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

Galaxies: complexity in process and simplicity in outcomes Physical processes in individual objects: Complexity ) * The 859.:7;-,831<..=>,0-,6?43@ ! " # !"$ 7 !"% evolving !"& !"' !"( ! " ) popula.on: ! ! " * Simplicity ! 7 #"!..............................7!"!..............................77"!................... 859..A1?? !"! !"* !"( !"& !"$ 7"! +,-./0123456

The Main Sequence: what do we mean by “quenching”? Brinchmann et al 2004 Whitacker et al 2014 Quenching refers to what makes some galaxies lie in the “red cloud” rather than the Main Sequence, with sSFR << sSFR MS 3

Main Sequence evolu.on from a gas regulator Lilly+2013 halo$ $$gas$inflow$into$halo$ wind$ou6low$ gas$inflow$into$galaxy$ � Φ $ galaxy$ system$ variable$gas$reservoir$ return$ star, forma?on$ long,lived$stars$ inflow ouSlow Ψ = (1 − r ) SFR + Φ + dm gas dt star-forma.on change in Wind-loading SF efficiency reservoir ⎛ ⎞ m gas ε = SFR λ = Φ − 1 ⎟ = sSFR × ε − 1 = τ dep ⇔ ⎜ SFR m gas m star ⎝ ⎠ 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 – m gas /m star via ε 4

Metallicity as a test of the gas-regulator Lilly et al (2013) ! ! “Chemical evolu.on” as ! !" = ! ! + 1 + ! 1 − ! ! ! + ! ! ! ∙ !"#$ + 1 − ! ! ! !"#$ changing state of regulator x !" ( à FMR) = ! ! + ! !"#$ ! ! • explains (s)SFR as second parameter in Z(m) • Z(z) reflects the state of regulator system ( ε , λ ,sSFR) not its history (because τ gas << τ H ) • Z(m,SFR) which will change with z only if ε or λ change (i.e. expect redshid-independent FMR) η Z ∝ m star η Z ∝ m star • Three-way links: Z(m), m star /m halo , and sSFR/sMIR 1(1 − η ) ~ m halo 1 1.7 m star ∝ m halo 1(1 − η ) ~ m halo 0.15 9.1 !"#$% ~ 1 − ! !"#$ ! ~ ! 2 ! !"#$ ! 1.7 m star ∝ m halo 0.1 0.5 9 0.05 0.5 0 log(SFR) 8.9 0 − 0.5 Fieng this surface: − 0.05 8.8 0 − 1 τ dep ~ 2.5m 10 -0.3 Gyr − 0.1 log(SFR) 8.7 − 1.5 λ ~ 0.3 m 10 -0.8 9 9.5 10 10.5 11 11.5 log(m * ) 0.15 8.6 − 0.5 0.1 0.5 8.5 0.05 0 log(SFR) − 1 0 8.4 − 0.5 − 0.05 8.3 − 1 − 0.1 − 1.5 − 1.5 8.2 9 9.5 10 10.5 11 11.5 9 9.5 10 10.5 11 11.5 log(m * ) log(m * ) Z(m,SFR) from Mannucci+ 2010 5

The “grow and quench” paradigm Factor of 20 decline s s a m in sSFR MS since z = 2 g o l SFR Growth of dark maher haloes s r a t S Rate of cosmic evolution driven by growth passive of haloes galaxies stellar mass 1970’s cartoon by Bruno Bingelli? Continuous “flow” of galaxies along the conveyer belt 6

Cosmological importance of quenching Behroozi et al 2013 Ω B / Ω cdm ceiling “Quenching” Inefficient star- forma.on (winds) m star /m halo vs. m halo Moster+ (2010) Birrer+ (2014) reproducing Behroozi+ (2013) Quenching Slowing down of x10 structure growth leading to decline in sSFR MS Half due to decrease of SF MS Inefficient SF due to winds and half due to quenching 7 From Birrer et al 2014 Madau+Dickinson 2014

Is quenching real? c.f. Pre-ordained evolu.on 2000 galaxies Abramson+ 2016, Gladders+2013 8

Pre-ordained or cosmological conveyor belt? Does it maher? Yes! The ques.ons you ask depends on what (picture) is in your head: SF histories Conveyer belt “grow-and-quench” • What drives the evolu.on of the 10 10 -10 11 -10 12 Main Sequence? • What quenches SF in galaxies and on what .mescale? “Pre-ordained” • Why is there apparent “down-sizing” • Why is the SFR history in individual galaxies apparently log-normal? 10 10 -10 11 -10 12 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? 9

Quenching

Separability of f q ( m star , ρ ) Baldry+ 2007, Peng+ 2010 Peng et al 2010 ) “environment quenching” (= satellites) * log (1+ δ ) over-density ( ) ( ) × 1 − ε ρ ( ρ ) f blue ( m , ρ ) = 1 − ε m ( m ) “mass 859.:7;-,831<..=>,0-,6?43@ !"# quenching” !"$ 7 ! " % ! " & !"' This terminology gives scope for ! " ( confusion: which mass is relevant? ! " ) • Stellar mass? ! !"* • Black hole mass? • Halo mass? But this is also “environment”!! ! 7 #"!..............................7!"!..............................77"!................... 859..A1?? log stellar mass !"! !"* !"( !"& !"$ 7"! 11 +,-./0123456

Mass-quenching Peng+ 2010 Only “mass quenching” depends from Ilbert+13 on mass, and therefore it is that process that controls the shape of the mass func.on of the surviving star-forming galaxies, i.e. Schechter M*. M* is ~ constant since z ~ 4 1 P ( m ) = exp m M * ( ) M * ⋅ SFR η = ⇔ Caplar+15 fieng Ilbert+13 Possible physical processes for “mass-quenching”: Halo effects linked to halo mass m halo “Halo quenching” • AGN feedback effects linked to black hole mass m BH “AGN-quenching” • Internal effects linked to m star or structural effects such • as surface density Σ , or B/T, or σ V etc “Morphological-” or “gravita.onal-quenching”

Environment- (satellite-) quenching Peng+ 2012 “Environment/satellite-quenching” is that process that quenches a galaxy because it is a satellite of another galaxy, described by f q satellites ε sat =( 1 - f q,sat )/( 1 - f q,cen ). ε sat is strikingly ε sat independent of satellite galaxy mass vd Bosch+ 2008, Peng+ 2012, Wetzel+2012 f q centrals Physical possibili.es: Ram-pressure stripping of gas • Tidal stripping • Strangula.on (starva.on of gas • inflow) Harrassment by neighbours • + others … • 13

How does ε sat vary with different parameters? Knobel et al 2014 1 1 δ and m h 2 0.5 0.8 0 1.5 0.6 0.6 log( δ + 1) 1 ε sat 0.4 0.4 ε sat 0.5 0.2 0.2 m sat m halo 0 0 0 -0.2 -0.5 9 9.5 10 10.5 11 11.5 12.5 13 13.5 14 14.5 12 12.5 13 13.5 14 14.5 log m sat log m h log m h 0.5 1 1 R and δ 0.5 0.8 0 0 0.6 0.6 log R ε sat 0.4 0.4 ε sat -0.5 0.2 0.2 m central 0 δ 0 -1 -0.2 -0.5 0 0.5 1 1.5 2 10.5 11 11.5 12 0 0.5 1 1.5 2 log( δ + 1) log m cen log( δ + 1) 0.5 1 R and m h 1 0.8 0.5 0 0 0.6 log R ε sat 0.6 0.4 0.4 -0.5 ε sat 0.2 0.2 sSFR central radius 0 0 -1 12 12.5 13 13.5 14 14.5 -0.2 log m h -1 -0.5 0 0.5 -1 0 1 ∆ sSFR cen log R Why is ε sat independent of satellite mass?

How does ε sat vary with different parameters? Knobel et al 2014 1 m h 0.5 0 0.6 0.4 ε sat 0.2 m sat m halo 0 -0.2 9 9.5 10 10.5 11 11.5 12.5 13 13.5 14 14.5 log m sat log m h 1 0.5 δ 0 0.6 0.4 ε sat 0.2 m central 0 δ -0.2 10.5 11 11.5 12 0 0.5 1 1.5 2 log m cen log( δ + 1) 1 0.5 m cen 0 0.6 0.4 ε sat 0.2 sSFR central radius 0 -0.2 -1 -0.5 0 0.5 -1 0 1 ∆ sSFR cen log R Distribu.on of “drivers” (i.e. m halo , R, and δ ) are R largely independent of m sat . But why is the “response” to these drivers evidently independent of satellite mass?

Timescales of environment/satellite quenching? Wetzel+ (2012) Woo+ (2017) Several clues: • ε sat independent of m sat ? • ε sat ~ 0.5, and increasing with SF 20% m halo suppressed in satellites • small Δ sSFR between satellites and field (≤ 0.08 Wetzel+ (2013) dex) Delayed then rapid evolu.on adapted from Wetzel+2013 Wetzel+ (2013) t quench t delay Es.mates of t delay ~ 2-4 Gyr (cf 1 Gyr t ff ) See also Cibinel+2013, Carollo+2016

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

The difficul.es of establishing causality constant Σ e Omand+ (2014) A cau3onary example: Very strong observed correla.ons between f Q quenched state (sSFR) of a galaxy and • surface mass density Σ (Kauffmann+ 2003), with Σ crit 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 ) or indirectly (e.g. via black hole mass following m bulge etc) But we know r e (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? 18

Appearance of strong Σ threshold Lilly & Carollo 2016 f Q Very strong correla.on with Σ e (and difference between centrals and satellites) both reproduced by a toy- model in which h SF evolves as m star 0.3 (1+z) -1 and quenching depends only on mass and not at all on structure! Success