Dual-peak electroluminescence spectra generated from Al n /12Ga1-n/12N (n = 2, 3, 4) for AlGaN-based LEDs with nonflat quantum wells

Abstract

Abstract The metastability of Al n /12Ga1-n/12N (n = 2, 3, and 4) was investigated by the statistical analysis of electroluminescence (EL) spectra having dual peaks with a peak-to-peak distance (pp) of >10 nm generated from nonflat Al x Ga1- x N (x ∼ 0.2) quantum wells (QWs) fabricated on c(0001) sapphire substrates with a miscut of 1.0° towards the m[1 1 ˉ 00] axis. To explain the origins of the dual-peak EL (DPEL) spectra, which are often observed for AlGaN-QWs with Ga content of greater than 0.7, a nonflat QW model incorporating two metastable compositions, Al(n-1)/12Ga1-(n-1)/12N and Al n /12Ga1-n/12N (n: integer), is proposed. By the statistical analysis of peak wavelengths in DPEL spectra and the verification of EL spectral shapes, two series of featured EL peak wavelengths with intervals of 2–3 nm were obtained from five out of six LED wafers. The two series of featured EL peak wavelengths were assigned by comparison with the calculated EL wavelengths. Then, Al2/12Ga10/12N and Al3/12Ga9/12N were determined to be the origins of peaks with the longer and shorter wavelengths in the DPEL, respectively, in addition to the metastable Al n /12Ga1-n/12N (n= 4–9) compositions observed in our previous studies. When DPEL (pp> 10 nm) appeared, the difference in QW thickness between Al2/12Ga10/12N and Al3/12Ga9/12N tended to be larger than one monolayer (ML), indicating a significant amount of Ga or GaN mass transport. Furthermore, the Al2/12Ga10/12N and Al3/12Ga9/12N QWs are considered to have thicknesses of m ML (m: consecutive integers), suggesting the 1 ML configuration of Al and Ga atoms on the c(0001) plane. In addition, the DPEL obtained from nonflat Al x Ga1- x N (x ∼ 0.25) QWs by another research group was shown to be related to two metastable Al n /12Ga1-n/12N (n = 3, 4), similarly to our one exceptional LED wafer, which also agreed with the model proposed in this work.