Characterization of the local charge environment of a single quantum dot via resonance fluorescence

We study the photon-statistical behavior of resonance fluorescence from self-assembled InAs quantum dots (QDs) as a function of the density of free charge carriers introduced by an above band-gap laser. Second-order correlation measurements show bunching behavior that changes with above-band laser power and is absent in purely above-band excited emission. Resonant photoluminescence excitation spectra indicate that the QD experiences discrete spectral shifts and continuous drift due to changes in the local charge environment. These spectral changes, combined with the tunneling of charges from the environment to the QD, provide an explanation of the bunching observed in the correlations.

Figure 1

Quantum dot sample with optical excitation and collection geometry. The resonant laser is focused on the cleaved face of the sample in order to couple into the waveguide mode of the cavity. The fluorescence is collected normal to the sample surface. The above-band laser at 633 nm is focused through the collection lens onto the QD location. ${\lambda }_0$ is the cavity resonance wavelength in vacuum (930 nm) and $n$ is the refractive index of GaAs ($\sim 3.5$).

Figure 2

Time-resolved fluorescence from a single QD under pulsed resonant excitation and without above-band excitation. The data (orange dots) are fit by an exponential decay model convolved with the measured instrument response function of the SPCM (black curve). The blue curve is the convolved result. The inset shows the lifetimes extracted from similar time-resolved fluorescence measurements with different levels of above-band laser power represented as a fraction of the saturation power ${\mathrm{P}}_1=28.5\mu \mathrm{W}$. The lifetime varies little with above-band laser power, with an average value of $T_1=518\pm 3$ ps. The average $T_1$ is depicted by the dark line in the inset, while the gray area is the standard uncertainty range.

Figure 3

(a)–(c) Three examples of normalized resonant photoluminescence excitation spectra (RPLE) at above-band laser powers of zero, $3.1\times {10}^{-7}{\mathrm{P}}_1$ and $7.7\times {10}^{-5}{\mathrm{P}}_1$, respectively. The filled curves are the individual Voigt peaks used to do the fitting; the blue curve along the orange data points is the sum of these individual peaks. Zero detuning is defined as the middle point of the two Voigt peaks with largest amplitude (green) in (a), which corresponds to 928.3713 nm. Each curve is normalized to its own maximum. (d) 2D plot of 16 RPLE spectra taken at different above-band laser powers plotted in a logarithmic scale on the vertical axis. The color-scaled spectral intensity is normalized to the overall maximum of the measured fluorescence intensity. The grey lines denote the three spectra in (a), (b), and (c). The black dots are the positions where correlation data are collected. Box A denotes the data shown in Fig. 5 and box B those in Fig. 5. The white dashed lines indicate the boundaries for the different above-band power regimes. (e) An example of polarization-dependent RPLE without above-band laser. Two RPLE spectra were recorded using a linear polarizer oriented at ${114}^{\circ }$ (blue) and ${27}^{\circ }$ (red) from the horizontal. Their sum is displayed as the orange curve with grey filled area. The red and blue peaks are the same shape but displaced and with different amplitudes, which implies that the two peaks are the orthogonally polarized emission from the two fine structure split states of a neutral QD.

Figure 4

(a) Center detunings of resonance peaks. For the RPLE spectra that comprise Fig. 3, the center detuning of each Voigt profile in the fit is plotted vs the corresponding above-band laser power. The green square curve corresponds to trap configuration (00), the red circle for trap configuration (11), the blue up-triangle for trap configuration (10) and black down-triangle for trap configuration (01). The lines are guides for the eye. The horizontal dashed lines give the boundaries of different above-band power regimes. (b) Spectrally integrated intensity of RPLE spectra in Fig. 3. The baseline offset due to the above-band excitation has been subtracted so that the curve represents the emission solely due to resonant excitation; this effect is only significant in the ultrahigh-power regime. The error bars correspond to experimental fluctuation and shot noise. The curve is normalized to its maximum value. The grey vertical dashed lines correspond to the boundaries of different above-band power regimes.

Figure 5

Second-order correlation function of the fluorescence, $g^{\left(2\right)}\left(\tau \right)$. The legends list the above-band laser power in fractions of the saturation power ${\mathrm{P}}_1$. (a) The $g^{\left(2\right)}\left(\tau \right)$ measured at the black dots in box A in Fig. 3. The resonant laser is tuned to $\sim 1.96$ GHz detuning with power of 0.1 ${\mathrm{P}}_0$ and the above-band laser power is varied from zero to $6.32\times {10}^{-7}{\mathrm{P}}_1$. A significant increase of bunching amplitude and small shrinking of bunching time can be seen. The inset is the correlation data measured with only above band-gap laser excitation at $0.23{\mathrm{P}}_1$. (b) The $g^{\left(2\right)}\left(\tau \right)$ from box B in Fig. 3. The resonant laser is at $\sim 0.56$ GHz detuning with a power of 0.1 ${\mathrm{P}}_0$. The power of above-band laser is varied from $1.24\times {10}^{-6}{\mathrm{P}}_1$ to $7.69\times {10}^{-3}{\mathrm{P}}_1$.

Figure 6

Possible charge trap locations consistent with the measured Stark shifts. Red and blue correspond to positive and negative trap polarity, respectively; the QD is represented schematically at the origin. The solid lines are for trap $\alpha$; the dashed lines are for trap $\beta$. The shaded regions denote locations consistent with the confidence range of $\mathrm{\Delta }{\nu }_2$; the confidence range of $\mathrm{\Delta }{\nu }_1$ is small enough that the corresponding region is hidden by the solid lines.