Deriving galaxy ages and metallicities using 6dF

Deriving galaxy ages and metallicities using 6dF

Deriving galaxy ages and metallicities using 6dF

6dFGS Workshop April 2005

Rob Proctor (Swinburne University of Technology)

Collaborators:Philip Lah (ANU)

Duncan Forbes (Swinburne University of Technology)

Warrick Couch (UNSW)

Matthew Colless (AAO)

6dFGS Workshop April 2005

Rob Proctor (Swinburne University of Technology)

Collaborators:Philip Lah (ANU)

Duncan Forbes (Swinburne University of Technology)

Warrick Couch (UNSW)

Matthew Colless (AAO)

Aim and OutlineAim and Outline• Aim: • To test theories of galaxy formation

using galactic-archeology.

• Outline: • The challenges.• Our approach to them using 6dFGS data.• Some preliminary results.• The future.• Some conclusions

• Aim: • To test theories of galaxy formation

using galactic-archeology.

• Outline: • The challenges.• Our approach to them using 6dFGS data.• Some preliminary results.• The future.• Some conclusions

The challengesThe challenges• The age-metallicity degeneracy:

• Young, metal-rich populations strongly resemble old, metal-poor populations.

• The age-metallicity degeneracy:• Young, metal-rich populations strongly

resemble old, metal-poor populations.

Age=6 Gyr , [Fe/H]=0.2

Age=12Gyr,

[Fe/H]=0.0

15 Gyr

1.0 Gyr

1.5 Gyr

[Fe/H]=-0.4

[Fe/H]=-2.252.0 Gyr 7 Gyr

Models: Bruzual & Charlot (2003) Models: Sanchez-Blazquez (Ph.D. thesis); Vazdekis et al. 2005 (in prep)

The challengesThe challenges• Abundance-ratio variations (e.g.

[Mg/Fe] †)

†[X/Y]=log(NX/NY)*-log(NX/NY)

• A new opportunity?

•‘’-element abundance ratios in stellar populations are indicators of the time-scale of star formation.

Lick indices (Worthey 1994)Lick indices (Worthey 1994)

• 25 spectral features with a variety of sensitivities to age, overall metallicity ([Z/H]) and ‘’-element abundance ratio ([Mg/Fe]).

• Models of Thomas, Maraston & Korn (2004) used here.•Model simple stellar populations (SSPs) with ages up to 15 Gyr and [Z/H] from -2.25 to +0.4 dex.•‘’-element abundance ratios of from -0.3 to +0.5 dex modelled using the spectral synthesis of Tripicco & Bell (1995)

• 25 spectral features with a variety of sensitivities to age, overall metallicity ([Z/H]) and ‘’-element abundance ratio ([Mg/Fe]).

• Models of Thomas, Maraston & Korn (2004) used here.•Model simple stellar populations (SSPs) with ages up to 15 Gyr and [Z/H] from -2.25 to +0.4 dex.•‘’-element abundance ratios of from -0.3 to +0.5 dex modelled using the spectral synthesis of Tripicco & Bell (1995)

• Differences in sensitivities leads to the breaking of the age/metallicity degeneracy.

Breaking the degeneracy with Lick

indices.

Breaking the degeneracy with Lick

indices.N=1200

• Data require extrapolation of models in metallicity

• A population apparently older than 15 Gyr.

• Observational error.

• Modelling uncertainties.• Horizontal-branch morphology?

Age =1 GyrZ

=0.5

Age=15 Gyr

Z=-2.25

Our approach.Our approach.

• Employ as many indices as possible (up to 25) in the derivation of galaxy properties using a 2-fitting procedure (Proctor & Sansom 2002; Proctor et al. 2004a,b).• This:

• Minimises effects of most reduction and calibration errors (sky-subtraction, flux calibration, stray cosmic rays, poor calibration to Lick system etc).

• Minimises effects of modelling errors.

• Utilises the fact that ALL indices contain SOME information about age, [Fe/H], [/Fe] and [Z/H] (Proctor et al. 2005).

• Provides some of the most reliable age and metallicity estimates from integrated spectra to-date (I.e. work to low S/N).

• Employ as many indices as possible (up to 25) in the derivation of galaxy properties using a 2-fitting procedure (Proctor & Sansom 2002; Proctor et al. 2004a,b).• This:

• Minimises effects of most reduction and calibration errors (sky-subtraction, flux calibration, stray cosmic rays, poor calibration to Lick system etc).

• Minimises effects of modelling errors.

• Utilises the fact that ALL indices contain SOME information about age, [Fe/H], [/Fe] and [Z/H] (Proctor et al. 2005).

• Provides some of the most reliable age and metallicity estimates from integrated spectra to-date (I.e. work to low S/N).

• Use Lick indices to estimate luminosity-weighted age, [Fe/H], [/Fe] and [Z/H] for ~5000 6dFGS DR1 galaxies (Already ~50x larger than any previous study of its kind).

Results from 6dFGS spectra: Emission

Results from 6dFGS spectra: Emission

Ref………..

H, OII and NII emission strengths supplied by Philip Lah.

• Use emission to isolate a sample dominated by early-type galaxies.

• From ~35,000 DR1 galaxies with index measurements we find:

• 9000 with S/N>15• 5000 emission free

• 2000 with HII region emission

• 2000 with AGN emission

• 3000 with S/N>22• 1800 emission free

• 600 with HII region emission

• 600 with AGN emission

HII regions AGN

6dFGS: Age with velocity dispersion

6dFGS: Age with velocity dispersion

• Both AGN and HII region galaxies lower velocity dispersion (mass) than the emission free.

• Emission line galaxies dominate at low velocity dispersion.

• Consistent with the notion that we are excluding late-type galaxies.

N=7500

6dFGS: Age with velocity dispersion

6dFGS: Age with velocity dispersion

• Suggests a mass-age correlation in opposite sense to hierarchical collapse models of Kauffmann (1996).

• i.e. Highest mass galaxies tend to be old.

• Range of ages inconsistent with models of primordial collapse.

BUT………..

N=3000

6dFGS: Age with velocity dispersion

6dFGS: Age with velocity dispersion

N=3000

“Frosting”• A busrt of SF of only a

few % of galaxy mass can easily provide the majority of the sampled luminosity.

e.g. NGC 821: Proctor et al. 2005

NGC 821

N=2500

6dFGS: Age with velocity dispersion

6dFGS: Age with velocity dispersion

• Sampling effects probably cause apparent age-mass relation.

• Recall sample is essentially

luminosity limited..

• Can infer Forbes & Ponman (1999) finding that young galaxies tend to have high luminosity for their velocity dispersion

MB=-21MB=-19

• Lines of constant luminosity estimated using FJ-relation and [M/L] models of BC03.

The Faber-Jackson Relation

The Faber-Jackson Relation

• Confirms Forbes & Ponman (1999) finding that residuals to the FJ-relation correlate with galaxy age.

• Suggests age/metallicity degeneracy has been broken.

• Confirms Forbes & Ponman (1999) finding that residuals to the FJ-relation correlate with galaxy age.

• Suggests age/metallicity degeneracy has been broken.N=1500

Red: Young

The Colour-Magnitude Relation (CMR)

(The ‘red-sequence’)

The Colour-Magnitude Relation (CMR)

(The ‘red-sequence’)

• However, the sample is limited to high luminosity galaxies.

•(photometric bimodality becomes significant R>-17)

• Nevertheless, argues against common belief that low scatter in CMR implies old ages.

(At least in high luminosity galaxies)

• However, the sample is limited to high luminosity galaxies.

•(photometric bimodality becomes significant R>-17)

• Nevertheless, argues against common belief that low scatter in CMR implies old ages.

(At least in high luminosity galaxies)

• Normally assumed to indicate a mass/metallicity relation and to imply a small range of ages.• Data suggest true picture not so clear-cut

6dFGS: Results for [Z/H]

6dFGS: Results for [Z/H]

An age-metallicity relation A mass-metallicity relation

[Z/H]=0.7log()-0.6log(age)-1.0

(a mass-metallicity relation that evolves with time)

Red: Low mass Red: Young

6dFGS: -element abundance ratios.

6dFGS: -element abundance ratios.Red: Young Red: Low mass

N=3500

An [/Fe]-age relation Suggests less continuous SF than solar neighbourhood

Pure Fe

Pure Fe

The future.The future.• Refine age/metallicity measurements (This is a work in

progress).

• Probe ages and metallicities in emission line galaxies (Consider ages<1.0 Gyr).

• Investigate emission line characteristics (HII/AGN,

Balmer decrements, gas metallicities).

• Quantify trends in galaxy parameters (FJ-relation, CMR and age/mass/metallicity planes).

• Test idea of ‘frosting’ (Compare spectroscopic results for central regions to global photometry).

• Investigate variations with environment.

• DR2 and DR3.

• Refine age/metallicity measurements (This is a work in

progress).

• Probe ages and metallicities in emission line galaxies (Consider ages<1.0 Gyr).

• Investigate emission line characteristics (HII/AGN,

Balmer decrements, gas metallicities).

• Quantify trends in galaxy parameters (FJ-relation, CMR and age/mass/metallicity planes).

• Test idea of ‘frosting’ (Compare spectroscopic results for central regions to global photometry).

• Investigate variations with environment.

• DR2 and DR3.

Conclusions.Conclusions.• We have used Lick indices to break the age-metallicity

degeneracy in by far the largest study of its kind to-date.

• Results show trends in ALL metallicity parameters with both mass AND age.• These provide challenges to both primordial and monolithic

collapse models of galaxy formation.

• The 6dFGS will prove to be an invaluable testing ground for galaxy formation models.

• The addition of reliable age and metallicity estimates for a large number of galaxies will significantly enhance the value of the 6dFGS.

• We have used Lick indices to break the age-metallicity degeneracy in by far the largest study of its kind to-date.

• Results show trends in ALL metallicity parameters with both mass AND age.• These provide challenges to both primordial and monolithic

collapse models of galaxy formation.

• The 6dFGS will prove to be an invaluable testing ground for galaxy formation models.

• The addition of reliable age and metallicity estimates for a large number of galaxies will significantly enhance the value of the 6dFGS.

Abrat issues 1 - ?Abrat issues 1 - ?

Lick indices (Worthey 1994)Lick indices (Worthey 1994)

• Properties of single stellar populations (SSPs) are estimated using:

• Stellar spectral libraries (Teff, log g and [Fe/H]).• Isochrones (age and [Fe/H]).• A Stellar Initial Mass Function (IMF: No. with

mass).

• Properties of single stellar populations (SSPs) are estimated using:

• Stellar spectral libraries (Teff, log g and [Fe/H]).• Isochrones (age and [Fe/H]).• A Stellar Initial Mass Function (IMF: No. with

mass).• Integration of stellar properties (weighted by IMF)

along isochrones of given age and metallicity yields model properties for an SSP.

• Spectral synthesis of Tripicco & Bell (1995) models ‘’-elements (Models used here : Thomas, Maraston & Korn 2004)

15 Gyr

1.0 Gyr

1.5 Gyr

[Fe/H]=-0.4

[Fe/H]=-2.25

Age-metallicity degeneracy

1. Photometry

Age-metallicity degeneracy

1. Photometry

- Tight locus of all combinations of age and metallicity in the range 2.0 -15 Gyr, -2.25≤[Fe/H]≤-0.4 (Models: Bruzual & Charlot 2003)

2.0 Gyr

7 Gyr

6dFGS: [Fe/H] results6dFGS: [Fe/H] results

Our approach.Our approach.• Estimate Age, [Fe/H],

[/Fe] and [Z/H]• Use as many indices as

possible (up to 25)• Thus:• Minimise effects of

most errors (reduction and calibration)

• Utilise the fact that ALL indices contain SOME information about age, [Fe/H] and [E/Fe].

• Estimate Age, [Fe/H], [/Fe] and [Z/H]

• Use as many indices as possible (up to 25)

• Thus:• Minimise effects of

most errors (reduction and calibration)

• Utilise the fact that ALL indices contain SOME information about age, [Fe/H] and [E/Fe].