To this aim, we have obtained deep photometric images with the WFPC2 camera on board of the Hubble Space Telescope of a few galaxies, where single stars can be resolved and measured with good accuracy. In the case of NGC 1569, data, analysis and interpretation are described in detail by Greggio et al. (1998, hereinafter G98).
NGC 1569 is an outstanding object among active galaxies (e.g. Gallagher et al. 1984); it hosts super-star-clusters (SSCs), i.e. high-density stellar aggregates with masses typical of Galactic globular clusters (O'Connell et al. 1994; Ho & Filippenko 1996), but ages of only a few Myr (González-Delgado et al. 1997, GD97); and it is one of the very few galaxies for which there is direct observational evidence of galactic winds (e.g. Waller 1991; Heckman et al. 1995; Della Ceca et al. 1996).
Aside from the above features, the properties of NGC 1569 are typical of dwarf irregulars. With a distance modulus of (m-M)0 = 26.71, its total absolute B magnitude is MB.0 ∼ -17 (Israel 1988), which is intermediate between the two Magellanic Clouds. Total and hydrogen mass are Mtot ≅ 3.3·108 Msun and MH ≅ 1.3·108 Msun respectively (Israel 1988), indicating that NGC 1569 is a gas rich system. The metal and oxygen abundances are also similar to those of the Small Magellanic Cloud: Z ∼ 0.25 Zsun (Calzetti et al. 1994; GD97), 12 + log(O/H) = 8.19 - 8.25 (Hunter et al. 1982; Kobulnicky & Skillman 1997).
The general morphology of the CMD of NGC 1569 is similar to that of Local Group irregulars observed from ground (see, e.g., Freedman 1988; Tosi et al. 1991; Greggio et al. 1993; Marconi et al. 1995; Gallart et al. 1996), with a quite scattered distribution and a prominent concentration of stars in the blue plume. The CMD shows a smaller fraction of red stars, but this is mostly due to the selection effect of the magnitude limit for object detection in the F439W frame. Indeed, the I vs. (V-I) diagram of NGC 1569 obtained by Vallenari & Bomans (1996, hereinafter VB96) shows a significant population of red (mostly intermediate age) stars.
Panel c of Fig. 1 shows the Geneva evolutionary tracks with Z = 0.001 (Schaller et al. 1992) transformed in the HST photometric system and with the reddening and distance of NGC 1569. By comparing the empirical CMD of panel b with the tracks in panel c, one can immediately understand that: 1) the blue plume is populated not only by main-sequence (core-H-burning) stars, but also by a significant fraction of stars at the hot edge of the blue loop (core-He-burning) evolutionary phase, as suggested by several authors (e.g., Freedman 1988; Tosi et al. 1991; Gallart et al. 1996); 2) the stars visible with our selected photometry are more massive than ∼12 Msun on the main sequence and than ∼5 Msun on the blue loops; theferore they are at most as old as the lifetime of a 5 Msun star, i.e. ∼0.15 Gyr. This is then our maximum lookback time with these data.
To interpret the empirical CMD and the corresponding luminosity functions (LF) in terms of SF history, we have applied the method of synthetic CMDs developed by Tosi et al. (1991). There is no need to recall it here, since the procedure is very similar to that described at this meeting by A. Aparicio. The method has proven to be a powerful tool to investigate the star formation history of galaxies with young stellar populations (see Tosi 1994; Greggio 1994). Given the many uncertainties affecting both the observational data and the theoretical models, one cannot currently pretend to obtain unique results for the data interpretation, but we can significantly reduce the range of acceptable parameter values and therefore of the evolutionary scenarios attributable to the examined regions.
Despite the large spread in the stellar distribution in the empirical CMD, it may turn out impossible to reproduce the observed CMD and LF if the set of adopted stellar models is not appropriate for the examined stellar population. This is, for instance, the case with both the Geneva tracks with Z = 0.004 and Z = 0.008 (see G98 for details); whereas those with Z = 0.001, displayed in Fig. 1c, fit fairly well the observed distributions. The other set of stellar models able to satisfactorily reproduce the CMD and LF of NGC 1569 is that by Fagotto et al. (1994), with Z = 0.004.
Fig. 2 shows three representative cases of the synthetic CMDs. In panel a the SF is still active at the present epoch and provides a number of bright objects on top of the blue plume that have no counterpart in the observational CMD of Fig. 1b. This excess is visible also in the LF (see G98 for a through discussion) and is unavoidable with an ongoing SF, even changing drastically the IMF, the galaxy distance or any other parameter. Exactly the same result is obtained with the Schaller et al. (1992) tracks, except that the stopping time is 7 Myr ago, rather than 10, due to the different luminosities and time scales of the stellar models. Notice that these results refer to the galactic field: in one of the SSCs the SF was still active about 4 Myr ago (GD97) and it must have been still working very recently also in the areas filled with several H II region complexes.
Fig. 2b displays one of the CMD in better agreement with the data. It assumes two episodes of SF, separated by 5 Myr. The SF activity has stopped 10 Myr ago for the same reason as above. The quiescent phase between the two SF episodes is quite short, but many simulations have shown that it cannot be significantly longer, independent of the adopted stellar models, of the IMF and of the other assumptions. Fig. 2c shows, for instance, that if one allows the quiescent interval to last 10 Myr instead of 5, and leave all the other parameters as in panel b, a gap, totally inconsistent with the observed distribution, appears in the synthetic CMD. Such gaps are unavoidable if the two formation periods are separated by more than 8-10 Myr.
It must be emphasized that such a high SF activity cannot be extrapolated back to the rest of the galaxy lifetime, since at such rate NGC 1569 would have run out of gas in much less than 1 Gyr (see G98 and also Hunter et al. 1989). We must then conclude that in the past the SF rate must have been much lower and that long quiescent phases must have occurred.
Some indication in this direction is already available from VB96, who suggested from the V-I CMD that a much shallower SF activity has taken place in NGC 1569 around 1 Gyr ago. The infrared data that we are going to acquire with the NICMOS camera on HST should reveal (if present) the stars at the tip of the red giant phase, with an age of up to 10 Gyr. This will allow us to derive the SF history over most of the galaxy lifetime and infer a much more complete scenario for the evolution of this intriguing system.
| First version: | 03rd | August, | 1998 |
| Last update: | 28th | September, | 1998 |