On the Scalar Scattering Theory for Thin-Film Solar Cells

Abstract

Nano-textured interfaces between two media of different refractive indices scatter light. The angular distribution and the intensity of the scattered light are deter- mined by the geometry of the nano-textures and the difference of the refractive indices of the two media. Thin-film silicon solar cells (TFSSC), which convert sunlight directly into electricity, have nano-textured interfaces. These interfaces scatter the light incident on the solar cell. The scattering leads to a longer average path length of the photons in the absorber layer of the solar cell. Therefore more light can be absorbed and thus converted to electricity. To introduce nano-textured interfaces into the solar cells, usually transparent conductive oxide (TCO) layers are used. Some TCO materials obtain nano-textured surfaces during the production process, while others are made rough by post processing, e.g. by etching. Nano-textures have been successfully implemented in TFSSC for almost 30 years by academia and industry; however, theoretical investigations on the relation between the nano-textures and the scattered electromagnetic fields have only been performed for about ten years. It is very important to investigate how the nano-textures can be optimized for scattering. In this thesis a scattering model is developed to tackle this important problem. The scattering model is based on the scalar scattering theory, i.e. it neglects the vector- character of the electromagnetic field and thus the light. Despite this strong assumption we have demonstrated that the model is suitable for simulating descriptive parameters of the scattered field in both reflection and transmission. The model is based on the fact that the transmitted field behind the nano-texture and the scattered field are related via Fourier transforms. By making simple assumptions for the transmitted field the model can be implemented using Fast Fourier trans- form algorithms, i.e. the model is very fast. The scattering model is formulated such that in principle it works for rough interfaces between arbitrary materials. We successfully evaluated it for several of these interfaces. We further showed that the model is also able to produce first predictions for the scattering parameters at oblique incidence. However, in this case the deviations between measured and simulated values are larger. Combining the scattering model with the ASA opto-electrical device simulator allows predicting how the nano-textures affect the performance of solar-cells. This combination can also be used to perform the major motivation for the development of scattering models: To investigate how the morphology of the nano- textures can be optimised. For this optimisation we use the “simulated annealing” optimisation algorithm. The optimisation and a subsequent evaluation reveal that the lateral feature size of the nano-textures is crucial for scattering into large angles: The smaller the lateral feature size, the more light is scattered into large angles. If, however, the lateral feature size becomes too small, less light is scattered since the nano texture then appears as effective medium. The vertical feature size hardly influences the shape of the scattered field. Nonetheless, it determines the fraction of the total light that is scattered away from the specular direction. If the rms-roughness, a measure for the vertical modulation of the texture, is kept constant, a nano- texture with the optimal lateral feature size is preferable to a texture that consists of a superposition of textures with different lateral feature sizes. However, due to the effect of the nano-textures on the electrical properties of the solar cells, a superposition of a texture consisting of large lateral and vertical features with another texture with small lateral and vertical features is preferable to a texture consisting of small lateral but large vertical features, i.e. sharp spikes. The results of our work give the direction to push absorption in solar cells towards the theoretical limits.