Impact of $\mathit{ex}\mathit{situ}$ rapid thermal annealing on magneto-optical properties and oscillator strength of In(Ga)As quantum dots

We discuss the influence of a rapid thermal annealing step on the magneto-optical emission properties of In(Ga)As/GaAs quantum dots. We map out a strong influence of the growth and annealing parameters on the excitons' effective Landé $g$ factors and in particular on their diamagnetic coefficients, which we directly correlate with the modification of the emitters' shape and material composition. In addition, we study the excitons' spontaneous emission lifetime as a function of the annealing temperature and the dot height and observe a strong increase of the emission rate with the quantum dot volume. The corresponding increase in oscillator strength yields fully consistent results with the analysis of the diamagnetic behavior. Specifically, we demonstrate that a rapid thermal annealing step of $850{}^{\circ }\mathrm{C}$ can be employed to increase the oscillator strength of as-grown InAs/GaAs QDs by more than a factor of 2.

Figure 1

(a) SEM image of surface QDs before the capping process. (b) STEM image of a QD after partial capping ($d=3\text{nm}$) and annealing and overgrowth by 50 nm of GaAs.

Figure 2

Plot of the normalized intensity of the right- and left-circular polarized part of the emission of a QD for different magnetic fields (reference sample with a cap thickness of $d=3\text{nm}$). An increasing Zeeman splitting between ${\sigma }^{+}$ and ${\sigma }^{-}$ is observable as well as a superimposed diamagnetic shift.

Figure 3

Zeeman splitting vs the magnetic field for cap thicknesses $d$ of 2 nm (a), 3 nm (b), and 4 nm (c) as well as the determined effective $g$ factors in dependence of the ex situ annealing temperature (d). The decrease of the absolute $g_X$ value for higher annealing temperatures implies an increasing lateral expansion of the QDs and a modification of the composition and a strain variation in the QDs.

Figure 4

Diamagnetic shift vs square of the magnetic field for cap thicknesses $d$ of 2 nm (a), 3 nm (b), and 4 nm (c) as well as the determined diamagnetic coefficient $\chi$ as a function of the ex situ annealing temperature (d). For increasing cap thicknesses and higher annealing temperatures a distinct increase of $\chi$ is observable.

Figure 5

(a) Exemplary emission of the QD ensemble of the reference sample with $d=2\text{nm}$. In addition the lifetimes of the ensemble for different wavelengths are shown. (b) Overview of the lifetimes acquired at the spectral positions used for determination of the diamagnetic shift and the Zeeman splitting vs the cap thickness $d$ and the RTA temperature. Higher annealing temperatures lead to a decrease of the mean lifetime of the ensemble and thereby to an increase of the OS.

Figure 6

Overview of the oscillator strength determined by the lifetime measurements using Eq. (3) (red, circles) and by the diamagnetic shift $\chi$ using Eq. (5) (black, squares), respectively. Both methods show a significant increase of the oscillator strength for higher annealing temperatures and almost all values for $f$ agree within the error bars. The maximum OS $f=25.7\pm 5.7$ from the lifetime measurements as well as the maximum OS $f=34.7\pm 5.2$ from the diamagnetic shift have been obtained for a cap thickness of $d=3\text{nm}$ and an annealing temperature of $850{}^{\circ }\mathrm{C}$.

Figure 7

Ensemble photoluminescence spectra recorded at 10 K for the sample with a 2-nm-high partial cap, which is discussed in the main text. The spectra correspond to the four selected annealing conditions.