Empirical optimization and implementation of boron emitter on n-type silicon solar cells using BBR3 liquid source

R. Silva
Abstract:The present work focus on the study and optimization of a BBr3 diffusion process for a p layer formation on n-type Si wafers, as it is currently one of the main limitations for the implementation of cheap high-efficiency solar cells based on this material with a boron emitter. A state of the art review is presented, covering the following topics: n-type silicon properties and its historical context; p layer metallization and passivation; boron emitter properties and boron diffusion based on a BBr3 liquid source. Next, an experimental study of several diffusion process parameters and their impact on the resulting emitter is presented, being the parameters: BBr3 bubbler temperature (and desirable stability); N2 flow; N2-BBr3 flow; wafer spacing in the boat structure and predeposition step duration. For each experiment it was verified the presence of a BRL while evaluating the diffusion homogeneity through a visual evaluation of the formed oxide and sheet resistance measurements. Afterwards, a selected single-wafer process was tested for batch process conversion, which then required a parameters optimization. The removal of the BRL through both chemical and thermal oxidation was also tested. In conclusion, it was verified that a good control of the bubbler temperature is required and that it should be at least at 20 oC in order to obtain an effective doping. All other parameters showed to have an impact in the formed emitter and resulting oxide, although no homogeneous doping nor oxide thickness were achieved. The formation of a homogeneous oxide is a crucial requirement as it can indicate a good homogeneity in both the doping and BRL formation, impeding an effective removal of this recombination-inducing layer. This was verified as both oxidation techniques tested resulted in only a partial removal of the BRL. It was also verified that a single-wafer process is not immediately convertible into a batch process as an ineffective doping was obtained, therefore requiring further parameter optimization. Finally, future work lines are suggested in order to go on with the present work.
Materials Science,Engineering,Physics
What problem does this paper attempt to address?