New results on sin2b with charmonium and penguin modes

1
New results on sin2b with charmonium and penguin
modes

• David J. Lange
• Lawrence Livermore National Laboratory
• Representing the BABAR Collaboration

2
CP Violation in Standard Model from non-zero
phase in CKM matrix

• Coupling for Q ?W q is VQq

Three generations 4 fundamental parameters 1
phase
Test unitarity of matrix with B decays. Does
3
Overconstrain angles and sides of Unitarity
Triangle to test the Standard Model
Unitarity Triangle
a(f2)
b(f1)
g(f3)
Measure b with golden modes B?J/yK0
4
CPV in B?J/yK0 Interference of decay and mixing
amplitudes
mixing
decay
Amplitude ratio
5
CPV in B?J/yK0 Interference of decay and mixing
amplitudes
mixing
decay
Amplitude ratio
(DG0)
C0 and S/- sin2b for B?J/yKS and B?J/yKL
6
2002 BABAR and Belle experiments conclusively
observe that sin2b is not 0

• World average
• 0.731/-0.056
• Perfect agreement between constraints of apex
  of the Unitarity Triangle.

7
Penguin modes also measure sin2b

• b?sss decays also measure sin2b in the SM
• Pure internal and flavor-singlet penguin
  diagrams
• Smaller SM amplitudes More sensitive to new
  physics amplitudes

Both decays dominated by single weak phase
Tree
Penguin
New Physics?
3?
?
8
Naïve expectation sin2b in some penguin modes
agrees at 5 with charmonium
Dominant
Suppressed
Color-suppressed tree

• Modes with suppressed C-S tree diagram have
  smallest uncertainty fKS

9
Penguin results from BABAR
10
Peak Pep-II luminosity gt3x design
(as of July 31, 2004)
11
BABAR Detector
Excellent high momentum particle ID performance
crucial for these analyses.
EMC 6580 CsI(Tl) crystals
e (3.1GeV)
DIRC (PID) 144 quartz bars 11000 PMs
Drift Chamber 40 layers
1.5T solenoid
e- (9GeV)
Instrumented Flux Return iron / RPCs (muon /
neutral hadrons)
Silicon Vertex Tracker 5 layers, double sided
strips
12
Run 5 starts October 15 with 1/3 of IFR barrel
RPCs replaced with LSTs

• Goal for 8 month run is 130 fb-1 with peak
  luminosity of 1.5x1034 achieved by June 2005

13
Typical B reconstruction variables
Variables for signal/BG discrimination
Reject background with
event shape information.
Event shape discriminators usually combined into
neural network (NN) or Fisher discriminant
14
Experimental procedure to measure time-dependant
CP Violation parameters
ee- ? ?(4S) ? B B
Boost bg 0.56
?(4S)
B0
B0
Coherent L1 state
15
Experimental procedure to measure time-dependant
CP Violation parameters
ee- ? ?(4S) ? B B
Boost bg 0.56
?(4S)
B0
B0
Coherent L1 state
16
Experimental procedure to measure time-dependant
CP Violation parameters
ee- ? ?(4S) ? B B
m-
Flavor tag and vertex reconstruction
Boost bg 0.56
K-
?(4S)
m-
B0
B0
Coherent L1 state
Fully reconstruct one B meson
17
Tagging and Dt resolution parameters are
determined from data
perfect tagging time resolution
typical mistagging finite time resolution
(f-)
(f)
18
Determine w and R parameters from more plentiful
B0-B0 decays to flavor eigenstates.
signal region
signal region
MES GeV
MES GeV
19
Boosted center-of-mass plus silicon vertex
detector required for Dt determination
K
J/Y
U(4S)
K0

• Reconstruct Brec vertex from charged Brec
  daughters
• Determine BTag vertex from
• All charged tracks not in Brec
• Constrain with Brec vertex, beam spot, and ?(4S)
  momentum
• Remove high c2 tracks (to reject charm decays)
• High efficiency 95
• Average Dz resolution 180 mm (dominated by
  BTag)
• ltDzgt 260 mm
• B mesons produced just above threshold ltDzgt
  30 mm if no boost

20
B decay properties used to determine if tagging B
decayed as a B0 or B0
Measure of tagging performance Q Qe(1-2w)2

• 5 (relative) improvement in tagging algorithm.

21
Tagging algorithm improvements

• New tagger based on same idea/framework as
  previous one.
• Physics changes
• Improved use of correlations
  between Kaons in event
• L?pp as source of tagging
• Secondary electrons
• New way to categorize events
• Category 1 Primary leptons
• Categories 2-6 Split remaining events based on
  estimated mistag rate (from NN).
• Not all analyses use new tagger yet.

MC Data NN estimate
22
sin2b with Charmonium modes
23
Event sample for Golden channels
signal region
MES GeV

• 3900 hCP -1 tagged signal events

24
B?J/yKL and B?J/yK0(KSp0)

• 400 J/yK0 tagged signal events

signal region

• 1600 J/yKL tagged signal events

MES GeV
25
New sin2b Results 227 BB events
hCP-1
hCP1
sin2ß 0.722 ? 0.040 (stat) ? 0.023 (sys)
(2002 measurement sin(2ß) 0.7410.0670.034)
26
Best of the best Lepton tagged hCP-1 events
Lower background Close to perfect
tagging Better Dt determination

• sin2b0.75/-0.08

27
Consistent results when data is split by decay
mode and tagging category
?21.9/5 d.o.f. Prob (?2)86
?211.7/6 d.o.f. Prob (?2)7
28
Decreasing systematic error sin2b measurement
still statistics limited.
s(sin2b) Description of background
events 0.012 CP content of peaking
background Background shape uncertainties Mistag
differences between BCP and Bflav
samples 0.007 Composition and content of J/y KL
background 0.011 Dt resolution and detector
effects 0.011 Silicon detector alignment
uncertainty Dt resolution model Beam spot
position 0.007 Fixed ?m, t, ?G/G,
? 0.005 Tag-side interference/ DCSD
decays 0.003 MC statistics/bias 0.003 TOTA
L 0.023
Steadily reducing systematic error July 2002
0.033 July 2001 0.05
29
CKM picture with new sin2b measurement
cos(2ß)lt0
cos(2ß)gt0

• 1 of 4 solutions for b overlays allowed region by
  other constraints.

30
B?J/yK channel sensitive to cos2b if angular
variables are included in analysis

• CP even (L0,2) and odd (L1) amplitudes averaged
  over in nominal sin2b analysis.
• Terms proportional to cos2b also in full
  amplitude
• Sign of cos2b mathematically ambiguous
• Two-fold ambiguity in determination of strong
  phases

angular amplitudes
decay angles
angular amplitudes in transversity basis
31
Ambiguity solved via S-wave P-wave interference
BABAR (L82fb-1) Kp- invariant mass
P-wave intensity
P-wave
S-wave intensity
S-wave

• Wigner causality
• Resonance phase rotates counter-clockwise
• P-wave moves fast, S-wave moves slow
• Look at interference term in amplitude analysis
• dS-dP vs. m(Kp) Which solution is physical?

m(Kp)
32
Clear solution to strong phase ambiguity
solution 1
? solution 1 unphysical solution
Preliminary
solution 2
? solution 2 physical solution
33
Clear solution to strong phase ambiguity
solution 1
? solution 1 unphysical solution
Preliminary
solution 2
? solution 2 physical solution
34
Comparison with LASS Kp scattering data
solution 1
? solution 1 unphysical solution
Preliminary
solution 2
? LASS data
? solution 2 physical solution
35
Measure cos2b with angular and time dependent
analysis

• Current results on 88 million BB events.
• 104 tagged signal events.

Preliminary
Distribution of cos(2ß) results from data-sized
Monte Carlo samples, generated with cos(2ß)0.68
(with sin(2ß) fixed to 0.731)
Standard Model sign of cos(2b) favored by our
data.
36
Conclusion for sin2b with Charmonium

• Updated measurement of sin2b with B?J/yK0 decays
  using full BABAR data sample
• Novel method to break strong phase ambiguity in
  measurement of cos2b in B?J/yK decays
• cos2b-0.68 excluded at 86.6 level. More data
  to be included in this analysis.

sin2b 0.722 0.040 (stat) 0.023 (syst)
37
BABAR charmless analysis requirements

• DIRC for separation of high
  momentum p and K.
• Continuum rejection
• Neural network or Fisher to
  optimally combine event
  shape discriminants.
• Design high efficiency selection. Maximum
  likelihood fit to untangle signal from background
  in optimal way.
• Variables mES, DE, (NN or Fisher), resonance
  mass, decay angle, tagging, and Dt
• Contributions Signal, continuum, B background(s)

38
B?fKS and B?fKL

• f?KK- dE/dx DIRC information
• B?fKL mode like B?J/yKL
• Add continuum suppression variables
• New for updated analysis
• New tagger
• Event yields determined along with CP parameters
• Improved B background treatment

39
Event yield results for 227x106 BB
114 12 signal events
Projection onto mES after likelihood ratio (w/o
mES)
Likelihood ratio
40
Event yield results for 227x106 BB
98 18 signal events
Projection onto DE after likelihood ratio (w/o DE
)
Likelihood ratio
full background
continuum bkg
41
New CP asymmetry results confirm previous
measurement
B0??KS
B0??KL

• Systematic errors dominated by
• opposite-CP background
• PDF modeling
• Tag-side CP violation

Previous result 114x106 BB
42
B?KK-KS

• B?fKS only 15 of B?KK-KS events
• We analyze the rest excluding the
  B?fKS contribution.
• Determine the CP content via
  angular moments analysis of
  KK- helicity angle
  distribution.
• Dominantly CP-even

43
B?KK-KS event sample
Likelihood projection onto mES (after LR cut)
Likelihood ratio
452 28 signal events (excluding ? KS events)
44
B?KK-KS (227x106 BB pairs)

• Systematic errors dominated by
• Fit bias
• Tag-side CP Violation

CP content
45
B?hKS

• Definitely not a rare decay mode
• Reconstruct in multiple final states
• ? ? ???, ?0?
• ? ? ?? , ? ??0
• KS ??? ,?0?0
• Likelihood projection

Background
Signal
46
Projections of B?hKS (227x106 BB)

• Likelihood projections onto mES and DE.
• Most modes have very low background.
• Yield from fit 819 38 signal events

47
S coefficient is 3s from ccK sin2b
3s
48
B?f0(980)KS (f0?pp-)

• f0(980) is broad. Use quasi two-body approach
• Analyze f0?pp- region of pp-KS Dalitz plot
• Account for other B?pp-KS contributions
• Vary size and relative phase of contributing
  amplitudes as part of systematic error
• Mass and width parameters of relativistic BW
  floating in likelihood fit
• Not sensitive to different lineshapes
• No analysis changes for updated results

49
B?f0KS from 206x106 BB pairs
Likelihood projection
152 19 signal events
Signal
Likelihood projection
Continuum
B back.
50
CP results from B?f0(?pp-)KS

• Larger than expected improvement in errors as
  well as shift in S largely due to new lepton
  tagged event with high signal probability and
  good Dt.
• Systematic error dominated by unknown ppKS
  contributions in f0 region of Dalitz plot.
  (Q2B approach)

51
Time-dependant analysis of B?p0KS using novel
vertexing technique

• Lifetime of KS requires additional information to
  be used to determine Dt with adequate precision
• Require at least 4 SVT hits on each KS?pp-
  daughter track.
• 40 of events failing this
  criterion are still used to
  determine the direct CP coeff.
• Include beam energy and beam
  spot (determined run by run)
  constraints and fit full
  U(4S) decay tree.

52
209x106 BB results for B?p0KS
Replace ?E and mES Reduced correlation,
improved resolution (mmiss)
mass-constrained
Projections after LR cut
300/-23 signal events
53
New CP results for B?p0KS
sPlot

• Systematic errors dominated by
• Background tagging asymmetry
• SVT alignment, vertexing

Previous result 124x106 BB
54
No indication of significant direct CP violation
(cos(DmDt) term)
Still 81 fb-1 results
Charmonium average
Penguin average
55
Penguin and tree measurements of sin(DmDt)
coefficient differ at 2.7s
Still 81 fb-1 results
Charmonium average
Penguin average
56
Belle results also show a difference with respect
to sin2b from ccK0
Prob. of c2 between experiments is 5
sin2b from ccK
S(ccK0) Spen. 0.31 0.09
3.6s from expectation?
57
Are we seeing hints of new physics?

• Maybe you believe the glass is ½ full or
  that the glass is ½ empty.
• 3s indication of additional CPV amplitude
  contribution.
• Expect new physics effects to appear differently
  in different modes
• Averaging most relevant in Standard Model case
• For BABAR, B?hKS drives average away from ccKS
  sin2b.
• Do we worry about BABAR vs Belle agreement?

Current situation gives interesting hint. But it
is too early to draw any conclusion
58
Future expectations
f0KS KSp0 jKS hKS KKKS
Luminosity expectations
2004240 fb-1 2006500 fb-1
Kg
5s discovery region if non-SM physics is 30
effect
2006
2004
59
Conclusion

• BABAR has updated its ccK0 and penguin sin2b
  measurements to include its latest data
• Most up to 227 BB pairs.
• BABAR sin2b with ccK0 now a 5 measurement.
• Hints that sin2b measured in penguin modes is not
  the same as in golden B?ccK0 modes.
• 3s deviation in B?hKS.
• B?hKS is also the most precise penguin
  measurement (/- 0.18).
• B?fK0 agrees with ccK0 within 1s.

sin2b 0.722 0.040 (stat) 0.023 (syst)
Stay tuned for increasingly precise results as B
Factory samples increase
60
How do we display the results of our likelihood
fits?

• Plot of likehood ratio L(sig)/(L(sig)L(backgroun
  d))
• Projection onto mES (or other variable) after cut
  on likelihood ratio.
• Plotted variable not used to calculate likelihood
  ratio
• Superimpose signal and background PDFs from
  primary likelihood fit.
• sPlots (ref physics/0402083)
• Weighted histogram of mES
• Weights determined from other variables in fit
• Weights chosen so histogram is unbiased estimator
  of mES for signal contribution. (or background)