5. Gas-to-dust mass ratio
where is the m surface brightness in Jy/beam.
Due to the proximity of NGC 891 the neutral gas in this galaxy has been reasonably well investigated. Rupen et al. (1991) mapped out the atomic hydrogen with a high sensitivity and an effective beam of . The atomic gas distribution can be described as a `plateau' which extends approximately out to the but is sharply truncated at this radius. Within the inner few kpc of the stellar disk there is a relative dearth of HI. The molecular gas, on the other hand, appears to be fairly centrally concentrated, with a distribution similar to the submm emission detected by SCUBA. Scoville et al. (1993) employed high resolution interferometry () to map NGC 891 out to radius of 10 kpc in the CO (J=1-0) line. Emission appears to be restricted to a thin disk ( FWHM) with a radial distribution reminiscent of our own Galaxy (Young & Scoville 1991).
In deriving the gas-to-dust mass ratio in NGC 891 we make use of these neutral gas measurements but ensure that all the relevant data are smoothed to a common spatial resolution. 3 Thus, we took the CO interferometry data of Scoville et al. and the dust map based on the m observations, and convolved them with a gaussian of and FWHM respectively. This produced images of comparable resolution to the Rupen et al. map at 21cm. We then profiled along the major axis of each image using a bin width of and a sampling interval of . To convert the CO emission to a H2 column density we used a value of cm-2 K km/s for the conversion factor X. There is considerable controversy surrounding the exact value of X (estimates vary from about 1.5-6 cm-2 K km/s for quiescent, giant spiral disks) but seems appropriate for the general ISM in the Milky Way (Maloney 1990). Ultimately, we wish to compare the gas-to-dust ratio in NGC 891 with that prevailing in the Galaxy so, in lieu of any direct measurement of NGC 891, we adopt cm-2 K km/s. At this stage, we refrain from allowing X to vary with radius (as might be the case for a steep metallicity gradient between the centre and edge of the NGC 891 disk). Maloney (1990) provides a cogent argument as to why corrections to X may only be necessary for environments with a metallicity below solar. Virtually nothing is known about the metallicity in NGC 891 (its edge-on orientation precludes optical spectroscopy) but, so long as NGC 891 constitutes a typical giant spiral galaxy, the heavy element abundance is unlikely to be lower than solar for those parts of the disk where Scoville et al. detect CO emission (Garnet 1998). We check our assumptions about the metallicity in NGC 891 in the next section (Sect. 6).
In Fig. 4, we plot the gas-to-dust ratio along the major axis of NGC 891. For comparison, we also show the corresponding ratio if either of the neutral gas components (HI or H2) is excluded from the calculation. We can see that, at most radii, the molecular hydrogen makes the dominant contribution to the gas mass. Futhermore, the distribution of H2 in NGC 891 seems to mimic that of the grain material very closely (see Fig. 5). There are significant fluctuations in the gas-to-dust ratio brought about by an imperfect matching between the submm and CO major axis profiles. Secondary maxima present in both the H2 and dust profiles, possibly attributable to spiral arms seen edge-on, are somewhat displaced with respect to each other. `Misalignments' between CO emission and dust extinction lanes in more face-on spirals such as M83 and M100 have already been noted by other observers, although it is still unclear whether this is a more subtle effect of density wave propogation or simply CO acting as a poor tracer of H2 (Rand & Kulkarni 1990; Rand 1995; Rand et al. 1999). In the case of NGC 891, Fig. 5 seems to indicate that the overall grain distribution is also more radially extended than the layout of molecular gas. Indeed, it is possible that the submm traces the distribution of atomic gas for radii . Submm observations closer to the edge of the optical disk would be necessary to confirm whether this is a continuing trend since our m data extend to only whilst 21cm emission is detected out to (). The gas-to-dust ratio at the `edge' () of spiral disks remains something of a mystery. However, background galaxies viewed through such outlying regions appear to be systematical reddened by atomic gas possessing a relatively high gas-to-dust ratio (; Lequeux and Guelin 1996). Likewise, m imaging of nearby spirals, using ISO, are also indicative of quite large dust scale-lengths, suggesting an association between cold dust and atomic hydrogen at large radii (Alton et al. 1998b). Our present observations indicate a gas-to-dust ratio towards the optical edge of spiral disks.
The prominent association between CO emission and submm flux density within NGC 891 (and its implied correspondance between H2 and galactic dust), is echoed by recent results from the SCUBA nearby galaxy survey. Dunne et al. (1999) have imaged objects of typical optical size known to possess strong m flux densities ( 5.24 Jy). Notably, they establish a strong correlation between global m and CO emission. Whilst the same trend is observed in both NGC 891 and the nearby, face-on spiral NGC 6946 (Bianchi et al. 2000), it is possible that an even stronger correlation with HI+H2 will emerge as submm observations become more sensitive.
The error-weighted gas-to-dust ratio in NGC 891 is 260 (Fig. 4). The corresponding value for the solar neighborhood is 150, where the dust mass is estimated from either stellar reddening or precipitation of grain elements from the gas phase (Spitzer 1978; Whittet 1992). Sodroski et al. (1994) analyzed the large-scale 140 and m from the Galaxy, as detected by COBE, and thereby established a mean gas-to-dust ratio of 160 in the diffuse ISM. Very recently, a gas-to-dust ratio of 200-500 has been determined in several nearby galaxies using either radiation transfer modelling (Xilouris et al. 1999; Block et al. 1994) or m ISO imaging observations (Alton et al. 1998b). Our determination for NGC 891 is consistent with all these values given that sizeable uncertainties exist in the derived dust masses (via the FIR/submm emissivity) and the quantity of molecular gas (via the X parameter).
The total dust mass in NGC 891, for that part of the disk covered by our SCUBA maps (), is . This is towards the low end of the range given in ABR ( ).
© European Southern Observatory (ESO) 2000
Online publication: April 17, 2000