Axion-like particles (ALPs) can induce localised oscillatory modulations in the spectra of photon sources passing through astrophysical magnetic fields. Ultra-deep Chandra observations of the Perseus cluster contain over $5 \times 10^5$ counts from the AGN of the central cluster galaxy NGC1275, and represent a dataset of extraordinary quality for ALP searches. We use this dataset to search for X-ray spectral irregularities from the AGN. The absence of irregularities at the O(30%) level allows us to place leading constraints on the ALP-photon mixing parameter $g_{a\gamma\gamma} \lesssim 1.4 - 4.0 \times 10^{-12} {\rm GeV}^{-1}$ for $m_a \lesssim 10^{-12}$ eV, depending on assumptions on the magnetic field realisation along the line of sight.

Axion-like particles (ALPs) and photons inter-convert in the presence of a magnetic field. At keV energies in the environment of galaxy clusters, the conversion probability can become unsuppressed for light ALPs. Conversion of thermal X-ray photons into ALPs can introduce a step-like feature into the cluster thermal bremsstrahlung spectrum, and we argue that existing X-ray data on galaxy clusters should be sufficient to extend bounds on ALPs in the low-mass region $m_a \lesssim 1 \times 10^{-12}\,{\rm eV}$ down to $M \sim 7\times 10^{11}\, {\rm GeV}$, and that for $10^{11}\, {\rm GeV} < M \lesssim 10^{12}$ GeV light ALPs give rise to interesting and unique observational signatures that may be probed by existing and upcoming X-ray (and potentially X-ray polarisation) observations of galaxy clusters.

It has been proposed that an excess in soft X-ray emission observed from many galaxy clusters can be explained by conversion into photons of axion-like particles (ALPs) in the cluster's magnetic field. Previously it has been shown that conversion of this primordially-generated background of relativistic ALPs---a cosmic ALP background (CAB)---can explain the observed soft X-ray excess in both the centre and the outskirts of the Coma cluster. Here we extend this work to the three clusters A665, A2199 and A2255. We use a stochastic model of the cluster magnetic field to numerically calculate ALP-photon conversion probabilities to predict the CAB-generated soft X-ray flux in these clusters. We compare this flux to ROSAT PSPC observations of the three clusters, and use these observations to constrain the CAB parameter space. We find the CAB can reproduce the magnitude of the observed excess in A2199 and A2255 for the same CAB parameters that match the observed soft excess in the Coma cluster. We also find good fit to the morphology of the excesses in these clusters. Simulation of CAB conversion in the cluster A665 is in mild tension with the other clusters due to producing a small but observable excess at large radii where none is observed. This tension is alleviated considering the uncertainty on predicting the count rate in the ROSAT detector, and on the systematics affecting the magnetic field determination. Overall we find good agreement between the CAB parameters for the four clusters studied so far.

We further study a scenario in which a 3.55 keV X-ray line arises from decay of dark matter to an axion-like particle (ALP), that subsequently converts to a photon in astrophysical magnetic fields. We perform numerical simulations of Gaussian random magnetic fields with radial scaling of the magnetic field magnitude with the electron density, for both cool-core `Perseus' and non-cool-core `Coma' electron density profiles. Using these, we quantitatively study the resulting signal strength and morphology for cool-core and non-cool-core clusters. Our study includes the effects of fields of view that cover only the central part of the cluster, the effects of offset pointings on the radial decline of signal strength and the effects of dividing clusters into annuli. We find good agreement with current data and make predictions for future analyses and observations.

We show that the soft X-ray excess in the Coma cluster can be explained by a cosmic background of relativistic axions converting into photons in the cluster magnetic field. We provide a detailed self-contained review of the cluster soft X-ray excess, the proposed astrophysical explanations and the problems they face, and explain how a 0.1-1 keV axion background naturally arises at reheating in many string theory models of the early universe. We study the morphology of the soft excess by numerically propagating axions through stochastic, multi-scale magnetic field models that are consistent with observations of Faraday rotation measures from Coma. By comparing to ROSAT observations of the 0.2-0.4 keV soft excess, we find that the overall excess luminosity is easily reproduced for $g_{a\gamma\gamma} \sim 2 \times 10^{-13}$ GeV$^{-1}$. The resulting morphology is highly sensitive to the magnetic field power spectrum. For Gaussian magnetic field models, the observed soft excess morphology prefers magnetic field spectra with most power in coherence lengths on ${\cal O}(3 {\rm ~kpc})$ scales over those with most power on ${\cal O}(12 {\rm ~kpc})$ scales. Within this scenario, we bound the mean energy of the axion background to $50\, {\rm eV}\lesssim \langle E_a \rangle \lesssim 250\, {\rm eV}$, the axion mass to $m_a \lesssim 10^{-12}\,\hbox{eV}$, and derive a lower bound on the axion-photon coupling $g_{a\gamma\gamma} \gtrsim \sqrt{0.5/\Delta N_{\rm eff}}\, 1.4 \times 10^{-13}$ GeV$^{-1}$.

In large volume models reheating is driven by the decays of the volume modulus to the visible sector, while the decays to its axion partners result in dark radiation. In this article we discuss the impact of loop corrections on the only model-independent visible decay channel: the decay into Higgs pairs via a Giudice-Masiero term. Including such radiative effects leads to a more precise determination of the relative fraction of dark radiation, since by contrast all loop corrections to the volume axion decay mode are Planck suppressed. Assuming an MSSM spectrum and that the Giudice-Masiero coupling is fixed at the string scale by a shift symmetry in the Higgs sector, we arrive at a prediction for the effective number of neutrinos. The result turns out to be too large to be consistent with data, highly disfavouring the minimal model.