der Paul van der Werf Sterrewacht Leiden Starburst galaxies at low and high redshift ESO, Santiago April 2006

Starburst galaxies at low and high redshift2 Credits  starbursts at low redshift  Leonie Snijders (VISIR)  Liesbeth Vermaas (SINFONI)  Kate Isaak (Cardiff)  Padelis Papadopoulos (ETH Zürich)  starbursts at high redshift  Kirsten Kraiberg Knudsen (PhD October 2004)  Tracy Webb (McGill University, Montreal)  Lottie van Starkenburg (SINFONI) and the FIRES team:  Marijn Franx  Pieter van Dokkum (Yale)  Ivo Labbé (Carnegie)  Greg Rudnick (NOAO)  Hans-Walter Rix (MPIA)  Mariska Kriek  Natascha Förster Schreiber (MPE)  Alan Moorwood (ESO)

Starburst galaxies at low and high redshift3 « The stellar systems are scattered through space as far as telescopes can penetrate. We find them smaller and fainter, in constantly increasing numbers, and we know that we are reaching out into space, until, with the faintest nebulae than can be detected with greatest telescopes, we arrive at the frontiers of the known Universe. » Edwin Hubble, The Realm of the Nebulae (1936)

Starburst galaxies at low and high redshift4 What is a starburst galaxy? correction factor for mass return from stars A starburst galaxy is a galaxy with such a high star formation rate that it will turn all of its gas into stars in t b << t Hubble Milky Way:  gradual buildup of disk

Starburst galaxies at low and high redshift5 A luminous infrared galaxy: the Antennae  Crossing orbits give gas concentration in gas concentration in interaction zone interaction zone  Intense obscured star formation in star formation in overlap region overlap region  Most intense star formation is obscured formation is obscured (Webb, Snijders & Van der Werf, in preparation) SCUBA 850  m

Starburst galaxies at low and high redshift6 Arp220 An ultraluminous infrared galaxy: Arp220

Starburst galaxies at low and high redshift7 Starformation in ULIRGs correction factor for mass return from stars Arp220: NB: t b  few 10 8 yrs  merger timescale  dynamical timescale  dynamical timescale (crossing/rotation time) (crossing/rotation time)  free-fall time protogalaxy?  free-fall time protogalaxy? Such high SFRs can build up an entire galaxy in t << t Hubble  relation to galaxy formation?

Starburst galaxies at low and high redshift8 Simple-minded estimate of "maximum star formation rate" initial starburst – rapid formation of bulk of the stellar population In the absence of external pressure, the maximum star formation rate occurs when a gas mass is turned into stars on the free-fall timescale.  formation of spheroids?  Implication: high SFRs are found in the most massive galaxies

Starburst galaxies at low and high redshift9 Do “maximum starburst” galaxies exist? maximum starbursts normal galaxies luminous IR galaxies (LIRGs) (Kennicutt 1998) ultraluminous IR galaxies (ULIRGs)

Starburst galaxies at low and high redshift10 ULIRGs as “maximum starbursts” (cf., 30–50% for normal spirals) ULIRGs

Starburst galaxies at low and high redshift11 Starformation efficiency  Starbursts cannot be simply scaled up. be simply scaled up.  More intense starbursts are also more efficient are also more efficient with their fuel. with their fuel. (Gao & Solomon 2001) L IR  SFR L IR / L CO  SFR/ M H 2 SFE

Starburst galaxies at low and high redshift12 Role of dense gas  More dense gas means more efficient star formation. (Gao & Solomon 2001) L HCN / L CO  mass fraction of dense gas L IR / L CO  SFR/ M H 2 SFE

Starburst galaxies at low and high redshift13 NGC 4038/4039 Orion (M42) 30 Doradus NGC 4038/4039 detail Superstarclusters: does size matter? NGC4038/4039 cluster:  100 pc  100 pc Orion:  1.5 pc

Starburst galaxies at low and high redshift14 Superstarclusters: densities Objectd[pc] MM[M][M]MM[M][M]   [ M  pc —3 ] in 4.5pc n H [cm —3 ] in 4.5pc [W99]2100 2∙ ∙10 5 R ∙ M

Starburst galaxies at low and high redshift15 The Antennae with Spitzer and VISIR IRAC/Spitzer (Wang et al., 2004) [NeII] 12.8  m ESO/VLT VISIR (Snijders et al., in prep.) contours: dust continuum

Starburst galaxies at low and high redshift16 Probing dense molecular gas Line T ex [K] n crit [cm —3 ] CO 1  ∙10 3 CO 4  ∙10 4 CO 6  ∙10 5 HCN 1  ∙10 5 HCN 4  ∙10 6 molecular gas dense molecular gas temperature of dense gas

Starburst galaxies at low and high redshift17 Dense gas in the ULIRG Mrk231 RxW/JCMT (Papadopoulos, Isaak & Van der Werf 2006) (Papadopoulos & Van der Werf 2001) Mrk231 CO 6  5 and 4  3

Starburst galaxies at low and high redshift18 Again: dense gas in Mrk231 RxB3/JCMT (Papadopoulos, Isaak & Van der Werf 2006) Mrk231 CO 3  2 and 10  HCN 4  3

Starburst galaxies at low and high redshift19  Combine these data with CO 1  0, 2  1 and 13 CO 2  1 (upper limit): no model reproduces all CO lines: no model reproduces all CO lines: CO 4  3/6  5 and CO 1  0/2  1/3  2 probe different gas phases. CO 4  3/6  5 and CO 1  0/2  1/3  2 probe different gas phases. diffuse phase: T  50—85 K and n  300—10 3 cm —3, diffuse phase: T  50—85 K and n  300—10 3 cm —3, dense phase: T  50—65 K and n  10 4 cm —3. dense phase: T  50—65 K and n  10 4 cm —3.  Total mass of molecular gas from CO: M  3.5∙10 10 M  (or up to a factor 4—5 lower). M  3.5∙10 10 M  (or up to a factor 4—5 lower). Mass of dense molecular gas from HCN: Mass of dense molecular gas from HCN: M  1.5—3.5∙10 10 M . M  1.5—3.5∙10 10 M . Almost all molecular gas in Mrk231 is dense. Almost all molecular gas in Mrk231 is dense. Two phases of molecular gas

Starburst galaxies at low and high redshift20 ULIRGs and galaxy formation  Locally, ULIRGs are energetically unimportant: contribute only 2% of local bolometric energy density contribute only 2% of local bolometric energy density How important are ULIRGs at high z? How important are ULIRGs at high z?  observe redshifted far-IR emission in submillimetre  observe redshifted far-IR emission in submillimetre with JCMT/SCUBA with JCMT/SCUBA  Local ULIRGs almost always have a hidden AGN: causal link with the extreme starburst? causal link with the extreme starburst? Are high- z ULIRGs the sites of the coeval formation of spheroids and Are high- z ULIRGs the sites of the coeval formation of spheroids and supermassive black holes? supermassive black holes?  question for the coming decade  question for the coming decade

Starburst galaxies at low and high redshift21 Microwave Background IR/Optical Background X-Ray Background Big Bang Stars+BlackHoles AGN 1mm 1  m 1nm (Puget et al. 1996, Hauser et al Fixsen et al. 1998) Cosmic energy density

Starburst galaxies at low and high redshift22 Is obscured star formation important?  Direct starlight:  Far-IR background: I = erg / s / cm 2 / sr I = erg / s / cm 2 / sr I = erg / s / cm 2 / sr I = erg / s / cm 2 / sr Obscured star formation dominates.

Starburst galaxies at low and high redshift23 Submillimetre cosmology  Submm samples have high proportion of high- z galaxies  Galaxies up to z  10 detectable  Brightness-limited  volume-limited  Luminosities without precise redshifts  Sources do not fade much from z =1 to 10  Deeper surveys probe fainter galaxies, not higher z Large negative k -correction :

Starburst galaxies at low and high redshift24 Counts and backgrounds Submm background comes from small number of (ultra)luminous galaxies Resolved background at 850  m:  25% down to 2 mJy Connection to other populations at fainter flux levels. Lensing can probe these.

Starburst galaxies at low and high redshift25 SCUBA/JCMT lensed galaxy survey Knudsen PhD thesis Van der Werf  12 gravitationally lensing clusters  1 blank field (NTT Deep Field)  Survey area 72 arcmin 2  Deepest fields are confusion- limited (2 mJy)  55 sources detected at 850  m

Starburst galaxies at low and high redshift26 Faint submillimetre source counts  Counts imply strong evolution  First time the faint submm population is substantially probed population is substantially probed  Connection with optical population  Turnover in counts detected for the first time first time (Knudsen, Van der Werf & Kneib, in prep.)

Starburst galaxies at low and high redshift27 Submillimetre source counts and evolution Counts + background imply pure luminosity evolution proportional to (1+ z ) 3 luminosity evolution: L   (1+ z ) p L   (1+ z ) p density evolution: N  (1+ z ) q N  (1+ z ) q

Starburst galaxies at low and high redshift28 Submm galaxies in the NTT Deep Field Not an ERO (?) ERO (Knudsen et al. in preparation)

Starburst galaxies at low and high redshift29 FIRES: Faint Infrared Extragalactic Survey ISAAC/Antu HDF-S: Labbé et al., 2003 MS1054—03: Förster Schreiber et al., 2006

Starburst galaxies at low and high redshift30

Starburst galaxies at low and high redshift31 FIRES allows better mass selection  FIRES allows selection of high- z galaxies in the rest-frame V -band  much closer to a mass-selection than rest-frame UV  a new class of high- z galaxies: Distant Red Galaxies (DRGs) I 814 KsKs the 6 most massive z  3 galaxies in the HDF-S

Starburst galaxies at low and high redshift32 Rest-frame SEDs: Lyman break galaxies

Starburst galaxies at low and high redshift33 Rest-frame SEDs: distant red galaxies (Förster Schreiber et al., 2004)

Starburst galaxies at low and high redshift34 Selection of distant red galaxies (Franx et al., 2003)

Starburst galaxies at low and high redshift35 DRGs are (mostly) star forming galaxies star formation rates > 100 M  yr —1 (Van Dokkum et al., 2004)

Starburst galaxies at low and high redshift36 SCUBA observations of MS1054–03 M1383 S 850 = 5 mJy

Starburst galaxies at low and high redshift37 Submm galaxies and DRGs  Lyman Break Galaxies have no detectable submm emission even after massive source stacking massive source stacking  Distant Red Galaxies are massive and have high SFRs  DRGs: 3 arcmin –2 for K <22.5 same source density as submm galaxies with >0.8 mJy at 850  m same source density as submm galaxies with >0.8 mJy at 850  m  1 FIRES field observed with SCUBA: 1 DRG detected directly (5 mJy)  connection between sub-mJy submm galaxies and DRGs?

Starburst galaxies at low and high redshift38 Stacking DRGs DRGs, EROs random positions DRGs: S 850 = 1.11  0.28 mJy (Knudsen et al., 2005)

Starburst galaxies at low and high redshift39 Breaking the confusion barrier: gravitational lenses A2218

Starburst galaxies at low and high redshift40 A2218: SCUBA 850  m

Starburst galaxies at low and high redshift41 The submm triple behind A2218 Triple image, z =2.515 Magnification in total factor 40 Intrinsic 850  m flux: 0.8 mJy Redshifted H  : strong star formation strong star formation J–K =2.7: Distant Red Galaxy (Kneib et al., 2004a)

Starburst galaxies at low and high redshift42 CO 3–2 detection of the submm triple Detected both at Owens Valley and Plateau de Bure CO line width  less massive than bright SMGs but comparable to DRGs (Sheth et al., 2004; Kneib et al., 2005b)

Starburst galaxies at low and high redshift43 Conclusions and outlook: low z  Gas density is crucial in determining starburst determining starburst properties; also related to properties; also related to depth of potential well? depth of potential well?  Luminous starbursts tend to have both high L / M and high M have both high L / M and high M  Molecular lines probe density/temperature structure density/temperature structure  The future: SINFONI+LGS, APEX, HIFI/Herschel, ALMA APEX, HIFI/Herschel, ALMA

Starburst galaxies at low and high redshift44 Conclusions and outlook: high z  Starbursts are a key process in galaxy evolution  We are looking back towards a strongly obscured universe, energetically dominated by dusty starbursts energetically dominated by dusty starbursts  DRGs are massive and have high SFRs; as a population they are at least as important as LBGs for cosmic mass budget and SFR least as important as LBGs for cosmic mass budget and SFR  Connection between DRGs and faint submm galaxies  The future: SCUBA2, ALMA, HAWK-I, JWST, ELT

Starburst galaxies at low and high redshift45 To coldly go…: SCUBA-2  SCUBA-2 will bring rectangular array imaging to submm astronomy imaging to submm astronomy  8 40  32 sub-arrays – 4 at 850  m, 4 at 450  m 4 at 450  m  Mapping speed unequalled  Operational mid-2007  Legacy programs defined:  Cosmological survey  Full Galactic plane survey  Full Gould’s Belt survey  Shallow “all-sky” survey  Unbiased debris disk survey

Starburst galaxies at low and high redshift46 SCUBA-2 Cosmology Legacy Survey  4 PIs: Dunlop, Halpern, Smail, Van der Werf Van der Werf  Large survey (20 deg 2 ) at 850  m, deep survey at 450  m (0.5 deg 2 ) deep survey at 450  m (0.5 deg 2 )  450  m survey will resolve full submm background and detect all submm background and detect all LIRGs in the survey area out to z=2 LIRGs in the survey area out to z=2 (ULIRGs out to z=4) (ULIRGs out to z=4)  Connection with other populations: optical, near-IR, X-ray, radio,… optical, near-IR, X-ray, radio,…  Time allocation 102 good nights (including 50% of the best weather (including 50% of the best weather until mid-2009) until mid-2009) SCUBA-2 field-of-view 1 deg 2 field, will be mapped to confusion limit in 23 hours by SCUBA-2 at 850  m

Starburst galaxies at low and high redshift47 « We are, by definition, in the very center of the observable region. We know our immediate neighborhood rather intimately. With increasing distance, our knowledge fades, and fades rapidly. Eventually, we reach the dim boundary, the utmost limits of our Telescopes. There, we measure shadows, and we search among ghostly errors of measurement for landmarks that are scarcely more substantial. The search will continue. Not until the empirical resources are exhausted, need we pass on the dreamy realms of speculation. » Edwin Hubble, The Realm of the Nebulae (1936)