Transcript Stan Woosley (UCSC)

The Death of Massive Stars and
the Birth of the Elements
Stan Woosley
Cornell, June 2, 2006
Stars are gravitationally
confined thermonuclear reactors.
  T3
25 M Presupernova Star
900 R
H, He
O, Mg, Ne
1R
He
Si, S, Ar, Ca
Fe
He, C
0.1 R
0.01 R
The key role of enropy…
Entropy
S/NAk
With each progressive burning
stage the central entropy decreases.
Red giant formation leads to an
increased entropy in the outer
hydrogen envelope.
Si
O
O, Ne, Mg, Si
Fe
S
Ne
H
C
He
Mg
C
C
collapses to
a neutron star
20 Solar Masses
Mayle and Wilson (1988)
bounce = 5.5 x 1014 g cm-3
1
2
3
4
5
6
0 ms
0.5
1.0
3.0
100 ms
230 ms
Some of the contributions of Hans Bethe to our understanding
of core collapse supernovae:
• The idea of a cold, low entropy bounce at nuclear
density –
Bethe, Brown, Applegate and Lattimer (Nucl. Phys. A , 324,
487, (1979))
G  exp(akT )
E  (akT )2
Fowler, Englebrecht, and Woosley
(1978) not withstanding
a ~A / 9
• The prompt shock model and its failure - Baron, Brown, Cooperstein
works if iron core mass is < 1.1 solar
masses (but it isn’t)
and Kahana (PRL, 59, 736,
(1987))
simplified models for the nuclear equation of state
• Delayed neutrino-powered explosions –
Bethe and Wilson, ApJ, 295, 14, (1985) - see also Wilson (1982)
Wilson discovered in 1982 that neutrino energy deposition
on a time scale longer than previously suspected could
re-energize the shock. Hans joined with Jim in the first
refereed publication to state this..
• Analytic Models for Supernovae.
“entropy”
“gain radius”
“net ram”
“foe” – now the “Bethe” = B
Bethe, RMP, 62, 801, (1990)
ApJ, 412, 192, (1993)
ApJ, 419, 197, (1993)
ApJ, 449, 714 (1995)
ApJ, 469, 192 (1996)
ApJ, 473, 343 (1996)
Nuc. Phys A, 606, 195 (1996)
ApJ, 490, 765, (1997)
Energy deposition here drives convection
Bethe, (1990), RMP, 62, 801
Velocity
Neutrinosphere
Neutron
Star
(see also Burrows, Arnett, Wilson, Epstein, ...)
gain radius
radius
 3000 km s 1
Infall

Accretion Shock;
Ram pressure
Inside the shock, matter is in approximate hydrostatic equilibrium.
Inside the gain radius there is net energy loss to neutrinos. Outside
there is net energy gain from neutrino deposition. At any one time there
is about 0.1 solar masses in the gain region absorbing a few percent
of the neutrino luminosity.
TSI/ORNL
LANL
Univ.
Arizona
MPI
No one has yet done a 3-dimensional
simulation of the full stellar core including
neutrino transport that the community
would agree is “good”.
But they are getting there and the attempts
are providing insights.
Scheck, Janka, et al (2006)
Scheck et al. (2004)
Stationary Accretion Shock Instability (SASI)
Supernova shock wave will become unstable.
Instability will
1. help drive explosion,
2. lead to gross asphericities.
New ingredient in the explosion mechanism.
Confirmed by:
 Scheck et al. 2004
 Janka et al. 2005
 Ohnishi et al. 2006
 Burrows et al. 2006
Buras et al. (2003) Physics
Livne et al. (2004) Physics
Blondin, Mezzacappa, and DeMarino (2003)
Parameterized neutrino heating/cooling.
see also Foglizzo (2001,2002)
Blondin, Mezzacappa, and DeMarino (2003)
SASI induced flow is remarkably self similar, with an aspect ratio ~2
that is consistent with supernova spectropolarimetry data.
Burrows et al. (2006) find considerable
energy input from neutron star vibrations –
enough even to explode the star, and
surely enough to influence the r-process
And so – maybe – most massive stars blow up the way
Hans and others talked about:
Rotation and magnetic fields unimportant
in the explosion (but might be important
after an explosion is launched)
Kicks and polarization from “spontaenous symmetry
breaking” in conditions that started spherical.
Just need better codes on bigger faster computers
to see it all work.
But ….
Dana Berry (Skyworks) and SEW
Need iron core rotation at death to correspond to a
pulsar of < 5 ms period if rotation and B-fields are to matter
at all. Need a period of ~ 1 ms to make GRBs.
This is much faster than observed in common pulsars.
Total rotational kinetic energy for a neutron star
E rot ~ 2  1052 (1 ms/P) 2 (R/10 km) 2 erg
j  R 2 ~ 6.31015 (1ms/P) (R/10 km) 2 cm 2 s-1 at M  1.4 M
For the last stable orbit around a black hole in the collapsar
model (i.e., the minimum j to make a disk)
jLSO  2 3 GM / c  4.6 1016 M BH / 3 M cm 2 s -1
non-rotating
jLSO  2 / 3 GM / c 1.5 1016 M BH / 3M cm 2 s -1
Kerr a = 1
Stellar evolution including approximate magnetic torques gives
slow rotation for common supernova progenitors.
Heger, Woosley, & Spruit (2004)
using magnetic torques as derived in
Spruit (2002)
Still faster rotation at death is
possible for stars born with
unusually fast rotation –
Woosley & Heger (2006)
Yoon & Langer (2005)
The spin rates calculated for the lighter (more common)
supernovae are consistent with what is estimated for
young pulsars
So, one could put together a consistent picture …
GRB – Dana Berry - Skyworks
Cas-A - Chandra
“slow” pulsar
90%(?)
fast pulsar/magnetar?
10%(?)
millisecond magnetar or
accreting black hole
(< 1%)
ROTATION
Ordinary SN IIp
- SN Ib/c - GRB
Nucleosynthesis
in Massive Stars:
Work with Alex Heger (LANL) and
Rob Hoffman (LLNL)
Survey - Solar metallicity:
•
Composition – Lodders (2003); Asplund, Grevesse,
& Sauval (2004)
• 32 stars of mass 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22
23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33
35, 40, 45, 50, 55, 60, 70, 80, 100, 120
solar masses. More to follow.
• Evolved from main sequence through explosion with
two choices of mass cut (S/NAkT = 4 and Fe-core) and
two explosion energies (1.2 B, 2.4 B) – 128 supernova
models
• Use the Kepler implicit hydrodynamics code
(1D)
(Weaver, Zimmerman, and Woosley 1978;
see RMP, 74, 1015, (2002) for description
of physics)
• Use best current information on nuclear reaction rates
and opacities
• Include best current estimates of mass loss at all stages of
the evolution
• Use recently revised solar abundances
Lodders, ApJ, 591, 1220 (2003)
Asplund, Grevesse, & Sauval, ASP Conf Series, (2004)
• The Explosion Model
The explosion can be
characterized by a piston
whose location and speed
are free parameters.
Mayle and Wilson (1988)
The piston location is
constrained by:
• Nucleosynthesis
• Neutron star masses
15 M
Electron mole number
Ye
The piston energy is
constrained by:
• Light curves
• Fall back
Except near the "mass cut", the shock temperature to which the
explosive nucleosynthesis is most sensitive is given very well by
4
 R3 aT 4  Explosion energy
3
 1051 erg  1B
Density
20 M presupernova
Entropy
Density (g cm-3)
Entropy per baryon (S/NAkT)
The edge of the iron core sets a
lower bound to the mass cut.
Otherwise, too many neutron-rich
isotopes …
The location where the entropy
S/NAkT = 4, typically at the base
of the oxygen shell sets an upper
limit. Stars that explode in real
simulations typically develop
their mass cut here. A larger value
gives neutron stars that are too
massive.
1.2 B explosions;
mass cut at Fe core
(after fall back)
1.2 B explosions
Above 35 M
black holes form
in Z=0 stars
Thorsett and Chakrabarty, (1999), ApJ, 512, 288
If in the models the mass cut is
taken at the edge of the iron core
the average gravitational mass for
for stars in the 10 – 21 solar mass
range is (12 models; above this black
holes start to form by fall back):
1.38  0.16 M
Ransom et al., Science, 307,
892, (2005) find compelling
evidence for a 1.68 solar
mass neutron star in Terzian 5
If one instead uses the S = 4
criterion, the average from 10 –
21 solar masses is
1.45  0.18M
From 10 to 27 solar masses the
average is
Vertical line is at
1.35  0.04 M
1.53  0.22 M
15 M
2.4 B
1.2 B
25 M
2.4 B
1.2 B
1.2 B of kinetic energy at
infinity gives good light curves
in agreement with observations.
2.4 B gives too bright a
supernova making Type II
almost as brilliant as Type Ia.
Though not shown here 0.6
B would give quite faint
supernovae, usually with
very weak “tails”.
Isotopic yields for 31 stars
averaged over a Salpeter
IMF, G = -1.35
Intermediate mass elements
(23< A < 60) and s-process
(A = 60 – 90) well produced.
Carbon and Oxygen overproduced.
p-process deficient by a
factor of ~2 for A > 130
and absent for A < 130
Conclusions
•
Overall good agreement with solar abundances
– see also WW95.
• Lightest neutron star 1.16 solar masses; average 1.4
solar masses. Black holes a likely product for some
current generation stars in the 30 – 50 solar mass range
(more black holes at metallicities lower than the sun)
• Overproduction of C and O suggests that current estimates
of Wolf-Rayet mass loss rates may be too large (and/or
Lodders (2003) abundances for C and O too small).
Two Mysteries
•
The nature of the r-process site
• The origin of the p-process 90 < A < 130
r-Process Site: The Neutrino-powered Wind *
Nucleonic wind
1 – 10 seconds
Anti-neutrinos are "hotter" than
the neutrinos, thus weak equilibrium
implies an appreciable neutron excess,
typically 60% neutrons, 40% protons
favored
at late times
T9 = 5 – 10
T9 = 3 - 5
T9 = 1 - 2
Results sensitive to the (radiation)
entropy, T3/, and therefore
to aT4/, the energy density
Duncan, Shapiro, & Wasserman (1986), ApJ, 309, 141
Woosley et al. (1994), ApJ, 433, 229
The r - process is favored by high entropy, i.e., low
density at a given temperature, because the reactions that
assemble  - particles to heavy ( seed ) nuclei between 3
and 5 billion K increase rapidly with density
 ( n,  )7 Be( , n)12 C
 2
Keeping the density low thus keeps most of the mass
in  - particles and thus increases the ratio of free neutrons
to heavy seed .
Woosley et al (1994)
Integrated abundances in the late time wind resemble the
r-process abundance pattern.
But,
The entropy in these calculations by Wilson was not replicated
in subsequent analyses which gave s/kB about 4 times smaller
s/kB ~ 80 not 300.
Throughout nucleosynthesis epoch
L  L
Early on (t  about 1/2 second)
  
Late (t >1 s)
  
Wilson (1994)
With nearly equal fluxes of neutrinos,
each having about the same energy,
and with the lifting of degeneracy
the neutron-proton mass difference
favors protons in weak equilibrium.
Later, the neutrino energy difference
favors neutrons.
QIAN AND WOOSLEY (1996)
t ~ 10 s
t<½s
data from Wilson (1994)
Janka, Buras, and Rampp (2003)
15 solar mass star – 20 angle averaged trajectories
The neutrino-assisted rp-process
0.36 s
Froehlich et al (2005)
Pruett, Hoffman, and Woosley (2005)
T9  2.05
5  0.27
s / kb  77
Ye  0.562
X p  0.124
X   0.844
X n 1013
Unmodified trajectory number 6 from Janka et al.
Entropy times two
Summary – Neutrino Wind
• Very rich site for nucleosynthesis – about half
of all the
isotopes in nature are made here
• Can produce both (part of) the p-process and the
r-process nearly simultaneously in one site.
• In both cases, the nucleosynthesis suggests a higher
entropy (and outflow with more internal energy per baryon)
than traditional models have provided
•Nuclear physics very uncertain – need RIA
• This has important implications for the explosion model
Energy input from: (Qian & Woosley 1996)
vibrations - Burrows et al (2006)
Alfven waves or reconnection – Suzuki and Nagataki (2005)
magnetic confinement – Thompson (2003)