-The aluminum alloy process implemented in Sandia's Photovoltaic Device Fabrication Laboratory (PDFL) has major beneficial effects on the performance of commercial multicrystalline-silicon (mc-Si) substrates. Careful analysis of identically processed cells (except for the alloyed layer) in matched mc-Si substrates clearly indicates that the majority of the benefit arises from improved bulk minority carrier diffusion length. Based on spectral response measurements and PC-1D modeling we have observed improvements due to the alloy process of up to 400% in the "average" difision length in moderate-area cells and around 50% in large-area cells. The diasion length is dramatically improved in the interior of the silicon grains in alloyed substrates, resulting in the majority of the recombination occurring at the grain boundaries and localized areas with high defect densities.
.O The ChallengeThe challenge for all solar cell technologies is to lower the cost per watt. One approach is to improve the conversion efficiency without disproportionately raising the cost. High-efficiency enhancements are only useful if they can be implemented cost-effectively. A informative review of several efficiency enhancement techniques and their costs for mc-Si cell processing is found in [l]. One of the potentially viable techniques identified in that review is the aluminum-alloy back-surface-field process. Several groups have investigated the beneficial effects of this process and some of the best cell results using the process are reported in [2-51. The aluminum-alloy back-surface-field process can improve cell efficiency both through gettering of fast-diffusing impurities [6] and reflection of carriers from the back contact. The latter mechanism is important only if the minority carrier difision length (L) is comparable to, or larger than the cell thickness (W). For most mc-Si cells to date the benefit of the alloy process must come primarily from the gettering effects, since L/W is typically less than unity. This paper reports on our experience in Sandia's Photovoltaic Device Fabrication Laboratory (PDFL) using the aluminum-alloy process on mc-Si cells. In the past we have reported the results of a statistically designed gettering study on several commercial mc-Si materials which included both phosphorus difisions and aluminum-alloy treatments [7]. The present work looks more carefully at the effects of the aluminum alloy treatment alone on just one of those materials (cast mc-Si material manufactured by Solarex Corporation). In [7] we were unable to identi5 an optimal combination of times and temperatures for the phosphorus and aluminum treatments for this material. In addition, we observed no interactions between the two processes for Solarex material. However, the previous study did not include omitting the aluminum-alloy treatment completely (alloying temperatures from 700 to 900°C were used, all of which are above the AI-Si eutectic temperature). In this study we compare the This work was performed by Sandia National La...