Large-Area Atom Interferometry with Frequency-Swept Raman Adiabatic Passage

We demonstrate light-pulse atom interferometry with large-momentum-transfer atom optics based on stimulated Raman transitions and frequency-swept adiabatic rapid passage. Our atom optics have produced momentum splittings of up to 30 photon recoil momenta in an acceleration-sensitive interferometer for laser cooled atoms. We experimentally verify the enhancement of phase shift per unit acceleration and characterize interferometer contrast loss. By forgoing evaporative cooling and velocity selection, this method lowers the atom shot-noise-limited measurement uncertainty and enables large-area atom interferometry at higher data rates.

Figure 1

(a) Diagram of pulse timings and wave-packet trajectories for Mach-Zehnder interferometers with $2\hslash k$, $6\hslash k$, and $10\hslash k$ atom optics. Augmentation pulses (A) are either Raman $\pi$ or ARP pulses. (b) Stimulated Raman transition coupling the ground states in a $\mathrm{\Lambda }$ system. (c) Bloch sphere depiction of frequency-swept adiabatic rapid passage, with poles corresponding to alkali-metal clock states.

Figure 2

Phase shift per unit acceleration $\mathrm{\Delta }\varphi /a$ for various LMT orders. Interferograms were acquired by perturbing the chirp rate of the Raman difference frequency. The resulting Doppler shift mimics a relative acceleration $\delta a$ between the atoms and Raman beams. Points represent 10- or 15-shot averages, error bars indicate standard error, and lines are fitted sine waves. The table shows reasonable agreement between measured and predicted values ($26\hslash k$ was not measured).

Figure 3

(a) Histogram of transition probabilities from an $18\hslash k$ interferometer with fitted probability density function (red curve). (b) Contrast as a function of LMT order. Lines are Monte Carlo predictions, points are contrast fits to histograms [see (a)], and fit uncertainties are smaller than symbol sizes. For tan/tanh interferometers, spontaneous emission-limited contrast corresponds to the bottom of the grey region. (c) Spatial intensity profile of the central portion of one Raman beam.

Figure 4

(a) Contrast as a function of dwell time. Estimates for the $6\hslash k$ interferometer were based on a contrast measurement at $T=1\mathrm{ms}$ and the fairly uniform trend observed with all other LMT orders. (b) Contrast as a function of ARP duration. Uncertainties are smaller than the symbol sizes and lines are guides for the eye.