The apparent doping profile obtained by conventional C-V data analysis is limited by the Debye length and is not identical to the majority-carrier distribution.
Answers are grounded in this paper and cite their sources.
C-V doping profiles require Debye-aware corrections; leaves open reliable sub-Debye feature resolution in device characterization.
The doping profile of a semiconductor is given only approximately by the conventional analysis ofC-Vmeasurements. The present study employs computer simulation of semiconductors with one-sided doping profiles that consist of high and low doped sections joined by steps and linear ramps. The computation yields the apparent doping profile that would be obtained by the conventional use ofC-Vdata, and this result is compared with the actual profile, with the majority-carrier distribution, and with the outcome of a correction previously proposed in the literature. The results show that a step in the profile cannot be resolved satisfactorily to less than several Debye lengths corresponding to the doping on the high side of the profile. A ramp cannot be distinguished accurately from a step unless its width is appreciably greater than a Debye length. Furthermore, the apparent doping profile is not identical with the majority-carrier distribution with contacts far away, as has been suggested, and the discrepancy is shown to depend on the side from which depletion is done.
No takes yet. Share an insight, caveat, or question.
Johnson et al. (1971) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: