The Stellar Populations, Mass-to-Light Ratios and Dark Matter in Spiral Galaxies

1
The Stellar Populations,Mass-to-Light Ratios
andDark Matterin Spiral Galaxies
Roelof S. de Jong Steward Observatory
Eric Bell Rob Kennicutt Rob Swaters Rob
Olling Don McCarthy Cedric Lacey
2
Overview

• Introduction
• Ages and metallicities of stellar populations
• description of method
• scaling laws with structural parameters
• Galaxy evolution modeling
• Mass-to-light ratios of stellar populations
• correlation with population colors
• constraints from rotation curves
• application to Tully-Fisher relation
• Future work

3
Galaxy Formation and Evolution

• Huge progress, both observational and
  theoretical
• observational e.g. the star formation history of
  the Universe and of local group galaxies
• theoretical hierarchical galaxy formation models
  in CDM-like universes

• Something is missing
• We do not not where, when and especially why
  stars are forming in particular galaxies

4
Galaxy Evolution and Structural Parameters

• What drives the Star Formation History and the
  Chemical Evolution within disk galaxies?
• current star formation in disks semi-regular,
  related to morphology and structural parameters
• are spirals determined by initial conditions or
  are infall and outflow important?
• how is galaxy evolution related to the luminous
  and dark matter distribution and galaxy dynamics?
• What is the distribution of dark and luminous
  matter?
• can we explain the Tully-Fisher relation?
• does dark matter really follow NFW profile
  distributions?
• do we need alternative gravity (e.g. MOND)?

Structural parameters luminosity, scale size,
surface brightness, mass, velocity
distribution Statistical studies scaling
relations
5
How to measure stellar ages and metallicities

• Gas metallicity easy from HII regions, but only
  current metallicity
• Problem Age-metallicity degeneracy for stellar
  populations

(nanometer)
6
Optical Near-IR Spectra Comparison
Blue/red versus red/near-IR breaks age/metallicity
degeneracy
(nanometer)
7
Origin of age-metallicity degeneracy
Trager (astro-ph/9906396)
8
Stellar populations Color-Color diagrams
Gyr
Gyr
Bruzual Charlot models
9
Data Samples

• Face-on disk galaxies with
• data in at least 3 passbands (of which one IR)
• good colors over at least 2 disk scale lengths

Total sample of 121 galaxies
10
Radial Color-Color Observations
R-K
B-R
B-R
11
Maximum Likelihood Fitting

• Make model grid of e-t/t Star Formation History
  and metallicity
• parameterize SFH by average age ltAgt
• Determine minimum ?2 between models and data
• use all available passbands
• take calibration, flatfield and sky errors into
  account
• Repeat for all radii
• Use Monte Carlo simulations to determine
  uncertainties

12
Local Age Local Metallicity versusLocal
Surface Brightness
13
Age vs Surface Brightness Luminosity
14
Metals vs Surface Brightness Luminosity
15
What determines SFH and Metals?Surface
Brightness or Luminosity?
Remember luminosity and surface brightness are
correlated!
16
The Galaxy Space DensitySurface Brightness
Magnitude
Space density ofspiral galaxies corrected for
selection effects (de Jong Lacey
2000)
17
The Galaxy Space Density Scale Size and Magnitude
18
Are Ages mainly determined by Surface Brightness
or Luminosity?
19
Is metallicity mainly determined by Surface
Brightness or Luminosity?
20
Summary observations

• Ages are mainly determined by surface brightness,
  suggesting inside-out disk formation
• Metallicity is determined by surface brightness
  and total luminosity
• The observed scatter is larger than observational
  errors

• So what are the caveats?
• Changes in the IMF
• Other Stellar Population Synthesis models
• The effect of dust reddening

21
IMF uncertainty
Salpeter IMF
Scalo IMF
22
Spectral synthesis model uncertainty
Bruzual Charlot
Kodama Arimoto
23
The effect of Dust Extinction

• Extinction will mainly effect metallicity
  determinations i.e. reddening vector runs
  parallel to metallicity color gradients
• Reddening not the main cause of the observed
  trends because
• we are using face-on galaxies
• of the limits set by overlapping and edge-on
  galaxy

24
The effect of Dust Extinction

• Extinction will mainly effect metallicity
  determinations i.e. reddening vector runs
  parallel to metallicity color gradients
• Reddening not the main cause of the observed
  trends because
• we are using face-on galaxies
• of the limits set by overlapping and edge-on
  galaxy

• we see no dependence on galaxy inclination
• colors are mainly determined by least obscured
  stars
• patchy dust structure reduces reddening effect
• reddening is caused by absorption only, not by
  scattering

25
Dust modeling with scattering

• Scattering preferably to face on direction
• Reddening follows absorption curve, not
  extinction curve
• For low optical depth reddening insignificant

26
Conclusion Age Metallicity Caveats

• Only very unusual IMFs can mimic our results
• Other Spectral Synthesis Models will only change
  the absolute age and metallicity values
• Dust will at most effect metallicities a bit

The relative rankings of Ages Metallicities are
Robust
27
Simple Galaxy Evolution Models

• Simple closed box models
• Start with exponential gas disk
• Form stars according to Schmidt law (surface
  density)n
• Instantaneous recycling of metals
• Maximum likelihood fitting on resulting
  integrated colors

• Additional bells and whistles
• Mass dependent metal free gas infall
• Mass dependent enriched gas blowout
• Mass dependent epoch of formation
• Fluctuations due to small starbursts

28
Galaxy evolution models
Closed box model
Mass dependent formation epoch model with star
burst
Mass dependent formation epoch model
29
Modeling conclusions

• Simple closed box models with a star formation
  rate dependent on local gas density explains the
  basic observed trends between stellar ages
  metallicities and galaxy surface brightness
  parameters
• Enriched gas blowout or mass dependent formation
  epoch models are needed to explain the
  metallicity dependence on total luminosity of the
  galaxy
• Small burst of star formation explains the
  scatter on the observed relations

• What about masses instead of luminosities?

30
Why stellar M/Ls?

• Stellar M/Ls needed to do dynamics in situations
  where we have more matter than just stars, e.g.
• (baryonic) Tully-Fisher and other scaling
  relations
• rotation curve decomposition
• Dynamics is needed to model star formation and
  galaxy evolution

• How? Many approaches possible
• Milky Way kinematics
• galaxy kinematics
• streaming motions induced by bars or spiral arms
• vertical velocity dispersion in stellar disks
• stellar population synthesis

31
Galaxy evolution models
Closed box model
Mass dependent formation epoch model
Mass dependent formation epoch model with star
bursts
The optical color of a stellar population is a
good M/L indicator
Even in K mass-to-light ratio varies by factor of
2
32
Color-ML for hierarchical galaxy model

• Even a hierarchical galaxy formation model
  shows strong correlation between color and M/L

Cole et al. (2000) models
33
Different population synthesis models

• The slope of the color-M/L relation is
  independent of stellar population synthesis
  models used

34
Different Initial Mass Functions

• The slope of the color-M/L relation is
  independent of models and IMFs used

• The normalization of the relation depends on the
  IMF used, i.e. the amount of low mass stars

35
Rotation curve M/L constraint
36
Maximum disk constraints

• The color-M/L relation must be normalized below
  all maximum disk values

• Salpeter IMF

• A Salpeter IMF is too massive

• Distribution suggests IMF similar in most
  galaxies and close to maximum disk for a fraction
  of the galaxies

data Verheijen (1997)
bad data point due to beam smearing
37
Stellar Mass Tully-Fisher relation

• Stellar masses derived from different passbands
  using the color-M/L relation agree to within 10
  rms
• The Tully-Fisher relations derived from different
  passbands are identical to within the errors
• The slope is very steep Vrot M4.5

• Raw Tully-Fisher relation has different slopes
  and offsets in different passbands

• Tully et al. (1998) extinction corrections makes
  the slopes steeper, but do not bring them into
  agreement

38
Baryonic Tully-Fisher relation

• Add in the HI gas mass to calculate the baryonic
  Tully-Fisher relation
• The slope is less steep than stars only and less
  than
• Vrot Mbar3.5
• Slope problematic for MOND, but consistent wit
  hierarchical CDM galaxy formation models with
  some fine-tuning

39
Future work Stellar Velocity Dispersions

• An isothermal disk yields

40
Future work Rotation Curves
41
Future work stellar populations
Ages and metallicities of resolved stellar
populations in nearby disk galaxies

• Ages of young star clusters in merging
  galaxies

42
Conclusions

• Local star formation history in disks mainly set
  by local surface density, resulting in inside-out
  disk formation
• Metallicity regulated by both surface density and
  mass
• Realistic galaxy evolution models predict a
  strong correlation between population color and
  M/L
• Maximum disk constraints support this
  observationally and suggest that a Salpeter IMF
  is too massive
• The stellar mass Tully-Fisher relation is
  independent of originating passband
• The baryonic Tully-Fisher relation has a maximal
  slope of about 3.5 /- 0.2