History of the molecular component

1
History of the molecular component

• SAAS-FEE Lecture 9
• Françoise COMBES

2
Formation of H2
Formation of the H2 molecules, very important in
the early universe Main coolant, to allow the
collapse of structures (then HD..) at z0, the
dust acts as a catalyst As soon as dust is
present, and ions have decreased, H2 is again
formed on dust Formation on grains two
processes Langmuir-Hinshelwood, H atoms land on
a grain and diffuse on the surface, by either
tunneling, or hopping Eley-Rideal, H atoms are
fixed in position, chemisorbed, and reaction
occurs only when another H lands atop
3
Binding sites for adsorbers N106 on a 0.1µ
grain Binding energy varies H on olivine ED
372K (Katz et al 99) H on amorphous carbon ED
658K For H2 respectively 314 and 542K (Katz et
al 99) Since the desorption energy is larger for
heavier molecules, only H2 can be formed like
that, at 10K The grain, as a catalyst, absorbs
the energy of formation, that can help to desorb
another molecule At low temperature H-atoms
dont diffuse on the grain and the
second (Eley-Rideal) process is the only one gt
Still poorly understood (Herbst 2000,  H2 in
Space   CUP)
4
Formation on Coronene (PAH) Sidis et al 2000
Computations of H with C24H12 Formation of H- on
MgO or forsterite Mg2 SiO4 by charge transfer Ab
initio
Experimental formation of H2 on amorphous C or
olivine (Pirronello et al 1999) Recombination
efficiency versus T triangles during irradiation
(48s) squares during desorption circles total
5
Formation in gas-phase reactions
Lepp Shull 1984 Formation of the interesting
molecules H2, HD, LiH and HeH during the
post-recombination epoch (z300-30) In standard
nucleosynthesis, there exist only H, D, 4He, 3He,
7Li Cooling by Ly? is ineffective below T
8000K, after z500 H2 is formed in small
quantities from H-, and provides sufficient
cooling to trigger the cloud collapse Cooling
is necessary since a Jeans mass will increase in
adiabatic collapse gt no fragmentation
6
First Process (Bienik Dalgarno 79) H e- --gt
H- ? H-H --gt H2 e- Second Process (Karpas
et al 79) HH --gt H2 ? H2 H --gt H2
H Third Process (Palla et al 83) Three-body
process, at high density 109 cm-3 3H --gt H2
H Direct radiative association forbidden, since
no dipole of H2 Not the case for HD and LiH (HH
--gt H2 ?)
7
Results of chemistry
Standard Friedman O0 Ob 0.1 Channels indicated
Once Compton heating becomes less than expansion
cooling, radiation and matter cool adiabatically
Tr (1z) Tm (1z)2 Lepp Shull 1984
H2 reaches 10-5
8
Cooling rate per H2 mol thermal equilibrium LiH
dominates for HD/H210-4 LiH/H210-6
During cloud collapse
Self-similar free-fall collapse at z50, ?J
560pc, MJ105Mo More molecules formed
n 3 1010 cm-3
9
The cooling is enough for cloud to collapse,
without being stopped by adiabatic regime The
cooling rotational lines become optically thick
at N 1024 cm-2 ?v for H2 N 1020 cm-2 ?v
for HD N 1015 cm-2 ?v for LiH ?v 100km/s
estimated by Lepp Shull 1984 H2 becomes
optically thick at the same point as 3-body
reaction sets in F 1012 LiH remains thin and
effective coolant until higher densities LiH/H
10-10
10
Chemistry of early Universe
More detailed reactions (Galli Palla 1998) 87
gas phase reactions
H, D, He, Li chemistry Different radiative
association for LiH was overestimated by 100
(Stancil et al 96) Non equilibrium chemistry in
radiative shocks during collapse, etc..
11
The minimal network of chemical reactions, which
can reproduce the more complex network (Galli
Palla 98) Improvement better H recombination 2-3
less electrons at z1, e/H 3 10-4 H2/H 10-6
12
Cooling functions of H2 for H-H2 coll nh 0.1
and 106 cm-3 LS Lepp Shull 83 HMK Hollenbach
McKee 89 MSM Martin et al 96 FBDL Forrey et al 97
??o/(1nc/n)
Cooling function per molecule of H2, HD, LiH and
H2 in the low n (lt100cm-3) limit Galli Palla
98
13
Lepp Shull 84
Palla et al 95
Comparisons between models
Influence of cosmology H2 independent Li more
variable LiH and HD can vary by 2-3 orders Now
Oo 1
Puy et al 93
Black 91
Giroux Shapiro 96
14
First structures in the dark age
In the current best CDM model, the first
structures to form are the smallest (bottom-up)
dk2 P(k) kn, with n1 on large-scales n -3
on small scales tilt when ?r ?m at the horizon
scale dM/M M-1/2 -n/6 when n gt -3,
hierarchical clustering
Abel Haiman 00
15
Fluctuations can be adiabatic, but they are
damped below 3 1013 (Oh2)-5/4 Mo (Silk 1968) or
isothermal The largest masses to get non-linear
after recombination (z1500) are M 106-8 Mo
(according to n, and compatible to 1014Mo
becoming non-linear today)
Mass spectrum assuming P(k) k at large-scale and
P(k) tilted n -3 at small scales (Peebles 82)
16
Growth of adiabatic fluctuations at a scale of
1014Mo (8 Mpc) They grow until their mass
equal that contained in the horizon ct Cst after,
and calibrated now Fluctuations of matter ()
standard  follow radiation, if ionised
After R (recomb) they grow too slowly to reach
the observed level
When DM is non-baryonic, only gravity gt they
grow from E when the matter gravity dominates
radiation When HDM, particles decouple when they
are relativistic stream motions gt stabilise
17
Two possible forms of spectrum for fluctuations
at Recombination I 106-8Mo first, then
bottom-up A first scales are 1014Mo for HDM
(neutrinos) cut-off
A
I
Evolution of spherical inhomogeneity At zsep,
decoupling from the rest of universe
18
The Jeans mass is M 105 Mo (Ob/0.06) -1/2
(h/0.5) -1 This corresponds to Giant Molecular
Clouds at z0 Nature
of the collapse if efficient cooling, quasi
isothermal tcool tff (Hoyle 1953) Fragmentation
could be pursued until very low masses, in
very short times (Myr) Opacity limited
fragmentation gt M 4 10-3 T 1/4 µ-9/4 f
-1/2 Mo 0.1 lt f lt 1 radiation efficiency
(with respect to black-body)
19
Mcl 10-3 Mo with densities n 1010cm-3
(PC94) Is cooling efficient? Compton cooling, HI
cooling above 1000K, then H2 cooling (Yoneyama
1972, Hutchins 76, Carlberg 81, Palla et al. 83,
Lepp Shull 84, Haiman et al 96, Abel et al 97,
Tegmark et al 97) The
calculated masses of the fragments range between
0.1 and 100 Mo More realistic collapse non
spherical, but sheets and filaments (Larson
1985) for instance, in 1D pressure forces cannot
halt the collapse (Uehara et al 96, Inutsuka
Miyama 97)
20
Most studies are interested in star
formation Globular cluster formation (Peebles
Dicke 1968) (very efficient star formation) gt
wide mass spectrum of stars (from 0.1 Mo) With
rotation, low-mass as well as VMOs (Kashlinsky
Rees 83) It is however possible that most of the
fragments remain pressure supported, as soon as
optically thick and form
a fractal of molecular clouds All the
hierarchical structures remain at the TCMB
temperature (PC94)
21
H2 formation and cooling
Simple scheme of chemistry reactions (Tegmark et
al 97) H e- --gt H h? (recombination) H
e- --gt H- h? H- H --gt H2 e- H H --gt
H2 h? H2 H --gt H2 H HHH --gt
H2H HHH2 --gt H2H2 The cooling is essentially
due to H2 below 1000K Then HD intervene below
100K LiH has a non-significant abundance (Stancil
et al 96)
22
Density computed from the simple top-hat model
before virialisation
Then assuming efficient cooling, fragmentation
occurs, as a fractal structure D 1.7 N
(fragments) 8 compatible with the fractal of the
Milky Way ISM today (Larson 81, Scalo 85) gt
scales in the ratio N 1/D 3.4 densities in
the ratio rd N(3-D)/D 4.9 n(t) ni
rd(t-ti)/tffi Fragmentation stops when the clumps
become opaque to the H2 rotational IR lines,
i.e. n 1010 cm-3 Tvir 1420K (M/105Mo)2/3
(1zvir)/100 When T lt Tvir, fragmentation can
proceed
23
Evolution of the density inside the bound
structures compared to the background density
(dash)

The solid lines correspond to the recursive
fragmentation with time-scale as a function of
the local density with different initial
fluctuations intensities (Combes Pfenniger 98)
24
n3 density in 103cm-3 fH2 molecular fraction Tvir
virial temperature xe ion fraction
T6 gas temperature in 106K
25
n3 density in 103cm-3 fH2 molecular fraction Tvir
virial temperature xe ion fraction
T6 gas temperature in 106K Tvir (1zvir) n
larger
26
n3 density in 103cm-3 fH2 molecular fraction Tvir
virial temperature xe ion fraction
T6 gas temperature in 106K n lower, since dense
structures form first
27
What are the first structures
This old question has received many different
answers, according to assumptions on the IMF and
efficiency of star formation could be gas
clumps, brown dwarfs, 0.5 Mo MACHOs, standard IMF
stellar systems, or massive black
holes Numerically, often spherical symmetry
(Bodenheimer 96) with hydrodynamics, chemistry,
(but of course no fragmentation) Big problem of
the high dynamical range (8-15 orders) non-equilib
rium chemistry, non -linear dynamics, Problem
when the structures become optically thick Abel
et al (95-98) in CDM, blobs form at filaments
crossing
28
Abel et al 2000 Top log overdensity cut in the
highest peak 320pc, 32pc velocities Bottom
log temperature Less than 1 of the gas will
form stars
29
Baryonic density around the fragment mass
enclosed M( r ) compared to Bonnort-Ebert f(H2)
and electron fraction H2 cooling time, tcross,
tff Temperature (vertical is min) Virial
radius of 5.6 106 Mo halo is 100pc cell size
0.024pc simulation box 6.4 kpc AMR
30
The first stars or quasars are likely to appear
in clumps of Tvir 104K at z 20, and H2 cooling
plays a major role The structures are probably at
filaments crossing The direct H2 radiation will
not be detectable, but as soon as there is star
formation with gt 1 efficiency, or quasar
radiating at Eddington limit, it will be
detectable with NGST at z15 (Haiman Loeb
98) In case of mini-quasars, there will be
reionisation by z 10 The average SF efficiency
can be probed through the enrichment of Lya
forest It is not larger than 2 in collapsed
regions, not far from the 1 deduced from
numerical simulations But based on the normal
IMF (very uncertain!)
31
Efficiency of star formation
Always very low efficiency, since there is not
yet a giant potential At high redshift
(z10-100), the physical conditions of the gas is
similar to the present one in the outer parts of
galaxies No deep potential wells, since
galaxy-size structures have not yet virialized
No critical surface density for star formation
Collapsing mass always of the order of the Jeans
mass gt not far from pressure support since
the Jeans mass decreases gradually as
fragmentation proceeds When objects of
galaxy-size collapse, M gtgt MJ gt violent
instabilities, starburst unless rotation
stabilises
32
Only sporadic star-formation in
small-mass GMC at high z
Can be sufficient to reionize and reheat the
diffuse inter-galactic medium Lya absorbers
(Tegmark et al 1984) HeII gas (Jacobsen et al
1994, Davidsen et al 96) Clumps resistance to
reionization N(H2) 1025 cm-2 self-shielding from
radiation fractal structure more gregarious
than homogeneous distribution Some erosion at
interfaces Extragalactic background estimation
of the ionization rate of ? 2 10-14 /s (Madau
92) gt sharp transition (Stromgren sphere
theory) in galactic disk at N(HI) 1018 cm-2
(Corbelli Salpeter 93)
33
Galaxy formation
Could range between z50 and z0 (at z200,
normal 10kpc galaxies would overlap) the
protogalaxy collapse is violent ( M gtgt
MJ) starburst in their centers In the outer
parts, there can remain cold self-gravitating
gas then hierarchical merging
Cluster formation
Another peak of star-formation, when larger
structures begin to collapse at z2, outer
gaseous haloes are stripped, heated Formation of
the hot IGM, at Tvir of the cluster Multiphase
medium survives (Ferland et al 94) Cooling gas at
the center (David et al 95)
34
Cooling flows in clusters
The hot X-ray gas pervading the cluster cools and
condenses toward the center (generally cD
galaxy) Gas phase at intermediate temperature 5
105 K has been seen (EUVE radiation, Lieu et al
96, 99) Final state of the gas still a mystery,
not sufficient stars formed Recent detection of
CO molecules in large amount (Edge et al 01,
Salomé Combes 02) Cooling filaments extend
over kpc Multi-phase gas Generally associated
with starbursts, and also AGN (feedback? Stop the
cooling flow?)
35
Latest Chandra results on Abell 1795 (ACIS for
11h) Fabian et al. 2001 size is 75" on a side
Ha NII (6583 A) emitting gas surrounding the
Abell 1795 cluster central Galaxy from Cowie et
al 1983
36
Evolution of gas content
Rao et al. 1995
Through the Damped Lya systems, the evolution of
the gas in galaxies could be traced DLA traced
by MgII, between z0 and 1.65 43
systems n(DLA)(z) (1z)2.27
Strong evolution, however depends strongly on a
few rare systems at t0 square O lum, circle
O(HI)
37
Conclusions
After recombination, GMCs of 10 5-6 Mo collapse
and fragment down to 10-3 Mo, efficient H2
cooling The bulk of the gas might not form
stars but a fractal structure, in statistical
equilibrium with TCMB Low level, sporadic star
formation gt after first stars, Reheating and
reionization The cold gas survives and will be
assembled in larger scale structures to form
galaxies A way to solve the cooling catastrophy
(Blanchard et al 92) Regulate the consumption of
gas into stars (reservoir)