Using the finite-difference time-domain pulse propagation method to simulate time-resolved THz experiments

Abstract

The finite-difference time-domain (FDTD) method has been applied to time-resolved THz spectroscopy (TRTS) experiments. Time-resolved THz spectroscopy utilizes an optical pump pulse to excite the sample, followed by a far-infrared (FIR) probe pulse with frequency components that span from 10 to 100 cm−1. The subpicosecond evolution of the FIR spectrum is obtained as a function of time after the visible photoexcitation event. Significant challenges arise in interpreting these experimental results due to the very different frequencies of the pump and probe pulses. Therefore, it is essential to simulate the experiment. The method described entails numerically propagating both the THz probe pulse and the visible pump pulse simultaneously, keeping track of the transiently induced polarization from absorption of the visible pulse. Group velocity mismatch between the visible and THz pulse and a transiently changing response function are completely accounted for in the calculation. Furthermore, a spatially varying polarization can be included to account for a nonuniform excited region of the sample under investigation. The response function of the material is described as a multimode Brownian oscillator that can describe dispersive media in a very general sense. In particular, the overdamped, underdamped, and critically damped cases are all included, as well as special cases such as a Debye or Drude response. As a specific example, we present results of modeling a TRTS experiment of photoexcitation of a dye in solution, namely, 2,11,20,29-tetra-tert-butyl-2,3-napthalocyanine, dissolved in toluene. We carry out a nonlinear least squares fit of a parameterized model to the measured data to show that the FDTD–TRTS method is able to accurately reproduce the features observed in the measured data set.