Very shallow p+/n junctions (less-than-or-equal-to 0.1 mum) are required for 0.25 mum complementary metal oxide silicon technologies. Spreading resistance probe (SRP) analysis can quickly give an electrical resistance profile, but the real problem is the conversion of the resistance profile into a concentration profile. On the one hand, a bevel effect has been seen. It modifies the resistance profile shape near the surface region and disappears far away from the bevel edge. In our work, we try to correct for this effect making the assumption that this region gives rise to a parasitic resistance that we can subtract from the spreading resistance. On the other hand, it is necessary to make some assumptions about the variation of the electrical contact radius versus the dopant concentration, especially in the presence of high gradient profiles. We propose a semi-empirical method which allows for the correction of the concentration profiles by taking into account this radius variation without any calculation instabilities. Thus, taking into account these effects and the corresponding corrections made, we can obtain SRP profiles comparable to the secondary ion mass spectroscopy (SIMS) profiles.
