1. Reflections on Beauty: CP Asymmetries in B Meson Decay K. Kinoshita University of Cincinnati Weak interactions & the b-quark: CKM matrix B (eauty) mesons & CP B meson production: e + e – ->  (4S) Belle/Babar experiment

2. 2 Symmetry of Physical Laws In interaction-free universe (4-d, relativistic QM) massless particles symmetric in transformations P(r -r), C(particle antiparticle), T(t -t) Add interactions : emission/absorption of field quantum mass via self-interaction interaction strength/probability  “ charge ” g 2  “ coupling constant ” symmetry info in vertex Forces : Strong, Electromagnetic, Weak, Gravitational coupling ~ 10 -5, quanta W ±, Z 0

3. 3 Weak Interaction The only force known to allow particle to change identity violate P symmetry (maximally) right-handed particles, left-handed antiparticles. no coupling to LH particles, RH antiparticles. violate CP symmetry (a little) What is source of observed CP asymmetry? Why is CP violation of interest? matter-antimatter asymmetry in universe requires CP violating interactions (Sakharov 1967)

4. 4 fermions: 3 generations x 2 types x 2 ea (doublets) all stable, if not for weak interaction distinguished ONLY by mass (?) We have an interesting possibility... Standard Model = 12 fermion flavors (+antifermion) + strong, EM, weak forces, unification of EM+weak

5. 5 Weak couplings W ± "charged current”  Q = ±1 l – q +2/3 q' – 1/3 (small) generation x-ing, quark only all different strengths Large # of fundamental "charges" – can this be simplified? Z 0 "neutral current”  Q = 0 l ± q +2/3 q' +2/3 NO generation x-ing –> no flavor-changing single coupling strength

6. 6 GIM mechanism Picture strong doublets, “ degenerate ” generations, perturbed by weak force: new doublets no generation x-ing, universal W-coupling (=g F, seen in leptons) d', s', b' are linear combinations of d, s, b: Cabibbo-Kobayashi-Maskawa (CKM) matrix complex preserves metric “ orthogonality = unitary } Explains suppression of flavor-changing neutral currents multiplicity of charged current couplings for >2 generations, CP violation uct d's'b' For 3 x 3, unitarity constrains {9 real+9 imaginary} dof to 4 free parameters, incl. 1 irreducible imaginary part

7. 7 (  "unitarity triangle" Unitarity of CKM V ji *V jk =  ik {i=1,k=3}: V ub *V ud +V cb *V cd +V tb *V td =0 V ub *V ud V cb *V cd V tb *V td V cb *V cd + 1 += 0 =>  Self-consistent if CKM is correct (Wolfenstein parametrization): d s b u V ud V us V ub c V cd V cs V cb t V td V ts V tb 1  2 /2 3 A(  i  )  1  2 /2 2 A 3 A(1  i  )  2 A 1  1/g F x W-couplings: Unitarity condition: from decay rates,  0.220 ± 0.002 A  0.81 ± 0.08 |  i  |  0.36 ± 0.09 |  i  |  0.79 ± 0.19  

8. 8 – t-integrated rates   | | 2 => not sensitive to phase: CP{ } = – need interference between processes: Complex couplings revealed via CP asymmetry e.g., decays to CP eigenstate - paths w/wo mixing interfere -> CP-dependent oscillation in decay time distributions {cc}+{K s,K L,π 0 }

9. 9 CP Asymmetry of B -> J/  K s * No theoretical uncertainty tree (real V ij ) mixing+tree arg(V td 2 ) = 2  1 ->

10. 10 Measure time dependence - what ’ s needed? B pair production  e + e – ->  (4S) -> BB Measure decay-time difference Asymmetric energy e + e –  (@KEKB:  c  200  m) good vertexing  silicon strip vertex detector Find CP eigenstate decays high quality ~  detector  Belle/Babar Tag other B’s flavor good hadron id  dE/dx, Aerogel, TOF, DIRC good lepton id  CsI, multilayer µ Lots of B mesons ~10 8 (Br (B  f CP ) ~ 10  3 ) very high Luminosity  KEKB/PEP2

11. 11 BB pair production: Upsilon e + e – ->  (4S) -> BB + 20 MeV B + :B 0 ~1:1 Event rate dN dt =  x L ~ 10 s – 1 ~10 8 yr – 1 Cross section ~ 1 nb = 10 -33 cm 2 Luminosity (collision rate) 10 34 cm – 2 s – 1 (design; currently 5.5x10 33 @KEKB) Currently@Belle: 3x10 7 BB events (published), 4.8x10 7 on tape CLEO 8.0 GeV e – + 3.5 GeV e + IP size = 77µmx2.0µmx4.0mm KEKB:

12. 12 Time measurement at  (4S) e-e-e-e- e+e+e+e+   B2B2B2B2 t=0  z≈  t  c B1B1B1B1 ~200 µm J/  KsKsKsKs flavor tag: e, µ, K ±,...  (4S): CP=-1, conserved   + e - ->  (4S) identify b/b {flavor tag} until first B decay (t=0) Reconstruct CP=±1 mode @ t=  t

13. 13 Detector: e.g. Belle Charged tracking/vertexing - SVD: 3-layer DSSD Si µstrip (~55 µm) – CDC: 50 layers (He-ethane) Hadron identification – CDC: dE/dx (~7%) – TOF: time-of-flight (~95 ps) – ACC: Threshold Cerenkov (aerogel) Electron/photon – ECL: CsI calorimeter (1.5%@1 GeV) Muon/KL – KLM: Resistive plate counter/iron Designed to measure CP asymmetry

14. 14 Belle Collaboration 274 authors, 45 institutions many nations many nations

15. 15 B 0  J/  K s (  +   ) “golden mode” 1lepton+1”not-hadron” K s  +    ~4MeV/c 2 K s mass  4  J/  l + l  CP mode reconstruction

16. 16 B 0  J/  K s (continued)  ~3MeV/c 2  ~10MeV Kinematics for final selection:  E  E* cand –E* beam  0 (E* beam  s 1/2 /2) 10-50 MeV res, depends on mode M bc (Beam-constrained mass) M bc  (E* beam 2 -p* cand 2 ) 1/2 Signal region 457 events ~3% background EE M bc

17. 17 Other charmonium+K B 0  J/  K s (  0  0 ) J/  K S (  +   &  0  0 )  (2S)(  l + l  & J/  +   ) K S  c1 (  J/  ) K S  c (  K S K   , K + K    ) K S J/  K L J/  K* 0 (  K L    (mostly)   f = -1   f = +1 76 events ~ 9 bkg

18. 18 Other charmonium M(l  l   )  M(l  l  ) 1st observation of inclusive B   c2 X  c2  c1

19. 19  J/  K L J/  : {tight mass cut} 1.42<p  *<2.00 GeV/c K L : {KLM/ECL cluster w/o track, >1 KLM superlayers (resolution~ 3° (1.5° if ECL) } within 45˚ of expected lab direction Require cand to have B mass, calculate momentum in CMS (p B *) (~0.3 GeV for signal) backgrounds: random (from data), “feeddown,” known modes - estimate via MC

20. 20 CP candidates Fully reconstructed modes N sig = 346 events N bkg = 223 evts J/  K L 750 candidates ~58 bkg

21. 21 Flavor taggingbcs l-l-l-l- l+l+l+l+ K–K–K–K– D0D0 π+π+ D *+ – high-p lepton (p*>1.1 GeV): b-> l - – net K charge b->K – – medium-p lepton, b->c-> l + – soft π b->c{D *+ ->D 0 π + } * multidimensional likelihood,  >99% Significance of CP asymmetry depends on – tagging efficiency  – wrong-tag fraction w (measured w data) - effective efficiency =  (1-2w)

22. 22  K-K- K-K-  z: vertex reconstruction  Constrained to measured IP in r-  B CP :  z ~75 µm (rms) use only tracks from J/     B tag :  z ~140 µm (rms) remaining tracks, excluding K s ; iterate, excluding tracks w. poor  2 /n resolution includes physics (e.g. charm) Overall eff. = 87% zz

23. 23 Raw  t distributions CP is violated!!! seen in raw data large effect distribution in  t~  z/  c, unbinned max. likelihood fit

24. 24 Prepare to fit for sin2  1 multiply by  for lab length (decay in flight) } B 0 lifetime = 1.548±0.032 ps, c  =464±10 µ m mixing  m = 0.47±0.02 ps – 1 ; cT~4.0 mm only ~ 1 cycle of oscillation measurable True CP asymmetry is diluted: background to CP reconstruction incorrect flavor tag rate vertex resolution - not exactly as modeled all need checks in data -> Use same methods to make other (better known) physics measurements: B 0 mixing, B lifetime, D lifetime, null CP

25. 25 Wrong tag fraction via mixing Same fit method, but CP->flavor-specific B  D *- l +, D (*)- π +, D *-  +  +flavor tag separate same-, opp-flavor events fit to  z: mixing asymmetry, w: "effective tagging efficiency"  eff =  (1-2w l ) 2  tag, l =(27.0±2.2)% 99.4% of candidates tagged (good agreement w MC) l Flavor tags classified by (MC) Purity - 6 bins

26. 26  t resolution function Double Gaussian, parameters calculated event- by-event, includes effects of - detector resolution - poorly measured tracks - bias from e.g. charm - approximation of  t=  z/  c form, parameters from - Monte Carlo - fits for D 0  K - π +, B  D * l lifetimes validate: B lifetime, same fitting  0 =1.55±0.02 ps (PDG2000: 1.548±0.032 ps)  + =1.64±0.03 ps (PDG2000: 1.653±0.028 ps) tail fraction: 1.8%

27. 27 Fitting  t distribution distribution in  t~  z/  c unbinned max. likelihood fit, includes - signal root distribution (analytic) - wrong tag fraction (const) - background: right & wrong tag (MC, parametrized) - detector & tagging  z resolution (parametrized,evt-by-evt)

28. 28 Results All modes combined: sin2  1 =0.99±0.14(stat)+0.06(sys) binned in  t curve from unbinned fit N B  N B N B +N B ff weighted by CP sensitivity

29. 29 Control sample: B 0  non-CP states (statistical error only) use: B 0  D (*)- π +, D *-  +, D *- l +, J/  K*(K + π - ) “sin2  1 ” 0.05  0.04

30. 30 fit CP –1 and CP+1 separately: sin2  1 0.84  0.17 1.31  0.23 (statistical errors only) CP=-1 CP=+1 opposite!! N B  N B N B +N B weighted by CP sensitivity

31. 31 Systematic errors Vertex algorithm  0.04 Flavor tagging  0.03 Resolution function  0.02 K L background fraction  0.02 Background shapes  0.01  m d and  B0 errors  0.01 Total  0.06

32. 32 Compare with other experiments

33. 33 Result in context Belle ’ s 1  band (two sol ’ ns) V ub BaBar (locating tip of unitarity triangle)

34. 34 Summary/Prospects Successful run of Belle in 2000-1 sin 2  1 : 30.5 fb – 1 on  (4S), 1137 tagged events sin 2  1 : 30.5 fb – 1 on  (4S), 1137 tagged events 19 papers published or submitted 19 papers published or submitted Next higher precision on sin2  1 data as of 1/23/02 - 48 fb – 1 ; anticipate 100 fb – 1 by summer Lum: peak 5.5x10 33 cm – 2 s – 1 ; 24 hrs 311 pb – 1 ; month 6120 pb – 1 other angles - need >300 fb – 1 - within sight!