Annu. Rev. Astron. Astrophys. 1981. 19: 77-113
Copyright © 1981 by . All rights reserved

Next Contents Previous


5.1. Irregular and Blue Compact Galaxies

The gas-rich Irregular galaxies and blue compact galaxies or extragalactic HII regions (Searle & Sargent 1972) are in many ways the simplest and most satisfying systems to study because they usually contain bright HII regions of high excitation in which abundances can be determined by standard methods. Abundance gradients are small or absent (Pagel et al. 1978, Pagel, Edmunds & Smith 1980) and the average abundances in the various systems mostly conform to simple concepts of galactic evolution, so that interesting parameters like the heavy-element "yield" and the primordial helium abundance can be deduced with a relatively high degree of confidence. Data are now available for about 35 of these systems (Lequeux et al. 1979, French 1980, Talent 1980, Kinman & Davidson 1981, Tully et al. 1981) ranging in oxygen abundance (12 + log O/H) from 7.14 (Mk 116 = IZw 18) to 8.44 (IC 3258) if temperature fluctuations are taken to be zero.

A correlation between helium and oxygen abundances is found in some of these investigations but not in others, which is hardly surprising because the whole effect is only 25 percent and this is easily swamped in errors of observation and ionization corrections, particularly with regions of differing ionization superposed on the spectrograph slit. The latter effect is least likely to arise in Local Group galaxies, where individual HII regions are well resolved, and we would therefore recommend the numbers given by Lequeux et al., i.e. Y0 = 0.23 for the primordial helium mass fraction and dY/dZ appeq 2, except that these numbers change to 0.22 and 3 respectively if temperature fluctuations are equated to zero, which is close to the results found by French, and this value of Y0 (for which we estimate that the error does not exceed ± 0.01) is in good agreement with less precise values deduced from the luminosities of subdwarfs (Carney 1979) and from the relative populations of the RGB and HB in globular clusters (Iben 1968, Renzini 1977, Caputo, Castellani & Wood 1978). Such a low value imposes severe constraints on big-bang cosmology and elementary particle physics (Tayler 1980). The existence of a dY/dZ effect is in agreement with that of a He/H gradient found in our Galaxy, for example, by Thum, Mezger & Pankonin (1980).

The "simple" one-zone model of galactic chemical evolution predicts the relationship

Z = p ln(m / g)

(Searle & Sargent 1972) between the heavy-element abundance Z (by mass) and the masses m and g of the entire system and of residual gas, respectively, where p is a constant known as the "yield" and is the ratio of mass of heavy elements newly synthesized and ejected to mass locked up in compact remnants and long-lived stars, in each stellar generation. Oxygen abundances in the vast majority of Irregular and blue compact galaxies studied conform to this relationship (but not quite all: Kinman & Davidson 1981) and give a yield of 0.003 ± 0.001 which places constraints on the initial mass function, while the large dY/dZ favors evolutionary models of massive stars with substantial mass loss in the hydrogen and helium-burning phases (Lequeux et al. 1979, Serrano & Peimbert 1981).

Both Z[ appeq 25(O/H)] and m/g increase with the total mass of the system, the data of Lequeux et al., Talent, and Kinman & Davidson being roughly represented by the relationship (assuming zero temperature fluctuations)

Z appeq 0.0025 log M8;     0.0 < log M8 leq 2.6

where M8 is the mass in units of 108 solar masses. (Lequeux et al., who include a temperature fluctuation correction, derive a coefficient of 0.004.) A composite mass-abundance diagram is shown in Figure 7.

The Ne/O ratio appeq 0.2 in all HII regions and Ar/O is probably between 5 × 10-3 (Meyer 1979) and 10-2 (Pagel et al. 1978, Pagel, Edmunds & Smith 1980); S/O has been found to decrease with increasing O/H (French 1980, cf. Talent & Dufour 1979), but French's result is affected by the use of an incorrect ionization correction scheme (cf. Pagel 1978, Natta, Panagia & Preite-Martinez 1980) and the true state of affairs is still uncertain. C/O may have about the same (or slightly lower) value in the SMC as in the Sun or Orion (Gilra et. al. 1980). Thus, observed elements attributed to helium, carbon, and oxygen burning apparently vary in lockstep, with the possible exception of sulphur.

The fact that nitrogen is over-deficient in the oxygen-deficient Magellanic Clouds, compared to the solar neighborhood, appeared some years ago to support the classification of N as a secondary nucleosynthesis product with (roughly) N/O propto O/H (Talbot & Arnett 1974), but a comparison of N/O with O/H in different objects in different galaxies leads to a much more complicated picture (Pagel et al. 1978, Edmunds & Pagel 1978, Alloin et al. 1979). Figure 3 shows the data as they are at present. (1)

Figure 3

Figure 3. Relation between log(N/O) and log(O/H) in galactic and extragalactic HII regions. The following sources of data are used: Galaxy, Talent & Dufour (1979); M101 and M33, Smith (1975) reanalyzed using the ([O II] + [O III]) / Hbeta calibration of Pagel, Edmunds & Smith (1980); M83, Dufour et al. (1980) using electron temperatures consistent with their oxygen abundances; NGC 300 and 1365, Pagel et al. (1979), M. G. Edmunds and B.E.J. Pagel (in preparation); NGC 7793, M.G. Edmunds and B.E.J. Pagel (in preparation); Irregular and compact galaxies, Lequeux et al. (1979), Pagel, Edmunds & Smith (1980), Talent (1980), French (1980); A0035, Fosbury & Hawarden (1977); Groombridge 1830, Tomkin & Bell (1973), Sneden, Lambert & Whitaker (1979), Galactic nuclear regions are marked by circles round the corresponding symbols. The error bars are schematic overall average estimates.

Overall, some correlation exists, with a 45° line forming a lower envelope to a good fraction of the data; but the scatter is large and is unlikely to be purely observational since, for example, the Magellanic Clouds and NGC 6822 are among the best-studied objects available (Peimbert & Torres-Peimbert 1974, 1976, Dufour 1975, Pagel et al. 1978, Aller, Keyes & Czyzak 1979, Lequeux et al. 1979, Pagel, Edmunds & Smith 1980, Talent 1980) and neither the irregulars (taken as a group) nor the HII regions of our Galaxy show a significant trend. We believe that varying amounts of primary nitrogen are needed to account for the data, as discussed in Section 3.2.

1 We have not included I Zw 18 as only an upper limit is available, log N/O < -1.3 (French 1980), CG 1116+51 for which the temperature is uncertain (French 1980), and VII Zw 403 for which only photographic spectrophotometry is available (Tully et al. 1981), but we do include data for Spirals in which it may or may not be relevant that the Simple galactic chemical evolution model does not apply. Back.

Next Contents Previous