Two-term spectral asymptotics in linear elasticity on a Riemannian manifold

Genqian Liu

2024-01-01

Abstract:In this note, by explaining two key methods that were employed in \cite{Liu-21} and by giving some remarks, we show that the proof of Theorem 1.1 in \cite{Liu-21} is a rigorous proof based on theory of strongly continuous semigroups and pseudodifferential operators. All remarks and comments to paper \cite{Liu-21}, which were given by Matteo Capoferri, Leonid Friedlander, Michael Levitin and Dmitri Vassiliev in \cite{CaFrLeVa-22}, are incorrect. The so-called "numerical counter-examples" in \cite{CaFrLeVa-22} are useless examples for the two-term asymptotics of the counting functions of the elastic eigenvalues. Clearly, the conclusion and the proof of \cite{Liu-21} are completely correct.

Spectral Theory,Mathematical Physics,Analysis of PDEs,Differential Geometry

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is the spectral asymptotics of eigenvalue problems on Riemann manifolds in linear elasticity theory. Specifically, the author attempts to prove the two - order asymptotic behavior of the counting function of the eigenvalues $\tau_k$ of the Navier - Lamé operator $P_g$ on a compact smooth Riemann manifold $(\Omega, g)$ with a smooth boundary. The core result of the paper is Theorem 1.1, which gives the asymptotic expansion of the trace of eigenvalues when the time $t$ approaches 0:

\[ \sum_{k = 1}^{\infty}e^{-\tau_k t}=\text{Tr}(e^{-tP_g})=\left[ \frac{n - 1}{(4\pi\mu t)^{n/2}}+\frac{1}{(4\pi(2\mu+\lambda)t)^{n/2}} \right]\text{Vol}(\Omega) \]

\[ \mp\frac{1}{4}\left[ \frac{n - 1}{(4\pi\mu t)^{(n - 1)/2}}+\frac{1}{(4\pi(2\mu+\lambda)t)^{(n - 1)/2}} \right]\text{Vol}(\partial\Omega)+O(t^{1 - n/2}) \]

where:
- $\text{Vol}(\Omega)$ represents the $n$-dimensional volume of $\Omega$,
- $\text{Vol}(\partial\Omega)$ represents the $(n - 1)$-dimensional volume of $\partial\Omega$,
- $\mu$ and $\lambda$ are Lamé parameters, satisfying $\mu>0$ and $\mu+\lambda\geq0$.

In addition, the paper also discusses in detail two key methods: the "mirror method" and the "analytic symbol method", which are used to derive the above asymptotic expansion. The author proves Theorem 1.1 through these methods and refutes the doubts raised by other researchers in [CaFrLeVa - 22], believing that those so - called "numerical counterexamples" have no practical significance for the study of two - order asymptotic behavior.