Collaborative Research: Elements: GPU-accelerated First-Principles Simulation of Exciton Dynamics in Complex Systems

University Of Illinois At Urbana-Champaign, Urbana IL

Investigators

Abstract

The development of societally-important energy harvesting approaches, such as photocatalysis or photovoltaics, requires detailed understanding of fast dynamics of electrons coupled with ions in novel materials. Modern supercomputers can help obtain such an understanding using sophisticated quantum-mechanical simulations. However, such a scientific effort requires accurate simulation techniques to be developed and efficient use of the underlying supercomputer hardware is crucial. This project meets these outstanding challenges by implementing novel techniques to describe the quantum-mechanical electron-electron interaction and interactions of electron dynamics with ions. Using and testing these new developments on graphical processing units advances science as it prepares quantum-mechanical simulations for next-generation supercomputers. Applying these advanced simulations to model complex systems of great importance for energy harvesting furthers the computational science community towards the goal of advancing national prosperity and welfare. The project makes these techniques freely available for a broad community, including documentation and tutorials, and trains the next generation of computational researchers through organizing summer schools and workshops. This project leverages a multi-organizational team to benefit from synergies that emerge from two possible solutions to current scientific barriers: Descriptions of exchange and correlation based on a long-range correction and on hybrid functionals that can scale favorably even for large systems with thousands of electrons are implemented and applied. Descriptions of non-adiabatic dynamics are implemented and applied to study long-term dynamics of excitons during which the interaction with the nuclei becomes important. Doing so within a cutting-edge electronic-structure code that runs efficiently on graphics processing units, provides a unique opportunity to compare accuracy, applicability to a broad range of systems, and computational cost. Knowledge on the reliability of these approximations and their computational cost for extended systems of practical relevance, including complex heterogeneous systems like semiconductor-molecule interfaces, are advanced by this research. The efforts include building, increasing, and growing a skilled community especially of US based researchers in a recurrent summer school. This proposal receives funds through the Office of Advanced Cyberinfrastructure in the Computer and Information Science and Engineering Directorate and the Division of Materials Research in the Mathematical and Physical Sciences Directorate. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.