Our Universe,
Age 380,000 years
Galaxy
Formation
Cosmology Star
Formation
Stellar
Evolution
Chemical
Enrichment
Dark
Matter
Black
Holes
?
Hubble GOODS field
Hierarchical
Galaxy Formation
• Galaxy formation =
Halo growth +
Gas accretion +
Feedback
• Halo growth:
Cosmology
• Gas Accretion:
Gravity + cooling
• Feedback: ???
Gravitational instability: Cosmic Web
From Max Tegmark
Infall
from
IGM
disk
Rvir
Rcool
Dekel+ 09 z=2, AMR
How Does Gas Accrete Into Galaxies?
• Hot mode: Heated to Tvir; ~spherical, slow.
• Cold mode: Filamentary, rapid, smooth, T ~ 104
K.
• Transition mass: Hot halo when Mhalo >~ 1012
M.
White & Rees 1978Gabor, RD+11
Feedback Regulates SFFeedback Regulates SF
Baldry+ 08
Halo mass function,
scaled by Ωb/Ωm.
Baldry+ 08
Milky Way
Schematic
Multiwavelength M31
Multiwavelength Antennae
Gas Processing Factories
η/(1+η)
★
1/(1+η)
αΖ
Mgrav
ζMgrav
SFR = ζMgrav/(1+η)(1-αZ)
Equilibrium Relations
 Finlator+RD 08
Scaling Relations &
Scatter
• First order: Scaling rel’ns
– Mgrav ∼ fb Mhalo
1.1
(1+z)2.25
That gas comes in “lumps” is,
to first order, irrelevant.
• Second order: Scatter
– Mergers, environment, satel-
lites, etc are 2nd
order effects.
– Accreting a lump higher
fgas & SFR, lower Z. Mannucci+10

Preventive Feedback:
Photo-ionization, AGN,
gravity, winds, …?
Stars can form
RD+11
Intuition from the Equilibrium Scenario
Main Sequence Galaxies as
“Gas Processing Engines”
• Obtain gas via cold, smooth accretion
– Creates tight evolving M*-SFR relation.
• Process some gas into stars
– Produces cosmic evolution of SF, main seq.
• Ejects most gas into outflows
– SFR, gas content, metallicity set by balance of
inflow vs outflow.
• Ejected material recycles
– Critical at low-z, sets e.g. stellar mass fcn shape.
Cold accretion dominates
• Star formation is supply-limited.
• Mergers are a small contribution to accretion.
Keres+09
Fundamental driver: Halo growth
 e.g. 1012
M halo
 … at z=0, Min = 6 M/yr
 … at z=2, Min = 80 M/yr
Prediction: A given mass
galaxy forms stars faster at
high redshifts.
Dekel et al 2009


from D. Elbaz
Observations
Quenching
star formation
• Use SAM intuition to make “red and
dead” galaxies:
Heat halo gas when fhot>0.6
• Produces correct:
– Red sequence
– Bright-end luminosity/mass fcn
– Does not change faint end
• Physics uncertain, but it likely has
something to do with hot halo gas!
Gabor, RD+11
“Holistic” Approaches: SAMs
and Sims• Semi-analytic models (SAMs)
parameterize baryon dynamics,
constrain via observables.
+ Fast; tunable; builds on N-body.
- Non-unique; builds in assumptions.
• Hydrodynamic simulations directly track gas.
+ Physics more robust; convergence tests possible
- Slow; limited dynamic range; subgrid physics parameters
• Compare to widest possible range of data: “holistic”.
• Sims develop physical insight; SAMs explore,
Our Cosmological Hydro Code
N-body plus hydro:
- Gravity using Tree-PM
- Gas dynamics using EC-SPH
- Cooling (H,He,metal)
Parameterized subgrid physics:
- Star formation
- Galactic outflows
- (Quenching feedback)
Typical simulation parameters:
- Spatial resolution: ~1-5 kpc (Kennicutt Law
scale)
- Mass resolution: ~106-8
M.
- Box size: ~10-100 Mpc.
- Evolve from linear regime (z>>100) to today.
IGM enrichment: Outflow TracerIGM enrichment: Outflow Tracerwindspeed
mass loading
Too few
metals in
IGM
IGM too hot
Diffuse IGM
unenriched
Too few
metals
produced
Momentum-driven
wind scalings!
Oppenheimer & RD 2006
Oppenheimer & RD 2008
Oppenheimer & RD 2009a,b
Lyman alpha forest Metal absorbers
(mostly CIV)

Galaxy Formation: An Overview

  • 1.
    Our Universe, Age 380,000years Galaxy Formation Cosmology Star Formation Stellar Evolution Chemical Enrichment Dark Matter Black Holes ? Hubble GOODS field
  • 2.
    Hierarchical Galaxy Formation • Galaxyformation = Halo growth + Gas accretion + Feedback • Halo growth: Cosmology • Gas Accretion: Gravity + cooling • Feedback: ??? Gravitational instability: Cosmic Web From Max Tegmark
  • 3.
    Infall from IGM disk Rvir Rcool Dekel+ 09 z=2,AMR How Does Gas Accrete Into Galaxies? • Hot mode: Heated to Tvir; ~spherical, slow. • Cold mode: Filamentary, rapid, smooth, T ~ 104 K. • Transition mass: Hot halo when Mhalo >~ 1012 M. White & Rees 1978Gabor, RD+11
  • 4.
    Feedback Regulates SFFeedbackRegulates SF Baldry+ 08 Halo mass function, scaled by Ωb/Ωm. Baldry+ 08
  • 6.
  • 7.
  • 8.
  • 9.
  • 10.
  • 11.
    Scaling Relations & Scatter •First order: Scaling rel’ns – Mgrav ∼ fb Mhalo 1.1 (1+z)2.25 That gas comes in “lumps” is, to first order, irrelevant. • Second order: Scatter – Mergers, environment, satel- lites, etc are 2nd order effects. – Accreting a lump higher fgas & SFR, lower Z. Mannucci+10 
  • 12.
  • 13.
    Intuition from theEquilibrium Scenario
  • 14.
    Main Sequence Galaxiesas “Gas Processing Engines” • Obtain gas via cold, smooth accretion – Creates tight evolving M*-SFR relation. • Process some gas into stars – Produces cosmic evolution of SF, main seq. • Ejects most gas into outflows – SFR, gas content, metallicity set by balance of inflow vs outflow. • Ejected material recycles – Critical at low-z, sets e.g. stellar mass fcn shape.
  • 15.
    Cold accretion dominates •Star formation is supply-limited. • Mergers are a small contribution to accretion. Keres+09
  • 16.
    Fundamental driver: Halogrowth  e.g. 1012 M halo  … at z=0, Min = 6 M/yr  … at z=2, Min = 80 M/yr Prediction: A given mass galaxy forms stars faster at high redshifts. Dekel et al 2009   from D. Elbaz Observations
  • 17.
    Quenching star formation • UseSAM intuition to make “red and dead” galaxies: Heat halo gas when fhot>0.6 • Produces correct: – Red sequence – Bright-end luminosity/mass fcn – Does not change faint end • Physics uncertain, but it likely has something to do with hot halo gas! Gabor, RD+11
  • 18.
    “Holistic” Approaches: SAMs andSims• Semi-analytic models (SAMs) parameterize baryon dynamics, constrain via observables. + Fast; tunable; builds on N-body. - Non-unique; builds in assumptions. • Hydrodynamic simulations directly track gas. + Physics more robust; convergence tests possible - Slow; limited dynamic range; subgrid physics parameters • Compare to widest possible range of data: “holistic”. • Sims develop physical insight; SAMs explore,
  • 19.
    Our Cosmological HydroCode N-body plus hydro: - Gravity using Tree-PM - Gas dynamics using EC-SPH - Cooling (H,He,metal) Parameterized subgrid physics: - Star formation - Galactic outflows - (Quenching feedback) Typical simulation parameters: - Spatial resolution: ~1-5 kpc (Kennicutt Law scale) - Mass resolution: ~106-8 M. - Box size: ~10-100 Mpc. - Evolve from linear regime (z>>100) to today.
  • 20.
    IGM enrichment: OutflowTracerIGM enrichment: Outflow Tracerwindspeed mass loading Too few metals in IGM IGM too hot Diffuse IGM unenriched Too few metals produced Momentum-driven wind scalings! Oppenheimer & RD 2006 Oppenheimer & RD 2008 Oppenheimer & RD 2009a,b Lyman alpha forest Metal absorbers (mostly CIV)