Mean-field dynamics for Ginzburg-Landau vortices with pinning and forcing (Q667935)

The paper studies the collective dynamics of many vortices in a 2D section of a type-II superconductor with applied current and impurities leading to complicated glassy effects. The authors establish in various regimes the correct mean-field equations describing the vortex matter. This work could be treated as a first step to identify proper questions towards a mathematical understanding of the glassy properties of such systems. Restricting to a 2D section of a superconducting material, it is considered rather the simpler 2D Ginzburg-Landau model, and vortex filaments are replaced by ``point vortices''. In this work, the authors focus on the question, which is to derive a mean-field equation for the vortex liquid directly from the mesoscopic model. Namely, to derive a mean-field equation describing the evolution of a large number of vortices, either by taking the limit of a fixed number of point vortices $N \uparrow \infty$ in the discrete problem, or preferably by taking the limit directly in the mesoscopic model, when the number of vortices $N_\varepsilon$ blows up as the inverse Ginzburg-Landau parameter $\varepsilon$ goes to $0$, thus investigating the commutation of the limits $\varepsilon \downarrow 0$ and $N \uparrow \infty$. This naturally leads the authors to the question on derivation of effective equations in the regime, when the impurities are scattered at a small scale, that is, when the pinning weight oscillates rapidly, by starting either from the mean-field equation, from the discrete problem, or preferably from the mesoscopic model. This question, however, remains largely open: in Section 1.5 the authors state various conjectures and give a few preliminary results. First in the paper, there is recalling the behavior of a fixed number $N$ of vortices in the asymptotic regime $\varepsilon \downarrow 0$. Vortices are subjected to three forces: (i) their mutual repulsive Coulomb (logarithmic) interaction; (ii) the Lorentz-like force due to the applied current and (iii) the pinning force. In this work, the authors rather focus on the situation when the number $N_\varepsilon$ of vortices in the mesoscopic model is not fixed but depends on $\varepsilon$ and blows up as $\varepsilon \downarrow 0$, which is a physically more realistic situation in many regimes of applied fields and currents. Then it is described the evolution of the density of the corresponding vortex liquid. In dilute regimes (that is, when $N_\varepsilon$ does not blow up too quickly with respect to $\varepsilon$), the correct limiting equation is naturally expected to coincide with the mean-field limit of the discrete vortex dynamics. The main goal of the present work is to adapt the modulated energy approach, developed by the second author to the setting with pinning and applied current, thus extending preliminary results to the case with $N_\varepsilon \gg 1$ vortices -- in the whole plane for simplicity. Several complications are caused, in particular due to the lack of sufficient decay at infinity of the various quantities, and also to the fact that the self-interaction energy of each vortex varies with its location due to the pinning weight. Next to the parabolic and conservative cases, it is also considered the mixed-flow case. The convergence to suitable fluid-like mean-field evolution equations is established. Different regimes for the intensity of the pinning and applied current lead to different limiting equations, and there is a discussion of all of them in the paper. At the first stage, the authors give a precise formulation of the problem under consideration, present modulated energy approach and underline the main new difficulties, state precise assumptions, discuss the various regimes that the approach allows one to consider, and then state main mean-field results. In order to establish the convergence of the rescaled supercurrent, it is adapted the modulated energy approach used by the second author. The main results of the paper consist of the following. In the dissipative mixed flow case, the dilute and the critical vortex regimes are studied with critical scalings: dissipative case -- dilute vortex regime (GL1), dissipative case -- critical vortex regime (GL2) and with non-critical scalings: dissipative case -- dilute vortex -- very weak interactions regime (GL1'), dissipative case -- dilute vortex -- weak interactions regime (GL2'), and also dissipative case -- nondilute vortex regime (GL3). The obtained result generalizes those in preliminary consideration to the case with pinning and applied current. The mean-field equations are fluid-like, but the incompressibility condition is lost in the critical vortex regime. For these treatments, the authors prove Theorem 1, defining the mean-field limit results in the dissipative mixed-flow case in the regimes (GL1), (GL2), (GL1'), and (GL2'). Moreover, it is proved Theorem 2, defining the mean-field limit result in the dissipative case in the nondilute vortex regime (GL3). The following result is only proven to hold in the parabolic case in the weakly nondilute regime $|\log\varepsilon| \ll N\varepsilon \ll |\log\varepsilon|\log|\log\varepsilon|$, and gives rise to a new degenerate mean-field equation. Since in the present paper the authors are mostly interested in pinning effects, they focus on the regimes (GL1') and (GL2'). At the same time, for regimes: dissipative case -- dilute vortex regime -- strong interactions (GL3') and dissipative case -- critical vortex regime -- strong interactions (GL4') the pinning effects vanish in the limit and the situation is thus much easier. For conservative case -- nondilute vortex regime (GP), the authors prove Theorem 3, defining the mean-field limit result in this regime. For $N_\varepsilon \gg |\log\varepsilon|$, the well-preparedness condition is naturally simplified, as the vortex self-interaction energy is no longer dominant. The pinning force is absent from the limiting equation since in the regime GP the interaction and the applied current dominate. The pinning weight nevertheless remains in the incompressibility condition. The mean-field equation is then a variant of the 2D Euler equation and is known as the lake equation in the context of 2D shallow water fluid dynamics. Finally, some natural homogenization questions raised by this study are briefly discussed.