Semiclassical eigenvalue bounds for compact homogeneous irreducible Riemannian manifolds
Luigi Provenzano, Joachim Stubbe · 2025
Read Semiclassical eigenvalue bounds for compact homogeneous irreducible Riemannian manifolds by Luigi Provenzano, Joachim Stubbe (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-2503-02716.txt.
Introduction
Luigi Provenzano and Joachim Stubbe wrote a mathematics paper in two thousand twenty-five on vibration values for a special class of shapes. The shapes are compact homogeneous manifolds, spaces that look the same at every point. The values they study are eigenvalues of the Laplacian, which describe the tones such a shape would ring with.
The authors use an identity that links the slopes of the vibration patterns on these spaces. From it they derive bounds that stay sharp as the values grow large. They also get sharp bounds on averaged forms of the values, called Riesz means. Their route avoids the variational methods that are often used here.
Catalog record
| CREATOR | Luigi Provenzano, Joachim Stubbe |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 42 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-02716
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Semiclassical eigenvalue bounds for compact homogeneous irreducible Riemannian manifolds Luigi Provenzano, Joachim Stubbe Abstract . We exploit an identity for the gradients of Laplacian eigenfunctions on compact homogeneous Riemannian manifolds with irreducible linear isotropy group to obtain asymptotically sharp universal eigenvalue inequalities and sharp Weyl bounds on Riesz means. The approach is non variational and is based on identities for spectral quantities in the form of sum rules. 1991 Mathematics Subject Classification 58C40, 58J50, 35P15, 35P20 1. Introduction and statement of the main results Let \mathcal{M} be a d -dimensional, compact homogeneous Riemannian manifold, and let \Delta be the (negative) Laplacian on \mathcal{M} . The spectrum of -\Delta consists of non-negative eigenvalue levels (1.1) 0=\Lambda_{0}<\Lambda_{1}<\cdots<\Lambda_{l}<\cdots\nearrow\infty with multiplicities (1.2) 1=M_{0}<M_{1}\leq\cdots\leq M_{l}\leq\cdots An orthonormal basis of the corresponding eigenspaces consists of M_{l} eigenfunctions in L^{2}(\mathcal{M}) which we denote by Y_{l}^{k} , k=1,\ldots M_{l} . We may assume that Y_{l}^{k} are real-valued. In [ 6 ] the following addition formula was shown: for x,y\in\mathcal{M} the function (1.3) Z_{l}(x,y)=\sum_{k=1}^{M_{l}}Y_{l}^{k}(x)Y_{l}^{k}(y) is a zonal eigenfunction of the Laplacian in the corresponding eigenspace. In particular, (1.4) Z_{l}(x,x)=\sum_{k=1}^{M_{l}}|Y_{l}^{k}(x)|^{2}=\frac{M_{l}}{|\mathcal{M}|}, where |\mathcal{M}| is the volume of \mathcal{M} . The second identity in ( 1.4 ) expresses the homogeneity of the manifold on the spectral level: the density of a Laplacian multiplet is constant over the manifold. Formula ( 1.4 ) has several consequences. For example, taking the Laplacian on both sides of ( 1.4 ) we get (1.5) \sum_{k=1}^{M_{l}}|\nabla Y_{l}^{k}(x)|^{2}=\frac{M_{l}\Lambda_{l}}{|\mathcal{M}|} where \nabla denotes the gradient operator and |V| the length of the vector V in the tangent space at a point x\in\mathcal{M} . In the present note we consider compact homogeneous Riemannian manifolds for which the linear isotropy group is irreducible on the tangent space (see Definition A.2 ) and investigate the effect of this property at a spectral level. If \mathcal{M} is a compact, rank one symmetric space (CROSS), it falls in this class. We recall that CROSS are all the compact isotropic spaces, namely the spaces for which the stabilizer of x (the subgroup of isometries fixing x ) acts transitively on T_{x}\mathcal{M} , see e.g., [ 19 , §8.12] . While homogeneity means that the manifold looks the same at each point, isotropy means that it looks the same in any direction (at any point). Hence isotropy implies homogeneity, which is a well-known result. The manifolds that we consider, however, satisfy a weaker assumption than isotropy (which is quite restrictive). Roughly speaking, for a d -dimensional compact homogeneous Riemannian manifold with irreducible isotropy group, given any direction, there exist at least other d-1 linearly independent directions along which the space looks the same. This vague statement can be clarified by means of the simple example of the 2 -dimensional flat torus for which we present some explicit computations in Appendix B . In fact, any flat torus is homogeneous but not isotropic. Nevertheless, the square torus and the equilateral torus are isotropy irreducible (they are the only such tori). The crucial observation leading to our main result (Theorem 1.2 ) is the following “addition-type formula”, or identity on tangent spaces, which holds for compact homogeneous spaces with irreducible linear isotropy group. Theorem 1.1 . Let \mathcal{M} be an isotropy irreducible compact Riemannian manifold of dimension d . Let x\in\mathcal{M} and V\in T_{x}\mathcal{M} . If we denote by \langle\cdot,\cdot\rangle the scalar product in the tangent space T_{x}\mathcal{M} induced by the Riemannian metric on \mathcal{M} then (1.6) \sum_{i=1}^{k}\langle V,\nabla Y_{l}^{k}(x)\rangle\nabla Y_{l}^{k}(x)=\frac{M_{l}\Lambda_{l}}{d|\mathcal{M}|}V. In particular, (1.7) \sum_{k=1}^{M_{l}}|\langle V,\nabla Y_{l}^{k}(x)\rangle|^{2}=\frac{M_{l}\Lambda_{l}}{d|\mathcal{M}|}\,|V|^{2}. Let us briefly comment on identity ( 1.6 ). It states that the gradients of the eigenfunctions of any orthonormal basis of an eigenspace form a tight frame for T_{x}\mathcal{M} . We remark that identity ( 1.6 ) is related with the theory of minimal isometric immersions of compact manifolds into spheres using eigenfunctions as coordinates, see e.g., [ 16 ] (see also [ 12 ] ). It is worth noting that an analogous identity holds in \mathbb{R}^{d} , even though the spectrum is not discrete. In fact, the spectrum of the Laplacian in \mathbb{R}^{d} is [0,+\infty) , and for \lambda>0 a family of generalized eigenfunctions is given by u_{\lambda,\xi}(x)=e^{i\sqrt{\lambda}\langle\xi,x\rangle} , with \xi\in\mathbb{R}^{d} , |\xi|=1 . We have then \int_{\mathbb{S}^{d-1}}|f_{\lambda,\xi}(x)|^{2}\,d\xi=|\mathbb{S}^{d-1}|, for all x\in\mathbb{R}^{n} (this is reminiscent of the classic addition formulas, in particular, of ( 1.4 )). Now, for a vector V\in\mathbb{R}^{d} we compute \int_{\mathbb{S}^{d-1}}\langle V,\nabla f_{\lambda,\xi}\rangle\overline{\nabla f}_{\lambda,\xi}\,d\xi=\frac{|\mathbb{S}^{d-1}|\lambda}{d}V, which is the analogous of ( 1.6 ). We include a proof of Theorem 1.1 in Appendix A . The use of Theorem 1.1 was motivated by the trace identities for eigenvalues of the Laplacian or more general operators H with purely discrete spectrum derived in [ 9 , 10 ] which the authors called sum rules refering to early quantum mechanics (see e.g., [ 2 , 9 ] ). The basic idea of [ 9 , 10 ] is that of exploiting algebraic relations among the first and second commutators of H with an auxiliary self-adjoint operator G , which is usually a multiplication operator by a suitable function (which by abuse of notation we still denote by G ). Concretely, if H is the Dirichlet Laplacian on a domain \Omega of \mathbb{R}^{d} , exploiting the trace identities in [ 10 ] one ends up with terms of the form \displaystyle\int_{\Omega}|\langle\nabla\phi,\nabla G\rangle|^{2}\,dx , where \phi is a Dirichlet Laplacian eigenfunction on \Omega . One then chooses G=x_{j} , where x_{j} , j=1,...,d are the Cartesian coordinate functions, and averages over j . As we shall see below, in our situation the natural choice (analogous to the Euclidean case) will be to take G=Y_{l}^{k} as coordinates and to average over the multiplet Y_{l}^{k} , k=1,...,M_{l} . At this point formula ( 1.7 ) will be crucial. The trace identities found in [ 9 ] lead to quadratic inequalities of the form of ( 1.11 ) here below. In turn, these kind of quadratic inequalities imply asymptotically sharp universal eigenvalue bounds and, by the discovery made in [ 8 ] , sharp Weyl-type bounds for Riesz-means or sharp Lieb-Thirring inequalities [ 15 ] for Schrödinger operators. Sum rules for manifolds immersed in the Euclidean space were first derived in [ 4 , 7 ] and later in [ 10 ] , where also an algebraic identity for the spectrum of an abstract selfadjoint operator H defined on a Hilbert space \mathcal{H} was shown. In [ 4 ] the authors discuss universal inequalities derived from sum rules on compact manifolds isometrically immersed in \mathbb{R}^{n} . In the present note we derive universal inequalities for compact homogeneous irreducible Riemannian manifolds, which can be isometrically immersed in \mathbb{S}^{n} by eigenfunctions, and for their domains. We can now discuss the main application of Theorem 1.1 . Let \Omega be a domain in \mathcal{M} and let \lambda_{k}=\lambda_{k}(\Omega) be the k -th eigenvalue of the Dirichlet Laplacian on \Omega . When \Omega=\mathcal{M} (and we have no boundary) the eigenvalues \lambda_{k}(\mathcal{M}) are just the Laplace eigenvalues on \mathcal{M} and we have, according to our previous notation, (1.8) 0=\lambda_{1}(\mathcal{M})<\lambda_{2}(\mathcal{M})=\cdots=\lambda_{M_{1}+1}(\mathcal{M})<\lambda_{M_{1}+2}(\mathcal{M})=\cdots with \lambda_{2}(\mathcal{M})=\Lambda_{1} , \lambda_{M_{1}+1}(\mathcal{M})=\Lambda_{2} , etc. In the following we shall omit the dependence of \lambda_{k} on \Omega when unnecessary. We define the quadratic polynomials P_{N}(z) and Q_{N}(z) by (1.9) \begin{split}P_{N}(z)&=\sum_{j=1}^{N}(z-\lambda_{j})(z-\Lambda_{1}-\frac{d+4}{d}\,\lambda_{j})\\ &=Nz^{2}-2\,\frac{d+2}{d}\big(\sum_{j=1}^{N}\lambda_{j}\big)z-\Lambda_{1}Nz+\frac{d+4}{d}\,\sum_{j=1}^{N}\lambda^{2}_{j}+\Lambda_{1}\sum_{j=1}^{N}\lambda_{j},\\ \end{split} (1.10)