Acoustic Black Hole in a Stationary Hydrodynamic Flow of Microcavity Polaritons

We report an experimental study of superfluid hydrodynamic effects in a one-dimensional polariton fluid flowing along a laterally patterned semiconductor microcavity and hitting a micron-sized engineered defect. At high excitation power, superfluid propagation effects are observed in the polariton dynamics; in particular, a sharp acoustic horizon is formed at the defect position, separating regions of sub- and supersonic flow. Our experimental findings are quantitatively reproduced by theoretical calculations based on a generalized Gross-Pitaevskii equation. Promising perspectives to observe Hawking radiation via photon correlation measurements are illustrated.

Figure 1

(a) Sketch of the experimental configuration. (b) Laser excitation conditions in the wave vector–energy plane. (c),(f) Linear scale image of the spatially resolved PL emission for excitation powers $p=7\mathrm{mW}$ and $p=100\mathrm{mW}$, respectively. (d),(g) Semilogarithmic plot of the PL emission integrated over the transverse direction. Inset of (g),(h): zoom of the PL emission profile and confinement potential in the vicinity of the defect. (e),(h) Generalized Gross-Pitaevskii numerical simulation of the quantities in (d),(g).

Figure 2

Wave-vector-resolved PL intensity for different excitation powers. (a)–(c) Upstream region, spatially filtered from $y=-25\mu \mathrm{m}$ to $y=-5\mu \mathrm{m}$. (d)–(f) Downstream region, filtered from $y=5\mu \mathrm{m}$ to $y=50\mu \mathrm{m}$. Blue dots: experimental points; red lines: Gaussian fits.

Figure 3

(a) Interaction energy in the upstream (blue circles) and downstream (orange squares) regions as a function of excitation power. (b),(c) Flow speed and speed of sound in the (b) upstream and (c) downstream regions. (d)–(f) Results of a generalized Gross-Pitaevskii numerical simulation.

Figure 4

Normalized spatial correlation function of photon density fluctuations, $g^{(2)}(y,y^{\prime })-1$, as predicted by an out-of-equilibrium Wigner Monte Carlo simulation [24] using the experimental parameters. A Gaussian real-space smoothing with ${\ell }_{\text{av}}=1\mu \mathrm{m}$ has been used.