1. Results and plans of the KamLAND experiment
Let me talk about `Results and plans of the KamLAND experiment’
Yoshihito Gando (RCνS, Tohoku Univ.)
for the KamLAND Collaboration
Chung-li, Taiwan, 4 October 2005

2. Various Physics Targets with wide energy range
0.4
1.0
2.6
8.5
Visible energy　[MeV]
neutrino electron elastic scattering
inverse beta decay
Neutrino Astrophysics
verification of SSM
Neutrino Geophysics
verification of earth evolution model
Neutrino Physics
Precision measurement of oscillation parameters
Neutrino Cosmology
verification of universe evolution
7Be solar neutrino
geo-neutrino
reactor neutrino
supernova relic neutrino
etc.
future
2nd phase
Nature 436, 28 (2005)
1st results
PRL 90, (2003)
2nd results
PRL 94, (2005)
Solar
PRL 92, (2004)

3. KamLAND detector Photo - coverage: 34% ~ 500 p.e. / MeV 13m 1000m
Cosmic ray 's are suppressed
by 1/100,000.
20 inch : 225
13m
1,000 ton liquid scintillator
Dodecane : 80%
Pseudocumene : 20%
PPO : 1.5g/l
Mineral oil
Dodecane : 50%
Isoparaffin : 50%
1.75m thickness
KamLAND is located 1000m underground in Kamioka mine, and muon event rate is about 0.34Hz.
The detector consist of 1000 tons of ultra-pure liquid scintillator and 1879 PMTs
17 inch :1325
20 inch : 554
~8000 photons / MeV
λ: ~10m
Photo - coverage: 34%
~ 500 p.e. / MeV

4. e ν detection in KamLAND e- Greatly removes backgrounds e+ Position
e+ + n e
(0.51)
Prompt e+ signal e e- e+
Te++annihilation
=Eν - 0.8MeV
Te+ p
E1.8MeV
(0.51) n
 (2.2 MeV) p
Delayed γ by
neutron capture
~210μs
Position
Time correlation
delayed energy information
Neutrinos are detected by the inverse beta-decay reaction.
Space and time correlations of prompt and delayed signal provide effective background reduction. d
Greatly removes backgrounds

5. Reactor Neutrino
As a first physical topic, I want to talk about reactor neutrinos.

6. Reactors near the KamLAND
80% of total contribution comes from 130~220km distance
effective distance ~180km
This map shows the location of the Japanese power reactors and KamLAND.
And this figure shows distances from KamLAND to reactors and thermal power of reactors.
80% of total contribution comes from 130 ~ 220km distance.
KamLAND group also calculate effects from reactors of other countries, Taiwan effect is about 0.1%
Reactor neutrino flux,
~95.5% from Japan
~3% from Korea
(2nd result period)
Reactors in Taiwan have ~0.1% contribution.

7. Event Selection μ 9Li Delayed Coincidence: 0.5 < ΔT < 1000μsec
ΔR < 200 cm
1.8 < Edelayed < 2.6 MeV
Prompt Energy Window:
2.6 < Eprompt < 8.5 MeV
Fiducial Volume:
Rprompt < 550 cm (500 :1st result)
Rdelayed < 550 cm (500 : 1st result) μ
Spallation Cuts:
ΔTμ < 2 msec
　ΔTμ < 2 sec (showering muons)
　　　　　　　　 or
　　　　ΔTμ < 2 sec (showering muons)
　　　　ΔL 　< 300 cm (non-showering)
In order to select reactor neutrinos, we applied these event selection,
delayed coincidence, fiducial volume, 3m
9Li
Isotope Halflives Decay Mode
9Li/8He 178.3ms/119.0ms β－ + n
Efficiency : 89.82%(I,II), 89.83%(III)

8. Time Variation of Reactor ν
Expected event rate
First result
Expected signal : 86.8±5.6
BG :1±1
Observed : 54
Neutrino disappearance at 99.95%
Observed event rate
I want to show you the results of reactor neutrino analysis.
This figure shows time variation of observed reactor neutrinos and expected event rate.
We observed 258 events, expected signals are about 365 and estimated backgrounds are about 18.
Observed / expected ratio is and this shows neutrino disappearance at % confidence level.
R = ±0.044(stat) ± 0.042(syst) ⇒ neutrino disappearance at % C.L.
High statistic from 1st result oscillation study

9. Energy Spectrum
This figure shows energy spectra of KamLAND data, no oscillation expected, scaled no oscillation expected, and backgrounds.
From the hypothesis test of scaled no-oscillation, spectral distortion is 99.6% confidence level.
And we use rate with shape information, no oscillation is excluded at % confidence level.
Hypothesis test of scaled no-oscillation: χ2/ndf = 37.3/18 ⇒ spectral distortion at > 99.6% C.L.
Rate + Shape: no oscillation is excluded at % C.L.

10. L/E plot with data for geo-ν analysis
(759 days, 5m fiducial)
low energy window
best fit reactor +
geo-neutrino model prediction
Oscillation pattern
with real reactor
distribution
Lo = 180 km is used for KamLAND
There is clear Oscillatory behavior (peak and dip)
oscillation parameter is determined.

11. several orders -> less than 10%
Oscillation Analysis
KamLAND best-fit (rate + shape)
KamLAND + Solar
assuming CPT invariance
Next is oscillation analysis. Left figure shows allowed region contour for oscillation parameters.
When we combine KamLAND data and solar global data, allowed region is singled out to LMA1 under the assuming CPT invariance.
several orders -> less than 10%
Precise determination of oscillation parameters made possible to use neutrinos as a new probe.

12. Geo - Neutrino Next topic is Geo – Neutrino analysis.
As I presented before, we have oscillation parameters.
Therefore we can analyze geo-neutrinos without oscillation uncertainty.

13. Earth Energetics
Terrible earthquakes, eruptions, etc. are originally caused
by mantle convection driven by heat.
Terrestrial magnetism is caused by a core movement,
it requires some heat source.
Observed Surface Heat Flow :
～44TW (31TW : re-evaluation)
Radiogenic Heart : ~20TW ?
U-chain 8TW / Th-chain 8TW / 40K 3TW??? (BSE model)
Radioactive heat sources contribute to about the half of the total heat outflow of the earth.
Geo-Neutrino is Produced by β-decay of radioactive element in the earth
I said radiogenic heart is 20TW, but it is also important topic.
Mantle conversion, terrestrial magnetic field etc. are driven by heart.
And it is said that radioactive heart sources contribute to about the half of the total heart outflow of the earth.
If we observe geo-neutrinos, we can identify one of the major heart source.

14. Methods of research about inside of the earth
Seismic analysis
Meteorite analysis
Composition of the earth
(Proto material ) is presumed
by meteorite analysis
BSE (bulk silicate Earth) model
McDonough et al.(1995)
Th/U mass ratio ~ 3.9
It expect that 20TW comes
from radioactivity
There are no direct measurements
In order to understand the earth, these methods have been used at geophysics.
For the physical composition, seismic analysis is most important method.
This analysis shows density, elastic parameter etc, in the earth, and indicate mantle conversion. But, it don’t tell chemical composition.
For the composition of the earth is presumed by meteorite analysis.
It is called BSE model and this model predict these information, average
composition on the earth, Uranium / Thorium mass ratio, 20TW comes from radioactivity.
But, there are no direct measurement.
Next methods is boring. It is direct measurement. But, it is only up to 7 km.
So, we want direct measurement in the earth.
Physical parameter
(density, elastic parameter etc…)
It does not tell chemical composition
Direct measurement is desired!!

15. Geophysics with Neutrino
Determination of the amount and distribution of U, Th in the earth from geo-ν observation
- Test for BSE model
Verification of basic paradigm of geochemical
earth formation and generation
- Determination of heat balance
Information for earth dynamics, evolution,
terrestrial magnetism
- Understanding of chemical composition of
deep interior of the earth
Determination of chemical structure model
(mantle model)
From the geo-neutrino measurement, we can provide these information.
If we can determine quantity and distribution of U / Th ratio, we can test for BSE model.
BSE model is verification of basic paradigm of geochemical.
And from this verification, we can understand the earth generation process.
Next is determination of heart balance.
It provide understanding of dynamics, earth development, mechanism for production of magnetic field.
And from the understanding of chemical composition of deep interior of the earth,
We can determine chemical structure model.

16. Mantle = Meteorite (BSE model) - Crust
Reference Earth Model
Upper continental crust U : 2.8ppm / Th: 10.7ppm
Middle continental crust U : 1.6ppm / Th: 6.1ppm
Lower continental crust U : 0.2ppm / Th: 1.2ppm
Rudnick et.al. (1995)
continental crust
Oceanic crust U: 0.08ppm / Th: 0.32ppm
Th/U ~3.9
Radiogenic heat ~16TW
To estimate geo-neutrino expectation, we used this earth model.
Uranium / Thorium ratio of each crust is calculated as these value, and
Mantle composition is calculated by BSE model and crust model.
U: 0ppm / Th: 0ppm
Ionic radius of U/Th are large
Core is very high density
do not exist
mantle
Core
U: 0.012ppm / Th: 0.048ppm
Mantle = Meteorite (BSE model) - Crust

17. U/Th distribution maps in Japan
Average component of upper continental crust
Geological map + rock sample (Togashi et al.)
To estimate uncertainty of geology heterogeneity near the detector, rock sample and map is useful information.
From the assumption that the surface uranium / thorium distribution extends to 5km in depth,
Uncertainty is estimated as 3%.
U : 2.32 ppm
Th : 8.3 ppm
Assume the surface U, Th distribution extends to 5km in depth
Geo-neutrino flux is calculated from global
and local U, Th composition

18. Geo-Neutrino spectrum
From the BSE model, expected geo-neutrino fluxes are shown as this figure.
Red line shows neutrino flux from uranium-series, green is thorium-series, and blue is potassium.
The energy threshold of inverse-beta decay reaction is here, so we can detect these energy region.

19. Event Selection (Geo-ν)
Delayed Coincidence:
0.5 < ΔT < 1000μsec
ΔR < 100 cm
0.9 < Eprompt < 2.6 MeV
1.8 < Edelayed < 2.6 MeV
Fiducial Volume:
Rprompt < 500 cm
Rdelayed < 500 cm
ρxy > 120 cm
Spallation Cuts:
ΔTμ < 2 msec, total volume (for all muons)
ΔTμ < 2 sec, total volume (showering muons) or
ΔTμ < 2 sec, ΔL < 300 cm (Non-showering muons)
Efficiency U-Series : 69.2% , Th-Series : 68.0%

20. (α, n) Background αcomes from 210Po decay (daughter nuclei of 222Rn)
Unfortunately, we inputted 222Rn
at the construction
After applying event selection, most of backgrounds could be rejected.
But backgrounds from alpha n reaction of carbon 13 could not be rejected,
because these backgrounds make delayed neutron capture event.
Therefore we estimated these backgrounds statistically.
For the reactor neutrino analysis, this backgrounds are estimated as 10.3 events.
On the other hand, these events are background for geo-neutrino.
Recent paper shows few % lower cross section of alpha n reaction of 13C,
So we could reduce about background estimation.
And our collaborator plan to measure this reaction, so we may be able to reduce uncertainty of this reaction.
Recent paper shows
few % lower cross section of 13C (α,n) 16O
(Harissopulos et al, nucl-ex/ )
We could reduce about B.G. uncertainty

21. Expected spectrum reactor BG + Geo-ν BG total Reactor ν (α,n) reaction
Th-chain geo-ν
U-chain geo-ν
Accidental coincidence

22. Expected + observed spectra
749.1 live days
Observed
152
B.G.
127.4±13.3
Signal
24.2±17.9

23. Rate + Shape analysis N U+ N Th :
C.L. contours for detected
U and Th geo-s.
Th/U Mass ratio=3.9
Th/U mass
Ratio=3.9
90%CL
NU+NTh
2
54.2
4.5
(NUNTh)/(NU+NTh)
Prediction from
the BSE model
NU+NTh
N U+ N Th :
Consistent with prediction of BSE model.
We observed geo-neutrinos with 90%C.L
99% C.L. upper limit ：70.7 events

24. Geo-ν after purification
Assume 210Pb : 10-5 level
(α,n) reaction and other
radioactive backgrounds are negligible
749days data
error ： 54% %
(statistical error of reactor neutrino
is dominant)
Significance : 99.96%
precise measurement
fiducial volume :
R < 5m m
detection efficiency : 90%

25. Signal v.s. heat We will contribute to geology
Fiorentini et al. (hep-ph/ )
Re-calculation with new cross section for (α,n) reaction for 13C
99% C.L. upper limit
from KamLAND data
Signal (U+Th) [TNU]
Relationship line from geochemical
and geophysical constraints
Analysis improvement
B.G. reduction
More statistics
BSE
Fully radiogenic
Heat (U+Th) [TW]
We will contribute to geology

26. Future Plan

27. Next target of KamLAND 7Be ν : neutrino electron elastic scattering
0.4
1.0
2.6
8.5
Visible energy　[MeV]
neutrino electron elastic scattering
inverse beta decay
Neutrino Astrophysics
verification of SSM
Neutrino Geophysics
Neutrino Physics
Neutrino Cosmology
7Be solar neutrino
geo-neutrino
reactor neutrino
supernova relic neutrino
etc.
7Be ν : neutrino electron elastic scattering e
(We couldn’t use delayed coincidence methods)
Very low level background is required

28. KamLAND-II : toward solar 7Be neutrino detection
4 m radius fiducial
1.2 m cylindrical cut
Total
210Po
85Kr
210Bi
7Be
11C
Required Improvements : Pb : 10-4~10-5
85Kr, 39Ar: ~10-6

29. LS Purification 2, 3, … , times distillation (1time : ~ 1 month)
Distillation System : Test Bench
N2 gas purge (N2/LS = 25)
Rn: ~1/10 Kr : ~1/100
Distillation (110 ℃, 37 hPa, 1time) Pb:
Rn: ( ) ×10-3
Kr : <10-5
2, 3, … , times distillation (1time : ~ 1 month)
We will achieve required performance

30. Purification Outline We will start purification at next year
The specification of
the purification system
was already decided.
And the tender of the
system was started.
We will start purification at next year
and 7Be neutrino observation!!

31. After the purification…
Solar 7Be neutrino observation with few % accuracy
Solar 8B neutrino observation (<5MeV)
Solar pep , CNO neutrino (with 11C tagging)
Geo-neutrino improvements
- no backgrounds from (α,n) reaction of 13C
- accidental coincidence will be reduced
- larger fiducial volume

32. Summary Rector neutrino
- Rate + Shape analysis excluded no-oscillation
at % C.L.
- Spectrum distortion (L/E) shows oscillatory behavior.
- Oscillation parameters are precisely measured:
Geo-neutrino
- It was proven that KamLAND can detect
Geo-Neutrino for the first time.
- We observed geo-neutrinos with 90%C.L.
KamLAND-II
- For the solar 7Be neutrino detection,
purification studies have been advanced.
- We will start purification at next year.

33. LS Purification and Radioactive Impurity
In order to reduce radioactive backgrounds, liquid scintillator was purified by water extraction.
After the purification, uranium, thorium and potassium were reduced to be these levels.
And, these background level can be measured by only KamLAND itself.
before
U: ~10-10 g/g, Th: <10-12 g/g, K: 7×10-11 g/g
after
U: 3.5×10-18 g/g, Th: 5.2×10-17 g/g, K: 2.7×10-16 g/g
measurable only by KamLAND itself !

34. Detector Calibration Radio-Active Source Deployment Muon Spallation
Products
Vertex Resolution
Detector calibration is performed by radio-active source deployment and muon spallation products.
Energy Resolution
Fiducial Volume Error: 4.7%

35. Detector Activity (Singles Spectrum)
Normal Trigger Range
Low Energy Region
Major Background Sources: LS impurity (210Pb, 85Kr, 39Ar)
extrinsic gamma (40K, 208Tl)
muon spallation (10C, 11C, 12B, ...)

36. Event Display : Low Energy Event

37. Event Display : Muon Event

38. Event Selection(1) Delayed Coincidence: 0.5 < ΔT < 1000μsec
ΔR < 200 cm
1.8 < Edelayed < 2.6 MeV
Prompt Energy Window:
2.6 < Eprompt < 8.5 MeV
Fiducial Volume:
Rprompt < 550 cm
Rdelayed < 550 cm
12C captured γ
In order to select reactor neutrinos, we applied these event selection,
delayed coincidence, fiducial volume,

39. Event Selection(2) μ 9Li Spallation Cuts: 3m ΔTμ < 2 msec
Isotope Halflives Decay Mode
6He 806.7ms β－
7Be 53.24day EC
8Li 838ms β－
8B 170ms β－
9C 126.5ms β＋
10C 19.25sec β＋
11Be 13.81sec β－
11C 20.39min β＋
9Li/8He 178.3ms/119.0ms β－ + n μ
And spallation cuts.
9Li
Spallation Cuts:
ΔTμ < 2 msec
　ΔTμ < 2 sec (showering muons)
　　　　　　　　 or
　　　　ΔTμ < 2 sec (showering muons)
　　　　ΔL 　< 300 cm (non-showering) 3m

40. (α, n) Background Recent paper shows
After applying event selection, most of backgrounds could be rejected.
But backgrounds from alpha n reaction of carbon 13 could not be rejected,
because these backgrounds make delayed neutron capture event.
Therefore we estimated these backgrounds statistically.
For the reactor neutrino analysis, this backgrounds are estimated as 10.3 events.
On the other hand, these events are background for geo-neutrino.
Recent paper shows few % lower cross section of alpha n reaction of 13C,
So we could reduce about background estimation.
And our collaborator plan to measure this reaction, so we may be able to reduce uncertainty of this reaction.
Recent paper shows
few % lower cross section of 13C (α,n) 16O
(Harissopulos et al, nucl-ex/ )
We could reduce about B.G. estimation

41. Accidental Coincidence Background
Off - time coincidence spectrum
⇒ ± 0.02 events

42. (α, n) Background 222Rn 210Pb 210Bi 210Po 206Pb α 13C (α,n) 16O
22.3 y
210Bi
210Po
206Pb
5.013 d
138.4 d
stable α
(5.3 MeV)
13C (α,n) 16O
13C (α,n) 16O*
16O*(6.13) → 16O + γ (6.1MeV)
16O*(6.05) → 16O + e+ + e－(6.0MeV)
14N (α,n) 17F
15N (α,n) 18F
n + p → n + p (B.G for Geo neutrino)
17O (α,n) 20Ne n
n + 12C →
18O (α,n) 21Ne
n + 12C*
12C + γ(4.4MeV)

43. Backgrounds Summary

46. Correlation with Reactor Power
constrained to expected BG
at present statistics is not enough to state something

51. (α, n) Background

52. Energy Scale Determination

53. Fiducial Volume Calibration
With Muon Spallation (12B)

54. Systematic Errors Summary (Reactor-ν)

55. Systematic Errors Summary (Geo-ν)
Cross section
Livetime
Fiducial volume
Trigger efficiency
(U / Th / Reactor) / 0.09 / 0.007
Spatial Cut Efficiency
Timing Cut Efficiency
Total

56. νdetection efficiency (Reactor)
Space correlation
Time correlation
MC simulation
Vertex resolution: 30cm/√E(MeV)
Capture time of spallation neutron
211.2±2.6μs
99.84%
ΔR(<2m) cut
Parameter
Efficiency(%)
Space correlation
91.32±1.49
Time correlation
98.89±0.05
Trigger efficiency -
Delayed energy
99.98
Neutron capture
99.48(I,II),99.48(III)
Total
89.82(I,II),89.83(III)
91.32±1.49%
Fiducial cut

57. Detection efficiency (Geo-ν)
Neutron capture
99.5 %
Trigger
U-Series: %
Th-Series: %
Spatial
Correlation
U-Series: %
Th-Series: %
Reactor: %
(α,n): %
Time correlation
90.4%
Energy of delayed event
99.97%
Spatial Correlation (MC)
MC/Data Comparison
total
U-Series: % Th-Series: 68.0%

58. Event Selection (Geo-ν)
Delayed Coincidence:
0.5 < ΔT < 1000μsec
ΔR < 100 cm
0.9 < Eprompt < 2.6 MeV
1.8 < Edelayed < 2.6 MeV
Fiducial Volume:
Rprompt < 500 cm
Rdelayed < 500 cm
ρxy > 120 cm
Spallation Cuts:
ΔTμ < 2 msec, total volume (for all muons)
ΔTμ < 2 sec, total volume (showering muons) or
ΔTμ < 2 sec, ΔL < 300 cm (Non-showering muons)

59. Backgrounds (Geo-ν) total 127.4±13.3
Cosmic ray muon
Neutron (inner of detector) negligible
Fast neutron (external) < 0.1
Spallation (9Li) ±0.047
Radioactive contamination
accidental coincidence ±0.0077
spontaneous fission < 0.1
correlated fission negligible
(α, n) reaction ±11.1
(γ, n) reaction negligible
Reactor neutrino ±7.2
Long lived nuclear (spent fuel rod) ±0.2
total ±13.3

60. Time variation of reactor neutrino flux
1 / 2
Time
Neutrino flux from
distance of ~160km
decreased.
190km
Oscillation pattern
depend on this
variation.
160km

61. L/E Analysis χ2/ndf GOF spectrum shape test 24.2/17 11.1% 35.8/17 0.7%
24.2/ %
35.8/ %
32.2/ %
Next figure shows L/E analysis. Horizontal axis is L/E and vertical axis is observed/expected.
Red dots show KamLAND data and this line is best-fit oscillation expected. You can see oscillatory behavior.
Observed / expected

62. 11C Tagging

63. Neutrino Propagation through the Earth

64. Mantle or Oceanic crust?
Seismic wave velocity anomaly
KamLAND
Subducting plate
Low speed
(high temp.)
Accumulation of
cold slab?
Subducting plate thickness ~50km
(oceanic crust ~6km)
Cold slab
Oceanic crust : mantle = 1 : 9
For the calculation of geo-neutrino flux, we should consider some uncertainties.
BSE model assume uniform chemical composition, but there is seismic wave velocity anomaly under the Japan.
This analysis, high speed region exist under the Japan, and it is similar to subducting plate.
Uranium / Thorium composition of oceanic crust is higher than mantle, so it makes uncertainty.
We estimated this uncertainty is about 2%
high speed
(low temp.)
Effect of the high speed region gives
~2% uncertainty of the total neutrino flux

65. Distance and Cumulative Flux
Total
<500km
50%
Crust
Mantle
This figure shows distance and cumulative neutrino flux.
50% of the total neutrino flux originates within 500km.
So, for the discussion of deep interior of the earth, we need understanding about surface geological features within 500km.
Sediment
50% of the total flux originates within 500km.
For the discussion of deep interior of the earth,
we need understanding about surface geological features within ~500km

66. Result of KamLAND and Geochemical model
KamLAND result is consistent with prediction of BSE model.
Fully-Radiogenic (44TW) is within 99%C.L., but out of 1σ.
99%C.L. limit is corresponding to 60 TW.

68. Spectrum Shape Analysis
number of events ： (corresponding to TNU)
99% C.L. upper limit ： 70.7 events (corresponding to 145 TNU)
No sensitivity for U/Th ratio
＋15.6
－14.6
＋32.0
－30.0

69. Extrinsic Gammas Screening
Current KamLAND Rate
MC of extrinsic gammas (40K, 208Tl)
7Be ν: ~1μHz
40K: < 3.4μHz
208Tl: < 5.6μHz

70. Solar Neutrino Prospects
7Be neutrinos will be seen between 14C and 11C background
11C can be reduced with neutron tagging (pep and CNO neutrinos extractable???)
11C