CP Asymmetry in B 0 ->π + π – at Belle

CP Asymmetryin B0->π+π– at Belle • B0->π+π– and CP asymmetry in CKM • e+e– -> (4S) at KEKB and Belle • • Belle data • Measurement of CP asymmetry in B0->π+π– • Interpretation vis-a-vis CKM • •Future Kay Kinoshita University of Cincinnati Belle Collaboration

B0->π+π– involves f2 (a) of CKM: CKM: matrix of W-quark couplings - 3x3, unitary Vub*Vud Vcb*Vcd Vtb*Vtd Vcb*Vcd One condition of unitarity: + 1 + = 0 -(rih) -(1-rih) Represented in complex plane as "unitarity triangle" a  

B0->π+π– 2 paths, each w/wo mixing: VtdVtb* –VudVub* f2 =arg Tree Penguin aVub*Vud aVtb*Vtd mixing+ " aVtb*2Vtd2VubVud* aVtb*2Vtd2VtbVtd* Bottom line: CP-asymmetric time-dependent rate from x-terms "direct" asym

+0.38+0.16 –0.27–0.13 +0.25+0.09 –0.31 –0.09 Sππ= –1.21 Aππ= +0.94 Each 2.9s from zero; note physical region is Uncertainty: relative amplitudes of Tree, Penguin • if T dominates, Aππ=0, Sππ=sin2f2 • if P, T comparable, Aππ≠0, Sππ~sin(2f2+2q)•2/(|l|2+1) Direct CP violation difference of strong phase ≠1 if direct CP violation Previous Belle result {PRL 89, 071801 (2002)} (42 fb–1 ~45M B pairs) Now: • more data - 78 fb–1 • improved analysis - tracking, Dt resolution, event selection • statistical analysis (total 126 fb–1, ~1.3x108 B events)

+e- -> (4S) {bg = 0.425} t=0 Dz≈Dtbgc CP mode @ t=Dt ~200 µm B2 e- e+ B1 flavor tag @t=0: e, µ, K±, ...  CP=–1, conserved first B decay (t=0), break CP B production: } B As with sin2f1 via J/yK: • reconstruct CP mode • tag flavor • reconstruct vertices • unbinned max. likelihood fit to Dt BB threshold

Belle detector Charged tracking/vertexing - SVD: 3-layer DSSD Si µstrip – CDC: 50 layers (He-ethane) Hadron identification – CDC: dE/dx – TOF: time-of-flight – ACC: Threshold Cerenkov (aerogel) Electron/photon – ECL: CsI calorimeter Muon/KL – KLM: Resistive plate counter/iron

…the people 274 authors, 45 institutions many nations

B0->π+π–reconstruction final selection: DE E*cand–E*beam: 0±0.057 GeV(E*beams 1/2/2) {Kπ shift –45 MeV} Mbc (E*beam2-p*cand2 )1/2: 5.271 –5.287 GeV/c2(Beam-constrained) … but less clean than B0->J/Ks: • "physics bg" B0->K+π– => hadron ID, kinematics dE/dx, TOF, Aerogel –“positive ID”eπ=91%, eK=10% • continuum=> event shape {qq “jet-like” vs BB “spherical”) Fisher discriminant from modified Fox-Wolfram moments B candidate direction relative to beam axis Construct Likelihood ratio LR=LBB/[LBB+Lqq], 2 selections: LR > 0.825 {eBB=53%, eqq=5%} 0.825 > LR > LRmin(cut depends on flavor tag classification)

ππ:106±16 Kπ:41±10 qq:128±6 ππ:57±8 Kπ:22±6 qq:406±17 Total signal 163±24 B0->π+π– Candidates 760 in signal box - 391 B0, 369 B0 0.825>LR>LRmin LR>0.825 ππ ππ

Flavor tagging: same as for sin2f1 or π– • 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*+->D0π+} • hard π b->{c}π–X * multidimensional likelihood, e>99% l- l+ b c s K– D*+ D0 π+ • wrong-tag fraction w classify events based on expected w (MC) - 6 bins. (B0 mixing amplitude in data) => • effective efficiency = e(1-2w): net (28.8±0.5)% mixing amplitude <-> w

Dz vertex reconstruction: same as for sin2f1 Resolution function: validate via lifetime <= tB0 = 1.551±0.018 ps (PDG02: 1.542±0.016) B0->D+p-, D*+p-, D*+r-, J/yKS, J/yK*0 sDt~1.43 ps (rms) z m+ m- K- Dz K-

More checks of Dt resolution+flavor tag B0->K+π– mixing Dmd = 0.55±0.07 ps–1 (PDG02: 0.489±0.008) tB0 = 1.42±0.14 ps (PDG02: 1.542±0.016) tB0 = 1.46±0.08 ps Fitted bg agrees w sideband

Check for flavor bias Look where zero asymmetry expected: "Sππ"= –0.045±0.033 "Aππ"= -0.015±0.022 SKπ= –0.03±0.11 AKπ= +0.08±0.16

Fitting for CP asymmetry Same technique as with sin2f1 • unbinned maximum likelihood fit • resolution function event-by-event: tracking, misreconstruction, physics, approximation of Dt=Dz/bgc • wrong tag fraction w, backgrounds Fit for Aππ, Sππ : root Diluted +Kπ (set AKπ = 0) +resolution + bg

Fit Results 0 0 -5 -5 5 5 +0.08 –0.07 Sππ= –1.23±0.41 Aππ= +0.77±0.27±0.08 (stat) (sys) still outside physical region => investigate … display Dt projection (78 fb–1 ~85M B pairs) LR>0.825 Likelihood not parabolic -> statistical errors estimated numerically via MC ensemble Background subtracted raw asymmetry

Fit Results: statistical analysis check linearity: generated vs fitted Aππ, Sππ • MC ensemble - 30k expts, 760 events ea, Aππ=0.569, Sππ=–0.822 probability of being outside physical boundary =60.1% " further (in s) from (0,0)=16.6% data physical boundary Aππ x Sππ

Confidence regions • Feldman-Cousins frequentist approach. • Acceptance regions from MC ensembles. • Systematic errors included. • Confidence Level (CL) calculated at each point. interpret as evidence for CP non-conservation in B0->π+π– } (3.4s) } hint of direct CP non-conservation (2.2s)

Find: 78°≤ f2≤ 152° (95% C.L.) insensitive to d From other CKM (CKM fitter group, 2002): 78.3°≤ f2≤ 121.6° (95% C.L.) => consistent Constraints on the CKM angle f2 • Sππ, Aππ depend on 4 parameters: f2, f1[21.3°-25.9°], |P/T|[0.15-0.45], d -> plot confidence contours in (f2, d)for various |P/T| e.g. |P/T|=0.3 f1 =23.5° f2 d

Summary Belle, 2000-2: • peak L= 9.5x1033cm–2s–1 - nearly at design (1x1034cm–2s–1) • passed 100 fb-1 in Oct. 2002 • with 78 fb–1 on (4S), sensitive to large values of sin22 • measure CP asym in B0->π+π– constraints on 2, consistent with other CKM constraints. hint of direct CP non-conservation result submitted to PRD. Next • ->150 fb–1 by summer, 500 fb–1 by 2005 • Luminosity >@ design • the CP challenge: stay tuned on 2

additional slides

Systematic uncertainties* * blind analysis: actual estimations done before seeing fit result.

Constraints on the CKM angle f2 convention taken from M.Gronau & J.L.Rosner Phys Rev D65, 093012 (2002) 4 parameters |P/T| 0.15-0.45 (representative) Theory ~0.3 f1 21.3 - 25.9deg (Belle & BaBar combined)

CP Asymmetry in B 0 ->π + π – at Belle