Evaluation of Infrared Intensities Using Diffusion Monte Carlo.

Abstract

Approaches for evaluating excited state wave functions and energies using diffusion Monte Carlo (DMC) with guiding functions (guided DMC) are discussed. For this work, the guiding functions are functions of a subset of the 3N - 6 coordinates that are needed to describe the structure of the molecule of interest. The DMC wave functions are used to evaluate intensities using two approaches. In the trial wave function approach, the product of the molecular wave function for one of the states involved in the transition and the guiding function for the second state is used to evaluate the matrix elements of the dipole moment. In the descendant weighting approach, descendant weights are used to evaluate the value of the wave function for one of the states involved in the transition at the geometries sampled by the DMC wave function for the second state. The descendant weighting approximation is shown to be more accurate as well as computationally more expensive compared to approximations that are based on various forms of the trial wave function approach. Strategies are explored, which combine results of different forms of the trial wave function approximation to minimize the errors in this approach. The trial wave function and descendant weighting approaches are applied to a study of a harmonic oscillator, where the sensitivity of the calculated energies and intensities to the quality of the trial wave function is explored. The two approaches are also applied to calculations of frequencies and intensities of transitions in water, H₃O₂^-, a four-dimensional (4D) model based on H₃O₂^- and H₅O₂⁺. We also show how comparisons of the results obtained using several forms of the trial wave function approach allow us to explore how couplings among vibrational motions are reflected in the intensities.