next up previous contents index
Next: The Host Galaxies Up: QSOsGalaxies, and Previous: Quasar Absorption Lines

Science with the Hubble Space Telescope -- II
Book Editors: P. Benvenuti, F. D. Macchetto, and E. J. Schreier
Electronic Editor: H. Payne

Elliptical Galaxies: The Impact of HST

Garth D. Illingworth
UCO/Lick Observatory and Board of Astronomy and Astrophysics
University of California, Santa Cruz, CA 95064 USA



HST data are providing new insights into the nature of elliptical galaxies, their relationship to S0 galaxies, and their evolutionary history. In particular, the characteristics of the stellar population, the nature of the globular cluster systems, the structure of the cores and the dust/gas component, and the size/scale of massive dark objects or galactic black holes are all aspects of early-type galaxies that are being studied by HST. Some recent results are summarized here. In addition, there is the serendipitous component of these studies of cores and nuclei that is particularly exciting. The inner parts of galaxies, with their deep potential wells and short timescales, are host to fascinating physical and dynamical processes. This is new and exciting territory, the exploration of which is still in its infancy.

Keywords: ellipticals, early-type galaxies, accretion, mergers, dust, evolution


The richness of the phenomena that we find in elliptical galaxies continues to surprise, in part because this richness contrasts so dramatically with the impression of simplicity conveyed by images of elliptical galaxies. The complexity of these systems has been emphasized once again with recent results from HST. As we look at the HST WFPC2 images of the inner regions of elliptical galaxies, and at the ubiquitousness of dust/gas in those regions, one is left wondering whether there are any classical ellipticals! It may be that our classical separation into E and S0 is due for replacement by a division into ``boxy'' and ``disky'' groupings, with their associated physical kinematical distinctions, even though such effects are quite subtle in many cases.

That notwithstanding, our future understanding of the nature of early-type galaxies, and their formation and evolutionary history are inexorably linked to HST. HST currently provides the only means of studying the stellar and globular cluster populations to faint levels over large areas, as well as providing the only means of establishing the characteristics of the core regions on sub-arcsec scales. A number of recent results from HST highlight the value and the potential of HST imaging and spectroscopy, and whet our appetite for the substantial enhancements in capability expected from the three new instruments, STIS, NICMOS and the ACS. The recent results that will be mentioned here are in the areas of stellar population characteristics, the properties of the globular cluster systems, and the structure of cores. In addition, there is a serendipitous aspect of imaging a sample of early-type galaxies, since our knowledge of the inner regions (--) of galaxies is very fragmentary and incomplete. Given that such cores are deep potential wells, occupied presumably, in many cases, by black holes, and clearly the site of much energetic phenomena, their properties are of considerable interest.

Stellar Populations

The resolution and sensitivity of HST with WFPC2 to unresolved sources make it the most powerful telescope available today for studying the evolved stellar populations directly in nearby early-type galaxies (ellipticals and disk galaxies with ``old'' bulges and disk components). Important characteristics of the intrinsic population properties like mean metallicity and mean age can be determined; estimates of the distribution of such properties can also be made. In addition, color magnitude diagrams and luminosity functions can be used to derive distances.

Figure: Position of the WFPC2 field in the halo of NGC 5128 from the WFPC2 team study of the halo stellar population (Soria et al. 1996). The image is from the STScI Digitized Sky Survey; the field size is .

A recent paper from the WFPC2 team exemplifies what can be done with HST, even with relatively short integrations. Four F555W images (3200 s total) and four F814W images (3800 s total) were taken of NGC 5128 (Centaurus A) and used to derive a color-magnitude diagram (Soria et al. 1996) of about 10,000 stars. The field location in the halo of NGC 5128 is shown in Figure 1. They detect an intermediate age population (Gyrs), but the fraction of stars at to () indicates that this intermediate age population is only about 10% of the total. From the color distribution at constant I mag they derive a mean [Fe/H] dex, higher than that seen in the dwarf elliptical NGC 185 and probably comparable to that in the halo of M31. There are indications as well, from the width of the color distribution, that the metallicity range is also large (like M31).

The change in slope in the luminosity function at mag enables them to identify the tip of the red giant branch quite accurately (error mag). Comparison with C-M diagrams for globular clusters and the NGC 185 allows them to obtain a distance modulus of , or a distance of Mpc, in good agreement with determinations based on surface brightness fluctuations and the planetary nebulae luminosity function.

Figure: Distributions of globular cluster V-I colors for the elliptical galaxies in the Forbes et al. (1996) sample. The mean color for the combined sample is , corresponding to a mean [Fe/H]. The indications of bimodality in IC 1459 and NGC 4494 are not statistically-significant; higher S/N data are needed to verify the reality of the apparent non-Gaussian structure.

Globular Clusters

The effectiveness of HST's imaging capability for studying the distribution and colors of globular clusters in elliptical galaxies is probably even more striking than its utility for stellar population studies. At small radii, where the galaxy surface brightness is quite high, the detectability of clusters with HST is far beyond what is possible with ground-based telescopes.

The parameters that can be derived from HST images for the globular cluster population are the surface density distribution, and the derived length scales and core size, the cluster colors, and the cluster luminosity function, as well as the associated radial variation of these properties. From the colors, the mean metallicity and metallicity distribution can be inferred, as can their radial variation (given adequate sample size), while the luminosity functions are potentially useful for distance estimation. The differential properties (clusters vs. galaxy) in the colors (and hence metallicities), surface density distribution, length scales and core properties are also very important constraints on the nature of the formation process for both the globular cluster population and the parent galaxy, and on the relative timescales of their formation and evolution.

Figure: Mean globular cluster metallicity vs. absolute V magnitude of the parent galaxy. The galaxies include dwarf ellipticals, the SMC and the LMC, bulges of nearby spirals and giant ellipticals. The weighted fit is [Fe/H] = 0.17M (the unweighted fit is similar). Circles and squares are photometrically-derived [Fe/H] and spectroscopically-derived [Fe/H], respectively. Filled and open symbols are different samples (see Forbes et al. 1996). The small offset between the photometrically and spectroscopically derived samples may not be significant; the errors are internal, and likely to be unrealistically small.

A recent paper by Forbes et al. (1996) demonstrates the value of HST for defining the characteristics of the globular cluster population in a sample of nearby ellipticals (see also Ajhar et al. 1996 for a WFPC2 study of several globular clusters in M31). This sample of 14 ellipticals with kinematically-distinct cores was imaged in F814W and F555W primarily to study their core properties. The integrations were short, typically 1000 s total in F555W and 460 s total in F814W, with the nucleus centered on the PC CCD, as appropriate for a study of the core properties. Yet the images proved to be extremely valuable for globular cluster measurements as well. Typically 200 globulars could be measured per galaxy (the range was 39--328), to about after the galaxy was modelled and subtracted. Colors were determined over a radial range that was typically more than 1 dex (usually 0.5--5+ kpc; km s Mpc).

The globular cluster systems in these galaxies have several properties: i) the mean color of the globulars is bluer than that of the galaxy, corresponding to a mean metallicity decrement of [Fe/H] dex; ii) there is no significant variation of the mean color with radius; iii) the mean colors are very similar from galaxy to galaxy, and show no statistically-significant deviations from unimodal distributions in color (Figure 2); iv) the mean cluster metallicity is found to correlate with galaxy luminosity as , similar to that of the galaxy population, over a range of nearly 10 mag in luminosity (Figure 3); iv) the globular cluster systems are, in general, aligned with but slightly rounder than the galaxy; and v) the globular cluster density distributions have a distinct core, the radius of which is weakly correlated with luminosity.

Figure: Luminosity density profiles for 42 galaxies (Gebhardt et al. 1996--- their Figure 3). Luminosity densities were derived by inversion of the surface brightness profiles in the Lauer et al. (1995) sample and scaled to the radius and luminosity at the radius of maximum curvature in the surface brightness.

The significance of these results from such short integrations indicates that future studies with higher S/N will contribute very substantially to our knowledge of the nature of globular cluster systems in galaxies, and their use as diagnostics of the history of early-type galaxies.

Core Structure

While studies of the stellar populations and globular clusters in early-type galaxies did not begin to flourish until the aberration-corrected WFPC2 became available, a number of very important studies were made of the core properties based on deconvolution of the aberrated WF/PC and FOC data (see, e.g., Lauer et al. 1992, Crane et al. 1993, Jaffe et al. 1994, Ferrarese et al. 1994, Forbes et al. 1995, and the comprehensive study by Lauer et al. 1995). The early work with HST in this area is discussed by Kormendy & Richstone (1995) in their Annual Reviews article on black hole searches in galaxy nuclei. A very comprehensive study of the dust properties of early-type galaxies was carried out by van Dokkum & Franx (1995), in which they utilized archival WF/PC data for 82 galaxies. References to earlier studies can also be found in that paper. Even with the aberrations, the pre-refurbishment HST data has allowed characterization of the generic dust (and gas) properties of early-type galaxies, as well as of the surface brightness profiles in the cores.

For example, van Dokkum & Franx (1995) detected dust in 31 of 64 galaxies (those of the sample of 82 with adequate S/N images), and inferred that % of early-type galaxies contain nuclear dust. The dust+gas masses range from to . The actual detection rate in radio-loud galaxies (72%) is double that in radio-quiet galaxies (33%), and the dust is usually oriented orthogonal to the radio axis. The dust appears generally not to be settled, even on small scales (r <250 pc), nor does the dust (and its associated gas) ``rotation'' axis generally coincide with that of the stars.

Figure: Deconvolved image of the central region of the cluster A2199 from Lauer et al. (1995). The large, low surface brightness object with the dust lane is the central cD galaxy NGC 6166. The contrast with the compact, high surface brightness cores of the other, lower-luminosity, cluster members is quite striking. Since such dense, high-surface brightness galaxies must sometimes be accreted, and would survive in the core of the cD, this raises interesting issues about the dynamical processes (destruction) that must accompany and follow such accretion events (see, e.g., Lauer et al. 1995 and Faber et al. 1996)

With surface brightness profiles from a very large sample of ellipticals now available from WF/PC and WFPC2, the generic properties of cores are beginning to be defined. One of the surprising results is that essentially all early-type galaxies studied to date have surface brightness profiles that have non-zero slopes at radii , to the limit of HST's resolution (in the PC). The logarithmic gradients in , where is the surface brightness, fall into two ranges, and .5--1. Two classes of objects appear to be represented by these non-overlapping ranges of values (see, e.g., Lauer et al. \ 1995, where the core structure in a sample of 45 early-type galaxies is discussed; also Forbes et al. 1995). This is shown very dramatically in the luminosity density profiles determined by Gebhardt et al. (1996) for 42 galaxies in the Lauer et al. (1995) sample. Gebhardt et al. find that virtually all have significantly non-zero slopes at small radii (). For only two of the 42 could the logarithmic gradient be consistent with zero slope. The luminosity density profiles from Gebhardt et al. (1996) are shown in Figure 4.

While this result suggests that the separation into two groups seems to be a characteristic property of early-type galaxies, Carollo et al. (1996) note that the values in their sample of 15 kinematically-distinct galaxies from WFPC2 images do not show such a clear separation into two groups. Interestingly (see, e.g., Carollo et al. 1996 and Faber et al. \ 1996), these core profiles are correlated with global properties; large, luminous, ``boxy'', non-rotating, anisotropic systems have smaller slopes to their cores and a distinct break in their surface brightness distributions (i.e., they are ``cuspy'' in the Lauer/Gebhardt definition), while the small, lower luminosity, ``disky'', high rotation, isotropic galaxies/bulges are those with steeper slopes with small or minimal break (the ``power-law'' profiles). It will be interesting to see if such a dichotomy is retained as larger samples of WFPC2 data on early-type galaxies are analyzed.

Figure: The cores of two elliptical galaxies, NGC 4278 and NGC 5813, from model-subtracted PC F555W ()images taken with WFPC2 (from Forbes et al. 1996 and Carollo et al. 1996). The region shown is about in extent. The dust (dark) structures are complex and unsettled. The unresolved objects are globular clusters in the inner region of these ellipticals.

Visually some of these distinctions can be seen in the striking image of the core of the cluster A2199 in which the low central surface brightness, large core of the central cD galaxy NGC 6166 can be seen surrounded by several high surface brightness, compact neighbors (see Figure 5---from Figure 2 of Lauer et al. 1995). The core regions of two ellipticals imaged with the PC of WFPC2 are also shown in Figure 6. The dust structures and globular clusters can be clearly seen after subtraction of the galaxy (from Forbes et al. 1996 and Carollo et al. 1996).

Additional interesting results on elliptical cores have arisen from these data. Nuclear star clusters (with luminosities -- ) appear to be commonly seen in the ``power-law'' galaxies (Lauer et al. 1995). Furthermore, about 50% of the sample of ellipticals with kinematically-distinct cores imaged with WFPC2 by Carollo et al. (1996) show photometric evidence for stellar disks. The lack of any measurable V-I color difference between these disks and the surrounding nuclear regions sets an upper limit to the age difference between these populations; it must be . In one case, NGC 4365, the inner core has a region about 15 pc in size that is bluer than the surrounding region, corresponding to an age difference of --5Gyr, or a small metallicity change of [Fe/H] dex.

The dynamical implications of these core structures are only just now beginning to be considered (see, e.g., Faber et al. 1996), but a key aspect of the theoretical development will be the existence and scale of any massive dark objects/black holes that reside in the nuclei. Kormendy & Richstone (1995) discuss in considerable detail the history, issues and implications of these new results, and the importance of future spectroscopy with HST.

Figure: WFPC2 PC images of three early-type galaxies and the residual maps after subtraction of a mean galaxy surface brightness distribution (see Phillips et al. 1996b; see also Phillips et al. 1996a for discussion of the procedures used). These examples are from a ``Snapshot'' survey of nearby galaxy nuclei. The dust distributions are more regular in these examples than many that are seen in other early-type galaxies (see, e.g., Figure 6 and Carollo et al. 1996).

Cores in Nearby Galaxies

The WF/PC and WFPC2 results on elliptical galaxies are being broadened now to include samples of nearby galaxies that extend across the whole Hubble sequence. These samples include many more S0 and Sa galaxies. Such surveys are rather more serendipitous in nature, particularly the ``Snapshot'' surveys from Cycle 3 (Phillips et al. 1996a) and Cycle 4 (Phillips et al. \ 1996b). In this latter program, WFPC2 PC images were taken of over 200 nearby, S0-Irr galaxies at recession velocities V<1000 km s, plus some in the Virgo and Fornax clusters. From this short-integration, gyro-pointed sample, some 130 galaxies are adequately centered in the PC and of high enough S/N to define structure and establish dust morphology. The core regions of a few early-type galaxies in the Cycle 4 sample are shown in Figure 7 (see Phillips et al. 1996b). A more limited Cycle 5 ``Snapshot'' survey is being made of a sample of 40 S0 galaxies with WFPC2 images being taken with two filters, F555W and F814W. As the Forbes et al. (1996) and Carollo et al. (1996) results indicate for ellipticals, the availability of color information adds a very valuable dimension to the program. In particular, corrections can be made for extinction to improve the reliability of the core profiles (see Carollo et al. 1996). Preliminary results already show some striking dust structures and core properties, and it is clear that the S0 sample will be as interesting as the elliptical sample studied to date.

HST and Future Observations

We are still exploring new territory when it comes to understanding the cores of nearby galaxies. The first WFPC2 results are now being published, and consideration given to the interesting dynamical issues raised by these data, in addition to their implications for the mass distribution in these deep potential wells. In particular, questions concerning the formation mechanisms for cores and their associated massive dark objects, the dynamical evolution of the cores and the role played by massive dark objects/black holes, and the associated timescales for that evolution, will be central to much theoretical work on galactic nuclei in the coming years. Observationally, HST's role will become even more important with the advent of the next generation of instruments: STIS for its crucial long-slit spectroscopic capability; NICMOS, for its ability to minimize the effect of dust on the determination of core structure; and the Advanced Camera, the ACS, for its well-sampled, high spatial resolution imaging capability (combining the best of both the FOC and the WFPC2) with the further advantages of extensive narrow-band filters and much higher throughput than the current instruments. Scientifically, we are reaching out towards understanding the centers of galaxies, and clarifying the relationship of AGNs to the cores and nuclei of ``normal'' galaxies. HST is key to the resolution of these issues.


I would like to acknowledge the contributions from a number of groups (cited in the text), as well as very interesting discussions with Sandra Faber. Funding for this work was provided by NASA grants GO-3551.01-91A and GO-4644.01-92A.


Ajhar, E. A. et al. 1996, AJ, 111, 1110

Carollo, C. M., Franx, M., Illingworth, G. D., & Forbes, D. A. 1996, ApJ, in press

Crane, P. et al. 1993, AJ, 106, 1371

Faber, S. M. et al. 1996, in preparation

Ferrarese, L., van den Bosch, F. C., Ford, H. C., Jaffe, W., & O'Connell, R. W. 1994, AJ, 108, 1598

Forbes, D. A., Franx, M., & Illingworth, G. D. 1995, AJ, 109, 1988

Forbes, D. A., Franx, M., Illingworth, G. D., & Carollo, C. M. 1996, ApJ, in press

Gebhardt, K. et al. 1996, AJ, in press

Jaffe, W., Ford, H. C., O'Connell, R. W., van den Bosch, F. C., & Ferrarese, L. 1994, AJ, 108, 1567

Kormendy, J. & Richstone, D. 1995, ARA&A, 33, 581

Lauer, T. R. et al. 1992, AJ, 103, 703

Lauer, T. R. et al. 1995, AJ, 110, 2622

Phillips, A. C., Illingworth, G. D., MacKenty, J. W., & Franx, M. 1996a, AJ, in press

Phillips, A. C. et al. 1996b, AJ, in preparation

Soria, R. et al. 1996, ApJ, June 20

van Dokkum, P. G. & Franx, M. 1995, AJ, 110, 2027

next up previous contents index
Next: The Host Galaxies Up: QSOsGalaxies, and Previous: Quasar Absorption Lines