Although the capability of scanning capacitance microscopy (SCM) for pn junction imaging has been qualitatively demonstrated, quantification of dopant profiles in two‐dimensions for pn junctions has proven to be a challenging problem. One reason is that the SCM result is seldom reproducible and this is generally believed to be due to sample preparation technique. In the first part of this work, we made a detailed study of sample preparation methods for SCM measurements. The purpose of the study was to establish an optimum sample preparation technique. Experiments were performed on two known dopant profiles: n type and p type staircase structures with concentration ranging from 1014to 1019cm−3. These two samples were cross‐sectioned and prepared using different oxidation techniques, which caused variations of interface states and oxide quality. We studied the effect of wet oxidation, thermal oxidation, as well as combined wet and thermal oxidation. Experimental parameters : temperature, baking duration and oxidation methods were studied in detail. To evaluate the oxide quality, we measured the flat band voltage change &Dgr;VFBby sweeping dC/dV vs. voltage in forward and reverse direction (from accumulation to depletion & vice versa). &Dgr;VFBrepresents the amount of the oxide trap charge. Since the staircase structure sample has a large doping range, the contrast reversal effect was obvious. We could minimize this effect by using the optimum experimental condition. In the second part, we investigated samples prepared under this optimum experimental condition. © 2003 American Institute of Physics