Molecular Gas in Dwarf Galaxies

1
Molecular Gas in Dwarf Galaxies
Adam Leroy UC Berkeley At Dwarf Galaxies as
Astrophysical and Cosmological Probes March 15,
2006
L. Blitz (UCB) A. Bolatto (UCB) E. Rosolowsky
(CfA) F. Walter (MPIA) J. Simon (Caltech) The
NANTEN Group at Nagoya Y. Fukui, A. Kawamura, N.
Mizuno
2
Astrophysical Probes of the effects of
environment on star formation
Nearby dwarf galaxies are ideal laboratories to
investigate how stars form out of gas in
environments deficient in heavy elements
1. In what dwarfs do we detect molecular gas? 2.
How does the molecular gas content of
dwarfscompare to that of large spirals? 3. Do
giant molecular clouds (GMCs) in dwarfsresemble
the GMCs found in large spiral galaxies? 4. What
does (and doesnt) CO tell us in lowmetallicity
systems?
CO emission with HI contours in IC 10
3
Which dwarfs do we see?
A survey looking for CO 1 ? 0 in 121 galaxies
small (vrot
s-1)detected by IRAS (actively star
forming) integration times 1 - 2 hours per target
ARO 12m
28 Detections Average intensity 1.1 K km s-1
16 Marginal Detections Average intensity 0.5 K km
s-1
77 Non-detections (67 shown) Average intensity 0.13 K km s-1
Leroy et al. (2005)
4
Which dwarfs do we see?
Median characteristics of a detection Distance
11 Mpc Hubble type Sc Irr Dynamical mass
1010 Msun LCO 2 107 K km s-1 pc2
(108 Msun at Galactic XCO) We marginally
detect CO stacking all the non-detections Noise
1 mK Integration time 100
hours Detection/Upper Limit 0.13 0.04 K km
s-1 Median dynamical mass 3 109 Msun Median
distance 11 Mpc - The LMC for us is likely a
marginal detection. - We can not detect the SMC
beyond 1 Mpc. - The galaxies we do detect are
more molecule rich than LG dwarfs.
Photo credit Fred Espenak
5
Down to LMC mass, molecular gas mass correlates
with many other galaxy properties
Magenta dots are dwarfs (with Vrot Blue dots are large spirals (with Vrot 100 km
s-1)
Young et al. (1995) Elfhag et al. (1996)
Taylor, Kobulnicky, Skillman (1998) Boker,
Lisenfeld, Schinnerer (2003) Leroy et al.
(2005)
6
Most of this behavior goes away if we normalize
by stellar mass (LK)
Magenta dots are dwarfs (with Vrot Blue dots are large spirals (with Vrot 100 km
s-1)
These dwarfs are just scaled down versions of
more massive spiral galaxies.
Data Young et al. (1995) Elfhag et al. (1996)
Taylor, Kobulnicky, Skillman (1998) Boker,
Lisenfeld, Schinnerer (2003) Leroy et al.
(2005)
7
Differences at very low mass
Local Group dwarfs and non-detections show much
less CO emission per stellar mass than more
massive systems. The molecular ISM in these very
low mass dwarfs is not just a scaled down version
of that found in spirals. With better
sensitivity, this plot should show a clear trend
with mass. Non-detections 70 galaxies. Marginal
detections 16 galaxies.
8
Overwhelmingly atomic ISMs
The ratio of molecular to atomic gas in dwarfs is
much lower than in large spirals.
q.v. Young and Scoville (1991)
9
Efficiency of Star Formation from H2

• - Includes nearby systems (H? or FIR).
• - Blue and magenta SFRs from radio continuum
  (NVSS)
• RC includes corrections to the radio continuum
  SFR (Bell 2003) based on size (to pin to FIR
  rate).
• XCO still Galactic (2 1020)
• Small galaxies form more stars per H2?
• Some may be varying calibrations (SFR too high,
  H2 too low) is all of it?
• See Albertos Talk

after Murgia et al. (2002)
10
Giant Molecular Clouds in dwarfs...

• In nearby dwarfs, CO resolves into GMCs
• Found on HI filaments/peaks.
• Sizes a few 10s of parsecs.
• Masses 104 Msun or more
• Are dwarf GMCs the same as spiral GMCs?
• Size, line width, and luminosity for each GMC.
• Correct for biases from sensitivity
  resolution.
• - Decompose emission self-consistently.

GMCs on the HI in IC 10 (HI from Wilcots Miller
1998)
11
Care in Decomposing N83 in the SMC at three
different resolutions 60, 20, and 10 parsecs.
(NANTEN, SEST) Our approach try to find robust,
physically meaningful local maxes. Key on
separation (physical and resolution units),
significance, and discontinuities in cloud
properties.
12
Correcting for Sensitivity Biases Finite
sensitivity leads to an underestimate of cloud
properties. Correct via an extrapolation to
perfect sensitivitiy.
13
How good a job do we do? Performance
recovering radius
14
How good a job do we do? Performance
recovering line width
15
Dwarf GMCs look Galactic

• The Luminosity-Line Width Relation
• GMCs need not be resolved.- Other galaxies
  show the same relation as inner Milky Way GMCs
  (Solomon et al. 1987).
• - A single population of GMCs?
• - Magenta points are BIMA/OVRO observations of
  galaxies from first half of talk (NGC 2976, 3077,
  4214, 4449, 4605)

Data Methods Blitz et al. (2006), Bolatto et
al. (2006, in prep), Leroy et al. (2006),
Rosolowsky (2006), Rosolowsky and Leroy (2006),
Rosolowsky et al. (2003), Walter et al.
(2001,2003a,2003b), Mizuno et al. (2001, 2006),
Fukui et al. (2001)
16
GMC mass function

• The GMC mass function
• Fraction of clouds above a given mass as a
  function of mass.
• - Normalized to the same value at 2 105 Msun.
• - Include small corrections to X by galaxy from
  virial masses.
• - Slope agrees across the Local Group (M 33 is
  slightly steeper) -1.75.
• Agreement includes IC 10, NGC 2976, and the LMC.
  Not enough clouds in the SMC.

17
Differences in GMC populations

• Luminosity vs. Radius
• Real differences in surface brightness galaxy by
  galaxy.
• No straightforward trend with metallicity or
  mass.

18
Differences in GMC populations

• Line width vs. Radius
• The Magellanic clouds lie slightly below other
  systems
• Lower density clouds?
• - Lower turbulent energy density?

19
How does CO trace H2?
Luminosity vs. Dynamical Mass (Virial Theorem,
Mdyn 1040 R ?v2) Galaxy XCO 1020 cm-2 (K
km s-1)-1 IC 10 3.0 LMC 5.0 SMC 9.5 Survey
Dwarfs 1.7 M33 3.1 M31 3.9 Based on virial
mass calculations, variations in XCO are small
(though present).
20
CO-to-H2 variations explain differences?

• Luminosity vs. Radius
• Considerably lower scatter after XCO correction.
• - Differences between galaxies, though small,
  remain.

21
But is CO telling us the whole story?
-
H2
FIR SED (MIPS, IRAS, ISO)
Atomic Gas Column(Stanimirovic et al. 2004)
Dust opacity from fitting the FIR SED.

• Use FIR and HI to infer molecular gas columns
  (Israel 1997)
• Measure the dust opacity along each line of sight
  from FIR SED fitting.
• Calibrate the dust (opacity) to gas ratio away
  from the molecular gas (but still near enough to
  be applicable) using HI and the FIR emission.
• Near the molecular gas (CO emission) infer a
  total column from the FIR and D/G ratio and
  subtract the observed (HI) column. Assume that
  the balance is H2.

Bolatto et al. (2006)
22
FIR and CO Yield Different Results
No CO,measure D/G here

• Calibrate the D/G ratio locally.
• - Resulting gas columns, less the HI, imply
• XSMC 4 1021 cm-2 (K km s-1)-1
• - This is 20 times the Galactic value 4 times
  the result from GMC properties. Uncertain by a
  factor of 3.
• FIR-based approaches consistently find evidence
  for more molecular gas than CO-based ones.
• (e.g. Israel 1997, Stanimirovic 2000)

CO emission, apply D/G to measure XCO
Our largest aperture on the SMC HI map by
Stanimirovic et al. (2004).
23
FIR and CO Yield Different Results
- H2 without CO due to diminished dust shielding
in low metallicity environments? - FIR (Israel
1997) and C (e.g. Madden et al. 1997) support
the idea of an envelope of H2 without CO (at
higher T?). - What is the significance of this
envelope if star formation is occurring within a
(pretty much) Galactic-looking GMC at the center?
Z
Maloney Black (1988), Pak et al. (1998),
Bolatto et al. (1999)
24
Conclusions

• 1. In what dwarfs do we detect molecular
  gas?
• Detecting CO in a galaxy like the SMC beyond the
  Local Group is very challenging at present,
  observations are limited to nearby low mass
  systems (still difficult) or molecule rich LMCs
  out to 10 Mpc.
• 2. How does the molecular gas content of dwarfs
  compare to that of large spirals?
• Below stellar masses around 109 Msun, dwarfs show
  much less CO emission relative to their other
  properties. Above this mass most galaxies look
  like scaled down spirals.
• Small dwarfs show much higher star formation
  rates for a given CO surface brightness.
• 3. Do giant molecular clouds (GMCs) in dwarfs
  resemble the GMCs found in large spiral galaxies?
• GMCs from Local Group galaxies fall on a single
  luminosity-line width relation and have similar
  GMC mass distributions. There are small but real
  deviations about GMC property scaling laws.
• The SMC, the lowest metallicity galaxy with
  studied GMC properties, shows the most deviation
  from scaling laws. It deviates in the sense of
  having large, low density, low luminosity clouds.
• 4. What does (and doesnt) CO tell us in low
  metallicity systems?
• Estimates of the CO-to-H2 conversion factor from
  dynamical (virial) masses and the FIR disagree.
  The former find values similar to Galactic, while
  FIR estimates yield much higher conversion
  factors.

25
What to look forward to

• CARMA ALMA
• Lots of collecting area
• -SMCs beyond the Local Group.
• - Thorough mapping of the rest of the LG dIrrs?
• - Other lines high density/optically thin
  tracers!
• Fantastic resolution
• - GMC properties to the distance of Virgo.
• - How do the size-line width relation and the GMC
  mass spectrum vary with environment?
• How does the molecular fraction found in dwarfs
  vary with environment?
• Spitzer, ASTRO-F? SOFIA ?
• - Flesh out the FIR side of nearby GMCs
  properties (and map them in C, C). Resolved
  comparisons with molecular line emission.

CARMA
1 ALMA dish