Transcript Present

Millimeter detection of Spitzer selected
hyperluminous infrared starburst galaxies
Alain Omont (IAP)
The SWIRE Spitzer survey has detected by far the largest set of Sub-Millimeter Galaxies (SMG)
But their identification is difficult and requires validation with (sub-)millimeter observations
With MAMBO/IRAM we detected at 1.2mm 20 SWIRE z~2 starburst HLIRGs with LFIR~1013Lo
They belong to a rather rare subclass of SMGs, with large PAH/FIR ratio and large stellar mass
Millimeter detection of Spitzer selected hyperluminous infrared starburst galaxies
OUTLINE
• General goals
• Background: High z submillimeter galaxies (SMGs)
• Summary of millimeter dust detection in SMG hosts of high z QSOs
• Identification of hyperluminous SMGs in SWIRE Spitzer survey
The SWIRE survey
Detection of SMGs in SWIRE
1.2mm Mambo observations of SWIRE HLIRGs
Properties of starburst SWIRE HLIGRs  a subclass of SMGs rich in PAH and stars
Search for AGN SWIRE HLIGRs
Prospects
• Other projects in progress with CFHTLS & Spitzer data and IRAM observations
z
Dphot
(Gpc)
1000
--------------------
20
12
230
z ~ 7 – 20 ?
Main z ranges in the Cosmic History of galaxies - Reionization
PopIII stars +1st
galaxies
-Formation of 1st
galaxies
Pop. II stars
~ 300 million
- First AGN
z ~ 4–7:
Current frontier
- Galaxy and Black-Hole
early assembly
- End of reionization
130
z ~ 1.5 -4:
- Peak of star formation
submm sources + LBGs
-Peak of QSO activity
--------------------
z= 6 60
--------------------
z=2
16
-------------------
0.5
3
--------------------
0
~ 3.5 billion
z ~ 0.5-1.5 :
Final phase of active SF
- Weak X-ray AGN
- Cluster formation
High-z Galaxies, Star Formation and AGN
• Much progress in the last 5-10 years about the exploration of the high-z Universe
High-z galaxies: HDF, Lyman-Break Galaxies, Lyman-a Galaxies, etc.
z ~ 2–3  6–7
High-z QSOs and radiogalaxies: peak at z ~2 – 3  6.4
SCUBA/MAMBO submillimeter galaxies (SMGs): peak at z ~2 – 3
• Good view of history of star formation in the Universe:
- First starbursts at z > 7
-Early star formation in early type (elliptical) galaxies  major mergers/starbursts
- Late type (spiral) galaxies formed most of their stars at z < 1
• Strong connection between AGN black-hole growth and galaxies (bulge/spheroid)
and their starbursts
SMGs: strongest starbursts in the Universe
Important players of star formation at z >~ 2
 Giant starbursts at the peak of elliptical formation z ~ 2-3  1-4
 Most of the energy in the far-infrared (FIR) dust at ~40K
 At least Ultra-Luminous Infra-Red Galaxies (ULIRGs: Sanders & Mirabel 1996):
• LFIR >~ 1012 Lo, star formation rate SFR > 100 Mo/yr
• Relatively rare, but ~1 per arcmin2
• Generally not isolated; strongly biased along high-z Large Scale Structures
• Probably progenitors of massive elliptical galaxies
• A few of these objects are powerful QSOs or radiogalaxies  MBH >~ 108 Mo
Spectral Energy Distribution (SED) of ULIRGs/SMGs
FIR
1.2mm obs
lobs=lrest(1+z)
PAH
Starbursts
SiO
AGN
AGN+starburst
SMGs: strongest starbursts in the Universe
 Giant starbursts at the peak of elliptical formation z ~ 2-3  1-4
 At least Ultra-Luminous Infra-Red Galaxies (ULIRGs):
• LFIR >~ 1012 Lo, SFR > 100 Mo/yr
• Relatively rare, but ~1 per arcmin2
• Generally not isolated; strongly biased along high-z Large Scale Structures
• Probably progenitors of massive elliptical galaxies
• A few of these objects are powerful QSOs or radiogalaxies  MBH >~ 108 Mo
 Most exceptional Hyper-Luminous IR Galaxies (HLIRGs):
• LFIR >~ 1013 Lo, SFR ~ 1000 Mo/yr
• Nothing equivalent in the local Universe
• Very rare ~1 per 50-100 arcmin2
• CO already detectable
• Probably in most massive DM halos progenitors of central cD galaxies of clusters
• A fraction of them are very powerful QSOs or radiogalaxies  MBH >~ 108 Mo
Dust detection: The Magic of the high-z submm window
FIR emission of cold dust
(Td ~ 35-50 K) :
- steep submm spectrum
- compensates for distance
- Sn practically independent
of z from z ~ 0.5 to 10
Redshift degeneracy
Effect also known as
« negative K correction »
LFIR = 4x1012 S250(mJy) Lsun
from Bertoldi, Voss, Walter
Dust detection: the magic of the submm window
SCUBA(-MAMBO) census of high-z ULIRGs
• Take advantage of steep submm spectrum
• Account for good part (most) of submm background
• z at Keck for radio ones (~50%) (weak AGN ?)
 History of star formation up to z~3-4
• Small but uncertain number at z > 4
SCUBA (+MAMBO) submm counts
SCUBA(-radio) redshift distribution
Chapman, Blain, Smail, Ivison 2005
Submm counts
from the SHADES survey
Coppin et al. 2006
High-z millimeter studies at IRAM
1. MAMBO detection of redshifted far-IR/submm dust emission from high-z QSOs
A. Beelen
F. Bertoldi
C. Carilli
P. Cox
+ J. Bergeron, K. Isaak, R. McMahon, R. Priddey, etc
+ X.H. Fan, M. Strauss, Ran Wang, etc.
( 2. CO detections with the Plateau de Bure Interferometer )
Omont et al. 1996
Carilli et al. 2001
Omont et al 2001
Omont et al 2003
Bertoldi etal. 2003
Beelen et al. 2007
Wang et al. 2007
MAMBO/IRAM detection of redshifted far-IR/submm dust (and CO) emission
from high-z QSOs
• To establish correlations between major starbursts and black-holes at high z
• It is the easiest way to find (biased) cases of HLIRGs at very high z, since the
redshifts of SCUBA/MAMBOsources are practically unknown at z > 4
• IRAM 30m Telescope ( Spain) + MAMBO bolometer cameras (Bonn)
• 200 bright QSOs
Lbol ~ 1013 -1014 Lo
• LFIR ~ 1013 Lo  HLIRGs
SFR ~ 1000 Mo/yr (if pure starburst)
• Concommittant major starbursts
and AGN. But the far-IR luminosity is
weakly correlated with Lbol
• CO has been detected in 18 high-z QSOs
with the IRAM interferometer
with z~2-6 : 55 detections
High-z millimeter studies at IRAM. 3.
Millimeter detection of Spitzer selected
hyperluminous infrared starburst galaxies
• The SWIRE survey
• Detection of SMGs in SWIRE and identification strategy
• 1.2mm Mambo observations of SWIRE HLIRGS
• Properties of starburst SWIRE HLIGRS  a subclass of SMGs rich in
PAH and stars
• Prospects
Main collaborators: Carol Lonsdale, Mari Polletta (UCSD)
+ D. Farrah, D. Shupe, R. Zylka, S. Berta, B. Siana, N. Bavouzet, G.Lagache,
H. E. Smith, F. Bertoldi., P. Cox, C. De Breuck, H. Dole, D. Lutz, L. Tacconi,
I. Perez-Fournon, H. Aussel, H. McCracken, D. Clements, M. Rowan-Robinson
A. Franceschini, D. Frayer
Importance of extensive studies of larger samples of high-z SMGs
• Major starbursts of massive (elliptical) galaxies: initial collapse or major mergers
• Parallel growth of super-massive black holes
• Trace density peaks, first massive dark matter halos and
early Large Scale Structures
Importance of larger samples of high-z SMGs
• Various statistical studies
• Trace their spatial distribution along early LSS
• Identify the rare most extraordinary, mm-brightest ones
 strong lensed cases
or most luminous (far-IR) galaxies LFIR > 1013 Lo, HLIRGs
Present SMG samples are very limited
Only ~400 sources provided by SCUBA/MAMBO surveys( <~ 0.5 deg2)
Waiting for very large samples/areas (~>2007) of SCUBA2 (850µm),
ASTRO-F/AKARI (50-200 µm) and (~>2008) Herschel (100-500 µm)
Much larger (x~20-50) samples already exist (buried) in Spitzer wide field surveys
in particular SWIRE: 50 deg2, 10000 - 15 000 SMG
Spitzer provides the first wide field (mid-)IR surveys
Spitzer is:
• Very well matched to the detection of red galaxies (elliptical, LIRGs…) up to z~1
• As well as AGN (Type 1 and 2 QSOs, etc.) up to z~3
• Even pretty sensitive for high-z SMGs although most energy is in FIR outside
of Spitzer sensitive l range
Wide-field Spitzer surveys (FLS, GTO, SWIRE) cover in total about 70 deg2:
SWIRE Legacy Project: 50 deg2
(PI Carol Lonsdale)
µm
3.5
4.5
5.8
8.0
24
µJy
3.7
10
4.5
10
48
45
38
50
106
250
( ? ~100 mJy at 160 µm, ~20 mJy at 70 µm)
nominal limits Lonsdale et al. 2004:
present catalogs
Identification of SMGs in SWIRE
Direct Spitzer detections of ~90 known SCUBA/MAMBO SMGs at z~2
with SWIRE sensitivities, HLIRGs with S(850µm) > 7 mJy or S(1.2mm) > 2.5 mJy:
are detected at >90% at 3.8 and 4.5 µm
and at ~45% in most IRAC (3-8µm) bands and MIPS 24µm
 150-300 per deg² (0.05-0.1 per arcmin2)
 0.7-1.5 104 in total SWIRE
 ~20-40 times more than all existing
MAMBO-SCUBA surveys
HLIRGs
Strategy for Identification of SMGs in SWIRE
• Careful identification criteria of HLIRG candidates from SWIRE+ optical
data + SMG SED templates
• Difficult extrapolation from 24µm to 1.2mm (need to simultaneously
determine the SED and redshift from five SWIRE bands (3.6-24µm)
Spectral Energy Distribution (SED) of ULIRGs/SMGs
Starbursts
AGN
AGN+starburst
SWIRE
detected with SCUBA or MAMBO
SCUBA-MAMBO
Spitzer infrared SED of
Strategy for Identification of SMGs in SWIRE
• Careful identification criteria of HLIRG candidates from SWIRE+ optical
data + SMG SED templates
• Difficult extrapolation from 24µm to 1.2mm (need to simultaneously
determine the SED and redshift from five SWIRE bands (3.6-24µm)
Four successive MAMBO projects scheduled at IRAM-30m
(Fall 2005 – Winter 2007) for detecting SWIRE HLIRGs candidates at 1.2mm
(S1.2mm >~2-5 mJy):
- 2005-2006: 24µm-bright starburst best candidates: 20 detections
- November 2006: AGN best candidates
- Winter 2007: 24µm-bright starbursts, unbiased sample
1.2mm observations of 24µm-bright starburst SWIRE HLIRGS
Observation of a sample of ~60 best candidates with:
- starburst-dominated SED
- S(24µm) >~ 400 µJy
- maximum in 5.8 µm
IRAC band
(redshifted 1.6 µm
stellar bump  z ~ 2)
1.2mm observations of 24µm-bright starburst SWIRE HLIRGS
Observation of a sample of ~60 best candidates with:
- pure starburst SED
- S(24µm) >~ 400 µJy
- maximum in 5.8 µm IRAC band (redshifted 1.6µm stellar bump  z ~ 2)
One third are detected with S(1.2mm) >~ 2 mJy
Average value < S(1.2mm) > = 1.6 mJy
Majority of the sources are ~HLIRGs with LFIR close to 1013 Lo
Follow-up at 350µm at CSO in January 2007  Tdust , accurate LFIR
Properties of 24µm-bright starburst SWIRE HLIRGS
Mid-IR spectra obtained with Spitzer IRS for 5 sources detected at 1.2mm
 practically pure PAH spectra
 confirmation of redshift ~ 2 (1.7 – 1.9)
Observed spectrum
Properties of 24µm-bright starburst SWIRE HLIRGS
Mid-IR spectra obtained with Spitzer IRS for 5 sources detected at 1.2mm
 practically pure PAH spectra
 confirmation of redshift ~ 2 (1.7 – 1.9)
Average spectrum
9 starburst sources
Weedman et al. 2006
Properties of 24µm-bright starburst SWIRE HLIRGS
Redshifts in the range 1.7 – 2.5 ( z-spectro 1.7 – 1.9 )
Ratio S(1.2mm)/S(24µm) much smaller than most SMGs whose typical SED
is relatively cold, similar to Arp 220
Their SED is rather
similar to M82 or
NGC6090
Properties of 24µm-bright starburst SWIRE HLIRGS
1. Large ratio PAH/FIR emission
Ratio S(1.2mm)/S(24µm) much smaller than most SMGs whose typical SED
is relatively cold, similar to Arp 220
Their SED is rather similar to M 82 or NGC 6090
Strong ratio of PAH/FIR emission

Maybe they are more extended starbursts
 less opacity and less absorption of
PAH emission
With perhaps strong superwind as in
M82 ?
But scaled by a factor > 10
 enrichment of intergalactic
medium in heavy elements??
Extended PAH emission in the super wind of M82
Engelbracht et al. 2006
Properties of 24µm-bright starburst SWIRE HLIRGS
2. Large stellar masses
Masses in red stars range from 8x1010 to 6x1011 Mo, if the mean stellar age is
below 109 years, and significantly higher masses if the stars are older.
 They are probably young massive ellipticals in the final stages of star
formation, where much of their stellar mass is already in place, yet very
high rates of star formation are still proceeding.
The rare starburst episode is likely to be very short lived, to avoid that the total
accumulated stellar mass by the end of the burst exceed that observed in the
largest local ellipticals.
 Probably one of the latest gas-rich major mergers of a massive elliptical
MAMBO observations of SWIRE HLIRG AGN candidates
October-November 2006: bad weather; 1/3 of proposal observed
XMM-LSS and Lockman SWIRE fields
• 14 Sources fitted with composite starburst-AGN SEDs
- average flux: 1.53 +/-0.40 mJy
- the majority should be HLIRGs
• 11 sources with « Torus » featureless SEDs
- average flux: 0.46 +/-0.38 mJy
- few HLIRGs
Spectral Energy Distribution (SED) of ULIRGs/SMGs
Starbursts
AGN
AGN+starburst
MAMBO observations of SWIRE HLIRG AGN candidates
October-November 2006: bad weather; 1/3 of proposal observed
XMM-LSS and Lockman SWIRE fields
• 14 Sources fitted with composite starburst-AGN SEDs
- average flux: 1.53 +/-0.40 mJy
- the majority should be HLIRGs
• 11 sources with « Torus » featureless SEDs
- average flux: 0.46 +/-0.38 mJy
- few HLIRGs
• Confirmed detection of two exceptional sources
- u-dropouts in D1 CFHTLS Deep Field:  z ~3 – 4
- probably red QSOs
- S24µm ~ 2–3 mJy
- S1.2mm ~ 5 mJy
- Nature of IR emission still unknown
Confirmed detection of two exceptional sources
- u-dropouts in D1 CFHTLS Deep Field:  z ~3 – 4 (to be confirmed)
- probably red QSOs
- S24µm ~ 2–3 mJy
- S1.2mm ~ 5 mJy
- Nature of IR emission still unknown
- Very large ratios 24/8.0µm and 8.0/5.8µm
Band
u g
r
i
z
3.6 4.5
5.8 8.0 24um 1.2mm 20cm
-----------------------------------------------------------------------------------------------------------------------SW49208 z ~ 3 – 3.5
m_AB
<26 24.6 23.4 23.0 22.5
20.9 20.7
Flux(µJy)
3.6 15.8 19.7 <50 185 2406 ~5000 350
-----------------------------------------------------------------------------------------------------------------------SW55275 z ~ 3.5 - 4
m_AB
Flux(µJy)
25.5 23.0 21.7 22.0 22.4
20.9 20.5
4.0
15.5 23.0
<50 245 3271
~5000 140
------------------------------------------------------------------------------------
PROSPECTS
(2005-2006: 24µm-bright starbursts, best candidates: 20 detections
November 2006: AGN best candidates)
- Winter 2007: 24µm-bright starbursts, unbiased sample
Goal to characterize this subclass of strong starbursts with 5.8µm bump and
S(24µm) > 400µJy (50 per sq. deg.) and its average star formation rate.
Previous samples were biased towards the identification of the most luminous
objects, so that the average starburst properties of the bulk of this class remain
undetermined.
We have been awarded 39h of MAMBO observations to observe an unbiased
complete random sample, which has excellent X-ray and optical coverage,
to determine the average value of S(1.2mm) and inferred quantities
PROSPECTS
• (New 1.2mm Mambo observations:
- Winter 2007: 24µm-bright starbursts, unbiased sample)
• Follow-up at 350µm at CSO in January 2007  Tdust , accurate LFIR

• Observations in SWIRE southern fields with APEX/LABOCA
• CO search with IRAM interferometer when accurate redshifts
• Morphology studies:
- HST?
- Look for relatively nearby similar objects; at least ULIRGs

• Full identification of SWIRE z~2 HLIRGs (+ ULIRGs) warranted with Scuba 2
(2007) and Herschel (2008)
• Detailed studies with ALMA (2012)
Other projects in progress
with CFHTLS & Spitzer data and IRAM observations
• Identification of Type 1 and 2 QSOs in CFHTLS/Spitzer fields
Pei Yu, J. Bergeron, X. Liu; N. Bavouzet
• 1.2mm MAMBO maps of deep Spitzer + multi-l fields
A. Baker, C. Lonsdale, F. Owen; F. Bertoldi
• 1.2mm observations of high z QSOs and IRS Spitzer sources
M. Jarvis; C. Willott; Ran Wang; D. Lutz, L. Yan
• CO observations of high z sources
I. Smail, P. Cox; D. Lutz, L. Yan; Ran Wang; F. Bertoldi; R. Genzel
CO detection in 18 high-z QSOs
Plateau de Bure IRAM interferometer
CO detection in
18 high-z QSOs
CO detection in QSO SDSS J1148 at z=6.42
Dense gas
density
~105 cm-3
T ~ 100 K
size
R ~400-1400 pc
gaz mass 1-2 1010 Mo
Multi-line excitation model
Bertoldi et al. 03 PdBI
Resolution of CO in 2 sources VLA Walter et al. 04
Walter et al. 03 VLA
Dust detection: the magic of the submm window
SCUBA(-MAMBO) census of high-z ULIRGs
• Take advantage of steep submm spectrum
• Account for good part (most) of submm background
• z at Keck for radio ones (~50%) (weak AGN ?)
 History of star formation up to z~3-4
• Small but uncertain number at z > 4
SCUBA (+MAMBO) submm counts
SCUBA(-radio) redshift distribution
Chapman, Blain, Ivison, Smail 2003