Purely magnetic tunneling conjecture for elliptical domains

From papers

Let Ω\Omega be an ellipse with boundary Γ\Gamma of length 2L2L, and let ss denote the curvilinear coordinate. Let λ1(h)\lambda_1(h) and λ2(h)\lambda_2(h) be the first two eigenvalues, let κ(s)\kappa(s) be the boundary curvature, and write κmax\kappa_{\max} and κmin\kappa_{\min} for its maximum and minimum. Let C1C_1, μ1(ξ0)\mu_1”(\xi_0), and ξ0\xi_0 be the constants from the magnetic boundary spectral reduction, and set

k2=κ(L2),γ0=ΩΓ=Ω2L.k_2=-\kappa”\left(\frac L2\right),\qquad \gamma_0=\frac{|\Omega|}{|\Gamma|}=\frac{|\Omega|}{2L}.

Define

S=2C1μ1(ξ0)L2L2κmaxκ(s)ds,\mathsf{S}=\sqrt{\frac{2C_1}{\mu_1”(\xi_0)}}\int_{-\frac L2}^{\frac L2}\sqrt{\kappa_{\max}-\kappa(s)}\,\mathrm{d}s,

and

A=exp([L2,L]sκmaxκ(s)k22κmaxκ(s)ds).\mathsf{A}=\exp\left(-\int_{[\frac L2,L]}\frac{\partial_s\sqrt{\kappa_{\max}-\kappa(s)}-\sqrt{\frac{k_2}{2}}}{\sqrt{\kappa_{\max}-\kappa(s)}}\,\mathrm{d}s\right).

The points s=L2s=-\frac L2 and s=L2s=\frac L2 correspond to the right and left points of maximal curvature, respectively. Purely magnetic tunneling conjecture. There exists α0R\alpha_0\in\mathbb{R} such that

λ2(h)λ1(h)=h138A252C134π(k2μ1(ξ0))14(κmaxκmin)12cos(L(γ0hξ0h12α0))eS/h14+o(h138)eS/h14as h0.\lambda_2(h)-\lambda_1(h)=h^{\frac{13}{8}}\mathsf{A}\frac{2^{\frac52}C_1^{\frac34}}{\sqrt{\pi}}\left(k_2\mu_1”(\xi_0)\right)^{\frac14}\left(\kappa_{\max}-\kappa_{\min}\right)^{\frac12}\left|\cos\left(L\left(\frac{\gamma_0}{h}-\frac{\xi_0}{h^{\frac12}}-\alpha_0\right)\right)\right|\mathrm{e}^{-\mathsf{S}/h^{\frac14}}+o\left(h^{\frac{13}{8}}\right)\mathrm{e}^{-\mathsf{S}/h^{\frac14}}\quad\text{as }h\to0.

This gives an explicit asymptotic formula for the exponentially small splitting between the first two eigenvalues in a purely magnetic tunneling problem for an ellipse. The formula was numerically checked in the paper and is recalled from an earlier conjecture; the theoretical formula is presented as an open question rather than a proved result.

Progress summary

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Sources & referencesView supporting material

Primary source

Virginie Bonnaillie-Noël, Frédéric Hérau and Nicolas Raymond, “Purely magnetic tunneling effect in two dimensions”, arXiv:1912.04035 (2021).

Solutions 0

No solutions have been posted yet.