Analysis of a single-atom dipole trap

We describe a simple experimental technique which allows us to store a single ${}^{87}\mathrm{Rb}$ atom in an optical dipole trap. Due to light-induced two-body collisions during the loading stage of the trap the maximum number of captured atoms is locked to one. This collisional blockade effect is confirmed by the observation of photon antibunching in the detected fluorescence light. The spectral properties of single photons emitted by the atom were studied with a narrow-band scanning cavity. We find that the atomic fluorescence spectrum is dominated by the spectral width of the exciting laser light field. In addition we observe a spectral broadening of the atomic fluorescence light due to the Doppler effect. This allows us to determine the mean kinetic energy of the trapped atom corresponding to a temperature of $105\mu \mathrm{K}$. This simple single-atom trap is the key element for the generation of atom-photon entanglement required for future applications in quantum communication and a first loophole-free test of Bell’s inequality.

Figure 1

(Color online) Experimental setup of the dipole trap and fluorescence detection. The dipole trap laser is focused at the intersection of three pairs of counterpropagating laser beams for optical cooling. Fluorescence light is collected with a microscope objective, coupled into a single mode optical fiber, and detected with a Si APD.

Figure 2

Single atom detection. (a) Number of photons counted by an avalanche photodiode per $100\mathrm{ms}$. (b) Histogram of the photon-counting data. Due to a collisional blockade effect [5, 21] only counts corresponding to zero or one atom are observed.

Figure 3

(Color online) (a) Hanbury-Brown-Twiss setup for the measurement of the photon pair correlation function $g^{\left(2\right)}\left(\tau \right)$. The fluorescence light is sent through a beamsplitter (BS) onto two single photon detectors D1, D2 to record detection time differences $\tau =t_1-t_2$. (b) On long time scales, $g^{\left(2\right)}\left(\tau \right)$ shows a small bunching effect for $\mid \tau \mid ⩽3\mu \mathrm{s}$. (c) On short time scales, clear photon antibunching at $\tau =0$ and oscillations due to Rabi flopping are observed. The dashed line corresponds to accidental coincidences caused by the dark count rate of the detectors. Experimental parameters: $I_{\mathrm{CL}}=103\mathrm{mW}∕{\mathrm{cm}}^2$, $I_{\mathrm{RL}}=12\mathrm{mW}∕{\mathrm{cm}}^2$, ${\Delta }_{\mathrm{CL}}∕2\pi =-31\mathrm{MHz}$, dipole trap depth $U=0.38\mathrm{mK}$.

Figure 4

(Color online) Intensity correlation function $g^{\left(2\right)}\left(\tau \right)$ (background corrected) of the resonance fluorescence from a single ${}^{87}\mathrm{Rb}$ atom in the dipole trap for two different trap depths and partial level scheme of ${}^{87}\mathrm{Rb}$ (inset). Bold line: numerical calculation. Thin line: measured correlation function. Experimental parameters: $I_{\mathrm{CL}}=103\mathrm{mW}∕{\mathrm{cm}}^2$, $I_{\mathrm{RL}}=12\mathrm{mW}∕{\mathrm{cm}}^2$, ${\Delta }_{\mathrm{CL}}∕2\pi =-31\mathrm{MHz}$, dipole trap depth (a) $U=0.38\mathrm{mK}$, (b) $U=0.81\mathrm{mK}$.

Figure 5

(Color online) Setup for the measurement of the resonance fluorescence spectrum of light scattered by a single ${}^{87}\mathrm{Rb}$ atom. Both the atomic fluorescence and the laser light are analyzed alternately with the same scanning FPI. The spectra exhibit a width of $0.90\pm 0.02\mathrm{MHz}$ and $1.00\pm 0.02\mathrm{MHz}$ (FWHM) for the excitation (-엯-) and the fluorescence light (-●-), respectively. Experimental parameters: $I_{\mathrm{CL}}=87\mathrm{mW}∕{\mathrm{cm}}^2$, $I_{\mathrm{RL}}=12\mathrm{mW}∕{\mathrm{cm}}^2$, ${\Delta }_{\mathrm{CL}}∕2\pi =-19\mathrm{MHz}$, dipole trap depth $U=(0.62\pm 0.06)\mathrm{mK}$.