Tölle Johannes, Niemeyer Niklas, Neugebauer Johannes
Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125, United States.
University of Münster, Organisch-Chemisches Institut and Center for Multiscale Theory and Computation, Corrensstraße 36, Münster 48149, Germany.
J Chem Theory Comput. 2024 Mar 12;20(5):2022-2032. doi: 10.1021/acs.jctc.3c01264. Epub 2024 Mar 4.
We present a simple and accurate GW implementation based on a combination of a Laplace transform (LT) and other acceleration techniques used in post-self-consistent field quantum chemistry, namely, natural auxiliary functions and the frozen-core approximation. The LT-GW approach combines three major benefits: (a) a small prefactor for computational scaling, (b) easy integration into existing molecular GW implementations, and (c) significant performance improvements for a wide range of possible applications. Illustrating these advantages for systems consisting of up to 352 atoms and 7412 basis functions, we further demonstrate the benefits of this approach combined with an efficient implementation of the Bethe-Salpeter equation.
我们提出了一种基于拉普拉斯变换(LT)与自洽场后量子化学中使用的其他加速技术(即自然辅助函数和冻结核心近似)相结合的简单且精确的GW实现方法。LT-GW方法具有三个主要优点:(a)计算缩放的前置因子较小;(b)易于集成到现有的分子GW实现中;(c)对于广泛的可能应用具有显著的性能提升。通过对由多达352个原子和7412个基函数组成的系统展示这些优点,我们进一步证明了该方法与Bethe-Salpeter方程的高效实现相结合的好处。