Modeling Kinetics and Dynamics of Spin-Dependent Processes

Understanding the kinetics and dynamics of spin-dependent processes involving transitions between electronic states of different spin multiplicities is critical for the development of artificial photochemical systems, low-cost transition metal catalysts, and molecular magnets for quantum information science applications. Spin-dependent processes can be modeled using either potentially very accurate but computationally expensive nonadiabatic molecular dynamics,¹ or more approximate but also much more computationally efficient nonadiabatic statistical methods.² We will discuss the advantages and disadvantages of these two approaches when applied to fast and slow spin-dependent processes. 

First, we will introduce our implementation of the computationally efficient nonadiabatic statistical theory (NAST) that is well suited for predicting the rates of slow spin transitions.² NAST relies on the information about molecular properties at only a few points on the potential energy surfaces, and therefore, can be coupled with very accurate high-level electronic structure methods. Accounting for the quantum effects, such as tunneling and zero-point vibration energy is relatively straightforward.³ The molecular properties required by NAST, including electronic energies, energy gradients, hessians, and spin-orbit couplings between the states are obtained from electronic structure calculations. We will highlight the modular structure of our NAST code and the computational tools used to loosely couple this code with different electronic structure packages. We will also describe several applications of NAST to predict the rates of intersystem crossings and “spin-forbidden” reactions in complex systems.^3-5

Second, we will discuss the generalized ab initio multiple spawning (GAIMS) nonadiabatic molecular dynamics, which is suitable for modeling fast intersystem crossing transitions.^6-8 This method, coupled with the GPU accelerated TeraChem electronic structure code, can be used to model small- and moderate-size molecular systems, accounting for the multidimensional effects that are missing in NAST. We will compare the intersystem crossing rate constants obtained with the GAIMS molecular dynamics and NAST for the small GeH₂ molecule.⁸ Finally, we will demonstrate the application of GAIMS-TeraChem to the excited states relaxation in 2-cyclopentenone, involving multiple intersystem crossings and internal conversions.

 

¹S. Mukherjee, D.A. Fedorov, S.A. Varganov, Annu. Rev. Phys. Chem. 72, 515-540 (2021)

²A.O. Lykhin, D.S. Kaliakin, G.E. dePolo, A.A. Kuzubov, S.A. Varganov, Int. J. Quantum Chem. 116, 750-761 (2016)

³A.O. Lykhin, S.A. Varganov, Phys. Chem. Chem. Phys. 22, 5500-5508 (2020)

⁴A.J. Pohlman, D.S. Kaliakin, S.A. Varganov, S.M. Casey, Phys. Chem. Chem. Phys. 22, 16641-16647 (2020)

⁵D.S. Kaliakin, D.G. Fedorov, Y. Alexeev, S.A. Varganov, J. Chem. Theory Comput. 15, 6074-6084 (2019)

⁶B.F.E. Curchod, C. Rauer, P. Marquetand, L. González, T.J. Martínez, J. Chem. Phys. 144, 101102 (2016)

⁷D.A. Fedorov, S.R. Pruitt, K. Keipert, M.S. Gordon, S.A. Varganov,  J. Phys. Chem. A 120, 2911-2919 (2016)

⁸D.A. Fedorov, A.O. Lykhin, S.A. Varganov, J. Phys. Chem. A 122, 3480-3488 (2018)