Dynamic Stark Effect in Strongly Coupled Microcavity Exciton Polaritons

We present experimental observations of a nonresonant dynamic Stark shift in strongly coupled microcavity quantum well exciton polaritons—a system which provides a rich variety of solid-state collective phenomena. The Stark effect is demonstrated in a $\mathrm{GaAs}/\mathrm{AlGaAs}$ system at 10 K by femtosecond pump-probe measurements, with the blueshift approaching the meV scale for a pump fluence of $2\mathrm{mJ}{\mathrm{cm}}^{-2}$ and 50 meV red detuning, in good agreement with theory. The energy level structure of the strongly coupled polariton Rabi doublet remains unaffected by the blueshift. The demonstrated effect should allow generation of ultrafast density-independent potentials and imprinting well-defined phase profiles on polariton condensates, providing a powerful tool for manipulation of these condensates, similar to dipole potentials in cold-atom systems.

Figure 1

(a) Calculated polariton levels using full Hamiltonian diagonalization [Eq. (2)] with no pump (solid lines) and pump coupling energy of 15 meV (dashed lines) for the UP (dark blue) and the LP (light red). (b) Calculated dynamic Stark shift vs pump coupling energy using full Hamiltonian diagonalization (dark blue) and exciton Stark shift approximation (light red) for the UP (solid lines) and the LP (dashed lines). (c) Schematic diagram of the experimental setup. (d) Measured normalized reflectivity spectra at normal incidence for different cavity-exciton detuning. The curves are offset for convenience.

Figure 2

(a) Calculated normalized differential spectra $\Delta R/R$ for various pump-probe time delays $\Delta t$, for a pump fluence of $0.2\mathrm{mJ}{\mathrm{cm}}^{-2}$. Measured normalized differential reflection $\Delta R/R$ spectra for various pump-probe time delays for a pump fluence of (b) $0.2\mathrm{mJ}{\mathrm{cm}}^{-2}$, (c) $0.6\mathrm{mJ}{\mathrm{cm}}^{-2}$, (d) $1.2\mathrm{mJ}{\mathrm{cm}}^{-2}$, (e) $2\mathrm{mJ}{\mathrm{cm}}^{-2}$.

Figure 3

Dynamic Stark blueshift vs pump fluence. The calculated dependence is given by the solid lines: light red for LP and dark blue for UP, with pump coupling efficiency as a fitting parameter. Inset (a) normalized reflection spectra for the initial (solid light red) and the Stark-shifted (dashed dark blue) polariton levels at a constant delay. Inset (b) is a $2\mathrm{mJ}{\mathrm{cm}}^{-2}$ fluence differential reflection at a constant delay for measured (solid line light red) and calculated (dashed black line). Inset (c) shows the shift vs differential reflection amplitude $\Delta R/R$ as a conversion curve from $\Delta R/R$ (Fig. 2) to the energy shift.

Figure 4

(a) Measured normalized reflectivity spectra at normal incidence of the initial (solid) and the Stark-shifted (dashed) polaritons, for different cavity-exciton detuning and constant delay. The curves are offset for convenience. The inset is the Stark shift dependence on cavity-exciton detuning-measured points and calculated curves show the LP (light red) and UP (dark blue). (b) Measured normalized differential reflection $\Delta R/R$ spectra for a ps-scale Stark shift.