Using Doping Variation to Determine J(0s) I-V Test

The most common measurement method for the separation of lifetime data into bulk and surface recombination is the Kane and Swanson method, originally developed for undoped n-type silicon. While it has been generalized for use in doped material, it depends on achieving high excess carrier densities relative to the doping level to yield unambiguous separation of bulk lifetime and surface recombination. This paper demonstrates an alternative technique to achieve this separation for heavily doped samples where reaching a very high injection level is not feasible. Rather than using the lifetime dependence on excess carrier densities to find the J(0s) term, for large datasets the natural variation in substrate doping offers an alternative fitting scheme that uses the same formula to determine J(0s). A dataset of monocrystalline PERC cells is used as a case study.