millimeter/submillimeter studies of high-redshift galaxies deciphering the ancient universe with...

Millimeter/Submillimeter Studiesof High-Redshift Galaxies

Deciphering the Ancient Universewith Gamma-Ray BurstsApril 20, 2010, Kyoto, Japan

Kotaro KOHNOIoA/BBC, Univ. of Tokyo

[email protected]

Title image: Yoichi Tamura

Outline• Why we observe high-z galaxies at

millimeter/submillimeter wavelengths?• Mm/submm selected galaxies (SMGs), one of the

most massive populations in the early universe, with highly elevated star formation rates (SFRs)– Recent hot topics, such as clustering and dark halo

masses, redshifts, ultra-bright populations, etc.– AzTEC on ASTE survey of SMGs

• Mm/submm studies on GRB host galaxies, a metal poor, less massive population in the early universe– Probing a molecular gas and hidden star formation

Why we observe high-z galaxies at millimeter/submillimeter

wavelengths?

Because a large portion of the cosmic star formation is obscured by dust

More hidden star formation

@higher redshift (>70% at z~1)

Takeuchi, Buat, & Durgarella 2005, A&A, 440, L17

Based on the comparisonof Lum. Functions at FUV with GALEX & IR with IRAS/Spitzer

Sta

r fo

rmat

ion

rate

den

sity

[M

o/yr

/Mpc

3]

redshift

DustTrue SFR

@IR

@UV

Observed SFR

And because mm/submm is the best to unveil the dust obscured star formation in the early universe

Rapid dimming at optical/IR/radio wavelengthsfor higher z

Almost constantflux at mm/submm

K correction as a function of λ• Almost constant flux for

~1 < z < ~10

at 850um/1.1mm

due to strong negative K correction

• 1.4 GHz intensity drops very rapidly ： S(z) ~ (1+z)-(4+α)

– α ~ 0.8 (Condon 1992 ARAA, 30, 575)

redshift

Flu

x de

nsit

y [m

Jy]

Blain et al., 2002, Physics Reports, 369, 111-176

850um1.1mm

24um

Deep survey in optical wavelengths

Deep survey with ALMA (simulation)

AzTEC/ASTE 1.1mm map of ADF-S

1σ < 0.8 mJy (970 arcmin2)1σ < 0.6 mJy

0.5 deg

Hatsukade et al., 2010, poster #26

SNR

AzTEC/ASTE 1.1mm map of ADF-S

1σ < 0.6 mJy

0.5 deg

Hatsukade et al., 2010, poster #26

1σ < 0.8 mJy (970 arcmin2)

• 181 detections (above 3.5 σ)

• 101 of them exceed 5 σ

SNR

Constraints on redshifts of AzTEC/ASTE sources in ADF-S

• 90um/1.1mm flux ratio

most of the AzTEC sources: z > 1 – AKARI 90um sources : low-z, AzTEC 1.1mm sources: high-z

• L(FIR) ~ (3-14) x 1012 Lo, SFR ~ 500-2400 Mo/yr

90um/1.1mm flux ratio

AzTEC

AKARI

AzTEC-on-ASTE surveysof Submillimeter galaxies

(SMGs)

AzTEC Camera, 144 pix, 1.1mm, FOV=8’, θ=28”(Wilson et al. 2008)ASTE10m

(Ezawa et al. 2008)

Collaborators

• Kohno, K., Hatsukade, B., Ikarashi, S., Tsukagoshi, T., Inoue, H. (Univ. of Tokyo),

• Kawabe, R., Tamura, Y., Oshima, T, Nakanishi, K., Ezawa, H., (NAOJ), Komugi, S. (ISAS/JAXA), Tanaka, K. (Keio Univ.), & ASTE team

• Cortes, J., (JAO), Bronfman, L. (Univ. of Chile)

• Wilson, G.W., (PI. of AzTEC; UMASS), Aretxaga, I., Hughes, D.H., (INAOE), Yun, M.S., Austermann, J., Scott, K.S. (UMASS), Perera, T. (Univ. of Chicago), & AzTEC team

• ADF-S/SXDS/SDF/SSA22 collaborations

AzTEC-ASTE 1.1 mm deep surveys• wide (~1.6 deg2) & deep (1σ ~ 0.4 – 1.2 mJy ⇔ ULIRGs @z>1)

surveys of blank fields: yielding >750 robust detections

• Biased regions survey: ~1 deg2, >680 detections– High-z radio galaxies, X-ray and optically selected proto-clusters; ~160

arcmin2 x ~40 fields:

Field name Survey area (30-50% coverage)

Noise level (1 sigma) Num. of sources(S/N > 3.5)

SXDF 960 arcmin2 0.5 – 0.9 mJy 174

SSA22 810 arcmin2 0.6 – 1.2 mJy 113

ADF-S 970 arcmin2 0.4 – 0.8 mJy 191

SDF 210 arcmin2 0.7 – 1.0 mJy 25 SXDF z=5.7 clump 300 arcmin2 ~ 0.8 mJy 30

COSMOS 2700 arcmin2 ~ 1.1 mJy 193

GOODS-S 270 arcmin2 0.5 – 0.7 mJy 48

（ + HDF-S, ECDF-S ）

>1400 detections in total, ~x3 of the known (published) SMGs

Highlights from our SMG surveys

• Clustering properties of SMGs; relation to the underlying dark halo distribution

• Infrared-dark SMGs: extreme starburst populations that are invisible at optical/near-infrared wavelengths

• “Flat spectrum” SMGs: another candidates for high-z SMGs? (z>4)

• Extremely bright SMGs

Bright SMGs in SSA22 by AzTEC/ASTE

LAEs SMGs

SFR ~ a few Mo/yr

No one-by-one correspondenceto LAEs, but clustered at the LAE density peak

SFR ~ a few 100 – a few 1000 Mo/yr

Tamura et al., 2009,Nature, 459, 61

SUBARUASTELyman Alpha Emitters

Distribution of LAEsAround z~3.1Nakamura et al.

Clustering of SMGs toward the biased region traced by LAEs

1.1mm image

Bright SMGson Lyα emittersat z~3.1

Y. Tamura et al., 2009,Nature, 459, 61

Angular cross correlation between SMGs & LAEs– Landy & Szalay’s estimator– probability of finding a source of the other population at distance θ

• A clear detection of cross correlation signal with z=3.1 LAEs detected bright SMGs in SSA 22 are also clustered around z~3 !

correlation signal at θ < 5 ‘

Tamura et al., 2009, Nature, 459, 61

Tracing the heart of massive dark halo with SMGs

• Structure formation w/ Λ cold dark matter model– young, less massive galaxies (LAEs) are falling into

the heart of the massive dark halo– Massive starburst galaxies (SMGs) are grown there

LAEs

SMGs

LSS

Dark halo

Clustering of SMGs(auto correlation of SMGs)

• Previous studies on SMGs– Tentative detections

• amp. = 4.4±2.9

• amp. = 4. 9±2.8

• AzTEC Sources in ADF-S and SXDF– All sources

– >3mJy sources L(IR) ~ 1013 Lo

N source ADF-S SXDF Combine

All 174 191 365

>3mJy 64 70 134

(using 50 sources; Webb+03)

(using 126 sources; Weiss+09)

(Webb+03)

(Weiss+09)

Hatsukade 2010, PhD thesisSee also Poster #26

Distribution of all/bright SMGs in ADF-S

• Distribution of bright SMGs deviates from random distribution!

Bright sources only

Clustering of AzTEC SourcesField Source N source Amplitude

ADF-S all 174 0.82±3.0

>3mJy 64 14±5.8

SXDF all 191 0.50±2.0

>3mJy 70 3.1±5.6

Combine all 365 0.91±2.1

>3mJy 134 9.0±4.0

>3mJy sources in combined data

• Evidence for clustering – Bright sources (L(FIR) ~

1013) are more strongly clustered

– DH mass: 1013-14 Mo

• Field-to-field variance – ADF-S > SXDF

ω（

θ）

θ [arcsec]

Hatsukade 2010, PhD thesis; see poster #26

Optical/infrared properties of SMGs

• Optically faint/often invisible

• Very red color in infrared (NIR/Spitzer MIR bands also)

SXDF 850.6, SMA contours on optical/infrared/radio imagesHatsukade et al. 2010, ApJ, in press

Extremely dark SMGs

• The brightest guy in SSA22 is invisible/faint in near infrared down to KAB~25 mag !

despite of their extremely elevated SFRs (~a few 1000 Mo/yr), they are not identified by existing deep optical/NIR surveys with 4-8 m class telescopes !!!

• Recnet AzTEC-ASTE surveys as well as previous SCUBA surveys unveil the presence of such dark populations of massive starbursts

Multi-wavelengths ID of the brightest SMG in SSA22

• very “red” in MIR(IRAC) bands !

• K-drop out !? even with SUBARU

Color ： 2um Subaru/MOIRCS(Uchimoto et al., 2008) Contour ： SMA 860μm

Color ： Spitzer/IRACContour ： VLA 20cm

SMA 860μm

Ks > 24.9 mag(AB), 2σ upper limit

Y. Tamura et al. 2010, submitted

“K-drop SMG” with elevated SFR >1000 Mo/yr

• Detection of a deeply obscured (NH~1024cm-2) hard X-ray source proto-quasar phase? Growing SMBH?

• SED Mstar ~7x1010 Mo growing bulge?

StellarAGN

SpitzerIRAC

SUBARUMOIRCS

VLA

Nobeyama (NMA)

AzTEC/ASTE

SMA

ULIRG SED templatesat z = 3.1

MIPS

Y. Tamuraet al. 2010submittedIt is very surprisingthat this kind of monstercan’t be uncoveredwith existing NIR deepsurveys!

Another K-band drop SMGs in literature

• No K band counterpart even after deep integration using MOIRCS/SUBARU

Wang et al.2009, ApJ, 690, 319

“GOODS-N850.5”or “GN10”

Daddi et al., 2009, ApJ, 695, L176(z_spec=4.04)

Further example of K-drop SMGs• HDF850-1: Cowie et al., 2009, ApJ, 697, L122

SED of HDF850.1• z~4.1±0.5

is suggested

Cowie et al. 2009,ApJ, 697, L122

S850um

~7.8 ± 1.0 mJy

Submm/mm flux ratiosas a rough indicator of redshift

Steeper low zFlatter high z

High-z SMG candidates in SXDF• “Flat spectrum” SMGs: 850um/1100um flux ratio <

1.8 (i.e., flatter than a typical slope) – Combined analysis of SHADES and AzTEC-ASTE sources

– ~20 % of the total SCUBA sources are “flat” (<1.8)

– Many of them are 20cm faint

○: AzTEC○: SCUBA

AzTEC/ASTESN map

Kohno, Ikarashiet al. 2010, in prep.

High-z SMG candidates in GOODS-S• 870um; LABOCA on APEX: Weiss et al. 2009, ApJ, 707, 1201

• 1.1mm: AzTEC on ASTE: Scott et al., 2010, MNRAS, in press.

• Low ratio (flat spectrum) SMGs z>5 SMGs !?○AzTEC×Laboca

The highest redshift SMGs known• z=4.76, LESS J0332-2756, by LABOCA/APEX

– Coppin et al. 2009, MNRAS, 395, 1905

• z=4.54, COSMOSJ1000+0234 by AzTEC/JCMT– Capak et al. 2008, ApJ, 681, L53 for optical spec.– Schinnerer et al. 2008, ApJ, 689, L5 for CO spec.

• z=4.042, GOODS-N850.5, by SCUBA/JCMT– Daddi et al. 2009, ApJ, 695, L176

z=4.76 SMG:Optical (left)IRAC (middle)24 um + 870um contour (right)

Too many high-z SMGs already !?• Known number of

the high-z (z>4) SMGs are already consistent to a L-CDM model prediction !?

• Number of high-z (z>4) SMGs can put tight constraint on models

Redshift

Vol

ume

dens

ity

Φ [

Mpc

^-3]

z~2 SMGs(Chapman et al. 2005)

z>4 SMGs

Coppin et al. 2009, MNRAS, 395, 1905

Solid line:Model predictionby Baugh et al. 2005MNRAS, 356, 1191

Further evidence for the presence of high-z SMGs

• Infrared faint SMGs– Reports on SMGs without K-band counterpart– ~1/4 of the AzTEC/ASTE SMGs in SXDF seems K-

band faint, i.e., another high-z SMG candidates!

• Flat 850/1100um spectrum SMGs– Higher-z flatter spectrum (small flux ratio)

• Radio faint SMGs (see Younger et al. 2007)– High angular resolution submillimeter imaging with

SMA pinpoint the accurate position of SMGs

5 of 7 bright SMGs are radio faint (i.e., high z)

Brightest end of number count• Very wide area survey rare population

– Detection of 33 mJy@1100um source@SXDF

⇔ ~70 mJy (!) @850um• A dozen of >10 mJy sources

(⇔ >20 mJy source@850um)

– cf. the brightest SMGs in SHADES (@850um)• [22 mJy], 14 mJy, 13 mJy, 13 mJy, 11 mJy, …

• Majority: less than 10 mJy

• New observational constraints on:– Brightest end of galaxy number count– Extreme starburst (up to ~10000 Mo/yr ?!)

The brightest SMG in SXDF• ~80 mJy source @870 um

– Brighter than Cloverleaf quasar, comparable to recently reported LABOCA source (at z=2.3, Swinbank et al. 2010, Nature, in press)

– No evidence for strong amplification by a gravitational lens

– low Tdust, gas rich object, moderate SFR a young SMG? (if z~1)

Ikarashi et al., in prep.

A genuine, intrinsically brightextreme starburst? (if z~3)Maximum SB? (e.g., Thompson 2005)

Ultra-bright SMGs @low-z clusters??• Bullet cluster (z=0.297)

The brightest one:Close to the mass peak z~2 SMG amplified by lens?Collaboration with Weak lensing How about another Ultra-brights?

1 arcmin

Wilson et al., 2008, MNRAS, 390, 1061

Ezawa, Oshima, et al., in prep.

Millimeter/submillimeterStudies of GRB host galaxies

Molecular gas in GRB host galaxies

• A search for molecular gas via CO emission at mm/submm in GRB host galaxies: fairly tricky (or crazy) business because CO is difficult to detect in low metallicity environment.

• Nevertheless, we were motivated because there was evidence for a dust-obscured star formation in some of GRB host galaxies.– CO detection yields: Mgas, Mdyn, (CO abundance), – star formation efficiency (= SFR/Mgas)

cf. SSFR (= SFR/M*)

SFRs in GRBs: obscured star formation?

GRB SFR [Mo/yr]

Submm radio optical

GRB000210 560±165 90±45 3

GRB000418 690±195 330±75 15

GRB010222 610±100 300±115 1.5

SFRs from extinction corrected Halpha luminosity　　　　　　　　　　　　　　

SFRs from submm (SCUBA/JCMT) measurements

Berger et al. 2003, ApJ, 588, 99

huge discrepancy!

Overestimation @ submm/radio?or obscured star formation invisible in optical wavelengths?

Molecular gas ?

[OII]

[OII]

GRB000418: ULIRG like host?

• One of the most strongest SCUBA sources among GRB host galaxies (?)

• Radio detection

SFR @ submm 690±195 Mo/yrSFR @ radio 330±75 Mo/yrSFR @ optical 55 Mo/yr

Berger et al. 2003, ApJ, 588, 99

- z = 1.118- sub-luminous (L ～ 0.6 L* )-ULIRG like LFIR /L1600 ratio 1”=8.2kpc

CO(2-1) in GRB000418

• PdBI Observations, no detection after 2 hours integration• CO(2-1) luminosity 3σ upper limit: L’co(2-1) < 1.9 x 1010 K km/s pc2

• M(gas) < 2x1010Mo, SFR<90 Mo/yr if this is ULIRG-like – i.e., velocity width = 300 km/s, CO(2-1)/CO(1-0) ratio = 0.9, Xco=0.8 Mo/(K

km/s pc2) + KS law, also assumes CO emitting region size = optical diameter

Hatsukade et al. in prep.

PdBI

1 sigma = 4.2 mJy/beam

Inconsistent with SFR from submm, consistent with optical SFR

Another unsuccessful CO search: CO(1-0) in GRB030329 (z=0.17)

• No CO emission; 1 sigma ~2mJy/beam for dv=50 km/s

Endo, Kohno, et al. 2007, ApJ, 659, 1431

NMA

Summary on CO emission search in GRB hosts• GRB000418: search for CO(2-1) by PdBI

– M(H2) < 2×1010 M, SFR<90 M/yr

– Hatsukade et al., in prep.

• GRB030329: search for CO(1-0) by NMA– M(H2) < 2.8×1010 M, SFR < 4-38 M/yr

– Endo et al. 2007, ApJ, 659, 1431; Kohno et al. 2005, PASJ, 57, 147

• GRB980425: search for CO(3-2) by ASTE– M(H2) < 3×108 M, SFR<0.1 M/yr

– Hatsukade et al. 2007, PASJ, 59, 67

• No CO detection (in emission) in GRB hosts so far.

• Consistent with the idea that they are metal poor.

Is [CII] in GRB hosts promising ?

Maiolino et al., 2009, A&A, 500, L1

CII/FIR= 0.1 %

CII/FIR= 1 %

0.01 %

LMC

M82

Arp220

• [CII] 157 um/1.9 THz line becomes very bright (i.e., high L[CII]/LFIR ratio) in low metal galaxies– CO is selectively

destructed in these low metal star forming galaxies

GRBs in ALMA era

• Huge gain of ALMA sensitivity enables us to detect emission lines from very high-z GRBs

• Primordial molecules like HD can also be detected at z>10?

Inoue et al., 2007,MNRAS, 380, 1715

Summary• Mm/submm surveys are essential to unveil dust-

obscured star formation in the early universe.• AzTEC-on-ASTE surveys: >2 deg2, with a sensitivity of

0.4-1.2 mJy (1σ), sufficient for detection of z>1 ULIRGs, yielding >1400 detections of SMGs.– Clustering in bright SMGs dark halo mass ~1013-14 Mo

– SMGs at z>4 possibly too many already? Further evidence for high-z SMGs by radio/K-band faintness and submm/mm flatness

• GRB host galaxies: no CO emission so far, consistent with metal poor nature of them (i.e., too faint to see). ALAM will drastically improve the situation.

millimeter/submillimeter studies of high-redshift galaxies deciphering the ancient universe with...

Documents