Molecules in ULIRGS, and high density tracers

Download Report

Transcript Molecules in ULIRGS, and high density tracers

Molecules in ULIRGS, and high
density tracers
SSAS-FEE Lecture 6
Françoise COMBES
Ultra-Luminous Galaxies
Interacting galaxies appear to have more H2 content
or at least much more CO emission
The H2 gas is also more concentrated
In average, the H2 content is multiplied by 4-5
(Braine & Combes 1993)
This can be explained by the gravitational torques of the interactions
driving gas very quickly to the centers
triggers a starburst
The condition of starburst: accumulating gas in a time short enough
2
that feedback mechanisms have no time to regulate
More M(H2) in proportion
for disturbed=interacting
More star-formation, too
Would it be the conversion X?
Problems, since high density
X ~n1/2/Tr
Star formation efficiency
SFE=LFIR/M(H2)
SFE too large?
3
High end of the luminosity function
At L > 1011 Lo infrared galaxies
are the dominant population z<0.3
more abundant than QSOs
Energy from starbursts
at L> 1012Lo, all major mergers
In some cases, an AGN is
superposed
4
Spectral Energy Distribution
SED
The ratio LIR/LB varies
considerably
and is an indication of starbursts
The brightest objects are the
more obscured ones
Sanders & Mirabel 96
The ratio F60/F100 also increases
with LIR: the brightest objects are
hotter (more star formation)
5
Borne et al 99
HST WFPC2
ULIRGS
6
Excellent correlation
Radio -FIR
q=log(FIR/Radio21cm)
some exceptions are
the radio-loud AGN
Origin of the correlation
starburst, SN
ULIRGS have very
high SF efficiency
7
Molecular gas in ULIRGS
These ultra-luminous galaxies have huge quantities of H2 gas
The gas is dense and hot 105 to 107 cm-3, 60-80K
similar to the star forming regions in GMC
Large sample observed in Solomon et al (97)
Tight correlation between the CO and 100μ luminosity
==> black body emission
Small sizes of the emission, example Arp200, 300pc
justifies the optical thickness
usually 100μ emission is thin τ ~ νβ
with β ~2, but begins at 60μ
8
The molecular emission is highly
concentrated within 1kpc or
even smaller, cf Arp220
Two disks are merging, as seen in
the dispersion, and mm continuum
CO on the optical
HST image
Downes & Solomon 98
9
Gas is concentrated in central nuclear disks or rings
Stability in these central disks?
Q = 2.2 for gas only, Q= 1 for gas+stars (Downes et al 98)
==> formation of giant clusters
If the dispersion is larger, the Jeans mass is larger
Jeans length λJ ~σ2/ Σg
τff ~ σ/ Σg instability as soon as Q ~σκ/ Σg =1
For the same ratio σ/ Σg, complexes of masses
M ~ Σg λJ2
M~σ4/ Σg will condense on the same time-scale
10
Slope 1
Tight correlation CO-100μ, supports
Black-Body model
Small sizes, N(H2) > 1024 cm-2,
==> τ ~1 at 100μ for the dust
11
Black- Body model
LFIR = 4 π R2 σ Td4
(no term in τ ~νβ)
LCO = 4 π2 R2 (2k/λ2) σƒ Tbdν
LFIR/LCO ~ Td3/(fv ΔV)
predicted curve
fv filling factor in velocity
The relation departs slightly, because of Td different than Tb
CO and FIR not exacty the same regions (CO size larger?)
filling factor not unity
12
AGN or starburst?
Molecular disks of 500pc, V = 300km/s, Periods 10 Myr
LFIR = 1012Lo
50 Mo/yr formation rate, in 100 Myr (or 10 rotations)
half of the gas is turned in new stars,
5 109Mo of H2 gas => stars
M*/LFIR= 5 10-3 Mo/Lo
(L/M ~200)
If 1012Lo comes from accretion onto a black hole, at
the efficiency of L = 0.1 dm/dt c2, the accretion must be
only 1 Mo/yr, and therefore only 1% of the gas would
be accreted on the same time-scale
The gas would remain available at 99% to form stars
13
Dynamical Triggering
Numerical simulations (Barnes & Hernquist 92)
Mihos & Hernquist 1994 including star formation recipes
Galaxy interactions produce strong non-axisymmetry and
torques that drive the gas towards the center, with the help
of a small rate of dissipation
This depends essentially on the stability of the disk prior the
interaction, therefore on the bulge-to-disk ratio
Finally the role of the geometry of the interaction is secondary
direct or retrograde (provided there is a merger)
14
Without bulge, disk more unstable
At the end, the same SFR
Several burst of SF
according to the pericenters
Star formation can be
delayed
15
Mihos & Hernquist 96
Simulations of
disk/halo galaxies
Gas and young stars
are plotted
16
High Density Tracers
Nuclei of Galaxies possess denser gas
GMC to survive to tidal forces must be denser
High-J levels of CO
higher critical density to be excited (>105cm-3)
as well as
HCN, HCO+, CS, CH3OH, H2CO, OCS, etc..
SiO traces shocks (for instance supershells in starbursts)
Isotopic studies: primary or secondary elements can trace the age
of the star formation events
17
M82
Mao et al (2000)
High J-levels of CO
Images are roughly similar in morphology
although somewhat less extended than CO(1-0)
Two hot spots on either side of the nucleus
Part of the molecular torus seen edge-on
Ring due to the bar (or also void due to starburst?)
18
LVG model
N(H2) ~1023cm-2
M(H2) a few 108Mo
n(H2) ~104cm-3
close to the tidal limit
Emission comes primarily
from PDR
photon-dominated regions
quite different from the other high
density tracers
Two components in the molecular gas: dense cores, + intercloud
A diffuse component intervenes in the CO emission, also CI/CO is high
Is this representative of starburst at high z?
19
Kinetic temperatures derived are 20-60K, rather low
Heating: star formation, cosmic rays, turbulence
Consistent with the weakness of CH3OH or SiO high temp tracers
SiO mapped by Garcia-Burillo et al 01
SiO traces the walls of the supershells
not the star forming regions
Vertical filament: SiO chimney
coincident with radio cm emission
Gas ejected by the starburst
Shock chemistry
20
M82, CO3-2, 2-1, 1-0
_|¯ __ - Isotopic ratio of about
10-15 for 12CO/13CO
==> Opticall thick gas
TA* = (Tex -Tbg) (1 - e-τ)
If optically thin R(21/10) --> 4
Survey of CO(3-2) in 30 spiral galaxies (Mauersberger et al 99)
R(32/10)= 0.2-0.7, predicted if Tkin < 50K and n(H2) < 103cm-3
21
High density tracers, at low temperatures
CS, HCN
The ratios CS/CO and HCN/CO are correlated with LFIR
(1/6 in ULIRGs, 1/80 normal, as MW)
Starbursts have a larger fraction of dense gas
22
Downes et al 92
HCN in IC342
Same morphology than in CO
2 spiral arms winding up in a ring
CO/HCN ratio from 7 to 14 going
outwards
The 3mm continuum is free-free, not
thermal dust emission
(no starburst emission)
Not very high density (except dense cores, high resolution)
23
Isotopic molecules
12C/13C
in the MW, from 50-90 at the Sun radius
towards 10-20 in the center
Tracer of the astration, 13C is secondary
In the Galactic Center, also deficiency of deuterium
In Starbursts and ULIRGS (Arp220 type), CO/13CO larger
Not due to a low optical depth, since C18O is normal with
respect to 12CO
But 12C is overproduced in the nucleosynthesis of a recent burst
(Casoli et al 1992)
24
12C/13C
ratios determined in M82 and IC342 by Henkel et al 98
from CN, HCN, HCO+ observations
Always
16O/18O
12C/13C
>40 (not as low as in the Galactic Center)
> 100, 14N/15N > 100
HC15N detected in LMC and N4945 (Chin et al 99)
14N/15N = 111
lower than in the Milky Way
==>15N is synthesized by massive stars
Controversial about this formation: destruction in H-burning
formation in SN-II, 14N more secondary, and the ratio increase
with time and astration
25
Deuterated species
LMC: DCN, DCO+
Ratios about 20
strong fractionation
D/H = 2 10-5, but the deuterated molecules have lower energies
At low temperature HD +HCN --> H2 +DCN
Here temperature is 20K
Chin et al 1996
26
The HNC/HCN ratio
Useful to disentangle abundances,
excitation, density or temperature
HNC (hydrogen isocyanide)
is a high density tracer as well
HNC is weaker than HCN, except
in ULIRGS such as Arp 220
where it is > 1
Not very clear however, since in
NGC 6240, it is 10 times lower
Huttemeister et al 95
27
Other molecules
Other molecules, which trace different physical conditions
OCS in NGC 253, M82 (Mauersberger et al 95)
NH3 in Maffei 2 and others (Henkel et al 00)
rotational temperatures of 85K
H2CO and CH3OH tracing high-density
subthermally excited,
clumpy structure
=> point out very different physical conditions
and various chemistry, from one galaxy to the
next (Huttemeister et al 97)
28
Methanol
asymmetric top
A-type, E-type
lines blended
n(H2) > 105cm-3
Atomic Carbon CI fine structure line 3P1-3P0 at 492 GHz
important tracer of non-ionising radiation
In Arp 220 CI is strong, as predicted from its FIR flux,
while CII emission is depleted
This could be due to higher density, optical thickness of the C+ line
29
and dust opacity
CI/CO = 0.2 (in Kkm/s)
cooling comparable
Normally smaller than C+
(except Arp220 and Mkn231!)
Gerin & Phillips302000
Conclusions
The molecular component is much more important in starbursts
and ULIRGs; it is not the case for the HI component
It explains the considerable enhancement in star forming efficiency
Not only a problem of gas excitation, density or temperature, since
all gas density tracers confirm the large H2 abundance and density
Dynamical origin of the gas flow
Explains the transformation of HI --> H2
31
Only in strong starbursts is the H2 gas dense enough to
emit sufficiently high-J CO lines
this has important consequence for high-z galaxies
Various molecules help to constrain the physical conditions
(density, temperature, excitation, clumpiness, chemical
abundances)
At least two components: hot dense cores where stars form
+ intercloud, more diffuse medium, subthermal?
32