Milky Way as a kiloparsec-scale axionscope

Very high energy gamma rays are expected to be absorbed by the extragalactic background light over cosmological distances via the process of electron-positron pair production. Recent observations of cosmologically distant gamma-ray emitters by ground based gamma-ray telescopes have, however, revealed a surprising degree of transparency of the universe to very high energy photons. One possible mechanism to explain this observation is the oscillation between photons and axionlike particles (ALPs). Here we explore this possibility further, focusing on photon-ALP conversion in the magnetic fields in and around gamma-ray sources and in the magnetic field of the Milky Way, where some fraction of the ALP flux is converted back into photons. We show that this mechanism can be efficient in allowed regions of the ALP parameter space, as well as in typical configurations of the galactic magnetic field. As case examples, we consider the spectrum observed from two HESS sources: $1\mathrm{ES}1101\mathrm{\text{−}}232$ at redshift $z=0.186$ and $\mathrm{H}2356\mathrm{\text{−}}309$ at $z=0.165$. We also discuss features of this scenario which could be used to distinguish it from standard or other exotic models.

Figure 1

The probability of ALP-photon conversion in three different models of the galactic magnetic field (PS, TT, HMR from top to bottom), for $g_{11}=5$ and assuming that the conditions of Eq. (3) are satisfied. Also shown are the locations of the known very high energy gamma-ray sources at redshift greater than 0.1, labeled according to the symbols found in Table .

Figure 2

Suppression of the gamma-ray spectrum due to pair production for sources at redshifts of $z=0.1$, 0.2, 0.3, 0.4, and 0.5, from top to bottom.

Figure 3

The gamma-ray spectrum from a source with an injected spectrum of $dN/dE\propto E^{-2}$, after propagation over a distance of $z=0.2$ (top) and $z=0.5$ (bottom). Results are shown with and without the effect of photon-ALP oscillations. The ALP-photon mixing mitigates the impact of absorption via pair production and leads to a plateau in the spectrum at high energies. We have calculated the effects of ALP-photon mixing assuming a source conversion probability of 0.3 and a Milky Way reconversion probability of 0.1. The vertical axis is in arbitrary units.

Figure 4

The points shown with error bars represent the observed gamma-ray spectrum of $\mathrm{H}2356\mathrm{\text{−}}309$ (top) and $1\mathrm{ES}1101\mathrm{\text{−}}232$ (bottom), from the data of Ref. 2. The other points shown denote the spectra required to be injected at the source in order to generate the observations after accounting for the absorption by the EBL. Using a conventional EBL spectrum and density, the reconstructed source spectra are required to be extremely hard ($dN/dE\propto E^{-0.6}$ and $dN/dE\propto E^{0.2}$). The mechanism of photon-ALP mixing, however, can soften the required spectral slope at injection to acceptable values. See text for more details.