Effects of Sample Inhomogeneity and Geometry on Photoconductivity Decay

Excess minority carrier decay is simulated by a finite element method in cylindrical, rectangular (including epitaxial structures), and wafer samples with consideration of limited absorption of excitation light and spatial variation of lifetime in multicrystalline materials. It is found that, as long as a lifetime is derived from a later part of a photoconductivity decay where asingle-exponential signal is obtained, (1) the ASTM correction factors are valid for large bulk samples without the need to roughen the surfaces; (2) the 1-D asymptotic solution commonly used for lifetime measurements on wafers does not agree with the simulation for intermediate surface recombination velocities and thin wafers; (3) for multicrystalline samples, the final decay represents overallquality of the material; and (4) when grain sizes are much less than the effective carrier diffusion length, carrier injection is uniform, and no surface nor grain boundary recombination is present, the effective lifetime of a multicrystalline material is the inverse of the volume-weighted inverse sum of local lifetimes.