pyUNIxMD: Python-based excited state molecular dynamics simulation package with multiple algorithms

Abstract

When multiple adiabatic electronic states are coupled to nuclear degrees of freedom, Born-Oppenheimer (BO) approximation breaks down. Especially when a molecular system absorbs a photon or multiple photons, we should treat nuclear motion beyond BO approximation to investigate nuclear wave packet splitting as well as quantum coherence. Statistical analysis based on multiple classical nuclear trajectories shows a nice description of BO populations. However, incorrect account for electron-nuclear correlations gives a wrong description for quantum coherence. In this talk, I will present a python-based package for excited state molecular dynamics simulations, so-called pyUNIxMD, which can deal with multiple algorithms including Ehrenfest dynamics, various surface hopping algorithms with decoherence, and coupled-trajectory mixed quantum-classical dynamics aiming for describing correct nuclear propagation with quantum coherences. Based on recent developments on exact factorization of a molecular wave function, a rigorous derivation of a correct electron-nuclear correlation is available which enables us to deal with proper quantum coherence and nuclear wave packet splitting. This correct electron-nuclear correlation contains not only the conventional nonadiabatic couplings with classical nuclear momenta but also couplings with quantum mechanical features of nuclear density. The pyUNIxMD has a unique feature to deal with quantum momenta of nuclei for quantum coherence. In the first part of my talk, I will present an algorithm to account for coupling between quantum nuclei and electronic degrees of freedom based on coupled trajectories or independent trajectories [1] based on Born-Huang expansion in pyUNIxMD. In addition, I provide several examples of direct simulations for excited state molecular dynamics [2,3,4]. Furthermore, I propose a real-space real-time propagation method with correct electron-nuclear correlations in the framework of real-time time-dependent density functional theory toward extended molecular systems.