3.1. UGC 2855
UGC 2855 is a large spiral galaxy, almost the same size as the Milky Way, classified as SBc (RC3, de Vaucouleurs et al. 1991). At 20 Mpc distance, its linear diameter to the radius is 25 kpc. Assuming a Galactic extinction in the B band of (Burstein & Heiles 1984), the absolute magnitude of UGC 2855 is , also comparable to the Milky Way. As it is typical for barred spirals, its main spiral arms start at the ends of the bar, and strong molecular emission can be traced in both the bar and the arms.
3.1.1. The distribution of 12CO
We present the total integrated intensity map, the velocity field and the map of the velocity dispersion from our interferometric 12CO observations in Fig. 1. From Fig. 1a, it is clear that the bar in UGC 2855 is filled with molecular gas over its entire length. The position of the peak total intergrated intensity at and agrees almost exactly with the radio continuum peak ( and , Condon et al. 1996). The overall distribution of the molecular gas in the interferometer map is somewhat asymmetric: The northwestern part of the bar outside the center is brighter in CO by about a factor of 1.5 than its southeastern counterpart. This effect can be explained, at least partially, by the fact that the interferometer misses more flux in the southeast of the bar than in the northwest (see Sect. 3.1.2).
The bar major axis has a position angle of 110o and is thus aligned very closely with the line of nodes (the major axis) of the galaxy, which is at PA according to the I-band surface photometry of Héraudeau & Simien (1996). This implies that the bar is seen side-on, at its maximum length. Assuming this viewing geometry and an inclination of , the CO bar can be deprojected, following the formalism given by Martin (1995). The deprojected ratio of bar length to diameter is 0.3, while we determine the axis ratio of the minor and major bar axis to be 0.37. According to the classification of Martinet & Friedli (1997), the bar of UGC 2855 thus qualifies as both `long' and `strong'.
The velocity field along the bar (panel b in Fig. 1) appears regular and suggests a solid body rotation. The isovelocity contours are roughly perpendicular to the bar major axis. The meaning of slight deviations from this behaviour will be explored further in Sect. 4.1.2.
The peak of the velocity dispersion (displayed in Fig. 1, panel c) coincides with the intensity peak. Along the bar, the dispersion drops by a factor of about 5 from km s -1 to km s -1. At the bar ends, rises slightly to km s -1. The narrowness of the lines along the bar is also demonstrated by sample spectra in Fig. 1c.
3.1.2. Line fluxes and the molecular mass
The total flux in the map is Jy km s -1. It is evident from optical images that the spiral arms extend far beyond our map. Thus, this CO flux is a lower limit only and demonstrates that UGC 2855 is a gas-rich galaxy. Assuming a galactic `standard' I(CO)-N(H2) conversion factor (SCF, cm-2 (K km s -1)-1, e.g. Strong et al. 1988), this corresponds to a total H2 mass of . This conversion factor is suspect, especially in the centers of galaxies (Dahmen et al. 1998). It may vary greatly between galaxies and also within a galaxy, depending on the properties of the molecular gas, e.g. the amount of diffuse gas, and on the metallicity (Arimoto et al. 1996). We give a rough estimate of how far the SCF might overestimate the molecular mass by comparing it to the result we derive from the 13CO line intensity. We assume LTE, optically thin 13CO emission, a kinetic temperature of 20 K, a 12CO/13CO ratio of 30 (thought to be typical for centers of galaxies) and a 12CO/H2 ratio of . Then, the H2 column density derived from 13CO for the center of UGC 2855 is lower by a factor of 8 than the column density calculated using the SCF. Interestingly, this value is the same for the interferometric and the single-dish observations, even though the absolute values for the column densities are higher for the far smaller interferometer beam by a factor of more than 10. To force agreement between the column density derived from 13CO and the SCF, we would have to raise to 190 K. For positions off the center, the correction implied by 13CO is smaller by as much as a factor of two. Corroborating evidence from our galaxy as well as external galaxies (e.g. Wall et al. 1993) confirms that gas column densities determined from the SCF and 13CO generally agree better away from the nuclear regions. Of course, the estimates based on 13CO depends on the assumed value of : If there is a gradient in with galactocentric radius (with hotter gas closer to the nucleus), this effect may be cancelled. Thus, the assumptions made for the 13CO emission are vastly oversimplified, but the estimate demonstrates the range of uncertainty of the molecular masses.
Throughout this paper, we will give masses based on the SCF to facilitate comparisons with other work and also because a determination of a `correct' conversion factor that goes beyond the estimate given above would require detailed information on the gas properties not yet available. It should be kept in mind, though, that the masses may be lower by almost an order of magnitude.
As a further complication, the interferometer is likely to filter out flux from an extended gas component due to missing zero spacing coverage. We find a shallow negative trough surrounding the strongest features in our map, indicating that the array may indeed miss some extended emission. To get a rough estimate of the amount of missing flux, we have compared the flux contained in single Onsala beams (see Fig. 3) to the flux in equivalent regions of the interferometer beam, weighted by the shape and size of the single dish beam (assumed to be gaussian). From this, toward the center and the northwestern part of the bar, we see at least 80% of the single dish dish flux in the interferometer map. For the southeastern part of the bar, there are indications that a somewhat lower amount of flux is picked up by the interferometer, but still at least 65% (and possibly as much as 80%) of the single dish flux is seen. These numbers imply that not much gas is distributed in a very diffuse component that has little structure on scales exceeding kpc. However, all pointing and calibration errors enter in this estimate, which therefore has to be considered to be very tentative. Clearly, in the mass estimates, the error introduced by the missing flux is dominated by the uncertainty of the conversion factor.
The total flux the interferometer detects along the bar, including the bar end regions, but excluding the beginning of the spiral arms, is 895 Jy km s -1. This corresponds to a molecular mass of , under the assumptions described above. The central beam contains a flux of 137 Jy km s -1 ( ).
The results of the OVRO 13CO measurements are presented in Fig. 2. Panel a shows the total integrated intensity distribution. 13CO is only detected toward the center of UGC 2855. Assuming a constant 12CO/13CO line intensity ratio (see below), this is, however, expected, since the sensitivity in the 13CO map is not sufficient to detect emission away from the intensity peak. The positions of the 13CO peak and the 12CO peak are identical. To within the noise, the linewidth and the line shape of the two isotopomers also agree (Fig. 2, panel b). The observations taken at the OSO 20 m telescope (Fig. 3) confirm the good agreement between the line shapes of 12CO and 13CO.
Fig. 4 (panel a) shows our H image of UGC 2855. Since the image lacks an absolute scale, absolute H fluxes or star formation rates cannot be derived. However, it gives information on the relative strength of H emission and thus star formation activity, in the center, along the bar and in the spiral arms. H emission is clearly detected in the center, but the central peak is not the strongest one in the galaxy. The peak in the southeastern spiral arm, outside the limits of our CO map, is almost twice as strong as the central emission, and the emission from three more peaks, all at the very edge or outside our CO mosaic, has roughly the same strength as the central emission. This demonstrates that, while there is some central activity, the nucleus of UGC 2855 is not (yet?) in a starburst phase. Of course, the nuclear H emission may be subject to more extinction than the emission from the spiral arms (Phillips 1996). However, starburst galaxies show nuclear H emission that is clearly enhanced over that seen in their disks, despite having stronger extinction effects in their nuclei than in their disks (e.g. Contini et al. 1998, who find that the H luminosity in the starburst nuclei of their sample is higher by a factor of than in extranuclear H II regions; another example is UGC 2866, see Sect. 3.2).
Along the bar, in particular the southeastern part, faint H emission is detected. Even though the subtraction of the background is problematic, we estimate that the total emission (not corrected for extinction) from the bar and the emission from the central region are about equal.
3.2. The companion: UGC 2866
The picture presented by UGC 2866, the smaller, distant companion of UGC 2855, is very different. The classification of this galaxy is not clear: The UGC catalogue (Nilson 1973) tentatively calls it an early spiral, while it is (mis)classified as an elliptical by Hau et al. 1995. It is one of only 21 galaxies from the IRAS PSC that are stronger than 1.5 Jy at both 12 µm and 25 µm with IRAS LRS spectra allowing its identification as an H II galaxy in the sample of Cohen & Volk (1989). All the molecular gas of UGC 2866 seems to be concentrated in one feature of slightly more than 2 kpc length that has an oval, possibly bar-like appearance (Fig. 5a). The emission peaks at and , unusually far () offset from the radio continuum peak at and (Condon et al. 1996).
The generally smooth, almost classical `spider' shape of the velocity field (Fig. 5b) indicates gas that is predominantly in circular motion, though some deviations, especially close to the edges of the structure, are present. The velocity dispersion is, once more, highest in the center, but the drop along the major axis is less extreme than in UGC 2855. For an offset along the major axis of UGC 2855 of and , the dispersion drops from the central peak of 73 km s -1 to 23 km s -1 and 19 km s -1 respectively, while the drop for corresponding offsets in UGC 2866 (central peak at 75 km s -1) is only to 36 km s -1 and 44 km s -1.
We detect a total flux of 543 Jy km s -1 in our map of UGC 2866, which, again assuming a `standard' conversion factor, translates to an H2-mass of , not much smaller than what is found in the bar of UGC 2855. The central beam contains of H2. If the radiocontinuum peak is identified with the dynamical center of the galaxy, about 55% of the CO emission arises in the northeastern half, which also contains the intensity maximum.
The H emssion from UGC 2866 (Fig. 4, panel b, a part of the same CCD frame as the image of UGC 2866) is very bright and very compact: it extends over only or pc. The intensity of the emission is more than 6 times higher than what is found toward the center of UGC 2855. The center position of H coincides with the CO peak in UGC 2866 to better than 1", i.e. to within the measuring uncertainty. The strength and compactness of H in UGC 2866 and the FIR properties (see below) indicate that this galaxy is experiencing a nuclear starburst.
© European Southern Observatory (ESO) 1999
Online publication: May 6, 1999