Pauli blocking dynamics in optically excited quantum dots: A picosecond excitation-correlation spectroscopic study

State-filling dynamics in self-assembled InAs/GaAs quantum dots (QDs) is studied through their steady-state photoluminescence (PL) using a variant of picosecond excitation-correlation (EC) spectroscopy. Steady-state PL showed an interesting transition from bimolecular recombination at low-excitation fluence to excitonic recombination at higher fluence. As for the EC signal, while the ground-state response is always snubbed when the two excitation pulses are temporally nearly coincident, the excited-state response can either be enhanced or reduced, depending on the excitation fluence. The time evolution of this response is studied for the first three levels in a QD ensemble. A minimal theoretical model, which combines carrier loss kinetics with the principle of detailed balance and the Pauli exclusion principle, quantitatively reproduces the observations.

Figure 1

Schematic of our EC setup: BS, beam splitter; RR, retroreflector; M1 and M2, mirrors; L1, L2, and L3, lenses; S, sample; SP, spectrometer; and D, array detector.

Figure 2

Photoluminescence spectra at different excitation fluences $I$ showing multiple peaks at higher fluence. Curves i–v correspond to $I=13$, 45, 128, 256, and 448 $\mu$J/cm${}^2$, respectively. Three clearly visible peaks are assigned, respectively, to the ground (G), the first excited (X1), and the second excited (X2) excitonic states. The inset shows the excitation fluence dependence of the spectrally integrated (between 0.95 and 1.3 eV) PL ($\square$) and the PL intensity at the peak G ($◯$) on a log-log plot. Data are fitted with a quadratic (linear) $I$ dependence in the low- (high-) excitation fluence regime. While linearity continues for the integrated PL up to the highest fluence, the PL signal at the peak G begins to show sublinearity at high fluence.

Figure 3

Normalized PL and EC spectra [as defined by Eq. (1)] for (a) $I=2\times 38$ $\mu$J/cm${}^2$ and (b) $I=2\times 192$ $\mu$J/cm${}^2$. Note that the sign of the EC signal for the X1 state changes from negative to positive between (a) and (b).

Figure 4

Delay dependence of the EC signal for excitation fluence (a) $I=2\times 38$ $\mu$J/cm${}^2$ and (b) $I=2\times 192$ $\mu$J/cm${}^2$ measured at energies indicated by arrows in Fig. 3. The sign of the EC signal from the X1 state is reversed at higher fluence. The solid lines are fits to the data by the model described in the text.