Direct Nonadiabatic Simulations of the Photoinduced Charge Transfer Dynamics

Abstract

We apply direct nonadiabatic dynamics simulations to investigate photoinduced charge transfer reactions. Our approach is based on the mixed quantum-classical fewest switches surface hopping (FSSH) method that treats the transferring electron through time-dependent density functional theory and the nuclei classically. The photoinduced excited state is modeled as a transferring single-electron that initially occupies the LUMO of the donor molecule/moiety. This single-particle electronic wave function is then propagated quantum mechanically by solving the time-dependent Schrodinger equation in the basis of the instantaneous molecular orbitals (MOs) of the entire system. The nonadiabatic transitions among electronic states are modeled using the FSSH approach within the classical-path approximation. We apply this approach to simulate the photoinduced charge transfer dynamics in a few well-characterized molecular systems. Our results are in excellent agreement with both the experimental measurements and high-level (yet expensive) theoretical results.

Ab initio nonadiabatic molecular dynamics of charge carriers in metal halide perovskites

“…Accelerated MD techniques, such as extended Lagrangian dynamics and force field simulations, can generate sufficiently long nuclear trajectories for NAC calculation and subsequent NA-MD. [244][245][246][247] Semiempirical approaches, i.e., the extended Hückel theory (EHT), 244,245,248,249 semi-empirical wavefunction methods, 81 and DFTB+ [250][251][252][253] notably accelerate the electronic structure calculations, compared to ab initio DFT, and have been adopted for semiclassical NA-MD simulations of systems comprised of 1000 atoms or more, with promising results achieved. Larger systems can be studied with fragment-based techniques, such as the divide-andconquer (DC) or fragment molecular orbital (FMO) methods developed by the groups of Nakai, 254 Akimov, 251 Kitaura, 255 Blumberger, 256 and Prezhdo, 257 through interfacing with different quantum chemistry codes.…”

“…103,110 Machine learning (ML) is a promising strategy to accelerate NA-MD simulations by predicting the excited state properties for longer timescales based on costeffective training sets, without involving underlying quantum mechanics calculations. 130,[249][250][251][252][253][254][255][256][257][258][259][260][261][262][263][264][265][266] The overall idea of QSH/ NA-MD is qualitatively similar to the ML/NA-MD simulation. So far, works on ML/NA-MD have focused primarily on molecular systems.…”

Tensor-Train Split-Operator KSL (TT-SOKSL) Method for Quantum Dynamics Simulations

“…Simulations of quantum phenomena in chemical and biological systems , typically require time-dependent methods. For example, photoinduced reactions, − as well as processes that involve energy transfer, electron transfer, − simulations of molecular spectroscopy, and coherent control, require rigorous descriptions of quantum effects, including tunneling, interference, entanglement, and nonadiabatic dynamics. , Simulations in the time-dependent picture require integration of the time-dependent Schrödinger equation (TDSE) explicitly, which can be efficiently performed for small molecular systems, for example, by using the split-operator Fourier transform (SOFT) method, which is numerically exact. − However, SOFT is limited to systems with very few degrees of freedom (DOFs) (i.e., molecular systems with less than four or six atoms), , since it is based on a full basis set representation requiring storage space and computational effort that scale exponentially with the number of coupled DOFs. Utilizing an adaptive grid that evolves simultaneously with the wavepacket, the capability of SOFT is extended to successfully treat the dynamics of an eight-dimensional Henon-Heiles model .…”

Tensor-Train Split Operator KSL (TT-SOKSL) Method for Quantum Dynamics Simulations

“…Simulations of quantum phenomena in chemical and biological systems 1,2 typically require time-dependent methods. For example, photoinduced reactions, [3][4][5][6] as well as processes that involve energy transfer, 7 electron transfer, [8][9][10][11][12] simulations of molecular spectroscopy, 13 and coherent control, 14 require rigorous descriptions of quantum effects, including tunneling, interference, entanglement, and non-adiabatic dynamics. 15,16 Simulations in the timedependent picture require integration of the time-dependent Schrödinger equation (TDSE) explicitly, which can be efficiently performed for small molecular systems for example by using the split-operator Fourier transform (SOFT) method which is a numerically exact method.…”