north carolina state university go.ncsu.edu/astrodata

Carla Fröhlich

North Carolina State University

SLAC Theory Seminar 15 May 2020

Neutrinos in supernovae

go.ncsu.edu/astrodata

Time

Ma

ss

www.nasa.gov

SuccessfulCCSNe

FailedCCSNe

Stars and Supernovae

Instabilities and Supernovae

Langer (2012)

CCSN

PISN

SN Ia

Core-collapse supernovae (CCSNe)

• Need enhanced efficiency of neutrino heating to revive

the shock and drive successful explosion

• Explosion mechanism:• Details are not fully understood

• Convection and/or SASI shock oscillations

enhanced neutrino-heating

• Need to show that this works

Fe-core is gravitationally unstable and collapses

“core bounce” & formation of shock

Shock stalls ~150 kmFormation of heating region behind shock

Janka+2012

The physics in CCSNe

Nuclear / SN:

• Progenitor structure

• Equation of state

• Shock position and

velocity

• SASI

• LESA

• Nucleosynthesis

conditions

• Neutrino opacities

� If you want to understand supernovae,

you have to understand neutrinos

Neutrinos

• Neutrino mass ordering

• Number of ν flavors

• Self-interaction effects

• MSW effects

• Turbulence effects

• Non-standard

interactions

Simulations of CCSNe

• Computational challenges:

• Multi-dimensional problem

(SNe are not spherically symmetric!)

• Gravitation: general relativistic

• Nuclear physics of dense matter

(not very well known)

• Neutrino transport

(diffusion and free streaming regimes)

• Multi-scale problem

(shock formation: several 100km;

entire star: 108 km)

Cas A; Chandra, NASA

SN EOSs by Hempel

CSIRO

Simulation Status

• Dimensionality:

• 1D: in general no self-consistent explosions

• 2D: models have converged

• 3D: mixed results

• Gravity:

• Newtonian

• Post-Newtonian

• Full general relativity

• Neutrino transport:

• Leakage, diffusion, Boltzmann, Monte Carlo

• Grey or spectral

• Multi-scale:

• Resolution; size of computational domain

(similar input gives similar results)

(models are close to threshold; explosions can be code dependent)

(Effective GR potential)

Simulation Status (continued)

• Nuclear processes:

• Energy generation from nuclear burning

• Detailed nucleosynthesis: in-situ vs post-processing

• Neutrino oscillations:

• Post-processing on snapshots

Outline (original plan)

• Coupling neutrino oscillations and

supernova simulations

• What are the effects of coupling a free-streaming

neutrino oscillation code with the hydro?

• Predictions from an effective CCSN model

• Connection betw progenitor, explosion and remnant

• Prediction of multi-messenger signals

• Neutrinos from pair-instability supernovae

All published data available from go.ncsu.edu/astrodata

Outline (modified plan)

• Coupling neutrino oscillations and

supernova simulations

• What are the effects of coupling a free-streaming

neutrino oscillation code with the hydro?

• Predictions from an effective CCSN model

• Progenitor-explosion-remnant connection

• Prediction of multi-messenger signals

• The neutrino-p process


• Method and first results for the collapse of a 20Msun star

• Stapleford, CF, Kneller (arXiv:1910.04172)

Neutrino Oscillations

Neutrinos in CCSNe

• How does the shock get revived after stalling?

• In the neutrino-driven mechanism:

• scattering and absorption in the gain region

• Different neutrino flavors interact differently:

• Electron neutrinos:

• strongest interactions but lower mean energy

• Electron anti-neutrinos:

• weakest interactions but higher mean energy

� Mixing the flavors could affect the heating

• There is a need for self-consistent neutrino

flavor oscillations in supernova simulations

Where does the field stand?

• Various arguments suggested that dense matter

and collisions suppress flavor transformation

below the shock

• Snapshots from simulations without oscillations

and postprocessed them with flavor conversion

code based on the Bulb Model

• Solve QKE in supernovae

• Very hard problem (need micro-m resolution, not km)

• First attempts are made

• Our approach:

• Couple free-streaming neutrino oscillation code with

GR hydro code with Boltzmann neutrino transport

Chakraborty+11

Suwa+11

Dasgupta+12

Capozzi+19

Richers+19

Stapleford, CF, Kneller arXiv:1910.04172

Our simulation setup

Agile-BOLTZTRAN:• 1D Lagrangian GR hydrodynamics• order v/c Boltzmann equation for 4 flavors• EOS: LS220• Transport processes:

• (anti)neutrino absorption on n,p• EC/PC on n,p• (anti)neutrino absorption on nuclei• EC/PC on nuclei• Isoenergetic scattering• Pair-production and annihilation

SQA:• Multi-energy, single-angle, free-streaming

oscillation code for 6 flavors• Solve Schroedinger equation for evolution

operator in a quasi-adiabatic basis

• Hamiltonian: vacuum, MSW, and SI• Includes GR corrections


Agile-BOLTZTRAN:• Progenitor: 20Msun star

• Same as in 1D code comparison paper• 20 energy groups, 8 angle bins• 192 spatial zones

SQA:• Energy resolution:

• < 0.5MeV below 50MeV• < 1MeV below 100MeV

• Inverted mass ordering• PDG values for masses and mixing angles• Use “single-angle approximation” due to

computational costs


Data passed to SQA:• Density profile, electron fraction, enclosed

gravitational mass• Neutrino luminosities, mean energies, mean

energies squared• Radius of neutrino sphere• Spatial grid

SQA: computes the transition probability across the spatial grid

Data returned from SQA:• Convert transition probabilities to effective

opacity for zone i and energy k• Effective opacities are added into the transport

Comparison: with and without oscillations

luminosity

Mean energy

neutrinos Anti-neutrinos

Stapleford+ (1910.04172)

Comparison: with and without oscillations

Stapleford+ (1910.04172)

Extra heating is not enough to cause an explosionShock is slightly behind compared to the case without oscillations

Slightly less heating (~1%) at early timesSome extra heating (~4%) after 300ms

Summary (Part I)

• Full quantum description of neutrinos in CCSNs is

necessary and a challenging problem

• We have taken a first step in this direction:

• Coupling Agile-Boltztran with SQA

• Found:

• ~4% changes in the neutrino heating in a simulation of

20Msun progenitor

• Not enough to change the outcome of a 1D simulation

• Matches the conclusion reached in previous,

post-processed studies

• But: effect may be larger in multi-D simulations

• Explosions and Nucleosynthesis at low/zero metallicity

• Ebinger, Curtis, Ghosh et al (ApJ 2020)

• Single star vs binary-merger progenitors

• CF et al (JPG 2019)

• Nucleosynthesis yields at solar metallicity

• Curtis, Ebinger, et al (ApJ 2019)

• Explosion properties at solar metallicity

• Ebinger, Curtis, et al (ApJ 2019)

• The Method

• Perego, Hempel, CF, et al (ApJ 2015)

Predictions from an effective CCSN model


Effective CCSN models

• Questions we would like to answer:

• Connection between progenitor and remnant?

• Which massive stars explode successfully?

Which ones do not?

• Prediction of nucleosynthesis yields

• Need (many) successful, long-term explosions

• Strategies:

• Ideal: self-consistent, detailed, long-term 3D models

• Realistic: effective models

• Simplify part of the problem, but have free parameters

• Computationally efficient, physically reliable


• General relativistic hydrodynamics: Agile

• Neutrino transport:

• Isotropic diffusion source approximation (IDSA)

• Advanced spectral leakage (ASL)

• Nuclear EOS: HS(DD2)

• Nucleosynthesis: Postprocessing of tracer

particles with nuclear reaction network

CF+06

Hempel+02 Typel+10

Lieberdoerfer+09

Liebendoerfer+02

Perego+16

Our approach: The PUSH method

• Parameterization of the

neutrino-driven mechanism

• Basic idea: tap heavy-neutrino luminosity

inside the gain region to mimic the net

enhanced heating efficiency of νe due to

convection and late accretion in multi-D

Perego, Hempel, CF, et al (2015)




• Basic idea: tap heavy-neutrino luminosity …

• Key features:

• Nuclear EOS and PNS evolution included

• Consistent Ye evolution

(electron-flavor transport not modified)

• Predict Eexpl and mass cut*, nucleosynthesis yields

* Mass cut emerges from the simulation consistent with

explosion energy (not put in by hand)





• Basic idea: tap heavy-neutrino luminosity …

• Key features:

• Nuclear EOS and PNS evolution included

• Consistent Ye evolution

(electron-flavor transport not modified)

• Predict Eexpl and mass cut*, nucleosynthesis yields

• Calibration: Reproduce observed properties of

SN 1987A


Explosions with PUSH (overview)

• Lower explosion energies at lower metallicity

• More models forming black holes at lower metallicity

Solar metallicity

Solar metallicity

Low metallicity

Zero metallicity

Ebinger, Curtis+ 20

Explosions with PUSH (details)

• Explosion energy• Explosion time• Nickel mass• Ejecta mass• Remnant mass

� See go.ncsu.edu/astrodata for ascii tables

NS mass distributions

Ebinger+19

Solar metallicity

Solar metallicity

Low metallicity

Zero metallicity

Ebinger, Curtis+ 20

BH mass distributions

Ebinger+19

Solar metallicity

Solar metallicity

Low metallicity

Zero metallicity

Ebinger, Curtis+ 20

Explosion energy and Ni mass

Ebinger+19

Ebin

ger,

Curt

is+

20

Summary (Part II)

• We developed an effective CCSN model (called

PUSH)

• Applied to hundreds of models for predictions

and correlations

• Explosion properties

• Remnant properties

• Nucleosynthesis (see next part of the talk)

• Some successes of our models

CCSN Nucleosynthesis

Status of CCSN Nucleosynthesis

• 2D models

• 12, 15, 20, 25Msun (at Zsun):

comparing postprocessing vs in-situ network

• 8.8, 11, 15, 27Msun (at Zsun),

8.1Msun (at Z/Zsun=10-4), 9.6Msun (at Z=0):

innermost 10-3Msun neutrino-processed ejecta

• 11.2 and 17Msun (at Zsun):

detailed processing of representative tracers,

extrapolating to other tracers (focus on p-nuclei)

• 3D models

• Postprocessing of ~100k tracers (focus on 44Ti and 56Ni for Cas A) Wongwathanarat+17

Eichler+17

Wanajo+17

Harris+17

Status of CCSN Nucleosynthesis

• 1D models

• Grids of models using piston or thermal/kinetic bomb

• Metallicities (Z/Zsun): 10-5 to 1

• ZAMS masses: ~10 – 40 Msun

plus some < 10 Msun and

plus some > 40 Msun

Woosley&Weaver 95, Rauscher+02,Heger+07, Heger+10Thielemann+96, Nomoto+06, Umeda+08,Nomoto+13, Nomoto+17Limongi & Chieffi 06, Limongi+12, Chieffi+13, Chieffi+17

• But open questions:

� How much energy?

� Where is mass cut? Ni yields?

� Neutrino physics? PNS evolution?

� Physics of collapse, bounce, onset of explosion?

Explosive nucleosynthesis

Explosive nucleosynthesis

s18.8 (RSG) Curtis+19

Isotopic and elemental correlations

Curtis+19 Ebinger, Curtis+ (accepted)

56Ni

58Ni

Elemental Mn

Elemental Nickel

Complete yields …

Curtis+19; Ebinger, Curtis+20

� See go.ncsu.edu/astrodata for ascii tables

… and production paths

Curtis+19

Metal-poor stars

Observational data for HD84937: Sneden+16

Curtis+19 Ebinger, Curtis+ 20

Neutrino-driven wind nucleosynthesis

Neutrino-driven winds

• Strong neutrino flux

from PNS

• Drives matter-outflow

behind shock wave

• Nucleosynthesis:

• NSE (T=10-8GK)

• Charged-particle

reactions (8-2GK)

• r-process and νp-process

nucleosynthesis (3-1GK)

Conditions in wind (Ye, entropy, timescale) determine details of nucleosynthesis

Figure: Janka

Fig

ure

: B

ruenn

Nucleosynthesis in neutrino-driven windsProton-rich conditions:• Elements from Zn to Sn• (p,γ) and (n,p) reactions� ννννp-process

Neutron-rich conditions:• Elements up to uranium

(depending on entropy)• (n,γ) reactions and β-decays� r-process

Proton-rich winds: The νp-process

• proton-rich matter is ejected under the

influence of neutrino interactions

• true rp-process is limited by slow β decays,

e.g. τ(64Ge)

• Neutron source:

• Antineutrinos help bridging long waiting points

via (n,p) reactions:

64Ge (n,p) 64Ga; 64Ga (p,g) 65Ge

�̅� + �→ � +

• With neutrinoso Without neutrinos

The νp-process path

64Ge

56Ni

Almost all reaction rates are from statistical model predictions(inputs are nuclear masses, level densities, spin, parity, particle and

gamma transmission coefficients, etc)

� Which ones are important? Uncertainties?

Individual (n,p) and (p,γ) reactions

• Reactions on light nuclei

• Reactions on heavy nuclei: 56Ni, 64Ge, 96Pd

• 56Ni(n,p): Seed nucleus for νp-process but also

neutron poison

• 64Ge(n,p): Bottle neck

• 96Pd(n,p): Predicted as second seed, but not

confirmed

• Systematic study of (p,g) reactions on Ni

isotopes

• Confirmed previous results for 64,62,58Ni

• Resolved previous disagreement for 61,60Ni

• New estimate for 56Ni(p,g)57Cu (lower by factor 0.37)

Wanajo et al (2012)

Simon et al (2013)

Wanajo et al (2012); Frohlich+ (2012)

Systematic sensitivity study

• Vary each reaction rate from Ni to Sn

• Reaction types: (n,p), (n,g), (p,g)

• Factors: 10 and 0.1

• Conditions: 2 different νp-process trajectories

(“standard” and “strong”)

59Cu(p, g) is competition with another channel

Monte Carlo Study

• Simultaneous variation of 10,000 rates for 23

different trajectories

Beyond the nuclear physics

• What is the nucleosynthesis in neutrino-driven

winds impacted by?

• hydrodynamics / reverse shock

• Neutrino properties (energy and fluxes)

Arcones, Frohlich, Martinez (2012)Wanajo et al (2012)

Summary (Part III)

• Explosive nucleosynthesis:

• Progenitor model matters

• Explosion method matters for iron-group

• Our predictions are consistent with observations of

EMPs

• Neutrino-driven wind nucleosynthesis:

• Proton-rich or neutron-rich?

• Neutrino-p process under proton-rich conditions

• Impact of nuclear reaction rates

• Neutrinos matter! They set the conditions for

nucleosynthesis

Summary

We coupled a neutrino oscillations code to a hydrodynamic simulation with Boltzmann neutrino transport

We have a tool (PUSH method) that allows us to compute many CCSNe …

Neutrino-drivencore-collapse supernovae

… and make observable predictions

… and do sensitivity studies

The Teams

PUSH Method:

• Sanjana Curtis (NCSU)

• Noah Wolfe (NCSU undergrad)

• Somdutta Ghosh (NCSU)

• Kevin Ebinger (NCSU � GSI)

• Albino Perego (now at Trento)

• Matthias Hempel (Basel)

Funding agencies

• Department of Energy

• Research Corporation for Science Advancement

Neutrino oscillations:

• Charles (CJ) Stapleford (NCSU)

• James Kneller (NCSU)

Neutrino-p Process:

• Daniel Hatcher (NCSU � CMU)

• Nobuya Nishimura (Japan)

Thank you!

north carolina state university go.ncsu.edu/astrodata

Documents