We found out that the Hall resistivity data (ρ H all ) of our original paper should be corrected due to an error in the wiring in the experimental set-up. The correction implies a change in the sign of the Hall resistivity data (−ρ H all ) presented in the original publication. This correction does not affect the interpretation of the experiments and our quantitative results remain valid, except that the third carrier band, required to account for the magnetic field evolution of the magnetoresistance and Hall resistivity, is a surface hole band (and not a surface electron band as it was reported in our original paper). This modification in the character of the surface carriers does not influence the conclusions made in the paper.In order to illustrate these points we show revised versions of Figs. 5,6,and 8, respectively. Figures 5(a) and 5(b) shows the corrected ρ H all isotherms from room temperature down to 2 K for the Bi films with thicknesses t = 100 nm and 10 nm, respectively. The remarkable change both in the magnitude and the field dependency of ρ H all with film 199904(E) (2013) thickness is shown in Fig. 6, in which the corrected ρ H all as a function of magnetic field is compared for different film thickness down to 10 nm at 2 K. Figure 8 displays the magnetic field dependence of the resistance and the corrected ρ H all at 50 K for 100-nm-thick film. Black solid line is calculated from the two band model (electron bulk and hole bulk band). Red solid line includes the third carrier band which has been interpreted as an effective band which contains surface carriers characterized by an effective density and an effective mobility. The values of the mobility μ i and the carrier density n i obtained for each band are not affected by the correction of the sign of the ρ H all and the results listed in the original Table I remain valid. However, the correction in the sign of ρ H all implies a change in the sign of the Hall coefficient R s of the surface band, which becomes positive (instead of negative, as was reported in our original paper). This implies that the surface carriers are holes with p s and μ p s (instead of electrons with n s and μ n s ). This minor change does not affect our conclusions.We would like to thank M. C. Martínez Velarte for drawing our attention to the wiring problem.
199904-2Role of the surface states in the magnetotransport properties of ultrathin bismuth films We have investigated the magnetotransport properties of ultrathin films of Bi grown on thermally oxidized Si͑001͒ substrates with thickness ranging from 10 to 100 nm at temperatures down to 2 K and in magnetic fields up to 90 kOe. Remarkable differences both in temperature and field dependence of the Hall resistivity are found for the films with thickness above and below 20 nm. These observations can be explained due to the presence of surface states, which play an important role in determining the electronic transport properties of the thinnest films. The estimated surface carrier density 4 ϫ 10 13 cm −2 at room temp...