"Simulating photochemistry in complex environments"
Chemical processes initiated by light—such as those underlying vision and photosynthesis—play a central role in chemistry and biology. Simulating the mechanisms of these processes requires an accurate representation of both the chromophores and their environment. While many methods have been developed for gas-phase simulations, adapting those methods to disordered, condensed systems requires a careful treatment of the environment.
Relaxation rates and quantum yields are often simulated by an ensemble of independent nonadiabatic dynamics trajectories. Accurate results depend on an accurate sampling of the initial conditions for the swarm of trajectories. I show how sloppy sampling methods can lead to erroneous dynamics in solvated systems and how to avoid those errors. I introduce an automated, iterative scheme for optimizing molecular mechanics (MM) force fields to match quantum mechanical (QM) forces. With optimized force fields, sampling can be carried out at the MM level.
The size of the light-harvesting complex LH2 presents unique challenges for photodynamics simulations. Treating the 27 chromophores (>2,000 atoms) with a GPU-accelerated, distributed exciton model permits ab initio calculation of excited states. The norm-preserving interpolation within ab initio multiple spawning (AIMS) models population transfer between those excited states during dynamics.
I present the results of three 1-ps AIMS trajectories on LH2. Atomistic simulations illuminate the structure-function relationship, revealing how chromophore vibrations enable energy transfer within LH2. A kinetic model based on the observed mechanism produces an efficiency and relaxation rate in agreement with experiment. The delocalized nature of the acceptor B850 band enables rapid, efficient energy transfer. With these insights, better artificial light-harvesting complexes can be designed.
Zoom Information:
Meeting ID: 936 4339 0205
Password: 879397