Astrophysical limits on very light axion-like particles from Chandra grating spectroscopy of NGC 1275

Abstract

Axions/axion-like particles (ALPs) are a well motivated extension of the Standard Model and are generic within String Theory. The X-ray transparency of the intracluster medium (ICM) in galaxy clusters is a powerful probe of light ALPs (with mass $<10^{-11}\mathrm{eV}$); as X-ray photons from an embedded or background source propagate through the magnetized ICM, they may undergo energy-dependent quantum mechanical conversion into ALPs (and vice versa), imprinting distortions on the X-ray spectrum. We present Chandra data for the active galactic nucleus NGC1275 at the center of the Perseus cluster. Employing a 490ks High-Energy Transmission Gratings (HETG) exposure, we obtain a high-quality 1-9keV spectrum free from photon pileup and ICM contamination. Apart from iron-band features, the spectrum is described by a power-law continuum, with any spectral distortions at the $<3\%$ level. We compute photon survival probabilities as a function of ALP mass $m_a$ and ALP-photon coupling constant $g_{a\gamma }$ for an ensemble of ICM magnetic field models, and then use the NGC1275 spectrum to constraint the $(m_a,g_{a\gamma })$-plane. Marginalizing over magnetic field realizations, the 99.7% credible region limits the ALP-photon coupling to $g_{a\gamma }<6-8\times 10^{-13}{\mathrm{GeV}}^{-1}$ (depending upon magnetic field model) for masses $m_a<1\times 10^{-12}\mathrm{eV}$. These are the most stringent limit to date on $g_{a\gamma }$ for these light ALPs, and have already reached the sensitivity limits of next-generation helioscopes and light-shining-through-wall experiments. We highlight the potential of these studies with the next-generation X-ray observatories Athena and Lynx, but note the critical importance of advances in relative calibration of these future X-ray spectrometers.

Published in The Astrophysical Journal, 2020, vol 890, pp59. This posting corrects typos in equation (2) and the likelihood function just prior to equation (3). Typos were introduced into the draft at late stage and the analysis is correct