osk meeting - indico.cern.chโฌยฆย ยท ๐พ distribution 15 ๐ ๐ ๐๐ ๐ ๐ธ ๐ ๐ธ...
TRANSCRIPT
2019 8/2
1
Search for ๐ฒ๐ณ โ ๐ ๐๐ธ decay in the KOTO experiment
Nobuhiro Shimizu Sep. 9th
๐ฒ๐ณ โ ๐ ๐๐ธ decay?
โข No measurements up to nowโ ๐ โ ๐๐พ transition, forbidden by FCNC
โ Violates CP
โข neither for ๐ฒ๐บ โ ๐ ๐๐ธโข Nice prove of New Physics??
โข violates an angular conservationโข ๐ฒ and ๐ = spin zeros: the spin of ๐ธ must be โcompensatedโ by
an orbital angular momentum, but back-to-back configuration cannot produce ๐ฟ๐ง (along the decay axis).
โข IF v > โspeed of lightโ, itโs allowed. (see PRD 59 116008)
2
Oppositely to say, good test of the Lorentz invariance in the realm of short distances
New Physics search via ๐ฒ๐ณ โ ๐ ๐๐เดค๐โซโฌSM ๐พ๐ฟ โ ๐0๐ าง๐ = (3.0 ยฑ 0.3) ร 10โ11
โซโkeyโ๏ผ not to miss any photonsHermetic Veto counters surrounding the decay volume
The KOTO experiment
3
JHEP 1511 033 (2015)
๐ธ
๐ธ
เดฅ๐๐
โ High intensity ๐พ๐ฟ beam๏ผ low inefficiency KOTO detector is ideal tool to search for ๐ฒ๐ณ โ ๐ ๐๐ธ
KOTO detector 4
Measure energy and position of photons by undoped CsIcalorimeter
Decay volume: surrounded by 18 types of veto detectors.
Two barrel counters to veto photons: main barrel (MB) and inner barrel (IB)
Charged veto in front of CsI
Data taking
Data taken during 2016 โ 2018: โผ 19 ร 1018 Protons On Target (POT)
Trigger: 3-cluster w/ energy deposit at CsI calorimeter
โ Data corresponding to ๐. ๐ ร ๐๐๐๐ POT was analyzed.
5
Inner barrel installation(2016, June)
Inner barrel
Main barrel
Reconstruction of ๐ฒ๐ณ โ ๐ ๐๐ธ
1. Find events which have exactly 3 clusters
2. Reconstruct a ๐ ๐ : ๐๐02 = 2๐ธ1๐ธ2 1 โ cos๐๐พ๐พ โ ๐๐๐๐
๐ ๐
3. Reconstruct a ๐ฒ๐ณ: ๐๐พ๐ฟ2 = ๐๐พ1 + ๐๐พ2 + ๐๐พ3
2โ ๐๐๐๐
๐ฒ๐ณ
7
๐ ๐๐ฒ๐ณ beam๐ธ
Two types of vertex position, which should be close.
4. Define ๐ซ๐๐๐๐ = ๐๐๐๐๐ ๐ โ ๐๐๐๐
๐ฒ๐ณ to suppress various BGs
clusters
๐
Event selections 9
1. Online trigger (veto and cluster counting)
2. Requirement of ๐ต๐๐๐ == ๐ offline
3. Vetoes
4. Cluster shape discrimination (CSD)
5. Kinetic cuts
โ๐ซ๐๐๐๐ = ๐๐๐๐๐ ๐ โ ๐๐๐๐
๐ฒ๐ณ
โ ๐ฌ๐ธ๐๐๐
6. Signal box1. (๐๐๐๐
๐ ๐ , ๐ด๐ ๐๐ธ)
Possible background sources
โข ๐พ๐ฟ โ 2๐0
1. a photon missed by some way
2. two photons fuse
โข ๐พ๐ฟ โ 3๐0
mechanisms similar to 2๐0
โข Neutron induced BG๐0 produced from n and a single acc. cluster
โข ๐พ๐ฟ โ 2๐พ
combines with neutrons
โข ๐พ๐ฟ โ ๐+๐โ๐0
combines with neutrons
โข ๐พ๐ฟ โ โ๐๐ (โ = ๐ or ๐)
e.g., ๐+๐detectorโ โ 2๐พ and ๐โ missed by CV
10
๐ธ ๐ธ ๐ธ ๐ธร๐ธ ๐ธ ๐ธ ๐ธ
๐ธ ๐ธ ๐ธ ๐ธ ๐ธ ๐ธร
๐ธ ๐ธ ๐
๐ธ ๐ธ ๐ ๐ +๐ โรร
Large
1
<1
<0.01
<0.01
<0.001
<0.01
Veto counters
11
๐๐ ๐
๐๐๐ญ๐
Main barrel veto energy
๐๐ ๐
๐๐๐ญ๐
Inner barrel veto energy
Main barrel
Inner barrel
Two barrel veto counters play crucial roles to suppress ๐พ๐ฟ โ 2๐0 and ๐พ๐ฟ โ 3๐0 decay modes.
Cluster shape discrimination (CSD) 12
โ neutron clusterโก fused cluster
minimum CSD (the smallest one out of 3)
๐๐ ๐๐๐ ๐๐๐๐ญ๐
๐๐ธ
๐ ๐๐ธ(scale is arbitrary)
๐ธ likenot ๐ธ like
Using neural network, the shower shape development in the CsI calorimeter is evaluated.
โ Separate clean photon clusters from those of BG
Energy deposition of photon cluster
MC Data
๐ซ๐๐๐๐ distribution 13
Under vetoes, CSD
โ๐๐๐ < ๐ซ๐๐๐๐/mm < ๐๐๐
One of the important cut to reduce 2๐0 decay
๐ง๐ฃ๐ก๐ฅ๐0๐ง๐ฃ๐ก๐ฅ
๐พ๐ฟ๐พ
๐พ
๐พ๐ฟ
๐ซ๐๐๐๐
๐ธ
๐0
๐๐ ๐๐๐ ๐๐๐๐ญ๐
๐ ๐๐ธ(scale is arbitrary)
In the signal case ๐ซ๐๐๐๐~๐
๐ด๐ ๐๐ธ and ๐๐๐๐๐ ๐ distributions 14
๐๐ ๐๐๐ ๐๐๐๐ญ๐
๐ ๐๐ธ(scale is arbitrary)
Under vetoes, CSD, ๐ซ๐๐๐๐ cuts
๐๐ ๐๐๐ ๐๐๐๐ญ๐
๐๐ธ
๐ ๐๐ธ(scale is arbitrary)
๐ด๐ ๐๐ธ ๐๐๐๐๐ ๐
These variables are used for defining the signal box.
๐๐๐ < ๐ด๐ ๐๐ธ/๐๐๐ < ๐๐๐, ๐๐๐๐ < ๐๐๐๐๐ ๐ /mm < ๐๐๐๐
๐ธ๐๐๐๐พ
distribution 15
๐๐ ๐๐๐ ๐๐๐๐ญ๐
๐๐ธ
๐ ๐๐ธ(scale is arbitrary)
Under vetoes, CSD, ๐ด๐ฒ๐ณ, ๐ซ๐๐๐๐ cuts
๐ฌ๐๐๐๐ธ
Under ๐ด๐ฒ๐ณcut,
๐ฌ๐๐๐๐ธ
is useful
variable to remove ๐๐ ๐.
To measure ๐๐ ๐ and
๐๐ ๐ a loose ๐ฌ๐๐๐๐ธ
cut is
also used.
loose
nominal
๐ธ๐๐๐๐พ
: minimum photon energy
Signal box
16
Signal MC
Signal region
Control region 2(CR 2)๐๐ ๐ dominant
Control region 1 (CR 1)3๐ ๐ and 2๐ ๐
โ โฌ(๐พ๐ฟ โ ๐0๐พ) is obtained as a ratio towards โฌ(๐พ๐ฟ โ 2๐0)โ CR2 is used to normalize statistics/estimate ๐๐ ๐ BGโ Signal region (SR) is blinded before determining the selection critaria to avoid
human bias.
Under the condition of 1. vetoes, 2. CSD, 3. ๐ซ๐๐๐๐,
4. ๐ฌ๐๐๐๐ธ
cuts
Signal box, defined in the 2D plane:
(๐๐๐๐๐ ๐ , ๐ด๐ ๐๐ธ)
Summary of the event selection 17Uncertainties are statistical only.
Step (after)
Signal ACC
(๐๐โ๐)๐๐ ๐ ACC(๐๐โ๐)
Nobs (Data)
Ncls==3 and online veto*
219 31 1150585
Shape ๐๐ 204 24 233568
Fiducial cut 189 22 214043
Veto 71 2.0 15732
Cluster shape 51 1.3 9483
๐๐๐๐๐ 49 0.41 2544
(loose) ๐ฌ๐๐๐๐ธ (24.4) 5.0 (0.1) 0.0041 (635) 36
loose CR2 ๐. ๐๐๐๐ ยฑ ๐. ๐๐๐ 528
SR ๐. ๐๐ ยฑ ๐. ๐๐ ๐. ๐๐ ยฑ ๐. ๐๐ ร ๐๐โ๐ ?
Single event sensitivity of signal: ๐
๐ต๐ฒ๐ณ๐ ๐ฒ๐ณโ๐ ๐๐ธ
= ๐. ๐ ร ๐๐โ๐
Z-๐๐0๐พ correlation plot under loose cut 18
MCDataAfter applying veto, CSD, ๐ซ๐๐๐๐, loose min๐ฌ๐ธ cut ๐ฒ๐ณ โ ๐๐ ๐
๐ฒ๐ณ โ ๐๐ ๐
๐ฒ๐ณ โ ๐ ๐๐ธ๐ธ
normalize
We use CR2 region with loose min๐ธ๐พselection to measure ๐พ๐ฟ โ 2๐0
contribution of data.
Control region 2 (CR 2)
๐๐ ๐๐๐ ๐๐๐๐ญ๐
๐๐ธ
๐ ๐๐ธ(scale is arbitrary)
loose
After applying veto, CSD, ๐ซ๐๐๐๐, cut
min๐ฌ๐ธ
Z-๐๐0๐พ correlation plot of under tight cut19
The expected number of events in the signal region is 0.34 ยฑ 0.10
MCData๐ฒ๐ณ โ ๐๐ ๐๐ฒ๐ณ โ ๐๐ ๐
๐ฒ๐ณ โ ๐ ๐๐ธ๐ธ
BG contribution (summary) 20
Sources Expected in signal box
Comment
๐ฒ๐ณ โ ๐๐ ๐ ๐. ๐๐ ยฑ ๐. ๐๐ Evaluated by MC
๐ฒ๐ณ โ ๐๐ ๐ <0.5 (68% C.L) Data driven estimation
๐ฒ๐ณ โ ๐๐ธ << 0.06 (68% C.L) Evaluated by full MC
Neutron induced BG <<0.02 (68% C.L) Evaluated by dedicated MC
Other ๐พ๐ฟ decays <0.02
Sum 0.85 (<1.0 at 68% C.L.)
Dominant background: the ๐ฒ๐ณ โ ๐๐ ๐ decay due to inefficiency of the barrel veto
๐ฒ๐ณ โ ๐๐ ๐ decay can be BG due to โ production of the overlapped ๐พ clusters (fusion)โก wrong paring of ๐ ๐ โ ๐๐ธ conbinations.Production of the sufficient amount of MC sample is difficult,but data driven approach gives an upper limit of 0.5 at 68% C.L.
Other decays โ negligible.
Systematic uncertainty of S.E.S 21
โฌ ๐พ๐ฟ โ ๐0๐พ =๐๐๐๐ (๐๐ )
๐2๐0๐๐๐ (๐ถ๐ 2) โ
๐2๐0๐ถ๐ 2
๐๐0๐พ๐๐ โ โฌ ๐พ๐ฟ โ 2๐0
Source ๐๐/๐ (in %) Commentโฌ ๐พ๐ฟ โ 2๐0 0.6 From PDG
Geometry 1.5 Estimated by varying beam E (+1%) and position (x,y=1mm) for signal
Veto cuts 17 By comparing data and MC in CR2 with ๐ซ๐๐๐๐ cut
Online veto 6.4 From the detector bits in the minimum bias data
Kinematic cuts 12 100% error of |1-๐๐ท๐๐ก๐2๐0 /๐๐๐ถ
2๐0|
Clustering 1.0 Compare five and six cluster events
CSD cuts 1.5 Use five cluster events
Reconstruction 0.3 By comparing ๐๐๐๐ 3๐0๐๐๐ 6๐๐๐ and ๐๐๐๐ 2๐0
๐๐๐ 6๐๐๐
Trigger 1.8 Difference of CDT efficiency from unity
Statistics 4.4 Statistics of normalization
Total 22
Syst. uncertainty due to ๐๐ฃ๐๐ก๐ 22
๐๐ฃ๐๐ก๐ =๐๐๐๐
๐๐๐๐ ๐ค๐/ ๐๐กโ ๐ฃ๐๐ก๐is compared
between data and MC: ๐ = ๐๐ถ๐ 2๐ท๐ /๐๐ถ๐ 2
๐๐ถ
๐๐๐ฃ๐๐ก๐
๐๐ฃ๐๐ก๐=
๐:๐๐๐ ๐๐๐ก๐๐๐ก๐๐๐
๐ ๐ โ 1 2 = 17%
MC
๐๐๐ญ๐
Inefficiency of veto detector is well understood by MC.
Open signal box 23
Single event sensitivity of signal:๐
๐ต๐ฒ๐ณ๐ ๐ฒ๐ณโ๐ ๐๐ธ
= ๐. ๐ ยฑ ๐. ๐๐ฌ๐ญ๐๐ญ ยฑ ๐. ๐๐๐๐๐ ร ๐๐โ๐
Open signal box 24
Single event sensitivity of signal:๐
๐ต๐ฒ๐ณ๐ ๐ฒ๐ณโ๐ ๐๐ธ
= ๐. ๐ ยฑ ๐. ๐๐ฌ๐ญ๐๐ญ ยฑ ๐. ๐๐๐๐๐ ร ๐๐โ๐
Open signal box 25
โฌ ๐พ๐ฟ โ ๐0๐พ < 1.7 ร 10โ7 at 90% C.L.
Taking into account the systematic uncertainty, the obtained upper limit is
Summary Neither ๐ฒ๐ณ โ ๐ ๐๐ธ nor ๐ฒ๐บ โ ๐ ๐๐ธ decay have been measured yet.
These decays violate the Lorentz invariance and are forbidden by the SMโ Lorentz invariance should be tested also in the high energy regime.
Studying backgrounds, we fixed the selection criteria withS.E.S = 6.8 ร 10โ8. The expected number of background is 0.3.
We opened the signal box and could not find any signal candidate.
The upper limit of the branching fraction is obtained for the first time!
26
โฌ ๐พ๐ฟ โ ๐0๐พ < 1.7 ร 10โ7 at 90% C.L.