Observing the very low surface brightness dwarfs in a deep field in the VIRGO cluster: constraints on dark matter scenarios ^⋆

INAF–Osservatorio Astronomico di Roma, via Frascati 33, 00078 Monteporzio, Italy
e-mail: menci@oa-roma.inaf.it

Received: 24 May 2017
Accepted: 13 June 2017

Abstract

We report the discovery of 11 very faint (r ≲ 23), low surface brightness (μ_r ≲ 27 mag/arcsec²) dwarf galaxies in one deep field in the Virgo cluster, obtained by the prime focus cameras (LBC) at the Large Binocular Telescope (LBT). These extend our previous sample to reach a total number of 27 galaxies in a field of just ~0.17 deg² located at a median distance of 390 kpc from the cluster centre. The association of such galaxies with the Virgo cluster is supported by their separate position in the central surface brightness – total magnitude plane with respect to the background galaxies of similar total magnitude. For a significant fraction (26%) of the sample, the association to the cluster is confirmed by spectroscopic follow-up. We show that the mere abundance of satellite galaxies corresponding to our observed number in the target field provides extremely tight constraints on dark matter models with suppressed power spectrum compared to the cold dark matter case, independently of the galaxy luminosity distribution. In particular, the requirement that the observed number of satellite galaxies not exceed the predicted abundance of dark matter sub-halos yields a limit of m_X ≥ 3 keV at 1-σ and m_X ≥ 2.3 keV at 2-σ confidence level for the mass of thermal warm dark matter particles. Such a limit is competitive with other limits set by the abundance of ultra-faint satellite galaxies in the Milky Way, is completely independent of baryon physics involved in galaxy formation, and has the potentiality for appreciable improvements with future observations. We extend our analysis to dark matter models based on sterile neutrinos, showing that our observations set tight constraints on the combination of sterile neutrino mass m_ν and mixing parameter sin²(2θ). We discuss the robustness of our results with respect to systematics.