Rare kaon decays at NA62 Evgueni Goudzovski (University of - - PowerPoint PPT Presentation
Rare kaon decays at NA62 Evgueni Goudzovski (University of Birmingham) on behalf of the CERN-NA62 collaboration Outline: 1) The NA62 experiment at CERN 2) Measurement of the K + + decay 3) Searches for lepton number/flavour
CLFV 2019 conference Fukuoka, Japan 18 June 2019
(University of Birmingham)
1) The NA62 experiment at CERN 2) Measurement of the K++ decay 3) Searches for lepton number/flavour violating K+ decays 4) Searches for HNL production in K+ decays 5) Summary
SPS NA48/NA62:
ECN3 hall
Jura mountains Geneva airport
France Switzerland
LHC N NA48
1997: ’/: KL+KS 1998: KL+KS 1999: KL+KS
KS HI
2000: KL only
KS HI
2001: KL+KS
KS HI
2002: KS/hyperons 2003: K+/K− 2004: K+/K−
tests
NA62
2007: K
e2/K 2
2008: K
e2/K 2
discovery
CPV
Earlier: NA31
NA62
RK run
NA48/1 NA48/2
tests
2015: commissioning
201618: physics run
Main NA62 goal: K++ measurement to 10% precision with a novel decay-in-flight technique. Currently ~200 participants from 31 institutions.
Un-separated hadron (p/+/K+) beam. SPS protons: 400 GeV , nominally 3.3×1012/spill. K+: 75 GeV/c (±1%), divergence < 100 rad. Nominal beam rate: 750 MHz, K+ rate 45 MHz; ~5 MHz K+ decays in fiducial volume
KTAG: Cherenkov kaon tagger, t=70ps
Anti-counters GTK: beam tracker Spectrometer: straw chambers
LAV: large-angle photon veto (12 stations) Muon detector (MUV) Small-angle
photon veto
LKr EM calorimeter
Dump
Z [m]
NA62 collaboration, JINST 12 (2017) P05025
t=70ps Hadronic Calorimeter (HAC)
O(106) mbar
Currently, 1 year of operation 2×1018 protons on target; 3×1012 K+ decays. Single event sensitivities for K+ decays: down to BR~1012. Kinematic rejection factors: 1×103 for K++0, 3×104 for K+. Hermetic photon veto: 0 decay suppression (for E0>40 GeV) = 3×108. Particle ID (RICH+LKr+HAC+MUV): ~108 muon suppression.
300 m3
Commissioning run 2015: minimum bias data (~3×1010 protons/pulse). Physics run 2016 (30 days, ~1.3×1012 ppp): 1011 useful K+ decays. Physics run 2017 (161 days, ~2.0×1012 ppp): ~2×1012 useful K+ decays. Physics run 2018 (217 days, ~2.3×1012 ppp): expect 3×1012 useful K+ decays. Resuming data taking after Long Shutdown 2 in 2021.
Result based on the 2016 data:
Theoretically clean, almost unexplored, sensitive to new physics. Mode BRSM1011 K++() 8.41.0 KL0 3.40.6 SM precision surpasses any other FCNC process involving quarks. Measurement of |Vtd| complementary to those from e.g. BB mixing. Main focus of kaon physics: measurement
The uncertainties are largely parametric (CKM) SM branching ratios Buras et al., JHEP 1511 (2015) 033
SM: box and penguin diagrams Ultra-rare decays with the highest CKM suppression:
A ~ (mt/mW)2|V tsVtd| ~ 5
*
CKM unitarity triangle with kaons
Kaon measurements alone
can fully constrain the unitarity triangle. Complementary to B physics in the description of NP flavour dynamics.
NP with CKM-like flavour structure
BR(KL0) vs BR(K++)
Current experimental uncertainty (1)
mmiss
2=(PKP)2 vs track momentum
Region II
Region I
K+++ K++00 K++0 K++
Further background suppression:
PID (calorimeters & Cherenkov detectors): suppression 108. Hermetic photon veto: suppression of 0 decays 3×108.
Main K+ decay modes (>90% of BR) rejected kinematically. Resolution on mmiss
2:
=1.0103 GeV4/c2. Measured kinematical background suppression: K++0: 1×10−3; K++: 3×10−4.
DATA 2016
Blinded Region 2 Blinded Region 1
Data sample: 30 days at 1.3×1012 ppp. Number of kaon decays: NK = (1.210.02syst)×1011. Background estimates are mostly data-driven. Signal acceptance: A = (4.00.1)%. Single-event sensitivity: SES = (3.150.24)×1010.
K+++ region K++0 region K++ region 2016 data
PLB791 (2019) 156
One K++ candidate observed: BR(K++) < 11×1010 at 90% CL. BNL-E949 (K+ decay at rest): BR(K++) = (1.73 )1010 SM prediction: BR(K++) = (0.840.10)×1010
+1.15 1.05
The NA62 decay-in-flight technique works. A non-trivial result obtained with 2% of the total statistics collected.
x [mm] y [mm]
The RICH ring for the track e
2016 data PLB791 (2019) 156 K+++ region K++0 region K++ region
Analysis of the 2017 data Data sample: 161 days at 2.0×1012 ppp. Analysis procedure is similar to the 2016 one. Number of kaon decays: NK = (1.30.1)×1012. Single-event sensitivity: SES = (3.40.4)×1011. Expected signal: 2.50.4 SM K++ events. Expected background: 0.760.10 events, excluding upstream decays. Result expected later this year, surpassing present best sensitivity. Analysis of the 2018 data The largest sample collected so far: 217 days at 2.3×1012 ppp. Analysis procedure being optimized to improve acceptance. Further data collection necessary to reach the 10% precision Data taking will resume after LS2 (in 2021). Developing a strategy to collect 100 SM events by 2024. Possible beam dump operation (3 months of data taking = 1018 pot): competitive searches for hidden sector (long-lived HNL, DP , ALP).
(CERN SPSC open session, 2 April 2019)
NA62 precision by 2025
(with 100 SM events)
KOTO sensitivity goal following beam & detector upgrades (~2025?)
KOTO 2015 data:
BR<3.0×109 @90% CL PRL122 (2019) 021802
KOTO sensitivity with 201518 data. Expect result in 2019.
Downscaled di-lepton trigger chains are in operation. First results based on 80% of the 2017 data: CERN-EP-2019-104, arXiv:1905.07770, submitted to PLB. Analyses completed so far: searches for K+ℓ+ℓ+ decays.
Major background for 3-track decays: the K+++ decay (BR=5.6%). Studied with data-driven methods and dedicated simulations. Background to decays into leptons: 1) via decays in flight; 2) via misidentification as e. Pion/electron identification: 1) by energy deposit in LKr (E/p); 2) by the RICH signal pattern. Pion mis-ID probability vs momentum
LNV selection: m(e+e+) SM selection: m(+e+e)
Signal region
K+++ K+++ K++e+ K++e+
Auxiliary selection: LKr only is used for pion/electron ID Validation of the background estimates using control mass regions. Sensitivity is limited by K++0
D background.
Therefore RICH is used for positron ID (10% loss of SES, 6 times lower background)…
K++0
D
LNV selection: m(e+e+) SM selection: m(+e+e)
Signal region
K++e+ Candidates observed: 2484 BR(K++e+e) = (3.000.09)×107 K+ decays in FV: (2.140.07)×1011 SES = (0.940.03)×1010 Expected background: 0.160.03 evt Candidates observed: 0 K+[e+e]0e+
Result: BR(K+e+e+)<2.2×1010 at 90% CL
LNV selection: m(++) SM selection: m(++)
Candidates observed: 8357 Background: 0.07% BR(K+++) = (0.9620.025)×107 K+ decays in FV: (7.940.23)×1011 K+++ K+++ K+++ K+++ SES = (1.280.04)×1011 Expected background: 0.910.41 evt Candidates observed: 1
Signal region Signal region
BR(K+++)<4.2×1011 at 90% CL
Backgrounds to K+++ decay
Data-driven estimates (inversion of PID criteria) Biased fast simulations with forced pion decays: dominated by pion decays in the spectrometer Full MC simulations tuned using measured misID probabilities and measured pileup rates
Final result with the 2017 data set:
One candidate observed
Single-event sensitivities achieved with the 2017 data For K++e+ [LNV] and K++e+ [LFV]: SES5×1011 (factor ~5 improvement on BNL-E865). For K+e++ [LFV], SES5×1011 (the first search for this mode); For K+e+e+ [LFV], SES1×1010 (factor 100 improvement on PDG). The full 201618 dataset is ~3 times the size of 2017 dataset. Not competitive yet for K+++e and 0e∓.
Six analyses are in progress, none is limited by background
ULs obtained with 80% of the 2017 data set: (arXiv:1905.07770)
(at 90% CL, assuming uniform phase space distribution). Factor 23 improvement on the earlier results (BNL-E865 and NA48/2).
Results based on the pilot 2015 data:
Neutrino minimal SM (MSM) = SM + 3 right-handed neutral heavy leptons.
[Asaka et al., PLB631 (2005) 151]
Masses: m1~10 keV [DM candidate]; m2,3~1 GeV. HNLs observable via production and decay.
Shaposhnikov, JHEP 0808 (2008) 008 Boyarsky et al., Ann.Rev.Nucl.Part.Sci.59 (2009) 191
|Uℓ4|2
Baryon asymmetry of the Universe Big-Bang nucleosynthesis
Accessible in K+ℓ+N decays
PLB96(1980)159
m1 [keV] m2,3 [GeV] Astrophysical & cosmological constraints on m1, m2,3
(K+ℓ+N) = (K+ℓ+) ℓ(mN)|Uℓ4|2 (mN)
RK = (K+e+)/ (K++) 2.5×105
e(mN)×RK
Kinematic enhancement factor
Minimum bias data (1% intensity); 12k SPS spills (=5 days) in 2015. Numbers of K+ decays in fiducial volume: NK=(3.010.11)×108 in positron case; NK=(1.060.12)×108 in muon case. Beam tracker not available: beam average kaon momentum is used. HNL production signal: a spike above continuous missing mass spectrum. Squared missing mass: (PKPe)2
K++, +e+ K+e+, BR=1.6×105:
1.7k candidates
HNL search region: low background due to photon veto and kaon ID
Squared missing mass: (PKP)2
K++, BR=64%:
24M candidates
HNL search region
HNL mass resolution m vs mass
Selection for each HNL mass hypothesis (mHNL ) includes the “mass window” condition: |mmHNL|<1.5m: background is proportional to mass resolution. Also, resolution is crucial to resolve possible HNL mass splitting.
[Baryogenesis: 2 quasi-degenerate mass states; Canetti et al., PRD87 (2013) 093006]
Signal selection acceptance vs mass
Relaxed selection at high mN Standard selection
Expected background (and stat.error) estimated from fits to the sidebands. Numbers of observed and expected events converted into limits for the signal. Background simulations used to certify the absence of peaking structures. Full MC background estimate would allow searches for K+ℓ+.
Upper limits on BR(K+ℓ+N)
90%CL
Local signal significance never exceeds 2.2: no HNL signal is observed. Reached 106107 limits for |Uℓ4|2 in the 170448 MeV/c2 mass range.
|Uℓ4|2 limits from production searches
Expected Observed
Data sample 201618 in comparison to data sample 2015: Beam tracker (GTK) in operation: a factor ~2 improved HNL mass resolution m,
therefore lower background and broader mass range accessible;
a factor ~3 lower background in the K+e+N mode (K++, +e+: muon decays in flight rejected geometrically); lower background from upstream decays in the K++N mode. Much larger datasets: In the K+e+N mode, the main K++ trigger is used (with reduced signal acceptance: max calorimetric energy = 30 GeV): expect at O(106) K+e+ events, i.e. a factor ~1000 improvement. In the K++N mode, downscaled control trigger (D=400): expect O(109) K++ events, i.e. a factor ~100 improvement. Expected sensitivities to |Uℓ4|2 with 201618 data: at least 108 for both |Ue4|2 and|U4|2
CERN-NA62 collected a sample of ~5×1012 K+ decays in flight during the first physics run in 201618. Focused on the K measurement (currently SES~1011). A broader programme of rare decay measurements, LF/LN violation and hidden sector mediator searches is pursued. Analysis of the 2016 data (~2% of the main data set): BR(K++) <11×1010 at 90% CL. [PLB791 (2019) 156] Analysis of the 2017 data (~30% of di-lepton data sets): BR(K+e+e+)<2.2×1010, BR(K+++)<4.2×1011 at 90% CL. [arXiv:1905.07770] Analysis of the 2015 minimum bias data: sub-106 limits on |Uℓ4|2 from HNL production searches. [PLB778 (2018) 137] All results to be updated/expanded soon with larger data sets.
Events Events K++0() background vs + momentum K++() background vs + momentum K++() K++0() K++ K++
K++ acceptance vs + momentum +/+ separation (RICH) +/+ separation
(calorimeters)