or the @xmath14-terms . in our string theory setup
, they can be read off from the wess - zumino ( wz ) action of the d - branes , by taking one of the d - branes as a background for the other .
one can view these gauge - interacting tis / tscs from d@xmath8-d@xmath9-systems as an analogue of the projective ( parton ) construction of the ( fractional ) qhe @xcite .
our string theory realization of tis / tscs sheds light on extending the projective construction of the qhe to more generic tis / tscs ; it tells us what type of gauge field is `` natural '' to couple with fermions in topological phases , and guarantees the topological stability of the system . .
[ charge ] classification of topological insulators and superconductors @xcite ; @xmath6 is the space dimension ; the left - most column ( a , aiii , @xmath15 , ci ) denotes the ten symmetry classes of fermionic hamiltonians , which are characterized by the presence / absence of time - reversal ( t ) , particle - hole ( c ) , and chiral ( or sublattice ) ( s ) symmetries of different types denoted by @xmath16 in the right most three columns .
the entries `` @xmath17 '' , `` @xmath0 '' , `` @xmath18 '' , and `` @xmath19 '' represent the presence / absence of topological insulators and superconductors , and when they exist , types of these states ( see ref .
@xcite for detailed descriptions ) .
[ cols="^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ]
the main conclusion of this paper is that there is a one - to - one correspondence between the tachyon - free d@xmath8-d@xmath9 systems and the ten classes of topological insulators in @xmath20 dimensions .
indeed , we explicitly constructed the corresponding ten classes of d@xmath8-d@xmath9 brane configurations in superstring theory .
two out of ten are realized in type ii string theory without orientifolds , while the other eight require orientifolds .
the k - theory charges of the d@xmath9-branes agree with that of the topological insulators .
one may wonder if there are other tachyon - free d@xmath8-d@xmath9 systems which have not been considered in this paper .
however , it turns out that their low - energy theories just correspond to multiple copies of the ten classes of topological insulators ( for details refer to @xcite ) .
since a topological insulator has a mass gap in the bulk , the distance between the d@xmath8 and d@xmath9 is taken to be non - vanishing in its corresponding d - brane system .
when we discuss the boundary of a ti / tsc , however , the d@xmath8-brane is bent toward the d@xmath9-brane and thus d@xmath8 and d@xmath9 are intersecting with each other .
therefore massless fermions appear at the intersection and they are identified with the boundary modes ( or edge modes ) .
we can also consider holographic descriptions of these systems by extending the constructions in @xcite in principle , though it is not possible to take the large-@xmath21 limit of @xmath0 charged d - branes .
we acknowledge `` quantum criticality and the ads / cft correspondence '' miniprogram at kitp , `` quantum theory and symmetries '' conference at university of kentucky .
we would like to thank k. hori , p. kraus , a. ludwig , j. moore , m. oshikawa , and s. sugimoto for useful discussion , and a. furusaki for his clear lecture at `` development of quantum field theory and string theory '' ( yitp - w-09 - 04 ) at kyoto university .
sr thanks center for condensed matter theory at university of california , berkeley for its support .
tt is supported in part by jsps grant - in - aid for scientific research no . 20740132 , and by jsps grant - in - aid for creative scientific research no . | realization of topological insulators ( tis ) and superconductors ( tscs ) , such as the quantum spin hall effect and the @xmath0 topological insulator , in terms of d - branes in string theory is proposed .
we establish a one - to - one correspondence between the k - theory classification of tis / tscs and d - brane charges .
the string theory realization of tis and tscs comes naturally with gauge interactions , and the wess - zumino term of the d - branes gives rise to a gauge field theory of topological nature .
this sheds light on tis and tscs beyond non - interacting systems , and the underlying topological field theory description thereof .
topological insulators and superconductors from d - brane + shinsei ryu@xmath1 and tadashi takayanagi@xmath2 |
biological channels are responsible for regulating the fluxes of ions and molecules ( hereafter referred to as particles for short ) across membranes and , therefore , are critically important for the cell functioning @xcite . as well - known , these protein channels are very efficient in the sense that they support a very fast , selective , and robust across - membrane transport , regardless of environment fluctuations @xcite . surprisingly , such privileged properties have been observed even in the case of large water - filled channels , where the particle transport does not involve the use of metabolic energy or conformational changes and was assumed to be simply diffusive @xcite .
understanding the nature of this channel - facilitated particle movement ( cfpm ) is crucially important from the fundamental molecular biology as well as the application point of view ( many modern drugs are developed in the way of using the ion - channels to enhance their efficiency , see for example refs.@xcite ) . experimentally , there are accumulative data showing that the observed cfpm is really resulted from some interaction between the moving particle and the channel - interior lining @xcite
. recent advancements of high - resolution current recording enable single - channel measurements that provide directly a living picture of how an individual channel functions and , therefore , shed light on the characteristics of channel current in dependence on different ( channel and environment ) parameters @xcite .
however , revealing exactly the nature of channel - particle interaction as well as the mechanism of cfpm is still very experimentally problemic due to the puzzled complexities related to both the channel structure and the measurement systems . theoretically , to describe the cfpm several models have been suggested .
considering the one - dimensional ( 1d ) diffusion model with a position - dependent diffusion coefficient , berezhkovskii et al .
supposedly introduced a square potential well , spanning the whole channel length , that brings about a channel - particle interaction @xcite .
it was then shown that at a given solute concentration difference there exists an optimum potential well depth that can maximize the particle current , facilitating the channel function . in this model
@xmath0 the channel is assumed large enough so that all the effects related to the particle size can be omitted , @xmath1 a single - particle diffusion is considered , neglecting all many particle correlations , and , particularly , @xmath2 no realistic potential was assigned as the source for the square potential well introduced .
bauer and nadler considered a similar 1d diffusion model with a square potential well that is however associated locally with only the particle bound temporarily inside the channel @xcite . using the macroscopic version of fick s equation
, it was then demonstrated that a transport increase always occurs for any square potential wells .
however , as already noted by the authors , the square potential well exploited in this model is also rather crude and a more realistic potential should be found @xcite . from the very other point of view , kolomeisky models the channel as a set of discrete binding sites arranged stochastically @xcite .
in such the discrete - state model the particles are assumed to hop along the binding sites in translocations across the channel and the optimum current may be achieved depending on the spatial distribution of binding - sites and the site - particle interactions @xcite .
this model is so simple that the main dynamic properties of the problem can be calculated exactly .
it was also demonstrated that the discrete - state model @xcite and the continuum diffusion model @xcite are closely related and can be effectively mapped into each other @xcite . nevertheless , like the square potential well in the continuum models @xcite , the nature of the binding sites ( a kind of channel - particle interaction ) and the hopping mechanism of particles in the discrete - state model @xcite still need to be identified .
importantly , in all the models mentioned @xcite the channel - particle interaction ( which was expressed by a square potential well or a binding site ) is generally viewed as the crucial condition for the transmembrane transport to be facilitated ( see also @xcite ) .
note again that in these models the particle motion is merely considered one - dimensional .
recently , dettmer et al . have measured the diffusivity of spherical particles in closely - confining , finite length channels @xcite .
measurements demonstrated a strongly anisotropic diffusion in the channel interior : while the diffusion coefficient parallel to the channel axis remained constant throughout the entire channel interior , the perpendicular diffusion coefficient showed an almost linear decrease from the axis towards the channel wall .
these observations put forward a need for the two - dimensional ( 2d ) description with direction - dependent diffusion coefficients when studying the movement of particles inside a large channel .
furthermore , experimentally , the single channel kinetics was extensively studied at different external voltages @xcite . and , the experimental sublinear current - voltage ( i - v ) characteristics reported in refs.@xcite is often used as one of the basic requirements for theoretical models @xcite . in the present paper
we consider a 2d diffusive movement of particles through a large water - filled channel , taking into account an anisotropy of diffusion coefficients as observed in ref.@xcite and an influence of external voltage as discussed in refs.@xcite .
the channel is modeled as a cylinder characterized by three structure parameters : radius , length , and surface density of negative charges of channel interior - lining .
the potential created by this charged interior - lining inside the channel is exactly calculated .
it causes the channel - particle interaction " that plays a key role in facilitating the transmembrane particle movement . solving the 2d stochastic langevin equation for the model
suggested we systematically analyze the typically dynamical characteristics of particles such as the translocation probabilities , the translocation times , the currents , and the channel ion conductance under the influence of various factors : the channel - induced potential , the external voltage , or the difference in reservoir particle concentrations .
it was particularly shown that to facilitate the transmembrane particle movement the channel should be reasonably self - optimized with appropriate structure parameters so that its potential coincides with the resonant one .
in addition , this facilitation is very selective in the sense that a channel of definite structure parameters can facilitate the transmembrane movement of only particles of proper valence at corresponding temperatures .
so , the model suggests that facilitating the transmembrane particle movement is an intrinsic property of biological channels .
this property is independent of the external factors such as the external voltage or the bulk particle concentration gradient , though these factors may strongly influence the magnitude of various particle dynamical characteristics .
the paper is organized as follows .
sec.ii introduces the 2d diffusion model for the problem under study , including the motion equation with an exact expression of the channel - induced potential , and describes the calculating method .
sec.iii presents the main numerical results obtained .
these results are discussed in great detail , showing the influence of various factors on the particle dynamical characteristics .
a particular attention is given to the self - optimized property of the channels in facilitating the transmembrane particle movement .
the paper concludes with a brief summary in the last sec.iv
we consider a cylindrical channel of length @xmath3 and radius @xmath4 that connects the two reservoirs with particle concentrations @xmath5 and @xmath6 as schematically drawn in fig.1@xmath7 .
the channel interior - lining carries negative charges which are for simplicity assumed to be continuously and regularly distributed with a surface density @xmath8 .
( the cation channels are believed to contain a net negative charge in the pore lining region of the protein @xcite . in the case of potassium and gramicidin channels
this is due to the partially charged carbonyl oxygens @xcite ) .
these negative surface charges create an electrostatic potential @xmath9 which affects the movement of particles inside the channel .
particles are assumed to diffuse independently , neglecting any many - particle correlation .
in addition , the diffusivity of a particle inside the channel is assumed anisotropic with the two different diffusion coefficients , @xmath10 ( parallel with ) and @xmath11 ( perpendicular to the channel axis ) .
following ref.@xcite , we assume that @xmath0 @xmath10 is constant throughout the channel cylinder [ @xmath12 and @xmath13 and somewhat smaller than the diffusion coefficient @xmath14 in the bulk , @xmath15 with @xmath16 [ we choose in the present work @xmath17 for definition ] and @xmath1 @xmath11 linearly decreases as @xmath18 going from the channel axis ( where the diffusion is isotropic ) to the channel wall , @xmath19 d_z$ ] .
the model also involves a longitudinal voltage @xmath20 , i.e. the difference in electrical potential between the two channel ends , that may include the intrinsic membrane potential @xcite and/or some externally applied voltage @xcite .
this voltage drives the particles moving along the channel . for definition
, we assumed that the voltage @xmath20 is directed from the left to the right [ in fig.1@xmath7 ] and the charge @xmath22 carried by a particle is positive .
actually , due to the cylindrical symmetry of the channel model suggested , the motion of a particle inside a channel can be effectively described by the 2d stochastic differential equation ( langevin equation for overdamped motion ) : @xmath23 [ t ] model of the cylindrical channel under study ; @xmath24 channel - induced potential @xmath25 of eq.(3 ) is plotted for the channel with @xmath26 , and @xmath27 [ note : @xmath28 is symmetrical with respect to the sign of @xmath18 ] ; @xmath30 the @xmath31-potential well ( red - solid line , see the left and bottom axes ) and the @xmath32-potential barrier ( blue - dashed line , see the right and top axes ) for the potential @xmath28 in @xmath24 .
the potential @xmath33 in this case.,title="fig : " ] where @xmath34 and @xmath35 are the 2d - coordinates of the particle at @xmath36-time , @xmath37 , @xmath38 is the drag coefficient in the @xmath39-direction , @xmath28 is the potential created by the charged channel lining , @xmath40 is the voltage - induced force acting on a particle of charge @xmath22 in the @xmath41-direction , and @xmath42 is the random force in the @xmath43-direction which is as usual assumed to have a zero mean and a white noise correlation : @xmath44 it is here worthy to mention the stokes - einstein relation between the diffusion coefficient @xmath45 , the drag coefficient @xmath46 , and the absolute temperature @xmath47 of a medium : @xmath48 , where @xmath49 is the boltzmann constant . in eq.(1 ) we need to identify the potential @xmath28 inside the channel . within the model considered , as mentioned above , @xmath9 is the electrostatic potential created by the charged lining of a cylindrical channel . by solving the fundamental electrostatic problem for a charged cylinder of finite sizes , we can exactly derive an analytical expression of @xmath9 as a function the @xmath50-coordinates [ @xmath51 and @xmath52 : @xmath53 \nonumber \\ & - & \ln [ \frac { z + \sqrt{(x + r)^2 + z^2}}{z - l + \sqrt{(x + r)^2 + ( z - l)^2 } } ] \ \ } , \end{aligned}\ ] ] where @xmath54 , and @xmath8 are the channel structure parameters defined above , @xmath55 is the vacuum permittivity , and @xmath56 is the dielectric constant of the water in the interior of the channel @xcite . note that the potential @xmath28 is symmetrical with respect to the sign of @xmath18 . as an example
, fig.1@xmath24 shows the potential @xmath28 of eq.(3 ) for the channel with @xmath57 , and @xmath27 . at a given @xmath18-coordinate
, @xmath58 behaves as a symmetrical potential well with the absolute minimum at @xmath59 . on the contrary ,
given a @xmath41-coordinate , the @xmath60-curve describes a symmetrical potential barrier with the absolute maximum at @xmath61 [ see , for example , @xmath31 as a function of @xmath41 ( bottom and left axes ) and @xmath32 as a function of @xmath18 ( top and right axes ) in fig.1@xmath30 ] .
while the well shape of the channel potential @xmath28 in the @xmath41-direction directly affects the movement of particles across the channel ( as will be seen below ) , its barrier shape in the @xmath18-direction demonstrates a noticeable role of the transverse motion in the anisotropic 2d - diffusion model considered . as a consequence of the observed symmetrical shape
, the potential @xmath28 can be characterized by its value at the center of the channel , @xmath63 , where @xmath64 \ \equiv \ u_0 .\ ] ] this potential value @xmath65 is uniquely determined by the channel structure parameters ( @xmath66 , and @xmath8 ) and can be used to characterize the potential @xmath28 on the whole : each channel creates a unique @xmath28 and each @xmath28 has a unique @xmath65 . as an intrinsic characteristics of the channel , the quantity @xmath65 will be used below as a typical measure of the channel potential @xmath28 .
fig.1@xmath30 indicates the potential @xmath67 for the channel potential @xmath28 examined in this figure .
thus , as an extension of the model suggested by berezhkovskii et al.@xcite , the present model is distinguished by the main factors as follows : @xmath0 the diffusion is anisotropically two - dimensional ( see eq.(1 ) ) , @xmath1 the negatively charged channel interior - lining creates inside the channel a potential that leads to the first term in the right hand of eq.(1 ) and that can be exactly identified as a function of only channel structure parameters ( see eq.(3 ) ) , and @xmath2 the external voltage causes a driving force expressed by the second term in the right hand of eq.(1 ) .
further , the study will be focused on showing how these factors affect the dynamical characteristics of particles moving through the channel .
the dynamical characteristics we are here interested in include the translocation probabilities , the translocation times , the particle current , and the ion conductance .
to calculate these quantities we have to solve eq.(1 ) .
reasonably , this stochastic equation can be solved numerically by using the molecular dynamics method @xcite .
a particle enters the channel from either the left ( @xmath68 ) or the right ( @xmath69 ) at random with the probabilities proportional to the reservoir particle concentration @xmath5 or @xmath6 , respectively [ fig.1@xmath7 ] .
the initial @xmath18-coordinate ( @xmath70 ) and the initial velocity components ( @xmath71 and @xmath72 ) are randomly given , following the standard molecular dynamics simulation procedure @xcite .
started from the given initial conditions , a discrete trajectory of the particle is step by step constructed .
given the channel potential @xmath28 and the external voltage @xmath20 , in each time - step ( @xmath73 ) the random forces , @xmath74 , are independently generated and then the final coordinates and velocity of the particle are determined from eq.(1 ) using the well - known euler scheme @xcite . in the @xmath18-direction
the full reflection condition is applied every time when a particle runs into the channel wall , @xmath75 . in the other direction ,
once the @xmath41-coordinate is out of the range @xmath76 $ ] , the data for the simulated particle is fixed and this particle is no longer followed .
the next particle enters the channel and undergoes a diffusion process in the same way as described above .
the number of particles involved in getting each of average values of studied dynamical quantities is so large that for all the data points presented below the error bar nowhere exceeds the symbol size [ @xmath77 to @xmath78 particles depending on the quantity and the direction of movement investigated ] .
the time step is taken to be @xmath79 , which is believed small enough .
the dynamical quantities we are interested in , as mentioned above , include the translocation probabilities , the average translocation times , the net particle current , and the ion conductance . actually , the calculating method we exploit in this study is the brownian dynamics . by solving the langevin equation
, this method is rather appropriate for the problem of interest .
a systematical classification of computational approaches proposed and employed for studies of ion channels can be found in the review paper @xcite . here , in solving numerically eq.(1 ) , for convenience we choose @xmath3 as the unit of length , @xmath80 as the unit of time , and @xmath81 as the unit of energy .
so , for example , if @xmath82 and @xmath83 @xcite , then @xmath84 . remind that @xmath85 for @xmath86 .
in presenting simulation results we introduce for short the symbols @xmath87 ( referred to as the effective channel potential ) and @xmath88 ( referred to as the effective external voltage ) .
so , the defined parameters @xmath89 and @xmath90 also contain the particle charge @xmath22 and the medium temperature @xmath47 .
we should keep this in mind when discussing the role of the channel potential in facilitating the transmembrane particle movement . additionally , for definition , in all the figures relating to the translocation probabilities and the average translocation times the parameters @xmath4 and @xmath91 are kept constant : @xmath92 ( in unit of @xmath3 ) and @xmath93 @xcite .
influences of these parameters will be later discussed when analyzing the net current [ fig.6 ] .
let us first examine obtained results for the translocation probabilities which are separately calculated for the particles moving through the channel from the left to the right ( @xmath94 ) and for those moving in the opposite direction ( @xmath95 ) [ see fig.1 with the @xmath20-direction indicated ] . in simulations ,
the probability @xmath94 ( or @xmath95 ) is determined as the ratio of the number of particles that passed through the channel to the total number of particles that entered the channel from the left ( or right ) .
[ t ] ( blue dash - dotted lines ) and @xmath95 ( red dashed lines ) are plotted versus @xmath96 for the channels of the same @xmath97 , and @xmath6 , but at different voltages @xmath88 : @xmath98 and @xmath99 , and @xmath100 and @xmath101 .
the points are the simulation results , whereas the lines are drawn as a guide for the eyes [ @xmath102 .
, title="fig : " ] fig.2 shows @xmath94 ( blue dash - dotted lines ) and @xmath95 ( red dashed lines ) plotted versus @xmath89 for the channels at different effective voltages @xmath90 : @xmath98 and @xmath99 , and @xmath100 and @xmath101 .
generally , this figure demonstrates that with increasing @xmath89 both the translocation probabilities , @xmath94 and @xmath95 , increase steadily first [ see main figure ] and then become saturated [ see the inset ] .
such the @xmath103-versus-@xmath89 behavior is observed at any voltage @xmath90 . in the case of zero @xmath90 , due to the left - right symmetry of the potential @xmath28 of eq.(3 )
the two curves , @xmath94 and @xmath95 , are totally coincidental and the common curve may be in a qualitative comparison with fig.3 in ref.@xcite(where the considered diffusion is one - dimensional and the potential well is square ) . note that with the chosen direction of @xmath20 [ fig.1 , @xmath22 is positive ] the external voltage raises @xmath94 ( two higher curves ) while suppressing @xmath95 ( two lower curves ) , compared to the case of @xmath104 ( the middle curve ) .
[ t ] ( blue dash - dotted lines ) and @xmath105 ( red dashed lines ) are plotted versus @xmath90 for the channels with different effective potentials @xmath89 : @xmath106 and @xmath107 and @xmath99 , and @xmath100 and @xmath101 .
other parameters and symbols are the same as in fig.2 . , title="fig : " ] the external voltage effects can more clearly be seen in fig.3 where the probabilities @xmath103 are presented as the functions of @xmath90 for the channels with different @xmath89 : @xmath108 and @xmath109 ) ; @xmath110 and @xmath111 ) ; and @xmath100 and @xmath112 ) . at zero @xmath90
the two probabilities @xmath103 associated to the same @xmath89 are of equal value [ two corresponding curves are started from the same point ] . with increasing @xmath90
the probability @xmath94 smoothly rises , while the probability @xmath95 strongly descends . at @xmath113
the probabilities @xmath95 become practically vanished for all the channels under study [ no particle can move through the channel in the right - to - left direction ] .
the probability @xmath94 , on the contrary , continues to grow with the tempo that gradually slows down at higher @xmath90 .
calculations reveal that even at @xmath114 the channels are still not perfectly transparent for the positively charged particles moving along the external voltage direction [ @xmath115 or 0.95 for @xmath116 or 1 , respectively ] .
next , we consider another fundamental characteristics - the average translocation time . in accordance with the probabilities @xmath103 studied in figs.2 - 3 ,
we separately calculated the average translocation times for the particles moving through the channel from the left to the right ( @xmath117 ) and for those moving in the opposite direction ( @xmath118 ) . in simulations , we count the time each of simulated particles spends inside the channel . the average translocation time @xmath117 ( or @xmath118 ) is then obtained by averaging these spending times over all the particles that passed through the channel from the left to the right ( or from the right to the left ) .
[ t ] ( blue dash - dotted lines ) and @xmath118 ( red dashed lines ) are plotted versus @xmath89 at different @xmath90 : @xmath119 and @xmath107 and @xmath99 , and @xmath100 and @xmath101 @xmath7 and @xmath120 versus @xmath90 at different @xmath89 : @xmath106 and @xmath107 and @xmath99 , and @xmath100 and @xmath101 ] @xmath24 .
other parameters are the same as in fig.2 .
, title="fig : " ] fig.4 shows how obtained translocation times @xmath120 vary with the effective potential @xmath89 [ fig.4@xmath7 ] or the effective voltage @xmath90 [ fig.4@xmath24 ] .
interestingly , in all the cases studied in both the figures , fig.4@xmath7 and fig.4@xmath24 , the two points , corresponding to @xmath117 and @xmath118 , are practically coincided .
so , our 2d - simulations suggest a general equality , @xmath123 , that should be always valid in the model studied regardless of the shape of the channel potential @xmath28 as well as the presence of the external voltage @xmath20 .
this really causes some surprise , noting on the directed influence of the voltage @xmath20 .
actually , a similar equality of the two average translocation times has been previously suggested in ref.@xcite , but it was there relating to the 1d diffusion model without any external voltage .
fig.4 thus allows us to deal with the two times @xmath117 and @xmath118 as a single average translocation time that will be below denoted simply by @xmath124 .
the fact that the channel potential @xmath89 raises the translocation time @xmath124 in fig.4@xmath7 , while it also raises the translocation probabilities in fig.2 , might cause some surprise . actually ,
as will be seen below , it turns out that a competition between these two seemingly contrary effects of the potential @xmath89 leads to the most important phenomenon in the ion - channel physics - the cfpm . comparing the points from three curves with different voltages @xmath90
, we learn that in the region of large @xmath125 in fig.4@xmath7 ] the time @xmath124 decreases almost linearly as @xmath90 increases from @xmath126 to @xmath127 . in a wider range of @xmath90
, fig.4@xmath24 shows that the larger the effective potential @xmath89 , the stronger the relative effect of @xmath90 on @xmath124 becomes . in the limit of high external voltage when the @xmath90-induced driving force becomes to dominate the right hand in eq.(1 ) , the translocation time should depend on @xmath90 as @xmath128 .
while the question of the particular time that most relevantly describes the transmembrane transport and that can be directly measured is still under discussion @xcite , the net current has always served as the most important quantity that should be determined theoretically in close comparison with experimental measurements . for the problem under study , the net particle current is determined as the average number of particles the two reservoirs actually exchanged via the channel in a unit of time ( @xmath129 ) .
[ t ] as a function of @xmath89 and @xmath90 .
note on the resonant behavior of the @xmath130 versus @xmath89 curves at different voltages @xmath90 .
other parameters are the same as in fig.2 , title="fig : " ] fig.5 presents a 3d - plot of the current @xmath130 in dependence on the effective channel potential @xmath89 and the effective external voltage @xmath90 .
remarkably , contrary to the monotonic behaviors of @xmath103 and @xmath124 in figs.2 - 4 , fig.5 shows clearly a resonant behavior of the current @xmath130 : for a given voltage @xmath90 in the @xmath130 versus @xmath131 curve there always has an impressively absolute maximum at some resonant channel potential , @xmath132 .
remind that @xmath133 with @xmath65 uniquely determined by the channel structure parameters ( @xmath66 , and @xmath8 ) .
so , the maximum observed in fig.5 implies that for given @xmath22 and @xmath47 , to successfully facilitate the transmembrane particle movement the channel has to be optimized with the appropriate structure parameters so that its potential @xmath28 coincides with the resonant one .
for example , for @xmath134 and @xmath135 , to own the resonant potential of @xmath136 as seen in fig.5 , the channel should be self - optimized with the following structure parameters : @xmath137 , and @xmath27 ( given @xmath138 @xcite ) .
[ t ] is an intrinsic characteristics of the channel .
@xmath7 @xmath130 versus @xmath89 curves extracted from fig.5 for some values of @xmath90 [ from bottom : @xmath139 , and 5 ] ; all these curves show their maximum at the same resonant potential , @xmath132 ( indicated by the arrow ) .
@xmath24 @xmath130 as a function of @xmath89 at @xmath140 for various values of the ratio @xmath141 [ from top : @xmath142 , and @xmath143 ; the resonant potential @xmath132 ( indicated by the arrow ) is independent of reservoirs particle concentration ratio and coincides with @xmath144 determined in fig.6@xmath7 .
note : in the case of @xmath145 the current is negative ( flowing from right to left in fig.1@xmath7 ) and reaches the largest magnitude at the same @xmath132 .
, title="fig : " ] to see whether the resonant potential @xmath144 depends on the external voltage @xmath90 , we depict in fig.6@xmath7 some @xmath146-curves extracted from fig.5 at various @xmath90 .
surprisingly , the resonant channel potential @xmath144 ( indicated by the arrow ) is practically the same for all the curves at different voltages @xmath90 .
actually , the fact that @xmath144 is independent of @xmath90 can be seen right in fig.5 for all the values of @xmath90 under study .
further , we check if the resonant potential @xmath144 depends on another important external parameter , the difference in particle concentration between the two reservoirs .
fig.6@xmath24 presents the @xmath146-dependence for several values of the ratio @xmath141 . in the cases of @xmath147 , and @xmath148 ( taken from table 12.1 in ref.@xcite ) , all the particle concentration gradients are directed along the external voltage @xmath20 , i.e. from the left to the right in fig.1@xmath7 , and , therefore , the currents are always positive [ see the higher three curves ] . on the contrary , in the case of the lowest curve in fig.6@xmath24 for @xmath145 ( e.g. for @xmath149-channels @xcite ) , the particle concentration gradient is directed from the right to the left in fig.1@xmath7 and , consequently , the current becomes negative ( note that in this case the concentration gradient is strong while the external voltage is relatively small , @xmath140 ) .
importantly , all the curves for various @xmath141 in fig.6@xmath24 show the maximums in magnitude at the same value of @xmath89 that exactly coincides with the resonant potential @xmath144 determined in fig.6@xmath7 .
thus , we arrive at an important remark : at a given @xmath22 and @xmath47 , the resonant potential is entirely determined by the channel structure parameters .
it is an intrinsic property of the channels and can not be affected by the external factors such as the external voltage or the particle concentration gradient .
[ t ] the current @xmath130 is plotted versus the voltage @xmath90 [ i - v characteristics ] for channels with different potentials @xmath89 ( indicated in the figure ) .
all i - v curves show a sublinear behavior .
@xmath24 the channel ion conductance @xmath150 as a function of the channel potential @xmath89 at various voltages @xmath90 ( given in the figure ) . at any @xmath90
the conductance always has the maximum at the same @xmath151 .
inset : @xmath150 versus @xmath90 for the channel with resonant potential @xmath132 .
, title="fig : " ] thus , in the present model , cfpm appears to be a self - optimized property of biological channels : to facilitate the transmembrane particle movement , the channels should be intrinsically optimized with appropriate structure parameters .
additionally , facilitating the transmembrane transport is very selective in the sense that a channel of definite structure parameters can facilitate the transmembrane transport of only particles of proper valence at corresponding temperatures .
furthermore , we demonstrate in fig.7@xmath7 the current - versus - voltage curves , @xmath152 ( i - v ) - characteristics , extracted from fig.5 for several channels of different @xmath89 .
the highest curve describes the i - v characteristics of the resonantly self - optimized channel with @xmath132 .
it is clear that all the @xmath152-curves presented in this figure are nonlinear , indicating the non - ohmic property of the channel model studied . in this case
, the channel ion conductance , defined as the ratio of @xmath130 to @xmath90 @xcite , becomes dependent on the applied voltage .
the @xmath152-curves in fig.7@xmath7 reveal that as @xmath90 increases the conductances @xmath153 decrease fast first at small @xmath90 , reach a minimum at @xmath154 , and then slightly increase at higher @xmath90 ( see , for example , the inset in fig.7@xmath24 for the case @xmath132 ) .
the voltage , where the conductance gets minimal , depends on the potential @xmath89 and the reservoir particle concentrations . to look for a possible relation between the channel ion conductance and the resonant channel potential @xmath144 associated with the current [ fig.6 ] , we present in fig.7@xmath24 the conductances @xmath150 calculated for channels of different potentials @xmath89 at the same voltage @xmath90 .
remarkably , at any @xmath90 the conductance @xmath150 always has the maximum at the same @xmath132 as that identified in fig.6 for the current .
so , fig.7@xmath24 gives one more demonstration for the resonantly self - organized property of channels in facilitating the transmembrane particle movement .
the @xmath90-dependence of @xmath150 in this figure is related to the sublinear behavior of the i - v curves in fig.7@xmath7 as just discussed above ( see the inset in fig.7@xmath24 ) .
note that such the sublinearity of calculated @xmath155 curves qualitatively resembles experimental data reported refs.@xcite .
[ t ] -curves for different @xmath11 ( indicated in the figure ) are compared to show the role of the transverse diffusion in the 2d - diffusion model considered [ @xmath156 , @xmath157 , other parameters are the same as in fig.2 ] .
, title="fig : " ] finally , fig.8 compares the @xmath146-curves obtained for different @xmath11 in showing the role of the transverse diffusion in the 2d - diffusion model under study .
four cases presented are : ( 1 ) @xmath158 , implying the 1d - diffusion , ( 2 ) @xmath159 , implying an anisotropically 2d - diffusion with @xmath11 constant and smaller than @xmath10 , ( 3 ) @xmath160 , implying an isotropically 2d - diffusion , and ( 4 ) @xmath161 d_z$ ] used in this work ( see corresponding symbols given in the figure ) .
obviously , the curve in the case of 1d - diffusion is largely separated from the rest , showing an essential role of the transverse diffusion . in this limiting case
there has also the maximum in the @xmath146-curve , however , the current peak is lower and the resonant potential is much larger ( @xmath162 ) , compared to those for 2d - diffusion models .
interestingly , all the three 2d - diffusion @xmath146-curves with @xmath11 different on the behavior or the value show very similar forms with the same resonant potential @xmath136 .
in addition , the isotropic 2d - diffusion model , @xmath160 , provides the highest current peak .
we have considered an anisotropic 2d - diffusion of a charged molecule ( particle ) through a large biological channel under an external voltage .
connecting the two reservoirs with different particle concentrations , the channel is modeled as a rigid cylinder characterized by the three structure parameters : the radius , the length , and the surface density of the negative charges of channel interior - lining .
these negative charges induce inside the channel a potential that is uniquely determined by the channel structure parameters and that critically affects the transmembrane particle movement .
the suggested model is rather phenomenological so that the channel - induced potential can be calculated exactly .
nevertheless , it serves well to gain an understanding of the physical mechanism of the channel - facilitated particle movement .
more detailed quantitative models are required to describe concrete realistic biological channels ( see for example , @xcite ) .
our study is concentrated on showing the influences of this channel - induced potential and the external voltage on the typical dynamical characteristics of particles such as the translocation probabilities , the average translocation times , the net current , and the channel ion conductance .
it was shown that while the external voltage does not cause any especial effect , the channel potential increases both the translocation probabilities and the average translocation times . and ,
surprisingly , studies demonstrated a single average translocation time that is equally applied for the particles passing the channel in two contrary directions , regardless of even the directed influence of the external voltage .
the most interesting result was appeared in examining the particle current .
it was shown that at a given temperature the channel with appropriate structure parameters can induce the resonant potential that effectively facilitates the transmembrane movement of the particles of a given valence , resulting in a very large net particle current across the channel .
in other words , to facilitate the transmembrane particle movement the channel should be naturally self - optimized so that its potential coincides with the resonant one .
the resonant potential is an intrinsic characteristics of the channel and facilitating the transmembrane particle movement is an intrinsic property of biological channels , independent of the external factors such as the external voltage or the particle concentration gradient .
in addition , the observed cfpm is very selective in the sense that a channel of definite structure parameters can facilitate the transmembrane movement of only particles of proper valence at corresponding temperatures .
calculated current - voltage characteristics also show that the channel model is non - ohmic .
the full characteristics of conductance exhibit an absolute maximum at the same resonant channel potential as that identified in the currents .
it should be conclusively noted that all the results presented above are principally related to the considered single particle model , neglecting all the effects associated with the many - particle couplings , the particle size , and the potential induced by particle itself .
so , these results might be served as an argument for further studies .
+ + * acknowledgments*. v.l.n . gratefully acknowledges a generous hospitality from institute for bio - medical physics in hochiminh city , where this work has been done .
99 b. alberts , d. bray , k. hopkin , a. johnson , j. lewis , m. raff , k. roberts , p. walter , _ essential cell biology _ , @xmath163 edn .
( garland science , taylor & francis , new york - london , 2014 ) . b. hille , _ ionic channels of excitable membranes _ , @xmath164 edn .
( sinauer associates , sunderland , 2001 ) .
w. wickner and r. schekman , _ science _ * 310 * , 1452 ( 2005 ) . h. lerche , k. jurkat - rott , and f. lehmann - horn , _ am .
genet . _ * 106 * , 146 ( 2001 ) .
e. marban , _ nature _ * 415 * , 213 ( 2002 ) .
j. payandeh , t. scheuer , n. zheng , and w.a .
catterall , _ nature _ * 475 * , 353 ( 2011 ) .
a. meller , _
j. phys . : condens .
matter _ * 15 * , r581 ( 2003 ) .
mohammad , s. prakash , a. matouschek , l. movileanu , _ j. am . chem . soc . _ * 130 * , 4081 ( 2008 ) .
e. neher and b. sakmann , _ nature _ * 260 * , 799 ( 1976 ) .
a.m. berezhkovskii , m.a pustovoit , and s.m .
bezrukov , _ j. chem
. phys . _ * 116 * , 3943 ( 2003 ) .
a.m. berezhkovskii , m.a pustovoit , and s.m .
bezrukov , _ j. chem .
* 119 * , 9952 ( 2002 ) .
a.m. berezhkovskii , s.m .
bezrukov , _ biophys .
j. _ * 88 * , l17 ( 2005 ) .
a.m. berezhkovskii , s.m .
bezrukov , _ chem .
phys . _ * 319 * , 342 ( 2005 ) .
w. r. bauer and w. nadler , _ pnas _ * 103 * , 11446 ( 2006 ) a.b .
kolomeisky , _ phys .
lett . _ * 98 * , 048105 ( 2007 ) .
a.b . kolomeisky and k. uppulury , _ j. stat .
phys . _ * 142 * , 1268 ( 2011 ) .
s. m. bezrukov , a. m. berezhkovskii , and a. szabo , _ j. chem
. phys . _ * 127 * , 115101 ( 2007 ) .
s. pagliara , c. schwall , and u.f .
keyser , _ adv .
mater . _ * 25 * , 844 ( 2013 ) .
s. l. dettmer , s. pagliara , k. misiunas , and u. k. keyser , _ phys .
* 89 * , 062305 ( 2014 ) a.m. correa , f. bezanilla , and r. latorre , _ biophys .
j. _ * 61 * , 1332 ( 1992 ) .
r. erdem and e. aydiner , _ phys .
e _ * 79 * , 031919 ( 2009 ) .
o.s . anderson , r.e .
koeppell , _ physiol .
rev . _ * 72 * , s89 ( 1992 ) .
busath , c.d .
thulin , r.w .
hendershot , l.r .
phillips , p. maughan , c.d .
cole , n.c .
bingham , s. morrison , l.c .
baird , r.j .
hendershot , m. cotten , t.a .
cross , _ biophys .
j. _ * 75 * , 2830 ( 1998 ) .
p. graf , m.g .
kurnikova , r.d .
coalson , and a. nitzan , sl j. phys .
b * 108 * , 2006 ( 2004 ) .
b. corry and s .- h .
chung , _ cell .
life sci . _ * 63 * , 301 ( 2006 ) .
whereas the dielectric constant of the water in the channel interior is often assumed to be lower than the bulk value , generally , it is still remains an issue due to the lack of experimentally solid data @xcite .
ignoring all these unsolved complications , for definition , we choose for @xmath56 the same value of the dielectric constant of the bulk water , @xmath165 .
it is important to note that though the magnitude of the potential @xmath9 in eq.(3 ) does depend on @xmath56 , the @xmath28-behavior does not , and therefore , all qualitative conclusions of this work are unaffected by a choice of @xmath56-value .
d. frenkel and b. smit , _ understanding molecular dynamics simulation _ , @xmath166 ed .
( academic press , san diego - tokyo , 2002 ) .
w.f . van gunsteren and h.j.c .
berendsen , _ molec .
phys . _ * 45 * , 637 ( 1982 ) .
m. fathi and g. stoltz , _ arxiv _ : 1505.04905v1 c. maffeo , s. bhattacharya , j. yoo , d. wells , and a. aksimentiev , _ chem .
rev . _ * 112 * , 6250 ( 2012 ) .
wilson , t.h .
nguyen , and a. pohorille , _ j. chem .
phys . _ * 141 * , 22d519 ( 2014 ) .
w. treptow and m. tarek , _ biophys .
j. _ * 91 * , l26 ( 2006 ) ; * 91 * , l81 ( 2006 ) . | we consider an anisotropically two - dimensional diffusion of a charged molecule ( particle ) through a large biological channel under an external voltage .
the channel is modeled as a cylinder of three structure parameters : radius , length , and surface density of negative charges located at the channel interior - lining .
these charges induce inside the channel a potential that plays a key role in controlling the particle current through the channel .
it was shown that to facilitate the transmembrane particle movement the channel should be reasonably self - optimized so that its potential coincides with the resonant one , resulting in a large particle current across the channel . observed facilitation appears to be an intrinsic property of biological channels , regardless the external voltage or the particle concentration gradient .
this facilitation is very selective in the sense that a channel of definite structure parameters can facilitate the transmembrane movement of only particles of proper valence at corresponding temperatures .
calculations also show that the modeled channel is non - ohmic with the ion conductance which exhibits a resonance at the same channel potential as that identified in the current . |
the thirring model has been the subject of a very intense research in the last fifty years : it is one of the very few qft models for which non - perturbative informations can be obtained and it shares with more realistic @xmath0 models , like @xmath1 , many features , as it is apparent from a classical perturbative feynman graph analysis @xcite . in the _ massless case _ ,
a complete set of correlations has been obtained via an exact solution @xcite ; they verify the wightman recontruction axioms @xcite so that a qft corresponding to the massless thirring model can be constructed from them . in the _ massive case _ , the equivalence ( at a perturbative level ) with the sine - gordon model is known @xcite and some eigenstates by bethe - ansatz analysis have been found @xcite ; but a complete rigorous construction has never been performed . in this paper
we fill this gap by considering a set of grassmann integrals regularized via suitable cutoffs and with a contact current - current interaction and proving that , removing cutoffs and for any value of the mass , they converge to a set of schwinger functions verifying the osterwalder - schrader reconstruction axioms for euclidean qft @xcite ; this provides the first rigorous construction of the massive thirring model .
moreover , even if in the massless case other constructions were known , we find in any case interesting to reach a complete non - perturbative construction of the thirring model relying only on a functional integral approach , which could be the only possible one at higher dimensions or for more realistic models .
we stress that our results are _ non - perturbative _ , in the sense that the grassmann integrals are expressed in terms of series expansion whose _ convergence is proved _ ( that is we resum the naive perturbation theory and we prove the convergence of the resummed expansion ) .
our results are smooth in the mass and the massless limit can be investigated ; the natural question is wether or not the correlations of the massive thirring model , which we can compute from a functional integral approach , pass over smoothly into the known correlations of the exact solution in the massless limit .
contrary to what was find in the well known paper @xcite , based on zimmermann s version of bph subtraction scheme ( the analysis was purely perturbative ) , we get that the correlations in the massless limit have _ different properties _ with respect to the exact solution .
the ward identities ( wi ) have anomalies violating the _ anomaly non - renormalization _ property ; moreover , _ additional anomalies _ , apparently unnoticed before , are present even in the closed equation for the interacting propagator , obtained by combining a schwinger - dyson equation with the ward identities .
it is worth to recall the johnson solution @xcite of the massless thirring model , describing @xmath2 massless dirac fermions with coupling @xmath3 , in order to compare its properties with the results from the functional integral approach .
the solution is essentially based on a _ self - consistency _ argument : a number of reasonable requirements on the correlations is assumed , from which their explicit expression can be determined .
calling @xmath4 , @xmath5 and @xmath6 the truncated vacuum expectations in the minkowski space of the @xmath7-product of @xmath8 , @xmath9 and @xmath10 respectively , where @xmath11 is the _ wave function renormalization _ , the first assumption is the validity of wi of the form @xmath12g(\xx,\yy)\label{w1 } \\
\dpr_\zz^\m g^{\m,5}(\zz;\xx,\yy)= -i\bar a[\d(\zz-\xx)-\d(\zz-\yy)]\g^5g(\xx,\yy)\label{w2}\end{aligned}\ ] ] where @xmath13 are parameters to be determined and @xmath14 are operators formally defined , via a point splitting procedure , from @xmath15 and @xmath16 ; @xmath17 are the minkowski gamma matrices . assuming also the validity of a _ schwinger - dyson _ equation and using the wi for @xmath18 , a _ closed equation _ for the two point schwinger function is obtained , which reads , if @xmath19 denotes the fourier transform of @xmath6 : @xmath20\;.\ ] ] this equation can be solved ( at a very formal level , as one has to take a vanishing wave function renormalization ) and @xmath21
is found to be equal to @xmath22 , where @xmath23 is an arbitrary constant with the dimension of a length and @xmath24 is a critical index related to the coefficients @xmath13 by @xmath25 finally , by a self - consistency argument involving also the four point schwinger function , the explicit values for the anomalies were found : @xmath26 the above equation is particularly significant , as it says the anomalies do not receive contributions from higher orders ; this property is called _ anomaly non - renormalization _ and it holds , as a statement valid at all order in perturbation theory , in realistic models like @xmath1 ( it is the content of the well known _ adler - bardeen theorem _ @xcite ) .
the order by order analysis of @xcite can be adapted to the thirring model @xcite and ( [ xx11 ] ) is indeed obtained ; hence the validity of ( [ xx11 ] ) has been considered _ non - perturbative _ verification of the _ adler - bardeen theorem _ applied to the thirring model .
however , it should be noticed that the regularization and the assumptions in the exact solution or in the perturbative ( or functional integral ) approach are different , hence there is no guarantee that the same results should be found in the two approaches , see for instance @xcite .
our starting point is the generating functional of the truncated schwinger functions , @xmath27 , defined so that @xmath28 is the following _ grassmann integral _ : @xmath29 } \label{fi}\ ] ] where @xmath30 is a renormalization constant , @xmath31 are external fields , @xmath32 are grassmann variables , @xmath33 and @xmath34 is the grassmann integration with propagator @xmath35 where the _ smooth cutoff function _ @xmath36 selects the momenta @xmath37 , with @xmath38 , @xmath39 . finally @xmath40 and @xmath41 are the bare wave function renormalization and mass and @xmath42 , @xmath43 are the euclidean gamma matrices .
the schwinger functions are defined as @xmath44 in particular @xmath45 and @xmath46 .
the presence of the ultraviolet cutoff @xmath47 and the infrared cutoff @xmath48 makes the functional integral ( [ fi ] ) well defined ; to carry out the renormalization program at non - perturbative level we have to prove that there exist @xmath47-depending bare parameters such that , in the limit @xmath49 , the schwinger functions verify the osterwalder - schrader axioms , ( os ) , @xcite .
our basic result @xcite is the following theorem .
given @xmath50 small enough and @xmath51 , there exist bare parameters @xmath52 with @xmath53 and @xmath54 , @xmath55 , such that @xmath56 exist at coinciding points and verify the osterwalder - schrader axioms . in particular , in the massless case @xmath57 , @xmath58 with @xmath59 indipendent from @xmath60 .
the above theorem says that , by choosing properly the bare wave function and mass renormalization , with a singular behavior as the cutoffs are removed , one gets a set of finite schwinger functions , for which a qft for the massive thirring model is obtained via the reconstruction theorem in @xcite .
the proof is based on the new methods introduced in @xcite , which allow us to overcome the well known technical problem posed by the combination of a non - perturbative setting based on multiscale analysis @xcite with the necessity of exploiting cancellations due to the local symmetries .
such cancellations are established by suitable wi valid at each scale and , contrary to the wi formally valid when all the cutoffs are removed , they have corrections due to the cutoffs . the crucial role of wi in the construction of the theory is a feature that the thirring model shares with realistic models like @xmath61 or the electroweak theory in @xmath0 , requiring wi even to prove the perturbative renormalizability . note that this feature is absent in the models previously rigorously constructed by functional integral methods , like the massive _ yukawa _ model @xcite or the massive _ gross - neveu _ model @xcite . a delicate point in the proof in @xcite of the above theorem
is the verification of the positivity axiom , as the cutoff we have chosen destroy the positivity definiteness and it would be quite difficult to prove it directly after the cutoffs are removed .
we overcome this problem by considering a functional integral with a _ lattice regularization _ containing a wilson term ( to avoid a spurious singularity ) ; in this case the positivity property is automatically satisfied and one has to choose the parameters @xmath62 , @xmath63 , @xmath64 , if @xmath65 is the lattice step , in order to control the limit @xmath66 .
we have then showed that it is possible to choose them so that @xmath67 , hence positivity follows .
note finally that it has been claimed that the functional integral ( [ fi ] ) can be exactly computed for massless fermions and no cutoffs , see for instance @xcite , by introducing a boson field and using that that fermionic determinants are _ quadratic _ if the bose field is regular . in this way one
recovers the same results of the exact solution , but mathematically this procedure is not justified ( the functional integral is over all the possible fields ) and , as we will see below , it leads to wrong conclusions .
we shall only consider the wi in the massless case , @xmath68 ; the presence of the cutoff function @xmath69 breaks the continuous symmetries , @xmath70 and @xmath71 , so that there are corrective terms with respect to the formal wi ( obtained formally neglecting all cutoffs ) : @xmath72 the function @xmath73 is given by @xmath74 ; @xmath75 s single out the cluster of fields w.r.t . which the truncation of the expectation is taken .
the above expression can be perturbative checked at lowest orders using the ( trivial ) identities @xmath76 % \notag\\ + c^{\m}_{\k , k}(\kk,\kk-\pp)\widehat g_{\k , k}(\kk-\pp ) \g^\m \g\widehat g_{\k , k}(\kk)\;,\label{iv}\end{aligned}\ ] ] for @xmath77 .
the last addenda in the above wi involves the average of an higly non - local and complex operator , but removing cutoffs , as proven in @xcite , they can be written in a remarkable simple form . in the same hypothesis of theorem 1 ,
@xmath78 where @xmath79 and @xmath80 are suitable functions of @xmath50 such that @xmath81 and @xmath82 strictly negative ; moreover , for fixed non - zero @xmath83 , @xmath84 hence the wi we get in the functional integral approach have the same form as those found in the exact approach , see ( [ w1 ] ) and ( [ w2 ] ) , but the anomaly coefficients , instead of by ( 5 ) , are given by @xmath85 .
the anomaly coefficients are not linear in the bare coupling ( _ the anomaly non - renormalization is violated _ ) , contrary to ( 5 ) ; hence the theory found in the massless limit starting from the functional integral ( 6 ) has different properties with respect to the one constructed by the exact solution .
the presence of such anomaly renormalization can be checked in standard perturbation theory , by calculating the two graphs of fig .
[ fig2 ] , but the proof of the non - perturbative bounds of theorem 2 requires a careful mathematical analysis , see @xcite .
first and second order conribution to the anomaly ; the small circle represents @xmath86
the two point function @xmath87 verifies the sd equation @xmath88 inserting the explicit expression of @xmath89 obtained form the wi , we obtain @xmath90\;.\notag\end{aligned}\ ] ] _ if _ the last term in ( [ vvv ] ) were vanishing in the limit @xmath91 , one would get a closed equation for @xmath92 , which is identical to the closed equation obtained in @xcite , that is ( 3 ) .
_ this is not what happens _ ; as already noticed , @xmath93 and @xmath94 are vanishing in the limit @xmath91 at @xmath83 fixed , but _ not _ if @xmath95 is integrated .
intuitively this can be understood by noting that the integral involves momenta close to the u.v .
cutoff scale @xmath47 , where @xmath96 is not small at all . in other words : _ even if the wi and the sd equation are true , in the limit @xmath91 , the closed equation obtained by combining the two identities is not verified ; this is a new anomaly _ which is hard to see in a purely perturbative approach and in fact it was never noticed before .
one could guess that the fact that the last term in ( [ vvv ] ) is not vanishing in the limit of removed cutoffs should imply that there is no closed equation for @xmath19 .
instead , we proved in @xcite another crucial identity . in the same hypothesis of previous theorems ,
the integral in ( [ vvv ] ) is equal to @xmath97 with o@xmath98 , non - zero @xmath99 and @xmath100 ; furthermore , for fixed non - zero @xmath101 , @xmath102 . by inserting ( [ hhhg ] ) in ( [ vvv ] )
we get a closed equation which is _ different _ from the one assumed in @xcite ; in particular , the relation ( [ xxx ] ) is replaced by @xmath103
we stress that one could construct a qft corresponding to the thirring model also starting from a _ non - local _
interaction @xmath104 with @xmath105 .
if the cut - offs are removed in the order @xmath106 , @xmath107 , the results are the same as depicted it the previous sections . on the contrary , if the cut - offs are removed in the opposite order , @xmath107 , @xmath106 , the results of the exact solution are recovered .
this means that the adler - bardeen theorem is not in contrast with our result : if one assumes , as in@xcite , that the interaction is mediated by a boson field and removes the fermionic u.v .
cutoff before the bosonic one , then the anomaly non - renormalization holds ; in the opposite case new features appear .
finally , the combination of wi and sd equations is a rather general technique ; it is used , for instance , in qed in @xcite and in condensed matter physics in @xcite .
we have seen that such a method can really be implemented in a full non - perturbative approach , but taking care of unexpected anomalous features .
it would be very interesting to see if such features occur in 4-dimensional models .
_ acknowledgments _ we are indebted with k. gawedzki and g. gallavotti for enlightening discussions .
p.f . gratefully acknowledges the support of the erwin schrdinger institute for mathematical physics ( vienna ) . | we present the first rigorous construction of the qft thirring model , for any value of the mass , in a functional integral approach , by proving that a set of grassmann integrals converges , as the cutoffs are removed , to a set of schwinger functions verifying the osterwalder - schrader axioms .
the massless limit is investigated and it is shown that the schwinger functions have different properties with respect to the ones of the well known exact solution : the ward identities have anomalies violating the anomaly non - renormalization property and additional anomalies , apparently unnoticed before , are present in the closed equation for the interacting propagator , obtained by combining a schwinger - dyson equation with ward identities . |
black holes through quantum outcome indicate that they radiate due to the hawking effect @xcite . in the absence of established theories of quantum gravity ,
black holes have become a main playground to divulge quantum gravity effects through their thermodynamics .
black holes have been used as theorists laboratories in many other relevant fields .
thermodynamic properties of black holes have been studied for many years , but established statistical explanations of black hole thermodynamics are still lacking .
it shows that black holes also have the standard thermodynamic quantities , such as temperature , entropy and heat capacity and so on , and even possess abundant phase structures like the hawking - page phase transition @xcite and similar critical phenomena in ordinary thermodynamic systems .
recent years witness the renewed interest towards the study of black hole solutions specially in modified theories of gravity @xcite , as besides theoretical results , cosmological evidence , e.g. dark matter and dark energy , suggests possibility of changing the einstein gravity . on the other hand ,
the einstein gravity can not be quantized ( non - renormalizable ) , so it is believed that it is a low energy effective theory and could be modified with higher derivative terms at high energy @xcite .
modifications to the einstein gravity theory , for instance the lovelock theory @xcite , the @xmath1 gravity theory @xcite , etc . have been studied extensively .
those models in higher spacetime dimensions have very different features .
for example , nature of stability in higher dimensions is quite different . extending spacetime dimensionality in gravity theories
has been one possible way to combine other interactions with gravity or often seems to be even required in many theories , e.g. kaluza - klein theory , a string theory , brane world scenarios , etc . @xcite . in these context , apart from the standard einstein - hilbert action , there also exist interesting theories of gravity in dimensions greater than four involving higher powers of the curvatures such that the field equations for the metric are at most in second - order . among the higher curvature gravity theories ,
the most extensively studied theory is the so - called lovelock gravity @xcite , which naturally emerges when we wish to generalize the einstein theory in higher dimensions by keeping all characteristics of usual general relativity excepting the linear dependence of the riemann tensor .
the lovelock gravity is one of the most general second order theories in higher dimensions , which is free from ghosts
. the lovelock theory may play an important role in a string theory where the low energy effective field theory of gravity contains higher curvature terms . in this sense
the lovelock gravity @xcite is a natural extension to the einstein gravity .
it is constructed by sum of all the euler densities of a @xmath2-dimensional manifold .
the lagrangian is given by @xmath3 where @xmath4 a spacetime dimension @xmath5 can be written as @xmath6 for even dimensions and @xmath7 for odd dimensions .
the @xmath8th order higher derivative terms @xmath9 become a surface term when @xmath10 .
non - trivial extra terms contribute to equations of motion in higher dimensions but not in dimensions less than @xmath2 .
moreover , higher derivative terms can cancel a ghost term .
for instance , the reference @xcite shows that the second order lovelock ( gauss - bonnet ) terms cancel a ghost term .
boulware and deser @xcite first found a static , spherically symmetric black hole solution with the gauss - bonnet corrections . using the gauss - bonnet gravity static , spherically symmetric solutions are obtained later @xcite with thermodynamic properties @xcite .
static , spherically symmetric black hole solutions in the lovelock gravity with general energy momentum tensors in any arbitrary dimension can be found in @xcite and @xcite and its thermodynamics in @xcite .
further extensive studies on the gauss - bonnet black holes with a focus on thermodynamic properties have been found in @xcite .
the special third order lovelock gravity also received a significant attention , e.g. for a black hole solution and its thermodynamics in this theory with born - infield source @xcite and for a causality violation @xcite .
also , topological properties of the general lovelock black holes in the context of thermodynamics have been investigated @xcite . in this paper
, we begin with finding static , spherically symmetric black hole solutions for a string cloud background for a specific case , i.e. , @xmath11 and examine thermodynamic properties in the third order lovelock gravity .
the solution in there can be utilized to calculate mass , temperature , entropy and heat capacity of black holes and explicitly study effects of a string cloud background .
it turns out that the horizon and thermodynamic properties of lovelock black hole in conjunction with a string parameter could have some interesting features .
it may be pointed out that gravity coupled to clouds of strings may be very useful and important as the universe can be considered as a collection of extended objects , like , one dimensional strings .
the study of black holes in a cloud of strings was initiated by letelier modifying the schwarzschild solutions for a cloud of strings as a source @xcite , which was recently extended to the gauss - bonnet gravity @xcite and also to the lovelock gravity @xcite .
we show that a string cloud background makes a profound influence on horizon structures and thermodynamic quantities but entropy is not changed .
the paper is organized as follows : in sec .
[ lovelock action and equations of motion ] we begin examining the third order lovelock action , which is a modification of the einstein - hilbert action , and also derive energy momentum tensors of a cloud of strings .
the thermodynamics of a static , spherically symmetric black hole solution in this theory is explored in sec .
[ thermodynamics of black holes ] . before that we find an exact static , spherically symmetric black hole solution in sec .
[ spherically symmetric solution in lovelock gravity ] .
the paper ends in sec .
[ conclusion ] , which gives concluding remarks .
we have used units that fix @xmath12 and the metric signature , @xmath13 .
the lovelock theory is the most general theory of gravity that gives second order field equations in arbitrary dimensions .
the recent interest in the lovelock theory arose because its action appears as a low energy limit of a heterotic superstring theory .
the simplest third order lovelock action reads @xcite : @xmath14 + \mathcal{i}_{s } \label{ig}\ ] ] where @xmath15 is a matter action due to cloud of strings .
the einstein term @xmath16 is @xmath17 , the second order lovelock ( gauss - bonnet ) term @xmath18 is @xmath19 and the third order lovelock lagrangian is @xmath20 here @xmath21 @xmath22 and @xmath23 are the ricci scalar , the riemann and the ricci tensors , respectively .
the coupling constants @xmath24 and @xmath25 in eq .
( [ ig ] ) have dimensions , @xmath26^{4-d}$ ] and @xmath26^{6-d}$ ] , respectively and will help us to explicitly bring out changes in the general relativity equations . in the limits @xmath27 ,
one recovers the einstein - hilbert action .
the variation of the action with respect to the metric @xmath28 yields modified field equations for the third order lovelock gravity , @xmath29 where @xmath30 is the einstein tensor , while @xmath31 and @xmath32 are given explicitly in @xcite , respectively : @xmath33 and @xmath34 and @xmath35 is the energy momentum tensor of a matter field which we consider here as clouds of strings .
note that for the third order lovelock gravity , the non - trivial third terms require a spacetime dimension @xmath5 to satisfy @xmath36 .
next we turn attention to calculate the energy - momentum tensor of a cloud of strings ( see @xcite , for further details ) .
the nambu - goto action of a string evolving in spacetime is given by @xmath37 with the lagrangian for a cloud of strings @xcite : @xmath38^{1/2}.\ ] ] the string worldsheet is associated with a bivector of the form @xmath39 where @xmath40 is the two - dimensional levi - civita tensor and @xmath41 . here @xmath42 is a constant for each string and @xmath43 is the determinant of an induced metric on the string world sheet @xmath44 given by @xmath45 @xmath46 with @xmath47 and @xmath48 which are a timelike and a spacelike parameter , respectively , is a parametrization of the world sheet @xmath44 , @xcite .
further , since @xmath49 , then the energy - momentum tensor for one string reads @xmath50 hence , a cloud of strings has the energy - momentum tensor @xmath51 where @xmath52 is a proper density of a cloud of strings .
the quantity @xmath53 is a gauge - invariant density .
as will be seen in next discussions the presence of a string cloud plays a main role in the horizon structure of the theory together with other parameters . for instance , a string cloud parameter @xmath54 can change the number of horizons and singularities and make singularity covered by a horizon with fixed other parameters .
given parameters @xmath55 a positive mass condition imposes either mass bound or range of the horizon radius .
rich analysis along this line for absence of energy momentum tensor has been made in @xcite .
vanishing string cloud is a transition point for singularities to be created @xmath56 and one for the number of horizons for @xmath57 . in general
the energy momentum tensor @xmath58 is expressed for static spherically symmetric spacetime , @xmath59 where @xmath60 is a constant , @xcite .
the dominant energy condition allows only @xmath61 and @xmath62 and the causality condition further constrains @xmath60 to @xmath63 @xcite .
thermodynamic quantities are determined by mass expression , which is given for the lovelock theory in terms of a horizon radius @xmath64 by @xmath65 where @xmath66 $ ] and the curvature constant @xmath67 , @xcite .
the mass for a string cloud is @xmath68 as noticed the a string cloud contributes to the mass with the highest possible power of @xmath64 , i.e. the energy condition upper bound @xmath69 but violates causality . for even dimensions the presence of a string cloud effectively changes the highest coupling @xmath70 , @xmath71 while for odd dimensions it is a unique source term thermodynamically , @xcite . in this paper
we take a string cloud only as a classical background .
the reference @xcite claims that causality violation in the third order lovelock theory can be fixed by adding an infinite tower of massive higher spin particles .
unless we look for dynamics instead of static background of strings , it is not certain how a string cloud contributes to the causality problem .
however , it is worthwhile to study whether causality violation in classical sense can be weaken or removed when the external massive particles are combined with a string cloud into a background .
here we wish to obtain static , spherically symmetric black hole solutions to eq .
( [ ee ] ) for the energy momentum tensors , eq .
( [ em ] ) , and investigate their horizons and thermodynamic properties .
hence , we assume the metric of the form : @xmath72 where @xmath73 is a metric of a @xmath74-dimensional constant curvature space @xmath75 or -1 , representing spherical , flat and hyperbolic spaces , respectively .
but , in this paper we shall confine ourselves to @xmath76 . to find the metric function @xmath77
, we should solve eq .
( [ ee ] ) . using this metric ansatz ,
an @xmath78 component of the field equations of motion reduces to : @xmath79 f^{\prime}\left ( r\right ) + \nonumber\\ & & \left ( n-1\right ) r^{4}\left ( f\left ( r\right ) -1\right ) -\left ( n-3\right ) \tilde{\alpha}_{2}r^{2}\left ( f\left ( r\right ) -1\right ) ^{2}+\nonumber\\ & & \left ( n-5\right ) \tilde{\alpha}_{3}\left ( f\left ( r\right ) -1\right)^{3 } = \frac{2r^6 } { n}t^r_r , \label{egbl rr}\end{aligned}\ ] ] where a prime denotes a derivative with respect to @xmath80 @xmath81 @xmath82 and @xmath83 . in general , the eq .
( [ master ] ) has one real and two complex solutions . but
it can also have three real solutions as well under appropriate conditions .
we are seeking static , spherically symmetric real solutions , which restrict a density @xmath52 and a bivector @xmath84 as a function of @xmath85 only .
further , the only possible non - zero component of a bivector @xmath44 is @xmath86 .
thus @xmath87 and we obtain @xmath88 , which implies : @xmath89 for some real constant @xmath54 .
clearly the third order lovelock gravity is non - trivial only for spacetime dimensions @xmath90 . henceforth , to extract information from our analysis
, we shall confine ourselves to @xmath91 in which the eq .
( [ egbl rr ] ) can be easily integrated and the solution reads @xmath92 where @xmath93 and @xmath94,\ ] ] with @xmath95 here we shall impose the condition @xmath96 , which simplifies @xmath77 significantly with all coefficients intact and hence it is worthwhile to consider this case .
this condition reduces @xmath77 to the following form @xmath97 ^ 2}\right\}^{1/3},\label{f*}\ ] ] where @xmath98 is given with the notation @xmath99 , @xmath100 note that the square root here should be defined including a sign of @xmath101 , i.e. , @xmath102 ^ 2}=-\alpha g(r)$ ] .
one can confirm it by noticing that the solution has been obtained from solving a cubic equation , @xcite .
( [ egbl rr ] ) can be rewritten as @xmath103=r^{-n+6}\frac{\partial}{\partial r}\int dr\frac{2r^n}{n}t^r_r(r),\ ] ] where @xmath104/r^2 $ ] . for @xmath105 and @xmath106
this reduces to , with an integration constant @xmath107 , @xmath108 thus @xmath77 reads : @xmath109.\label{f-}\ ] ] we consider only positive mass , i.e. , @xmath110 . @xmath77
asymptotically behaves as @xmath111 as shown in fig .
[ fig : f ] with various parameter values . in fact , for @xmath112 the einstein gravity can be thought of as a particular case of the lovelock gravity since the einstein - hilbert term is one of several terms that constitute the lovelock action .
hence , for @xmath112 and @xmath113 , the higher dimensional schwarzschild solution in a string cloud model reads @xcite : @xmath114 which goes to eq .
( [ fasym ] ) for @xmath91 .
( [ f- ] ) , in the limit @xmath115 , again leads to @xmath116 which can be also obtained from eq .
( [ hdschwarzchild ] ) when @xmath91 . with @xmath117 , @xmath77 in eq .
( [ f- ] ) can be identified with one in the reference @xcite with @xmath118 , i.e. the vanishing born - infield field .
the solution can be exactly verified through the reference @xcite .
since @xmath107 represents mass it should be positive , @xmath119 .
due to a fractional power on @xmath98 , @xmath120 leads to a curvature singularity at @xmath121 , where @xmath122 as in the gauss - bonnet case @xcite . using the expression of the ricci scalar @xmath123 $ ] from @xcite
, one sees @xmath124 .
a horizon radius @xmath64 is defined by @xmath125 . to see how @xmath64 and @xmath126 can be related ,
it is convenient to define @xmath127 , where @xmath128 .
note that @xmath129 and @xmath130 .
let us first consider @xmath56 and @xmath131 case .
it is worthwhile to notice that @xmath132 and @xmath133 .
the latter tells us that only one horizon can exist .
a necessary and sufficient condition for existence of one and only one horizon can be found to be a negative @xmath134-intercept , i.e. , @xmath135 .
@xmath136 has a minimum at @xmath137 . in case that @xmath136 has a negative minimum , hence giving two singularities , @xmath138
, a horizon can stay in either @xmath139 or @xmath140 due to @xmath132 , but the case @xmath140 leads to @xmath141 , which is not true .
this means that a horizon can exist only in @xmath139 , not being covered by a horizon . for @xmath56 and
@xmath142 case @xmath143 has one minimum , which can lead to two horizons if @xmath144 . for @xmath61 @xmath136
is a monotonously increasing function with a positive a @xmath134-intercept and hence there is no singularity . when @xmath117 there exists one horizon if @xmath135 and is no singularity .
in short for @xmath56 , @xmath145 is a necessary and sufficient condition for existence of one and only one horizon for all @xmath54 .
non - negative @xmath54 allows only one horizon while negative @xmath54 two horizons at most , whose necessary condition is @xmath144 .
positive @xmath54 gives a naked singularity while @xmath142 does not give a singularity . without a string cloud @xmath146
there is no singularity .
next , consider @xmath57 case .
there is one and only one singularity in this case .
it is useful to notice that a horizon can not exists between @xmath147 and @xmath126 .
this can be checked by @xmath148 for @xmath149 and @xmath150 for @xmath151 .
we are particularly interested in whether singularities are covered by horizons . because horizons can not stay between @xmath152 and @xmath126 as mentioned above , the inequality between @xmath152 and @xmath64 is equivalent to one between @xmath126 and @xmath64 , i.e. the criteria for whether singularities are covered by horizons or not .
the solutions for horizons are intercepts between the curve @xmath153 and the straight line @xmath154 . for @xmath131
, @xmath155 can have at most two zeros , leading up to maximum three horizons . if a singularity is covered by a horizon , the horizon exists in @xmath156 and is the only one . for @xmath142 @xmath155
has one and only one zero , leading to maximum two zeros in @xmath5 .
let us consider possibility to have two intercepts for @xmath156 .
the necessary conditions are @xmath142 and @xmath157 .
the slopes of @xmath158 and @xmath159 are equal to @xmath160 when they have one intercept .
thus @xmath161 , i.e. , @xmath162 . however , the condition @xmath163 with the first one @xmath164 conflicts with the positive mass condition @xmath110 , i.e. , @xmath165 .
therefore , there can exist only one horizon in the region @xmath156 and the condition for the existence of one horizon is @xmath166 .
it is important to notice that this condition is just @xmath167 .
once one horizon exists in @xmath156 , the other exists in @xmath168 and a singularity exists in @xmath169 , i.e. a singularity stays closer to the greater horizon @xmath170 . for @xmath117
there is a similar property to case @xmath131 , i.e. there are at most two horizons .
only when a singularity is covered by a horizon , it is the only horizon .
the sign of @xmath171 , equal to @xmath172 , tells us both whether @xmath64 is ahead or behind from @xmath152 and whether @xmath126 is ahead or behind from @xmath152 .
therefore , along with the fact that there is no horizon between @xmath152 and @xmath126 , only configuration , @xmath173 or @xmath174 is allowed , in other words , @xmath126 stays closer to @xmath64 than @xmath152 . in short for @xmath57 , there is only one singularity . for @xmath175
when a singularity is non - naked there should be only one horizon .
for @xmath142 , when a singularity is non - naked , the singularity must be between two horizons . [ cols="^,^,^ " , ]
the lovelock theory is a natural extension of the einstein theory of general relativity to higher dimensions and it is of a great arena for theoretical physics research .
the lovelock theory describes string - inspired corrections of the einstein - hilbert action and hence admits the general relativity as a particular case . in this paper , we have obtained exact static , spherically symmetric black hole solutions to the third order lovelock gravity in a string cloud background in seven dimensions with help of carefully choosing coefficients of the curvature correction terms , thereby generalizing the static , spherically symmetric black hole solutions for these theories .
these solutions possess rich properties of black holes and in the limits go over to black holes in einstein s gravity .
the string cloud parameter @xmath54 changes the number of horizons and singularities . for @xmath56 ,
@xmath54 can change the maximum number of horizons and the positivity of @xmath54 creates a singularity . in this case a singularity , if any , is naked .
for @xmath57 there is an interesting property .
a horizon can not exists between @xmath176 and a singular point @xmath126 and hence we need know only whether a horizon is ahead or behind from @xmath152 instead of @xmath126 in order to see whether a singularity is naked .
there is only one singularity . for @xmath175
when a singularity is non - naked there should be only one horizon .
for @xmath142 , when a singularity is non - naked , the singularity must be between two horizons .
we proceeded to find exact expressions for the thermodynamic quantities like the black hole mass , the hawking temperature , entropy and heat capacity and in turn also analysed the thermodynamic stability of black holes .
in addition we explicitly brought out the effect of a string cloud background on black hole solutions and their thermodynamics .
we found that a positive mass condition leads to positive temperature for @xmath177 , which is a condition for a non - naked singularity for @xmath57 .
the entropy does not depend on a string cloud .
this result can be extended for any spherical and static source in the lovelock theory @xcite .
we also see the entropy does not obey the horizon area formula . for heat capacity ,
@xmath56 a critical(singular ) point @xmath178 is greater than a transition ( zero ) point @xmath179 and for @xmath57 when a positive mass and a non - naked singularity conditions applied , @xmath180 and there is no @xmath178 . in this case
the heat capacity is negative , telling us that the black hole is thermodynamically unstable like the schwarzschild black hole .
d. lovelock , j. math .
* 12 * , 498 ( 1971 ) .
h. a. buchdahl , mon . not .
soc . * 150 * , 1 ( 1970 ) .
d. j. gross and e. witten , nucl .
b * 277 * , 1 ( 1986 ) .
r. emparan and h. s. reall , living rev .
rel . * 11 * , 6 ( 2008 ) .
b. zwiebach , phys .
b * 156 * , 315 ( 1985 ) .
boulware , s. deser , phys .
lett . * 55 * , ( 1985 ) 2656 ; j.t .
wheeler , nucl .
b * 268 * , ( 1986 ) 737 ; r.c .
myers , j.z .
simon , phys .
d * 38 * , ( 1988 ) 2434 .
j. t. wheeler , nucl . phys .
b * 273 * , 732 ( 1986 ) .
j. t. wheeler , nucl .
b * 268 * , 737 ( 1986 ) .
r. c. myers and j. z. simon , phys .
d * 38 * , 2434 ( 1988 ) .
s. h. mazharimousavi , o. gurtug and m. halilsoy , int .
j. mod .
d * 18 * , 2061 ( 2009 ) .
m. h. dehghani , n. alinejadi and s. h. hendi , phys .
d * 77 * , ( 2008 ) 104025 ; s. h. hendi and m. h. dehghani , phys .
b * 666 * , ( 2008 ) , 116 ; m. m. anber and d. kastor , jhep * 05 * , 061 ( 2008 ) ;
r. g. cai , l. m. cao , y. p. hu and s. p. kim , phys .
d * 78 * , ( 2008 ) , 124012 ; s. h. mazharimousavi and m. halilsoy , phys .
b * 665 * ( 2008 ) , 125 ; c. garraffo , g. giribet , e. gravanis and s. willison , j. math . phys .
* 49 * ( 2008 ) , 042502 ; g. kofinas and r. olea , jhep * 11 * , ( 2007 ) , 069 ; q. exirifard and m. m. sheikh - jabbari , phys .
b * 661 * ( 2008 ) , 158 .
r. g. cai , phys .
d * 65 * , 084014 ( 2002 ) [ hep - th/0109133 ] .
r. g. cai and q. guo , phys .
d * 69 * , 104025 ( 2004 ) [ hep - th/0311020 ] .
p. li , r. h. yue and d. c. zou , commun .
phys . *
56 * , 845 ( 2011 ) .
x. o. camanho and j. d. edelstein , class .
* 30 * , 035009 ( 2013 ) [ arxiv:1103.3669 [ hep - th ] ] .
x. o. camanho , j. d. edelstein , j. maldacena and a. zhiboedov , arxiv:1407.5597 [ hep - th ] .
r. g. cai and k. s. soh , phys .
d * 59 * , 044013 ( 1999 ) b. zwiebach , phys .
b * 156 * ( 1985 ) 315 ; b. zumino , phys .
* 137 * ( 1986 ) 109 ; n. deruelle and l. farina - busto , phys .
d * 41 * ( 1990 ) 3696 ; g. a. menamarugan , ibid .
* 46 * ( 1992 ) 4320 ; * 46 * ( 1992 ) 4340 ; r. g. cai and n. ohta , phys .
d * 74 * ( 2006 ) 064001 .
m. salgado , class .
grav . * 20 * , 4551 ( 2003 ) [ gr - qc/0304010 ] .
tae - hun lee , sushant g. ghosh , sunil maharaj and dharmanand baboolal `` thermodynamics of the lovelock black holes '' in preparation . | we present an exact static , spherically symmetric black hole solution to the third order lovelock gravity with a string cloud background in seven dimensions for the special case when the second and third order lovelock coefficients are related via @xmath0 .
further , we examine thermodynamic properties of this black hole to obtain exact expressions for mass , temperature , entropy and also perform the thermodynamic stability analysis .
we see that a string cloud background makes a profound influence on horizon structure , thermodynamic properties and the stability of black holes .
interestingly the entropy of the black hole is unaffected due to a string cloud background .
however , the critical solution for thermodynamic stability is being affected by a string cloud background . |
the low - energy properties of multilayer graphene@xcite depend crucially on the stacking order of the constituent graphene layers@xcite . in the case of trilayer graphene , there are two stable stacking orders : ( i ) aba or bernard stacking and ( ii ) abc or chiral stacking . recent advances in sample fabrication methods have resulted in high - quality trilayer samples which can be used to probe many of the theoretical predictions@xcite .
abc stacked trilayer graphene appears to be particularly exciting because it is expected to host a wealth of interesting phenomena , such as chiral quasiparticles with berry phase @xmath3@xcite , a lifshitz transition of electronic bands due to trigonal warping@xcite , band - gap opening in an external electric field@xcite , and broken symmetry phases at low electron densities@xcite , to name a few .
although there are a number of theoretical@xcite and experimental@xcite studies on the electronic properties of aba and abc stacked trilayer graphene , the spin - orbit coupling ( soc ) in these systems has received much less attention .
aba trilayer graphene was considered in ref . within a framework of an effective low - energy theory , whereas the case abc stacking was only briefly mentioned in ref . .
the understanding of spin - orbit interaction would be important to study other interesting and experimentally relevant phenomena such as spin relaxation@xcite , weak - localization@xcite or even spin - hall effect@xcite in trilayer graphene .
the recent report of ref .
on the fabrication of high mobility double gated abc trilayer graphene may open very promising new avenues for trilayer graphene spintronics as well , similarly to the monolayer case where highly efficient spin transport has recently been reported@xcite , but with the additional advantage that external gates can open a band gap in abc trilayer graphene .
in this paper we aim to make the first steps towards a detailed understanding of the spin - orbit coupling ( soc ) in chirally stacked trilayer graphene .
we start by investigating the case when the system has inversion symmetry , i.e. in the absence of external electric fields , adatoms or a substrate .
this is the case of _ intrinsic _ soc .
the intrinsic soc opens a band - gap at the band - degeneracy points without introducing spin polarization .
ab initio _
calculations on monolayer@xcite and bilayer@xcite graphene provided strong evidence that the key to the understanding the soc in _ flat _
graphene systems is to take into account the ( nominally unoccupied ) @xmath0 orbitals in the description of electronic bands .
here we take the same view and by generalizing the work of ref .
derive the intrinsic soc hamiltonian for abc trilayer graphene .
it turns out that the most important @xmath0 orbitals to take into account are the @xmath4 , @xmath5 and @xmath1 orbitals . while the former two have been considered in ref . in the context of monolayer graphene ,
the latter one is important to understand the soc in ab and abc graphene .
we obtain explicit expressions for the soc constants in terms of slater - koster@xcite hopping parameters .
we also rederive the intrinsic so hamiltonian for bilayer graphene@xcite . through the comparison of our trilayer and bilayer analytical results with the recent _ ab initio _ calculations of ref .
we are able to make predictions for the actual values of the soc parameters in abc trilayer graphene .
the theory involves electronic bands which are far from the fermi energy @xmath6 but are coupled to the physically important low - energy bands close to @xmath6 and hence complicate the description of the electronic properties .
therefore , we derive an effective low - energy hamiltonian and calculate its spectrum .
this helps us to understand how soc lifts certain degeneracies of the electronic bands . generally speaking , due to the low atomic number of carbon , the intrinsic soc in single and multilayer graphene is weak ( according to our prediction , the soc parameters are of the order of @xmath7 in abc trilayer , the same as in monolayer@xcite and bilayer@xcite graphene ) .
recently however , there have been exciting theoretical proposals to enhance the strength of soc in monolayer graphene and hence , e.g. , make the quantum spin hall state@xcite observable .
these proposals suggest deposition of indium or thallium atoms@xcite or to bring graphene into proximity with topological insulators@xcite .
indeed , very recently the combined experimental and theoretical work of ref . has provided evidence of a large ( @xmath8 ) spin - orbit gap in monolayer graphene on nickel substrate with intercalated gold atoms .
motivated by these studies we also discuss what might be a minimal model to describe the case where the the soc is strongly enhanced in only one of the layers of bilayer and abc trilayer graphene . our work is organized as follows . in sect .
[ sec : tb - model ] we present the tight - binding ( tb ) model of abc - stacked graphene and introduce certain notations that we will be using in subsequent sections . in sect .
[ sec : intrinsic - soc ] , employing group - theoretical considerations and the slater - koster@xcite ( sk ) parametrization of transfer integrals , we derive the soc hamiltonian in atomistic approximation at the @xmath2 point of the brillouin zone .
we repeat this calculation for bilayer graphene in sect . [
sec : bilayer ] so that in sect .
[ sec : so - abc - sl ] we can make predictions for the actual values of the soc parameters .
next , in section [ sec : eff_so_ham ] , using @xmath9 theory and the schrieffer - wolff transformation@xcite , we derive an effective low - energy soc hamiltonian which is valid for wave - vectors around the @xmath2 ( @xmath10 ) point . finally , in sect .
[ sec : induced - soc ] , we consider the case when soc is enhanced in one of the graphene layers with respect to the other(s ) .
the basic electronic properties of abc trilayer are well captured by the effective mass model which is derived assuming one @xmath11 type atomic orbital per carbon atom .
this model has been discussed in detail in refs . and , ; therefore we give only a very short introduction here ( see also fig . [ fig1 ] ) . and
@xmath12 are the two lattice vectors .
b ) schematic side view of the unit cell with the most important hopping amplitudes .
c ) schematic of the brillouin zone with reciprocal lattice vectors @xmath13 , @xmath14 , the high - symmetry points @xmath15 , @xmath2 and @xmath10 and the momentum @xmath16 measured from @xmath2 .
d ) schematic of the band structure at the @xmath2 point of the brillouin zone .
the velocity @xmath17 is given by @xmath18 , where @xmath19 is the lattice constant : @xmath20 . ]
there are six carbon atoms in the unit cell of abc trilayer graphene , usually denoted by @xmath21 , @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 , where @xmath27 and @xmath28 denote the sublattices and @xmath29 is the layer index .
the parameters appearing in the effective model are : @xmath30 for the intra - layer @xmath31 nearest - neighbour hopping , @xmath32 for the interlayer hopping between sites @xmath33 and @xmath34 , @xmath35 ( @xmath36 ) describes weaker nearest - layer hopping between atoms belonging to different ( the same ) sublattice , and finally @xmath37 denotes the direct hopping between sites @xmath21 and @xmath26 that lie on the same vertical linein the outer layers @xmath38 and @xmath39 .
these hoppings can be obtained by e.g. fitting the numerically calculated band structure with a tb model@xcite .
the six @xmath11 orbitals in the unit cell give rise to six electronic bands@xcite . as shown in fig.[fig1](d ) , at the @xmath2 ( @xmath10 ) point of the brillouin zone ( bz ) two of these bands lie close to the fermi energy @xmath40 and we will refer to them as " low - energy `` states .
in addition , there are four ' ' split - off
states far from @xmath6 at energies @xmath41 . to obtain the intrinsic soc hamiltonian of abc trilayer graphene we generalize the main idea of ref . where the soc of monolayer graphene was discussed . using group - theoretical considerations and density functional theory ( dft ) calculations
it was shown in ref . that in the case of monolayer graphene the main contribution to the intrinsic soc comes from the admixture of @xmath11 orbitals with some of the ( nominally unoccupied ) @xmath0 orbitals , namely , with the @xmath4 and @xmath5 orbitals .
the other @xmath0 orbitals , @xmath1 , @xmath42 and @xmath43 play no role due to the fact that they are symmetric with respect to the mirror reflection @xmath44 to the plane of the graphene layer , whereas @xmath11 is antisymmetric .
the importance of the @xmath0 orbitals for the understanding of soc in monolayer graphene was also pointed out in ref . .
the symmetry group of abc stacked trilayer graphene is @xmath45 ( @xmath46 ) which does not contain the mirror reflection @xmath47 .
therefore , from a symmetry point of view , in an approach similar to ref . , all @xmath0 orbitals need to be taken into account . to derive a tb model one should therefore use as basis set the bloch wavefunctions @xmath48 where the wavevector @xmath49 is measured from the @xmath15 point of the bz ( see fig . [ fig1 ] ) ,
@xmath50 is a composite index for the sublattice @xmath51 , @xmath52 and layer @xmath53 indices and @xmath54 denotes the atomic orbitals of type @xmath55 in layer @xmath56 .
the summation runs over all bravais lattice vectors @xmath57 , whereas the vectors @xmath58 give the position of atom @xmath59 in the two - dimensional unit cell .
we use a coordinate system where the primive lattice vectors are @xmath60 and @xmath61 , the positions of the atoms in the unit cell are @xmath62 , @xmath63 and @xmath64 where @xmath65 is the lattice constant .
the @xmath2 and @xmath10 points of the brillouin zone , which are important for the low energy physics discussed in this paper , can be found at @xmath66 , @xmath67 .
the symmetry group of the lattice contains threefold rotations by @xmath68 , about an axis perpendicular to the graphene layers .
since the atomic orbitals @xmath69 themselves do not possess this symmetry , instead of @xmath70 given in eq .
( [ bloch - functs ] ) we will use bloch states which depend on @xmath71 , @xmath72 ( rotating orbitals ) , where @xmath73 are spherical harmonics .
taking into account that @xmath74 and @xmath75 , the bloch states we use as basis will be denoted by @xmath76 , where @xmath77 and @xmath78 , whereas @xmath79 can take all allowed values @xmath80 .
often , we will need a linear combination of two of these basis functions where both of the basis functions have the same quantum number @xmath81 but one of them is centered on an @xmath27 type atom and the other one is on a @xmath28 type atom , e.g. @xmath82 $ ] . as a shorthand notation , we will denote the symmetric combination of two such basis functions by @xmath83 $ ] and the antisymmetric one with @xmath84 .
$ ] the first upper index @xmath56 in @xmath85 always denotes the layer index of the atomic orbital centered on the @xmath27 type atom , the second upper index @xmath86 is the layer index for the orbital centered on the @xmath28 type atom , the first lower index @xmath81 is the common angular momentum quantum number , and finally , the second and the third lower indices @xmath87 , @xmath88 give the magnetic quantum number in the same manner as the upper indices give the layer index . to lighten the notation
, we will usually suppress the dependence of the bloch functions on @xmath89 and use the bra - ket notation , e.g. @xmath90 , @xmath91 .
the derivation of the spin - orbit hamiltonian proceeds in the same spirit as in ref . :
( i ) first , we neglect the spin degree of freedom . using the slater - koster parametrization to describe the hopping integrals @xmath92 ( @xmath93 is the single particle hamiltonian of the system ) at a high symmetry point ( the @xmath2 point ) of the brillouin zone and group - theoretical considerations we obtain certain effective bloch wavefunctions which comprise @xmath11 and @xmath0 atomic orbitals centered on different atoms ; ( ii ) using these effective wavefunctions we calculate the matrix elements of the spin - orbit hamiltonian in atomic approximation , and ( iii ) employing the @xmath9 theory we obtain the bands around the @xmath2 point and then we derive an effective low - energy hamiltonian .
if , in addition to the @xmath11 orbitals , we include also the @xmath0 orbitals into our basis , there will be six basis functions @xmath94 centered on each of the six carbon atoms in the unit cell and hence the tb hamiltonian @xmath95 , which is straightforward to calculate in the sk parametrization , is a @xmath96 matrix . by e.g. numerically diagonalizing this matrix one would find that the @xmath11 orbitals hybridize with some of the @xmath0 orbitals and one could see how the low energy and the split - off states , obtained in the first instance by neglecting the @xmath0 orbitals , are modified . according to band theory
each state at the @xmath2 point should belong to one of the irreducible representations of the small group of the @xmath2 point@xcite , which is @xmath97 ( @xmath98 ) in this case .
this group has two one - dimensional irreducible representation , denoted by @xmath99 and @xmath100 respectively , and a two - dimensional one denoted by @xmath101 ( see appendix [ sec : irrep - basis - funct ] ) .
the matrix elements of @xmath95 between basis states corresponding to different irreducible representations of @xmath98 are zero@xcite . in other words
, @xmath95 can be block - diagonalized by choosing suitable linear combinations of the basis functions @xmath94 such that the new basis functions transform as the irreducible representations of the group @xmath98 because the hybridization between @xmath11 and @xmath0 orbitals will preserve the symmetry properties .
a group - theoretical analysis of the problem shows that in a suitable basis @xmath95 is block - diagonal having ( i ) two @xmath102 blocks which we denote by @xmath103 and @xmath104 , they correspond to basis states with @xmath99 and @xmath100 symmetry , and ( ii ) there is one @xmath105 block @xmath106 corresponding to states with @xmath101 symmetry .
( the basis vectors with @xmath99 , @xmath100 and @xmath101 symmetries are listed in appendix [ sec : irrep - basis - funct ] , table [ tbl : symmetry - basis - full - trilayer ] ) . as a concrete example we will consider @xmath103 and discuss how one can extract an effective orbital in which @xmath11 atomic orbitals with large weight and @xmath0 orbitals with small weight are admixed .
the calculation for @xmath104 and @xmath106 cases is analogous and will be presented only briefly .
the basis states transforming as the irreducible representation @xmath99 are @xmath107 , @xmath108 , @xmath109 , @xmath110 , @xmath111 and @xmath112 .
the tb hamiltonian @xmath103 can be further divided into @xmath113 blocks : @xmath114 explicitly , the upper left block @xmath115 , corresponding to the basis states @xmath116 and @xmath109 , reads @xmath117 where the upper indices @xmath21 , @xmath26 on the sk parameters indicate the atomic sites between which the hopping takes place .
the parameter @xmath118 describes hopping between @xmath27 and @xmath28 type atoms within the same graphene layer and we assume that its value is the same in all three layers .
the matrix elements in @xmath119 of hamiltonian ( [ a1_tb_ham ] ) are either zero@xcite or describe skew hoppings between the @xmath11 and @xmath0 orbitals located on different atoms .
we assume that these skew hoppings are much smaller than both the vertical hopping @xmath120 and @xmath118 .
this is not a crucial assumption and the neglected skew - hoppings can be taken into account in a straightforward manner .
however , it simplifies the lengthy algebra that follows and we believe it yields qualitatively correct results ( see section [ sec : bilayer ] ) . with @xmath121
we see that @xmath122 ( corresponding to the basis functions @xmath110 , @xmath111 and @xmath112 ) is decoupled from @xmath123 and that this latter matrix describes the hybridization between the @xmath11 orbital based basis vector @xmath107 and the basis vectors @xmath108 , @xmath109 involving @xmath4 , @xmath5 and @xmath1 orbitals . by diagonalizing @xmath115 one could find out how the energy @xmath124 of one of the low - energy states is modified by the @xmath0 orbitals .
the secular equation @xmath125 leads to a cubic equation in @xmath126 and the solutions can only be expressed using the cardano formula .
instead , we next perform a schrieffer - wolff transformation ( lwdin partitioning ) @xmath127 to approximately block - diagonalize @xmath115 into a @xmath128 and a @xmath129 block by eliminating the matrix elements between @xmath107 on one hand and @xmath108 , @xmath109 on the other hand .
( a detailed discussion of this method can be found in e.g. refs . .
) the matrix @xmath130 is anti - hermitian : @xmath131 and only its nondiagonal blocks @xmath132 and @xmath133 are non - zero . in first order@xcite of the coupling matrix elements @xmath118 and @xmath120 one finds that @xmath134 where @xmath135 , @xmath136 and @xmath137 .
the @xmath138 block of @xmath139 reads @xmath140 , this means that the energy @xmath141 of the low energy state is shifted by @xmath142 while the schrieffer - wolff transformation is usually used to obtain effective hamiltonians , one can also obtain the new basis in which @xmath139 is blockdiagonal . making the approximation @xmath143 ( see ref . )
we find that the purely @xmath11-like state @xmath107 is transformed into @xmath144 i.e. it is admixed with two other basis vectors containing @xmath145 and @xmath1 orbitals .
since @xmath146 corresponds to a @xmath128 and hence diagonal block of @xmath147 , it is an approximate eigenvector of @xmath115 with energy @xmath148 .
the upper index @xmath11 in @xmath146 is meant to indicate that in this state @xmath11 orbitals have the largest weight .
there are two other states with @xmath99 symmetry which could be obtained by diagonalizing the the remaining @xmath129 block of @xmath139 . in these states
@xmath108 and @xmath109 would have large weight .
they are however far remote in energy from @xmath146 and therefore play no role in our further considerations
. the situation will be similar in the case of the two other irreducible representations , @xmath100 and @xmath101 , therefore we will suppress the upper index @xmath11 henceforth in the notation of the physically important approximate eigenstates .
we now briefly discuss the symmetry classes @xmath100 and @xmath101 .
the calculation for the other @xmath149 block of @xmath150 with @xmath100 symmetry is completely analogous to the @xmath99 case , the resulting approximate eigenvector , @xmath151 is shown in the left column of table [ tbl : slatkost - soc ] .
its energy , apart from the shift due to the @xmath0 orbitals , which will be neglected , is @xmath152 .
the matrix block corresponding to states with @xmath101 symmetry can be written as @xmath153 here the @xmath154 block @xmath155 contains the matrix elements between the basis vectors @xmath156 , @xmath157 , @xmath158 , @xmath159 , the @xmath160 block @xmath161 is the coupling matrix between the above shown @xmath11 orbital based basis vectors and the @xmath0 orbitals based basis vectors .
( the full set of basis vectors for each of the three symmetry classes is listed in appendix [ sec : irrep - basis - funct ] . )
a direct calculation shows that @xmath162 is a diagonal matrix with @xmath163 entries .
earlier we have referred to these states as split - off states . since @xmath155 is diagonal , its approximate eigenvectors @xmath164 , @xmath165 , @xmath166 and @xmath167 , which are listed in table [ tbl : slatkost - soc ] , can be obtained in exactly the same way as @xmath168 .
we will refer to the basis formed from the physically important approximate eigenvectors @xmath169 as the symmetry basis henceforth .
( the symmetry basis for the @xmath10 point can be obtained by complex - conjugation . ) looking at these basis vectors we see that in contrast to monolayer graphene@xcite , where only @xmath4 and @xmath5 orbitals hybridize with the @xmath11 orbital , here also the @xmath1 orbitals are admixed . as it will be shown below
, the admixture of @xmath1 orbitals is crucial to obtain the non - diagonal elements of the soc hamiltonian .
we can now proceed to calculate the soc hamiltonian .
this can be done in the atomic approximation , whereby the spin - orbit interaction is described by the hamiltonian @xmath170 here @xmath171 is the spherically symmetric atomic potential , @xmath172 is the angular momentum operator and @xmath173 is a vector of spin pauli matrices @xmath174 ( with eigenvalues @xmath175 ) . introducing the spinful symmetry basis functions by @xmath176 , where @xmath177 denotes the spin degree of freedom and noting that @xmath178 , where @xmath179 and @xmath180 , it is straightforward to calculate the matrix elements @xmath181 in the symmetry basis introduced earlier . using the notation @xmath182 , where @xmath183 corresponds to the @xmath2 ( @xmath10 ) point of the bz , the result is shown in table [ intrinsic - so - trilayer ] . referring to fig .
[ fig1](a ) , the atomic sites participating in the formation of the split - off bands are denoted by @xmath22 and @xmath23 , whereas they are labeled as @xmath21 and @xmath24 in ref . , and similarly for the sites contributing to the low energy bands .
therefore , to arrive at the soc hamiltonian shown in table [ tbl : intrinsic - so - bilayer ] , ( i ) one has to rotate the symmetry basis into the on - site basis , ( ii ) re - label the sites as @xmath186 , ( iii ) and make the identification @xmath187 , @xmath188 , @xmath189 , @xmath190 .
the matrix elements are real numbers in our calculations because we use different lattice vectors than in ref . .
here we collect some useful formulae : we give explicitly the transformation between the symmetry basis and the effective on - site @xmath11 orbital basis for the trilayer case and present the low - energy electronic hamiltonian for bilayer graphene . for completeness and for comparison with the trilayer case given in eq .
( [ inv_sym_el_eff ] ) , we show here the low - energy electronic hamiltonian of bilayer graphene in the on - site basis @xmath21 , @xmath24 .
this hamiltonian has been discussed in many publications before , see e.g. the recent review of ref . .
as in sect .
[ sec : induced - soc ] , we assume that atoms @xmath21 and @xmath22 in the layer adjacent to the substrate have a different on - site energy @xmath192 than atoms @xmath23 , @xmath24 in the second layer . the most important terms are found to be : @xmath193 \left ( \begin{array}{cc } 1 & 0\\ 0 & 0 \end{array } \right ) .
\label{}\end{aligned}\ ] ] the approximation @xmath143 can be justified by the numerical calculations of ref . .
there it was shown that in the case of monolayer graphene @xmath194 .
since @xmath37 is a small energy scale compared to @xmath195 , we find that @xmath196 .
furthermore , since @xmath197 and @xmath198 correspond to hopping between @xmath21 and @xmath26 atoms which are at larger distance than the @xmath27 and @xmath28 atoms within the same graphene layer , one can expect that @xmath199 and @xmath200 , hence both matrix elements of @xmath201 in eq .
( [ s - approx ] ) are much smaller than unity . ref
. considers an effective one - layer model and obtains a soc hamiltonian which is wavenumber independent and can be characterized by two soc constants .
one of the so parameters , denoted by @xmath202 in ref . , corresponds to our @xmath203 .
this can be seen by folding down the full hamiltonian @xmath204 to obtain a low - energy effective model [ see eq .
( [ so - ham - compact ] ) in sect .
[ sec : eff_so_ham ] ] .
further wavenumber - independent soc terms can also be obtained from the folding down procedure , but they are higher order in the soc parameters . therefore we think that they are negligible with respect to the effective soc hamiltonian shown in eq .
( [ so - ham - compact ] ) .
more precisely , as we have already mentioned in connection with the intrinsic soc@xcite , when one folds down the full hamiltonian one can obtain additional momentum - independent terms for the low - energy hamiltonian but they are proportional to products and higher powers of the soc parameters .
we neglect these terms . | we present a combined group - theoretical and tight - binding approach to calculate the intrinsic spin - orbit coupling ( soc ) in abc stacked trilayer graphene .
we find that compared to monolayer graphene ( s. konschuh , m. gmitra , and j. fabian [ phys . rev .
b 82 , 245412 ( 2010)])@xcite , a larger set of @xmath0 orbitals ( in particular the @xmath1 orbital ) needs to be taken into account .
we also consider the intrinsic soc in bilayer graphene , because the comparison between our tight - binding bilayer results and the density functional computations of ( ref . allows us to estimate the values of the trilayer soc parameters as well .
we also discuss the situation when a substrate or adatoms induce strong soc in only one of the layers of bilayer or abc trilayer graphene . both for the case of intrinsic and externally induced soc
we derive effective hamiltonians which describe the low - energy spin - orbit physics .
we find that at the @xmath2 point of the brillouin zone the effect of bychkov - rashba type soc is suppressed in bilayer and abc trilayer graphene compared to monolayer graphene . |
a superconductor in electric contact with a normal metal induces superconducting correlations on the normal side .
this _ proximity effect _ has been studied extensively , both theoretically and experimentally ( see ref . and references therein ) .
the superconducting properties of the normal metal show up , e.g. , in the conductance or the magnetization ( meissner effect ) .
the vanishing of the resistance of a normal wire in contact to superconducting islands has been observed . in recent experiments ( typically with wires consisting of a superconducting core covered by a normal metal ) by oda and nagano @xcite , mota _ et al . _
@xcite , and bergmann _ et al .
_ @xcite the magnetization was measured . in each of these experiments ,
the magnetic field was expelled from the normal - metal part ( n ) of the sample by superconducting screening currents induced by the presence of the superconductor ( s ) . in the present work ,
we study theoretically the proximity and meissner effect in an ns sandwich using the quasiclassical eilenberger formalism @xcite . in materials with a high concentration of impurities ( dirty limit ) , the quasiclassical theory can be reformulated leading to the usadel equation @xcite . in both approaches the induced pair amplitude in the normal metal , the density of states , the critical current , etc .
, have been calculated ( see , e.g. , refs . ) . in general , for a realistic geometry the solution of the eilenberger or usadel equation together with the maxwell equations can be performed only numerically .
we have solved the combined system of equations in a wide range of parameters ( temperature , external magnetic field , layer thickness ) .
in addition we have paid special attention to the non - linear response that shows interesting hysteretic behavior at low temperatures . in section
ii we introduce the quasiclassical green s function formulation for the clean and dirty limit .
the geometry and the characteristic length scales are also defined there . in section iii
we evaluate the meissner screening current and present some results on the space dependence of the various quantities .
finally , in section iv results for the susceptibility in a wide range of temperatures and values of the magnetic field are presented and discussed . throughout the paper we use units with @xmath0 .
our theoretical description is based on the quasiclassical green s function technique @xcite .
we will study two limiting cases , the clean limit with complete absence of impurities and the dirty limit with high concentration of impurities such that the elastic scattering rate @xmath1 is large compared to the superconducting order parameter @xmath2 and the temperature @xmath3 .
we introduce two different coherence lengths for both the superconducting and the normal side of the double - layer structure . in a superconducting material
we have in the clean limit @xmath4 and in the dirty case @xmath5 here @xmath6 is the fermi velocity , @xmath2 the pair potential and @xmath7 the diffusion constant .
these lengths are temperature - independent for @xmath8 . in a normal metal
we define the temperature - dependent coherence lengths in the clean limit @xmath9 and in the dirty limit @xmath10 for definiteness we first consider a one - dimensional geometry with a bulk superconductor in the region @xmath11 and a normal metal layer in the region @xmath12 in perfect electric contact with the superconductor .
( later we will also present some results for the cylinder geometry of the experiments . )
since we assume @xmath13 , we can neglect the spatial dependence of the pair potential in the superconductor .
this has also been confirmed numerically @xcite .
therefore , we assume the following form for the pair potential @xmath14 for the vector potential we use the coulomb gauge . together with the boundary conditions it is written as @xmath15 where @xmath16 is the external applied field . we have neglected the magnetic field on the superconducting side , i.e. , we assume that the penetration depth on the superconducting side is much smaller than the other relevant length scales .
we first consider the clean limit defined by @xmath17 the eilenberger equations for this system read @xmath18=0\ ; , \ ] ] which is to be combined with the normalization condition for the green s functions @xmath19 the matsubara frequencies are @xmath20 , @xmath21 are the pauli matrices and @xmath22,@xmath23 are the @xmath24- and @xmath25-components of the fermi velocity .
we assume ideal transmission , i.e. there are no surface potentials at the n - s - boundary and the fermi velocities are equal on both sides .
the boundary condition reads @xmath26 . at the interface to the vacuum
we assume specular reflection , @xmath27 .
the current density in the @xmath25-direction is given by @xmath28 where @xmath29 is the density of states at the fermi level and @xmath30 denotes averaging over the fermi surfaces that are assumed to be spherical . in systems with high concentration of nonmagnetic impurities , such that @xmath31 the eilenberger equation reduces to the usadel equation for the isotropic part of the green s function , @xmath32 .
for our geometry it can be written as @xmath33 \nonumber\end{aligned}\ ] ] we take the same boundary conditions as in the clean case , i. e. we neglect the magnetic field on the superconducting side . in the dirty limit the current density is given by the london - like expression @xmath34a(x)\,,\end{aligned}\ ] ] which together with the maxwell equation @xmath35 defines a local penetration depth @xmath36 .
in the clean limit an analytic solution of eq .
( [ eilenberger.clean ] ) for the normal metal layer has been found by zaikin @xcite .
it turns out that the @xmath37 component of the green s function in the normal metal layer is spatially constant .
consequently the current density ( [ clean.current.general ] ) in the normal metal layer is spatially constant as well and can be expressed as @xmath38^{-1}\ ; .\end{aligned}\ ] ] here we have introduced a dimensionless constant @xmath39 where @xmath40 is the normal - metal penetration depth , defined analogously to the london penetration depth with the normal electron density @xmath41 replacing the superfluid density .
furthermore , @xmath42 is the length of a classical trajectory divided by the thermal coherence length @xmath43 and @xmath44 is the aharonov - bohm phase connected with this trajectory .
the last equation shows , that the relationship between current and vector potential is completely nonlocal .
this phase factor leads to a shift in the energies of the andreev levels in normal metal layer @xcite .
the fact that these bound states are extended over the whole thickness of the n - layer and that the density of andreev levels is spatially constant , leads to a constant current density .
we can add , that in cylindrical geometries if the normal layer is not thin compared to the radius , this second condition is not satisfied . hence , as was pointed out by nazarov @xcite , the current density is not constant in space .
to evaluate the meissner effect , we have to solve the maxwell equations with the boundary condition given in eq .
( [ boundary.condition ] ) .
the solution in the n - layer is @xmath45 for the susceptibility of the n - layer we find @xmath46 since the solution of the problem may lead to @xmath47 the magnetic field in the region @xmath48 may change its sign relative to the applied field .
this overscreening effect was first found by zaikin @xcite . in fig .
[ clean.current ] we have plotted the current density and the magnetic field in the n - layer for different temperatures in the limit of small magnetic fields . below @xmath49 we find the overscreening .
below this temperature the screening quickly reaches its saturation value ( solid curve ) and a susceptibility of @xmath50 as compared to complete screening .
thus , even at @xmath51 the screening is incomplete .
next we study the nonlinear response . for this purpose
we write the self - consistency equation for the integrated dimensionless vector potential @xmath52 @xmath53 solving eq .
( [ vp.integral ] ) for the applied field @xmath16 we get , together with eq .
( [ susc.clean ] ) , the magnetization curve @xmath54 @xmath55 parameterized by the integrated vector potential . here
( [ usadel.equation ] ) can not be solved analytically anymore . to proceed
numerically we reduce it to the scalar equation @xmath56,\end{aligned}\ ] ] where @xmath57 is the off - diagonal part of the green s function and @xmath58 is fixed by the normalization ( [ normalisation ] ) .
the maxwell equation is given by eq .
( [ maxwell.dirty ] ) with the boundary conditions given in ( [ boundary.condition ] ) .
we write for the local penetration depth @xmath59 where we have defined an dimensionless constant @xmath60 .
we have solved this system of equations numerically for a range of parameters .
some results are shown in fig .
[ dirty.current ] for a layer thickness of @xmath61 and @xmath62 in a weak magnetic field , where we neglect the term @xmath63 in the usadel equation . at the highest temperature , @xmath64
, the field penetrates through the whole n - layer . near the n - s - boundary ,
where the induced superconductivity is strong , the field decays rapidly to zero . at lower temperatures the field - free region increases , because the induced superconductivity extends to a region of linear size @xmath65 . at the lowest temperature shown here , @xmath66 ,
the field expulsion is london like and the field as well as the current density fall off exponentially .
the reason is that at this temperature the local penetration depth near the n - vacuum - boundary is practically constant . in the intermediate temperature regime
the current flows in a well defined region inside the n - layer , and consequently the screening takes place in this region . to calculate the nonlinear response we have to solve the complete usadel equation eq .
( [ usadel.equation ] ) together with the maxwell equations including the local penetration depth ( [ local.pendepth ] ) numerically .
from the solution of the maxwell equations , we find the susceptibility of the normal - metal part @xmath67 results for the clean and the dirty case are discussed in the next section .
first we study the case of linear response . in the clean limit we linearize eq .
( [ clean.jn ] ) with respect to the aharonov - bohm phase @xmath68 .
the resulting expression has been analyzed analytically by zaikin in some limits @xcite and numerically by higashitani and nagai @xcite .
we have performed the remaining integration over @xmath69 numerically for all temperatures . in fig .
[ clean.linear ] some results for different layer thicknesses are shown . the temperature at which screening sets in , as well as the slope of the temperature dependence , changes drastically when the layer thickness increases . at @xmath51
the integrals can be solved analytically . for @xmath70 ,
the saturation value is equal to @xmath50 of a perfect diamagnet , independent of the thickness @xcite .
here we should mention that the temperature dependence of the susceptibility can not be fitted to a power law @xmath71 in any temperature range , in contrast to earlier theoretical predictions @xcite which suggest powers in the range @xmath72 depending on the impurity concentration .
rather it satisfies an exponential law .
this explains qualitatively some experimental results @xcite , where exponents @xmath73 with values up to @xmath74 had been found . in the dirty limit
, we neglect the term @xmath63 in eq .
( [ usadel.equation ] ) and solve the resulting differential equation numerically . in fig .
[ dirty.linear ] some results of our calculation are shown . in comparison with the clean case , the screening sets in at higher temperatures and the saturation values can be larger ( depending on @xmath75 ) .
the temperature dependence is well described by @xmath76 in the intermediate temperature regime well below @xmath77 and above the saturation temperature .
our results agree with previous works in the applicable limits . in earlier work @xcite , a generalized ginzburg - landau approach was used .
narikiyo and fukuyama @xcite linearized eq .
( [ usadel.equation ] ) with respect to @xmath78 and solved the system of equations for an infinite system numerically . however , our calculation is free of the limitations of the ginzburg - landau theory and is valid at any temperature .
furthermore , we have taken into account the nonlinearity of the usadel equation and finite - size effects . the behavior shown in fig .
[ dirty.linear ] has also been observed experimentally in very dirty samples @xcite . in these experiments ,
a temperature dependence described by @xmath79 was found and saturation values of the susceptibility of @xmath80 of a perfect diamagnet . in fig .
[ dirty.exp ] we show a comparison between theory and experiment .
the experimental data @xcite are measured on a cylindrical geometry with a superconducting core surrounded by a normal metal layer of thickness @xmath81 . the theory has been generalized to this kind of geometry and also been solved numerically .
the parameters for the theory were calculated from experimental values for @xmath81 and @xmath82 and from theoretical values for @xmath83 and @xmath84 .
no fitting parameter was used .
the agreement is quite remarkable at low temperatures . at higher temperatures ( @xmath85 )
the dirty limit theory can not be applied anymore and this explains the disagreement in this temperature regime .
we will now turn to the nonlinear response . in fig .
[ clean.nonlinear ] some numerical results for the clean case are shown .
at the highest temperature shown here , the susceptibility drops to zero around @xmath86 . at lower temperatures the solution is not unique
there is one solution with the maximal field expulsion , the value of which depends on temperature ( lower thick lines ) , and one solution , for which the field completely penetrates in the n - layer and there is no screening ( upper thick lines ) .
the third solution with the negative derivative is unstable ( thin lines ) .
the region in which the solution is non - unique grows , as the temperature is lowered , but the lower boundary of this region stays at the same value of the field . since only one solution can be stable , there must be a jump in the susceptibility at a certain point . in principle
we could determine this point by comparing the free energies of the n - layer for the two different solutions .
unfortunately the free - energy functional proposed by eilenberger @xcite can not be applied for @xmath87 ( normal region ) . in experiments
the situation is often quite different .
on lowering or raising the field , the jump in the susceptibility would not occur at the value predicted theoretically , but effects like `` superheating '' or `` supercooling '' are likely to occur .
these effects were observed in very clean samples @xcite .
when the field is increased , the susceptibility will remain on the lower branch up to a certain field value and then will jump to the upper branch .
on the other hand , when the field is decreased , the system will stay on the upper branch with complete field penetration . at a certain value of the field , the field
will suddenly be expelled and the system will jump to the lower branch .
these jump fields do not necessarily coincide with the boundaries of the instability zone .
the boundaries of the hysteretic regions are shown in fig .
[ clean.breakdown ] for different layer thicknesses .
the superheating field depends exponentially on temperature , whereas the supercooling field depends not very strongly on temperature . at low temperatures
the superheating field tends to saturate at a certain value , which is independent of the layer thickness .
further calculations show that the saturation value is proportional to the material constant @xmath88 . in the dirty limit
there is also a temperature - dependent hysteretic effect , as shown in fig .
[ dirty.nonlinear ] . above temperatures of about @xmath89
the magnetization curve is unique . below these temperatures there
is a region with constant lower boundary and increasing upper boundary as in the clean case .
the absolute values of the susceptibility are drastically different
. the saturation value at low fields and low temperatures can reach complete field expulsion for appropriately chosen values of @xmath90 . on the other hand
the jump in the susceptibility is less than in the clean case . on increasing the field
the susceptibility is reduced only by @xmath91 at the jump and reaches slowly zero , if the field is increased further . on lowering the field ,
the jump is a slightly smaller .
this hysteretic behavior was observed @xcite . in fig .
[ dirty.breakdown ] the limits of the region with hysteretic behavior for different layer thicknesses in the dirty limit are shown .
as in the clean case the superheating field depends exponentially on the temperature , but not as strongly . also , the supercooling field tends to become a constant at low temperatures and does not very strongly depend on temperature .
in contrast to the clean case the saturation values of both fields at low temperatures depend strongly on the layer thickness . finally , we would like to comment on some of our assumptions .
we have assumed ideal interfaces between normal metal and superconductor , specular reflection at the interface between normal metal and vacuum , and spherical fermi surfaces in both materials . the formalism presented in this paper
can be modified to describe more general physical situations . including e.g. non - ideal interfaces
will weaken the proximity effect and diminish the diamagnetic response of the normal metal . in this paper , we have concentrated on presenting a model calculation in which the main aspects of the diamagnetic response of ns sandwiches can be studied . despite our simplifying assumptions ,
the calculation was shown to be in reasonable agreement with experiment , see fig .
[ dirty.exp ] .
further calculations will be necessary to describe the high - temperature behaviour correctly and to study arbitrary concentrations of impurities , i.e. , cases that are neither clean nor dirty . in conclusion
, we have applied the quasiclassical ( eilenberger or usadel ) theory to calculate the meissner effect in a proximity sandwich .
we have evaluated the full non - linear magnetic response at all temperatures and for various thicknesses of the normal layer .
we have shown that both in the clean and in the dirty limit a hysteretic behavior of the magnetization of a normal metal layer in proximity with a superconductor is possible , as has been observed experimentally in both limits . | the magnetic response of a proximity - coupled superconductor - normal metal sandwich is studied within the framework of the quasiclassical theory .
the magnetization is evaluated for finite values of the applied magnetic field ( linear and nonlinear response ) at arbitrary temperatures and is used to fit recent experimental low - temperature data .
the hysteretic behavior predicted from a ginzburg - landau approach and observed in experiments is obtained within the quasiclassical theory and shown to exist also outside the ginzburg - landau region . |
the search for new physics beyond the standard model is continuing at the large hadron collider ( lhc ) .
although we are yet to see any clear hint , there is reason for exhilaration in another way , namely , the discovery of a boson of mass @xmath2 125 gev @xcite . the properties of this particle are very similar to the higgs boson predicted in the standard model ( sm ) , but possibilities of some new physics information contained in it can not yet be ruled out .
thus a great deal of attention has shifted to the exploration of physics beyond the standard model ( bsm ) in the electroweak symmetry breaking sector .
one vexing issue , often mentioned as a motivation for bsm physics , is the identification of a mechanism for neutrino mass generation .
this basically means finding some explanation for the smallness of neutrino masses as compared to those for the other fermions , and also the very different nature of mixing evinced in the neutrino ( or more precisely , lepton ) sector @xcite .
it is thus natural that efforts to unravel new physics in the higgs sector will sometimes be guided by considerations related to neutrino masses @xcite .
the lhc signals of any scenario proposed in this context are also of undeniable interest @xcite . in this paper , we consider a model which not only accounts for the tiny neutrino masses , but also plays a role in the electroweak symmetry breaking mechanism .
we are interested in a two higgs doublet model with right - handed neutrinos proposed by gabriel and nandi @xcite .
the essential idea is that neutrinos , like all other fermions , have dirac masses , but are much lighter than the others because their masses come from a different higgs doublet .
the yukawa couplings of the neutrinos can still be @xmath3 .
this is ensured by giving a very tiny vacuum expectation value ( vev ) @xmath2 ev to the neutral component of one of the higgs doublets , which due to a @xmath4 symmetry couples only with the neutrino sector .
the charged higgs in this model therefore has very different properties when compared to other standard two higgs doublet models ( 2hdm ) @xcite .
it is found to couple very weakly with the quarks while a large coupling with charged leptons and right - handed neutrinos is allowed .
thus , the leptonic mode is the most promising decay channel of this charged higgs for @xmath3 yukawa couplings in the neutrino sector @xcite . depending upon the mass of the charged higgs other decay modes
are also possible . in a similar model proposed by davidson and logan @xcite , which considers a @xmath5 global symmetry in place of the discrete @xmath4 symmetry ,
only leptonic decay modes of the charged higgs were allowed .
the branching probabilities for the decay of the charged higgs are very sensitive to the small vev of the additional doublet , which couples to the neutrinos @xcite .
as recent astrophysical bounds on the neutrino yukawa couplings in such extensions of the sm suggest that the vev of the second doublet can not be in the sub - ev range @xcite , the charged higgs can no longer decay dominantly into a leptonic final state .
it turns out that the main decay mode for the charged higgs is into a @xmath0 boson and a light neutral scalar present in the model . in this work we mainly focus on the challenges presented for a charged higgs search at the lhc , when it decays through the @xmath0 mode .
we study the pair production of the charged higgs at the lhc and consider its decay to @xmath0 boson and the neutral scalar .
the signal is identified by two isolated leptons and a large missing energy .
we have analyzed the most dominant sm background subprocesses that affect the signal , to estimate the signal significance .
the study is carried out at both the 8 tev and 14 tev center - of - mass energies for the lhc .
we find that it is practically impossible to achieve any significant signal for the charged higgs in this model with the available integrated luminosity at 8 tev .
the situation is much more optimistic at 14 tev , if one waits for a sizeable luminosity to accumulate .
there , on applying appropriate kinematical event selection procedure , a signal significance of 5@xmath1 at 14 tev can be reached with an integrated luminosity of 4000 fb@xmath6 for @xmath7 gev .
the observation of the signal is found to be more promising for heavier charged higgs masses .
the paper is organized as follows . in section [ sec : model ]
we describe the model briefly .
various constraints on the model parameters are discussed in section [ sec : constraints ] . in section [ sec : analysis ] , we discuss the charged higgs pair production and major backgrounds for the signal at the lhc .
results are presented in section [ sec : results ] and we conclude in section [ sec : summary ] .
the model under consideration is based on the symmetry group @xmath8 , where @xmath9 .
in addition to the matter fields in the sm , the model includes two scalar doublets @xmath10 and @xmath11 , and three @xmath12 singlet right - handed neutrinos @xmath13 , @xmath14 .
all the sm fermions and the scalar doublet @xmath10 are even under the discrete symmetry @xmath4 , while the right - handed neutrinos and the scalar doublet @xmath11 are odd under @xmath4 . the most general scalar potential and the yukawa interaction of leptons with the scalar doublets which respect the @xmath8 symmetry are @xcite , @xmath15 , \label{eq : pot } \\ { \cal l}_y & = & y_l^{ij}\ ; { \bar{\rm \psi}}^{l , i}_l l_r^j \chi + y_{\nu_l}^{ij}\ ; { \bar{\rm \psi}}^{l , i}_l \nu_r^j { \tilde \phi } + h.c . , \label{eq : lyuk } \end{aligned}\ ] ] where , @xmath16 and @xmath17 are the usual @xmath12 lepton doublet and singlet fields , respectively and @xmath18 represent the matrix elements of the lepton yukawa matrices .
the standard electroweak symmetry is broken spontaneously by giving a vev , @xmath19 246 gev to the @xmath10 doublet , while the @xmath4 symmetry is broken by a vev , @xmath20 for the @xmath11 doublet .
the spontaneous breaking of the @xmath4 symmetry is arranged for generating small neutrino masses , @xmath21 @xmath20 which can be in the sub - ev / ev range for @xmath3 yukawa couplings .
we note that we are assuming lepton number conservation so that the majorana mass terms for the right handed neutrinos , @xmath22 , @xmath23 are not allowed .
thus the light left - handed neutrinos can not acquire masses via the usual see - saw mechanism @xcite .
dirac mass , as obtained from eq .
[ eq : lyuk ] from the tiny vev of @xmath11 is the only possibility . as a result of the symmetry breaking ,
the physical higgs sector includes charged scalars @xmath24 , two neutral cp - even scalars @xmath25 and @xmath1 and a neutral pseudoscalar @xmath26 .
the masses for these particles are given by , @xmath27 where , @xmath28 .
we have neglected the subdominant terms in @xmath20 when deriving these relations .
we note that in the case of exact @xmath4 symmetry , the @xmath1 will be exactly massless .
the breaking of this @xmath4 symmetry with a tiny vev @xmath29 gives mass to the @xmath1 , as well as tiny dirac masses to the observed neutrinos .
therefore in this model , the neutral scalar field @xmath1 is very light and the field @xmath25 behaves like the sm higgs boson.the cp - even scalars ( @xmath30 ) are the mass eigenstates and they are related to the weak eigenstates @xmath31 by the mixing angle @xmath32 : @xmath33 where , @xmath34 this mixing can be neglected because @xmath35 .
it is also clear from the above equations that @xmath36 lies around the electroweak scale . in the lepton yukawa sector the above symmetry breaking leads to neutrino masses given by , @xmath37 , where @xmath38 are the eigenvalues of the neutrino yukawa matrix .
the yukawa interaction of the charged higgs with the leptons can then be written down following eq .
[ eq : lyuk ] as , @xmath39 in the above equation , @xmath40 represents the flavour index while @xmath41 is the index representing neutrino components in mass eigenstate .
the @xmath42 are the charged lepton yukawa couplings while @xmath43 represent the elements in the pmns matrix @xcite for the mixing of the neutrino flavours .. ] note that the second term in eq .
[ eq : yukawa ] is clearly sub - dominant and negligible ( suppressed by the factor @xmath44 ) when compared to the first term and can therefore be safely neglected when considering the interaction strength of the charged higgs with the leptons . as the couplings of charged higgs with sm quarks are also generated through terms similar to the second term in eq .
[ eq : yukawa ] , the charged higgs is very weakly coupled to quarks and in all practicality behaves as a `` chromophobic '' field .
this property of the charged higgs plays a crucial role in avoiding strong constraints on its mass , otherwise evident in other 2hdm , from low energy physics experiments such as weak meson decays and mixing .
thus in the yukawa sector , only the decay @xmath45 becomes relevant .
other main decay modes of the charged higgs include @xmath46 and @xmath47 which have gauge coupling interaction strengths .
ev and ( b ) @xmath48 ev .
the mass of @xmath1 particle is related to @xmath20 as in eq .
[ eq : masses ] .
we have chosen the other relevant variables @xmath49 and @xmath50 gev for calculating the above branching ratios.,title="fig:",width=302 ] ev and ( b ) @xmath48 ev .
the mass of @xmath1 particle is related to @xmath20 as in eq .
[ eq : masses ] .
we have chosen the other relevant variables @xmath49 and @xmath50 gev for calculating the above branching ratios.,title="fig:",width=302 ] the branching ratios of charged higgs decay is quite sensitive to the value of @xmath20 . as the leptonic channel @xmath51 is dictated by the coupling strength given by @xmath52 ( see eq .
[ eq : yukawa ] ) , smaller values of @xmath20 for a fixed neutrino mass increase the branching probability . to highlight this , we consider two sets of values for @xmath20 and show the branching probabilities of the charged higgs decay as a function of its mass in fig . [ fig : br ] .
note that the neutrino data as shown in table [ table : nudata ] have been incorporated when calculating the partial decay widths of the charged higgs decaying into the three generations of leptons .
if @xmath20 is in the sub - ev range as shown in fig .
[ fig : br]a , it is found to decay mostly through the leptonic mode for @xmath53 gev , while if @xmath20 is increased to about an ev ( fig .
[ fig : br]b ) , it decays dominantly into @xmath54 . as the value of @xmath20
is increased further , we find that the leptonic channel becomes completely negligible and the @xmath54 becomes the only significant mode available for the charged higgs decay for @xmath53 gev .
as the model under consideration is quite different from the generic 2hdm and is envisioned to account for the observed neutrino masses and mixing angles , it becomes imperative to first check how the experimental constraints affect the parameters of the model .
a brief discussion on several constraints on the model parameters is already present in ref .
we choose to accommodate them with new and updated results that have modified these constraints as well as supplement them with additional constraints , if any .
.neutrino mass - mixing parameters with 3@xmath1 uncertainties @xcite .
the allowed ranges of parameters for the normal hierarchy ( nh ) and inverted hierarchy ( ih ) cases are shown separately . [
cols="^,^,^",options="header " , ] the effects of applying various cuts on the signal and background events have been summarized as a cut flow scheme in table [ table:150 ] for the two benchmark values of the charged higgs mass of 150 gev and 200 gev .
we have selected only those cuts that increase the signal significance @xmath55 defined as @xmath56 where @xmath57 and @xmath58 are number of signal and background events respectively .
this significance estimator is useful for events with low statistics @xcite . when the background is large , the formula for @xmath59 reduces to the more familiar @xmath60 form used for estimating the signal significance .
although the number of events that satisfy all the applied cuts are low , the significance is nevertheless promising . with the 3000 fb@xmath6 integrated luminosity ,
the same set of cuts lead to a signal significance of about 6.5 for @xmath61 gev .
we can further improve the significance by optimizing various cuts .
for example , pushing the minimum @xmath62 cut on the higher side does help in achieving better significance .
as mentioned before , a suitable large cut on @xmath63 instead of the cut on @xmath64 can also be used to suppress the background efficiently . however , in that case the minimum @xmath62 cut should be relaxed slightly to maintain a high signal significance . ) as a function of minimum cuts on @xmath65 ( y - axis ) and @xmath66 ( x - axis ) for ( a ) @xmath67 fb@xmath6 , ( b ) @xmath68 fb@xmath6 , ( c ) @xmath69 fb@xmath6 , and , ( d ) @xmath70 fb@xmath6 .
the blue shaded regions in the above plots refer to 2@xmath1 statistical significance.,title="fig:",width=288,height=288 ] ) as a function of minimum cuts on @xmath65 ( y - axis ) and @xmath66 ( x - axis ) for ( a ) @xmath67 fb@xmath6 , ( b ) @xmath68 fb@xmath6 , ( c ) @xmath69 fb@xmath6 , and , ( d ) @xmath70 fb@xmath6 .
the blue shaded regions in the above plots refer to 2@xmath1 statistical significance.,title="fig:",width=288,height=288 ] ) as a function of minimum cuts on @xmath65 ( y - axis ) and @xmath66 ( x - axis ) for ( a ) @xmath67 fb@xmath6 , ( b ) @xmath68 fb@xmath6 , ( c ) @xmath69 fb@xmath6 , and , ( d ) @xmath70 fb@xmath6 .
the blue shaded regions in the above plots refer to 2@xmath1 statistical significance.,title="fig:",width=288,height=288 ] ) as a function of minimum cuts on @xmath65 ( y - axis ) and @xmath66 ( x - axis ) for ( a ) @xmath67 fb@xmath6 , ( b ) @xmath68 fb@xmath6 , ( c ) @xmath69 fb@xmath6 , and , ( d ) @xmath70 fb@xmath6 .
the blue shaded regions in the above plots refer to 2@xmath1 statistical significance.,title="fig:",width=288,height=288 ] as both @xmath65 and @xmath64 play an important role in enhancing the signal significance , it will be useful to study a possible correlation between the minimum cuts that can be applied on these kinematic variables .
we have shown this correlation in fig .
[ fig : contours ] using the contour plots for @xmath71=150 , 200 gev .
the plots are shown for two values of the integrated luminosity , _ viz _
@xmath72 fb@xmath6 and @xmath73 fb@xmath6 .
the contour plots have been obtained after applying the minimum @xmath74 gev cut along with the acceptance cuts . in these plots
we show the minimum cuts on @xmath64 and @xmath65 required to achieve a signal significance @xmath75 2 , keeping the total number of events @xmath76 5 .
the signal significance is clearly seen to increase substantially with more optimal choices of cuts for the two correlated kinematic variables .
the variation in the signal significance is represented with different color codes . with the help of these plots
it is easier to find the optimal values of the cuts on @xmath65 and @xmath64 which can maximize the significance .
we should also point out that our analysis for the two benchmark values of the charged higgs mass achieved a signal significance in accordance with the contour plots shown in fig .
[ fig : contours]a and [ fig : contours]b . however , the efficiency of these correlated cuts for the case of @xmath61 gev which is clearly visible in fig .
[ fig : contours]b , suggests that a significance , as high as 5@xmath1 , should be well within the reach with more optimal choices of the cuts when compared to those listed in table [ table:150 ] . as expected , with higher integrated luminosity , much better signal significance can be achieved ( see figs .
[ fig : contours]c and [ fig : contours]d )
. it is worth pointing out that in fig .
[ fig : contours ] , the constraint @xmath76 5 plays a major role in modifying the shape of common significance contours .
as the discrete cut - off for events will not be uniform for both @xmath77 3000 fb@xmath6 and @xmath78 5000 fb@xmath6 , it gives an impression of a non trivial scaling at different luminosities .
.,width=403 ] in fig .
[ fig : discovery ] , we estimate the signal significance for various charged higgs masses assuming 1000 , 3000 and 5000 fb@xmath6 integrated luminosities at the lhc with @xmath79 tev . we have applied the same set of cuts as listed in table [ table:150 ] . since we have applied a large missing energy cut , the significance increases for higher charged higgs masses . however , beyond a certain value of charged higgs mass , the significance goes down .
this is mainly due to the small pair production cross section of the charged higgs .
we note that for @xmath80 gev the significance is seen to become maximum in fig .
[ fig : discovery ] .
this is just the artifact of the choice of @xmath62 cut , which is more effective at that charged higgs mass .
note that beyond the charged higgs mass of about 180 gev , the @xmath47 decay channel is also open for @xmath81 100 gev .
this reduces the charged higgs branching ratio in the @xmath82 decay mode and hence reduces the signal cross section further for @xmath83 180 gev .
the upper bound on @xmath84 which is related to an upper bound on the coupling @xmath85 can be about 470 gev @xcite .
it means , by considering heavier @xmath26 mass we can ensure a 100% branching ratio of the charged higgs decay to @xmath86 for the heavier charged higgs masses .
it is needless to say that optimization of kinematic cuts is required to estimate the actual signal significance for different charged higgs masses .
for example , despite larger signal cross section for @xmath87 gev , the significance is maximized for a minimum @xmath62 cut of 90 gev . on the other hand , for @xmath88 gev
we have smaller signal cross section but a minimum cut of 130 gev on @xmath62 is more helpful in achieving a larger signal significance .
we note that ( @xmath89 ) combination which we have used to discriminate the signal from the background for @xmath90 gev is not very useful in increasing the significance for lower @xmath71 values .
this is related to the fact that the efficiency of these cuts is closely related to the high @xmath62 cut which in turn is related to the mass difference between the charged higgs and the @xmath0 boson .
when the mass difference between the charged higgs and the @xmath0 boson is not large enough , the application of high @xmath62 cut also kills the signal along with the background .
it is also worth pointing out that for larger charged higgs masses a suitable @xmath65 cut with large enough @xmath62 cut is sufficient . in other words
, if it is possible to apply a very high @xmath62 cut , the additional @xmath64 cut becomes less relevant .
we have studied the signatures of charged higgs boson at the lhc in a two higgs doublet model with right - handed neutrinos .
the model , aiming to derive neutrino masses via preferential yukawa couplings @xmath91(1 ) with an additional higgs doublet , also implies large yukawa coupling of the charged higgs with the light leptons and neutrinos .
however , if cosmological constraints are taken into account , the leptonic decay mode of the charged higgs is highly suppressed and @xmath92 is the dominant decay mode . in this study , the charged higgs pair production via the drell - yan process and its further decay leading to opposite sign di - leptons+missing energy in the final state is considered as the signal .
the major sm background to the signal comes from the process , @xmath93 .
we have done a complete signal - background cut based analysis for both the 8 tev and 14 tev center - of - mass energies at lhc .
the charged higgs masses of 150 gev and 200 gev serve as the benchmark points for our study .
since the signal has additional sources of missing energy , we find that a large @xmath62 cut helps in suppressing the sm background .
also , a combination of minimum cuts on the angle @xmath94 and @xmath64 plays an important role in enhancing the signal significance . due to the lack of sufficient data and the low signal cross section as compared to the background , the observability of the signal is not possible at the lhc with @xmath95 tev and we therefore carry out our analysis for the 14 tev run of the lhc .
we find that even at the 14 tev run of lhc , a charged higgs with the characteristics of a
fermiophobic " field will prove elusive without a very optimized kinematic selection of events even with high integrated luminosity . we show this by identifying the kinematic variables sensitive to specific selection cuts which with a large ( 3000 fb@xmath6 ) integrated luminosity yields a signal significance of 4.6@xmath1 for @xmath96 150 gev . for the case of @xmath96 200 gev
, a signal significance of 5@xmath1 can be easily achieved with better statistics .
we highlight the significance of the optimized cuts in the analysis through a correlation plot for event selection that enhances the signal significance in a very robust way .
our analysis indicates that the observation of this otherwise elusive fermiophobic charged higgs boson is quite promising at the high energy and high luminosity run of the lhc , provided a proper event selection criterion is applied .
we thank arindam chatterjee and sandhya choubey for fruitful discussions .
the work of um , bm , skr and as was partially supported by funding available from the department of atomic energy , government of india , for the regional centre for accelerator - based particle physics , harish - chandra research institute .
the work of sn was was supported in part by the u.s .
department of energy grant number de - sc0010108 .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 716 * , 1 ( 2012 ) ; s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * , 30 ( 2012 ) .
m. c. gonzalez - garcia and y. nir , rev .
phys . * 75 * , 345 ( 2003 ) [ hep - ph/0202058 ] .
e. ma , phys .
lett . * 86 * , 2502 ( 2001 ) [ hep - ph/0011121 ] .
s. gabriel and s. nandi , phys .
b * 655 * , 141 ( 2007 ) [ hep - ph/0610253 ] .
f. wang , w. wang and j. m. yang , europhys .
* 76 * , 388 ( 2006 ) [ hep - ph/0601018 ] . s. m. davidson and h. e. logan , phys . rev .
d * 80 * , 095008 ( 2009 ) [ arxiv:0906.3335 [ hep - ph ] ] ; s. gabriel , b. mukhopadhyaya , s. nandi and s. k. rai , phys .
b * 669 * , 180 ( 2008 ) [ arxiv:0804.1112 [ hep - ph ] ] .
s. m. davidson and h. e. logan , phys .
d * 82 * , 115031 ( 2010 ) [ arxiv:1009.4413 [ hep - ph ] ] .
g. c. branco , p. m. ferreira , l. lavoura , m. n. rebelo , m. sher and j. p. silva , phys .
rept . * 516 * , 1 ( 2012 ) [ arxiv:1106.0034 [ hep - ph ] ] and the referrences there in .
s. zhou , phys .
d * 84 * , 038701 ( 2011 ) [ arxiv:1106.3880 [ hep - ph ] ] ; m. sher and c. triola , phys .
d * 83 * , 117702 ( 2011 ) [ arxiv:1105.4844 [ hep - ph ] ] .
r. n. mohapatra and g. senjanovic , phys .
* 44 * , 912 ( 1980 ) and the referrences there in .
z. maki , m. nakagawa and s. sakata , prog .
* 28 * , 870 ( 1962 ) ; b. pontecorvo , sov .
jetp * 26 * , 984 ( 1968 ) [ zh .
* 53 * , 1717 ( 1967 ) ] .
g. l. fogli , e. lisi , a. marrone , d. montanino , a. palazzo and a. m. rotunno , phys .
d * 86 * , 013012 ( 2012 ) [ arxiv:1205.5254 [ hep - ph ] ] .
g. steigman , k. a. olive and d. n. schramm , phys .
lett . * 43 * , 239 ( 1979 ) ; k. a. olive , g. steigman and t. p. walker , phys .
rept . *
333 * , 389 ( 2000 ) [ astro - ph/9905320 ] .
p. a. r. ade _ et al .
_ [ planck collaboration ] , arxiv:1303.5062 [ astro-ph.co ] ; p. a. r. ade _ et al . _
[ planck collaboration ] , arxiv:1303.5076 [ astro-ph.co ] ; c. l. bennett _ et al . _
[ wmap collaboration ] , astrophys .
j. suppl .
* 208 * , 20 ( 2013 ) [ arxiv:1212.5225 [ astro-ph.co ] ] .
atlas collaboration conference note [ atlas - conf-2013 - 011 ] .
cms collaboration [ cms - pas - hig-13 - 028 ] .
j. beringer et al .
( particle data group ) , phys .
d86 , 010001 ( 2012 ) .
j. pumplin , d. r. stump , j. huston , h. l. lai , p. m. nadolsky and w. k. tung , jhep * 0207 * , 012 ( 2002 ) [ hep - ph/0201195 ] .
j. alwall , m. herquet , f. maltoni , o. mattelaer and t. stelzer , jhep * 1106 * , 128 ( 2011 ) [ arxiv:1106.0522 [ hep - ph ] ] . n. d. christensen and c. duhr , comput . phys .
commun . * 180 * , 1614 ( 2009 ) [ arxiv:0806.4194 [ hep - ph ] ] .
t. sjostrand , s. mrenna and p. z. skands , comput .
commun .
* 178 * , 852 ( 2008 ) [ arxiv:0710.3820 [ hep - ph ] ] .
g. aad _ et al . _ [ atlas collaboration ] , phys .
d * 87 * , 112001 ( 2013 ) [ arxiv:1210.2979 [ hep - ex ] ] .
g. aad _ et al . _
[ atlas collaboration ] , arxiv:0901.0512 [ hep - ex ] ; bruce mellado , private communication . | we study the signals for a `` fermiophobic '' charged higgs boson present in an extension of the standard model with an additional higgs doublet and right handed neutrinos , responsible for generating dirac - type neutrino masses .
we study the pair production of the charged higgs at the large hadron collider ( lhc ) , which can be relatively light and still allowed by experimental data .
the charged higgs decays dominantly into a @xmath0 boson and a very light neutral scalar present in the model , which decays invisibly and passes undetected .
we find that the signal for such a charged higgs is overwhelmed by the standard model background and will prove elusive at the 8 tev run of the lhc .
we present a cut - flow based analysis to pinpoint a search strategy at the 14 tev run of the lhc which can achieve a signal significance of 5@xmath1 for a given mass range of the charged higgs . |
recent proposals in theoretical physics have suggested that our four - dimensional universe could lie on a brane embedded in a higher - dimensional space - time @xcite .
one of the most studied is the randall sundrum ( rs ) model @xcite where our four - dimensional universe is a brane embedded in five - dimensional anti - de sitter space - time ( ads@xmath0 ) .
indeed it is possible in this framework to have only one brane in a non - compact space and this is the scenario we will consider in this letter .
this brane - world perspective may dramatically change our picture of the early universe . in this model , the friedmann equation
is modified @xcite to @xmath1 so that the hubble parameter @xmath2 at high energy - density .
bulk gravity is strongly coupled to the dynamics on the brane at high energy , which changes the evolution of cosmological perturbations @xcite .
thus , it is important to establish whether the evolution of the perturbations generated during inflation could be significantly different from standard four - dimensional slow - roll inflation . in a recent paper
@xcite we calculated these effects for the simplest realization of slow - roll inflation on a brane .
the purpose of this letter is to illustrate the effect of these brane - world corrections on the spectral index of density perturbations created during inflation in some specific models . the simplest way to realize
inflation in the rs model is to have a single scalar field , the inflaton , confined to the brane @xcite and only gravity in the bulk .
the amplitude of the resulting scalar curvature perturbations is given by @xmath3 where @xmath4 is the time derivative of the inflaton @xmath5 and @xmath6 is the inflaton fluctuation on a spatially flat hypersurface .
if one makes the assumption that back - reaction due to metric perturbations in the bulk can be neglected , the quantum expectation value of the inflaton fluctuations on super - horizon scales in the de sitter space - time is @xmath7 .
it has been shown that the curvature perturbation @xmath8 on comoving hypersurfaces is constant outside the horizon in this model @xcite and , indeed , independently of the gravitational theory so long as energy - momentum is conserved @xcite .
one can then use eq .
( [ curvature0 ] ) as initial conditions to calculate observables like cosmic microwave background ( cmb ) temperature anisotropy .
although the formula ( [ curvature0 ] ) for the curvature perturbation is exactly the same as in the four - dimensional case , the relation between the hubble parameter , @xmath9 , and the scalar field potential , @xmath10 , is different @xcite due to the modified friedmann equation ( [ modfriedmann ] ) .
this changes the prediction of the spectrum of scalar perturbations for a given potential and observational predictions of brane - world inflation have been made using this approach @xcite .
a crucial assumption is that back - reaction due to metric perturbations in the bulk can be neglected . in the extreme slow - roll limit this is necessarily correct because the coupling between inflaton fluctuations and metric perturbations vanishes ; however , this is not necessarily the case when slow - roll corrections are included in the calculation .
previous work @xcite has shown that such bulk effects can be subtle and interesting ( see also @xcite for other approaches ) .
in particular , sub - horizon inflaton fluctuations on a brane excite an infinite ladder of kaluza
klein modes of the bulk metric perturbations at first order in slow - roll parameters , and a naive slow - roll expansion breaks down in the high - energy regime once one takes into account the back - reaction of the bulk metric perturbations , as confirmed by direct numerical simulations @xcite .
however , an order - one correction to the behaviour of inflaton fluctuations on sub - horizon scales does not necessarily imply that the amplitude of the inflaton perturbations receives corrections of order one on large scales ; one must consistently quantise the coupled brane inflaton fluctuations and bulk metric perturbations .
this requires a detailed analysis of the coupled brane - bulk system @xcite .
the rs model is a brane - world model where the background spacetime is ads@xmath0 and matter fields are confined to a 3-brane .
we will be considering the single - brane model first proposed in ref .
@xcite . the brane - world inflation model under consideration
has the action @xmath11 where @xmath12 is the trace of the extrinsic curvature of the boundary @xmath13 , the ads curvature scale is defined by @xmath14 and the tension @xmath15 is tuned to make the brane flat if it has no other energy density .
the brane is our four - dimensional universe and , because we are interested in studying early - universe inflation , we will make the slow - roll approximation , where the induced metric is assumed to be de sitter but with a slowly changing hubble parameter
. we will describe this in coordinates where the background line element has the form @xmath16\,,\ ] ] with the function @xmath17 given by @xmath18 @xcite . here , @xmath9 is the hubble parameter , which is constant in a pure de sitter universe , and the coordinate @xmath19 is conformal time on the brane .
the brane is located at @xmath20 , which would be exactly constant for a pure de sitter universe . during inflation
the universe is approximately de sitter in that the hubble parameter @xmath9 changes slowly as the field rolls slowly down the potential .
this slow - rolling behaviour is usually expressed in terms of slow - roll parameters defined as @xmath21 where the overdot denotes derivatives with respect to proper time on the brane , not the conformal time @xmath19 .
we will use the convention that @xmath19 is positive , so that early times corresponds to @xmath22 and late times to @xmath23 .
the equations of motion for the scalar metric perturbations are most conveniently expressed in terms of a master variable , a technique first used in the study of brane - world perturbations in ref .
we define a master variable @xmath24 , whose equation of motion is @xcite @xmath25 where @xmath26 note that eq .
( [ eom : chi ] ) is separable .
the other equations are simplified by defining a new variable , @xmath27 , describing the scalar field and metric perturbations on the brane , and given by @xmath28 in terms of the field @xmath24 evaluated on the brane and the curvature perturbation @xmath29 on comoving hypersurfaces .
this new variable obeys the following equation of motion @xcite @xmath30 and the junction condition at the brane is @xmath31_{_{\rm b } } = \kappa_5 ^ 2 \dot\phi h \tau \xi \,.\ ] ] here we have neglected all terms suppressed by the slow - roll parameters except for the terms responsible for the coupling between @xmath27 and @xmath24 , i.e. , the terms that induce the mixing between the brane and bulk modes .
it is useful to define the time - independent wronskian of two states @xmath32 and @xmath33 by @xcite @xmath34 which is conserved . at late times ,
@xmath35 , a growing bound state solution is @xcite @xmath36 where @xmath37 is the overall amplitude , while @xmath38 can be determined through @xmath37 .
a decaying bound state is given by @xcite @xmath39\ , , \nonumber\end{aligned}\ ] ] where @xmath40 is linearly related to @xmath41 .
at early times ( and high energies ) , @xmath42 , the growing and decaying solutions both take the form @xcite @xmath43 the growing and decaying solutions have a different amplitude and phase , with the phase difference being approximately equal to @xmath44 .
we see that the amplitude of @xmath45 increases as @xmath46 while the amplitude of @xmath27 tends to a constant at large @xmath19 , becoming increasingly dominated by higher modes and more closely bound to the brane .
the early - time solution ( [ earlytimesolution ] ) differs from the solution for the scalar field in 4d at leading order , where @xmath27 satsifies the equation of motion @xmath47 and has solutions at late times given by @xmath48 where @xmath49 , and at early times by @xmath50 where @xmath51 and @xmath52 for standard 4d inflation .
one might expect that the @xmath53 difference between the early - time solutions would lead to a significant difference to the amplitude of quantum perturbations generated during inflation .
in fact this does not happen .
we now illustrate how the quantum effects can be calculated and use the result to determine the slow - roll corrections to the spectral indices of scalar perturbations for some monomial potentials .
in ref . @xcite we calculated the growing and decaying mode solutions ; these were found to differ significantly from the solution in the standard 4d inflationary paradigm . however , the amplitude of large - scale perturbations from the quantum vacuum state in the early universe is only affected at the first order in the slow - roll parameters .
we follow the method developed in refs .
@xcite to relate the amplitude of the growing mode solution to the creation and annihilation operators , which is summarized here . in principle , this method involves the construction of properly normalised positive- and negative - frequency solutions at early times and the evaluation of the inner product ( wronskian ) between these solutions and the bound states . however , there is a short - cut : all we need are the values of the real coefficients of the positive and negative frequency parts .
we first explain this short - cut by making use of the familiar four - dimensional example , and then proceed to the brane - world model .
let us begin with the standard four - dimensional case .
the wronskian is defined as @xmath54 we are interested in the quantum field that behaves at late times as @xmath55 , where @xmath56 is a time - independent quantum operator .
this field is quantised on small scales where we can expand @xmath57 in terms of negative- and positive - frequency modes , @xmath58 where @xmath59 and @xmath60 are annihilation and creation operators , respectively , which define the vacuum @xmath61 .
note that because @xmath19 decreases as proper time increases , the negative- and positive - frequency functions are @xmath62 and @xmath63 .
the mode functions should be normalised as @xmath64 to ensure the canonical commutational relation between @xmath65 and its conjugate momentum .
we use the constancy of the wronskian to express @xmath56 in terms of @xmath59 and @xmath60 . at late times , @xmath66 where the numerator can be calculated using the early - time solutions and the denominator using the late - time solutions ( [ late4dxi ] ) as @xmath67 up to the first order in slow - roll parameters .
we expand the growing mode and the decaying mode solutions at early times by @xmath68 this allows us to evaluate the expectation value of the square of the operator @xmath56 as @xmath69 which fully characterizes the gaussian random field .
thus , the problem reduces to finding @xmath70 , the expansion coefficient for the decaying mode solution .
the positive- and negative - frequency modes have the form @xmath71 , where the normalisation factor @xmath72 is not needed here , and @xmath73 is some phase ( which is irrelevant in the four - dimensional case ) . comparing the two expressions ( [ xi4d ] ) and ( [ xisplit ] ) , and using the wronskian @xmath74 , @xmath75 is determined to be @xmath76 thus , the quantity @xmath75 entering ( [ fin4d ] ) is expressed in terms of the ratio @xmath77 of the amplitudes and difference @xmath78 of the phases of the decaying and growing modes evolved back in time .
these are evaluated analytically as @xmath79 and @xmath80 , giving @xmath81 the quantity that is related to observables is the comoving curvature perturbation @xmath29 , which is defined as @xmath82 \left ( \frac{\phi_{,\tau}}{h } { \cal r}_c \right).\ ] ] thus , one finds the power spectrum of the curvature perturbation to be @xmath83 \left.\frac{h^2}{2\pi\dot\phi}\right|_{k = ah } \,,\ ] ] where @xmath84 and @xmath85 , using the fact that conformal time is @xmath86 up to the first order in slow - roll parameters .
the reason for preferring this method for dealing with the brane - world case is that one does not need to know the properly - normalised positive- and negative - frequency modes at early times , merely the quantities @xmath77 and @xmath87 . for the brane - world
a similar calculation presented in our previous work @xcite allow us to include the effect of interactions between the brane and bulk . the wronskian method , using the formula in eq .
( [ wronskian ] ) , allows us to deal with complicated early - time behaviour .
the expectation value of the square of the operator @xmath56 is obtained as @xmath88 \,,\ ] ] where @xmath89 is determined by numerical solution .
the left panel of fig . 1 shows the dependence of k on the energy scale of the inflation @xmath90 . at low energies , @xmath91 ,
the correction @xmath12 is negligible and we recover the standard vacuum ( [ zz ] ) . in the high - energy limit , @xmath92
, the correction can be approximated as a linear function of @xmath93 , which we write as @xmath94 .
the right panel of fig .
1 shows @xmath95 in the high energy limit .
this modifies eq .
( [ powerspectrum ] ) for the power spectrum by changing the coefficient @xmath96 to @xmath97 while @xmath98 .
one of the strengths of the slow - roll inflation paradigm is that the perturbations generated by quantum fluctuations have a power spectrum that is almost scale - invariant . in 4d inflation
the first - order correction to the tilt , @xmath99 , is well known to be @xmath100 .
the second - order correction was calculated by stewart and lyth @xcite .
this was extended to high energy inflation in the rs model in ref .
@xcite , including the modified effect of the friedmann equation ( [ modfriedmann ] ) in eq .
( [ powerspectrum ] ) but without including bulk gravity effects on the initial vacuum state .
this gives @xmath101 where @xmath102 .
the actual values of @xmath93 and @xmath103 for any given potential will change .
alternative slow - roll parameters @xmath104 and @xmath105 are defined in terms of the potential by @xmath106 for brane - world inflation in the high - energy limit and @xmath107 for standard 4d inflation .
these are related to the slow - roll parameters given in eq.([slowrolldef ] ) by @xmath108 and @xmath109 .
note that the different friedmann relation means that the field , @xmath5 will have a different value at horizon crossing .
the calculations presented in the preceding section show that we can determine the spectral index to second order for brane - worlds , including bulk gravity effects at high energies as well as the modified friedmann equation , by using the formula given in eq .
( [ powerspectrum ] ) , replacing @xmath96 with @xmath110 as given in eq .
( [ modifiedc1 ] ) .
tables [ table : monomialpotentials1 ] and [ table : monomialpotentials2 ] show the slow - roll parameters and spectral indices for some monomial potentials .
the differences at first order arise from the modified evolution on the brane , not from the brane - bulk mixing .
the second - order corrections in brane - worlds have a contribution from the brane - bulk mixing @xcite in addition to the familiar stewart lyth correction , and these are of the same order of magnitude .
for both the standard and brane - world inflation models the second - order correction is less than 1 part in @xmath111 so the first - order result should be sufficient for observational cosmology .
.showing the values of the slow - roll parameters @xmath104 and @xmath105 for three monomial potentials @xmath112 and 50 e - folds of inflation . [
cols="^,^,^,^,^,^ " , ] in the standard inflation paradigm slow - roll corrections change the dynamics on sub - horizon scales by introducing an effective mass that is negligible at early times ( though this is not negligible on super - horizon scales ) .
the problem thus reduces to quantizing a massless scalar field in de sitter space .
by contrast , in the brane - world model it is essential to use the slow - roll approximation even to formulate the problem in a way where the bulk wave equation is separable . in the brane - world model
we have shown that the coupling to bulk metric perturbations can not be ignored in the equations of motion .
indeed , we have shown that there are order - unity differences between the classical solutions without coupling and with slow - roll induced coupling . however , the change in the amplitude of quantum - generated perturbations is at next - to - leading order @xcite because there is still no mixing at leading order between positive and negative frequencies when scales observable today crossed the horizon , so the bogoliubov coefficients receive no corrections at leading order .
the amplitude of perturbations generated is also subject to the usual slow - roll corrections on super - horizon scales .
thus , the correction we calculate is in addition to the usual stewart lyth correction @xcite and we find it to be of the same order of magnitude .
our results also show that the ratio of tensor - to - scalar perturbation amplitudes are not influenced by brane - bulk interactions at leading order in slow - roll .
it is not possible for us to calculate the next - order corrections without also deriving the slow - roll correction to tensor amplitudes .
this correction has not yet been calculated because it requires one to go beyond the approximation of a separable wave equation in the bulk for a de sitter brane . from an end - user s perspective
our results are significant in establishing rigorously the validity of assumptions made in earlier brane - world literature .
the most significant difference between the prediction for the spectral indices in the brane - world at high energy and in standard inflation with the same potential arises at first order in slow - roll from the modified evolution of the hubble parameter and scalar field , leading to different values of the slow - roll parameters when the number of e - foldings of inflation is fixed @xcite .
it is remarkable that the predictions from inflation theories should be so robust that this result hold in spite of the leading - order change to the solutions of the classical equations of motion .
the authors are supported by stfc .
a. cardoso , t. hiramatsu , k. koyama and s. s. seahra , arxiv:0705.1685 [ astro - ph ] and references therein .
k. koyama , a. mennim , v. a. rubakov , d. wands and t. hiramatsu , [ arxiv : hep - th/0701241 ] .
r. maartens , d. wands , b. a. bassett and i. heard , phys .
d * 62 * , 041301 ( 2000 ) [ arxiv : hep - ph/9912464 ] .
d. langlois , r. maartens , m. sasaki and d. wands , phys .
d * 63 * , 084009 ( 2001 ) [ arxiv : hep - th/0012044 ] .
d. wands , k. a. malik , d. h. lyth and a. r. liddle , phys .
d * 62 * , 043527 ( 2000 ) [ arxiv : astro - ph/0003278 ] .
a. liddle and a. n. taylor , phys .
d * 65 * , 041301 ( 2002 ) ; a. r. liddle and a. j. smith , phys .
d * 68 * ( 2003 ) 061301 ; s. tsujikawa and a. r. liddle , jcap * 0403 * ( 2004 ) 001 .
d. seery and a. taylor , phys .
d * 71 * , 063508 ( 2005 ) [ arxiv : astro - ph/0309512 ] .
e. ramrez and a. r. liddle phys . rev .
d * 69 * 083522 ( 2004 ) k. koyama , d. langlois , r. maartens and d. wands , jcap * 0411 * , 002 ( 2004 ) [ arxiv : hep - th/0408222 ] .
k. koyama , a. mennim and d. wands , phys .
d * 72 * ( 2005 ) 064001 [ arxiv : hep - th/0504201 ] .
k. koyama , s. mizuno and d. wands , jcap * 0508 * ( 2005 ) 009 [ arxiv : hep - th/0506102 ] . c. de rham , phys .
d * 71 * , 024015 ( 2005 ) [ arxiv : hep - th/0411021 ] ; c. de rham and s. watson , arxiv : hep - th/0702048 . t. hiramatsu and k. koyama , arxiv : hep - th/0607068 .
k. koyama , a. mennim and d. wands , phys .
d * 72 * , 064001 ( 2005 ) [ arxiv : hep - th/0504201 ] ; a. cardoso , k. koyama , a. mennim , s. s. seahra and d. wands , phys .
d * 75 * , 084002 ( 2007 ) [ arxiv : hep - th/0612202 ] . s. mukohyama , phys .
d62 : 084015 ( 2000 ) [ arxiv : hep - th/0004067 ] d. s. gorbunov , v. a. rubakov and s. m. sibiryakov , jhep * 0110 * , 015 ( 2001 ) [ arxiv : hep - th/0108017 ] ; t. kobayashi and t. tanaka , phys .
d * 71 * , 124028 ( 2005 ) [ arxiv : hep - th/0505065 ] ; t. kobayashi and t. tanaka , phys .
d * 73 * , 044005 ( 2006 ) [ arxiv : hep - th/0511186 ] ; t. kobayashi , phys .
d * 73 * , 124031 ( 2006 ) [ arxiv : hep - th/0602168 ] . m. v. libanov and v. a. rubakov , phys .
d * 72 * , 123503 ( 2005 ) [ arxiv : hep - ph/0509148 ] . | we quantify the slow - roll corrections to primordial density perturbations arising from inflation driven by a four - dimensional scalar field with a monomial potential in a five - dimensional non - compact bulk spacetime .
although the difference between the classical brane - world solutions and standard four - dimensional solutions is large at early times , the change to the amplitude at late times of perturbations generated from quantum fluctuations is first - order in slow - roll parameters , leading to second - order slow - roll corrections to the spectral index .
this confirms that the leading - order effects are correctly given by previous work in the literature . |
the formation mechanism of jets is a key issue in the study of the magnetospheric structure of compact astrophysical objects .
indeed , jets are observed in most compact sources , ranging from active galactic nuclei ( agns ) , quasars , and radio galaxies @xcite to accreting neutron stars , solar - mass black holes ( ss 433 , x - ray novae ) @xcite , and young stellar objects @xcite .
moreover , jets have also been recently discovered in young radio pulsars @xcite . at the same time , in most studies devoted to the magnetohydrodynamic ( mhd ) model of such objects @xcite , in which the formation of jets is coupled with the attraction of longitudinal currents flowing in the magnetosphere , the attention was focused on intrinsic collimation in the sense that the effect of the external medium was assumed to be marginal .
however , such a situation is possible only for a nonzero total current @xmath2 flowing within the jet @xcite , so the question of its closure in the outer parts of the magnetosphere arises . on the other hand , the longitudinal current
is often constrained by the regularity condition at the fast magnetosonic surface , which by no means always leads to the sufficiently large longitudinal currents required for collimation @xcite . in other words , with the exception of the force - free case @xcite , as yet no working model of a jet in which , on the one hand , the total electric current would be zero and , on the other hand , the total magnetic flux @xmath3 in the jet would be finite , has been constructed .
however , the force - free approximation , in which , by definition , the particle energy density is disregarded , does not allow the fraction of energy transferred by the outflowing plasma to be determined .
at the same time , the question of collimation can not be solved in isolation from the external conditions ( see , e.g. , @xcite ) . in particular , this is clear even from a popular example of the magnetosphere of a compact object with a monopole magnetic field , because for any arbitrarily weak external regular magnetic field , the monopole solution ( for which the magnetic field falls off as @xmath4 ) can not be extended to infinity . moreover , as is well known from an example of moving cosmic bodies , such as jupiter s moons @xcite or artificial earth satellites @xcite , as well as radio pulsars @xcite , the external magnetic field can serve as an effective transfer link , which occasionally determines the general energy losses of the system .
for this reason , constructing a consistent magnetospheric model for compact objects immersed in an external magnetic field is , in our view , of undeniable interest , especially since , as was noted above , such a jet model was previously constructed in the force - free approximation @xcite . undoubtedly , the existence of an external regular magnetic field in the vicinity of compact objects is largely open to question .
the regular magnetic field in our galaxy , i.e. , the field that is constant on scales comparable to the sizes of our galaxy , is known to be @xmath5 and essentially matches the random magnetic - field component , which varies even on scales of several parsecs @xcite .
however , if the collimation is assumed to be actually produced by an external magnetic field , it becomes possible to estimate the jet radius .
indeed , assuming the magnetic field in the jet to be similar to the external magnetic field ( [ 1 ] ) , we obtain from the condition for the conservation of magnetic flux @xmath6 where @xmath7 and @xmath8 are the radius and magnetic field of the compact object , respectively .
for example , for agns ( @xmath9 , @xmath10 ) , we have @xmath11 which corresponds to the observed jet radii @xcite
. one might expect such a picture to be also preserved for an external medium with pressure @xmath12 ; therefore , it seems of interest to consider the internal structure of a one - dimensional jet immersed in an external uniform magnetic field .
however , a discussion of the more realistic case of a medium with pressure is beyond the scope of this study . nor
do we discuss the collimation itself but only consider the internal structure of observed one - dimensional jets .
this issue has become particularly urgent because of the new possibilities offered by space radio interferometry , which enables the internal structure of such jets to be resolved .
the effect of an external medium on the internal structure of relativistic jets in the mhd model discussed here was previously studied only by appl and camenzind @xcite .
they considered only a special case with a constant angular velocity of the plasma , in which the solution with a zero total electric field flowing inside the jet could not be constructed . as we show below
, it is for the case of an angular velocity decreasing toward the jet periphery ( which , incidentally , is typical of all models with a magnetic field passing through the accretion disk ) that the solution with finite magnetic flux @xmath3 and zero total current @xmath13 can be constructed .
on the other hand , many authors @xcite obtained a universal solution with a central core for a cylindrical jet : @xmath14 where , in the relativistic case , @xmath15 is the size of the central core , @xmath16 is the angular velocity of the compact object , and @xmath17 is the characteristic lorentz factor of the outflowing plasma . as can be easily seen , such a solution results in a rapid falloff of the poloidal field @xmath18 far from the rotation axis @xmath19 .
however , this solution is in conflict with the force - free approximation , in which the poloidal magnetic field remains essentially constant @xcite .
indeed , when the energy density of the electromagnetic field exceeds appreciably the plasma energy density ( and it is this case that was considered ) , it would be natural to assume that the internal jet structure must be similar to the force - free one .
the examples given above show that a more detailed study with allowance for all possible solutions is required even in the simplest case of a one - dimensional cylindrical jet considered in terms of ideal magnetohydrodynamics .
our study aims at a consistent investigation of this issue .
several features that we use when studying the structure of relativistic jets typical of agns and radio pulsars should be immediately noted .
first of all , the jet radius @xmath20 in all real cases proves to be considerably larger than the light - cylinder radius @xmath21 .
this implies that , when the internal structure of jets is investigated , the corresponding equations must be written in complete relativistic form . on the other hand ,
the gravitational forces can be disregarded in them far from the compact object . finally , for simplicity , we consider below a cold plasma , which is justifiable because the thermal processes in the magnetospheres of radio pulsars play no crucial role . as for the jets from agns , this approximation is applicable here in those magnetospheric regions in which the plasma density is low . in any case , this is true for the field lines passing through the surface of a black hole .
the one - dimensional solutions describing collimated jets are obtained in sect .
2 ; analytic and numerical solutions for the basic physical quantities characterizing the structure and physics of jets are given in sect .
the problem is solved in straightforward statement ; i.e. , all jets characteristics are determined by a set of parameters in the compact source and , most importantly , by the physical conditions in the external medium . as a result , we have found the conditions under which most of the energy in actual relativistic jets must be transferred by the electromagnetic field , while a region with subsonic flow exists in the central jet regions .
we also show that the solution with a central core ( [ 4 ] ) and ( [ 5 ] ) can not be realized in an external magnetic field . finally , some astrophysical implications of the theory developed for one - dimensional jets are discussed in sect .
let us consider the structure of a one - dimensional jet where all quantities depend only on radius @xmath22 ; in what follows , the temperature of the matter is assumed to be zero , and @xmath23 . as in the general axisymmetric case , it is convenient to describe the magnetic - field structure in terms of magnetic - flux function @xmath24 , which is related to the longitudinal magnetic field by b_z()=. [ 6 ] accordingly , it is convenient to write the toroidal magnetic field , the electric field , and the 4-velocity vector of the matter as b_()=- , [ 7 ] = - , [ 8 ] = + _ f , [ 9 ] where @xmath25 is the total current within @xmath26 . in the case of a cold plasma , at the cylindrical magnetic surfaces @xmath27 , four `` integrals of motion '' @xcite can be introduced , which should be considered precisely as functions of magnetic flux @xmath28 in the most general statement .
these are primarily @xmath29 and @xmath30 in the definitions ( [ 8 ] ) and ( [ 9 ] ) , as well as the @xmath31-component of angular momentum @xmath32 and the energy flux @xmath33 . here , @xmath34 is the relativistic specific enthalpy , which is equal to the mass of particles for a cold plasma .
the specific form of the integrals of motion must be determined from boundary conditions in the compact source and from critical conditions at the singular surfaces . as a result ,
the equilibrium equation for magnetic surfaces far from gravitating bodies ( grad shafranov s equation ) can be written as ( see , e.g. , @xcite ) ( ) + _ f()^2 + ( ) -=0 , [ 10 ] where @xmath35 @xmath36 @xmath37 here , @xmath38 is the square of the mach number with respect to the alfven velocity @xmath39 , and the derivative @xmath40 acts only on the integrals of motion .
the remaining jet parameters are given by the well - known algebraic relations ( see , e.g. , @xcite ) : @xmath41 equation ( [ 10 ] ) contains four integrals of motion ; this equation has no singularity at the fast magnetosonic surface , because it depends only on coordinate @xmath22 . as for the alfven surface , @xmath42 , the problem of the boundary conditions generally requires a further study beyond the theory of ideal magnetohydrodynamics . at the same time , for the fairly large currents @xmath43 considered here , a solution continuous across this surface can always be constructed by a small change in the integrals of motion near the alfven surface , @xmath44 .
consequently , equation ( [ 10 ] ) requires six boundary conditions .
these boundary conditions primarily include the external uniform magnetic field b_z(r_j)=b_ext , [ 14 ] and the regularity condition at the magnetic axis @xmath45 @xmath46 in addition , all four integrals @xmath47 , @xmath48 , @xmath49 , and @xmath50 must be specified . as for the remaining quantities characterizing the flow , such as the jet radius @xmath20 and the outflowing plasma energy ,
they must be determined as a solution of the problem formulated above .
similarly , the solution of the problem must also give an answer to the question of whether the flow in the jet is supersonic .
let us now consider the determination of the integrals of motion in more detail
. it would be natural to assume that , at the jet boundary where there is no longitudinal motion of the matter , all four integrals of motion become zero _
f(_0 ) = 0 , e(_0 ) = 0 , l(_0 ) = 0 , ( _ 0 ) = 0 .
[ 16 ] here , @xmath3 is the finite total magnetic flux concentrated in the jet .
this case corresponds to the absence of tangential discontinuities at the jet boundary ; according to ( [ 11 ] ) , the total electric current within the jet is automatically equal to zero .
we use the integrals of motion @xmath29 , @xmath51 , and @xmath52 derived by beskin _
et al . _
@xcite for the force - free magnetosphere of a black hole , which satisfy the conditions ( [ 16 ] ) and , consequently , can be directly used to study the jet structure .
the only but very important change here is the fact that , for a finite magnetization parameter @xmath53 @xcite , = , [ 17 ] which tends to infinity in the force - free approximation , the particle contribution must be added to the energy integral @xmath54 , because the energy flux of the electromagnetic field near the rotation axis must inevitably vanish .
as a result , we have with accuracy up to @xmath55 @xmath56 @xmath57 @xmath58 below , we assume , for simplicity , that @xmath59 we emphasize that @xmath60 in expression ( [ 20 ] ) has the meaning of the injection lorentz factor in the region of the compact object and it is not equal to the lorentz factor of the jet particles .
thus , we see from the formula for the energy flux @xmath54 that the contribution by the electromagnetic field becomes dominant only at @xmath61 , where @xmath62 at low values of @xmath28 , most of the energy is transferred by the relativistic particles ; as directly follows from relation ( [ 20 ] ) , their lorentz factor is constant and equal to their initial value @xmath60 . as for the integral @xmath30 , the particle - to - magnetic flux ratio , we chose it in the form ( ) = _ 0(1-/_0 ) , [ 23 ] which satisfies the condition ( [ 16 ] ) .
we emphasize that the very possibility of using the integrals of motion obtained by analyzing the inner magnetospheric regions , is not trivial . indeed ,
the flow outside the fast magnetosonic surface is completely determined by four boundary conditions at the surface of a rotating body . at the same time , a one - dimensional flow can be produced by the interaction with the external medium , which gives rise ( see , e.g. , @xcite ) to perturbations or shock waves propagating from `` acute angles '' and other irregularities .
therefore , in regions where the conditions for the validity of ideal magnetohydrodynamics are violated , a significant redistribution of energy @xmath48 and angular momentum @xmath49 is possible ( e.g. a part of them can be lost via radiation ) .
nevertheless , we assume here , for simplicity , that the integrals of motion @xmath54 and @xmath63 , functions of flux @xmath28 , remain exactly the same as those in the inner magnetospheric regions . in the one - dimensional case we consider ,
it is convenient to reduce the second - order equation ( [ 10 ] ) to a set of two first - order equations for @xmath24 and @xmath64 . multiplying equation ( [ 10 ] ) by @xmath65
, we obtain + a()^2 + ( ) -(^2 ^ 2)=0 , [ 24 ] with the derivative @xmath66 acting only on the integrals of motion . finally , we use `` bernoulli s relativistic equation '' @xmath67 , which , given the definitions of the integrals of motion @xmath54 and @xmath63 , can be written as a^2()^2= + - - , [ 25 ] where we introduced the dimensionless variables @xmath68 as a result , substituting the right - hand part of ( [ 25 ] ) into the first term of ( [ 24 ] ) and performing differentiation , we obtain the first first - order differential equation @xmath69\frac{dm^2}{dx}= \frac{m^6}{x^3a}\frac{\omega_{\rm f}^2(0)l^2}{\mu^2\eta^2}- \frac{xm^2}{a}\frac{\omega_{\rm f}^2}{\omega_{\rm f}^2(0 ) } \left(\frac{e^2}{\mu^2\eta^2}-2a\right ) \nonumber
\\ + \frac{m^2}{2}\frac{dy}{dx}\left [ \frac{1}{\mu^2\eta^2}\frac{de^2}{dy}+ \frac{x^2}{\omega_{\rm
f}^2(0)}\frac{d\omega_{\rm f}^2}{dy}- 2\left(1-\frac{\omega_{\rm f}^2}{\omega_{\rm f}^2(0)}x^2\right ) \frac{1}{\eta}\frac{d\eta}{dy}\right ] .
\label{28}\end{aligned}\ ] ] the second first - order differential equation is bernoulli s equation ( [ 25 ] ) , which should now be considered as an equation for the derivative @xmath70 .
the set of equations ( [ 25 ] ) and ( [ 28 ] ) allows a general solution to be constructed for a one - dimensional jet immersed in an external magnetic field .
we emphasize one important advantage of the set of first - order equations ( [ 25 ] ) and ( [ 28 ] ) over the initial second - order equation ( [ 10 ] ) .
the point is that the relativistic equation ( [ 10 ] ) , which is basically the force balance equation , contains the electromagnetic force _ em = _ e*e * + * j * , [ 29 ] in which the electric and magnetic contributions virtually cancel each other far out from the rotation axis @xmath71 . using bernoulli s equation ( [ 25 ] )
, we can derive @xcite ~. [ 30 ] when analyzing ( [ 10 ] ) , we therefore must retain all higher order terms @xmath72 , while the zero - order quantities @xmath73 and @xmath74 in ( [ 28 ] ) are analytically removed using bernoulli s equation , so all terms of this equation are of the same order . finally , it is also important that the exact equation ( [ 28 ] ) has no singularity near the rotation axis .
in other words , its solution contains no @xmath75-shaped current @xmath76 flowing along the jet axis ; several authors pointed out to the necessity of it @xcite .
let us consider the basic properties of the set of equations ( [ 25 ] ) and ( [ 28 ] ) .
as can be easily verified , in the relativistic case under consideration , we may assume @xmath77 with high accuracy .
far from the rotation axis , @xmath78 ( @xmath79 ) , equation ( [ 25 ] ) can be rewritten in the limit @xmath80 as @xmath81 or , equivalently , @xmath82 as we see , equation ( [ 31 ] ) does not contain @xmath83 at all and can therefore be integrated independently
. this must be the case , because equation ( [ 31 ] ) must coincide with the asymptotics of the force - free equation , which can be derived from ( [ 25 ] ) by going to the limit @xmath84 .
assuming now that @xmath85 in ( [ 32 ] ) , we obtain , in particular , for the jet radius @xmath86 consequently , the jet radius is determined by the limit of the @xmath87 ratio as @xmath88 . in particular , for @xmath54 and @xmath29 given by ( [ 18])([20 ] ) , we have @xmath89 so the limit ( [ 33 ] ) does actually exist . as a result
, we obtain @xmath90 which essentially coincides with estimate ( [ 2 ] ) .
this is no surprise , because we show below that equations ( [ 25 ] ) and ( [ 28 ] ) for the integrals of motion ( [ 18 ] ) , ( [ 20 ] ) , and ( [ 23 ] ) have a constant magnetic field as their solution over a wide range of @xmath22 .
let us now consider in more details the behavior of the solution in the inner jet region , where @xmath91 and , hence , the integrals of motion can be approximately written as @xmath92 with @xmath93 and @xmath94 . as a result , @xmath95 , and we can rewrite equations ( [ 25 ] ) and ( [ 28 ] ) as @xmath96 @xmath97 equations ( [ 39 ] ) and ( [ 40 ] ) describing the internal jet structure can be solved analytically .
it can be verified by direct substitution that we have the following asymptotics for @xmath98 : m^2(x ) = m_0 ^ 2 = const , [ 41 ] y(x ) = x^2 , [ 42 ] which correspond to a constant magnetic field b_z = b_z(0)= = b(r_l ) = const , [ 43 ] where @xmath99 . here , we assume that @xmath100 , which is the typical for jets from agns and radio pulsars .
the solution of ( [ 39 ] ) and ( [ 40 ] ) for @xmath101 , i.e. , at @xmath79 , depends on the relationship between @xmath60 and @xmath102 .
for example , at @xmath103 , when , according to ( [ 43 ] ) , the axial magnetic field is fairly weak , the total magnetic flux within @xmath104 @xmath105 can be written as @xmath106 where the flux @xmath107 is given by ( [ 22 ] ) .
we see that , if the condition @xmath103 is satisfied , then the total magnetic flux within @xmath104 is lower than @xmath107 ; so , outside this region , the particles also make the main contribution to @xmath54 as before , while the contribution of the electromagnetic field may be neglected . as a result , at @xmath79 , the solution of ( [ 39 ] ) and ( [ 40 ] ) has a quadratic rise of @xmath83 and a power - law falloff of the magnetic field @xcite : m^2(x ) = m_0 ^ 2x^2 , [ 46 ] @xmath108 consequently , the magnetic flux increases very slowly ( logarithmically ) with the distance from the rotation axis : @xmath109 such a behavior of the magnetic field , in turn , shows that the transition flux @xmath110 is reached exponentially far from the rotation axis , in conflict with the estimate ( [ 2 ] ) corresponding to the assumption of jet collimation .
we may thus conclude that an external constant magnetic field limits the mach number at the rotation axis above @xmath111 accordingly , as follows from ( [ 43 ] ) , the magnetic field at the rotation axis can not be weaker than @xmath112 if , however , the mach number at the rotation axis does not exceed @xmath60 ( i.e. if @xmath113 ) , then , as for the similar asymptotics @xmath114 , the solution of ( [ 39 ] ) and ( [ 40 ] ) for @xmath115 gives a constant magnetic field ( [ 43 ] ) , which corresponds to the solution y(x ) = x^2 . [
51 ] at the same time , in this case , we have only a linear increase in the square of the mach number m^2(x ) = m_0 ^ 2x^2 . [ 52 ] then , according to ( [ 27 ] ) and ( [ 51 ] ) , the jet radius can be written as @xmath116 which is equivalent to ( [ 35 ] ) [ and in agreement with ( [ 2 ] ) ] .
moreover , as can be easily verified , the constant magnetic field @xmath117 for the invariants ( [ 18])([20 ] ) proves to be an exact solution of ( [ 31 ] ) in the entire jet up to the jet boundary , @xmath118 . here
, we may therefore assume @xmath119 .
consequently , according to ( [ 43 ] ) , we obtain m_0 ^ 2 = .
[ 54 ] using relation ( [ 54 ] ) , we can also express all the remaining jet parameters in terms of the external magnetic field .
note that the absence of a declining solution @xmath18 [ see ( [ 4 ] ) ] is associated with the first term in the right - hand part of ( [ 28 ] ) proportional to @xmath120 .
this term , which changes appreciably the behavior of the solution , appears to be missed previously . as it was already emphasized above
, this is not surprising because the corresponding term in the second - order equation ( [ 10 ] ) is of high order and small . on the other hand , far from the rotation axis @xmath121 , equation ( [ 28 ] )
can be rewritten as + l^2 = 0 , [ 55 ] in which both terms are of the same order . neglecting the term proportional to @xmath120 , we arrive at the solution ( [ 46 ] ) , @xmath122 , for @xmath123 and @xmath124 .
the conservation of function h = [ 56 ] was first found by heyvaerts and norman @xcite for conical solutions , when all quantities depend only on spherical coordinate @xmath125 , but has also been repeatedly discussed when analyzing cylindrical flows .
however , as we see , @xmath126 is generally not conserved in the cylindrical geometry for relativistic jets . to be more precise , the second term in ( [ 55 ] )
turns out to be significant for all models with a nearly constant density of the longitudinal electric current in the central jet region , where the invariant @xmath63 linearly increases with magnetic flux @xmath28 if @xmath127 .
thus , we conclude that the solution with a central core ( [ 4 ] ) can not be realized in the presence of an external medium with a finite regular magnetic field .
this conclusion appears to be also valid in the presence of a medium with finite pressure @xmath128 .
indeed , since the magnetic flux ( [ 48 ] ) increases very slowly ( logarithmically ) , the solution ( [ 4 ] ) yields an exponentially large jet radius @xmath129 .
accordingly , the magnetic energy density must also be low at @xmath130 .
however , this configuration can not exist in the presence of an external medium with finite pressure @xmath128 , irrespective of whether it is produced by a magnetic field or by a plasma .
we may therefore conclude that the solutions with a central core can be realized only for a special choice of the integral @xmath63 , which increases only slightly with the magnetic flux , and only in the absence of an external medium .
for the most natural ( from our point of view ) models with a constant current density in the central jet regions , the solution with a central core can not be realized even in the absence of an external medium . in order to derive now the energy distribution in the jet and the particle lorentz factor , it is convenient to introduce the quantity g(x)=. [ 57 ] since at large distances @xmath79 , according to ( [ 12 ] ) , we have = , [ 58 ] @xmath131 is simply the ratio of the energy flux transferred by particles @xmath132 to the energy flux of the electromagnetic field . as a result , from relation (
[ 52 ] ) for @xmath101 , we obtain x^-11 , [ 59 ] accordingly , from ( [ 57 ] ) and ( [ 58 ] ) for the particle lorentz factor at @xmath101 , we derive @xmath133 finally , expression ( [ 13 ] ) for the @xmath134-component of the 4-velocity vector @xmath135 yields the following toroidal velocity @xmath136 at @xmath78 : @xmath137 we see that the particle energy approaches the universal asymptotic limit ( [ 60 ] ) at @xmath79 .
naturally , such a simple asymptotics can also be derived from simpler considerations . indeed , using the `` frozen - in '' equation @xmath138
, we obtain for the drift velocity @xmath139 in our case , however , according to ( [ 6 ] ) and ( [ 7 ] ) , we have @xmath140 @xmath141 as a result , relations ( [ 62])([64 ] ) immediately lead to the exact asymptotics ( [ 60 ] ) .
it thus follows that , for example , for electron positron jets from agns , the jet particle energy is typically @xmath142 on the other hand , according to ( [ 57 ] ) , we reach a very important conclusion that , far from the light cylinder , @xmath143 , @xmath144 consequently , according to ( [ 58 ] ) , the particle contribution to the general energy flux balance proves to be minor .
for example , at @xmath145 for @xmath146 , we have @xmath147 while in the general case , we obtain @xmath148^{1/2}. \label{68}\ ] ] thus , we reach a fairly nontrivial conclusion that the fraction of energy transferred by particles in a one - dimensional jet must be determined by the parameters of the external medium .
let us now discuss the results of exact calculations obtained by numerical integration of equations ( [ 25 ] ) and ( [ 28 ] ) with the integrals of motion ( [ 18])([20 ] ) and ( [ 23 ] ) . in figures [ fig_1](a ) and [ fig_1](b ) , the mach number and the energy flux concentrated in particles @xmath149
are plotted against @xmath150 for @xmath151 , @xmath152 , and @xmath153 .
the dashed lines indicate the behavior of these quantities that follows from the analytical asymptotics ( [ 52 ] ) and ( [ 60 ] ) . as we see , at sufficiently small @xmath154 when the integrals of motion ( [ 18])([20 ] ) and ( [ 23 ] ) are similar to ( [ 36])([38 ] ) , the analytical asymptotics match the exact numerical results . on the other hand ,
as expected , @xmath149 and @xmath83 are zero at @xmath155 , i.e. , at the jet edge .
figure [ fig_1](c ) shows the dependence of the poloidal field , @xmath156 , for the inner parts of the jet , @xmath157 .
we see from this figure that the magnetic field is nearly constant at @xmath158 , in agreement with the analytic estimates ( [ 43 ] ) and ( [ 51 ] ) . of course , in general , the structure of the poloidal magnetic field is determined by a specific choice of the integrals @xmath54 and @xmath63 . in conclusion
, it is of interest to compare the energy of jet particle with the limiting energy acquired by the particles as they outflow from the magnetosphere of a compact object with a monopole magnetic field . according to calculations by beskin et al .
@xcite , the particle lorentz factor outside the fast magnetosonic surface @xmath159 in the absence of an external medium can be written as @xmath160 where @xmath161 is given by ( [ 27 ] ) .
on the other hand , relations ( [ 51 ] ) and ( [ 60 ] ) for the jet yield @xmath162 as shown in figure [ fig_2](a ) , for @xmath163 , i.e. , for @xmath164 , where @xmath165 the lorentz factor of the jet particles ( [ 71 ] ) is always larger than the lorentz factor acquired by the particles as they outflow from a magnetosphere with a monopole magnetic field , but , of course , is always smaller than the critical lorentz factor @xmath166 which corresponds to the complete transformation of the electromagnetic energy into the particle energy .
this implies that , at @xmath164 , the particles must be additionally accelerated during the collimation coupled with the interaction of the outflowing plasma with the external medium . if , alternatively , @xmath167 , then , in the inner jet regions , at @xmath168 the particle energy on a given field line turns out to be even lower than that for a monopole magnetic field , as shown in figure [ fig_2](b ) .
the latter result can be easily explained .
indeed , for the integrals of motion ( [ 36])([38 ] ) we consider , the factor @xmath169 , whose zero value determines the location of the fast mhd surface ( see @xcite for more detail ) , can be rewritten in the case of a cold plasma as @xmath170 it is easy to show that expression ( [ 76 ] ) for @xmath161 and @xmath83 given by ( [ 51 ] ) and ( [ 52 ] ) is negative at @xmath171 i.e. , at @xmath154 corresponding to ( [ 75 ] ) .
consequently , we may reach another important conclusion that , for sufficiently strong external magnetic fields @xmath167 ( [ 73 ] ) when @xmath172 , a region with a subsonic flow inevitably emerges in the inner jet regions @xmath173 , where @xmath174^{3/2}r_{\rm l}. \label{78}\ ] ] at the same time , a region with the subsonic flow can be produced far from the compact object either by a shock wave or by a strong distortion of the magnetic field within the fast magnetosonic surface located in the vicinity of the compact object . in both cases ,
the magnetic - field perturbation causes the particle energy to decrease .
thus , we conclude that the exact equilibrium equations ( [ 25 ] ) and ( [ 28 ] ) do actually allow a construction of a self - consistent model for a jet immersed in an external uniform magnetic field .
the advantage of these equations over equation ( [ 10 ] ) results from the fact that all terms in ( [ 28 ] ) are of the same order . in this case , the uniformity of the poloidal magnetic field within the jet ( [ 43 ] ) results from the choice of integrals ( [ 36])([38 ] ) . in general , the poloidal magnetic field depends on the specific form of the integrals .
a full analysis of the possible solutions is beyond the scope of this paper .
we have shown that the fraction of energy transferred by particles @xmath175 must be largely determined by the parameters of the external medium . in case @xmath176 , where @xmath177^{1/2 } , \label{79}\ ] ] the energy transferred by particles is only a small fraction of the energy flux @xmath178 transferred by the electromagnetic field .
consequently , the jet is strongly magnetized ( @xmath179 ) only at sufficiently large @xmath53 .
if , however , the magnetization parameter does not exceed @xmath180 , then , in this case , an appreciable part of the energy in the jet is transferred by particles .
this , in turn , implies that a considerable part of the energy must be transferred from the electromagnetic field to the plasma particles during jet collimation .
it is interesting that @xmath180 turns out to be approximately the same both for agns and for fast radio pulsars : _ cr 10 ^ 5 - 10 ^ 6 .
[ 80 ] we have shown that the central part of the jet must be subsonic for sufficiently strong external magnetic fields . thus , the theory gives direct predictions whose validity can be verified by observations .
it should also be noted that the results obtained above are applicable both to electron positron and to electron proton jets .
however , a direct evidence that the jets in agns are actually electron
positron ones has recently appeared @xcite . in our view , an important result is that , if the external regular magnetic field is taken into account , the mhd equations allow a self - consistent model to be constructed for a jet with a zero total longitudinal electric current , @xmath181 . in this case , a uniform magnetic field that matches the external magnetic field can also be a solution for the inner jet regions . as was already emphasized above , the radii of the jets from agns can thus be also explained in a natural way .
in addition , since only a small fraction of the electromagnetic - field energy is transformed into the particle energy , the energy transfer from the compact object in the region of energy release can be explained as well .
at the same time , extending the mhd solution to the jet region requires very high particle energies ( @xmath182 ) , which have not been recorded yet .
however , a consistent discussion of the outflowing - plasma energy requires a proper allowance for the particle interaction with the surrounding medium ( for example , with background radiation ) , which may cause a significant change in particle energy . as for the quantitative predictions about the real physical parameters of jets , they , as we showed above , essentially depend only on the following three quantities : the magnetization parameter @xmath53 ( [ 17 ] ) , the lorentz factor @xmath60 in the generation region , and the external magnetic field @xmath183 . in this case , the main uncertainty for electron positron jets from agns ( @xmath9 , @xmath184 ) is the value of the magnetization parameter .
indeed , this quantity depends on the efficiency of pair production in the magnetosphere of a black hole , which , in turn , is determined by the density of hard gamma - ray photons . as a result ,
if the density of hard gamma - ray photons with energies @xmath185 near the black hole is high enough , then the particles will be produced by direct collisions of photons @xmath186 @xcite .
this causes an abrupt increase in the multiplicity parameter @xmath187@xmath188 , where @xmath189 is the characteristic particle density required to shield the longitudinal electric field . using the well - known estimate ( see , e.g. ,
@xcite ) @xmath190 we obtain @xmath191 on the other hand , for low densities of gamma - ray photons when an electron positron plasma can be produced only in regions with a nonzero longitudinal electric field , which are equivalent to the outer gaps in the magnetospheres of radio pulsars @xcite , the multiplicity of the particle production is fairly small : @xmath192@xmath193 . in this case
, we obtain @xmath194 finally , for fast crab- or vela - type radio pulsars ( @xmath195 , polar - cap radius @xmath196 , and @xmath197 ) in which jets are observed , we have @xcite @xmath198 relations ( [ 82 ] ) and ( [ 83 ] ) show that the properties of the jets from agns considerably depend on the magnetization parameter @xmath53 .
for example , according to ( [ 79 ] ) , the jet particles for sources with large @xmath199 transfer only a small fraction of the energy compared to the electromagnetic flux , so the flow within the jet differs only slightly from the force - free flow .
in addition , in this case the external magnetic field @xmath200 exceeds the critical magnetic field @xmath201 . according to ( [ 43 ] ) and ( [ 73 ] )
, this implies that a subsonic region must exist in the inner regions of such jets . on the other hand ,
a substantial part of the energy in sources with magnetization parameter @xmath202 during jet collimation must be transferred by plasma particles , and no subsonic region is formed near the rotation axis . as for the fast radio pulsars , the condition @xmath203 is satisfied for them , so a subsonic region in the central parts of pulsar jets is not achieved either . on the other hand ,
the estimate ( [ 79 ] ) shows that an appreciable part of the jet total energy must be coupled with particles .
we emphasize that , since the jet radius ( [ 53 ] ) for agns always exceeds the light - cylinder radius by several orders of magnitude , @xmath204 the toroidal magnetic field @xmath205 within the jet must exceed the poloidal magnetic field @xmath206 in the same proportion , @xmath207 consequently , detection of such a strong toroidal component would be a crucial argument for the picture discussed here .
unfortunately , determination of the actual physical conditions in jets currently involves considerable difficulties .
nevertheless , not only data on the direct detection of such a structure @xcite but also evidence for the existence of magnetic fields @xmath208 , closely matching the estimate ( [ 86 ] ) @xcite , have recently appeared .
we wish to thank a.v .
gurevich for interest in this study , a useful discussion , and support .
we also wish to thank s.v .
bogovalov , l.i .
gurvits , and s.a .
lamzin for fruitful discussions .
this study was supported in part by the intas grant no . 96154 and the russian foundation for basic research ( project number 99 - 02 - 17184 ) .
l. malyshkin is also grateful to the international science foundation . | a magnetohydrodynamic model is constructed for a cylindrical jet immersed in an external uniform magnetic field .
it is shown that , as in the force - free case , the total electric current within the jet can be zero .
the particle energetics and the magnetic field structure are determined in a self - consistent way ; all jet parameters depend on the physical conditions in the external medium .
in particular , we show that a region with subsonic flow can exist in the central jet regions . in actual relativistic jets ,
most of the energy is transferred by the electromagnetic field only when the magnetization parameter is sufficiently large , @xmath0 .
we also show that , in general , the well - known solution with a central core , @xmath1 , can not be realized in the presence of an external medium
. = -2.0 cm = -3.0 cm = 17.5 cm |
Subsets and Splits