4. Kinematics of stars and gas in NGC 7217
4.1. Circumnuclear region
In the previous paper (Zasov & Sil'chenko 1997) we noted clear signatures of the dynamical distinctness of the nucleus in NGC 7217: the rotation axis of the ionized gas inside turns by almost 90o. Now we present further proofs of the existence of a circumnuclear `polar' gaseous ring (disk?) in this galaxy.
We now proceed to determine the parameters of the stellar and gaseous rotation around the nucleus of NGC 7217. Under the assumption of planar axisymmetric rotation, the azimuthal dependence of central line-of-sight velocity gradients is:
where is the deprojected central angular rotation velocity, i is the inclination of the rotation plane and is the orientation of the line of nodes, coinciding in the case of an axisymmetric ellipsoid (or a thin disk) with the photometric major axis. So by fitting azimuthal variations of the central line-of-sight velocity gradients with a cosine curve, we can determine the orientation of the dynamical major axis by its phase and the central angular rotation velocity by its amplitude. To apply this procedure, we need two-dimensional velocity fields.
The new MPFS observations allow us to construct two-dimensional velocity maps for the central region of NGC 7217. Fig. 3 shows such maps for the stars (left) and for the ionized gas as traced by the [NII ]6583 emission (middle), as well as the stellar velocity dispersion map (right). One can see that two velocity maps look different. At first glance, the dynamical major axis of the stellar component seems to be close to the global line of nodes, west-east. However it is strange that the brightness center coincides with a flat area in the velocity map. If the dynamical center coincides with the brightness center, we would obtain a central projected angular rotation velocity of only () km/s/arcsec. If we shift the dynamical center 3" to the east, to the point of maximum isovelocity crowding, we can increase to () km/s/arcsec. Fortunately, the direction of the maximum line-of-sight velocity gradient, or the dynamical major axis, remains the same for both dynamical center locations: it is within , some 30o different from the line-of-nodes direction, but coinciding with the PA of the low-contrast bar reported by Buta et al. (1995). Outside the dynamical major axis turns to , implying an axisymmetric stellar rotation parallel to the symmetry plane of the galaxy. The velocity map for the ionized gas obtained by measuring the baricenters of the [NII ] emission line (H is less relevant because of the underlying absorption line) looks even more striking than that for the stars. The isovelocities in the very center demonstrate a strong so-called "S-shape" distortion, whose dynamical major axis direction changes by 90o in the radius range of 3" -7". By fitting a cosine curve to the line-of-sight velocity gradients within 3" from the center, we obtain a projected angular rotation velocity of () km/s/arcsec and a dynamical major axis , orthogonal to the dynamical major axis of the stars in this radius range. Again, we find the signature of a circumnuclear polar gaseous disk.
The two-dimensional map of the stellar velocity dispersion in the central 16" region (Fig. 3, right) looks somewhat peculiar: though rather smooth, it demonstrates an elongated `saddle' of relatively low , at about km s-1, in and a slight increase of along the kinematic minor axis on both sides of the nucleus. We have never seen anything like this and cannot give an interpretation of this map. The dynamical center which we define here as a center of symmetry of the stellar velocity dispersion map again seems to be shifted to the east of the brightness center.
We can check our result on the decoupled rotation of the circumnuclear ionized gas with the long-slit spectral data from the ING Archive. Fig. 4 shows line-of-sight velocities measured from [NII ]6583 and H emission lines in and . In , not too far from the line of nodes, one can see a flat velocity profile segment in the very center, till from the nucleus, implying that the dynamical major axis of the circumnuclear ionized gas is aligned orthogonally, in (or -22o). Outside this central region the gas shows much larger projected velocities with respect to the nucleus. In , consistent with the previous cross-section, the central part of the velocity profile shows fast decoupled rotation within from the nucleus with a slope of 24 km/s/arcsec - evidently, this direction is close to the dynamical major axis of the circumnuclear gas; outside the radius of 4" the projected rotation velocity falls to a half of the maximum value reached at the 4" radius. Therefore, the long-slit cross-sections confirm the dynamical distinctness of the central region of NGC 7217 within a radius of 3" -4" and the existence of a `polar' circumnuclear gaseous disk rotating in a plane orthogonal to the global plane of the galaxy.
4.2. The whole galaxy
Kinematics of the ionized gas in NGC 7217 has been studied more than once. Long-slit cross-sections along the major axis were obtained by Peterson et al. (1978), Rubin et al. (1985), and Buta et al. (1995), and our team also observed NGC 7217 with a long-slit spectrograph and a scanning Fabry-Perot interferometer at the 6m telescope (Zasov & Sil'chenko 1997). We have now analyzed the H and [NII ]6583 emission lines in the three additional long-slit cross-sections taken from the ING Archive. The results obtained over the full radius range are quite consistent with the velocity curves published previously, and we do not present the raw velocity data here.
The stellar kinematics in NGC 7217 have received much less attention than the gas kinematics. Merrifield & Kuijken (1994) cross-correlated absorption spectra taken along the major and minor axes in the spectral range near Mgb5175 with those of template stars, and reported a clear evidence of the presence of a counterrotating stellar subsystem in the disk of NGC 7217: LOSVDs of stars look double-peaked in the full radius range considered by Merrifield & Kuijken (1994), i.e., up to 60" from the center. However, their work lacks direct measurements of the velocities of the co-rotating and counter-rotating components, after their separation in the spectra, as well as estimates of velocity dispersions, so their conclusion that the counter-rotating component is a part of the global disk of the galaxy is not very convincing, as it is only based on an ambiguous photometric decomposition. We have calculated stellar velocity profiles for two spectra in , one near the NaI 5890,5896 lines and another near the CaII Ir triplet, and for one spectrum near NaI in through an interactive Gauss analysis of the cross-correlation peaks (as a template star, we have taken HR 661). The results are presented in Fig. 5 and Fig. 6. We confirm the existence of the counter-rotating stellar component seen in the cross-sections not far from the major axis, in . But its velocity profile (Fig. 5, left) does not represent a simple mirror picture of the main co-rotating component. While the main rotation curve is rather flat between and 30", the counter-rotating component has a peak near and then already falls back to the systemic velocity at . If the result obtained by Merrifield & Kuijken (1994) for major-axis cross-section differing by 30o in PA from ours is true, i.e., if two maxima of LOSVD are located symmetrically with respect to the systemic velocity, this may signify that the planes of rotation of the co-rotating and counter-rotating stars are different. The counter-rotating stellar subsystem may rotate faster than the co-rotating one; the reason for this can be found in Fig. 6. The stellar velocity dispersion in NGC 7217 is strongly peaked near the nucleus where it reaches about 160 km s-1 (previous estimates: km s-1, Whitemore et al. 1979; km s-1, Dressler 1984). In both directions, in and , it falls rapidly with radius and becomes less than 100 km s-1 at . But the velocity dispersion of the counter-rotating component seems to be systematically lower than that of the co-rotating one, so since both rotate in the same potential, the former has the faster rotation. As the stellar velocity dispersion of the counter-rotating component is of order of 50 km s-1, we can surely state that it is in a disk, so the guess of Merrifield & Kuijken (1994) has been correct.
In Fig. 7 we compare the stellar and gaseous rotation curves deprojected with the orientation parameters and from the data noted in the legend. All three gas rotation curves, the one from Peterson et al. (1978) as approximated by the analytic formula in the paper of Merrifield and Kuijken (1994) and our two, from the WHT-91 cross-section in and from the WHT-97 cross-section in , are in general agreement with each other: they are flat, with a maximum velocity level of km s-1 and a possible decrease by some 20 km s-1 in the radius range of 4.5-6 kpc. The stellar rotation looks more irregular though we have taken into consideration the main component only, i.e. the brighter which is co-rotating with the gas. The highest values of stellar rotation velocities are comparable to the gas ones, thus supporting the idea of disk predominance at already kpc - in favour of Kent's (1986) photometric decomposition (see the next section). But in many places the stellar rotation velocities fall below the gas rotation curve. The most striking feature of this kind is seen around kpc: at different directions and wavelengths (implying: for stars of different ages?) the stellar velocity is 50% -100% less. Interestingly, a similar detail at the same radius was found by us in the azimuthally averaged rotation curve obtained for the ionized gas with the scanning Fabry-Perot interferometer (Zasov & Sil'chenko 1997). We noted that this feature is located at the outer edge of the inner pseudoring and perhaps related to some non-circular motions; but no detailed interpretation was given. Now, when we detect an even stronger response to the ring in a dissipationless dynamical subsystem, we are even more astonished about the nature of the inner pseudoring.
© European Southern Observatory (ESO) 2000
Online publication: January 29, 2001