Spectral Indices

1
Spectral Indices
Broad Band Colors are affected by the
AGE-METALLICITY DEGENERACY Spectral indices have
been introduced to overcome this problem
2
Lick Indices
Definition Worthey, Faber, Gonzales,
Burstein 1994
EW, e.g.
MAG, e.g.
Measurement
Kuntschner and Davies 1997
3
Lick Indices measurementNOT TRIVIAL
Tragers Thesis
1. Define Pseudocontinuum Flux in the two side
bands
2. Fc(?) is the straight line through the 2
adjacent FP Then index is
4
Stellar Population Indices are Synthetizednot
analyzed
INDICES FOR SSPs Start from simplyfied
definitions
with some algebra you get
In the SSP, each star contributes its index
weighted by the stars contribution to the total
continuum flux
5
SSP Indices first derive the SED, then measure
the index
Kurucz models have a resolution of 20 A Lick
indices resolution is 8-10 A Need a high
resolution stellar SED
Vazdekis 1999 500 stars (Jones, 97) with spectral
resolution of 2 A
Only in the range 3820 4500 A 4780 5460 A
AND Basically all stars at solar Z
6
Lick Indices from SED
Bertone, Buzzoni et al 2004 Models of atmospheres
7
Alternatively Fitting Functions
Gorgas et al. 1993
8
Lick Indices
Worthey et al.1994
9
SSP Lick Indicesan example
Metallic line strenghts increase with both AGE
and Metallicity
Hß gets weaker as age increases and as
Metallicity increases
Use combination of metallic and Balmer line
strenghts to solve the AGE-METALLICITY degeneracy
10
First, but very interesting, resultsalpha
elements overabundance in Es
Worthey, Faber Gonzales 1992 At given Fe
index, the data Mg index is stronger than the
model predictions Interpreted as a supersolar
Mg/Fe ratio Among various possibilities Short
Formation timescales for Es
Notice that ZFe,o1.3e-03
ZO,o 5.8e-03 ?lpha overabundance is more an Fe
underabundance
11
First, but very interesting, resultsAge spread
among Es
Gonzales et al. 1992 E galaxies span a small
range in metallicity and a wide range in age The
most metal rich are also the OLDEST The alpha
overabundance syndrome is also evident
2.5
1.6
1.3
12
Indices which diagnostic power?
For magnitues consider 10-0.4(INDEX)
Most of the Index growth occurs in the (upper) MS
and in the (lower) RGB No wander these are the
phases which Most contribute to the continuum
flux In the optical
The various fitting functions appear to agree
in the relevant evolutionary phases the final
values of the indices happen to agree (at least
for this isochrone)
13
Model Lick Indices compared with GCs data
Puzia et al. (2002) measure Lick Indices for 12
MW GCs, 9 in the Bulge Maraston et al. (2002)
derive metallicity by comparing with 12 Gyr old
SSP models Spectroscopic/Photometric metallicity
is available for these clusters
COMPARE THE TWO
The metallicity derived from the Mg index is
approx. OK The metallicity derived from the Fe
index is systematically low Again effect of
alpha overabundance (Fe depletion)
14
Indices with alpha enhancement
15
Tripicco and Bell (1995)
Stellar sample with solar Z

• TB compute high resolution stellar
• spectra using model atmospheres
• for (g,Te) combinations along the
• M67 isochrone.
• Model Indices are computed for solar
• Mixture and for other mixtures in which
• Each element abundance is doubled
• So as to measure the partial derivative.
• Find that some indices trace abundance
• Some do not Fe4668 is very sensitive to
• C abundance
• Compare model indices to
• Stellar values
• b) Worthey fitting functions

16
TB95
DWARFS
W ff
GIANTS
17
Effect of elemental variation
Tripicco and Bell 1995 compute variation of
Index in response to doubling the abundance of
one element,
which leaves (almost) constant the total
metallicity
Computed for Dwarf Turn-Off
star Red Giant
18
Indices Correction for alpha overabundancefollow
ing Thomas, Maraston Bender (2003)
0.3
Tripicco and Bell 95 give
for D, TO and G
For each class compute the corrected index with
Get SSP corrected Index as
19
Thomas, Maraston and Bender 2003
12 Gyr old, Solar Metallicity SSP Models
At constant Z, when a/Fe increase Mg indices
get stronger Fe indices get weaker MgFe stays
constant
20
Thomas, Maraston and Bender 2003comparison to GC
data
From model Mgb and Fe52 get Fe and as
abundance ? Total Z
Brodie and Huchra 1990 metallicity Calibration
of MW and M31 GCs
Data Puzia et al GCs
21
What have we learnt

• In an SSP each stellar index is weighted with the
  contribution of the
• Star to the Total Continuum Flux of the SSP
• Cool Dwarfs have a high Mg index, but cannot be
  efficiently used to enhance the
• SSP Mg index
• Metallic Line Strenghts are stronger in older and
  more metal rich stellar populations
• Balmer Line Strenghts are stronger in younger and
  more metal poor stellar populations
• Index Index Diagrams offer a
  diagnostic for
• age AND metallicity
• Element Abundance Ratios affect the indices in
  different ways

The Optimists View
One Balmer Index One Mg Index One Fe Index
AGE METALLICITY a/Fe RATIO
22
But Tantalo and Chiosi (2004) SSP Models with a
enhancement
Based on Salasnich et al. a enhanced models TB
95 response functions
Based on patching literature isochrones TB 95
response functions
Based on isochrones FCT TB 95 response
functions
Ages depend on abundance pattern As a
consequence, Z also does
23
Seek better diagnostic for AGE other Balmer
Indices (H? , Hd) unaffected by gas emission
Worthey and Ottaviani 1997
Models high sensitivity to AGE (but Hß is still
the more sensitive) Jump from Fe/H-1 to -0.5
due to Variation of HB morphology
For each index 2 definitions F for wide (40 A),
A for narrow (20 A) central bandapasses
24
New Generation of ModelsThomas,Maraston and
Korn 2004
Korn et al. (2005) compute response functions of
atmospheres with various (g,Teff) for a wide
range of metallicities (1/200 to 3.5 solar).
Based on these, TMK re-compute Indices with
variable alpha/Fe ratios.
Contrary to Hß (?!) higher order Balmer Indices
are sensitive to a/Fe ratio (at high Z) The
pseudocontinua contain Fe lines
25
Effect on Galaxies AGE Dating
When a/Fe0.2 is used the Ages indicated by Hß
and H?F become consistent When using solar ratio
models high order Balmer lines lead to
underestimate the age
26
New Response FunctionsKorn,Maraston,Thomas 2005
Model atmospheres with high resolution for
typical MS,TO and RG stars on Isochrones with
Z-2.25 0.67
Determine the response function of 21 Lick indices
27
Korn,Maraston,Thomas 2005
Solid new response functions Dotted old
response functions Black Dots GCs data Open
Square Bulge field Grey dots Es
The new models are almost the same As the old. We
have much more confidence on the metallicity
dependence of the correction for non solar a/Fe
ratio Still not explored the behaviour along
the Isochrone
28
Puzia et al. 2006 AGES, METALLICITIES and
ABUNDANCE RATIOsof a sample of extragalactic GCs
Quality of the index as age indicator depends
on Mean error of the data Transformation
accuracy to the Lick system Error on the original
Lick spectra Accuracy of the Lick FF Dynamic
range of the index Degeneracy parameter
Use MgFe as total Z indicator Weighted ave of
Balmer Lines as AGE
indicator ltFegt and ltMg2gt to measure a/Fe

Notice that many GCs fall out of the Grid. This
should be further investigated
29
CAVEATS BALMER LINE STRENGTHSdepend on HB
morphology
Maraston 2003
Hß depends on HB morphology Models have 10 and
15 Gyr Solid no mass loss
30
CAVEATS BALMER LINE STRENGTHSdepend on HB
morphology
from Puzia et al 2006 OBSERVED BALMER
INDICES OF MW GCs (squares) M31 GCs (inv.
Triangles) LMC GCs (triangles) STAR high Z GCs
with blue HB
HBR(B-R)/(BVR)
31
CAVEATS SPURIOUS CORRELATIONS
From Thomas et al. 2005 (SFH of Ellipticals)
Monte carlo Simulation - take indices of one
SSP with t10.7, Z/H0.26,a/Fe0.25 - apply
observed errors (gaussian) on the diagnostic
indices - derive (t,Z,a/Fe)
Error of the procedure can be Quantified into 0.1
dex for metallicity 0.03 for abundance ratio 1.5
Gyr of age
ERRORS ON Hß INDUCE A SPURIOUS ANTICORRELATION
OF AGE AND METALLICITY