Free induction decay (FID) in atomic magnetometers measures the Larmor precession frequency of atoms, but the useful signal depends on the sensor's orientation relative to the field. At certain angles the signal drops to nearly zero, which is the dead zone. Neighboring works in our archive tackle related problems: one removes the dead zone by modulating the polarization, another uses multiple sensors, and a third folds the beam with a mirror inside the cell. This work stands apart: it is the first to quantitatively test a model with cross-talk in a three-mirror cell and pushes the sensitivity to a level suitable for geomagnetic measurements.
Three cylindrical mirrors forming a Herriott cavity were built into a cell containing rubidium-87 vapor and nitrogen buffer gas. The middle mirror rotates the beam by 90°, creating two orthogonal pump regions. Atoms are pumped by an amplitude-modulated laser on the D1 line; then the pump is turned off, and a probe beam records the decaying precession. The signal was analyzed using the Hilbert transform to extract the instantaneous frequency. Dead zones were scanned by rotating the field while keeping the sensor fixed and maintaining resonance via a closed loop. A total of 266 orientations were measured.
The amplitude of the AC signal varies by no more than a factor of three across all orientations, confirming the elimination of the dead zone. A simple model with independent regions does not describe the data: it predicts a dependence only on the polar angle, whereas the experiment shows an azimuthal dependence as well. Introducing a cross-talk parameter η = 0.28 (the fraction of atoms transferring from one region to the other) yields agreement with experiment. The field sensitivity was 58 fT/√Hz in the optimal orientation and remains better than 80 fT/√Hz over the full sphere. The orientation error in a 50 μT field does not exceed 0.7 nT, and the field tracking speed reaches 10,000 nT/s.
This work shows that cross-talk is not merely an artifact but a necessary element of the model for three-mirror cells. This distinguishes it from neighboring solutions where the regions are treated as independent. The quantitative agreement between theory and experiment for the orientation error means that the systematic error can be corrected to within 0.7 nT if the approximate field direction is known. This level is sufficient for geomagnetic surveys and navigation.
The sensor could become the basis for a vector magnetometer with a switchable bias field. Two tasks need to be solved for that: eliminating dead zones (already done) and reducing the orientation error (shown to be possible). Further improvement in tracking speed could come from shortening the pump time or broadening the resonance line.
Geomagnetic prospecting, GPS-free navigation, and biomagnetic imaging in unshielded environments.
Quantitatively describe the cross-talk parameter η in terms of cell geometry and atomic diffusion. Develop a vector mode with a switchable field and test real-time orientation error correction.
For atomic magnetometry, a key unresolved problem is simultaneously achieving high sensitivity, low orientation error, and no dead zones in a compact sensor. This work eliminates dead zones and provides a model for error correction, but the ultimate accuracy of vector measurements in motion remains an open question.
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
The paper describes a scalar atomic magnetometer with a three-mirror cell, where orthogonal pump beams eliminate dead zones. The authors showed that describing the signal requires accounting for cross-talk between regions, and they achieved a sensitivity better than 80 fT/√Hz at any orientation.
Where its strength is
A quantitative model with cross-coupling η=0.28 explains the azimuthal dependence of the signal in a three-mirror cell for the first time, which makes it possible to correct the orientation systematic error down to 0.7 nT.
Where it could go next
Worth trying
Why this may matter
Eliminating dead zones while simultaneously accounting for cross-talk opens the way to vector FID magnetometers with accuracy sufficient for geomagnetic surveys and navigation without complex orientation systems.
Nearby works in our archive
The archive is sparse around this work: the closest review is only 0.53, against a usual 0.64. It looks like few have drilled nearby.
The diffusion length of rubidium atoms during the probing time is only 0.6 mm, so cross-talk arises not from long-range flights but from local mixing near the middle mirror.