CAREER: Reduced Scaling Multireference Perturbation Theory with Polarizable Multipole Force Field Embedding for Simulating Photoreactions in Biological Environments

Chenchen Song of University of California, Davis is supported by an award from the Chemical Theory, Models and Computational Methods program in the Division of Chemistry to develop new theoretical methods for simulating photoreactions in life. Since the early days of chemistry, it was well known that light energy can be harnessed to drive chemical reactions in ways that are not possible using other energy sources such as heat or electricity.

Photoreactions are at the heart of essential life functions such as vision and vitamin D synthesis in humans, and photosynthesis and phototropism in plants. In these reactions, specialized proteins control the chemical environment around the light absorbing molecules and steer the reactions toward precise outcomes. Improving our understanding of photoreactions in biological environments will not only advance medical and agricultural applications, but will also assist in developing new technologies such as organic solar cells and photocatalysts.

However, progress towards these outcomes has thus far been limited by a lack of understanding of how photoexcited molecules behave in protein environments. To address this need, Dr. Song and her research group will develop both novel excited state quantum chemistry methods as well as new classical embedding approaches.

These combined approaches will make it possible to simulate the behavior of photoexcited molecules with spatial and temporal resolutions difficult to access experimentally. Simulation results from the research will be incorporated as course materials to introduce photochemistry in general chemistry, and the computational algorithms will be taught in a graduate level quantum chemistry course.

Dr. Song will develop new theoretical methods to study photoreactions in life through a multi-scale QM/MM description in which the chromophore is treated quantum mechanically and surroundings are treated classically. This will involve developments in excited state quantum chemistry methods as well as in classical embedding approaches.

The excited chromophore will be described with multi-state multi-reference second order perturbation theory (MSPT2), one of the gold standards for photochemistry. To make MSPT2 feasible for large systems, the computational prefactor will be reduced through parallelization and efficient implementation on graphical processing units (GPUs), and the computational scaling will be reduced with the supporting subspace factorization.

This strategy will be extended to the evaluation of forces, nonadiabatic couplings, and properties (e.g. transition dipole moments and spin-orbit couplings). The biological environment will be treated with a polarizable all-atom force field (AMOEBA) that can capture a wide range of environmental effects important for photoreactions. To embed MSPT2 in the AMOBEA model, an extended dynamic weight scheme will be used to capture the nonequilibrium polarization effects from the protein environment.

New developments will also be made to enable non-adiabatic molecular dynamics simulations on the MSPT2/AMOEBA potential energy surfaces using generalized ab initio multiple spawning. The significant reduction of computational cost will make it possible to perform these simulations on individual desktops and commodity-grade servers, which will be much more affordable and accessible to a broader range of research groups doing either theoretical or experimental studies.

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.