5.4. Environmental Effects?
Spiral galaxies are known to be modified by the environment within rich clusters of galaxies (see reviews by Haynes 1990 and Whitmore 1990). Tidal stripping, ram pressure stripping, and ablation by the intracluster medium can all play a role and may lead to significant effects on galaxy evolution. The cluster environment clearly influences galaxy morphology (Dressler 1980), which leads to the natural question of whether it also affects the star formation and chemical evolution histories of galaxies. Does the cluster environment predetermine galaxy morphology by solely influencing initial conditions or does it continually influence the evolution of a member galaxy? Comparing the chemical abundances of field and cluster members may provide clues to these fundamental questions. Moreover, a systematic abundance differential between cluster and field galaxies would have implications for the stellar populations and could impact a number of currently popular distance indicators.
My collaborators and I have been pursuing this question by studying the Virgo cluster (Shields, Skillman, & Kennicutt 1991; Skillman et al. 1996). The Virgo cluster, due to its proximity, provides the logical starting point for the study of environmental effects on abundances in cluster spirals. The Virgo cluster is irregular, and in some respects this compromises the study of environmental effects. However, the extensive catalog of the positions, morphologies, and radial velocities of 1277 galaxies by Binggeli, Sandage, & Tammann (1985) allows a definition of the cluster's structure. In addition, the structure of the hot gas component, which has been revealed through sensitive mapping of the X-ray emission made possible by ROSAT (Böringer et al. 1994), is very similar to the galaxy distribution.
Substantial evidence points to differences between the spiral galaxies in the core of the Virgo cluster and field spiral galaxies. Davies & Lewis (1973) first pointed to a deficiency in HI gas in Virgo spirals when compared to field spirals. Giovanelli & Haynes (1985) showed that galaxies in the outer parts of the Virgo cluster (and eight other clusters) were less deficient in HI than the galaxies in the core of the cluster. From HI synthesis observations of Virgo cluster spirals, Warmels (1986) concluded the HI deficiencies could be attributed to a depletion of the HI at large galactocentric radii. This was supported by Cayatte et al. (1994, CKBG).
Here I will review the findings from observations of chemical abundances in HII regions in spiral galaxies of the Virgo cluster and a comparison of Virgo galaxies and field spirals (Skillman et al. 1996b). With these new data there now exist nine Virgo spirals with abundance measurements for at least four HII regions. Our sample of Virgo galaxies ranges from HI deficient objects near the core of the cluster to galaxies with normal HI properties, far from the cluster core. We looked at the relationship between the HI disk characteristics and the chemical abundances to determine whether dynamical processes that remove gas from the disk, such as ram pressure stripping by the intracluster medium, also affect the chemical abundances.
The radial HI surface profiles of the Virgo spirals are shown in Figure 21 (upper). The points represent the azimuthally averaged HI surface densities, while the solid lines represent the mean distributions for field spirals of the same morphological type (from CKBG). CKBG demonstrate that the radial HI surface density distribution of a field spiral galaxy is chiefly determined by its morphological type, if it is normalized to the optical size of the galaxy. We have normalized the observed profiles of the Virgo spirals in the same manner, using the DO diameters from the RC3. We have ordered the panels in Figure 21 in increasing DH / DO relative to the mean for the appropriate T-type, with the result that the galaxies divide into 3 deficiency groups: those with strong HI deficiencies, intermediate cases, and those with no HI deficiencies. Note the large variation in HI deficiency within the Virgo sample, especially at the extremes (e.g., compare NGC 4501 and NGC 4689 with NGC 4651).
Figure 21. (upper) A comparison of the radial HI profiles of the Virgo spiral galaxies in our sample with the average radial HI profile for the field galaxies of corresponding Hubble type as derived by Cayatte et al. (1994). The galaxies are ordered in increasing DH / DO ; the top row shows the HI-deficient (cluster core) galaxies and the bottom row shows the HI-normal (peripheral) galaxies. The individual points represent the radially averaged HI column densities, while the solid lines represent the average radial HI profile for the corresponding Hubble type. The filled circles represent the WSRT observations of Warmels (1988), while the filled triangles represent the VLA observations of Cayatte et al. (1990), and, in the case of NGC 4571, van der Hulst et al. (1987). (lower) Oxygen abundances from HII regions are plotted versus galactic radius for the nine Virgo spiral galaxies. The oxygen abundances are determined from measurements of [O II] and [O III] in individual HII regions and calibrated empirically as described in the text. The radial positions are normalized to the effective optical radius as defined in the RC3. The filled triangles are taken from Henry et al. (1992, 1994). (From Skillman et al. 1996b).
For comparison to field galaxy abundances, we have chosen the large sample of Zaritsky, Kennicutt, & Huchra (1994, ZKH). To ensure consistency among the abundances measured for the various field galaxies and the ZKH field galaxy sample, we have applied the same empirical R23 abundance calibration as used by ZKH for all 70 of the Virgo cluster HII regions. The derived oxygen abundances are plotted as a function of galactocentric radius (normalized to the effective radius, following McCall 1982) in Figure 21 (lower) and the ordering of the galaxies is identical to that used in the upper panel.
Comparing the HI data with the abundance patterns in Figure 21, we note that the peripheral galaxies have strong radial gradients in O/H and reach rather low values, 12 + log (O/H) 8.7, at the largest radii. In contrast, the three most HI deficient galaxies show high abundances with little evidence for radial gradients (but note that the radial range of HII regions is very limited in the HI deficient spirals).
This is further illustrated in Figure 22, which shows the characteristic O/H values and gradients for the Virgo galaxies as a function of DH / DO . In the top panel of Figure 22, a trend of decreasing O/H with increasing DH / DO is evident for the Virgo galaxies. It also shows that our grouping into three categories is somewhat arbitrary as all nine Virgo spirals display a continuum of decreasing mean O/H with decreasing HI deficiency. The bottom panel of Figure 22 illustrates the relationship between oxygen abundance gradient and HI deficiency. Here we tentatively see weak evidence for a trend of stronger gradient with decreasing HI deficiency (with the caveat that the radial sampling is small).
Figure 22. (upper) A plot of mean O/H as a function of DH / DO for the Virgo spirals. (lower) Oxygen abundance gradient versus DH / DO. The filled circles represent the HI deficient galaxies, the filled squares represent the intermediate galaxies, and the filled triangles represent the galaxies with normal HI disks. (From Skillman et al. 1996b)
The main result of the intra-Virgo comparison is that the three most HI deficient Virgo spirals have larger mean abundances (0.3 to 0.5 dex in O/H) than the spirals on the periphery of the cluster. This suggests that dynamical processes associated with the cluster environment are more important than cluster membership in determining the current chemical properties of spiral galaxies.
Next we compared the abundance properties of our Virgo sample to the large sample of field spirals studied by Zaritsky, Kennicutt, & Huchra (1994). Since the mean abundances of disks are systematically correlated with galaxy type, luminosity, and circular velocity it is important to check that the patterns in Figure 22 are not due to underlying variations in those properties. For example, many of the Virgo spirals are very luminous galaxies, and thus, they might be metal-rich objects solely on that basis, independent of cluster environment.
Following ZKH, we have fitted a linear relation in log O/H versus radius to the abundances of the individual HII regions to determine the "mean" value of O/H at a fiducial radius. Like ZKH, we use a fraction of the isophotal radius, 0.4 RO, as the fiducial radius. In Figure 23 we added the Virgo spirals to the field spirals from Figure 10 of ZKH. We excluded strongly barred spirals (RC3 "B" type) from the comparison because bars have been shown to affect the gradient slope (Pagel et al. 1979, VCE, ZKH, Martin & Roy 1994, Friedli, Benz, & Kennicutt 1994). In Figure 23(a) we plot the mean O/H as a function of MB. The HI deficient Virgo galaxies have higher oxygen abundances than the field galaxies. This is seen again in Figure 23(b) where the mean O/H value is plotted against VC and in Figure 23(c), where the galaxies are plotted as a function of T-type.
Figure 23. Mean O/H as a function of (a) absolute blue magnitude, (b) maximum rotation curve velocity, and (c) Hubble type for the Virgo spirals and the non-barred spirals of the ZKH sample. The open circles correspond to the field sample of ZKH, while the points for the Virgo spirals have been coded as in Figure 22. (From Skillman et al. 1996b).
The outstanding impression rendered by Figure 23 is that the three HI deficient Virgo spirals are all at the top of the abundance distributions of galaxies with similar properties. The dispersion in the properties of field galaxies and the small size of the Virgo sample make it difficult to draw definitive conclusions about any systematic difference between the field and Virgo spirals. Nevertheless, the HI deficient Virgo galaxies have larger mean abundances than field galaxies of comparable luminosity or Hubble type, while the spirals at the periphery of the cluster are indistinguishable from the field galaxies.
Having empirically established the abundance differential and provisionally quantified the magnitude of the effect, it is reasonable to ask if the observed effect is consistent with the physical processes known to be operating in the Virgo cluster environment. Our framework is that spiral galaxies are falling into the Virgo cluster (Tully & Shaya 1984), and that removal of gas from the infalling spirals by the cluster environment is primarily responsible for the HI deficiencies observed (Warmels 1986).
Simple, illustrative chemical evolution models with infall of metal-poor gas were constructed and compared to models in which the infall is terminated. The models were constrained by comparison with observed gas mass fractions, current star formation rates, and gas consumption times (e.g., Kennicutt, Tamblyn, & Congdon 1994). The model results indicated that the curtailment of infall of metal-poor gas onto cluster core spirals may explain part of the enhanced abundance. However, additional work is needed, particularly modeling of the effects of truncating the outer gaseous disk within the context of models with radial gas transport.
I think that this work has implications beyond a better understanding of the chemical evolution of field and cluster spiral galaxies. Foremost are implications for distance determination studies. The elevated abundances of HI deficient cluster spirals may impact both Tully-Fisher and Cepheid variable distance determinations. For example, if the dust content of spiral galaxies scales with abundance (van den Bergh & Pierce 1990), and the opacity corrections for galaxy luminosity are important (see Giovanelli et al. 1994 for a recent discussion), then there will be a systematic offset between cluster and field spiral galaxy Tully-Fisher relationships. Additionally, if the distances derived from Cepheid variable stars have a metallicity dependence (cf. Gould 1994), then there could be a systematic offset between distances derived for core cluster galaxies and field galaxies. It is also possible that chemical abundance studies will provide insight into the processes by which gas is removed from spiral galaxies as they enter the cluster environment. This may, in turn, lead to a more secure interpretation of current studies of the evolution of cluster galaxies (e.g, Dressler et al. 1994), and, thus, a better understanding of the chemical abundance patterns in both field and cluster spirals.