On the Progenitors of Collapsars

Alexander Heger & Stan Woosley

Department of Astronomy and Astrophysics, University of California, Santa Cruz, CA 95064

Introduction

nermost 1.6 M of the collapsing star would produce

a neutron star of 1.45 M gravitational mass, while
conserving the total angular momentum contained
in the pre-collapse model. The neutron star was assumed to rotate rigidly.
Table 1: Presupernova properties of (initially) 15 M helium

tation rates of massive stellar cores when they

collapse?

Progenitor Models

Figure 3: Equatorial angular momentum (thick black line)

of a 15 M helium core of initially 10 % Keplerian rotation

as a function of the interior mass coordinate, m. The blue
line shows the specic angular momentum required for a test
particle at the last stable orbit around a Schwarzschild black
hole of mass equal to the mass coordinate, m. The green line
shows the same for a extreme Kerr black hole (spin parameter
a = 1).

e simulate the evolution of bare helium cores

with an initial mass of 15 M . Two dierent initial
surface rotation rates, 10 % and 30 % of Keplerian
rotation, are investigated. The former may be either
the result of a massive single star ( 40 M ) that
has lost its envelope early during helium burning or
a close binary that lost its envelope to a companion. The latter might result from a binary merger.
The evolution of the helium core and its rotation is
followed as described by Heger, Langer, & Woosley
(2000). Fine surface zoning is employed.
We investigate two dierent evolutionary paths: The
rst neglects mass loss while in the second we take
it into account.

We use the WR mass loss rate given by Wellstein & Langer (1999), but
reduce it by a factor 3 to account for eect of \clumping" (Hamann
& Koesterke 1998). We take an initial stellar metallicity of 1/10 solar and assume a that the WR mass loss rate scales as the square
root of metallicity, reducing the mass loss rate by an
 additional 1factor

_ =M yr
of 3. In total, we assume a mass loss rate: log M
=
10:95 + 1:5 log (L=L ). This is equal to  1:710 5 M yr 1
when the star dies.

Results

e present the rst progenitor models for collapsars that follow the evolution of the angular momentum in the stellar interior till the onset of core collapse.

{
{
{
{
yes
yes
yes
yes

10
10
30
30
10
10
30
30

{
yes
{
yes
{
yes
{
yes

0.09
0.23
0.06
0.18
0.45
3.4
0.26
1.9

mass coordinate
ranges for which
an equatorial
disk could form
(M )

BH (3 M )

BH (4 M )
a

Two essential ingredients for the collapsar model are

a suciently large core to form a black hole and
a sucient rotation rate at the time of collapse to
allow the formation of a disk. The question we address here is: What can be expected for the ro-

magnetic elds
rotation
(% Keplerian)
mass loss

The traversal time for the relativistic jet through typical hydrogen envelopes of massive stars is hundreds to thousands of seconds. Thus, at
the time of the GRB, bare helium stars which have radii of only a few
light seconds (about a solar radius) are necessary if the lifetime of the
engine is to match approximately the burst duration and not be short
compared to the time it takes the jet to drill through the star.

BH (2.5 M )

, at onset of core collapse.

BH (2 M )

coordinate,

Figure 2: Angular velocity as a function of the interior mass

core models. The rst three columns dene the initial model

and physics employed (magnetic elds according to Spruit
(2001), amount of rotation, mass loss by winds). Next we
give the period a pulsar would have if it formed in this star,
then the non-dimensional spin parameter, aBH , a black hole
would acquire, if all the angular momentum below the mass
coordinate indicated were to go into the black hole of that
mass (the upper limit for aBH is 1; here we show in brackets
the values that would formally result solely to give a measure
for the angular momentum available in the model), and in the
last column we show the mass ranges in which the equatorial
mass could form a centrifugally supported accretion disk
around a central compact object.

ne of the most promising models for Gamma-Ray

Bursts (GRBs), especially of the \long variety", is
the so-called collapsar model (Woosley 1993). This
model assumes that a suciently massive stellar core
collapses into a black hole and the infalling outer layers form a disk around it. Energy dissipated in the
disk or the rotation of the black hole itself is assumed
to power a jet of high Lorenz factor (  100) that
escapes from the engine along the polar axis to large
distance ( 1015 cm) and powers a GRB by interaction with the circumstellar medium or by internal
shocks.

pulsar period
(ms)

Calculations that used the eects of magnetic elds

as described by Spruit & Phinney (1998; not shown
here) resulted in much too low core rotation for the
collapsar model of GRBs.

I
Though mass loss signicantly reduces the angular
momentum at core collapse (see Figures 3 and 4),
enough remains in the equatorial regions of the star
to form a centrifugally supported disk.

n a suciently close binary system the envelope can

be tidally locked to the orbit of the star. Depending
on separation and mass ratio, this rotation can reach
several 10 % of Keplerian. In Figure 6 we show that
maintaining 10 % Keplerian surface rotation are can
be sucient for collapsars. Employing the dynamo
process by Spruit (2001), however, the resulting presupernova rotation is too slow even for a system with
20 % Keplerian rotation.

Magnetic Fields

e also did test calculations employing magnetic

eld torques that might result from a dynamo process recently developed by Spruit (2001), as well as
calculations following the approximation of Spruit &
Phinney (1998). The torques lead to considerable
spin-down of the core, in particular when combined
with wind mass loss.

Figure 6: Final angular momentum distribution for close bi-

nary scenario. The outer 1 M of the star is kept at 10 %

Keplerian rotation from helium ignition till core collapse.

Case C Binary Mergers

Figure 1: Angular momentum in the equatorial plane as a

function of the interior mass coordinate, m, at dierent evolutionary stages: green: helium ignition (our initial model),
red: end of central helium burning, magenta: carbon ignition,
blue: carbon depletion, cyan: oxygen depletion, and black:
onset of core collapse.

Rotation Prole and Disk Formation

o decide whether a centrifugally supported accretion disk can form around a central black hole (which
we assume to emerge directly or by fallback), we
compare with the angular momentum a test particle would require at the last stable orbit around a
Schwarzschild or Kerr black hole (see Figures 3 and
4).

Figure 5: Same as Figure 4, but an initial rotation rate of

30 % Keplerian rotation was assumed and angular momentum transport by magnetic elds according to a prescription
by Spruit (2001) was included.

For comparison, preliminary calculations (Heger,

Woosley, Spruit) of complete stellar models of 15,
20, and 25 M stars with initial surface rotation velocities of 200 km s 1 were followed from hydrogen
burning until core collapse using the same prescription for torques as suggested by Spruit. They resulted in rotation periods for pulsars at birth of 7.65,
5.50, and 3.99 ms, respectively. Without magnetic
elds but treatment of rotation according to Heger
et al. (2000), the same calculations gave rotation periods of 0.2 ms in all these stars (which is, however,
beyond breakup rotation). We assumed that the in-

gle stars may already retain sucient angular momentum for collapsar progenitors. In the Case C
merger scenario probably too much angular momentum would remain to allow the formation of an ecient accretion disk.

Decoupling of Core Rotation

t what evolution stage needs the core rotation

decouple from the surface? Figure 8 shoes that this
need to happen before central carbon burning { assuming angular momentum is locally conserved from
this time on and no further transport occurs. In the
model star of Figure 8 the decoupling would need
to happen no later than when a central density of
10,000 g cm 3 is reached.

(2.5)(3.4)(3.8)(3.7)
0 15
(1.1) 0.90 0.98 (1.2) 0 2:5, 2:7 8:6
(4.1)(4.8)(5.6)(5.9)
0 15
(1.7)(1.3)(1.3)(1.7)
0 8:4
0.58 0.76 0.80 0.88 2:4 2:8, 5:7 15
0.08 0.09 0.11 0.10
8:5 8:8
(1.2)(1.4)(1.4)(1.5)
0 15
0.16 0.14 0.16 0.17
7:4 8:8

Co-rotating Close Binaries

Figure 4: Same as Figure 3, but mass loss due to stellar

winds is included. The red line shows the specic angular momentum a test particle requires at the last stable orbit around
a black hole that has formed with the mass and integrated angular momentum below the given mass coordinate. Where the
red line is missing, sucient angular momentum is available
to form a Kerr black hole.

Without magnetic elds, as shown above, even sin-

o simulate a Case C merger of a binary system (a

merger after or at the end of central helium burning) we study a model where we assumed rigid rotation with 50 % Keplerian surface rotation at core
helium depletion and employ the dynamo process of
Spruit (2001). We nd (Figure 7) that even with
this extreme assumption on rotation magnetic stress
removed too much angular momentum from the core
for the collapsar model to work. The magnetic stress
kept the star in rigid rotation till carbon burning.

Figure 7: Presupernova specic equatorial angular momen-

tum distribution in a Case C merger scenario, using the dynamo process of Spruit (2001).

Figure 8: Average specic angular momentum distribution for

a rigidly rotating 15 M star with Keplerian surface rotation

at dierent evolution stages. The specic angular momentum
has been scaled by m 2=3 to remove the singularity at m = 0
from the plot. Dotted and dashed lines show the specic angular momentum required for the last stable orbit around a
Schwarzschild and a Kerr black hole

Conclusions

bare helium star of low metallicity can retain

enough angular momentum to form a centrifugally
supported disk around a central black hole of 
3 M, as required by the collapsar model for GRBs.
Without the inclusion of magnetic elds, the angular
momentum is sucient to form a Kerr black hole and
support most or all of the star in an accretion disk.
However, if we include current approximative treatments of angular momentum transport by magnetic
elds, the resulting spin rates become too low to
form centrifugally supported disks in the inner part
of the core. We speculate that even a binary helium star merger at the end of central helium burning might not be able to avoid this fate. Mass loss
can lead to an additional signicant spin-down of
the core, especially if magnetic elds couple it effectively to the envelope. Even in case of Keplerian
surface rotation, the core rotation needs to decouple before carbon ignition in order to make a Kerr
black hole. The dynamo process recently proposed
by Spruit (2001) seems too ecient to form collapsar
progenitors from single stars or helium star mergers.
This is even more so for the magnetic eld modeling
suggested by Spruit & Phinney (1998).
Acknowledgements: We are grateful to Henk Spruit for a

preview of his work while it was still in progress and many

very helpful discussions. This work has been supported by
the NSF (AST-9731569), NASA (NAG5-8128), the DOE
(B347885), and the AvH (FLF-1065004).

References
Hamann, W.-R., Koesterke, L. 1998, A&A, 335,
1003
Heger, A., Langer, N., Woosley, S.E. 2000, ApJ, 528,
368
Heger, A., Woosley, S.E., Spruit, H.C. 2002, ApJ, in
preparation
Spruit, H.C., Phinney, E.S. 1998, Nature, 393, 139
Spruit, H.C. 2001, accepted by A&A; astroph/0108207
Wellstein, S., Langer, N. 1999, A&A, 350, 148
Woosley, S.E. 1993, ApJ, 405, 273

mailto: alex@collapsars.org, woosley@collapsars.org