Modeling Exciton Transport in Organic Semi-conductors Using Machine Learned Hamiltonian and its Gradients

In this study, we present a multiscale method to simulate the propagation of Frenkel singlet excitons in Organic Semi-conductors(OSCs). The approach uses advanced neural network models to train Frenkel-type Hamiltonian and its gradient, obtained by the Long-Range Correction version of Density Functional Tight-Binding with Self-Consistent Charges (LC-DFTB2). Our models accurately predict site energies, excitonic couplings, and corresponding gradients, essential for the non-adiabatic molecular dynamics simulations. Combined with Fewest Switches Surface Hopping (FSSH) algorithm, the method was applied to four representative OSCs: Anthracene (ANT), Pentacene (PEN), Perylenediimide (PDI), and Diindenoperylene (DIP). The simulated exciton diffusion constants align well with experimental and reported theoretical values, and offer valuable insights into exciton dynamics in OSCs.