The M 31 double nucleus probed with OASIS and HST. A natural m=1 mode?

Abstract

We present observations with the adaptive optics assisted integral field spectrograph OASIS of the M 31 double nucleus at a spatial resolution better than 0.5 arcsec FWHM. These data are used to derive the two-dimensional stellar kinematics within the central 2 arcsec. Archival WFPC2/HST images are revisited to perform a photometric decomposition of the nuclear region. We also present STIS/HST kinematics obtained from the archive. The luminosity distribution of the central region is well separated into the respective contributions of the bulge, the nucleus including P1 and P2, and the so-called UV peak. We then show that the axis joining P1 and P2, the two local surface brightness maxima, does not coincide with the kinematic major-axis, which is also the major-axis of the nuclear isophotes (excluding P1). We also confirm that the velocity dispersion peak is offset by ~ 0.2 arcsec from the UV peak, assumed to mark the location of the supermassive black hole. The newly reduced STIS/HST velocity and dispersion profiles are then compared to OASIS and other published kinematics. We find significant offsets with previously published data. Simple parametric models are then built to successfully reconcile all the available kinematics. We finally interpret the observations using new N-body simulations. The nearly keplerian nuclear disk of M31 is subject to a natural m=1 mode, with a very slow pattern speed (3 km/s/pc for M_BH = 7 10^7~\Msun), that can be maintained during more than a thousand dynamical times. The resulting morphology and kinematics of the mode can reproduce the M~31 nuclear-disk photometry and mean stellar velocity, including the observed asymmetries. It requires a central mass concentration and a cold disk system representing between 20 and 40% of its mass. Abridged..