Studying the Atomic-Molecular Transition in the Local Group

Erik Rosolowsky

Radio Astronomy Lab, UC Berkeley

Ringberg - May 19, 2004

CollaboratorsCollaboratorsThe BossThe Boss: Leo Blitz: Leo BlitzCollaboratorsCollaborators::

Dick PlambeckDick PlambeckGreg EngargiolaGreg EngargiolaJulianne Dalcanton (UW)Julianne Dalcanton (UW)

Star Formation

• A fundamental problem– Solution required for a time evolution of stellar

populations in disk.

• With fundamental complications:

The Schmidt Law

Kennicutt (1998)

Resolved Schmidt Law Studies

Wong & Blitz (2002) studied CO and star formation in a sample of 7 galaxies.

A slight case of déjà vu.

CO Only

Kormendy & Kennicutt (2004)

CO & HI

Photo Credits: R. Gendler ,the FORS Team, D. Malin, SAO/Chandra, D. Thilker

The Gas Cycle in the ISM

Molecular Clouds Star Forming Regions

Stars

Supernova RemnantsStellar Ejecta

Atomic ISM

Photo Credits: R. Gendler, D. Malin, D. Thilker

Molecular Clouds

Stars

Atomic ISM

This TalkThis Talk

Schmidt LawSchmidt Law

Stellar EvolutionStellar Evolution& &

Turbulent ISMTurbulent ISM

InfallInfall

Toward a simple model….

What is a Giant Molecular Cloud?• Large cloud of molecular

gas: M > 104 Msun

• Self gravitating [?]

– -T ~ Ugrav/2

• In MW, nearly all the molecular mass is in GMCs.

• Since SFR scales with MH2, then GMCs

populations set star formation history.

PASJ, vol 56, no. 3, cover

The Orion Molecular Cloud in 12CO(1-0)

How do GMCs vary across the Local Group?

(which is secretly a question about how GMCs form…)

Macroscopic Cloud Properties

• Resolved observations give cloud radius (R)– Correct for beam convolution!Correct for beam convolution!

• Get linewidth (V) from spectral lines

• Luminous mass from MCOXLCO

• Virial Mass for resolved observations

Constant X Factor?• Comparing Virial and

CO masses over a range of galactic radii in M33

• No significant trend with radius

• No change in X due only to:– Metallicity (0.6 dex)– ISRF (1 dex)– Midplane hydrostatic

pressure (1 dex)

Larson’s Laws• Larson (1981) noted correlations among the simplest

characteristics of molecular clouds.

• The linewidth-size relationship is expected for turbulent motions.

• If the clouds are virialized, the mass-linewidth relationship follows from linewidth-size and V.T.

• Caveat: How well do these characterize GMC properties?

The Linewidth-Size Relationship

The Linewidth-Mass Relationship

The LG-GMC Population

• Individual GMCs in MW, LMC, M31, M33 are consistent with being drawn from the SAME statistical population– 1 Parameter Clouds

• These macroscopic properties of GMCs set average internal properties (, Pint, tdyn)

• A constant IMF would not be surprising for a common population of molecular clouds.

• Parameterize with cumulative mass distribution:

• Binned approximations are only accurate for sample sizes larger than ~300 (only 1 sample of GMCs)

The GMC Mass Distribution

M33

The Local GroupGMC Mass Distribution.

Mass Spectra are different!

• Re-fit all catalogs of GMCs available that have reliable data

• Changing index is likely the signature of different formation mechanisms.

• Enter: the importance of dynamics.

Object Inner MW -1.60 to -1.72

LMC -1.63 to -1.92

Outer MW -1.91 to -2.11

M33 -2.10 to -2.60Inc r

e asi

ng H

D S

tab i

lity

Inferences about GMC formation

• Physics intrinsic to GMCs establishes their macroscopic properties (e.g. self gravity).

• GMCs appear to unify the star formation process across a variety of environments.

• Suggests important factor in SF is the conversion of gas into GMCs.

• Conversion efficiency (and process?) varies across environments.

Where does H2 form?

(and what physics makes that so?)

(and is this the same as making GMCs?)

Why go extragalactic?

• Top-down perspective• No blending!• Association with other

components in the ISM

• Spatially complete studies

• Wide range of galactic radii

From Dame, Elmegreen, Cohen & Thaddeus (1986)

M33 in H

Cheng et al. (1993)

• 850 kpc distant• Sc spiral• 1 of 3 Local Group

spirals

The D-array Survey

The GMCs in M33

Correlation with HI

Deul & van der Hulst (1987)

• BIMA SONG (Helfer et al., 2003) CO(,)

• Literature Maps of HI HI (single value)

• 2MASS K-band maps (Jarret et al. 2003) *(,)

What determines fmol(R)?

*=120 Msun / pc2

The Physics of *=120 Msun/pc2

• Constant value of ISRF– Sets H2 dissociation rate

• Constant Midplane Pressure

• Constant volume density (nH)

– Sets H2 formation rate

Work in Progress

1. Include spatial distribution of HI

2. Include rotation curves

Assembling a Big Picture1. Filaments of HI (H2) collected by [M]HD

processes2. Another factor [f(R)] determines what fraction of

these clouds are converted to molecular gas3. Different environments create different mass

distributions of bound molecular clouds.4. Self-gravity (or other physics) establishes uniform

Larson Law scalings across environments.5. Macroscopic properties of GMCs set their internal

properties, which are the initial conditions of star formation.

Future Efforts: NGC 4826

• Extreme surface density of molecular gas.

• No sign of discrete 12CO clouds.

• 13CO clouds have similar properties as MW GMCs and show signs of star formation.

Dwarf Ellipicals

• CO emission seen in dEs NGC 185 and NGC 205.

• Gas appears to be intrinsic, not from infall or stripping

• Presence of cool ISM and star formation without:– Spiral arms

– Ordered B-field

– Shearing disks

– High HI column densities

NGC 185 - L. Young (2001)

Requirements for Formation

• Consider a 106 Msun GMC with D=80 pc

• Requires enhancing the surface gas density from gas=10 Msun pc-2 (ISM) to

GMC = 200 Msun pc-2

• Implies accumulation scale of l >350 pc.

• If atomic, the conversion to molecular gas is reasonably quick for typical densities (3-10 Myr).