Proceedings of the Workshop
"The Magellanic Clouds and Other Dwarf Galaxies"
of the Bonn/Bochum-Graduiertenkolleg

Preliminary Results on the Resolved Stellar Population of I Zw 18

A. Aloisi1, L. Greggio1, M. Tosi2, M. Clampin3, A. Nota3, and M. Sirianni3

1Dipartimento di Astronomia, Università di Bologna, Bologna, Italy
2Osservatorio Astronomico di Bologna, Bologna, Italy
3Space Telescope Science Institute, Baltimore, USA

Received 06th March 1998
Abstract. To infer the star formation history and initial mass function (IMF) of the resolved stellar population in the Blue Compact Dwarf Galaxy I Zw 18, we have applied the method of comparing the observed colour-magnitude diagram (CMD) with synthetic ones based on homogeneous sets of stellar evolutionary tracks. To this purpose we have retrieved and re-reduced archive HST/WFPC2 data. Here we present preliminary results on our data reduction and CMD simulations. Models seem to indicate a quite flat IMF, with an exponent in the range α ∼ 1.5 - 2.0, and a relatively old star formation activity stopped or drastically reduced 3 - 5 Myr ago. It is not yet possible to definitively distinguish between the single-burst or two-burst scenarios proposed for this galaxy, but the latter seems to provide better results.

1. Data Reduction

Our data on I Zw 18 are taken from HST archive. I Zw 18 has been observed in the broadband filters F336W, F555W and F814W of HST/WFPC2 in November 1994 (the galaxy was centered in the PC1 CCD), and CMDs obtained from photometric reduction have been published by Hunter & Thronson (1995). We have performed a new photometric reduction of archive data to obtain a direct estimate of all parameters involved in the CMD simulations (distribution of photometric errors, incompleteness, etc.). Figure 1a shows the CMD in the HST F555W and F814W bands containing all the stars with photometric error σ < 0.2 mag in both bands (247 objects).

2. Synthetic CMDs

In order to deduce the star formation history and IMF of I Zw 18 from the CMD obtained by data reduction, we are carrying on several simulations (see Tosi et al. 1991 and Greggio et al. 1998 for more details on the simulation code) considering two sets of stellar evolutionary tracks: Geneva tracks with Z = 0.001 (Schaller et al. 1992) and Padova tracks with Z = 0.0004 (Fagotto et al. 1994). These are the only ones available with a metallicity compatible with that of I Zw 18, the most metal poor galaxy ever known with Z∼1/50 Zsun. In order to obtain the transformation from the theoretical to the observational plane, we have assumed a distance modulus (m-M)0 = 30 mag, corresponding to a distance of 10 Mpc, and we have adopted a reddening E(B-V) = 0.04 (Hunter & Thronson 1995).

3. Preliminary Results

In the CMD of I Zw 18 we have sampled stars with a magnitude V < 27.5. At the assumed distance modulus of the galaxy this means MV < -2.5, corresponding to stars more massive than 4 Msun in their post-main sequence phase, and therefore with lifetimes <0.2 Gyr. A comparison of the observed stellar distribution with synthetic diagrams has not allowed us so far to understand if in this galaxy the star formation activity has occurred in one single burst or in two bursts separated by a quiescent period: both these star formation scenarios give diagrams which are not in perfect agreement with the observations. Anyway we have recognized the following general trends in the star formation activity:
  1. both in a single-burst and in a two-burst scenarios, simulations indicate an IMF flatter than Salpeter (α = 2.35) in order to avoid too many faint red stars. Values of the IMF exponent more consistent with data are in the range α = 1.5 - 2.0;
  2. in a single-burst scenario star formation cannot have started later than 30 Myr ago to reproduce the observed faint stars (both blue and red), but it should have stopped or been drastically reduced around 3 - 5 Myr ago to avoid stars brighter than V < 22 not observed in I Zw 18. In Fig. 1b we show an example of a single young burst started 10 Myr ago and still ongoing (Padova tracks with α = 1.5): the simulation is not able to predict stars fainter than V ∼ 26, to obtain both faint and bright red stars and to avoid stars brighter than V ∼ 22. In Fig. 1c we present instead one of the single-burst models which better seems to reproduce the observed CMD (Padova tracks with α = 1.5) with an exponentially decreasing and still ongoing star formation activity started 0.5 Gyr ago and with an e-folding time of 0.3 Gyr. There are still too many red stars of intermediate brightness in the synthetic CMD;
  3. in a two-burst scenario we generally obtain the best consistency with observations, especially if we consider an old burst started 0.5 - 0.3 Gyr ago and finished 50 Myr ago and a more recent burst started 30 Myr ago and stopped 3 - 5 Myr ago. In Fig. 1d we show our best two-burst model obtained considering a first episode from 0.5 Gyr to 50 Myr ago, and a second one from 30 Myr to 5 Myr ago (Padova tracks with α = 1.5).

[Click here to see Fig. 1!]

References

Links (back/forward) to:
First version: 24thJuly,1998
Last update: 28thSeptember,1998
Jochen M. Braun   &   Tom Richtler
 (E-Mail: jbraun|richtler@astro.uni-bonn.de)