Cu electroplating is of a great deal of interest for the metallization of silicon solar cells due to its comparable resistance with the current silver metallizaion, and its lower cost for both the material and the operation.[1,2] However, the notorious Cu diffusion into Si is of great concern to the devices performance degradation during the life cycle of the cells. Ni silicide and Ni metal layers have been recently reported as part of the integration scheme to enable the Cu metallizaion and greatly improved the solar cell reliability.[3,4] The formation of this Ni silicide and Ni stack structure is to provide a low-resistance contact between the silicon and metal grid, a good adhesion of metal grid, and a diffusion barrier for the Cu. Among them, the Ni silicide formation needs to be well controlled so that it is not too thick to shunt the solar cell.

This paper will present the development of light assisted plating processes for the formation of nickel silicide and copper grids on silicon solar cells, including beaker scale, bench top tool, and pilot tool processing[5,6,7]. The solar cell substrates for plating comprise ~ 60 ohm/sq emitter formed by POCl diffusion, PECVD SiN passivation layer, and evaporated Al back surface field. Electrodeposition of Ni and copper was studied on 4 cm x 4 cm solar cells in small scale studies and 6 in x 6 in cells on tools.

Figure 1 shows the absorption spectra of different electrolytes for deposition and the relation between the applied potential and the light intensity at different constant current densities. Nickel solution has a strong absorption at wavelength between 350 to 450 nm and between 600 to 800 nm. Two regions of operation were clearly observed with a threshold light intensity when a high intensity wide spectrum halogen light source is used for illumination during electroplating. At high illumination, when enough carriers were generated by the light, the plating occurs at the forward bias mode of the diode. As the light intensity decreases below a threshold value, the potential increases significantly and stays at a high value indicating the diode operates at reverse bias mode. The threshold light intensity is dependent on the illumination energy spectrum, solution absorbance, anode shadowing and applied current density. The learnings from beaker scale experiments were directly used to define the tool requirements and process specifications.

REFERENCES

1. D. A. Yates, US Patent 4,609,565 (1986).
2. K. Holdermann, US patent 5,543,333 (1996).
3. J Bartsch, A Mondon, K Bayer, C Schetter, M Horteis, and S W Glunz, J. Electrochem. Soc., 157, H942 (2010).
4. J Bartsch, A Mondon, C Schetter, M Horteis, S W Glunz, IEEE 35^th PVSC, Honolulu HI (2010).
5. Q Huang, S S Papa Rao, K. Fisher, ECS J. Solid State Sci. Technol., 4, Q75 (2015).
6. Q Huang, ECS J. Solid State Sci. Technol., 5, P51 (2016).
7. Q Huang, K B Reuter, Y Zhu, V R Deline, ECS J. Solid State Sci. Technol., 5, Q24 (2016).