Dark Matter and Formation of Galaxies

1
Dark Matter and Formation of Galaxies

• Françoise Combes
• Observatoire de Paris
• Les Houches, August 2005

2
Overview

• Necessity of non-baryonic dark matter for galaxy
  formation
• Successes of the CDM and hierarchical scenario
• What are the candidates for DM (CDM and BDM)
• Main problems of CDM model at galactic scales
• Alternative model MOND
• Or is it a problem of baryonic physics?
• Feedback mechanisms?
• Evidence for secular evolution and cold gas
  accretion

3
Dark matter essential for the formation of
galaxies
In a universe in expansion, structures develop
only linearly In comoving coordinates, dr /r
replaces r in Poisson equation Primordial
fluctuations dr /r ltlt 1 by definition dr /r
d d R(t) (1 z) For baryons, which can
develop Only after recombination at z 1000 The
growth factor would only be 103, ?
insufficient, since fluctuations are a few
10-5 Last scattering surface (COBE, WMAP) ?T/T
a few 10-5
4
This problem is solved by the existence of
non-baryonic dark matter, which is not coupled
with radiation, except by gravity This dark
matter can grow in density before baryons, at
any scale after equality, but only for
perturbations larger than horizon before equality
(free streaming, damping) z gt z eq
z lt zeq Radiation
Matter l gt ct ? (1
z) -2 ? (1 z) -1 l lt ct ? cste ?
(1 z) -1
5
Growth of adiabatic fluctuations at a scale of
1014Mo (8 Mpc) They grow until they contain the
horizon mass Then stay constant (calibration
t0, arrow)
? Baryons fluctuations () "standard model"
follow the radiation, and grow only after
Recombination R ? CDM fluctuations grow already
from point E equality matter -radiation
6
Power spectrum
Scale-independent spectrum, power-law such that
perturbations enter the horizon with the same
amplitude ?r /r ?M/M A M-a a 2/3, or
?(k)2 P(k) kn with n1 prediction of
inflation models P(k) k at large scales but
P(k) tilted n -3 at small scale (Peebles
82) Come from the streaming effect below horizon
7
Hierarchical formation
In the CDM model, the first structure to form
are the smallest ones, then they merge to form
larger ones (bottom-up)
dk2 P(k) kn, with n1 at large scales,
then n -3 at small scales dM/M M-1/2
-n/6 when n gt -3, hierarchical formation (dM/M
negative slope) Abel Haiman 00
8
Fluctuations of density
Tegmark et al 2004
9
Numerical Simulations
With assumed gaussian fluctuations at the
start, the non-linear regime can be followed
But gas and baryons ? (CDM easily taken
into account by semi-analytical models à la
Press-Schechter)
10
Hierarchical Formation of galaxies
The smallest structures form first, from dwarf
galaxies and globular cluster sizes From
successive merger and accretion more and more
massive systems form (Lacey Cole, 93, 94) They
are less and less dense M µ R2 and r µ 1/R
11
Hypotheses for CDM
Cold Particules no longer relativistic at
decoupling Particules WIMPS (weakly interactive
massive particles) Neutralinos the lightest
supersymmetric particle LSP Relic of Big-Bang,
should disintegrate in gamma rays (40 Gev-
5Tev) May be even lighter particules, or with
more interaction? non-gravitational ? (Boehm,
Fayet, Silk 04, 511kev INTEGRAL) Actions
(solution to the strong-CP problem, 10-4
ev) Primordial black holes?
12
Hypotheses for dark baryons
Baryons in compact objects (brown dwarfs, white
dwarfs, black holes) are either not favored by
micro-lensing experiments or suffer major
problems MACHOS --gt MACDOS (Disk instead of
Halo Objects) (Alcock et al 2001, Lasserre et al
2000, Tisserand et al 2004) ?Best hypothesis is
gas, Either hot gas in the intergalactic and
inter-cluster medium Or cold gas in the vicinity
of galaxies, and filaments
13
First structures of gas
After recombination, GMCs of10 5-6 Mo collapse
and fragment until 10-3 Mo, H2 efficient cooling
Either a burst of star formation followed by
strong feedback Or a fraction of the gas does
not form stars But a fractal structure, in
statistical equilibrium with TCMB Sporadic star
formation ? after the first stars,
Re-ionization The cold gas survives to be
assembled in structures at larger scale to form
galaxies A way to solve the  cooling
catastrophy" Self-regulates the gas consumption
into stars (reservoir)
14
Dark Age andReionization
Big-Bang Recombination 3 105yrs Dark Age 1st
Stars and QSO 0.5Gyr Cosmic Renaissance End
of the dark age End of reionization
1Gyr Galaxies evolve Solar system forms
9Gyr Astronomers today 13Gyr
QSO z6. SDSS Gunn-Peterson Effect Double
reionisation? WMAP polarisation
15
Evidence of reionization
Line of sight in front of a quasar Absorption
spectrum Lyman-alpha forest Or total absorption
Djorgovski et al 01
16
Cold H2 clouds and baryonic dark matter
Mass 10-3 Mo density 1010 cm-3 size 20
AU N(H2) 1025 cm-2 tff 1000 yr Adiabatical
character Life-time much larger than free-fall
Fractal collisions lead to coalescence,
heating, and to a statistical equilibrium (Pfenni
ger Combes 94)
Around galaxies, baryonic matter
dominates The stability of cold H2 is due to
its fractal structure
17
D1.8
Formation by recursive Jeans fragmentation ?a
hierarchical fractal ML N ML-1 rLD NrL-1D a
rL-1/rL N-1/D
D2.2
18
Projected mass log scale (15 mag) N10,
L9 The surface filling factor depends strongly
on D lt 1 for D1.7
19
Radial gas distribution
Radial distribution of a few components in the
spiral NGC6946

• CO follows an exponential, as all
• other component related to star
• formation
• Radio Continuum
• Blue luminosity
• H-alpha
• Only HI is different
• N(H2) is 10 times N(HI) in the centre

20
Dark gas in the solar neighborhood
Dust detected in B-V (by extinction) and in
emission at 3mm Gamma-rays associated to the
dark gas (interaction with Cosmic Rays)
Largely a factor 2 (or more) Grenier et al (2005)
21
ISO -Signal of dark matter N(H2) 1023 cm-2 T
80 90 K 5-15 X N(HI) NGC 891 Grey
matter Valentijn Van der Werf 99
22
Hot Gas in filaments
Detection of OVI in X-ray?
WHIM
ICM
DM
23
Main problems of the L-CDM paradigm

• Prediction of DM cusps in galaxy centers, in
  particular absent in
• dwarf Irr, dominated by dark matter
• ? Low angular momentum of baryons, and consequent
  too small
• radius of disks
• ? High predicted number of small haloes (not
  observed)
• Solution could come from physical processes (SF,
  feedback)
• More realistic? Lack of resolution in
  simulations?
• Cold gas accretion might help to solve the
  problems?

24
Predictions LCDM cusp versus core
Dwarf Irr galaxies are dominated by dark
matter In average, their dark density is
characterised by a core (de Blok et al 2003,
Swaters et al 2003, Gentile et al
2004) Power-law for the density a 1-1.5,
observations a 0-0.5
25
Angular momentum and disk formation
Baryons lose their angular momentum on the
CDM Usual paradigm baryons at the start ? same
specific AM than DM The gas is hot and shock
heated to the Virial temperature of the halo But
another way to accrete mass is cold gas mass
accretion Gas is channeled through filaments,
moderately heated by weak shocks, and radiating
quickly Accretion is not spherical, gas keeps
angular momentum Rotation near the Galaxies, more
easy to form disks
26
Avoidance of dynamical friction
GAS
If the gas flows slowly in a cold phase on
galaxies, the hierarchical merging will lose less
angular momentum through dynamical
friction Late (instead of early)
accretion Same process as feedback, but can be
more efficient (Gnedin Zhao 02)
The gas, stripped, does not experience friction
27
Too many small structures
Today, CDM simulations predict 100 times too
many small halos around giant Galaxies like the
Milky Way
28
Disruption of small structures

• The radial density profiles are key
• to the control disruption and merging
• More cold gas in dwarf haloes
• Much less concentration
• Fragmentation
• Baryonic clumps heat DM through
• dynamical friction and smooth any cusp
• in dwarf galaxies?
• May change the mass function for
• low-mass galaxies

core (Mayer et al 01)
cusp
29
Influence of Feedback
5 1015erg/g adiabatic during 30 Myr Preventing
star formation Gas above the curve cannot cool
Thacker Couchman (2001) Conclusion does not
solve the problem not enough resolution?
30
Feedback solutions
Feedback due to star formation The gas is
reheated, galactic winds eject the gas out
galaxies Avoid too fast central
concentrations Feedback due to the AGN Black
hole in the center, in every spiral galaxy with
bulge Radio jets eject mass, heat the gas, and
avoid its concentration Regulates formation of
black hole ?Feedback without too many stars
formed
31
Dark Matter in Galaxy Clusters
In clusters, the hot gas dominates the visible
mass Baryons are heated and become visible fgas
tends to fb Wb / Wm 0.15 universal
ratio The radial distribution dark/visible is
reversed The mass becomes more and more visible
with radius (David et al 95, Ettori Fabian 99,
Sadat Blanchard 01) The gas mass fraction
varies from 10 to 25 according to clusters
32
Radial distribution of the hot gas fraction fg in
clusters The abscissa is the mean density in
radius r, normalised to the critical density ?
The dark matter is more towards the centre (dark
baryons?)
33
Cooling Flows in Galaxy Clusters
Cooling time lt Hubble time at galaxy cluster
center ? Gas Flow, 100 to 1000 Mo/yr Mystery
the cold gas or newly formed stars not
detected? Today, flow rates have been divided by
10 and the gas is detected 23 galaxies detected
(Edge 2001 Salomé Combes 2003) Results
Chandra XMM cooling flows self-regulated Re-he
ating processes, feedback from the Active Nucleus
or Central black hole shocks, jets, sound waves,
bubbles...
34
Perseus Ha (WIYN) and CO (IRAM)
CO, cold molecular gas Salomé, Combes, Edge et al
05
Ha, Conselice 01
35
Perseus Cluster
Fabian et al 2003
36
Sound Waves in Perseus
The relativistic plasma of the radio jets
compress the hot gas and generates sound
waves Fabian et al 03
37
Abell 1795 cooling wake
T(cool) 300 Myr (Fabian et al 01) 200 Mo/yr
for R lt 200kpc (Ettori et al 02) dynamical
time of oscillation
60kpc filament Ha (Cowie et al 85) à
V(amas) ?Cooling Wake The cD galaxy at V374km/s
w/o cluster
38
A1795 CO(2-1) integrated map
Tight correspondance between CO(2-1) emission and
the lines Ha NII (grey scale) Radio jets
contours 6cm van Breugel et al 1984 The active
nucleus creates cavities in the hot gas ? Cooling
on the border of cavities, where CO Ha are
observed (Salomé Combes 2004)
39
Relation Black Holes-Galaxies
Mbh 0.2 Mbulge
Blue stellar velocities Green gas
velocities Red disk of masers H2O, OH.. Gebhardt
et al 00, Ferrarese Merritt 00
40
MOND MOdified Newtonian Dynamics
Gravity law modified, or inertial law (Milgrom
1983)
NGC 1560
m(x) x if x ltlt 1 1 if x gtgt1 When
acceleration is below a0 2 10-10
m/s-2 gM (a0 gN)1/2 Logarithmic potential V
log r, instead of 1/r
gas
stars
Sanders McGaugh 02
41
MOND theoretical developments
This is an empirical hypothesis from the start,
but with a large success at galactic scale To
ensure momentum conservation Lagrangian
formulation AQUAL With a modified Poisson
equation (Bekenstein Milgrom 1984) Then
multiple trials (like Phase Coupling Gravity,
PCG, 1988) But problems with causality, etc.. In
2004, Bekenstein publishes its TeVeS
Tensor-Vector-Scalar theory, which replaces
general relativity, and predicts
properly gravitational lensing
42
MOND critical acceleration
gM2 V4/R2 GM/R2 ? Tully-Fisher law M
V4
Mdyn/L versus acceleration (radial) at that point
Mdyn/L versus radius at last measured point
43
Fit of HSB and LSB as well
Only M/L is the free parameter, and it varies
within reasonable ranges, with what we know from
stars
LSB N1560
HSB N2903
44
Multiple rotation curves
Sanders Verheijen 1998, all types, all masses
(Uma)
45
Mass-to-light ratios
The M/L obtained is in some sort a prediction of
MOND, and corresponds to what is expected from
stellar populations Uma spirals Sanders
Verheijen 98 Solid lines show predictions from
SPS by Bell de Jong (2001)
46
Pressure-supported systems
GC Clobular clusters, GMC Giant Molecular
Clouds CE Compact ellipticals, dwSph dwarf
Spheroidals gE giant ellipticals, X-ray
clusters of galaxies
Solid line is s2/r a0 10-10ms-2
X-ray
CE
gE
GC
In MOND all objects are bound
dwSph
GMC
47
Problems of MOND in clusters

• Inside galaxy clusters, there is still a lot of
  dark matter, that
• cannot be explained by MOND, since the cluster
  center
• is only weakly in the MOND regime (0.5 a0)
• Explored by X-ray gas in hydrostatic equilibrium,
  and gravitational
• Lensing
• MOND reduces by a factor 2 the missing mass
• There is still some mass to be discovered
  (Sanders 1999)
• baryons, neutrinos.
• Observed the baryon fraction is the universal
  one only in the outer
• parts of the clusters

48
MOND and clusters

• In CDM cosmological simulations, more baryons
  fall inside the cluster
• (Kravtsov et al 2005)
• Strong variations of the baryon fraction
• according to the baryon physics (adiabatic,
  cooling, SF..)

49
MOND fit of WMAP data
Fit by MOND (with no exotic DM) of Acoustic
peaks (S. Mc Gaugh 03)
Includes a neutrino mass of 1 ev
Fit with CDM L
50
A crucial test
Skordis et al
51
Recent developments for MOND

• Dynamical friction enhanced (Ciotti Binney 04)
  Black holes and globular clusters spiral to the
  center in tdyn galaxy clusters should show
  segregation
• Theory with Lorentz covariance TeVes, which tends
  to MOND at the limit (J. Bekenstein, 2004) ?
  allows to consider MOND and CMB, structure at
  large scale
• Theory that replaces GR, and tends to Newton, or
  MOND according to the value of acceleration,
  takes into account gravitational lenses
• Extends the theory AQUAL, that solved the
  conservation of momentum (lagrangian
  formulation), without superluminic propagation

52
LCDM problems solved by baryon physics?
Paradigm of gas heated to virial temperatures of
structures ?Gas accretion is then slow, and
galaxy mergers dominate This could be wrong or
oversimplified, and at small scales (the galaxy
scales), cold gas accretion instead The fraction
of baryons at galactic scale would be much
larger, there will be core instead of cusps, and
satellites haloes will then be fragile to
merging ? Solves the main problems of CDM?
53
Cold Gas AccretionBars and secular evolution
Dynamical instabilities are responsible for
evolution With self-regulation ?Bars form in a
cold unstable disk ?Bars produce gas inflow, and
?Gas inflow destroys the bar
gas accretion Recent debate about this
cycle -- is bar destruction efficient? -- can
bars reform? Central Mass Concentration (CMC)
54
With accretion (and star formation)
Without gas accretion
Bournaud Combes 2002
55
Gas accretion to maintain SFR and bars
?Galaxies in the middle of the Hubble sequence
have about constant SFR (Kennicutt 1983,
Kennicutt et al 1994) Even taking into account
the stellar mass loss, an isolated galaxy should
have an exponentially decreasing SFH ?Small
companions are not sufficient, like systems
falling now on the MW (Sag dw, Canis major,
etc..) And large systems will destroy disks The
galaxy must double its mass in 10 Gyr, for bar
frequency ?Secular evolution can also produce
starbursts, the gas infalls intermittently
according to the bar cycle
56
Angular momentum transfer

• Bars are waves with negative angular momentum
  they are created by
• a transfer to the outer disk (through a transient
  spiral)
• In collisionless systems (no gas), the exchanges
  can only be with the
• outer stellar disk
• Or with the dark matter halo
• (Athanassoula 2002, Curir Mazzei 2000)
• But, too massive DM-halo slow down bars
  (Debattista Sellwood 2000)
• When the DM-halo is not dominant in the inner
  galaxy,
• angular momentum transfer is much more efficient
  with gas
• Bar formation and destruction occurs through gas
  infall

57
Evolution along the Hubble Sequence
B/T is an essential parameter of the sequence
although it generally increases through
evolution, it can also decrease ? cycle
58
Statistics on bar strength (OSU) Quantification
of the accretion rate Block et al 2002
Observed
Doubles the mass in 10 Gyr
No accretion
59
Merging of companion and gas accretion
To have bars, cold gas is required to increase
self-gravity of the disk Dwarf companions not
more than 10 of accretion (interaction between
galaxies heat the disk, Toth Ostriker
92) Massive interactions develop the
spheroids Required a source of continuous cold
gas accretion from the filaments in the near
environment of galaxies ? Cosmological accretion
can explain bar reformation
60

• 4  phases 
• 4 Zoom levels
• from 20 to 2.5 Mpc.
• z 3. (from. z10.)

61
Multi-zoom Technique

• Objective
• Evolution of a galaxy
• (0.1 to 10 kpc)
• Accretion of gas
• (10 Mpc)

62
Galaxies and Filaments
Multi-zoom
63
History of star formation
Isolated galaxy
Galaxy with accretion and mergers
?Accretion is compatible with doubling the mass
in 10 Gyr
64
Cold Gas AccretionLopsided Galaxies
Peculiar galaxies without any companion Richter
Sancisi (1994) 1700 galaxies, 50 asymmetric
Late-types 77 Matthews et al 98 Stellar disk
also Zaritsky Rix 97 About 20 of
galaxies have A1 gt 0.2 In NIR distribution
(OSUB sample) 2/3 have A1 required by an
external mechanism
ltA1gt 1.5rd lt r lt2.5rd
65
Frequency of m1 perturbation
Baldwin et al 80 kinematic waves have long
life-time, but not sufficient to explain the A1
frequency ?Mergers ?Gas accretion Bournaud et
al 2005 The parameter A1 (density) does not
correlate with the tidal index Tp M/m
r3/D3 Most galaxies are isolated (Wilcots
Prescott 04) A1 and A2 are correlated, for each
type Interactions and mergers cannot explain
The m1 of isolated galaxies, the correlation
with type and with m2 ? a large number of m1
by accretion

66
Simulations m1 accretion
Only gas accretion (here with 4 Mo/yr) can
explain the observed frequency of m1 and the
long life-time of the perturbation
NGC 1637 simulation observations NIR
67
Conclusion (1)

• Parameters of the Universe Wm0.3, 15 baryons,
  85 ??
• The model of dark matter CDM, with L 0.7 is the
  best fit to
• observations, including large scale structures
• Still unsolved problems at galaxy scale
• ? CDM is predicted to dominate in galaxies, with
  a central cusp
• Problem of angular momentum for baryons, lost to
  the benefit
• of CDM, and too small disk formation
• Prediction of a multitude of small halos, not
  observed
• The physics of baryons could solve a large part
  of the problems
• Essentially cold gas accretion for instance
• Or else MOND??

68
Conclusion (2)
The physics of the baryonic gas is a crucial clue
to the formation of galaxies The usual
assumption that gas is shock heated to the virial
temperature of the dark haloes might not be
true Cold gas accretion instead, with the
consequence of more baryons accreted at a given
time ? dominance in the center of galaxies
masking the cusps ? large gas extent around
galaxies, less angular momentum lost by dynamical
friction ? more disruption and merging of the
small masses