1. Gagan Mohanty Charm of the charmless – Three-body Hadronic B decays at BaBar RAL PPD Seminar December 12, 2007

2. 2 Outline of the talk Theory and Motivation Dataset and Detector Analysis Strategy –Particle Identification –Continuum Suppression –Kinematical Variables Results  B + → K + K - π + /K + π + π - /π + π - π +  B 0 → K S π + π - /K S K + K - /π + π - π 0 Summary and Outlook Preliminary

3. 3 History: Timeline (1993) 1 st observation of charmless B decays by CLEO  Since then… PRL 71, 3922 (1993)

4. 4 Today: LP2007 http://www.slac.stanford.edu/xorg/hfag/rare/index.html

5. 5 Theory b → s loop (penguin) transition contributes only to the final states with odd number of kaons due to presence s quark e.g. Kππ, KKK b → u b → s/d Final states with even number of kaons, such as KKπ get contributions from b → u tree and b → d penguin diagrams. Odd number kaon states are further Cabibbo suppressed [~ sinθ c ] Typical diagrams for charmless three-body B decays

6. 6 Motivations Interfering tree and penguin amplitudes good place to search for direct CP violation New particles can appear in loop diagrams (signature of physics beyond the SM) Probes flavor sector, particularly by measuring –sin(2β) or just β in the K S h + h - (K/π) Dalitz plot –α in the modes: ππ, ρπ and ρρ –γ using flavour symmetries (isospin, U-spin, etc.) Low energy spectroscopy Testing ground for perturbative QCD, factorization, SU(3) flavor symmetry … Klempt, Zaitzev Phys Rept 454, 1 PRL 93, 131801

7. 7 PRO : - Larger BF than two-body decays - Correct way to study interference - Some modes in a well-defined CP eigenstate PRO : - Larger BF than two-body decays - Correct way to study interference - Some modes in a well-defined CP eigenstate CON : - large phase space with low event density; hard to identify all phase-space structures - mixture of CP-even, CP-odd final states - more complicated analysis needed CON : - large phase space with low event density; hard to identify all phase-space structures - mixture of CP-even, CP-odd final states - more complicated analysis needed Why three-body? Gershon, Hazumi PLB 596, 163 3-body 2-body

8. 8 Dataset L int ~ 500 fb -1 (10% off-peak) Run 6 ended on this Sept. New results based on run 1-5 data (other on smaller set) Final run 7 is about to begin and would continue till next September 9 GeV e - →Υ(4S)← 3.1 GeV e + γβ = 0.56 ~ 260 μm  Peak luminosity > 3 X Design

9. 9 BaBar Detector Cerenkov Detector (DIRC) 144 Quartz bars and 11000 PMTs 1.5T solenoid Electromagnetic Calorimeter 6580 CsI(Tl) Crystals Drift Chamber 40 layers Instrumented Flux Return Resistive Plate Chamber & Limited Streamer Tube Silicon Vertex Tracker five layers, double strip e + (3.1GeV) e - (9GeV) e  ID  reco.   & K L ID High quality tracking (fiducial volume: 0.41 <  < 2.54)  (p T )/p T = 0.13% P T  0.45%  E /E = 2.32% E -1/4  1.85% Particle ID

10. 10 Analysis Strategy  Signal extraction (kinematics)  Background fighting: Continuum (event topology) Other types of B decays (PID, charm and charmonia veto)  Dalitz plot technique (three-body decays having reasonable signal size)  Time-dependent analysis in neutral B meson decays to determine CP violation parameters at each point of the phase space Inclusive Full (3body)/partial (Q2B) Time-dependent DP (3body) uds:cc:bb = 2.1:1.3:1.1 Complexity

11. 11 Particle Identification PID is crucial for these analyses –distinguish K vs. π –veto the leptons >3σ Longitudinal shower depth from an unsegmented EMC Always room to improve (submitted to NIM ) D.N. Brown et al., arXiv:0711.1593

12. 12 Continuum Suppression Event topology –B produced at rest (spherical) –Jet-like udsc events Signal Background Flavour and decay- time of the B meson candidate (not used in TD analysis)

13. 13 Typical Performance We do not live in an ideal world But we try to get as close as possible A. Hocker et al., physics/0703039

14. 14 Kinematical Variables Utilize precise beam energy information and (E,p) conservation Sideband Signal box Background signal background Inclusive/Dalitz plot

15. 15 Dalitz-plot Analysis Powerful technique relying on Lorentz invariant phase-space variables in a three-body decay B 1 2 3 B 2 1 3 + Resonance {13} s ij = m 2 ij Zemach, PL133 B1021 (1964) Flatte, PLB63, 224 (1973)

16. 16 Dalitz-plot Analysis [2] Extract c k,NR and θ k,NR by performing a maximum likelihood fit Measure CP violation asymmetries by comparing B and B amplitudes Fit fraction is the ratio of the integral of a single decay amplitude squared to the coherent sum of all

17. 17 Time-dependent DP  Time-dependent decay rate of B 0 (B 0 ) → three-body  Determine mixing-induced CP [sine coefficient] and direct CP [cosine coefficient] at each point in the DP  direct CP  Include detector effects (mistagging and resolution) PEP-II  (4S) e-e- Coherent evolution e+e+  z≈  m mixing Flavor tagging Q ≈30% B 0  K S π + π - Eff.~25% -- -- ++ ++ B0B0 K-K- e-e-

18. 18 B + → K + K - π + inclusive

19. 19 Motivations Potentially rich Dalitz structure Good place to look for direct CP (interference between b→u tree and b→d penguin amplitudes) Little experimental information exists new physics effects not excluded Rate and asymmetry in B + → K *0 (892)K + are inputs to a method to extract γ Same Q2B state helps on understanding observed discrepancy of sin(2β eff ) in the B → φ K S mode Grossman et al., PRD 68, 015004 Surprises do happen And... Soni and Suprun, PRD 75, 054006

20. 20 Current Status Theory Numbers are from BaBar, competitive limits also available from Belle and CLEO for the modes indicated by (*) Experiment 0.1 0.3 0.5 0.7 0.1 0.6 3. 0 Factorization pQCD SU(3) Interesting Lower Limit: Fleischer and Recksiegel PRD 71, 051501 (2005)

21. 21 B + → K + K - π + inclusive  An unbinned maximum likelihood fit of [m ES, ΔE, NN] to 16143 candidate events finds a signal yield of 429±43  12.6σ (statistical only) and 9.6σ including systematic uncertainty 383M B pairs BB qq PRL 99, 221801

22. 22 B + → K + K - π + inclusive Half of the events originates from Reminiscent of similar structures in K S K + K - and K + K + K - Dalitz plots PRD 74, 032003  Nature of this state around 1.5GeV/c 2 is not very clear Efficiency-corrected distribution Rate reasonably consistent with the Q2B results on K *0 K Efficiency-corrected distribution KSK+K- KSK+K- K+K+K- K+K+K- PRL 99, 221801 PRL 99, 161802 PRD 76, 071103

23. 23 Dalitz plot analyses of B + → K + π + π - and π + π - π +

24. 24 B + → π + π + π - Dalitz plot Phase-space Coupled BW ρ(770) f 2 (1270) Folded Single BW qq BB PRD 72, 052002 (2005) 232M B pairs 468±35 Fit

25. 25 B + → π + π + π - : Summary ρ 0 (770) is the dominant component  3σ indication for f 2 (1270) and NR  Little evidence for σ (seen by BES in the decay J/ψ→ωπ + π - )  No contribution from χ c0 not feasible to measure γ with analysed dataset PRD 72, 052002 (2005) σ f 0 (980) f 2 (1270) PLB 597, 39 (2004) Bediaga et al., PRL 81, 4067 (1998)

26. 26 B + → K + π + π - : Dalitz plot B+B+ B+B+ qq BB Fit ρ(770) f 0 (980) χ c0 K *0 (Kπ) *0 PRD 72, 072003 (2005) 226M B pairs 1078±56 B±B± B+B+

27. 27 B + → K + π + π - : Summary  (Kπ)* 0 K* 0 (1430) resonance + Effective range nonresonant component (Belle uses K *0 (1430) only)  Evidence for direct CP violation in the ρ 0 (770)K ± mode PRD 72, 072003 (2005) PRL 96, 251803 (2006)  Total BF differs significantly from Belle (48.8±1.1±3.6)10 -6

28. 28 Belle CP in charged B decays? Chiang et al., PRD 69, 034001 (2004)  Interesting to see the results with higher statistics… BABAR B+B+ B-B- Large A CP in agreement with predictions based on flavour SU(3) symmetry (19-24)% PRL 96, 251803 (2006) PRD 72, 072003 (2005)

29. 29 Time-dependent Dalitz plot analyses of B 0 → K S π + π - and B 0 → K S K + K -

30. 30 Motivations Dominantly b → s penguin transition prone to NP effect Provides a test if mixing-induced CP asymmetry equals to that of tree-level transition b → ccs Measure β eff in Q2B modes unambiguously interference term allows determination of cosine term (beauty of DP) We can determine the relative phase between B 0 → K *+ (892)π - and B 0 → K *- (892)π + access to CKM angle γ Deshpande et al., PRL 90, 061802 (2003) Ciuchini et al., PRD 74, 051301 (2006) Gronau et al., PRD 75, 014002 (2007)

31. 31 Existing Measurements  Both agree reasonably well Discrepancy in the nonresonant contribution Belle also observes structure near 1.3 GeV/c 2 in the π + π - spectrum Time-dependent Q2B Time-integrated Q2BTime-integrated DP

32. 32 Signal Yield Simultaneous fit including –m ES, ΔE, NN, Δt and tagged (B 0 /B 0 ) DP variables 383M B pairs arXiv:0708.2097 Contrib. LP2007 BB qq  Signal: (2172±70) in total candidate sample of 22525

33. 33 Dalitz plot Content qq BB f 0 (980)  ρ 0 (770)K S Gounaris-Sakurai  f 0 (980)K S Coupled BW  K * (892)π Single BW  K *0 (1430)π LASS shape  f x (1300)K S Single BW  f 2 (1270)K S,,  χ c0 K S,, ρ(770) f x (1300)+f 2 (1270) χ c0 K * (892) K *0 (1430) arXiv:0708.2097 Contrib. LP2007

34. 34 Time-dependent CP violation Charmonium  Time-dependent CP asymmetry measured at each point in the K S π + π - Dalitz plot for the first time f 0 (980)K S value 2.1σ above charmonium ρ 0 K S consistent with the world-average arXiv:0708.2097 Contrib. LP2007

35. 35 CP violation in DP amplitudes  Advantage of time-dependent Dalitz plot probes CP violation from two orthogonal directions No CP Phase diff. in K * π mode is significantly different from zero disfavoured at 3.7σ arXiv:0708.2097 Contrib. LP2007

36. 36 B 0 → K + K - K s : Dalitz plot  φ(1020)K 0  f 0 (980)K 0  f x (1500)K 0  χ c0 K S  Nonresonant Fit including m ES, ΔE, NN, Δt and tagged (B 0 /B 0 ) DP variables 383M B pairs Signal model arXiv:0706.3885, accepted by PRL  Low mass: m KK < 1.1 GeV/c 2

37. 37 B 0 → K + K - K s : Results CP violation at 4.8σ level (5.1σ for high mass region)  β eff is about 2σ lower than the SM prediction (~0.370 rad)  Direct CP is consistent with SM disfavoured at 4.5σ arXiv:0706.3885, accepted by PRL

38. 38 Time-dependent Dalitz plot analysis of B 0 → π + π - π 0

39. 39 Motivations Contributions from b→u tree and b→d penguin transition (through intermediate ρ resonance) Possible to extract of CKM angle α and strong transition amplitude simultaneously Complex isospin analysis 12 parameters to be determined Rather utilize the variation of strong phases of interfering ρ resonances over the DP Time dependent Dalitz plot Synder and Quinn, PRD 48, 2139 (1993)

40. 40 B 0 → π + π - π 0 time-dependent DP  Parameterize B 0 (B 0 ) → π + π - π 0 amplitude in terms of ρ + (→π + π 0 )π -, ρ - (→π - π 0 )π + and ρ 0 (→π + π - )π 0 }  Use these in the time-dependent decay rate: f κ (κ = +,-,0) are BW functions Determine U κ and I κ (27 → 16 parameters for small ρ 0 π 0 contribution) in the likelihood fit

41. 41 DP Fit Results Likelihood built using PDFs for the discriminating variables: ΔE, m ES, NN, Δt & DP variables Extract physics parameters from the fitted U κ and I κ : { PRD 76, 012004 (2007) 375M B pairs Event yield: 2067 ± 86 BB qq misreconstruction ρ(770) D+π -D+π - KSπ0KSπ0

42. 42 Angle α and direct CP { Direct CP violation ~ 3σ level  11 11 22 22 11 22 33 No CP PRD 76, 012004 (2007) {

43. 43  using b  uud: WA

44. 44 Conclusions  First measurement of the inclusive mode B + → K + K - π +  DP measurements in the charged Kππ and πππ modes  Evidence of direct CP violation in the ρ 0 (770)K ± mode of charged Kππ final state  β eff measured without any sign ambiguity (thanks to the time-dependent DP technique)  Measured CP violation parameters agree reasonably well with SM predictions  Time-dependent DP measurement of B 0 → (ρπ) 0  direct CP violation at ~3σ level and measured α

45. 45 Bonus slides

46. 46 - parameters a,r taken from LASS experiment (*) - valid up to 1.8GeV (*) LASS, K  scattering at 11GeV at SLAC Scalar Kπ near 1.4 GeV/c 2  BaBar:  BW for K * (1430) plus an effective range NR component  Flat NR component  Belle:  BW for K * (1430)  exponential NR

47. 47 sin2  in Penguins New naïve HFAG average <1  from the charmonium mode sin2  value  New/Updated BaBar/Belle Result                 sin2  S f = -sin2  eff 1% CL for the average Slide from LP2007 (Dave Brown)