Non-equilibrium divide-and-conquer method for nanoscale simulations

点击翻译按钮获取中文摘要

In this project funded by the Chemical Theory, Models and Computational Methods program in the Division of Chemistry, Professor Michael Galperin, of the University of California, San Diego, is developing new methods to provide the theoretical framework to simulate the behavior of nanoscale devices. Interest in such systems stems from their technological promise and fundamental physical properties. The goal of the project is to resolve one of the central challenges of theoretical chemistry: accurate prediction of the electronic properties of nanoscale systems and their response to external perturbations. This can lead to an understanding of such devices at the molecular level with predictive theory. The results are central in developing memory and logic molecular devices, molecule-based sensors, photovoltaics for solar conversion and new types of electronics. The ability of the research to interconnect different scientific communities allows students involved in the project to participate in a flexible curriculum, thus preparing them to work in either fundamental or applied sciences, in academy or in national labs, in one of the fields or at interfaces of many-body theory, molecular spectroscopy, and computational quantum chemistry.Theoretical simulations of nanoscale open non-equilibrium systems mostly employ single-particle basis functions (e.g. orbitals as in the non-equilibrium Green function density functional theory (NEGF-DFT) approach). These approaches allow treatment of systems of realistic sizes, but become inconvenient in treating strong intra-system interactions. An alternative approach, the non-equilibrium atomic limit (e.g. the pseudoparticle (PP) or Hubbard NEGF), utilizes many-body states of the isolated system as a basis, which allows accounting for the system interactions. The applicability of such formulations is limited to relatively small systems. The project develops a methodology, the non-equilibrium divide-and-conquer method, which capitalizes on the strong sides of both methodologies: the NEGF-DFT ability to treat big systems and the PP and Hubbard NEGF ability to account for strong local interactions exactly. Partitioning of the original system (either in real or state space) is utilized with the parts treated by either single-particle or many-body states approaches. This extends the usual divide-and-conquer methodology to the realm of open non-equilibrium quantum systems. The results are central in developing memory and logic molecular devices, molecule-based sensors, photovoltaics for solar conversion and new types of electronics.