Precise parameterization of the recombination velocity at passivated phosphorus doped surfaces

Abstract

We investigate the surface recombination velocity Sp at the silicon-dielectric interface of phosphorus-doped surfaces for two industrially relevant passivation schemes for crystalline silicon solar cells. A broad range of surface dopant concentrations together with a high accuracy of evaluating the latter is achieved by incremental back-etching of the surface. The analysis of lifetime measurements and the simulation of the surface recombination consistently apply a set of well accepted models, namely, the Auger recombination by Richter et al. [Phys. Rev. B 86, 1-14 (2012)], the carrier mobility by Klaassen [Solid-State Electron. 35, 953-959 (1992); 35, 961-967 (1992)], the intrinsic carrier concentration for undoped silicon by Altermatt et al. [J. Appl. Phys. 93, 1598-1604 (2003)], and the band-gap narrowing by Schenk [J. Appl. Phys. 84, 3684-3695 (1998)]. The results show an increased Sp at textured in respect to planar surfaces. The obtained parameterizations are applicable in modern simulation tools such as EDNA [K. R. McIntosh and P. P. Altermatt, in Proceedings of the 35th IEEE Photovoltaic Specialists Conference, Honolulu, Hawaii, USA (2010), pp. 1-6], PC1Dmod [Haug et al., Sol. Energy Mater. Sol. Cells 131, 30-36 (2014)], and Sentaurus Device [Synopsys, Sentaurus TCAD, Zurich, Switzerland] as well as in the analytical solution under the assumption of local charge neutrality by Cuevas et al. [IEEE Trans. Electron Devices 40, 1181-1183 (1993)]

Recombination processes in passivated boron-implanted black silicon emitters

“…2. The surface doping, which can significantly affect boron and phosphorus emitter surface passivation, 23,41 decreases in implanter emitters from (4.5 6 0.1) Â 10 19 cm À3 to (1.90 6 0.07) Â 10 19 cm À3 under the effect of increasing anneal temperature. In the same time, the emitter depth increases from 0.7 lm to about 1.5 lm [ Fig.…”

“…finite diffusion coefficients may lead to J 0e underestimation in the poor quality emitters, e.g., those annealed at 850 C. Apart from this error source, a large uncertainty in the extracted J 0e can arise from its dependence on the minority carrier excess density. 23 The error bars in the graphs thus indicate the J 0e variation in the range Dn ¼ SMCDð1630%Þ.…”

“…35 Note that J 0e tends to be underestimated with the models previously implemented. 35 Contrarily to other studies, 23,36 no J 0e rescaling based on the surface area was performed in the bSi results, because the effective value is more relevant from the solar cell point of view. The J 0e was averaged over the values extracted at 630% of the specified minority carrier excess density (SMCD) taken as Dn % 10N D with N D the bulk doping.…”

“…33 Despite the lack of knowledge on the current distribution inside bSi needles when it is injected through metal probes, the four-point probe results are usually consistent with the values measured from the dark conductance. 10,22,23 The Sinton tool was also used for minority carrier lifetime and J 0e measurements. The J 0e was extracted at 298 K from the injection level-dependent carrier lifetime curve, based on the method originally developed by Kane and Swanson, 34 which was later improved by Blum et al and is now used in WCT-120.…”

“…20,21 We discuss the possibility to solve the common issues encountered with emitter diffusion and analyze the recombination process in implanted emitters. Although few studies exist on ion-implanted textured silicon 22,23 and to our knowledge none on bSi, ion implantation can provide a number of advantages in comparison to diffusion. It may allow controlling recombination in bSi without tradeoff in dimensions, since the total impurity dose injected in the nanostructures is fixed regardless of surface area.…”