Mutual coupling of two semiconductor quantum dots via an optical nanocavity

We present an experimental and theoretical study of a system consisting of two spatially separated self-assembled InGaAs quantum dots strongly coupled to a single optical nanocavity mode. Due to their different size and compositional profiles, the two quantum dots exhibit markedly different dc Stark shifts. This allows us to tune them into mutual resonance with each other and a photonic crystal nanocavity mode as a bias voltage is varied. Photoluminescence measurements show a characteristic triple peak during the double anticrossing, which is a clear signature of a coherently coupled system of three quantum states. We fit the entire set of emission spectra of the coupled system to theory and are able to investigate the coupling between the two quantum dots via the cavity mode, and the coupling between the two quantum dots when they are detuned from the cavity mode. We suggest that the resulting quantum V-system may be advantageous since dephasing due to incoherent losses from the cavity mode can be avoided.

Figure 1

(Color online) (a) PL spectra of the double dot-cavity system as a function of bias voltage (false-color plot). (Inset) Schematic of the layer structure of the device. (b) Peak positions of cavity mode (black circles), QD1 (red squares), and QD2 (blue diamonds) for different bias voltages, emphasizing the different dc Stark shifts of QD1 and QD2. (inset) Schematic of the energy levels when QD1 and QD2 are in resonance and both are detuned from the cavity mode, forming a V-system (indicated by the gray ellipse at $V_{\text{app}}=0.45\text{V}$).

Figure 2

(Color online) (a) High-resolution PL spectra of the investigated system as function of $V_{\text{app}}$ (false-color plot). (b) Calculated spectral function of the same system. Parameters were obtained from best fits. (c) Calculated eigenvalues ${\lambda }_0$, ${\lambda }_1$, and ${\lambda }_2$ (solid black lines), clearly showing the three branches of the double anticrossing and the triple peak in resonance. The predicted anticrossing of the two QDs when they are detuned from the mode is magnified in the inset. For the gray shaded area, we compare in (d) the observed (open circles) and the fitted, theoretical (black solid lines) spectral functions, showing the good agreement.

Figure 3

(Color online) Calculated eigenvectors of the investigated system for the eigenvalues ${\lambda }_0$ (upper panel), ${\lambda }_1$ (middle panel), and ${\lambda }_2$ (bottom panel). The plotted curves show the contributions of QD1 (red dotted lines), QD2 (blue dashed lines), and the cavity mode (green solid lines) to the individual states for different detunings as a function of $V_{\text{app}}$.

Figure 4

(Color online) Resonant spectra $\left(V_{\text{app}}=0.41–0.46\text{V}\right)$ obtained from fitting of an alternative theoretical model to the experimental data (blue circles), assuming that (a) the two quantum dot states cannot coexist (e.g., neutral exciton and charged exciton of the same quantum dot), and (c) QD2 is only weakly coupled to the cavity mode with a fixed $\hslash g_2=20\mu \text{eV}$. (b) and (d) Contour plots (false color) of the calculated spectra of the same data for a wider range showing the qualitative evolution of the peaks during the crossing.