Bulk and surface recombination properties in thin film semiconductors with different surface treatments from time-resolved photoluminescence measurements

Abstract

The knowledge of minority carrier lifetime of a semiconductor is important for the assessment of its quality and design of electronic devices. Time-resolved photoluminescence (TRPL) measurements offer the possibility to extract effective lifetimes in the nanosecond range. However, it is difficult to discriminate between surface and bulk recombination and consequently the bulk properties of the semiconductor cannot be estimated reliably. Here we present an approach to constrain systematically the bulk and surface recombination parameters in semiconducting layers and reduces to finding the roots of a mathematical function. This method disentangles the bulk and surface recombination based on TRPL decay times of samples with different surface preparations. The technique is exemplarily applied to a CuInSe 2 and a back-graded Cu(In,Ga)Se 2 compound semiconductor, and upper and lower bounds for the recombination parameters and the mobility are obtained. Sets of calculated parameters are extracted and used as input for simulations of photoluminescence transients, yielding a good match to experimental data and validating the effectiveness of the methodology. A script for the simulation of TRPL transients is provided.

Single-graded CIGS with narrow bandgap for tandem solar cells

“…These results indicate that the improvements seen are a combination of reduced back surface recombination and improved bulk absorber quality. A more detailed discussion of the TRPL behavior of these samples together with a model to estimate for the relative recombination rates at the back contact compared to the bulk recombination are shown elsewhere [20]. …”

Time-resolved photoluminescence on double graded Cu(In,Ga)Se2 – Impact of front surface recombination and its temperature dependence

“…The first configuration consists of a chemical bath deposited CdS buffer layer (14 min deposition) of approximately 30-nm thickness (subsequently labeled: 14ʹ CdS). The CdS buffer layer has been shown to yield stable TRPL measurements [15] with a rather low front surface recombination velocity [16]. For the second configuration, the CdS layer was removed by etching the absorber in 5 wt.% HCl for 1 min (subsequently labeled: after HCl).…”

“…For the second configuration, the CdS layer was removed by etching the absorber in 5 wt.% HCl for 1 min (subsequently labeled: after HCl). It is reported that the free CIGS surface degrades in ambient air [10,17] and thus leads to an increased front surface recombination velocity [16]. The third configuration was done by a 10 wt.% KCN etch for 3 min followed by a chemical bath deposited CdS buffer layer of roughly 19 min (subsequently labeled: KCN, 19ʹ CdS) of the HCl-etched absorber.…”

“…The PL emission is measured spectrally integrated. Additional information of the TRPL system can be found in [16]. The temperature of the absorber was varied by using a Peltier element below the sample.…”

“…The optical generation was calculated by RayTracing with an absorption coefficient of 8:12 μm À1 as measured for a CuInSe 2 absorber [19] and used in previous studies [16]. Thus, in the simulations, the excitation of electron-hole pairs is independent of the GGI grading in the CIGS front region, which simplifies a comparison of the simulated PL decay curves.…”