Carrier-diffusion corrected J0-analysis of charge carrier lifetime measurements for increased consistency

• We show that the apparent reduction of the recombination parameter J0 obtained by carrier lifetime measurements at high injection densities origins in the finite carrier diffusion towards the surfaces.
• We present an analysis taking into account the finite carrier diffusion and band-gap narrowing due to injection in the substrate.
• The presented analysis substantially improves the prediction of J0 independently from the applied injection density and the base substrate doping.
• Numerical simulations taking into account band-gap narrowing in the substrate are in excellent agreement with the measurement.

We investigate the origin of the apparently reduced recombination parameter J0 of highly doped regions obtained from e.g. QSSPC lifetime measurements at high injection densities as well as the influence of recombination at bulk defects on the recently updated J0-analysis of QSSPC measurements in the Sinton lifetime tester software. Using the example of crystalline silicon, we show that the frequently observed reduction and underestimation of J0, which is strongly pronounced at high injection levels, originate from neglecting the finite carrier diffusion coefficient. This work presents an analysis taking into account the finite carrier diffusion coefficient and thus enabling to evaluate charge carrier lifetime data at high injection levels leading to strongly increased consistency of the determined J0. Furthermore, we give a simple correction term for the recently published analysis leading to a total independence of a reasonably injection-independent Shockley–Read–Hall (SRH)-lifetime in the substrate. Combining the proposed methods, the determination of J0 from numerically simulated lifetime-data is accurate within 3% of the input J0=100 fA/cm2 over an injection range of 1014–3·1016 cm−3. The application of the proposed method to measured data of crystalline silicon devices is in excellent agreement with the simulations and shows a strongly improved linearity and thus constant extracted J0 compared to the former and recently published methods, which underestimate J0 at high injection levels, nearly independently from doping type and -level of the substrate. The new and simple method is implemented in a spreadsheet calculator enabling fast and robust data evaluation similar to the Sinton lifetime tester software.