Probing of single quantum dot dressed states via an off-resonant cavity

In recent experiments on coupled quantum dot (QD) optical cavity systems, a pronounced interaction between the dot and the cavity has been observed even for detunings of many cavity linewidths. This interaction has been attributed to an incoherent phonon-mediated scattering process and is absent in atomic systems. Here, we demonstrate that despite its incoherent nature, this process preserves the signatures of coherent interaction between a QD and a strong driving laser, which may be observed via the optical emission from the off-resonant cavity. Under bichromatic driving of the QD, the cavity emission exhibits spectral features consistent with optical dressing of the QD transition. These cavity emission measurements are more akin to absorption measurements of a strongly driven QD rather than resonance fluorescence measurements. In addition to revealing new aspects of the off-resonant QD-cavity interaction, this result provides a new, simpler means of coherently probing QDs and opens the possibility of employing off-resonant cavities to optically interface QD nodes in quantum networks.

Figure 1

Experimental setup and numerical simulations. (a) The schematic shows the relative position of the QD and the cavity on a wavelength axis. For the particular QD-cavity system considered, the QD is red detuned from the cavity, though off-resonant coupling is observed for both red and blue detuned QDs. In experiments, a strong pump laser dresses the QD while a weak probe laser is scanned across the QD. QD emission is incoherently coupled to the cavity. The cavity emission is monitored as a function of probe laser wavelength ${\lambda }_p$. (b) The experimental setup is a confocal cross-polarization setup. The PBS (polarizing beam splitter) is used to perform cross-polarized reflectivity measurements, as in previous work (Ref.3). The powers are measured in front of the objective lens (OL). The output is dispersed in a single-grating monochromator and measured by a nitrogen-cooled CCD. We employ a linear three hole defect PC cavity (a scanning electron micrograph is shown in the inset). (c) Normalized off-resonant cavity emission obtained by numerical simulation is plotted as a function of ${\lambda }_p-{\lambda }_{QD}$, ${\lambda }_{QD}$ and ${\lambda }_p$ being the QD resonance and probe laser wavelengths, respectively, for different pump powers $P$ (normalized units), while the probe power is kept at 1. The pump laser is resonant with the QD. (d) For a pump power of $P=25$, the cavity emission is plotted as a function of ${\lambda }_p-{\lambda }_{QD}$ for different pump-QD detunings $\Delta {\lambda }_{\mathrm{pump}}$ (nm). In both (c) and (d), spectra are vertically offset for clarity.

Figure 2

Power dependence of the peaks and dips in the cavity emission spectra and effect of coherent dot-cavity interaction strength $g$: (a) The separation between the two peaks [as shown in Fig. 1c] as a function of the laser Rabi frequency. The slope of the linear fit is $\sim$4. (b) The separation between the two dips [as shown in Fig. 1c] as a function of the laser Rabi frequency. The slope of the linear fit is $\sim$2. (c) Cavity emission as a function of probe laser wavelength, for different dot-cavity coupling $g$.

Figure 3

Characterization of the QD-cavity system in photoluminescence (PL) and probing of the off-resonant dot-cavity coupling. (a) PL spectrum of the system. From the Lorentzian fit to the cavity we estimate a cavity linewidth $\Delta {\lambda }_{\mathrm{cav}}=0.1$ nm. (b) The laser is scanned across the QD-cavity system. Emission from the cavity is observed when the laser is resonant with the QD. Similarly, emission from the QD is observed when the laser is resonant with the cavity. (c), (d) The QD (cavity) linewidth is measured by monitoring the cavity (QD) emission as a function of the probe wavelength ${\lambda }_p$.

Figure 4

Coherent interaction between the QD and the laser observed through cavity emission. (a) Normalized cavity emission as a function of the probe laser wavelength for different pump powers (measured before the objective lens). We observe that a single QD resonance splits into two peaks. The splitting is linearly proportional to the Rabi frequency of the pump laser. Each peak is fitted with a Lorentzian. (b) Rabi frequency $\Omega$ of the laser (estimated from the splitting) as a function of the square root of the pump power $P$. A linear relation exists between $\Omega$ and $\sqrt{P}$. (c) Normalized cavity emission for a pump power of 190 nW for two different QD-cavity detunings $\Delta \lambda =0.22$ and $0.4$ nm. (d) Cavity emission for a pump power of 290 nW at two different QD-cavity detunings $\Delta \lambda =0.26$ and $0.4$ nm. We observe that the splitting increases for smaller detuning (i.e., when the pump laser is closer to the cavity), which suggests that the input laser power is enhanced by the presence of the cavity. For all experiments the probe laser power is kept constant at 20 nW. The QD-cavity detuning is defined as $\Delta \lambda ={\lambda }_{QD}-{\lambda }_{\mathrm{cav}}$. In (a), (c), and (d) the spectra are vertically offset for clarity.

Figure 5

Dependence of the result on pump-QD detuning. Off-resonant cavity emission as a function of the probe laser wavelength for different pump-QD detunings $\Delta {\lambda }_{\mathrm{pump}}={\lambda }_{\mathrm{pump}}-{\lambda }_{QD}$. We observe that the QD linewidth broadens when the pump is present and detuned from the QD resonance. As the pump is tuned through the QD resonance, we observe the emergence of two peaks in the cavity emission spectrum. This two-peak spectrum is consistent with the observation of the anticrossing of Rabi sidebands. The pump and probe power are kept at 290 nW and 20 nW, respectively. The spectra are offset for clarity.