Borosiloxane boron diffusion for p-emitter formation on n-type silicon wafers

Thermal boron diffusion, which forms highly doped and shallow p-emitters on phosphor-doped silicon wafers, is one of the primary processes in commercial-scale production of n-type cells. Here, we report on the use of nontoxic and nonvolatile borosiloxane sols as the spin-on boron source. In comparison to the tribromide (BBr₃) boron diffusion that is applied in the production of most commercial n-type cells, borosiloxane boron diffusion may have potential advantages in terms of cost, safety, operability, etc. The borosiloxane sols studied here were formed from a cross-linking reaction between boron acid and a mixture of two types of silicon alkoxides, methyltriethoxysilane (MTEOS) and dimethyldiethoxysilane (DMDEOS). The MTEOS/DMDEOS ratio was found to be the key factor determining the synthetic period, net structure, thermal transformation property, and coating behavior of the borosiloxane sols. In combination with the spin-coating method, the performance of borosiloxane boron diffusion into phosphor-doped silicon wafers (1–3 Ω·cm) was systematically investigated by varying the borosiloxane composition (boron/silicon or MTEOS/DMDEOS ratio) and the diffusion process parameters (temperature, ambience, and period of individual sub-steps), and by relating them to the resulting sheet resistance and boron doping profile. The results show that under the diffusion conditions in conventional cell production lines, the formed p-emitters can have variable sheet resistance in the range of 35–582 Ω/□ and boron doping depth of 100–300 nm.