AGN spectral variability across activity states and searches for axion-like particles

TL;DR

The study tackles ALP-induced spectral modulations in gamma-ray spectra from AGNs seen through galaxy clusters and the risk that intrinsic AGN variability can bias such searches. It builds on a stacking approach that yields a predictable survival probability $P_{\gamma\gamma}(E)$, and demonstrates a robust anti-correlation between AGN flux and spectral hardness that can mimic ALP signatures. It shows that this variability can produce false detections but not false exclusions, and proposes flux-state binning to mitigate the effect, thereby sharpening ALP bounds. The results improve the reliability of ALP searches in astrophysical spectra and provide practical guidelines for handling variability in stacking analyses, with implications for constraining the ALP parameter space in the nano-eV mass range.

Abstract

Axion-like particles (ALPs) are compelling candidates for dark matter and potential portals to new physics beyond the Standard Model. Photons traversing magnetized regions can convert into ALPs, producing characteristic, energy-dependent absorption features in astrophysical spectra. The probability of such conversions depends sensitively on both the photon energy and the properties of the intervening magnetic fields. Most existing searches have focused on individual astrophysical sources, but uncertainties in the structure and strength of cosmic magnetic fields have limited their reach. Recently, we have demonstrated that active galactic nuclei (AGNs) observed through galaxy clusters provide especially promising targets for ALP searches. By stacking multiple AGN-cluster sightlines, one can average over poorly known magnetic field configurations in galaxy clusters and recover a distinctive ALP-induced spectral suppression, thereby significantly enhancing sensitivity. In this work, we investigate a possible systematic uncertainty in such analyses: the intrinsic time-variability of AGN spectra. We demonstrate that AGN flux variability is correlated with spectral hardness, and that time-averaging over flaring and quiescent states can potentially mimic the suppression features imprinted by ALP-photon mixing. Our findings imply that the recent constraints remain conservative, and that incorporating detailed spectral variability into stacking analyses can further sharpen the search for axion-like particles.

Figures (4)

• Figure 1: Photon survival probability and its averages.Left: Photon transparency (survival probability $P_{\gamma\gamma}$) for photons traversing a galaxy cluster, shown for different realisations of the cluster magnetic field with identical global characteristics (blue lines). The turbulent magnetic field is modelled as a sequence of cells with constant but randomly oriented fields along the photon trajectory. The transparency has a complex shape, with little similarity between different realizations. Right panel:Black lines demonstrate the effect of averaging over 30 (blue) or 100 (red) randomly selected realizations. The self-similar shapes emerge, with the green dot-dashed line showing the asymptotic analytical approximation to these lines (Equation \ref{['eq:p_ga']}). The ALP parameters for all curves are the same.
• Figure 2: Correlation between photon index $\Gamma$ and flux $F$ (0.1-100 GeV) in monthly bins for ten AGNs. Points of different colours correspond to different sources as indicated in the legend; statistical uncertainties are shown as error bars in both coordinates. The black dashed line represents the best-fit power-law model to the combined data, with the fitted parameters given in the legend.
• Figure 3: Weekly-binned relation between the photon index $\Gamma$ and photon flux $F$ for two AGN from our sample, selected due to their extensive observational statistics. Blue points represent the Fermi-LAT measurements with statistical uncertainties (error bars) on both flux and photon index. Black dashed lines show the best-fit power-law models, while gray dotted lines indicate the best-fit constant indices. The corresponding best-fit parameters (see Eq. \ref{['fig:gamma_flux']}) and statistical uncertainties are listed in each panel.
• Figure 4: Potential false detection and false exclusion of ALPs due to flux-index correlation. This figure demonstrates that the index-flux correlation can induce spurious detection but cannot hide a genuine ALP signal, if one were present. The solid green line represents the ratio of simulated data to the best-fit power-law model. If the power-law index were independent of the flux, this ratio would be equal to unity within statistical uncertainties. However, the figure clearly demonstrates a significant deviation of the green curve from a horizontal line, well outside the nominal $\pm 5\%$ uncertainty region. If we modify the best-fit model to include ALP-induced suppression, the ratio becomes closer to unity (blue dot-dashed line), except in the lowest and highest energy bins, where systematic uncertainties are larger. This indicates that ALP-like spectral features can spuriously emerge from the correlation bias alone, as including them in the model artificially "improves" the quality of fit. Conversely, the red dotted line ("false exclusion" case) displays the ratio of data containing a genuine ALP signal to a pure power-law model. This ratio deviates more strongly from unity than the green curve, indicating that the ALP signal remains detectable despite the bias.