Transcript mg_colloq - University of Massachusetts Amherst

The Progenitors of the Compact EarlyType Galaxies at High Redshift
or, the evolution from z=∞ to z≈2
Mauro Giavalisco
University of Massachusetts Amherst
With Christina Williams, BoMee Lee, Paolo Cassata, Elena Tundo,
Yicheng Guo.
Galaxy properties largely
bi-modal
“Population” features:
Early types: passive, evolved, spheroidal, v/σ<1
Late types: star forming, young, disk, v/σ>1
When did this differentiation begin?
Schawinski+
Lee, MG+, 2012, press release
Clear Trends in the ways galaxies form stars:
Main Sequence Galaxies: continnuous mode of SF
Starburst Galaxies: large SFR from major merging
(they have not quenched yet, however)
Noeske+ 2005
Daddi+ 2010
Rodighiero+2011
Elbaz+ 2011
Many others…
Lee, MG+ 2014
Quenching
Passive galaxies
Rodighiero et al. 2011
Peng et al. 2010, 2013; Renzini 2009
Environment Quenching
Gas strangulation?
Tidal stripping?
Shock heating?
Mass Quenching
AGN feedback?
Star formation Feedback?
Ellipticals: key testing grounds
• Include the oldest, most massive galaxies
• Formed the bulk of their stellar mass at high redshift,
on short time scale: ≈90% at z>2 (Renzini 2006)
• Probes of the physics of early star formation
• Evolved passively since
• Dispersion of properties (color/age gradients, light
profiles)
• probes how environment drives galaxy evolution
Early Type Galaxies at High Redshift
•
Observations showed that early-type galaxies, i.e. non star-forming galaxies with low
sSFR and “spheroidal looking” morphology (sSFR<10-2 Gyr-1; nSersic>2.5) , can indeed be
massive, but they are also much more compact than their local counterpart of the same
mass: x5 smaller re, x102 higher ρstar (Daddi, Cimatti, van Dokkum, Trujillo, Nipoti,
Saracco, Valentinuzzi, Poggianti, Huerdas-Company, Cassata, Guo, Williams….)
Compact:
Σ50≥3x109 M kpc-1
Ultra compact: Σ50≥1.2x1010 M kpc-1
The Evolution of compact ETG
Compact:
Σ50≥3x109 M kpc-1
Ultra compact: Σ50≥1.2x1010 M kpc-1
•Compact massive ETG first galaxies to passivize
•
They dominate passive galaxies at high z
•cETG appear to peak at z≈1
•Less compact galaxies evolve monotonically
•
The formation of ETG continues at all times
•The formation of cETG stopped at z≈1
•
cETG only destroyed afterward?
•Today, ultra-compact galaxies are very rare (but see
Kormendy+ 2009)
Cassata, MG+ 2013
The nature of the compact ETG at z≈2
– Are they really passive? Are they spheroids?
– Little disagreement they are passively evolving (sSFR<10-11 yr-1)
– Light profile fit to Sersic generally yields n≥2 (very often n>3); axial ratio≈1
– Compact ones are barely resolved by HST
– Today, fraction of disk dominated SDSS galaxies with sSFR≤10-11 yr-1 and M*≥1010 M
with n>2 is <5%
– Dynamical properties being explored; certainly consistent with being massive, compact
spheroids (Onodera et al. 2012; van Dokkum et al. 2011)
– Some suggest that they include a significant (25-50%) fraction of compact disks (Bruce et
al. 2012) or even that they are mostly compact disks (van der Wel et al. 2011)
– The details of the selection vary. One must make sure they are equivalent
– Especially at ground-based resolution, it remains very difficult to recognize the
kinematical signature of compact disks and spheroids at high redshift
Panchromatric SED and spectra
consistent with passively evolving
(or quenching) stellar populations
Today’s ellipticals occasionally
show some emission lines (e.g.
[OII]) indicative of SF activity
(either residual or due to episodes
of rejuvination)
The nature of the compact ETG?
Onodera et al. 2012
Early Type Galaxies at High Redshift
•
Key questions:
•
what are the cETG and how do they relate to the evolution of ETG in general?
•
Do all ETG go through the compact phase and then become “normally-sized”?
•
Or are these galaxies a separate class of ETG? Why are they so compact?
•
•
Compactness seems a good predictor of passivity (e.g. Bell et al. 2012). Why?
Are they telling us about a different formation mechanism for massive galaxies
(e.g. Dekel et al. 2009; Wuyts et al. 2011; Sales et al. 2012)?
•
Compactness suggests that a highly dissipative mechanism was at work during the
assembly of their stellar mass. How?
•
Lots of efforts trying to understand the evolution from z≈2 to the present.
A possibility is that these objects formed through a highly dissipative process
that involved essentially only gas
e.g. accretion of cold gas directly from the cosmic Web
Are compact galaxies (both passive and SF) direct evidence of galaxy
formation by cold accretion?
Expected to have strong dependence on environment
:
Cold accretion ends sooner in denser environment
Dekel et al. 2009a,b
Do compact ETG evolve by growing inside-out
(e.g., by accretion, star formation)?
Inner core seems to evolve at
constant size
Total size and mass increase by
accretion of, SF in, outer envelope
Van Dokkum+ 2011
ETG also grow in size because of the addition of newer,
larger members of the population of quenched galaxies
Carollo+ 2013
The morphology of cETG
Cassata, MG+ 2013,
Williams, MG+ 2013
All but passive sources masked out
Possibly, there is an excess of ETG companions (≈2.7σ)
There are NO low SB structures or extended
halos around the compact core of cETG at
z≈2 (e.g. tidal tails, companions)
Stack of all ETGs in the CANDELS sample
in GOODS
Residuals consistent with NORMAL surface
density of intervening sources
•
The z≈2 compact ETGs with Mstar>3x1010 M do not appear
to have extended structure or halos around them
•
Subtracting the residuals of the best-fit Sersic model only
reveals intervening galaxies.
How did the cETG form?
•
So far most efforts focused on how cETG
evolved from z≈2 to z≈0
•
But perhaps more importantly is how they
evolved from z≈∞ to z≈2
•
These galaxies might be the best evidence so
far that cold accretion (in ways we do not fully
understand) was actually a key mode of galaxy
formation
Light profile of stack show no evidence for diffuse light
(cfr. Mancini et al. 2011, who do find evidence
of “normal sizes” and “halos” around some
ultra massive, Mstar>2x1011 M, ETG at z≈1.6)
Williams, MG et al. 2013, ApJ, in press
Stacks show no evidence of
diffuse light (halo) or structures
•
No diffuse light light around compact ETG, both around individual galaxies
and from stacks
•
No evidence that extended halos and/or tidal debris are common
•
If these objects formed via wet mergers, a diffuse light profile is expected
from the violent relaxation of the dissipationless component (e.g. Hopkins et
al. 2008)
•
The physical message here is that high-z ETG have generally NOT
experienced any major merging of galaxies with a sizeable stellar
component.
•
Whatever process has put the baryons in such small volume, it was
characterized by a very high amount of dissipation, i.e. the baryons were
mostly in the form of gas when that happened. The stars formed after the
gas was in place (Dekel et al. 2009, Wuyts et al. 2010)
How did such objects form?
Is wet merging (fgas≥50%) viable?
Mergers needs to be wet (fgas>40%) and compact.
Still, sims cannot reproduce the observations,
remnants are too large, have too much light at
large radii.
Wuyts et al. et al. 2010
simulations by T.J. Cox
Radial profiles:
Results from the Sersic
fits to data and sims
Sims of merging
produce remnants
that are too large
Williams, MG et al. et al. 2012
New sims show that neither merging nor other dissipative
processes seem to work… very puzzling!
Hopkins+, in prep.
How do SB galaxies
look like?
z≈2 cETG stack
z≈2 SB stack
z≈2 SBs are our best candidates for
merger remnants
Core-normalized difference
Provide empirical information on morphology
of major merger remnants
We staked all 5x SB above MS
We also stacked those that are “compact”
Stack of SBs is “disky” (n=1.7) and large (re=
2 kpc). <M> = 6.3x1010 M
Stack of cETG is “spheroidal” (n=3.6) and
small (re= 1 kpc). <M> = 6.4x1010 M
MG+ in prep.
cETG and their candidate progenitors
reasonably well described by steep,
“spheroidal” (n>3) Sersic profile
Sersic profile of core of SB galaxies is
“disky” (n=1.7)
Removal leaves residual (halos)
Mergers scatter stars and form halos
(stars are not dissipative)
The progenitors of compact ETG
•
If cETG formed in situ via some highly dissipative process (cold accretion?), then their
progenitors must be among compact star-forming galaxies at z≥3. Are there any
reasonable candidates?
•
We used pure UV selection (LBG), UV/Opt selection (VJL) as well as SED selection.
Here results for LBG at z≈3 in GOODS-S. Search criteria are:
1.
2.
3.
4.
Redshift such that SF ended ≈1+ Gyr prior observation epoch of cETG (consistent
with estimated age of stars, see Onodera et al. 2012)
Compact morphology; (from the WFC3 CANDELS images)
•
We estimate the projected stellar mass density from the WFC3 H-band images
(rest frame optical): Σ50 = M*/2πre2
•
Cassata et al.’s stellar density criteria for normal, compact, ultra-compact
SFR and stellar mass such that, after including the extra mass formed during
quench at z<zobs, they reproduce the stellar mass distribution of ETG observed at
1.5<z<2.5
Must be passive, i.e. have SSFR<10-11 yr-1, at <z>≈1.6
Modeling the cessation of SF
•
For each galaxy, we used a simple exponentially declining SFH, e -t/τ
•
Quenching starts right after zobs
•
We measured SFR from UV continuum + Calzetti dust
•
How do we constrain τq, the quenching time scale?
•
τq set by the difference between the redshift of the cETG and that of the
progenitor candidates
•
To end the SF activity across the whole structure, the minimum τq must be
of the order of the sound crossing time:
•
The maximum τq set by the requirement SSFR<10-11 yr-1 at z=1.5
The progenitors of the massive compact ETG
Because of the LBG simplicity and high
efficiency, we looked for candidates first
among z~3 LBG (U-band dropouts).
We will extend the search to other redshift
epochs and selection criteria
Williams, MG et al. 2013
•
Candidates must have the right SFR,
stellar mass, and SFH, to be
observed as cETG at z≈2
•
Must have right morphology, too: a
SF disk at z≈3 is still visible at z≈2
even after quenching
•
We used conservative choice of τq.
•
Larger τq result in larger mass and
more candidates
The progenitors of the compact ETG:
star-forming galaxies at z≈3
Galaxy Type
Compact ETG
Compact z≈3 LBG
Co-moving volume density
3.5 x 10-4 [Mpc-3]
2.3 x 10-4 [Mpc-3]
The number depends on the assumptions on the quenching history, τq
More candidates expected when other types of SF galaxies will be
included
•
WFC3 H-band images of the candidate progenitors
(compact z≈3 LBG)
UV/Opt SED of ETG progenitor candidates vs. non candidates:
The UV SED of
candidates is
redder
Repeating the stacks after eliminating
all galaxies with [OII] and [OIII] in any of
the used bands yields the same result
•
SED differs only in the UV part
•
Candidate progenitors have
slightly larger D4000 and redder
UV
•
Optical part is virtually identical
•
Consistent with higher metallicity
or an older, more evolved burst
(not supported by emission lines)
IR SED of ETG progenitor candidates vs. non
candidates (MIPS 24 mm & Herschel 160 mm)
-9.4
-5.9
-2.4
1.0
4.5
Average Flux Density [uJy/pixel]
Candidate ETGs at <z> ~ 3
20
7.9
11
Non-candidates
Arcseconds
10
0
•
Both candidates and non candidates seem to have
-10
similar MIR luminosity. i.e. similar dust-emission
properties (24 mm at z≈3 is about 6 mm, i.e. thermal IR
-20
emission by warm/hot dust)
•
Non-candidates seem to have higher luminosity
0
20
40
Pixels
60
80
Effects of AGE and DUST on the UV/Opt SED of
Star-Forming Galaxies: models
Effects of AGE and DUST on the UV/Opt SED of
Star-Forming Galaxies: observations
Difference between the average points of candidates
and non-candidates is along the the age line
Williams, MG+ 2013
Stacked spectra show that:
Candidates have stronger UV features, consistent with higher
metallicity (see Rix et al. 2004).
They also seem to have larger galactic winds
Williams+ in prep.
Candidates have both wider and more blueshifted
interstellar lines:
More powerful outflows and more turbulent ISM
MG+ in prep.
AGN quenching?
Barro et al. report that the AGN fraction (X-ray detection) in the
compact SF galaxies is 30% vs. 1% in non compact ones at
2<z<3.
At z>3, we do not observe the same
Very few detections of AGN in our sample (<6%)
Similar rates for candidates and non-candidates
No individual galaxy is detected in 4M Chandra image
We used stacks
Stacks of the CXO 4-Msec X-ray images (Tundo+ in prep.)
Candidates Hard band
Non candidates Hard band
Candidates Soft band
Non candidates Soft band
Stacks using the CXO 4-Msec X-ray images revealed no
obvious AGN activity (Tundo et al. in prep.)
Galaxy Type
Average X-ray flux
Compact z≈3 LBG, soft
1.9 ± 2.5 cnt/source
Compact z≈3 LBG, hard
0.5 ± 2.5 cnt/source
Non compact z≈3 LBG, soft 2.0 ± 0.9 cnt/source
Non compact z≈3 LBG, hard
-0.2 ± 1.1
cnt/source
•
•
•
Detection rate of individual sources is about 6% for both compact, 4% for
non compact
Consistent with the small incidence of AGN among LBG
Absence of evidence is NOT evidence of absence, but presence of AGN
does not seem obvious
We do not know how galaxies
quench, i.e. their s.f.h. to
passivity.
Remember that compact
galaxies quench sooner than
non compact ones
All our SF candidates
progenitors sit high on the MS.
These galaxies should be the
one with the fastest rise in the
SFR, followed by a quick
quenching time-scale τq
(Renzini 2009)
Quick τq  internal process
Slow τq  environmental
process
Maybe the best predictor of a
galaxy future S.F.H. is its
position on the MS
(if measured well)
Better than fitting a S.F.H. and
then extrapolating it…
Compactness (n, S*) a good predictor of passivity
U-B
Bell+ 2012
Cheung+ 2012
McGrath+ 2013
The Quenching rate of
compact galaxies
At any given epoch, the
number of cETG depends on
1) the rate at which compact
star forming galaxies
appear and
2) the rate at which they
become passive
(quenched)
Only galaxies with Mstar>1010 M
and Σ50≥1.2x1010 M kpc-1
shown
(ultra-compact and massive as per
Cassata et al.’s definition)
MG et al. in prep.
Conclusions
•Compact
Early-Type Galaxies unlikely to form by merging of pre-existing galaxies
•Their
compactness require highly dissipative gas process; they may be the telltale of galaxy
formation by “cold accretion”. However, current models seem too crude.
•They
are the first passive galaxies to appear in the universe. The place to go to understand
underlying physics of quenching. Compactness seems to play a role
•Dependence
of compact systems (ETG & SF) on environment as a function of redshift is key (Lani+ 2013;
Poggianti+ 2013)
•
Will help constrain the cold accretion idea
•
Will help testing if merging and interaction are responsible for the disappearance of compact
galaxies
candidates: SF galaxies at z≈3 whose stellar density is as high as cETG at z≈2 and have the
same (projected) mass
•Progenitor
•Progenitors
•They
appear to have higher metallicity. Consistent with properties of ETG
sit high on the MS: more likely to quench sooner, more rapidly.
•AGN?
no obvious difference of X-ray properties b/w compact and non-compact SFG (candidate and noncandidate progenitors)
•ISM
kinematics seems more extreme. Stellar feedback, i.e. transfer of energy, momentum to ISM efficient
quenching agent in high density environment? (e.g. see Krumolz, Quataer, other)