Positive binding energy of a biexciton confined in a localization center formed in a single ${\text{In}}_x{\text{Ga}}_{1-x}\text{N}/\text{GaN}$ quantum disk

We report microphotoluminescence spectroscopy performed on individual and ensemble InGaN/GaN quantum disks (Q-disks). The typical spectrum of a single Q-disk exhibited the contribution of localization centers (LCs) formed in the InGaN active layer of the Q-disks, characterized by sharp lines appearing on the low energy side of the spectra. In addition, a broader emission peak identified as the luminescence of the quasi-two-dimensional (2D) InGaN active layer surrounding the LCs appears systematically at higher energy. Time-resolved photoluminescence experiment performed on single Q-disks exhibited the excitonic transfer, from the 2D InGaN active layer to LCs, at the submicroscopic scale. Excitation power dependence studies and linear polarization analysis allowed us to identify a biexciton complex confined in a LC in a single Q-disk with a surprising positive binding energy of 13 meV. The absence of screening effect by increasing the excitation power density and the fast excitonic radiative lifetime of a few hundred picoseconds that we measured on several individual Q-disks indicate that the absence of internal electric field in the structure can explain the observed positive biexciton binding energy.

Figure 1

(Color online) (a) Macroscopic PL spectrum. (b) a $\mu \text{PL}$ image performed on an ensemble of Q-disk.

Figure 2

(a) $\mu \text{PL}$ spectrum of a single Q-disk (called ${\text{Q-disk}}_1$) at 20 K. Inset image: schematic illustration of a Q-disk band structure according to our experimental results. (b) Scanning electron microscopy image of the ${\text{Q-disk}}_1$.

Figure 3

Integrated intensity of the emission lines $X_1$, $X{X}_1$, and $X_2$ as a function of the excitation power density plotted in bilogarithmic scale. The solid lines show the linear power-law fitting results. Inset image: spectra of the ${\text{Q-disk}}_1$ under three different excitation power densities: 120, 220, and $2.5\times {10}^3\text{W}/{\text{cm}}^2$.

Figure 4

(a) $\mu \text{PL}$ spectra of the ${\text{Q-disk}}_1$ analyzed at three different angles $\theta$. (b) Polar representation of the integrated intensity of the emission lines $X_1$, $X{X}_1$, and $X_2$ as a function of the analyzer angle over $180°$. The angle $\theta =0°$ is a priori not related to any crystallographic direction of the Q-disk structure.

Figure 5

(Color online) Time dependence of the PL intensity of an ensemble of Q-disks emitting around 2.9 eV [see Fig. 1a]; the vertical axis is in logarithmic scale. The data had been fitted by a biexponential function.

Figure 6

(Color online) Time dependence of the $\mu \text{PL}$ intensity of (a) the emission peak $E$, and the emission lines (b) $X_1$ and (c) $X{X}_1$. The dashed line is the biexponential fit obtained for each decay curve. The vertical axis is in logarithmic scale.

Figure 7

Excitation power density dependence of the $\mu \text{PL}$ spectra of a single InGaN/GaN Q-disk (called ${\text{Q-disk}}_2$) acquired at 10 K.