Trapping and Observing Single Atoms in a Blue-Detuned Intracavity Dipole Trap

A single atom strongly coupled to a cavity mode is stored by three-dimensional confinement in blue-detuned cavity modes of different longitudinal and transverse order. The vanishing light intensity at the trap center reduces the light shift of all atomic energy levels. This is exploited to detect a single atom by means of a dispersive measurement with 95% confidence in $10\mu \mathrm{s}$, limited by the photon-detection efficiency. As the atom switches resonant cavity transmission into cavity reflection, the atom can be detected while scattering about one photon.

Figure 1

The blue intracavity dipole trap from the perspective of an entering atom. The slow atom is restricted to the field minima of the blue light fields that coincide with the antinodes of the near-resonant probe mode no. 1 at the cavity center. Persistent axial confinement is provided by a ${\mathrm{TEM}}_{00}$ mode no. 2, “pancakes.” Combined with the transverse nodal line of a ${\mathrm{TEM}}_{10}$ mode no. 3, funnels are formed to guide the atom to a strong-coupling region. Full three-dimensional confinement is achieved by adding a ${\mathrm{TEM}}_{01}$ mode to complete a transverse “doughnut” mode no. 4.

Figure 2

Experimental setup. Slow ${}^{85}\mathrm{Rb}$ atoms are injected from below into a high-finesse cavity. The cavity is excited by a weak near-resonant probe field at 780.2 nm and strong blue-detuned dipole fields. Behind the cavity, a grating directs the probe light onto single-photon counting modules (SPCM) whereas all dipole beams are detected by photomultiplier modules (PMM). The cavity length $L=0.122\mathrm{mm}$ is independently stabilized by a weak laser at 785.2 nm. Midway between the mirrors, the trap center coincides with an antinode of the probe field. The mode profiles in the dash-dotted and the dashed central plane are discussed in Fig. 1. The waist of the 780.2 nm fundamental transverse mode is $w_0=29\mu \mathrm{m}$, and $F=4.4\times {10}^5$ is the cavity finesse. The maximum atom-cavity coupling constant and the decay rates of the atomic polarization and the cavity field are $(g_0,\gamma ,\kappa )/2\pi =(16,3,1.4)\mathrm{MHz}$, respectively.

Figure 3

A blue-trapping event. (a) Experimental intensity patterns of the different modes (cp. Figure 1). (b) Sample trace: Transmitted probe power in units of intracavity photon number, $⟨a^{†}a⟩$, and the maximum radial trap height. Upon detection of an atom (A), the probe intensity is decreased, and the trap is closed. When the atom leaves (B), the empty cavity transmission is observed. Experimentally, the doughnut mode no. 4 is a controlled superposition of two nondegenerate eigenmodes defined by our cavity. The trace is taken for detunings of continuous axial cavity cooling $({\Delta }_{\mathrm{ac}},{\Delta }_c)/(2\pi )=(-35,0)\mathrm{MHz}$, where the average storage time for single trapped atoms is about 30 ms.

Figure 4

Normal-mode spectrum: transmission as a function of detuning for a well-coupled system (squares and crosses). The bare atom (A) is detuned from the bare cavity resonance (C) by ${\Delta }_{\mathrm{ac}}/2\pi =-35\mathrm{MHz}$. A transmission of 1 pW corresponds to 1.2 intracavity photons. Intervals contribute to the spectrum if the transmission in the neighboring cooling intervals is $<10\%$ of the bare cavity value $⟨n_0⟩$. An analytical fit (solid line) for a fixed coupling $g$ at low excitation results in $g=0.83(12)\times g_0$ and a residual Stark shift of ${\Delta }_s/2\pi =0.7(1.3)\mathrm{MHz}$. The empty cavity transmission (pluses, dashed line) is shown for reference.