JPARC MLF RCNP on behalf of the JPARC

J-PARC MLFにおけるステライルニ ュートリノ探索実験 平岩 聡彦 阪大RCNP on behalf of the J-PARC P 56 collaboration (We requested the 1 st stage approval at the 19 th PAC on Dec 2014. ) Contents • Introduction - LSND - Current status of sterile neutrino searches. • Sterile neutrino search at MLF (J-PARC P 56) - Experimental principle. - Experimental features. • Background measurement at candidate sites - Results - Sensitivity • Summary and outlook 21 st ICEPP Symposium 1

Introduction 21 st ICEPP Symposium 2

LSND anomaly • LSND: signal: nm ne (appearance) • Using m+ decay at rest (m+ DAR): PRD 64 (2001) 112007 Oscillations ? • Detected by Liq. Scinti. : nep e+n (IBD), followed by neutron capture g (2. 2 Me. V) • Excess events: 87. 9 ± 22. 4 ± 6. 0 events. 3. 8 s evidence for oscillation Sterile neutrino(? ) 21 st ICEPP Symposium 3

Status of sterile neutrino search (Dm 2 ~ 1 e. V 2) • Positive results: Experiments Neutrino source signal type Significance σ LSND μ Decay-At-Rest νμ→νe appearance 3. 8 Mini. Boo. NE π Decay-In-Flight νμ→νe appearance 3. 4 νμ→νe appearance 2. 8 combined 3. 8 Gallium e capture νe→νX disappearance 2. 7 Reactors Beta decay νe→νX disappearance 3. 0 • There are several negative results: - Mini. Boo. NE (disappearance). - KARMEN etc. • A definite search is awaited. (high statistics and low background)

Sterile neutrino search at J-PARC MLF (J-PARC P 56 experiment) 21 st ICEPP Symposium 5

Material and Life science Facility (MLF) 400 Me. V Linac Candidate site (3 F) L = 24 m 3 Ge. V Synchrotron Rapid Cycle Synchrotron Energy: 3 Ge. V Repetition: 25 Hz Design Power: 1 MW 40 ms 600 ns 100 ns x 2

J-PARC P 56 experiment • Search for the LSND anomaly using m+ decay at rest (m+DAR) : - nm ne (appearance). • Detector: - Gd-loaded liquid scintillator. (25 tons x 2 ~ total 50 tons) • Measurement principle: - “Delayed Coincidence”: - ne + p e+ + n (IBD) prompt signal - 8 Me. V g from n-capture by Gd delayed signal (capture time ~ several tens msec) - En = Ee(visible) + 0. 8 Me. V 21 st ICEPP Symposium 7

Experimental features beam bunch • Pulsed beam and muon long life time enable the separation of m. DAR. (top fig. ) • Due to nuclear absorption, neutrinos from p- and m- decay (main BG) are highly suppressed to the same level of the signals. Signature of oscillation by spectrum shape. (bottom fig. ) • Well-defined energy spectrum shape of n from m. DAR. • Well-known cross section for IBD (ne + p e+ + n). m. DAR Dm 2 = 5. 5 e. V 2 21 st ICEPP Symposium Neutrino energy 8

Background measurements at candidate sites 21 st ICEPP Symposium 9

Background measurement • BKG measurements were performed at the candidate sites (MLF 3 F). (Apr-Jun 2014) • Detectors: - 500 kg plastic scintillation counter (yellow): main detector veto eff: - Inner veto counter (red). - Outer veto counter (green). > 99. 9 % • 2 different data sets: - beam-on - beam-off (to subtract the beam-unrelated BKGs. ) • 3 different points: point-1, 2, 3. • The results for “point 2”(L ~ 20 m) are presented here. 21 st ICEPP Symposium Point 1: L ~ 17 m Point 2: L ~ 20 m Point 3: L ~ 40 m 10

BKG(1): Michel-e from beam fast n On bunch e+ , νe + p→ e+ + e- → 2γ IBD No activities n+C→X+π →m Michel-e by beam n Clipping muons (Cosmic) delayed Prompt + Selection criteria n +Gd →γ time 〜 30μs 〜 2. 2μs μ→e • Prompt signal: 1 < Tp [ms] < 10 20 < Ep [Me. V] < 60 • Delayed signal: Tp < Td [ms] < 100 7 < Ed [Me. V] < 12 Thermalized n captured by Gd time Huge, but rejected by charged veto (eff > 99. 9 %) • Beam-associated fast neutrons (T > 200 Me. V) can produce pions, followed by Michel electrons. (p m e) • Michel electrons from beam fast neutrons: - Michel-e from beam fast neutron mimics the IBD signals. • The “Michel-e” signals have activities on bunch timing, whereas the “IBD” signals have no activities on bunch timing.

Before “on-bunch activity cut” Energy vs Hit Time Beam bunch projection after veto Before charged cosmic veto After “on-bunch activity cut” (require Eonbunch>4 Me. V) “beam on” /spill/300 k. W “beam off” /spill/300 k. W subtraction Before cosmic veto (1. 68± 0. 03)× 10 -4 (1. 64± 0. 03)× 10 -4 (4. 0± 4. 2)× 10 -6 After veto (1. 58± 0. 09)× 10 -5 (1. 52± 0. 09)× 10 -5 (0. 6± 1. 3)× 10 -6 (4. 91± 0. 28)× 10 -7 (expectation) (-0. 3± 1. 6)× 10 -7 (90%C. L. UL；= <13 /5 y/50 t/MW） 20<E<60 Me. V 1. 75<t(ms)<4. 65 p bunch After on-bunch cut (4. 60± 1. 53)× 10 -7 No Michel-e from beam fast n !! = beam off data (veto) x accidental on-bunch 12

BKG(2): Accidental backgrounds • Accidental background rate: Racc = Rprompt x Rdelay x Dvtx x Nspill - Rprompt: BKG rate for prompt signal. - Rdelay: BKG rate for delayed signal. - Dvtx: Rejection factor of spatial correlation cut (= 1/50) - Nspill: # of spills (= 1. 5 x 109 /5 years) • Rprompt and Rdelay were measured: - Prompt: cosmic gammas and neutrons. - Delayed: - Beam associated gammas. - Beam neutrons 21 st ICEPP Symposium 13

Cosmic g and n (Prompt): • Measurements using small NE 213(< 1 kg) and Na. I (2’’f x 2’’) @ Tohoku. (identify g and n) • Scaled to 500 kg scinti. at MLF 3 F. (right fig) Consistent within 6%. • Gammas and neutrons are dominant. (neutrons can be removed by PID of the P 56 detector. ) Beam-associated g (Delayed): • Beam neutrons are thermalized and captured by the concrete floor, and g’s are emitted. 21 st ICEPP Symposium 12. 5 cm thickness lead under the detector 1/1000 g’s 14

BKG summary and Sensitivity Source Contents Number of Event/50 t/5 y Comments BG νe from μ- 237 Main BKG. L = 24. 12 C(ν e, e-) 12 N g. s 16 Michel-e from beam fast n <13 (90%Cl UL) Fast neutron (cosmic) 37 Accidental 32 Based on measurement. 480 Δm 2=3. 0 sin 2θ=0. 003 342 Δm 2=1. 2 sin 2θ=0. 003 Signal Sensitivity (MW x 5 years, L = 24 m) Based on measurement. 5 s sensitivity as a function of MW x years Δm 2>2. 0 e. V 2 LSND 90%CL Allowed region (lower edge) (high Dm 2 region) We can discuss the all LDND allowed region (90% C. L. ) with 3 s (MW x 5 years). Especially for Dm 2 > 2. 0 e. V 2, we can conclude with 5 s (MW x 4 years).

Summary and outlook • We plan to perform a definite search for sterile neutrinos at J-PARC MLF. • Background measurements at the candidate sites were performed, and the experimental feasibility was examined. • We can conclude all the LSND allowed regions (90 % C. L. ) with 3 s. (1 MW x 5 years) • We can start the experiment within 1 -2 years after getting the budgets. New challengers, especially young scientists, are very welcome. Please join us !! 21 st ICEPP Symposium 16

J-PARC P 56 collaboration 21 st ICEPP Symposium 17

Backup slide 21 st ICEPP Symposium 18

Success in RCS 1 -MW trial NOTE: This is a very short term test. 21 st ICEPP Symposium 19

BKG(2): Accidental BKG for “Prompt” • Acccidental BKG: Racc = Rprompt x Rdelay x Dvtx x Nspill - Rprmpt, Rdelay: BG rates for prompt and delay. - Dvtx: spatial correlation (rejection power: 1/50) - Nspill: # of spills: 3 x 108/year • Measurement @ Tohoku (Left figure): - Using Na. I and NE 213 (w/ PID capability), surrounded by cosmic veto counters. - ratio: g : n = 3 : 1 (20 < E [Me. V] < 60) • Measurement @ MLF 3 F (right figure): - Consistent with the rate predicted by the Tohoku results within 6 %. • g’s and neutrons are dominant. (neutrons can be rejected by PID) x 30 larger than that in proposal !! 21 st ICEPP Symposium 20

BKG(3): Accidental BKG for “Delayed“ (beam g) • Event rate @ “point 2”: > 1 k. Hz (E > 1 Me. V) • 10 times larger than that @ “point 3” • Assumption: beam associated neutrons are thermalized and captured by the concrete floor and g’s are emitted. Energy spectra for “delayed” • It can well reproduce the measured spectrum. • Beam g’s can be reduced by putting 12. 5 cm thickness lead under the detector down to 1/10. (Checked by small plastic scintillator counter. ) 21 st ICEPP Symposium 21

BKG(4): Accidental BKG for “Delayed“ (beam Beam bunch • Beam associated neutron (Tn > 10 Me. V) can reach the fiducial volume and are thermalized and captured by Gd. Delayed BKG: 0. 016/spill/MW/25 t On-bunch n • Strong spatial correlation between “onbunch neutron” and “delayed captured g” • DVTXOB-delayed cut: Delayed n rate: 4 x 10 -4 /spill/MW/25 t • Signal inefficiency due to accidental onbunch hit: < 2. 0 % 21 st ICEPP Symposium 22