Strain tuning of quantum dot optical transitions via laser-induced surface defects

We discuss the fine-tuning of the optical properties of self-assembled quantum dots by the strain perturbation introduced by laser-induced surface defects. We show experimentally that the quantum dot transition redshifts, independently of the actual position of the defect, and that such frequency shift is about a factor five larger than the corresponding shift of a micropillar cavity mode resonance. We present a simple model that accounts for these experimental findings.

Figure 1

Sketch of the micropillar samples with embedded quantum dots (see text for details). A laser beam (about 500 mW power, $\lambda =532$ nm) is tightly focused on the sample to create small defects, far away from the cavity area. Such defects (see AFM image in the inset) appear as holes, about 2 $\mu$m wide and 2 $\mu$m deep, with some material removed and deposited on the edges.

Figure 2

Spectrally resolved photoluminescence as a function of the applied bias voltage for increasing number of holes burned around $20\mu$m away from the dot. The energy-shifting lines correspond to quantum dots (Stark shift), while the constant ones refer to the emission from cavity modes. Due to the effect of the holes burned, the QD transitions (labeled as QD-1 and QD-2 in the figure) shift to lower and lower energies (about about 50–100 $\mu$eV for each hole burned).

Figure 3

The top plot shows the energy of four QD optical transitions with respect to their original energy for 6 holes burned in total on the sample in different positions at the same distance from the cavity region (see inset for hole location). All transitions always redshift ($\Delta E_{QD}$ keeps decreasing for each hole burned), independently of the orientation of the hole position with respect to the cavity and the crystal axes. On the bottom plot the corresponding energy difference for the two orthogonally polarized submodes M${}_{00}^x$ and M${}_{00}^y$ of the fundamental cavity mode with respect to their original (0 holes) resonance energy. M${}_{00}^x$ and M${}_{00}^y$ can either blueshift or redshift depending on the position of the hole. The dependence of the cavity mode shift on the hole position is described by the model in Sec. IIB. Resonant frequencies were determined by a Lorentzian fit of the photoluminescence spectrum peaks, with an accuracy of the order of 4–5 $\mu$eV.

Figure 4

Energy shifts for quantum dot and optical cavity mode for six different quantum dots. The dotted red line is a linear fit. On the inset histogram of the ratios between the shift of the dot and the shift of the cavity mode. On average the dot frequency shifts about 5 times more than the cavity mode resonance.

Figure 5

Parabolic confinement band-edge approximation for the quantum dot potential.

Figure 6

Values of $\eta$, ratio between the frequency shifts of the quantum dot optical transitions and cavity mode resonance, for different dot dimensions.