Analysis on the Internal Quantum Efficiency of Deep-Ultraviolet Emitting AlGaN Nanowires

Abstract

Abstract Body: In recent years, AlGaN nanowires grown by molecular beam epitaxy (MBE) have emerged as an appealing platform for semiconductor deep ultraviolet (UV) light-emitting devices [1, 2]. Although the progress, a better understanding of some of the fundamental physical properties of such nanowires is necessary. For example, the internal quantum efficiency (IQE) of AlGaN nanowires is often investigated by the temperature-dependent photoluminescence (PL) technique that assumes a 100% IQE at low temperatures. This assumption is, however, not always valid. This contrasts with the comprehensive understanding of IQE of AlGaN quantum wells and epilayers. Moreover, the efficiency droop of the deep UV emitting AlGaN nanowires is much less studied. In this regard, we analyzed the room-temperature IQE of AlGaN nanowires emitting at ~260 nm, being considered as the most efficient wavelength for disinfection, through a theoretical analysis on the room-temperature excitation dependent PL. A peak IQE of ~55% was derived. Moreover, a carrier delocalization mechanism was pinpointed as a droop mechanism. In this work, the AlGaN nanowires were grown on Si (111) substrates by MBE in the nitrogen-rich condition at a relatively low substrate temperature. Detailed growth conditions can be found in Ref. [3]. The power-dependent PL experiments were performed in a setup consisting of a 213 nm laser and a deep UV spectrometer. We further used a theoretical model that considers the major recombination channels in semiconductors, i.e., An as the Shockley-Read Hall (SRH) nonradiative recombination, Bn2 as the bimolecular radiative recombination, and Cn3 as the higher-order nonradiative recombination, where n is the carrier density and A, B, C are the respective recombination coefficients, to analyze the excitation power dependent integrated PL intensity (IPL). In this framework and assuming at steady state, the generation rate G follows G = An+Bn2+Cn3. Further expressing IPL = γBn2, where γ is an experimental parameter, one can have, G = A(IPL)1/2/(Bγ)1/2 + IPL/γ + C(IPL)3/2/(Bγ)3/2 (1) Here, G can be experimentally determined by G = P(1-RF)α/(Aspothv), where P is the peak excitation power, RF is the Fresnel reflection, α is the absorption coefficient, Aspot is the laser spot size, and hv is the excitation photon energy. The experimentally determined G – IPL dependence can be further fitted to Eq. (1), allowing one to determine the parameter γ; and the IQE is further calculated by IQE = IPL/γG. The scattered symbols and the solid line in Fig. 1 show the experimentally determined G – IPL dependence and the fitting curve according to Eq. (1), respectively. It is seen that a good fitting is obtained. The calculated IQE as a function of G is shown in the inset of Fig. 1. It is seen that a peak IQE of ~55% is reached, followed by a droop. To further analyze the droop mechanism, the excitation dependent PL peak energy and full-width-at-half-maximum (FWHM) were analyzed. Shown in Fig. 2, a blueshift of the PL peak energy is observed accompanied by an overall increasing trend of the FWHM, suggesting the role of the carrier delocalization in the efficiency droop. Current studies are on the Al content dependent efficiency droop behavior and will be reported.