Assessing limitations to the two-level approximation in nonlinear optics for organic chromophores by ab initio methods

Abstract

The use of a two-level approximation to simply characterize the nonlinear optical properties of organic materials is well known. Usually only electronic ground states are significantly populated; higher levels are engaged only in the capacity of virtual states, and it is frequently assumed that just one such state dominates in determining the response. Calculating nonlinear optical susceptibilities on this basis, excluding all but the ground and one excited state in a sum-over-states formulation, is a technique widely deployed in the calculation and analysis of nonlinear optical properties. However, the necessity for such an approach is diminishing as, particularly within the last decade, the accuracy of ab initio calculations has reached unprecedented levels. This offers new opportunities for a vigorous test of existing models using real molecular structures. Here we report the results of our recent work on testing the general validity of two-level calculations in nonlinear optics. Firstly, through the extension of approximation to a three-level model we demonstrate that the neglect of additional excited states can lead to substantially erroneous results for the hyperpolarizability elements. Secondly, using high levels of theory and basis set we report the results of ab initio calculations for both ground and electronically excited states of the optimised structures, for selected merocyanine dyes. The results are used for the calculation of hyperpolarizabilities by a rigorous sum-over-states formulation. A systematic comparison with the two-level approach provides a means for identifying the limits of the model and the criteria for its validity.