Many solar cell developers may fancy a machine where you can put a solar cell in, push a button, and receive a full diagnostic report of the solar cell – instead of spending considerable time on the measurement equipment park and understanding the methods. This article describes how a complete analysis of a solar cell can be made a routine practice.
Trying to get a comprehensive understanding of a solar cell is not the most convenient approach in the quest for higher efficiencies, but it is probably the most successful in the long run, as it focuses on exploiting the full potential of the solar cell concept.
A simulation is necessary to determine what that potential actually is. The purpose of a loss analysis is to discover which assumptions of the simulation differ from reality and to reveal the weakest points that limit the performance.
Three-dimensional device simulators use sophisticated models that account for even tiny effects. However, the losses are not easy to take apart and the dominant mechanisms are often not obvious. Because they are rather tedious, these device simulators are also not suited to be used in fit routines.
Simple analytical, one-dimensional models are much faster to calculate and can therefore be used in least-square fits. To keep the models handy it is best to reduce the number of model parameters to a minimum, which can be achieved by decoupling the interdependent problems, finding scaling laws, making simplifying assumptions, and finding one-dimensional equivalents to describe three-dimensional device features.
Some simplifications and assumptions will restrict the applicability of the models, so it is important to apply them only in the allowed range.
The big advantage in understanding your solar cell – in the sense that you get the simulated current-voltage behaviour, quantum efficiency, and reflectance spectra, etc. to agree with experimental data – is that you can use the model to calculate current or power losses in the individual parts of the solar cell. It also allows you to calculate the improvement that can be achieved when model parameters are improved, such as when a back surface reflectance can be improved from 70 per cent to 95 per cent, for example.