Decoherence of Rabi Oscillations in a Single Quantum Dot

We develop a realistic model of Rabi oscillations in a quantum-dot photodiode. Based in a multiexciton density matrix formulation we show that for short pulses the two-level model fails and higher levels should be taken into account. This affects some of the experimental conclusions, such as the inferred efficiency of the state rotation (population inversion) and the deduced value of the dipole interaction. We also show that the damping observed cannot be explained using constant rates with fixed pulse duration. We demonstrate that the damping observed is in fact induced by an off-resonant excitation to or from the continuum of wetting layer states. Our model describes the nonlinear decoherence behavior observed in recent experiments.

Figure 1

Schematic band structure and level configuration of QD photodiode. (a) An electromagnetic pulse creates an exciton in the dot, and an applied gate voltage forces the electron and hole to tunnel out, generating the photocurrent signal. (b) A schematic representation of processes and levels involved in this system: $|x⟩$ and $|b⟩$ are exciton and biexciton states.

Figure 2

Contour plot of the biexciton average occupation as a function of pulse duration $t_p$ and pulse area $\Theta$. Notice long pulses do not produce biexciton population. The inset compares exciton and biexciton contribution for a 1 ps pulse.

Figure 3

Photocurrent signal as a function of pulse area $\Theta$ for 1 ps pulse showing different contributions of Eq. (5). Blue squares show results neglecting biexcitons and without the second integral in Eq. (5); yellow circles include the latter. Triangles show results including biexcitons and all terms in Eq. (5). The dashed line at $I_{\mathrm{P}\mathrm{C}}=fq\simeq 13.1\mathrm{p}\mathrm{A}$ indicates optimum value for photocurrent if only exciton channel is considered.

Figure 4

(a) Level structure and couplings including nonresonant excitation to WL. (b) Occupation probability of states considered in model as a function of pulse area $\Theta$. (c) Photocurrent signal versus $\Theta$ for 1 ps pulse. Squares show the results of our model and circles are experimental data from Ref. 4. Notice this model reproduces experimental results.