Nuclear-spin-dependent corrections to the transition polarizability in cesium

The Stark-interference technique is commonly used to amplify the feeble parity-violating signal in atomic experiments. As a result, interpretation of these experiments in terms of electroweak observables requires knowledge of the Stark-induced E 1 transition amplitudes or, equivalently, transition polarizabilities. While the literature assumes that these transition polarizabilities do not depend on the nuclear spin, here we prove the contrary. The nuclear-spin dependence arises due to the hyperfine mixing of atomic states and requires a third-order perturbation theory (one hyperfine interaction and two electric-dipole interactions) treatment. We demonstrate that the so far neglected tensor contribution appears in the transition polarizability and present numerical results for the nuclear-spin-dependent corrections to the 6S(1/2)-> 7S(1/2) transition polarizability in Cs-133. We investigate the effect of these corrections to transition polarizabilities on the extraction of the Cs-133 anapole moment from the Boulder experiment [Science 275, 1759 (1997)]. We also consider their effect on the extraction of the ratio between the scalar and vector transition polarizabilities from the measurements [Phys. Rev. A 55, 1007 (1997)]. While the corrections are minor at the current level of experimental accuracy, our analysis provides a framework for future experiments.