Development and analysis of silicon solar cells with laser-fired contacts and silicon nitride laser ablation

Abstract

The goal of this thesis was the development and analysis of crystalline silicon solar cells processed by laser radiation. Solar cells with n+pp+ structure on p-type, CZ-Si solar grade substrate were developed, analysed, and evaluated, based on two laser processing techniques: laser-fired rear contacts (LFC) and laser ablation of the front surface silicon nitride by means of laser chemical processing (LPC) or using a mirror galvanometer laser system (SCA). The LFC method was employed to form the rear contacts of crystalline silicon solar cells after the deposition of an aluminium layer. The LCP and SOA methods were used to develop a silicon nitride ablation process. The laser ablation process was employed to open regions of the devices antireflection coating, followed by selective chemical deposition of Ni/Ag to form the front metal grid. The best laser processing parameters found for LFC solar cells were: 33.0 A pumping lamp current, 20.0 kHz q-switch frequency, and 0.50 mm contact distance. LFC solar cells with screen printed front metallization and Si02 rear passivation layer achieved an average efficiency of 14.4 % and best value of 15.3 %, after an annealing step at 400 00 with a belt speed of 50 cm/min. lncreasing the rear aluminium layer thickness from 2 um to 4 um did not improve the performance of the devices significantly. The best laser processing parameters found for the silicon nitride laser ablation process based on the LCP technique were: 15.3 uJ laser pulse energy, 16.0 kHz q-switch frequency, and 100 mm/s processing speed. The best laser processing parameters found for the silicon nitride laser ablation process based on the SOA technique were: 5.0 uJ laser pulse energy, 130.0 kHz q-switch frequency, and 813 mm/s processing speed. Solar cells with silicon nitride laser ablation, front side metallization by Ni/Ag selective electrochemical deposition, and screen-printed rear side metallization achieved an average efficiency of 16.1 % and best value of 16.8 % for the LCP technique and an average efficiency of 16.3 % and best value of 16.6% for the SOA technique.