In our paper, the calculation of the dressed ground state surface for the case of a strong gradient field B=1 G and G =5 Tm−1; results shown in Fig. 2 b and d of the original publication is incomplete. In detail, emergent two-photon couplings between the 5S1/2 ,F=mF=2 ground state and 5S1/2 ,F=2,mF 2 levels via the hyperfine levels of the 5P3/2 excited state due to the first excitation laser have been omitted in the original publication. Including these states in our calculations, it turns out that the spatial dependence of the light shift is not strong enough to entail the ring-shaped trapping potential as originally presented in Figs. 2 b and d . The results presented in the regime of strong Ioffe fields B=10 G and G=5 Tm−1; results shown in Fig. 2 a and c of the original publication do not change: In this case, the couplings with the 5S1/2 ,F=2,mF 2 levels are negligible and the simplified 3-level scheme as described in the original publication holds. Consequently, the conclusion that the specific signatures of the Rydberg trapping potential can be probed by means of ground state atoms that are off-resonantly coupled to the Rydberg state via a two photon laser transition does not change. The conclusion of our original publication that the composite nature of Rydberg atoms significantly alters their trapping properties opposed to pointlike particles with the same magnetic moment is not affected by the changes described in this erratum. Moreover, we would like to correct two minor typos of our original publication. Firstly, the off-diagonal matrix elements of Eq. 7 should read /2 instead of − /2. The eigenvalues of the matrix and hence the dressed surfaces are not affected by this change of sign. Secondly, Ẽp= B R instead of Ẽp=2 B R ; in our calculations Ẽp was implemented correctly. PHYSICAL REVIEW A 80, 029903 E 2009
Read full abstract