QIPIRC

An InAs/GaAs quantum dot is a 20-nm diameter by 3-4 nm high disk of InAs embedded in GaAs. This acts as the proverbial three dimensional finite quantum well, trapping electrons and holes in discrete energy levels. Resonant excitation with a circularly polarized laser creates a spin-polarized electron-hole pair in the dot, known as an exciton. We study a 4-level system as shown in figure 1. The 4 states are: no exciton “0”, spin-(up/down) exciton “X”, and biexciton “XX”. There are two laser pulses: the “control” couples the spin-up exciton and biexciton states, whilst the weaker “probe” pulse monitors the no-exciton to spin-up exciton states.  A photocurrent detection technique is used to the total number of excitons created by the sequence of laser pulses.

In the first experiment the two pulses are temporally overlapped. The control is tuned to the exciton-biexciton transition, and the photocurrent is measured as a function of the detuning of the probe laser, and the pulse-area of the control. As can be seen in figure 2(a) the control pulse splits the exciton transition into an Autler-Townes doublet with a Rabi-splitting proportional to the pulse-area of the control.  A splitting of up to 0.5 meV was observed.

In the second experiment, the pulse-area of the control is fixed, and the photocurrent is measured as a function of the detuning of both the probe and control pulses. In fig 2(b), an anti-crossing between the sequential absorption of an exciton (0-X) followed by a biexciton (X-XX), and creation of the biexciton through direct two-photon absorption (0-XX) is observed.  This shows that the control laser has mixed the exciton and biexciton states to form dressed states. We now study the Autler-Townes splitting in the time-domain.

To time-resolve the Autler-Townes doublet, the control pulse is tuned to the exciton-biexciton transition, and its pulse-area fixed. The photocurrent is measured as a function of the detuning of the probe, and the inter-pulse time-delay. As can be seen in figure 3(left) the Rabi splitting follows the Gaussian envelope of the control pulse, shown as a red-line, demonstrating that the Rabi splitting and hence the composition of the dressed states adiabatically follows the control laser, on a picosecond timescale. This is encouraging for the prospects of using coherent control schemes based on adiabatic following in semiconductor quantum dots.

The grand finale is to observe a beat between the excitonic dressed states. The time-duration of the probe pulse is reduced. Figure 3(right) shows the photocurrent versus inter-pulse time-delay for a series of different control pulses of different pulse-areas. The clearest interpretation of the data is a time-resolved measurement of the Rabi rotation between the exciton and biexciton states. Where the probe creates an exciton triggering the X-XX Rabi oscillation at a time equal to the negative time-delay, and thus by varying the time-delay one can observe the time evolution of the Rabi rotation.   Because of the experimental difficulties of varying the time duration of a picosecond laser pulse, excitonic Rabi rotations are usually measured as a function of pulse-area by varying the incident laser power. However, Rabi oscillations are generally understood in the time-domain, so it is aesthetically pleasing to observe the Rabi oscillation in the time-domain by optically gating the control field.