Formation of the First Stars Under Protostellar Feedback

Athena Stacy

First Stars IV 2012

Collaborators

• Volker Bromm (U.Texas)• Andreas Pawlik (U. Texas)• Thomas Greif (MPA)• Avi Loeb (Harvard/CfA)

Dark Ages

Open Questions

- What role did they play in reionization and metal enrichment?

- What feedback did they exert on later star formation? (Pop III to Pop II transition)

This depends on the Pop III IMF, SFR, and rotation rates…

- What were their typical masses?

- What was their typical multiplicity?

- When will a Pop III protostar’s accretion become shut off by feedback ?

or ???or

I. Pop III Star Formation Without Feedback

Stacy, Greif, & Bromm, 2010 MNRAS, 403, 45

???

Cosmological simulation:

- Gadget (SPH + N-body)

-initialized at z=100 according to CDM model

- followed formation of protostar (sink particle) and subsequent 5000 yr of accretion

- msph (gas) = 0.015 M - Mres ~ 1.5Nneighmsph ~ 1 M

= minimum allowed Jeans mass

Initial Collapse

IGM

minihalo sink

3-body reactions and H2

formation

(time)

Sink Particles

By using sink particles, we can continue following evolution of star-forming gas for thousands more years (~ 100 freefall times)!

• Msink = 1 M

• n = 1012 cm-3

• racc ~ 50 AU ~ 1015 cm• R ~ 1011 cm Density no longer evolved• Accrete gas particles

that fall within racc of

sink

50 AU

Pop III stars can form in multiples!

Multiple stars form within a disk that has grown to ~ 40 M (tacc = 5000 yrs)

5000 AUDensity[cm-3]

Binary and Multiple Formation

• Toomre Fragmentation criterion:• Q ~ 0.4 < 1

• tcool < trot and

(Gammie 2001, Kratter et al. 2010, 2011)• Multiple sinks form through disk fragmentation

Rapid Pop III Accretion Rates

Sink B:Msink ~ t0.25

dM/dt ~ t-0.75

Sink A:Msink ~ t0.5

dM/dt ~ t-0.5

B&L 2004

sink B

sink A

Mfinal > 100 M

II. Pop III Star Formation With Radiative Feedback

Stacy, Greif, & Bromm, 2012, MNRAS, 422, 290

Protostellar Feedback

• Repeat previous cosmological simulation, but with updated H2 cooling rates (Sobolev approximation)

• Model LW radiation and growth of surrounding HII region

• Also performed a comparison “no-feedback” simulation

• How will radiation alter the growth of the Pop III star?

-200 radial segments-105 angular sements-107 bins

The I-front Tracker

M* = Msink

R* = ?

The Protostellar Model

Hosokawa et al. 2010

Hosokawa et al. 2010

Adiabatic accretion

KH contraction

ZAMSAdiabatic expansion

KH contraction

ZAMS

The Protostellar Model

Adiabatic expansion

KH contraction

ZAMS

Slowcontractionmodel

Simulation model

M*=Msink

I-front breakout

Ionization rate

Mass infall rate

M* = 15 M

I-front Evolves in Morphology

1500 yr 2500 yr 4500 yr

Temperature Structure

2500 yr

3000 yr

With feedback

Sink potential well heating

I-front

Warm neutral bubble

Without feedback5000 yr

4000 yr

2500 yr

2500 yr

With Feedback

* = main sink

+ = secondary sink

Feedback Slows Disk Growth

No feedback

With feedback

Envelope = gas with n>108 cm-3

WithoutFeedback

Density,

x-y plane

Density,

x-z plane

Temperature,

x-y plane

Temperature,

x-z plane

Box = 10,000 AU

Disk disruptedby N-body dynamics

Reduced Accretion Rate

With feedback

Msink ~ t0.09

Without feedback

Msink ~ t0.13

2nd largest sink(with feedback)

Mfinal ~ 30 M

III. Pop III Formation Under Dark Matter Effects?

Stacy, Pawlik, Bromm, & Loeb 2012, MNRAS, 421, 894

WIMP annihilation important for Pop III stars?

(www.nasa.gov, Sky and Telescope, Gregg Dinderman)

Can it heat SF gas, or replace/supplement nuclear fusion?

Pop III stars form in regions of high DM density

May lead to extremely massive and luminous Pop III stars

(e.g., Freese et al. 2008, Spolyar et al. 2008, Ioccoet. al 2008, Natarajan et al. 2009)

-R* ~ 1 AU-Teff too low to ionize-Accretion unimpeded for long time

a.k.a. “dark stars”

DM heating and capture rates

Higher DM density greater effect on gas and stars

1. DM heating delayed protostellar contraction prolonged accretion (M* reaches 105 M?)

2. DM capture by MS star burn DM instead of hydrogen prolonged stellar lifetime (to z=0?)

But will this work in a Pop III MULTIPLE system?

New DM Initialization

1. Begin simulation immediately after the first sink has formed2. Align DM peak with main sink3. Continue simulation for 20,000 yrHow will these DM profiles evolve? Will density stay peaked?

Decline of DM Density on Stellar Disk Scales

10,000 yr

20,000 yr

10,000 yr

20,000 yr

Minimum DM density for DM capture to support a star

Sinks and DM density peak UNALIGNED

Rapid Decline of DM Effect on Gas and Stars

Blue – DM capture rate within sinks (-> DM density)Black – DM heating rate within sinks (-> DM density2)

Pop III Mass Growth Relatively Unaffacted

Sim A

Sim B

DM unrefined

Conclusions• Range of Pop III masses is likely very broad.

• Multiple mechanisms, particularly disk fragmentation, will contribute to formation of low mass stars.

• Fragmentation and broad mass range likely to describe Pop III stars even under radiative feedback! Possibly massive binaries

• Pop III stars can likely reach tens of solar masses, but hundreds of solar masses may be harder (see also Susa and Hosokawa’s talks)

Maybe explains why PISN signature has not been

observed (requires 140M < M* < 260M)

• Pop III multiplicity will strongly mitigate effects of DM annihilation - ‘dark stars’ unlikely (see also Smith’s poster and Iocco’s talk)

• Growing understanding of Pop III stars will ultimately increase physical realism of models of later star and galaxy formation

Questions?

THE END

Thank you!