Gravitational-wave Physics&Astronomy Workshop, January 28, 2011
Neutrino observatories
M.Nakahata
Kamioka observatory
ICRR/IPMU, Univ. of Tokyo
SN1987A
Contents
 Supernova burst neutrinos




How they are produced
SN1987A
Current detectors in the world
What information neutrino detectors can provide
 High energy astrophysical neutrinos
 How they are produced
 Detectors in the world
Core-collapse supernova
Standard scenario of the core-collapse supernova
Core-collapse
Neutrino trapping
Core bounce
C+O
He
H
n
Si
Fe
n
n
Supernova burst
n
n
n
Shock wave propagation
n
Shock wave at core
n
Neutron star
Figure from K.Sato
n
n
n
n
Neutrino emission from supernova burst
Released gravitational energy: ~3x1053 erg
Expected time profile
Neutrinos carry almost all (99%) of the energy.
Energy for explosion and optical emission is
only ~1%(~1051erg).
Livermore simulation
(x=m,t)
Mean neutrino energy
(x=m,t)
T.Totani, K.Sato, H.E.Dalhed and J.R.Wilson, ApJ.496,216(1998)
SN1987A: supernova at LMC(50kpc)
Feb.23, 1987 at 7:35UT
Kam-II (11 evts.)
IMB-3 (8 evts.)
Baksan (5 evts.)
24 events total
IMB-3
Total Binding Energy
Kamiokande-II
from G.Raffelt
BAKSAN
95 % CL
Contours
Theory
_
Spectral ne Temperature
Supernova burst detectors in the world
(running and near future experiments)
Super-Kamiokande
Borexino LVD
Baksan
SNO+
KamLAND
(under construction)
HALO
IceCube
The Baksan underground scintillation telescope
(Russia)
E.N.Alexeyev, L.N.Alexeyeva, astro-ph/0212499
Total number of standard detectors…………..3150
Total target mass…………………….…...330 tons of oil-based scintillator
~100 nep e+n events expected for 10 kpc SN.
Running since 1980 with ~90% live time.
From E.N.Alexeyev
LVD detector
LVD consists of an array of
840 counters, 1.5 m3 each.
Total target:
1000 t liquid scintillator
4MeV threshold
With <1MeV threshold for
delayed signal (neutron tagging
efficiency of 50 +- 10 %)
E resolution: 13%(1s) at 15MeV
~300 nep e+n events
expected for 10 kpc SN.
From W.Flugione
Twenty
Years after SN1987A
At Gran SassoFulgione@to.infn.it
Lab., Italy
Single volume liquid scintillator detectors
KamLAND
(Kamioka, Japan)
Borexino
(Gran Sasso, Italy)
SNO+
(SNO Lab.,Canada)
300ton liq.sci.
Running since 2007.
1000ton liq.sci.
1000ton liq.sci.
Running since 2002.
From K.Inoue, G.Bellini, M.Chen
Under construction.
Liquid scinitillator detectors
Expected number of events(for 10kpc SN)
Events/1000 tons

Inverse beta( ne+p→e++n) :
～300 events
Spectrum measurement with good energy resolution, e.g. for
spectrum distortion of earth matter effect.
CC on 12C (ne+12C→e+12N(12B)) :
～30 events
Electron scattering (n+e-→ n+e-) :
 NC g from 12C (n+12C→n+12C+g):
～20 events
～60 events

Tagged by 12N(12B) beta decay

Total neutrino flux, 15.11MeV mono-energetic gamma

n+p scattering（ n+p→ n+p）:
～300 events
Sensitive to all types of neutrinos.
(Independent from neutrino oscillation)
Spectrum measurement of higher energy component.
From K.Inoue, G.Bellini, M.Chen
HALO - a Helium and Lead Observatory
SNO
3He
(SNO Lab., Canada)
neutron detectors with lead target
CC:
NC:
HALO-1 is using an available 76 tonnes of Pb
In HALO-1 for a SN @ 10kpc†,
• Assuming FD distribution with T=8 MeV for
nμ’s, nτ’s.
• 65 neutrons through ne charged current channels
• 20 neutrons through νx neutral current channels
~ 85 neutrons liberated;
with ~50% of detection efficiency, ~40 events expected.
HALO-2 is a future kt-scale detector
From C.Virtue
IceCube: The Giga-ton Detector Array
(South pole)
IceTop
Design Specifications
•
•
•
•
InIce
AMANDA
Fully digital detector concept
Number of strings – 75
Number of surface tanks – 160
Number of Optical modules
(DOMs)
– 4820
• Instrumented volume – 1 km3
Supernova neutrinos coherently
increase single rates of PMTs.
Construction finished on
Dec.18, 2010.
From L.Koepke, S.Yoshida
IceCube as MeV n detector
10 kpc to SN
Simulation based on Livermore model
Advantage:
 high statistics
(0.75% stat. error
@ 0.5s and 100ms bins)
Good for fine time
structures (noise low)!
Disadvantage:
 no pointing
 no energy
 intrinsic noise
Significance:
Galactic center: ~200 s
LMC
: ~5 s
SMC
: ~4 s
Galactic
Center
IceCube
LMC
SMC
Amanda
From L.Koepke
Super-Kamiokande detector
(Kamioka, Japan)
LINAC
Electronics hut
Water and air
purification system
Control room


Atotsu
entrance

ID
41.4m

OD
Ikeno-yama
1km

Kamioka-cho, Gifu (2700mwe)
Japan
3km
2km

39.3m
Mozumi
SK
Atotsu


50,000 ton water
32,000 ton photosensitive volume
~2m OD viewed
by 1885 8-inch
PMTs
32kt ID viewed by
11,100 20-inch
PMTs
22.5kt fid. vol.
(2m from wall)
~4.5MeV energy
threshold
SK-I: April 1996~
SK-IV is running
Super-K: Expected number of events
Neutrino flux and energy spectrum from Livermore simulation
(T.Totani, K.Sato, H.E.Dalhed and J.R.Wilson, ApJ.496,216(1998))
~7,300 ne+p events
~300 n+e events
~360 16O NC g events
~100 16O CC events
(with 5MeV thr.)
for 10 kpc supernova
Summary of current supernova n detectors
# of events expected for 10kpc.
Baksan
(1980-)
LVD
(1992-)
Super-K
(1996-)
KamLAND
(2002-)
ICECUBE
(2005-)
Directionality
330 ton liquid scintillator
~100 nep e+n events.
No
1000 ton liquid scintillator. 840 counters 1.5m3 each. 4 MeV
thres., ~50% eff. for tagging decayed signal.
~300 nep e+n events.
No
32,000 tons of water target.
~7300 nep e+n, ~300 nene scattering events.
Yes
1000 ton liquid scintillator, single volume.
~300 nep , several 10 CC on 12C, ~60 NC g, ~300 np  np
Gigaton ice target.
By coherent increase of PMT single rates.
High precision time structure measurement.
No
BOREXINO 300 ton liquid scintillator, single volume.
(2007-)
~100 nep , ~10 CC on 12C, ~20 NC g, ~100 np  np
HALO
SNO 3He neutron detectors with 76 ton lead target.
~40 events expected.
(2010-)
No
No
No
Super-Kamiokande: Water Cherenkov detector
Cherenkov light
Emitted when particle
travels faster than speed
of light in water(c/1.33）.
Emission angle =
e
cos-1(1/nb)=42°
q
• Using hit PMT timing, vertex position is
reconstructed.
• Then, direction is reconstructed using hit
PMT pattern.
One of the SN1987A events in Kamiokande
Neutrino interaction in water
Cross section (H2O target)
Supernova n
Angular distribution
ν＋e－→ν＋ｅ－
νｅ＋ｐ→ｅ++n
νｅ+１６O→e-+１６F
νｅ+１６O→e++１６N
Neutrino
COSqSN
Super-K: simulation of angular distribution
n+e
n+e
SN at 10kpc
ne+p
ne+p
n+e
n+e
Direction of supernova can
be determined with an
accuracy of ~5 degree.
ne+p
ne+p
Neutrino flux and spectrum
from Livermore simulation
Future possible improvement in Super-K
Identify nep events by neutron tagging with Gadolinium.
ne
n
p
+
e
Gd
γ
8 MeV
ΔT~20μs
Vertices within 50cm
The main purpose of this
project is to detect diffuse
supernova neutrinos.
Captures on Gd
Gadolinium has large neutron capture cross section and emit
8MeV gamma cascade.
100%
80%
0.1% Gd gives
>90% efficiency
for n capture
In Super-K this means
~100 tons of water soluble
GdCl3 or Gd2(SO4)3
60%
40%
20%
0%0.0001% 0.001% 0.01%
0.1%
Gd in Water
1%
Super-K: Angular distribution (with Gd)
n+e
n+e
With nep
identification by
neutron tagging
ne+p
ne+p
n+e
ne+p
ne+p
n+e
Direction accuracy can be
improved to ~3 degree.
SN at 10kpc
Burst onset time
Simulation of initial stage of a burst (10kpc supernova)
Super-Kamiokande
IceCube
Halzen, Raffelt, arXiv:0908.2317
IceCube
example
10~40 events in the first 20msec.
±3.5 msec at 95% C.L.
Super-K and IceCube have a few msec precision on onset time.
Triangulation on supernova direction
Beacom, Vogel, Phys.Rev.D60, 033007, 1999
Super-K (36˚N)
cosq = t / d
q
IceCube (90˚S)
If Super-K and IceCube has ~2 msec rise time resolution,
cos(q) = ~0.1 (i.e. about 25 deg.).
Super-K resolution by electron-scattering is much better.
SuperNova Early Warning System
snews.bnl.gov
Details: arXiv:astro-ph/0406214
arXiv:0803.0531
Individual supernova-sensitive experiments send burst
datagrams to SNEWS coincidence computer at
Brookhaven National Lab(backup at U. of Bologna)
Email alert to astronomers if coincidence in 10 seconds
Participating experiments:
SuperKamiokande
(Japan)
From K. Scholberg
Large
Volume
Detector
(Italy)
IceCube
(South Pole)
SNO
(Canada)
until end of 2006
High Energy Astrophysical Neutrinos
There are high energy accelerators in the universe.
We know cosmic rays(p, He, ..) exist.
So, there much be cosmic high energy neutrinos.
Possible sources: AGN, supernova remnants, ….
ne
e+
nm
m+ nm
+
gsync
p
e-
gsync
Black hole
accretion disk
Cosmic Ultra High Energy neutrinos
Ultra High Energy Cosmic Ray(UHECR) exists.
“Horizon” of UHECR is ~50Mpc.
Neutrinos
galactic
Extragalactic
cosmic rays interact with the
microwave background
p g    n 

GZK neutrinos

  m   m  {e   m  e }   m
n e :n m :n t  1 : 1 : 1
(due to n osc.)
Detectors for High Energy Astrophysical Neutrinos
Antares
Mediterranean sea
0.2x0.2x0.4km
Volume: 0.01 km3
Running since
May 2008
South pole
Volume:1 km3
Construction finished in Dec.2010
NT200+
Lake Bikal
0.2 kmf 0.2 kmh
Volume: 0.01 km3
Running from
2006
Diffuse n limit
Now below the Waxman-Bahcall limit
IceCube Preliminary
nm only
Atmospheric
n
This work
From S.Yoshida
Sean Grullon (UW-Madison)
GZK
n
search
All flavor
limits
(ne  nm  nt)
Systematic errors included
IceCube Preliminary
Aya Ishihara (Chiba)
From S.Yoshida
Conclusions
• Supernova burst neutrinos
–
–
–
–
Many detectors in the world with various types of signals.
~8,000 events expected at Super-K for 10kpc SN.
High precision time profile by Icecube.
~5 deg. accuracy in direction of supernova by Super-K
neutron-electron scattering events.
– Onset time with ~2 msec resolution by Super-K and
IceCube.
• High energy neutrinos
– Construction of the IceCube was finished.
– High energy neutrino events are expected in near future.
Backups
Distance to Galactic supernova
Mirizzi, Raffelt and Serpico, JCAP 0605,012(2006),
astro-ph/0604300
Based on birth location of neutron stars
Type Ia
0
7% probability
< 3.16 kpc
> x10 statistics
10kpc
16% probability
< 5 kpc
> x 4 statistics
Core collapse type
mean: 10.7 kpc
r.m.s.: 4.9 kpc
20kpc
3% probability
> 20 kpc
< 1/4 statistics
Super-K: Angular distribution (without Gd)
n+e
n+e
Without nep
identification by
neutron tagging
ne+p
ne+p
n+e
ne+p
n+e
ne+p
SN at 10kpc
Accuracy of supernova direction with neutron tag
Thomas, Sekikoz, Raffelt, Kachelriess, Dighe, hep-ph/0307050v2
Accuracy of SN direction with
40000 simulated supernova
Accuracy can be improved by a factor
of two with neutron tagging.
Tagging efficiency vs.
accuracy (95% CL)
G: Garching model, L: Livermore model
a: normal hierarchy sin2q13>10-3
b: inv. hierarchy sin2q13>10-3
c: any hierarchy sin2q13<10-3
Angular distribution of νｅ＋ｐ→ｅ++n
νｅ＋ｐ→ｅ+＋ n
5MeV
10MeV
neutrino energy
20MeV
30MeV
A. Strumia and F. Vissani,
Phys.Lett.B564:42-54,2003.
COSqSN
Super-K: Neutronization burst
(e-+pn+ne)
SN at 10kpc
Neutrino flux and spectrum from Livermore simulation
Event rate of ne+p events
Event rate of neutronization burst
(forward peaked n+e scattering events)
No oscillation
Normal PH=1 or
Inverted hierarchy
Normal hierarchy PH=0
PH: crossing probability at H resonance
(PH=0: adiabatic)
Number of events from neutronization burst is 0.9~6 events for SN@10kpc.
nep events during this 10msec is about 8 - 30 events.
N.H. +adiamacitc case: neutronization=0.9ev., nep = 14 ev.(1.4 for SN direction).
n+p elastic signal（ n+p→ n+p） at liq. Scintillator
Beacom, Farr, and Vogel, PRD66, 033001(2002)
detectors
Expected spectrum
ne
Sensitivity of temperature
measurement
ne
~300 events/kt above 200keV
~150 events/kt above 500keV
Determine original nm, nm, nt, nt temperature
with ~10% accuracy.
(free from neutrino oscillation.）
Current Borexino threshold: 200keV
Current KamLAND threshold: 600~700keV(will be lowered after 2008 distillation.)