1. Molecules in galaxies at all redshifts
1. How to observe the H2 component?
2. Molecular component of the Milky Way
3. Fractal Structure
4. Formation of the fractal, shear, turbulence
5. H2 in external galaxies
6. H2 in ULIRGs, Dense tracers
7. Molecules in absorption
8. CO at high redshift
9. Primordial H2, history

2. How to observe the H2 component?
SAAS-FEE Lecture 1
Françoise COMBES

3. The H2 molecule Symmetrical, no dipole Quadrupolar transitions ΔJ = +2
Light molecule => low inertial moment and high energy levels
Para (even J) and ortho (odd J) molecules (behave as two different species)

4. H2 is the most stable form of hydrogen at low T
dominant in planetary atmospheres?
Formation: on dust grains at 10K
However formation still possible in primordial gas
(H + H- Palla et al 1983)
Destruction: through UV photons (Ly band)
Shielded by HI, since the photodissociation continuum starts at
14.7eV, and photo-ionization at 15.6 eV
(HI ionization at 13.6 eV)
Self-shielding from low column densities
1020 cm-2 in standard UV field
H2 will be present, while other molecules such as CO
would be already photo-dissociated

5. Potential curves involved in the Lyman and Werner bands (Roueff 00)

6. Ortho-Para transitions?
Formation in the para state not obvious
Large energy of formation 2.25 eV/atom
ortho-para conversion in collisions H++H2
n(O)/n(P) = 9.35 exp(-170/T)
Anormal ratios observed (ISO)
IR lines J=2-1 at 42 μ, 1-0 at 84 μ ?
A = cm3/s (Black & Dalgarno 1976)

7. Infrared Lines of H2 Ground state, with ISO (28, 17, 12, 9μ)
S(0), S(1), S(2), S(3)
From the ground, 2.2 μ, v=1-0 S(1)
excitation by shocks, SN, outflows
or UV-pumping in starbursts, X-ray, AGN
require T > 2000K, nH2 > 104cm-3
exceptional merger N6240: 0.01% of L in the 2.2 μ line (all vib lines 0.1%?)

8. H2 distribution in NGC891 (Valentijn, van der Werf 1999)

9. NGC 891, Pure rotational H2 lines S(0) & S(1)

10. H2 v=1-0 S(1) 2.15μ in NGC 6240
van der Werf et al (2000) HST

12. UV Lines of H2 Absorption lines with FUSE
Very sensitive technique, down to column densities of NH cm-2
Ubiquitous H2 in our Galaxy (Shull et al 2000, Rachford et al 2001) translucent or diffuse clouds
Absorption in LMC/SMC reduced H2 abundances, high UV field (Tumlinson et al 2002)
High Velocity Clouds detected (Richter et al 2001)

13. Ly 4-0
FUSE Spectrum of the LMC star Sk (Tumlinson et al 02)
NH2 = cm-2

14. R0/3 R0 Io
Io*20
Column densities and molecular fraction compared to models

15. Detection of H2 in absorption by FUSE in HVCs

16. Sembach et al 2001

17. The CO Tracer In galaxies, H2 is traced by the CO rotational lines
CO/H2 ~10-5
CO are excited by collision with H2
The dipole moment of CO is relatively weak
 ~0.1 Debye
Spontaneous de-excitation rate Aul  2
Aul is low, molecules remain excited in low-density region about 300 cm-3

18. Competition between collisional excitation and radiative transitions, to be excited above the 2.7K background
J=1 level of CO is at 5.2K
The competition is quantified by the ratio Cul/Aul
varies as n(H2)T1/2 /( 3 2)
Critical density ncrit for which Cul/Aul = 1
Molecule CO NH CS HCN
 (Debye)
ncrit (cm-3) 4E4 1.1E E6 1.6E7

19. Various tracers can be used, CO for the wide scale more diffuse and extended medium, the dense cores by HCN, CS, etc..
The CO lines (J=1-0 at 2.6mm, J=2-1 at 1.3mm) are most often optically thick
At least locally every molecular cloud is optically thick
Although the "macroscopic" depth is not realised in general, due to velocity gradients
Relation between CO integrated emission and H2 column density?
Is it proportional? How to calibrate?

20. NGC CO(2-1) map 13" beam
IRAM 30m
Spectra, Weliachew et al 1988

21.  Isotopic molecule 13CO, UV lines
 Statistics of "standard" clouds
 The Virial relation
1- Use the isotope 13CO much less abundant
at the solar radius: Ratio ~90
therefore 13CO lines more optically thin
A standard cloud in the MW has CO ~10
and 13 ~ 0.1
The average ratio between integrated CO and 13CO
intensities is of the order of 10

22. Successive calibrations
knowing 13CO/H2 ratio in the solar neighbourhood
(direct observations of these lines in UV absorption
in front of stars, with diffuse gas on the line of sight)
2- Statistically "standard" clouds
For extragalactic studies, numerous clouds in the beam
Typical mass of a cloud 103 Mo
something like 104 or 105 clouds in the beam
No overlap, since they are separated in velocity
Filling factor fs fv << 1 (hypothesis)
Usually TA* ~ 0.1K for nearby galaxies, 10K for a cloud
constant factor between ICO and NH2

23. 3- More justified method: the virial
Each cloud contributes to the same TA* in average
reflecting the excitation temperature of the gas
the width of the spectrum gives the cloud mass
through the virial hypothesis
V2 r ~ GM
The conversion ratio can then be computed as a function
of average brightness TR and average density of clouds n

24. Milky Way
Virial mass
versus LCO
Mvt=39LCO.81
Slope is not 1
Solomon et al 1987

25. Area A of the beam A = /4 (D)2
N clouds, of diameter d, projected area a= /4 d2
velocity dispersion V
ICO = A-1 N (/4 d2)TR V
Mean surface density
NH2 = A-1 N (/6 d3) n
NH2 / ICO = 2/3 nd/( TR V)
from the Virial V ~ n1/2 d
and the conversion ratio as n1/2 /TR

26. This factor is about 2.8 E20 cm-2 /(km/s) for
TR ~10K and n~200cm-3
This simple model expects a low dependence on metallicity,
since the clouds have high optical thickness
and are considered to have top-hat profiles
(no changes of sizes with metallicity)
However, for deficient galaxies such as LMC, SMC,
where clouds can be resolved, and the virial individually applied,
the conversion factor appears very dependent on metallicity

27. The size of clouds, where  = 1, is varying strongly
Models with  ~r-2, NH2 ~ r-1
Diameter of clouds d ~ Z (or O/H)
Then filling factor in Z2
The dependence of the conversion ratio on metallicity
could be more rapid than linear
(the more so that C/O ~O/H in galaxies,
and CO/H2 ~(O/H)2)
In external galaxies, the MH2/MHI appears to vary indeed
as (O/H)2 (Arnault et al 88, Taylor et al 1998)

28. Arnault, Kunth, Casoli & Combes 1988
LCO/M(HI) α (O/H)2.2

29. On the contrary, in the very center of starbursts galaxies,
an overabundance of CO could overestimate the molecular content
Not clear and definite variations, since TR is larger, but nH2 too, and
NH2 / ICO varies as n1/2 /TR
Possible chemical peculiarities in starbursts
12C primary element, while 13C secondary
Isotopic ratios vary
Can be seen through C18O

30. Another tracer: cold dust
At 1mm, the emission is Rayleigh-Jeans
B(, T) ~ 2 k T / 2
flux quasi-linear in T (between 20 and 40K)
In general optically thin emission
Proportional to metallicity Z
Z decreases exponentially with radius

31. When the molecular component dominates in galaxies,
the CO emission profile follows the dust profile
(example NGC 891)
When the HI dominates, on the contrary, the dust does not
fall as rapidly as CO with radius, but follows more the HI
(example NGC 4565)
CO might be a poor tracer of H2

32. Radial profiles N891 (Guélin et al 93) & N4565 (Neininger et al 96)

33. The excitation effects combine to metallicity
Explains why it drops more rapidly than dust with radius
CO(2-1) line tells us about excitation
Boarder of galaxies, CO subthermally excited
When optically thick CO21/CO10 ratio ~1
If optically thin, and same Tex, could reach 4
But in general < 1 in the disk of galaxies
Tex (21) < Tex (10) upper level not populated
even if Tkin would have allowed them

34. Braine & Combes 1992, IRAM Survey

35. Gradient of excitation
in the LMC vs MW
Sorai et al (2001)
Average value of 0.6 for MW from
Sakamoto et al 1995

36. CO(2-1)/CO(1-0) vs IRAS, and vs CII in LMC (grey band = MW)

37. Conclusion The H2 molecule is invisible, in cold molecular clouds
(the bulk of the mass!)
CO is not a good tracer, both because metallicity effect (non -linear,
since depending on UV flux, self-shielding, etc.
Very important to have other tracers
dense core tracers, HCN, HCO+, isotopes..
H2 pure rotational lines, also a tracer of the "warm" H2, always
present when cold H2 is there