& their NucleosynthesisWilliam Raphael Hix (ORNL/U. Tennessee)
Questions about Core-Collapse Supernovae
Grefenstette, Harrison, Boggs, Reynolds, … 2014
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
What is a CCSN model?Our code, CHIMERA, has
Spectral Neutrino Transport (MGFLD-TRANS, Bruenn) in Ray-by-Ray ApproximationShock-capturing Hydrodynamics (VH1, Blondin)Nuclear Kinetics (XNet, Hix & Thielemann)Plus Realistic Equations of State, Newtonian Gravity with Spherical GR Corrections.
Other models use a variety of approximationsSelf-consistent models use full physics to the center.Parameterized models replace the core with a specified neutrino luminosity. Leakage & IDSA models simplify (oversimplify?) the transport. Ray-by-Ray Approximation
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Self-Consistent Models using Discrete Ordinates, VTEF, M1 and MGFLD can produce quite similar results when used
in one dimensionwith limited opacities & EOS.
Can we agree on anything?
0
20
40
60
80
100
Elec
tron
neut
rino
lum
inos
ity (1
051 e
rgs/
s)0
20
40
60
80
100
Elec
tron
anti-
neut
rino
lum
inos
ity (1
051 e
rgs/
s)
10 100 10000
20
40
60
Radius (km)
/ n
eutri
no lu
min
osity
(1051
erg
s/s)
10
100
Elec
tron
Neu
trino
RM
S en
ergy
Agile-Boltztran (Liebendörfer et al 2005)
CoCoNUT-VERTEX (Müller et al 2010)PROMETHEUS-VERTEX Pot. A (Müller et al 2010)
VERTEX (Liebendörfer et al 2005) =
PROMETHEUS-VERTEX Pot. R (Müller et al 2010)
GR1Dv2 (O'Connor 2015)
CHIMERA C (Bruenn et al 2016)
10
100
Elec
tron
Anit-
Neu
trino
RM
S en
ergy
(MeV
)
10 100 100010
100
Radius (km)/
Neu
trino
RM
S en
ergy
(MeV
)
0 50 100 1500
50
100
150
200
Time (ms after bounce)
Shoc
k Ra
dius
(km
)
Agile-Boltztran (Liebendörfer et al 2005)
CoCoNUT-VERTEX (Müller et al 2010)
PROMETHEUS-VERTEX Pot. A (Müller et al 2010)VERTEX (Liebendörfer et al 2005) = PROMETHEUS-VERTEX Pot. R (Müller et al 2010)GR1Dv2 (O'Connor 2015)
CHIMERA C (Bruenn et al 2016)
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Success should be determined by comparison to observations, but at what level of completion?Shock velocity reaching 105 km/s?Explosion energy (or surrogate) reaching ~1B?Ejecting ~0.1 M☉ of Nickel?Looking like Cas A?
What is a Completed Model?
Time after Bounce [ms]
0
Explo
sion
Ener
gy [B
] B12-WH07B15-WH07B20-WH07B25-WH07
0 200 400 600 800 1000 1200 1400
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6E+ = Energy sum over positive energy zonesE+
ov = E+ + Overburden
E+ov,rec = E+
ov + Nuclear recombination
10 15 20 25ZAMS Progenitor Mass [M☉]
0
0.02
0.04
0.06
0.08
0.1
0.12
56Ni
Mas
s [M ☉]
SN 2012awSN 2004A
SN 2004dj
SN 2004et
SN 1993J
SN 1987A
SN 2004cs
0 200 400 600 800 1000 1200 1400Time after Bounce [ms]
0
5000
10000
15000
20000
Shoc
k Ra
dius
[km
]
B12-WH07B15-WH07B20-WH07B25-WH07
0 200 400 600 800 1000 1200 14000
5000
10000
15000
20000
mean shock radiusmaximum shock radiusminimum shock radius
Bruenn, Lentz, Hix … (2016)
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Even in our most fully developed model, the explosion energy has not leveled off 1.3 seconds after bounce.The reason is that accretion continues at an appreciable rate, showing no sign of abating.
When does the Explosion End?
0 200 400 600 800 1000 1200 1400Time from Bounce [ms]
0.01
0.1
1
10
Mas
s Acc
retio
n Ra
te [M
O. s-1
] B12-WH07B15-WH07B20-WH07B25-WH07
Mass Accretion Rate Through the Gain Layer vs Post-Bounce Time12 - 25 W-H Progenitors
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Even in our most fully developed model, the explosion energy has not leveled off 1.3 seconds after bounce.The reason is that accretion continues at an appreciable rate, showing no sign of abating.
When does the Explosion End?
0 200 400 600 800 1000 1200 1400Time from Bounce [ms]
0.01
0.1
1
10
Mas
s Acc
retio
n Ra
te [M
O. s-1
] B12-WH07B15-WH07B20-WH07B25-WH07
Mass Accretion Rate Through the Gain Layer vs Post-Bounce Time12 - 25 W-H Progenitors
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Even in our most fully developed model, the explosion energy has not leveled off 1.3 seconds after bounce.The reason is that accretion continues at an appreciable rate, showing no sign of abating.This extends the “hot bubble” phase and suppresses the development of the PNS wind.
When does the Explosion End?
0 200 400 600 800 1000 1200 1400Time from Bounce [ms]
0.01
0.1
1
10
Mas
s Acc
retio
n Ra
te [M
O. s-1
] B12-WH07B15-WH07B20-WH07B25-WH07
Mass Accretion Rate Through the Gain Layer vs Post-Bounce Time12 - 25 W-H Progenitors
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
What is 2D good for?In both 2D and 3D, explosions are preceded by the development of large scale convective flows that span the heating region.However, in 2D the convective plumes develop too rapidly, leading to an earlier onset of explosion.What can these accelerated, but much cheaper, models teach us about CCSN?
0 100 200 300 4000
100
200
300
400
500
600
700
Time after Bounce (milliseconds)
Mea
n Sh
ock
Rad
ius
(km
)
Melson (2015) 3DsMelson (2015) 2DsMelson (2015) 3DnMelson (2015) 2DnLentz (2015) 3DLentz (2015) 2DLentz (2015) 1D
15 M☉
20 M☉
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Is 2D Turning down the heat?The Rayleigh-Taylor Instability, driven in CCSN by neutrino heating, favors large scale plumes, regardless of dimensionality.In 2D, the turbulent cascade also favors organizing small scale motion into larger scale flows.However, in 3D, the cascade favors tearing apart large scale flows. Thus in 3D, R-T requires more time and more heating to develop.This implies that successful 2D models will tend to have lower entropy in the heating regions.This likely impacts the degree of alpha-richness in the ejecta.
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Multi-D introduces stochastic flow, raising uncertainty in the range of variations if the same model is run multiple times.Cardall & Budiardja (2015) ran 160 3D hydrodynamic simulations mimicking SASI-dominated and convectively-dominated CCSN.
This gives some hope that convective models, at least, are predictive.
How predictive are the Models?
Convectively-dominated models show low stochasticity
SASI-dominated models show high stochasticity
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Observations tell us that the explosion, and the ejected elements, are asymmetric. Yet we rely on spherically symmetric models to understand supernova nucleosynthesis.
Ni, O+Ne+Mg, C
1D
Still Exploding an Onion?
Hughes, Rakowski, Burrows & Slane 2000
Fe, Si O, Reality
≠
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Observations tell us that the explosion, and the ejected elements, are asymmetric. Yet we rely on spherically symmetric models to understand supernova nucleosynthesis. This colors our discussion, for example the notion that the matter created closest to the neutron star is most sensitive to the “mass cut”.
Ni, O+Ne+Mg, C
1D
Still Exploding an Onion?
Hughes, Rakowski, Burrows & Slane 2000
Fe, Si O, Reality
? =
A. Wongwathanarat et al.: 3D CCSN simulations
of two smaller than in the constant wind model (see Tab. 2, anddiscussion in Sect. 5.2 and 5.3.4).
To follow the evolution beyond shock breakout we embed-ded our stellar models in a spherically symmetric circumstel-lar environment resembling that of a stellar wind. In this envi-ronment, the density and temperature distribution of the matter,which is assumed to be at rest, is given for any grid cell i withr
i
> R⇤ by
⇢e(r) = ⇢0
✓R⇤r
◆2, (6)
Te(r) = T0
✓R⇤r
◆2(7)
with ⇢0 = 3⇥10�10 g cm�3 and T0 = 104 K. The stellar radius R⇤is given in Tab. 1.
3. Comparison with HJM10
Before discussing the set of ”standard” 3D simulations (seeSect. 5.1), we first consider two additional 3D simulations thatwe performed specifically to compare the results with those ofthe 3D simulation of HJM10. The numerical setup and the inputphysics di↵er slightly from the standard one used in all our othersimulations presented here, so that they closely resemble thosedescribed in HJM10, except for the utilization of the Yin-Yanggrid in our simulations.
3.1. Simulation setup
The simulations are initialized from the 3D explosion model ofScheck (2007) that results from the core collapse of the BSGprogenitor model B15. Scheck (2007) simulated the evolution in3D from 15 ms until 0.595 s after core bounce using a sphericalpolar grid with 2� angular resolution and 400 radial grid zones.To alleviate the CFL time step constraint he excised a cone of 5�half-opening angle around the polar axis from the computationaldomain. The explosion energy was 0.6 B at the end of the simula-tion, but had not yet saturated. Scheck (2007) neglected nuclearburning and used the EoS of Janka & Muller (1996) with fournuclear species (n, p, 4He, and 54Mn), assumed to be in nuclearstatistical equilibrium.
We mapped the explosion model of Scheck (2007) onto theYin-Yang grid using two grid configurations with 1200(r) ⇥92(✓) ⇥ 272(�) ⇥ 2 and 1200(r) ⇥ 47(✓) ⇥ 137(�) ⇥ 2 zones. Thiscorresponds to an angular resolution of 1� (model H15-1deg)and 2� (model H15-2deg), respectively. Since a cone around thepolar axis was excised in the explosion model of Scheck (2007),we supplemented the missing initial data using tri-cubic splineinterpolation. The radial grid extends from 200 km to near thestellar surface, the fixed outer boundary of the Eulerian grid be-ing placed at 3.9 ⇥ 107 km. We imposed a reflective boundarycondition at the inner edge of the radial grid, and a free-outflowboundary condition at the outer one. During the simulations werepeatedly moved the inner boundary outwards, as described inSect. 2.2.
As in HJM10 we artificially boosted the explosion energy toa value of 1 B by enhancing the thermal energy of the post-shockmatter in the mapped ”initial” state (at 0.595 s). We did neithertake self-gravity nor nuclear burning into account. We advectedeight nuclear species (n, p, 4He, 12C, 16O, 20Ne, 24Mg, and 56Ni)redefining the 54Mn in the explosion model of Scheck (2007) as56Ni in our simulations.
Fig. 3. Isosurfaces of constant mass fractions at t⇡9000 s formodels H15-1deg (left) and H15-2deg (right), respectively.The mass fractions are 7% for 56Ni (blue), and 3% for16O+20Ne+24Mg (red) and 12C (green). The morphology is al-most identical to that shown in the lower left panel of Fig. 2in HJM10, except for some additional small-scale structuresin the better resolved model. There are two pronounced nickelplumes (blue) visible on the right, which travel at velocities upto 3800 km s�1 and 4200 km s�1 in model H15-2deg and H15-1deg, respectively, and two smaller nickel fingers on the left.
The setups employed for our two H15 simulations and thesimulation of HJM10 di↵er only with respect to the grid config-uration. HJM10 used a spherical polar grid excising a cone of5� half-opening angle around the polar axis as Scheck (2007),while we performed our present simulations with the Yin-Yanggrid covering the full 4⇡ solid angle. Our model H15-1deg hasthe same angular resolution as the 3D simulation of HJM10. Wenote that in the simulation of HJM10 the reflecting boundarycondition imposed at the surface of the excised cone might havea↵ected the flow near this surface, while our simulations basedon the Yin-Yang grid avoid such a numerical problem.
3.2. Results
Fig. 3 shows isosurfaces of constant mass fractions of 56Ni,”oxygen”, and 12C about 9000 s after core bounce for modelH15-1deg (left) and H15-2deg (right), respectively. Note thatas in HJM10, we denote in this section by ”oxygen” the sumof the mass fractions of 16O , 20Ne, and 24Mg. At first glance,both simulations exhibit similar RT structures. Two pronouncednickel (blue) plumes, a few smaller nickel fingers, and numer-ous ”oxygen” (red) fingers burst out from a quasi-spherical shellof carbon (green). The maximum radial velocity of the pro-nounced nickel plumes is about 4200 km s�1 in model H15-1degand about 3800 km s�1 in model H15-2deg (Fig. 4). However,while at the tips of these nickel plumes well-defined mushroomcaps grow in model H15-1deg, they are less developed in modelH15-2deg, because the responsible secondary Kelvin-Helmholtz(KH) instabilities are not captured very well in the run with thelower angular resolution.
There are also more ”oxygen” fingers in model H15-1degthan in model H15-2deg. Nevertheless, these fingers grow alongexactly the same directions in both simulations. Comparing thespatial distribution of RT fingers in Fig. 3 and the lower left
6
Wongwathanarat, Müller & Janka (2015)
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Unlike 1D, Nickel and Titanium have higher velocities than Silicon and Oxygen, thus they are not preferentially sensitive to fallback.
Slow Ni?
Radial Velocity [km/s]
0 4,000 8,000 12,000 16,000 20,000 24,000
Mass
[M-]
10!6
10!5
10!4
10!3
10!2
10!1B20-WH07 M+
56Ni 28Si 44Ti
Radial Velocity [km/s]
0 4,000 8,000 12,000 16,000 20,000 24,000
Mass
[M-]
10!6
10!5
10!4
10!3
10!2
10!1B25-WH07 M+
56Ni 28Si 44Ti
Radial Velocity [km/s]
0 4,000 8,000 12,000 16,000 20,000 24,000
Mass[M
-]
10!6
10!5
10!4
10!3
10!2
10!1B12-WH07 M+
56Ni 28Si 44Ti
Radial Velocity [km/s]
0 4,000 8,000 12,000 16,000 20,000 24,000
Mass
[M-]
10!6
10!5
10!4
10!3
10!2
10!1B15-WH07 M+
56Ni 28Si 44Ti
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Distance along symmetry axis [#103 km]
-6.0 -5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0
Distance
from
symmetry
axis[#103km]
1.0
2.0
3.0
4.0
5.0
6.0
PeakTem
pera
ture
[GK]
2
4
6
8
10
B12-WH07
1D Mcut
2D Mcut
Fe/SiSi/Si+O
The Lagrangian view provided by tracer particles reveals the complexity of the mass cut, with discontiguous patches of ejecta (color dots) and bound matter (black dots).
How distorted is the Mass Cut?
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Where is the νp-process?The νp-process is very weak in our models, even at 1.2-1.4 seconds.
The suppression of the PNS wind is delaying or preventing a strong νp-process from occurring.
100
101
102
103
X=X
-
B12-WH07B15-WH07B20-WH07B25-WH07
40 50 60 70 80 90 100 110
Mass Number (A)
0
0.05
0.1
0.15
j/ 8j
8-rates o,8-rates on
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
One way to view the limitations of the tracer resolution is the distribution in the electron fraction of the ejecta.Tracer resolution clearly limits the production of more exotic species.For the CHIMERA B-series, run to 1.2-1.4 s after bounce, this is the largest uncertainty, though it only affects α-rich freezeout.
How many Tracers is Enough?
Model Particles Mtracer [M⊙]B12-WH07 4000 1.87 × 10-4
B15-WH07 5000 2.86 × 10-4
B20-WH07 6000 3.55 × 10-4
B25-WH07 8000 3.49 × 10-4
B12-WH07tpb = 1:336 s
Mass
[M-]
10!8
10!6
10!4
10!2
100
B15-WH07tpb = 1:200 s
Mass
[M-]
10!8
10!6
10!4
10!2
100
B20-WH07tpb = 1:383 s
0.35 0.4 0.45 0.5 0.55 0.6M
ass
[M-]
10!8
10!6
10!4
10!2
100
B25-WH07tpb = 1:399 s
Ye
0.35 0.4 0.45 0.5 0.55 0.6
Mass
[M-]
10!8
10!6
10!4
10!2
100
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Can we make Ti without Ni?The observations of Cas A by Grefenstette, … (2014), and follow-ups at other wavelengths, put significant limits on the amount of Fe (Ni) that is co-resident with 44Ti, which 1D models can’t replicate.
Distance along symmetry axis [#103 km]
-20 -15 -10 -5 0 5 10 15 20
Dista
nce
from
sym
met
ryaxis
[#103km
]
0
5
10
15log10(X44Ti=X56Ni)
Particle Size / ;X44Ti
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
B25-WH07tpb = 1.399 s
Distance along symmetry axis [#103 km]
-20 -15 -10 -5 0 5 10 15 20
Dista
nce
from
sym
met
ryax
is[#
103km
]
0
5
10
15log10(X44Ti=X56Ni)
Particle Size / ;X44Ti
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
B20-WH07tpb = 1.383 s
Distance along symmetry axis [#103 km]
-20 -15 -10 -5 0 5 10 15 20
Dista
nce
from
sym
met
ryaxis
[#10
3km
]
0
5
10
15log10(X44Ti=X56Ni)
Particle Size / ;X44Ti
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
B15-WH07tpb = 1.200 s
Distance along symmetry axis [#103 km]
-20 -15 -10 -5 0 5 10 15 20
Dista
nce
from
sym
met
ryaxis
[#10
3km
]
0
5
10
15log10(X44Ti=X56Ni)
Particle Size / ;X44Ti
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
B12-WH07tpb = 1.336 s
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Can we make Ti without Ni?The observations of Cas A by Grefenstette, … (2014), and follow-ups at other wavelengths, put significant limits on the amount of Fe (Ni) that is co-resident with 44Ti, which 1D models can’t replicate.
Distance along symmetry axis [#103 km]
-20 -15 -10 -5 0 5 10 15 20
Dista
nce
from
sym
met
ryaxis
[#103km
]
0
5
10
15log10(X44Ti=X56Ni)
Particle Size / ;X44Ti
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
B25-WH07tpb = 1.399 s
Distance along symmetry axis [#103 km]
-20 -15 -10 -5 0 5 10 15 20
Dista
nce
from
sym
met
ryax
is[#
103km
]
0
5
10
15log10(X44Ti=X56Ni)
Particle Size / ;X44Ti
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
B20-WH07tpb = 1.383 s
Distance along symmetry axis [#103 km]
-20 -15 -10 -5 0 5 10 15 20
Dista
nce
from
sym
met
ryaxis
[#10
3km
]
0
5
10
15log10(X44Ti=X56Ni)
Particle Size / ;X44Ti
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
B15-WH07tpb = 1.200 s
Distance along symmetry axis [#103 km]
-20 -15 -10 -5 0 5 10 15 20
Dista
nce
from
sym
met
ryaxis
[#103km
]
0
5
10
15log10(X44Ti=X56Ni)
Particle Size / ;X44Ti
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0
B12-WH07tpb = 1.336 s
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
are1D results reasonable?Until we can replace 1D CCSN models in all of their applications, we can use the 2D models to identify areas of concern.Intermediate mass elements, up to A=50, are similar, though significant isotopic differences exist. Mass Number (A)
0 10 20 30 40 50 60 70 80 90 100 110
Productionfactorrelativeto
16O
10!2
10!1
100
101
102
tpb = 1 yr12 M-
15 M-
20 M-
25 M-
WH07B-Series
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
are1D results reasonable?Until we can replace 1D CCSN models in all of their applications, we can use the 2D models to identify areas of concern.Intermediate mass elements, up to A=50, are similar, though significant isotopic differences exist.Iron peak and heavier, up to A=90, the differences get larger.
Mass Number (A)0 10 20 30 40 50 60 70 80 90 100 110
Productionfactorrelativeto
16O
10!2
10!1
100
101
102
tpb = 1 yr12 M-
15 M-
20 M-
25 M-
WH07B-Series
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
How does Multi-D impact ejecta?Multi-dimensional dynamics allows the ejecta to experience a wider variety of temperature, density, electron fraction and neutrino exposure.
Deeper Mass Cut results in modest increase in intermediate mass and iron-group elements.
N (Neutron Number)
2 8 20 28 50 82 126
Z(P
roto
nNumber)
28
20
28
50
82
HBeNNe
AlSK
TiMn
NiGa
SeRb
ZrTc
PdIn
TeCs
CePm
GdHo
YbTa
OsAu
PbAt
Ra
1014
2230
4147
5460
6773
8086
9299
105
112
118
124
131
137
144
150
156
163
169
176
182
189
195
201
208
214
221
log10(X2D=X1D)
B12-WH07tpb = 1 yr
-2 -1 0 1 2
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Can we make 48Ca in a CCSN?Argument has been that ejecta in parameterized spherically symmetric models is all too high in entropy to make 48Ca.
In the self-consistent, multi-dimensional models, accretion streams occasionally dredge neutron-rich matter from near the neutron-star.If this matter is not heated too much by subsequent interactions, such matter can be the source of 48Ca.
What Else Can we Find?
W.R.Hix(ORNL/UTK) BrainstorminginBasel,October2016
Answers, so farExamining the nucleosynthesis of CCSN with models that self-consistently treat the explosion mechanism is possible but it requires running models to times > 1 second for uncertainties like the mass cut, thermodynamic extrapolation, etc. to become tractable. Even then, low post-processing resolution is a significant uncertainty. Differences from 1D models are seen in differing amounts of iron peak and intermediate mass elements as a result of changes in the explosion timing and mass cut. The ejection of significantly more proton-rich matter as well as small quantities of neutron-rich matter can change the production of individual isotopes by orders of magnitude. Neutrino-Driven wind is strongly suppressed by accretion. There is a lot of work yet to be both on the mechanism (especially in 3D) and on the nucleosynthesis.