Hyperpolarizabilities of the hydrogen fluoride molecule: A discrepancy between theory and experiment?

Abstract
Due to the recent availability of an experimental value for the second- and third-order electric polarizabilities of the HF molecule, possible errors in the previous prediction of these quantities have been considered. These include basis sets, vibrational corrections, frequency dependence, infinite-order correlation corrections, and the effect of triple excitations. Despite the inclusion of all of these effects, the discrepancy between experiment and theory remains. Our best results predict χ(2)∥ to be −3.3 to −3.8×10−32 and χ(3)∥ to be 45 to 48×10−39 esu which are at best 81% and 69% of the experimental values, respectively. Possible additional corrections are considered with emphasis on basis set completeness. Considering the difficulty in predicting such sensitive quantities, this is rather good agreement. The paper also addresses questions of reliability of ab initio calculations for such high-order properties as hyperpolarizabilities, while identifying some places where the experimental results are susceptible to error.