Final measurement of e

1
Final measurement of e/e by NA48
G.Unal, LAL-Orsay On behalf of the NA48
Collaboration Cagliari,Cambridge,CERN,Dubna,Edinb
urgh,Ferrara,Firenze, Mainz,Orsay,Perugia,Pisa,Sac
lay,Siegen,Torino,Vienna,Warsaw

• Direct CP violation in neutral kaon decays
• History of the e/e measurement by NA48
• Analysis of the 2001 data sample
• Final result and conclusions

2
CP violation in neutral kaon decays

• CP violation in mixing ( indirect )
• KL K2 e K1 KS K1 e K2
• (K1,K2 CP eigenstates) e
  (2.280.02)10-3
• CP violation in p p decay
• A(KL?pp-)/A(KS?pp-) e e
• A(KL?p0p0)/A(KS?p0p0) e 2 e
• e  direct  CP violation (interference
  between I0 and I2 amplitudes)
• Standard Model both e and e
• Quantitative predictions difficult e/e ? (-10
  to 30) 10-4

3
The double ratio R

• need to measure small deviation of R from 1
• reduce to event counting if at least 2 modes
  taken simultaneously

• NA48 method
• take the 4 modes
• simultaneously (?cancellation of dead time,
  inefficiencies, )
• from same decay region
• KL events are weighted to have same decay
  distribution as KS (?minimise detector
  acceptance correction)
• high resolution detectors ?minimise residual
  backgrounds
• KS/KL identification by  tagging  the proton
  creating the KS

4
The Simultaneous KS and KL beams
not to scale
5
The NA48 detector
p0p0 detection ( ?4 g) LKr calorimeter
s(E)/E0.032/?E?0.09/E?0.0042 lt 1
for E25 GeV
pp- detection magnetic spectrometer s(p)/p
0.5?0.9?(p/100 GeV)
6
History of the e/e measurement by NA48
Year 1997 1998
1999 2000 2001 N(KL?2p0)
0.5 M 1 M 2 M
- 1.5 M
improvements in LKr, triggers, DAQ,efficiency
Beam tube implosion Cross-checks with
neutrals (rare Ks decays)
Drift chambers rebuilt Lower beam intensity
e/e (15.3?2.6) 10-4 (published in 2001)
Result today !
7
Summary of uncertainties on R for 98-99 data
pp- trigger ? 5.2
KL accidental tagging as KS ? 3.4
KS tagging inefficiency ? 3.0
Accidental activity ? 4.2
KS in time activity ? 1.0
p0p0 reconstruction ? 5.8
pp- reconstruction ? 2.8
p0p0 background ? 2.0
pp- background ? 3.0
beam scattering bkg ? 2.0
Acceptance (stat) ? 4.1
(syst) ? 4.0
AKS inefficiency ? 0.4
related to beam intensity effects
smaller uncertainties for 2001
data Example pp- trigger efficiency increased
from 97.78 to 98.70
smaller uncertainty in efficiency measurement
8
1998-1999 spill2.4s every 14.4s 2001
spill5.2s every 16.8s gt 80 higher duty
cycle Instantaneous beam intensity reduced by
?30 Dead time in drift chamber readout 20 ?
11 (this dead time condition is recorded and
applied in the analysis to all events)
9
2001
1998-1999
slightly better efficiency Performance
s of rebuilt DCH similar pp- mass
resolution
10
Analysis of the 2001 data sample
(some selected topics)
p0p0
pp-
11
p0p0 reconstruction
Decay region definition ? Energy scale 1 cm on
decay vertex ? 10-4 on E-scale
D(vertex-LKr) 1/M(K) ??Ei Ej dij (from K mass
constraint)
Cross-check use p0p0 hadronic production from p-
beam striking two thin targets during special runs
Adjust E-scale to reproduce nominal position of
AKS
KS?p0p0
Total uncertainty (E-scalenon linearities) ?
5.3 10-4 on R
12
Acceptance
Weight KL events to equalize decay vertex
distribution and make detector acceptance the same
Residual correction (beam geometry) D(R)
(21.9?3.5?4.0) 10-4 Does not rely on detailed
detector simulation
13
KS-KL identification
R sensitive to pp- p0p0 differences DaSL
(0?0.5)10-4 ? DR (0?3)10-4 DaLS
(3.4?1.4)10-4 ?DR (6.9?2.8)10-4
From vertex-identified pp- aSL(1.120.03)10-4
(KS tagging inefficiency) aLS(8.1150.010) (KL
accidental tagging)
Higher losses related to beam intensity in
pp- Can be predicted using overlay technique
Time in spill (s)
14
Accidental effects

• Accidental effect?event losses induced by (KL)
  beam activity
• Miminised by simultaneous data collection in 4
  modes
• DR D(p0p0-pp-) D(KL-KS)
• D(p0p0-pp-) minimised by applying to all events
  the recorded
• dead time conditions
• main tool overlay pp events with random
  events ( ? beam intensity)
• D(KL-KS) small by design of the experiment
• simultaneous beam
• KS and KL decays see the same beam
  intensity
• deviation  intensity difference
  effect 
• lifetime weighting
• KS and KL decays illuminate the same
  part of the detector
• residual effect  illumination
  difference effect 

15
Event losses (from overlay) D(pp--p0p0)(1.00.5
) for 2001 data Better check of linearity of
losses in 2001 (better beam monitors)
Measure accidental activity ( ? KL beam
intensity) in KS and KL events ? DI/I (01)
16
Uncertainties from accidental effects

• Intensity difference effect
• from estimates of D(pp--p0p0) and DI/I
• DR 1.1 10-4
• (was 3 10-4 for 98-99 data)
• Illumination difference effect
• overlaying  random  events to KS and KL
  decays
• DR 3.0 10-4
• (limited by statistical uncertainty of
  overlay sample)

17
Summary of corrections and uncertainties on R for
2001 data
in 10-4 units
pp- trigger inefficiency 5.2  3.6 (stat)
p0p0 reconstruction 5.3
pp- reconstruction 2.8
p0p0 background - 5.6 2.0
pp- background 14.2 3.0
Beam scattering background - 8.8 2.0
KS tagging inefficiency 3.0
KL accidental tagging as KS 6.9 2.8 (stat)
Accidental activity intensity difference 1.1
illumination difference 3.0 (stat)
KS in time activity 1.0
Acceptance correction 21.9 3.5 (stat)
4.0
AKS inefficiency 1.2 0.3
Total 35.0 11.0
18
Cross-checks of the stability of the result
2001 data
published 98-99 data
19
The double ratio result (2001 data)
The analysis is performed in Kaon energy bins to
be insensitive to KS-KL differences in energy
spectra
R 0.99181 0.00147 stat 0.00110 syst
20
Final result
From 2001 data e/e (13.72.51.8)10-4
(13.7 3.1) 10-4 in very good agreement
with 97-98-99 published result e/e
(15.32.6)10-4 Final combined result from
NA48 e/e (14.7 2.2)10-4 (correlated
systematic uncertainty is 1.4 10-4)
21
e/e vs data taking year
22
Comparison of experimental results
NA31 (23.06.5)10-4 E731 (7.45.9)10-4 KTeV
(20.72.8)10-4 (preliminary) NA48
(14.72.2)10-4
World average e/e (16.61.6)10-4 c26.2/3
(prob10)
23
Conclusions
NA48 measurement of e/e is completed
e/e (14.72.2) 10-4
proposed accuracy is reached papers - V.Fanti
et al, Phys. Lett. B465, 335(1999) 97 data
result - A.Lai et al, Eur. Phys.
Jour.C83,22(2001) 98-99 data -
coming out soon on 2001 data KTeV still to
analyse 1999 data (?same stat as 96-97 data) Kloe
with different method (need luminosity) The
ball is now on the theory side