Structure of the circumnuclear region

The linear analysis of the kinematics indicates the existence of a possible iILR at 2 arcsec. Do we see a correlation between this dynamical feature and any morphological or spectroscopic properties?

Morphologically, the H1#1, colour, and HST optical images all show structure within the inner 2 arcsec; i.e., it is not a point like nucleus. Friedli et al. (1996) have looked without success for the existence of a secondary bar, that could be possibly associated with the existence of the iILR. There is no indication of such an inner bar either in our images. In fact, the HST optical image shows a clear inward spiraling structure down to 0.5 arcsec radius (46 pc).

Spectroscopically, several diagnostics are spatially resolved within the inner 2 arcsec radius, such as the [Nii]/H1#1 ratio (Fig. ), the FWHM of both H1#1 and [Nii] (Fig. ), and the equivalent with of the CaT lines indicating the presence of an important population of red supergiants in the nucleus (Fig. ). The gaseous and stellar velocity curves also show features within the iILR: for the three PA, the two stellar components merge at the iILR and are no longer distinguishable; the gas velocity curves at PA=84 and 138 flatten within a 1.5 arcsec radius, while at PA=48 the gas velocity shows a peculiar structure within a 2 arcsec radius, that is not present in the stellar velocity.

This peculiar structure along the minor axis could be due to an off-centered slit, to a gaseous nuclear outflow, or to a structure related to the iILR. The first possibility is ruled out because this velocity structure is present only in the ionized gas but not in the stars, that implies that it is a physically existing feature, given that the two spectral regions were observed simultaneously and through the same slit. Of the two other possibilities, our data do not have sufficient spatial resolution to make a more definitive conclusion. However, we can make a few comments. A nuclear outflow could be produced by the combined effect of the winds and supernova explosions of the recent past generation of star formation indicated by the red supergiant stars we have detected. Saikia et al. (1994) give a total flux for the nuclear beam component of NGC 6951 at 6 cm of 1.1 mJy; they argue that this non-thermal flux is most likely due to supernovae. This flux implies a luminosity of 0.47 43#43 watt Hz^-1, that can be converted to a supernova rate using the models of Colina & Pérez-Olea (1992), to give 0.002 yr^-1. A similar calculation with the 20 cm radio flux in the nuclear component given by Vila et al. (1990), that implies a luminosity of 1.03 43#43 watt Hz^-1, yields the same value for the SN rate. This is a low SN rate when compared with bright nuclear starbursts such as NGC 7714, for which González Delgado et al. (1999a) find 0.07 yr^-1, but is qualitatively consistent with the faintness of the nucleus in H1#1. If a nuclear outflow were the explanation for the kinematic features in the nucleus, we would only be seeing the approaching, blueshifted, side of the outflow with a speed of 13#13100 km s^-1; the typical expansion velocity of a superwind bubble blown by a nuclear starburst in dwarf galaxies is of the order of 50 km s^-1 (Marlowe et al. 1995); but is significantly larger, a few 100 km s^-1, for nuclear superwinds in Starburst galaxies (González Delgado et al. 1998). The receding side would be presumably occulted by a combination of nuclear obscuration and spatial and spectral resolution.

A third possibility suggested by the data is that the nuclear gas dynamics inside the iILR is partially decoupled from that of the circumnuclear region which, in turn, is partially decoupled from that at larger scales. This would be a scenario of nested disks within disks, where a circumnuclear disk accumulates mass from the outer main galaxy disk via the torques produced by the bar, and becomes sufficiently massive to decouple from the main disk dynamics at the oILR; part of this infall proceeds further into the nucleus, where a similar rotating structure (a torus or a nuclear disk such as those seen in HST images, Ferrarese & Ford 1999; Ford et al. 1998) decouples from the circumnuclear disk at the iILR. We notice that the orientation of the ring in NGC 6951 is leading with respect to the inflow of material from the bar, i.e. it is advanced from the perpendicular to the bar, and this relative orientation is seen in numerical simulations only when the system is relatively old (see, for example, figure 3 of Byrd et al. 1994). In this context, it would be possible to interpret the change of the velocity curve in the three PA, within the inner 2 arcsec radius, as produced by rotation in a disk with a minor axis different from the 48 of the main disk, and different from the 30 of the circumnuclear disk (cf. section 4.3); it would have a value somewhere between 84 and 48, so that at PA=48 we could observe the change of velocity slope with respect to that at 84.

The appearance of the HST images would support this scenario. Figure  shows the HST V and H sharp divided images; the H image shows a very uniform stellar light distribution in the transition zone between the ring and the nucleus. However, the V image shows a structure spiralling inwards, delineated by the dust that can be traced right into the nuclear 0.5 arcsec. Thus, these high resolution images show both the lack of any inner bar-like structure and the existing of a continuing spiralling into the nucleus, similar to that found in numerical simulations by Piner et al. (1995, cf. their fig. 4). A similar conclusion has been recently reached by Regan & Mulchaey (1999), who have looked for evidences of strong nuclear bars in WFPC2 and NICMOS2 images of a sample of Seyfert galaxies, and find evidence of the existence of this nuclear bar in only 3 out of 12 galaxies studied, while the majority of the galaxies show a spiral morphology similar to what we find in NGC 6951. This does not necessarily implies that the 'bars within bars' mechanism proposed to fuel the nucleus has not been at work in these galaxies (Shlosman et al. 1989; Friedli & Martinet 1993); at least in NGC 6951 we have found several indications that point to a system that is dynamically old, in particular, the large percentage of molecular mass accumulated in the nucleus, 2#2%, could have already occasioned the dissolution of a nuclear bar, and be working towards the desassembling of the large scale bar.

This idea is inferred from extrapolations of three slit PA. While we understand that a complete 2D spectroscopic mapping of this and other galaxies is required to confirm or otherwise dismiss it, it is also true that such a 2D mapping at a spectral and spatial resolution equal to or better than our data is very difficult to obtain with present day instrumentation. Recent developments like SAURON (Miller et al. 1999) and similar instruments will have an important impact on such studies in the next few years.