Ab initio determination of the nuclear relaxation contribution to the second hyperpolarizability of carbon disulfide

Although basis set saturation, electron correlation and frequency dispersion have been addressed thoroughly, the electronic second hyperpolarizability of carbon disulfide computed by K. Ohta, T. Sakaguchi, K. Kamada and T. Fukumi (Chem. Phys. Lett. 274 (1997) 306) is not in agreement with experiment. In this Letter the potentially substantial nuclear relaxation contribution is evaluated within the Møller-Plesset scheme limited to second order by using the 6-31G^* basis set augmented by three diffuse functions (1p and 2d). Within the enhanced approximation, the nuclear relaxation contribution to the static, dc-Kerr and ESHG second hyperpolarizability turns out to amount to 26.5%, 6.8% and -0.8% of the pure static electronic counterpart, respectively. The remaining gap between theory and experiment suggests new experiments should be carried out.