Evidence for a supermassive black hole in the S0 galaxy NGC 3245

The S0 galaxy NGC 3245 contains a circumnuclear disk of ionized gas and dust with a radius of 1.1 (110 pc), making it an ideal target for dynamical studies with the Hubble Space Telescope (HST). We have obtained spectra of the nuclear disk with the Space Telescope Imaging Spectrograph, using a 0.2 wide slit at five parallel positions. Measurements of the H alpha and [N II] emission lines are used to map out the kinematic structure of the disk in unprecedented detail. The data reveal a rotational velocity field with a steep velocity gradient across the innermost 0.4. We construct dynamical models for a thin gas disk in circular rotation, using HST optical images to map out the gravitational potential due to stars. Our modeling code includes the blurring due to the telescope point-spread function and the nonzero slit width, as well as the instrumental shift in measured wavelength for light entering the slit off-center, so as to simulate the data as closely as possible. The H alpha + [N II] surface brightness measured from an HST narrowband image is folded into the models, and we demonstrate that many of the apparent small-scale irregularities in the observed velocity curves are the result of the patchy distribution of emission-line surface brightness. Over most of the disk, the models are able to fit the observed radial velocity curves closely, although there are localized regions within the disk that appear to be kinematically disturbed relative to the overall rotational pattern. The velocity dispersion of [N II] lambda 6584 rises from sigma approximate to 50 km s(-1) in the outer disk to similar to 160 km s(-1) at the nucleus, and most of this line width cannot be attributed to rotational or instrumental broadening. To account for the possible dynamical effect of the intrinsic velocity dispersion in the gas, we also calculate models that include a correction for asymmetric drift. This correction increases the derived black hole mass by 12% but leads to slightly poorer fits to the data. A central dark mass of (2.1 +/-0.5) x 10(8) M-circle dot is required for the models to reproduce the steep central velocity gradient. This value for the central mass is consistent with recently discovered correlations between black hole mass and bulge velocity dispersion.