On Minima of Difference of Epstein Zeta Functions and Exact Solutions to Lennard-Jones Lattice Energy

Abstract

Let $\zeta(s,z)=\sum_{(m,n)\in\mathbb{Z}^2\backslash\{0\}}\frac{(\Im(z))^s}{|mz+n|^{2s}}$ be the Eisenstein series/Epstein Zeta function. Motivated by widely used Lennard-Jones potential \begin{equation}\aligned\nonumber \mathcal{V}(|\cdot|^2):=4\varepsilon\Big( (\frac{\sigma}{|\cdot|})^{12}-(\frac{\sigma}{|\cdot|})^{6} \Big), \endaligned\end{equation} in physics, in this paper, we consider the following lattice minimization problem \begin{equation}\aligned\nonumber \min_{z\in\mathbb{H}}\Big(\zeta(6,z)-b\zeta(3,z)\Big), \;\;b=\frac{1}{\sigma^6} \endaligned\end{equation} and completely classify the minimizers for all $b\in \R$. Our results resolve an open problem in Blanc-Lewin \cite{Bla2015}, and a conjecture by B\'etermin \cite{Bet2018}. Furthermore, our method of proofs works for general minimization problem \begin{equation}\aligned\nonumber \min_{z\in\mathbb{H}}\Big(\zeta(s_1,z)-b\zeta(s_2,z)\Big), \;\;s_1>s_2>1 \endaligned\end{equation} which corresponds to general Lennard-Jones potential \begin{equation}\aligned\nonumber \mathcal{V}(|\cdot|^2):=4\varepsilon\Big( (\frac{\sigma}{|\cdot|})^{2s_1}-(\frac{\sigma}{|\cdot|})^{2s_2} \Big),\;\;s_1>s_2>1. \endaligned\end{equation}