SURVEYS: THE MASS ASSEMBLY AND STAR FORMATION HISTORY · SURVEYS: THE MASS ASSEMBLY AND STAR FORMATION HISTORY Lecture #4 Observational facts . Olivier Le Fèvre – ON Rio de Janeiro

SURVEYS: THE MASS ASSEMBLY AND STAR FORMATION HISTORY

Lecture #4

Observational facts

Olivier Le Fèvre – ON Rio de Janeiro School 2014

Putting it all together

Clear survey strategies Instrumentation and observing procedures Selection function estimates

Let’s measure galaxy evolution !

Lecture plan

1. What are the main contenders to drive galaxy SFR and mass growth ?

2. The luminosity function and its evolution 3. The star formation history: luminosity

density and SFRD 4. The mass function and the stellar mass

density evolution 5. Mass assembly from merging 6. A scenario for galaxy evolution ?

What may drive galaxy evolution ?

A rich theory/simulation literature… Identify key physical processes When ? On which timescales ?

Beware: fashion of the day (e.g. from simulations) may fade quickly…

…Stick to facts !

Main physical processes driving evolution

Hierarchical assembly by merging Increases mass “catastrophically”

Gaz accretion Cold / Hot Fuels star formation Increases mass continuously along the cosmic web

Feedback: sends matter back to the IGM AGN (jets, …) Supernovae (explosion)

Star formation and stellar evolution Luminosity / color, lifetime Star formation quenching

Environnement, f(density) Quenching, Harassement, Stripping,…

5

Hierarchical merging

6

• The basics: hierarchical growth of structures

• Merging of DM halos • Galaxies in DM halos merge by

dynamical friction • Major mergers can produce

spheroids from disks • Merging increases star

formation (but maybe short lived)

• Increases mass (minor, major) • Merger Rate ∝ (1+z)m

Stellar mass growth from star formation and evolution of stellar populations

In-situ gas at halo collapse transforms into stars

Accreted gas along lifetime transforms into stars

Stars evolve (HR diagram) Luminosity evolution Color evolution

Stellar population synthesis models: (Bruzual&Charlot, Maraston,…)

7

Along the filaments of the cosmic web

Steady flow for some billion years can accumulate a lot of gas

Gas transforms into stars

Produces important mass growth

From Press-Schechter theory

8

Simulations

Dekel et al., 2009 At z~2

Cold gas accretion

Feedback Takes material out of a galaxy

back to DM halo

May quench star formation ?

AGN feedback

εf=0.05 (thermal coupling efficiency) εr=0.1 (radiative efficiency)

SNe feedback ψ: instantaneous SFR

feedback efficiency

Vhot=485km/s and αhot=3.2

9

Example: combined effect of feedback and cooling on mass function

10

A lot of “definitive” theories and simulations

Hopkins et al., 2006

White and Rees, 1978

White & Frenk, 1991

Dekel, 2013

Cool simulations, but… need to measure galaxy evolution !

A short summary of previous lectures…

With deep galaxy surveys Imaging & Spectroscopy

In large volumes Minimize cosmic variance

For large numbers Statistical accuracy

Measure properties at different epochs to trace evolution

Use these measurements to derive a physical scenario

13

Main evolution indicators

Luminosity function, luminosity density Star formation rate density Stellar mass function Stellar mass density Merging Accretion …

The luminosity function

From lecture #1

The reference at z~0.1: SDSS

Blanton, 2001 10000 galaxies

Blanton, 2003 150000 galaxies

Galaxy types vs. color

Evolution ! Canada-France Redshift Survey back in 1995

600 zspec

First evidence of evolution over ~7 Gyr

M* brightens by ~1 magnitude

Global LF Lilly et al., 1995

Le Fèvre et al., 1995

1 mag

CFRS: LF evolution per type to z~1

The LF of red galaxies evolves very little since z~1 Red early-type galaxies are

already in place at z~1 Consistent with passive

evolution (no new star formation)

Strong evolution of the LF for blue star-forming galaxies Luminosity or number

evolution ?

Little evolution

Strong evolution

A jump to z~2-4: UV LF from LBG samples

Using the LBG samples of Steidel et al. ~700 galaxies with redshifts

Continued evolution in luminosity L*

Steeper faint end slope α

From Reddy et al., 2008

Probing the LF to z~4 with the magnitude-selected VVDS

Steep slope for z>1 Continuous evolution

in luminosity Evolution in density

before z~2

Cucciati et al. 2012

1 mag

2.5 mag

Downsizing

The most massive / luminous galaxies form first, followed by gradually lower mass galaxies

The most massive galaxies stop forming stars first, with lower mass galaxies becoming quiescent later

This is ‘anti-hierarchical’ !

SFR(z) vs. Halo mass

De Lucia et al., 2006

Quenching

Star formation is stopped

But what produces quenching ? Merging Mass-related (feedback ?) Environment

Peng et al., 2010

The Star Formation Rate Evolution: the ‘Madau diagram’ back in 1996

Putting together several measurement: the strong evolution in

luminosity density observed by the CFRS from z~0 to z~1

Lower limits on SFRD from LBG samples at z~3

Lower limits on SFRD from HST LBG samples 2.7<z<4

A peak in SFRD at z~1-2 ?

From CFRS

From Steidel et al.

Let’s call it the “et al. diagram”…

From HST Hubble Deep Field

SFRD from the UV

Direct observation of UV photons produced by young stars

But absorbed by dust: need to estimate dust absorption

SFRD from the IR

UV photons produced by young stars are warming-up dust

Dust properties: calibration of UV photons to IR flux

Comparing Luminosity density from UV and IR

Same shape: transformation is extinction E(B-V)

Star formation rate evolution: today

Cucciati et al., 2012 • SFRD rise to z~2, then flat, then decreases • Considerable uncertainties at z>3

Stellar mass function evolution

Get stellar mass of galaxies from SED fitting Uncertainties ~x2 (Initial

Mass Function, Star formation history, number of photometric points on the SED, …)

Compute the number of galaxies at a given mass per unit volume

Stellar mass function evolution

Use double Schechter function Because of the different

shape of the MF for different galaxy types (next slide)

Massive galaxies are in place at z~1.5

Strong evolution of the low-mass slope

Evolution in number density

Redshift

MF evolution per type Star-forming galaxies Strong evolution in M* Strong evolution of α

Quiescent galaxies Strong evolution in M*

to z~1.5, then no-evolution

Strong evolution in number density

Ilbert et al., 2013

The mass growth of galaxies: stellar mass density ρ* evolution

Integrate the MF Global and per type

Smooth increase of the global ρ*

z=1-3: the epoch of formation of quiescent/early-type galaxies Almost x100 from z~3 to z~1

Specific star formation rate evolution: SFR/M*

Tasca et al. 2014

Galaxy mass assembly: Cold gas accretion or merging ?

Cold gas accretion: The main mode of gas/mass assembly ? « This stream-driven scenario for the formation of disks and spheroids is an alternative to the merger picture » (Dekel et al., 2010)

Merging major merging ? minor merging ? Occasional but large mass increase

Over time mergers can accumulate a lot of mass

Need to measure the GMRH since the formation of galaxies Mergers more/less frequent in the past Integral mass accrued from mergers

36

?

Method 1, A priori: pairs of galaxies

Method 2, A posteriori: merger remnants, shapes

Both methods require a timescale Timescale for the pair to merge

(vs. mass and separation) Timescale for features visibility

(vs. redshift, type of feature…)

At high redshifts z>1: pairs Faint tails/wisps lost to (1+z)4

surface brightness dimming

37

Measuring the evolution of the galaxy merger rate

Merging rate from pair fraction

38

Merging rate Pair count Number density

Merger probability in Tmg

Merging Timescale

Tmg depends on separation rp and stellar mass Kitzbichler & White 2008 computed timescales ~x2 larger than previously assumed ~1Gy vs. 500My

39

z=0.35

z=0.63

z=0.93

Spectroscopy enables to identify real pairs

Both galaxies have a spectroscopic redshift No contamination issue

Mergers at z~1.5 from MASSIV survey

80 galaxies selected from VVDS

Observed with SINFONI: 3D velocity fields

Straightforward classification: 1/3 galaxies are mergers

10kpc

Mergers at z~1.5

40 Lopez-SanJuan, 2013

What about merging at early epochs ? Merging pairs at higher z from VUDS

41

Merging pair at z~2.96

HST/ACS VIMOS spectra

Tasca et al, 2014

Galaxy Merger Rate History since z~3 from spectroscopic pairs

Peak in major merger rate at z~1.5-2 ?

Integrate the merger rate: >40% of the mass in

galaxies has been assembled from merging with >1/10 mass ratio since z~1

Doubling of the mass since z~3

Merging is an important contributor to mass growth

42

Cold gas accretion ?

First evidence in 2013 ?

Building a galaxy evolution scenario ?

Several key processes have been identified, Direct: mergers, stellar evolution Indirect: accretion, feedback, environment

Properties have been quantified over >12Gyr Observationnal references exist to confront models

Semi-analytical models Take the DM halo evolution Plug-in the physical description of processes Get simulated galaxy populations

Semi-successful… some lethal failures Over-production of low-mass/low-z and under-production of

high-mass/high-z galaxies Reproducing low-z LF/MF AND high-z LF/MF

More to be done ! 44

Circa 2002

Hopkins et al., 2008