Cosmological Neutron Stars Produce Diffuse Axion X-Ray Signatures

TL;DR

The study probes diffuse axion-induced X-ray signatures from the cosmological neutron-star population in two channels: ultralight axions converting to photons in NS magnetospheres and heavy axions decaying to photons in the cosmic past. Using NSCool to model NS cooling and core temperatures, along with a magnetosphere framework and a cosmological NS formation history, the authors compare predicted fluxes to the cosmic X-ray background measured by NuSTAR, HEAO, Swift, and INTEGRAL, deriving 95% upper limits on the couplings $|g_{ann} g_{a\gamma\gamma}|$ for ultralight axions and $|g_{a\gamma\gamma}|$ for heavy axions across broad $m_a$ ranges. They find no evidence for axions and set robust constraints that reach new parameter space, effectively ruling out the axion explanation for the Magnificent Seven X-ray excess. The analysis also covers astrophysical uncertainties (superfluidity, EOS, formation rates, and NS demographics) and extends to a cosmological regular-star population, highlighting the importance of high-energy CXB measurements for constraining axion models and motivating future high-energy X-ray missions.

Abstract

Axion-like particles can be abundantly produced through scattering processes in the cores of neutron stars (NSs). If they are ultralight ($m_a \lesssim 10^{-4}$ eV), then they can efficiently convert to detectable photons in the external NS magnetospheres, and if they are heavy ($m_a \gtrsim 1$ eV), then they can decay into photons before reaching Earth. In this work, we search for the resulting X-ray signatures from both of these channels summing over the $\textit{cosmological}$ NS population. We compare the predicted axion-induced X-ray signal to the cosmic X-ray background today as measured by a number of instruments such as NuSTAR, HEAO, Swift, and INTEGRAL. We model the axion-induced signal using NS cooling simulations and magnetic field evolution models. We find no evidence for axions and derive strong constraints for both ultralight and heavy axion scenarios, covering new parameter space for the axion-photon and axion-nucleon couplings. Our results rule out the axion-explanation of the Magnificent Seven X-ray excess from nearby isolated NSs.

Figures (22)

• Figure 1: Axions are produced from the cores of NSs in our cosmological NS population. If the axions are ultralight, they can convert to photons in the magnetospheres of the host NS (upper process), and if the axions are heavy, they can spontaneously decay to photons at some later cosmic epoch (lower process). We search for both of these channels in cosmic X-ray background data, summing over the collective contributions from the cosmological NS population. The NS temperature distribution is illustrated for an example $t \sim 10^5$ yr NS.
• Figure 2: An illustration of how our axion-induced signal compares to the measured NuSTAR CXB data, which extends up to $\sim$20 keV. (Top) We show our best-fit axion and CXB background model (gray), as well as an example axion signal with the indicated coupling, all in the low $m_a$ limit. The residuals of the data compared to the fitted background model under the null hypothesis are shown directly below. (Bottom) The same but for a heavy axion with the indicated mass, and indicated $E/N$. We see that in both cases the axion spectra is morphologically different than the CXB, and that the data are largely consistent with the null hypothesis.
• Figure 3: The 95% upper limits on $|g_{ann}\times g_{a \gamma \gamma}|$ obtained in this work from the non-observation of ultralight axions from cosmological NS populations. Our fiducial result (black, solid), uses observations of NuSTAR CXB data, while we also show constraints using CXB measurements from the HEAO instrument, Swift, and INTEGRAL. We illustrate projections for constraints on $g_{ann}\times g_{a \gamma \gamma}$ from future telescopes which have smaller systematic uncertainties and reach to higher energies than those probed with NuSTAR (see text for details). Existing constraints are in gray, noting stringent constraints from SN 1987A Manzari:2024jns in particular, and the parameter space favored to explain the M7 X-ray excess is in red Buschmann:2019pfpDessert:2019dos, which is completely excluded by this work.
• Figure 4: The 95% upper limits on $g_{a \gamma \gamma}$ obtained in this work from the non-observation of heavy axions from cosmological NS populations, assuming a GUT anomaly coefficient ratio as indicated (see main text), and for the observed NuSTAR (fiducial, black, solid), HEAO, Swift, and INTEGRAL-measured CXBs. Our fiducial result can also be compared to the subdominant constraints obtained from $g_{a \gamma \gamma}$-only processes occurring in the cosmological regular star (RS) population (see text). We compare these constraints to existing constraints in gray, valid for our fiducial choice of $E/N = 8/3$.
• Figure S1: The instantaneous massless axion emissivity from a 1.4 $M_{\odot}$ NS with the BSk24 EOS at an age of $10^4$ years with the indicated axion-nucleon couplings. Results are shown for nucleon bremsstrahlung and PBF processes under superfluid model SFB-SCGF-BS07. In this model, the neutron $^1S_0$ pairing is computed using the model of Ref. Schwenk:2002fq, the proton $^1S_0$ pairing is computed using the model of Ref. Baldo:2007jx, and the neutron $^3P_2$ pairing is computed using the model of Ref. Ding:2016oxp.