On the importance of initial conditions for excited-state dynamics

Photodynamical simulations are increasingly used to explore photochemical mechanisms and interpret laser experiments. The vast majority of ab initio excited-state simulations are performed within semiclassical, trajectory-based approaches. Apart from underlying electronic-structure theory, the reliability of simulations is controlled by a selection of initial conditions for the classical trajectories. We discuss appropriate choices of initial conditions for simulations of different experimental arrangements: dynamics initiated by continuum-wave (CW) laser fields or triggered by ultrashort laser pulses. We introduce a new technique, CW-sampling, to treat the former case, based on the ideas of importance sampling, combined with the quantum thermostat approach based on the Generalized Langevin Equation (GLE) that allows for efficient sampling of both position and momentum space. The CW-sampling is particularly important for photodynamical processes initiated by absorption at the tail of the UV absorption spectrum. We also emphasize the importance of non-Condon effects for the dynamics. We demonstrate the performance of our approach on the photodissociation of the CF 2 Cl 2 molecule (Freon CFC-12). A quantitative agreement with the experimental data is achieved with the use of empirical correlation energy correction (CEC) factor on top of FOMO-CASCI potential energy surfaces. The vast majority of ab initio excited-state simulations are performed within semiclassical, trajectory-based approaches. Apart from the underlying electronic-structure theory, the reliability of the simulations is controlled by a selection of initial conditions for the classical trajectories. We discuss appropriate choices of initial conditions for simulations of different experimental arrangements: dynamics initiated by continuum-wave (CW) laser fields or triggered by ultrashort laser pulses.