Automated parameter optimization in modeling absorption spectra and resonance Raman excitation profiles

An automated method is described for optimizing the molecular parameters in simultaneous modeling of optical absorption spectra and resonance Raman excitation profiles. The method utilizes a previously developed Fortran routine that calculates absorption spectra and Raman excitation profiles for polyatomic molecules in solution from a model for the potential energy surfaces and spectral broadening mechanisms. It is combined here with an optimization routine from the commercial MATLAB package that iteratively adjusts the parameters of the molecular model to minimize the least-squared error between calculated and experimental spectra. Optimizations that typically require days to weeks of human time when performed interactively can be accomplished automatically in less than an hour of computer time. The method can handle large molecules (we show results for as many as 23 Raman-active modes) and mixtures of spectral broadening mechanisms (lifetime, Brownian oscillator, and inhomogeneous), and is robust toward noise or missing data points.