On the quantification of Auger recombination in crystalline silicon

Abstract

Quantification of Auger recombination in crystalline silicon is usually challenging because it requires distinguishing Auger recombination from extrinsic recombination processes, especially at the surface. In this work we exploit the known injection dependence of surface recombination at surfaces passivated by highly charged dielectric films to assess the value of the ambipolar Auger coefficient Camb in crystalline silicon from carefully calibrated photoconductance decay measurements of high-lifetime Si wafers. A value of Camb = (2.11±0.02) × 10−30 cm6 s−1 is determined for both n- and p-type Si, independent of the dopant concentration for moderately doped samples near 300 K. Furthermore, by exploiting the expected equality of the surface saturation current density J0s for identically processed n- and p-type substrates, we are able to estimate the relative contributions of ehh and eeh Auger processes to Camb under the same doping and injection regime. The resulting ratio Cn/Cp=2.81 is in close agreement with the value of 2.91 derived from the data of Dziewior and Schmid at high dopant concentrations, which we reassess in light of modern mobility models. Based on our own data and literature data at higher concentrations, we offer a revised parameterisation of the Auger recombination rate in c-Si as a function of dopant concentration and injection level which is consistent with these findings. This revised parameterisation is also shown to be consistent with the most recent record lifetimes reported by other authors. Finally, we suggest experimental considerations relevant for future efforts to further refine the Auger parameterisation.