arXiv:2610.27731
Axion detection in the non-Markovian noise
Jing Dong, Shuang-Quan Ma, Rong-Hang Chen, Bin-Kai Luo, Tao Liu, Qing Ai, Jun-Hong An
Abstract
The detection of axion dark matter poses a significant challenge due to its exceptionally weak coupling to standard-model particles. Here we develop a noise-tolerant quantum-metrology scheme based on coupled $^{129}$Xe spin ensembles to search for such weakly interacting particles. Within an experimentally accessible parameter regime, our scheme achieves an ideal exclusion limit on the axion coupling constant that is several orders of magnitude more stringent than that of conventional detection methods. We further investigate the performance of our scheme under dissipative noise and find that its robustness is fundamentally governed by the energy-spectrum structure of the composite spin-environment system. Remarkably, when a bound state emerges in the spectrum, the exclusion limit asymptotically recovers its ideal value in the non-Markovian dynamics, whereas under the Markovian approximation the limit deteriorates significantly due to noise. Our results establish a universal mechanism for noise-resilient exploration of beyond-standard-model physics, which can be practically implemented by engineering bound-state formation via tuning the Larmor frequency of the spin ensembles.
AI Summary
The paper asks whether non-Markovian dissipation can preserve sensitivity to axion-mediated neutron–neutron interactions in coupled $^{129}$Xe ensembles. Precessing spins in a source cell generate an oscillating pseudomagnetic field in a sensor cell; its amplitude encodes the pseudoscalar coupling. Quantum Fisher information (QFI) quantifies sensitivity, with transverse-polarization measurements providing an attainable readout at suitable interrogation times. For independent sensor spins without noise, the field uncertainty is $\delta B_{\rm ac}=2/(\gamma_n t\sqrt{MN_{\rm se}})$. The noise model couples each sensor spin to an initially vacuum bosonic environment with an Ohmic-family spectral density. Laplace-transform analysis relates persistent coherence to an isolated spin–environment bound state. Its existence requires $\omega_A<\eta\Gamma(s)\omega_c/4$. Without a bound state, coherence vanishes asymptotically; with one, the long-time QFI grows as $\gamma_n^2t^2Z^4/4$, retaining the ideal time scaling with an uncertainty penalty $1/Z^2$. Numerical solutions confirm the bound-state threshold and sensitivity scaling. Using $N_{\rm se}=2.1\times10^{17}$ and $M=4.32\times10^4$, the projected coupling bound is reported to improve on previous measurements by approximately $10^6$ near $m_a=3.5\,\mu$eV. The treatment addresses dissipative noise, assuming dephasing can be suppressed by dynamical decoupling. Numerical examples use an Ohmic spectrum; applying the proposal requires characterizing the actual environmental spectrum and accessing a bound-state regime, proposed through Larmor-frequency tuning. The quoted exclusion limits are theoretical projections, not experimental constraints.
Central claim may not hold: The asymptotic sensitivity retains a relative penalty of 1/Z^2, approximately 1.204 for the stated Ohmic example, rather than recovering the ideal value exactly. The noise-protection result is established only for an initially vacuum bath with an excitation-conserving interaction, not for general dissipative environments relevant to implementation. Missing absolute Larmor frequency, source population, and source–sensor separation prevent reproduction of the numerical exclusion limits and claimed million-fold improvement. No apparatus-specific environmental spectrum or physical coupling and cutoff values establish that Larmor-frequency tuning can reach the required bound-state regime in the proposed Xe apparatus. The displayed rotating-wave and fixed-unitary transformations do not produce the asserted stationary spin–bath Hamiltonian, leaving the physical basis of the proposed Larmor-tuned protection mechanism unestablished.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.