The quenching of star-forma3on in galaxies Simon Lilly ETH Zurich - - PowerPoint PPT Presentation

The quenching of star-forma3on in galaxies

Simon Lilly

ETH Zurich

The evolving popula.on: Simplicity Physical processes in individual objects: Complexity

Galaxies: complexity in process and simplicity in outcomes

The Main Sequence: what do we mean by “quenching”?

3

Brinchmann et al 2004

Quenching refers to what makes some galaxies lie in the “red cloud” rather than the Main Sequence, with sSFR << sSFRMS

Whitacker et al 2014

4

variable$gas$reservoir$ long,lived$stars$ star, forma?on$ gas$inflow$into$galaxy$ return$ $$gas$inflow$into$halo$

Φ$

halo$

• wind$ou6low$

galaxy$ system$

Ψ = (1− r) SFR +Φ + dmgas dt

inflow

• uSlow

star-forma.on change in reservoir

ε = SFR mgas = τ dep

−1

⇔ mgas mstar ⎛ ⎝ ⎜ ⎞ ⎠ ⎟ = sSFR ×ε −1 λ = Φ SFR

Main Sequence evolu.on from a gas regulator

Wind-loading SF efficiency Key feature of this system: if λ and ε are constant and if the system is fed at some specific accre.on rate (sMIR) then the system produces sSFR=sMIR, independent of the values of λ and ε. Control is reversed: sMIR – sSFR – mgas/mstar via ε

Lilly+2013

log(SFR) log(m*)

9 9.5 10 10.5 11 11.5 −1.5 −1 −0.5 0.5 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 9 9.1

5

mstar ∝ mhalo

1(1−η) ~ mhalo 1.7

Z ∝ mstar

η

Metallicity as a test of the gas-regulator Lilly et al (2013)

• explains (s)SFR as second parameter in Z(m)
• Z(z) reflects the state of regulator system (ε,λ,sSFR)

not its history (because τgas << τH)

• Z(m,SFR) which will change with z only if ε or λ

change (i.e. expect redshid-independent FMR)

• Three-way links: Z(m), mstar/mhalo, and sSFR/sMIR

!"#$%~ 1 1 − ! !"#$!~!2!!"#$!

mstar ∝ mhalo

1(1−η) ~ mhalo 1.7

Z ∝ mstar

η

Z(m,SFR) from Mannucci+ 2010

“Chemical evolu.on” as changing state of regulator (à FMR)

!!" = !! + !! 1 + ! 1 − ! !! + !!! ∙ !"#$ + 1 − ! !! !"#$ !" = !! + !

!"#$ !!

x

Fieng this surface: τdep ~ 2.5m10

• 0.3 Gyr

λ ~ 0.3 m10

• 0.8

The “grow and quench” paradigm

SFR stellar mass

Factor of 20 decline in sSFRMS since z = 2

Continuous “flow” of galaxies along the conveyer belt

Growth of dark maher haloes

l

• g

m a s s

Rate of cosmic evolution driven by growth

• f haloes

S t a r s

passive galaxies 6 1970’s cartoon by Bruno Bingelli?

Cosmological importance of quenching

7

Birrer+ (2014) reproducing Behroozi+ (2013)

ΩB/Ωcdm ceiling

Moster+ (2010)

mstar/mhalo vs. mhalo

Inefficient star- forma.on (winds) “Quenching” Madau+Dickinson 2014 Quenching Slowing down of structure growth leading to decline in sSFRMS x10 Half due to decrease of SFMS and half due to quenching Inefficient SF due to winds Behroozi et al 2013 From Birrer et al 2014

Is quenching real? c.f. Pre-ordained evolu.on

8

2000 galaxies

Abramson+ 2016, Gladders+2013

Pre-ordained or cosmological conveyor belt? Does it maher?

9

Yes! The ques.ons you ask depends on what (picture) is in your head:

Also likely to give rather different perspec.ves on

• Gas content – driver or consequence?
• Chemical evolu.on – modified-closed-box or flow-through?
• Environment effects – imprinted at birth or accident later?
• Structure (spheroid vs. disk) – spheroids from short τ or at

early epoch?

Conveyer belt “grow-and-quench”

• What drives the evolu.on of the

Main Sequence?

• What quenches SF in galaxies and
• n what .mescale?

1010-1011-1012 1010-1011-1012

“Pre-ordained”

• Why is there apparent “down-sizing”
• Why is the SFR history in individual

galaxies apparently log-normal?

SF histories

Quenching

!"# !"$ ! " % ! " & !"' ! " ( ! " ) !"* +,-./0123456 7"! !"$ !"& !"( !"* !"! ! * 7 !7 859.:7;-,831<..=>,0-,6?43@ ) #"!..............................7!"!..............................77"!................... 859..A1??

Separability of fq(mstar,ρ) Baldry+ 2007, Peng+ 2010

fblue(m,ρ) = 1−εm(m)

( ) × 1−ερ(ρ)

( )

“mass quenching”

“environment quenching” (= satellites)

This terminology gives scope for confusion: which mass is relevant?

• Stellar mass?
• Black hole mass?
• Halo mass? But this is also

“environment”!!

log (1+δ) over-density log stellar mass Peng et al 2010

11

Mass-quenching

Peng+ 2010

Only “mass quenching” depends

• n mass, and therefore it is that

process that controls the shape

• f the mass func.on of the

surviving star-forming galaxies, i.e. Schechter M*. M* is ~ constant since z ~ 4 Possible physical processes for “mass-quenching”:

• Halo effects linked to halo mass mhalo

“Halo quenching”

• AGN feedback effects linked to black hole mass mBH

“AGN-quenching”

• Internal effects linked to mstar or structural effects such

as surface density Σ, or B/T, or σV etc “Morphological-” or “gravita.onal-quenching”

η = 1 M * ⋅ SFR ⇔ P(m) = exp m M *

( )

Caplar+15 fieng Ilbert+13 from Ilbert+13

Environment- (satellite-) quenching

13

“Environment/satellite-quenching” is that process that quenches a galaxy because it is a satellite of another galaxy, described by εsat=(1-fq,sat)/(1-fq,cen). εsat is strikingly independent of satellite galaxy mass vd Bosch+ 2008, Peng+ 2012, Wetzel+2012

Peng+ 2012

fq centrals fq satellites

εsat Physical possibili.es:

• Ram-pressure stripping of gas
• Tidal stripping
• Strangula.on (starva.on of gas

inflow)

• Harrassment by neighbours
• + others…

Why is εsat independent of satellite mass?

12 12.5 13 13.5 14 14.5

• 0.5

0.5 1 1.5 2 log mh log(δ + 1) εsat 0.2 0.4 0.6 0.8 1

• 0.5

0.5 1 1.5 2

• 1
• 0.5

0.5 log(δ + 1) log R εsat 0.2 0.4 0.6 0.8 1 12 12.5 13 13.5 14 14.5

• 1
• 0.5

0.5 log mh log R εsat 0.2 0.4 0.6 0.8 1

δ and mh R and δ R and mh

How does εsat vary with different parameters?

9 9.5 10 10.5 11 11.5

• 0.2

0.2 0.4 0.6 log msat εsat 0.5 1 10.5 11 11.5 12

• 0.2

0.2 0.4 0.6 log mcen εsat 0.5 1

• 1
• 0.5

0.5

• 0.2

0.2 0.4 0.6 log R εsat 0.5 1 12.5 13 13.5 14 14.5 log mh 0.5 1 1.5 2 log(δ + 1)

• 1

1 ∆ sSFRcen

msat mhalo mcentral δ radius sSFRcentral

Knobel et al 2014

Distribu.on of “drivers” (i.e. mhalo, R, and δ) are largely independent of msat. But why is the “response” to these drivers evidently independent of satellite mass?

How does εsat vary with different parameters?

9 9.5 10 10.5 11 11.5

• 0.2

0.2 0.4 0.6 log msat εsat 0.5 1 10.5 11 11.5 12

• 0.2

0.2 0.4 0.6 log mcen εsat 0.5 1

• 1
• 0.5

0.5

• 0.2

0.2 0.4 0.6 log R εsat 0.5 1 12.5 13 13.5 14 14.5 log mh 0.5 1 1.5 2 log(δ + 1)

• 1

1 ∆ sSFRcen

msat mhalo mcentral δ radius sSFRcentral δ mh mcen R

Knobel et al 2014

Timescales of environment/satellite quenching?

tquench tdelay Several clues:

• εsat independent of msat?
• εsat ~ 0.5, and increasing with

mhalo

• small ΔsSFR between

satellites and field (≤ 0.08 dex)

Es.mates of tdelay ~ 2-4 Gyr (cf 1 Gyr tff) See also Cibinel+2013, Carollo+2016

Wetzel+ (2013)

SF 20% suppressed in satellites

Delayed then rapid evolu.on

adapted from Wetzel+2013

Wetzel+ (2013) Wetzel+ (2012) Woo+ (2017)

Trying to iden.fy the rela.ve importance of different physical quenching processes

fQ

The difficul.es of establishing causality

18

Omand+ (2014)

A cau3onary example: Very strong observed correla.ons between quenched state (sSFR) of a galaxy and

• surface mass density Σ (Kauffmann+ 2003),

with Σcrit evolving as (1+z)2 (Franx+ (2008).

• velocity dispersion (Smith+ 2009, Wake+

2012).

• Sersic indices (Blanton+ 2003, Wuyts+ 2011).
• Bulge mass (Bluck+ 2015).

Structure could produce quenching directly, e.g. via disk stability (Mar.g+2009, Genzel+ 2014 )

• r indirectly (e.g. via black hole mass following

mbulge etc) constant Σe But we know re(m) at fixed m evolves roughly as (1+z)-1. Are quenched galaxies dense because density quenched them, or simply because they stopped evolving at high z when all galaxies were denser?

fQ

Appearance of strong Σ threshold Lilly & Carollo 2016

Very strong correla.on with Σe (and difference between centrals and satellites) both reproduced by a toy- model in which hSF evolves as mstar

0.3 (1+z)-1 and quenching depends only on mass and not at all on structure!

Success

average star-forming Re(m) average all Re(m) average quenched Re(m)

centrals satellites

Where Σthresh comes from and why satellites are different than centrals

Conclusion: The strikingly strong correlation of quenched fraction and surface mass density Σe is likely a coincidental result of quenching, not a driver of it. But what about spheroids?

contours of constant fq

constant Σ

Explana.on

21

Two (poorly understood) clues about quenching

• Galac.c conformity
• The coincidence of quenching

• 1. Conformity

22

• 0.2

0.2 0.4 0.6 0.8 1

εsat fq|sat

# Gal: 1155 quenched centrals SF centrals

9 9.5 10 10.5 11 11.5

• 0.2

0.2 0.4 0.6 log msat εsat 0.5 1 10.5 11 11.5 12

• 0.2

0.2 0.4 0.6 log mcen εsat 0.5 1

• 1
• 0.5

0.5

• 0.2

0.2 0.4 0.6 log R εsat 0.5 1 12.5 13 13.5 14 14.5 log mh 0.5 1 1.5 2 log(δ + 1)

• 1

1 ∆ sSFRcen

“Conformity” (Weinmann+ 2006) Satellite quenching is 2.5 .mes stronger with quenched centrals even when you match the satellites in all five of Mhalo, mcen, msat, R, δ

quenched central SF central

(One-halo) conformity suggests that there is a very close connec.on between mass-quenching and environment-quenching via “halo-wide” effects Knobel et al (2015)

• 2. The coincidence of quenching Birrer et al 2014

23

Why does “quenching” happen just as mstar/mhalo approaches to within a factor

• f a few of the maximum efficiency (cosmic baryon frac.on)? What is this

telling us?

• bserved rela.on

from Moster (2010) satura.on of baryon conversion without quenching

cosmic baryon limit

naïve extrapola.on of conversion efficiency

Increasingly efficient conversion

• f stars to baryons in galaxies

due to decreasing winds (Mass-) quenching as required by constant M*SF

The explana.on? a link between SN- and AGN-feedback?

24

Black holes are able to grow to “dangerous” levels only when SN-driven feedback becomes less efficient at expelling gas from the galaxy or halo. 1. Hydro simula.ons (Seth: Dubois+2015, EAGLE: Bower+2017) 2. Semi-analy.c models - see Henriques et al, 2017 (in prep.) 3. Phenomenological analysis of AGN and galaxy popula.ons see Caplar et al, in prep 2017.

Is AGN feedback responsible for quenching?

25

OuSlows of gas in AGN produced by “feedback” from ac.ve AGN: There is good evidence for major

• uSlows of ionized and molecular gas (e.g.

Cicone+2014, Fiore+2017) but:

• this is not the same as quenching (e.g. high mass-

loading λ in star-forming galaxies) and

• These ouSlows may well not affect gas in star-

forming disk (e.g. Gabor+Bornaud 2014) “Radio-mode” energy injec.on from quiescent BH into the hot gas in the CGM/ICM (e.g. Croton 2006)

• There is good evidence of interac.on (bubbles etc).
• Main argument: can’t think of anything else

energe.c enough, so we put it in by hand (propor.onal to BH mass).

But is the “quenching mechanism ques.on” well-posed?

26

What if there is a gate-keeper? i.e. Process A keeps trying to quench the galaxy, but Process B doesn’t allow it un.l some other condi.ons are sa.sfied Which of A or B is then actually “quenching”?

Back to the role of gas: major ques.ons

27

What is the rela.ve importance in quenching of

• fast removal of gas
• fast consump.on of gas
• cessa.on of inflow followed by slow deple.on (“starva.on”)
• suppression of star-forma.on efficiency

Prospec.ve/promising diagnos.cs:

• Rela.ve metallici.es of stars in SF and quenched galaxies

(see Peng+ 2015)

• Timescales of quenching (e.g. Wetzel 2013)
• (Molecular) gas content of quiescent galaxies

Points to take away

28

General points The outcome of complex galaxy evolu.on seems to have been remarkably

• simple. Focusing on these simple outcomes:
• constrains output of more physical models
• can provide new insights (e.g. reversed causality of regulator)
• can show what should not be causally inferred from observa.ons

More specific points:

• The constant quenching mass of mass-quenching (stellar, also halo?)
• ver a wide range of cosmic .me.
• The separability of mass- and environment- quenching may well hide

deep connec.ons between the two, as shown by conformity and shows that quenching is a halo-wide phenomenon

• The “coincidence” of quenching links the failure of SN-feedback with

quenching à BH driven quenching via the halo?

• Major task ahead: understanding role of gas in quenching (starva.on,

consump.on, ejec.on, etc).