Daniela Bortoletto

1
The measurement of sin(2?)

• Daniela Bortoletto
• Purdue University
• Introduction
• SM expectations
• Previous measurements
• The measurement of sin 2? at CDF
• Signal reconstruction
• Flavor tagging methods
• Fit results and cross checks
• Future prospects

University of Southampton 25-29 July 199
2
Introduction

• SM with 3 generations and the CKM ansatz can
  accomodate CP
• if the complex phase ? is ? 0?CP. Only
  ?0.2196?0.023, A0.819?0.035 are measured
  precisely.
• CP is one of the less well-tested parts of SM (?
  , ? ?/ ? in the Kaon system)
• CP asymmetries in the B system are expected to be
  large. Independent observations of CP in the B
  system can
• test the SM

• lead to the discovery of new physics

3
B Physics and CKM matrix

• The goal of B-physics is to over-constrain the
  unitarity triangle to test the CKM ansatz or to
  expose new physics

Vud VubVcd VcbVtdVtb0
B???
B?J/?K0s

Unitarity triangle
4
CP violation in B decays

• Possible manifestations of CP violation can be
  classified as
• CP violation in the decay It occurs in
  B0/Bdecays if A(f)/A(f)?1
• CP violation in mixing It occurs when the
  neutral mass eigenstates are not CP eigenstates
  (q/p?1)
• CP violation in the interference between decays
  with and without mixing
• Mixing Vtd introduces a complex phase in the
  box diagram
• Interfering amplitudes
• direct decay B0 ? f
• B0 ? B0 mixing followed by B0 ? f

Box Diagram
Vtd
5
Determination of sin(2?)

• Color suppressed modes b?ccs. Dominant penguin
  contribution has the same weak phase ? Negligible
  theoretical uncertainty
• Cabibbo suppressed modes b?ccd such as B0/B0?
  DD,DD. ?Large theoretical uncertainties due to
  the penguin contribution
• Penguin only or penguin dominated modes b?sss or
  dds. Tree contributions absent or Cabbibbo and
  color suppressed ? penguin diagrams dominate ?
  even larger theoretical uncertainties

6
Experimental considerations

• B-factories at the ?(4S)
• B0 and B0 mesons are produced in a coherent C-1
  state? time integrated CP asymmetry 0.
• Determination of CP needs A(?t ) where ?t
  t(CP)-t(tag) or ?z ??c ?t
• Need good ?z resolution
• pp and pp colliders time integrated asymmetry
  does not vanish
• Since xd0.732? 0.0032 (PDG98)

ACP is Maximum at t2.2 lifetimes

• Measurement of the asymmetry as a function of
  proper time ACP(t) is more powerful
• Combinatoric background dominates small ct region

7
B0/B0?J/?K0s

• For B0/B0? J/?K0S we have CP(K0s)1 and CP(J/?K0S
  ) -1. To reach a common final state the K0 must
  mix ? additional phase
• Asymmetry is directly related to sin2?.
• ACP(t)sin2(?M- ?D)sin ?mdt sin2? sin ?mdt
  and
• sin2?
  

K0-K0 mixing
B0 B0 Mixing
Ratio of
8
Indirect determination of sin2 ?

• In SM the asymmetries in the B system are
  expected to be large

• Vub/Vcb0.093 from semileptonic decays
• ?K2.28?10-3
• B0-B0 mixing ?md0.472 ps-1
• Limit on Bs-Bs mixing ?ms gt12.4 ps-1

• S. Mele CERN-EP-98-133, 1998 finds sin2?0.75?
  0.09
• Parodi et al. sin2?0.725? 0.06
• Ali et al. 0.52ltsin2?lt0.94

9
Measurement accuracy

• Measurement of ACP(t) requires
• Reconstruct the signal B0/B0?J/?K0S
• Measure proper decay time (not critical in pp
  colliders but useful)
• Flavor tagging to determine if we have a B0(bd)
  or B0(bd) at production
• Tagging algorithms are characterized by an
  efficiency ? and a dilution D. The measured
  asymmetry is AobsCPD ACP
• Ntot total number of events
• NW number of wrong tags
• NRnumber of right tags
• D2P-1 (Pprob. of correct tag) and D1 if NW0
  D0 if NWNR
• Best tagging methods has highest ? D2

Crucial factor
10
Tagging

• Assume you have 200 events ? N200
• 100 are tagged ? Ntag100
• tagging efficiency ? Ntag/Ntot50
• Of those 100 events
• 60 are right sign ? NR60
• 40 are wrong sign? NW40
• Dilution
• D(NR-NW)/(NRNW)(60-40)/10020
• Effective tagging efficiency
• ?D2( 0.5)(0.2)22
• Statistical power of this sample
• N?D22000.024 events

11
Previous Measurements

• Opal Z?bb D. Ackerstaff et al. Euro. Phys. Jour.
  C5, 379 (1998) (Jan-1998)

24 J/?K0S candidates Purity ? 60

• Flavor tagging techniques
• Jet charge on opposite side jet
• Jet charge on same side B
• Vertex charge of a significantly separated vertex
  in the opposite hemisphere

sin2?3.2? ?0.5
1.8 2.0
12
Previous Measurements

• CDF pp?bb Abe et al. PRL. 81, 5513 (1998) (June
  1998)

• 198 ?17 B0/B0 ?J/?K0S candidates with both muons
  in the SVX ( S/B ? 1.2). Measure asymmetry with
  Same side tagging
• Dsin2?0.31? 1.1 ? 0.3.
• Using D0.166 ? 0.018 (data) ? 0.013 (MC) from
  mixing measurement MC

sin2?1.8? 1.1 ? 0.3
13
Run I CDF detector

• Crucial components for B physics
• Silicon vertex detector ? proper time
  measurements
• impact parameter resolution
• ?d(1340/pT) ?m
• typical 2D vertex error ?(r-?) ? 60
  ?m
• Central tracking chamber ? mass resolution.
  B1.4T, R1.4m (?pT/pT)2(0.0066)2?(0.0009pT)2
• typical J/?K0S mass resolution ? 10 MeV/c2
• Lepton detection (triggering and tagging)

14
CDF updated measurement

• Add candidate events not fully reconstructed in
  the SVX
• Double the signal to 400 events but additional
  signal has larger ?(ct)
• Use more flavor tag methods to establish b flavor
  at production
• Check ?D2 with mixing analysis
• Use a maximum likelihood method to combine the
  tags. Weight the events
• in mass (B peak versus sidebands)
• in lifetime (more analyzing power at longer
  lifetimes)
• in tagging probability
• Account for detector biases

background
B
??c?
?(B0)1.56?10-12 s
15
J/?K0S Event selection

• Signal
• J/? ? ?-? require two central tracks with
  matching hits in the muon chambers
• K0S ? ?-? use long lifetime c?(K0S)2.7 cm to
  reject background by requiring Lxy/?(Lxy)gt5
• Perform 4-track fit assuming B? J/? K0S
• Constrain ?-? and ?-? to m(K0S) and m(J/?)
  world average respectively
• K0S points to B vertex and B points to primary
  vertex
• Background
• cc production ? prompt J/? ( not from b decays)
  random K0S or fake
• bb production ? J/?X, random K0S or fake

16
J/?K0S Signal sample

• CDF run1, L110 pb-1
• 202 events with both muons in SVX ? ?(ct)? 60 ?m.
  
• 193 with one or both muons NOT in SVX ? ?(ct)?
  300-900 ?m

Both ? in SVX
202 ?18 events
395 ?31 events
S/B0.9
One or Both? not in SVX
193?26
S/B0.7
S/B0.5

• Plot normalized mass
• M????-MB/ error on M

17
Flavor tagging methods

• We must determine if we had a B0 or a B0 at the
  time of production.
• Opposite-side flavor tagging (OST)? bb produced
  by QCD? Identify the flavor of the other b in the
  event to infer the flavor of the B0 /B0? J/?K0S.
  At CDF? 60 loss in efficiency due the acceptance
  of the other B0.
• Lepton tagging
• b?? X ?b
• b?? -X ?b
• Jet charge tag
• Q(b-jet) gt 0.2 ?b
• Q(b-jet) lt- 0.2 ?b

B0(bd)? J/?K0S
?
?-
?
K0S
?-
Opposite side b
?
Q(b-jet)gt0.2
18
Jet Charge Flavor tagging
Qjet in B??J/?K?

• Identify the flavor of the B0/B0?J/?K0S through
  the charge of the opposite b-jet
• Jet definition allows for wide low PT jets
• Cluster tracks by invariant mass (Invariant mass
  cutoff? 5 GeV/c2 )
• remove track close to primary B
• Weight tracks by momentum and impact parameter
• pT track momentum
• TP probability track comes from primary vertex
  (low Tp more likely track comes from B )

-QKQJet

• Qjetgt0.2 ? b
• Qjetlt-0.2 ? b
• Qjetlt0.2 ? no tag
• ? (40.2 ?3.9)

19
Soft Lepton Flavor tagging

• Identify the flavor of the B0/B0?J/?K0S through
  the semileptonic decay of the opposite B.
• b? ? -? X b? ? ? X
• Electron central track (PTgt1 GeV/c) matched to
  EM cluster
• Muon central track (PTgt2 GeV/c) matched to muon
  stub
• Efficiency ? 6
• Source of mistags
• Sequential decay b ? c ? ??X
• Mixing
• Fake leptons
• Opposite side tagging was used at CDF to study B0
  B0 mixing

?md0.50 ?0.05(stat)0.05(sys)? ps-1 ?md0.464
?0.018 ? ps-1 (PDG)
Ph. D. Thesis O. Long and M. Peters
20
Same side tagging

• Problems with opposite side tagging
• Opposite b-hadron is central only ? 40 of the
  time
• If opposite b-hadron is B0d or B0s mixing
  degrades tagging
• Same side flavor tagging (SST). Exploits the
  correlation between the charge of nearby ? and
  the b quark charge due to fragmentation or B
  production (Gronau,Nippe,Rosner)

No K/? separation ? higher correlation for
charged B
21
Same side tagging

• Correlation due to excited B production
• B (I1/2) resonance B- ?B0 ?-
• Implementation of SST Search for track with
  minimum Ptrel in b-jet cone
• SST has higher efficiency (? 70 ) than OST

d
B direction
22
Tagger calibration

• Use B??J/? K ? sample to determine the efficiency
  ? and the dilution D of the sample
• Charge of the K? ? b or b
• Decay mode and trigger analogous to B? J/? K0S
• B/B- does not mix

23
Calibration Jet Charge Tagging

• Sample of 988 J/?K? events
• 273 right-sign events
• 175 wrong-sign events
• Tagging efficiency ?Ntag/Ntot(44.9 ? 2.2)
• Tagging dilution
• DNR-NW/NRNW (21.5 ? 6.6)
• Mistag fraction w(39.2?3.3)

24
Calibration of Soft Lepton Tagging

• Sample of 988 J/?K? events
• 54 right-sign events
• 12 wrong-sign events
• Tagging efficiency ?Ntag/Ntot(6.5 ? 1.0)
• Tagging dilution
• DNR-NW/NRNW(62.5 ? 14.6)
• Mistag fraction w(18.8?7.3)

25
Same Side Tagging Calibration

• Use inclusive ? D sample. This sample was used
  for the determination of B0/B0 mixing in F. Abe
  at al Phys. Rev. Lett. 80, 2057(1998) and Phys.
  Rev. D 59 (1999)

D0.27?0.03(stat)0.02(syst)
D00.18?0.03(stat)0.02(syst)
D0.166?0.022 both muons in SVX D0.174?0.036
one/both muons NOT in SVX

• Use MC to scale for different PT spectrum in J/
  K0S wrt ? D/D sample

26
Flavor Tagging Summary

• Same side SVX ?(35.5?3.7) D (16.6 ?2.2 )
  
• Same side non-SVX ?(38.1?3.9) D (17.4
  ?3.6 )
• Soft lepton all ? (5.6?1.8) D (62.5 ? 14.6)
  ?D2 (2.2 ?1.0)
• Jet charge all ? (40.2 ? 3.9) D (23.5 ?6.9
  ) ?D2 (2.2 ?1.3) (if SLT do not use Jet
  charge)

?D2 (2.1 ?0.5)

• Combined flavor tagging power including
  correlations and multiple tags A
  sample of 400 events has the statistical power of
  25 perfectly tagged events

?D2 (6.3 ?1.7)

• Combining Dilution Define DqD where q-1
  (b-quark), q1 (b-quark) and q0 (no tagging)
  then Deff(D1D2)/(1D1D2)
• Tags agree Deff(D1D2)/(1D1D2) Example
  SST and JCT D36.8
• Tags disagree Deff(D1-D2)/(1-D1D2) Example
  SST and JCT D5.1
• Each event is weighted by the dilution in the fit

27
Results

• Muons from J/? decay in Silicon vertex detector ?
  High resolution ct ? Asymmetry vs ct
• Data with low resolution ct measurement ? Time
  integrated ACP
• ACP0.47 sin2?
• If ?md is fixed to the PDG world average
  (?md0.464?0.018 ps-1), the minimization of the
  likelihood function yields
• sin2?0.79?0.39(stat)?0.16(syst) Statistical
  error gtsystematics.

28
Systematic errors and cross checks

• Systematic errors
• Dilution 0.16 (limited by the statistics of the
  calibration sample)
• Other sources ? 0.02
• Cross checks
• Float ?md
• Measure time integrated asymmetry sin2?0.71
  ?0.63
• Only SVX events and SST sin2?1.77?1.02
• Verify errors and pulls with toy MC

1 ? contours
error
Pull
Mean0.44
?1.01
29
Cross checks

• As a check we can apply the multiple flavor
  tagging algorithm to the measurement of mixing in
  B0?J/?K0 decays.
• The data is consistent with the expected
  oscillations
• Measurements
• ?md(0.40?0.18) ps-1
• DK0.96 ?0.38 dilution due to incorrect K-?
  assignments
• Expectation
• ?md(0.464?0.018) ps-1
• DK0.8 ?0.3

30
Confidence Limits on sin(2?)

• Measurement
• Feldman-Cousin frequentist (PRD 57, 3873, 1998)
• 0ltsin2?lt1 at 93 CL
• Bayesian (assuming flat prior probability in
  sin2?)
• 0ltsin2 ? lt1 at 95 CL
• Assume true value sin2?0. Probability of
  observing sin 2 ? gt0.79 3.6 .

Scan of the likelihood function
sin2?
31
Results in ? and ? plane
1? bounds

• CDF sin2 ? measurements ? fourfold ambiguity ?,
  ?/2- ?, ??, 3?/2-?
• Solid lines are the 1 ? bounds, dashed lines two
  solutions for ? for ?lt1, ?gt0 (shown)
• two solutions for ?gt1, ?lt0 (not-shown)

32
B-factories at?(4S)
pp colliders
BABAR estimates J/?K0S
33
CDF reach in run II for sin2?

• Run I value with Run II projected error
• sin2?0.79 ?0.084

34
Summary

• CDF measures

• Mixing mediated CP will be measured precisely by
  CDF/D0 /BaBar/Belle/HeraB by the beginning of the
  new century
• Precise determination of sin2? is a key step
  towards understanding quark mixing and CP