ON THE DETERMINATION OF MINORITY-CARRIER DIFFUSION LENGTH IN THE BASE REGION OF N+-P-P+ SILICON SOLAR-CELLS USING PHOTORESPONSE METHODS

A comparative study of the measurements of the minority carrier diffusion length L in the base region of n+-p-p+ silicon solar cells by the LWSR, the MWSR, and the JSK methods, all of which exploit the photoresponse of cell in one way or the other, has been made. The LWSR and the MWSR methods utilize the long wavelength (e.g., 0.90 < lambda < 1.05 mum) spectral response and middle wavelength (e.g., 0.75 < lambda < 0.90 mum) spectral response, respectively, when the cell is illuminated from the n+-front side. However, the JSK method utilizes the rate of increase of short-circuit current density of the cell with the intensity of a single suitable middle or long wavelength radiation when the cell is illuminated from the p+-back side. Among these, the MWSR method and the JSK method each require the surface recombination velocity Sb of the minority carriers at the p+-back end of the p-base region to be known. We fabricated a few n+-p-p+ structure based bifacial silicon solar cells from 250 mum and 420 mum thick CZ silicon wafers and carried out measurements of L for them. The JSK method was applicable to both the thin and thick cells and gave accurate values of L. For thick cells the LWSR method was found to be more accurate than the MWSR method, the error in the value of L being, < 4% for the former and < 6% for the latter when compared with the JSK method. On the other hand, for thin cells, the LWSR method was found to be highly erroneous; the errors being 17-31% for the LWSR method in comparison with an error of 4% for the MWSR method. The results further showed that, generally, for thin front illuminated n+-p-p+ cells having a shallow n+-p front junction, L sufficiently large and d/L < 2.5, the MWSR method would give more accurate value of L than the LWSR method.