search for nucleon decay in super-kamiokande m.miura kamioka observatory, icrr icrr seminar icrr...
TRANSCRIPT
Search for Nucleon DecaySearch for Nucleon Decay in Super-Kamiokande in Super-Kamiokande
M.MiuraM.Miura
Kamioka Observatory, ICRRKamioka Observatory, ICRR
ICRR SeminarICRR Seminar
Nov 26, 2014 in Kashiwa campusNov 26, 2014 in Kashiwa campus
1
ContentsContents1)1) IntroductionIntroduction
2)2) Super-Kamiokande detectorSuper-Kamiokande detector
3)3) p p e e++++00 mode mode
4)4) p p K K++++ mode mode
5)5) pp ++++, n, n ++00
6)6) Other modesOther modes
7)7) Summary of searchesSummary of searches
8)8) Future prospectsFuture prospects
2
1. Introduction
3
The Standard Model
• In July 2012, LHC reported they found Higgs (-like) particle.• It was a missing piece of the Standard Model.• Experimental data of particle physics can be well explained by the Standard Model.• End of explore for particle physics ?
Wait, Wait, Wait, Wait
• There are several important questions which There are several important questions which the Standard Model has not been answered. the Standard Model has not been answered. Why quarks and leptons ?Why quarks and leptons ? Why three generations ?Why three generations ? Why three gauge interactions ?Why three gauge interactions ? Why mass differences ?Why mass differences ? What is the origin of charge ?What is the origin of charge ? ……… ………..
4
Hint of beyond the Standard Model
5
Running Coupling Constants
log E(GeV)
Inve
rse
of c
oupl
ing
Electromagnetic
Weak
Strong (o
nly quark)
• Strength of Electromagnetic, Weak, and strong interaction are function of energy.• By extrapolating to high energy, they seems to meet around ~1015 GeV.• Merging three interactions Grand Unify Theory.• Obviously, this energy is out of range by accelerators.
Not conserve baryon and lepton number: GUTs allowed Proton decay
Examples of proton decay
Model Mode Prediction (years)
Minimal SU(5) pe+0 1028.5 ~ 1031.5 [1]
Minimal SO(10) pe+0 1030 ~ 1040 [2]
Minimal SUSY SU(5) pK+ ≤ 1030 [3]
SUGRA SU(5) pK+ 1032 ~ 1034 [4]
SUSY SO(10) pK+ 1032~1034 [5]
[1] P. Langacker, Phys. Reports 72, 185 (1981)[2] D.G. Lee, M.K. Parida, and M. Rani, Phys. Rev. D51, 229 (1995)[3] H.Murayama and A. Pierce, Phys. Rev. D65, 55009 (2002)[4] T. Goto and T. Nihei, Phys. Rev. D59, 115009 (1999)[5] V. Lucas and S. Ruby, Phys. Rev. D55, 6986 (1997)
Proton lifetime predictions
u
dp
3
~
c~
W
sc
c
~
u uK+
Minimal SU(5) model SUSY SU(5) model
d
u
u e+
d
dc
X
p0
> 1030 years !
How measure proton lifetime?
• Impossible to keep watching a proton for > 10Impossible to keep watching a proton for > 103030 years: Age years: Age of universe~ 1.4x10of universe~ 1.4x101010 years years
• Definition of lifetime Definition of lifetime : N = N: N = N00 exp(-t/ exp(-t/))
• If initial number of proton, NIf initial number of proton, N00 is large, we may observe is large, we may observe
proton decay.proton decay.
• Need massive detectorNeed massive detector, same as neutrino observation., same as neutrino observation. Kamiokande (3kt), IMB (8kt): water cherenkov detectorKamiokande (3kt), IMB (8kt): water cherenkov detector Soudan (963t), NUSEX (150t), Frejus (900t): Iron calorimeterSoudan (963t), NUSEX (150t), Frejus (900t): Iron calorimeter
• Super-Kamiokande (SK): 7.5x10Super-Kamiokande (SK): 7.5x103333 protons in fiducial protons in fiducial volume.volume.
• SK is the most sensitive detector for nucleon SK is the most sensitive detector for nucleon decay search!decay search!
7
2. Super-Kamiokande Detector2. Super-Kamiokande DetectorLocation: Kamioka mine, Japan. ~1000 m under ground.Size: 39 m (diameter) x 42 m (height), 50kton water. Optically separated into inner detector (ID) and outer detector (OD, ~2.5 m layer from tank wall.)Photo device: 20 inch PMT (ID), 8 inch PMT (OD, veto cosmic rays).Mom. resolution: 3.0 % for e 1 GeV/c (4.1%: SK-2).Particle ID: Separate into EM shower type (e-like) and muon type (-like) by Cherenkov ring angle and ring pattern.
-like () e-like (e
Nucleon Decay Experiment
8
K+ paper 2005: > 2.3x1033 years (92kton ・ year)
SK-3Live time: 518 daysExposure: 31.9kton ・ yrID PMT:11146Photo coverage: 40 %
SK-2Livetime:- 798.6 daysExposure:- 49.2 kton·yrInner PMT:5182Photo coverage:19 %
96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13
Accident reconstruction electronics upgrade
SK-4 (~13’Feb)Live Time: 1417 daysExposure: 87.3kton ・ yr
History of Super-Kamiokande
e+0 paper 2007: > 8.2x1033 years (141kton ・year)
This talk: 260 kton ・ year exposure in total
Same
SK-1Livetime: - 1489.2 daysExposure:- 91.7 kton·yrInner PMT: 11146Photo coverage:40 %
9
How to Search 1
Proton decay MC Efficiency, Selection criteria 8 bounded protons in O:
Fermi momentum, binding energy, various nuclear effects are taken into account.
2 free protons: simple two body decay.Atmospheric MC BKG for proton decay Flux : Primary cosmic rays make and e.
M.Honda et .al., Phys.Rev. D75 043006(2007)
interaction: NEUT Y.Hayato, Nucl.Phys.Proc.Suppl. 112,171(2002)
• Analysis method: two typesAnalysis method: two types 1) Select events by cuts (less BKG)1) Select events by cuts (less BKG)
Selection efficiency Selection efficiency Estimated by signal MC. Estimated by signal MC. Number of background Number of background Estimated by BKG MC. Estimated by BKG MC. Criteria are tuned as Sig/Criteria are tuned as Sig/√√BKG becomes max.BKG becomes max. Check agreement between data and BKG in sideband region.Check agreement between data and BKG in sideband region. Open data and calculate lower limit of lifetime.Open data and calculate lower limit of lifetime.
2) Fit data distribution by Signal and BKG MC and 2) Fit data distribution by Signal and BKG MC and estimate lower limit of lifetime.estimate lower limit of lifetime.
11
How to Search 2
Data BKG Sig= x
+ x
3. p3. pee++00 mode mode
0
P
e+
Event features;• e+ and 0 are back-to-back (459 MeV/c) •0 2 s : all particles can be detectable. Reconstruct proton mass and momentum.
Selection;• Fully contained, VTX in fiducial volume.• 2 or 3 ring • all e-like, w/o decay-e• 85 < M0 < 185 MeV (for 3-ring event) .• 800 < MP < 1050 MeV & Ptot < 250 MeV/c
12
e+ 0
Signal MC
13
What causes inefficiency ?
Signal MC
Signal box
• If a proton in oxygen decays, 0 is suffered from nuclear interactions.
Absorption: 1 ring Scatter: Change invariance mass and total momentum. Charge exchange: Change Nring and PID.
• Also nuclear interactions are dominant source of systematic error. •Free protons (H) are free from them (eff.~90%). High efficiency for water cherenkov detector.
Results of pe+0
14
Blue: Signal MC Green: BG MC(ATM Red: Data
Total invariant mass (MeV/c2)
Tot
al m
omen
tum
(M
eV/c
)Total mass vs Total momentum
Total mass (1D)
Dot: DataBlue: Signal MCGreen: BKG MC
• No data found.• Data agrees with BKG (Atm.) MC.
15
Exp.(kt ・yr)
Eff(%)
BKG
b
Data
n
SK1 91.7 39.2±0.7 0.27 0
SK2 49.2 38.5±0.7 0.15 0
SK3 31.9 40.1±0.7 0.07 0
SK4 87.3 39.5±0.7 0.22 0
Total 260.1 0.71 0
Lifetime limit (90%C.L,): > 1.4x1034 years
Error for eff. (%)
-N int. 15
Frmi-mom 8
Correlated decay
7
Other 4
Total 19
Error for BKG (%)
-N int. 8
p-int. water 36
-flux, crs 11
Other 22
Total 44
dbddbPPPPn
be
AnP
nb
)()()()(!
)(1)|(
)(
it
dnPLClim
0)|(..
Life time limit can be calculated based on Bayes’ therem:
P(x): probability function for x: Exposure: Efficiencyb: Expected BKG
: decay rate, lim = 1/lim
• Regarded as a dominant proton decay mode in SUSY-GUTs.Regarded as a dominant proton decay mode in SUSY-GUTs.
• Momentum of KMomentum of K++ is is below Cherenkov thresholdbelow Cherenkov threshold, stop in water , stop in water and decay and decay use decay products from Kuse decay products from K++:: KK++ ++ (63 %) (63 %) KK++ ++ 00 (23 %) (23 %) have monochromatic momentumhave monochromatic momentum..
• Nuclear Nuclear (6 MeV) is emitted from remained nucleus when (6 MeV) is emitted from remained nucleus when deexcitation (41 % in probability).deexcitation (41 % in probability).
• Following three methods are used to search for Following three methods are used to search for pp + K+ K++ . .
• Published to Phys. Rev. D 90, 072005 (2014),Published to Phys. Rev. D 90, 072005 (2014),
16
4. p4. p + K+ K++ mode mode
17
… with Synopsis ( selected by editors).
e+e
16O->15N
6.3MeV
K+
+
Method (A) K+ ++ with nuclear
t
T(dN/dt=max)
Tstart
12ns window
e
Hits
Event features;• K+ is invisible, stops and 2 body decay (P = 236 MeV/c).• Proton in 16O decays and excited nucleus emits 6 MeV (Prob. 41%, not clear ring).
=> Tag to eliminate BKG.
Selection:• 1 -like ring with decay-e.• 215 < P < 260 MeV/c• Proton ring rejection.• Search Max hit cluster by sliding time window (12ns width); - 8 < N < 60 hits for SK-1,3,4 4 < N < 30 hits for SK-2 - T-T < 75 nsec
visibleinvisible
18
19
What causes inefficiency ?
• hit cluster is much smaller than .• hits spread due to PMT time resolution (2.2 nsec @ 1pe). • To avoid tail, time window starts to slide from dN/dt <0 (t0).• As a result, tagging efficiency is low if is close to .
t-t(nsec)
Results of Method (A) Black: DataRed: ATM MCBlue: PDK MC
No candidates have been observed.
Exp.(kton ・ yr)
Eff(%) BKG Data
SK1 91.7 7.9±0.1 0.08 0
SK2 49.2 6.3±0.1 0.14 0
SK3 31.9 7.7±0.1 0.03 0
SK4 87.3 9.1±0.1 0.13 0
Total 260.1 0.38 0
N distributions
20
40% coverage
19% coverage
NOTE• PMT density was half in SK2.• Decay electron tagging has been improved in SK4 by new electronics.
21
Black: DataRed: ATM MCBlue: PDK MC
P
Method (B) K+ ++, spectrum
• Proton decay signal should have monochromatic momentum.• Make momentum distributions of rejected events by cut.• Data agrees with atmospheric MC and cannot see any bump in signal region.• Fitting data by signal and BKG MC to estimate lower limit of lifetime limit with other two method (see later page).
Method (C) K+ ++
K+
+
e+
0
+
e
205 MeV/c
Event features;• Br. 21 %.•and are back-to-back and have 205 MeV/c.•P+ is just above Č thres.(not clear ring).
=> Search for monochromatic 0 with backward activities.
Selection:• 1 or 2 e-like rings with decay-e.• 85 < M0 < 185 MeV.• 175 < P0 < 250 MeV/c.• Ebk: visible energy sum in 140-180 deg. of 0 dir, Eres: in 90-140 deg, Lshape: Likelihood based on charge profile 10 < Ebk < 50 MeV Eres < 12 MeV (20 MeV for 1ring) Lshape > 2.0 (3.0 for 1ring)
Eres
Ebk
visibleinvisible
22
Blue: SignalRed: BKG
Improvement: pImprovement: pKK++,K,K++++00
# of Ring: K++0
Judge as 1 ring (~20%) if opening angle of 2 s is small or momentum of one is small.
Use “0 fitter”• Make likelihood assuming 0 and search for missing ring.• It is used for e appearance analysis of T2K to reduce BKG.
It makes 1 ring sample available It makes 1 ring sample available for this analysis!for this analysis! Efficiency increased.Efficiency increased.
M0
P0
23
1 2 3 4
0 50 100 150 200 250 300
0 50 100 150 200 250 300 350 400
No candidates.
Exp.(kton ・yr)
Eff(%) BKG Data
SK1 91.7 7.8±0.1 0.18 0
SK2 49.2 6.7±0.1 0.17 0
SK3 31.9 7.9±0.1 0.09 0
SK4 87.3 10.0±0.1 0.18 0
Total 260.1 0.62 0
Results of Method (C)
p K+ Lifetime limit (90% CL) combining Method (A) ,(B), and (C): > 5.9 x1033 yrs
-
24
Black: DataRed: ATM MCBlue: PDK MC
Ebk (MeV)
Lshape
• Published to Phys.Rev.Lett. 113, 121802 (2014).Published to Phys.Rev.Lett. 113, 121802 (2014).• Predicted by minimal SUSY SO(10) model;Predicted by minimal SUSY SO(10) model;
(n (n ++00 )=2 )=2(p (p ++++) ) 5.7 ~ 13x10 5.7 ~ 13x103232 years years
25
4. p4. p + + ++, n, n + + 00,,
0
n
+
p
n +0 p ++
Two e-like rings one -like ring
P0 ~ 460 MeV/c P+ ~ 458.8 MeV/c
Affected by nuclear interaction, same as pe+0 case
Large background Fit momentum by Signal and BKG MC
26
n +0 p ++
Data: SK1 ~ SK3, 173 kt ・ year exposure
(n +0 ) > 1.1 x1033 years (p ++) > 3.9x1032 years
Almost ruled out the predicted range.Almost ruled out the predicted range.
5. New modes: |5. New modes: |(B-L)| =2(B-L)| =2
27
★ Decay to trilepton p e+, + Phys.Rev.Lett. 113,101801 (2014) Predicted by SO(10) (Pati and Salam): 1030~1033 years [6] Single charged lepton in final state check excess in P distribution.
Pe P
No significant excesses. : Lifetime limit (90%CL) pe+: > 1.7x1032 years p+: > 2.2x1032 years (IMB-3:1.7x1031 years) (Frejus:2.1x1031 years)
Cross: dataHist: Atm MC : Signal MC
Cross: dataHist: Atm MC : Signal MC
[6] J.C. Pati, Phys.Rev.D29,1549 (1984), P.H. Gu and U. Sarkar (2011), 1110.4581
28
★ Dinucleon decay: pp++, pn+0, nn 00
Search for -pair in back-to back. For nn00, use total mass and momentum cuts. For other modes, difficult to reconstruct total mass and momentum. Use multi-variable analysis (Boosted Decision Tree).
Eff(%) BG(evts) Obs(evts) Limit(yr) (Frejus)
pp++ 5.9 4.5 2 >7.3x1031 7.0 X 1029
pn+0 10.2 0.8 1 >1.8x1032 2.0 X 1030
nn 00 21.1 0.1 0 >4.1x1032 3.4 X 1030
New modes: New modes: B =2B =2
Total invariant mass Total invariant mass
Tot
al m
omen
tum
Tot
al m
omen
tum
Signal MC
BG MC
Data
Consistent with expected backgrounds.
Signal region
6. Summary of searches6. Summary of searches• We have We have not fond any evidencesnot fond any evidences
of nucleon decay, so far.of nucleon decay, so far.• We calculated nucleon lifetime We calculated nucleon lifetime
limits with 90 % C.L.limits with 90 % C.L.
pp e e++00: > 1.4x10: > 1.4x103434 yrs yrs
KK++: > 5.9x10: > 5.9x103333 yrsyrs
• Other new modesOther new modes have also been have also been studied and published in this studied and published in this year.year.
• Those are the most stringent Those are the most stringent limit in the world.limit in the world.
29
nn 0 0pn + 0
pp + +
pp K +K +
n 0
n + -
n e + -
n + -
n e + -
p +
p e +
p +K 0
p K +
p +
p +
p e +
p + 0
p e + 0
p + 0
p e + 0p + 0
p e + 0
1031
1032
1033
1034
1035
/B (years)
30
7. Future prospect7. Future prospectOn going project: neutron tagging
pe+0: ATM MC• Backgrounds of proton decay are atmospheric interactions.• Sometimes they are accompanied with neutron.• Fow example, dominant background of pe+0 is Chargted Current 1 production;ep e+n0
en e-n+ +charge exchange• neutron tag is useful to reduce background.
31
n+p d+ (2.2 MeV), =204s (enable for SK4)• Search for hit cluster N>7 in 10 ns window after prompt signal.• Eff. 25 %, BKG 1.4 %.• MC including neutron capture is developing.
How to tag neutron
Agree well!
Preliminaryneutron after all pe+0 cut (SK4)
Almost 50 % BKG can be rejected!
Super-K+Gd option is under study, also.
• Longer exposure = increase expected background. Longer exposure = increase expected background. Any room to improve cut criteria?Any room to improve cut criteria?
• Improvements of systematic errors contribute better Improvements of systematic errors contribute better sensitivities.sensitivities.
• Not all searches use data including SK1-SK4. Should Not all searches use data including SK1-SK4. Should be updated.be updated.
• Any other good decay modes ? Feedback from Any other good decay modes ? Feedback from theorists are welcome!theorists are welcome!
• Model independent analysis ? Look for discrepancies Model independent analysis ? Look for discrepancies between data and atmospheric between data and atmospheric MC. MC.
32
Other items to be studied
33
Further more: Hyper-Kamiokande
1 Mega ton Water Chrenkov Detector
Base design
• Total Volume: 0.99 Mton• Inner Volume: 0.74 Mton• Fiducial volume: 0.56 Mton
(0.056 Mton x 10 compartments)
• Outer volume: 0.2 Mton• Photo-sensors:
99,000 20 inch PMT (ID)(20 % photo coverage)
25,000 8 inch PMT (OD)
34
Hyper-K for Nucleon Decay • Sensitivity (90%CL) with HK
10 years run;
e+0: 1.3x1035 year
K+: 2.5x1034 year
• 33 discovery potential discovery potential
e+0: 5.6x10 5.6x10 3434 years years
K+: 1.2x101.2x103434 year year
(with base design)(with base design)
pe+0
pK+
--
Letter of Intent: Hyper-Kaiokande Experiment arXiv:1109.3262 [hep-ex] 35
-
Photo sensor R&D for Hyper-K
20 inch PMT in baseline design- It is almost hand-made even in electrics parts.- Expensive and not suitable for mass production.- But we need 99,000 PMTs.
Other possibility: Hybrid Photo Detector (HPD), High QE Other possibility: Hybrid Photo Detector (HPD), High QE PMT (box&line)PMT (box&line)
e
DynodeDynode
~2kV
e
~8kV
Avalanche Diode (AD)Avalanche Diode (AD)
e
Avalanche Gain ~100@250V
Bombardment Gain ~400 @8kV
PMT HPD36
Advantage of HPD• Avalanche Diode is suitable for mass production.
lower cost.• Better photon counting.• Better collection efficiency .
(20 inch PMT~75 %HPD ~95%)• Better timing resolution
(20 inch PMT ~2.2 nsec HPD ~1.1 ns in TTS)
Expect to improve nuclear tagging efficiency in pK+
90ns8mV1p.e.
2p.e.3p.e.
Pulse Height [mV]
1p.e.2p.e.
3p.e.4p.e.
ped
t
Hits
Hidden by tail
37t
Hits
Better separation
Better t-resolution
38
Test: HPD time resolution with SK4 setup Efficiency curve of tagging
T –T(nsec)
Black: 2.2 nsecRed: 1.1 nsec @1pe
• tagging efficiency close to is increased. K+ decay ~exp(-12ns/t).• Selection efficiency of pK 9.1 % 13.1 % (44% increased)• Background rate: 1.5 1.0 evts/Mt ・ yr (33% decreased) Even though half density of sensor, performance of HK with HPD may be better than SK !
New photo sensor can improve physics sensitivity.
39
Hyper-K open meeting will be held on Jan.29-31, 2015 at IPMU.
Your participations are welcome !http://indico.ipmu.jp/indico/conferenceDisplay.py?confId=52
Backup
40
41
Why efficiencies decreased (45 % 40 %) Because we have updated -interaction in Oxygen.
(1) Charge exchange increased to match experimental data.
(2) Kinematics of scattered changed in P > 500 MeV/c
Final momentum Final angle (degree)
Black: old MCRed: new MC
Increased in large angle
Input: 550 MeV/c
NOTE: Sys.error for -int 15 %
42
Recent improvements on BKG rejection
Fix momentum bias for events with decay-e • Hits in (-50:250nsec) used for momentum.• If decay-e is closer to , P is overestimated generate hit template for VTX fit with larger P VTX is shifted to forward. TOF over-subtracted for backward hits, make mimic signal.• Solution: Time window size is shortened when decay-e is closer
to parent . BKG is reduced by 30 %.
43
Recent improvements on BKG rejection (2)
Proton ID• Dominant BKG is CCQE with high mom. recoiled proton and
invisible , the proton is identified as -ring. VTX is shifted to forward to adjust C-ang. TOF over-subtracted for backward hits, make mimic signal.• Proton and muon ring can be separated using likelihood
function based on opening angle and hit pattern.
Proton
True VTX
Proton
260 <P<400 MeV
Black: DataRed: ATM MC
Agree well.
Fitted VTX
BKG is reduced to ~50 %.
Recent improvement(2): pRecent improvement(2): pKK++,K,K++++00
Charge distribution in angle
Opposite of o dir= + dir
0 dir
0+ zoom
Red: PDK MCBlue: BKG
Conventional method: Use corrected charge sum in <40°EbkNew method: 1. Define Ebk by charge sum in < 35° 2. Make expected charge distribution for signal and BKG, and compare observed charge by likelihood. Compare shape of dist.
44
0 10 20 30 40 50 60
),(Prlog(),(Prlog( exp
40
0exp
40
0
BKGobs
sigobsshape qqobqqobL
45
Input variables: ppInput variables: pp++++
nring2 2.5 3 3.5 4 4.5 5 5.5 6
Even
ts
0
5
10
15
20
25
30
35
signal
background
data
number of rings
nmue0 0.5 1 1.5 2 2.5 3 3.5 4
Even
ts
0
2
4
6
8
10
12
14
16
18
20
22
24
number of decay electrons
number of pi candidates2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4
Even
ts
0
10
20
30
40
50
number of non-showering rings
energy (MeV)0 200 400 600 800 1000 1200 1400 1600
Even
ts
0
2
4
6
8
10
visible energy
rtot (corrected p.e.)0 500 1000 1500 2000 2500 3000
Even
ts
0
1
2
3
4
5
6
7
8
magnitude of rtot vector sum
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Even
ts
0
2
4
6
8
10
rtot[0]/rtsum
distance (cm)0 200 400 600 800 1000 1200 1400 1600 1800 2000
Even
ts
0
2
4
6
8
10
12
max decay-e distance
angle (degrees)0 20 40 60 80 100 120 140 160 180
Even
ts
0
2
4
6
8
10
pion decay-e opening angle
angle (degrees)120 130 140 150 160 170 180
Even
ts
0
1
2
3
4
5
6
7
8
maximum opening angle
Number of rings Number of decay-e Number of -like
Visible energy Rtot vector sum Fraction of dominant ring
Max decay-e length -decay-e ang Max opening ang
46
Input variables: pnInput variables: pn++00
momentum (MeV/c)0 200 400 600 800 1000 1200 1400
Even
ts
0
5
10
15
20
25
30
35 signal
background
data
pi0 momentum
angle (degrees)120 130 140 150 160 170 180
Even
ts
0
2
4
6
8
10
12
14
16
18
pi0-pi+ opening angle
momentum (MeV/c)0 100 200 300 400 500 600 700 800 900 1000
Even
ts
0
5
10
15
20
25
30
35
40
45
pi+ momentum
energy (MeV)0 200 400 600 800 1000 1200 1400 1600
Even
ts
0
2
4
6
8
10
12
14
visible energy
charge ratio0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Even
ts
0
2
4
6
8
10
12
14
16
18
20
22
rtot[0]/rtsum
mass (MeV/c^2)0 50 100 150 200 250 300 350 400 450 500
Even
ts
0
5
10
15
20
25
30
35
40
45
pi0 mass
nmue0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Even
ts
0
20
40
60
80
100
120
number of decay electrons
0 momentum 0-+ opening ang. + momentum
Visible energy Fraction of dominant ring 0 mass
Number of decay-e
47
Outputs of BDT (SK4 only) Outputs of BDT (SK4 only)
MVA-1 -0.5 0 0.5 1
Even
ts
0
2
4
6
8
10
12
14
16
MVA output
signalbackgrounddata
MVA-0.6 -0.4 -0.2 0 0.2 0.4 0.6
Even
ts
0
10
20
30
40
50
60
MVA output
signalbackground
data
pppp++++ pnpn++00