High-mass star formation Paola Caselli (Leeds), Francesco Fontani (IRAM), Izaskun Jimenez-Serra (CfA), Mark Krumholz (UCSC), Christopher McKee (UCB), Francesco Palla (Arcetri), Jan Staff (LSU), Leonardo Testi (ESO), Barbara Whitney (SSI) Michael Butler, Audra Hernandez, Bo Ma, Yichen Zhang Sven Van Loo, Peter Barnes, Elizabeth Lada Charlie Telesco Orion Nebula Cluster (VLT; JHK) (McCaughrean) IR Dark Cloud Ext. Map G28.37 (Spitzer/GLIMPSE) (Butler & Tan) Jonathan Tan (University of Florida)

Why work on massive star formation? Galaxies form and evolve by forming star clusters, where the influence of massive stars is paramount. Massive stars are what tend to be seen in distant galaxies. Planets form from the crumbs left over from star formation. Planet & star formation in star clusters can be influenced by massive star feedback. Supermassive black hole formation may be via massive star clusters or Pop III stars. Supermassive black hole accretion is likely to be regulated by star formation. The First (Pop III) Stars were likely massive, some potentially supermassive stars, reionizing the universe and producing the first metals.

Why not to work on massive star formation... Wide range of scales (~12 dex in space, time) and multidimensional. Uncertain/unconstrained initial conditions/boundary conditions. Complete theory of star formation Numerical models Analytic theory Observations Physics: Gravity vs pressure (thermal, magnetic, turbulence, radiation, cosmic rays) and shear. Heating and cooling, generation and decay of turbulence, generation (dynamo) and diffusion of B-fields, etc. Chemical evolution of dust and gas. A complicated, nonlinear process Some notation: Core -> star or close binary Clump -> star cluster

Outline Physical properties of massive star-forming regions Theoretical scenarios - core accretion, competitive accretion, mergers, etc. The “Turbulent Core Accretion” Model Initial conditions: IRDCs; how are they generated? Does the clump reach pressure equilibrium? Timescale of star cluster formation? Collapse of the core: fragmentation? Massive protostars: star, disk, outflow formation and evolution. Radiative transfer modeling. [Feedback: outflows, ionization, rad. pressure. On core & clump.]

Overview of Physical Scales

A V =7.5 A 8μm =0.30 N H =1.6x10 22 cm -2  =180 M  pc -2 A V =1.4 N H =3.0x10 21 cm -2  =34 M  pc -2

Overview of Physical Scales A V =7.5 A 8μm =0.30 N H =1.6x10 22 cm -2  =180 M  pc -2 A V =1.4 N H =3.0x10 21 cm -2  =34 M  pc -2 A V =200 A 8μm =8.1 N H =4.2x10 23 cm -2  =4800 M  pc -2 Mueller, Shirley, Evans, Jacobsen (2002)

Overview of Physical Scales A V =7.5 A 8μm =0.30 N H =1.6x10 22 cm -2  =180 M  pc -2 A V =1.4 N H =3.0x10 21 cm -2  =34 M  pc -2 A V =200 A 8μm =8.1 N H =4.2x10 23 cm -2  =4800 M  pc -2

Overview of Physical Scales A V =7.5 A 8μm =0.30 N H =1.6x10 22 cm -2  =180 M  pc -2 A V =1.4 N H =3.0x10 21 cm -2  =34 M  pc -2 A V =200 A 8μm =8.1 N H =4.2x10 23 cm -2  =4800 M  pc -2 M82 SSCs (McCrady & Graham 2007)

Overview of Physical Scales A V =7.5 A 8μm =0.30 N H =1.6x10 22 cm -2  =180 M  pc -2 A V =1.4 N H =3.0x10 21 cm -2  =34 M  pc -2 A V =200 A 8μm =8.1 N H =4.2x10 23 cm -2  =4800 M  pc -2 n H ~2x10 5 cm -3 t ff ~1x10 5 yr SSCs in dwarf irregulars (K. Johnson, Kobulnicky, J. Turner et al.) These are the environments where massive stars form: can we scale-up low-mass SF theory? These are the environments where massive stars form: can we scale-up low-mass SF theory?

Turbulent core model (McKee & Tan 2002, 2003) Schematic Differences Between Massive Star Formation Theories time disk fragmentation core fragmentation t=0 protostar formation massive star m *f >8M  m * =8M  pre-massive-stellar core massive-star-forming core [protostar+gravitationally-bound gas] massive-protostar (MP) LIMP-MP Beuther, Churchwell, McKee, Tan (2007); Tan (2008) Radiation pressure likely to prevent accretion of dusty, unbound gas (e.g. Edgar & Clarke 2004) Competitive Bondi-Hoyle accretion model (Bonnell ea. 2001; Bonnell & Bate 2006; Dobbs+, R. Smith+, P. Clark+) Rare evolution from magnetically subcritical state? Is there any isolated massive star formation? Prestellar core mass function?

Strong magnetic fields in star-forming regions Girart et al. (2009) Correlation of field orientations from ~100pc to <1pc scales (Hua-bai Li et al. 2009) Taurus (Heyer et al. 2008) Strength of B-field vs.  (Crutcher 2005; Falgarone et al. 2008) Supercritical Subcritical  = 1 g cm -2

From Cores to Stars: Individual Stars Appear to Form from Cores Nutter & Ward-Thompson (2007) Beuther & Schilke (2004) See also: e.g. Motte et al. 1998; Testi & Sargent 1998; Motte et al. 2001; Mike Reid & Wilson 2005; Alves et al. 2007; Li et al. 2007; Enoch et al. 2008; Pineda et al. 2009; Ragan et al. 2009; André et al Rathborne et al. (2009) ε core = m * /m core -> ~

Overview of Physical Scales A V =7.5 A 8μm =0.30 N H =1.6x10 22 cm -2  =180 M  pc -2 A V =1.4 N H =3.0x10 21 cm -2  =34 M  pc -2 A V =200 A 8μm =8.1 N H =4.2x10 23 cm -2  =4800 M  pc -2 n H ~2x10 5 cm -3 t ff ~1x10 5 yr Turbulent Core Model of Individual Massive Star Formation (McKee & Tan 2003)

How many Massive Starless Cores in the Galaxy? Number in the Galaxy: (see also Zinnecker & Yorke 2007) If lifetime of this phase is ~ t *f ~ 10 5 yr, then for a Galactic SFR of 3M  yr -1 and an IMF yielding 1 massive star per 130 M  (Salpeter M  ) and 2/3 of massive stars are forming in binaries, we expect 1500 Massive Starless Cores in the Galaxy.

Mid-IR Extinction Mapping of Infrared Dark Clouds 16’ Spitzer - IRAC 8  m (GLIMPSE) Extinction map to derive  Distance from molecular line velocities (GRS) -> M(  ) (Butler & Tan 2009; see also Peretto & Fuller 2009; Ragan et al. 2009; Battersby et al. 2010) MJy sr -1 Median filter for background around IRDC; interpolate for region behind the IRDC Correct for foreground emission - tricky-> choose nearby clouds g cm -2 G

Application to Filamentary IRDCs 3’ I 8μm (MJy sr -1 )Σ (g cm -2 ) G035.39−00.33 Comparison to mm dust emission (Rathborne et al. 2006) and 13 CO and C 18 O line emission (Hernandez & Tan, submitted), give agreements at ~factor of 2 level

Formation of IRDCs Some evidence that filamentary IRDCs are not yet virialized: Extended SiO emission along one IRDC (Jimenez-Serra et al. 2010) Filamentary virial analysis of 2 IRDCs (Hernandez & Tan, submitted) But, the regions closer to virial equilibrium do appear to be those forming stars Comparing to models of Fiege & Pudritz (2000)

Massive Starless Cores MIPS 24μm IRAC 8μm Extinction Map Butler & Tan (2009), Butler & Tan, in prep. Σ = 0.26 g cm -2 m core = 205 M  Σ = 0.12 g cm -2 m core = 94 M  Σ = 0.12 g cm -2 m core = 50 M  Cores show central concentration; can fit power law radial density profiles, index ~-1.5. They contain many thermal Jeans masses. B-fields may be suppressing fragmentation within the core. 10” n H ~10 5 cm -3, B~1mG -> M B ~100 M 

Overview of Physical Scales A V =7.5 A 8μm =0.30 N H =1.6x10 22 cm -2  =180 M  pc -2 A V =1.4 N H =3.0x10 21 cm -2  =34 M  pc -2 A V =200 A 8μm =8.1 N H =4.2x10 23 cm -2  =4800 M  pc -2 n H ~2x10 5 cm -3 t ff ~1x10 5 yr Butler & Tan 2009

We expect massive star forming environments exist for >1t ff and so can achieve approx. pressure equilibrium (proto star clusters take > 1t ff (central) to form) IRDC cores have t ff ~10 5 yr, which is short Some (most?) star clusters appear to have age spreads >10 6 yr, e.g. Orion Nebula Cluster median age of 2.5-3Myr (Da Rio et al. 2010) A plausible mechanism has been identified to maintain turbulence over many t ff : protostellar outflow feedback (Norman & Silk 1980; Nakamura & Li 2007) Tan, Krumholz, McKee (2006) While the issue of star cluster formation timescales is still debated (e.g. Elmegreen 2000, 2007; Hartmann & Burkert 2007), it seems likely that t form >t ff (central).

Collapse of the Core - Core Fragmentation? We expect most of these structures will fragment to form star clusters. Most mass -> low-mass stars. Fragmentation will be reduced by radiative feedback from the central star (Krumholz, Klein & McKee 2007; c.f. Dobbs, Bonnell, Clark 2005). Magnetic field support should increase the “magneto-Jeans” mass and reduce fragmentation: (Machida et al. 2005; Price & Bate 2007, Hennebelle & Teyssier 2008, Duffin & Pudritz 2009). However, see Li, Wang, Abel, Nakamura (2010). Fragmentation should be reduced by radiative feedback from surrounding accreting low-mass stars (Krumholz & McKee 2008).

Overview of Physical Scales A V =7.5 A 8μm =0.30 N H =1.6x10 22 cm -2  =180 M  pc -2 A V =1.4 N H =3.0x10 21 cm -2  =34 M  pc -2 A V =200 A 8μm =8.1 N H =4.2x10 23 cm -2  =4800 M  pc -2 n H ~2x10 5 cm -3 t ff ~1x10 5 yr Butler & Tan 2009 Fragmentation stopped by radiative heating (Krumholz & McKee 2008) But B-fields likely to also suppress fragmentation

The later stages of individual massive star formation Final mass accretion rate Core core bounded by pressure of clump

Support by combination of large & small scale B-fields, and turbulent motions. Core boundaries fluctuate. Final mass accretion rate Protostellar evolutionDisk structureOutflows r*r* m*m* The later stages of individual massive star formation Outflow-confined HII Region

Collapse from Core to Disk Observational evidence for rotating toroids on scales ~1000AU, perpendicular to bipolar outflows, e.g. G A1 Also claims from maser observations (e.g. Wright et al., Greenhill et al., Goddi et al. in Orion KL; Pestalozzi et al in NGC7538 IRS1N) A1 A2 Theory: Analytic study of disk accretion and fragmentation (Kratter, Matzner, Krumholz 2007) Radiation-hydro simulation of turbulent core collapse: modest disk fragmentation. (Krumholz, Klein, McKee 2007a). Simulated ALMA observations (Krumholz, Klein, McKee 2007b). CH 3 CN (12,0,0,13)->(11,0,0,12) [220GHz, 69K] Beltrán et al. (2004 )

Protostellar Evolution Radius Luminosity protostar+boundary layer+disk Ionizing L. Outflow momentum flux (scaled from Najita & Shu 94) Total outflow momentum [see also Hosokawa & Omukai 2009] Tan & McKee 2002

Core Star Formation Efficiency from Outflow Feedback Assuming angular distribution of momentum flux of standard X-wind or disk-wind models (Matzner & McKee 1999), work out the maximum angle from the rotation/outflow axis at which core gas is expelled, assuming a steady wind that either stays coupled beyond the core radius or decouples. This sets ε core (Matzner & McKee 2000) and Tan & McKee, in prep.

Overview of Physical Scales A V =7.5 A 8μm =0.30 N H =1.6x10 22 cm -2  =180 M  pc -2 A V =1.4 N H =3.0x10 21 cm -2  =34 M  pc -2 A V =200 A 8μm =8.1 N H =4.2x10 23 cm -2  =4800 M  pc -2 n H ~2x10 5 cm -3 t ff ~1x10 5 yr Butler & Tan 2009 Bontemps et al. (2010)

Radiative Transfer Modeling Zhang & Tan, in prep.

Radiative Transfer Models Zhang &Tan, in prep. see also: Robitaille et al. 2006; Molinari et al Rotation and outflow axis inclined at 60˚ to line of sight. Σ = 1 g cm -2 M core = 60 M m * = 8 M m disk = m * /3 L bol = 6x10 3 L

Radiative Transfer Models Rotation and outflow axis inclined at 60˚ to line of sight. Σ = 1 g cm -2 M core = 60 M m * = 8 M m disk = m * /3 L bol = 6x10 3 L Zhang &Tan, in prep. see also: Robitaille et al. 2006; Molinari et al d=1kpc convolved with telescope beam

G35.2N (De Buizer 2006) L MIR ~ 1.6x10 3 L Mid IR Emission - Outflow Cavity 11.7μm 18μm Observations: Rotation and outflow axis inclined at 60˚ to line of sight. m * = 8 M L bol = 6x10 3 L

Outflow-Confined HII Regions (Thermal Radio Jets) 15GHz A number of ionized HCHIIs seen in other nearby sources (e.g. van der Tak & Menten 2005 Orion source I: Tan & McKee 2003) G35.2N (De Buizer 2006) 18μm IRAS Guzmán et al. (2010) 10μm 8.6GHz

Highly-collimated outflows from massive protostars e.g. Beuther et al. (2002) Evidence for similar accretion processes as low-mass SF

Conclusions Is massive star formation a scaled-up version of low-mass star formation? “Turbulent Core Model”: normalize core surface pressure to surrounding clump pressure, i.e. self-gravitating weight. The cores are probably are marginally magnetically super critical, limiting their fragmentation. If there is a different mechanism, e.g. competitive accretion, stellar mergers, then one would expect some break in the IMF at the mass scale it takes over. How can competitive accretion, i.e. accretion of dusty gas initially unbound to the protostellar core, overcome radiation pressure feedback for m * >10M sun ? Mergers require unrealistic stellar densities. We see massive pre-stellar and star- forming cores; rotating toroids; ordered B- fields; collimated outflows; outflow- confined HII regions (thermal radio jets)....and, yes, I too have a theory for Orion!

Herschel: Find us the Massive Pre-Stellar Cores......and, if you are brave, do multi-layered kinematics to understand their formation. Find us the Massive Protostars......we probably already know (HCHIIs, UCHIIs) -> L bol ALMA: Kinematics of the core envelopes and disks... Magnetic field strengths and orientations... e-VLA and ALMA: more outflow-confined HII regions, please...