Transcript Observational Data

A Wide Area Survey for HighRedshift Massive Galaxies
Number Counts and Clustering
of BzKs and EROs
Kong et al. (2006), Astro-ph/0510299, ApJ in press
N. ARIMOTO (NAOJ)
X.Kong, M.Onodera, C.Ikuta (NAOJ), K.Ohta (Kyoto),
N.Tamura (Durham), A.Renzini, E.Daddi, L. Da Costa (ESO),
A.Cimatti (Arcetri), T.Broadhurst (Tel’Aviv), L.F.Olsen (Cote d’Azur)
Formation of Giant Ellipticals
Massive ellipticals are the products of recent hierarchical merging of disk
galaxies taking place largely at z<1.5 with moderate SFRs (Cole et al. 2000),
fully assembled massive galaxies with M*>1011Mo at z>2 are extremely rare.
Massive Galaxies in the Redshift Desert (z>1.3)
Glazebrook et al. (2004)
Cimatti et al. (2004)
Previous Spectroscopic Surveys
1)
2)
3)
4)
5)
6)
K20 (Cimatti et al. 2002)
HDFN (Ferguson et al. 2000)
GOODS (Giavalisco et al. 2004)
HST/ACS UDF (Yan et al 2004)
GDDS (McCarthy et al 2004)
LBGs@z～2 (Steidel et al 2004)
52 arcmin 2
5.3 arcmin 2
160 arcmin2
12 arcmin 2
121 arcmin 2
100 arcmin2
Massive galaxies are quite rare and likely highly clustered
at all redshifts, hence small areas such as those explored
so far are subject to large cosmic variance.
EIS Deep 3a Survey
Kong et al . (2006) astro-ph/0510299
We have
a fairly
deep, wide-field
imaging with
Theundertaken
prime aim of
this survey
is to understand
the Subaru/Suprime-Cam
of two fields
of 900 arcmin2
how and when the present-day
massive
each galaxies
for part formed.
of whichTo
near-IR
data
areimaging
available from
this end,
the
observationsESO
haveNTT
beenobservations.
optimized for the use
of optical/near-infrared multi-colour selection
criteria to identify both star forming and
passive galaxies (BzK selection).
7. EIS3a-F (Subaru/NTT, Ks=20.8)
320 arcmin2
8. Daddi-F (Subaru/NTT, Ks=19.0)
600 arcmin2
Subaru/Sup-Cam Observation
Daddi Field
RA=14:49:29, DEC=09:00:00 (J2000.0)
Subaru/Suprime-Cam BIz’: 2003/03/02-04
WHT
R : 1998/03/19-21
NTT/SOFI
K : 1999/03/27-30
BRIz’ (940 arcmin2) 3σ in 2”(AB)
B(AB)=26.59
R(AB)=25.64
I(AB)=25.62
z’(AB)=25.31
K (600 arcmin2) 3σ in 2”(AB)
Ks(AB)=20.91
600arcmin2
940arcmin2
Subaru/Sup-Cam Observation
ESO Imaging Survey (EIS Deep 3a) Field
RA=11:24:50, DEC=-21:42:00 (J2000.0)
Subaru/Suprime-Cam BRIz’: 2003/03/02-04
NTT/SOFI
JK : 2002/03/28-31
BRIz’ (940 arcmin2) 3σ in 2”(AB)
B(AB)=27.46
R(AB)=26.87
I(AB)=26.56
z’(AB)=26.07
JK (320 arcmin2) 3σ in 2”(AB)
J(AB)=23.40,
Ks(AB)=22.70
320arcmin2
940arcmin2
Differential K-band Galaxy Counts
K-band Galaxy Number Counts
BzK-Selected Galaxies (K20)
(z-K)>2.5
BzK=(z-K)-(B-z)>-0.2
(Daddi et al 2004, ApJ 617, 746)
Why BzK-selection if efficient for culling
star-forming and passive galaxies at 1.4<z<2.5?
B
z
K
star-forming BzK galaxy at z=1.6
Photometric vs Spectroscopic Redshifts
BzKs
K20 Daddi et al (2004)
High-z galaxies Deep 3a field
Star-forming galaxies at
z>1.4 (sBzKs)
Old galaxies at
z>1.4: (pBzKs)
BzKs
stars
BzK(ERO)
ERO
BzK
ERO
BzK
ERO
BzK
ERO
387 sBzK
121 pBzK
513 ERO
108 sBzK
48 pBzK
337 EROs
Star/Galaxy Separation
(z-K)AB-0.3(B-z)AB<-0.5
Sky densities of sBzKs, pBzKs, EROs
arcmin-2
Number Counts of sBzKs, pBzKs, and EROs
galaxies
pBzKs
EROs
sBzKs
Number Counts of sBzKs, pBzKs, and EROs
•
•
•
•
•
For EROs, the slope of the number counts is variable, being steeper at
bright magnitudes and flattening out towards faint magnitude.
The pBzKs number counts have a similar shape, but the break in the
count slope is shifted to 1-1.5 magnitude fainter.
Both EROs and pBzKs have fairly narrow redshift distribution: peaked at z
～1 (EROs) and at z～1.7 (pBzKs).
The number counts are direct probes of their respective luminosity
functions. The shift in the counts is consistent with the different typical
redshift of the two populations.
The counts of sBzKs have roughly the same slope at all K-band
magnitude, which reflects the much wider redshift distribution of this
class of galaxies.
Photo-z Distribution
Two Point Correlation Functions w(Θ)
Landy & Szalay (1993)
Daddi-F
Deep 3a-F
Angular Clustering Amplitude
EROs, sBzKs, and pBzKs distribute in
a very inhomogeneous way in
the sky.
EROs and sBzKs appear to be
strongly clustered, but pBzKs
clustered most strongly in
both fields.
The clustering strengths of all
the three populations
increase with K-band luminosity.
Physical Properties of sBzKs and pBzKs
• Supposing <z>～2 for sBzKs, we have derived their
physical properties, such as the reddening, star
formation rate, and the stellar mass.
(While errors by a factor of 2 or more may affect individual
estimates, the average quantities should be relatively robust.)
• Reddening : E(B-V)=0.25(B-z+0.1)AB ←UV Continuum slope
(Calzetti law)
• SFR : SFR(Mo/yr)=L1500[erg/s/Hz]/8.85x1027
• Stellar Mass : log(M*/1011Mo)=-0.4(Ktot-20.14Vega)
The field area
the histogram for sBzKs which associated with X-ray sources (25%).
11is
Above 10The
Mo
the lines
numbers
andhistograms
pBzKs are
similar.
dashed
are for of
the sBzKs
stellar mass
of pBzKs.
Correlation between Colour Excess E(B-V), SFR
and stellar mass for sBzKs
• There is evidence for an intrinsic correlation between SFR
and reddening at z～2 star-forming galaxies, with galaxies
with higher star formation having more dust obscuration (>5σ
level).
• The correlation between E(B-V) and stellar mass Is likely to
be intrinsic, with more massive galaxies being also more
absorbed (>7σ level).
• Given the previous two correlations, not surprisingly we also
find a correlation between SFR and stellar masses (>4σ
level).
• The upper edge in the SFR vs M* appear to be intrinsic,
showing a limit on the maximum SFR that can be present in a
galaxy of a given mass.
SFRs/mass @ z～2 were ～10 times larger than today.
Brinchmann et al (2004)
Downsizing Effects?
• At z=0 the vast majority of massive galaxies
(M*>1011Mo) are passively evolving “red” galaxies,
while at z～2 actively star forming (sBzKs) and
passive (pBzKs) galaxies exist in similar numbers, and
most massive galaxies tend to be the most actively
star forming galaxies.
• This can be seen as yet another manifestation of the
downsizing effect, with massive galaxies completing
their star formation at an earlier epoch compared to
less massive galaxies, which instead have more
prolonged star formation.
Contribution of sBzKs to SFRD
SFRD=0.06 Mo/yr/Mpc3
for sBzKs in Deep3a-F
(KVega<20)
SFRD=0.013 Mo/yr/Mpc3
for sBzKs in Daddi-F
(KVega<19.2)
SFRD=0.044±0.08 Mo/yr/Mpc3
for sBzKs in GOODS-S
(KVega<20; Daddi et al 2004)
for the volume in the redshift
range 1.4<z>2.5
Substantial contribution to the total SFRD
25% AGN
Contamination
is likely
come from
KVega>20 sBzKs.
cosmic variance?
Contribution of sBzKs and pBzKs to Stellar
Mass Density
ρ*(sBzKs)=2.45x107 Mo/Mpc3
ρ*(pBzKs)=1.79x107 Mo/Mpc3
for Deep3a-F(KVega<20)
for the volume in the redshift
range 1.4<z>2.5
logρ*(total)=7.7 Mo/Mpc3
logρ*(total)=7.86 Mo/Mpc3
(1.5<z<2.0, Fontana et al 04)
logρ*(total)=7.65 Mo/Mpc3
(2.0<z<2.5, Fontana et al 04)
logρ*(total)～7.5 Mo/Mpc3
(@z～2, Dickinson et al 03)
25% AGN contamination
Images of BzKs at z～2
K>20 HST/ACS F435W, F850LP & K-band (VLT+ISAAC)
A sample of 9 galaxies at 1.7<z<2.23 with bright
K-band magnitudes 18.7<K<20 has recently been
discovered (Daddi et al. 2003, astro-ph/0308456).
Summary and Conclusions (I)
BzK selection is a quite powerful way to separate
high-z galaxies such as sBzKs, pBzKs and EROs
at 1.4<z<2.5.
1) Down to the K-band limit of the survey the log of the
number counts of sBzKs increases linearly with the
K-magnitude, while that of both EROs and pBzKs flattens
out by Kvega～19.
EROs are in a modest redshift shell (z～1),
while pBzKs are also in a relatively narrow
redshift shell but at higher redshift (z～1.7).
sBzKs are drawn from a large range of redshifts,
and their relative numbers increase sharply with redshift.
Summary and Conclusions (II)
2) The clustering properties of EROs and sBzKs are
very similar, clustering amplitudes ～10 times
higher than generic galaxies in the same magnitude range.
This suggests an evolutionary link between sBzKs
at z～2 and EROs at z～1, with star formation
on sBzKs quenching by z～1 thus producing
passively evolving EROs.
The clustering amplitude of pBzKs is even higher
than that of sBzKs and EROs, suggesting that
quenching epoch of star formation in massive
galaxies depends on environmental density.
Summary and Conclusions (III)
3) sBzK galaxies (KVega<20) have median reddening
E(B-V)～0.40, average SFR ～ 190 Mo/yr,
typical stellar mass ～1011 Mo,
and ～solar metallicity.
The high SFRs, large masses and high metallicities
of sBzKs suggest that these z～2
star forming galaxies are the precursors of
z=1 passive EROs and z=0 early-type galaxies.
Summary and Conclusions (IV)
4) The number density of massive pBzKs
(KVega<20, M*>1011 Mo) is about 1/2 of similarly
massive early-type galaxies at z=0.
The quenching of star formation in massive
star-forming galaxies must result in a
doubling since <z>～1.7 in the number of massive,
passive galaxies.
It is indeed quite reassuring that the number of
M*>1011 Mo sBzKs is very close to that of pBzKs.
We argue that most of this star-formation
quenching is likely to take place between z～2 and z～1.
Massive Early-type Galaxies
Evolutionary Tracks (M*>1011Mo)
z～0
z～1
z～2
z>2
E(B-V)～0.4
SFR～190Mo/yr
Z～Zo
Passive
EROs
Early-type Galaxies
strong clustering
sBzKs
number density 1/2
strong clustering
0.5-1Gyr
SMGs
40-200Myr
pBzKs
number density 1
strong clustering
number density 1/2
?
sRjLs
very very strong clustering