A statistics modeling approach for the optimization of thin film photovoltaic devices

The growing interest in exploring thin film technologies to produce low cost devices such as n-i-p silicon solar cells, with outstanding performances and capability to address the highly relevant energy market, turns the optimization of their fabrication process a key area of development. The usual one-dimensional analysis of the involved parameters makes it difficult and time consuming to find the optimal set of conditions. To overcome these difficulties, the combination of experimental design and statistical analysis provides the tools to explore in a multidimensional fashion the interactions between fabrication parameters and expected experimental outputs. Design of Experiment and Multivariate Analysis are demonstrated here for the optimization of: (1) the low temperature deposition (150 degrees C) of high quality intrinsic amorphous silicon (i-a-Si:H); and (2) the matching of the n-, i-, and p-silicon layers thickness to maximize the efficiency of thin film solar cells. The multiple regression method applied, validated through analysis of variance and evaluated against exact numerical simulations, is shown to predict the overall intrinsic layer properties and the devices performance. The results confirm that experimental design and statistical data analysis are effective approaches to improve, within a minimum time frame and high certainty, the properties of silicon thin films, and subsequently the layer structure of solar cells.(C) 2017 Published by Elsevier Ltd.