Giant Molecular Clouds and Star Formation in Dwarf Galaxies

1
Giant Molecular Clouds and Star Formation in
Dwarf Galaxies
Leo Blitz Alberto Bolatto Greg Engargiola Adam
Leroy Dick Plambeck UC Berkeley
Yasuo Fukui Akiko Kawamura Norikazu
Mizuno Nagoya University
Erik Rosolowsky Center for Astrophysics
Josh Simon Caltech
Tony Wong ATNF
2
Comparison of Millenium simulation with Data
Springl White (2006)
3
Questions
How can we understand the star formation history
of a heirarchical Universe?
Star formation must have always occurred in
molecular clouds (except for primordial black
holes) because stellar mass objects require
densities and temperatures only found in
molecular clouds.
Steidel et al. 1999
4
Questions
We wish to extrapolate what we know about local
star formation to the early Universe and thus
interested in dwarfs and starbursts.

• How might star formation in primeval galaxies be
  different from today? Effect of metallicity on
  CO/H2?
• What determines the formation of GMCs (and
  therefore stars) in dwarfs?
• Why is CO used as a star formation reservoir when
  it has nothing to do with star formation?

5
CARMA
6
CARMA
Lets answer last question first Does CO have
anthing to do with star formation?
7
E Lada (1992)
CS 2-1
2µm sources
8
98 of the embedded sources (and thus the star
formation) in the Orion B cloud are in the red
areas.
Most of the CO (gt 90) is inert.
9
Does CO have anything to do with star formation?
Starbursts and ULIRGS
Gao Solomon 2004
SFR vs. HCN
SFR vs. CO
Green line has a slope of 1.44
10
Does CO have anything to do with star formation?
Gao Solomon (2004)
Wu et al. (2006)
SFR vs. HCN
SFR vs. CO
Green line has a slope of 1.44
11
Conclusion

• CO does indeed have little directly to do with
  star formation, but it is well-correlated with
  the dense SF gas that does
• Thus, if we trace H2 with HCN, star formation
  goes S(H2)1.0
• If we trace H2 with CO, star formation goes as
  S(H2)1.4
• This appears to be the origin of the 1.4 power
  law index in the Kennicutt-Schmidt Law, since K
  used CO as a tracer of H2

12
So, how do we determine how much molecular gas
you have? Kennicutts Paradox
SSFR 0.16 Sgas1.4
Kennicutt 1998
13
The Role of Pressure in GMC Formation

• Lets assume that
• S(H2)/S(HI) f (Pext) only
• PEXT (2G)0.5Sgvgr0.5 ((p/4) rg)0.5
• but, almost everywhere, r gtgt rg
• PEXT 0.84(GS)0.5Sgvgh-0.5
• but, vg and h are constant in disk galaxies
• PEXT g (SSg) only S(H2)/S(HI) f (SSg)

14
Pressure vs. H2/HI
PEXT 0.84(GS)0.5Sgvgh-0.5 but, vg and h are
constant in disk galaxies
15
Pressure vs. H2/HI
S(H2)/S(HI) fmol 2.9 x10-5 (Pext/k)0.92
16
This means that if you know the total gas
content, you can determine the molecular gas
fraction, and from that determine the star
formation rate. Therefore one can derive a star
formation prescription based on the molecular gas
only.
17
Star Formation Prescription

• Using the relation of Gao Solomon with FIR as a
  SF tracer
• Kennicutt relation
• SSFR 0.16 Sg1.4

18
We know a lot about large spiral galaxies
NGC 5055
19
We know a lot about large spiral galaxies

• We know that large spirals are disks with v/s
  20 40. The gas is concentrated in thin disks.
• Molecular clouds, are often concentrated in
  spiral arms and thus star formation takes place
  there, too.
• Are the GMCs and therefore star formation
  different in dwarfs? Dwarf spirals vs. DIrr
  DSph?

20
HI in four Local Group dIrrs
Are these face-on galaxies?
21
HI in four Local Group dIrrs
Are these face-on galaxies?
22
LGS 3
Young Lo 1997
23
(No Transcript)
24
V/s 1 Dispersion is gtgt than in large, face-on
spirals
25
(No Transcript)
26
V/s 1
27
Velocity dispersions are constant with radius out
to large R.
28
Conclusions

• DIrrs can have very different, even spherical HI
  distributions.
• V/s can be 1, or less not rotationally
  supported
• Dispersion is independent of radius out to large
  distances implies that these galaxies are not
  embedded in halos with larger VC.

29
Approach
Does metallicity affect molecular cloud
properties? Star formation?

• Berkeley, Nagoya have the first four completely
  mapped galaxies with enough resolution to
  determine GMC properties

30
CO in the Large Magellanic Cloud
Fukui et al. (2001, 2006)
31
The GMCs in M33
Ha Cheng et al. (1993)
CO Engargiola, et al. (2003)
32
CO in the Small Magellanic Cloud
Mizuno et al. 2001
33
CO on Optical Image of IC10
Leroy et al. (2006)
34
CO on HI in the LMC
Total molecular mass (10 of HI) 7107 M?
HI Kim et al. (1998), CO Fukui et al. (2001)
35
CO on HI in M33
Engargiola et al. (2002)
36
CO on HI in the SMC
Mizuno et al. (2006)
37
CO on HI in IC10
Leroy et al. (2006)
38
(No Transcript)
39
Conclusions

• CO forms on HI filaments
• How much H2 forms is determined by the pressure
• Filaments form first (filaments exist even where
  there is no CO)
• Filaments sometimes formed by gravity, sometimes
  by explosions

40
Now, lets look at the resolved cloud population

• Molecular Clouds all reduced the same way
• Takes account of different sensitivities and
  beamwidths used in different surveys
• Uses the same algorithm to determine radii,
  velocity dispersions, masses, etc. for all of the
  clouds
• Would like to see if metallicity affects derived
  H2 mass. i.e. X factor, CO/H2 conversion

41
X-factor in M33
42
Viral vs. Luminosity
X factor
43
Maximum Disk Fit
How much mass does the gas add?
NGC 2976
HI
H2
Simon et al. 2005

• Even with no disk, dark
• halo density profile is
• r(r) 1.2 r -0.27 0.09 M?/pc3

44
Maximum Disk Fit
NGC 2976
Simon et al. 2005

• Maximal disk M/LK
• 0.19 M?/L?,K
• After subtracting stellar
• disk, dark halo structure is
• r(r) 0.1 r -0.01 0.12M?/pc3

45
Answers

• Why is CO used as a star formation reservoir when
  it has nothing to do with star formation?
• Because there is a good, tight relation between
  CO and HCN, and because HCN is a linear tracer of
  star formation over many orders of magnitude in
  LHCN.

46
Answers

• What determines the formation of molecular clouds
  (and therefore stars) globally in both dwarfs and
  large spirals?
• Apparently it is hydrostatic pressure. Also
  explains e.g. M31, H2 in ellipticals.
• We keep in mind that stars form from molecular
  gas only. We can even explain the origin of the
  1.4 exponent in the Schmidt SF law. We cant
  know yet whether pressure is just a substitute
  for density.

47
Answers

• What is the affect of metallicity in determining
  GMC properties, and use of CO to get H2 mass?
• This is unclear at present. Strong effect of
  metallicity on CO luminosity seen only in SMC.
  Apparently, there is not a one-one relation of
  metallicity on CO as a function of H2 mass.