Compositional disorder in $\mathrm{Ga}{\mathrm{As}}_{1-x}{\mathrm{N}}_x:\mathrm{H}$ investigated by photoluminescence

Compositional disorder is investigated by means of photoluminescence (PL) and PL excitation (PLE) measurements in as-grown and hydrogen-irradiated $\mathrm{Ga}{\mathrm{As}}_{1-x}{\mathrm{N}}_x$ samples $(x⩽0.21\%)$. The dependence of the linewidth of the PLE free-exciton band on N concentration agrees well with that predicted by a theoretical model developed for a purely random alloy. We also find that hydrogen irradiation and ensuing nitrogen passivation reduce significantly the broadening of the free-exciton band. This result is consistent with a removal by hydrogen of the static disorder caused by nitrogen. Finally, an analysis of the dependence of the Stokes shift on the free-exciton linewidth shows that free carriers are thermalized even at low temperature, another indication of a low degree of disorder in the investigated samples.

Figure 1

Peak-normalized photoluminescence (PL, dotted line) and PL excitation (PLE, solid line) spectra at $T=10\mathrm{K}$ of a selection of the $\mathrm{Ga}{\mathrm{As}}_{1-x}{\mathrm{N}}_x$ samples studied in this work. Nitrogen concentrations are given. For the hydrogenated samples, the effective N concentration $x_{\mathrm{eff}}$ is also specified. Upward and downward arrows mark the free-exciton energy for PLE and PL spectra, respectively. $(e,C)$ indicates the free-electron to neutral-carbon recombination, and NC indicates carrier recombination from nitrogen complexes. From topmost to bottommost PLE spectrum the detection energies are $E_{\det }=1.423\mathrm{eV}$ $(x=0.21\%)$, $1.472\mathrm{eV}$ $(x=0.095\%)$, $1.466\mathrm{eV}$ $(x_{\mathrm{eff}}=0.076\%)$, $1.448\mathrm{eV}$ $(x=0.043\%)$, and $1.511\mathrm{eV}$ $(x_{\mathrm{eff}}<0.001\%)$.

Figure 2

Comparison between the photoluminescence (PL, circles) and PL excitation (PLE, squares) spectra at $T=10\mathrm{K}$ of: (a) an untreated $x=0.095\%$ sample (PLE detection energy $E_{\det }=1.472\mathrm{eV}$); (b) a hydrogenated $x=0.095\%$ sample resulting in an effective N concentration $x_{\mathrm{eff}}=0.076\%$ $(E_{\det }=1.466\mathrm{eV})$ [$(e,C)\text{−}\mathrm{Ga}\mathrm{As}$ in the PL spectrum indicates the free-electron to neutral-carbon recombination in the GaAs buffer layer]; (c) a hydrogenated $x=0.095\%$ sample resulting in an effective N concentration $x_{\mathrm{eff}}=0.001\%$ $(E_{\det }=1.510\mathrm{eV})$. The result of a fit of the free-exciton (FE) PLE band using two Gaussians is also shown (solid line), along with the FE light-hole (lh, dotted line) and heavy-hole (hh, dashed line) components.

Figure 3

Dependence of the experimental free-exciton linewidth $w$ (full symbols) on the effective nitrogen concentration $x_{\mathrm{eff}}$. Full circles refer to as-grown samples and full triangles refer to the hydrogenated samples. The theoretical predictions of Eq. (1) are also shown in the figure (open squares; the dashed line is a guide to the eye). Notice that the lowest $x$ value used in the theoretical model is $x=0.001\%$.

Figure 4

Ratio between the photoluminescence (PL) and PL excitation (PLE) spectra of the sample with $x=0.043\%$ taken at $10\mathrm{K}$ and with excitation power density $P_{\mathrm{ex}}=3\mathrm{W}∕{\mathrm{cm}}^2$ (full diamonds, right axis). The dashed line is the result of a fit performed with the Boltzmann distribution function $(T_{\mathrm{c}}=14.3\mathrm{K})$, see Eq. (3). The PL and PLE spectra used in this procedure are also shown (solid lines, left axis). PLE detection energy $E_{\det }=1.448\mathrm{eV}$.

Figure 5

Dependence of the product of the Stokes shift SS and the carrier thermal energy $k_{\mathrm{B}}{T}_{\mathrm{c}}$ on the experimental linewidth squared $w^2$. $w$ values as derived from PLE, see text, are reported in Table . Full circles (triangles) refer to as-grown (hydrogenated) samples. The dashed line gives the behavior described by Eq. (4). Data from Ref. 38 (empty squares) are also shown for comparison.