Novel sulfur surface passivation for n- and p-type silicon
Date
2020
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
The reduction of recombination at the surface of a silicon wafer is an important factor in improving the efficiency of solar cells. Thermal oxidation is the most common method for Si surface passivation and has excellent performance, but the high temperature process can degrade the bulk quality of silicon, which adversely affects solar cell performance. Deposition of a thin aluminum oxide layer is another common passivation method that works well on p-type silicon thanks to its high negative fixed charge that repels electron minority carriers away from the surface and it is very stable under light, but this same aspect makes it worse on n-type silicon for hole minority carriers. The objective of this research is this to create a lower temperature passivation layer than the thermal SiO2 that can withstand an industry-standard screen-printed metallization process, while being applicable to both types of surfaces unlike Al2O3. Sulfur and selenium are proposed theoretically to be good candidates for this purpose because of their ability to cap multiple surface bonds at once, as well as their similar bond angles and lengths to Si. [1] Experimental work done by S. Liu et al. [2] at IEC has demonstrated a lifetime >2000 μs and an Surface Recombination Velocity (SRV)< 3 μs by H2S gas phase reaction, which can match thermal SiO2. This thesis will expand upon that work, offer some explanations of passivation mechanism and apply the process to passivate dopant-diffused Si surface. ☐ A custom-built sulfurization/selenization reactor is used for the passivation of HF-cleaned n- and p-type silicon wafers with and without a diffused n+ or p+ emitter. The samples were reacted in H2S and H2Se for S and Se passivation, respectively, with a background Ar gas flow for dilution. Reactions were performed at a range of temperatures between 400-650°C, and with reaction times between 15 to 210 minutes. The best lifetimes and SRVs are achieved at 550°C (>2000 μs, < 3 cm/s, similar to previous work , ([2]) with reaction duration not having a significant effect on as-passivated lifetime. S passivation decays rapidly in air, decaying by an order of magnitude within half an hour. A longer reaction time (≥ 60 min) can reduce the rate at which the S passivation decays in air, though further processing is needed to fully stabilize the passivation. A deposited a-SiNx layer on top of the S layer prevents degradation for some months. Some attempts have been made at Se and S/Se mixture passivation, but current results suggest that Se passivation is inferior compared to S passivation. In terms of how the passivating gas interacts with the Si surface, a higher net Ar flow clearly improves the performance of the S passivation, while there is a certain minimum concentration of H2S for significant passivation to occur. Dopant diffused n-n+ and n-p+ wafers follow a similar passivation model to the undiffused n and p-type Si wafers. ☐ Modelling of the lifetime data to estimate surface state density (Dit) and fixed charge density (Qf) suggests that higher lifetime is related to a lower Dit, but a linear relationship between Dit and Qf is present and unexplained. In order to fully understand the S passivation mechanism, further studies, including capacitance-voltage (CV) testing, are necessary to precisely determine Dit and Qf and confirm modelling trends.
Description
Keywords
Electronic devices, Photovoltaics, Silicon, Sulfur, Surface passivation