Massive disk galaxies at z1

1
Massive disk galaxies at z?1

• Joël Vernet (ESO)
• Collaborators
• S. di Serego Alighieri, A. Cimatti, A Rettura
• E. Daddi, P. Cassata, E. Pignatelli, G. Fasano,
• A. Franceschini, L. Pozzetti, A. Fontana, A.
  Renzini

2
Outline of the talk

• Introduction
• The programme
• Observations and data analysis
• Results the Tully-Fisher relation in B and K
• Conclusions

3
Context

• How and when disk galaxies formed ?
• Hierarchical merging models provide global
  understanding...
• Dark haloes develop hierarchically
• They acquire angular momentum via interactions
• Baryons cool and dissipate gravitational
  potential energy conserving angular momentum ?
  disk
• But they are not yet able to reproduce all
  observables because of complicated physics
• main issues are cooling efficiency, star
  formation and feedback

4
Why kinematics ?

• What do we need to observe ?
• Imaging only surface brightness profiles
• magnitudes (MB, MK...) ? follow evolution of
  luminosity functions, stellar mass functions
• sizes (exponential disk scale length rd)
• Not enough to disentangle the evolution in mass
  and size from the evolution in mass-to-light
  ratio...
• ? need to measure the total mass
• Moderate resolution spectroscopy
• rotation curves (Vmax) ? dynamical mass
• constrain total M/L ? directly linked to cooling
  and feedback
• push to high redshift ? reconstruct formation
  history
• Problem
• quite challenging measurements at high redshift
• difficult to do on large samples

5
The programme

• Goal kinematics of a sample of emission line
  galaxies at z?1
• Follow-up of K20 ESO large programme (PI A.
  Cimatti)
• 52 arcmin2 2 fields CDFS and Q0055
• Groundbased photometry UBVRIzJHK
• Complete to Kslt20 low res. spectroscopy
  (spectroscopic redshifts, galaxy type)
• for free GOODS HSTACS imaging for CDFS
  (morphology)
• Strategy
• Measure the extended OII emission line with
  VLT FORS2-MXU with red upgraded CCD... Why ?
• OII is on average 2.5 times weaker than H?...
  but MXU more efficient than long slit IR
  instrument like ISAAC
• Difficult part of spectrum (7000 to 9000 Ang.)
  fringing strong sky lines... but FORS2 red CCD
  very efficient and low fringing (much better
  than KeckLRIS red channel)

6
Observations

• Efficient target selection
• extended OII emission
• avoid strong sky emission lines
• Instrumental setup limits z range 0.8ltzlt1.4
• The run
• 2 nights
• 2 masks (1 per field) ? deep 6-7.5 hours
  integrations
• 27 late-type galaxies (and same number
  early-type gal.)
• 1 slitlets aligned with target major axis (a few
  tilted slits)
• moderate resolution 600z grism ? R1400
• dithering for better background subtraction
• good seeing 0.6-0.7 arcsec (effective)

7
The data
Spatial direction
Wavelength
8
Good rotation curves
9
and the others
10
Quality Control

• Three cases
• Disturbed velocity field or i lt 40º ? rejected
  (5/27)
• Rotation curve but flat part not reached or
  inconsistent velocity dispersion ? low quality
  (7/27)
• Good rotation curve ? good quality (15/27)
• Keep all RC with indication of a flattening
• And that have a consistent velocity dispersion

11
Good rotation curves
Vmax ?
12
Kinematics data analysis 1

• Problem at high z, observed disk are small with
  typical scale length close to seeing disk and
  slit width ? dilution
• Recovering the true Vmax requires modelling
• assume thin disk with exponential luminosity
  profile
• velocity field simple flat rotation curve (what
  is observed locally) with flattening at Rd or
  step function
• allow misalignment slit wrt galaxy major axis

13
Kinematics data analysis 2

• ?2 mininization on Vmax

14
Kinematics data analysis 3
15
Structural parameters

• Obtain disk scale length Rd, inclination i,
  position angle on the sky ? and estimate B/T
• CDFS
• analysis of public deep HSTACS F850LP from GOODS
  using GalFit
• Bulge/Disk decomposition
• Result B/T lt 20
• Additional test Sérsic profile fit
  (I(r)?exp(-r1/n))
• Result nSérsiclt2
• Late-type morphology although spectroscopic
  selection

16
Magnitudes

• B band magnitudes
• k-correction at z1.1, restframe B close to
  observed z ? robust
• Internal extinction correction main source of
  uncertainty
• We correct to face-on orientation using
• Tully et al. 1998 adds dependence on Vmax
  (luminosity dependence)
• Most important to be consistent when comparing
  to other samples !
• K band magnitudes
• k-correction a bit more uncertain because
  extrapolated
• ... but extinction correction is very small

17
Sample characterization (1)

• The main limitation...
• By definition Ks lt 20 ? at z1.1 MK lt -23.5
• We are looking at massive disks

L(B)
L(K)
18
Tully-Fisher relations
19
B band Tully-Fisher
z1.1 B zero point -3.52?0.28z0 zero point
-2.81?0.22 Brightening by 0.71 mag
?
20
Oii Luminosity

• The kinematics sub-sample is brighter in OII
  luminosity corresponding to 0.75 mag

21
K Offset vs. B Offset
22
Correlation B offsets-EW
Kannappan Fabricant
Good correlation between offsets from the z0 TF
in B and the OII EW
23
Conclusions B band

• Apparent brightening of 0.7 mag
• BUT it can be entirely explained by selection
  biases
• No significant evolution of the TF in the B band

24
K band Tully-Fisher
z1.1 K zero point -0.37?0.24z0 zero point
-0.01?0.20 No significant evolution (0.36?0.44)
25
Change in slope (downsizing)?
No evidence for mass dependant evolutionfrom
theseobservations
26
Stellar Mass TF
27
Stellar Mass Fraction

• Ideally compare Mvir with baryonic mass of the
  galaxy (MstellMgas) ? cooling, feedback
• Dynamical masses ?
• Spherical model M(R)RV2(R)/G
• Model based estimates of the total virial mass
• van den Bosch (2002) Mvir2.54 106 Rd V2max
• Within 10 of M(100kpc) (Conselice et al. 05)
• Mstell /Mvir comparable to local values

28
Conclusions

• We have observed 27 emission line galaxies from
  the K20 sample
• This has provided 15 measurements of Vc for
  objects 0.8ltzlt1.4 with ltzgt1.1
• The apparent brightening of the B band TF is
  entirely due to selection effects
• This is probably the cause of disagreement
  between different high-z studies
• No evolution of the K band TF relation
• Corollary no evolution in stellar mass TF
• No change in slope i.e. no evidence for
  downsizing in this sample
• Stellar mass fraction of does not evolve but this
  is very uncertain!