GALAXIES, CHEMICAL EVOLUTION MONICA TOSI The study of the chemical evolution of galaxies includes the various processes which have led to the present chemical abundances observed in the stars and the interstellar gas of galaxies. This study started essentially in 1963 with a seminal paper by Maarten Schmidt. The evolution of a galaxy of any type is due to several complicated and interconnected processes. For the sake of simplicity astronomers tend to divide them into three major categories and to treat the photometric, dynamical, and chemical evolution separately. However, to fully understand the actual evolution of galaxies we will need to take all of them into account at the same time. OVERVIEW The main goal in modeling the chemical evolution of galaxies is to provide a correct interpretation of the element abundance distributions observed in their stars and interstellar media. In addition, comparison between observed galactic properties and theoretical predictions has frequently put important constraints on the theories of star formation, stellar nucleosynthesis, evolution, etc. The main processes governing chemical evolution are summarized in Fig. 1, which is adapted from fundamental review by Beatrice Tinsley. As indicated in the sequence of the figure, when the protogalactic cloud reaches a critical density which may vary from one galaxy to another, it starts to fragment and form stars of various masses (ranging essentially from 0.1 to 100 times the mass of the Sun). These stars evolve and bum their nuclear fuel, synthesizing in their interiors chemical elements of progressively higher atomic weight. The original chemical composition of the stars, mostly hydrogen and helium, changes in favor of heavier elements like carbon, oxygen, and iron. During their life and mainly at their death, stars eject most of their initial mass, thus polluting the circumstellar medium with the new elements produced in their interiors. From this medium, modified in mass and metallicity, new stars form, with an initial chemical composition and, therefore, with a subsequent evolution somewhat different from that of the previous stellar generation. In principle this cycle may last forever, leading to a continuous decrease of the abundance of hydrogen and to a corresponding increase of that of heavier elements. However, as it has been found that in every stellar generation the fraction of formed stars is a decreasing function of their mass (i.e., in a single generation there are more low-than high-mass stars) most of the gas remains trapped in low-mass, long-lived stars and cannot contribute to the cycle for a long time. Besides, a dying star never returns all its mass back to the medium and always leaves a dead remnant (white dwarf, neutron star, or black hole, depending on the initial mass) unable to contribute to the productive cycle. After a few stellar generations, the situation can remain frozen for several billion years (or even forever through the exhaustion of the gas reservoir). Meanwhile, other phenomena can effect the cycle and thus the chemical evolution of a galactic region. Some gas can fall in from outside, either a residual of the collapse of the protogalactic cloud, or intergalactic gas trapped by the gravitational force of the galaxy. Some gas, on the other hand, can also leave the region, for instance, when supernova explosions produce winds effective enough to sweep away the surrounding medium. Finally, the presence of nearby galaxies can provoke strong interactions that may even produce a merger of two or more systems and in any case that will significantly alter the star formation rate of the galaxy. Modelers of galactic chemical evolution take into account all these possible phenomena by treating them as free parameters of the problem. Modelers borrow from stellar evolution theory the stellar lifetimes and the amount of elements synthesized or burnt by stars of different mass, "cook" all the ingredients together, and search for the recipe that provides objects as similar as possible to actual galaxies. The most important parameters for building a model of a galactic region are the star formation rate (SFR, i.e., the amount of gas mass which goes into stars per unit time), the initial mass function (IMF, i.e., the fraction of stars of each mass formed in a stellar generation), the amount and the rate of gas flows in and out of the region, and the amount of newly formed elements ejected by stars (stellar yields). Some parameters of the stellar evolution theory are still rather uncertain and may consequently be treated as variables in the chemical evolution models. The free parameters are therefore quite numerous, and the only case where they do not outnumber the observational constraints is that of our own galaxy. For external galaxies more observational data are still necessary to properly constrain models. MODELS FOR THE VARIOUS TYPES OF GALAXIES As is well known, galaxies are usually classified in three major types: ellipticals, spirals, and irregulars. This classification was originally based only on morphological characteristics, but it actually corresponds to profound differences in chemical evolution as well. The main features characterizing these three types of galaxies that a chemical evolution model must account for are the following. (a) Elliptical galaxies are more metal rich than the Sun, have very little or no gas left, and show an average stellar population older than that around the Sun. All these facts suggest that they are evolved objects, with no, or very little, recent activity of star formation. (b) Spiral galaxies show two distinct populations: (1) metal poor, old stars, associated with a gas-poor halo; (2) stars of any age and with metallicity of the same order of magnitude as the Sun, imbedded in a significant amount of gas (ranging roughly between 2 and 20% of the total disk mass) in the disk (1) is generally interpreted in terms of a first rapid phase of star formation that has generated the halo population. Most of the gas, however, must have escaped this star formation activity, and has concentrated into a disk where the star formation seems to be in a sort of stationary regime. (c) Irregular galaxies have a much higher gas fraction (mass in interstellar gas relative to the total mass in gas and stars) than the other galaxies, have chemical abundances significantly lower than solar, and show a large fraction of young stars, thus indicating that they are still in an early stage of evolution. The most up-to-date scenarios for the evolution of the different types of galaxies can be summarized as follows. ELLIPTICAL GALAXIES All the models that are in agreement with the observational features of ellipticals predict, right after galaxy formation, a rapid very strong episode of star formation, which consumes almost all the gas. After a certain amount of time, which increases with the galactic mass but never exceeds 1 billion years, the explosion of a large number of supernovae in a relatively short time triggers galactic winds, which remove all the residual gas. This prevents any further star formation and explains the average age of the stars presently observed. Despite the rapidity of these events, the star formation activity has been so high until this point that the system has already reached a metallicity even larger than solar (it has taken 10 times longer for our galaxy to produce the Sun, with its metallicity). After this major episode of galactic wind, the only gas available in the galaxy is that which is returned little by little by dying stars. Since there is no diluting medium, this gas is very metal rich. It can reach a present mass around ******** solar masses (i.e., ******* the mass of the galaxy) and is presumably recognizable in the hot gaseous coronae sometimes observed around ellipticals. SPIRAL GALAXIES Given the old age of all the halo stars of our galaxy and nearby spirals, we can easily presume that early star formation activity in these galaxies was even more rapid than in ellipticals. The presence of the disk, however, strongly modifies the situation, because the major gas reservoir lies in this case in a potential well deep enough to protect it from the supernova winds in the halo. A steady state of moderate star formation can then start in the disk, with more activity in the central, denser regions and less in the outer ones. In fact, the SFR presently derived in nearby spirals from indicators like supernovae and pulsars is roughly exponentially decreasing with increasing distance from the galactic center. As time goes by, this difference in the SFR of different regions produces relevant effects on their properties: Central regions show lower values of gas-to-total-mass ratios and larger chemical abundances, whereas in the outermost layers of the disk the fraction of gas is still up to 50% and the metallicity is lower than solar. The immediate results of this phenomenon are the radial abundance gradients of heavy elements observed in the disks of spirals. These gradients are always negative (i.e., any element heavier than hydrogen is more abundant in the inner than in the outer regions), but their slope varies from galaxy to galaxy and from element to element. No tight correlation has been found between the variation with time of the SFR and the galaxy morphological subtype, but there is some indication that in early-type spirals like M31 the SFR has slowly decreased, whereas in late-type objects like M33 it has been practically constant since galaxy formation. Another important parameter in the chemical evolution of spiral galaxies is infall. An infall of gas from outside the system was first suggested by Richard Larson in 1972 to explain why so many galaxies have not yet exhausted their gas through star formation processes. The observational evidence of high velocity clouds, gas condensations falling toward our galaxy, supported his idea of gas replenishment. Currently infall is still the best way to interpret not only the gas survival in galactic disks, but also the steep abundance gradients of some spirals and local features like the so-called G-dwarf problem. In fact, if the infall density is roughly uniform and its metallicity low, as would be reasonable in the case of intergalactic gas attracted in the disk potential well, its ratio to the SFR increases toward the outer regions, thus diluting them more efficiently and leading to a steepening of the metallicity gradient. Moreover, the observational evidence of gas clouds falling into a galaxy, which was formerly restricted to the Milky Way, is now found in other spirals as well (e.g., M101). There are only about 10 nearby spirals studied in enough detail at the appropriate wave-lengths to put reliable constraints on the models. For these objects it is found that a conspicuous amount of infall (around 1-2 solar masses per year) is predicted for the more massive galaxies like M101 and NGC 6946, but very little or none is required for the smaller ones like M33 and NGC 2403. This is naturally consistent with the idea that the infall is due to diffuse gas captured by the gravitational force of the disk. IRREGULAR GALAXIES Gas flows seem to be very important also in the evolution of these small diffuse objects, the irregular galaxies. In this case, however, the main flow is outwards and the galaxy tends to lose most of its gas, because the potential well of an irregular is low enough to allow a strong galactic wind triggered by supernova explosions. In fact, the models predict stronger winds for the lower mass irregulars. The reason these gas outflows seem to be necessary in the evolution of irregulars is that these galaxies show such low metal abundances that we are led to conclude that most of the enriched gas ejected by the dying stars is not retained by the galaxies. The low mass and metallicity of these galaxies have also suggested to some theorists that star formation cannot have been continuous as in spiral galaxies, but has been confined to short-lived episodes or bursts, or that it has increased with time and has only recently reached the present value. There is no definite conclusion on this subject, but certainly small galaxies like the closest dwarf irregulars in the Local Group cannot have sustained a star formation rate of the intensity observed now for their entire lifetimes. There are even some galaxies, called blue compact galaxies because of their appearance, that show incredibly strong star formation activity but have metal abundances so low as to suggest that they are perhaps now experiencing their first burst (e.g., the galaxy No. 18 in Fritz Zwicky's catalogue). Additional Raeding Diaz, A.I. and Tosi, M.(1986). The origin of nitrogen and the chemical evolution of spiral galaxies. Astron. Ap. 158 60. Gilmore, G., Wyse, F.G., and Kuijken, K.(1989). Kinematics, chemistry, and structure of the Galaxy. Ann. Rev. Astron. Ap. 27 555. Larson, R.B.(1972). Infall of matter in galaxies. Nature 236 21. Matteucci, F.(1984). The chemical enrichment of galaxies. ESO Messenger (No. 36) 17. Schmidt, M.(1963). The rate of star formation. II. The rate of formation of stars of different mass. Ap. J. 137 758. Tinsley, B.M.(1980). Evolution of the stars and gas in galaxies. Fundam. Cosmic Phys. 5 287. See also Galactic Structure, Stellar Populations; Galaxies, Disk Evolution; Galaxies, Formation.