Mixing and CP Violation in neutral D meson

Mixing and CP Violation in neutral D meson system BESIII 物理分析讲习班 2008年 6月26日 郑阳恒 中国科学院研究生院 Y. Zheng (GUCAS) 06/26/2008

Some notations èCPV: CP violation èDCS: Doubly Cabbibo Suppressed èCF: Cabbibo Favored ( CA: Cabbibo Allowed ) èSM: Standard Model èMC: Monte Carlo èRS: Right Sign D 0→K-π+ èWS: Wrong Sign D 0→K+π2 Y. Zheng (GUCAS) 06/26/2008

Outline èIntroduction (formalism and experimental view) èExperimental Apparatus & Analysis Strategy èMeasurement status èSummary and Future perspective 3 Y. Zheng (GUCAS) 06/26/2008

Motivation èUniverse in 15 billion years ago: high degree of symmetry between matter and antimatter èMatter encounters antimatter annihilated èPresent-day Universe: matter >> antimatter, Why? èStandard Model: CP violation is the KEY èNeutral meson mixing CPV 4 Y. Zheng (GUCAS) 06/26/2008

A little history è Before 1956, discrete space-time symmetries were considered to be exact è 1956, T. D. Lee and C. N. Yang suggested P violation. (Nobel prize) è 1957, C. S. Wu et al. , discovered P violation experimentally. è 1964, Christenson et al. , discovered CP violation in the neutral K system. (Nobel prize) è 1973, Kobayashi and Maskawa: KM model with 6 quark flavors CP violation è 1980, Carter, Sanda and Bigi: Large CP asymmetries in B decays. è 2001, B-factories observed CP asymmetries in B decays. è How about D decays? 5 Y. Zheng (GUCAS) 06/26/2008

Starting point: Standard Model è Described the elementary building blocks of matter and interactions. è Account for all experimental phenomena to a high degree of precision. è Many predictions verified experimentally. è A successful theory. è CP violation is one of the least constrained sector. è measurements of Mixing and CP violation in neutral D meson system provide a sensitive testing method. 6 Y. Zheng (GUCAS) 06/26/2008

CKM Picture of CP Violation è In Standard Model: è Cabibbo, Kobayashi & Maskawa: CKM matrix è SU(2) doublets in standard electroweak model: u, c, t d’, s’, b’ 7 Y. Zheng (GUCAS) 06/26/2008

Lagrangian è In a physics system: Lagrangian the dynamics èQuark flavor transitions: q Wq’ q W q’ èLagrangian of charged weak current via W Under the CP transformation èVij Vij* (at least 1 non-real) CP-violation (only source in SM) 8 Y. Zheng (GUCAS) 06/26/2008

Parameterization of CKM matrix free parameters 9 complex matrix elements 9 2 = 18 Unitarity of the matrix, V+V = I -9 5 arbitrary phases of 6 quark field -5 total 4 è Wolfenstein parameterization(A, ρ, η, Cabibbo angle ~ 0. 22): 9 Y. Zheng (GUCAS) : complex phase CPV 06/26/2008

Mixing in neutral meson system “Box diagrams” for second order weak processes: d, s, b u c __ D 0 W W D 0 _ _ c u d, s, b b(s) __ __ B 0(s) (K 0) W _ _ d(s) 10 u, c, t Time evolution: d(s) W u, c, t Mass and flavor eigenstates: B 0(s) (K 0) __ b(s) Y. Zheng (GUCAS) 06/26/2008

Time dependent decay amplitudes è For any decay final state f, |f CP|f 11 Y. Zheng (GUCAS) 06/26/2008

Time dependent decay rates è For any decay final state f, |f CP|f è If CP is conserved: 12 Y. Zheng (GUCAS) 06/26/2008

CP violation categories è CP violation in decay (direct CP violation) : observed in B decays è CP violation in mixing (indirect CP violation) : observed in neutral K system è CP violation in interference between decay and w/o mixing (mixing-induced CP violation) : observed in neutral B system 13 Y. Zheng (GUCAS) 06/26/2008

neutral D decays (to common states) D 0 phys(t) f D 0 For convenient: 14 Y. Zheng (GUCAS) 06/26/2008

D→Kπ (non-CP eigenstates) + u D 0 K - + is a Cabibbo-allowed favored decay mode (Br=3. 8%) d W c u s u K- There are 2 ways that a D 0 can decay to the opposite combination K+ -: 1) Doubly Cabibbo-suppressed 2) decays (DCS) u 2) D 0; D 0 K+ - decays + K d s c 0 D u W Br ~ 0. 014% 15 d u - u u 0 c D mixing Y. Zheng (GUCAS) c D 0 u W s u K+ Cabibbo-allowed D 0 decay (CA) 06/26/2008

D→Kπ (non-CP eigenstates) è f = K -π + , f = K + π δ: strong phase diff. (assume CPV via weak phase only), φ: weak phase RS WS 16 Y. Zheng (GUCAS) 06/26/2008

Mixing and CPV: D→Kπ è If CP is conserved: fit WS distribution → Mixing è For CPV case: fit two WS distribution separately èDirect CP violation in DCS Decay èCP violation in mixing èCP violation in interference between decay and mixing: 17 Y. Zheng (GUCAS) 06/26/2008

CP eigenstates: K+K-, π+πè CP eigenstates, a simpler case: |f CP|f = ±|f è CP Asymmetries: (Integrated time-dependant decay rates over time, assume no CPV in decay, AD=0) 18 Y. Zheng (GUCAS) 06/26/2008

Requirements for Measurements è Large D meson source ( Br(D f) ~ 10 -2 - 10 -3) èvery high luminosity e+e- collider B-factories, BEPCII è B meson reconstruction èhigh quality ~4 detetor Belle, Ba. Bar, BESIII è Tag flavor of the D meson ègood particle id d. E/dx, Cherenkov, TOF, EMC è Measure proper-decay-time difference (Belle, Ba. Bar) èhigh precision vertexing (Δz) silicon strip vertex detector èLikelihood fit to the t distributions è Measure Branching Ratios (BESIII) 19 Y. Zheng (GUCAS) 06/26/2008

Colliders: KEKB and PEP-II Coherent BB production KEKB: 8 x 3. 5 Ge. V, bg=0. 43 PEP-II 9 x 3. 1 Ge. V, bg=0. 55 s(BB)/shad=0. 28 B B production threshold 20 Asymmetric e-e+ Colliders@ U(4 S) Y. Zheng (GUCAS) 06/26/2008

Detectors: Ba. Bar and Belle ECL KLM e. V . 1 G 3 e + e. V e- 9 G DIRC: Quartz bar + water tank SVT: 5 -layer … ACC: aerogel Cherenkov counters SVD: 3 -layer. (4 -layer soon) … high performance vertexing and PID 21 Y. Zheng (GUCAS) 06/26/2008

A typical collision event è Signals: èD*+ D 0 + K - + K +K - + è Variables used to separate signal and background è D invariant mass: M(K , KK, ) è m = M(K tag) – M(K ) 22 Y. Zheng (GUCAS) 06/26/2008

Proper decay time è Signals: D*+ → D 0(→Kp, KK, pp) p+tag èt=(ldec/p)(m/c) 23 Y. Zheng (GUCAS) 06/26/2008

Analysis Strategy in B-factories è Signals: D*+ → D 0(→Kp, K K, pp) p+tag è Backgrounds: shapes from MC, fractions from data è Fitting method: Unbinned maximum likelihood fit. è Proper-time distribution: èResolution function: from RS data fitting èFit WS data event-by-event è Several proper time fits are performed. èlifetime (no mixing) èmixing, no CP violation èmixing, CP violation è Monte Carlo: search for systematics and validate statistical significance of results. 24 Y. Zheng (GUCAS) 06/26/2008

Signal selection criteria è Beam-constrained vertex fits of K, , tag tracks. è tag charge gives D flavor at production. è Require fit probability > 0. 001 è D 0 selection èCMS p(D*) > 2. 5 Ge. V/c to eliminate D 0’s from B decays è K, particle identification è DCH hits > 11 è 1. 81 < M(K ) < 1. 92 Ge. V/c 2 è decay time error < 0. 5 ps è -2 < decay time < 4 ps è ptag è CMS p* < 0. 45 Ge. V/c è lab p > 0. 1 Ge. V/c è SVT hits > 5 25 Y. Zheng (GUCAS) 06/26/2008

Background shapes è True D 0 combined with a random πs èpeaked at M(K ) èdoes not peak in m è Mis-recon D 0 èpeaked at m èdoes not peak in M(K ) è Combinatorial background èdoes not peak in both M(K ) and m è Shapes: from MC è Yields: from 2 -D fit 26 Y. Zheng (GUCAS) 06/26/2008

RS(top)/WS(bottom) Datasets After Event Selection Integrated Luminosity Approximately 384 fb-1 x 103 Ba. Bar Data 64, 000 WS candidates 27 Y. Zheng (GUCAS) 06/26/2008 2 events/0. 1 Me. V/c 2 events/1 Me. V/c 1, 229, 000 RS candidates

Resolution function è There is no perfect measurement! è Likelihood for proper decay time measurement: è Resolution function models: (from RS signal fit) èSignal: triple-Gaussian model èrandom πs and Mis-recon D 0 : triple-Gaussian model (same as signal) è Combinatorial background model: double. Gaussian (from sideband) 28 Y. Zheng (GUCAS) 06/26/2008

Validation: RS lifetime Ba. Bar Data The D 0 lifetime is consistent with the Particle Data Group value, within the statistical and systematic errors of the measurement. Plot selection: 1. 843<m<1. 883 Ge. V/c 2 0. 1445< m< 0. 1465 Ge. V/c 2 29 Y. Zheng (GUCAS) 06/26/2008

WS Mixing Fit: No CP Violation èVaried fit parameters Ba. Bar Data èMixing parameters èFit class normalizations èCombinatoric shape Mixing minus No mixing PDF Data minus No mixing PDF Ba. Bar Data Plot selection: 1. 843<m<1. 883 Ge. V/c 2 0. 1445< m< 0. 1465 Ge. V/c 2 30 Y. Zheng (GUCAS) 06/26/2008

Mixing Contours: No CP Violation èy’, x’ 2 contours computed by change in log likelihood Ba. Bar Data èBest-fit point is in non-physical region x’ 2 < 0, but onesigma contour is in physical region ècorrelation: -0. 95 èAccounting for systematic errors, the no-mixing point is at the 3. 9 -sigma contour RD: (3. 03 0. 16 0. 06) x 10 -3 x’ 2: (-0. 22 0. 30 0. 21) x 10 -3 y’: (9. 7 4. 4 3. 1) x 10 -3 31 Y. Zheng (GUCAS) 06/26/2008

Results for K Analysis 32 Y. Zheng (GUCAS) 06/26/2008

D 0 reconstruction and lifetime fit 33 Y. Zheng (GUCAS) 06/26/2008

ycp , AΓ results 34 Y. Zheng (GUCAS) 06/26/2008

What about Charm factory? èCharm events at threshold are very clean èRatio of signal to background is optimum èLots of systematic uncertainties cancellation while applying double tag method èMixing at threshold èBad news: no time-dependent information èGood news: Quantum coherence, CP tags èThe coherence of two initial D allows simple methods to measure DDbar mixing, strong phase and CP violation 35 Y. Zheng (GUCAS) 06/26/2008

BEPCII/BESIII experiment èWill collect collision data in July! èWill operate at 36 Y. Zheng (GUCAS) 06/26/2008

Coherent D 0 – D 0 states è (3770) D 0 D 0 / (3770) D+D- 50/50 (3770) : spin=1, cc bound state, Mass: 3. 771 Ge. V D 0 : spin=0, Mass: 1. 864 Ge. V D mesons created at rest in CM DD orbit angular momentum L=1 è Bose statistics D 0 D 0 state anti-symmetric D 0 D 0 and D 0 D 0 are prohibited è At any time : one D 0 until one D decays 37 Y. Zheng (GUCAS) 06/26/2008

D Pairs at Different Experiments 128 M is expected at BES-III with 4 years’ luminosity. 5 M is expected at CLEO-c until 2008. 700 M 500 M 0. 2 M 38 5 M 128 M (3770) peak (4 S) Peak Background free Higher statistics Y. Zheng (GUCAS) 06/26/2008

Charm tags èSingle tags èreconstruct one D meson èDouble tags èBoth D and Dbar are reconstructed èFlavor tags in Mixing language èSemileptonic modes: K( )en èCP tags èCP even èCP odd 39 Y. Zheng (GUCAS) 06/26/2008

K, Identification at BESIII 40 Y. Zheng (GUCAS) 06/26/2008

RM measurements @3. 773 Ge. V èGolden channel èD 0 K èSemileptonic channel èKen, Kmn, etc 2 -body identical final states are Required in both D hadronic decays 41 theory experiment Double tag measurements. Number of R. S. tags at BESIII are expected to be 104 -105, the sensitivities of Rmix will @ 10 -4— 105 Y. Zheng (GUCAS) 06/26/2008

CP eigenstate Tags èCP – èCP + èKS 0(0. 012) èKsh (3. 9 X 10 -3) èKS h' (0. 0094) èKSr 0 (0. 0078) èKsw (0. 012) èKS (4. 7 X 10 -3) èK+K- (3. 89 X 10 -3 ) è + -(1. 38 X 10 -3 ) èKs 0 0 è 0 0(8. 4 X 10 -4) èKSKS (7. 1 X 10 -4) èr 0 0(3. 2 x 10 -3) Dalitz Analysis In 20 fb -1 (3770) data, we can get > 4. 5 x 10 5 CP+ tags and > 3. 6 x 10 5 CP- tags With large sample of CP tags, we may improve the measurements of strong phase, probe the direct. Y. CP, and other mixing Zheng (GUCAS) 42 For Ks modes: CPV effect of Ks need to be considered! (Prof. Xing/Zhizhong’s suggestion) 06/26/2008

CP Violation 1. Direct CP Violation (in decay) 2. Indirect CP Violation (in mixing) 3. CP violation in the interference between decays with/without mixing 43 Y. Zheng (GUCAS) 06/26/2008

Quantum Coherence Suppose Both D 0 decay to CP eigenstate f 1 and f 2. Thus if a final state such as (KK)( ) observed, we immediately have evidence of CP violation In 20 fb-1 y(3770) data, > 1000 double CP+ and CPtags can be obtained. if 100%CPV, it lead to ACP~10 -3 level 44 Y. Zheng (GUCAS) 06/26/2008

Unitarity Triangle B 0 pp B 0 rp - (r, h) Vub*Vud+Vcb*Vcd+Vtb*Vtd = 0 (B system) 45 B 0 D(*) p B+ DK Y. Zheng (GUCAS) B 0 J/y. Ks B 0 f. Ks B 0 D(*) 06/26/2008

3 from B- D 0 K- è No hadronic uncertainty è Methods è Gronau-Wyler original method è Atwood-Dunietz-Soni Method è Dalitz method è Problem: statistics 46 Y. Zheng (GUCAS) 06/26/2008

Gronau-Wyler original method 3 3 δB è Theoretically clean è Experimentally challenging èHadronic D decay modes: hard for D flavor tagging èSemi-leptonic D decays : Background too high èCP eigenstate decays of D: small Branching ratio 47 Y. Zheng (GUCAS) 06/26/2008

Atwood-Dunietz-Soni Method è Use interference between èB+ DK+ and B+ DK+ follows by D (D) f è To get a common final state f, we need èDouble Cabibbo Suppression (DCS): f = K+ - , K+ Kè K - K mixing: f = KS 0 , KS + - è D hadronic parameters: è Decay rates: è r. D, δD : measured from Charm factory (see next slides) è (r. B, δB, 3 ) 3 unknowns, 4 measurements 3 48 Y. Zheng (GUCAS) 06/26/2008

δD from Charm-factory è Get r. D from the large tagged D decay samples (B- factory or Charm factory (CLEO-c sensitivity: ~0. 05 from 3 fb-1)) è δD Charm factory on (3770) accurately measured (Soffer hep-ex/9801018) è Reconstruct Double Tags: CP and f è CP+: K+ K-, + -, Ks 0 0 è CP- : Ks 0 , Ks è Asymmetry in CP+ and CP- of D decays: è Input RD= r. D 2 from PDG èBESIII sensitivity: <0. 06 from 20 fb-1 for cos D 49 Y. Zheng (GUCAS) 06/26/2008

Dalitz method èThree body D decays: KS + -, + - 0, KSK+K-… èEffect of D – D interference 50 Y. Zheng (GUCAS) 06/26/2008

Formalism (Giri, Grossman, Soffer, Zupan) è B (KS + -)D K (hep-ph/0303187) è D hadronic parameters è Partition the Dalitz plot to 2 k bins è Label bins below symmetry axis i, above axis i S 13 unknown Measurable from tagged D S 12 51 Y. Zheng (GUCAS) 06/26/2008

3 extraction è 2 k bins 2(B modes) = 4 k equations èFor the ith bin: è 2 k+3 unknowns: ci, si, r. B, δB, 3 Solvable for k 2 èBelle results from Dalitz method. D Decay in 2005: model Systematic Uncertainty 52 Y. Zheng (GUCAS) 06/26/2008

ci , si from Charm-factory è D double tag: (KS + - vs General state: g) è If g= KS + - and j=i c 2 i +s 2 i è If g=CP sgj=0 , g g g T j = c j ci è Belle studied relationship between systematic error on 3 and # of CP tagged KS + - events in Charm factory (BESIII) è 2000 CP+ and CP- tagged events δ 3 (sys)~ 1 o -2 o 53 Y. Zheng (GUCAS) 06/26/2008

Summary è Ba. Bar using Kπ final state, finds a mixing signal at the 3. 9 sigma confidence level (assuming CP conservation and including systematic effects): RD: (3. 03 0. 16 0. 06) x 10 -3 x’ 2: (-0. 22 0. 30 0. 21) x 10 -3 y’: (9. 7 4. 4 3. 1) x 10 -3 No evidence is seen for CPV. è Belle using CP eigenstate (K+K-, π+π-) final states, finds mixing evidence (>3 sigma) : ycp=1. 31± 0. 32± 0. 25% AΓ=1. 31± 0. 32± 0. 25% (no evidence for CPV) è BES-III contributions will be coming soon (getting more interesting)! 54 Y. Zheng (GUCAS) 06/26/2008

Future persepctive Sensitivities (20 fb-1): èMixing parameters =(x 2+y 2)/2 è < K and Ken channels èProbe y: Δy. CP < 0. 7%, èΔcos K < 0. 06 10 -4 in BESIII èCP Violation èΔACP~10 -3 in D+ decays (direct CPV), ècontributions to 3/ errors: <2 o (CLEO-c: ~5 o) 55 Y. Zheng (GUCAS) 06/26/2008

谢谢！ Y. Zheng (GUCAS) 06/26/2008