Previously, on Neutrino Physics… •  explained propagation through matter using optical theorem

phase shift of the νe wavefunction •  in 2-flavour simplified: diagonalized the “Hamiltonian” to get the

mass squared eigenvalues (in the matter mass basis) •  determined the rotation transformation from flavour basis to

matter mass basis •  solved for θm in terms of θ, Δm2, E, GF, Ne

•  examined resonance condition •  noted matter effects change sign for antineutrinos •  plotted matter effect versus electron density; discussed solar

neutrino evolution through smoothly changing density •  considered how to think “numerically” about neutrinos

propagating through matter

Previously…cont’d •  looked at atmospheric neutrino oscillations in depth:

history and evidence for, glance at where we are now •  surveyed solar neutrino experiments up to and including

SNO •  including source calibration anomaly in gallium •  Including new Super-K solar neutrino day-night effect

•  ran out of time to talk about SNO spectrum or global solar neutrino oscillation analysis (historical possible solutions LMA, SMA, VAC)

•  ran out of time to talk about Borexino

Fast Forward: All we Know about Neutrino Oscillations

• mass splittings • mixing angles

What remains to be measured? • mass hierarchy • CP phase(s)

Neutrino Mass Hierarchy

ν3

Δm atmospheric ~50 meV

�normal�

�inverted�

Δmsolar ≈ 9 meVΔmatm ≈ 50 meV

Mass Splittings – Current Understanding ! neutrino oscillation experiments have measured the mass splitting

between the three neutrino mass states

!  solar neutrino oscillations (MSW matter effect) depend on the sign of ; thus we know the mass of ν2 is greater than ν1

!  atmospheric neutrino and long baseline neutrino beam experiments do not (yet) determine the sign of

Δm122

Δm232

13. Neutrino mixing 45

Table 13.7: The best-fit values and 3σ allowed ranges of the 3-neutrino oscillationparameters, derived from a global fit of the current neutrino oscillation data,including the T2K and MINOS (but not the Daya Bay and RENO) results(from [140]) . The PDG average of the results of the three recent reactorexperiments [26,27,59] is given in the last line [142]. The values (values inbrackets) of sin2 θ12 and sin2 θ13 are obtained using the “old” [35] (“new” [34])reactor ν̄e fluxes in the analysis.

Parameter best-fit (±1σ) 3σ

∆m2⊙ [10−5 eV 2] 7.58+0.22

−0.26 6.99 − 8.18

|∆m2A| [10−3 eV 2] 2.35+0.12

−0.09 2.06 − 2.67

sin2 θ12 0.306 (0.312)+0.018−0.015 0.259 (0.265)− 0.359 (0.364)

sin2 θ23 0.42+0.08−0.03 0.34 − 0.64

sin2 θ13 [140] 0.021 (0.025)+0.007−0.008 0.001 (0.005)− 0.044 (0.050)

sin2 θ13 [142] 0.0251 ± 0.0034 0.015 − 0.036

A combined analysis of the data on θ13 from the T2K, MINOS, Double Chooz, DayaBay and RENO experiments was performed in Ref. 145. The authors find that θ13 ̸= 0at 7.7σ:

sin2 2θ13 = 0.096 ± 0.013 (±0.040) at 1σ (3σ) (13.80)

In this analysis the positive or negative sign of ∆m2A was used as input and the values

of sin2 2θ23 and |∆m2A| were varied imposing Gaussian priors based on the results of the

atmospheric neutrino [139] and MINOS [130] experiments. The value of sin2 2θ13 thusobtained showed a statistically insignificant dependence on the Dirac phase δ.

It follows from the results given in Table 13.7 that θ23∼= π/4, θ12

∼= π/5.4 and thatθ13

∼= π/20. Correspondingly, the pattern of neutrino mixing is drastically different fromthe pattern of quark mixing.

Note also that ∆m221, sin2 θ12, |∆m2

31(32)|, sin2 θ23 and sin2 θ13 are determined from

the data with a 1σ uncertainty (= 1/6 of the 3σ range) of approximately 2.6%, 5.4%,4.3%, 12% and absolute error 0.42 × 10−2, respectively.

The existing SK atmospheric neutrino, K2K and MINOS data do not allow todetermine the sign of ∆m2

31(32). Maximal solar neutrino mixing, i.e., θ12 = π/4, is ruled

out at more than 6σ by the data. Correspondingly, one has cos 2θ12 ≥ 0.27 (at 99.73%CL).

At present no experimental information on the Dirac and Majorana CP violationphases in the neutrino mixing matrix is available. Thus, the status of CP symmetry inthe lepton sector is unknown. With θ13 ̸= 0, the Dirac phase δ can generate CP violationeffects in neutrino oscillations [40,52,53]. The magnitude of CP violation in νl → νl′ and

June 18, 2012 16:19

Δm122 =

Δm232 =

×10−5 eV2

×10−3 eV2

KamLAND (and solar)

MINOS, T2K (and SK)

Neutrino Oscillation Analysis Global Solar plus KamLAND

KamLAND Discovery K. Eguchi et al., Phys. Rev. Lett. 90, 021802 (2003)

KamLAND observes disappearance of reactor antineutrinos! (61 ± 9)% expected flux announced Dec. 6, 2002 reactor oscillations consistent with solar ν$LMA solution$

νe

νe + p→ e+ + n

KamLAND Discovery K. Eguchi et al., Phys. Rev. Lett. 90, 021802 (2003)

KamLAND observes disappearance of reactor antineutrinos! (61 ± 9)% expected flux announced Dec. 6, 2002 reactor oscillations consistent with solar ν$LMA solution$

νe

νe + p→ e+ + n

#*���'�%'$)$#��'$((����) $#�"  ���� #+�'(����)������-��

# ��!�-����$ #� ��#���� () #�* (��(�( �#�!�"  � '()�#�*)' #$���)��) $#��-��� #�(��#���$,�#��� ��

*(���)� (�'���) $#�"  #�*)' #$�)�'�(�$!������������"#�/�"%���"���" σ��������.��0����%������"���&*��'�) ��, )���#�'�-�

"  �)��ν��� ������σ����.��0���"�

"  �'$((�(��) $#�'�!�)���)$��'���#�*)'$#�����-�! ��) "���(-()�"�) ��*#��')� #)-� (�("�!!�

�

νe + p→ e+ + n

Location ! where the old

Kamiokande detector was

! 2,700 mwe overburden

! 70 GWth (7% world total nuclear power) generated at about 175 km distance

! total effective power at 180 km: ~82.5 GWth

KamLAND Detector 80% dodecane 20% pseudocumene (1,2,4-trimethylbenzene) 1.52 g/L PPO (2,5-diphenyloxazole) 8,000 photons/MeV 10 m @ 400 nm r = 0.78 g/cm3

buffer oil 50% dodecane 50% isoparaffin

density matched 1.0004 (balloon is heavier) 1325 17-inch PMT’s 22% coverage 320 photoelectrons/MeV

Kunio Inoue NOON2003 Talk

position resolution: σx = 25 cm @ 1 MeV

KamLAND Source Systematics

! reactor power: 2% ! burn-up: 1.0% ! Korean reactors: 0.25% ! other reactors: 0.35% ! neutrino spectra: 2.46% ! cross section: 0.2% ! time lag of beta decay: 0.28%

total neutrino source systematic: 3.4%

first result uncertainties to illustrate reactor neutrino systematics

KamLAND Detector Systematics

! liquid scintillator mass: 2.13% ! fiducial volume / total: 4.06% ! energy threshold: 2.13% ! delayed tag efficiency: 2.06% ! livetime: 0.07% ! source systematics: 3.4%

total rate systematic uncertainty: 6.4%

first result uncertainties to illustrate reactor neutrino systematics

KamLAND and Other Reactor Expts

original figure KamLAND circa 2003

14

PMNS Neutrino Mixing Matrix

atmospheric solar plus KamLAND reactor/T2K Majorana phases

ν f = U fi ν ii∑

Pontecorvo, Maki, Nakagawa, Sakata

Neutrino Flavour Mixing

ν f = U fi ν ii∑

atmospheric solar plus KamLANDreactor/T2K

Majorana phases

0.407 ≤ sin2θ23 ≤ 0.583 sin2θ12 = 0.308 ± 0.014sin2θ13 = 0.025−0.015

+0.018

sin2θ13 = 0.025 ± 0.004New!

from Daya Bay last monthand RENO a few days ago

Abstract

The Tokai to Kamioka (T2K) experiment is a long-baseline neutrino oscillation ex-

periment based in Japan. An o↵-axis, high purity muon neutrino (⌫µ) beam is produced

at the Japan Proton Accelerator Research Complex by colliding a proton beam with a

graphite target. The neutrinos are detected at a near detector complex (ND280), situated

280 m from the neutrino production target, and the Super-Kamiokande (SK) far detector

at 295 km. This thesis first outlines current knowledge and challenges in the physics of

neutrinos, then describes the T2K beam and detectors, including a novel optical transi-

tion radiation monitor for precisely measuring the proton beam in order to determine the

neutrino beam direction. A framework for evaluating the uncertainties in neutrino inter-

actions and pion hadronic interactions in ND280 and SK is presented. A new SK event

reconstruction algorithm is described and the SK detector systematic errors are evalu-

ated based on atmospheric neutrino and cosmic ray muon data. These developments are

used in a Markov Chain Monte Carlo neutrino oscillation analysis of the T2K Run 1´ 4

data corresponding to 0.657 ˆ 1021 protons on target. The analysis simultaneously con-

siders the 2013 ND280 ⌫µ samples, and SK single muon and single electron samples,

producing a measurement of ⌫µ disappearance and ⌫µ Ñ ⌫e appearance. The estimated

oscillation parameters and 68% Bayesian credible intervals (CI) at �CP “ 0 are as follows:

sin2 ✓23 “ 0.520`0.045´0.050, sin

2 ✓13 “ 0.0454`0.011´0.014 and �m2

32 “ ´2.57 ˘ 0.11, where the nega-

tive sign indicates a preference for the inverted neutrino mass hierarchy (IH) with 55%

probability compared to the normal hierarchy (NH). Recent measurements of ✓13 from

reactor neutrino experiments are combined with the T2K data resulting in the following

estimates: sin2 ✓23 “ 0.528`0.055´0.038, sin

2 ✓13 “ 0.0250 ˘ 0.0026 and �m232 “ 2.51 ˘ 0.11,

where the NH is preferred with 68% probability compared to the IH. Furthermore, the

data hint toward �CP « ´⇡{2 with a 90% CI excluding values between 0.14⇡–0.87⇡.

ii

SK atmospheric + T2K Daya Bay 2014 SNO 3-Phase Combined sin2θ13 = 0.0215 ± 0.0015

Fast Forward: All we Know about Neutrino Oscillations

• mass splittings • mixing angles

What remains to be measured? • mass hierarchy • CP phase(s)

•  but, before we do…let’s look at evidence for neutrino oscillations from LSND

Liquid Scintillator Neutrino Detector VOLUME 77, NUMBER 15 P HY S I CA L REV I EW LE T T ER S 7 OCTOBER 1996

Evidence for nm ! ne Oscillations from the LSND Experimentat the Los Alamos Meson Physics Facility

C. Athanassopoulos,12 L. B. Auerbach,12 R. L. Burman,7 I. Cohen,6 D.O. Caldwell,3 B.D. Dieterle,10 J. B. Donahue,7A.M. Eisner,4 A. Fazely,11 F. J. Federspiel,7 G. T. Garvey,7 M. Gray,3 R.M. Gunasingha,8 R. Imlay,8 K. Johnston,9H. J. Kim,8 W.C. Louis,7 R. Majkic,12 J. Margulies,12 K. McIlhany,1 W. Metcalf,8 G.B. Mills,7 R.A. Reeder,10

V. Sandberg,7 D. Smith,5 I. Stancu,1 W. Strossman,1 R. Tayloe,7 G. J. VanDalen,1 W. Vernon,2,4 N. Wadia,8 J. Waltz,5Y-X. Wang,4 D.H. White,7 D. Works,12 Y. Xiao,12 S. Yellin3

LSND Collaboration1University of California, Riverside, California 92521

2University of California, San Diego, California 920933University of California, Santa Barbara, California 93106

4University of California Intercampus Institute for Research at Particle Accelerators, Stanford, California 943095Embry Riddle Aeronautical University, Prescott, Arizona 86301

6Linfield College, McMinnville, Oregon 971287Los Alamos National Laboratory, Los Alamos, New Mexico 87545

8Louisiana State University, Baton Rouge, Louisiana 708039Louisiana Tech University, Ruston, Louisiana 71272

10University of New Mexico, Albuquerque, New Mexico 8713111Southern University, Baton Rouge, Louisiana 70813

12Temple University, Philadelphia, Pennsylvania 19122(Received 9 May 1996)

A search for nm ! ne oscillations has been conducted at the Los Alamos Meson Physics Facility byusing nm from m1 decay at rest. The ne are detected via the reaction ne p ! e1 n, correlated with ag from np ! dg (2.2 MeV). The use of tight cuts to identify e1 events with correlated g rays yields22 events with e1 energy between 36 and 60 MeV and only 4.6 6 0.6 background events. A fit tothe e1 events between 20 and 60 MeV yields a total excess of 51.0120.2

219.5 6 8.0 events. If attributedto nm ! ne oscillations, this corresponds to an oscillation probability of (0.31 6 0.12 6 0.05)%.[S0031-9007(96)01375-0]

PACS numbers: 14.60.Pq, 13.15.+g

We present the results from a search for neutrino os-cillations using the Liquid Scintillator Neutrino Detector(LSND) apparatus described in Ref. [1]. The existence ofneutrino oscillations would imply that neutrinos have massand that there is mixing among the different flavors of neu-trinos. Candidate events in a search for the transformationnm ! ne from neutrino oscillations with the LSND de-tector have previously been reported [2] for data taken in1993 and 1994. Data taken in 1995 have been included inthis paper, and the analysis has been made more efficient.Protons are accelerated by the Los Alamos Meson

Physics Facility (LAMPF) linac to 800MeV kinetic energyand pass through a series of targets, culminating with theA6 beam stop. The primary neutrino flux comes from p1

produced in a 30-cm-long water target in the A6 beam stop[1]. The total charge delivered to the beam stop whilethe detector recorded data was 1787 C in 1993, 5904 Cin 1994, and 7081 C in 1995. Neutrino fluxes used in ourcalculations include upstream targets and changes in targetconfiguration during these three years of data taking.Most of the p1 come to rest and decay through

the sequence p1 ! m1nm, followed by m1 ! e1nenm,supplying nm with a maximum energy of 52.8 MeV. Theenergy dependence of the nm flux from decay at rest

(DAR) is very well known, and the absolute value isknown to 7% [1,3]. The open space around the targetis short compared to the pion decay length, so only 3% ofthe p1 decay in flight (DIF). A much smaller fraction(approximately 0.001%) of the muons DIF, due to thedifference in lifetimes and that a p1 must first DIF. Thetotal nm flux averaged over the detector volume, includingcontributions from upstream targets and all elements ofthe beam stop, was 7.6 3 10210nmycm2yprotony.A ne component in the beam comes from the sym-

metrical decay chain starting with a p2. This back-ground is suppressed by three factors in this experiment.First, p1 production is about 8 times the p2 produc-tion in the beam stop. Second, 95% of p2 come to restand are absorbed before decay in the beam stop. Third,88% of m2 from p2 DIF are captured from atomic or-bit, a process which does not give a ne. Thus the rela-tive yield, compared to the positive channel, is estimatedto be , s1y8d 3 0.05 3 0.12 ≠ 7.5 3 1024. A detailedMonte Carlo simulation [3] gives a value of 7.8 3 1024

for the flux ratio of ne to nm.The detector is a tank filled with 167 metric tons of dilute

liquid scintillator, located about 30 m from the neutrinosource and surrounded on all sides except the bottom

3082 0031-9007y96y77(15)y3082(4)$10.00 © 1996 The American Physical Society

Liquid Scintillator Neutrino Detector VOLUME 77, NUMBER 15 P HY S I CA L REV I EW LE T T ER S 7 OCTOBER 1996

Evidence for nm ! ne Oscillations from the LSND Experimentat the Los Alamos Meson Physics Facility

C. Athanassopoulos,12 L. B. Auerbach,12 R. L. Burman,7 I. Cohen,6 D.O. Caldwell,3 B.D. Dieterle,10 J. B. Donahue,7A.M. Eisner,4 A. Fazely,11 F. J. Federspiel,7 G. T. Garvey,7 M. Gray,3 R.M. Gunasingha,8 R. Imlay,8 K. Johnston,9H. J. Kim,8 W.C. Louis,7 R. Majkic,12 J. Margulies,12 K. McIlhany,1 W. Metcalf,8 G.B. Mills,7 R.A. Reeder,10

V. Sandberg,7 D. Smith,5 I. Stancu,1 W. Strossman,1 R. Tayloe,7 G. J. VanDalen,1 W. Vernon,2,4 N. Wadia,8 J. Waltz,5Y-X. Wang,4 D.H. White,7 D. Works,12 Y. Xiao,12 S. Yellin3

LSND Collaboration1University of California, Riverside, California 92521

2University of California, San Diego, California 920933University of California, Santa Barbara, California 93106

4University of California Intercampus Institute for Research at Particle Accelerators, Stanford, California 943095Embry Riddle Aeronautical University, Prescott, Arizona 86301

6Linfield College, McMinnville, Oregon 971287Los Alamos National Laboratory, Los Alamos, New Mexico 87545

8Louisiana State University, Baton Rouge, Louisiana 708039Louisiana Tech University, Ruston, Louisiana 71272

10University of New Mexico, Albuquerque, New Mexico 8713111Southern University, Baton Rouge, Louisiana 70813

12Temple University, Philadelphia, Pennsylvania 19122(Received 9 May 1996)

A search for nm ! ne oscillations has been conducted at the Los Alamos Meson Physics Facility byusing nm from m1 decay at rest. The ne are detected via the reaction ne p ! e1 n, correlated with ag from np ! dg (2.2 MeV). The use of tight cuts to identify e1 events with correlated g rays yields22 events with e1 energy between 36 and 60 MeV and only 4.6 6 0.6 background events. A fit tothe e1 events between 20 and 60 MeV yields a total excess of 51.0120.2

219.5 6 8.0 events. If attributedto nm ! ne oscillations, this corresponds to an oscillation probability of (0.31 6 0.12 6 0.05)%.[S0031-9007(96)01375-0]

PACS numbers: 14.60.Pq, 13.15.+g

We present the results from a search for neutrino os-cillations using the Liquid Scintillator Neutrino Detector(LSND) apparatus described in Ref. [1]. The existence ofneutrino oscillations would imply that neutrinos have massand that there is mixing among the different flavors of neu-trinos. Candidate events in a search for the transformationnm ! ne from neutrino oscillations with the LSND de-tector have previously been reported [2] for data taken in1993 and 1994. Data taken in 1995 have been included inthis paper, and the analysis has been made more efficient.Protons are accelerated by the Los Alamos Meson

Physics Facility (LAMPF) linac to 800MeV kinetic energyand pass through a series of targets, culminating with theA6 beam stop. The primary neutrino flux comes from p1

produced in a 30-cm-long water target in the A6 beam stop[1]. The total charge delivered to the beam stop whilethe detector recorded data was 1787 C in 1993, 5904 Cin 1994, and 7081 C in 1995. Neutrino fluxes used in ourcalculations include upstream targets and changes in targetconfiguration during these three years of data taking.Most of the p1 come to rest and decay through

the sequence p1 ! m1nm, followed by m1 ! e1nenm,supplying nm with a maximum energy of 52.8 MeV. Theenergy dependence of the nm flux from decay at rest

(DAR) is very well known, and the absolute value isknown to 7% [1,3]. The open space around the targetis short compared to the pion decay length, so only 3% ofthe p1 decay in flight (DIF). A much smaller fraction(approximately 0.001%) of the muons DIF, due to thedifference in lifetimes and that a p1 must first DIF. Thetotal nm flux averaged over the detector volume, includingcontributions from upstream targets and all elements ofthe beam stop, was 7.6 3 10210nmycm2yprotony.A ne component in the beam comes from the sym-

metrical decay chain starting with a p2. This back-ground is suppressed by three factors in this experiment.First, p1 production is about 8 times the p2 produc-tion in the beam stop. Second, 95% of p2 come to restand are absorbed before decay in the beam stop. Third,88% of m2 from p2 DIF are captured from atomic or-bit, a process which does not give a ne. Thus the rela-tive yield, compared to the positive channel, is estimatedto be , s1y8d 3 0.05 3 0.12 ≠ 7.5 3 1024. A detailedMonte Carlo simulation [3] gives a value of 7.8 3 1024

for the flux ratio of ne to nm.The detector is a tank filled with 167 metric tons of dilute

liquid scintillator, located about 30 m from the neutrinosource and surrounded on all sides except the bottom

3082 0031-9007y96y77(15)y3082(4)$10.00 © 1996 The American Physical Society

Decay-at-Rest Neutrino Beam “simple” to understand, intrinsic beam contamination is orders of magnitude smaller � 7.8 × 10–4 times lower, according to LSND beam Monte Carlo

νe

νµ ,νµ in redνe,νe in blue

DAR Beam Concept Diagram

from H. Ray!

Breakdown of LSND Neutrino Flux units [ν/cm2] TABLE III: Average neutrino fluxes in LSND. Both decay at rest (DAR) and decay in flight (DIF)

are shown in ν/cm2. The νµ and ν̄µ DIF fluxes are above µ production threshold.

Source Type 1993-1995 Flux 1996-1998 Flux Total Flux

µ+ DAR ν̄µ and νe 7.38 × 1013 5.18 × 1013 1.26 × 1014

µ− DAR νµ and ν̄e 5.96 × 1010 4.87 × 1010 1.08 × 1011

π+ DIF νµ 1.37 × 1012 8.26 × 1011 2.20 × 1012

π− DIF ν̄µ 1.45 × 1011 1.11 × 1011 2.56 × 1011

π+ DIF νe 5.56 × 108 5.01 × 108 1.06 × 109

µ+ DIF νe 4.13 × 109 2.44 × 109 6.57 × 109

TABLE IV: Cross section uncertainties for the neutrino reactions with two-body final states that

occur in LSND. The cross sections for these processes are known accurately because either related

measurements can be used to constrain the matrix elements or only fundamental particles are

observed. Also shown are the corresponding neutrino flux constraints.

Process σ Constraint σ Uncertainty Flux Constraint

νe → νe Standard Model Process 1% µ+ → νeν̄µe+ DAR

12C(νe, e−)12Ng.s.12Ng.s. 5% µ+ → νeν̄µe+ DAR

12C(νµ, µ−)12Ng.s.12Ng.s. 5% π+ → νµµ+ DIF

p(ν̄µ, µ+)n neutron decay 5% π− → ν̄µµ− DIF

31

Main Signal •  electron antineutrino appearance from a muon

antineutrino beam, between 20-60 MeV •  energies are low enough you can use nuebar-proton with the

delayed coincidence signal •  observed the delayed 2.2 MeV signal from •  at higher energies, nuebar CCQE interaction may dominate; but

DIS and resonant inelastic starts to become important (MiniBooNE)

SOURCE * TARGET * INTERACTION � SIGNAL beam off to measure backgrounds calculate possible beam-on backgrounds

νe + p→ e+ + n

n + p→ d + γ

Main Backgrounds •  beam-off cosmic ray (positron, neutron)

•  beam is only on 7% of the time so should be able to estimate these

•  beam-on backgrounds with correlated neutrons •  beam-on backgrounds with accidental gammas

•  µ– DAR (intrinsic beam electron antineutrinos) •  π– DIF which decay to muon antineutrinos which makes µ+ and n,

correlated in the detector, (the µ+ is missed before) it decays to e+ in the detector

νe + p→ e+ + n

personal comment: should be harder to fake this signal than for CCQE in MiniBooNE b/c the neutron is correlated, spatially and in time

Evidence for Appearance •  22 events between 36-60 MeV •  backgrounds 4.6 ± 0.6 events

VOLUME 77, NUMBER 15 P HY S I CA L REV I EW LE T T ER S 7 OCTOBER 1996

electrons from muon decay. Selection VI required Dtpgreater than 20 ms, and no activities between 20 and34 ms before the event trigger time with more than50 PMT hits or reconstructed within 2 m from thepositron position. The selection I and VI efficiencies are0.50 6 0.02 and 0.68 6 0.02, respectively. The time toany subsequent triggered event, Dta, is required to be. 8 ms to remove events which are misidentified muonswhich decay (0.99 6 0.01 efficiency). The reconstructedpositron location was required to be a distance D .35 cm from the surface tangent to the faces of the PMTs(0.85 6 0.05 efficiency). This assures that the positronis in a region of the tank in which the energy and PIDresponses vary smoothly and are well understood. The 35cm cut also avoids the region of the tank with the highestcosmic ray background.To suppress cosmic ray neutrons, the number of associ-

ated g with R . 1.5 is required to be no more than two forselection I (0.99 6 0.01 efficiency) and no more than onefor selection VI (0.94 6 0.01 efficiency). Recoil neutronsfrom the nep ! e1n reaction are too low in energy toknock out additional neutrons. The number of veto shieldhits associated with the events is no more than one for se-lection I (0.84 6 0.02 efficiency) and no more than threefor selection VI (0.98 6 0.01 efficiency).Beam-off data surviving these cuts were found to in-

clude cosmic ray events entering the detector tank fromoutside. We have found two new criteria which are effec-tive at reducing this background. One is the distributionof angles between the e1 direction and its position vec-tor relative to the tank center—background events tendto head inwards. The other is in the distribution of vetohits—cosmic ray events tend to have more of them. Thesetwo distributions are used in a way analogous to the R pa-rameter discussed earlier in defining a likelihood ratio, S[5]. For selection VI, but not I, we require S . 0.5, acut that loses 13% of the expected neutrino signal whileeliminating 33% of the beam-off background. Includinga 0.97 6 0.01 data acquisition efficiency gives overall ef-ficiencies of 0.26 6 0.02 for selection I and 0.37 6 0.03for selection VI.The backgrounds to nep ! e1n followed by n cap-

ture fall into three general classes: beam-off events (cos-mic ray induced), beam-related events with correlatedneutrons, and beam-related events with an accidental g.As outlined above, the cosmic ray background to beam-on events is 0.07 times the number of beam-off eventswhich pass the same criteria. The major sources of beam-induced backgrounds are from m2 DAR, discussed above,and from p2 DIF in the beam stop. The latter results ina background from nm p interactions where the final m1

is missed, and its Michel decay positron is mistaken fora primary ne p event. These nm backgrounds are esti-mated using the detector Monte Carlo simulation [1,5].The backgrounds with accidental g overlap are greatlyreduced by selection on the R parameter. Details of allbackgrounds considered are presented in Ref. [5].

Table I lists the number of signal, beam-off-background, and neutrino-background events for thetwo selections with 36 , Ee , 60 MeV—to avoid largeaccidental-g backgrounds. The likelihood ratio R is usedto determine whether a candidate 2.2 MeV g is correlatedwith an electron or from an accidental coincidence.Requiring R . 30 (correlated-g efficiency ≠ 0.23) weobserve 22 events beam-on and 36 3 0.07 ≠ 2.5 eventsbeam-off. The estimated beam-related background con-sists of 1.72 6 0.41 events with correlated neutrons and0.41 6 0.06 without. The probability that the beam-onevents are entirely due to a statistical fluctuation of the4.6 6 0.6 event expected total background is 4.1 3 1028.Figure 1(a) shows the energy distribution of all primaryelectrons which pass selection VI with associated R $ 0.Figure 1(b) shows the electron energy distribution forselection VI with R . 30.Kolmogorov tests have been done to check for unex-

pected concentrations of events in position (e.g., in regionsof high cosmic ray or g backgrounds), energy, or time(year). No consistency check yields a probability so lowas to demonstrate a serious inconsistency [5]. A restrictivegeometric cut, removing the 55% of the selection VI accep-tance with highest cosmic ray rates [5,7], also demonstratesno inconsistency; its results are labeled VIb in Table I.To determine the oscillation probability we fit the overall

R distribution, for events satisfying selection VI, in thefull energy range 20 , Ee , 60 MeV. The larger energyrange is used in this and the following fit to utilize themaximum amount of data and is made possible by ourincreased understanding of the background processes. The1763 beam-on and 11981 beam-off events were fit bya x2 method which took spatial variations in accidentalphoton rates into account by averaging the appropriate Rdistributions at the positions of each positron. The result of

FIG. 1. The energy distribution for events which pass selec-tion VI with (a) R $ 0 and (b) R . 30. Shown in the fig-ure are the beam-excess data, estimated neutrino background(dashed), and expected distribution for neutrino oscillations atlarge Dm2 plus estimated neutrino background (solid).

3084 first PRL final result in Phys. Rev. D 64 112007 (2001)

other

p(ν_e,e+)n

p(ν_µ→ν_

e,e+)n

L/Eν (meters/MeV)

Beam

Exc

ess

Beam Excess

0

2.5

5

7.5

10

12.5

15

17.5

0.4 0.6 0.8 1 1.2 1.4

FIG. 24: The Lν/Eν distribution for events with Rγ > 10 and 20 < Ee < 60 MeV, where Lν is

the distance travelled by the neutrino in meters and Eν is the neutrino energy in MeV. The data

agree well with the expectation from neutrino background and neutrino oscillations at low ∆m2.

62

excess 87.9 ± 22.4 ± 6.0 events 3.8σ evidence of appearance Peµ = (0.264 ± 0.067 ± 0.045)%

νe

Oscillation Allowed Regions

10-2

10-1

1

10

10 2

10 -3 10 -2 10 -1 1sin2 2θ

∆m2 (e

V2 /c

4 )

BugeyKarmen

NOMAD

CCFR

90% (Lmax-L < 2.3)99% (Lmax-L < 4.6)

FIG. 27: A (sin2 2θ,∆m2) oscillation parameter fit for the entire data sample, 20 < Ee < 200

MeV. The fit includes primary ν̄µ → ν̄e oscillations and secondary νµ → νe oscillations, as well

as all known neutrino backgrounds. The inner and outer regions correspond to 90% and 99% CL

allowed regions, while the curves are 90% CL limits from the Bugey reactor experiment, the CCFR

experiment at Fermilab, the NOMAD experiment at CERN, and the KARMEN experiment at

ISIS.

65

•  read their final paper! •  very detailed analysis •  why is this result not widely

accepted?

•  present day: MiniBooNE was supposed to test and rule out LSND (to 5σ) •  instead MiniBooNE neutrino

result rules it out but has unexplained low-energy background

•  while MiniBooNE antineutrino mildly supports LSND (possible excess consistent with appearance) νe