Filename
stringlengths 22
64
| Paragraph
stringlengths 8
5.57k
|
|---|---|
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Magnesium is composed of three stable isotopes 24Mg, 25Mg and 26Mg, which can be formed in massive stars (e.g. Woosley & Weaver 1995; hereafter WW1995). The lightest isotope is formed as a primary isotope from H, while 25,26Mg are formed as secondary isotopes. The heaviest Mg isotopes are also produced in intermediate-mass AGB stars (Karakas & Lattanzio 2003), so the isotopic ratios 25,26Mg/24Mg increase with the onset of AGB stars. Therefore, Mg isotopic ratios in halo stars could be used to constrain the rise of AGB stars in our Galaxy. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | the California Institute of Technology, the University of California and the National Aeronautics and Space Administration. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | although the latter option may be unlikely because those stars may not have had time to enrich the halo due to their longer lifetime. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Mg isotopic abundances can be obtained from the analysis of MgH lines in cool stars. After the early work of Boesgaard (1968) and Bell & Branch (1970), other studies have increased the coverage in metallicity down to [Fe/H] = −1.8 (Tomkin & Lambert 1980; Lambert & McWilliam 1986; Barbuy 1985, 1987; Barbuy, Spite & Spite 1987; McWilliam & Lambert 1988; Gay & Lambert 2000). |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | In order to reach lower metallicities ([Fe/H] < −2), very metal-poor cool dwarfs have to be discovered. This work was undertaken by Yong & Lambert (2003a,b), who found a number of metal-poor ([Fe/H] < −2) cool dwarfs (Teff < 5000 K) useful for Mg isotopic studies. Employing that sample, Yong, Lambert & Ivans (2003, hereafter YLI2003) were able to study 25,26Mg/24Mg ratios down to [Fe/H] = −2.5. Surprisingly, they found metal-poor stars with relatively high 25,26Mg/24Mg ratios, suggesting thus an important contribution by intermediate-mass AGB stars even at such low metallicities (Fenner et al. 2003, hereafter F2003). |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | The sample was selected from previous spectroscopic analyses of metal-poor cool dwarfs (Yong & Lam- bert 2003a,b). Five metal-poor stars were chosen covering the range −2.6 ≤ [Fe/H] ≤ −1.3: G 69-18 (LHS 1138), G 83-46 (LHS 1718), G 103-50, G 63-40 (LHS 2765) and the well-known moderately metal-poor dwarf HD 103095, with [Fe/H] = −2.6, −2.6, −2.2, −1.9 and −1.3, respectively. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | The observations were obtained with HIRES (Vogt et al. 1994) at the Keck I telescope. The first set of spectra was taken in August 2004, just a few days after a HIRES upgrade, taking thus advantage of improvements in efficiency, spectral coverage and spectral resolution. A resolving power of R ≈ 105 was achieved using a 0.4′′-wide slit. Additional observations were obtained in November 2004 and June 2005. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Three wavelength regions at 5134.6, 5138.7 and 5140.2 ˚A are usually employed to determine the isotopic abundance ratios 25,26MgH/24MgH (e.g. McWilliam & Lambert 1988; YLI2003). For these regions we adopted laboratory FTS measurements of the isotopic 24,25,26MgH lines obtained by Bernath, Black & Brault (1985). |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Bernath et al. (1985), and the corresponding 25,26MgH line positions were computed by adding the theoretical isotopic shifts to the laboratory 24MgH wavenumbers. The 25,26MgH isotopic shifts were calculated using the relative reduced mass of 25,26MgH to 24MgH. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | The energy levels were calculated using molecular constants by Shayesteh et al. (2004) and we adopted a dissociation energy of 1.27 eV (Balfour & Lindgren 1978). Oscillator strengths were obtained from transition probabilities given by Weck et al. (2003). |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Molecular C2 lines are also present in the same region, so a line list of C2 lines was also implemented. Laboratory wavenumbers were taken from Amiot (1983) and Prasad & Bernath (1994). The rotational strengths (H¨onl-London factors) were computed following Kovacs (1961), and we adopted an oscillator band strength of f00 = 0.03 (see Grevesse et al. 1991). Excitation potentials were computed using the molecular constants by Prasad & Bernath (1994) and a dissociation energy of 6.297 eV (Urdahl et al. 1991) was adopted. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Atomic lines present in the region were also included. The initial line list was based in the work of Barbuy (1985), and lines were added or discarded based on spectral synthesis of both the Sun and Arcturus spectra. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | In previous works the macroturbulence has been determined mainly using two lines: Ni I 5115.4 ˚A and Ti I 5145.5 ˚A (e.g. McWilliam & Lambert 1988; YLI2003). For the two more metal-poor stars in our sample these lines became too weak, so we additionally used lines of Fe I and Ca I present in the 5569-5601 ˚A region. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | The stellar parameters (Teff , log g, [Fe/H], vt) were initially adopted from Mel´endez & Barbuy (2002) for HD 103095, and from Yong & Lambert (2003b) for the other two stars. A check of the stellar parameters was done employing the IRFM Teff calibrations by Ram´ırez & Mel´endez (2005), Hipparcos parallaxes, Y2 isochrones (Demarque et al. 2004) and our HIRES spectra. E(B-V) was estimated both using interstellar Na I D lines and reddening maps (§4.1 of Mel´endez et al. 2006). Reasonable agreement was found with respect to the stellar parameters given in the above references. Our final adopted values are given in Table 1. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | The contribution of C2 lines was constrained by spectral synthesis of the weak feature around 5135.7 ˚A, which is a blend of C2 lines (5135.57 and 5135.69 ˚A) sometimes blended with an unidentified line in the red side (Gay & Lambert 2000). Fortunately the observations are of such high resolution that it is possible to constrain the contribution of C2 employing the blue side of this feature, imposing thus an upper limit to blends by C2 lines. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Initially the Mg isotopic ratios were determined by an eye-fit of the synthetic spectra to the HIRES observed data of the three recommended regions (see §3). The results are shown in Table 1. After the fits by eye were completed we performed a χ2 fit by computing χ2 = Σ(Oi − Si)/σ2, where Oi and Si represents the observed and synthetic spectrum, respectively, and σ = (S/N)−1. As an example of the χ2 fits we show in Fig. 1 the fits for the recommended region at 5140.2 ˚A in the most metal-rich and most metal-poor stars of our sample. The results of the χ2 fits are shown in Table 1. As can be seen, the eye fit compares well to the χ2 fit. The errors given in Table 1 are due to statistical errors (the standard deviation between the isotopic ratios of the three recommended regions) and systematic errors of 1% (due to errors in the atmospheric parameters, see e.g. YLI2003). Our results for HD 103095 (24Mg:25Mg:26Mg = 94.8:2.4:2.8) compare very well with previous visual (eye-fit) determinations in the literature. For HD 103095 both Tomkin & Lambert (1980) and Barbuy (1985) obtained isotopic ratios of 94:3:3. More recently Gay & Lambert (2000) determined 93:4:3. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | ratio 26Mg/24Mg, since the isotopic ratio for 25Mg is more uncertain due to the smaller isotopic shift. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | In Fig. 2 (left panel) we compare our results with the models computed by F2003, both including and neglecting the contribution of intermediate-mass AGB stars. Another model (Alib´es, Labay & Canal 2001, hereafter ALC2001) which includes only massive stars is also shown. A comparison with other models (Ashenfelter, Mathews & Olive 2004, hereafter AMO2004; Goswami & Prantzos 2000) is shown in the right panel. As can be seen, our low isotopic ratios can be explained mostly by massive stars, thus we find no need to invoke the contribution of intermediate-mass AGB stars at low metallicities. HD 103095 lies slightly above the predicted F2003 curve of massive stars nucleosynthesis (although it is in perfect agreement with the ALC2001 model), so this may indicate that at [Fe/H] & −1.5 the contribution from AGB stars begins. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | The high isotopic ratios found in metal-poor stars by YLI2003 were interpreted by F2003 as an important contribution of intermediate-mass AGB stars at low metallicities. However, most of the stars with high isotopic 25,26Mg/Mg ratios in YLI2003 are not bona fide halo dwarfs. We computed the probability of halo membership following Bensby, Feltzing & Lundstr¨om (2003) and found that a fraction of the metal-poor stars in YLI2003 are actually thick disk stars. Furthermore, some halo stars have abundance anomalies (e.g. CH stars) or their spectra are abnormal (e.g. double lined), and they should be removed for a fair comparison with chemical evolution models. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Thus both our results and YLI2003 suggest a small (or none) 25,26Mg contribution of intermediate-mass AGB stars to the Galactic halo. Perhaps the 25,26Mg yields from AGB stars are lower than in current models (Karakas & Lattanzio 2003). If this is the case, then intermediate-mass AGB stars can not be invoked to explain the possible variation of the fine-structure constant α (AMO2004). The chemical evolution models used to explain variations in α require an ad-hoc AGB-enhanced IMF in order to produce large amounts of 25,26Mg (AMO2004). If the correct yields are lower than present calculations, then much larger ad-hoc modifications to the IMF would be required. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Calculations of Karakas & Lattanzio (2003) show that the AGB stars that contribute significant amounts of 25,26Mg are stars with initial masses of 3-6 M⊙. Since these stars have lifetimes considerably shorter than the age of the universe, they can be used to constrain the timescale for the formation of the Galactic halo. According to the Padova evolutionary tracks4 the lifetime of 3-6 M⊙ metal-poor stars ([Fe/H] = −1.5) are 0.1-0.3 Gyr, so the halo timescale formation should be of the order of 0.3 Gyr. This short timescale probably explains why recent studies of age spread in Galactic globular clusters have shown that most clusters from intermediate to low metallicity are coeval within the uncertainties (e.g. Rosenberg et al. 1999; De Angeli et al. 2005). |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Further high S/N high spectral resolution observation of a larger sample will help constrain the rise of AGB stars in the Galaxy and will be useful to better constrain the formation timescale of the Galactic halo. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | JM thanks C. Chiappini for useful discussions and D. Yong for providing data to check our MgH line list, as well as for useful discussions on Mg isotopic ratios. JGC is grateful for partial support to NSF grant AST-0507219. We have made use of data from the SIMBAD database operated at CDS. The authors 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 are most fortunate to have the opportunity to conduct observations from this mountain. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Fig. 1.— Fits for the 5140.2 ˚A region in the stars HD 103095 and G 69-18. Observed spectra are represented with filled circles, and synthetic spectra with solid lines. The calculations were performed for 25,26Mg/Mg ratios of 0-5%. The relative variation of the χ2 fits are shown as a function of the isotopic abundance. |
Processed_Magnesium_Isotopes_in_Metal-Poor_Dwarfs,_the_Rise_.txt | Fig. 2.— 26MgH/24MgH as a function of [Fe/H] in halo dwarfs. Filled and open circles represent our results and those by YLI2003, respectively. All models include yields of massive stars (mostly by WW1995). Models including massive stars and intermediate-mass AGB stars (F2003 and AMO2004) are also shown. Note that AMO2004 extrapolated AGB yields for Z < 0.004, so their results may be unrealiable. The model that agrees better with the observed data is the ALC2001 model, which does not include intermediate-mass stars. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | In this paper, we study the influence of an Allee effect on the spreading rate in a local reaction-diffusion-mutation equation modeling the invasion of cane toads in Australia. We are, in particular, concerned with the case when the diffusivity can take unbounded values. We show that the acceleration feature that arises in this model with a Fisher-KPP, or monostable, non-linearity still occurs when this non-linearity is instead bistable, despite the fact that this kills the small populations. This is in stark contrast to the work of Alfaro, Gui-Huan, and Mellet-Roquejoffre-Sire in related models, where the change to a bistable non-linearity prevents acceleration. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | The invasion of cane toads in Australia has interesting features different from the standard spreading observed in most other species. The experimental data [29, 33] show that the invasion speed has steadily increased during the eighty years since the toads were introduced in Australia. In addition, the younger individuals at the edge of the invasion front have a significantly different morphology compared to other populations – their legs tend to be on average much longer than away from the front. This is just one example of a non-uniform space-trait distribution – see, for instance, a study on the expansion of bush crickets in Britain [34]. Several works have addressed the front invasions in ecology, where the trait is related to the dispersal ability [3, 15]. It has been observed that selection of more mobile individuals can occur, even if they have no advantage in their reproductive rate, due to the spatial sorting [1, 23, 29, 30]. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | spatial variable, with the diffusion coefficient θ, representing the effect of the trait on the spatial spreading rates of the species. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | In the model proposed in [4], the reaction is of non-local, monostable type. In the case where Θ is bounded, the problem has been shown to be well-posed [35] and exhibit travelling-wave solutions [9]. When Θ is unbounded, sharp results have been proven for the local analogue while weaker results have been obtained for the non-local equation [8, 12]. In particular, the Cauchy problem exhibits propagation of of the order t3/2. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | With a monostable non-linearity, acceleration, i.e. super-linear spreading of level sets, has been shown in various settings in mathematical biology [7,11,13,14,16,18,20–22,26,32]. We mention, in particular, the existence of acceleration for Fisher-KPP with fat-tailed initial data and for Fisher- KPP with the fractional Laplacian. In these models, acceleration is related to the notion of pulled fronts. In particular, we see that small populations far from the origin grow exponentially, causing the acceleration. Recently, in these settings, it has been shown that when the non-linearity is ignition or bistable type, there is no acceleration [2, 19, 27]. Indeed, the small populations that drive the acceleration are killed when the non-linearity is bistable. In other words, acceleration is a tail phenomenon, so when a bistable non-linearity kills the tails, there is no acceleration. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | In this article, we obtain results in stark contrast with the results discussed above. Indeed, we demonstrate propagation of the order t3/2 in the local cane toads equation with a bistable non- linearity. This shows that, for the cane toads equation, acceleration is a bulk phenomenon, i.e. that it is not driven by small populations far from the origin. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Our interest is in understanding where the “front” of u is. In other words, we will fix a level set and understand its propagation. Using the same equation with a monostable non-linearity as a super-solution to (1.1), the results from [8, 12] show that no level set can move faster than O(t3/2). Our main result is to show that a lower bound of the same order holds as well. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Before we continue, we note that we expect that this theorem holds in much greater generality. However, our interest is in showing that the acceleration in the cane toads equation is a also bulk phenomenon, as opposed to a tail phenomenon. Hence, we seek to simply show that acceleration occurs despite the bistable non-linearity, and we do not attempt to find the most general setting or the sharpest bounds. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | To prove this theorem, we will follow a similar strategy as in a previous paper by the authors and Ryzhik [12] and that also appears in [6, 31]. In rough words, we will slide a suitable “bump” along suitable trajectories in the phase plane R × Θ, making sure that the ball remains below the solution of the original Cauchy problem (1.1). Once the trajectories are well chosen, this will imply the acceleration phenomena claimed in theorem 1.1. However, the “bump” sub-solution is significantly more complicated to create in this setting than in the monostable one. Indeed, in [12], the sub-solution is created almost entirely with the linearized (around zero) problem. In our setting, of course, the linearized problem decays to zero on any traveling ball. We describe how to overcome this difficulty below. This is the main technical difficulty in the present article. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | where Λ is a positive constant to be chosen later that encodes the sizes of this two objects. For now, our only assumption on Λ is that Λ ≤ λθ/8. Our goal is to build a sub-solution to (1.1) on the bigger ellipse At,Λ ∪ Et,Λ, what we called a “bump” above. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | where we define cr = (1 − α)(r − α)−1. It is easily verified that fr(u) ≤ f (u) for all u, and it is clear that fr is Lipschitz continuous. For technical reasons, we are required to take r to be slightly less than 1 in the sequel. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Figure 1.1: Schematic plot of the flying saucer like sub-solution w that is slid along the trajectory. The bolded green line denotes the set At, the red dotted one the set Et. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | where ∂n is the (outward) normal derivative to the boundary of Et,Λ. This is the main technical issue at hand in the proof. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Let us enlighten theorem 1.1 by showing in fig. 1.2 some numerical insights of the acceleration behavior. One can compare these to [8,10] that are numerics related to the study of the acceleration in the nonlocal cane toads equation. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | It is easy to see that one many bound the propagation rate above by using the monostable model with growth rate r = supρ∈R+ ((ρ − α)(1 − ρ)). This implies that the propagation can be no faster than O(t3/2). However, obtaining a lower bound is significantly more complicated. We are led to the following conjecture, which we are unable to prove at this time. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Conjecture 1.2. The model (1.9) exhibits acceleration. That is, the level sets of ρ move super- linearly in time. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | We provide some numerics in fig. 1.3 to support this conjecture. However, the proof of such a result is far beyond the scope of this paper. Indeed, it requires to study the dynamics of the zone where ρ is greater that α, which is not at the edge of the front at all, so that an argument by contradiction as in [12] is impossible. Moreover, to our knowledge, it is not possible to derive a probabilistic framework such as in [8] (and the references therein) to be able to study the nonlocal model (1.9). As new tools will be needed to study this model, we thus leave this conjecture for further investigation. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | In section 2, we give the main outline of the proof of theorem 1.1 assuming the main proposition 2.1. In section 3, we prove this proposition by finishing the construction of the subsolution (1.7), and reducing it to showing a proposition 3.2 that gives (1.8). This main lemma involves understanding a related steady state problem, which we analyze in sections 4 and 5. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Figure 1.2: Numerical simulations of the Cauchy problem of equation (1.1) at a fixed time, in the phase space R × Θ. From top left to bottom right: the initial data, t = 30, t = 60, t = 90, t = 120. One can track the accelerated behavior, for example on the space axis. The invasion at the back in the θ-direction is expected to be linear in time. This pattern is very similar as for the monostable cane toads equation, see also [8, 10, 12]. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Figure 1.3: Numerical simulations of the Cauchy problem of equation (1.9) at a fixed time, in the phase space R × Θ. From top left to bottom right: the initial data, t = 30, t = 60, t = 90, t = 120. Here we see the distinctive ”up-and-over” behavior that leads to acceleration for the non-local, monostable model. One can track the accelerated behavior, for example on the space axis. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | In this section, we prove theorem 1.1. For this purpose, we first define carefully the trajectory that has been introduced in the previous section. Second, we take for granted the existence of a sub-solution of the form (1.7) (this will the object of section 3) and we show how this implies the acceleration feature stated in theorem 1.1. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | This implies that w will initially sit below u, allowing us to apply the maximum principle. With this in mind, we state in (a very minimalist fashion) a proposition that will imply our main theorem 1.1. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | In addition, all constants λ0, Λ0, c0, and T0 are bounded for m bounded away from 1, and λ0 depends on Λ. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Figure 2.1: Plot of the trajectory in the phase plane (x, θ). We emphasize that the first part of the trajectory corresponds to a movement towards large traits only in the θ-direction, whereas the second part of the trajectory corresponds to a movement towards large space positions only in the x-direction, at an accelerated spreading rate. The red dotted line is the support of the initial condition. The red bold line is the initial support of the subsolution w : it sits inside the support of the initial condition. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Remark 2.2. To find optimal constants, one could write a general framework for the trajectories such as for the cane toads equation in [12] and try to optimize afterwards. However, in this paper, the suboptimal trajectories are not the only reason for sub-optimality. One would have to quantify all the constants in various lemmas in this paper, that are not strongly related to the definition of the trajectories themselves. Since this is not our main interest, we opt for the simplest possible presentation. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | w+(t, x, θ) = α + v+ (t, y, η) , w−(t, x, θ) = α − v− (t, y, η) . |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | The maximum principle assures us that 0 ≤ v− ≤ α holds everywhere in time and space as long as the initial conditions of v− satisfy the same bound. Similarly, The maximum principle assures us that 0 ≤ v+ ≤ r − α holds everywhere in time and space as long as the initial conditions of v+ satisfy the same bound. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | We now need to fix the initial conditions. It is easy to see that the non-constant coefficients in the equation vanish as λ tends to infinity. Hence, it behooves us to look at the steady state solutions to (3.1) and (3.2) where the dependence on λ disappears. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | where we choose r(1) = 1 and we leave r(2) ∈ (max{m, 2α}, 1) to be determined. We now state the lemma regarding these steady solutions that we require. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | With the initial conditions we have fixed for v±, it is clear that as Λ tends to infinity, v± should tend to ϕ± uniformly in the C1 topology. The following lemma ensures that this convergence holds and is uniform in T . |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | We postpone the proof of this Lemma until the dedicated section 5. We now show how the combination of lemma 3.1 and lemma 3.3 finally implies proposition 2.1. After changing variables it is equivalent to proving the same result for v+. This is exactly now that we are going to travel along the trajectory (XT , ΘT ) and play with parameters. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | on the set EΛ/2. This concludes our sliding over the first part of the trajectory. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | First, by increasing Λ0 if necessary, we may find r(2) := r ≥ (cid:15) + m in the interval (2α, 1 − 2ε). |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | By the maximum principle any solution to (1.4) will be less than or equal to r. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Proof of proposition 3.2. Fix any unit vector ˆv ∈ S1(R2) and look at any point Λˆv on the boundary of EΛ. Then (3.6) follows from lemma 3.3 and lemma 3.4. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | In this section, we prove lemma 3.1 and lemma 3.4, which give the desired behavior for the steady states ϕ± (i). |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | C = {(y, η) | |y − (y0 − yL)| = L or |η − (η0 − ηL)| = L} . |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | for all u ∈ [0, γ]. The last inequality follows by decreasing c0 and increasing L, if necessary. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Now we claim that ϕ+ tends to r − α locally uniformly in the set EΛ−L as Λ and L tend to ∞. To prove this, we argue by contradiction. If this is not true, then we may find a positive (cid:15) > 0, sequences Lk ≤ Λk/2 tending to infinity, and a sequence of points (yk, ηk) ∈ EΛk−Lk such that ϕ+(yk, ηk) ≤ r − α − (cid:15). |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | k converges locally uniformly in C2 to a function ϕ+ ∞(0, 0) ≤ r − α − (cid:15). By our work above, we have that ϕ+ ∞ ≤ r − α everywhere. Finally, we have that ϕ+ ∞ satisfies (3.4). We may assume that (0, 0) is a positive global local minimum for ϕ∞ since, if not, we may simply re-center the equation again, taking limits if necessary. It is apparent, however, that a solution to (3.4) can only have a minimum at the zeros of fr(α + ·), which are 0 and r − α. Thus, we arrive at a contradiction. Hence, it must be that ϕ+ tends uniformly to r − α, finishing the proof of the claim. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | By the compactness of the unit circle, ˆvk converges (up to a subsequence) to some unit vector ˆv∞. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | In addition, z±(y, η) is uniformly positive away from this boundary. Furthermore, z+(y, η) converges to r − α and z−(y, η) converges to α as the distance between (y, η) and ∂P tends to infinity. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | results from lemma 5.1. Lemma 4.1. There exists δ > 0 such that if (cid:107)c∞(cid:107) ≤ δ then ∂nz+(0, 0) > ∂nz−(0, 0). |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Since we have reached a contradiction, it must be that (3.8) holds for all Λ sufficiently large. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Before we prove lemma 3.3, we prove a few useful lemmas. The first of these is the uniqueness of solutions to the related steady half-plane problem. We require this to prove lemma 4.1, stated above. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | P = {(y, η) : 0 ≤ (y + yb, η + ηb) · ˆv}. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | which satisfies Dirichlet boundary conditions on ∂P and which tends uniformly to r − α as dist((y, η), ∂P ) tends to infinity. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | which satisfies Dirichlet boundary conditions on ∂P and which tends uniformly to α as dist((y, η), ∂P ) tends to infinity. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Proof of lemma 5.1. We show the argument for z+ but the argument for z− is identical. By rotation and translation, we may assume that ˆv = (1, 0) and that (yb, ηb) = (0, 0). We will use a sliding method [5]. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | is monotonically decreasing for u ≥ u0. Since z1 and z2 tend uniformly to r − α as they tends to infinity, we may find L such that if y > L, then zi(τ, y, η) ≥ u0 for i = 1, 2. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | We handle the former case first. In this case, elliptic regularity ensures that z1, z2 are uniformly (cid:0) Lipschitz continuous, and that we may then decrease τ to a positive constant τ (cid:48) < τ and preserve the fact that zτ (cid:48) 2 ≥ z1 on P . This is contradicts the definition of τ . |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Hence, it must hold that τ = 0. By the definition of τ , this implies that z2 = zτ =0 contradicts our original assumption that z2(y0, η0) < z1(y0, η0). This finishes our proof. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | The following lemma addresses the issue of uniqueness for the steady problem (3.4) on the bounded ellipse EΛ. While we expect uniqueness to hold in this setting, we are unable to prove it because fr need not satisfy the condition that fr(α + u)/u that is usually used for uniqueness results of this type. As such, we settle for a type of asymptotic uniqueness, stated below. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Lemma 5.2. For any (cid:15) > 0, there exists Λ(cid:15) such that if Λ ≥ Λ(cid:15) any two positive solutions to (3.4) must be (cid:15)-close in the L∞ norm. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Proof of lemma 5.2. Suppose the opposite, that there is a positive constant (cid:15), a sequence of Λk tending to infinity, and two sequences of functions ϕ1,k and ϕ2,k which both solve (3.4) such that (cid:107)ϕ1,k − ϕ2,k(cid:107)∞ ≥ (cid:15). Then we may find a sequence of points (yk, ηk) such that ϕ1,k(yk, ηk) ≥ ϕ2,k(yk, ηk) + (cid:15). We may select (yk, ηk) to be a local maximum of ϕ1,k − ϕ2,k. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | By elliptic regularity, it follows that ˜ϕi,k converge, along a subsequence, if necessary, locally uni- formly to ϕi. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | for some unit vector ˆv and some translation (y∞, η∞), with Dirichlet boundary conditions on ∂P . From lemma 5.1, we know that (3.4) has a unique positive solution on P . From this, it follows that ϕ1 = ϕ2. This contradicts the fact that ϕ1(0, 0) ≥ ϕ2(0, 0) + (cid:15). This contradiction finishes the proof of the claim. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | to v± to ensure that when taking limits in the arguments, we do not lose positivity of v±. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Lemma 5.3. There exists Λ0 and λ0 such that if Λ ≥ Λ0 and λ ≥ λ0, then v+ and v− are uniformly positive for any subset K of E and A, respectively. Moreover, the lower bound may be chosen in such a way to depend only on the distance between the subset K and the boundary ∂E and ∂A, respectively. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Proof of lemma 5.3. This follows easily by using the time dependent eigenvalues of Lemmas 5.1 and 5.2 from [12] along with the strategy from lemma 3.1. As such we omit the proof. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Proof of lemma 3.3. Before we begin, we point out that, by parabolic regularity, it is enough to show that v± converges to ϕ± uniformly in L∞. In addition, we prove the claim for v+ and ϕ+, but the proof is identical for v− and ϕ−. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | By the compactness of E, we have that (yk, ηk) tends, along a sub-sequence if necessary, to a point (y, η). In addition, lemma 5.3 ensures that vk is uniformly positive on any compact subset of E. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | There are two cases: either tk is bounded or not. If it is bounded, then the choice of initial conditions for v+ along with parabolic regularity, see e.g. [25], implies that vk converges to ϕ+. Hence it must be that tk tends to infinity. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | k falls within 3(cid:15)/4 of a steady state of (3.4). Indeed, ∞(t, y, η). First, notice that this must solve the parabolic analogue of (3.4). |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | by the maximum principle. By our choice of initial conditions, it must be that wτ > 0 for all times and for all y and η. By the boundedness of w, it follows that wτ tends uniformly to zero and hence, to a solution of eq. (3.4). We may similarly find an upper bound for v∞(0, y, η) in terms of a non-trivial solution of eq. (3.4). Hence, choosing t0 large enough and applyinglemma 5.2 with the choice (cid:15)(cid:48) = (cid:15)/8, we have that v∞ is within (cid:15)/4 of a steady state. |
Processed_Super-linear_spreading_in_local_bistable_cane_toad.txt | Applying lemma 5.2 again implies that v+ and ϕ+ are within (cid:15)/2, using again our choice of Λ above. However, this contradicts our assumption that |v+(0, y, η) − ϕ+(y, η)| ≥ (cid:15), finishing the proof. |
Processed_Space-Time_Variation_of_Physical_Constants_and_Rel.txt | Detection of high-redshift absorption in the optical spectra of quasars have provided a powerful tool to measure spatial and temporal variations of physi- cal “constants” in the Universe. It is demonstrated that high sensitivity to the variation of the fine structure constant α can be obtained from a comparison of the spectra of heavy and light atoms (or molecules). We have performed calculations for the pair FeII and MgII for which accurate quasar and labora- tory spectra are available. A possibility of 105 times enhanced effects of the fundamental constants variation suitable for laboratory measurements is also discussed. |
Processed_Space-Time_Variation_of_Physical_Constants_and_Rel.txt | The first ideas about possible variation of the fundamental physical constants in the expanding Universe were suggested by Dirac [1]. Now this subject is of particular current interest because of the new possibilities opened up by the structure of unified theories, like the string theory and M-theory, which lead us to expect that additional compact dimensions of space may exist. The “constants” seen in the three-dimensional subspace of the theory will vary at the same rate as any change occurring in the extra compact dimensions (see, e.g. [2–4]). |
Processed_Space-Time_Variation_of_Physical_Constants_and_Rel.txt | Quasar absorption systems present ideal laboratories where to search for any temporal or spatial variation of fundamental constants by comparing atomic spectra from the distant objects with the laboratory spectra (see, e.g. [5] and references therein). |
Processed_Space-Time_Variation_of_Physical_Constants_and_Rel.txt | It would seem natural to find this change from measurements of the spin-orbit splitting within a fine-structure multiplet. However, this way is not the most efficient, and it may even give incorrect results, since other relativistic effects are ignored. The aim of this letter is to demonstrate that the change in α produces an order of magnitude larger effect in the difference between transition frequencies in heavy and light atoms (or molecules). We have calculated the dependence of the transition frequencies on α for FeII (see eq.(8)) and MgII (see eq.(6)) where accurate data exist both for laboratory and quasar spectra. Other possibilities include comparisons of different optical transitions, for example s − p and p − d, in the same atom or molecule, or comparisons of microwave transitions in molecules which contain rotational and hyperfine intervals. |
Processed_Space-Time_Variation_of_Physical_Constants_and_Rel.txt | We also propose another interesting possibility: to use transitions between “accidentally” degenerate levels in the same atom or molecule. This degeneracy would disappear after a minor change in α. For example, in the Dy atom there are two degenerate opposite-parity levels [7,8]. The frequency of the E1-transition between them is smaller then the hyperfine splitting of each level. As a result, the relative effect of the change in α is enhanced by five orders of magnitude (the ratio of the size of the relativistic effect to the transition frequency). This case seems to be more suitable for laboratory experiments. Similar experiments with “accidentally” degenerate molecular levels belonging to different electron terms are also sensitive to the smallest changes in the ratio of the nucleon to electron masses since in this case the difference in electron energies is compensated by the difference in the vibrational and rotational energies of the nuclei. As is known, the nuclear mass is a function of the strong interaction constants and vacuum condensates. |
Processed_Space-Time_Variation_of_Physical_Constants_and_Rel.txt | The second term in the square brackets is presented to provide a continuous transition from the hydrogen-like ion Eq. In multi-electron ions (1) to the multi-electron ion Eq. (Z ≫ Za) this term is, in fact, a rough estimate based on the direct calculation of hV i. We should neglect this small term since there are more important many-body corrections (the accurate many-body calculations discussed below give approximately constant correction term in square brackets with the value about −0.6). |
Processed_Space-Time_Variation_of_Physical_Constants_and_Rel.txt | We see that the relativistic correction is largest for the s1/2 and p1/2 states, where j = 1/2. |
Processed_Space-Time_Variation_of_Physical_Constants_and_Rel.txt | We changed slightly the values of the calculated coefficients K1 and K2 to reproduce exactly the experimental value of the fine splitting ∆ls = 91.6cm−1. The laboratory frequency of the 3s1/2 − 3p3/2 transition in the main isotope 24Mg (abundance 79 %) has been taken from [12]. The increase of the frequencies (isotope shifts) for the transition in 25Mg (abundance 10 %) and 26Mg (abundance 11 %) are equal to 0.053 and 0.102 cm−1, respectively [12]. The laboratory frequency of the 3s1/2 − 3p1/2 transition has been taken from [11]. |
Processed_Space-Time_Variation_of_Physical_Constants_and_Rel.txt | Now consider FeII. We are interested in the E1-transition from the ground state 3d64s 6P . Due to the selection rule ∆J ≤ 1 the transitions from the ground state J = 9/2 can involve only few components of the excited multiplets. The splitting between the nearby components of the same multiplet (e.g. J = 9/2 and J = 7/2) are about 100 - 200 cm−1. On the other hand, the total relativistic correction to the mean frequency of the 4s − 4p transition is about 1800 cm−1, i.e. ∼ 10 times larger! Unfortunately, we can not measure the effect of this large correction by measuring the frequency of 4s − 4p transitions in one element (FeII) since it will be hidden in the definition of the red shift parameter 1 + z. Therefore, to make use of this large effect we need to consider two elements with different relativistic shifts, e.g. MgII and FeII, or different transitions in the same element, e.g. s − p and d − p. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.