1 K L →π 0 νν 探索実験 KEK-PS E391a における Run3 データ解析の現状 JPS 2008 Spring Meeting Hideki MORII (Kyoto Univ.)

1

KL→π0νν 探索実験 KEK-PS E391a におけるRun3 データ解析の現状

JPS 2008 Spring Meeting

Hideki MORII (Kyoto Univ.)

2

Contents• Introduction

– E391a Run3– Detector Upgrades in Run3

• BA (calibraion plot)• APC (calibration / g-tagging)

• Current Status & Strategy for Run3 analysis– calibration– MC development– Run3 data quality check– Optimize cuts

• Plan• Summray

Overview

3

E391a Run3• E391a

– KL→π0νν measurement @ KEK 12GeV PS– pilot experiment for J-PARC E14– Three data taking

• Run1 : Feb 2004 – Jul 2004• Run2 : Feb 2005 – Apr 2005• Run3 : Nov 2005 – Dec 2005

• Run3– Detector Upgrade

• new Back Anti (in-beam veto)• Aerogel Photon Counter (-tagger)

– Data Taking• (almost) stable DAQ condition

Introduction

4

Detector Upgrade in Run3

• aaa

Introduction

Aerogel Photon Counter(APC)

Upgraded

Added

Back Anti

5

Back Anti UpgradeIntroduction

PWO crystalconstruction

• In-beam veto counter

Back AntiUpgraded

6

Aerogel Photon CounterIntroduction

• Prototype of E14 BA

Aerogel Photon Counter(APC)

Added

γ

e+

e-

Cerenkov light

7

Run3 Data• Full Intensity Run (C only mode)

– with Be aborber – ~583000 spills– ~70% of Run2 data

• Half Intensity Run (A&C mode)– w/o Be absorber– ~98000 spills– ~10% of Run2 data

c.f. ) Run2– with Be absorber– ~860000 spills

Introduction

KEK 12GeV PS East Counter Hall

E391aC-line

A-line

Be absorber to reduce neutron (betteer n/K ratio) # of KL ~60% / # of n ~ 40%

※

※

※

8

Strategy for Run3 AnalysisCalibration

kdecay,halo-n, eta

Strategy for Run3 Analysis

Results

Step1Confirmation

Step2 Optimization

Step0Preparation

Step3Physics Output

[MC] Develop Run3 MC

[Data] Data quality check

[Data] Cut optimization

[MC] MC mass production

9

Step0 : PreparationCalibration

kdecay,halo-n, eta

Step0 : Preparation

Results

Step1Confirmation

Step2 Optimization

Step0Preparation

Step3Physics Output

[MC] Develop Run3 MC

[Data] Data quality check

[Data] Cut optimization

[MC] MC mass production

10

Step0 : Preparation

• Calibration– completed

(including upgraded / new detectors)

• MC development– detector upgrades are implemented– now under middle-size production : confirming results– preparing for mass-production

Step0 : Preparation

11

Back Anti Upgrade• Upgrade Back Anti

– lead plate + plastic scinti. + quartz

PWO crystal + quartz– segmentation :

longitudinal

transverse

• Benefits– better n/ separation

(shower shape analysis)– lower rate

(typ. 1/2 @ center crystal)

Run2 BA

beam

Run3 BA

Step0 : Preparation

12

Calibration of Back Anti

• Calibration– Calibration has been done

with Muon Run

ADC count

MIP peak with Muon Run

Muon Run : use from upstream (with beam shutter closed)

MIP peak

Step0 : Preparation

13

Step1 : ConfirmationCalibration

kdecay,halo-n, eta

Step1 : Confirmation

Results

Step1Confirmation

Step2 Optimization

Step0Preparation

Step3Physics Output

[MC] Develop Run3 MC

[Data] Data quality check

[Data] Cut optimization

[MC] MC mass production

14

Step1 : Confirmation• Confirm Run3 data quality

– compare with Run3 MC : MC middle size production– compare with Run2 data

KL 30 6sample, 4(KL 20 4, 2

Step1 : Confirmation

Run2 Run3

Mass (GeV/c2) Mass (GeV/c2)

Invariant Mass of 6 sample

Matchs well CsI calibration is good in Run3

15

Step2 : OptimizationCalibration

kdecay,halo-n, eta

Step2 : Optimization

Results

Step1Confirmation

Step2 Optimization

Step0Preparation

Step3Physics Output

[MC] Develop Run3 MC

[Data] Data quality check

[Data] Cut optimization

[MC] MC mass production

16

Step2 : Optimization• Develop complete BA veto (algorithm, threshold, etc…)

• Study with Run2 opened box– to get more acceptance– optimize veto & event selections

Step2 : Optimization

Veto kinematic cut

Acceptance Loss / Rejection Power in Run2

17

PlanCalibration

kdecay,halo-n, eta

Plan

Results

Step1Confirmation

Step2 Optimization

Step0Preparation

Step3Physics Output

[MC] Develop Run3 MC

[Data] Data quality check

[Data] Cut optimization

[MC] MC mass production

Completed

Almost done

Ongoing (~1-2 month)

OK

~4-5 month

In parallel withMC mass prod.

18

Summary• E391a Run3

– Data : ~70% of Run2 (~80% with A&C mode)– Detector Upgrade

• Upgraded Back Anti (BA) : in-beam -veto• New Aerogel Photon Counter : prototype of E14 BA• both worked well

• Current Status– calibration is completed– finalizing MC development– Now checking data quality

• Future Plan– Develop BA veto– Precise study with Run2 opened box for more acceptance

Summary

19

backup

20

The E391a experiment

• KL production with KEK 12GeV PS– Slow extraction– K0 beamline in the East Counter Hall

• Intensity– 2 x 1012 protons on target (POT) per 2sec spill, 4sec cycle

• production angle: 4°, KL peak momentum 2GeV/c, n/KL ratio: ~40

• Physics runs– Run I: February to July of 2004

• “Express” analysis with 10% data published in PRD (2006)

– Run II: February to April of 2005• The main topic of this seminar• Full data analysis

– Integrated protons: 1.4x1018 POT» ~ 32 days without break

– Run III: October - December of 2005• Calibration ready, MC development in progress

21

Principle of the experiment

1. require 2 photons– Hermetic veto system

2. measure the photon energies and positions

3. reconstruct the decay vertex on the beamline assuming M2γ = Mπ0

halo/core ~10-5

5cm

4. require missing PT and the vertex in the fiducial region• “Pencil” beam line

to improve PT resolution- 8cm diameter @ 16m

from the target

22

Features of E391a apparatus

• Decay region– High vacuum: 10-5 Pa

• to suppress the backgroundfrom interactions w/ residual gas

• Detector components– Set in the vacuum: 0.1 Pa

• separating the decay regionfrom the detector regionwith “membrane”: 0.2mmt film

CsI calorimeterCsI calorimeter

Charged Veto (CV)Charged Veto (CV)Main Barrel (MB)Main Barrel (MB)

Front Barrel (FB)Front Barrel (FB)

23

Aerogel Photon Counter• Aerogel Photon Counter (APC)

– Aerogel Cherenkov counter : only sensitive to fast particle• insensitive to neutrons / sensitive to shower

– Can be used as photon tag counter (for BA study)– prototype of E14 BA

Pb convertor : 2mm thick (~0.3 X0)Aerogel : 30cm(x) x 30cm(y) x 5cm(z)

γ

e+

e-

Cerenkov light

Step0 : Preparation

24

Calibration of Aerogel Photon Counter

• Light yield– calibrated with Muon Run– MIP peak appears clearly

• Photon Tagging– checked with KL decay– 2 MIPs (= e+e-) peak seen

Muon Run Physics Run (KL decay sample)

ADC count ADC count

MIP peak ~300count 2 MIPs peak

~ 600 count

Step0 : Preparation

25

Response to photons (3)

N-cluster trigger

Accidental trigger (TMON)

(Black) – (Red)

15 p.e.

• Clear peak around 15 p.e. is observed.

• Response to photon is well reproduced by MC.

• Tagging quality is 94 % (#p.e. > 10). will be improved by using (for example) 5 sample.

26

Mechanism of CV Background

• even+extra odd

and 2 extra

fusion 1 from 0 + extra

27

Mechanism of CV background• removing Veto : odd & 01+extra• removing g-selection : even+extra• with bifurcation for each mechanism, even+extra is dominant

setup + box + selection + veto

-veto - selection all cuts

all 3277 44 6 0 (.081)

even+extra 306 4 2 0 (.026)

odd 210 26 0 0 (0.0?)

fusion 226 3 1 0 (.013)

1+ extra

2525 11 3 0 (.013)

extra 2clustr 10 0 0 0 (0.0?)

tighten

loosen

28

Electronics and DAQ• Number of channels

– CsI calorimeter: ~600ch– Veto counters: ~400ch

• “AmpDiscri” Module– Discrimination for TDC– Set near the detector

• low noise– min. threshold: ~0.5 mV

(ex. ~0.7MeV for CsI)

– 8ch sum for the trigger

• Trigger– Logic

• CsI hardware clustering (thres. 80MeV) + Veto (20-100MeV)

– ~300 events / 2 sec spill = 150Hz

• DAQ live time– ~90%

PMT

PMT

8ch

Analog each out

Analog sum out

Digital outfor

timing

Trigger logic

CsI

FASTBUS-VME

FASTBUS ADC

Veto

FASTBUS ADC

AmpDiscri

FASTBUS-VME

TKO TDC

TKO-VME

GbE

Event BuilderStorag

e

29

Problems in Run-I• core neutron background

– hitting on the membranesagging into the beam-line

30

Result from Run-I 1week

• Using 10% of Run-I data

• set new limit– Br < 2.1x10-7 (@90%C.L.)

(PRD 74:051105, 2006)

31

Can we speed up MC mass production?

• Halo-n needs large amount of MC– needed 3 months in Run2 halo-n MC– in Run2 analysis, we used bifurcation method for CV bg

• Recycling Method– collect only BG-like events with strong online-veto

(discard “safe” events in production stage)– then, full simulation for “dangerous” events

Strategy for Run3 Analysis