An arXiv preprint models a neutron star spiraling into a stellar-mass black hole surrounded by an ultracompact minihalo, a tightly packed cloud of QCD-axion dark matter. Under the paper's benchmark assumptions, the encounter could leave two linked traces: a gravitational-wave waveform whose phase separates from the vacuum prediction, and a narrow-band radio signal from axions converting to photons in the neutron star's magnetosphere.
How the halo is built
The halo is reconstructed numerically with a secondary-infall calculation, which follows how material builds up around an initial black-hole seed. The setup combines relic-abundance and isocurvature conditions from a pre-inflationary Peccei-Quinn scenario with a conservative inner core of constant density. This reconstructed halo then supplies the density used in both the gravitational-wave and radio projections.
Allowing the density-profile slope to vary independently gives gamma = 2.244, the steep fall-off expected for the reconstructed outer halo. The authors also report that axion-isocurvature fluctuations and the initial departure of shells from Hubble expansion do not discernibly change the final profile in the model. It remains predominantly governed by the central black-hole seed.
At its center, the fiducial halo is both compact and dense. The estimated core radius is about 7 x 10^-9 to 1 x 10^-8 parsecs, while the core density is about 2 x 10^-4 to 6 x 10^-4 grams per cubic centimeter for black-hole masses from 1 to 10 solar masses. In the paper, these estimates provide the inner-halo input for the two signal calculations.
A waveform that does not quite match
The gravitational-wave calculation compares the binary in vacuum with a version that carries two effects from the minihalo: dynamical friction, a drag-like influence of surrounding matter, and the extra mass enclosed along the orbit. The authors treat the mass contribution as negligible beside dynamical friction, making the friction prescription the main source of the modeled waveform change.
To set a line between similar and distinguishable waveforms, the paper uses matched-filter overlap, a score for how closely a modeled signal matches a template. Its threshold is R = 0.5, corresponding to an accumulated phase difference of |Delta phi| = pi/2. That threshold is the paper's stated benchmark for assessing whether the modeled dephasing is significant.
For black-hole seed masses of 1 to 10 solar masses, significant modeled dephasing extends across the LISA frequency range from 0.1 mHz to 1 Hz. In practical terms, the inspiral with the minihalo included can move far enough away from a vacuum waveform to matter across that band, provided the core, friction and detector-noise assumptions hold.
A second signal in radio
The radio calculation follows a different route. It estimates the axion density at the radius where conversion is modeled, using the neutron star's gravitational focusing and a phase-space mapping that carries the inner-core distribution to that radius. The projections use representative neutron-star benchmarks: three nearby members of the Magnificent Seven and a source at extragalactic distance motivated by an accreting X-ray pulsar in M31.
At a 10-hour observation time and a 1-solar-mass black-hole benchmark, the projected GBT sensitivity to the axion-photon coupling ranges from 7 x 10^-21 to 6 x 10^-20 GeV^-1 for RX J1856.5-3754. For 3XMM J004232.1+411314, the range is 2 x 10^-16 to 1 x 10^-15 GeV^-1. The sensitivity estimate uses the radiometer equation and sets its signal-to-noise threshold at 5.
Nearby benchmark targets are projected to reach at least two orders of magnitude below the QCD-axion band, while the more distant benchmark reaches into that band under the model assumptions. These are sensitivity projections, not radio detections.
Two messengers, one test
The paper's central idea is that the two messengers would answer different questions. In the authors' interpretation, gravitational waves probe the surrounding dark-matter structure, while the radio counterpart probes the axion component and its coupling to photons. Combining them could therefore test the compact object, its halo and the particle component together, creating a model-based check of a mixed dark sector.
What remains uncertain
The result rests on several assumptions. The innermost density profile is unresolved in the analysis, and the constant-density core is a phenomenological fiducial; a more concentrated profile could strengthen the projected signals. The halo calculation treats baryons and axion dark matter as one effective collisionless component and leaves out delayed baryonic infall. The neutron-star magnetosphere uses an idealized Goldreich-Julian prescription, while the M31 benchmark's magnetic field and plasma profile are also idealized inputs.
The analysis also uses a simplified radial-patch treatment of gravitational focusing and quasi-static radio snapshots. The survival of the minihalos and the rate at which they encounter neutron stars still need better modeling. The remaining questions include more realistic angular-momentum-supported inner halos, axion phase-space distributions, neutron-star magnetospheres, galactic environments and a dedicated multimessenger search strategy.
The document is an arXiv version-1 preprint dated 26 August 2026. Its reported results are projections for representative systems and future or benchmark instruments, making the work a feasibility and sensitivity study for multimessenger searches.
Paper data and sources
Original title: Multi-messenger signal from QCD-Axion Ultracompact Minihalos
Authors: Dorian W. P. Amaral, Enrico D. Schiappacasse, Karla Tapia-Rebolledo
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text