Characterization of thin-film silicon materials and solar cells through numerical modeling

Abstract

At present most commercially available solar cells are made of crystalline silicon (c-Si). The disadvantages of crystalline silicon solar cells are the high material cost and energy consumption during production. A cheaper alternative can be found in thin-film silicon solar cells. The thin-film silicon used in this type of solar cells is in a different phase than c-Si and usually alloyed with hydrogen. The most common thin-film silicon phases are hydrogenated amorphous silicon (a-Si:H) and hydrogenated micro-crystalline silicon (Âμc-Si:H). Thin-film silicon solar cells are made very thin, they use less than 5 micrometers of silicon, leading to less material consumption in comparison to bulk 250 micrometer thick c-Si solar cells. Furthermore, the production of thin-film silicon solar cells consumes less energy than crystalline silicon solar cell technologies. Both a-Si:H and Âμc-Si:H are disordered semiconductors, which in general require complex models for describing material and device properties. A device simulator specially designed for such disordered semiconductor materials, can be a useful tool for interpretation of experiments and thus analyze the properties of these materials and solar cells. In this thesis we address the characterization of several key parameters of a-Si:H and Âμc-Si:H for the use of these materials in solar cells. To this end we used and adapted the Advance Semiconductor Analysis (ASA) program, which is a device simulator developed at the Delft University of Technology. The density-of-states distribution in a-Si:H is of central importance for the understanding and modeling of a-Si:H devices. In order to characterize the density-of-states distribution in a-Si:H we developed a model for Charge Deep-Level Transient Spectroscopy (Q-DLTS). Q-DLTS is a technique which can be used to characterize deep-levels in a-Si:H, however, interpretation of these experiments on a-Si:H is not straight forward. By implementing our models for Q-DLTS in ASA we developed a tool for extraction of the density-of-states distribution in a-Si:H. In the case of Âμc-Si:H the value of the mobility gap is controversial. We present a simple analytical model that describes the temperature dependent behavior of Âμc-Si:H solar cells, and show that using this model an accurate value for the mobility gap can be determined. We verified our analytical model with detailed numerical simulations.