rotation, supernovae and gamma-ray bursts · grb 050709 0.060 0.229 0.161 late dwarf grb 050724...
TRANSCRIPT
Rotation, Supernovae
and Gamma-Ray Bursts
When Massive Stars Die,
How Do They Explode?
Neutron Star
+
Neutrinos
Neutron Star
+
Rotation
Black Hole
+
Rotation
Colgate and White (1966)
Arnett
Wilson
Bethe
Janka
Herant
Burrows
Fryer
Mezzacappa
etc.
Hoyle (1946)
Fowler and Hoyle (1964)
LeBlanc and Wilson (1970)
Ostriker and Gunn (1971)
Bisnovatyi-Kogan (1971)
Meier
Wheeler
Usov
Thompson
etc
Bodenheimer and Woosley (1983)
Woosley (1993)
MacFadyen and Woosley (1999)
Narayan and Piran (2004)
etc
All of the above?
a question that is at
least 65 years old
(Baade and Zwicky 1939)
Ian Strong – left Ray Klebesadel – right
September 16, 2003
Gamma-ray bursts (GRBs) discovered 1969 - 72 by Vela
satellites. Published by Klebesadel, Strong and Olson (1973)
GRBs come in at least two flavors: “SHB”s and “LSB”s
Paciesas et al (2002)
Briggs et al (2002)
Koveliotou (2002)
Shortest 6 ms
GRB 910711
Longest ~2000 s
GRB 971208
“Short Hard Bursts”
Burst duration E50 z host
(sec)
GRB 05009B 0.032 0.0275 0.225 Ellipitical
GRB 050709 0.060 0.229 0.161 Late dwarf
GRB 050724 0.20 1.0 0.258 Early galaxy
GRB 050813 0.35 1.7 0.719 Elliptical
Prochaska et al (2006)
Star formation rate in hosts < 0.3 solar masses yr-1
(fac 10 less than LSBs)
Burst energy at least one order of magnitude less
Hard to make in massive star models
People think most SHBs are
merging compact objects
LSBs ~ 20 > 5 ~ 2 SF galaxies
GRB070201
Intense short hard burst observed by
Konus-Wind, Integral, and Swift
1045 erg if in Andromedabetween March 5, 1979 (2 x 1044 erg)
and SGR 1806-20 giant flare (2 x 1046 erg)
Error box 0.325 sq deg 1.1 deg from center of Andromeda
Previous bright short burst near M81
No gravitational radiation (LIGO)
peak flux 10-3 erg cm-2 s-1
• Most bursts discovered so far (though not necessarily
per fixed volume) are LSB’s at cosmological distances.
Today’s talk focuses on these.
As of April, 2008
131 bursts
SWIFT gives
an average z
about twice as
great and the
farthest GRB,
so far, is at
z = 6.3
www.astro.ku.dk/~pallja/GRBsample.html
Most distant so far is
GRB050904 z = 6.29
Frail et al. (2001)
As a relativistic jet decelerates we see a
larger fraction of the emitting surface until
we see the edges of the jet. These leads to
a panchromatic break slope of the
the afterglow light curve.
LSBs are beamed and their total energy in relativistic
ejecta is ~1051 erg.
Fruchter et al (2006)
The green circles show
GRB locations to an
accuracy of 0.15 arc
sec.
Conclusion: GRBs trace star
formation even more than the
average core-collapse supernova.
They are thus to be associated
with the most massive stars.
They also occur in young, small,
star forming galaxies that might
be metal poor.
LSBs occur in star-forming regions
Pian et al. (2006)
In general, these 4 events:
• Are lower in overall gamma-ray and x-ray
energy than typical LSBs (except for 030329)
• Have a spectrum just as hard as ordinary LSBs
(except for 060218)
• Can be made by jets with ~ 10 - 30 (Kaneko et al. 2007)
• Have much more energy in non-relativistic ejecta than in
the burst
All SN Ic - BL
Environmental Clue
Kelly, Kirshner, & Pahre (2008, submitted)
SN Ic-bl
SN Ic
LGRB
Local SN Ic and GRB have similar locations
compared to host galaxy light
Fruchter etal (2006)
-Similar (large)
progenitor
masses
Fynbo et al (2006)
Gal-Yam et al (2006)
Della Valle et al (2006)
But not all LSBs (or at least LBs) have bright supernovae
Perhaps another mechanism
at work (WD+BH?), or
the SN did not eject any56Ni.
Observations suggest that Z is low!
Fynbo, Jakobsson, & Moeller, et al, 2003, A&ApL, 406, L63
“On the Ly-alpha emission from gamma-ray
burst host galaxies: evidence for low metallicities”
Gorosabel J., Perez-Ramirez D., Sollerman J., et al.,
2005, A&A, 444, 711
“The GRB 030329 host: a blue low metallicity subluminous
galaxy with intense star formation”
Sollerman J., Ostlin G., Fynbo J. P. U., et al.,
New Astronomy, 2005, 11, 103
“On the nature of nearby GRB/SN host galaxies”
Fruchter et al. 2006, Nature, 441, 463 “Long -ray bursts and core-collapse supernovae have
different environments”
Stanek et al., 2006, Acta Astron., 56, 333 “Protecting Life in the Milky Way: Metals Keep the GRBs Away”
Local abundances of GRB-SN and
broad-lined SN Ic
Local SDSSgalaxies(Tremonti etal 2004)
Modjaz et al (2008)
SMC
LMC
Maiolino et al (2008)
astroph0806-2410
AMAZE Survey
ESO-VLT
Z ~ 2 - 3 is an era of
intense metallicity
evolution
Metallicity in low M
galaxies rises slower
than in high M
• How common are SN Ib/c? Local rate:– ~15-20% of all SN
– ~30% of CC-SN
– Broad-lined SN Ic (SN Ic-BL): ~5-10% of all SN Ib/c(Cappellaro et al 1999, Guetta & Della Valle 2007, Leaman et al. in prep)
So SN Ic-BL are 1 - 2% of all supernovae.
GRBs are a much smaller fraction
Not all SN Ic - BL are GRBs
(though they still may be “active” at some level.
Madau, della Valle, &
Panagia, MNRAS, 1998
Supernova rate per 16
arc min squared per year
~20
This corresponds to an
all sky supernova rate
observed at the earth of
6 SN/secFor comparison the
universal GRB rate is
roughly 3 /day * 300 for
beaming or
~ 0.02 GRB/sec
The rate at which massive stars die in the universe is very
high and GRBs are a small fraction of that death rate.
GRB
SN0.3%
Kocevski and Butler (2008)Some detection bias but note
absence of nearby energetic bursts
To Summarize the Observations -
For Common “Long Soft GRBs” (LSBs)
• LSBs are the deaths of massive stars. Most of the ones
studied so far are typically at redshift z >~ 1
• The stars involved may be a rare subset of those stars that make
Type Ic-BL supernovae ~~10%. Most of those that do make
GRBs are not pointed at us
• Many if not most LSBs have bright supernovae that go
along with them but the supernovae sampled so far are
not necessarily a representative sample of (distant) LSBs
• LSBs are favored by low metallicity
• LSBs in general require the production of a relativistic jet
of matter, fields and radiation ( ~ 100; E ~ 1051 erg)
The Place of LSBs in Stellar Evolution
• The observations favor (in fact were predicted by) a
model in which LSBs occur with increasing violence
in stars of lower metallicity
• The conditions that distinguish a LSB from
and ordinary core-collapse supernova are: a) the loss of
the star’s hydrogen envelope, and b) rapid rotation
(to be demonstrated). In fact here, low Z is a surrogate
for high
Today, there are two principal models being discussed
for GRBs of the “long-soft” variety:
• The collapsar model • The millisecond magnetar
Need iron core rotation at death to correspond to a
pulsar of < 5 ms period if rotation and B-fields are to matter
to the explosion. Need a period of ~ 1 ms to make GRBs.
This is much faster than observed in common pulsars.
52 2 2
ro
12 5
t
2 2 -1
E ~ 2 10 (1 ms/P) (R/10 km) erg
j
Total rotational kinetic energy for a neutr
~ (1ms/P) (R/10 km)
on st
cm s at M 1.4 M
a
0
r
6.3 1R=
For the last stable orbit around a black hole in the collapsar
model (i.e., the minimum j to make a disk)
1 2 -1
2
6
11 -6
2 3 / / 3 cm s non-rotatin4.6 10
1
g
2 / 3 / / 3 cm s Kerr a = .5 110
LSO BH
LSO BH
j GM c M M
j GM c M M
= =
= =
It is somewhat easier to produce a magnetar model!
Heger, Langer, & Woosley (2002)
This is plenty of angular
momentum to make either a
ms neutron star or a
collapsar.
Calculations agree that without
magnetic torques it is easy to make
GRBs – too easy.
Similar results by Hirishi, Meynet, &
Maeder (2004)
Spruit (2002)
Braithwaite (2006)
Denissenkov and Pinsonneault
(2006)
Zahn, Brun, and Mathis
(2007)
Torque BrB
B from differential winding
Br from Tayler-Spruit dynamo
"Any pulely poloidal field should be unstable to instabilities
on the magnetic axis of the star" (Tayler 1973)
Approximately confirmed for
white dwarf spins (Suijs et al
2008)
15 solar mass helium core born rotating rigidly at f times break up
If include WR mass loss and magnetic
fields the answer is greatly altered....
with mass loss with mass loss and B-fields
no mass loss or
B-field
15 M rotating helium star
Heger, Woosley, & Spruit (2004)
using magnetic torques as derived in
Spruit (2002)
Stellar evolution including approximate magnetic torques gives
slow rotation for common supernova progenitors. (solar metallicity)
magnetar
progenitor?
This is consistent with what is estimated for
young pulsars
from HWS04
Implication:
Rotation and magnetic fields are a minor (though
perhaps interesting) component in the explosion of
most supernova (Type IIp and Ib)
Notes:
• Ott et al. (ApJS, 2006) - 2D simulations of collapse,
bounce and post-bounce.
“… the magnitude of the precollapse iron core angular
velocities is the single most important factor in determining
the PNS spin.”
• Recent survey of Ib progenitors by Yoon and
Woosley gives P(n*) = 10 to 30 ms.
Much of the spin down occurs as the star evolves from
H depletion to He ignition, i.e. forming a red supergiant.Heger, Woosley, &
Spruit (2004)
solar metallicity
1 2 -1
2
6
11 -6
2 3 / / 3 cm s non-rotatin4.6 10
1
g
2 / 3 / / 3 cm s Kerr a = .5 110
LSO BH
LSO BH
j GM c M M
j GM c M M
= =
= =
Aside:
If the core of a massive star collapsed to a black hole, could
it simply disappear?
Probably not. Either a faint SN if envelope comes off
or a transient ULX if it doesn’t.
Possibly relevant
ApJ, 652, 518 (2006) - McClintock et al.
(spectral analysis of x-ray continuum)
Extreme spin of black hole in a Galactic microquasar
GRS 1916+105 a > 0.98
Two others quite high
Spin natal, not acquired by later accretion,
but mass ~14 solar masses.
See also Liu et al (2008) M33-X7 a = 0.77
But must all massive stars pass through
a red or blue supergiant phase?
Consider stars with the upper 5% of rotation
on the main sequence and standard rotational
mixing parameters.
Further restrict our attention to low metallicity.
Woosley & Heger (2006)
Yoon and Langer (2005)
Yoon, Langer, and Norman (2006)
earlier work by Maeder
-1v 400 km s
eq
H He
C,Ne
O
O
Si
C,Ne
He –burning
C,O produced
never a red giantvrot = 400 km/s
R = 4.8 x 1010 cm
L = 1.9 x 1039erg s-1
0.86
-6 -1 ZM = 2.4 x 10 M yr
0.01 Z
WO-star16 M
main sequence
Vink & de Koter (A&A, 442, 587, (2005))
M(WC) 10 M
M(WN) = 20 M
=
The mass loss rate can be quite low!
A typical He-burning lifetime is 0.5 My.
0.86M Z
Z is ZFe
1/2
1% solar metallicity
M Z
Solar Metallicty
(became RSG)
= 400 km/srot
v
= 400 km/srot
v
H
He-depl
C-depl
PreSN
8 ms pulsar
GRB
1% solar metallicity
10% solar metallicity
Solar metallicity
Yoon, Langer,
and Norman (2006)
Woosley and Heger (2006) find similar results but estimate a
higher metallicity threshold (30% solar) and a higher mass
cut off for making GRBs.
i..e., 1/8 solar
NGRB / NSN << 1%
out to redshift 4
saturates at 2% at
redshift 10
Caveats:
• Magnetic torques (Spruit) uncertain. Certainly
the final angular momentum could be off by
a factor of 2 or more.
• Metallicity means iron in the vicinity of the GRB,
not CNO averaged over the galaxy
• If more mass is lost along the polar axis as
theory suggests, higher metallicities can be
tolerated
• Rotation requirements for making a GRB may
have been overestimated for the millisecond magnetar
model
But how are the GRBs made in
a rapidly rotating massive star?
Dana Berry (Skyworks) and SEW
The Collapsar Model: Basics
•Wolf-Rayet Star – no hydrogen envelope – about 1 solar radius.
• Collapse time scale tens of seconds
• Rapid rotation – j ~ few x 1016 erg s
• Black hole ~ 3 solar masses accretes about a solar mass
Woosley (1993)
MacFadyen & Woosley
(1999)
In the vicinity of the rotationalaxis of the black hole, by avariety of possible processes,energy is deposited.
7.6 s after core collapse;high viscosity case.
The star collapses and forms a disk (log j > 16.5)
Zhang (2008, in progress)
MacFadyen and Zhang, in preparation (this calculation by A. MacFadyen)
Model T16A3 with
jet and disk wind
included parametrically
Slide from N. Bucciantini
Bucciantini, Quataert, Arons, Metzger and
Thompson (MNRAS; 2007) and refs
therein
Assume a pre-existing supernova
explosion in the stripped down core
of a 35 solar mass star.
Insert a spinning down 1 ms magnetar
with B ~ 1015 gauss.
Two phase wind:
Initial magetar-like wind contributes to
explosion energy. Analog to pulsar wind.
Sub-relativistic
Later magnetically accelerated neutrino
powered wind with wound up B field
makes jet. Can achieve high field to
baryon loading.
Density Pressure
4 s
5 s
6 s
Generic Jets
The jet must
maintain its direction
(< 3o) as well as its
energy input for at least
20 s or the jet dies
and there is no GRB.
No GRBs expected from
red or blue supergiants
or from central engines
that precess.
Jets precessed on a period of 2 s with half angle 3, 5, and 10 degrees
Zhang et al (2004)
3D studies of relativistic jets
by Zhang & Woosley (2008, in prep.)
As the energy of the jet is turned
down at the origin, the jet takes an
increasingly long time to break out.
The cocoon also becomes smaller
and the jet more prone to instability.
Jets were inserted at 1010 cm in a WR star with
radius 8 x 1010 cm. Jets had initial Lorentz factor
of 5 and total energy 40 times mc2.
Going to ever lower jet
energies, eventually, one
must confront the
possibility that the jet
either “dies” before
breaking out or shortly
after break out.
Relativistic input turned
off at the base at 22 s.
all calculations by
Weiqun Zhang
Maximum Lorentz factor
here (red) is about 5
or …
The supernova explodes
before the jet breaks out.
If the jet can stay on for
100’s of seconds a bright
transient may still emerge.
SN shock
1048 erg jet introduced following
a 1051 erg isotropic explosion.
Calculations on a 2275 x 2275
cylindrical grid
Implication:
Low energy jets may fail to emerge or emerge with low
Lorentz factors. One still gets a very asymmetric explosion and
possibly intermediate Lorentz factors ( ~ 10 instead of ~ 100)
Common case in stars with less rotation - e.g., nearby
GRBs with nearly solar metallicity?
Can we distinguish the collapsar
and magnetar models?
The biggest shortcoming of the collapsar model is
the large amount of angular momentum necessary to
its operation. The physics of black hole formation
also needs clarifying.
The biggest problem for the magnetar model is the
first second of evolution of a collapsing very
massive star. The accretion rate is about a solar
mass per second. What turns it around and how is
the 56Ni made?
Diagnostics:
• Low mass progenitor (e.g., Mazzai et al 2006)
• Late time magnetar activity
Generally one expects neutron stars in low mass progenitors
and black holes in heavier ones, but can we really infer the
mass of such an asymmetric explosion and is it impossible to make
a GRB with a 20 solar mass progenitor?
If a magnetar is made, one might expect energetic soft gamma-ray
burst repeater activity for years afterwards. GRB 070201 might
(barely) have been visible at the distance of GRB 980425
pulsar collapsar
Makes 56Ni MHD blast Disk wind
neutrinos?
Bright SN Deposition of Typical accretion rates
~1052 erg in 1 sec make 56Ni and not
(hard?) other Fe isotopes
Faint SNa Slow energy Fall back or “unusual”
depositionb, accretion physics
fall back
b But then need some other mechanism to explode the star. MRI?
a possibly GRB 060505 or GRB060614
56Ni
There are two ways 56Ni might be produced.
• A quasi-spherical blast from a millisecond
magnetar, or
• A wind off a disk where temperatures greater
than 5 x 109 K are achieved.
Suppress either one and you lose the 56Ni
The disk wind: MacFadyen & Woosley (2001)
Neglecting electron capture in the disk
Also recent work by Narayan and Piran (2004), Igumenshchev et al (2003), “NDAF”
GRB 070110 - Troja et al. (2007)
z = 2.35
x-ray plateau
~ 20,000 sec
~ 1052 erg
Flares.
log t 2 3 4 5 6
GRB 050904 - SWIFT
redshift 6.29
One of the most
luminous and long
lasting GRBs ever
observed.
Complex light curve.
Explosive outbursts
continued for over two
hours.
Eiso = 3.8 x 1053 erg
T90 = 225 s
Continuing pulsar activity or fall back?
MacFadyen, Heger,
and Woosley (2001)
25 M star
8 M helium core
see also Kumar, Narayan,
and Johnson (2008, in prep)