Numerical simulations of core-collapse supernovae J´ erˆ ome Novak ([email protected]) Laboratoire Univers et Th´ eories (LUTH) CNRS / Observatoire de Paris / Universit´ e Paris-Diderot 10 th Rencontres du Vietnam, Very high energy phenomena in the Universe ICISE Quy Nhon, August, 8 th 2014
23
Embed
Numerical simulations of core-collapse supernovaevietnam.in2p3.fr/2014/vhepu/transparencies/FridayMorning/... · 2014. 8. 8. · Numerical simulations of core-collapse supernovae
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Core-collapse supernovaeObservations of “supernovae” reported sinceAntiquity (SN185) ⇒explosions commonCategories, depending on the spectrum:
Type I: no hydrogen. Ia: ionized silicon,Ib/Ic : strong/weak presence of helium.
Type II: presence of hydrogen. IIP/IILdepending on the light curve shape, IIb:spectrum changing from II to Ib. ALMA/Hubble/Chandra
Milisavljevic et al. 2013
Two kinds of theoretical models:
Thermonuclear supernovae:explosion of a white dwarf(runaway nuclear reactions).
Core-collapse supernovae:collapse-bounce-explosion of amassive main-sequence star. . .
Open questions. . .
Why / How do massive stars explode?
What are the properties of the final compactobject at the center?
How / Where do heavy elements form?
What are observable signals?
What can we learn for fundamental physics? Ugliano et al. 2012
Growing number of groups:
MPA Garching (H.-T. Janka), Princeton (A. Burrows), Oak Ridge (T.
Mezzacappa), Univ. Basel (M. Liebendorfer), Tokyo (S. Yamada),
NAOJ/Fukuoka (K. Kotake), Caltech (C. Ott), Los Alamos (C. Fryer),
France (T. Foglizzo), Univ. Valencia (M.-A. Aloy), . . .
Open questions. . .
Why / How do massive stars explode?
What are the properties of the final compactobject at the center?
How / Where do heavy elements form?
What are observable signals?
What can we learn for fundamental physics? Ugliano et al. 2012
Growing number of groups:
MPA Garching (H.-T. Janka), Princeton (A. Burrows), Oak Ridge (T.
Mezzacappa), Univ. Basel (M. Liebendorfer), Tokyo (S. Yamada),
NAOJ/Fukuoka (K. Kotake), Caltech (C. Ott), Los Alamos (C. Fryer),
France (T. Foglizzo), Univ. Valencia (M.-A. Aloy), . . .
Model description
Collapse and bounce
Massive (& 10M�) main-sequence star with onion-likestructure:
Iron core becomes unstable (electron degeneracy pressure)
Collapse with electron captures on nuclei/free protonsp+ e− → n+ νe
Central density ∼ nuclear density ⇒nuclear repulsion
Shock wave expanding outwards. . .
�
��
�
�
���������������������
������������ �� ������������������
��
��
�
��
��
���������������������������������
�
�
��
���������
�������� ��������
���� ����
������ ������ �����������
��
Janka et al. 2012
Shock evolution
Shock stalls, due to energy loss by iron nuclei photodissociation
��
�!�
���������
���� ���"������
���� ����
�������� ��������
��
Janka et al. 2012
⇒there must be some mechanism to revive the shock: transferenergy from the gravitational well to the shock.
Physical ingredients
Progenitor model⇒evolution for massive starsGravitation – hydrodynamics⇒relativistic?Microphysics: equation of state for hot, dense matter, outof β-equilibrium : p(ρ, T, Ye)⇒very large range of densities, temperatures andasymmetryMicrophysics: electron capture and neutrino reaction rates(opacities)Neutrino transport⇒6+1 dimensions ?Magnetic field evolution⇒resistivity?
Endeve et al. 2009
Numerical challengesMost of supernovacore-collapse simulationsrun on HPC centers, withmillions of CPU-hoursused (sometimes) for asingle run:⇒need for exaflop? PRACE/Curie
Hydrodynamics: high-resolutionshock-capturing methods need CPU andhave poor convergence properties near theshock (unavoidable?)
Some physical processes may need manyhydro time scales to appear (e.g. SASI)⇒millions of time-steps (implicit for νs)
Blondin & Mezzacappa
2007
Gravitational wavessee also talk by M. Was
Core-collapse supernovae are good sources of GW, although notthe best ones (talk by U. Sperhake) : too close to sphericalsymmetry!
3 phases of GW emissionassociated with currentsimulations:
bounce,
post-bounce convectiveinstabilities,
long-term instabilities(SASI-like)
Ott 2009
What can we learn from core-coolapse GW?⇒multiple bounces?⇒importance of convection?⇒development/saturation of long-term instabilities. . .
What is missing?
Recent 2D/3D runs show at best weak explosions : the releasedenergy is too small!
Kuroda et al. 2014
Dimensionality: spherically symmetricsimulation known not to succeed. . .2D: weak explosions . . . 3D?
Resolution: are all features resolved?(turbulence, instabilities) or simulated longenough?
Neutrino transport: too much simplified?
More physics: Relativistic gravity (andhydro)? Magnetic field (MRI) ? Progenitormodels (rotation)?
Microphysics: are reaction rates and EoScontrolled?
. . .
“Recent” developments
Analogue SASIStanding Accretion Shock Instability:advective-acoustic cycle between theproto-neutron star surface and thestalled shock Blondin & Mezzacappa 2007
Foglizzo et al. 2012
Provides energy to the shock and helpsneutrino heating
Exhibited in the simplified adiabaticcase (pure hydro)
Analogy with shallow water 2D model,studied in laboratory with a simpleexperiment.
⇒very strong indication for the existence of SASI insupernovae: not numerical artifact⇒study of growth, saturation and properties of the instability,much faster than with a code⇒public outreach. . .
Progenitor dependence
Couch & Ott 2013
Models for progenitors verysophisticated but not enough :rotation, magnetic field, stellaratmosphere, advanced nuclearburning stages. . .
Possibility of asymmetriesfrom nuclear burning :deviation from sphericalsymmetry
Study of genericasymmetries
Evidence of more energytransmitted to the shock
Help in explosion?
Magnetic field
Magnetic field is present inmassive main-sequence stars
By conservation of magneticflux, can reach (usual) pulsarobserved values . . . what aboutmagnetars?
0
1
2
3
4
5
0 100 200 300 400 500 600 700
τa/τ
h
time after bounce [ms]
(a) BG∆in=100 m
50 m
25 m
Sawai & Yamada 2014
Magneto-rotational instability (MRI) is very likely to appearduring core-collapse : differential rotation, magnetic field arepresent⇒Very poor knowledge on initial magnetic field and rotationprofile⇒Need for very high numerical resolution in global simulations⇒What is the resistivity of matter at such densities andtemperatures?
General relativity
0
0.5
1
1.5
2
2.5
3
τadv/τ
heat[s
]
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
τheat[s
]
0.1 0.2 0.3 0.4 0.5time [s]
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
τadv[s
]
G15M15S15N15G11
Mueller et al. 2012
Relativity is needed both for
gravity
(2GM
Rc2∼ 0.1
)and
hydrodynamics(vc∼ 0.3
)But difficult to build a code:coupling of dimensions in hydro,many non-linear equations forgravity, nightmare forneutrinos. . .
⇒Deepening of gravitational well:more energy available Vs. moredifficulties for the shock to escape . . .
Unexpected : GR helps the shock !
3D simulations
0.1 0.2 0.3 0.4 0.5 0.6
1.6
1.8
2
2.2
2.4
2.6
2.8
3
3.2
1D
2D3D
M(M⊙ s−1 )
Lνe(1052ergs−
1)
Nordhaus et al. 2010
Dimensionality can have great influenceon the explosion mechanism:
2D (axial symmetry) : inverseenergy cascade for turbulence
For neutrinos: exchange of energy inthe radial directions
Hanke et al. 2013
Debate on the necessity of full 3Dsimulations : does it help the shock toescape?⇒Recent simulations in 3D / GR,with sophisticated neutrino transportdo not show any such trend. . .
MicrophysicsInput from the nuclear and particle physics communities:
New equations of state, considering full nuclear statisticalequilibrium distribution (e.g. Hempel & Schaffner-Bielich2010)
or additional particles, as hyperons (e.g. Oertel et al. 2012)
New electron capture rates change the shock energy(Fantina et al. 2012)
Peres et al. 2013
Full nuclear distribution Vs.mean nucleus : not much effect⇒what about neutrino rates ?
Phase transition to quarks orhyperons ⇒second shock revivingthe first? (Sagert et al. 2009)⇒Not expected in core-collapseto a neutron star; relevant for thecollapse to a black hole?
Neutrino transportNeutrino transport uses a distribution function f(t, xi, pi) :almost impossible to solve without symmetry / approximation
Leakage scheme (in GR: Sekiguchi 2010)
Flux-limited diffusion (not relativistic, Bruenn et al. 1978)