5.1.2 On the Age of IZw18
Izotov and Thuan (1999) come to the conclusion that all galaxies with 12 + log(O/H) < 7.6 must be younger than 40 Myr, thus in practise newly born galaxies; and this would be very much so for IZw18 with its record low abundance (12 + log(O/H) = 7.18). There are however other possible interpretations of the abundance ratios of metal-poor galaxies (as we already pointed out in Sect. 4.4.8). Moreover there is some independent evidence suggesting that IZw18 does indeed host an old underlying population.
The age of IZw18 has been debated ever since the early seventies when the intriguing properties of this galaxy were first realised. Being the most metal-poor galaxy, it is of course one of the most promising candidates for a genuinely young galaxy. The absence of an outer regular envelope made IZw18 a good young galaxy candidate. Thuan (1983) argued from near infrared aperture photometry that the galaxy was old, but comparing his colours with the recent Bruzual and Charlot (2000) leads to a different conclusion. Moreover the apertures only cover the central part of the galaxy which is dominated by young stars irrespective of the possible presence of an underlying population.
Pantelaki and Clayton (1987) claimed a high age for IZw18 based on the observed C/O and N/O ratios, but these were later revised downwards. A similar argument was given by Garnett et al. (1997) from their new C abundance, that however was revised downwards recently by Izotov and Thuan (1999) weakening the conclusion. In view of the uncertainties on the N/O and C/O ratios and yields, and that these ratios are subtle to interpret anyway, it is clear that they give very limited constraints on the star formation history in IZw18. The chemical evolution model used by Kunth et al. (1996) and the spectro-chemical evolution model of Legrand (1998) also suggest that IZw18 is not young. The latter one predicts that IZw18 could have experienced a low but continuous star formation rate for several Gyr prior to the present burst. Of course the difficulties outlined in Sect. 3 make all constraints on galaxy ages derived from chemistry very uncertain.
IZw18 has been imaged in the optical by HST by two different groups (Hunter and Thronson 1995; Dufour et al. 1996), both finding a resolved population of young massive stars. There was no evidence for old stars, but none against, since the observations were not sufficiently deep to allow old stars to be detected. These data sets were recently reanalysed by Aloisi et al. (1999) and give support for an age of at least 0.5 Gyr. Östlin (1999a) studied the resolved stellar population in the near infrared with NICMOS onboard HST, and found that while the NIR colour magnitude diagram was dominated by stars 10-20 Myr old, numerous red AGB stars require a much higher age in agreement with Aloisi et al. (1999). The NICMOS data require stars older than 1 Gyr to be present, and an age as high as 5 Gyr is favoured. This holds even if a distance slightly higher than the customary 10 Mpc is adopted.
In Fig. 8 we show surface brightness and colour (B - R and B - J) profiles of IZw18 from deep CCD images obtained at the Nordic Optical Telescope and infrared images from the UKIRT (Östlin 1999c). The colours rise continuously with increasing radius and reach B - R = 0.6 and B - J = 1.6 at a radius of 10 arcseconds. If due to purely stellar emission, a single stellar population model with a metallicity of 1/50 Z (Bruzual and Charlot 2000) indicates an age of log (age) = 9.1 ± 0.1 for both colours, irrespective of IMF (Salpeter, Scalo or Miller-Scalo). With a more realistic assumption of a more or less continuous star formation rate (e.g. for an exponentially decaying SFR with e-folding time = 3 Gyr) we predict an age of at least 5 Gyr. One possible caveat is that there might be a substantial contribution from ionised gas to the colours, but we believe that this is not a dominant effect. The B-band profile has an exponential shape for radii greater than 10 arcseconds; fitting a disc to the outer parts yields a central surface brightness of ~ 23 mag/arcsec2 and an integrated disc luminosity MB = -11.8. With a mass to light ratio for a 109 years old single stellar population, the disc luminosity is equivalent to a stellar mass of ~ 5 . 106 M, but may be a factor of two higher if the SFR has been continuous (cf. the predictions by Legrand 1998). With a mass of the young population of ~ 106 M this means that while difficult to detect, the old population dominates the stellar mass. This suggests that IZw18 could be hosted by a low surface brightness galaxy.
Thus, although not foolproof, photometry (of individual stars and surface photometry) now indicates that the galaxy is in fact old, although Izotov et al. (1999a) in a last attempt to resurrect its youth, suggested that the distance of IZw18 has been severely underestimated.
Figure 8. Luminosity and colour profile of IZw18. Left panel: the radial surface brightness profile of IZw18 in B. Note the apparent exponential shape of the profile for radii greater than 10 arcseconds, indicating an underlying low surface brightness component with µB,0 = 23 mag/arcsec2. Right panel: B - R (filled squares) and B - J (crosses) profiles of IZw18; note the red colours at large radius. From deep images obtained at the Nordic Optical Telescope (NOT) and the UKIRT (Östlin 1999c, in prep).