Chapter 2

Chapter 2, The Science Case

Section 2.3: Resolved Stellar Populations

NOAO Logo    Gemini Logo


Over the past 100 years, the study of stellar populations in nearby galaxies has radically changed our outlook on the universe. At the turn of the last century, the Kapteyn model held that the Sun lies close to the center of the Galaxy, then considered to be a flattened system of stars ~ 10 kpc in diameter and comprising the entire universe. Shapley's study of the globular cluster system of the Galaxy in the early 1900s showed that the Sun in fact lies several kiloparsecs from the Galactic center (GC), which in Copernican fashion removed the Sun from a special place in the universe. In the 1920s, Hubble's use of Cepheid variables in nearby galaxies to extend the distance ladder established that the galactic nebulae are independent "island universes," a view that dramatically changed the scale and concept of the universe.

In modern times, the study of stellar populations has continued to advance our understanding of the universe we live in. It has been critical in establishing that luminous matter accounts for only ~ 10% of the mass of the universe. In addition, gravitational microlensing studies of stars in the Large Magellanic Cloud suggests that at least half of the dark matter in the universe consists of something other than ordinary matter. All of these breakthroughs were accomplished with the use of ground-based, seeing-limited telescopes. Future breakthroughs will depend on the availability of more powerful telescopes and instruments. Over the last decade, the success of the Hubble Space Telescope (HST) has shown that increasing telescope resolution is a powerful capability for the study of stellar populations. With the HST, we were able to detect white dwarf cooling sequences in globular clusters; pick out recently formed massive clusters in interacting galaxies such as the Antennae; and resolve the oldest main sequence stars in nearby dwarf galaxies, which revealed that these galaxies have complex histories involving multiple episodes of star formation. None of this work could have been done under seeing-limited conditions, because the crowding limit in these stellar systems is severe.

The GSMT (Giant Segmented Mirror Telescope), with the leap in angular resolution and light gathering power that it offers, is expected to make similar advances through its capability for studies of resolved stellar populations in galaxies beyond the Milky Way. These studies will provide critical complementary evidence with which to reconstruct the formation and evolutionary history of galaxies. Although the studies described in the previous section use ensembles of galaxies spread out over cosmic time to reconstruct history, we can also analyze the ensemble properties of resolved stars in order to recover the historical record of individual galaxies. Given that stellar kinematics, ages, and metallicities can be measured for individual stars, correlations between these quantities and the existence of clustered subpopulations will reveal the star formation, mass assembly, and chemical enrichment history of the galaxy.

In the case of the Milky Way, for example, recent studies have uncovered the existence of a dwarf galaxy that is currently merging with our Galaxy (the Sagittarius Dwarf),1 thereby providing definitive evidence that mergers have contributed to the formation of the Milky Way. The possible existence of moving groups in the Galactic halo indicates that multiple mergers may have contributed to the formation of the Galaxy.2 Correlations between stellar abundance ratios, such as alpha/Fe vs. Fe/H, are being used to understand the interplay between star formation and Galactic gas dynamics (infall, ejection, and mixing) that have driven the chemical enrichment history of the Galaxy.

Although in the coming decade we expect to use these techniques on existing telescopes to obtain a more complete understanding of the formation history of the Milky Way, the challenge of future decades will be to understand whether the formation history of the Milky Way thus obtained is representative of the histories of other galaxies, i.e., of galaxies spanning a range of masses and morphologies. We already have tantalizing hints of diversity in galaxy formation histories that suggest the need for studies of galaxies beyond the Milky Way. For example, the metal-rich M31 halo3 suggests a different evolutionary history for M31, a galaxy that appears otherwise quite similar to our own.

The GSMT will provide the powerful combination of unprecedented collecting area and unprecedented angular resolution needed to carry out these studies. By equipping the GSMT with innovations in instrumentation, specifically multi-conjugate adaptive optics (MCAO), coupled with large format infrared arrays and advanced integral field units (IFUs), we will be able to carry out diffraction-limited imaging and spectroscopy over fields of ~ 1'. With these advances, we will be able to probe the star formation, merging, and chemical enrichment histories of galaxies in dense environments (e.g., the GC and the inner regions of Local Group galaxies) where confusion restricts current studies to only the most luminous stars. In less crowded regions (e.g., outskirts of galaxies and intracluster regions), similar observations can be carried out at the distance of the Virgo cluster and beyond.

Although excellent image quality has been available on space-based platforms (e.g., HST) for some time, recent developments with ground-based adaptive optics AO are now beginning to produce exciting results that are scientifically competitive with those obtained in space. Because ground-based AO observations are only now being applied to the study of stellar populations, the bright future for this field may not have received much attention thus far. To address this, in Section 2.3.2 we describe the current state-of-the-art in the study of resolved stellar populations with ground-based 8-10-m telescopes. In Section 2.3.3, we use simulated data to predict the performance of GSMT for studies of resolved stellar populations.


The University of Hawaii AO module (Hokupa'a) and infrared camera (QUIRC) deployed on the Gemini North telescope since early 2000 are producing high quality images in the near-IR. This is particularly encouraging because the data have already yielded exciting results, yet they have been obtained with a system that is known to be less than fully optimized. (Because Hokupa'a was developed for the UH 88" telescope, it has fewer actuators than would normally be deployed on an 8-m telescope.) A fully optimized system is expected to produce even more spectacular results. In 2000, Gemini sponsored a demonstration science program, featuring Hokupa'a, that was centered around stellar populations in the GC. Images of the central few parsecs of the Galaxy were obtained in H, K, and additional narrow bands. H and K images of the mini-starburst Arches cluster, which lies ~ 30 pc in projection to the northeast of the GC, were also obtained (see Figure 1). The scientific results that have emerged from these data are described briefly below.

Star Formation in the Galactic Center: Previous work4,5 has established the existence of a very young stellar component (< 10 Myr) in the central parsec of the GC via the detection of the energetic near-infrared emission lines that signal the presence of the most massive members of this population (~ 100Msolar). Recent efforts to further characterize this young population have been sensitive only to the very massive evolved stars, and have not resulted in the identification of the lower mass main sequence OB stars. This population has now been identified with Gemini Demo Science data obtained in narrow-band filters centered on the CO bandhead (2.3 µm) and the nearby continuum.6 Figure 2 shows the (CO continuum) index for the central 20", which includes the well-known emission line star cluster IRS 16.7 Two sequences are clearly identified. The older population has the steadily increasing CO strength expected for AGB stars and M supergiants, while the young sequence has the approximately constant (CO continuum) color, which reflects only the continuum slope. By comparing these results with observations of a control field, we can obtain a statistical estimate of the total number of stars associated with the young sequence (see Figure 2).

Figure 1   An image of the cluster in the GC obtained in a narrow-band filter centered on the 2.3 um CO bandhead (left) and a K image of the Arches cluster (right), both obtained as part of the Gemini demonstration science program. The Arches cluster shows less isoplanatism because the guide star was located in the field. Both images are 20 arcsec = 0.8 pc on a side.

The IMF in the Arches Cluster: Like its better-known counterpart in the LMC, R136, the Arches cluster is a local example of a mini-starburst. Its high mass(>104Msolar), large number of high mass stars (~ 150 O stars), and high density (~ 10 times as dense as R136)8 make it a touchstone for our understanding of the cluster mode of star formation in the nuclear regions of galaxies. Using the Gemini Demo Science data, Blum et al.6 have measured the mass function of the Arches. They find a mass function slope (Gamma = -0.93 ± 0.1; see Figure 3) that is flatter than Salpeter (Gamma = -1.35), but steeper than the mass function obtained by Figer et al.8 (Gamma = -0.7 ± 0.1) for the same region from HST/NICMOS (near-infrared camera and multi-object spectrometer) data. Compared to the NICMOS data, the Gemini AO data exhibit a better-behaved PSF (point spread function) and provide superior image quality. As a result, the mass function obtained from the Gemini Demo Science data is more complete, reaching lower masses (~ 2Msolar) than were reached with NICMOS (> 5Msolar). The flat IMF (initial mass function) slope is consistent with the expectation of Serabyn and Morris,9 who suggested that star formation may be naturally skewed toward higher masses in the inner Galaxy due to the higher surface density, temperature, magnetic field strengths, and turbulent velocity in the molecular clouds within 200 pc of the GC.

Figure 2  The (CO-continuum) index for the central 20 arcsec of the GC field (left). The black sequence represents the lower mass main sequence OB stars associated with a recent burst of star formation in the GC. Comparison with a control field 20 arcsec north of the GC (right). The red control field stars match the same population in the GC, consisting of older AGB stars. Because significant anisoplanatism, due to the use of a single off-axis guide star, is present in the data, the use of narrow-band filters is a significant advantage. The closely spaced filters have PSF residuals that are similar as a function of position in the frame. As a result, the CO-continuum index cancels the effect of the spatially varying residuals when the photometry is extracted using an average PSF.

Figure 3 The mass function for the Arches cluster obtained from the Gemini Demo Science data. The Hokupa'a/QUIRC image quality (0.12 arcsec FWHM at K) allows for a deeper mass function determination than did previous work using HST.

Is the flatter IMF a reflection of the initial conditions for star formation in the vicinity of the GC, as hypothesized by Serabyn & Morris,9 or is it the result of dynamical evolution, i.e., mass segregation? Recent dynamical evolution models for the Arches cluster indicate that significant mass segregation is possible.10 A complete census of the GC population, extending beyond the region currently surveyed, is needed to distinguish between these two possibilities.


The studies described above are representative of the current frontier of (ground-based?) stellar populations research. In all of the studies, the depth to which the frontier can be pushed is limited by confusion, in the case of imaging, and photon flux, in the case of spectroscopy. The tremendous impact that a diffraction-limited GSMT will have on stellar populations research is, therefore, a consequence of both its resolution and its light gathering power, each of which gains as D2, where D is the telescope diameter. The GSMT will be able to image and take spectra of stars three magnitudes fainter than is possible with Gemini S + MCAO. We have used simulated data to explore the consequences of this ability.

GSMT Studies of Mini-Starburst Clusters: The GSMT will be able to carry out detailed studies of mini-starburst clusters, located in the Milky Way and other Local Group galaxies that span a range of metallicity, age, and galactic environment. The results for these templates will be invaluable in interpreting extra-galactic starburst phenomena. Significant developments are expected, even in advance of the GSMT, with the advent of multi-conjugate adaptive optics (MCAO) on Gemini South, which will allow diffraction-limited observations within the next three years. For example, this capability will push the crowding limit in the GC down by 1.5 magnitudes relative to the Gemini North data discussed above. In the Arches cluster, the mass detection limit will be lowered to ~1Msolar.

A diffraction-limited GSMT will further reduce the crowding limit by three magnitudes, allowing complete sampling of the IMF to near the hydrogen burning limit (0.35Msolar) in the Arches cluster. Similar gains are expected for R136 in the LMC, where the larger distance compared to Arches is compensated for by a less dense cluster core. While current HST data reach about 2Msolar at the crowding limit,11 GSMT observations of R136 will be able to extend the measurement of the mass function to well below a solar mass. Another local massive cluster for which similarly detailed data could be obtained is NGC3603 (6.5 kpc distant), which is less massive than R136 and less dense than the Arches. Thus, with the GSMT, we expect to be able to measure the stellar (and perhaps substellar) mass function for nearby starburst-like clusters spanning a range of metallicity, age, and galactic environment. With the diffraction limit of the GSMT (0.015" FWHM at K), we will also be able to study massive young clusters in more distant galaxies. It will be possible to resolve individual bright stars and thus study the upper end of the mass function in a wide range of star formation environments in the Local Group and beyond. We present below the results of a simulation in which stellar photometry is obtained for Arches-like clusters in M33.

The Central Parsec: In the central parsec of the Milky Way, the GSMT will be able to zero in on the accretion source surrounding the nuclear black hole. An unambiguous near-infrared counterpart to the associated radio source, SgrA*, could be identified. The putative accretion flow (perhaps with a disk) is anomolously faint and current models advect most of the inflowing matter and energy into the black hole where it vanishes unseen (see the review by Melia & Falcke).12 Current limits on the near-infrared source put it below the crowding limit (K ~ 17), but current theoretical models allow a huge range of near-infrared emission from the accretion flow, as low as K ~ 24! Constraining the brightness of SgrA*, or detecting it, will provide powerful constraints on the models. Only the angular resolution of GSMT can lower the observed limit, which is currently set by the crowding limit for Keck and Gemini.

Recent theoretical predictions on microlensing in the GC suggest that lensing by SgrA* of background bulge stars could in turn be used as a probe of stellar remant blackholes, which should accumulate in great number due to dynamical friction over the life of the Galaxy. Such observations require deep K-band imaging, and thus angular resolution, to overcome the crowding limit and reach faint magnitudes. Chanam'e et al.13 present calculations which show that it will be possible to detect the black hole population in the GC if observations down to K ~ 21 can be made, the limit at which crowding will start to dominate and only reachable in the near term by GSMT.

M32 and Beyond: As the closest example of an elliptical galaxy, M32 is a benchmark for population synthesis studies of more distant elliptical galaxies. In work done to date, no existing telescope has had the angular resolution to resolve stars in M32 fainter than the red horizontal branch, even in the lowest surface brightness regions of the galaxy. Deeper CMDs (color magnitude diagrams) are needed in both the outer and inner regions to confirm the mean age and metallicity, and the radial gradients thereof, that are predicted by population synthesis models. For example, population synthesis analyses find that the inner regions of M32 are dominated by a population of ~ 5 Gyr old stars of solar metallicity, with outwardly increasing gradients in age and metallicity.14 High-resolution spectroscopy of M32 giants is also needed to test the predicted pattern of abundance enhancements. We have used simulated images in order to explore the ability of the GSMT and its MCAO system to resolve stars in dense environments such as M32. These simulations are described in detail in Section 2.3.5. In brief, the M32 simulation assumed a mix of 1 Gyr ([Fe/H] = 0), 5 Gyr ([Fe/H] = 0), and 10 Gyr ([Fe/H] = -0.3) old populations, each drawn from a Salpeter IMF and extending to masses well below the crowding limit. The 1, 5, and 10 Gyr old populations were assumed to comprise 10%, 45%, and 45% of the total by mass, respectively.

Figure 4 Simulated J-K vs. K CMD of the inner region of M32 displayed in four annuli to provide a direct comparison with the CMD of Davidge et al. (ref 15)

In the resulting J-K vs. K color magnitude diagram of the inner region of M32 (see Figure 4), the photometric errors are dominated by confusion, even at the resolution of the GSMT. However, the main features-the horizontal branch, the red giant branch, and the asymptotic giant branch-are sharply defined. Indeed, the color of these features is more accurately known than are their single-band magnitudes, a paradox explained by the correlation of the brightness of the background fluctuations through the different infrared bands. In comparison with Davidge et al., who were able to measure reliable photometry only for AGB stars (MK < -5) in the outer annuli, the GSMT is able to reach stars as faint as MK ~ +1.5, close to the turnoff of the 1 Gyr old population. Moreover, horizontal branch, red giant, and AGB stars are detected all the way into the nucleus of M32 with GSMT.

In the inner field, we recover stars with 10% photometric accuracy to K ~ 20 (K = 18 in the center and K = 22 at the edge of the field). To this depth, the density of recovered stars is, on average, ~ 20 stars per sq. arcsec. If the photometric errors due to crowding are correlated between J and K, as they are in the simulation, the limit for 10% J-K photometry is 2 magnitudes deeper (K = 22), where there are on average 80 stars recovered per sq. arcsec. Due to the severe crowding in the inner field, we recover only a tiny fraction of 1% of the total number of stars, but an appreciable 64% of the total light.

Figure 5 examines in more detail the CMD for the outer annulus (7.4-14.1") of the inner field. The RGB, horizontal branch, and AGB are clearly defined. The separation in color of the 5 and 10 Gyr old RGB stars is due almost entirely to their 0.3 dex difference in metallicity, showing that the GSMT can detect narrow metallicity spreads even in the densest regions of M32. Also striking is the AGB bump, a feature produced by a brief pause in the evolution of AGB stars.16

The excellent definition of this feature, a result of the high spatial resolution and large number of stars, raises the possibility of using it, in combination with the metallicity distribution determined from spectroscopy, to estimate the star formation history of nearby galaxies.

Figure 5 Detail of the CMD for the 7.4-14.1 arcsec annulus in the inner field of M32. Figure 6 Detail of the CMD for the outer field in M32.

Figure 6 shows the CMD measured in a field 90" from the center of M32, in a region of low surface brightness, similar to that studied by Grillmair et al.17 In this outer field, we recover stars with 10% photometric accuracy to K = 24. The density of recovered stars is, on average, ~ 20 stars per sq. arcsec. If the photometric errors due to crowding are correlated between J and K, the limit for 10% photometric accuracy in J-K is 2 magnitudes deeper (K = 26), where we would recover, on average, 80 stars per sq. arcsec. This represents only 1% of the total number of stars, but 82% of the total luminosity. In this less crowded field, the GSMT will be able to detect the main sequence turnoff for ages as old as 5 Gyr. With the GSMT, M32 may thus be the first elliptical galaxy to have an age directly measured from its turnoff.

Using the results for M32 as a guide, we can estimate the ability of GSMT to produce CMDs and take spectra of stars in more distant galaxies, such as those in the M81 group, the Sculptor group, Cen A, and the Virgo cluster. At the following distances, the GSMT could reach the following imaging depths, set by confusion:

The ability to study resolved stars in such a statistical sample of galaxies would truly represent a Copernican revolution for modern stellar populations research, much of which remains trapped in our low-density corner of the universe.

We can also use these simulations to illustrate the extent to which the GSMT is expected to produce significant gains relative to JWST (James Webb Space Telescope) in the realm of stellar populations. Figure 7 shows the CMD for the 7.4-14.1" annulus in the inner field of M32 as observed with JWST. When compared with the GSMT CMD shown in Figure 5, the great difference in the ability of the GSMT and JWST to study crowded stellar fields is dramatically apparent.

Figure 7 CMD for the annulus 7.4-14.1 arcsec in radius from the center of M32 as measured from a simulated NGST image. The same set of stars simulated in Figure 5 and the NGST PSF of Krist (1999) were used to produce the simulated image. Note the inability of NGST to distinguish the three input stellar populations.

Massive Star Clusters in Nearby Galaxies: We also used simulated data to investigate the ability of the GSMT to explore the upper end of the young main sequence in dense starburst-like clusters. Three clusters with properties similar to those observed for the Arches (Section 2.3.2) were placed on the stellar field of the M33 nucleus, plus disk components sampling a range in background density against which to extract the cluster stellar sequences (see the AppendixSection 2.3.5 for again, this is what used to be 3.5 details of the simulation). The simulated clusters were assumed to be 66 Myr old (set by the choice of Girardi et al.18 isochrones), considerably older than the 2-4 Myr age of the Arches cluster, but with the same brightness and number distribution as the Arches at K (for the brightest stars) in order to simulate the expected crowding conditions.

Figure 8   Measured K-band luminosity function for the simulated clusters compared to the input luminosity function. There is an intrinsic gap in the input cluster K-band luminosity function (KLF) at K = 20.

Figure 8 shows the combined KLF extracted in regions around each of the three simulated clusters, compared with the input luminosity function for the clusters. From the comparison of the input and output star lists, the counts are found to be 77% complete in the interval 16 < K < 19. Although at 0.7 Mpc distance the starburst clusters appear extremely compact on the sky, the GSMT is able to resolve individual stars over the brightest ~ 3 magnitudes. The position of the cluster in the galaxy is found to have little effect on the observed KLF; this is likely due to the very high density of the clusters. For lower density clusters, and more extended ones, the position of the cluster on the background galaxy will be more significant.

Massive starburst clusters such as those described here would be ideal candidates for diffraction-limited IFU spectroscopy. Moderate resolution (R = 2000-5000) K-band spectroscopy would provide accurate infrared spectral types,19 which can be used to estimate the masses and ages of the individual main sequence OB stars, as is done routinely in the Galaxy at infrared wavelengths.20,6 With this approach, the upper mass cut-off of the IMF could be explored directly in local galactic environments spanning a wide range in metallicity and stellar density. The GSMT is the only facility capable of making measurements for individual stars in extra-galactic clusters beyond the Magellanic clouds.

It is tempting to speculate how the GSMT would add to the study of young clusters in galaxies even more distant than M33. There are some 50 super star clusters (SSC) in the Antennae (merging pair) galaxies,21 and many more of lower mass. At 10 times the distance to M33, an Arches-like cluster would span just 0.035" in the Antennae. These are seven K-band pixels. The Antennae clusters studied to date with VLT and Keck show CO absorption at 2.3µm.22,23 Thus they contain evolved luminous M supergiants and/or AGB stars(>10 Myr) and several just show spatial structure at the angular resolution of the VLT. About 40% of this sample of SSCs can be observed with near-infrared integral-field spectrograph (NIFS) (the facility IR IFU on Gemini North) in integrated light. The GSMT will be able to observe more highly buried (thus extinct) clusters in these and other nearby galaxies, and the brighter counterparts in even more distant galaxies. Clearly this science case should be explored in detail for follow-up studies of the GSMT and its instrument complement.


The key enabling instruments are a near-IR imager (Section 4.7.4) fed by an MCAO system (Section 4.6.2) and an IFU spectrograph fed by an MCAO system.


Images for three simulations have been produced: M32, the GC, and M33 with embedded massive star clusters. They test different regimes in crowding and dynamic range.

Input Data: The input data were generated with a slightly modified version of the star formation history code written by K. Olsen. This code uses a specified set of isochrones,18 IMF, and spatial distribution parameterization to produce the artificial population. The key parameters are:

There are two GC fields, one centered on the central cluster and one offset by a full FOV along the diagonal. There are three M32 fields, one each at 0, 30, and 90 arcseconds radial distance from the center. The M33 case employs a nuclear component,24 a disk component (exponential with R0 = 61"; Kent 1987), and the Arches cluster (spatial distribution based on HST data). The total luminosity was set for M33 by the aperture photometry of Gallagher et al.25

The Arches and M32 input data were iterated until they matched the luminosity functions (in the bright end) of existing Gemini/Hokupa'a data. No attempt was made to model true aggregate stellar populations. Multi-component models were adopted which produced the "right" bright K-band luminosity function. The faint end was estimated by determining the crowding limit, then going some ~ 4 magnitudes deeper before cutting off the input mass. The last input mass bin contributes < 5% of the light to the background surface fluctuations, so going deeper adds only to the uniform background. The crowding limit for the GC and M32 were estimated based on the formulation of Olsen 2001 (see also Renzini 1998).26,27 The limit is inversely proportional to the square of the length of a resolution element. Thus the observed limit in the Hokupa'a data and the image quality difference between the observed data and that expected for the diffraction-limited GSMT lead to a direct estimate of the simulation crowding limit. It is necessary to include fainter stars in the simulation because they contribute to the fluctuations against which the limiting magnitude stars are observed (extracted).

As an example, consider the GC case. The Hokupa'a data suggest (conservatively) that the crowding limit is K = 16 for 0.2" FWHM (full width half-maximum) images. For a diffraction-limited 30-m telescope, the core image size will be 0.015". This suggests a crowding limit of K = 20.6 for the GSMT on the GC. The actual input KLF for the GC study case was 24+ for a lower mass cutoff of 0.4Msolar. The stars at K = 24+ contribute fewer than 10% of the fluctuations in a resolution element when compared to the total for all stars between K = 20.5 and KK = 24. The input parameters are summarized in Table 1.

Caseµsub K
Radial ProfileLower Mass
(Gyr) (%tot)
IMF Slope
M32 inner10.1((R/a)2 +1)-0.6 1.24,1.11,0.881,5,13(.1,.45,.45)0.02,0.02,0.012.35
M32 middle 16.8 0.2,0.2,0.21,5,13(.1,.45,.45)0.02,0.02,0.012.35
M32 outer 18.1 0.15,0.15,0.151,5,13(.1,.45,.45)0.02,0.02,0.012.35
M33 Nucleus 7.89((R/a2 +1)-0.75
M33 Disk 17.0e(R/a)
Table 1 Simulation Input Parameters

Image Generation: The input stellar lists were put into 4096 x 4096 (actually a bit larger, to account for edge effects, then trimmed) and convolved with the MCAO PSF appropriate to each wavelength (JHK). F. Rigaut of Gemini computed the MCAO psfs, and these are described in detail in Section 4.6.2. The scale was taken as 0.005"/pixel which gives three pixels for a diffraction-limited PSF at K. After convolution, sky flux and unresolved stellar light (the light attributed to the part of the stellar population below the lower mass limit in Table 1) were added to the images. The latter contribution was less than 10-15% of the total stellar luminosity in the most crowded regions. Poisson noise was added in each pixel for the flux components, and readnoise was added as well (15 e-). The images were also scaled to several integration times and for each time, a series of frames was created and co-added into the final astronomical image.

The noise and background characteristics are given in Table 2 for each case. The system throughput was estimated by scaling up an 8-m class telescope to 30-m, including all mirrors, the atmosphere, and an MCAO module. An additional background was also computed for the MCAO module (see Table 2). The total system throughput is 0.31 and 0.40 at J and K respectively.

CaseM32 Inner M32 Outer M33 Center GC
Exposure Time10s x 510s x 510s x 5


& coadds 120s x 15120s x 15120s x 15


Read Noise 15 e- 15 e- 15 e- 15 e-
J sky / MCAO (mag/sq")16.2/-16.2/-16.2/-16.2/-
H sky / MCAO14.1/20.814.1/20.814.1/20.814.1/20.8
K sky / MCAO13.7/13.713.7/13.713.7/13.713.7/13.7
Unresolved light(%)10 16


Table 2 Image Noise Characteristics


  1. Ibata, R.A., Gilmore, G., & Irwin, M.J. "A Dwarf Satellite Galaxy in Sagittarius". Nature, 370, 194I (1994)

  2. Majewski, S.R., Munn, J.A.,& Hawley, S.L. "Absolute Proper Motions to B approximately 22.5: Large-Scale Streaming Motions and the Structure and Origin of the Galactic Halo" . ApJ, 459, L73 (1996)

  3. Rich, M., Mighell, K.J., & Neill, J.D. "The Metal Rich Halo of M31". ASP Conf. Ser. 92, 544 (1996)

  4. Forrest, W. J., Shure, M. A., Pipher, J. L., & Woodward, C. E. "Brackett Alpha Images". AIP Conf Proc. 155, 153 (1987)

  5. Allen, D. A., Hyland, A. R., & Hillier, D. J. "The source of luminosity at the Galactic Centre". MNRAS 244, 706 (1990)

  6. Blum, R. D., Conti, P. S., Damineli, A., & Figuerado, L. To be published in the proceedings of "Hot Star Workshop III: The Earliest Phases of Massive Star Birth," ed P. Crowther, forthcoming from ASP.

  7. Krabbe, A., Genzel, R., Drapatz, S., & Rotaciuc, V. "A cluster of He I emission-line stars in the Galactic center". ApJ 382, L19 (1991)

  8. Figer, D. F., Kim, S. S., Morris, M., Serabyn, E., Rich R. M., McLean, I. S. "Hubble Space Telescope/NICMOS Observations of Massive Stellar Clusters near the Galactic Center". ApJ 525, 750 (1999)

  9. Serabyn, E. & Morris, M. "The Galactic Center Environment". ARA&A,34, 645 (1996)

  10. Portegies Zwart, S. F.; Makino, J.; McMillan, S. L. W.; Hut, P. "How Many Young Star Clusters Exist in the Galactic Center?". ApJ 546, L101 (2001)

  11. Zinnecker, H., Brandl, B., Brandner, W., Moneti, A., & Hunter, D. "30 Doradus: The Low-Mass Stars". IAU Symp. 190, 222 (1999)

  12. Melia, F. & Falcke, H. "The Supermassive Black Hole at the Galactic Center". ARA&A, 37, 309 (2001)

  13. Chanamé, J., Gould, A., & Miralda-Escudé, J. "Microlensing by Stellar Black Holes Around Sgr A*". ApJ 563, 793 (2001)

  14. Trager, S.C., Faber, S.M., Worthey, G., Gonzalez, J.J. "The Stellar Population Histories of Local Early-Type Galaxies. I. Population Parameters". AJ, 119 (4), 1645 (2000)

  15. Davidge, T., Rigaut, F., Chun, M., Brandner, W., Potter, D., Northcott, M., Graves, J. E. "The Peak Brightness and Spatial Distribution of Asymptotic Giant Branch Stars Near the Nucleus of M32". ApJ 545, 89 (2000)

  16. Gallart, C. "Observational Discovery of the Asymptotic Giant Branch Bump in Densely Populated Color-Magnitude Diagrams of Galaxies and Star Clusters". ApJ, 495, L43 (1998)

  17. Grillmair, C. et al. "Hubble Space Telescope Observations of M32: The Color-Magnitude Diagram". AJ 112, 1975 (1996)

  18. Girardi, L., Bressan, A., Bertelli, G., Chiosi, C. "Evolutionary tracks and isochrones for low- and intermediate-mass stars: From 0.15 to 7 M¤, and from Z=0.0004 to 0.03". A&AS 141, 371 (2000)

  19. Hanson, M. M., Conti, P. S., & Rieke, M. J. "A Spectral Atlas of Hot, Luminous Stars at 2 Microns". ApJS 107, 281 (1996)

  20. Hanson, M. M., Howarth, I., & Conti, P. S. "The Young Massive Stellar Objects of M17". ApJ 489, 698 (1997)

  21. Whitmore, B. C., et at. "The Luminosity Function of Young Star Clusters in "the Antennae" Galaxies (NGC 4038-4039)". AJ 118 (4), 1551 (1999)

  22. Gilbert, A., et al. "Infrared Spectroscopy of a Massive Obscured Star Cluster in the Antennae Galaxies (NGC 4038/9) with NIRSPEC". ApJ, 533, L57 (2000)

  23. Mengel, S., et al. "K-Band Spectroscopy of Compact Star Clusters in NGC 4038/4039". ApJ 550, 280 (2001)

  24. Lauer, Tod R.; Faber, S. M.; Ajhar, Edward A.; Grillmair, Carl J.; Scowen, Paul A. "M32 +/- 1". AJ 116 (5), 2263 (1998)

  25. Gallagher, J. S.; Goad, J. W.; Mould, J. "Structure of the M33 nucleus". ApJ 263, 101 (1982)

  26. Olsen, K. 2001, in preparation.

  27. Guarnieri, M. D.; Ortolani, S.; Montegriffo, P.; Renzini, A.; Barbuy, B.; Bica, E.; Moneti, A., "Infrared array photometry of bulge globular clusters. I. Combined ground based JK and HST VI photometry of NGC 6553". AA 331, 70 (1998)

November 2002