Developing the multilayer multiconfiguration time-dependent Hartree theory

New Mexico State University, Las Cruces NM

Investigators

Abstract

Haobin Wang, of New Mexico State University is supported by an award from the Chemical Theory, Models and Computational method in the Chemistry Division and the Computational and Data-Enabled Science and Engineering (CDS&E) program to develop and apply computational approaches to study electron transfer processes in several technologically-important materials and devices such as dye sensitized solar cells, molecular junction devices, and new electronic/magnetic materials. Because electrons are very small objects, they obey the laws of quantum mechanics. To get a sufficiently accurate description of many of these systems, quantum mechanical methods must be applied to both the electrons and the atomic nuclei. Professor Wang and his colleagues have developed an accurate and efficient quantum mechanical approach that is designed for this purpose called the multilayer-multiconfiguration time-dependent Hartree (ML-MCTDH) theory. The fundamental understanding resulting from this research contributes to the rational design of new materials such as for solar energy conversion and molecular electronics. The numerical techniques and software developed in this project have potential applications to a broad range of problems in science and engineering, e.g., numerical methods, signal compression, and parallel computation. Both undergraduate and graduate students contribute to this research. With the extension of ML-MCTDH to the second quantization representation for treating identical particles, the theory is formally complete. The purpose of the current research is: (i) to investigate as many important physical models as possible that can be treated accurately via the current ML-MCTDH implementation; and (ii) to develop new algorithms within the ML-MCTDH framework to treat more general models. Specifically, the principal investigator and his coworkers apply the current theory to a variety of important and challenging problems in physics and chemistry, e.g., heterogeneous electron transfer using the Anderson-News model, spin relaxation where the electronic and nuclear spins are treated on equal footing, and the Kondo effect using the Anderson impurity model. Beyond those straightforward applications the current implementation of the ML-MCTDH theory is being extended to more general forms of potential energy functions. Further goals are to improve the relevant ML-MCTDH algorithms, e.g., using multilayer improved relaxation to generate the initial density matrix at low temperatures, grouping and streamlined evaluation of a series of operators, and modifying ML-MCTDH equations of motion that remove the regularization of the numerical singularities. The software developed in this research will be made freely available to the research community.