Transcript Diapositiva 1 - INAF-Osservatorio Astronomico di Bologna

The size evolution of early-type
galaxies since z=2
P. Saracco1, M. Longhetti1,
with the contribution of
S. Andreon1, A. Mignano1, G. Feulner 2, N. Drory 2,
U. Hopp 2, R. Bender 2
1 INAF
– Osservatorio Astronomico di Brera, Milano
2 Max Planck Institute and University of Munchen
Bologna 22.01.2009
Outline of the talk
 Small/compact Early-Type Galaxies (ETGs) at z>1: first
evidence
 A morphologycal study of a sample of 10 ETGs at 1.2<z<1.7:
size evolution of ETGs required
 The population of ETGs at 1<z<2: new clues on their
formation and evolution ?
 Summary and conclusions
Bologna 22.01.2009
Small size, high-density ETGs: first evidence
Daddi et al. (2005)
Hubble UDF - 7 ETGs z>1.4
HST-ACS obs., FWHM~0.12”,
F850W filter, λrest<3000 Ǻ
Bologna 22.01.2009
Further evidence
Cassata et al. (2005)
Trujillo et al. (2006)
Re [Kpc]
Re [Kpc]
K20 + GOODS data
Mass
IR ground based observations
FWHM~1.0 arcsec
redshift
HST-ACS observations, F850W
λrest<3000 Ǻ
Bologna 22.01.2009
Are ETGs at z>1 really more compact/denser than local
counterparts ?
These results were based on
• HST optical observations sampling the blue and UV rest-frame of the
galaxies sensitive to k-correction and star formation and/or
• seeing limited ground-based observations
Doubts on the reliability of the estimate of Re
Doubts on the reliability of the comparison high-z vs low-z
High-resolution near-IR obs. sampling λrest~6500 Ǻ for a reliable
comparison between high-z and low-z ETGs.
Bologna 22.01.2009
HST-NICMOS observations in the F160W (λ~1.6 µm) filter of a
sample of 10 ETGs at 1.2<z<1.7.
0.075 “/pixel
(Longhetti et al. 2007)
Data sampling the rest-frame R-band
(λrest~6500 Ǻ) at z~1.4, at a spatial
resolution <0.8 kpc (FWHM~0.12 “)
NIC2
images
• Effective radius re (arcsec)
and mean surface brightness
(SB) <>e within re from
Sersic profile fitting
I (r )  I e e
models
bn [( r / re )1 / n 1]
residuals
n=4 de Vaucouleurs profile
n=1 exponential profile
• galfit (Peng et al. 2002) to
perform the fitting after the
convolution with the NIC2 PSFs.
z=1.34
z=1.40
z=1.7
n=3.2
n=4.5
n=2.7
Bologna 22.01.2009
The Kormendy relation in the R-band
It is a scaling relation between the
effective radius Re [Kpc] and the
mean SB <>e [mag/arcsec2]
  e     log(Re )
The ETGs follow this tight relation with
~3 up to z~1.  is found to vary
reflecting the luminosity evolution.
  e  M  ( z)  5 log(Re )  38.57
Expected KR at
z=1.5
Any deviation from
the KR at z=0
should reflects the
evolution of <>e
due luminosity
evolution .
passive luminosity
evolution (maximum
evolution expected
for early-types).
Observed KR at
z=0.
Expected locus for
z<1.5 early-type
galaxies in case of
luminosity evolution.
Bologna 22.01.2009
The Kormendy relation in the R-band
It is a scaling relation between the
effective radius Re [Kpc] and the
mean SB <>e [mag/arcsec2]
  e     log(Re )
The ETGs follow this tight relation with
~3 up to z~1.  is found to vary
reflecting the luminosity evolution.
  e  M  ( z)  5 log(Re )  38.57
Expected KR at
z=1.5
The SB exceeds by
~1 mag the one
expected in the
case of PLE for
constant Re, i.e.
passive luminosity
evolution (maximum
evolution expected
for early-types).
Observed KR at
z=0.
luminosity
evolution does not
account for the
observed SB of
ETGs at high-z.
Expected locus for
z<1.5 early-type
galaxies in case of
luminosity evolution.
(Longhetti et al. 2007)
Bologna 22.01.2009
Are ETGs at z>1 really more compact/denser than local
counterparts ?
These results are based on
• HST near-IR observations sampling the red rest-frame of the
galaxies NOT sensitive to k-correction and star formation and/or
• NO seeing limited ground-based observations
NO doubts
on the reliability of the estimate of Re
High-z ETGs (at least some of them) are more compact then
their local counterparts.
(Longhetti et al. 2007)
Bologna 22.01.2009
The Kormendy relation in the B-band
GMASS sample
13 ETGs 1.4<z<2
Spectroscopic data
Morphology based on
HST-ACS obs. F850W
(λrest~3000 Ǻ)
(Cimatti et al. 2008)
Bologna 22.01.2009
Literature and HST archive research
Aim – to collect a large (larger than 10…!) sample of ETGs at z>1 with
Sample
•10
spectroscopic
confimation
of(Saracco
the spectral
ETGs 1.2<z<1.7
from TESIS
et al.type;
2005; Longhetti et al. 2005)
+•10
ETGs 1.4<z<1.9
from GDDS in
(Abraham
et al.filter;
2004; McCarthy et al. 2005)
HST-NICMOS
observations
the F160W
+•6multiwavelength
ETGs z~1.27 from
RDCS 0848+4453
(Stanford et al.1997; van Dokkum et al.
coverage
(optical + near-IR)
2003
+ 3 ETGs 1<z<1.8 from HDF-N (Stanford et al. 2004)
in order to study the population of ETGs at 1<z<2 from an homogeneous
set of data and a uniform analysis
+ 2 ETGs z=1.4,1.9 from GMASS H-UDF (Daddi et al. 2005; Cimatti et al. 2008)
+ 1 ETGs z=1.55 53W091 (Dunlop et al. 1996; Waddington et al. 2002)
= 32 ETGs 1<z<2, 17.0<K<20, HST-NICMOS observations F160W
• covering a larger interval in luminosity;
NIC2 (0.075 ”/pixel) for 14 galaxies
• defining
the scaling
relations at z~1.5
NIC3
(0.2 “/pixel)
for 18 galaxies
(Kormendy,
size-luminosity/mass relations)
FWHM
~ 0.12 arcsec
Bologna 22.01.2009
Physical properties of ETGs
Morphological parameters
• effective radius and surface brightness derived as in Longhetti et al.
(2007);
• Simulations done also for NIC3 images
 rNIC 2  0.02arc sec  rNIC 3  0.04arc sec 0.16 and 0.32 kpc at z~1.5
e
e
Absolute magnitudes, stellar masses, ages
• Fit to the observed SEDs (BVRIzJHK F160W) at fixed z
Charlot and Bruzual models (2007, CB07)
IMF=Chabrier
SFHs SFR  e


t
τ=0.1,0.3,0.6 Gyr (best-fit τ<0.3 Gyr for 28 out of 32)
Metallicity Z☼,0.4 Z ☼
(best-fit Z☼ )
AV<0.6 mag
(best-fit AV<0.3 for 24 out of 32 )
Bologna 22.01.2009
The Kormendy relation in the R-band
    18.2  2.92log(Re )
R
e
  e     log(Re )
z=0
  eR  16.100..12  2.7200..52 log(Re )
z~1.5
The ETGs at z~1.5 are placed on
the [<µ>e,Re] plane according to the
KR.
z~1.5 ETGs follow the same KR of
ETGs at z=0 but with a different
zero-point.
Saracco et al. 2008
Bologna 22.01.2009
Luminosity evolution
Each ETG evolves from z=zgal to z=0 according to its own SFH.
  e  M R ( z  0)  5 log(Re )  38.57 M R ( z  0)  F160W  5 log DL ( z )  k R, F160W  E ( z )
E ( z )  [ R( Agegal )  R( Agegal  t )]template
Only 40% (13 gal) of the sample
occupies the KR at z=0.
The remaining 60% (19 gal) does
not match the local KR, the SB
exceeds by 1-1.5 mag the one
expected.
Two distinct populations ?
Saracco et al. 2008
Bologna 22.01.2009
Two distinct populations !?
Saracco et al. 2008
Bologna 22.01.2009
Two distinct populations of ETGs at 1<z<2
• Old ETGs , <Age>~3.5 Gyr, <z>=1.5  zf>5
Their stellar population formed in the early universe. Pure
luminosity evolution does not account for their high SB. The
evolution of their size must be invoked.
• Young ETGs , <Age>~1.2 Gyr, <z>~1.5  zf~2.5
Their stellar population formed much later than the stellar
population of Old ETGs. Pure luminosity evolution from zgal to
z=0 brings them onto the local KR.
Bologna 22.01.2009
Size-Luminosity/Mass relations
SDSS Shen et al. (2003)
Size-Luminosity
log Re [kpc]  0.26M R  5.02
Size- Mass
M 
Re [kpc]  3.47 105  * 
 M 
Bologna 22.01.2009
0.56
Size-Luminosity (S-L) relation
log Re [kpc]  0.26M R  5.02
M RR ( z)0F
W
5 log
z)L( zk)R,F160
) 160
 FW
160
 5D
log
kRW,F160W  E( z)
L (D
Young
Old
Re of oETGs is 2.5-3 times
smaller than
- the local ETGs and
- the yETGs with comparable
luminosity.
Saracco et al. 2008
Bologna 22.01.2009
Size-Mass (S-M) relation
M 
Re [kpc]  3.47 10  * 
 M 
0.56
5
Young - 9 out of 13 (70%)
follow the S-M relation
Old - 4 out of 19 (20%)
follow the S-M relation
Re of Old ETGs is 2.5-3 times
smaller than
- the local ETGs and
- the yETGs with comparable
stellar mass.
Old ETGs are 15-30 times
denser !
Saracco et al. 2008
Bologna 22.01.2009
Constraining the formation and the evolution of
ETGs
Two distinct populations of ETGs at z~1-2
1. How did these two populations evolve from z~2 to z=0
to match the properties of the local ETGs ?
2. Which assembly history did they follow to have the
properties shown at z~1.5-2 ?
Bologna 22.01.2009
Tracing the evolution at z<2
oETGs
Luminosity evolution DOES NOT bring them onto the local
Kormendy and S-L relations.
They DO NOT match the local S-M relation.
They are 2.6(±0.5) times smaller than their local counterparts.
They must change their structure.
Size evolution from z~2 to z=0 is required to move them
onto the local scaling relations.
Bologna 22.01.2009
Tracing the evolution at z<2
oETGs
M 
Size evolution often
the merging processes


R [kpcused
]  3.47 to
10 advocate

the ETGs should experience inM the
hierarchical paradigm of
galaxy formation.
0.56
5
*
e

Dissipation-less (“dry”) merging is the most obvious and
efficient mechanism to increase the size of galaxies.
The size of ETGs increases according to the relation
Re  M*
0.6    1.3
Boylan-Kolchin et al. 2006-08
Khochfar and Silk 2006
Nipoti et al. 2002
Ciotti et al. 2007
M2i .6M    1.3 1/ 
R f  2.6M
Rif M2f.1
M f  2.6 M i
i
Bologna 22.01.2009
Tracing the evolution at z<2
oETGs
- Merging would produce too much ETGs with M>1011 Msun: we
should observe 3 times more ETGs with M>4-5x1011 Msun .
- Why α=1.3 ?
Merging cannot be the mechanism with which oETGs
increase their size at z<2.
Alternative mechanism(s) leaving nearly unchanged the mass
and relaxing the system:
1. interactions between galaxies (e.g. close encounters)
2. minor or “satellite” merging (Naab et al. 2007):
M1:M2 = 0.1:1
Efficiency can be constrained from simulations.
Bologna 22.01.2009
Tracing the evolution at z<2
yETGs
Luminosity evolution brings them on the local Kormendy and
S-L relations.
They match the local S-M relation.
No size evolution is required.
To move them along the S-M, α~0.6  Mf~5Mi
No evidence of merging at z<2.
The build-up of yETGs was already completed at z~2.
Bologna 22.01.2009
Constraining the path at z>2 - Toward the
formation of ETGs
oETGs
<Age>~3.5 Gyr, <z>=1.5  zf>5 (Age Univ. 4.2 Gyr at z=1.5)
To build-up 1011 Msun SFR>>100 Msun/yr
Size 2.5-3 times smaller
 mechanism(s) acting at z>2 must be capable to produce
galaxies 5-10 times more compact (15-30 times denser) than
local ones
Gas-rich merging with high fraction of stars formed during the
merger in a violent starburst can produce highly compact ETGs
(Khochfar et al. 2008; Naab et al. 2007).
BUT
tmerger>3 Gyr
Bologna 22.01.2009
Constraining the path at z>2 - Toward the
formation of ETGs
yETGs
<Age>~1.2 Gyr, <z>~1.5  zf>2.5
Constraints on the mechanism(s) acting at z>2 less stringent:
They can increase their mass and enlarge their size by
subsequent mergers (major and minor/satellite) and through
starburts till z~2.5 (contrary to oETGs).
Different progenitors
oETGs: we should see them as they are (younger) till z~3-3.5
yETGs: in the phase of merging, or star forming and interacting
with other galaxies at z>2.5
Bologna 22.01.2009
Summary and conclusions
Two distinct populations of ETGs at z~1-2 whose stellar
populations differ in age by about 2 Gyr
Young ETGs: No size/mass evolution is required.
Old ETGs: Strong size evolution is required at z<2.
The system must relaxes from high to low redshift
 oETGs must show higher central velocity dispersion
than local ETGs and than yETGs.
Key observational test: measuring the velocity dispersion of
oETGs.
ESO-P82 VLT-FORS2: spectra of 10 oETGs, 10 hrs/spec
Observations started in November 2008…we shall see!
Bologna 22.01.2009
Mean age vs stellar mass
5% Stellar mass
Bologna 22.01.2009
The evolution of the zero point α
  e     log(Re )
Our sample
Zero point α of the
KR derived from
various samples at
different redshifts.
Luminosity evolution
SFH tau=0.6 Gyr,
solar metallicity,
Chabrier IMF
The curves show
the
expected
evolution of α for
different formation
redshift zf.
Longhetti et al. 2007
Luminosity evolution
+
Evolution of Re
Re ( z)  Re ( z  0)  ( z  0.5)
Bologna 22.01.2009
1
Luminosity evolution of Young and Old ETGs
M R ( z )  F160W  5 log DL ( z )  k R , F 160W  E ( z )
E ( z )  [ R( Agegal )  R( Agegal  t )]template
Saracco et al. 2008
Bologna 22.01.2009
Absolute magnitudes
M R  F160W  5 log[DL ( z gal )]  k R , F 160W
k R , F 160W  ( Rz 0  F160Wz gal ) template
Bologna 22.01.2009
Morphological study of a sample of 10 ETGs at 1.2<z<1.7 based on
HST-NICMOS observations in the F160W (λ~1.6 µm) filter
(Longhetti et al. 2007)
Sample - K<18.5, spectroscopic
confirmation of the spectral type from
TESIS (TNG EROs Spectroscopic
Identification Survey; Saracco et al. 2003,
2005; Longhetti et al. 2005).
NICMOS data - NIC2 camera (0.075
“/pixel) sampling the rest-frame Rband (λrest~6500 Ǻ) at z~1.4, at a
resolution <0.8 kpc (FWHM~0.12
arcsec)
Bologna 22.01.2009
Estimating the mean age of the stellar population
0.5 Gyr old
5% Stellar mass
95% stellar mass, 4 Gyr old
B V R
I
z
J
H
Bologna 22.01.2009
K
Size-density and mass-density relations
0.5M star
 50 
Re2
Saracco et al. 2008
0.5M star
50 
4 / 3Re3
Bologna 22.01.2009
Simulations
To assess the robustness of the results we applied the same fitting
procedure to a set of simulated galaxies
Real galaxies
Simulated De
Vaucouleurs profile
100 simulated galaxies
• magnitudes F160W and re assigned randomly in the ranges 19<F160W<21
and 0.1< re <0.5 arcsec (1-5 Kpc at z~1.4);
• axial ratio b/a and position angle PA in the ranges 0.4<b/a<1 and 0<PA<180
Bologna 22.01.2009
NIC3 images (0.2 “/pixel)
NIC3 images (0.2 “/pixel)
GDDS sample.
HDFS-NICMOS
z=1.65
z=1.73
z=1.85
z=1.55
Bologna 22.01.2009
zphot=1.94