In specimens beveled for spreading resistance measurements or copper staining, the ’’on‐bevel’’ carrier distributions may be distorted from the correspondng dopant atomic distributions. The problem becomes very serious for very shallow structures in the 0.1–0.2 &mgr;m range. We have studied this problem by solving the two‐dimensional Poisson equation for some 130 cases comprising several parameters. It is found that insofar as the deviation of the on‐bevel electrical junction from the metallurgical junction is concerned, dopant profiles generally fall into two groups. Group I profiles exhibit significantly shallower on‐bevel electrical junctions than corresponding metallurgical junctions, and the situation gets precipitously worse when junctions get shallower than 0.2 &mgr;m. Profiles in this group possess gentle tails, and include gaussian, complementary error function, and exponential profiles. Group II profiles exhibit a very small ’’positive deviation’’ of the on‐bevel electrical junction from the metallurgical junction, and this deviation remains essentially unchanged as the metallurgical junction gets shallower. Profiles in this group have nearly linear or superlinear gradients in the vicinity of the metallurgical junction; a high concentration diffused arsenic emitter is an example. The ’’negative deviation’’ of the on‐bevel electrical junction from the metallurgical junction is much more serious in the case of epi with buried layers, unless the epi‐doping level is ≳1×1017atoms/cm3. Various cases of surface charge effect have also been investigated. It is found that surface charge arising from deep‐level surface states exhibits a self‐moderating effect, in contrast to a fixed amount of surface charge from shallow states.