Gas and star flows

Infalls and outflows change the [X/Fe] ratios depending on the abundances and the amount of gas of the flows. Rich outflows with SN II material reduce [X/Fe] as opposed to the increment required by most of the observed trends. Rich outflows with SN Ia material increase [X/Fe] but they also decrease [Fe/H], preventing the formation of metal-rich stars. Infall of a metal-rich gas overabundant in elements present in metal-rich stars, such as Na, can reproduce the rise in [X/Fe], but how does the infalling gas acquire those [X/Fe] values?

Brook et al. (2005) explain the chemical properties of thick- and thin-disk stars with models that assume mergers and infalls mainly at redshifts lower than 1, but their results do not predict the abundance trends for [Fe/H] > -0.1.

Reddy (in this meeting; see Chapter 8) showed a secondary peak in that [Fe/H] distribution for [Fe/H] > 0. Inclusion of a significant amount of stars (or gas that triggered the star formation) from a merger event could explain the secondary peak. If the thin-disk metal-rich stars formed in one or several Galactic satellites that settled in the Galactic disk, how did the stars of those satellites reach supersolar [Fe/H] with (super)solar [X/Fe]?

According to the merger scenario, the bulge was also formed by satellites that fell early on into the Milky Way, therefore the old and metal-rich stars of the bulge and

Figure 43.1. Evolution of [X/Fe] versus [Fe/H] predicted by various models: for Al by Timmes et al. (1995), for Ca by Portinari et al. (1998), for Ti, Ni, and Zn by Francois et al. (2004), for O by Gavilan et al. (2005), for Mg, Si, and Cr by Prantzos (2005), for Ba and Eu by Cescutti et al. (2007), and for Na by Izzard et al. (2007). Filled circles: F and G dwarf disk stars (Bensby et al. 2005). Open squares: the most-metal-rich G dwarf bulge star (Johnson et al. 2007). Corrections to [X/H] to take account of the different photometric Solar values assumed by the authors have not been made.

Figure 43.1. Evolution of [X/Fe] versus [Fe/H] predicted by various models: for Al by Timmes et al. (1995), for Ca by Portinari et al. (1998), for Ti, Ni, and Zn by Francois et al. (2004), for O by Gavilan et al. (2005), for Mg, Si, and Cr by Prantzos (2005), for Ba and Eu by Cescutti et al. (2007), and for Na by Izzard et al. (2007). Filled circles: F and G dwarf disk stars (Bensby et al. 2005). Open squares: the most-metal-rich G dwarf bulge star (Johnson et al. 2007). Corrections to [X/H] to take account of the different photometric Solar values assumed by the authors have not been made.

Galactic disk originated from small galaxies, or metal-rich stars form of material from small structures, but again, how did those structures reach [X/Fe] > 0?

According to another scenario the bulge could have been formed by the stars that were born in the inner Galactic disk and were dynamically heated by the bar (Colin et al. 2006). Moreover, the same bar could have been able to produce radial flows of stars from the inner to the outer part of the Galactic disk. Therefore the metal-rich stars of the bulge and the Solar neighborhood formed in the inner disk, but how did the inner disk reach [X/Fe] > 0?

Radial gradients can be powerful tools allowing one to decide whether metal-rich stars observed in the Solar vicinity and the bulge formed in situ or alternatively were formed with in the inner Galactocentric radius or originated from merged satellites. In the most complicated case (or the most realistic), a combination of these three possibilities should be the answer.

Therefore, stellar or gaseous infalls might explain the abundance trends observed for [Fe/H] > -0.1, but they pose the question of how these infalls could acquire those (super)solar [X/Fe] values.

3.2 The star-formation rate

Important changes in the SFR affect the [X/Fe] ratios mainly after a significant star-formation burst (e.g. Carigi et al. 1999, 2002; Chiappini 2001).

The spiral wave is the most important inner mechanism that triggers star formation and that could recently have formed stars from a metal-rich gas. Rocha-Pinto et al. (2000) and Hernandez et al. (2000) inferred the SFH from the color-magnitude diagram for the Solar vicinity. They found (i) a decreasing exponential general behavior of the SFH during the last ~10 Gyr, and (ii) variations from the general behavior related directly to the spiral-wave passages. These results indicate that there were no significant bursts of star formation within the last ~6 Gyr, therefore it is unlikely that a burst could have modified the [X/Fe] slopes.

An important fact is that metal-rich stars with similar [X/Fe] values have been observed in the Solar neighborhood and in the bulge, galactic components with different SFHs. The Solar vicinity had a moderate SFR for 12 Gyr (e.g. Carigi et al. 2005) whereas the bulge formed very quickly, in less than 0.5 Gyr, with a high SFR (Ballero et al., Chapter 48, and Matteucci, Chapter 44, in this volume).

Therefore, I discard changes in the SFR as the explanation of the change in the [X/Fe] slopes for [Fe/H] > -0.1.

3.3 The initial mass function

The initial mass function (IMF) gives the mass distribution of the stars formed in a star-formation burst. This function is parametrized by the slope for various mass ranges and by the lower and upper mass limits of the stars formed. Since [X/Fe] depends strongly on the IMF, a dependence of the IMF on metallicity, density, or gas mass could explain the change of the [X/Fe] slope.

According to Kroupa (Chapter 24 in this volume) there is no evidence that the IMF changes with Z. Nevertheless, Bonnell suggested (in this meeting; see Chapter 39) that the IMF changes with Z: for supersolar metallicities the slope for massive stars could be steeper and the upper limit could be lower than for subsolar metallicities, producing subsolar [X/Fe] values, in contradiction with the observed values.

In a metal-rich gas it is more difficult to create massive stars due to the dependence of the Jeans mass on Z-2/3 and metal-rich stars truncate the star-formation process by the action of their stellar winds. This suggestion could explain the low ionization of the metal-rich H ii regions compared with that of H ii regions with subsolar metallicity: in a metal-rich gas the number of massive stars may be lower, leading to a smaller number of ionizing photons; on the other hand, the lower stellar temperatures of the metal-rich stars help to reduce the number of ionizing photons. Further work on this suggestion needs to be done.

The dependence of the IMF on the gas density should be less important, because the same abundance trends have been observed in Galactic components with different densities (bulge, open clusters, isolated dwarfs), but with a common property: supersolar metallicity.

According to Weidner & Kroupa (2005) the IMF depends on the gas mass. They found that the slope and the upper mass limit change with the gas mass available to form stars. In dwarf galaxies the upper mass limit is lower and the slope in the massive-star range is steeper, producing lower [X/Fe] values than those of normal galaxies for elements synthesized only by massive stars.

Moreover, Carigi & Hernandez (2008) found important effects on the abundance ratios when the IMF is stochastically populated. The [O/Fe] values varied within three orders of magnitude for a stellar population of 500 M© enriching a gas mass of 104M©. This effect could explain the dispersion observed in the abundance ratios, but not the abundance trends.

Therefore, possible modifications to the IMF cannot explain the abundance trends observed for [Fe/H] > -0.1.

3.4 Stellar yields

Since supersolar [X/Fe] values seem to be a common property of stars with [Fe/H] > 0 in Galactic components (thin disk, bulge, open clusters) with different formation histories, the abundance trends can be explained as being due to the stellar yields of (super) solar-metallicity stars. The observed abundances will provide strong constraints on the physical processes taking place in the stellar cores.

The models shown in Figure 43.1 consider different Z-dependent yields for massive stars, low- and intermediate-mass stars (LIMSs), and SN Ia. These stellar yields were computed for Z < Zcan with the exception of the Portinari et al. (1998) yields, but those authors never used their yields for Z — 0.05 because they stopped their computations at [Fe/H] — 0. Cescutti et al. (2006) and Francois et al. (2004) modified the stellar yields obtained from stellar-evolution models in order to reproduce the observed trends for [Fe/H] < +0.1.

Edmunds (during this meeting; see Chapter 46) suggested that stellar yields increasing with Z raise the [X/Fe] values for [Fe/H] > -0.1. The [O, Mg/Fe] values for bulge and thin-disk stars indicate that there is a Z dependence in the ratio of the O to Mg yield.

Meynet et al. (Chapter 36 in this volume) show that non-rotating massive-stars with Z = Zcan and a high mass-loss rate eject more C than O, and that these stars are an important source of He, C, Ne, and Al, but not so much of O. These facts could explain the decrease in [O/Fe] with increasing [Fe/H] while the [Al/Fe] values remain almost constant for [Fe/H] > 0.

The significant change in the [Na/Fe] slope suggests that there is an extra source of Na production. Assuming that SN II and AGB stars produce Na, Izzard et al. (2007) reproduce the [Na/Fe] values for [Fe/H] < -0.2, but fail to reproduce the increase in [Na/Fe] for [Fe/H] > 0. They suggest that the change in the [Na/Fe] slope may be explained by invoking secondary Na produced by SNII. Nevertheless, according to Frohlich (see Chapter 37) the effect of core-collapse supernovae cannot explain the increase in [Na/Fe] observed for [Fe/H] > 0.

Another channel that contributes to the enrichment of a metal-rich gas is provided by SN Ia. According to Yoon (Chapter 38 in this volume) there are various scenarios for SN Ia with different time delays, but the amount of heavy elements ejected is similar for all of the scenarios. The role of rotation might be important in the production of chemical elements, but this effect has not been studied yet. Therefore, new stellar yields for massive stars and intermediate-mass stars of Solar and supersolar metallicity are required.

3.4.1 The importance of stellar winds in metal-rich stars

One of the most important problems in the chemical evolution of galaxies is that of C production. Carigi et al. (2005, 2006) have studied the contributions to C enrichment by massive stars and LIMSs in various types of galaxies. We have found that the massive stars have contributed 48% and 36% of the total C produced in the Solar neighborhood and in the dIrr galaxy NGC 6822, respectively. The difference is due to the effect of Z on the stellar winds of massive stars. Massive stars of Solar Z eject more C than do those of subsolar Z through stellar winds (see Meynet et al., Chapter 36, and Crowther, Chapter 29, in this volume).

Carigi et al. (2005) made a chemical-evolution model of the Galaxy in which they assumed that the metal-rich stars behave like stars of Zcan. They are able to reproduce the C/O and O/H values of the Solar vicinity as well as the O/H and C/O gradients observed by Esteban et al. (2005) but cannot reproduce the decrease in [C/Fe] for [Fe/H] > -0.1 shown by Bensby & Feltzing (2006) and Allende-Prieto in Chapter 3 (see Figure 43.2).

Figure 43.2. Model predictions by Carigi (2000) (dashed lines) and Carigi et al. (2005) (continuous lines) considering yields of metal-rich massive stars published by Portinari et al. (1998) and Maeder (1992), respectively. Left panels: [C/O, Fe] evolution with [O, Fe/H] in the Solar vicinity. Right panels: present-day ISM abundance ratios as a function of Galactocentric distance. Open circles: galactic H ii regions, gas plus dust, by Esteban et al. (2005) and Carigi et al. (2005). Star: extragalactic H ii region (H1013) inM101 (Bresolin 2007). Filled triangles: F and G dwarf disk stars (Bensby & Feltzing 2006). Filled squares: dwarf stars (Akerman et al. 2004). Photometric Solar values from AGS05 are considered, except for the data published by Bensby & Feltzing (2006) because they assumed their own Solar abundances.

Figure 43.2. Model predictions by Carigi (2000) (dashed lines) and Carigi et al. (2005) (continuous lines) considering yields of metal-rich massive stars published by Portinari et al. (1998) and Maeder (1992), respectively. Left panels: [C/O, Fe] evolution with [O, Fe/H] in the Solar vicinity. Right panels: present-day ISM abundance ratios as a function of Galactocentric distance. Open circles: galactic H ii regions, gas plus dust, by Esteban et al. (2005) and Carigi et al. (2005). Star: extragalactic H ii region (H1013) inM101 (Bresolin 2007). Filled triangles: F and G dwarf disk stars (Bensby & Feltzing 2006). Filled squares: dwarf stars (Akerman et al. 2004). Photometric Solar values from AGS05 are considered, except for the data published by Bensby & Feltzing (2006) because they assumed their own Solar abundances.

Carigi (2000) made a model considering yields of massive stars published by Portinari etal. (1998)for Z = 2.5 Zcan and predicted a C/O gradient flatter than the observed one. The flattening of the gradient is due to the increase in the mass-loss rate with Z (aZ0 5) assumed by Portinari et al.; consequently metal-rich massive stars are stripped before C is synthesized and their C yields are lower than those of stars with Z = Zcan.

According to Meynet et al. (Chapter 36 in this volume) rotating stars with Z = Zcan are more efficient at ejecting C and O than are rotating stars with Z = 2Zcan. This could explain the decrease in [C/Fe] exhibited by metal-rich stars of the thin disk, but the C contribution due to LIMSs must be included also in order to have a complete picture of the evolution at high Z.

In order to reproduce the high C/O values for inner Galactocentric radii the mass-loss rate for metal-rich stars has to be lower than that assumed by Portinari et al. (1998). On the other hand, to reproduce the low C/Fe values for [Fe/H] > 0 in the Solar neighborhood the mass-loss rate has to be higher than that assumed by Portinari et al. Puis (during this meeting; see Chapter 31) gave limits for the mass-loss rate, which varies as Z0 62 ± 015.

Consequently, there is an inconsistency between theory and observations regarding the behavior of C/O and C/Fe for high metallicities and a more complex explanation is needed.

Was this article helpful?

0 0

Post a comment