Mechanical defects in solar cells can be found using advanced image processing
Mechanical defects in solar cells can be found using advanced image processing lead image
Cracks in solar cells are mechanical defects that can negatively affect the performance and reliability of the module. This is particularly important to applications in space—such as solar panels on satellites—which are constantly exposed to extreme conditions and need an efficient method for detecting defects. Currently, electroluminescence techniques are the predominant method of detecting cracks, but these are not ideal due to their lack of tunability. A new article introduces an imaging method to increase the visibility of these cracks.
Instead of using electroluminescence techniques, Zimmermann used a method based on photoluminescence, which allows the photocurrent through each subcell in a multijunction cell to be separately adjusted, making it possible to decrease background contrast to better see defects.
The system can be modeled as an equivalent circuit, where the cracks behave like current sinks that force the lateral current flow between the different subcells to the site of the crack. By changing the illumination conditions of two adjacent subcells, one can adjust the magnitude and even the direction of this current flow, which enables the control of the local junction voltage in each layer and in effect, its photoluminescence intensity.
Applying this property, Zimmermann reversed the lateral current flow between the subcells to create images with reversed contrasts. The background can then be reduced by subtracting two images of opposite contrast from one another, which enhances the visibility of the defects.
Variations in the backgrounds of these images make it difficult for image processing algorithms to automate crack detection. According to the author, the new method will make defect detection easier for automated software.
Source: “Photoluminescence-based detection of mechanical defects in multijunction solar cells,” by Claus G. Zimmerman, Journal of Applied Physics (2019). The article can be accessed at https://doi.org/10.1063/1.5106414 .
