Engineering aluminum oxide/polysilicon hole selective passivated contacts for high efficiency solar cells

Tunnel layer passivated contact technology is already highly efficient in case of selective electron extraction but not as efficient in case of selective hole extraction. Thus far, SiOx/p(+)-poly-Si contacts have resulted only in efficiencies above similar to 20.1% for rear-side deployed hole selective contacts. We investigate if hole extraction selectivity can be further improved by substituting the 'conventionally' used SiOx tunnel layer exhibiting moderate or even high positive fixed charge density by AlOx tunnel layers, exhibiting high negative fixed charge density. The merits of using atomic layer deposited ultrathin AlOx tunnel layers are investigated and compared with wet chemically formed SiOx tunnel layers to form AlOx/p(+)-poly-Si and SiOx/p(+)-poly-Si hole selective passivated contacts respectively. The AlOx thickness (0.13-2 nm) and its thermal budget including annealing time, temperature and ambient were varied. The quality of the resulting AlOx/p(+)-poly-Si passivated contacts was determined by measuring the recombination current density (J(c)) and the effective contact resistivity (rho(c)). Finally, using the measured values of J(c) and rho(c), we predict the efficiency potential and selectivity of the passivated contact using Brendel's model. We show that for 425 degrees C annealed AlOx samples prior to poly-Si capping, there is an improvement in passivation quality due to the high negative AlOx interface charge, which forms only for "thick" tunnel layers (>= 1.5 nm). However, after high-temperature poly-Si capping, enhanced boron in-diffusion and charge compensation are degrading the overall passivation quality of "thick" AlOx/p(+)-poly-Si passivated contacts. The best AlOx/p(+)-poly-Si passivated contacts use ultra-thin AlOx tunnel layers (efficiency potential of 26.9%), which is only marginally better than the SiOx reference samples, but still improves hole selectivity.