|Annu. Rev. Astron. Astrophys. 1996. 34:
Copyright © 1996 by . All rights reserved
3.2. Evolved Field Stars
3.2.1. ASYMPTOTIC GIANT BRANCH STARS For many years, "intermediate-age" in the Magellanic Clouds was taken to be nearly synonymous with the presence of luminous asymptotic giant branch stars. Numerous workers have identified and studied AGB stars in field and clusters of the MCs (e.g. Aaronson & Mould 1985, Frogel et al 1990, Mould 1992). Because the peak AGB luminosity increases steadily with decreasing age (due to the increase in core mass with increasing ZAMS mass), and because AGB stars are relatively easy to identify, the brightest AGB stars are useful as beacons for the intermediate-age populations throughout the Clouds and their environs. Infrared observations are required to determine AGB luminosities because these stars have low surface temperatures and correspondingly large optical bolometric corrections (e.g. Frogel et al 1990).
Magellanic Cloud clusters have played an important role in calibrating the luminosities of AGB stars in terms of age. Frogel et al (1990) provide the most recent calibration of AGB luminosities as a function of cluster age, but, as has been known for some time (Aaronson & Mould 1985, Hodge 1983), this correlation between AGB properties and age is very broad: For clusters of a given age, the observed peak AGB luminosities vary significantly in different objects. In practice, it becomes impossible to use the AGB to attain age resolution of better than a factor of two for populations older than about 1-2 Gyr. One important reason for this is that the AGB is a short-lived evolutionary phase (Iben & Renzini 1983). Thus, defining an AGB sequence or determining the peak AGB luminosity depends sensitively on the sample size. Metallicity differences between clusters at a given age and star-to-star variations in core mass and mass-loss rates also affect their AGB luminosities and lifetimes and contribute to the intrinsic scatter of AGB properties with age. The AGB is a blunt tool with which to constrain ages, but for many galaxies with only partially resolved stellar populations, it is often the only tool available.
While acknowledging these limitations, we can still use AGB stars to study intermediate-age populations in the MCs. Frogel & Blanco (1983, 1990) used the detailed features of AGB stars to estimate ages of the progenitor field populations in the Bar-West field of the LMC. They conclude that the luminosity and IR-color distribution of the AGB can be understood as two sequences. The bluer, more luminous AGB stars appear to be associated with stars as young as 108 yrs, while the redder, lower-luminosity AGB stars trace 1-5 Gyr populations (Frogel et al 1990, Feast & Whitelock 1994). Frogel & Blanco (1990) show that the young and intermediate-age populations are present in the ratio 1:4 in this inner LMC field. The results imply an episodic star-formation rate in the inner LMC coupled with a slow overall decline in star formation during the past few Gyr. These studies also illustrate the value of IR photometry as a way to identify AGB stars easily without the need for confirming optical spectroscopy.
Studies of carbon AGB stars - luminous, cool AGB stars with dredged-up carbon in their atmospheres - have recently been used to map the extent of intermediate-age populations in both MCs. Feast & Whitelock (1994) have analyzed the C stars identified by Demers et al (1993) to trace intermediate-age populations in the outer parts of the LMC, in the inter-cloud "bridge" connecting the two Clouds, and in the outer halo of the SMC. In all three cases, the mere presence of the C stars indicates the presence of a significant intermediate-age population in these regions.
The utility of AGB stars as intermediate-age population tracers is well illustrated by the Magellanic bridge. Demers and collaborators (Grondin et al 1990, 1992;, Demers & Irwin 1991;, Battinelli & Demers 1992;, Demers et al 1991) have demonstrated the existence of young ( 200 Myr) stars and loose clusters or associations in a stellar inter-cloud bridge connecting the two galaxies. This stellar bridge is closely associated with the H I bridge between the Clouds (Mathewson & Ford 1984, Westerlund 1990). The presence of luminous AGB stars in the bridge (Demers et al 1993) suggests that the young population - which may well have been born when the bridge itself formed (Demers et al 1991, Gardiner & Hatzidimitrou 1992) - maybe accompanied by older stars that possibly predate the bridge. Whatever mechanism formed the bridge and its youngest stars was also probably responsible for stripping some intermediate-mass stars from one or both of the Magellanic Clouds (see Section 3.5).
3.2.2. ANOMALOUS CEPHEIDS Anomalous Cepheids (ACs) are believed to form during the late evolution of intermediate-mass metal-poor stars (Hirshfeld 1980). ACs have periods of 0.25-1.5d and obey a period-luminosity relation (Nemec et al 1994). These stars have luminosities considerably higher than RR Lyrae stars, making ACs relatively easy to identify in external galaxies. The field of our Galaxy is not known to contain any ACs, but, even if present, they would be very difficult to identify due to confusion with RR Lyraes of similar periods (see Mateo 1996a), and with type II Cepheids and evolved HB stars (Sandage et al 1994). Because the standard model for ACs requires them to be more massive than the turnoff masses of ancient clusters such as NGC 5466 (Hirshfeld 1980), it is generally believed that NEC 5466-V19 formed as the result of mass transfer in a now-merged binary progenitor (Zinn & King 1982, Nemec & Harris 1987).
In contrast, dwarf spheroidal galaxies (Mateo et al 1995, Nemec et al 1994) and the SMC (Smith & Stryker 1986, Smith et al 1992) appear to be rich in ACs. In the case of the SMC, this conclusion rests on observations of only four bona fide candidates. In dwarf galaxies - where complex star-formation histories appear to be the norm rather than the exception (e.g. Da Costa 1992, Smecker-Hane et al 1994, Mateo 1996a) - ACs are generally believed to result from the evolution of single, relatively massive metal-poor stars as described by Hirshfeld (1980 ; see also Smith et al 1992). Anomalous Cepheids therefore represent another potential tracer of the intermediate-age populations in the Clouds. One candidate AC has been reported for the LMC (Sebo & Wood 1995); however, these authors also note that the light curve of this star could be understood as that of a normal 0.51-day RR Lyrae star with a nonvariable but unresolved companion of comparable brightness. Thus it remains uncertain whether the LMC contains any ACs at all. Ongoing and planned large-scale MACHO and OGLE photometric surveys should soon provide comprehensive censuses of the ACs in both Clouds.
It is tempting to attribute the higher frequency of ACs in the SMC as evidence that this galaxy contains intermediate-age progenitors while the LMC does not. The details of how anomalous Cepheids relate to their progenitor population appear to be quite complex, with metallicity and age probably playing nearly equal roles (Smith & Stryker 1986, Mateo et al 1995). One can also confuse ACs with other types of pulsating, short-period variables found above the HB but which are generally evolved from old-not intermediate-age-progenitors (Sandage et al 1994). In dwarf spheroidal galaxies the frequency of ACs does not appear to correlate strongly with the age of the predominant field population (Mateo et al 1995). The LMC's deficiency of ACs may simply reflect the absence of suitably metal-poor intermediate-age progenitors rather than the complete absence of that population. New theoretical models of ACs applicable to Cloud metallicities and age ranges are badly needed to address this issue.
3.2.3. INTERMEDIATE-AGE LONG-PERIOD VARIABLES As described in Section 2.6, LPVs with periods less than about 250 days are generally classified as "old." Hughes et al (1991) have identified such stars with an old spheroidal-like population of the LMC. For LPVs with periods in the range 225d-425d, which are found in abundance in the LMC, the putative progenitors have main-sequence lifetimes of about 1-3 Gyr (Wood et al 1983, 1985;, Hughes & Wood 1990). This conclusion is consistent with their kinematics, which are disk-like and intermediate between the kinematical properties of young LMC components (e.g. H I clouds, very young clusters) and of older components (the ancient clusters, old LPVs). Hughes & Wood (1990) note that the relative frequency of old and intermediate-age LPVs suggests that the older population is considerably more abundant, by up to a highly uncertain factor of five.
We see below that very few field stars or clusters seemed to have formed 4-12 Gyr ago. The period distribution of LPVs shows no sign of a break corresponding to this hiatus in star formation, suggesting to us that the scatter in the mass-period relation for LPVs is comparable to the full mass range corresponding to intermediate-age stars. Alternatively, the progenitors of LPVs may have continued to form during an epoch when star-formation activity was otherwise very low in the LMC.
3.2.4. PLANETARY NEBULAE Planetary nebulae (PNe) are associated with the late stages of evolution in intermediate-mass stars. About 150 and 50 PNe are cataloged in the LMC and SMC, respectively (Meatheringham 1991, Vassiliadis et al 1992). In the late 1980s, Meatheringham and collaborators studied the spatial, kinematic, and physical properties of PNe in the LMC; Westerlund's (1990) review describes this work in detail.
How old are the PNe in the Magellanic Clouds? Using a simple orbital diffusion model based on heating by massive molecular clouds in the LMC, Meatheringham et al (1988) concluded that the mean age of the PNe sample is about 2-4 Gyr. This value is also consistent with stellar evolutionary models for PNe formation. It should be stressed that these models only very weakly constrain the PNe progenitor masses and therefore their ages. More recently, Vassiliadis et al (1992) confirmed the earlier kinematic solution and mean age for LMC PNe after adding to their sample radial velocities of 11 new nebulae located in the outskirts of the galaxy. Both the kinematic and stellar evolutionary ages of LMC PNe suggest that these objects belong to the galaxy's intermediate-age population. Given the large uncertainties in how the ages of PNe are determined in the Clouds, we cannot currently rule out that some of the PNe come from the ancient population.
In the SMC, Dopita et al (1985) measured a large velocity dispersion and found no clear evidence for rotation in a sample of 44 PNe. Interestingly, Hardy et al (1989) likewise found a kinematic solution very similar to that of a true spheroid or halo for C stars in the SMC. Unlike in the LMC, these two tracers of intermediate-age populations seem to be kinematically ancient. Whatever caused the complex structure and chaotic kinematics of the SMC as a whole are undoubtedly behind this unusual kinematic signature as well (e.g. Hatzidimitriou et al 1993); see section 3.5. If, as in the LMC, the PNe and C stars correspond to 2-6 Gyr populations, then we can conclude that whatever event induced the spheroidal-like kinematics in these objects probably did so within the past few Gyr. A lower limit of about 0.5-1 Gyr for the time when this event occurred is imposed by the fact that the H I in the SMC has approximately disk-like kinematics.
3.2.5. RED GIANT CLUMP STARS Virtually all intermediate-age stars experience a stable He core-burning phase. Among the oldest stars, this evolutionary phase corresponds to the horizontal branch seen in ancient star clusters (Sections 2.2 and 2.5). Stars in intermediate-age populations evolve into an analogous region located just blueward of the base of the red giant branch in the HR diagram. This evolutionary phase was first systematically studied in Galactic open clusters by Cannon (1970), and its location in the HR diagram is referred to as the red clump (RC). Data presented by Sarajedini et al (1995) nicely illustrate that the distinction between the RC and HB is fuzzy: As a population ages, the RC evolves smoothly into a red HB such as observed in 47 Tuc.
Unlike globular cluster horizontal branches (e.g. Lee & Demarque 1990), the structure and locations of red clumps do not change drastically over a very large range of age. Using photometry from MC and Galactic clusters with well-known ages, Hatzidimitrou (1991) noted that the (B - R) color separation of the RC and the RGB at the same luminosity varies systematically as a function of cluster age. This separation, dB-R, varies systematically by about 0.15 mag as a population ages from 1 to 10 Gyr. This behavior is, to first order, independent of metallicity since RC and similar-luminosity RGB have similar surface temperatures and thus vary in color by similar amounts as line blanketing varies (though see Smecker-Hane et al 1994 for a counter-example of this simple evolutionary scenario observed in the Carina dwarf galaxy). For composite stellar populations it is nearly impossible to separate stars with different ages within this age range because the color spread of the RC itself is comparable to the full range of dB-R. However, one can determine the mean age by measuring the color of the RC centroid.
Gardiner & Hatzidimitriou (1992) applied this technique to seven SMC fields, deriving mean ages for each. The individual age determinations scatter between 3 and11 Gyr, with a mean age of 7.6 ± 1.2 Gyr. We see below that the LMC seems to have formed very few stars at this time. These authors also claim the existence of an ancient population in their SMC fields on the basis of the color distribution of RC stars (see Section 2.7). Hatzidimitrou et al (1989, 1993), Gardiner & Hawkins (1991), Gardiner & Hatzidimitrou (1992) have also used RC stars to determine the distribution of intermediate-age populations in the outer SMC, to study the line-of-sight depth of the SMC, and to measure the SMC kinematics as a function of depth.
Numerous photometric studies have identified a prominent RC population throughout the LMC (e.g. Hodge 1987, Brocato et al 1989, Bencivenni et al 1991, Bertelli et al 1992, Vallenari et al 1994a, c, Walker 1993b, Elson et al 1994, Reid & Freedman 1994, Gilmozzi et al 1994, Bhatia & Piotto 1994, Westerlund et al 1995) and the SMC (Brück 1980, Bolte 1987, Hilker et al 1995). In some cases, detailed photometric studies have subsequently revealed the intermediate-age main-sequence progenitors of these RC stars. The fact that the RC is visible in all fields suggests that the intermediate-age populations of the MCs are ubiquitous. With the possible exception of the inter-cloud Bridge (Grondin et al 1992) and some fields located quite far from the MCs (e.g. near NGC 1841; Walker 1990), we know of no MC fields with adequate photometry that do not reveal RC stars. Such stars are even seen as a pervasive background in regions located close to active sites of present-day star formation (e.g. Elson et al 1994).