Recent R esults from Super- Kamiokande

H.Sekiya, Jul 30th 2008 , Philadelphia, ICHEP2008 　

Recent Results from Super-Kamiokande

Hiroyuki SekiyaICRR, University of Tokyo

on behalf of the Super-Kamiokande Collaboration

ICHEP 2008Philadelphia, USA

H.Sekiya, Jul 30th 2008 , Philadelphia, ICHEP2008 　 2

Super-Kamiokande Collaboration

• 130 authors• 36

institutions• 5 countries


Super-Kamiokande• 50kton water

Cerenkov detector • 1km (2.7km w.e)

underground in Kamioka zinc mine.

• 11129 50cm PMTs in Inner Detector

• 1885 20cm PMTs in Outer Detector

~5-20 ~20-50 ~1

Solar ν

MeVRelic SN ν

GeV TeV

Atmospheric ν

~100

Physics targets of Super-Kamiokande

Proton decay


History1996 1997 1998 1999 2000 2002 2003 2004 2005 2006 2007 20080120 06 2009

SK-I (1996-2001) 11,146 ID PMTs (40% coverage) 1,885 OD PMTs

SK-II (2003-2005) 5182 ID PMTs (19% coverage) Acrylic shields added

SK-III (2006-2008) 11,129 ID PMTs (40% cov.) OD segmentation 　 (top/barrel/bottom)

Coming soon:SK-IV (2008- ... ) Replace DAQ electronics


Solar neutrino

• Total flux• Day/night• Seasonal flux variation• Spectrum distortion

8B neutrino measurementby elastic scattering:

(sensitive to all ν flavors)

Phase Energy response

Energy threshold

SK-I 6p.e./MeV 5.0MeV

SK-II 3p.e./MeV 7.0MeV

SK-III 6p.e./MeV 5.0MeV→4.5MeV

SK-IV 6p.e./MeV 4.0MeV

Measure/Observe

Reconstruct• Recoil electron energy• Recoil angle relative to the Sun


Solar neutrino flux

6

Live time (days)

Energy Range (MeV)

Number of signal events Flux x 106 cm-2 sec-1

SK-III 289 6.5-20.0 3378.9 (stat only) In preparation

SK-II 791 7.0-20.0 7212.8 (stat) (sys) 2.38± 0.05 (stat) (sys)

SK-I 1496 5.0-20.0 2240± 226 (stat) (sys) 2.35 ± 0.02 (stat) ± 0.08 (sys)

+82.7-81.1

+152.9-150.9

+483.3-461.6+784-717

+0.16-0.15

SK-III 289 days Full Final sample6.5 - 20 MeV, 22.5 ktonSignal:

Preliminary

Extract number of signal events by fit signal + background shapes to cosθsun distribution

SK-III flux seems consistent with SK-I & SK-II flux

Derive neutrino flux

cosθsun distribution


Energy Spectrum arXive:0804.4312

7

• Consistent with flat• SK-II is consistent with SK-I

SK-II spectrum:

SK-II 791 daysEnergy-correlated errors

SK-I average


Oscillation Analyses (SK only)

8

SK Exclusion Regions SK Allowed Regions

SK-I onlySK-II onlySK-I + SK-II

SK-I onlySK-II onlySK-I + SK-II

8B flux constrained to SNO Salt Phase NC flux

Based on SK energy spectrum shapeSK-II contributes to widen the region.

S.N. Ahmed et al., PRL92 (2004) 181301


Solar global analysis+KamLAND

99

SK-I + SK-II + SNO + radiochemicalKamLAND (arXiv:hep-ex/0801.4589v2)

SNO data: 371-day salt phase (CC & NC fluxes) 306-day pure D2O phase (AD-N)Radiochemical data: Homestake SAGE GALLEX

KamLAND is consistent with solar global

Combined experimental data allow us to measure the oscillation parameters in this framework...

...but we would like to observe predicted upturn at low energyif the parameters are in this region…

Best fit


Prospects for SK-III/SK-IV

In order to see it, we must:reduce statistical errorsreduce energy-correlated sys errors (0.5 x SK-I) → 　 lower backgrounds &energy threshold

Low energy upturn ~10% effect in Super-K

Work in progress...

Rat

io (d

ata/

ssm

) SK-III, 5years(sin2q, Dm2) = (0.30,7.9x10-5)

E (MeV)

10

H.Sekiya, Jul 30th 2008 , Philadelphia, ICHEP2008 　11

SK-I SK-IIISK-III background rate lower than SK-I in central region

Z

R20

SK-III Backgrounds

cosθsun

• Radon in the water from the material (Tank/PMT/FRP)

SK-ISK-III

…promise for future lowering of analysis threshold


Solar

95%99.73% KamLAND

Solar+KamLAND

SK-III/IV up-turn sensitivity

x10-4

Year

2

1

sign

ifica

nce

σ

Target:2 sigma level upturn discovery (or exclusion) for 3 years observation

Assuming: - achieved SK-III Background level- 0.5x energy correlated systematic error of SK-I


Atmospheric neutrino interactions in SK

13,000 km

15 kmAtmospheric Neutrinos

Fully-Contained Partially-Contained Upward Stopping Muon

Upward Through-going Muon

Event categories in Super-kamiokande


• zenith angle analysis– Use many subsamples of data– Look for zenith angle　 distortion

• L/E analysis– Use much more selective subsample of data– Require good L/E resolution– Look for first oscillation dip

Oscillation analyses downwards（ L=10~100 km）

upwards（ L=up to 13000 km）

Θ

cosΘ


Updates in AnalysesRe-analysis of SK-I and SK-II data due to many changes/improvements:

Changed to agree with K2K measurement.Effect: Increase number of events

Effect: Small change in lepton momentum distributions

Effect: Suppression in forward direction of lepton scattering angle

Effect: Reduction in number of multiple-π events

Effect: Better data/MC agreement for various quantities

Simulationatmospheric neutrino flux model: Honda06neutrino interaction model (neut)QE: MA = 1.2 GeV1π (resonant): MA = 1.2 GeV Add Δ → Nγ Add lepton mass effects in CC1π1π (coherent): Rein & Sehgal with lepton mass correctionDIS: GRV98 PDF with Bodek-Yang correctiondetector simulationmore detailed model of light reflections and scatteringbetter OD tuning

Reconstructionimproved ring counting

Otherhigher MC statistics re-evaluate and add systematic uncertainties

Effect: Reduced systematic errors

Increase from 100 yrs to 500 yrs


Updated Zenith Angle distribution

16

Monte Carlo (no oscillations)Monte Carlo (best fit oscillations)

cos θzenith cos θzenith

cos θzenith

DatasetsSK-I FC/PC: 1489 daysSK-I Upmu: 1646 daysSK-II FC/PC: 799 daysSK-II Upmu: 828 days


Zenith Angle Analysis: SK-I + SK-II

Best fit:Δm2 = 2.1 x 10-3 eV2

sin2 2θ = 1.02χ2 = 830.1 / 745 d.o.f.

χ2 fit in bins of zenith angle with 90systematic error pull termsto account for uncertainties in:Neutrino flux Cross sectionsEvent reconstruction Data reduction


Updated L/E distribution

18

DatasetsSK-I FC/PC μ-like: 1489 daysSK-II FC/PC μ-like: 799 days

Use only event categories with good L/E resolution:

Partially-contained muons Fully-contained muons

Compare against:Neutrino decoherence (5.0σ)Neutrino decay (4.1σ)Grossman and Worah: hep-ph/9807511Lisi et al.: PRL85 (2000) 1166Barger et al.: PRD54 (1996) 1, PLB462 (1999) 462


L/E Analysis: SK-I + SK-IIBest fit:Δm2 = 2.2 x 10-3 eV2

sin2 2θ = 1.04χ2 = 78.9 / 83 d.o.f.

90% C.L. allowed regionsin2 2θ > 0.941.85x10-3 < Δm2 < 2.65x10-3 eV2

χ2 fit to 43 bins of log10(L/E) with 29 systematic error terms


SK-IV: DAQ UpgradeSK-IV Installation begins August 2008

to be completed by mid-September

~6-month commissioning period before T2K beam


New DAQ readout scheme

Currentreadout module

Newreadout module

Triggerlogic12 PMT

signalspermodule

24 PMTsignalspermodule

Hitsum

Trigger (1.3 μsec x 3kHz)

Readout (backplane)

Hardware triggerby hit information(HITSUM)

1.3 μsec event window

clockPeriodic trigger

(17 μsec x 60 kHz)

Readout (Ethernet)

Record every hit by 60kHz periodic timing signal x 17 μs TDC window

Variable event windowby software trigger

SK-I,II,III DAQ scheme:

No hardware trigger. Instead record all hits and apply software triggers.

SK-IV DAQ scheme:


Summary• Super-Kamiokande (I+II+III) have been operated

successfully. More than 10 years dataset for atmospheric & solar neutrino data are accumulated.

• Study “Standard Model” oscillation physics - help constrain solar parameters - precisely measure atmospheric parameters• Will continue to observe every predicted effect and

measure mixing angles.– SK-IV will start from this September with upgraded

electronics

22


Extra Slides

23


๏ Simplified detector operationsunified readout scheme for ID and OD

๏ Increased reliability/performance- fewer discrete components

- improve energy resolutionwider dynamic range

- improve multiple-hit capabilityefficient ID of μ-decay electrons

- reduce SPE hit thresholdlow E solar ν’sγ-tagging for proton decay

- improve supernova burst capability

๏ Ethernet-based readoutincreased bandwidth and reduced dead timebuild DAQ system from commodity network devices!

SK-IV: DAQ Upgrade


SK-IV Installation begins August 2008

to be completed by mid-September

~6-month commissioning period before T2K beam


Low energy events in SK

26

SK-I10 MeV electron

SK-II10 MeV electron

Simulated event Simulated event

♱Using SK-II improved algorithm

Energy response Vertex resolution for 10 MeV electron

SK-I ~6 p.e./MeV ~70 cm → 60 cm♱

SK-II ~3 p.e./MeV ~100 cm

SK-III ~6 p.e./MeV in preparation


BG reduction for SK-III solar ν analysis

27

100% trigger efficiency at 5 MeVPreliminary SK-III reduction tools

Datasets:✤ Full Final (FF) sampleLivetime: 288.9 daysEnergy > 6.5 MeV

✤ Radon Reduced (RR) sample (shown)→ periods of high radon activity removedLivetime: 191.7 daysEnergy > 5 MeV

Run period shown: Jan. 24, 2007 - Mar. 2, 2008

Good agreement of SK-III with SK-I final data sample


Energy Spectra

28

• Consistent with flat• SK-II is consistent with SK-I

SK-I spectrum: SK-II spectrum:

SK-I 1496 daysEnergy-correlated errors SK-II 791 days

Energy-correlated errors


Time Variation of Flux

29

SK-I SK-IISeasonal Variation

Consistent with expected variations due to eccentricity of Earth’s orbit

SK-IISK-I SK-I (binned)

Consistent with zero

SK-I asymmetry:

SK-II asymmetry:

Day/Night Asymmetry


SK-I 1496 daysEnergy-correlated errors

Prospects for SK-III+SK-IV

3030

In order to see it, we must:reduce statistical errorsreduce energy-correlated sys errors (0.5 x SK-I)→ 　 lower backgrounds &energy threshold

arXiv:hep-ph/0405172v6 Low energy upturn ~10% effect in Super-K


Event Categories

31

Event rates consistent across all phases of SK

Event CategoryEvent Rate (events/day)

SK-I SK-II SK-III(Preliminary)

Fully Contained (FC) 8.18 ± 0.07 8.22 ± 0.10 8.31 ± 0.22

Partially Contained (PC) 0.61 ± 0.02 0.54 ± 0.03 0.57 ± 0.06

Upward-stopping μ (Upstop) 0.25 ± 0.01 0.28 ± 0.02 0.24 ± 0.03

Upward-thrugoing μ (Upthru) 1.12 ± 0.03 1.07 ± 0.04 1.11 ± 0.06

SK-III run period: July 29, 2006 - present

Fully-Contained Partially-Contained Upward Stopping Muon Upward Through-

going Muon


SK-I1 GeV electron

SK-I1 GeV muon

SK-II1 GeV electron

SK-II1 GeV muon

32

Atmospheric ν in SK


SK-III Atmospheric ν Zenith Distributions

No oscillation analysis yet, but zenith angle distortion clearly visible

SK-III dataMonte Carlo (no oscillations)

\

>25,000 atmospheric ν events in SK-I + II + III


Atmospheric ν Analyses

Oscillation:Zenith angle (2-flavor)L/ENon-standard interactionsPoster by G. Mitsuka:“Limit on Non-Standard Interactions from the Atmospheric Neutrino Data in Super-Kamiokande”

Zenith angle (3-flavor) (Phys. Rev. D 74, 032002 (2006))

ντ appearance (Phys. Rev. Lett. 97, 171801 (2006))

MaVaNs (Phys. Rev. D 77, 052001 (2008))

Exotic scenarios: LIV, CPT, Sterile3-flavor with solar term

Non-oscillation:Nucleon decay searchesPoster by H. Nishino:“Search for proton decays via p → e+ π0 and p → μ+ π0 in Super-Kamiokande”

WIMP searchPoster by T. Tanaka“Search for Indirect Signal of WIMPs in Super-Kamiokande”


Atmospheric ν Analyses

Exotic ScenariosModel Exclusion level or limitνμ → νs oscillation SK-I+II: 7.3σAdmixture (2+2 hierarchy) SK-I+II: 23% allowedDecay I (sin4θ + cos4θ e-αL/E) SK-I+II: 17σDecay II (sin2θ + cos2θ e-αL/2E)2 SK-I+II: 3.9σDecay Limit (GeV2) SK-I+II: 6.5 x 10-23

Decoherence ((1+e-βL/E)/2) SK-I+II: 4.2σDecoherence Limit (GeV) SK-I+II: 6.0 x 10-24

LIV Limit SK-I+II: 1.2 x 10-24

CPTV Limit (GeV) SK-I+II: 0.9 x 10-23

MaVaNs (various models) SK-I: 3.5-3.8σNon-Standard Interactions See poster by G. Mitsuka

Neutrinos frequently set stringent limits, although not usually testing exactly the same parameters.e.g., cosmic ray spectrum LIV < 10-15, NMR LIV < 10-22

K0K0bar CPTV < 10-18


Zenith Angle Analysis

36

400 bins for SK-I 350 bins for SK-II

χ2 fit in bins of zenith angle with systematic error pull terms:

Data binned according to:event type+momentum+zenith angle

where

90 systematic error terms to account for uncertainties in:Neutrino flux Cross sectionsEvent reconstruction Data reduction

DatasetsSK-I FC/PC: 1489 daysSK-I Upmu: 1646 daysSK-II FC/PC: 799 daysSK-II Upmu: 828 days


Zenith Angle Analysis: SK-I + SK-II


sin2 2θ = 1.02χ2 = 830.1 / 745 d.o.f.


L/E Analysis:SKI+SKII

38

χ2 fit to 43 bins of log10(L/E) with 29 systematic error terms

DatasetsSK-I FC/PC μ-like: 1489 daysSK-II FC/PC μ-like: 799 daysUse only event categories with good L/E resolution:

Partially-contained muons Fully-contained muons

Compare against:Neutrino decoherence (5.0σ)Neutrino decay (4.1σ)

Grossman and Worah: hep-ph/9807511Lisi et al.: PRL85 (2000) 1166Barger et al.: PRD54 (1996) 1, PLB462 (1999) 462


L/E Analysis: SK-I + SK-II


sin2 2θ = 1.04χ2 = 78.9 / 83 d.o.f.

90% C.L. allowed regionsin2 2θ > 0.941.85x10-3 < Δm2 < 2.65x10-3 eV2


SK-II Low E reconstruction

Vertex resolution improves with new algorithms developed for SK-II


Target:2 sigma level upturn discovery (or exclusion) for 3 years observation

SK Sensitivity to Solar Upturn

To achieve this, we must:enlarge fiducial volume while maintaining low backgroundreduce energy-correlated systematic errors


Oscillation Analysis

Spectrum Energy correlated systematic error Time variation

Function for energy correlated systematic errors

8B spec. shape

energy scale

energy resolution


Unbinned Time Variation Analysis

Solar Signal Shape

Solar ν FluxTime-Variation

Background Shape

EventEnergy

Event“Time”

# SignalEvents

# Backgroundsin each energy bins

21 Energy bins

Likelihood for solar neutrino extraction


Solar Neutrinos at Super-K

SK-III data agree well with SK-I- Analyses are in progress- Low background in central region of detector shows promise for future lowering of analysis threshold

SK-I + SK-II results are consistent

SK will continue work to hopefully observe low energy upturn

H.Sekiya, Jul 30th 2008 , Philadelphia, ICHEP2008 　45♱Using SK-II improved algorithm

Low energy events in Super-K

SK-I10 MeV electron

SK-II10 MeV electron

Simulated event Simulated event

Energy response Vertex resolution for 10 MeV electron

SK-I ~6 p.e./MeV ~70 cm → 60 cm♱

SK-II ~3 p.e./MeV ~100 cm

SK-III ~6 p.e./MeV in preparation


Ring likelihood

Old New

sub-GeV sub-GeV

multi-GeV multi-GeV

Phys. Rev. D 71, 112005 (2005)

Reduced systematic uncertainty: XX% (old) XX% (new)Better separation of single vs. multi-ring events


Better data/MC agreement for PC and FC datasets

Ring Counting

New

Old


ID/OD Optical Segmentation

PC

CornerClippers


L/E and Zenith Angle: SK-I + SK-II

Stronger constraint on Δm2 with high resolution sub-sample of data


L/E Analysis (SK-I re-analysis cf. published result)

Best fit oscillation parameters

PRL93, 101801 (2004) best fit:

(1.00, 2.4 x 10-3)


Zenith Analysis Systematic Errors


Zenith Analysis Systematic Errors


L/E Systematic Errors

NormalizationMulti-mu / PC relative normFlux: anti numu/numu ratioFlux: up/downFlux: horiz/verticalNeutrino flight lengthenergy spectrumSample-by-sample multi-GeVQE cross sectionSingle-pi cross sectionDIS cross sectionCoherent pi cross sectionNC/CCFC reductionPC reductionRing countingPID 1-ringEnergy calibrationUp-down asym of E calibrationNuclear effectsNon-nu BG (cosmic ray mu)PC stop/thru separationQE and Single-pi MADIS (Bodek-Yang correction)Solar activitySingle-meson pi0/pi+pi-PC stop/thru bottomPC stop/thru barrel


• Search for proton decays via pe+p0 and pm+p0 in Super-Kamiokande

54









3 flavor oscillation using SK-1

• Assume One mass scale dominance Δm212 ＝ 0

• Earth matter effect play an important role.• Normal mass hierarchy (90%CL)

– sin2θ13 < 0.14

– 0.37< sin2θ23 < 0.65

• Inverted mass hierarchy(90%CL)– sin2θ13 < 0.27

– 0.37< sin2θ23 < 0.69

62


Normal mass hierarchy

63


Inverted mass hierarchy

64


SN & GADZOOKS!

65


Super-K: Expected number of events

~7,300 νe+p events~300 ν+e events~360 16O NC g events ~100 16O CC events (with 5MeV thr.)for 10 kpc supernova

Neutrino flux and energy spectrum from Livermore simulation (T.Totani, K.Sato, H.E.Dalhed and J.R.Wilson, ApJ.496,216(1998))

H.Sekiya, Jul 30th 2008 , Philadelphia, ICHEP2008 　S.Ando, NNN05

Supernova Relic NeutrinosS.Ando, Astrophys.J.607:20-31,2004.


0

0.5

1

1.5

2

2.5

3

3.5

4

Constant SN rate

(Totani et al. 1996)

Totani et al. 1997

Malaney et al. 1997)

Hartmann et al. 1997)

Kaplinghat et al. 2004

Ando et al. 2005

Fukugita et al. 2003

Lunardini et al. 2006

SK-II limit = 3.68 /cm2/sec

SK-I limit = 1.25 /cm2/sec Combined limit = 1.08 /cm2/sec

preliminary(E>18MeV)

Super-K results so farFlux limit VS predicted flux

T.iida, poster #90.5

νe can be identified by delayed coincidence.

Neutron tagging in water Cherenkov detector

νe

e+

pn

g

g

Positron and gamma ray vertices are within ~50cm.

2.2MeV g-rayDT = ~ 200 msec

Another possibility

n+p→d + g

Number of hit PMT is about 6 in SK-III

p

Gd

n+Gd →~8MeV gDT = ~30 msec

GADZOOKS!

(J.Beacom and M.Vagins) Phys.Rev.Lett.93:171101,2004

Add 0.2% GdCl3 in water

This apparatus deployed in the SK tank.

BGO

13 cm

18 cm 18

cm

5 cm

BGO

0.2 % GdCl3Solution

Am/Be

GdCl3 test vesselTest neutron tagging at Super-K

α + 9Be → 12C* + n 12C* → 12C + g(4.4 MeV) n + p → …… → n + Gd → Gd + g (totally 8 MeV)

BGO signal (prompt signal (large and long time pulse))Look for Cherenkov signal (delayed signal)

H.Watanaba, poster #94

Cherenkov signal of Gd gamma rays

τ = 23.7±1.7μs

[msec]

Num

ber o

f Eve

nts

0

Time from prompt

100ms

Vertex position92% within 2m

dR [cm]

Energy spectrum

Measured time, vertex and energy distributions are as expected from the MC simulation.

Recon. Energy [MeV]

Num

ber o

f Eve

nts

Black: DataRed: MC

H.Watanaba, poster #94

10-410-310-210-1

110102103104

6 8 10 12 14 16 18 20

SK-I final data sample

SSM(BP2004) * 0.4(efficiencies are considered)

Selection criteria of delayed signal:(1) Vertex position within 2m(2) Energy of delayed signal > 3MeV(3) Time after the prompt within 60msec.(4) Ring pattern cutsSelection efficiency is ~74%.With 90% capture eff. by 0.2% Gd,

Tagging efficiency is 67% 92 %

Tagging efficiency and BG reduction

While the chance coincidence prob. is estimated to be ~2×10-4

Recon. Energy [MeV]

Energy spectrum

MC simulation

It almost satisfy the requirement to remove remaining spallation background at 10 MeV.

SRN predictions

Current single BG rate(mainly due to spallation and solar 8B)

H.Watanaba poster #94

Possibility of SRN detectionRelic model: S.Ando, K.Sato, and T.Totani, Astropart.Phys.18, 307(2003) with flux revise in NNN05.

If invisible muon background can be reduced by neutron tagging

Assuming invisible muon B.G. can be reduced by a factor of 5 by neutron tagging.

By 10 yrs SK data,Signal: 33, B.G. 27(Evis =10-30 MeV)

SK10 years (e=67%)

Assuming 67% detection efficiency.

0123456789

10

10 15 20 25 30 35 40 45 50

relic+B.G.(inv.mu 1/5)

B.G. inv.mu(1/5)atmsph.ν–e

Visible energy (MeV)

even

ts/10

years

/2Me

V

Items to be studied before introducing gadolinium to SK

Effect to water transparencyWater transparency should be long enough to do various physics at SK.Water purification systemCurrent water purification system remove ions. So, it must be modified to purify water without removing gadolinium.Material effectsCorrosion by gadolinium solution should be checked.How to introduce/removeHow to mix gadolinium uniformly in the tank. How quickly/economically/completely can the Gd be removed? Ambient neutron level in the tankDoes it cause significant increase singles in trigger rate[for solar analysis]?

In order to study those things, we will construct a test tank (6~10m size) in the Kamioka mine.

Documents

Recent R esults from Super- Kamiokande