Experimental Search for the Decay

1
Experimental Search for the Decay
K. Mizouchi (Kyoto University)
(1) Physics Motivation (2) Detector (3) Selection
Criteria (4) Branching Ratio (5)
Background Subtraction (6) Conclusions
2
Physics Motivation
1 Helicity suppressed decay
(A) Neutrino mass
implies .
(B) Neutrino type Majorana neutrino ? x2
larger branching ratio.
3
Event Detection Strategy
Charged particles from K decay at rest
Km2
Hermetic photon detection system
Kp2
4
E949 Detector
E949 detector side view (upper half)
E949 detector end view (upper half)
(1) Barrel Veto (BV) Pb-scintillator sandwich
(2) Barrel Veto Liner (BVL) Pb-scintillator
sandwich (3) Endcap Calorimeter CsI crystals
5
Analysis Strategy
Offline Data (Kp2 rich)
1/3 sample
2/3 sample
(1) Kp2 selection
p0 sample
p0 sample ( )
tuning
(2) Find the best photon veto parameters
Signal candidate (N)
Acceptance Cacc
6
Kp2 selection and none-Kp2 bkgnd
Real data (2/3 sample)
Impurity 10-9
7
Disruption Correction Factor Cdis
Overlapping g,e/- (from p0) may cause disruption
in the p track reconstruction.
Disruption correction

8
(1) Kp2 Tag Done.
(2) Hermetic Photon Veto
9
Acceptance Measurement Cacc
acceptance loss due to coincident
accidentals
p
accidental
n
n
Measure acceptance loss of Km2 decays (real data)
by the photon veto, after all m activities are
removed.
10
Maximization of the Sensitivity
Photon veto rejects events with Esum in
T1,T2 gt Ethreshold
Real data 1/3 sample
Hermetic photon veto
(1) p0 ? gg rejection
(2) p0 ? nn acceptance
Find the best parameters the largest rejection
with the given acceptance.
11
Opening the Box
Real data 2/3 sample
A total of 99 candidates were observed in the
signal box
p momentum (MeV/c)
Kaon decay time (ns)
12
Branching Ratio
Conservative upper limit
signal lt 113 (90CL) subtracting the non-Kp2
bkgnds
New upper limit A factor of 3 improvement
from the previous best result.
13
p0 ? gg Background subtraction
Measurement of the detector single photon
inefficiency
Kp2 w/ one photon missing event
Relaxed photon veto (acc 0.80)

1. Establish a background subtraction method
2. Understand the detector performance

14
Single Photon Inefficiency
15
p0 ? gg detection inefficiency
(2) Photon kinematics
(1) Single photon inefficiency
PSPI
from MC simulation (N events)
16
p0 ? gg background subtraction
Number of candidates with relaxed
photon veto
Arbitrary
4131 events
Singal (90 C.L) 2259
A factor of 1.8 improvement
17
Subtraction at various levels of photon veto
Improvement (Before/After)
A factor of two improvement at various photon
veto
18
Num of p0 ? gg backgrounds as a function of
cos(qp)
Signal candidates
Single photon inefficiency
Signal discrimination capability from backgrounds
19
Background Subtraction with dip angle distribution
Candidates sraw 4131 Best fit value s
1977 90 C.L. s90 2449
20
Conclusions
(1) search was performed with
3.02x109 Kp2 events, where impurity of 10-9
was achieved. (2) New upper limit of
was obtained
with a total number of 99 candidates in
the signal region x3 improvement
from the previous best limit.

• Single photon inefficiency was measured with
  special data
• p0 ? gg background subtraction was performed with
  the inefficiency
• (A) x1.8 improvement with simple subtraction
• (B) x1.7 improvement from cos(qp) shape
  discrimination

21
Thank you !
22
Early accidental hits
23
Two peaks in BVL
24
Background distribution
25
Kp2 photon kinematics
26
Unvetoed hits in Candidates
Unvetoed hits in BV
Outside the veto time window. Lower energy than
threshold.
27
Unvetoed hits
(6)
(5)
(4)
(1)
(3)
(2)
(2)
(3)
(4)
(5)
(6)
(1)
PV for single photon study
PV for search
28
Kp2 photon kinematics
29
Background understanding and detector inefficiency
Can we understand the remaining events from a
view of photon inefficiency ? ( if possible,
subtract them as backgrounds.)
An Idea

• Special trigger
• Kp2 but one photon is missed.

(2) Event reconstruction Missing photon kinematics
?
(3) Photon inefficiency as a function of its
energy and direction
NOTE

1. Different type of critical backgrounds.
2. Geometrical dependence Detector hole, dead
   material
3. Energy dependence Photonuclear
   interaction

30
Barrel Veto Liner
31
K ? p nothing in E949
(1) K ? pnn (above Kp2) Published in PRL, 93
031801 (2004)
Charged track momentum from various K decay modes
(2) p0 ?nn (on Kp2 peak) This report. Need
tighter photon rejection.
(3) K ? pnn (below Kp2) Analysis ongoing.
Require more sophisticated treatment in p
multiple scatterings.
32
Published in PRD as rapid communication
Phys. Rev. D72, 091102 (2005)
33
Optimized Photon veto parameters
34
Performance of the clustering Method
MC sample
theta
35
p0 ?gg backgrounds
Photon inefficiency 20ltEgMeVlt225
Low energy g sampling fluctuation High energy
g photonuclear interaction ( hard to simulate
reliably.)
Detector photon inefficiency (measured with real
data)
2040MeV
4060MeV
6080MeV
80100MeV
100120MeV
120140MeV
140160MeV
36
Phase space correction factors
Polar angle distribution
Correction factors
37
Self-vetoing effect due to split photon
Missing-side
MC simulation
Missing photon kinematics
38
Single Photon Inefficiency
Measure single photon inefficiency with real data.
Kp2 w/ one photon missing event
(1) raw photon distribution
(2) mis-detected photon
distribution
(3) Trigger prescale compensasion,
39
Analysis Strategy
Kp2 w/ one photon missing event

1. Reconstruct (tagging) photon
2. Extract kinematics of the mis-detected photon.
3. Correction factors

40
(1) Photon Clustering Method
Reconstruct photons and extract their (A)
positions (B) energies and
(C) timings.
41
(2) Kinematical Fitting
Lagrange Multiplier c2 minimization with
constraints.
(A) Four Constraints
(B) Five inputs
42
Correction factors

• CL1.1after unwanted trigger rejection embedded
  in
• online photon veto
• (2)Cacc over-rejection by photon veto
  with accidentals
• (3)Csplit self-vetoing effect by splitting
  tagging photon

CL1.1after 1.14
Cacc 0.80
43
High Purity Kp2 Identification
Dominant non-Kp2 backgrounds
(1) background rejection
(2) Single beam background rejection
(3) Two-beam background rejection
44
Top half of side view
m
n
Km2 backgrounds
45
High Purity Kp2 Identification
Dominant non-Kp2 backgrounds
(1) background rejection
(2) Single beam background rejection
(3) Two-beam background rejection
46
Top half of side view
p
B4
Cerenkov
Single beam backgrounds
47
High Purity Kp2 Identification
Dominant non-Kp2 backgrounds
(1) background rejection
(2) Single beam background rejection
(3) Two-beam background rejection
48
Beam wire chamber
Beam 1
Top half of side view
Beam 2
p
K
veto
veto
veto
Two-beam backgrounds
49
E949 Detector
E949 detector side view (upper half)
E949 detector end view (upper half)
Target
(1) Target Kaon decay at rest (2) Drift
chamber Momentum (3) Range Stack
(scintillator) Energy / Range
50
Error distribution w/ Daughter Table Method
w/ Binominal error
Convoluted inefficiency
Daughter tables produced by random number
generator
300 daughter tables
51
DAQ Summary
Platinum target used in 2002
of accumulated Kaons
Before data taking
After data taking
Accumulated K
52
Data Acquisition
Platinum target used in 2002
After data taking
Before data taking
Accumulated K
53
Single Photon Inefficiency
54
Subtraction at various levels of photon veto
p0 ? gg rejection at various photon veto
Improvement (Before/After)
A factor of two improvement at various photon
veto
search
Estimation from photon inefficiency
55
Disruption Correction Factor Cdis
Overlapping g,e/- (from p0) may cause disruption
in the p track reconstruction.
Estimation (Pure MC Study)
(1) Normal Kp2 decays
(2) Kp2 decays but was forced.
Difference in the p recon. efficiency ?
correction
Disruption correction