Structural Feature in Dynamical Processes Accelerated Transition State Calculations

TL;DR

This paper introduces an adaptive semi-rigid body approximation (ASRBA) that leverages chemical bond length similarities to accelerate transition state calculations in various dynamical processes, significantly reducing computational time.

Contribution

The study presents a novel ASRBA method that provides physically reasonable initial guesses for MEP structures, improving efficiency over traditional linear interpolation methods.

Findings

ASRBA produces robust transition state estimates across different systems.

Transition state calculations with ASRBA are significantly faster than existing methods.

The method maintains accuracy while reducing computational cost.

Abstract

Minimum energy path (MEP) search is a vital but often very time-consuming method to predict the transition states of versatile dynamic processes in chemistry, physics, and materials science. In this study, we reveal that the chemical bond lengths in the MEP structures, including those directly involved in the dynamical processes, largely resemble those in the stable initial and final states. Based on this discovery, we propose an adaptive semi-rigid body approximation (ASRBA) to construct a physically reasonable initial guess for the MEP structures, which can be further optimized by the nudged elastic band method. Examination of several distinct dynamical processes in bulk, on crystal surface, and through two-dimensional system show that the transition state calculations based on the ASRBA results are robust and significantly faster than those based on the popular linear interpolation and image-dependent pair potential methods.