Laplacian eigenvalues for large negative Robin parameters on domains with outward peaks
Konstantin Pankrashkin, Firoj Sk, Marco Vogel · 2025
Read Laplacian eigenvalues for large negative Robin parameters on domains with outward peaks by Konstantin Pankrashkin, Firoj Sk, Marco Vogel (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-04996.txt.
Introduction
Konstantin Pankrashkin, Firoj Sk, and Marco Vogel study how Laplacian tone values behave on domains with outward peaks when a boundary weight goes large and negative. Peaks are sharp tips that change how waves live in a shape. Past work dealt with round tip cuts only.
Their results allow a wide class of tip cross shapes. Growth forms for each tone reflect both tip sharpness and cross shape in clean terms. The work widens past peak results to far more shapes. Such domains model tips in acoustics, heat, and quantum wells, where edge terms steer the tones.
Catalog record
| CREATOR | Konstantin Pankrashkin, Firoj Sk, Marco Vogel |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 100 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-04996
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Laplacian eigenvalues for large negative Robin parameters on domains with outward peaks Konstantin Pankrashkin, Firoj Sk, Marco Vogel Abstract . We study the asymptotic behavior of individual eigenvalues of the Laplacian in domains with outward peaks for large negative Robin parameters. A large class of cross-sections is allowed, and the resulting asymptotic expansions reflect both the sharpness of the peak and the geometric shape of its cross-section. The results are an extension of previous works dealing with peaks whose cross-sections are balls. 1991 Mathematics Subject Classification 35P15; 35P20; 47A75; 49R05; 58C40 1. Introduction For \Omega\subset\mathbb{R}^{d} a bounded domain (non-empty connected open set) with suitably regular boundary, we are interested in the following Robin eigenvalue problem: (1) \left\{\begin{aligned} -\Delta u&=\lambda u\text{ in }\Omega,\\ \partial_{\nu}u&=\alpha u\text{ on }\partial\Omega,\end{aligned}\right. where \alpha>0 is a parameter and \partial_{\nu} means the outer normal derivative. The problem ( 1 ) is understood in a weak sense. For a strict formulation we denote by \mathcal{H}^{m} the m -dimensional Hausdorff measure, and for the sake of readability we denote a function in \Omega and its Sobolev trace on \partial\Omega by the same symbol. Under appropriate regularity assumptions on \Omega the symmetric bilinear form r^{\Omega}_{\alpha} defined on the domain \mathcal{D}(r^{\Omega}_{\alpha})=W^{1,2}(\Omega) by r^{\Omega}_{\alpha}(u,v):=\int_{\Omega}\langle\nabla u,\nabla v\rangle_{\mathbb{R}^{d}}\,\mathrm{d}\mathcal{H}^{d}-\alpha\int_{\partial\Omega}uv\,\mathrm{d}\mathcal{H}^{d-1} is closed and generates a self-adjoint operator R^{\Omega}_{\alpha} with compact resolvent in L^{2}(\Omega) . Informally, the operator R^{\Omega}_{\alpha} corresponds to the Laplacian u\mapsto-\Delta u acting on the functions u satisfying the boundary condition \partial_{\nu}u=\alpha u on \partial\Omega , and the eigenvalues \lambda in ( 1 ) are understood as the eigenvalues of R^{\Omega}_{\alpha} . In particular, a number \lambda is an eigenvalue of ( 1 ) with an eigenfunction u if and only if r^{\Omega}_{\alpha}(u,v)=\lambda\int_{\Omega}uv\,\mathrm{d}\mathcal{H}^{d}\text{ for all }v\in W^{1,2}(\Omega). Alternatively, one can consider u\mapsto r^{\Omega}_{\alpha}(u,u) as the energy functional for ( 1 ) and characterize the eigenvalues variationally using the min-max principle. Remark that the boundary term in the above expression for r^{\Omega}_{\alpha}(u,u) is negative for \alpha>0 , so this case is usually termed as the case of negative Robin parameters. In what follows, for a lower semibounded self-adjoint operator A we denote by \lambda_{j}(A) its j -th eigenvalue (if it exists), assuming that the eigenvalues are enumerated in the non-decreasing order by taking into account their multiplicities. The goal of the present paper is to study the asymptotic behavior of the individual eigenvalues of R^{\Omega}_{\alpha} , i.e. of \lambda_{j}(R^{\Omega}_{\alpha}) with fixed j , for \alpha\to+\infty and a special class of domains \Omega . The dependence of \lambda_{j}(R_{\alpha}^{\Omega}) on \alpha for various classes of \Omega has been given a considerable attention during the last decade, see the reviews in [ 8 , 21 ] . While the above problem is linear, it originally appeared in the study of non-linear problems, see e.g. [ 12 ] for a stochastic interpretation, [ 16 ] for a discussion of an interplay between the Robin eigenvalues and critical temperature estimates in surface superconductivity, and [ 23 ] for a link with the study of reaction-diffusion equations. If \Omega is a bounded Lipschitz domain, then all required regularity assumptions are satisfied, and it is known that for large \alpha one has a two-sided bound (2) -c\alpha^{2}\leq\lambda_{1}(R_{\alpha}^{\Omega})\leq-\alpha^{2} with some c\geq 1 , see e.g. [ 8 , Prop. 4.12] and [ 21 , Lem. 2.7] , and the very recent paper [ 13 ] shows that no asymptotics of the form \lambda_{1}(R_{\alpha}^{\Omega})\sim-c_{\Omega}\alpha^{2} with c_{\Omega}>0 is available in general. Under stronger regularity assumptions one can construct detailed asymptotic expansions for \lambda_{j}(R^{\Omega}_{\alpha}) with any fixed j involving various geometric properties of \Omega and \partial\Omega , see e.g. [ 4 , 7 , 11 , 17 , 20 , 24 , 27 ] and the reviews in [ 8 , 21 ] . On the other hand, it was observed in [ 24 ] that ( 2 ) fails for domains \Omega with outward peaks, which was later studied in greater detail in [ 22 ] . As the subsequent text is specifically devoted to the study of \lambda_{j}(R^{\Omega}_{\alpha}) for such \Omega , let us introduce an adapted language in order to continue the discussion. From now on let d\in\mathbb{N} with d\geq 2 . The vectors x\in\mathbb{R}^{d} will be written in the form x=(x_{1},x^{\prime}) with x^{\prime}\in\mathbb{R}^{d-1} . The following definition is in the spirit of [ 25 , Sec. 5.1.1] , see Figure 1 for an illustration: Definition 1 . Let q>1 and \omega\subset\mathbb{R}^{d-1} be a bounded Lipschitz domain. We say that a bounded domain \Omega\subset\mathbb{R}^{d} has an outward peak at the origin, of sharpness order q with cross-section \omega , if for some \delta>0 it holds \Omega\cap(-\delta,\delta)^{d}=\Big\{x\in\mathbb{R}^{d}:\ x_{1}\in(0,\delta),\ x^{\prime}\in x_{1}^{q}\omega\Big\} and \Omega is Lipschitz at all boundary points except at the origin. x_{1} x^{\prime} \Omega x_{1}=c c^{q}\omega Figure 1. An example of a domain \Omega with an outward peak with cross-section \omega . The main result of the present work is as follows: Theorem 2 . Let \Omega\subset\mathbb{R}^{d} be as in Definition 1 with some bounded Lipschitz cross-section \omega\subset\mathbb{R}^{d-1} and sharpness order q\in(1,2) , then for any fixed j\in\mathbb{N} one has (3) \lambda_{j}(R^{\Omega}_{\alpha})=\left(\dfrac{\mathcal{H}^{d-2}(\partial\omega)}{\mathcal{H}^{d-1}(\omega)}\right)^{\frac{2}{2-q}}\lambda_{j}(L_{1})\alpha^{\frac{2}{2-q}}+O\left(\alpha^{\frac{2}{2-q}-(q-1)}\right)\text{ for }\alpha\to+\infty, where L_{1} is the differential operator in L^{2}(0,\infty) defined first by \big(L_{1}f\big)(s):=-f^{\prime\prime}(s)+\bigg(\dfrac{q^{2}(d-1)^{2}-2q(d-1)}{4s^{2}}-\dfrac{1}{s^{q}}\bigg)f(s),\quad f\in C^{\infty}_{c}(0,\infty), and then extended using the Friedrichs extension. The result on the individual eigenvalues is then used to show that the corresponding eigenfunctions are localized near the peak tip, which is rigorously described using an Agmon-type estimate as follows: Theorem 3 . Let the assumptions of Theorem 2 be fulfilled. For any j\in\mathbb{N} and b>0 there exist B>0 and \alpha_{0}>0 such that for any \alpha>\alpha_{0} and any eigenfunction u of R_{\alpha}^{\Omega} for the eigenvalue \lambda_{j}(R_{\alpha}^{\Omega}) one has (4) \int_{\Omega}e^{2b\alpha|x|}\big|u(x)\big|^{2}\,\mathrm{d}\mathcal{H}^{d}(x)\leq B\|u\|^{2}_{L^{2}(\Omega)}. Let us make some comments on the assumptions and the relations with earlier works. All eigenvalues \lambda_{j}(L_{1}) are negative, see Subsection 2.2 and Appendix A below, so the eigenvalues \lambda_{j}(R^{\Omega}_{\alpha}) diverge to -\infty much faster than in the Lipschitz case. We are not aware of any value of q\in(1,2) for which the eigenvalues of L_{1} can be computed explicitly. The assumption q<2 is necessary to ensure that the Sobolev trace operator is well-defined and compact from W^{1,2}(\Omega) to L^{2}(\partial\Omega) , which guarantees the required properties of the above bilinear form r^{\Omega}_{\alpha} and the self-adjointness of the Robin Laplacian R^{\Omega}_{\alpha} for all \alpha>0 , see [ 10 , Sec. 5.2] or [ 3 , 26 ] . At the same time, the condition q>1 is required to actually have a peak, as \Omega would be covered by the analysis of Lipschitz domains for q\leq 1 . The case when \omega is a ball of radius \rho>0 was already studied in the earlier paper [ 22 ] , and its main result is recovered (with an improved remainder estimate) by using ( 3 ) and observing that in this case \dfrac{\mathcal{H}^{d-2}(\partial\omega)}{\mathcal{H}^{d-1}(\omega)}=\dfrac{d-1}{\rho}. It should be noted that the analysis in [ 22 ] was crucially depending on a separation of variables in the ball (and lengthy manipulations with asymptotic expansions of special functions), which is indeed unavailable for general \omega . The central observation in this paper is that a part of the arguments of [ 22 ] based on a separation of variables can be replaced by an argument based on the first-order perturbation theory of linear operators