With:

V. Smolcic, A. Karim,, B. Magnelli, A.Zirm, M. Michalowski, P. Capak, K. Sheth, K. Schawinski, S. Wuyts, D. Sanders, A. Man, D. Lutz, J. Staguhn, S. Berta, H. McCracken,

The highest redshift Sub-mm galaxies as progenitors of compact quiescent galaxies

Sune ToftDark Cosmology Centre

Dark GroupA. Zirm, A. Man, J.-K. Krogager, K. Olsen

“Connecting the Extreme”

Compact Quiescent Galaxies at z=2 (cQGs)

Photometric surveys:The fraction of quiescent galaxies increases rapidly between z=3 and 2

At z=2, half of the most massive galaxies are quiescent with little ongoing star formation and evolved stellar populations

ncQG=6.0±2.1 × 10-5 Mpc-3

(Brammer++, 2011), Ilbert et al 2012)

Restframe optical absorption line spectroscopy

Broad wavelength continuum fits: Post-Starburst with strong Balmer abs.

Line indices: Metal rich, 1-2 Gyr old Velocity dispersions 300-500 km/s

(e.g. Toft et al 2012; van de Sande et al 2012, Onodera et al, 2012 )

Spectroscopic sample of z=2 cQGS

COSMOS 3DHST+CANDLES• Strong 4000A break z-spec• SED fit M*, Av, age, zform• Galfit HST/F160W size

“Complete”

(Krogager, Zirm, Toft, Man & Brammer, ApJ (Submitted))

Mass-size relation (complete spectroscopic sample)

(Krogager, Zirm, Toft & Brammer, 2013)

3DHST/CANDELS:

The mass-size relation at z=2 is shifted to ~3 times smaller sizes at a given mass, with respect to the local relation(slope & scatter identical)

>10 times larger stellar mass densities

How did they form?

NIR spectroscopy:• Post-starburst spectra• Baryon dominated• Some dust (Av=0-1)• 1-1.5 Gyr -> zform>3(Toft++2012, van de Sande++2012, Onodera++2012, K13)

“Main sequence” star forming galaxies at z>3? • SFRmin>115 M/yr from z=10 to zform

>3 times higher than observed for z=3 LBGs (Carilli++ 2008)

• Number density of z>3 LBGs with M>1011 M << ncQG (Stark++ 2009)

-> Progenitors must be dust obscured starbursts

(Krogager++, 2013)

Dust obscured nuclear starbursts?

(Hopkins++ 2006)

(Wuyts ++ 2010) Remnants very compact, concentrated (high sersic n)

Sub-mm galaxies:• Prime examples of high-z dusty nuclear

starbursts• Many authors have suggested a

connection to cQGs (e.g. Tacconi++2006, Capak++2008, Toft++2009, Riechers++2013)

Simulations of gas rich major mergers

SMGs as progenitors?

cQG SMGs

M* >1011 M >1011 M

Int. Velocity (σ★)=300-500 km/s FWHMCO(1-0)=350-800 km/s

<velocity> <σ★>=363 ±30 km/s <Vc>=392 ±134 km/s

<size> <re>=2.0 ± 0.2 kpc <re>= 2.0 ± 0.3 kpc<Mdyn> (2.3 ± 1.4) ×1011 M(2.5 ± 1.3) ×1011 M

<Redshift> zform> 3 <z>obs =2

(Tacconi et al 2006, 2008, Ivison et al 2011 van de Sande et al 2012, Toft et al 2012, , Krogager et al in prep,)

At z=2 SMGs (probed through CO emission), have many similarities with cQGs (probed through their stars)

High redshift SMG sample

COSMOS Aztec/JCMT/SMA sample

“Statistical Sample”:Flux (F1.1mm>4.2mJy) & S/N (>4.5) –limited over 0.15☐o

Redshifts peak at <z>=3

11 galaxies with z>3 (5 with spec-z)

n(z>3) = 2.1 ± 0.4 × 10-6 Mpc-3

(Smolcic et al 2012)

Redshift distribution match

cQG (zform)SMG (zobs)

SMG surface brightness fits

Fit 2D surface brightness profiles with galfitStacked YJHK UltraVISTA images (WFC3/F160W where avail) Very compact, small sersic n>Half have bright multiple components

Mass-size relation

z=2 cQG z>3 SMG

Duty cycle of SMG starburst1: Assume :-SMGs are direct progenitors of cQGs-Each progenitor only undergo one SMG phase

2: Require number densities to match:

Timescale of SMG starburst tburst (duty cycle)

• Consistent with Independent estimates tburst= 40-200 Myr

(gas depletion, clustering analysis, merger simulations)

• Relatively independent of IMF

nSMG,z>3 = 2.1 ± 0.4 × 10-6 Mpc-3

nq,z=2= 6.0 ± 2.1 × 10-5 Mpc-3

MIR-FIR SED fits

Fit FIR SEDs with DL07 models

Data:Spitzer MIPS, Herschel PACS, SPIRE, AzTEC, LABOCA, MAMBO, SMA, CARMA, PdBI

Derive: LIR, SFR, Mdust, Mgas

Eddington Limited Bursts?

(Younger et al 2010)

z=2 cQG z>3 SMG

Maximum SFR of cQGs during formation (Eddington limited burst)

Additional Stellar mass growth

z=2 cQG z>3 SMG

Z>3 stellar mass distribution broader than that of z=2 cQGs

Ongoing starbursts in the SMG will increase their stellar mass

ΔM★ ~ Mgas× η

Star formation effeciency η~0.1-0.15(Hayward++ 2011)

Mgas from from Mdust (derived from FIR SED fits) assuming a mass and metallicity dependent dust-to-gas ratio

Quenching by Active Galactic Nucleii?

CDFS X-ray observationsDeepest X-ray observations (4 million seconds)

70-100% of z=2 quiescent galaxies host AGN

(Olsen, Rasmussen, Toft & Zirm, 2013)Stack: 50x4 Ms = 200 Ms

22%

“Connecting the Extreme”

(Toft et al , 2014)

Summary

Credit: HST press office

Extra slides

Star formation efficiency

Gas fraction decrease by 10-15% from peak of starburst to when it is quenched

Disc Merger

(Hayward++, 2011)

Transition population

(Barro++, 2012)

Population of compact starforming galaxies at z~3 with depressed SFR and enhanced AGN fraction.

Not quiet as massive, but sizes and number densities match z=2 cQGs

Descendants of z=2 SMGs

(Bezanson ++2013)

Population of compact post starburst galaxies at z=1.5, with high stellar and dynamical masses and zform~2