on the co- on the co- evolution of evolution of galaxies and - - PowerPoint PPT Presentation

• n the co-
• n the co-

evolution of evolution of galaxies and their galaxies and their black holes black holes

rac he l so me rville

Max Planc k I nstitute fo r Astro no my, He ide lbe rg

with thanks to E. Bell, X. Zheng, H.-W. Rix, & the GEMS team

• P. Hopkins, B. Robertson, T.J. Cox, L. Hernquist, Y. Li

mergers as drivers of galaxy and BH growth

trigge rs o f po we rful bursts o f star fo rmatio n? re spo nsible fo r transfo rming disks into sphe ro ids? drive rs o f gas into galac tic nuc le i, fe e ding the c e ntral BH, & pro duc ing an AGN?

questions

ho w signific ant are me rge rs in driving the c harac te ristic rise and fall o f the glo bal SF R and do wnsizing? ho w signific ant are me rge rs in driving the c harac te ristic rise and fall o f BH ac c re tio n and do wnsizing? are o bse rvatio ns o f me rge r rate s and QSO L F c o nsiste nt with the m be ing c lo se ly asso c iate d pro c e sse s?

COMBO-17 + GEMS + Spitzer

c o lo rs, pho to -z, ste llar masse s to z~1 fro m

COMBO-17; HST imag ing fro m GE MS

Spitze r 24m data fro m the MI

PS instrume nt te am (83Jy --> 3M yr

• 1 at

z~0.7)

Spitzer image Spitzer image COMBO-17 and COMBO-17 and GEMS coverage GEMS coverage Rieke Rieke et al. 2004; data described in et al. 2004; data described in Papovich Papovich et al. 2004 et al. 2004

Stellar mass function split by color

Borch Borch et al. et al. 2006 2006 Local Local MFs MFs Weak evolution Weak evolution in MF of blue guys in MF of blue guys (~ disks) (~ disks) Strong evolution Strong evolution in red guys at in red guys at L<2L* at least L<2L* at least (~ spheroids) (~ spheroids)

Blue cloud Blue cloud Red sequence Red sequence Bundy et al. Bundy et al. 2006 2006

global stellar mass evolution by color

• ste llar mass de nsity in blue

c lo ud ~c o nstant sinc e z~1

• ste llar mass de nsity in re d

se que nc e has inc re ase d by a fac to r o f 2-3 sinc e z~1

Chen et al. 03 Chen et al. 03

Borch Borch et al. 2006 et al. 2006

Blue cloud Blue cloud Red sequence Red sequence

Bundy et al. Bundy et al. 2006 2006

which galaxies contribute to the UV luminosity density at z=0.7?

0.65<z<0.75 Wolf et al. 2005

contributions to global SFR budget by type at z=0.7

Wolf et al. 2004

the IR view

Re d E / S0s are no n-star-fo rming Mo st SF is in large spirals 20-30% o f to tal SF at z=0.7 in manife stly inte rac ting syste ms

Bell et al. 2005 at z=0.7:

U-V rest V

Most IR luminous galaxies in GEMS are relatively normal looking spirals

LIRG, z = 0.24 LIRG, z = 0.48 LIRG, z = 0.46 ULIRG, z = 0.68 LIRG, z = 0.33 LIRG, z = 0.33

Evolution of IR LF

IR LF very strongly IR LF very strongly evolving evolving Almost all SF is Almost all SF is in blue disks in blue disks

Total Total Blue / disk Blue / disk Local Local Red Red Le Le Floc Floc’ ’h h et al. 2005 et al. 2005 Bell et al. 2005 Bell et al. 2005

SF and mass assembly split by color

c o mpare inte grate d SF R with o bse rve d gro wth in ste llar mass mo st star fo rmatio n is in blue s, but their mass de nsity do e s no t gro w! ne e d a pro c e ss that mo ve s galaxie s fro m blue to re d pile ...

Bell et al., in prep., Le Bell et al., in prep., Le Floc Floc’ ’h h et al. 2005 et al. 2005

total blue red

mergers?

direc t measure o f merger frac tio n fro m c lo se pair frac tio n in COMBO agrees fairly well with o ther results fro m mass- se le c te d samples; SAM

Patton et al. 2002 Patton et al. 2002 Bundy et al. 2004 Bundy et al. 2004 Le Le Fevre Fevre et al. 2000 et al. 2000 Conselice Conselice et al. 2003 et al. 2003 Somerville et al. in prep. Somerville et al. in prep.

* *

Bell et al. 06 Bell et al. 06

Bell, Phelps, rss et al. 2006

Mergers?

Patton et al. 2002 Patton et al. 2002 Bundy et al. 2004 Bundy et al. 2004 Le Le Fevre Fevre et al. 2000 et al. 2000 Conselice Conselice et al. 2003 et al. 2003 Somerville et al. in prep. Somerville et al. in prep.

* *

Bell et al. 06 Bell et al. 06

implied ‘transfo rmatio n’ rate is highe r, and ste e pe r, than the ‘dire c t’ e stimate s o f me rge r rate n.b. ke e p in mind me rge r rate s are hard to me asure by any means! Bell, Phelps, rss et al. 2006

co--evolution of global accretion rates

Zheng, Bell, rss, Rix, Jahnke et al. in prep

diamonds = accretion rate based on hard X-ray LF x BH(z=0)/ *,sph(z=0) ~2000

co-evolution in mass and accretion rate

Zheng et al. in prep

accretion onto massive BH assuming distribution of L/LEdd from Kollmeier et al.

SFR in all galaxies is too much... SFR in massive spheroids not enough SFR in (all) massive galaxies is about right

co-evolution by mass and accretion rate

take L

F

• f o ptic al QSOs.

c o nve rt L

• bs to L

bo l using standard bo lo me tric

c o rre c tio ns.

L

bo l --> ac c re tio n rate assuming e ffic ie nc y o f

c o nve rsio n o f re st mass to e ne rg y o f 10%

c o nve rt L

bo l to MBH using distributio n fro m

K

• llme ie r e t al. 2006 base d o n line widths

(<L

bo l/ L E dd>~0.25, ~0.3 de x)

plo t ac c re tio n rate c o ntribute d by BH abo ve

a g ive n mass limit

• pen circles:

black hole accretion filled: star formation Zheng, Bell, rss, Rix, Jahnke et al. in prep. mass limit mass & SF limit

AGN host morphologies in GEMS:

• 4/15 appear to be mergers
• 2/15 have exponential profiles (n<2.5)
• majority have deV-like (n>2.5)

profiles

Sanchez et al. 2004

colors of star forming galaxies vs. AGN hosts

Sanchez et al. 2004 Bell et al. 2005 V-band magnitude U

• V

c

• l
• r

1.5 galaxies z=0.7 AGN spirals mergers

Hydrodynamic simulations

• f galaxy mergers including

black hole growth and feedback

di Matteo, Springel & Hernquist Springel, di Matteo & Hernquist Robertson et al. Hopkins et al.

• sub-grid model of SF

in multi-phase ISM

• Bondi accretion onto

central BH

• thermal energy from stars

and AGN returned to ISM

characteristic QSO lightcurves

Hopkins et al. 2005

• QSO accretes

at << Ledd over most of its lifetime

• ’visibility time’

much longer at faint luminosities

• obscuration is

largest during peak of accretion

• optical QSO

becomes visible during ‘blowout’ phase

parameterization of lightcurves

differential time spent in a given logarithmic luminosity interval: dt/dlog(L) = tQ (L/LQ) exp[-L/LQ] where LQ propto peak lum or final BH mass, ~const Hopkins et al. 2005

mapping the observed merger LF and QSO LF

we c an use the se parame te rizatio ns

• f starburst & QSO lightc urve s to

c o mpute statistic al mappings be twe e n QSO lumino sity func tio ns, L F

• r mass func tio n o f me rging

galaxie s, galaxy-galaxy me rge r rate , ho st lumino sity, e tc .

Hopkins, rss, Hernquist, Cox, Robertson & Li 2006

QSO LF predicted from

• bserved merger LF

z=0.2 z=0.7

merger LF from

• bserved QSO LF

inferred (gas rich) merger mass function

• bservations: hard X-ray QSO LF (Ueda et al. 2003)

merger LF at z=0.2 from 2MASS (Xu et al.) & z=0.7 from GEMS

Hopkins, rss, Hernquist, Cox, Robertson & Li 2006

star formation rate function of mergers GEMS mergers predicted from QSO LF

Hopkins, rss et al. 2006 mergers contribute ~1% to the global SFR at z=0, 15-25% at z~1, 30-50% at z~2

primordial power spectrum star formation SN feedback chemical enrichment photo-ionization squelching collisional heating radiative cooling merger tree, sub-halo merging stellar populations dust absorption & emission galaxy galaxy

• bservables
• bservables

BH formation, AGN feedback

the need for AGN feedback in cosmological models

• verc o o ling pro blem: to o large a

frac tio n o f gas c o o ls into galaxies; huge exc ess o f ultra-massive/ lumino us blue galaxies at z~0 inverted c o lo r-magnitude relatio n & mass-age relatio n; dearth o f massive red galaxies at high redshift weak o r no c o lo r bimo dality dec rease in number density o f lumino us QSOs; AGN ‘ do wnsizing’

AGN feedback mechanisms I

• during pe rio ds o f e ffic ie nt

fe e ding (asso c iate d w/ me rge r?) we pro duc e a lumino us AGN/ QSO

• the rmal c o upling o f AGN

e ne rgy with I SM is pro bably fairly we ak (5%?), and duty c yc le sho rt

• BH gro wth se lf-re gulate d

(pro duc e s MBH- re latio n)

• AGN c an drive a wind

that ‘ swe e ps up’ galaxy Di Matteo, Springel & Hernquist 2005

AGN feedback mechanisms II

perio ds o f lo w ac c re tio n e ffic ie nc y (ADAF ?) asso c iate d with je t fo rmatio n ene rgy c o uple s with gas ve ry e ffic ie ntly (~100%?) and duty c yc le is lo ng resulting bubble s lo o k similar to tho se se e n in Chandra image s

• f so me c luste rs

QuickTime™ and a decompressor are needed to see this picture.

Bruggen, Ruszkowski & Hallen 2005

AGN feedback III

belo w a c ritic al halo mass ~few x 1011- 1012 Msun, infalling gas never sho c k heats to the virial temperature but ‘ falls in c o ld’ abo ve this c ritic al mass, gas sho c k heats to fo rm a ho t halo , then c o o ls in a ‘c o o ling flo w’ (Birnbo im & Dekel, K eres e t al.) ‘radio mo de’ feedbac k o nly affec ts ho t mo de c o o ling (Cattaneo e t al., Cro to n et al., Bo wer et al.)

‘Effervescent’ heating by giant radio jets

• re c e nt wo rk sugge sts

e ve n c o lumnate d je ts c an he at a large filling fac to r o f I CM

• re sulting bubble s lo o k

similar to tho se se e n in Chandra image s o f so me c luste rs Bruggen, Ruszkowski & Hallen 2005

• each top level halo seeded

with a 100 Msun BH

• c o o ling, quie sc e nt star

fo rmatio n, c hemic al evo lutio n and superno va feedbac k mo delled within e ac h galaxy Modelling AGN and BH growth in a Hierarchical Cosmological context

me rging o f galaxie s within DM halo s via dynamic al fric tio n me rge rs trigge r bursts o f star fo rmatio n and ac c re tio n o nto BH; e ffic ie nc y and time sc ale parame te rize d base d o n hydro me rge r simulatio ns (μ, B/ T , Vc, fg, z; Co x e t al., Ro be rtso n e t al.) BH ac c re te at E ddingto n until the y re ac h ‘ c ritic al mass’ , the n e nte r ‘ blo wo ut’ (po we r-law de c line ) phase dmac c/ dt = mE

dd/ [1+(t/ tQ)]

(Ho pkins e t al. 2005)

furthe r star fo rmatio n is suppre sse d as lo ng as mBH>mc rit

.

log Mh = 12.1, lg m*=10.6, mr=-20.6, lg MBH=7 B/T=0.22 total mass

stellar/ bulge mass

cold gas

black hole mass

star formation

BH accretion

log Mh = 13.2, lg m*=11.3, mr=-21.8, lg MBH=8.4 B/T=0.9 total mass

stellar/ bulge mass

cold gas

black hole mass

SFR

BH accretion rate

• number of major

mergers experienced

• ver a galaxy’s history

is a strong function of mass

predicted MBH-Mbulge relationship

large symbols: Haering & Rix data purple: H&R fit + scatter intrinsic scatter: 0.3 dex green: predicted median, 10th, & 90th percentile predicted scatter: ~0.15 dex

merger rates

m1/m2 >0.1 m1/m2 > 0.25 S-S major mergers S-E major mergers E-E major mergers major mergers with two BH Mbh>106 Msun major mergers with m*>2.5x1010 Msun z=5 z=2 z=1

stellar mass function

u

• r

c

• l
• r

r-band magnitude predicted color-magnitude relation z=0 Baldry et al. red sequence and blue cloud

color distributions

black=Baldry et al. SDSS; purple=model u-r with AGN feedback

stellar mass function evolution

data from Borch et al. (COMBO-17); Drory et al. (MUNICS, GOODS, FDF) rss et al. in prep

mass-assembly

data from

Borch et al. (COMBO-17) Drory et al. (GOODS, FDF) Glazebrook et al. (GDDS) Fontana et al. (K20)

still produce (at least) enough massive galaxies at all redshifts where we have observations

rss et al. in prep

z=1 rest-frame color-magnitude relation

COMBO-17 red sequence (Bell et al.)

z=2 rest-frame color-magnitude relation

GOODS data

extrapolated COMBO-17 red sequence line

downsizing?

data from Feulner et al. Papovich et al. Brinchmann et al. (z=0)

100 Msun/yr

massive galaxies: log m*=10.7 Msun expected downsizing for Kennicutt + constant gas fraction

star formation rate function of mergers GEMS mergers

SAM major mergers SAM all mergers

predicted from QSO LF

accretion rate x2000 SF in major mergers

SFR and BH accretion history in SAM

conclusions

de c line in ac c re tio n ac tivity o nto the mo st massive BH and de c line o f SF ac tivity in the mo st massive galaxie s trac k e ac h o the r in a manne r c o nsiste nt with ‘stric t’ c o -e vo lutio n de c line in the glo bal SF R sinc e z~1 o nly abo ut 1/ 3 due to de c lining me rge r rate

• nly abo ut 1/ 3 o f the re d-blue transitio n

sinc e z~1 due to me rge rs de c line in QSO ac tivity sinc e z~1 c o nsiste nt with o bse rve d de c line in me rge r rate at all (o bse rvable ) lumino sitie s

• pen questions

we kno w that the site s o f ac tive star fo rmatio n and BH ac c re tio n are no t

• c c urring in the same o bje c ts at

(e xac tly) the same time why the n do the glo bal SF R and BH ac c re tio n rate , and e ve n the mass- limite d quantitie s, trac k e ac h o the r so we ll? lo w-le ve l ac c re tio n that trac ks ‘quie sc e nt’ star fo rmatio n? bo th re gulate d by glo bal gas supply?

SFR vs. SFH

Borch Borch et al. 2006 et al. 2006

IR luminosity from 24m flux

Re st-frame 12-15m c o rre late s stro ngly with to tal I R lumino sity in the lo c al Unive rse , with < x2 sc atte r Will be able to te st I R flux e stimate s with Spitze r 70,160m, Ape x 350m and 870m and He rsc he l PACS and SPI RE

Chary & Chary & Elbaz Elbaz 2001; 2001; Papovich Papovich & Bell 2002; & Bell 2002; See also Dale et al. 2005 See also Dale et al. 2005

Log Log10

10 L

LIR

IR/L

/L

• Log

Log10

10

• L

L

• /L

/L

• <0.3

<0.3 dex dex scatter scatter

co-evolution of downsizing

Zheng et al. in prep