Transcript Slide 1

Selected references to this lecture:
Kamland papers:
K.Eguchi et al., Phys. Rev. Lett.90, 021802
(2003),
K.Eguchi et al., Phys. Rev. Lett.92, 071301
(2004),
T. Araki et al., hep-ex/0406035 .
General review:
C. Bemporad, G. Gratta, and P. Vogel., Rev. Mod.
Phys. 74, 297 (2002).
• Pontecorvo already in 1946 suggested to
use nuclear reactors in order to perform
neutrino experiments.
• Indeed, in 1953-1959 Reines and Cowan
showed that neutrinos are real particles
using nuclear reactors as a source.
• Since then, reactors, powerful sources
with ~6x1020 /s electron antineutrinos
emitted by a modern ~3.8 GWthermal reactor,
have been used often in neutrino studies.
• The spectrum is well understood….
Electron antineutrinos are produced by
the b decay of fission fragments
Reactor spectrum:
1) Fission yields Y(Z,A,t), essentially all known
2) b decay branching ratios bn,i(E0i) for decay branch i,
with endpoint E0i , some known but some (particularly
for the very short-lived and hence high Q-value)
unknown.
3) b decay shape, assumed allowed shape, known
P(En,E0i,Z) or for electrons Ee= E0 – En.
dN/dE = Sn Yn(Z,A,t) Si bn,i(E0i) P(En,E0i,Z)
and a similar formula for electrons.
If the electron spectrum is known, it can be `converted’
into the antineutrino spectrum.
Spectrum Uncertainties
Theory only
Klapdor and Metzinger,
1982
Beta calibrated
Schreckenbach, 1985
Hahn, 1989
Bemporad, Gratta, and Vogel, RMP 74, 297 (2002)
Results of Bugey experiment (1996)
Reactor spectra
Detecting reactor antineutrinos;
low detection threshold required
Detector reaction ne + p -> e+ + n, positron spectrum measured
Cross section
known to ~0.2%,
see Vogel & Beacom,
Phys. Rev. D60,053003
To study oscillations, use the disappearance test:
The survival probability of electron
antineutrinos of energy En produced
at the distance L from the detector is
Pee(En,L) = 1 – sin2(2q)sin2(Dm2L/4En)
The experiment become sensitive to
oscillations if Dm2L/En ~ 1,
proof of oscillations is Pee(En,L) < 1.
History of reactor neutrino oscillation search:
Probability of
oscillations is
proportional to
sin2(Dm2L/4E).
Since for the
reactors E~4MeV,
the sensitivity to
Dm2 is inversely
proportional to
the distance L.
• Discovery of oscillations of atmospheric
neutrinos implies Dm2 ~ (2-3)x10-3 eV2,
thus indicating that reactor experiments
with L ~ (1-3) km should be performed
(Chooz and Palo Verde).
• Also, the preferred `solution’ to the solar
neutrino deficit implies Dm2 ~ (5-10)x10-5 eV2,
thus indicating that reactor experiments
with L ~ 100 km should be performed
(KamLAND)
~180 km
~80 GW : 6% of world nuclear power
~25GW :
most powerful station
in the world
KamLAND Collaboration
13 institutions & 93 members
Ｃｏｌｌａｂｏｒａｔｏ
ｒ
KamLAND Experiment
300
antineutrinos from the Sun . . .
Live time
(days)
Jan 2002
-
Run A
(data-set of 1st paper)
Mar 9 – Oct 6
2002
145.4*
Electronics upgrade & 20”
PMT commissioning
Jan/Feb 2003
-
Run B
Oct - Jan 11
2004
369.7
Data-set presented here†
Mar 9, 2002 –
Jan 11, 2004
515.1
Start data taking
†
*Was
Dates
145.1 with
old analysis
A brief history of KamLAND
T.Araki et al, arXiv:hep-ex/0406035 Jun 13, 04
submitted to Phys Rev Lett
A limited range of baselines contribute to the flux
of reactor antineutrinos at Kamioka
Over the data period
Reported here
Korean reactors
3.4±0.3%
Rest of the world
+JP research reactors
1.1±0.5%
Japanese spent fuel
0.04±0.02%
m- Induced
Neutrons & Spallation-12B/12N
DL < 3m
12N
12B
DT (T-Tm ), DL
Tagged cosmogenics can be used for calibration
12B
τ=29.1ms
Q=13.4MeV
12N
τ=15.9ms
Q=17.3MeV
μ
Fit to data shows that
12B:12N ~ 100:1
Radioactivity in liquid scintillator
238U: 214Bi
→ 214Po → 210Pb
β+γ
τ=28.7 m
E=3.27 MeV
232Th: 212Bi
α
τ=237 μs
E=7.69 MeV
→ 212Po →
β+γ
τ=87.4 m
E=2.25 MeV
208Pb
α
τ=440 ns
E=8.79 MeV
238U:
(3.5±0.5)∙10-18 g/g
232Th: (5.2±0.8)∙10-17 g/g
τ=(219±29) μs Expected: 237 μs
needed 10-14 g/g
Note: The best background in 76Ge bb decay
detectors is at present ~0.2 counts/(keV kg y).
Expressing the background in the liquid scintillator
in KamLAND in the same units, and for
energies 2-3 MeV, one finds value ~10 times
smaller going out to 5.5 m radius and ~20 times
smaller for 5 m radius
Systematic Uncertainties
E > 2.6 MeV
Total LS mass
Fiducial mass ratio
Energy threshold
Tagging efficiency
Live time
Reactor power
Fuel composition
Time lag
ne spectra
Cross section
# of target protons
Total Error
%
2.1
4.1
2.1
2.1
0.07
2.1
1.0
0.28
2.5
0.2
< 0.1
6.4 %
5 % : goal
4. 6
Very clean measurement
Accidental
background
Second KamLAND paper
Expect 1.5 n-12C
captures
2003 saw a substantial dip in reactor antineutrino flux
Good correlation with reactor flux
Fit constrained
through known
background
χ2=2.1/4
90% CL
~0.03 for
3TW
hypothetical
Earth core
reactor
(But a horizontal line still gives a decent fit with χ2=5.4/4)
In CHOOZ
it was
possible
to determine
background
by this effect.
Results
(766.3 ton·yr,
~4.7 the statistics of the first paper)
Observed events 258
No osc. expected 365±24(syst)
Background
7.5±1.3
Background
Events
Accidentals
2.69±0.02
8He/9Li
4.8±0.9
μ-induced n
<0.89
Total
7.5±1.3
Inconsistent with simple 1/R2 propagation
at 99.995% CL
(Observed-Background)/Expected = 0.686±0.044(stat)±0.045(syst)
Caveat: this specific number does not have an absolute meaning in KamLAND,
since, with oscillations, it depends on which reactors are on/off
Second paper
Decay chain leading to
210Po:
222Rn
a(3.8d) 218Po a(3.1m) 214Pb b(27m) 214Bi,
214Bi b(20m) 214Po a(164ms) 210Pb b(22.3y) 210Bi,
210Bi b(5d) 210Po a(138d) 206Pb(stable)
The long lifetime of 210Pb causes its accumulation.
The a from 210Po decay then interact with 13C in
the scintillator by 13C(a,n)16O making unwanted
background. There is only ~10-11g of 210Pb in
fhe fiducial volume, enough however to cause
1.7x109a decays in 514 days.
Ratio of Measured to Expected ne Flux
from Reactor Neutrino Experiments
LMA:
Dm2 = 5.5x10-5 eV2
sin2 2Q = 0.833
G.Fogli et al., PR
D66, 010001-406,
(2002)
Energy spectrum now adds substantial information
Best fit to
oscillations:
Δm2=8.3·10-5 eV2
sin22θ=0.83
Straightforward
χ2 on the histo
is 19.6/11
Using equal
probability bins
A fit to a simple rescaled reactor spectrum
is excluded at 99.89% CL (χ2=43.4/19)
Second paper
χ2/dof=18.3/18
(goodness
of fit is 42%)
This result
Δm2=8.3·10-5
eV2
sin22θ=0.83
First KamLAND result
Dm2 = 6.9 x 10-5 eV2
sin2 2 q = 1.0
Combined solar ν – KamLAND 2-flavor analysis
0.6 105 eV 2
Dm122  8.2 
 0.5
0.09
tan2 q12  0.40 
 0.07
Includes (small) matter effects
KamLAND uses a range of L and
it cannot assign a specific L to each event
Nevertheless the ratio of detected/expected
for L0/E (or 1/E) is an interesting quantity, as it decouples
the oscillation pattern from the reactor energy spectrum
no oscillation expectation
p/2
p
Hypothetical
single 180km
baseline
experiment
Conclusions
KamLAND reactor exposure: 766.3 ton·yr (470% increase)
Data consistent with large flux swings in 2003
Spectrum distortions now quite significant, shape-only very powerful
Best KamLAND fit to oscillations Δm=8.3·10-5 eV2, sin22θ=0.83
LMA2 is now excluded
2
Together with solar ν Dm12
 8.2  0.6 105 eV 2 ; tan2 q12  0.40 0.09
 0.5
 0.07
Welcome to precision neutrino physics !
What next in reactor neutrino studies?
What’s next?
•Purification of liquid scintillator; remove 85Kr
and 210Pb (low energy background) and
attempt to observe solar 7Be ne
(feasibility studies under way).
•Determine or constrain the flux of solar
antineutrinos and of the geoneutrinos. Study
the neutron production by muons.
•In a different reactor experiment (two detectors,
one close and another at ~1-2 km) determine or
constrain the unknown mixing angle q13. That
is a whole different story.
Future Reactor Measurements
Chooz reached at ~1 km
2.8% statistical error
2.7% systematic error
Next generation search for
Theta-13 needs to achieve
~1% errors
Apollonio et al., EPJ C27, 331 (2003)
A high sensitivity search for ne from the
Sun and other sources at KamLAND
hep-ex/0310047, Phys. Rev. Lett. 92, 071301 (2004)
No events found between
8.3-14.8 MeV for 0.28kt-y
exposure. Assuming the
8B spectrum shape, this
limits the antineutrino flux
to 2.8x10-4 of the SSM 8B
flux.
This represents a factor of
30 improvement over the
best previous limit.
Thanks to Atsuto Suzuki, Patrick Decowski,
Gianni Fiorentini, Andreas Piepke and
Giorgio Gratta who made some of the
figures used in this talk.