Investigating efficiency droop in InGaN/GaN quantum well structures using ultrafast time‐resolved terahertz and photoluminescence spectroscopy

Abstract

The mechanisms governing efficiency droop in an In0.18Ga0.82N/GaN multiple quantum well structure were investigated using a combination of ultrafast time‐resolved terahertz and photoluminescence spectroscopy. From excitation fluence dependent studies, a reduction in the room temperature photoluminescence efficiency to 3% of its maximum value was observed for an excitation fluence of 0.96 mJcm‐2. A correlation was found between the onset of efficiency droop and the emergence of a peak on the high‐energy side of the quantum well emission with a 1/e decay time of 19.6 ps. These characteristics were attributed to the saturation of localised states and the population of higher energy delocalised states. Time‐resolved studies revealed different scaling behaviours between the terahertz and photoluminescence decay dynamics, suggesting that the saturation of localised hole states may be playing a part in the onset of efficiency droop. (© 2016 The Authors. Phys. Status Solidi C published by WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

Disentangling the Impact of Point Defect Density and Carrier Localization-Enhanced Auger Recombination on Efficiency Droop in (In,Ga)N/GaN Quantum Wells

“…By inserting the calculated B ( N ) and C ( N ) into eqs and , ϕ­( t ) may be fitted to the PL transients for each of the samples, as also shown in Figure a. In each case, the only fitting parameters are A and N 0 , the latter of which is allowed to vary since previous studies have shown that for high N 0 the value of N can decrease significantly within the first 100 ps after excitation, i.e., on a timescale less than the time resolution of the experimental apparatus. This fit yields estimates for A in the range from 2 × 10 6 to 4 × 10 6 s –1 , which is similar to previous reports. ,, As noted above, a previous study of the spatial variation of the PL of these samples has shown that the intensity can vary by as much as 40% from point to point; however, nonetheless, this range of values may be regarded as broadly representative of this sample set.…”

“…By inserting the calculated B(N) and C(N) into eqs 2 and 3, ϕ(t) may be fitted to the PL transients for each of the samples, as also shown in Figure 4a. In each case, the only fitting parameters are A and N 0 , the latter of which is allowed to vary since previous studies 22 have shown that for high N 0 the value of N can decrease significantly within the first 100 ps after excitation, i.e., on a timescale less than the time resolution of the experimental apparatus. This fit yields estimates for A in the range from 2 × 10 6 to 4 × 10 6 s −1 , which is similar to previous reports.…”