Dynamical models of NGC 3115

Citation
E. Emsellem et al., Dynamical models of NGC 3115, M NOT R AST, 303(3), 1999, pp. 495-514
Citations number
45
Categorie Soggetti
Space Sciences
Journal title
MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETY
ISSN journal
00358711 → ACNP
Volume
303
Issue
3
Year of publication
1999
Pages
495 - 514
Database
ISI
SICI code
0035-8711(19990301)303:3<495:DMON3>2.0.ZU;2-Z
Abstract
We present new dynamical models of the S0 galaxy N3115, making use of the a vailable published photometry and kinematics as well as two-dimensional TIG ER spectrography, The models are based on a detailed model of the luminosity distribution bui lt using an MGE fit on HST/WFPC2 and ground-based photometric data. We firs t examined the kinematics in the central 40 arcsec in the light of two-inte gral f(E, J) models. Jeans equations were used to constrain the mass-to-lig ht ratio, and the central dark mass, the existence of which was suggested b y previous studies, The even part of the distribution function was then ret rieved via the Hunter & Qian formalism. We thus confirmed that the velocity and dispersion profiles in the central region could be well-fitted with a two-integral model, given the presence of a central dark mass of similar to 10(9) M.. However no two-integral model could fit the h(3) profile around a radius of about 25 arcsec where the outer disc dominates the surface brig htness distribution, Three-integral analytical models were therefore built using a quadratic pro gramming technique. These models showed that three-integral components do i ndeed provide a reasonable fit to the kinematics, including the higher Gaus s-Hermite moments. Again, models without a central dark mass failed to repr oduce the observed kinematics in the central arcsec. This clearly supports the presence of a nuclear black hole of at least 6.5 x 10(8) M. in the cent re of NGC 3115. These models were finally used to estimate the importance o f the dark matter in the outer part of NGC 3115, suggested by the flat stel lar rotation curve observed by Capaccioli et al. This study finally points out the difficulty of integrating independently p ublished data in a coherent and consistent way, thus demonstrating the impo rtance of taking into account the details of the instrumental setup and the reduction processes.