Transcript Presentation

Neutrino Burst from Supernovae and
Neutrino Oscillation
-What is the effect of neutrino OSC on explosion and the detection?
-Can we extract OSC parameters from the neutrino observations ?
Katsuhiko Sato (Univ.Tokyo)
Collaborators:
K. Takahashi, S.Ando, T. Totani, K. Kotake, S. Yamada,
T. Shimizu, S. Ebisuzaki
J. Wilson, S. Dalhed, A. Burrows and T. Thompson
Plan of this talk



Introduction
A brief review of gravitational collapse-driven
supernova
Neutrino OSC in supernova and the detection
-Constraint on OSC parameters from the detection of
supernova neutrino burst –


Effects of rotation on explosion and neutrino burst
（Gravitational wave from supernovae）
Supernova１９８７A
in LMC ２３Feb.７：３５AM（UT），１９８７
10 trillion neutrinos
passed through your
body.
Huge water Cerenkov
counters could detect
this neutrino burst,
11 events by
Kamiokande &
8 events by IMB .
Direct evidence that SN is triggered by gravitational collapse of stellar cores.
Remarkable achievement which remains in history.
Nobel Prize was awarded to Dr. M. Koshiba, the
head of Kamiokande, Professor emeritus of the Univ. of Tokyo.
From http://www.nobel.se/
We have been waiting the prize more than 15 years ！
Now huge neutrino detectors are running!
If a supernova appears at the Galactic center, then almost
10,000 events at SK
and
350 events at SNO are expected.

Total mass:10,000t, Fiducial
mass:3,200t 30xKamII
800 events
1,000tD2O, 1.400t H2O
At LVD.
Now we must consider seriously what astrophysics/physics are obtained
from the detection of supernova neutrino burst.
Collapse of Stellar Cores and Neutrino Trapping
(K. Sato’75)
If M > 8-10 Msolar, Iron core is formed.
–10
9.5g/cm 3
–10
11g/cm 3
–10
12g/cm 3
ν
ν
Unstable and begins to collapse.
Neutrinos can escape from
the core without scattering.
56Fe+e->56Mn+ν
e
The mean free path becomes shorter than
the core radius, core, lmfp<R.
The diffusion time τdif=
R
c
2
mfp
becomes longer than the characteristic
collapsing time scale τff .. τdif >τff ..
lmfp
Neutrinos are trapped, and are degenerate in SN
cores.
Neutrino reactions in supernova cores
emission, absorption,
scattering on nucleons
scattering on
electrons
emission, absorption,
scattering on nuclei
n-pair creations,
annihilations
Neutrinos are trapped by the effect of
coherent scattering (Sato,’75) since the cross
section proportional to A2 ,and is larger.
Coherent scattering depends on the size and shape of nuclei .
How nuclei melt into supernova matter / neutron star matter ?
Just after the glitches of pulsars were
discovered.
How nuclei melt in the course of collapse? Important for opacity
“Nuclear Pasta “ Structure With increasing matter density, the shape
changes from sphere,cylinder,slab,
cylindrical bubble,spherical bubble and
eventually becomes homogeneous.
(Ravenhall & Pethick.,83, Hashimoto et al,84,
Oyamatsu et al., 84,…Maruyama et al., 98, Watanabe
et al., 00, and 01, Iida et al.,01)
Essentially this change is described by the surface
energy
surface _ area
g
(volume) 2 / 3
minimum principle.
Oyamatsu,93
From Oyamatsu et al., 84
QMD method is suitable for
investigating the melting by finite
temperature (Maruyama et al. 98).
How the structure Perturbation analysis
with the analogy of
changes with
Recently we improved this method
liquid crystal Pethick,
increasing
and succeeded to construct pasta
Potekhin,98, Watanabe et al., 00
temperature?
structure (Watanabe et. al .,01,02,03).
Results of QMD calculation
(Watanabe,Sato,Yasuoka,Ebisusaki; PRC ’02)
Model:
N= 2048
T~0.1MeV
X=p/(p+n)=0.3
Sphere(0.1ρ０）
slab (0.35ρ０）
cylinder (0.18ρ０）
Cylinder hole (0.５ρ０）
Spherical hole (0.55ρ０）
Preliminary result on the melting with increasing temperature:
Model:N= 2048,ρ~0.３５ρ０ , X=p/(p+n)=0.5
T=0.1MeV (cylinder+slab)
T=3MeV
T=1MeV
T=4MeV (almost homogeneous)
T=2MeV (slab)
T=5MeV
Two-point correlation function ξ of the
nucleon density fluctuations δ
with
&
disappearance of
long-range correlation
at T=5MeV
ξ(r)=0 at larger r
uniform phase
at T>4-5MeV
Nucleon Distributions for
X=p/(p+n)=0.3 and ρ=0.175ρ0
T=1MeV
T=2MeV
Cylinder phase at T=0
ρ=0.175ρ0
N=2048, Np=614, Nn=1434
box size=41.394fm
T=3MeV
T=4MeV
T=3, 4MeV : Nuclear surface
cannot be identified by an
isodensity surface.
Nucleon Distributions for
x=0.3 and ρ=0.35ρ0
T=0MeV
T=1MeV
Slab phase at T=0
ρ=0.35ρ0
N=2048, Np=614, Nn=1434
box size=32.85fm
T=2MeV
T=3MeV
T=3MeV : Nuclear surface
cannot be identified by an
isodensity surface.
Phase Diagrams for ｘ＝０．３
χ>0
‹H› > 0
χ=0
‹H› > 0
χ<0
‹H› > 0
χ=0
‹H› = 0
χ<0
‹H› < 0
χ=0
‹H› < 0
χ>0
‹H› < 0
Structure with χ<0 (“intermediate” phase) : Sponge-like
χ = (number of isolated
regions) – (number of
tunnels)
+ (number of cavities)
ｘ＝0.3
phase-separating region
limit for identification
of nuclear surface
Still preliminary, but systematic
investigation is in progress.
Euler characteristic
-10
15g/cm 3
shock
The core bounces and the
unshocked inner core
is formed.The shock is
generated at the surface.
Unshocked core playas
a role of spring for explosion.
Inner core
If the shock is sufficiently strong, the star explodes; Prompt explosion
However, most simulations show it is insufficient for explosion, and
stalled . No prompt explosion occurs in realistic sim.
ν
ν
ν
•νcore
Inner
ν
Eventually the shock revived by
ν deposition, and outer shells are
expelled. Delayed Explosion(Wilson)
ν
ν
ν
ν
An example of delayed explosion
late time explosion by ν-heating
shock
n
The stalled shock is revived by the neutrino
deposition from the proto-neutron star. Wilson ‘82
Neutrino Burst from SN
The latest
example of LLL
group: the general
relativistic corecollapse
simulation with
full νtransport
calculation
(Totani, Sato,
Dalhed, Wilson,’98)
Pre-supernova
Model:
Weaver & Woosley
20 Solar Mass.
109
108
1.The latest
neutrino burst
models of the
LLL group (pre-
e p   n e n
(νμ、ντ and their
antiparticles)
SN model:Weaver,
Woosely 20 Msolar )
(Totani,Sato,Dalhed,
Wilson,’98).
Time evolution
of ν luminosity
& the average
energies
<E>=23MeV
<E>=15MeV
<E>=10MeV
Latest simulations with
updated neutrino processes
and sophisticated neutrino
transfer show no explosion.
Liebendoerfer et al. ‘01
Rampp et al. ’00,’03
Thomson et al. ‘02
Why no explosion?
1. Microphysics (neutrino processes, EOS etc.) are still insufficient ?
Something important processes are missed?
2. Computational methods ( neutrino transfer , convection, etc) are
still unreliable?
3. Spherical symmetric simulation is inadequate. Stellar rotation
and/or magnetic field play essential role for explosion ?
In the present neutrino OSC analysis of supernova neutrino burst,
We employ two models
1. LLL models (Totani et al, ’98) as the full neutrino burst model (~ 15
sec). (only one full time neutrino burst model available today)
2. Burrows’s group model (Thomson et al, ’02) as an early phase
burst model (~ 0.2 sec.). (as a representative of latest simulations)
2.TBP (Thompson,Burrows,Pint) Model (early 0.2 sec burst)
Evolution of luminosity
(Takahashi, Sato, Burrows, Thompson, PRD‘03)
Evolution of average energies
<E>=12MeV
<E>=15MeV
<E>=20MeV
The early phase analysis has advantage in that it is not affected whether
the remnant is a neutron star or a black hole.
Neutrino OSC and Neutrino Burst from Supernovae
SK and SNO showed clearly neutrinos have masses, and oscillate.
1. What is the effect on Explosion ?
If we take values of oscillation parameters suggested by solar ν
and atmospheric ν obs., no resonances occur in the core, but
they occur in the mantle of SN (C+O shell, He shell) .
No effect on explosion.
Note: If Δm ~101-2ev, resonance happens in the hot
bubble region, energy deposition is greatly enhanced
because of the large cross section of high energy electron
type neutrinos. Explosion is greatly strengthened
(Shramm et al , …..)
2.What are the effects on the detection?
In order to get original information of cores and to extract the
explosion mechanism, it is essentially important to know how the
spectra of the neutrino burst are modified by neutrino OSC.
3. Can we extract osc parameters from the neutrino observation if a
Galactic supernova appears ?
Supernova is the strongest source of three type of neutrinos in the
universe. (Sun e- type only, atmospheric neutrinos e- and μ- type)
i) Can we obtain the implication on the parameter 13 ,which has not
yet determined?
ii) Can we solve the mass hierarchy problem ?
Inverted mass hierarchy model (mνμ>mνe>>mντ) has not yet ruled out
by experiments.
Resonance in Supernova Mantle
–normal hierarchy modelResonance Condition:
ne  nres
m 2 cos 2

2 2GF E
Dighe,Smirnov, ’00
Lunardini,A.Y.Smirnov,01
Minakata, Nunokawa.’01
Takahashi,Sato, ’01
Takahashi et al,’01
………….
H
He
C,
O
Ne,
Mg
Si
Fe
Neutrino OSC Models
sin 2 212 sin 2 2 23
sin 2 213
LMA-L
0.87
1.0
LMA-S
0.87
1.0
1.0  10
5.0 10
3
1.0
0.043
5.0 10
3
1.0
1.0  10 6 6.0  10 6 3.2 10 3
SMA-L
SMA-S
0.043
m132
m122
7.0  10 5 3.2 10 3
6
7.0  10
5
3.2 10 3
6.0  10 6 3.2 10 3
Inverted mass hierarchy models are denoted as
Inv-LMA-L, Inv-LMA-S, Inv-SMA-L, Inv-SMA-S.
12 m122 ：from solar neutrinos, 13 : upper limit from nuclear
2
m12  10 7
 23 m132 : atmospheric neutrinos
reactor
Time evolution of conversion probability
for LMA-L
and LMA-S
νe
νeνe
ντ
ννττ
Event rate at SK for LLL neutrino burst Model
We calculate the event rate and
the energy spectra at SK,
assuming
SN appeared at GC(10kpc).
_
n e p  e n
_
 
ne e  ne e

_
 
n x e  n x e
_

n e O  e  N
n e e   n e e
n x e   n x e
n e O  e F
_
Most of events come from
n e p  e n
Time evolution of event rate expected at SK
Time-integrated Energy spectra and
Event numbers
_
Most of events come from
 effect of vacuum OSC.
n e p   e . n
νe
νμ、τ
Models with larger mixing angle
deviate from no osc model.
Can be distinguished from
the ratio of event rate at the
peak region to the tail region.
Event rate at SNO for LLL neutrino burst Model
1,000t D2O, （1.400t H2O）
Important reactions
Electron type neutrinos can be
detected efficiently by
n e  d  e   p  p(CC )
_
n e  d  e   n  n(CC )
n x  d  n x  n  p(NC )
n x  e  n x  e
We discuss only CC, not NC.
Time evolution of event rate expected SNO
Energy spectra and Event numbers
_
Events come from the both
n e ,n e .
 both effects, vacuum OSC and MSW.
with increasing mixing angle, event number
increases.
Can be distinguished
from the ratio of event
rate at 15Mev region to
the E>30Mev region.
Case for inverted mass hierarchy
Crossing diagram for antineutrinos

m11
 Case
Case of
of Normal
invertedmass
masshierarchy
hierarchy ((m
m

m3 3) )
No Level crossing
H-resonance happens
for anti neutrinos
m3
m2
m1
ne
H-resonance
n e n1
n   n 2 is
If the resonance
adiabatic (large 13 ),
 n 3
n

n e  n  , conversion
occurs effectively.
ne
Event rate is greatly
increased !
237
The time-integrated
energy spectra
185
111
νe events are increased
by H-resonance: ντ νe.
68
13,084 events
10,245events
190
82
118
In order to extract information of mixing angle,
we define the ratios, R(SK) and R(SNO),
Events(30 ~ 70MeV )
R( SK ) 
Events(5 ~ 20MeV )
Events(25 ~ 70MeV )
R( SNO) 
Events(5 ~ 20MeV )
R(SK) and R(SNO) are good indicators for neutrino OSC.
Plots on RSK-RSNO plane (Error –bars represent only statistical errors.)
n e  d  e   p  p(CC )
Anti neutrino events
can be subtracted by
neutron detection.
nor-LMA-s and inv-LMA-s are degenerate, but inv-LMA-L is clearly
discriminated from nor-LMA-L. If the mixing parameter 13 is -L, mass
hierarchy problem is solved.
Analysis by using the TBP burst model


This simulation was done by using the updated
neutrino processes and sophisticate neutrino
transfer program.
Available data are only the initial 0.2 second of the
neutrino burst, but this early phase analysis has
advantage that it is not affected whether the remnant
is a neutron star or a black hole.
We investigated the dependence on the pre-supernova
mass. We found the results are almost independent
of the masses.
Evolution of the burst and time-integrated energy spectra
(TBP model)
Presupernova-mass dependence on R(SK)-R (SNO) Plots
R( SNO) 
Events(20MeV  E )
Events( E  20MeV )
Error bars
come from only
statistical errors,
which are
increased
because the
event numbers
becomes small.
Events(20MeV  E )
R( SK ) 
Events( E  20MeV )
nor-LMA-s and inv-LMA-s are degenerate, but inv-LMA-L is clearly
discriminated from nor-LMA-L. If the mixing parameter 13 is -L, mass
hierarchy problem is solved.
The Earth effects on the supernova neutrinos
Supernova neutrinos oscillate and are reconverted
each other in the earth. The spectra are greatly
modified if they pass through the earth.
(Dighe &Smirnov (00,01), Takahashi & Sato (00, 01) )
SK 9500 events
LVD
850 events
Earth
SNO
300 events
Pass length when SN occurs at Galactic center.
t=0 : the time at which the SN is aligned with the
Greenwich meridian.
In order to analyze the earth effect and to get information on
OSC parameters, we need to know supernova direction, which
is determined accurately by electron scattering.
 Electron scattering
has sharp forward
peak, but the fraction
of is ~ 3% (282
events/total 8441 for
Galactic Center
Supernova at SK)
Monte Carlo Simulation of recoiled_ e- (e+) direction

for SK: Most of the events are by n e p   e n
By using the leastsquare method, we
get the direction
within the accuracy
7degree (1σ）。
Ando & Sato, ‘01
Modification of the spectra
Spectra are greatly modified by MSW effects in the earth.
Case: LMA-S, nadir angle =0 (pass through the center)
The spectra depend sensitively on the nadir angle
2
and m12
Since the nadir angle can be determined from the scattering by electrons at
SK or SNO, m 2 could be determined more precisely by the earth
12
effects.
nadir angle dependence
m
2
12
dependence
Supernova Relic Neutrinos and its detectability
Ando, Sato, Totani ’02, Ando , Sato ‘03
Cosmic time
We are here.
z=0
z=1
z=5
n
n
There should be a diffuse background of
neutrinos emitted from past supernovae.
(Supernova Relic Neutrino Background, or
SRN)
Flux depends on the history of supernova
rate and neutrino oscillation parameters.
We investigated the dependence of the flux on the OSC parameters,
and effects on the detectability.
Flux is increased greatly if LMA, and if inverted mass hierachy.
History of Supernova Rate

The SN rate model is evaluated from
corresponding SFR model based on
optical/UV observation by HST.

Particularly, behavior at high redshift is
not known well. (Luminosity function is
not established and dust extinction is
unknown.)
 m  0.3,    0.7, h  0.7

However, owing to energy redshift,
neutrinos emitted at high-z contribute
only to low energy region (, where SK
does not have sensitivity).
Madau et al. (1996)
The uncertainty around here is not
important so much.
Flux for Various OSC Models



We obtain the hardest
spectrum for the INV-L
model.
The spectra for the other
LMA models are
degenerated.
We also set upper limit for
these oscillation models, by
analyzing the spectrum with
the SK observational result.
Theoretical prediction and Observational Limit (SK)
Malek et al,’03
model
Predicted flux
(cm2 s1)
SK limit
(90%C.L.)
Prediction/
Limit
NOR-S
12
< 35
0.34
NOR-L
11
< 34
0.33
INV-S
11
< 34
0.33
INV-L
9.0
< 12
0.74
No oscillation 12
< 73
0.17
The upper limit is more severe for the INV-L model.
(In spite of difficult observation, SK upper limit is approaching the theoretical prediction.
It is expected constraints on OSC parameters could be obtained near future.)
Effects of Rotation on the Supernova Explosion

Massive stars have large angular
momentum:
q= J/(GM2/C) ~ 10
Implications of rotational collapse
and Jet-like explosion from
SN1987A observations.
1.Observation of asymmetry of
expanding envelope by SPECKLE
(Papalios, et. al. 89)

2.Observation of linear polarization of
scattered photons (Cropper et al.88)
3. Rings suggest pre-supernova was
rapidly rotating.
Many groups have been challenging the simulation
of rotational collapse of stellar cores.
Mesh code
SPH (Smoothed Particle Hydrodynamics)
2dim.
Herant et al (’94)
LeBranc, Wilson
Fryer (’99), Fryer et al (’01)
Symbalisty
…………
Moenchmeier et al (91)
Yamada, Sato
Shimizu, Yamada, Sato(94,01)
…..
Difficulties in simulation
multi-dimension neutrino transport
general relativistic treatment
………
3-dim.
Shimizu, Yamada, Sato (94)
All simulations are still preliminary ones.
Asymmetric n-heating due to rotation
Shimizu, Ebisuzaki, Sato, Yamada ,’01
•If oblate proto neutron stars are formed due to centrifugal forces,
more neutrinos are emitted in the direction of rotation axis.
• Assuming an oblate proto neutron star is formed, we carried out
hydrodynamic simulation, and found that n-heating is enhanced
near the rotation axis, and global convections are induced in heating regions.
•As the result, Jet like explosion is induced.
oblate
proto
neutron
star
2D Rotational Collapse Simulations
Kotake, Yamada, Sato , ApJ595, 304 (03)
t = 256ms
Shape of neutrino
sphere becomes
spheroid.
entropy
Radius [cm]
Density
Temperature on the
sphere
５
３
90°
0°
Neutrino luminosity and the average energy
depend on what direction we observe.
We are investigating whether implication on
OSC parameter could be obtained or not.
Gravitational Radiation from Axisymmetric
Rotational Core Collapse Kotake, Yamada, Sato, PRD68, 044023 (03)
We calculated gravitational radiation by
using quadrupole radiation formula.
Most preceding works took simplified
EOS, i.e. , p=Kργ , and neutrino emission/
absorption/transport are neglected.
We carried out collapse simulation by
using realistic EOS (Relativistic MFA;
Shen et al. ‘98 ) and included neutrino
processes.
Preceding works and recent
works
 ……
 Moenchmeyer et al . ’91
 Yamada, Sato, ’97
 Zwerger & Mueller ’97
 Dimmelmeier et al. ’02
 Shibata ’03
 Kotake,Yamada, Sato ’03
 Ott et al., ’03
Theoretical prediction of “hTT” when SN appear at Galactic center and
detection limit
We carried out for various rotation models, and
found most of them are higher than TAMA
detection limit.
Example of wave pattern
Case of Moderate rotation
Case of strong differential rotation
Small fluctuations
disappear because
of centrifugal force
in the central core
region.
Wave patterns depend on rotational
speed and distribution of angular
momentum.
If the gravitational wave is
detected and wave pattern is
observed, information on the
rotation would be obtained.
Summary
● Now
huge neutrino detectors (SK, SNO,LVD,..)
and supersensitive GW detectors (TAMA,LIGO,..) are
working.
1.If SN appears at Galactic center, 10,000 events (SK) , and 350 events
(SNO) will be detected, and fruitful information on the explosion
mechanism and neutrino OSC parameters would be obtained.
More huge detector Hyper Kamiokande is proposed.
2.TAMA and LIGO would detect gravitational waves from Galactic
supernovae if precise time of explosion is informed by SK, and implication
on the rotational speed and the stiffness of EOS could be obtained.
Despite almost 40 years of intensive and extensive studies, we still do
not figure out how the collapse-driven supernova occurs 。
More extensive and systematic studies on gravitational collapse
including realistic EOS and neutrino transfer are necessary.