First principles simulations of optically activated processes in materials and molecules
Marco Govoni1,2
1Materials Science Division & Center for Molecular Engineering, Argonne National Laboratory
2Pritzker School of Molecular Engineering, University of Chicago
Abstract. —The robust description of excited states for complex heterogeneous systems is the cornerstone of a computational framework that enables the modelling of materials for sustainable energy and quantum information science applications [1]. I will present the simulation of optically activated processes in materials using a hierarchical modeling approach that relies on the combination of density functional theory, many-body perturbation theory, and multi-reference methods. I will discuss the use of orbital localization and machine learning to accelerate the calculation of optical spectra and exciton binding energies [2,3], and the validation of first principles simulations of photoluminescence spectra [4]. I will then discuss the use of a quantum embedding method to simulate strongly correlated excitations of defects in materials [5,6], e.g., analogs of the NV center in diamond [7]. Finally, I will discuss opportunities for computational materials science that are driven by emerging trends in the high-performance computing landscape, which include strategies to combine machine learning and quantum computing with electronic structure calculations on pre-exascale resources.
[1] M. Govoni et al., npj. Comput. Mater. 7, 32 (2021)
[2] N.L. Nguyen et al., Phys. Rev. Lett. 122, 237402 (2019)
[3] S. Dong et al., Chem. Sci. 12, 4970 (2021)
[4] Y. Jin et al., Phys. Rev. Mat. 5, 084603 (2021)
[5] H. Ma et al., J. Chem. Theory Comput. 17, 2116 (2021)
[6] H. Ma et al., npj Comput. Mater. 6, 85 (2020)
[7] H. Ma et al., Phys. Chem. Chem. Phys. 22, 25522 (2020)
Host: Walter Lambrecht