KS p pp0 CPconserving part

1
KS ? pp-p0CP-conserving part

• an easy analysis?
• Manfred Jeitler

2
overview

• the approach
• the details
• the magnetic field problem
• cowboys, sailors, and azimuth
• .p momentum distributions
• drift chamber illumination
• how to get a result?

3
The approach
4
the approach

• different from KL ? pp and KS ? p0p0p0
• those are purely CP-violating decays
• KS ? pp-p0 has CP-conserving and CP-violating
  part
• L1 is CP 1 (CP-conserving), has
  centrifugal barrier
• L0 is CP -1 (CP-violating)
• distinguish them by Dalitz-plot distribution

5
the approach

• decay KS ? pp-p0 (CP-conserving part) is
  described by complex parameter l
• branching ratio can be derived from l
• extract CP-conserving part by taking difference
  between distributions of Xgt0 and Xlt0 events
• X Dalitz plot variable

6
relative excess ?
the function to fit to the data
V-0 for literature values dilution
0.28 Re (l) 0.031 Im (l) - 0.006
ct in units of KS lifetimes ?
7
V-0 dilution 0.28 Re (l)
0.031 varying the imaginary part -0.1 lt Im (l)
lt 0.
relative excess
ct in units of KS lifetimes
8
the original Dalitz plot
p- kinetic energy in CMS
p0 kinetic energy in CMS
kinematically allowed region
p kinetic energy in CMS
9
the usual Dalitz plot variables

• si ... the invariant mass squared of the other
  two particles the higher the si, the smaller
  the kinetic energy of that (i-th) pion in
  the CMS system
• X ... the difference between the si of the p
  and the p- (in units of mp?2)
• Y ... the difference between the si of the p0
  and the mean
• of all the si (in units of mp?2)

10
ct spectra for the various kaon energy
ranges dilution (K0 / K0 ratio) is energy
dependent ? fit separately
11
distinguish CP-conserving from CP-violating
distribution is odd in X for CP-conserving
even
violating
X gt 0
X lt 0
Y
X
12
30 lt EK lt 47 GeV
Dalitz X gt 0
Dalitz X lt 0
fit ct distribution for Xgt0 - Xlt0 Xgt0 Xlt0 to
extract the l parameter and branching ratio
difference
difference normalized to sum
13
98 lt EK lt 115 GeV
Dalitz X gt 0
Dalitz X lt 0
fit ct distribution for Xgt0 - Xlt0 Xgt0 Xlt0 to
extract the l parameter and branching ratio
difference
difference normalized to sum
14
(No Transcript)
15
c2 surface (data, using standard cuts) Re l
and Im l are strongly correlated ! fit
converges nicely
? c2
0.1 ? Im l -0.1
-0.1 Re l ? 0.1
16
c2 surface (standard cuts)
17
an easy analysis (?)

• normalize data to themselves
• no worrying about trigger efficiencies
• no need for Monte Carlo corrections

• just take the data (2002), fit them, and publish!

• ... but ... the devil is in the detail!

18
result depends strongly on polarity of
spectrometer magnet ? !
19

• sailors show strong
• dependence on
• magnetic field
• orientation !
• different result
• bad c2

cowboy sailor pos.
neg. pos. neg.
20
sailors
cowboys
21
narrow outer radius cut or lower energy cut
improve the problem ... but dont solve it
! some link to drift chamber acceptance?
22
no strong variation with the vertex position
23
these are no good sailors!
24
Cowboys, sailors and azimuth
25

• kick out those sailors?
• they make up
• half the statistics
• one should understand the problem

26

• almost half the cowboys keep their feet crossed
• ? will compensate acceptance problems!
• but sailors stay all on one side

27
cowboys have lower momentum (because low-momentum
sailors get their feet outside the drift chamber
acceptance)
28
azimuth atan2(Dy,Dx) (absolute value, divided
by p) another way to separate these events
29
Dvtx (neutral-charged) depends on
azimuth shifting it by p (1800) does not
completely match positive and negative magnetic
field
30
Momentumdistributions
31
momentum distribution in laboratory system
positive field negative field
p p- ratio
32
(No Transcript)
33
p/p- for same field orientation
34
p/p- for same detector side
35
momentum distribution in center-of-mass system
positive field negative field
p p- ratio
36
p/p- for same field orientation
37
p/p- for same detector side
38
p/p- split up into cowboys and sailors for
same field orientation
39
p/p- split up into cowboys and sailors for
same detector side
40
everything flat in Monte Carlo
41
Drift chamber illumination
42
left-right asymmetry? 1) take symmetric
illumination (p p-, mag.pos. neg.) 2)
mirror x ? -x 3) orig mirror orig mirror
Jura (left) Saleve (right) asymmetry for
unbiased events just the same as for KS ? pp-p0
43
LKR clusters go to the other side for KS ?
pp-p0
44
detector shift (cm)
45
the difference between magnet up down p -
p-
46
KS ? pp-p0 mass over runs (and magnetic field
periods) opposite shift (0.5 MeV/c2) for
cowboys and sailors
47
p ghosts
they disappear when demanding 1 vertex,
2 tracks but why are there no p- ghosts ? d
rays?
48
weighting periodswith magnet up down ?
49
(No Transcript)
50
but why not take just cowboys?
or weighted sailors?
there is a discrepancy of a factor 4!
51
KS ? pp-p0 summary

• main (only?) remaining problem is the magnetic
  field effect
• before we publish
• we should understand it better
• be able to prove that our approach is correct

52

• even in an easy analysis .... one may
  stumble over something !

53
RESERVE
54
the experimental situation

• Fermilab experiment E621
• 1.1 million pp-p0
• 0.6 to 7.6 KS lifetimes
• BR ( 4.8 2.2-1.6 (stat.) ? 1.1 (syst.) ) ?
  10-7
• CPLEAR
• BR ( 2.5 1.3-1.0 (stat.) 0.5-0.6 (syst.) ) ?
  10-7
• Martin Wache
• BR ( 6.0 ? 2.5 (stat.) ? 2.5 (syst.) ) ? 10-7
• see his February presentation
• he thinks one can improve the systematic error

55
the CP-violating part

• rate expected to be about 300 times smaller than
  for the CP-conserving part
• only limit given by CPLEAR
• Re (h-0) ( -2 ? 7 (stat.) 4-1 (syst.) ) ?
  10-3
• Im (h-0) ( -2 ? 9 (stat.) 2-1 (syst.) ) ?
  10-3
• do not yet know if we could be competitive
• difficult analysis
• detailed MonteCarlo needed

56
(No Transcript)
57
(No Transcript)
58
magnet field sub-periods


-
-
59
asp dependence on X
60
LKR (X) dependence maybe not negligible
61
LKR (X) difference for Xgt0 and Xlt0 normalized to
sum