CP-Violating Asymmetries in Charmless B Decays: Towards a measurement of  James D. Olsen Princeton University International Conference on High Energy Physics Amsterdam, July 24-31, 2002 On behalf of the BaBar Collaboration CP asymmetries in     and K +   Decay rates for     and     CP asymmetries in     and   K  Submitted to Phys Rev (hep-ex/ ) hep-ex/ and hep-ex/ hep-ex/

ICHEP 2002 J. Olsen 2 CP Violation in the Standard Model Unitarity Triangle          V td V tb * V ud V ub * V cd V cb * B  ,  B   J/  K S B   DK CP symmetry can be violated in any field theory with at least one irremovable complex phase in the Lagrangian This condition is satisfied in the Standard Model through the three- generation Cabibbo-Kobayashi-Maskawa (CKM) quark-mixing matrix The angles ( , ,  ) are related to CP-violating asymmetries in specific B decays  One down, two to go… ICHEP 2002

J. Olsen 3 Observing CP violation at the  (4S) At the  S), BB  pairs are produced in a coherent P-wave Three observable interference effects:  CP violation in mixing (|q/p| ≠ 1)  (direct) CP violation in decay (|A/A| ≠ 1)  (indirect) CP violation in mixing and decay (Im ≠ 0) Observable in time evolution of B  B  system (assume  ) direct CP violation  C ≠ 0 indirect CP violation → S ≠ 0

ICHEP 2002 J. Olsen 4 CP Violation in B  →     With Penguins (P): mixing decay Tree (T) Level: Need branching fractions for    ,    , and     to get  from  eff → isospin analysis

ICHEP 2002 J. Olsen 5 Overview of Analyses Analysis issues: charmless B decays  Rare decays! BR ~ 10  -10   → need lots of data (PEP-II)  Backgrounds: Large background from e  e  → qq → need background suppression Modes with   suffer backgrounds from other B decays  Ambiguity between  and K → need excellent particle ID (DIRC) Time-dependent CP analysis issues:  Need to determine vertex position of both B mesons → silicon  Need to know the flavor of “other” B → particle ID We use maximum likelihood (ML) fits to extract signal yields and CP-violating asymmetries  Kinematic and topological information to separate signal from light- quark background  Particle ID to separate pions and kaons The data sample corresponds to 87.9 million BB pairs

ICHEP 2002 J. Olsen 6 K/  Separation with the DIRC Cherenkov angle  c used in the maximum likelihood fit to distinguish pions and kaons Resolution and K-  separation measured in data Kaon sample K hypothesis  hypothesis

ICHEP 2002 J. Olsen 7 Analysis of B → , K , KK Analysis proceeds in two steps:  Time-independent fit for yields and K  charge asymmetry  Time-dependent fit for S , and C  Kinematically select B candidates with m ES,  E Suppress qq background with Fisher discriminant Fit yields and charge asymmetry **   p*p*

ICHEP 2002 J. Olsen 8 Branching Fraction Results ModeYieldBR (10 -6 ) A CP (K  ) B  →     B  → K    B  → K  K   KK KK Projections in m ES and  E Preliminary KK 87.9  1.1 million BB Submitted to Phys Rev (hep-ex/ )

ICHEP 2002 J. Olsen 9 Vertex Reconstruction Resolution function parameters obtained from data for both signal and background  Signal from sample of fully reconstructed B decays to flavor eigenstates: D * ( , , a  )  Background from data sidebands Beam spot Interaction Point B REC Vertex B REC daughters Exclusive B rec reconsctuction B TAG direction TAG tracks, V 0 s z B TAG Vertex B →  Example in B →  e  e  → qq  z resolution dominated by tag side → same resolution function as charmonium (sin2  ) sample Average  z resolution ~ 180  m

ICHEP 2002 J. Olsen 10 B Flavor Tagging New tagging algorithm with physics-based neural networks  Inputs include leptons, kaons, slow-  (from D*), and high-momentum tracks  Outputs combined and categorized by mistag prob (w) 5 mutually exclusive categories:  Lepton – isolated high-momentum leptons  Kaon I – high quality kaons or correlated K  and slow-    Kaon II – lower quality kaons, or slow-   Inclusive – unidentified leptons, poor-quality kaons, high- momentum tracks  Untagged – no flavor information is used b sc K-K- ~7% improvement in Q  w  

ICHEP 2002 J. Olsen 11 Tagging in Charmless B Decays Tagging efficiency is very different for signal and bkg  Strong bkg suppression in categories with the lowest mistag prob (Lepton/Kaon)  Different bkg tagging efficiencies for , K , KK 81/fb B→ h  h  sample split by tagging category CategorySignal  KK KK Lepton Kaon I Kaon II Inclusive Untagged Tagging Efficiencies (%) Background

ICHEP 2002 J. Olsen 12 Validation of Tagging, Vertexing, and ML Fit K  decays are self-tagging  T = tag charge  Q = kaon charge Float  and  m d in same sample used to extract CP asymmetries: Fit projection in sample of K  -selected events

ICHEP 2002 J. Olsen 13 CP Asymmetry Results Fit projection in sample of  -selected events qq + K  Preliminary Submitted to Phys Rev (hep-ex/ )

ICHEP 2002 J. Olsen 14 Cross-checks Inspect  -selected sample  2-param fit consistent with full fit  asymmetry vs. m ES Asymmetry in yields consistent with measured value of C , but does not suggest large direct CP violation Toy MC generated over all allowed values of S  and C   Expected errors consistent with data  No significant bias observed Validated in large samples of signal and background MC events Systematic errors dominated by uncertainty in PDF shapes A  vs. m ES in sample of  -selected events signal bins

ICHEP 2002 J. Olsen 15 Taming the Penguins: Isospin Analysis The decays B    ,    ,     are related by isospin Central observation is that  states can have I = 2 or 0  (gluonic) penguins only contribute to I = 0  I = 1/2)      is pure I = 2  I = 1/2) so has only tree amplitude  (|A     A   ) Triangle relations allow determination of penguin- induced shift in  Gronau and London, Phys. Rev. Lett. 65, 3381 (1991) But, need branching fractions for all three decay modes, and for B  and B  separately

ICHEP 2002 J. Olsen 16 The Base of the Isospin Triangle: B  →     Analysis issues:  Usual charmless two-body; large qq background,  /K separation  Potential feeddown from     Minimize with tight cut on  E ModeYieldBR (10 -6 )A CP B  →     B  → K    Simultaneous fit to   /K   Preliminary Fit region e  e  → qq  hep-ex/

ICHEP 2002 J. Olsen 17 Next Side Please: B  →     Analysis issues:  Small signal!    feeddown Background suppression:  Event shape and flavor tagging to reduce qq  Cut on M(     ) and  E to reduce   background, then fix in the fit Significance including systematic errors = 2.5  Data after cut on probability ratio (  ) Preliminary    hep-ex/

ICHEP 2002 J. Olsen 18 Setting a Bound on Penguin Pollution Can still get information on  with only an upper bound on     :  For example: Grossman-Quinn bound (assume only isospin) Many other bounds on the market  Charles, Gronau/London/Sinha/Sinha, etc… Correlations and systematic errors included

ICHEP 2002 J. Olsen 19 CP-Violating Asymmetries in B  →    ,      Opportunity and challenges  In principle, can measure  directly, even with penguins  Much more difficult than     Three-body topology with neutral pion (combinatorics, lower efficiency) Significant fraction of misreconstructed signal events and backgrounds from other B decays Need much larger sample than currently available to extract  cleanly We perform a “quasi-two-body” analysis:  Select the  -dominated region of the        K      Dalitz plane  Use multivariate techniques to suppress qq backgrounds  Simultaneous fit for     and     R. Aleksan et al., Nucl. Phys. B361, 141 (1991)

ICHEP 2002 J. Olsen 20   direct CP violation → A CP and C ≠ 0 indirect CP violation → S ≠ 0  C and  S are insensitive to CP violation CP Not a CP eigenstate, (at least) four amplitudes contribute: Time evolution includes: CP Time-integrated asymmetry: Q is the  charge  K is self-tagging: Fit for:

ICHEP 2002 J. Olsen 21 Analysis Multi-dimensional ML fit  m ES,  E, Neural Net (NN),  c,  t Components  Signal  and  K  Misreconstructed signal events Mostly due to wrong photon(s)  B backgrounds from b → c and charmless B decays Same lifetime as signal  e  e  → qq Fix B background yields, fit for signal yields and CP asymmetries Validation: signal misreconstructed signal

ICHEP 2002 J. Olsen 22 Yields and Charge Asymmetries Preliminary hep-ex/

ICHEP 2002 J. Olsen 23 B 0   time-dependent asymmetry Systematic error dominated by uncertainty on B backgrounds Preliminary hep-ex/

ICHEP 2002 J. Olsen 24 Summary Hyperactive effort within BaBar to constrain, measure, and otherwise determine  Charmless two-body decays:  No evidence for large direct or indirect CP violation in   Beginning to piece together the necessary inputs to the isospin analysis Measurements of decay rates for   and     (upper limit) Too early for a significant constraint Charmless three-body decays  First measurement of CP asymmetries in  and  K The next few years will be interesting indeed!