|


DYNAMICAL TRIGGERS OF STAR FORMATION IN LUMINOUS INFRARED GALAXIES



Joshua E. Barnes (IfA)

Ken Chambers (IfA)
Lisa Chien (IfA)
John Hibbard (NRAO)
Lisa Kewley (IfA)
Dave Sanders (IfA)
|

Luminous Infrared Galaxies


Three nearby luminous infrared galaxies in different stages of merging (courtesy D. Sanders).
|

Inflow Dynamics: Disk Simulations


Response of a gas-rich disk to a strong tidal encounter; top row: gas; bottom row: disk stars (Barnes 1994).
Inflowing gas; top: angular momentum; bottom: torques (Barnes & Hernquist 1996).

Gravitational torques transfer angular momentum from gas to stellar bar
(Gerin et al. 1990; Barnes & Hernquist 1991).

Shocks play a crucial role in separating gas from stars.
|

Inflow Dynamics: Mergers


Final stages in a gas-rich merger; top row: gas; bottom row: disk stars (Barnes 1994).
Merger remnant; left: stars; right: gas (Barnes 1994).

Dense nuclei lose orbital angular momentum via dynamical friction
(Barnes & Hernquist 1991).

Nuclear starbursts can build cores of elliptical galaxies (Kormendy & Sanders 1992).
|

Inflow Dynamics: Global Relations


Energy dissipation (descending) and density increase (ascending) in three merger simulations (Barnes & Hernquist 1996).

Dissipation is necessary for density increase:

|

Inflow Dynamics: Dissipative Merger Simulation


DIRect encounter of equal-mass disks. Only gas is shown; colors indicate dissipation rate (Barnes 2002).
|

Inflow Dynamics: Dissipative Merger Simulation - Action!


BOO!
DIRect encounter of equal-mass disks. Only gas is shown; colors indicate dissipation rate (Barnes 2002).
|

Star-Formation Law


Schmidt (1959):

    

Kennicutt (1998):

    

Theory:

    
    
Relation between gas surface density and star formation rate (Kennicutt 1998).
|

Starburst Simulations


Star formation rates (left) and gas remaining (right) in merger simulations with a Schmidt law and n = 1.5 (Mihos & Hernquist 1996).

Starburst timing is governed by disk stability, not presence of a bulge per se.

If initial SFR is scaled to normal disk galaxies, peaks are comparable to ULIGs.

Star formation mostly confined to central regions.
|

Extended Star Formation: NGC 4038/9


Gas and star formation in NGC 4038/9 (Sanders et al. 1999).

Note intense star formation in overlap region and disks.
What is range of CO/mid-IR ratio?
|

Extended Star Formation: Arp 299


NIR image of Arp 299 (Neff et al. 2004)

This ``supernova factory'' contains multiple sites of star formation!
|

Extended Star Formation: Other Galaxies


NGC 4676

  • Spectrum dominated by A-stars, indicating ``rapid, widespread star formation that
    effectively ceased at least 5 × 107 years ago'' (Stockton 1974).

NGC 7252

  • A-star spectrum within nucleus and ~7 kpc from center (Schweizer 1982).

    Violent relaxation ineffective at redistributing products of central star formation
    => spatially extended star formation required.
|

New Star Formation Rules


Governing parameters

  • Shock strength: fast trigger

  • Gas density: slow trigger

Density-dependent SF:

n = 1.5, m = 0 Modified Schmidt law

Shock-induced SF:

n = 1, m = 0.5 Weak shock-driven SF
n = 1, m = 1.0 Strong shock-driven SF
|

A Model of the Mice


The Mice, NGC 4676, and the model (Barnes 2004; Hibbard & Barnes, in prep.).

NGC 4676 matched by parabolic encounter of equal-mass galaxies;
first passage at t1 = 150 to 200 Myr ago.
|

A Model of the Mice: Best Match


Model of NGC 4676 with weak shock-induced star formation (Barnes 2004). Rendering includes star formation and dust.
|

A Model of the Mice: Best Match - Action!


BOO!
Model of NGC 4676 with weak shock-induced star formation (Barnes 2004). Rendering includes star formation and dust.
|

A Model of the Mice: Star Formation Rates


Solid line: density-dependent SF; dashed line: weak shock-induced SF; dotted line: strong shock-induced SF (Barnes 2004).

Different star formation rules produce very different star formation histories.
|

A Model of the Mice: Distribution of New Stars


Left: old stars (contours) & gas (halftone); middle: density-dependent SF; right: weak shock-induced SF. Points show ages: red < ~11 Myr, green < ~44 Myr; blue < ~176 Myr.

|

A Model of the Mice: Orbit-Plane Views


density-dependent SF weak shock-induced SF strong shock-induced SF
New star particles, viewed looking down on the orbit plane at first passage.
|

A Model of the Mice: Orbit-Plane Views - Action!


density-dependent SF weak shock-induced SF strong shock-induced SF
BOO!
New star particles, viewed looking down on the orbit plane.

Final distribution of newly-formed stars is ~3× more extended in shock-driven models.
|

Dynamical/Photometric Models of Interacting Galaxies

(Chien & Barnes, in prep.)

  • Model ISM as isothermal gas:

    T ~ 104 K

  • Survey encounter geometries:

    DIRect, INClined, POLar, RETrograde

  • Vary star formation rules:

    density-dependent, weak & strong shock-induced

  • Compute luminosities with spectral
    synthesis code

  • Include dust opacity
DIRect (top) and RETrograde (bottom) encounters of equal-mass galaxies (Barnes 2002). Pericenter occurs at t = 1. Contours show old stars; greyscale shows gas.
|

Dynamical/Photometric Models: Luminosity Calculation


Bolometric (black), B (blue), and V (green) magnitudes for a Simple Stellar Population (SSP) as functions of age. Computed using Starburst99 and Padova tracks (courtesy C. Leitherer).
|

Dynamical/Photometric Models: Bolometric Luminosity


Luminosities of merger models with density-dependent SF (solid) and strong shock-induced SF (dotted); note variety of SF histories.
|

Star Formation Histories: Clusters as Chronometers


  • Select interacting starburst galaxies

  • Identify young clusters on ACS images

  • Obtain spectra with LRIS in multi-object mode

  • Determine cluster ages using SSP models

  • Compare age distributions with predictions of
    dynamical models

  • Also obtain:
    • Line-of-sight extinctions
    • Cluster masses
    • Cluster kinematics
    • Cluster metallicities
Starburst99 spectra of a SSP. Ages from to to bottom are 10, 30, 100, and 300 Myr.
|

Star Formation Histories: Proof of Concept. I

(Barnes, Chien, & Chambers, in prep.)

Use de Grijs et al. (2003) lists of young clusters in NGC 4676 (cz = 6600 km/s) and UGC 10214 (cz = 9410 km/s).

  • Correct positional errors of ~1'' or more

  • Select clusters spanning widest range possible on BVI color-color plot

  • Target 15 clusters in NGC 4676 and 13 in UGC 10214

  • First run on Keck I in May 2005

Image of NGC 4676 with clusters indicated by circles (de Grijs et al. 2003).
|

Star Formation Histories: Proof of Concept. II


A yery young cluster in UCG 10214 (bottom), a slightly older cluster in NGC 4676 (middle), and a still older cluster in NGC 4676 (top).

4.5 hr on NGC 4676, 2 hr on UGC 10214 => more than 30 clusters detected; many exhibit emission lines as well as continuum. Good S/N to mV = 23 or fainter.
|

Star Formation Histories: Young Cluster in UGC 10214


UCG 10214 and the very young cluster (Tran et al. 2003).
|

Star Formation Histories: Selection Bias


Monte-Carlo simulations of cluster visibility. Points above the lines are clusters with mV = 23 at D = 100 Mpc.

Luminosity evolution biases samples towards younger clusters, but clusters from bursts up to ~100 Myr old should still be observable if they are as massive as large globular clusters.
|

Star Formation Histories: Galaxy Sample

Requirements: tidal features conducive to dynamical modeling, HI interferometry
for velocity fields, ACS images for cluster identification.
|

Conclusions


  • Extended starbursts are real and require explanation

  • Hydrodynamic triggers of star formation:
    • Shocks are fastest
    • Density is slowest
    • Real answer may lie between
  • Shock-induced SF can produce extended bursts

  • Both kinds of SF can produce LIGs

  • Star formation history depends on encounter geometry
    • shock-induced SF produces greatest variety of histories
  • Multi-object spectroscopy of young clusters will allow
    reconstruction of SF history