next up previous contents index
Next: Hot Luminous Stars Up: StarsStellar Populations, Previous: Globular Clustersand

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

HST Observations of Stellar Populations in Galaxies

Robert C. Kennicutt, Jr.
Steward Observatory, University of Arizona, Tucson, AZ 85721



This paper reviews recent HST observations of stellar populations in galaxies, with emphasis on resolved field stellar populations in nearby galaxies, young populations, and starbursts.

Keywords: stellar populations,galaxies,starbursts


Observations of stellar populations in nearby galaxies provide vital information on galactic evolution and the physical processes driving that evolution. In this review, I will concentrate on two areas where HST is making an especially strong impact, the study of resolved stellar populations in nearby galaxies, and the observations of young stellar populations and starbursts (globular clusters are reviewed by Renzini elsewhere in this volume, p. gif). I will highlight several recent results---mainly post-refurbishment observations---that illustrate the power and versatility of HST for attacking these problems.

Resolved Field Stellar Populations in the Local Group and Beyond

Observations of the resolved stellar populations in the nearest galaxies provide detailed fossil records of the history of star formation and chemical enrichment in galaxies of different types, masses, and environments. Most of this information comes from deep color-magnitude diagrams (CMDs), and since stellar crowding is the limiting factor in most situations, the impact of HST for such work has been revolutionary, extending the depth of ground-based studies by 4--5 magnitudes or more.

Large Magellanic Cloud

The Magellanic Clouds offer a unique opportunity to study the detailed history of star formation in extragalactic systems. Several ground-based studies of the LMC show evidence for a stellar birthrate history that is distinct from that in any component of the Galaxy, with the bulk of the star formation occurring in the past few Gyr (Bertelli et al. 1992 and references therein). The age distribution of the LMC clusters is even more unusual, with most clusters having ages of either 0--3Gyr or >12Gyr (Da Costa 1991).

To investigate the LMC population further the WFPC2 IDT has obtained deep V and I imaging of several fields. These data make it possible to study the luminosity function and CMDs to well below the main sequence turnoff (to and ), and infer the star formation history directly. The first results have been analyzed for a field in the outer disk near the cluster NGC 1866 (Gallagher et al. 1996). There is evidence for active star formation throughout the past 0.1--3Gyr, confirming the previous ground-based work. The data show evidence for a pronounced burst of star formation about 2Gyr ago, but no other evidence for large bursts or lulls in star formation over the past 6Gyr. Further interpretation will require analysis of other fields and a more thorough separation of age and metallicity effects. It is also interesting that the main sequence luminosity function in this field very closely matches that of the solar neighborhood, suggesting that the IMF in the LMC (for intermediate stellar masses) cannot be very different from that in the Galactic disk.

Figure: CMDs for halo fields in M31, obtained with WFPC2 by Rich et al. (1996). Fiducial giant branches are for Galactic globular clusters M15, 47 Tuc, and NGC 6553.

M31 and M33

At the distances of M31 and M33 HST imaging can reach the horizontal branch, making it possible to study the metallicity distributions of the old populations. One of the most exciting recent results has come from deep imaging of four fields in the M31 halo by Rich, Mighell, & Neill (1996). Figure 1 shows CMDs of these fields, which range in radius from 5.3 to 19.4 kpc. The horizontal branch is clearly visible, and blue horizontal branches are seen in the outermost two fields. Shown, for reference, are fiducial giant branches for the Galactic globular clusters M15, 47 Tuc, and NGC 6553, shifted to the distance modulus and reddening of M31, and spanning the range of metallicities observed in the Galaxy. The most interesting result is the apparent detection of a red metal-rich giant component in even the outermost fields, showing evidence of a solar-metallicity component at radii of >19 kpc in the M31 halo.

The same group has obtained WFPC2 imaging of the nuclear region of M33. The giants in this region show a very broad dispersion in color, corresponding to a metallicity range between M15 to at least 47 Tuc (Mighell & Rich 1995,1996). More information on the M33 population larger radii is expected from a study of the M33 globular cluster system by Sarajedini et al.

Dwarf Spheroidal Galaxies

Some of the most exciting new results in this field have come from deep WFPC2 observations of the dwarf spheroidal galaxies in the Local Group. Ground-based CMDs have revealed a wide range of star formation histories in these objects. A few appear to be predominantly old, classic Pop II systems, but most possess significant intermediate-age populations, as evidenced by the presence of carbon stars or a composite CMD morphology in the horizontal branch region (Da Costa 1992). With HST it is now possible to resolve the turnoff region and probe the star formation histories of these galaxies directly. Most of the dSph members of the Local Group have been targeted for observations in Cycles 4 or 5, and several studies are in progress. Here I briefly mention two examples, the Leo II and Leo I galaxies.

Deep color-magnitude diagrams of Leo II have been analyzed recently by Mighell & Rich (1996). The data extend to and , and show a well populated horizontal branch and a broad distribution of stars in the turnoff region (). The large vertical spread in the subgiant branch is consistent with the oldest stars forming 13--15Gyr ago, but with the bulk of the star formation occurring 8--10Gyr ago, and some star formation extending to 6--7Gyr ago. Hence, Leo II appears to be an example of a dwarf spheroidal galaxy with a dominant intermediate-age population, not unlike that of the Carina dwarf (Mighell 1990, Smecker-Hane et al. 1994).

Probably the most extraordinary star formation history revealed by HST belongs to the Leo I dwarf spheroidal galaxy (Mateo et al. 1996). The CMD of this galaxy is illustrated in Figure 2. Immediately striking is the abnormally luminous and red clump of evolved giants, which confirms ground-based observations by Lee et al. (1993). The absence of blue horizontal branch stars in this metal-poor system ([Fe/H] ) suggests the presence of a predominantly young to intermediate age population. Once again the depth of the WFPC2 data allow one to infer the star formation history directly, by comparing the distribution of stars in the turnoff region to theoretical isochrones. This is shown in Figure 2, which overlays metal-poor isochrones for 3 and 12Gyr on the CMD. From this comparison Mateo et al. conclude that at least 70% of the stars in Leo I are younger than 7Gyr, and that star formation proceeded relatively smoothly until roughly 2Gyr ago, perhaps even more recently. There is evidence from ultraviolet FOC imaging for stars as young as 1Gyr (Deharveng et. al., this volume, p. gif). The epoch of star formation was extended, and certainly not characterized as a burst.

Figure: CMD for the Leo I dwarf spheroidal galaxy, from WFPC2 observations by Mateo et al. (1996). Metal-poor isochrones for 3 and 12Gyr are superimposed.

Leo I thus represents a case of a dwarf galaxy which evolved over most of its lifetime like a spiral disk or an irregular galaxy, with an extended period of nearly constant star formation over several billion years, but which abruptly terminated 1--2Gyr ago. Today there is no evidence of star formation or substantial amounts of interstellar gas. How the star formation was terminated so abruptly, without any apparent evidence of a starburst, remains unanswered. These two examples illustrate the diversity of the star formation histories found in dwarf systems, and emphasize that galaxy mass is at least as strong a determinant of the evolutionary history as galaxy type. Results for several more dSph systems should be available in the next two years.

Beyond the Local Group

Beyond the Local Group the horizontal and subgiant branches fall below the WFPC2 detection/crowding limit, so studies of the sort described above cannot be carried out. However, the giant and supergiant populations remain detectable to distances of at least 10 and 100 Mpc, respectively, and can be used to characterize the massive young populations and the evolved giant populations. An illustration of the type of data that can be obtained is given in Figure 3, which shows the CMD of a field in NGC 925, kindly provided by Paul Harding. The data are based on a deep series of exposures (total integrations 26400 sec in V and 8800 sec in I) obtained to identify Cepheids as part of the H Cepheid Key Project (Silbermann et al. 1996). The most luminous stars are massive supergiants and unresolved star clusters, and the background sheet of AGB stars and red giants is evident near 25--26.

Figure: Deep CMD of a field in the disk of NGC 925, obtained as part of the Cepheid Key Project (Silbermann et al. 1996)

Figure: I-band luminosity function for NGC 3379, from Sakai et al. (1996). Note the abrupt jump at the red giant tip, and the flat slope of the luminosity function below the tip. The lower plot shows the result of an edge filter applied to the data, to precisely determine the location of the red giant tip.

Our team is using these data for a variety of applications. Hughes et al. (1994) and Bresolin et al. (1996) have used photometry of the M81 and M101 to the CMDs, luminosity function, and initial mass function of the young stellar population, and this work will be extended eventually to the entire Key Project sample. Bresolin is also using the data to inventory the population of OB associations and populous clusters in the spiral sample (next section).

As Wendy Freedman describes elsewhere in this volume, the tip of the red giant branch is detectable with WFPC2 to distances of >10 Mpc, and this makes it possible to determine distances to nearby galaxies independently of the Cepheid distance ladder. The method has been used by Sakai et al. (1996) to measure a distance to the giant elliptical galaxy NGC 3379. The sharpness of the red giant tip is illustrated in Figure 4, which shows the I-band luminosity function and the response of a derivative edge detector for NGC 3379 (Sakai et al. 1996). The bottom panel demonstrates that the discontinuity in the luminosity function can be determined with a very high degree of precision. The clear detection of the giant tip at a distance of 11.6 Mpc in 8 hours total exposure time implies that it should be detectable to uncrowded fields as distant as the Virgo and Fornax clusters, providing a powerful standard candle and a potential stellar populations diagnostic.

Young Stellar Populations and Starbursts

Observations of the youngest stellar populations provide important information on star formation rates in galaxies, the dependence of these rates on the physical conditions in the ISM, and the systematics of the mass functions of the star forming regions and the stars themselves. Since the high stellar densities in these regions make it difficult to resolve individual stars---even with HST---the primary impact of HST in this field has been to provide a high resolution ultraviolet capability and expand the reach of visible imaging and spectroscopic techniques to encompass the full range of galaxy types and star formation environments.

Resolved Studies of OB Associations and Clusters

The superb spatial resolution of HST makes it possible to resolve the nearest OB associations and star forming regions in the Local Group. The 30 Doradus complex in the LMC is the nearest example of a giant HII region, with over 2000 OB stars, making it a bona fide starburst region (Kennicutt & Chu 1994). Pre-refurbishment HST observations resolved the central core of 30 Dor, R136, into an extraordinary concentration of massive stars (Campbell et al. 1992, Malumuth & Heap 1994). Recently Hunter et al. (1995) have used deep WFPC2 observations in UVI to study the intermediate-mass stellar population and IMF of this region. Figure 5 shows a CMD with 2--5Myr isochrones superimposed. Stars below about 2.8 have not yet contracted on to the main sequence but the form of the IMF above this mass can be readily determined. Hunter et al. find that the function is well represented by a power law with slope = -2.220.06, very similar to the Salpeter slope of -2.35. Thus while the IMF may be slightly shallower than normal there is no evidence for a grossly different or bimodal IMF in these regions. IMFs similar to the Salpeter function were also derived for the OB association NGC 604 in M33 (Hunter et al. 1996) and NGC 206 in M31 (Hunter, private communication). Follow-up observations of the central 0.5 pc core of R136 itself does show evidence for a somewhat top-heavy IMF (), though the large uncertainties do not rule out a constant IMF (Hunter et al. 1996).

Figure: CMD of the R136 star cluster in 30 Dor, from Hunter et al. (1995).

Similar observations are being obtained of the resolved stellar populations in several more distant star forming galaxies, including the extreme dwarf irregular I Zw 18 (Hunter & Thronson 1995, Dufour et al., this volume, p. gif). This prototype of an ``isolated extragalactic HII region'' remains among the most metal-poor Pop I galaxies known, and an important calibrator of the primordial helium abundance. Ground-based observations have also suggested that it may be a rare example of a genuinely young galaxy (Loose & Thuan 1985, Kunth et al. 1995). Hence, understanding the stellar population of this object is relevant to a wide range of important problems. The CMD of this galaxy reveals a dominant population of stars ranging in age from 1Myr to of order 20--50Myr, consistent with other observational constraints (Kunth et al. 1995, Martin 1996). Thuan et al. (1996a,b) have carried out similar analyses of the blue compact dwarfs Mk 996 and SBS 0335-052.

In principle, ultraviolet imaging of galaxies should provide a very powerful method for studying the young populations and upper IMF in different galactic environments. Unfortunately, the poor UV efficiencies of the existing cameras and the limited availability of the FOC/48 camera has limited this application. The potential of this technique is aptly illustrated by the FOC observations of the gas-rich S0 galaxy NGC 5102, presented at this conference by Deharveng et al. (1996). This galaxy shows a anomalously UV-bright nucleus, and FOC imaging at F175W and F342W resolve a population of young stars, which they attribute to a burst of star formation that ceased 15Myr ago. The opportunities for this type of observation should improve dramatically with the anticipated installation of STIS and the Advanced Camera.

UV Spectroscopy, Imaging and Synthesis Modelling of Starbursts

Beyond the Local Group, severe crowding in the brightest star forming regions makes it impossible to resolve the individual stars, even with HST. However, the combination of UV imaging and high resolution spectra has made it possible to derive quantitative constraints on the stellar content, ages, and mass functions in the brightest starburst regions.

UV imaging of starburst regions has been carried out by several groups, and an excellent compilation of much of these data was recently published by Meurer et al. (1996). The data span a diverse range of parent galaxy types and, perhaps not surprisingly, the morphology of the starburst regions show a wide range of irregular structures. Much of the UV light, typically 20% of the total, is produced by compact star clusters. Perhaps the most interesting new result from this work is the apparent detection of a peak UV surface brightness in the regions, which may indicate some sort of negative feedback mechanism that limits the concentration of the massive star formation (Meurer et al. 1996).

Attempts to diagnose the stellar content of starbursts from modelling of the integrated stellar and spectral synthesis have been frustrated by degeneracies between the age distribution and mass functions; for example, when very massive stars are absent it is difficult to distinguish between an evolved population with a low turnoff mass and a genuinely truncated IMF (Kennicutt 1991, Rosa & Benvenuti 1994). However, it has recently been demonstrated that synthesis of the ultraviolet absorption line spectra can break these degeneracies, especially when spectra with sufficient S/N and resolution to observe the line profiles are obtained. The most widely applied method utilizes fitting of the P-Cygni profiles of the C IV 1550 and Si IV 1400 features (Robert et al. 1993, Leitherer et al. 1995).

Figure 6 shows an example of this technique applied to FOS observations of a bright HII region knot in the nearby irregular galaxy NGC 4214 (Leitherer et al. 1996). This galaxy has also been classified as a Wolf-Rayet galaxy due to the presence of strong W-R features in the optical spectra of the HII regions. The observed spectrum through a 1 arcsec FOS aperture is shown in the top panel. Many of the narrow absorption features are interstellar lines, which emphasizes the need for sufficient S/N and resolution to separate these from the stellar features. The two synthetic spectra below correspond to a continuous starburst with age 5Myr, mass limit 80 , = -2.35 and an instantaneous burst of age 4Myr, mass limit 80 , and = -3.00. Robert (1996) derived a similar fit to the spectrum of a bright knot in the interacting starburst galaxy NGC 3690. This technique is being applied to a wide range of starburst regions in other galaxies, and should eventually provide valuable constraints on the behavior of the stellar mass function in different environments.

Massive Star Clusters in Starbursts and Mergers

One of the most exciting discoveries made in this field with HST has been the detection of compact massive star clusters in the cores of starbursts and in strongly interacting and merging galaxies. Observations with WF/PC prior to the refurbishment mission revealed clusters in the cooling flow galaxy NGC 1275 (Holtzman et al. 1992) and the merger systems NGC 7252 (Whitmore et al. 1993) and NGC 4038/9 (Whitmore & Schweizer 1995). In NGC 7252, roughly 40 compact blue clusters were detected with an average and effective radius 10 pc. In NGC 4038/9, over 700 blue clusters were identified, with luminosities up to and average radius 18 pc (all for = 50). In this case, the cluster luminosity function could be derived, and it follows a power law with slope . Since the refurbishment mission, several other merger systems have been targeted for observation. One of the most impressive cluster systems has been detected in the merger remnant NGC 3256 (Zepf et al. 1996). In this galaxy over 15% of the total B luminosity is produced by a rich cluster system. Again the clusters fit a power-law luminosity function with and the brightest clusters reach .

Figure: FOS spectrum of a UV-bright starburst knot in NGC 4214 (top). The two bottom panels show synthesis models for continuous and instantaneous bursts, with parameters as given in the text. From Leitherer et al. (1996)

HST imaging of the nearby starburst galaxies have revealed instances of similar compact and luminous ``super star clusters,'' such as in NGC 1569 and NGC 1705 (O'Connell et al. 1994), NGC 1140 (Hunter et al. 1994), M100 (Freedman et al. 1994), and M82 (O'Connell et al. 1995). These clusters appear to be structurally similar to the core clusters of 30 Doradus and other nearby giant HII regions, and probably represent evolutionary precursors to the populous clusters that are observed in the LMC, M33, and other late-type nearby galaxies (Kennicutt & Chu 1988).

The observations of large numbers of compact luminous clusters in the merging galaxies has fueled speculation that these objects may represent proto-globular clusters---the precursors to the surviving globular clusters observed today in the bulge and halo of the Galaxy and other spheroids (Ashman & Zepf 1992, Whitmore & Schweizer 1995). The evidence is impressive but the possibility that one is observing scaled up versions of the young cluster populations in normal galaxies cannot be excluded entirely. The luminosity function observed for the compact clusters in the merger remnants is remarkably close to that observed for star forming regions in active star forming spiral and irregular galaxies (Kennicutt et al. 1989). More analysis of the cluster populations in both peculiar and normal galaxies is needed. A study of the Key Project sample aimed at characterizing the systematics of star clustering in normal spirals is being carried out by Fabio Bresolin for his Ph.D. thesis, and this should help to address these issues.

Concluding Remarks

Just as the deep HST observations of distant galaxies kindle hopes of reconstructing the evolutionary history of galaxies, deep observations of the resolved stellar populations in the nearest galaxies will provide an independent reconstruction of the history of star formation and chemical evolution of the Galactic neighborhood. Furthermore, HST observations of the youngest stars in nearby galaxies help to define the physical processes that regulate this evolution, and help to uncover the underlying physical basis of the Hubble sequence. The convergence of these independent probes into the formation and evolution of galaxies may eventually prove to be one of the most enduring legacies of HST.


I am very grateful to several individuals for providing preprints and preliminary results to present in this review, Fabio Bresolin, Peter Conti, Jean-Michel Deharveng, Rosa Delgado-Gonzales, Reggie Dufour, Jay Gallagher, Don Garnett, Paul Harding, Tim Heckman, Deidre Hunter, Claus Leitherer, Mario Mateo, Gerhardt Meurer, Ken Mighell, Jeremy Mould, Ed Olszewski, Mike Rich, Carmel Robert, Shoko Sakai, Ata Sarajedini, Trinh Thuan, Anne Turner, Brad Whitmore, and Steve Zepf. I acknowledge the support of NASA grant GO-2227.06-87A.


Ashman, K. M. & Zepf, S. E. 1992, ApJ, 384, 50

Bertelli, G., Mateo, M., Chiosi, C.,& Bressan, A. 1992, ApJ, 388, 400

Bresolin, F. et al. 1996, in preparation

Campbell, B. et al. 1992, AJ, 104, 1721

Da Costa, G. S. 1991, in The Magellanic Clouds, R. Haynes & D. Milne, Dordrecht: Kluwer, 183

Da Costa, G. S. 1992, in The Stellar Populations of Galaxies, ed. B. Barbuy & A. Renzini, Dordrecht: Kluwer

Deharveng, J.-M., Jedrzejewski, R., & Rocca-Volmerange, B. 1996, this volume, p. gif

Freedman, W. L. et al. 1994, ApJ, 435, L31

Gallagher, J. S. et al. 1996, ApJ, in press

Holtzman, J. A. et al. 1992, AJ, 103, 691

Hughes, S. M. G. et al. 1994, ApJ, 428, 143

Hunter, D. A., Baum, W. A., O'Neill, E., & Lynds, R. 1996, ApJ, in press

Hunter, D. A., O'Connell, R. W., & Gallagher, J. S. 1994, AJ, 108, 84

Hunter, D. A., O'Neill, E. J., Lynds, R., Shaya, E. J., & Groth, E. J. 1996, ApJ, in press

Hunter, D. A., Shaya, E. J., Holtzman, J. A., Light, R. M., O'Neill, E. J., & Lynds, R. 1995, ApJ, 448, 179

Hunter, D. A. & Thronson, H. A. 1996, ApJ, in press

Kennicutt, R. C. 1991, in Massive Stars in Starbursts, ed. C. Leitherer, T. M. Heckman, & C. A. Norman, Cambridge: Cambridge University Press, 157

Kennicutt, R. C. & Chu, Y.-H. 1988, AJ, 95, 720

Kennicutt, R. C. & Chu, Y.-H. 1994, in Violent Star Formation, ed. G. Tenorio-Tagle, Cambridge: Cambridge Univ. Press, 1

Kennicutt, R. C., Edgar, B. K., & Hodge, P. W. 1989, ApJ, 337, 761

Kunth, D., Matteuchi, F., & Marconi, G. 1995, A&A, 297, 634

Lee, M. J. et al. 1993, AJ, 106, 1420

Leitherer, C., Robert, C., & Heckman, T. M. 1995, ApJS, 99, 173

Leitherer, C., Vacca, W. D., Conti, P. S., Filippenko, A. V., Robert, C., & Sargent, W. L. W. 1996, ApJ, in press

Loose, H.-H. & Thuan, T. X. 1985, in Star Forming Dwarf Galaxies, ed. D. Kunth, T. X. Thuan, J. T. Thanh Van, Paris: Editions Frontieres, 73

Malumuth, E. M. & Heap, S. R. 1994, AJ, 107, 1054

Martin, C. L. 1996, ApJ, in press

Mateo, M. et al. 1996, in preparation

Mighell, K. J. 1990, A&AS, 82, 1

Mighell, K. J. 1995, AJ, 110, 1649

Mighell, K. J. & Rich, R. M. 1996, AJ, in press

Meurer, G. R., Heckman, T. M., Leitherer, C., Kinney, A., Robert, C., & Garnett, D. R. 1996, AJ, in press

O'Connell, R. W., Gallagher, J. S., & Hunter, D. A. 1994, ApJ, 433, 65

O'Connell, R. W., Gallagher, J. S., Hunter, D. A., & Colley, W. N. 1995, ApJ, 446, L1

Robert, C., Leitherer, C., & Heckman, T. M. 1993, ApJ, 418, 749

Rich, R. M., Mighell, K., & Neill, J. D. 1996, in preparation

Robert, C. 1996, in The Interplay Between Massive Star Formation, the ISM, and Galactic Evolution, ed. D. Kunth et al., Paris: Editions Frontieres, in press

Rosa, M. R. & Benvenuti, P. 1994, A&A, 291, 1

Sakai, S., Madore, B. F., Freedman, W. L., Lauer, T. R., Ajhar, E. A., & Baum, W. A. 1996, in preparation

Silbermann, N. A. 1996, in preparation

Smecker-Hane, T. A., Stetson, P. B., Hesser, J. E., & Lehnert, M. D. 1994, AJ, 108, 507

Thuan, T., Izotov, Y. I., & Lipovetsky, V. A. 1996, ApJ, in press

Thuan, T., Izotov, Y. I., & Lipovetsky, V. A. 1996, in The Interplay Between Massive Star Formation, the ISM, and Galactic Evolution, ed. D. Kunth et al., Paris: Editions Frontieres, in press

Whitmore, B. C. & Schweizer, F. 1995, AJ, 109, 960

Whitmore, B. C., Schweizer, F., Leitherer, C., Borne, K., & Robert, C. 1993, AJ, 106, 1354

Zepf, S., Ashman, K., English, J., Freeman, K., & Sharples, R. 1996, in preparation

next up previous contents index
Next: Hot Luminous Stars Up: StarsStellar Populations, Previous: Globular Clustersand