The Signal
Researchers Build a Quantum Measurement Model for Axion Searches
Six researchers present a fully quantum model of continuous axion-dark-matter measurements in cavity haloscopes. Their calculations predict photon-count statistics that differ from the Poisson model often used for coherent signals, while fi
Quantum.world Team · Written with AI · fact-checked ·
With it, we compute the exact axion-induced number distribution for the first time, demonstrating that it is not Poisson.
Yunjia Bao, Dhong Yeon Cheong, Nicholas L. Rodd, Joey Takach, Lian-Tao Wang and Kevin Zhou, “Continuous Measurement of Quantum Axion Dark Matter”
In a paper submitted to arXiv on October 21, Yunjia Bao, Dhong Yeon Cheong, Nicholas L. Rodd, Joey Takach, Lian-Tao Wang and Kevin Zhou set out a quantum-mechanical account of how axion dark matter would register in a continuously monitored cavity haloscope. The work is a theoretical model, not a report of an axion detection or a new detector. Its central result is a way to calculate the full probabilities of measurement outcomes rather than infer the signal from averages alone.
A quantum account of the haloscope readout
A cavity haloscope looks for axions by coupling a resonant cavity to a transmission line. In the researchers’ model, the axion field, cavity and outgoing line all receive a quantum description, and the line’s output can be monitored through photon counting or quadrature measurements. The authors extend input-output theory to track how the axion’s state affects not just average signal power but also higher-order fluctuations.
They pair that frequency-domain method with a measurement-record formalism, which calculates the joint probability of outcomes collected over time. That approach is designed to handle time-dependent records and different choices of measurement. The two methods describe the same underlying system, but make different calculations more direct: input-output theory suits frequency-domain observables, while the measurement record gives access to time-domain statistics.
Why the expected photon counts differ
The difference emerges when the authors compare two models of the axion field. A coherent state gives Poisson photon-count statistics, while the paper argues that virialized axion dark matter is more likely to be described by a Gaussian quantum state. Under that Gaussian hypothesis, the count distribution changes with the length of the measurement bin: it is geometric for bins shorter than the output’s coherence time and approaches a negative binomial distribution for much longer bins.
The paper says a Poisson distribution does not accurately describe the full count distribution under the Gaussian hypothesis. The gap is most useful once a signal is resolved: when the expected count is far below one, the leading probabilities for zero or one photon are nearly the same across distributions. The authors estimate a conversion efficiency of about 10−22 for a benchmark set of axion-photon coupling, magnetic-field, cavity-quality and axion-mass values, illustrating how weakly the field couples to the cavity even as a large axion occupation can leave a measurable signal.
State comparisons demand much more data
The proposed statistics could help distinguish a virialized Gaussian field from a coherent-state model after a signal is found. For an illustrative case with a mean count of 10−2 per coherence time, the paper estimates that detecting a signal against no background would take roughly 100 coherence times, while separating the Gaussian prediction from the coherent, Poisson prediction would take about one million. In its example for the ADMX frequency band, those estimates correspond to less than a second for discovery and less than an hour for the state comparison.
Those timings assume no background noise, which the authors say would make real measurements take longer. They also depend on the stated model assumptions and a chosen statistical test; the paper discusses a more effective estimator that could reduce the illustrative state-discrimination time. The numbers describe calculations, not demonstrated performance from an operating experiment.
Nonclassical axion effects remain beyond reach
The distinction between Gaussian and coherent states is not itself evidence of nonclassical behavior: both can be represented by classical-like statistical mixtures. The researchers separately assess signatures such as Fock or squeezed axion states and conclude that the detector’s weak coupling suppresses those effects too strongly for practical observation. In their account, discovering a signal requires measuring nonzero mean power; establishing that the axion field is intrinsically nonclassical requires far more information.
The framework also has a defined scope. It focuses on cavity haloscopes with transmission-line readout, and the authors say it does not directly apply to broadband lumped-element detectors, where the resonant-cavity approximation breaks down. For the systems it does cover, its immediate contribution is a precise prediction: under the paper’s virialized Gaussian model, the full photon-count distribution is not Poisson.