Download Galaxy Formation, Reionization, the First Stars and Quasars

Ay 127
Galaxy Formation,
Reionization, the First Stars and Quasars
Coral Wheeler
Galaxy Formation
• The early stages of galaxy evolution — no clear-cut boundary — has two
principal aspects: assembly of the mass, and conversion of gas into stars
!
• Related to large-scale (hierarchical) structure formation, but very messy process,
much more complicated than LSS formation and growth
!
• Probably closely related to the formation of the massive central black holes as
well
!
• Star formation peaks in massive galaxies at z ~ 3, lower mass peak later. Cosmic
SFR peaks near z ~2
!
• Observations have found thousands of galaxy candidates at z > 6, with
spectroscopically confirmed galaxies out to nearly z = 9
!
• The frontier is now at z ~ 6 - 10, the so-called Era of Reionization (EoR)
A General Outline
• The smallest scale density fluctuations keep collapsing, with baryons falling into
the potential wells dominated by the dark matter, achieving high densities
through cooling
– This process starts right after the recombination at z ~ 1100
!
• Once the gas densities are high enough, star formation ignites
– This probably happens around z ~ 20
– By z ~ 6, UV radiation from young galaxies reionizes the universe
!
• These protogalactic fragments keep merging, forming larger objects in a
hierarchical fashion ever since then
!
• Star formation enriches the gas, and some of it is expelled in the intergalactic
medium, while more gas keeps falling in
– SNe feedback can dramatically affect dwarf galaxy morphology and star
formation
!
• If a central massive black hole forms, the energy release from accretion can also
create a considerable feedback on the young host galaxy
An Outline of the Early Cosmic History
(illustration from Avi Loeb)
!
Recombination:
Release of the CMBR
!
Dark Ages:
Collapse of Density
Fluctuations
!
Reionization Era:
The Cosmic
Renaissance
!
Galaxy
evolution
begins
Toy model of galaxy formation
1. Primordial abundance gas (hydrogen + helium) is dragged from
into potential wells set by dark matter
2. The gas shocks and is heated to the halo virial temperature. The
shock expands and fills the dark matter halo with hot gas that roughly
traces the dark matter potential.
3. The hot gas is supported by pressure in quasi-static equilibrium,
and cools radiatively. The cold gas falls in to form a disk.
Classical Galaxy Formation Picture
If the gas cools and contracts without angular momentum loss
to form a thin, angular momentum supported exponential disc,
the disc scale radius is given by Mo, Mao & White (1998):
Rhalo = min(Rcool, Rvir)
Simulated CDM haloes typically have
The function f (∼1) in equation (2) contains information on the
halo profile shape and contraction from baryonic infall, and
depends on the initial halo concentration c ≡ Rhalo/Rs and the disc
mass, md, in units of the total mass within Rhalo.
Physical Processes of Galaxy Formation
Galaxy formation is actually a much messier problem than
structure formation. Must also include:
!
•
•
•
•
•
•
•
The evolution of the resulting stellar population
Feedback processes generated by the ejection of mass and
energy from evolving stars
Production and mixing of heavy elements (chemical
evolution)
Effects of dust obscuration
Formation of black holes at galaxy centers and effects of
AGN emission, jets, etc.
… etc., etc., etc.
Also, may not be the same for all galaxy masses!
Hydrostatic Force Balance
For a gas in hydrostatic equilibrium, the gas profile, ρg(r), will be set by a competition
between the isothermal gas pressure,
, and the gravitational potential. If we
assume that the gravitational force is dominated by the NFW halo potential, the hydrostatic
force balance equation yields a pressure-support radius
Pressure support
radius
Angular
momentum
support radius
Kaufmann, Wheeler &
Bullock 2007
Oñorbe et al. 2015
The Evolving
Dark Halo
Mass Function
(Press-Schechter)
Perturbations on all scales
imprinted on the universe by
quantum fluctuations during
an inflationary era. Later, as
radiation redshifts away,
become mass perturbations,
grow linearly. From small to
large mass scales,
perturbations collapse to
form galaxies/clusters
z=0
z = 30
Barkana & Loeb
20
10
5
The Evolving
Dark Halo
Mass Function
(Press-Schechter)
Perturbations on all scales
imprinted on the universe by
quantum fluctuations during
an inflationary era. Later, as
radiation redshifts away,
become mass perturbations,
grow linearly. From small to
large mass scales,
perturbations collapse to
form galaxies/clusters
Cosmological (CDM)
z=0
z = 30
Barkana & Loeb
20
10
5
Read & Trentham 2005
CDM simulations
Galaxy luminosity fct
Read & Trentham 2005
CDM simulations
Galaxy luminosity fct
Read & Trentham 2005
CDM simulations
Galaxy luminosity fct
Cooling & AGN
Feedback (QSOs)
Read & Trentham 2005
CDM simulations
Galaxy luminosity fct
SNe Feedback
Cooling & AGN
Feedback (QSOs)
LCDM predicts 1000s of subhalos around the
Milky Way
Stadel et al. 2009
Currently ~30 known MW satellites
Drlica-Wagner et al. 2015
Koposov 2015, Leavens 2015, Martin 2015, Kim 2015, Kim & Jerjen 2015b
UFDs have ancient stellar pops
Hercules
Leo IV
CVn II
UMa I
Bootes I
Coma Beren
Hercules
Leo IV
CVn II
UMa I
Boo I
Com Ber
Fraction of stars formed
Magnitude
m814 (STMAG)
30 6
-0.4
-0.3
m606-m814 (STMAG)
3
2
Redshift
0.5
0.15
1
Hercules Cumulative SFH
Brown et al. 2014
Epoch of reionization
14 13 12 11 10 9
8
7
Age (Gyr)
6
5
4
3
Stopped forming stars ~10 Gyr ago
20
Simulated UFDs have ancient stellar pops
5
M* (Msun)
106
3
2
0.5
0.25
Simulated UFDs
UFD
105
104
Redshift
Simulated Classical Dwarfs
Centrals
Cumulative SFH
103
Satellites
109
1010
Mhalo (Msun)
12
10
8
6
Age (Gyr)
Both satellites & centrals
4
2
Wheeler et al. 2015
Simulated UFDs have ancient stellar pops
Infall times
5
M* (Msun)
106
3
2
0.5
0.25
Simulated UFDs
UFD
105
104
Redshift
Simulated Classical Dwarfs
Centrals
Cumulative SFH
103
Satellites
109
1010
Mhalo (Msun)
12
10
8
6
Age (Gyr)
4
2
Wheeler et al. 2015
Quenching unrelated to infall, likely reionization
Simulated
UFD continues
form stars when no
Mocking
thetoUniverse
reionizing background present
105
Cumulative Stellar Mass [M ]
Cumulative Star Formation History
1.0
0.75
0.5
0.25
Normal Reionization
No Reionization
0
1
2
3
4
104
103
0
Time [Gyr]
1
2
Time [Gyr]
Elbert et al. in prep
Garrison-Kimmel et al. 2014
Normal Reionization
No Reionization
3
4
Lyman Alpha Forest
Gunn & Peterson, 1965
QSO Observations Suggest the End of the Reionization at z ~ 6
Songaila (2004)
Evolution of transmitted flux from observations of 50 high signal-to-noise quasars (2 < z < 6.3)
Strong evolution of Lyα absorption at z > 5
•
Transmitted flux quickly approaches zero at z > 5.5
•
Complete absorption troughs begin to appear at z > 6
•
Gunn-Peterson optical depths >> 1, indicating a rapid increase in the neutral fraction
QSO Observations Suggest the End of the Reionization at z ~ 6
Ionized
Neutral
Songaila (2004)
Evolution of transmitted flux from observations of 50 high signal-to-noise quasars (2 < z < 6.3)
Strong evolution of Lyα absorption at z > 5
•
Transmitted flux quickly approaches zero at z > 5.5
•
Complete absorption troughs begin to appear at z > 6
•
Gunn-Peterson optical depths >> 1, indicating a rapid increase in the neutral fraction
Reionization
Tricky process to determine!
Depends on a number of poorly understood parameters:
!
•
•
•
Abundance and spectral shapes of early galaxies
Fraction of ionizing photons produced that are able to escape galactic
environment contribute to reionization, fesc
IGM gas density distribution
!
fesc in local Universe extremely low, but may increase at higher redshift
High resolution simulations show fesc = 0.01 - 0.20, with possible boosts
if one considers older stellar populations and binaries
Substantial Diversity of IGM
Absorption Seen
IGM transmission over same
redshift window along 4
different QSO lines of sight
A Considerable Cosmic
Variance Exists in the
transmission of the
Lyα forest at z > 5
Also possibly due to
“patchy” reionization
Cube = 200 million light-years
Time = first billion years of the universe
Blue = heated & ionized regions around galaxies Visualization credit: Marcelo Alvarez, Ralf Kaehler (Stanford), Tom Abel
When did reionization start?
Use the CMB:
•
Light from the CMB is Thompson scattered by free electrons, polarizing
the signal
•
Estimate the column density via this polarization
•
Planck measured the scattering optical depth, found best “instantaneous
reionization” model at z ~ 8
CMB Constraints on Reionization
Planck Collaboration 2016: τ = 0.058 ± 0.012 for instantaneous
reionization model, z = 8.2 ± 1.1 favored
Upper limit to the width of reionization period: ∆z < 2.8
Data points from
QSOs, Ly-alpha
emitters and Lyalpha absorbers
Looking Even Deeper: The 21cm Line
We can in principle image HI condensations in the still neutral, pre-reionization universe
using the 21cm line. Several experiments are now being constructed or planned to do this,
e.g., the Mileura Wide-Field Array in Australia, or the Square Kilometer Array (SKA)
(Simulations of z = 8.5 H I, from P. Madau)
Reionization Simulations
Pawlik et al. 2016
fesc ~ 0.6
fesc ~ 0.3
Escape fraction and SNe feedback efficiency tuned to match cosmic SFH
Simulation with the most realistic escape fraction matches observations best
Reionization in Simulations
Sims using common models
for reionization backgrounds
did not reproduce heating
claimed by those same
models!!!
This has important
considerations for many aspects
of galaxy formation, such as the
minimum halo mass that can
form a galaxy
Onorbe et al. 2016
What Reionized the Universe?
•
First stars in minihalos?
- initially thought to start reionization, now likely form too early
•
First galaxies?
- How low mass are needed? What happens to the L-fct at high
redshift? What happens to those galaxies today?
•
Active Galactic Nuclei / QSOs?
- Are there enough at high z?
Minihalos
First stars expected to form in dark matter (DM) minihalos of typical
mass ∼ 106 M⊙ at redshifts z ∼ 20.
•
Primordial density field randomly enhanced over the
surrounding matter; perturbation amplified by gravity until
it decouples from the Hubble flow, turns around, and
collapses into virial equilibrium.
•
Typical gas temperatures in minihalos are below ∼ 104 K
(efficient cooling due to atomic hydrogen).
•
H2 required for SF (no metal line cooling): fH2 ∝ Tvir^(1.5) •
Deeper potential well —> larger the Tvir —> higher
molecular abundance —> stronger cooling effect due to
molecules.
The First Stars
Gas infall into the potential wells of the dark matter fluctuations leads
to increased density, formation of H2, molecular line cooling, further
condensation and cloud fragmentation, leading to the formation of the
first stars
Krumholz, M. R., Klein, R. I., & McKee, C. F. 2007
Pop III Stars
Minimum mass at onset of collapse determined by Jeans mass:
!
!
!
But this is only a rough estimate. Mass also set by feedback processes:
• Photodissociation of H2 in accreting gas reduces cooling rate
• Ly-alpha pressure can reverse infall in polar regions at 20-30 Msun
• Expansion of HII regions can reduce accretion rate at 50-100 Msun • Photoevaporation-driven mass loss from the disk stops accretion
and fixes the mass of the star (60-300 Msun)
• Possibly also must account for: B-fields, cosmic rays, or other
ionizing sources in early Universe, etc. etc. …
Difficult process to understand because Pop III stars different from
contemporary stars (hotter, lack dust, weaker B-fields …)
Primordial Star Formation: a Top-Heavy IMF?
Expected in all modern models of Pop III star formation,
characteristic M ~ few x 10 M!
Bromm 2013
How to detect high redshift galaxy
Finkelstein 2016
Spectroscopy
Spectroscopic redshifts
highly accurate, but blind
surveys only reach out to
z < 0.5
The Lyman-Break Method
Absorption by the interstellar
and intergalactic hydrogen of
the UV flux blueward of the
Ly alpha line, and especially
the Lyman limit, creates a
continuum break which is
easily detectable by multicolor
imaging, confirmed later with
spectroscopy
Galaxies at different z
show Ly-alpha break at
different redshifts
Where they “drop-out” in
each photometric color
band gives redshift
Halo mass — Galaxy L-fct
discrepancy in place at early times
Finkelstein 2015
QSO / Quasar / AGN
"quasi-stellar radio sources," or "quasars,"
•
•
•
QSOs are thought to be
powered by supermassive black
holes at the center of galaxies,
accreting material (AGN)
z ∼ 2 − 3 has been often
designated as the “quasar
epoch”, i.e. the period where
QSOs were most active
An accurate determination of
high-z QSO LFs might enable
us to set more stringent limits
on the ionizing photon
production by these sources
during Reionization
Properties of high-z Quasar Population
•
•
•
•
UV through X-ray spectra generally have properties
indistinguishable across cosmic time — most distant quasars
establish their main emission characteristics by the time
Universe only 7% of current age
However, recent detection of high-z quasars with no emission
from hot dust (likely associated with the accretion disk) seen
in lower redshift quasar spectra may change this Possible decrease in the radio-loud fraction at the highest
redshifts could indicate a change in the relative accretion
modes and spin for the first quasars
Most distant QSO (ULAS J112001.48+064124.3), at z = 7.085
has no radio continuum emission — consistent with very
massive black hole (2 − 3 × 109 M⊙) with lower spin due to
isotropic accretion
The Nature of the BH “Seeds”
Some plausible choices include:
• Primordial, i.e., created by localized gravitational collapses in the early
universe, e.g., during phase transitions
– No evidence and no compelling reasons for them, but would be very
interesting if they did exist; could have “any” mass…
– See many good reviews by B. Carr
• Remnants of Pop. III massive stars: Mseed ~ 10 - 100 M◉
– Could be detectable as high-z GRBs
• Gravitational collapse of dense star clusters, or runaway mergers of
stars: Mseed ~ 102 - 104 M◉ • Direct gravitational collapse of dense protogalactic cores: continues
directly as a SMBH growth
More than one of these processes may be operating …
QSO Luminosity Fct Flattens at high z
Richards 2006
What Reionized the Universe?
Yan & Windhorst 2004
Bouwens et al. 2015
But is this the whole picture?
Boylan-Kolchin, Bullock & Garrison-Kimmel 2014
Faint galaxy counts required to
sustain reionization over-predict
the number of dwarf galaxy
satellites around the MW at z = 0
** for “reasonable escape fractions” **