On the first critical field in the three dimensional Ginzburg-Landau model of superconductivity (Q1737563)

The paper is devoted to the definition of a precise estimate of the strength of the applied magnetic field called the first critical field for type-II superconductors at which vortices arise. The author characterizes the behavior of the global minimizers of the Ginzburg-Landau (GL) functional below this value in three dimensions. The first critical field $H_{c1}$ is rigorously defined by the fact that below it, the global minimizers of the Ginzburg-Landau functional do not have vortices, while they do for applied fields whose strength is higher than its value. In the 2D setting, E. Sandier and S. Serfaty have provided an expansion of the first critical field, up to an error $o(1)$ as $\varepsilon \to 0$ (where $\varepsilon > 0$ is the inverse of the Ginzburg-Landau parameter usually denoted by $\kappa$, a nondimensional parameter depending only on the material; in this paper, the regime of small $\varepsilon$ is considered, corresponding to extreme type-II superconductors). They rigorously characterized the behavior of the global minimizers of the GL functional below and near this value. The paper performs a 3D consideration of this problem. First, the author recalls the 3D $\varepsilon$-level estimates for the Ginzburg-Landau functional provided by the authors before and playing a crucial role in the paper. The main results of the paper are based on the divergence-free assumption on the applied magnetic field, which is in accordance with the fact that magnetic monopoles do not exist in Maxwell's electromagnetism theory. Then, a special vortexless configuration is defined that turns out to be a good approximation of the so-called Meissner solution, i.e., the vortexless global minimizer of the GL energy below the first critical field $H_{c1}$, which is unique up to a gauge transformation. The divergence-free vector field is the analogue of the function considered by E. Sandier and S. Serfaty in the analysis of the first critical field in 2D. It is shown that the divergence-free vector field plays an important role in the author's 3D analysis. This approach allows to split the GL energy of a given configuration $(u, A)$. One of the achievements of this paper is the determination of the right pair to split the energy, which then allows one to implement (almost) the same strategies as in 2D. By combining this splitting with the optimal $\varepsilon$-level estimates, an inequality for the GL functional is found, where the 1-current is written in series, where the sum is understood in the sense of currents. Inserting this expression in the inequality allows to heuristically derive the leading order of the first critical field. The first main result (Theorem 1.2) characterizes the behavior of the global minimizers of the GL functional below the leading order of the first critical field. In the 2D setting, a similar result was proved by E. Sandier and S. Serfaty. In the proof of Theorem 1.2, the fact that solutions of the GL equations in the Coulomb gauge satisfy a clearing-out result proved by \textit{D. Chiron} [Commun. Contemp. Math. 7, No. 5, 597--648 (2005; Zbl 1124.35081)] is used. The proof given by Chiron [loc. cit.] relies on monotonicity formulas, and is very much inspired by previous work by \textit{F. Bethuel} et al. [J. Eur. Math. Soc. (JEMS) 6, No. 1, 17--94 (2004; Zbl 1091.35085)]. The second main result provides bounds from above and below for the first critical field in three dimensions (Theorem 1.3). These inequalities attain the value of the leading order of the first critical field as $\varepsilon \to 0$. The next main result (Theorem 1.4) shows that, beyond the first critical field $H_{c1}$, there exists a locally minimizing vortexless configuration. A similar result was proved by S. Serfaty in 2D. The last main result concerns the uniqueness, up to a gauge transformation, of locally minimizing vortexless configurations (Theorem 1.5). This uniqueness result allows one to conclude that the locally minimizing configuration of Theorem 1.4 is, indeed, up to a gauge transformation, the unique global minimizer of the GL energy below the first critical field. Therefore, Theorem 1.4 provides a detailed characterization of the behavior of the Meissner solution. Let us also note that, if $\alpha \geq 1/4$ for a uniform applied field $hex = o(\varepsilon^{-\alpha})$, then the strategy of the proof fails. For this reason, the uniqueness of the locally minimizing vortexless configuration of Theorem 1.4 could be guaranteed only if $\alpha < 1/4$. Thus, the author proves that below the first critical field, up to a gauge transformation, the Meissner solution is the unique global minimizer of GL functional. Beyond this value, at least up to $hex = o(\varepsilon^{-1/3})$, a Meissner-type solution continues to exists as a local minimizer of the Ginzburg-Landau energy. This solution is unique, up to a gauge transformation, at least up to $hex = o(\varepsilon^{-1/4})$. Since this branch of vortexless solutions remains stable in the process of raising $hex$, vortices should not appear at the first critical field $H_{c1}$, but rather at a critical value of hex called the superheating field $H_{sh}$, at which the Meissner-type solution becomes unstable. It is expected that $H_{sh} = O(\varepsilon^{-1})$.