We made an error in the designation of the voltage offset used for the calculation of the shift of the surface potential. The method that we used with the Vref having a minimum in δIds/δVref calculates the threshold voltage and not flat band potential. As both offsets are equivalent to follow the changes in ψ0, so there is no significant change in the interpretation of the data. The labels and legend for Figure 2 in the original article should be corrected as shown here. (Figure presented).