We construct a falsifiable, parameter-auditable framework for detecting ultralight vector darkmatter or generic new long-range forces through oscillatory, composition-dependent accelerationsmeasured by a co-located dual-species light-pulse atom interferometer. The observable is the differ-ential interferometric phase ∆ΦAB (t) between species A and B, which rejects common-mode inertialnoise while retaining equivalence-principle–violating accelerations. For a sinusoidal differential ac-celeration ∆a(t) = ∆a0 cos(ωt + φ), we derive the exact three-pulse Mach–Zehnder transfer function|∆ΦAB (ω)| = keff |∆a0| (4 sin2(ωT /2)/ω2), where T is the pulse separation time and keff the effectiveoptical wavevector. We specialize to an ambitious but well-defined operating point motivated by theoperating parameters: apparatus scale ∼ 10 m and interrogation time T = 6 s, which sets the intrin-sic low-frequency bandwidth and enables sensitivity to ultralight masses m ≃ hf in the sub-Hz band.Because virialized ultralight fields are coherent over ∼ 106 oscillation periods, the signal appears asa narrow spectral line with fractional width ∆f /f ∼ 10−6. We formulate an explicit semi-coherentmatched-filter statistic appropriate for a random-phase narrowband line and derive an analyticalmapping from a null result (“no line above X”, where X is specified as an SNR threshold or dB)to a constraint on the differential acceleration amplitude ∆a0, and hence to coupling parametersof a generic vector mediator or dark photon scenario. We provide an end-to-end experimental pro-tocol, including calibration, systematics vetoes, and coherence-time segmentation, yielding a cleandiscovery signature: a reproducible line with coherence-limited width whose amplitude scales withcomposition-dependent “charge-to-mass” contrast.
SIKX HILTON (Tue,) studied this question.