Effects of random alloy disorder, shape deformation, and substrate misorientation on the exciton lifetime and fine structure splitting of GaAs/AlxGa1−xAs (111) quantum dots

Using atomistic, million-atom screened pseudopotential theory together with configuration interaction, as well as atomically resolved structures based on experimental characterization, we perform numerical calculations on self-assembled $\mathrm{GaAs}/{\mathrm{Al}}_{x}{\mathrm{Ga}}_{1\ensuremath{-}x}\mathrm{As}$(111) quantum dots that we compare with our experimental data. We show that random alloy disorder in the barrier can cause a symmetry breaking at the single-particle level (distortions of wave functions and lifting of degeneracies) which translates into the appearance of a nonzero exciton fine structure splitting (FSS) at the many-body level. Nevertheless, our results indicate that varying the concentration of aluminum in the random alloyed barrier allows simultaneous tuning of the exciton fine structure splitting and emission wavelength without altering its radiative lifetime $\ensuremath{\tau}\ensuremath{\approx}200\phantom{\rule{0.28em}{0ex}}\text{ps}$. Additionally, the optical properties of these quantum dots are predicted to be very robust against both symmetric and asymmetric shape elongation (with $\text{FSS}\ensuremath{\le}2.2\phantom{\rule{0.28em}{0ex}}\textmu{}\text{eV}$), rendering postselection less essential under well-controlled growth conditions. On the other hand, the growth on miscut substrates introduces a structural anisotropy along the quantization axis to which the system is very sensitive: the FSS ranges between 5 and $50\phantom{\rule{4pt}{0ex}}\textmu{}\text{eV}$ while the radiative lifetime of the transition is increased up to $\ensuremath{\tau}=400\phantom{\rule{0.28em}{0ex}}\text{ps}$. The numerical results for the FSS are in perfect agreement with our experimental measurements which give $\text{FSS}=10\ifmmode\pm\else\textpm\fi{}9\phantom{\rule{0.28em}{0ex}}\textmu{}\text{eV}$ for ${2}^{\ensuremath{\circ}}$ miscut angle at $x=0.15$.