Lorentz space estimates for the Ginzburg-Landau energy (Q2472862)

The authors consider the functional representing the Ginzburg--Landau free energy of the model of superconductivity, that is \[ F_{\varepsilon}(u,A)=\frac12\int_{\Omega}|\nabla_Au|^2|+|\text{curl} A|^2+\frac{(1-|u|^2)^2}{2\varepsilon ^2}, \] where \(\Omega\) is a bounded regular domain in \(\mathbb R^ 2\), \(u\) is a complex-valued function and \(A\in\mathbb R^2\) is a vector field in \(\Omega\), representing the vector potential of the magnetic field. The authors focus on the regime for small \(\varepsilon\). Note that \(\varepsilon\) is a material constants whose smallness indicates the superconductivity of the so-called type II. In this setting, \(u\) can have vortices (zeros of non-zero topological degree). The authors also consider the case without magnetic field (\(A=0\)). The main objective is to obtain lower bounds for \(F_{\varepsilon}\), in particular, bounds that relate \(F_{\varepsilon}(u,A)\) to the norms of \(\nabla u\) in the Marcinkiewicz space \(L^{2,\infty}\). These spaces and their dual Lorentz spaces \(L^{2,1}\) have been studied in a similar context by \textit{F.-H. Lin} and \textit{T. Rivière} [Commun. Pure Appl. Math. 54, No. 2, 206--228 (2001; Zbl 1033.58013)]. Usually, the Ginzburg--Landau energy is unbounded as \(\varepsilon\) tends to zero, in fact, it blows up roughly like \(\pi n|\log n|\varepsilon\), where \(n\) is the number of vortices. The authors present several main results. First, they improve the \(L^2\)--lower bounds obtained by \textit{E. Sandier} and \textit{S. Serfaty} [Vortices in the magnetic Ginzburg-Landau model. Basel: Birkhäuser (2007; Zbl 1112.35002)] by a ball construction as follows: \[ \begin{aligned} \frac12\int_V&|\nabla_A u|^2+\frac1{2\varepsilon^2}(1-|u|^2)^2+r^2(\text{curl} A)^2\\ &\geq \pi D(\log\frac r{\varepsilon D}-C)+\frac1{18}\int|\nabla_{A+G}u|^2+\frac1{2\varepsilon^2}(1-|u|^2)^2, \end{aligned} \] where \(V\) is the intersection of the set \(\{x\in\Omega,\,\text{dist}(x,\partial\Omega)>\varepsilon\}\) with a union of a certain disjoint collection of closed balls \(B\), \(D=\sum_{B\subset\Omega}|d_B|\), and \(d_B=\deg(u,B)\). The improvement consists in the addition of the term \(\frac1{18}\int|\nabla_{A+G}u|^2\). Next, in their principal result, the authors give a bound involving the Marcinkiewicz norm, precisely \[ \begin{aligned} \frac12\int_V&|\nabla_A u|^2+\frac1{2\varepsilon^2}(1-|u|^2)^2+r^2(\text{curl} A)^2 +\pi\sum|d_B|^2\\ &\geq C \|\nabla_A\|^2_{L^{2,\infty}(V)}+\pi \sum|d_B|(\log\frac r{\varepsilon \sum|d_B|}-C), \end{aligned} \] where the sum is taken over all balls \(B\) in the above-mentioned collection. Various consequences are pointed out and a sequel paper in which the ideas are extended to an applied magnetic field is announced.