Chemical Evolution of Galaxies: a problem of

Astroarchaelogy

Francesca Matteucci,

Trieste University

Lubljana, February 24, 2014

Chemical Evolution of Galaxies

Beatrice Tinsley (27 January 1941- 23 March

1981) She started the field of

galactic chemical evolution

Collaborators:

Outline of the talk Definition of galactic chemical evolution Basic ingredients of chemical evolution The death of stars: planetary nebulae and

supernovae The effects of element production by stars

in galaxies Summary

Chemical Evolution

Chemical abundances tell us about the nucleosynthesis as well as the formation and evolution of galaxies

Light elements (H, D, He, Li) were synthesized during the Big Bang

All the elements with A>12 were formed inside stars

Stars produce new elements and then restore them into the ISM. This process is chemical evolution

Basic Ingredients of Chemical Evolution

Initial conditions (open, closed-box; initial chemical composition)

The stellar birthrate function: SFRxIMF The stellar yields (i.e. the mass restored

into the ISM by a star of a given mass in the form of a given chemical element) by stellar winds and supernovae

Gas flows: infall, outflow, inflow

The Star Formation History

The star formation history depends on the SFR (star formation rate) and IMF (initial mass function)

The IMF describes the distribution of stellar masses at birth, the SFR tells how much gas goes into stars per unit time

The IMF (power law) has been derived in the solar vicinity (Salpeter, 1955; Scalo1986; Kroupa & al. 1993 ; Kroupa 2000+ others..)

The parametrization of the SFR

The most common parametrization is the Schmidt (1959) law, where the SFR is proportional to some power (k=2) of the gas density

Kennicutt (1998) suggested k=1.4 from studying star forming galaxies. The quantity nu is the efficiency of star formation, expressed in t-1

Kennicutt’s law

Stellar Yields Low and intermediate mass stars (0.8-8

Msun): produce He, N, C and heavy s-process elements (e.g. Ba). They die as C-O WDs, when single, and can die as Type Ia SNe when binaries

Massive stars (M>10 Msun, core-collapse SNe): they produce alpha-elements (O, Mg..), some Fe, light s-process elements and perhaps r-process elements? Merging of neutron stars a possible source

Type Ia SNe produce mainly Fe (0.6-0.7 Msun per SN) and Fe-peak elements(?)

Metal dependent stellar yields (O, Fe) from core-collapse SNe (Nomoto & al.06)

Supernova Progenitors

Core-collapse SNe originate from massive stars (M> 8Msun), they can be of Type II,Ib and Ic.They leave neutron stars or black holes as remnants. Type Ib/c SNe, M>30-40Msun , if single and 12-40Msun if binaries, are connected to GRBs. PCSNe M>100Msun (no remnant)

Type I a SNe are thought to originate from C-O white dwarfs (WD) in binary systems. Exploding by C-deflagration when reaching the Chandrasekhar mass. They do not leave any remnant

Different supernovae

SN Ia (artistic view) SN II (Cassiopea A)

Type Ia Supernovae Single-degenerate scenario (e.g. Whelan & Iben

1974; Hachisu et al. 1996; Han &Podsiadlowski 2004): a binary system with a C-O white dwarf plus a MS star. Minimum time for explosion 35 Myr!

Double-Degenerate scenario (Iben & Tutukov, 1984): two C-O WDs merge after loosing angular momentum due to gravitational wave radiation. Minimum time 36 Myr! Merging of He WD and C-O WD also possible

Empirical delay time distributions (DTDs) (Mannucci & al. 2006:Strolger & al. 2004; Totani & al. 2008; Pritchett & al. 2008)

Gas flows

Gas can be accreted or lost from a galaxy The most common parametrization of the

accretion rate is an exponential law

Gas Flows

Outflows and galactic winds are seen in galaxies

The most common parametrization is

Wind=CxSFR

The wind rate is proportional to the star formation rate (Martin, 2000,2004)

Basic Equations

All these ingredients should be included in equations describing the evolution of the abundances of single chemical elements in the gas

These equations should be solved numerically if one wants to take into account in details stellar lifetimes

Basic Equations

The Milky Way

The Formation of the MW from Astroarchaelogy

The two-infall model of Chiappini, FM & Gratton (1997) predicts two main episodes of gas accretion

During the first one the halo, bulge and all or part of thick disk formed, the second gave rise to the thin disk

This model can fit most of the observational data

Main Assumptions for the Two-Infall Model

SFR- Kennicutt’s law with a dependence on the surface gas density (exponent k=1.5) plus a dependence on the total surface mass density (feedback). Threshold in the SF in the thin disk

Type Ia SN progenitors from the SD model (Greggio & Renzini 1983; FM & Recchi 2001). This scenario and the DD one are the best to reproduce chemical properties (FM & al. 2006; 2009). Yields from Woosley& Weaver 95 and van den Hoek & Groenewegen 97

IMF Scalo (1986) normalized over a mass range of 0.1-100 Msun

Predicted SN rates in the MW in the two-infall model

Type II SN rate (blue) follows the SFR

Type Ia SN rate (red) increases smoothly

(small peak at 1 Gyr) SD as Type Ia

progenitors Oscillations are due

to the threshold in SF

The G-dwarf Metallicity Distribution in Solar Vicinity

G-dwarf metallicity distribution compared with predictions of the two-infall model (Kotoneva et al. 2003)

The assumed timescale (8 Gyr) for disk formation is a very important parameter

Abundance ratios in the MW, Solar Vicinity (Francois, FM &al. 2004)

According to the time-delay model

the ratios [X/Fe] vs. [Fe/H] can be explained by the different timescales on which different elements are produced

The alpha-elements (O,Mg,Si, S, Ca) are produced in core-collapse SNe on short timescales (from 1 Myr to 35Myr), whereas Fe is mainly produced on longer timescales (from 35 Myr to a Hubble time), but a fraction of it is also produced in core-collapse SNe

This interpretation explain the behaviour shown in the figure and it applies to any chemical element

Effect of different yields

Predicted and observed [X/Fe] for two different sets of yields

Dashed lines:yields from Woosley & Weaver (1995) for massive stars and from van den Hoeck & Groenewegen (1997) for low and intermediate mass stars

Continuous lines: yields from Karakas (2010) for low and intermediate mass stars and from Geneva group for He, C, N,O with rotation and from Kobayashi & al. ( 2006) for heavy elements

Models from Romano & al. 2010

Romano & al. 2010. Best yields: dashed line. Solar ab. Asplund & al. 2009

The time-delay model and different star formation rates

Predicted [alpha/Fe] ratios for different SFR histories (a more modern version of FM & Brocato 1990)

A strong starburst (dashed line), a SFR like in the solar vicinity (dotted) and a slow SFR (continuous) like in Magellanic Irregulars or Dwarf Spheroidals

The Galactic Bulge

Model (black, Cescutti & FM 2011): fast Bulge formation (0.3 Gyr) and IMF flatter than in S.N.

Predicts large Mg to Fe for a large Fe interval

Turning point at larger than solar Fe (effect of the time-delay model)

Data from giants and dwarfs. Bensby & al. 2010; Alves-Brito & al. 2010

Two Bulge populations?

Two populations: i) classical bulge, fast formation (0.3 Gyr), ii) younger stars related to the bar, longer formation (3-4 Gyr)

Model by Grieco, FM & al. 12 compared to Hill & al’s (2011) data

Salpeter IMF MDF convolved with

an error of 0.25 dex

Abundance Gradients in the Galactic Disk Predicted and observed

abundance gradients from Chiappini, FM & Romano (2001). Disk forms inside-out:

tau(R)=1.033R(kpc) -1.27 Gyr. A SF threshold is considered (7Msunpc-2)

Data from HII regions, PNe and B stars, red dot is the Sun

The gradients slightly steepen with time (from blue to red)

Abundance Gradients in the Galactic Disk Different models with

and without radial inflows (Spitoni & FM 2011)

Black line, no flows no threshold inside-out, red line radial flows of speed 1 Km/sec

Blue line a variable speed for radial flows

Gradient inversion at z=3 with and without radial flows

Summary from Astroarchaelogy of the Milky Way The inner stellar Halo must have formed on a timescale

of 0.8-1.5 Gyr whereas the outer Halo must have formed on a longer timescale (inside-out Halo formation)

The local disk must have assembled by gas accretion on a time scale from 6-8 Gyr and the timescale increases with galactocentric radius

The Bulge stars of the classical component must have formed on a time no longer than 0.3-0.5 Gyr with flatter IMF. The stars related to the bar should have formed on longer timescales

A gradient inversion at z=3 is predicted by the two-infall model with and without radial flows. Observed by Cresci & al. (2010). It is due to the inside-out formation

Dwarf Spheroidals of the LG vs. Milky Way (Shetrone+02)

Dwarf Spheroidals vs. Milky Way

The [alpha/Fe] ratios in dSphs evolve differently as a function of [Fe/H]

There is some overlapping of the

[alpha/Fe] ratios only at low metallicity Can the dSphs have been the building

blocks of the Galactic halo? Chemical abundances should reveal it

Chemical Evolution of Dwarf Spheroidals Chemical models for dSphs (Lanfranchi &

FM, 2004) suggest that they evolved slowly with a low SF efficiency and suffered strong winds

SF lasted for long periods either continuous or in bursts

This explains the sharper decrease of the [alpha/Fe] ratios relative to the MW

Carina: data vs. models

Best Model for Carina by Lanfranchi+( 2006) compared with data from Koch + (2008)SF history taken from Color-Magnitude diagramFour bursts of SF with SF efficiency 0.15Gyr-1

Sculptor: data vs. models

Data for Sculptor (Kirby et al. 2009, Shetrone et al. 2003, Geisler et al. 2005)

One extended episode of SF, strong wind: models from Lanfranchi & FM (2009) in red (SF eff. 0.2 Gyr-1 )

The Ultra Faint Dwarfs-Koch & al. (2012) Better candidates for

the Galactic halo are theultra-faint-dwarfs (UFD) ?

Stars refer to the UFD Hercules, black points are the MW and blue points the dSphs

Models with very low SF efficiency (from 10-4 to 0.1Gyr-1 )

The Galactic Halo and First Stars Caffau & al. (2011) have discovered a halo star

with Z <4.5 10-5 Zsun! Keller+2014 a star with [Fe/H]<-7.0!

Is the existence of massive PopIII stars compatible with that metallicity?

PopIII stars should have formed at z=20, been massive (Bromm 2002) from 10Msun and possibly >200Msun (PCSNe) for Z< Zcrit=10-4 Zsun

The Caffau star is below the Zcrit for passing from PopIII to low mass PopII

Are there PopIII stars with low mass? (see Clark+ 2011)

Galactic Halo MDF: the effect of PopIII stars (Brusadin,FM&al.13) Predicted and observed

MDF in the Galactic halo (Brusadin, FM&al. 2013)

PopIII stars are present in both models with different IMFs

Case I: stars from 10 to 260 Msun, one slope x=1.7 for Z<Zcrit

Case II: IMF as in Salvadori & al. (2007) one slope x=1.35 but variable low cut-off

Summary Astroarchaeology by means of chemical abundances

can allow us to impose constraints both on stellar nucleosynthesis and star formation hystory

We can reasonably well reproduce the chemical abundances in the solar vicinity. For some elements the yields are still very uncertain

The roles of SNe of different Types (time-delay model) help in interpreting abundance patterns

Prompt Type Ia SNe seem to be necessary in all galaxies to fit chemical abundances

A crucial parameter in galaxy evolution is the efficiency of star formation

The MDF of the halo is best fitted without massive PopIII stars

Summary

Dwarf spheroidals and ultra faint dwarfs in the Local Group should have suffered a star formation rate with much lower efficiency than in the Milky Way, as well as strong galactic winds

It is difficult to assess whether these galaxies have been the building blocks of the MW halo which can be well reproduced by in situ SF

The bulge of the MW seems more complex than previously thought. Two different populations are identified, one of which is old and formed fast

The MW disk could have formed inside-out and this implies an inversion of the O gradient at high redshift, due to the strong inner infall at early times

Abundance gradients can put constraints on the mechanism of disk formation