The Galaxy Luminosity Function

1
The Galaxy Luminosity Function

• Simon Driver
• Research School of Astronomy and Astrophysics,
• Institute of Advanced Studies,
• Australian National University
• JENAM 2002

2
Overview

• The Field Luminosity Function - Consensus or
  Conflict ?
• 2dFGRS, SDSS, ESP, (APM, SSRS2, LCRS, UKST, NOG,
  CS)
• The Millennium Galaxy Catalogue and the GLOBAL LF
• Group and Cluster Luminosity Functions
• The Local Group/Sphere
• Virgo, Fornax and Coma
• Other low-z clusters
• The 2 degree-field galaxy redshift survey cluster
  and field LF
• LF dependencies ?
• Spectral type variations between field and
  cluster ?
• Hubble Space Telescope WFPC2 Observations
• Abell 0868 (A morphological dissection)
• Abell 2218 (The luminosity profile of giants and
  dwarfs)
• The Bivariate Brightness Distribution
• Overcoming/monitoring selection bias ?
• Expanding the toolbox ?

3
The Schechter Luminosity Function

• Derived from Press-Schechter theory (Schechter
  1976)
• f normalisation parameter
• L charachteristic turn-over
  luminosity
• a faint slope parameter

4
Measuring the Field LF

• Problems
• Completeness
• Imaging
• Spectroscopic
• Selection Bias
• Imaging
• Spectroscopic
• Limited statistics
• Bright-end (rare)
• Faint-end (dV)
• M, f,a
  correlated

N(z)
statistical limit ?
N(M)
5
The Field Luminosity Function !
Large uncertainties in all parameters ?
x2 at M
x5 at limit (M-16)
How do we choose which field LF to adopt ?
6
Constraining the GLOBAL LF

• The various field LFs predict dramatically
  differing number-counts

7
The Millennium Galaxy Catalogue

• A deep wide survey (36 sq deg, to mlim26 mags/sq
  arcsec, B24 mags) overlapping with the 2dF GRS
  and SDSS-EDR, fully eyeballed.

FIRST THREE POINTINGS
8
Constraining the GLOBAL LF

• Using the Millennium Galaxy Catalogue (Liske et
  al 2002) enables f to be recalibrated
  f(2dFGRS)x1.06, f(SDSS)x0.71, f(ESP)x0.90
• But which M and which faint-end slope to adopt ?
  (-0.9 gt a gt -1.3 ?)

9
Constraining the LF without zs ?

• For a cosmology, k(z) and e(z) the LF predicts
  the galaxy number counts.
• The shape of the counts depends on M a, e.g
• The Millennium Galaxy Catalogue represents a
  pristine CCD based galaxy count sample (Liske et
  al 2002) over 36 sq deg to m26 B mags

10
Deriving the GLOBAL LF from counts alone ?

• Yes but the constraints are weak
• Require some priors Adopt mean M from
• 2dFGRS SDSS redshift surveys
• Most recent
• Largest ( gt 100,000s)
• gt M-19.67 5 log(h) mags
• Fit to counts now yields
• gt a -1.23 /- 0.05
• gt f0.0158 h3/Mpc3

11
The GLOBAL LF (h1)

• Finally we have a consensus between the US,
  Anglo-Australian and European measurements plus
  an independent estimate of a
• M(B) f
  a
• (ESP -19.61 0.0179 -1.22)
• SDSS -19.77 0.0147 -1.26
• 2dFGRS -19.66 0.0171 -1.21
• MGC (Counts) -19.72 0.0158 -1.23
• GLOBAL (h0.65) -20.61 0.0043 -1.23
• Now for the group and cluster LFs ----gt

12
The Local Group
GLOBAL LF
INCOMPLETENESS ?
50 galaxies only
13
The Local Sphere (R10Mpc)
GLOBAL LF
Incompleteness or turn-over ?
300 galaxies
14
Virgo
Proposed LSBGs based on size
Membership determined by morphology and size
400 galaxies
15
Fornax
Proposed LSBGs based on background
profiling/ over-density
Membership based on morphological
appearance although criticised as excluding
dwarf spirals etc
300 galaxies
16
Coma
CORE REGION
PURE BACKGROUND SUBTRACTION
2000 galaxies
17
Combined Nearby Cluster LF

• Straight sum of all three
  clusters to
  improve statistics
• Overall cluster LF only
  discrepant at
  M gt -16
  where field LF invalid
• Conclude

GLOBAL FIELD LF and GLOBAL
CLUSTER LF identical gt STRONG INFALL ?
18
The 2dFGRS
Can we confirm the universality from within the
2dFGRS data ? (de Propris 2002)

• -many clusters
• -spectroscopic selection
• -identical biases
• -faint mag limit

19
The cluster sample

• 60 clusters with z lt 0.11 and gt40 spectroscopic
  members with high completeness.
• Composite LF's are computed, each cluster
  normalized to N(Mlt-19) and the number of clusters
  contributing to each bin weighted 1/Nclus.
• Adopt a single k-correction, valid for the E/S0
  types that dominate the cluster.
• Completeness corrections assume no bias
  for/against cluster members.

20
The overall cluster LF
M -19.93 ? 0.07 ? -1.25 ? 0.03
21
Cluster and field LFs are identical!!
Cluster M -19.93 ? 0.07 ? -1.25 ? 0.03
Field M -19.79 ? 0.04 ? -1.19 ? 0.01
22
LFs and richness/velocity dispersion
Low ?
Poor
M -20.01 ? -1.27
M -19.94 ? -1.26
?M ? 0.1 , ?? ? 0.05
High ?
Rich
M -19.93 ? -1.18
M -19.89 ? -1.20
23
LFs and cluster structure
B-M gt II
No substructure
M -19.92 ? -1.22
M -19.90 ? -1.26
B-M lt II
Substructure
M -19.99 ? -1.26
M -20.03 ? -1.43
24
Cluster LFs - inner vs outer regions
Only global variation found is for inner and
outer regions
Rgt300kpc
Rlt300kpc
M -20.09 ? -1.29
M -19.93 ? -1.32
Intermediate dip reminiscent of that seen in Coma
25
Spectral and morphological types

• Spectral types from principal component analysis
  ? ? parameter
• ? correlates strongly with H? EW (i.e. SFR)
• ? correlates weakly with morphological type -
  can only distinguish E/S0/Sa Sb/Scd/Im
• The distribution of ? differs between the field
  and clusters
• more low-? (low SFR) types in clusters
• more high-? (high SFR) types in field

26
Cluster field LFs by spectral type
M type 1 is 0.5 mag brighter than field types
2 and 34 are within 0.2 mag.
? all types have steeper faint ends in clusters
than in field!
27
Background Subtraction Method (Driver et al 1998)
SIMULATED DATA
FIELD
CLUSTER
VALID FOR RICHNESS gt2 CLUSTERS IN GOOD SEEING lt
1
RECOVERED LUMINOSITY FUNCTION
28
Cluster cores via Background subtraction

• Driver et al 1998 - 7 Abell clusters - variations
  seen in core regions.
• See also
• Lopez-Cruz et al (1997)
• Paolillo et al 2001

However variations MAY be partly due to cosmic
variance along the cluster sight-line (Valotto
et al 2000)
29
The dwarf-density relation

• Dwarf-to-giant ratio a better
  discriminant than a
• The DGR rises towards the core
• Substantial scatter at core
• Consistency at low density
• comparable to the field

COMA (Various)
DRIVER et al 1998
Lopez-Cruz et al 1997
30
A2218 Mosaic - Luminosity profiles
1.4 Mpc/h
The deep WFPC2 mosaic (Archival Data) - 22
pointings - F606W filter - 0.2 deg (1.4 Mpc/h) -
12 drizzled 700 sec exposures per pointing
(8400secs) - Probes LF to M(F606W) -14
31
The Inner and Outer A2218 LF

• Overall LF M(F606W) -20.4, a -1.29 /-
  0.05
• Inner LF a -1.21 /- 0.05
• Outer LF a -1.31 /- 0.05

32
The A2218 galaxy profile

• DGR profile of A2218 follows
  Phillipps et al 2000
• Higher DGR in core
• Smooth transition
• Profile of giants steep
• Profile of intermediate shallow
• Profile of dwarfs flat

33
A2218 - Morphology Radius relation
34
A868 A Morphological Dissection

• A868 initially showed
  a steep
  upturn via
  background subtraction
• z0.15, BMII-III, R2
• 24 orbits in F606W
  tilled to
  provide
  continuous coverage

CORE
Sub-Clump
35
A868 Morphological Classification
STARS E/S0s Sabcs Sd/Irrs

• Morphological classification by
  Artifical Neural
  Networks into
  stars, E/S0s, Sabcs, Sd/Irrs
• Identical analysis performed on
  
  HDFNHDFS53W02 for field reference
• Millennium Galaxy Catalogue classified
  using identical ANN
  classifier to provide
  very bright counts
• Statistical field subtraction to extract
  overall and
  morphological LFs

36
A868 Number-Counts
37
A868 Morphological LFs
M-20.14 a 0.51 f 41.2
COMBINED LF(DIPPING)
GLOBAL LF
M-16.6 a -1.14 f 86.0
M-20.50 a -1.16 f 14.8
38
Summary

• Consensus now reached on the GLOBAL LF (a-1.23)
• Overall cluster LF identical to the GLOBAL LF to
  M -16
• Significant recent infall from field to cluster
• Cluster haloes mostly inert
• LSBGs may dominate at M gt -14
• Implications for field LF ?
• Cluster cores deviate from the GLOBAL LF
  (intermediate dip)
• Evolutionary process at work in cluster core only
• Core LF characterised by an overdensity of E/S0s
  AND a steeper faint-end
• implies two process (Sabc --gt S0, Sd/Irr --gt dE
  ?)
• implies morphology-density relation
• Cluster LF dominated by eta type 1s
• Star-formation inhibited throughout cluster halo
  during infall
• Radial profile of giants steeper than dwarfs
• Implies luminosity-density relation as seen by
  Phillipps et al 1998

39
The BBD A new methodology

• Surface brightness versus magnitude
• Quantitative
• Physical basis ? (SB -gt ang. Mom., mag -gt mass)