In this article we present results of calculations and experiments concerning the characterization of pn junctions with depths x(J)<100 nm. Simulations of differential Hall effect profiling with model dopant profiles as input demonstrate that this method call be used to analyze dopant profiles in pn structures, if it is known how the p and n regions of the pn junction are contacted. Results of differential Hall effect simulations obtained with the same model dopant profile as input strongly depend on, the kind of the contact to the p and a regions of the junction. The results reveal that differential Hall effect analysis can be accurate with respect to the dopant profile and to the junction depth if Ohmic contacts are formed to both regions of the junction. The shift in the position of the apparent pn junction relative to the metallurgical pn junction is estimated for the case that contact is made to one of the regions. Spreading resistance, differential Hall effect measurements, and chemical staining of beveled samples are used to characterize pn junctions experimentally. Determining pn junction depths we find an error of about 5 nm using spreading resistance measurements and differences of more than 10 nm using conventional chemical staining.