Cornering the axion-like particle explanation of quasar polarisations

In a series of paper, it has been shown that the distribution of polarisation position angles for visible light from quasars is not random in extremely large regions of the sky. As explained in a recent article, the measurement of vanishing circular polarisation for such quasars is an important problem for a mechanism involving the mixing with axionlike particles in external magnetic fields. In this note, we stress that a recent report of similar coherent orientations of polarisation in radio waves further disfavours the need for such particles, as an effect at these wavelengths would be extremely suppressed or would directly contradict data. The existence of coherent orientations in the polarisation of visible light from quasars in large-scale (∼ 1 Gpc) regions of the sky [1, 2] is a very puzzling observation. To date, no satisfactory explanation of this effect, based on 355 high-quality measurements of linear polarisation from quasars, is available. Until recently, the best hope to explain these data involved axion-like particles (ALPs), extremely light spinless particles with a coupling to two photons similar to the Primakoff effect for neutral pions. In a nutshell, light from these quasars could mix with such ALPs inside the external magnetic fields encountered on their way towards Earth [4]. Given the dichroic property of the mixing, it was believed that coherent alignments could be reproduced from initial random distributions of polarisation position angles. However, in a recent publication [5], we have shown that measurements of null circular polarisation in V-filter for objects taken from the 355-quasar sample [6] is in severe contradiction with the phenomenology expected from ALP-photon mixing. The production of alignments of linear polarisation, while keeping the circular polarisation small, fails already at the qualitative level—even within a wave-packet treatment, including fluctuations, and considering more refined magnetic field configurations. As there is either too much circular polarisation, too much linear polarisation, or no alignment in general, we concluded in [5] that this mechanism was strongly disfavoured by data. Recently, a very interesting data analysis [7] has suggested that similar coherent orientations in the linear polarisation of quasars also exist in radio wavelengths (8.4 GHz). While the authors do agree with the original data analysis [8] in the restricted case which was ∗e-mail: [email protected]. Let us stress that this work only deals with the mixing of light with ALPs inside external magnetic fields. In particular, the case of the propagation of light inside an hypothetical anisotropic scalar background field, for which the phenomenology is very different [3], is beyond the scope of this note.