Three-photon Rydberg-atom-based radio-frequency sensing scheme with narrow linewidth

We demonstrate Rydberg atom-based radio-frequency sensing with a colinear three-photon scheme in a room-temperature cesium vapor cell that minimizes residual Doppler broadening of the probe-laser absorption feature. A sub-200-kHz spectral linewidth is observed and extends the self-calibrated Autler-Townes sensing regime to weaker fields by a factor of 18 compared to the theoretical limit of the most commonly used two-photon scheme. The sensitivity of the method to microsecond pulses is shown to be sufficient to detect radio-frequency Rabi frequencies of 2π × 0.44 MHz using a 480-kHz bandwidth at 108.9 GHz, demonstrating the ability to sense time-dependent signals suitable for radar and communications.

Figure 1

(a) Energy-level diagram and experimental setup. An rf E field is emitted by a horn antenna perpendicular to the optical axis, and is sensed by monitoring the probe laser’s transmission through the vapor cell. (b) The absorption feature produced by the three-photon excitation is spectrally narrower by over an order of magnitude relative to a standard two-photon 509- and 852-nm excitation scheme. Rabi frequencies of Ω₈₉₅= 2π × 0.6 MHz, Ω₆₃₆= 2π × 4.9 MHz, and Ω₂₂₆₂= 2π × 0.17 MHz with 1.2, 1.3, and 1.7 mm beam diameters are used for the three-photon excitation, while Ω₈₅₂= 2π × 1.7 MHz and Ω₅₀₉= 2π × 2.1 MHz with 300-µm beam diameters in a 3-cm-long cell are used for the two-photon excitation.

Figure 2

(a) Extracted Lorentzian linewidth (FWHM) of the three-photon signal with no rf E field present, measured at various intermediate coupling Rabi frequencies Ω₆₃₆. Error bars represent one standard deviation in the distribution of linewidths from individual traces. The inset shows averaged experimental spectral line shapes. (b) Probe absorption of the three-photon system simulated using a density-matrix model (dot-dashed line), producing a theoretical Lorentzian linewidth of 190 kHz that is comparable to experiment (solid line). Ω₈₉₅= 2π × 1.0 MHz and Ω₂₂₆₂= 2π × 0.04 MHz.

Figure 3

Autler-Townes splitting of the probe-laser absorption in the vapor cell in response to a 108.9-GHz continuous-wave rf E field whose amplitude is increased from top to bottom. The top trace is measured without an rf E field applied, and produces a FWHM of $2\pi \times 284$ kHz. The self-calibrated rf field strength measured from the atoms via the peak splitting is labeled next to each trace. Ω₈₉₅= 2π × 1.0 MHz, Ω₆₃₆= 2π × 7.2 MHz, and Ω₂₂₆₂= 2π × 0.04 MHz.

Figure 4

(a) Matched filter output in response to single 5365-µV/cm rf pulses. The inset shows an averaged atomic response to a 3724-µV/cm rf pulse, used as a filter template. (b) Average signal height output by the matched filter in response to individual pulses at different rf E-field strengths and pulse durations. The error bars represent one standard deviation from 200–400 pulses. The inset shows a fit (red) to the quadratic amplitude regime at weak rf E fields. Sensitivity is determined by the minimum field at which the average signal height exceeds a 1σ noise error bar (black dashed line in inset) at the measurement bandwidth.