5.1.1 Chemical Abundance of IZw18
The first estimate of the metallicity of IZw18 came from Searle and Sargent (1972) who showed that the oxygen abundance was below one tenth of solar, while the He abundance appeared to be normal. Alloin et al. (1978) derived 12 + log(O/H) = 7.2, which is close to most recent determinations, in particular by Izotov and Thuan (1999) who give 12 + log(O/H) = 7.18, or 1/50 of the Sun. Values of O/H in the range 1/30 to 1/60 of the solar value have been quoted over the last 20 years. The oxygen abundance seems to be constant, within the errors, over the optical face of the galaxy (e.g. Skillman and Kennicutt 1993, Vilchez and Iglesias-Páramo 1998, Legrand et al. 1999).
C/O was first derived by Dufour et al. (1988) who found [C/O] = -0.27 from IUE and ground based optical observations, an unexpectedly high value: higher than [C/O] in dwarfs like the LMC, indicating that the chemistry in IZw18 is anomalous. This value was later revised down by Dufour and Hester (1990) to [C/O] = -0.5. Garnett et al. (1995) used the Faint Object Spectrograph (FOS) on HST but only got a lower limit [C/O] > -1.3. Later, they re-observed IZw18 with the FOS and found [C/O] = -0.63 and -0.56 for the north-west (NW) and south-east (SE) regions respectively (Garnett et al. 1997). These values are again rather high when compared to other metal-poor dwarfs and Garnett et al. (1997) proposed that a carbon enrichment from an intermediate age population had occurred previous to the current star formation burst. Recently, Izotov and Thuan (1999) reanalysed the HST data by Garnett et al. (1997), now finding [C/O] = -0.77 and -0.74 for the NW and SE regions respectively. Their lower C/O follows mainly from a higher adopted electron temperature based on new ground based spectra. Izotov and Thuan (1999) also found indications of a temperature gradient, giving rise to the apparent abundance difference between the NW and SE components. These last [C/O] values for IZw18 follow nicely the trend for other very metal-poor galaxies (see Fig. 4). However, the fact that the [C/O] determination has changed so much with time is of course not satisfying. Hence the new value adopted by Izotov and Thuan (1999) should perhaps not be taken at face value since part of their disagreement with other authors might be associated with the imperfect match between HST and ground based apertures. A deep HST UV+optical spectrum with a single instrument like STIS could resolve this issue.
Similar problems were encountered with the study of the nitrogen to oxygen ratio in IZw18 that first yielded lower limits until Dufour et al. (1988) were able to derive [N/O] = -1.36. Izotov and Thuan (1999) recentlyderived a lower [N/O] of -1.60 in perfect agreement with their overall trend that [N/O] stays constant at low [O/H] in metal-poor galaxies, contrary to previous findings (see Fig. 4). The helium abundance of IZw18 has been a concern for some time. Its value is of special importance for deriving the primordial helium abundance, since IZw18 has the lowest known ISM heavy element abundance of galaxies. Searle and Sargent (1972) found a rather normal He abundance, that later was revised down, e.g. Y = 0.21 (Davidson and Kinman 1985), and Y = 0.226 (Pagel et al. 1992). These values were rather low in view of the standard big bang nucleosynthesis (SBBN) theory. Izotov and Thuan (1998a, b) find Y = 0.242, in comfortable agreement with SBBN.
Kunth et al. (1994) attempted to measure the oxygen abundance in the neutral gas cloud surrounding IZw18 by using UV absorption lines, observed with the GHRS onboard HST. They found that the oxygen abundance could be as low as 1/1000 of the solar value, indicating near pristine gas with abundances lower than what is found in QSO absorption line systems. However the use of saturated lines was criticised by Pettini and Lipman (1995). Van Zee et al. (1998) argue that the H I velocity dispersion together with the measured OI line of Kunth et al. (1994) imply an oxygen abundance 1/60 of the solar value, suggesting that the chemical enrichment products are well mixed.