Analysis and simulation of manufactured screen-printed solar cells

Abstract

A detailed study has been conducted of the performance of millions of screen-printed solar cells to analyse the various processing parameters and their interdependencies. Algorithms have been developed to quite accurately predict device structure and hence performance as a function of the many manufacturing processing parameters. The processes characterized include the saw-damage removal etch, anisotropic texturing, emitter diffusion, edge junction isolation, AR coating deposition and the screen-printing and firing of the front and rear metal contacts. From the developed algorithms, a simulator has been developed which includes the natural variations in device parameters that occur in a real production environment due to relatively random effects such as thermal gradients within furnaces, slight fluctuations in belt speeds, gas flows, plasma etching power, chemical concentrations, furnace temperature controllers, wafer quality, etc. These can have a significant effect on electrical yields. Even common factors affecting the mechanical yields in production have been programmed into the simulator. Since the virtual production line was primarily developed for the purposes of educating and training engineers, extensive in-line quality control testing has also been incorporated, such as the measurement of: minority carrier lifetimes; substrate resistivity; emitter sheet resistivity; wafer thickness; screen-printing paste thickness and width; AR coating thickness and refractive index; surface reflection; spectral response; I-V characteristics; series resistance; shunt resistance; ideality factor; doping profile; etc.