Front‐ and rear‐side photoluminescence: recombination, depth profiles of excess carriers and optical band gap of Cu(In,Ga)Se₂ in a three‐layer system

Abstract

Spectral photoluminescence emitted from the front compared to the rear side of a semiconductor layer like a photovoltaic absorber is shown to exhibit significant differences in the highenergy regime. This arises from the excess-carrier depth profile and the absorption of photoluminescence photons during their way through the semiconductor layer depending on photon energy, distance to the absorber exit, and absorption coefficient.We get access to surface-recombination velocities, the minority-carrier diffusion length, the excess-carrier depth profile and the optical band gap by fitting photoluminescence spectra via numerical modeling. The numerical modeling is based on an one-dimensional three-layer system that includes multiple reflection. This procedure is exemplarily demonstrated for a thin-film system based on Cu(In,Ga)Se 2 .

Avoiding arbitrarily wrong microluminescence statistics due to a non-quantitatively calibrated setup

“…Further information about the deposition, the structural, and opto‐electronic properties of these and similar samples can be found in Refs. (sample A) and Refs. (sample B), respectively.…”

“…Compared to CIGSe, the wide-gap semiconductor CIS shows a smaller efficiency and larger inhomogeneities in terms of integrated PL yield and optical band gap. In addition, lateral inhomogeneities of the stoichiometry, potential fluctuations, and band gap gradients in this ternary (CIS) or quaternary (CIGSe) semiconductor as well as the influence of grain boundaries are well reported [9,11,14,22,[25][26][27].…”

“…Due to experimental difficulties, microluminescence setups are often calibrated with respect to spectral alterations only, although recent publications also cover the quantitative calibration of a microluminescence PL setup and an unexplainable mismatch between the expected and the measured QFL‐splitting distribution determined with an uncalibrated microluminescence setup has already been published . Nonetheless, there are several publications which state explicitly or assume implicitly that for determination of the lateral variation of the QFL‐splitting it is sufficient to calibrate the setup spectrally but not with respect to absolute quantities. It is assumed that the determined mean only shifts by an offset and that this offset and hence the quantitative calibration could be neglected.…”