Astron. Astrophys. 356, 795-807 (2000)

6. Metallicity considerations

In this section, we attempt to establish the heavy element abundance along the disk of NGC 891. A determination of metallicity may cause us to modify the value of X, chosen in the previous section, to convert CO line emission to H₂ column density. Since the gas-to-dust ratio, at least in the inner half of the disk, is dominated by the molecular rather than atomic gas, some justification for our chosen value of X would seem in order. Due to the obscuring effects of the NGC 891 dust lane, we resort to FIR cooling lines from highly ionized metal species (e.g. OIII m) which are expected to be optically thin even for a galaxy viewed edge-on. Accordingly, we have searched the ISO archive for Long Wavelength Spectrometer (LWS) and Short Wavelength Spectrometer (SWS) observations of NGC 891 and, although scientifically-validated spectra of several HII regions were procured, we could not relate the line strengths nor their ratios to any of the diagnostic models currently available in the literature (Statinska 1990; Spinoglio & Malkan 1992).

We then explored more indirect means of inferring the metallicity in NGC 891. Garnet (1998) has shown that there is a clear trend, spanning 4 magnitudes in B-band brightness, between relative oxygen abundance O/H and absolute galaxy luminosity (see also Garnett & Shields 1987). Although extinction effects are severe in NGC 891 (due to its edge-on orientation), the radiative transfer model of XAD may be used to derive the intrinsic blue luminosity of the galaxy. The object can then be located on the O/H vs.  plot of Garnet. Following Garnet, we adopt =50 km/s(Mpc)^-1 in order to determine the distance modulus and hence absolute B magnitude of NGC 891. Thus, we derive a value of -21.0 for . On the Garnet plot this corresponds to an oxygen abundance of 12+log()=9.20.2 at 1 B-band scale-length from the nucleus (where 0.2 constitutes the dispersion in the metallicity-magnitude `relation'). The Milky Way has an O/H abundance of 9.1 at 1 B-band scale-length from the centre (Shaver et al. 1983), suggesting a very similar metallicity between NGC 891 and our own Galaxy.

As mentioned in the previous section, we might expect to modify our value chosen for the conversion factor X if the metallicity in NGC 891 falls below solar (O/H=8.9). For an average metallicity gradient of 0.08 dex/kpc in spiral disks (Vila-Costas & Edmunds 1992), we expect the oxygen abundance to be 8.90.3 at the maximum radius at which we convert CO line emission to H₂ column density (r=9.2 kpc). This suggests that the vast majority of molecular gas in NGC 891 has a heavy element abundance similar to, or indeed above, the solar level. Thus the value of cm^-2 K km/s, generally adopted for the general ISM inside the solar ring (Sect. 5), can be used with some justification for NGC 891. In Fig. 6, reproduce the plot of Issa et al. (1990) showing the trend of increasing metallicity with higher dust-to-gas ratio for galaxies close to or in the Local Group. We incorporate the corresponding quantities derived for NGC 891 which, although subject to appreciable errors, show pleasing agreement with the correlation.

Fig. 6. Metallicity against dust-to-gas ratio for NGC 891 compared to galaxies in or near the Local Group (the latter taken from Issa et al. 1990). Each quantity is derived at a distance of from the centre of the corresponding object and is normalized to the Milky Way. The position of NGC 891 is denoted by a solid triangle and the corresponding errorbar represents principally the uncertainty in the conversion factor X and the dispersion in the metallicity-magnitude relation of Garnet (1998).

© European Southern Observatory (ESO) 2000

Online publication: April 17, 2000
helpdesk.link@springer.de