Unraveling the Star Formation Scaling Laws in
Galaxies (review + 2 new results)
Robert KennicuttInstitute of Astronomy
University of Cambridge
M51: FUV, H, 24m
rants
• Galaxies exhibit an immense diversity in star
formation properties, varying by >107 in absolute SFR,
SFR/mass and SFR/area.
• Over this range the SFR/area is correlated with gas
surface density, following a truncated Schmidt power
law with index N = 1.4 +-0.1
– the correlation of with dense gas (e.g., HCN) is roughly linear
• The Schmidt law shows a turnover below a threshold
surface density that varies between galaxies.
– in gas-rich, actively star-forming galaxies this transition is
seen as a radial transition in the SFR/area
– some gas-poor disks reside in the threshold regime at all radii
Basic Observations: circa 1998
IR-luminous starbursts circumnuclear starbursts BCGs, ELGs
Kennicutt 1998, ApJ, 498, 541 Gao, Solomon 2004, ApJ, 606, 271
normal galaxies
starbursts
NGC 1291
Blue: Carnegie Atlas Sandage & Bedke 1994
H + R: SINGG survey Meurer et al. 2006
• Is the correlation really this good (or this bad)?
– Why all the discrepant results yesterday (and today)?
– Do all galaxies follow the same Schmidt law?
– Is the scatter driven by a second parameter?
• Is the Schmidt law the result of a more fundamental
underlying SF scaling law?
– over what range of physical scales is the law valid?
• Is the SFR correlated more strongly with the total
(atomic + molecular) surface density or with the
molecular surface density alone?
• What is the physical origin of the relation?
Questions: Schmidt Law
“Schmidt law”: SFR vs gas density power law
“Silk law”: SFR vs gas density/dynamical time
Blitz & Rosolowsky 2006, ApJ, 650, 933
“pressure law”
• Is the correlation really this good (or this bad)?
– Why all the discrepant results?
– Do all galaxies follow the same Schmidt law?
– Is the scatter driven by a second parameter?
• Is the Schmidt law the result of a more fundamental
underlying SF scaling law?
– over what range of physical scales is the law valid?
• Is the SFR correlated more strongly with the total
(atomic + molecular) surface density or with the
molecular surface density alone?
• What is the physical origin of the relation?
Questions: Schmidt Law
• Do the observed H edges of galaxies trace
proportional changes in the SFR/area?
• Does the SFR in the sub-threshold regime follow
a (modified) Schmidt law? Or is it triggered
entirely by local compression events?
• What is the physical nature of the threshold?
Questions: Thresholds
• The “star formation law” and “star formation rate” mean different
things on different linear scales
– disk-averaged scale (1-20 kpc, >100 Myr): useful empirical recipes, but
physical significance difficult to infer
– radial averages (~1 kpc, 20-300 Myr): breaks some parametric
degeneracies, but still smooths nonlinear phenomena over 10-100’s of cloud
scales
– cloud scale or below (50-500 pc, 3-10 Myr): probes “star formation
efficiency”, but with large observational scatter. “SFRs” really are measures
of cluster luminosities. SF law on larger scales may have very different
form.
Key Caveats, Considerations (aka Rant #1)
Rant #2: Beware of the local coincidence between
ISM phase vs gravitational instabilities
Log P/k (pressure)
Log
NH
grav bound
diffuse
• It is important to match the SF and gas tracers to the
application of interest
– color-magnitude diagrams: nirvana: range of ages, stellar masses
– H, P, 24m knots: massive stars, last 3-10 Myr
– FUV continuum: massive stars, last 0-200 Myr
– diffuse dust continuum emission, 20-200m + PAH emission:
massive and intermediate mass stars, last 0-2 Gyr
– CO, HI clumps: probably bound clouds, <10 Myr
– diffuse CO, HI: anybody’s guess, probably ~0.1-1 Gyr
Rant #3
Draine et al. 2007, astro-ph/0703213
GALEX FUV + NUV (1500/2500 A)
IRAC 8.0 m MIPS 24 m
H + R
M81
H + R
Calzetti et al. 2007, ApJ, 666, 870
M 81
24µm 70µm 160µm
PHOENIGS: from the SINGS `ashes’Project for Herschel On an Extragalactic Normal
Infrared Galaxies Survey
R. Kennicutt (IoA, Cambridge), L. Armus (SSC), A. Bollato (UMd), B. Brandl (Leiden), D. Calzetti (UMass), D. Dale (UWyo), B. Draine (Princeton), C. Engelbracht (UofA, USA), K. Gordon (UoA, USA), B. Groves (Leiden), L. Hunt (Oss Arcetri, Italy), J. Koda (Caltech), O. Krause, A. Leroy, H.W. Rix (MPIA), H. Roussel (IAP), M. Sauvage (CEA), E. Schinnerer (MPIA), J.D. Smith (Toledo), L. Vigroux (IAP), F. Walter (MPIA), M. Wolfire (UMd) + TBD
Broad Science Objectives:• Trace and characterize the flow of energy through the ISM in galaxies; • Link heaters-emitters: use Herschel spatial resolution to enable definitive modeling of radiative transfer of dust and gas cooling in galaxies;• Probe the nature/origin of extended cold dust envelopes; link warm-cold dust emission;• Improve dust and spectral diagnostics of star formation and ISM properties.
Approach:• An objectively selected sample of nearby galaxies (SINGS-inspired), optimized to cover a broad and representative range of properties, and broad range of local physical environments;• Exploit angular resolution for resolving infrared components and dust heating populations.• Leverage existing and new ancillary data: from UV to radio• Data and high-level data products would be delivered quickly to the broad community.
• Spatially resolved measurements, vs disk-
averaged or azimuthally-averaged data
• Extinction corrections (H, UV), and corrections
for unextincted star formation (infrared, radio)
• Accurate molecular mass measurements (how
reliable is CO?)
• IMF
Observational Wishlist
NGC 628(M74)
C. Tremonti
SINGS Galaxies
The Global Schmidt Law Revisited
• analyze galaxies with spatially-mapped star formation (H, P, FIR), HI, and CO
• enlarged, diversified samples – normal galaxy sample 3x larger– larger ranges in gas and SFR
densities– large subsamples of
circumnuclear starbursts, low-metallicity galaxies incorporated
• densities averaged within active SF regions
• explicit corrections for [NII], extinction
• point-by-point analysis of SINGS + BIMA SONG galaxies
Work in progress!
Kennicutt (1998) sample expanded sample constant extinction
expanded sample constant extinction H + 24m extinction corrections
PH
Hm
H
total FIR
HI + CO (const X)
HI + H2 HI H2
metal-poor dwarf galaxies
The Spatially Resolved Star Formation Law in M51 Kennicutt et al. 2007, ApJ, in press (astro-ph/0708.0922)
FUV, H, 24m
Calzetti et al. 2005, ApJ, 633, 871
- Use spatially-resolved measures of CO, HI, and SFR to characterize SFR vs gas surface density relation on a point-by-point basis
- Use combinations of H + P and H + 24 mm emission to correct for extinction in SFR measurements
- Probe scales from 300 - 1850 pc (IR/HII regions to unbiased sampling of the disk)
M51: BIMA SONG Survey Helfer et al. 2003
NGC 6946– THINGS VLA HI Survey F. Walter et al.
Local Schmidt Law in M51
Kennicutt et al. 2007
Tentative Conclusions
• The disk-averaged Schmidt law in galaxies is rooted in
a local relationship that persists to scales of <500 pc
• In M51 the SF density is tightly coupled to the local H2
surface density, and not with HI density
• A kinematic star formation law does not seem to
extend as well to local scales
• The disk-averaged SF law is confirmed with
more/better observations. Some metal-poor galaxies
lie systematically above the mean relation.
Tentative Conclusions
• The combination of H and 24 m imaging provides
a reliable method for obtaining extinction-corrected
ionizing fluxes of HII regions. The combination of
Ha and FIR luminosities can provide reliable
extinction-corrected SFRs of galaxies as a whole.
• As SFR estimators become more reliable, empirical
characterization of the SF law will become
increasingly limited by the accuracy and depth of
cold gas tracers, especially for molecular gas.
• Spitzer Infrared Nearby Galaxies Survey (SINGS) – resolved UV radio mapping of 75 galaxies
– selection: maximize diversity in type, mass, IR/optical
• 11 Mpc Ha/Ultraviolet Survey (11HUGS)– resolved H, UV imaging, integrated/resolved IR of 400 galaxies
– selection: volume-complete within 11 Mpc (S-Irr)
• Survey for Ionization in Neutral-Gas Galaxies (SINGG)– resolved H, UV imaging, integrated/resolved IR of 500 galaxies
– selection: HI-complete in 3 redshift slices
• Integrated Measurements– Ha flux catalogue (+IR, UV) for >3000 galaxies within 150 Mpc
- integrated spectra (+IR, UV) for ~600 galaxies in same volume (Moustakas & Kennicutt 2006, 2007)
Primary Datasets
• Beware the treacheries of correlation vs causation!
– For a Q ~ 1 disk: gas ~ c/G
so
gas
– Likewise: gas tot , so P 2
– Also, for local Galactic ISM pressures, crit for self-
gravitating clouds is approximately the same as crit
for self-shielding of molecular clouds
– And--- in SF regions much of HI may be a
photodissociation product of UV radiation on H2
Physical Origins of SF Law
Thanks to: S. Akiyama, J. Lee, C. Tremonti, J. Moustakas, C. Tremonti (Arizona), J. Funes (Vatican), S. Sakai (UCLA), L. van Zee (Indiana) + The SINGS Team: RCK, D. Calzetti, L. Armus, G. Bendo, C. Bot, J. Cannon, D. Dale, B. Draine, C. Engelbracht, K. Gordon, G. Helou, D. Hollenbach, T. Jarrett, S. Kendall, L. Kewley, C. Leitherer, A. Li, S. Malhotra, M. Meyer, E. Murphy, M. Regan, G. Rieke, M. Rieke, H. Roussel, K. Sheth, JD Smith, M. Thornley, F. Walter
normal galaxies
starburst galaxies
HI+H2 mass surface density
SFR
su
rface d
en
sit
y
Calzetti et al., ApJ, submittedKennicutt & Moustakas, in prep
HII regions galaxies (integrated fluxes)
Kennicutt 1998, ApJ, 498, 541
Martin et al. 2005, ApJ, 619, L59
Wang & Heckman 1996, ApJ, 547, 965
H + R
H + R