Sawtooth wave adiabatic passage in a grating magneto-optical trap

We demonstrate sawtooth wave adiabatic passage (SWAP) in a grating magneto-optical trap (MOT) operating on the $^1$S$_0$ $\rightarrow$ $^3$P$_1$ transition of neutral $^{88}$Sr. From numerical simulations of SWAP using our laser beam geometry, we find that SWAP provides greater cooling than triangle wave frequency modulation despite the complex polarization environment of a grating MOT. The simulation is confirmed by our experimental results, where we demonstrate a factor of two improvement in transfer efficiency between our $^1$S$_0$ $\rightarrow$ $^1$P$_1$ grating MOT and our $^1$S$_0$ $\rightarrow$ $^3$P$_1$ grating MOT. We trap up to $3\times10^6$ $^{88}$Sr atoms in the $^1$S$_0$ $\rightarrow$ $^3$P$_1$ grating MOT, at an average temperature of 4.9 $μ$K with a lifetime of approximately 0.7 s. Our results show that SWAP is effective in non-orthogonal laser beam geometries, allowing greater duty cycles or higher atom number in sensors based on narrow-line grating MOTs.

💡 Research Summary

The authors present a comprehensive study on implementing Sawtooth Wave Adiabatic Passage (SWAP) cooling in a grating magneto‑optical trap (MOT) operating on the narrow ¹S₀ → ³P₁ transition of neutral ⁸⁸Sr. The motivation stems from the need to simplify and miniaturize cold‑atom platforms that traditionally require two sequential MOT stages—a broad‑line (461 nm) MOT for capture and a narrow‑line (689 nm) MOT for sub‑Doppler cooling. Conventional six‑beam orthogonal MOTs are bulky, whereas compact geometries such as pyramidal, photonic‑integrated, Fresnel, and especially grating MOTs use non‑orthogonal tetrahedral beam configurations that reduce optical access and introduce mixed polarizations, limiting capture volume and cooling force.

To address these limitations, the paper investigates whether SWAP—originally demonstrated in 1‑D or six‑beam configurations—can enhance radiation pressure in the more complex polarization environment of a grating MOT. Using the PyLCP Python package, the authors construct a full three‑dimensional optical Bloch equation model that includes the σ⁺ input beam and the diffracted beams containing σ⁺, π, and σ⁻ components. By artificially suppressing spontaneous emission (Γ = 1 s⁻¹) they map all possible adiabatic transitions as a function of axial velocity v_z and magnetic field projection B_z, identifying eight distinct regimes with characteristic momentum transfers Δp_z that depend on the diffraction angle θ_d. Some regimes even exceed the canonical –2ħk per sweep expected for 1‑D SWAP, while reversing the sweep direction (anti‑SWAP) leads to heating.

Realistic simulations then re‑introduce the natural decay rate of the ³P₁ state (Γ≈2π × 7.5 kHz) and experimentally achievable Rabi frequencies (Ω = 2π × 2 MHz). Two frequency‑modulation waveforms are compared: a sawtooth sweep (t_s = 50 µs, detuning from –2 MHz to +2 MHz) and a triangular sweep (t_s = 25 µs, detuning from –4 MHz to 0 MHz). Across a range of axial and transverse velocities, as well as zero and finite Zeeman shifts (≈1 Mrad/s), the sawtooth waveform consistently yields larger velocity damping than the triangle, often surpassing the theoretical limit of ħk/m per sweep for a six‑beam MOT. This demonstrates that SWAP can provide a net cooling force even when the beams are non‑orthogonal and the polarizations are impure.

Experimentally, the authors upgrade a previously reported grating MOT apparatus. Permanent magnets are replaced by fast‑switching electromagnets, providing a 5.5 mT/cm gradient for the broad‑line MOT and up to 0.4 mT/cm for the narrow‑line stage, with independent control of a uniform bias field. A titanium mounting block eliminates eddy currents, improving vacuum to <10⁻⁷ Pa. The 689 nm narrow‑line input beam is combined with the 461 nm beam via a dichroic mirror, delivering a 1/e² radius of ~17 mm and a peak intensity of ~18 mW/cm² (≈6000 I_sat). The grating diffraction angle is measured at 27°.

Applying the SWAP protocol, the authors achieve a two‑fold increase in transfer efficiency from the broad‑line to the narrow‑line MOT. The final narrow‑line grating MOT captures up to 3 × 10⁶ ⁸⁸Sr atoms at an average temperature of 4.9 µK and a 1/e lifetime of approximately 0.7 s. These performance metrics surpass previous grating MOT results, which typically reported temperatures near 10 µK and lower atom numbers.

The work concludes that SWAP is robust against the challenges posed by tetrahedral beam geometries—reduced momentum projection, mixed polarizations, and spatially varying interference patterns. Consequently, SWAP enables higher duty cycles and larger atom numbers in compact, chip‑scale cold‑atom sensors that rely on narrow‑line transitions, such as portable optical clocks, gravimeters, and inertial navigation devices. Future directions include optimizing sweep parameters, exploring multi‑frequency SWAP sequences, and integrating larger‑area diffraction gratings to further increase capture volume while preserving the benefits of SWAP cooling.