dedup-isc-ft-v107-score
float64 0.32
1
| uid
stringlengths 32
32
| text
stringlengths 0
32.5k
| paper_id
stringlengths 1
14
| original_image_filename
stringlengths 5
222
|
---|---|---|---|---|
0.44076 | d2ec968144514d72ae8c79bb76bc7b12 | J(T) and b(T) coefficients of J_c(H) curves (see Eq. <ref>) for MgB_2 bulks. Symbols represent experimental data for high-pressure sintered (circles) and ball-milled (squares) samples. Solid lines represent fitting results, which is described by formulas (<ref>) and (<ref>) for sintered and ball-milled samples respectively. | cond | mat0307318-figure3.jpg |
0.467302 | b10ac9b4fd624ac1a855889725923d1d | Native-state configurations for cubes with five sites per edge; for ease of visualization the chains are shown as continuous tubes. | cond | mat0307319-fig_01.jpg |
0.423536 | b9d7c5d4abc34fcdab9507bb51a7e4de | Typical random early state (type D configuration); the chain is shown as a sequence of linked spheres. | cond | mat0307319-fig_02.jpg |
0.445588 | e82da0ec39a148e2848e7d4932dd77a2 | Examples of misfolded states, one for each type of configuration. | cond | mat0307319-fig_03.jpg |
0.540573 | f051e8f48a3d45d299782e25af695002 | A successfully folded configuration (type B) subject to the effects of thermal vibration. | cond | mat0307319-fig_04.jpg |
0.411082 | 55a999c3f0c04e78b37fa13c1c796274 | Mean internal energy (normalized) vs time for successful and unsuccessful folders of all configuration types for chains of length 125 (reduced units are used). | cond | mat0307319-fig_05.jpg |
0.499498 | 5e5ef0a612d04f058824a9944aac013e | Mean radius of gyration (normalized) vs time for successful (solid curves) and unsuccessful (dashed) folders of all configuration types for chains of length 125. | cond | mat0307319-fig_06.jpg |
0.498368 | 1db41f9055e14f01a8a5769f3380fedc | Contact pair fraction S vs time for all configurations (A to D ordered vertically) and cube sizes (4 on the left, 5 on the right); in each of these and subsequent plots of a similar kind, the time axis runs along the lower left edge of the grid, the value of the quantity whose distribution is being measured along the right edge, and the distribution at each instant is normalized. | cond | mat0307319-fig_07.jpg |
0.441368 | 93f950ca4ab24a0e95677ca3b36842bc | Separation of site pairs that are in contact when in the native state vs time (the sequence of plots is the same as in Fig. <ref>). | cond | mat0307319-fig_08.jpg |
0.54801 | 43937528a83d4497ac038e5eccb3a916 | Contact fractions for core and surface sites vs time for chains that fold successfully (plots are organized as before). | cond | mat0307319-fig_09.jpg |
0.502383 | 666b5dc0ab184c81947a6088bbc823bb | Contact fractions for site pairs within and between layers vs time for chains that fold successfully. | cond | mat0307319-fig_10.jpg |
0.405382 | 22d6ff6a819d43c29a010574fa227de8 | Generalized contact maps for cubes of size four (only chains that fold successfully contribute); shading is used to indicate the folding history, with darker spots corresponding to contacts that form earlier (the main diagonal represents the backbone neighbors). | cond | mat0307319-fig_11.jpg |
0.396434 | b48e2344161649978a6b334a0a644362 | Source-drain current I_sd versus drain voltage V_d measured at different values of V_g. The inset shows the dependence of log (I_sd) on V_g at fixed V_d, for a different device, which has a mobility μ = 0.05 cm^2/Vs and a threshold voltage V_t ≃ 0.3 V. From this plot we calculate the subthreshold slope to be 1.6 V/ decade. | cond | mat0307320-FETchar.jpg |
0.482354 | a007aa04206846768939c911a10d9c67 | Temperature dependence of the field-effect mobility for two different devices, measured at large negative gate voltage. The inset illustrates that at large negative gate voltage, -20 to -50 V, where the highest mobility is observed, μ is essentially independent of V_g. | cond | mat0307320-FETmobvstemp.jpg |
0.422114 | bdbf3964badb4531b23a83d198f89c83 | Schematic representation (a) and side view (b) of a tetracene single-crystal FET. (c) Optical microscope image of a tetracene single-crystal FET. In this device, the semi-transparent tetracene single crystal extends over several pairs of electrodes, which are clearly visible under it. In most cases smaller crystals have been used which extend over only one or two pair of contacts. These different configurations allow us to study transistors with different W/L ratios on the same crystal. (d) Molecular structure of the tetracene molecule. | cond | mat0307320-FETpicture.jpg |
0.472831 | 89ece4a51d43492ab3dc81e2247017e4 | Average degree c (top) and clustering coefficient q (bottom) from numerical simulations with η /λ =0.01 for populations of size n=1000 and 2000. Runs were equilibrated for a time t_eq=3000/λ before taking averages. | cond | mat0307321-figsim.jpg |
0.52002 | 24a364bb06e94d9bb48d437c8ebe0821 | Mean degree c (top) and clustering coefficient q (middle) as a function of ξ /λ for η / λ =0.19. Bottom: phase diagram in the mean field approximation. The behaviour of c and q along the dashed line is reported above. | cond | mat0307321-mfres.jpg |
0.501109 | 65d4f0e504c74cb8b6cdfc409a5c7a49 | Resistivity of as a function of temperature for x=1.0. For the fit we used α=2.33. Similar behavior for the resistivity is observed in all the other annealed samples where α did not significantly deviate from the FL value α=2. | cond | mat0307328-ucusige-fig1.jpg |
0.443952 | 8f86a384b6ec442ba3117d0162344616 | Experimental data for the residual resistivity of normalized by its value at 300 K (squares) and the fit corresponding to the binary alloy (solid line), according to the text. | cond | mat0307328-ucusige-fig2.jpg |
0.436464 | d29900bbc06a46d7aa4a575623de92eb | Curie temperature, T_c, as a function of the amount of Ge, x. Experimental results: squares; theoretical result from Eq. <ref>: solid line. | cond | mat0307328-ucusige-fig3.jpg |
0.438301 | c5e5626c482b4685925e8729cf4808f5 | Classes of diagrams discussed in the text. The grey blobs stand for any number of interactions. | cond | mat0307329-persfig1.jpg |
0.5072 | d9cda3ee434b4ca5bdffc860eee4dcfd | a) Diagrammatic expansion of Z(n) b) Diagrams contributing to Z(n) L̂^ | cond | mat0307329-persfig2.jpg |
0.599685 | 09b1b38f022a499d993629368ff4719b | ρ^*_0 () (Eq. (3.64)) as function of . Results of the ϵ–expansion: O(ϵ^0), short dashes; O(ϵ), long dashes; O(ϵ^2), full line. One loop crossover model: dot–dashed line. Sum rule: dotted line. | cond | mat0307329-persfig3.jpg |
0.496528 | 655bb7054f074118a0dd1544e621f33f | L̂(n/2, n)/6 ℓ̃^2 as function of ẑ = √(s_n n̂). Full line: O(ϵ^2); dashes: O(ϵ); dot-dashed line: one loop crossover model. | cond | mat0307329-persfig4.jpg |
0.452959 | a9124b91d040478eb276dbae4757af03 | Simulation results for ρ_0(,n) = n L̂(j,n)/R^2_e(n) as function of . Triangles: n = 20; crosses: n = 200; points: n = 2000. | cond | mat0307329-persfig5.jpg |
0.506011 | ccac4ab0208049548326a1d32fe244b9 | ρ_0(,n) = n̂ L̂(j,n)/R^2_e(n̂) as function of . Data: n = 200, large points; n = 2000, small points. Theory evaluated at the fixed point; full line: O(ϵ^2); dot-dashed line: one loop crossover model; dotted line: sum rule (3.63). | cond | mat0307329-persfig6.jpg |
0.462031 | 1d6dcf1ab7eb4490a07edfb670e77395 | log_10(L̂(n/2, n)/ 6 ℓ̃^2) as function of log_10(ṽ n^1/2) = log_10z̃, measured for w = 0.1, 0.4, 1.0. Dot-dashed lines: one loop crossover model. The bars indicate n_max = 2000 for the different w–values. | cond | mat0307329-persfig7.jpg |
0.45749 | 3c28353c98ee4f69a05975f05c431da1 | ρ(,n) as function of , measured on the square lattice. Data: crosses, n = 200; points, n = 2000. Excluded volume parameter: w = 0.4. The curve gives a fit to Eq. (4.2). | cond | mat0307329-persfig8.jpg |
0.450433 | 158e025a6e904edcba4adde8985faf14 | A sketch of a central core – ring system. | cond | mat0307335-fig1.jpg |
0.371594 | 8a319535d4a04fbb910b1324aa038f8e | Contour plot of vs and N, defined by the Eq. (<ref>) and constraint (<ref>) for ł=0.5 and x_0=10^-3. The energy decreases (increases by absolute value) from dark to light, and reaches its minimum at the largest possible N, N=1/x_0=1000 and the smallest a allowed by (<ref>), ≈ 0.49. The black area corresponds to values of and N which do not satisfy the localization constraint r≤ R (<ref>). | cond | mat0307335-fig2.jpg |
0.457902 | cbd1eee30d454229b91dd743ccbb0ac0 | Plot of the rescaled ground state energies of core-ring _c-r (solid line) and asymmetric binary e_b (dashed line) configurations vs angular momentum ł for x_0=10^-7. | cond | mat0307335-fig3.jpg |
0.999953 | 79ad5fe2f6dd4fab93811317a2842a8b | Schematic cross-section of the terminated coaxial line region showing the sample and reference materials suspended on sapphire plates in symmetric locations in the rf magnetic field. The sapphire plate is epoxied into the bore of a quartz tube which thermally isolates it from the copper holder, fixed at the temperature of the 1.2 K helium bath. | cond | mat0307340-endwallschematic3.jpg |
0.473509 | 394a602fc3ea491abd260f179195b771 | Example of a time dependence of the scattering length a_s during the experiment <cit.>. A condensate with ground state wave function ψ_0() is prepared at a_s=a_ init. Then the collapse is induced by going to a negative scattering length for a time τ_ evolve and finally the time evolution of the remaining particles is studied at large positive values of a_s. | cond | mat0307344-at.jpg |
0.46503 | 2a316ae1018c427d86452cee873c76ee | Number of remaining atoms after collapsing a ^85Rb condensate of N_0=16000 atoms with different values for the three-body loss rate δ_2^0. Collapse is achieved by ramping the scattering length linearly from a_ init=7a_0 to a_ collapse=-30a_0 in 0.1 [ms] as a function of time τ_ evolve [ms] (labelled as t). The `*' are taken from the experiment <cit.>. Curves are displayed in the order of decreasing N_ remnant^0 for: Cyan: δ_2^0=0.00005 (i.e. K_3^0=2.598× 10^-28 [cm^6/s]); Green: δ_2^0=0.00016 (i.e. K_3^0=8.315× 10^-28 [cm^6/s]); Blue: δ_2^0=0.0004 (i.e. K_3^0=2.079× 10^-27 [cm^6/s]); Black: δ_2^0=0.0013 (i.e. K_3^0=6.756× 10^-27 [cm^6/s]); Red: δ_2^0=0.003 (i.e. K_3^0=1.559× 10^-26 [cm^6/s]); Yellow: δ_2^0=0.008 (i.e. K_3^0=4.157× 10^-26 [cm^6/s]). | cond | mat0307344-del2tc.jpg |
0.432654 | c401aab8f4b240cd8b53f2841d82532c | Surface plot of the density function |ψ|^2 in BEC with a_ init=7a_0, a_ collapse=-30a_0 and N_0=16000. At times τ_ evolve (in [ms]) a). 0, b). 3.51, c). 10.53, d). 14.04, e). 17.55, f). 21.06. | cond | mat0307344-fig3a.jpg |
0.448251 | 73332136a5854f9398696caede5b389c | Jet images (i.e. image of |ψ(x,0,z,t)|^2) for a series of τ_ evolve values for a_ init=7a_0, N_0=16000 and a_ collapse=-30 a_0. The evolution times τ_ evolve were 2, 3, 4, 6, 8 and 10 [ms] (from a) to f)). a_ quench=0 and a_ expand=250 a_0. The time from the application of a_ quench until the acquisition of the images was equal to 5.2 [ms]. | cond | mat0307344-fig4c.jpg |
0.425048 | 9a58200843d34d7fb6f4fc688be7c860 | Jet images for time τ_ evolve at 2, 3, 4, 6, 8 and 10 [ms] (from a) to f)) when we shift the center of the initial data from the origin to (0.375,0,0). All other parameters as in Fig. <ref>. | cond | mat0307344-fig5c.jpg |
0.492648 | 367926b0993a47ddb466425ee2618c43 | Normalized gap amplitude Δ_0 (z=0,r)/Δ_0 (z=0,r=0) at T=0, as a function of the hopping ratio r=t^' /t, for different electron-electron couplings (λ = 0.9 - 1.1, increasing from bottom to top). Solid lines refer to the s-wave case, while dashed lines refer to the d-wave case. One can recognize the direct correlation between T_c^max∝Δ_0 (z=0) and r, as argued by Pavarini et al. <cit.>. Redrawn after Reference <cit.>. | cond | mat0307347-Tcr.jpg |
0.472862 | 0b7a03cba29d463882d25910312f40f3 | Critical temperatures T_c^(i,o) for inner and outer layers in the case of an n=3 layered complex, as a function of the overall doping δ. A nonuniform hole-content distribution among inequivalent layers has been assumed, according to the point-charge model. The solid curve corresponds to the uniform distribution. A crossover from superconductivity arising from inner or outer layers manifests itself as a kink in the doping dependence of T_c = max_ℓ T_c^(ℓ). Redrawn after Reference <cit.>. | cond | mat0307347-kink.jpg |
0.492601 | cfb63304e6834eafb74a41cb0d9334c2 | Circles denote the experimental specific heat coefficient of URhGe <cit.>. The solid lines show the theoretical expectation if only half of the electrons pair (a), and if there are two transitions within 5% of one another (b). See the text for additional information. | cond | mat0307348-C.jpg |
0.478751 | 70795fa76ee941b990b66dfe5639745f | The universal scaling function of order parameter and its fluctuations (lower right inset) for the CLG model, the CTTP as well as the Manna model for D=3 with β=0.840 and σ=2.069. The values of the non-universal metric factors are listed in Table <ref>. The upper left inset displays the non-universal scaling plots accordingly neglecting the non-universal metric factors. For all considered models the scaling plots contain at least four different curves corresponding to four different field values (see for <cit.> details). | cond | mat0307353-fig_01.jpg |
0.473489 | bd1e7930cfdc4c5d995f10ae4033fda0 | The universal scaling function of the order parameter and its fluctuations (inset) above the upper critical dimension D_c=4 with β=1 and σ=2. The numerical data agree perfectly with the universal mean-field scaling functions R̃(x,1) and D̃(x,1) (thick dashed lines). | cond | mat0307353-fig_02.jpg |
0.443391 | 9b0e0b90d2994fb697e4d3c891685960 | The universal scaling function at the upper critical dimension D_c=4. The right insets show the order parameter at the critical density and for zero field, respectively. The order parameter is rescaled according to Eqs. (<ref>,<ref>). Approaching the transition point (h→ 0 and δρ→ 0) the data tend to the function f(x)=x (dotted lines) as required. | cond | mat0307353-fig_03.jpg |
0.379771 | 1d012fd9532e4c6c94f55d398049836f | Schematic illustration of the proposed device. The two-dimensional hole gas (2DHG) in the GaAs part is attached to spin-polarized source and drain contacts, formed by 2DHGs in the MnGaAs parts. The gate electrode on the top controls both the carrier concentration and the Rashba spin-orbit coupling strength. | cond | mat0307354-figure1.jpg |
0.469022 | e357b913e09d4568a815409581c41759 | The modulation of the current density through the spin FET when the top-gate voltage is varied. The magnetization direction in the ferromagnetic 2DHGs is given by n̂=(0,0,1). The Fermi energy for V_G=0 is E_F=0.09 eV, the Rashba coupling term ⟨β E_z⟩ = 0.05 eV nm. | cond | mat0307354-figure10.jpg |
0.455762 | 58c42c265f894cb2ab7a8ebfd4cd876e | The modulation of the current density in the spin FET when the top-gate voltage and the bulk-gate voltage are simultaneously varied in a way such that E_F is unchanged. The contact magnetization points along n̂=(0,0,1). | cond | mat0307354-figure11.jpg |
0.475831 | 9d46cfc8d839402f94851915285d2fbd | Dispersion of the lowest quasi–2D subbands (HH1 and LH1) of a semiconductor quantum well when no exchange field and no spin-orbit coupling are present. The subband-bottom energies are computed using e E_z equal to 4 × 10^7 eV/m and vertical masses of m_hh z=0.38 m_0 and m_lh z=0.09 m_0. | cond | mat0307354-figure2.jpg |
0.45938 | e38104994058499d8aa109eacdefcc53 | The first subbands (HH1 and LH1) of a semiconductor quantum well in the presence of Rashba spin-orbit coupling, which removes the spin degeneracy. The coupling constant ⟨β E_z⟩ is 0.15 eV nm. | cond | mat0307354-figure3.jpg |
0.419738 | a56381175a0247e89b450bdccaadbe1a | The ferromagnetic-semiconductor 2D quantum-well subbands for the case of magnetization perpendicular to the 2DHG plane. The interaction constant is J_pd = 0.06 eV nm^3, the Manganese concentration is N_Mn=1 nm^-3, and ⟨β E_z⟩ = 0.15 eV nm. | cond | mat0307354-figure4.jpg |
0.415555 | d636adb884c541c39ad4b9993c53f818 | Spin components of the current density in a nonmagnetic 2DHG plotted as a function of the distance L from the interface with a ferromagnetic 2DHG. Results are shown for a half–metallic ferromagnetic contact having magnetization direction n̂ =(0,0,1). The Fermi energy is equal to 0.09 eV in a) and to 0.045 eV in panel b). It is apparent that the period of current oscillations is proportional to 1/E_F. For both spin directions, the current is normalized to the incident hole flux. | cond | mat0307354-figure5.jpg |
0.458656 | ec2410b196464198a61646b15f0deb06 | Same as Fig. <ref> a) but for the case where the magnetization direction in the ferromagnetic 2DHG (and, hence, the quantization axis for spin components of the current) is equal to n̂=(1/√(2),1/√(2),0), i.e., lies in the 2DHG plane. Note the diminished amplitude of current oscillations which would disappear altogether for n̂=(0,1,0). | cond | mat0307354-figure6.jpg |
0.423184 | efc39d440df2454b958cfadab7d02673 | The transmission probability of spin–up electrons for the case of two interfaces separated by a distance L. The high–frequency oscillations are due to resonances arising from multiple reflections between the two interfaces. It turns out (see below) that such features tend to be smeared out when the transmission is averaged over the injection angle. Parameters are the same as in Fig. <ref>. | cond | mat0307354-figure7.jpg |
0.505154 | 252e36fc1b6a4a1b9849bac27eaea6ec | The total current density as a function of the quantity L/L_so in the 2D system treatment, where L is the channel length and L_so is the total modulation length. The magnetization in the contacts is perpendicular to the plane of the 2DHG. | cond | mat0307354-figure8.jpg |
0.477595 | 7e29c557a0de4c449d059636a0ef824b | Linearized capacitance model for the influence of top and back–gate voltages on the spin FET. The 2DHG is located at point A, V_G is the voltage applied at the top gate, V_BG that of the back gate. The voltages at the drain (V_D) and source (V_S) contacts are assumed to be equal since we consider the linear response regime. | cond | mat0307354-figure9.jpg |
0.568656 | 65725f7b54dc47cea438f584bbe5ad34 | Raman spectra of SWNT samples after HP/HT-treatment at 8.0 GPa and temperatures of (a) 200, (b) 500, (c) 800, (d) 1200, and (e) 1500 (λ_ex=780 nm). Reproduced from Ref. <cit.>. | cond | mat0307356-KGBZ02-Fig-6.jpg |
0.497199 | ec2563461713487ca931a12bf9815783 | (left) Ambient-pressure Raman spectra of PLV bucky paper as prepared and after pressure release from a maximum quasi-hydrostatic pressure of 8.2 GPa (nitrogen pressure medium), showing irreversible broadening of the T band. (right) HRTEM image of graphitic needles that were observed at the base of SWNT bundles after application of hydrostatic pressure of 8–9 GPa. Reproduced from Ref. <cit.>. | cond | mat0307356-OLSK99-Fig-45.jpg |
0.502914 | 0183e346ef9240f9b1fc3db45e0e12c0 | Ambient-pressure Raman spectra in the spectral region of the tangential modes (∼1600 ) and the defect-related D band (∼1350 ) of (a) SWNT as grown and (b) shear-deformation-treated SWNT after pressure release. (c) R bands and (d) second-order Raman spectra of as-grown SWNT (lower traces) and shear-deformation-treated SWNT (upper traces). Reproduced from Ref. <cit.>. | cond | mat0307356-PKNK02-Fig-3-new.jpg |
0.418404 | 0b342946abca40f9a7c6351d93b3a775 | In-plane lattice constants vs pressure for a (10,10) nanorope. Insets: Cross sections of tubes at atmospheric pressure and 1.85 GPa. Reproduced from Ref. <cit.>. | cond | mat0307356-PMLK-Fig-8.jpg |
0.401313 | f5eb790e47bf43259b0996ed56d80edd | Raman shift vs pressure for tangential modes after Peters <cit.>. The three tangential peaks are represented by circles, squares, and diamonds, respectively. Open (closed) symbols indicate data taken on pressure increase (decrease). | cond | mat0307356-PMLK_Fig-6.jpg |
0.472494 | 843629ece5f0488382b9811d7f6ca2a6 | Raman spectrum of arc-discharge-grown SWNTs at ambient conditions (λ_ex=515 nm). | cond | mat0307356-SWNT-Raman-ambient.jpg |
0.539787 | c9b52c20a055417ead0d12e0476d61e6 | Raman spectra of (a) purified multi-wall carbon nanotubes as used as a starting material and after treatment at (b) 6.0 GPa/1600 and (c) 7.0 GPa/1800. The material was purified again after the HP/HT treatment <cit.>. Reproduced from Ref. <cit.>. | cond | mat0307356-TCWS00-Fig-2.jpg |
0.546925 | b92416b2d28e4d77a6e123b97b627ff1 | Raman spectra in the tangential-mode region of arc-discharge grown SWNTs for increasing hydrostatic pressures up to 10 GPa. Solid lines represent the experimental data and dotted ones the decomposition of the spectra into three peaks. Reproduced from Ref. <cit.>. | cond | mat0307356-TRJL99-Fig-1.jpg |
0.536383 | af846a8301474135bcc93b8885b2234f | Raman spectra of the high-energy mode of multi-wall carbon nanotubes under hydrostatic pressure. Solid lines represent the experimental data, dashed lines depict Lorentzian peaks fitted to the spectra. Reproduced from Ref. <cit.>. | cond | mat0307356-TRJL99-Fig-2.jpg |
0.453797 | 121d9793218c40268aad6c2725a31027 | Wavenumbers vs pressure of the three T-band components of the AD-SWNTs shown in Fig. <ref> and for the E_2g-like mode of AD multi-wall nanotubes. Reproduced from Ref. <cit.>. | cond | mat0307356-TRJL99-Fig-3.jpg |
0.440613 | 3fa1f4da696048079c91d63c3ea6a65a | Raman spectra of AD-SWNT bucky paper in the tangential mode region for (a) increasing pressures to 26 GPa and (b) decreasing pressures (λ_ex=515 nm). Dotted lines represent the decompositions into Lorentzian peaks fitted to the data. Reproduced from Ref. <cit.>. | cond | mat0307356-TSMS00-Fig-2.jpg |
0.478474 | bba5462788f34b8280051cae46a47e5c | Peak positions as a function of pressure for the Raman spectra of AD-SWNT bucky paper shown in Fig. <ref>. Peak positions of (a) three components of the T band up to 10 GPa, of (b) the second-strongest component (1568- at 0 GPa), and of (c) the most intense component (1594 at 0 GPa). Solid (open) symbols represent data for increasing (decreasing) pressures. The straight lines at 0–10 GPa are linear fits to the data. Models I–III in panel (c) refer to the calculations of Venkateswaran <cit.> in the range 0–5 GPa. Possible structural distortions of the nanotube cross section <cit.> are sketched in the inset. Reproduced from Ref. <cit.>. | cond | mat0307356-TSMS00-Fig-4.jpg |
0.448597 | 1b4f14dc5ee248beaf8d89368d35074b | Variation of the tangential mode wavenumbers of AD bucky paper with respect to pressure, using (a) a 16:3:1 methanol/ethanol/water mixture and (b) pure water as pressure media. The solid (open) symbols represent data for increasing (decreasing) pressures. The straight lines are linear fits to the data with the resulting pressure derivatives (in /GPa) given on the figure (λ_ex=515 nm). Reproduced from Ref. <cit.>. | cond | mat0307356-TSSK01-Fig-4.jpg |
0.582914 | 75285bdb3b5043628feb2d2d03c79679 | Raman spectra of AD-SWNTs as prepared and after pressure release from a maximum quasi-hydrostatic pressure of 26 GPa (methanol/ethanol/water pressure medium), showing changes of the R band and irreversible broadening of the T band. The measured spectra are represented by circles, solid lines are fits to the spectra with the dotted lines indicating the fitted Lorentzian peaks. Reproduced from Ref. <cit.>. | cond | mat0307356-TSSK01-Fig-6.jpg |
0.486048 | 4c1d67b709dd45318a71094a00d13da2 | R band integrated intensity of AD-SWNT bucky paper as a function of pressure, normalized to its value at 0.5 GPa (λ_ex=515 nm). The lines are guides to the eye. The inset shows the ratio of the R band to T band integrated intensity. The abrupt discontinuity near 2 GPa is apparent in the insert since the discontinuity in the T band intensity is not as pronounced as that of the R band. Reproduced from Ref. <cit.>. | cond | mat0307356-VBSR01-Fig-3.jpg |
0.464332 | 40e4b24513bd453d992fcb6f34ef026a | Pressure dependence, as calculated within model I of Ref. <cit.>, of (a) the lattice constants and (b) the hexagonal distortion of the cross section of an individual (9,9) nanotube in a bundle. The ratio r^(>)/r^(<) serves as a measure of the distortion. For clarity the distortion depicted in the inset of (b) is exaggerated compared to the actual deformation at 5 GPa. Reproduced from Ref. <cit.>. | cond | mat0307356-VRRM99-Fig-4.jpg |
0.487539 | 0e1534b8e0824304b55e5b1fef6045c3 | Cross-sectional view of a SWNT bundle showing schematically the forces on the bundle and the individual tubes in the triangular lattice when hydrostatic pressure is applied. In model I, the entire bundle is subjected to an external compression. In model II, the individual tubes are each compressed symmetrically, and intertubule coupling is ignored. The absence of vdW interactions between the tubes is shown schematically by the lightly shaded tubes surrounding the center tube. In model III, the pressure medium is allowed to penetrate into the interstitial channels between tubes and thereby exerts a sin^2 3 Θ force on the tubes. Models I and III include intertubule vdW coupling. Reproduced from Ref. <cit.>. | cond | mat0307356-VRRM99_Fig-2.jpg |
0.585156 | 505222eeb294427eb172a8af8c0c0a8c | Raman spectra of (a) the transparent material produced from MWNTs at ∼17 GPa and laser-heating to 2500–3000 K in a DAC and (b) a synthetic diamond crystal. Reproduced from Ref. <cit.>. | cond | mat0307356-Yus02-Fig-2.jpg |
0.479581 | 88c4f4a19ef241af9e80772eaf389302 | Rolling-up of a graphene sheet into a nanotube. The geometry of a single-wall nanotube is defined by the roll-up vector (n, m) or, equivalently, the chiral vector C⃗_h = na⃗_1 + ma⃗_2. The chiral vectors of a zigzag and an armchair tube are also shown. | cond | mat0307356-graphene-roll-up.jpg |
0.490199 | 3a54559fa8794885b37e7f5971b5620f | 3D energy levels (solid lines) together with 1D energy levels (asterisks) as a function of the two-body well depth d (b is fixed at b=0.13 a_⊥) for two spin-polarized fermions under highly-elongated confinement. Straight dotted lines indicate the energy levels of two non-interacting spin-polarized fermions, while straight dashed lines indicate those for two non-interacting bosons. Arrows indicate values of d for which the corresponding scattering volume V_p(E) has the particular values V_p=0 and V_p=V_p^crit, respectively. | cond | mat0307358-figure1.jpg |
0.421007 | 8e8c8085b93e4c6ab57fe8e6ef55984c | The conduction electron-pseudo fermion bubbles in Peierls (a) and Cooper (b) channels. Conduction electron propagators have lead-index , and pseudo fermion propagators carry the spin index . Latin indices refer to the Keldysh-matrix structure in (<ref>). | cond | mat0307365-bubble.jpg |
0.522888 | ab3c09ca177f4b0782bd15320b2a9b19 | Conductance as a function of V/B, for T=10^-2 B and D=10^3 B. Solid and dashed (dotted and long-dashed) lines correspond to 3.(2.) order perturbation theory. Dashed and long-dashed lines have M=1 and correspond to Appelbaum's result, which neglects the V-dependence of M shown in Fig. <ref>. Upper panel: ===0.05. Middle panel: ==0.05 and =0.2. Lower panel: ==0.2 and =0.05. | cond | mat0307365-condplotsbcdV.jpg |
0.549617 | 6f0fb9c3ec534aa89a7d522b882cda93 | Unlabeled Feynman diagrams for D_LR. Triangles denote the bare measurement vertices. | cond | mat0307365-current.jpg |
0.555892 | 65f7bd4e3e434bb6933f6e1e2858500b | Nonequilibrium magnetization (dashed and long-dashed lines) and susceptibility (full and dotted lines) as functions of B/V, for T=10^-2 V and D=10^3 V. Solid and dashed (dotted and long-dashed) lines correspond to 3.(2.) order perturbation theory. Upper panel: ===0.05. Middle panel: ==0.05 and =0.2. Lower panel: ==0.2 and =0.05 (note the different scale for χ). Corrections from third order are more pronounced for g_LR^2 ≠ g_LL g_RR. | cond | mat0307365-magnplotsbcdB.jpg |
0.536176 | 5cd5d259ed43425da20327a9fa2b399b | Nonequilibrium magnetization as a function of V/B, for T=10^-2 B and D=10^3 B. The two lower curves in each panel show the absolute value of the slope, i.e. |∂ M/∂ V|. This quantity enters the conductance shown in Fig. <ref>. Solid (dotted) lines correspond to 3.(2.) order perturbation theory. Upper panel: ===0.05. Middle panel: ==0.05 and =0.2. Lower panel: ==0.2 and =0.05. | cond | mat0307365-magnplotsbcdV.jpg |
0.496088 | ba8a0df775ab41a899010aac0bb4f341 | Unlabeled Feynman diagrams for the pf self-energies. | cond | mat0307365-selfen.jpg |
0.536399 | 633d54f288944d5eac957f8b943de066 | (a) The bare conduction electron-pseudo fermion vertex. (b) The absorption, and (c) the emission vertices. Propagators are dashed (pf) and full (ce) lines. Note that the absorption, and emission vertices could equally well have been drawn with pf-propagators. | cond | mat0307365-vertices.jpg |
0.437615 | b43b9d183a8440efb3ba3823333d8e65 | a) A bulk (2,0) mode in action. The curve is a sine wave of wavelength 1.33R_TF predicted for the (1,0) Tkachenko wave. b) Phase plot of δρ Eq. (<ref>). Notice the π phase change as you radially cross a node. c) Approximate region in the array from which atoms are removed to excite the (2,0) bulk or (1,0) Tkachenko modes in the JILA experiment. d) Close-up of b) on the location of a vortex. In both of the phase plots above, phase changes from 0 to 2π as you go from dark to light regions. | cond | mat0307368-bulk_scheme3.jpg |
0.469465 | 1b508e2b888f4b44a27a5b585adfdbd9 | Energy of the (1,0) mode vs. ξ^2/ϵ^2. In Ref. <cit.> the energy of this mode is given as 1.43ϵΩ. | cond | mat0307368-excitation.jpg |
0.396514 | ff1b561b9e8b4a98b94e7e09b208b620 | Both figures depict a snapshot of the (1,0) mode in action. a) Simulation. b) Result from the JILA experiment. The curve is a fitted sine wave of wavelength 1.33R_TF | cond | mat0307368-jila_rochester_10.jpg |
0.459287 | a79aa427e5434fcfbe52c8c16fa15798 | (2,0) mode in action. a) Simulation. b) Result from the JILA experiment. In both cases, the curves are drawn to guide the eye. | cond | mat0307368-jila_rochester_20.jpg |
0.480394 | e2b69f31af7c4980a1c4c0d7a161465a | Polarization operator π(ω) describing interband absorption in the spin-polarized low-density 2DEG in the zero LL n=0 with electron filling factor ν_e = 2π n_e l_B^2 ≪ 1. The exact selection rule is M_z=0; ν is the “principal” quantum number that labels different energy states; see text. The absorption at frequency ω is given by -2 Im π(ω)∝ n_e. | cond | mat0307370-Fig1.jpg |
0.432451 | 5bc2f9203568405e9c17899cf8a867cf | Schematic drawing of the hole-ExCR optical transitions in the low-density 2DEG from spin-polarized n=0 ↑ electron LL to the final three-particle 2e-h states in two circular polarizations σ^± (the hole is in its first LL). Open dots depict three-particle bound states of charged magnetoexcitons. The shaded area is the three-particle continuum. See also Fig. 3. | cond | mat0307370-Fig2.jpg |
0.455797 | df8a494a83554eea8ea4866c283b9f3e | Energies and dipole matrix elements of the hole-ExCR (ExCR_h) transitions from the spin-polarized n=0 ↑ LL to the first hole LL in two circular polarizations σ^±. The representative case E_0=ħω_ ch is shown. Spectra have been convoluted with a Gaussian of the 0.015E_0. Labeling of the peaks is explained in Fig. 2. The spin Zeeman energies are not included. | cond | mat0307370-Fig3.jpg |
0.561091 | 6bc180c2a77442109dd3a931c1656c8a | Experimental total structure factor for a-Co_57Ti_43. | cond | mat0307375-fig1.jpg |
0.487443 | eb8cc87d29b6432f908367338548678e | Experimental (solid line) and AHS (dashed line) G(r) functions for a-Co_57Ti_43. | cond | mat0307375-fig2.jpg |
0.466086 | 383cd6224d164ac6b4439a29c9ae012e | Experimental RDF(r) (solid line) and partial RDF^ AHS_ij(r) (dashed lines) functions for a-Co_57Ti_43. | cond | mat0307375-fig3.jpg |
0.518372 | 0effabd990614ed49a24f779525b32b8 | Experimental EXAFS spectra: (a) Co K edge and (b) Ti K edge. | cond | mat0307375-fig4.jpg |
0.407495 | 220a3e5a571341afbb004848099c3d5d | Fourier-filtered first shell (full line) and its simulation (squares) on (a) Co K edge, (b) Ti K edge. The inset at (b) shows the high-k EXAFS data and its simulation. | cond | mat0307375-fig5.jpg |
0.426812 | aebeb26cd03c444d891732a5fdcb4cb4 | Sketch of a particle interacting with a collection of passing spots, showing some of the quantities involved in defining the nonlocal SDE, Eq. (<ref>). | cond | mat0307379-SDE_defs.jpg |
0.575493 | c5e7eb890a9f4be9930360ca405a17a3 | The Spot Hypothesis I: Localized spots free interstitial volume, extending several particle diameters, diffuse upward from the silo orifice as particles drain out, causing the reverse reptation of thick chains of particles. | cond | mat0307379-rising_spots2.jpg |
0.416743 | 5bb4a3fc2402442d866c1dd2859dbda8 | The Spot Hypothesis II: Each spot displacement causes a block of neighboring particles to make a smaller displacement in the opposite direction, so as to roughly conserve the local volume. | cond | mat0307379-spot-pic2.jpg |
0.400759 | 3c3687f5b69d4145bca579f30e944d6f | The Spot Hypothesis III: Real spots also cause small internal rearrangements driven by packing constraints (relative to the average motion shown in Fig. <ref>). In a multiscale simulation algorithm, this could be done by a constrained relaxation following the mean spot-induced displacements. | cond | mat0307379-spot-pic3.jpg |
0.460044 | 09d5830c15fd432282a705055545a344 | A spot simulation of the experiment in Fig. <ref>, to be compared with the void simulation in Fig. <ref>. (Courtesy of Jaehyuk Choi.) | cond | mat0307379-spot_layers_compress.jpg |
0.402025 | 5d9225191ef149f28df8f0af16525560 | Simulation of the experiment in Fig. <ref> using the Void Model on a triangular lattice, which results in far too much mixing. (Courtesy of Chris Rycroft.) | cond | mat0307379-void_layers_compress.jpg |
0.461967 | aa073a6e56004e62a3dabf2b1871a2f4 | Sketch of nonlinear waves of free volume (gray regions) observed in funnels of draining sand <cit.>. For the same orifice width, the waves propagate upward for large opening angles (a) and sometimes downward for small opening angles (b). | cond | mat0307379-waves.jpg |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.