Ab Initio Molecular Vibrations Enable Reliable Exciton Dynamics Simulations

Abstract

In 2007, Engel’s proposal of long-lived coherence in photosynthesis precipitated substantial efforts to understand the role of quantum effects in biology. While much progress has been made in the last decade, the timescale, type, and functional importance of coherence coupled to thermal vibrations remains poorly understood. In this dissertation, I simulate and disentangle excitation dynamics in the presence of complex vibrational environments for a light-harvesting protein and a synthetic dimer, both of which are controversial. I develop new techniques and leverage substantial computational resources to employ accurate ab initio methods that were previously considered prohibitively expensive. I examine phycobiliprotein PC645, a light-harvesting complex found in cryptophyte algae and, contrary to previous literature, I show that the protein’s key function proceeds via an incoherent mechanism. For the first time, I extract spectral densities from ab initio molecular dynamics (AIMD) combined with density functional theory excited state calculations. I obtain the first-ever quantitative reproduction of experimental linear spectra from first principles for a pigment-protein complex and find that stronger than expected environmental vibrations control the pathways of excitation transport. I generalize my findings using a model vibronic dimer, defining regimes of vibronic transport and demonstrating that biological systems are outside of the coherent regime. I then investigate three controversial synthetic fluorescein heterodimers that were engineered with the goal of providing a model system for understanding coherent energy transfer in multichromophoric systems. I extract Hamiltonians and spectral densities using AIMD and find electronic couplings an order of magnitude larger than previously thought. I calculate coherence dynamics and a trajectory of 2D spectra and demonstrate that electronic coherence persists for just 100-150 femtoseconds, after which long-lived vibrationally driven coherences persist for hundreds of femtoseconds. Thus, I resolve the controversy and confirm that while electronic coherence can be engineered in artificial molecular systems, it will not persist for a picosecond, as previously claimed. Leveraging ab initio methods, I have extracted mechanistic detail and clarified the presence and functional relevance of coherence in two controversial systems. However, significant experimental and computational challenges remain to enable the rational design of next-generation materials for light-harvesting and energy transport.