Bubble-Induced Star Formationin Dwarf Irregular Galaxies

Daisuke Kawata1, Brad Gibson2, David Barnes1, Robert Grand1, Awat Rahimi3

1Mullard Space Science Laboratory, University College London, UK2Jeremiah Horrocks Institute, U. Central Lancashire, UK

3National Astronomical Observatory, Chinese Academy of Science, China

Mullard Space Science Laboratory

Kawata et al. arXiv:1306.6632

Why dIrr?

• Small, but forming stars continuously:Typical stellar mass ~ 107 M⊙◉☉⨀gas mass ~108 M⊙◉☉⨀

• An easier(?) system to simulate:require fewer particles to resolve smaller mass scale phenomena.

• Powerful laboratory to study star formation and feedback effects.(e.g. Mori et al. 1997; Carraro et al. 2001; Ricotti & Gnedin 2002; Kawata et al. 2006; Kaufmann et al. 2007, Stinson et al. 2007,09; Governato et al. 2010; Sawala et al. 2011; Revaz & Jablonka 2012; Bekki et al. 2013, Simpson et al. 2013; Teyssier et al. 2013; Schroyen et al. 2013)

Kudrtizki

Wolf-Lundmark Melotte (WLM) dIrr:Local Group dIrr, but relatively isolated

Jackson et al. 2007

Average SFR in the last ~10 Gyr: 0.0002 M⊙◉☉⨀ yr-1 (Dolphin 2000) Current SFR~0.006 M⊙◉☉⨀ yr-1 (Hunter et al. 2010)

Low level of star formation for a long time

Kepley et al. 2007

Hα image+ HI

contours

HI bubble/hook?R~250-450 pcage~35-70 Myr

Star forming region

around the bubbles

WLM-size galaxy simulation: initial condition

• NFW DM profile (fixed potential)Mtot 2.0x1010 M⊙◉☉⨀, c=12

• Stellar diskMd=1.5x107 M⊙◉☉⨀, Rd=1.0 kpc, zd=0.7 kpc

• Gas diskMd=3x108 M⊙◉☉⨀, Rd=1.5 kpc

ρd(R, z) =Md

4πh2dzd

exp(−R/hd)sech2(z/zd)

Original N-body/SPH Chemodynamics code, GCD+(Kawata & Hanami 1998, Kawata & Gibson 2003, Kawata et al. 2013a)

star particles gas particles

face-on

edge-on

Initial condition

2 Models for 1.5 Gyr evolution

• with stellar energy feedback vs.no stellar energy feedback

• baryon particle mass: 100 M⊙◉☉⨀, softening length limit: ~5 pc

• typical time step: ∆t~100-1000 yrfactor(<1) x ~pc / VSNbubble(100-1000 km s-1)(adopted also in recent Saitoh et al., Hopkins et al. sims)

• no gas accretion

Improving GCD+for dynamics(Kawata et al. 2013a)

• Rosswog & Price (2007), Price (2007)Artificial Viscosity (AV) switchArtificial thermal Conductivity (AC) ⇒ SPH can handle KHI

• Price & Monaghan (2006)Adaptive softening for both N-body and SPH.

• Saitoh & Makino (2009)individual time step limiter.

• Saitoh & Makino (2011) FAST scheme

point-like explosion testdensity profile

no individual timestep limiterwith limiter

rr

New version of GCD+ISM, star formation, feedback(Rahimi & Kawata 2012, Kawata et al. 2013b)

• Radiative cooling and heating generated by Cloudy (n, T, [Fe/H], redshift) (Robertson & Kravtsov 2008)

• New star formation and feedback modelkeeping the same particle mass for all the gas and star particles

• Stellar wind (>30 M⊙◉☉⨀), SNe II, Ia feedback, mass loss from IM stars (feedback particles)

• metal diffusion between SPH particlesGreif et al. (2009) scheme

• Pressure floor to avoid numerical Jeans instability (Hopkins et al. 2012)

Star formation criteria

Pgas < Peff

Peff

for pressure floor

unstable due to

resolution limit

Pgas

nH > 1000 cm-3

Star formation rate

C*=1 (e.g. Hopkins et al. 2013)resolution limit: nH~10 cm-3

<5Myr old stars

Bubble-induced star formation

Intermittent (~100 Myr) SF in feedback model (<SFR>=0.005 M⊙◉☉⨀ yr-1)

Steady increase in SFR in no feedback model

no feedback

with feedback

No feedback model: compact star forming region at the centre

shallow gravitational potential and inefficient cooling

no dense gas diskno star forming spirals

0

1

-1

(kpc)

0

1

-1

(kpc)

<5Myr old stars

More spread star formation(see also e.g. Stinson et al. 2006, Teyssier et al. 2012)

with feedback no feedback

mixing between the metal-poor ISM and metal-rich

SNe bubbles

too efficient enrichment in the central region

always low Z

too high Z

Feedback model: low stellar rotation velocitymany counter-rotating stars?

VSN>Vrot

<Vrot>max ~ 4 km s-1

~ WLMLeaman et al.

(2012)

new born starstSF=1-1.5 Gyr

mean <Vrot>

WLM stellar disk kinematics: small rotation and high velocity dispersion

counter-rotating stars?

Leaman et al. (2012)

Summary

• Bubble-induced star formation can maintainspread star forming regionlow metallicitylow stellar rotation velocity

• different from larger spiral galaxies?(Dobbs’ talk)

Ferreras et al. (2012) Grand, Kawata, Cropper (2012)

<50 Myr <100 Myr <200 Myr

Spitzer: NASA/JPL-Caltech

Swift UVOT: Ignacio Ferreras

galaxies like M33could be in-between?

NRAO/AUI and NOAO/AURA/NSF

Challenges=Fun ahead...

• cooling and heating in high density (nH > 1 cm-3 ) regimenon-equilibrium, radiation from stars

• star formation and feedback modelsneed to be calibrated with observationslearning from ISM and MC simulations (nH > 10 cm-3 )

• speeding upparallel efficiency for massive parallel computers→ cosmological simulations

mass loss from!intermediate mass stars"

ID

1

8

61

remnant starsold stars,

WDs, NSs, BHs

number density Temperature metallicity

Bubble induced star formation enhance the mixing

The gas properties at the disk planefeedback model at t=1.195 Gyr.

z=0, Log nH=-1 z=0, Log nH=2cooling rate

heating rate

mean molecular weight

Log T(K)

function of z, T, nH, Z

Log T(K)Cloudy + Haardt & Madau 05 UVB + CMB

http://www.nublado.org

GCMHD+ (David Barnes, DK, Wu 2012)cluster simulation

(Mvir=1.4x1014M⦿, 1.2M particles within rvir)density temperature

GCMHD+ (David Barnes, DK, Wu 2012)

magnetic field strength X-ray flux +

radio emission (contours)

Kelvin-Helmholtz Instability test(DK et al. 2013)

t=tKHI t=1.5tKHI t=2tKHI

GCD+

Athena