The Origin of Stars and Their Planetary Systems

Exploring the Universe with JWST II

24-28 Oct., 2016, Montreal

Ralph E. Pudritz

McMaster University

Students, postdocs, and collaborators McMaster: Ph.Ds. Mikhail Klassen, Corey Howard, Jason Fiege (U. Manitoba) Alex Cridland and Matt Alessi (planets) Helen Kirk (Banting Fellow, HIA), Bill Harris Samantha Pillsworth (undergrad) Hamburg/Heidelberg: Robi Banerjee, Daniel Seifried (Ph.D.) Ralf Klessen PPVI Andre et al team (Philippe Andre, James de Francesco, Derek

Ward-Thompson, Shu-ichiro Inutsuka, and Jaime Pineda): PPVI Li et al team (Zhi-Yun Li, Jes Jørgensen, Hsien Shang, Ruben

Krasnopolsky, Anaëlle Maury)

I. Clouds, filaments, and star formation

Filaments – and the road to cores…

Diffuse Atomic Hydrogen in Milky Way (Canadian Galactic Plane Survey CGPS) near midplane towards Perseus. Extinction map of Orion and Mon

clouds (Cambresy 1998); right - Scuba continuum 850 micron map of 10 pc portion of cloud (Johnstone & Bally 2006)

Shocks and filaments on many scales, in diffuse and self-gravitating gas

Turbulent structure in ISM

Star forming gas in GMCs: column density structure in molecular clouds (use near IR dust extinction maps)

No star formation: Star formation: Lupus V and Coalsack clouds Taurus and Lupus 1 clouds Lognormal column density Lognormal + power-law Turbulence produces lognormals – gravity the power-law..

Kainulainen +2010

Column density thresholds for star formation

Left – column density distribution for cores in Aquila (Könyves+2010)

Right – column density distribution for cloud (André+2011, Schneider+2013) Threshold: vertical line Av ~ 8, or N ~ 7 x 1021 cm-2

volume density n ~ 104 cm-3

Star forming cores in filaments: gravitational instability?

Filaments: lengths 0.1 – 100 pc; masses 1 – 105 Msun , line masses < 1000 Msun pc-1 (Bally+ 1987, Hacar+2013, Kainulainen+2013,2016, .. Review: Andre

+2014) Herschel observations: clouds filamentary down to sub pc (Andre+ 2010, 2014 (PPVI), Menshchikov+2011, Henning+2010..) - Cores strongly associated with filaments > 70%; Polychroni+ 2013)

- GI? White = mass / length (T=10K): m >mcrit = 2 cs

2 / G ~ 16 Msun pc-1

Aquila star forming cloud: Andre+ 2010

Core mass functions n  Well fit by lognormal

distribution.

n  Closely resembles the IMF

n  Peak at 0.6 Msun

n  Compare to peak of IMF:

conversion efficiency of core to stellar mass: ~ 0.2- 0.4 in Aquila Feedback effects?

Core mass function for Aquila cloud (500 starless cores: 0.01-0.1 pc in size) (Könyves+2010, Andre+2010, Andre+ review 2014)

Filaments and cores live in highly dynamic environments: Velocity structure in filaments

Accretion and filamentary flow: sketch of velocity field around Serpens-South cluster (Kirk et al 2013)

Total velocity dispersion vs central column density of filaments (Arzoumanian et al 2013)

B fields - channeled accretion flows or infall? (review Hua-bai Li + 2014) Infrared (H band) polarization overlaid on column density map + young stellar cluster

- Orientation – often perp to dense filament? - Field strength ~ 100 µG

Radiative feedback from massive stars in young clusters – affect filaments? IMF?

Filaments and OB clusters, and HII regions in G10.6-0.4 ( IRAM 30m MAMBO-2 bolometer array + SMA ) - 200 M¤ OB cluster - Ultracompact HII regions: A-E

H.B. Liu et al 2011

Theory: turbulence, gravity and fragmentation

Turbulence, filaments, and

turbulent fragmentation -Theory; eg. Larson 1981; Elmegreen & Scalo

(2003) -Reviews: eg. MacLow & Klessen 2004; McKee

& Ostriker 2007; Bonnell et al 2007 - Simulations; Porter et al 1994; Vazquez-

Semadeni et al 1995, Bate et al 1995, Klessen & Burkert 2001; Ostriker et al 1999, Padoan et al 2001; Tilley & Pudritz, 2004,2007; Krumholz et al 2007, Federrath et al 2010,…

Shocks dissipate turbulent support as t-1 (eg. Ostriker 2001)

Bonnell et al (2003)

1. Turbulent fragmentation: - Isothermal gas and supersonic turbulence, 3D simulations: PDF of gas density = log- normal (Padoan & Nordund 1997). Generalize to Include chemistry (Glover+2010). - Star Formation Rate- Integral of density fluctuations above threshold: each collapsing in 1 local free-fall time. Efficiency ε~ 0.3-0.7 - Different models for threshold (review Padoan+ PPVI, (Hennebelle & Chabrier 2011, Fedderath & Klessen 2012, Hopkins )

P(s) = 12πσ s

2exp −(s− so )

2

2σ s2

⎡

⎣⎢

⎤

⎦⎥

s ≡ ln(ρ / ρo )so = −σ s

2 / 2σ s2 = ln[1+ bMs

2 ]Ms =σ v / cs

SFR = εφt

t ff ,ot ff (s)scrit

∞

∫ sP(s)ds

2. Gravitational instability in filaments 1. Radial force balance for infinite cylinders, constant line mass m ~ ρ r2 ~ const i) Gravitational force: Fg = 2Gm / r ~ r -1

ii) Radial (thermal) pressure gradient for polytropic gas (index γ ) Fp = ρ-1 dP/dr ~ r 1- 2γ -> radial pressure / gravitational force ~ r 2 ( 1- γ)

POINT: γ > 1 stable (pressure wins for small r)

Υ<1 unstable (pressure loses to gravity ultimately) γ=1 constant force ratio -> critical condition. mcrit = 2cs

2 / G = 16 (T/10 K) M¤ pc-1 Values: 2 states, low and high density; γ ~ 0.73, 1.03 (Larson 1985, 2005; Koyama & Inutsuka 2000; low density – Benson & Myers 1989, Evans 1999)…

Radial density profiles – probing forces on filaments Not observed: isothermal filaments (Ostriker 1964; r -4 ). Herschel: r-1.5 – r-2.5 (Arzoumanian+2011) MHD: Poloidal field supports cylinder, toroidal component *squeezes it*. Fiege & Pudritz 2000a (FP2000ab) Observations

(Matthews & Wilson 2001) r -1.5 – r -2 Accretion: dynamic accretion not static equilibria: r-1.3 - r – 2

(R.Smith+2014; Kirk, Klassen, Pudritz, & Pillsworth 2015) Turbulence: Changes the critical mass to mcrit = 2σ2 / G ; σ2 = cs

2 (1 + M2/3) (FP2000a)

JWST: Column density extinction maps of cores/dense filaments much better than ground based ( Av=50 mags, at high res 10-15”) -> Av = 120 mag -> KAB = 29 mag or 9nJ @ 2µm

Fragmentation of the “Integral Filament” – ALMA Observations of Orion A (Kainulainen+2016)

Multi-scale fragmentation - Linear theory; periodic - fastest growing mode ~ several times filament width (eg. Larson.., Inutsuka * Miyama,.. Peak at 55,000 AU but growing power at spacings < 16,000 AU - Periodic fragmentation for 55,000 AU? (Fiege & Pudritz 2001b, Fischera & Martin 2012). Jeans type in spherical regions < 16,000 AU?

Surface density of dense cores – contours at 60,80…140 x 1021 cm-2

2 point correlation function of dense cores

II. Cluster formation and radiative feedback

IMF in star forming regions

IMFs from 8 star forming regions. Data: Scholz+2012, Alves de Oliveira+2013, Oliveira+ 2009, Sung & Bessell 2010, Luhman+2007, Pena Ramirex+2012, Bayo+2011, Lodieu2013. Compiled by Offner+2014 (ppvi)

Remarkable agreement between different regions: with IMF Chabrier (2005): lognormal - with peak 0.2-0.3 Msun and σ = 0.5 JWST: push down to 1 Mjup by observing in 3-10 µm: -  Broad band NIR and MIR

photometry - Classification spectroscopy

Cluster Formation Including Radiative Feedback Radiative feedback from

massive stars: raises Jeans Mass - filaments don’t fragment - gas drains into primary and its disk (eg. Krumholz et al 2007) - prevent fragmentation out to

1000 AU scales Suppression of brown dwarfs: by factor 4 (Bate 2009): get robust low mass part of IMF…?

2/3TM J ∝

Bate (2009)

Mass limitation during cluster formation: fragmentation-induced starvation? (Peters, Klessen, Mac Low, & Banerjee 2010)

Collapse into a disk of 1000 Msun clump. Ionization feedback. -  Fragments compete for gas in an unstable

disk: muliplicity for O stars… -  Radiative heating -> larger Jeans mass -  (100 AU resolution)

Disk mass levels out..

Building Clusters - ionization does not readily erode dense cold filaments (Dale & Bonnell, 2011)….. how does accretion stop? Ionize diffuse gas in the voids instead – filamentary structure has major implications for efficiency of feedback, and therefore SFR Break the filaments with outflows? (eg. Wang et al 2010)

Radiative feedback and cluster formation in simulated GMCs (Howard, Pudritz, & Harris, 2016, MNRAS)

n  Radiative feedback from evolving IMF in cluster sink particles – using ray tracing from sources (Peters + 2010), not FL. Cluster sink particles.

n  Clouds idealized: initial n=100 cm-3, box 32- 80 pc, flattop + r-3/2 density profile; low rotation (2 %), Burgers turbulence, Mcl = 106 M¤

n  Ionization, thermal heating, radiative momentum..

SEQUENCE in virial parameter α. Observed clouds: α from 0.5 – 5 (Rosolowsky2007, Hernandez & Tan 2015))

M = 106 Msol, α = 0.25, flatop,..

Cluster formation: initially poorly bound (α = 3), 106 MSun , GMC

α and cluster formation

Top: Stellar mass evolution of most massive star clusters – jumps are mergers.. Bottom: clump and star formation efficiencies Unbound model shows efficiencies that agree with obs. Radiative feedback effects negligible…! POINT: low SFE requires initially unbound clouds…

UV escape fraction from Giant Molecular Clouds (Howard, Pudritz, & Klessen 2016, archive next week)

- Hammer projections of UV ionizing photon fluxes from 106 Msun , α = 3 cloud; - Circles – 10 most luminous clusters, positions projected.

UV escape fraction- highly variable in time with peaks fesc,max ~ 30-35%, average fesc ~ 15%

III. Massive Star Formation

Disks around massive stars Keplerian disks observed around 8-30 Msun stars (review: Bertran & De Witt 2013) Disk radii ~ few 1000 AU, masses from 4 to a few 10s of MSun Core-accretion picture favoured? (Myers & Fuller 1997, McLaughlin & Pudritz 1997, McKee & Tan 2002)..

ALMA maps: Keplerian disk around forming O7 star in AFGL 4176 mm1: 25 Msun star, with 12 Msun disk. Top: outflow perp to disk Bottom: CH3CN (J=13-12) (Johnston + 2015)

Best fit is Keplerian disk; velocity peaks of different tracers: (Sanchez-Monge + 2013)

Massive star formation – how do you make a 150MSun ? Geometry: for spherical infall, Eddington limit for most massive star = 40Msun (Larson & Starrfield 1971, Kahn 1974, Yorke & Krügel 1977, Wolfire & Cassinelli 1987,..)

-1D sims give Mmax ~ 40 Msun independent of core mass > 60 Msun (Kuiper + 2010 ).

Disk accretion -> asymmetry to radiation field (Nakano 1989) - “flashlight” effect (Yorke & Bodenheimer 1997)

- 2D disk sim, multifrequency FLD but unresolved dust sublimation front; -  similar mass limit (Yorke & Sonnalter 2002):

Disk in 120 solar mass case quickly destroyed

The problem: need to resolve dust sublimation region!

Yorke & Sonnalter 2002

Pure FLD vs Hybrid Radiative Feedback (Krumholz+2009 – Kuiper + 2010)

Pure FLD in a 3D AMR code: initial 100 Msun core, r-1.5 density profile, no turbulence.. - R-T instabiltiies @ 41.7kyr, multiple fragmentation

Hyrbid code (ray trace + FLD) - 2D: 120 Msun shown over long time 460kyr – no R-T instabilities

Radiative Feedback and Massive Star Formation (Klassen, Pudritz, Kuiper, Peters, & Banerjee 2016, ApJ ) - using Hybrid AMR FLASH radiative transfer (Klassen+ 2014)

4 stages: - Gravitational collapse - Disk formation - smooth - Disk instability and massive episodic accretion - Radiative feedback – radiatively driven bubbles

No R-T instability seen (remove ray trace and RT develops, Kuiper+2012) No disk fragmentation up to 50kyr. 100 Msun mode: left face on, right edge on

Time evolution of central “sink” (massive star) for 3 sims - 30, 100, and 200 Msun cores

1. Initial phase of accretion: - core collapse dM/dt = mo cs

3 / G mo ~ NJ (number of Jeans masses) (Shu,1977 Giichidis + 2012)

2. Disk accretion - epsodic driven by onset of GI. (Vorobyov & Basu 2007). Disk mass levels out…

Edge- on view of collapse of 200 Msun core zoom in to central region

Fragmentation and cooling… mostly stable

-1000

0

100015 kyr

-1000

0

100025 kyr

-1000

0

100035 kyr

-1000 0 1000

-1000

0

100040 kyr

-1000 0 1000 -1000 0 1000

0.0 0.6 1.2 1.8 2.4 3.0 3.6 4.2 4.8 5.4 6.0β Stability Parameter

Fragmentation of disks requires both Q< 1, AND rapid cooling (Johnson & Gammie 2003) tcoolΩ = β < 1 Middle: β map ( cooling from flux radiative flux loss): Right column: Q<1 in dark blue: maybe one fragments ~ 1000 AU

JWST observations..

- No disk fragmentation observed on ~1000 AU scale for t < 50 kyr. Possibly later… - No R-T fallback (But see Rosen+2016) - Primary role of disk instability; drives high accretion rates that are highly episodic.

JWST: fragments in disks? Diffraction limited imaging at NIR? (3.8 and 4.8 microns. Av ~ 150 mag reduced to 4 and 6 mags. Diffraction limited imaging gives 0.15” res… equivalent to 7 fully sampled elements acros 1” = 1000AU @ 1kpc.

IV. Low mass star formation: disks, outflows, planet formation…

1. First hydrostatic stellar cores (FHSC) and Class 0 Protostars

- FHSC earliest object forms inner region becomes opaque to radiation (ρc > 10-13 gm cm-3) – collapse becomes adiabatic (Larson 1969). Collapse resumes when when T> 2000K (H2 dissociates). Mass ~ 1 Jupiter; radius ~ several AU, lifetime ~ 102 - 103 yr, ~ 9 candidates.. JWST- some signal at MIR wavelengths… - Class 0: protostar too deeply embedded for source detection, but has outflow.

Dunham+2014

2. Disk formation

Cloud to YSOs: Specific angular momentum j as a function of scale and YSO evolutionary stage: - j conserved on core infall scales - Keplerian inside 100 AU

Belloche 2013

Disk formation in turbulent, magnetized cores: 1. Quiescent rotating core with B field: Magnetic braking “catastrophe” NO DISK FORMS! (Mellon & Li 2008, Hennebelle & Fromang 2008, review Z-Y Li+ 2014) -  Turbulent reconnection? (Santos-Lima+ 2012,2013) -  Turbulent MHD – degrades torque, disk forms – accretion from a few

filaments covering only 10% of area (Seifried+2012, 2013, 2015)

1300 AU scale: 2.6 and 100 M¤, no initial rotation, velocity vectors, B (black lines), forming Keplerian disk (blue), filaments (green). Seifried, Banerjee, Pudritz, & Klessen ( 2015)

Observational Tests…

MIR: Class 0 observations: Simple envelope profiles ρ ~ r-2 , r-1.5 must be modified (Jørgensen+2005, Enoch+2009)

Complex environments on 1000 AU scales (Tobin+ 2010) - asymmetries and filaments… NOT in line with spherical collapse pictures. JWST MIR observations to identify/image filamentary build up of disks…

- Upper; magnetic tower flow - Lower; zoomed in by 1000, centrifugally driven disk wind

3. Outflows Early stages: gravitational collapse of rotating, magnetized cores produces disks and disk winds (Banerjee & Pudritz 2006):

2 components of the flow

MHD outflows associated with formation of first hydrostatic core (Bate+2014)

Solving angular momentum problem for disks and accreting YSOs :

Gressel+2014

Centrifugally driven disk winds carry angular momentum from disk (review Pudritz+2007); Accretion powered stellar winds spin down YSOs (Matt & Pudritz 2005, Zanni & Ferreira 2009)) [disk locking picture – Königl 1991] Disk turbulence (MRI) completely quenched in inner regions by non-ideal MHD (“dead zone”) – ang. momentum transport by disk winds (Bai & Stone 2013, Gressel + 2014..)

Matt & Pudritz (2005)

How star formation controls planet formation…

Stellar irradiation of evolving disks - temperature structure -> heat transition: viscous to radiative -> position of ice lines Stellar ionizing flux on evolving disks – ionization driven disk chemistry -> position of dead zones All 3 of these act as planet traps – slow planetary migration – affects planetary composition… (eg. Hasegawa & Pudritz 2011, 2012, Alessi, Pudritz, & Cridland 2016, Cridland, Pudritz, & Alessi 2016).

ALMA partnership 2015

Zhang+2015

Summary:

JWST can drive transformative research on: - Physics of core and filament structure and formation - IMF at BD and Jovian mass scales - HII regions/ UV escape from GMCs - Existence of earliest protostellar hydrostatic states (FHCs) - Filamentary flows and formation of disks - Structure and fragmentation of massive disks - Early outflows