Alyson Brooks Fairchild Postdoctoral Fellow in Theoretical Astrophysics Caltech

Alyson BrooksFairchild Postdoctoral Fellow in Theoretical Astrophysics

Caltech

In collaboration with the University of Washington’s N-body Shop™ makers of quality galaxies

The Formation of Realistic Galaxy DisksThe Formation of Realistic Galaxy Disks

Outline

I. Simulating Realistic Disk Galaxies is the Necessary Starting Point for I. Simulating Realistic Disk Galaxies is the Necessary Starting Point for Comparison to Observations Comparison to Observations

a) The Effect of Resolution a) The Effect of Resolution b) The Effect of Feedbackb) The Effect of Feedback

II. What we Learn II. What we Learn a) The Creation of a Bulgeless Disk Galaxy with a Dark Matter “Core”a) The Creation of a Bulgeless Disk Galaxy with a Dark Matter “Core”b) How to Create Large Disks despite Major Mergersb) How to Create Large Disks despite Major Mergers

• Fully Cosmological, • Parallel,• N-body Tree Code,• + smoothed particle hydrodynamics (SPH)

Stadel (2001)Wadsley et al. (2004)

Galaxy Formation: Now with AMR!

Outline

I. Simulating Realistic Disk Galaxies is the Necessary Starting Point for I. Simulating Realistic Disk Galaxies is the Necessary Starting Point for Comparison to Observations Comparison to Observations

a) The Effect of Resolution a) The Effect of Resolution b) The Effect of Feedbackb) The Effect of Feedback

II. What we Learn II. What we Learn a) The Creation of a Bulgeless Disk Galaxy with a Dark Matter “Core”a) The Creation of a Bulgeless Disk Galaxy with a Dark Matter “Core”b) How to Create Large Disks despite Major Mergersb) How to Create Large Disks despite Major Mergers

Low resolution: disks heat and lose angular momentum to halo

Rd=30% smaller

Kaufmann et al. (2007)

20 kpc 20 kpc 20 kpc

x1019

ato

ms

cm-2

N=1,000,000 100,000 30,000

• Isolated (non-cosmological) galaxy simulation• MW mass halo after 5 Gyr

The CDM Angular Momentum Problem

Navarro & Steinmetz (2000)

Disks are too small at a given rotation speed

Disks rotate too fast at a given luminosity

Mass (dark and luminous) is too concentratedL

og V

rot

MI

catastr

ophe!

“Zoom in” technique:high resolution halo surrounded by low resolution region

3 million particle simulation1 million particles

Computationally Efficient

Cosmic Infall and Torques correctly included

6 Mpc 100 Mpc

Katz & White (1993)

HI W20/2 velocity widths = Vmax

Geha et al. (2006)Brooks et al., in prep

Governato et al. (2008, 2009)

baryonic Tully-Fisher relation

(magnitudes from Sunrise)

Size - Luminosity Relation

Brooks et al., in prep.Data from Graham & Worley, 2008

Simulated Galaxies

Observed Sample

Naab et al. (2007) Mayer et al. (2008)

As resolution increases, Vpeak decreases Same mass

galaxies, but with and without feedback

With feedback = Disk galaxies!

Without feedback = Elliptical galaxies!

Outline

I. Simulating Realistic Disk Galaxies is the Necessary Starting Point for I. Simulating Realistic Disk Galaxies is the Necessary Starting Point for Comparison to Observations Comparison to Observations

a) The Effect of Resolution a) The Effect of Resolution b) The Effect of Feedbackb) The Effect of Feedback

II. What we Learn II. What we Learn a) The Creation of a Bulgeless Disk Galaxy with a Dark Matter “Core”a) The Creation of a Bulgeless Disk Galaxy with a Dark Matter “Core”b) How to Create Large Disks despite Major Mergersb) How to Create Large Disks despite Major Mergers

No feedback

Zavala et al. (2008)Scannapieco et al. (2008)

1) “Over-cooling” leads to loss of angular momentum similar to low resolution

Maller & Dekel (2002)

1) “Over-cooling” leads to loss of angular momentum similar to low resolution

2) “Over-cooling” leads to ONLY elliptical galaxies!

No feedback Thermal Disable Cooling Blastwave ?

• Star Formation: reproduces the Kennicutt-Schmidt Law; each star particle a SSP with Kroupa IMF

• Energy from SNII deposited into the ISM as thermal energy based on McKee & Ostriker (1977)

• Radiative cooling disabled to describe adiabatic expansion phase of SNe (Sedov-Taylor phase); ~20Myr (blastwave model)

• Only Free Parameters: SN & Star Formation efficiencies

Sub-grid physics & Blastwave Feedback Model

Stinson et al. (2006), Governato et al. (2007)

LMC HI distribution LMC HI distribution (Stavely-Smith 2003)(Stavely-Smith 2003)

Brooks et al. (2007)Erb et al. (2006)

Tremonti et al. (2004)Maiolino et al. (2008)

Stellar Mass (M)

12+

log(

O/H

)

Stellar Mass (M)12

+lo

g(O

/H)

The Regulation of Star Formation due to Feedback

Outline

I. Simulating Realistic Disk Galaxies is the Necessary Starting Point for I. Simulating Realistic Disk Galaxies is the Necessary Starting Point for Comparison to Observations Comparison to Observations

a) The Effect of Resolution a) The Effect of Resolution b) The Effect of Feedbackb) The Effect of Feedback

II. What we Learn II. What we Learn a) The Creation of a Bulgeless Disk Galaxy with a Dark Matter “Core”a) The Creation of a Bulgeless Disk Galaxy with a Dark Matter “Core”b) How to Create Large Disks despite Major Mergersb) How to Create Large Disks despite Major Mergers

Moving Forward: “Resolving” Star Formation Regions

Den

sity

X

low SF threshold

Den

sity

X

high resolution + high SF threshold

Feedback becomes more efficient.(more outflows per unit mass of stars formed)

The Formation of a Bulgeless Dwarf Galaxy

Mvir = 2x1010 M

M* = 1.2x108 M

15 kpc on a side

Green = gas

Blue/Red = age/metallicity weighted stars

The effect of altering the SF density threshold

Resolving Star Forming Complexes

The effect of altering resolution

Governato et al., 2009, Nature, accepted

arXiv:0911.2237

“Observed” Rotation Curve

Governato et al., 2009, Nature, accepted

arXiv:0911.2237

Diffuse Star Formation

“Observed” Surface Brightness Profile

“Resolved” Star Formation

Radius (kpc)0 1 2 3 4 5 6 7

18

20

22

24

Mag

/ars

ec2

0 1 2 3 4 Radius (kpc)

22

24

26

28Mag

/ars

ec2

Governato et al., 2009, Nature, accepted

arXiv:0911.2237

At high zoutflowsremove

low angularmomentum gas at a rate

2-6 timesSFR(t)

Time

J

Accreted material

Outflows

Gas removal from the galaxy center

Hot gas perpendicular to disk plane: 100km/sec

Cold Gas in shells = 30 km/sec

Outflows Remove Low Angular Momentum Gas

See also: van den Bosch (2002), Maller & Dekel (2002), Bullock et al. (2001)

Angular Momentum of Stellar Disk vs DM halo

van den Bosch et al. (2001)

The Angular Momentum Distribution of baryons must

be altered to match observed galaxies

Dark Matter with a Central Core

Diffuse SF: weak outflows

5 million particles

2 million particlesSF in high density regions:strong

outflows

DM

D

en

sity

Log Radius (kpc)

Clumpy Gas transfers orbital energy to DM via dynamical friction

DM expands as gas is rapidly removed

e.g., Mashchenko et al. (2007, 2008); El-Zant et al. (2004); Navarro et al. (1996); Mo & Mao (2004);

Tonini et al. (2006)

Outline

I. Simulating Realistic Disk Galaxies is the Necessary Starting Point for I. Simulating Realistic Disk Galaxies is the Necessary Starting Point for Comparison to Observations Comparison to Observations

a) The Effect of Resolution a) The Effect of Resolution b) The Effect of Feedbackb) The Effect of Feedback

II. What we Learn II. What we Learn a) The Creation of a Bulgeless Disk Galaxy with a Dark Matter “Core”a) The Creation of a Bulgeless Disk Galaxy with a Dark Matter “Core”b) How to Create Large Disks despite Major Mergersb) How to Create Large Disks despite Major Mergers

Mergers or Smooth Gas Accretion?

Mergers destroy or thicken disks

e.g., Toth & Ostriker 1992, Kazantzidis et al. 2007, Bullock et al. 2008, Purcell et al. 2008

Therefore, disk galaxies must grow rather quiescently

NO GAS

NO GAS

CDM mergers

Mergers or Smooth Gas Accretion?

CDM mergers

Mergers destroy or thicken disks

e.g., Toth & Ostriker 1992, Kazantzidis et al. 2007, Bullock et al. 2008, Purcell et al. 2008

Therefore, disk galaxies must grow rather quiescently

Baugh et al. 1996, Steinmetz & Navarro 2002, Robertson et al. 2006, Hopkins et al. 2008, Robertson & Bullock 2008

Not if the disks are gas rich (fgas > 50%)

Or do they?

NO GAS

NO GAS

NO GAS

ACCRETIO

N

NO GAS

ACCRETIO

N

How Do Galaxies Get Their Gas?

Dark Matter Halo + Hot Gas Dark Matter Halo + Hot Gas

Cold Disk

Infalling Gas

• Not all gas is shock heated!

• Fraction of shocked gas is a strong function of galaxy mass

• Cold flow gas accretion due to both:

1. Mass threshold for stable shock

2. Even after shock develops, there can be cold gas accretion at high z in dense filaments

Case 1

Standard

Case 2

This is already in the SAMs.

This is not.

Gas Accretion Rates at the Virial Radius

3.4x1010 M 1.3x1011 M

1.1x1012 M 3.3x1012 M

Gas Accretion Rates

Disk Star Formation Rates

3.4x1010 M 1.3x1011 M

1.1x1012 M 3.3x1012 M

1.3x1011 M

z = 1

3.4x1010 M

1.1x1012 M 3.3x1012 M

Massive Disks at z=1: Vogt et al. (1996); Roche et al. (1998); Lilly et al. (1998); Simard et al. (1999); Labbe et

al. (2003); Ravindranath et al. (2004); Ferguson et al. (2004); Trujillo & Aguerri

(2004); Barden et al. (2005); Sargent et al. (2007); Melbourne et al. (2007); Kanwar et

al. (2008); Forster-Shreiber et al. (2006); Shapiro et al. (2008); Genzel et al. (2008);

Stark et al (2008); Wright et al. (2008); Law et al. (2009)

Historic Problem : Disk Growth After z=1,dramatic change in scale lengths since z=1

The Formation of a Milky Way-Mass Galaxy to z=0

30 kpc on a side

Green = gas

Blue/Red = age/metallicity weighted stars

The Role of Cold Flows

Disk growth prior to z=1 due to cold flows

Brooks et al. (2009), Governato et al. (2009) Keres et al. (2008), Ocvirk et al. (2008), Agertz et al. (2009), Dekel et al. (2009), Bournaud & Elmegreen (2009)

The Role of Feedback

Bulg

e SF

R (M

/yr)

Develop a gas reservoir, yet fgas never > 25%

Age of Universe (Gyr)5 10Brooks et al. (2009), Governato et al.

(2009)

The Role of Feedback

Bulg

e SF

R (M

/yr)

SFR increases by ~2-3x in mergers

Age of Universe (Gyr)5 10

Heiderman et al. (2009), Jogee et al. (2008), Stewart et al. (2008), di Matteo et al. (2008), Hopkins et al. (2008), Cox et al. (2008), Daddi et al. (2007), Bell et al. (2005), Bergvall et al. (2003), Georgakakis et al. (2000)

Disk Regrowth

~30 % due to cold gas accreted prior to lmm; ~35% due to cold gas accreted after lmm; ~30% due to hot gas accretion

Young stellar disk, formed after last major merger (z < 0.8); 30% of z=0 disk mass (but dominates the light)

Old stellar disk, formed prior to last major merger (z > 0.8); 70% of z=0 stellar disk mass

Brightness not to scale!

Sunrise: Jonsson (2006)www.ucolick.org/~patrik/sunrise/

Disk Regrowth

B/Di = 1.1B/DM* = 1.2M*disk = 2.07x1010 M

B/Di = 0.49B/DM* = 0.87M*disk = 3.24x1010 M

Conclusions

• Simulations are improving! (due to resolution and feedback)

• Bulgeless galaxies with shallow DM cores are compatible with a CDM cosmology

• Strong gas outflows can selectively remove low angular momentum gas (but force resolution < 100pc is required)

• Although mergers are expected to be common in CDM, this is not at odds with the existence of disks

• Filamentary gas accretion leads to the building of disks at higher z than predicted by standard models

• Feedback regulates SFR in galaxies, building a gas reservoir and limiting gas consumption in mergers