Decoupling recombination mechanisms and trap state localization in direct bandgap semiconductors using photoluminescence decay
M. W. Gerber,R. N. Kleiman
DOI: https://doi.org/10.1063/1.5001128
IF: 2.877
2017-09-07
Journal of Applied Physics
Abstract:In this work, we show that extraction of the true bulk lifetime from the biexponential decay that follows from low initial carrier density photoluminescence decay experiments is not generally possible, and introduce new models to enable extraction of the bulk lifetime in the case where the initial carrier density exceeds the doping level. From measurements with high initial carrier density, we establish quasi-equilibrium between localized and free carrier states and accurately measure the bulk lifetime. Using our new models, we measure the time constants associated with localization processes as well as nonradiative and radiative bulk recombination in our GaAs double heterostructures grown with molecular beam epitaxy from experiments with varied excitation strength providing initial carrier densities that range from around 1014 to 1017 cm–3. We demonstrate that this approach can be applied to lightly doped (1016 cm–3) materials where the strongest excitation yields initial carrier densities that exceed the doping level. In our n-type sample, we report lifetime values of (22.7 ± 0.1) ns for bulk recombination, (73 ± 1) ns for trap-capture, (51 ± 2) ns for trap-emission, and (63 ± 2) ns for trap-decay, with a low-level injection effective radiative efficiency of (27.5 ± 0.7)%. In our p-type sample, we report lifetime values of (78.9 ± 0.3) ns for bulk recombination, (77.5 ± 0.7) ns for trap-capture, (530 ± 10) ns for trap-emission, and (177 ± 4) ns for trap-decay, with a low-level injection effective radiative efficiency of (47.0 ± 0.8)%. In comparison with the long and short lifetimes extracted from the biexponential decay with weak excitation, the mean bulk lifetime measured with strong excitation was (33 ± 2)% and (53 ± 1)% longer than the short lifetime, and (68 ± 4)% and (103 ± 3)% shorter than the long lifetime in our n-type and p-type samples, respectively. In our n-type sample, the extracted low-level injection nonradiative lifetime was (33 ± 1) ns, and it was observed to remain constant with the injection level. In our p-type sample, the high-level injection nonradiative lifetime was measured to be (30 ± 30)% larger than the low-level injection nonradiative lifetime of (140 ± 2) ns.
physics, applied