CDS&E: Bridging electronic structure theory and dynamics with applications to nonadiabatic processes

点击翻译按钮获取中文摘要

Richard Dawes of the Missouri University of Science and Technology is supported by an award from the Chemical Theory, Models and Computational Methods program for his work on methods that create highly accurate potential energy surfaces (PES). This award is cofunded by the Computational and Data-Enabled Science and Engineering program in the Division of Advanced Cyberinfrastructure. The PES are central to how chemists think about the structure and dynamics of molecular systems, in terms of minima and asymptotes connected by paths across landscapes or over energetic barriers. The potential energy surfaces that Dawes and his coworkers create are used to study molecular dynamics and spectroscopy data that is relevant to many chemical processes including atmospheric chemistry, astrochemistry, and combustion. A major thrust of the project is to develop a user-friendly, freely-distributed software package to support and enable members of a wider community, both experimentalists and theorists, to create their own potential energy surfaces to study a large variety of different systems. The goal of this project is to develop user-friendly software that acts as a bridge between codes that perform electronic structure calculations and those that compute the dynamics of interest. The idea is to represent the electronic structure data with such high-fidelity that the computed dynamics directly reflect the underlying level of electronic structure calculations. (The fit only serves to render affordable the millions of potential evaluations required by dynamics codes). Very high fitting accuracy is required in order to assess the performance of subtly different methods or the importance of small correction terms. Previous work by Professor Dawes has resulted in suitably accurate interpolative methods known as interpolating moving least squares (IMLS). The automated fitting methods that were developed have reduced the bottleneck of PES generation and produced a stockpile of spectroscopically accurate PES. The current plan is to develop a generalized, extendable software distribution that others in the community can use to generate and fit their own PES. The particular focus of this project is on processes involving several coupled PES. This involves systematic studies of systems which include various non-adiabatic interactions between electronic states such as conical intersections, Renner-Teller interactions, and spin-orbit coupling. Accurate treatment of these effects in real applications represents the frontier of theoretical dynamics.