Application of the spectral response of photoluminescence in photovoltaics

Abstract

Over the last 10 years, photoluminescence has been widely adopted as a common characterisation tool in photovoltaics owing to its low cost, fast acquisition time, and contactless nature. This adoption stimulated further investigations by research groups into numerous aspects of photoluminescence characterisation, including photoluminescence’s: excess carrier density dependence, temperature dependence, emission wavelength dependence, and spatial resolution of luminescence. The result of this research is that a myriad of device and material parameters can now be determined with photoluminescence measurements. However, little attention has been given to the illumination wavelength dependence of the photoluminescence emission from a sample. This is the focus of this thesis. The illumination wavelength dependence of a photovoltaic device is inherently important as the device is designed to operate under a broad spectrum source, i.e. the sun. Standard metrics used to quantify the performance of a device as a function of illumination wavelength include: the external quantum efficiency, which requires contact to be made to the sample; and the absorptance, which is typically determined as an area averaged value for the entire sample. This thesis investigates if these common metrics can be determined from measurement of the spectral response of photoluminescence. Well established theory is used to determine the conditions in which the spectral response of photoluminescence provides a relative measure of the external quantum efficiency and the absorptance. Experimental demonstration of the extraction of the external quantum efficiency and absorptance is then provided. In doing so, this thesis draws two conclusions: that a relative measure of a device’s external quantum efficiency can be obtained contactlessly, enabling measurements to be performed on partially processed samples; and that the spectrally and spatially resolved absorptance of a device can be determined from PL images taken with variable illumination wavelengths, providing fast access to spatially resolved data.