Transcript Part2

Nearby Galaxies
(mostly) at mm and IR wavelengths
Adam Leroy (MPIA Heidelberg)
Christof Buchbender (IRAM Granada)
Topics
A Broad Look at Nearby Galaxies
Observing Nearby Galaxies at Millimeter Wavelengths
Mapping nearby galaxies with the IRAM 30m
Working Group:
Mapping the bulk distribution of molecular gas in a bright nearby spiral galaxy.
Observing Nearby Galaxies at mm s
Goal: An overview of what you can observe in nearby galaxies at millimeter
wavelengths. Discuss typical intensities from several perspectives.
Specific Topics:
o What you see studying galaxies at mm wavelengths.
o Intensities at extragalactic distances: clouds, galaxies, chunks of galaxies.
o Millimeter continuum: origin and use (very briefly).
Apologies in advance (but not really): I’m going to be very “CO-centric” (though not “COexclusive”). The reason, which I hope you’ll appreciate is that CO - by far the brightest mm line is already very sensitivity limited at extragalactic distances.
What There Isn’t To See In The Milky Way: H2
H2 molecule lacks a dipole moment and most H2 is too cold to excite the
lowest rotational transitions. This makes tracing the star-forming part of the
interstellar medium a challenge.
H2
UV absorption studies: great! But requires background source, probe pencil beam.
Rotational line emission (IRS), sensitive only to gas down to T~100K (most H2 much colder).
 rays
Cosmic ray hits nucleus, produces  ray… modeled to yield total nuclei column density.
Major resolution and sensitivity challenge beyond Milky Way, modeling complex…
Dust
Probes total gas column modulo dust-to-gas ratio.
Absorption tough because of lack of clear background screen.
Emission limited by finite resolution of IR telescopes.
CO
Next most common molecule after H2.
Standard tracer of H2 in high-mass galaxies.
At low metallicities C and O less abundant and dust shielding weak. CO suppressed?
Comes with velocity information - dynamics offer another way to trace mass.
What There Is To See In The Milky Way: CO
Short answer for mm line work: Giant Molecular Clouds
Dame+ 01
Galactic Ring Survey Jackson+ 06
Orion Molecular Cloud
Wilson+ 05
(Giant) Molecular Clouds
Key properties of molecular clouds*:
o Distinct (in real and P-P-V space) dense clouds containing a
significant mass of H2.
o Clumpy (multiscale structure / cloud-clump-core) and turbulent.
o High mass (giant) clouds gravitationally bound, low mass clouds
confined by external pressure.
o Sites of all present-day star formation.
o Power law mass spectrum, dN/dM  M-1.5 N (M > M)  M-0.5
(top-heavy, but maybe variable across the Local Group)
o Obey a set of basic scaling relations (“Larson’s Laws”)
* Reasonably established, but there are open questions on each point here.
Review: Blitz ‘93
What There Is To See In Nearby Galaxies
M33 (Engargiola+ 03, BIMA)
LMC (Fukui+ 99, NANTEN)
Color: Atomic gas
Color: Stars and H
Green dots: Bright CO
Contours: CO
What There Is To See In Nearby Galaxies
M31 (Nieten+ 06; IRAM 30m)
Left: CO from Andromeda
Giant Molecular Clouds in Galaxies
o Typical sizes: few 10s of parsecs
o Line widths: few km/s
o Surface density (brightness): ~100 Msun pc-2 (10-20 K km s-1)
o Mass: ~105 to few times 106 Msun
Line Width [km s-1]
Virial Mass 2 R [Msun]
o To first order, cloud in other galaxies look like Milky Way GMCs:
Radius [parsecs]
Bolatto+ ‘08
CO Luminosity
Starbursts and Associations of GMCs
Resolution and sensitivity matched to a single GMC are largely limited to
the Local Group (especially for a single dish telescope).
Aalto+ ‘99: collections of GMCs in M51
On various scales, most mapping of
other galaxies relies on getting several
clouds per beam.
Shen & Lo ‘95 (M82, left), Roslowsky & Blitz (M64,
right); Wilson+ (Antennae, bottom)
Starbursts show GMC-like properties (or more
extreme) over scales of few hundred pc to kpc.
No longer analogs to collections of MW GMCs.
Well Beyond CO
The low rotational transitions of 12CO are the best available simple tracer of
distribution of H2 mass. More sophisticated mm spectroscopy can reveal
physical conditions (density, temperature) and refine estimates of H2.
R. Genzel (1991)
Well Beyond CO
Martin+ ‘06: IRAM 30m
NGC 253 (nearby starburst)
111 lines, 25 species
When things are bright and dense enough, there’s lots of information there! (see S. G-B)
But CO is Still Your Benchmark / Starting Point
Pointing towards the bright part of a nearby GMC:
12CO
J=2-1 : Tpeak ~ 110 K
Next brightest line:
13CO
J=2-1 : Tpeak ~ 35 K
Other lines down by larger factors.
…
The same thing is true in galaxies:
o 12CO J=2-1 (230 GHz) and J=1-0 (115 GHz) are
the brightest lines.
o Typically this is followed by the same transitions in
13CO (lower ; intensity down by a factor of ~6 in
the Milky Way).
o Then HCN, HCO+, CN, HNC, CS, C18O, etc.
o Intensities for these species seldom much higher
than 1/10 CO.
Sutton+ ‘85: 1mm line survey of Orion A
Example: HCN in Nearby Galaxies
log10 HCN to 12CO Line Ratio
Observations towards ~
kpc2 regions in the disks
of nearby, massive,
star-forming galaxies.
Courtesy A. Usero (in prep.)
The numbers on the
y-axis are low!
Line ratio in a normal
galaxy disk ~ 0.03.
log10 12CO Intensity [K km s-1]
What’s Been Done (very roughly)
Millimeter line work on nearby galaxies:
o CO pointings or sampling towards ~500 - 1000 galaxies.
o Complete/resolved CO maps of ~100 nearby galaxies.
o HCN/HCO+ pointings towards ~100 nearby galaxy centers.
o HCN/HCO+ maps of a handful of nearby galaxies.
o Multitransition studies of a handful of nearby galaxy centers.
o CO maps able to identify individual GMCs in a handful of galaxies.
Looking at Other Galaxies: Numbers
For the next few slides we’ll look at other the intensity of other galaxies as
observed with the 30m. We’ll do this in three ways:
o In terms of individual molecular clouds …
o In terms of integrated molecular mass (luminosity) …
o In terms of surface density …
We’ll focus on the 12CO J=1-0 transition. As we’ve just discussed, this (and
the corresponding J=2-1 transition) is the brightest available transition by a
factor of several.
Looking at Other Galaxies: Units
Unit
What Does It Measure?
Notes
K
brightness temperature
(a way to phrase
specific intensity)
• be careful when switching to and from Jy
beam-1 (factor of 2)!
integrated intensity
• brightness temp. integrated over velocity.
• with XCO yields a surface density.
• usually quoted as an average over a
telescope beam.
flux
• brightness temp. integrated over velocity
and angular area.
• independent of telescope beam.
luminosity
• brightness temp. integrated over velocity
and physical area
• with XCO yields an integrated molecular mass.
K km/s
K km/s
arcsec2
K km/s
pc2
Looking at Other Galaxies: Conversion Factors
XCO : assume a linear conversion between CO and H2 to hold on large scales.
This “X-factor” is the conversion.
Milky Way value ( rays, dust emission, virial mass):
XCO = 1.5 - 3  1020 cm-2 (K km s-1)-1
But be careful! XCO is an approximation. It does not hold perfectly within
Galactic clouds and different values are used in starburst and dwarf galaxies.
Using XCO and the units just described, we can then convert:
o Integrated Intensity [K km/s] to Surface Density [Msun pc-2]
H2 Msun


X co
pc  3.2 I co K km s  

21020 cm-2 (K km s-1 )-1 
-2
-1
o Luminosity [K km/s pc2] into Molecular Mass [Msun]



X co
M H2 Msun   3.2 Lco K km s pc  

21020 cm-2 (K km s-1 )-1 
-1
2
N.B., this is only H; apply another factor of ~ 1.36 to include He.
Looking at Other Galaxies: GMC Perspective
How bright are giant molecular clouds in other galaxies?
What is the line-average intensity (in K) you expect pointing the 30m at
a GMC in a nearby galaxy?
1. Luminosity of a GMC
K km s-1 pc2
2. Line width of a GMC
km s-1
3. Size of a cloud (or the beam)
pc2
Line-average intensity
K
Looking at Other Galaxies: GMC Perspective
Luminosity of Molecular Clouds
Highest mass Milky Way GMCs
~ 106 K km s-1 pc2
High Mass SF Region (like Orion)
~ 105 K km s-1 pc2
Low Mass SF Region (Like Taurus)
~ 104 K km s-1 pc2
Low mass, pressure-confined cloud
~ 103 K km s-1 pc2
Line Width [km s-1]
Line width (FWHM) ~ 3 - 20 km s-1
v varies systematically with mass.
For this exercise we’ll take
vFWHM ~ 10 km s-1
for all cases.
Radius [parsecs]
(N.B., plot at left is using RMS line width)
Looking at Other Galaxies: GMC Perspective
Physical area being integrated over:
o GMC area (via mapping) if bigger than telescope beam.
o Telescope beam if bigger than cloud size.
Angular area of 30m beam ~ 520 / 115 GHz2 arcsec2
In parsecs2 vs. distance:

  d 
2
4 115 GHz
Abeam pc   1.210 
 

 
 1 Mpc
2
2
Typical cloud size ~ 30 pc
ULIRGS

Nearest
Cluster
Other
Groups
Local
Group
Area of a typical GMC
Area ~ 103 pc2
30m beam relevant scale (i.e.
clouds unresolved) beyond Local
Group.
Different case for interferometers.
Looking at Other Galaxies: GMC Perspective
Take line width ~ 10 km s-1 for all cases.
Take cloud luminosity.
Take 30m physical beam size (function of distance).
2
Average Intensity Across Line [K]
I CO

 d   v 1
L co
K  0.8  5
 
-1
2  
-1 
10 K km s pc 1 Mpc 10 km s 
1 hour ON source with EMIR
(single pixel 30m receiver)
at 115 GHz, 10 km s-1 channel:

~3 in 1h ON
 ~ 3 mK
Recall radiometer formula:
 t-0.5
Ref. + overheads add x 2-4
Looking at Other Galaxies: GMC Perspective
Some things to take away:
o Individual CO-emitting structures are faint at extragalactic distances.
o Individual clouds only resolved inside the Local Group by the 30m.
o Individual low mass clouds only detectable inside Local Group.
o Individual high mass clouds detectable with effort in nearest other groups.
o Observing distant galaxies requires averaging many clouds inside a beam
(that’s okay, beam has large spatial area too).
We’ve used 12CO as our example, scale to get other lines.
Looking at Other Galaxies: Galaxy Perspective
How bright are whole galaxies in CO?
Spatial size of the 30m beam gets very big at extragalactic distances.
It is easy to find yourself with most of a galaxy inside a beam.
In this case, the relevant thing is the CO luminosity of the galaxy, LCO.
How to make a good guess at LCO?
o We’ve already seen that many galaxy properties are strongly covariant
with one another. This include CO luminosity.
o Both stellar luminosity and infrared luminosity (or some other tracer of star
formation rate) are good places to start.
CO per Stellar Luminosty (B-band)
Looking at Other Galaxies: Galaxy Perspective
Dots: relatively massive
star-forming galaxies.
Stellar Luminosity [Magnitudes]
For relatively massive (M* > 1010 Msun) star-forming (blue) galaxies, CO per
unit starlight is relatively fixed.
There are important exceptions to this, especially
o low metallicity dwarfs
o gas-poor ellipticals
o CO-bright mergers
… but it remains a good starting point.
Star Formation Rate [from IR]
Looking at Other Galaxies: Galaxy Perspective
Gao & Solomon ‘04
H2 Mass from CO
For actively star-forming galaxies, CO is clearly related to star formation
(above traced by IR, but this works for other tracers, too).
Details vary a bit:
o In normal disks, a fixed CO-to-IR ratio is probably pretty much OK.
o Very vigorous star-formers have less CO per IR.
o Quiescent and low-metal galaxies may also have less CO per IR.
Looking at Other Galaxies: Galaxy Perspective
A simple scaling from the Milky Way is often a good place to start. From Allen’s:
Property
Milky Way Value
H2 Mass
(CO Luminosity)
2-6 109 Msun
(0.5 - 1.4 109 K km s-1 pc2)
Star Formation Rate
~ 3 Msun yr-1
Stellar Mass
(B band Luminosity)
4-8 1010 Msun
(2.3 1010 Lsun)
Then (very) roughly:
o H2-per-SFR ~ 1-2 Gyr (nearby spirals also yield ~ 2 Gyr)
o H2-per-Stellar Mass ~ 0.03 - 0.1
Looking at Other Galaxies: Galaxy Perspective
Let’s run the same exercise we did with the clouds: LCO  <ICO>
o take H2-per-Stellar Mass ~ 0.05
o take XCO = 2 1020 cm-2 (K km s-1)-1
o assume the galaxy fits inside the 30m beam.
o assume FWHM = 200 km s-1*
Average Intensity Across Line [K]
* A (much) better way is to use TF and inclination to estimate line width.
1 hour ON source with EMIR
(single pixel 30m receiver)
at 115 GHz, 200 km s-1 channel:
 ~ 0.7 mK
~3 in 1h ON
Recall radiometer formula:
 t-0.5
Ref. + overheads add x 2-4
Distance [Mpc]
Looking at Other Galaxies: Galaxy Perspective
Some things to take away:
o When all CO emission from a galaxy is in the beam, the 30m can detect
galaxies out to a large distance, though the time investment is still not trivial.
o We have ignored metallicity effects, enhancement due to intereactions.
In reality, low mass galaxies are hard to see at all and mergers can be very
bright.
o Unlike GMCs, it isn’t always straightforward to get all of the galaxy inside
the beam. We’ll talk a bit about the intermediate case next…
We’ve used 12CO as our example, scale to get other lines.
Looking at Other Galaxies: Surface Density
Often, especially when mapping, you are in the intermediate regime:
Many GMCs per beam, but not a sizable piece of the galaxy.
Relevant quantity here is H2 surface density or CO surface brightness.
We’ve talk about how we might estimate the CO luminosity of a galaxy, but how
is this luminosity distributed?
HERA (30m) maps of some normal spiral galaxies at ~10 Mpc. The spatial resolution is a kpc and
the mass in each resolution element is more than 106 Msun, implying collections of GMCs.
Looking at Other Galaxies: Surface Density
A typical CO scale length in a
massive star forming galaxy (like the
Milky Way)
Size of a typical Milky Way Giant
Molecular Cloud
We’re in this regime from just beyond the Local Group (can see clouds in SMC,
LMC, M33, M31) to several 10s of Mpc (past which even big galaxies are
essentially point sources).
Looking at Other Galaxies: Surface Density
In star-forming galaxies, CO surface
brightness varies strongly with radius.
Right:
o Profiles of integrated intensity (black
points) vs. galactocentric radius for 7
galaxies mapped by HERA on the 30m.
o Lines show exponential fits.
o Higher (gray) profile shows peak
surface brightness in the ring.
Azimuthally averaged variations in CO
emission tend to track those in stellar surface
density or star formation.
Scale Length At Other  [kpc]
Typical CO scale length ~ 0.2 r25.
CO Scale Length [kpc]
CO Intensity (points) and Stellar Profile (line) + arbitrary offset
Looking at Other Galaxies: Surface Density
Regan+ 01
Radius
Looking at Other Galaxies: Surface Density
Recent Star Formation (IR or RC)
Kiloparsec-scale averages of CO emission tend to track IR emission or other tracers of recent
star formation. For a local guess of the CO surface brightness this is probably a reasonable
place to start (but beware metallicity issues).
CO Surface Brightness
HERA 30m (color, Bigiel+ ‘08) + Literature (gray, Young+ ‘95, Elfhag+ ‘96, Murgia+ ‘02, Leroy+ ‘05)
Looking at Other Galaxies: Surface Density
Some things to take away:
o Azimuthally averaged, CO emission looks pretty similar to stars (there are
important differences, but this is a good place to start): an exponential
decline with a scale length ~0.2 to 0.25 times the optical radius.
o The exponential decline is a mix of filling factor (e.g., arms vs empty
space) and decline in the peak integrated intensity (arms get fainter).
o The local star formation rate or IR surface brightness are reasonable ways
to guess at the surface brightness of CO on fairly large (~ kpc) scales.
We’ve used 12CO as our example, scale to get other lines or use the wellestablished HCN-IR (or an analogous HCO+-IR) relations.
Millimeter Continuum
We’ve focused on lines, what about the continuum at 1-3mm?
(though many 30m observing modes deliberately filter out continuum via references).
o 3mm is a low point in the SED of a (non-”monster”) galaxy. Emission from longer than ~ 1mm
makes up < 10-4 of the bolometric luminosity.
o Dust emission dominates at shorter : R-J tail + -dependent emissivity
o Synchrotron dominates at longer : declines with Fn ~ -0.8
o Thermal free free “fills in the gap”: emission from ionized gas, p-e collisions
M82
(Difference is nu*Fnu on left vs. Fnu on right)
Condon ‘92
Dust Continuum
Dust emission dominates at shorter : R-J tail + -dependent emissivity
2 2kT 
F   2 1 e   with   -
 c 
o usually (very) optically thin
o=1-2
o know/estimate T and  (the mass absorption coefficient relating  to mass) can get dust mass.
o more points: sophisticated fits to models or constrain , , T simultaneously

The Sombrero galaxy at optical (right/bottom) and millimeter wavelengths (top left: LABOCA 870 mm,
top right: MAMBO2/30m 1.2mm). Millimeter continuum traces dust, getting notably weaker with
wavelength (RJ tail + emissivity). Vlahakis+ 2008.
Dust Continuum
Dust emission utility:
o dust and gas well mixed: dust is an optically thin, relatively robust tracer of the ISM
o dust SED fueled by young star formation (mostly), estimate recent SF (better near 60 or 100 m)
o disentangling dust temperature and mass allows estimate of heating radiation field
o dust to gas ratio relevant to many aspects of star formation and interpreting observations
The Sombrero galaxy at optical (right/bottom) and millimeter wavelengths (top left: LABOCA 870 mm,
top right: MAMBO2/30m 1.2mm). Millimeter continuum traces dust, getting notably weaker with
wavelength (RJ tail + emissivity). Vlahakis+ 2008.

Free Free Continuum
Thermal free free emission (electrons accelerated by protons):
Temperature of electrons in the emitting (HII) region
Observed frequency
1.35
2.1
4




10
K
1
GHz
  3.3 10-7 

 
 Te    
 n n dl
e
p
TB  Te 1 e  
Observed brightness temperature.
Emission measure: integrated
density squared along the line of
sight ( # of recombinations).
(N.B. these equations already assume RJ approximation and h < kTe; okay for mm …
but there is a cutoff in the spectrum once h ~ kTe).
See Tielens ‘05
Millimeter Continuum
Physical information from thermal free free emission:
o Estimate the rate of ionizing photons hitting the region producing the emission.
o Emission measure  number of recombinations, so…
-0.45
N uv 


Te
52
 -1  6.310  4 
 s 
10 K 

  0.1 
Lff

  20
-1 
1 GHz  10 W Hz 
(Note the unit change on Lff!)
… why is this interesting? Ionizing photons are produced only by young stars, making
this an effective, easily interpreted tracer of recent star formation (in practice this is
slightly easier at longer wavelength than in the millimeter, e.g. K band).
See Condon ‘92
Wrap Up
1.
Millimeter spectroscopy of other galaxies:
• Best available tool to study distribution of H2.
• CO strongest line by several factors (ncrit, abundance).
• Weaker lines give physical conditions (or fraction of dense, excited gas).
2.
Sensitivity calculations / feasibility estimates:
• through the lens of GMCs.
• through the lens of H2 surface densities.
• for whole galaxies.
Optically thinner, high excitation, high ncrit lines: down by factor 5-40+ …
3.
Continuum:
• mixture of free-free and dust. Weak, often filtered out by design. Carries
info about ionizing radiation, dust temperature and dust mass (AW).