CP Violation in B Decays at Belle

Journal of Physics: Conference Series CP Violation in B Decays at Belle To cite this article: Masaya Iwabuchi J. Phys.: Conf. Ser. 335 37 View the article online for updates and enhancements. Related content - CP violation at the B factories John Walsh and the Babar and Belle Collaborations - CP Violation Results from B Decays at BABAR Pietro Biassoni and Representing the Babar Collaboration - CP violation at LHCb Olaf Steinkamp and the LHCb Collaboration This content was downloaded from IP address 46.3.94.58 on 3//7 at :4

Journal of Physics: Conference Series 335 () 37 doi:.88/74-6596/335//37 CP Violation in B Decays at Belle Masaya Iwabuchi Department of Physics, Yonsei University, Seoul -749, Korea E-mail: iwabuchi@post.kek.jp Abstract. IreviewtherecentBellemeasurementsonCP violation in B decays. In this review, I cover the experimental constraints on the Cabibbo-Kobayashi-Maskawa (CKM) angles of the Unitarity Triangle by introducing several experimental approaches. Search for new physics beyond the Standard Model in b sqq transitions is also presented.. Introduction The present status on the measurement of CP violation in the quark sector is mainly based on the results achieved by BaBar and Belle. Those B-factory experiments, already ended their operations, have been spectacularly successful at discovering and investigating CP violation in B-meson decays, thanks to the great success of PEP-II and KEKB asymmetric-energy e + e colliders operating at the center of mass energy corresponding to the mass of Υ(4S). Those experiments have significantly improved our knowledge on CP-violation phenomena over a decade. In the Standard Model (SM), CP violation arises from a complex phase in the Cabibbo- Kobayashi-Maskawa (CKM) quark mixing matrix [, ], where CP violation can be quantified in terms of three parameters, φ, φ and φ 3,representingthethreeanglesoftheso-calledUnitarity Triangle. Precise measurements of CKM angles are therefore of fundamental importance for the description of the weak interaction of quarks and the investigation for the new sources of CP violation. In these proceedings, measurements of CKM angles at the Belle experimentarereviewed. This review also includes the measurement of effective φ (φ eff )inb sqq transitions for new-physics search.. φ sin φ can be obtained using b ccs transitions, where the most precise measurement for φ comes from B J/ψK decays. With 535 million BB pairs, Belle updated the measurement of sin φ using B J/ψK decays [3]. Figure shows the results of a fit to the reconstructed candidates in the fit region. Several kinematic variables are used here; for B J/ψKS decay, the energy difference E EB E beam and the beam-energy constrained mass M bc (Ebeam ) (p B ) are used in the fit, where Ebeam is the beam energy, and E B and p B are the energy and momentum, The angles φ, φ and φ 3 are also sometimes known as β, α and γ, respectively. Published under licence by Ltd

Journal of Physics: Conference Series 335 () 37 doi:.88/74-6596/335//37 respectively, of the reconstructed B candidate, all measured in the center-of-mass system. p B is used in the fit for B J/ψK L candidates. Events/(.GeV/c ) (a) J/ψ KS 8 6 4 5. 5.5 5.3 M bc (GeV/c ) Events/(.5GeV/c) 3 (b) J/ ψ K L data J/ΨK L J/ΨK L X BG, K L detected J/ΨX BG combinatorial BG.5.5 p* (GeV/c) B Figure. (a) M bc distribution in the E signal region for selected B J/ψK S candidates and (b) p B distribution for selected B J/ψK L candidates. The signal yields are determined to be 7484 ± 87 for B J/ψK S and 65 ± 3 for B J/ψK L,whereerrorsarestatisticalonly. In the decay chain Υ(4S) B B f CP f tag,whereoneoftheb mesons decays at time t CP to a CP eigenstate f CP and the other decays at time t tag to a final state f tag that distinguishes between B and B,thedecayratehasatimedependencegivenby P( t) = e t /τ B 4τ B { [ +q S CP sin( m d t)+a CP cos( m d t)] }. Here S CP and A CP describe mixing-induced and direct CP-violation, respectively, and τ B is the B lifetime, m d is the mass difference between the two B mass eigenstates, t = t CP t tag, and the b-flavor charge q = + ( ) when the tagging B meson is a B (B ). Here, the SM predicts S f = ξ f sin φ and A CP = for both b ccs and b sqq transitions, where ξ f =+( ) corresponds to the CP-even (-odd) final states. t can be determined from the displacement in the electron beamline (z) betweenthef CP and f tag decay vertices: t (z CP z tag )/(βγc) z/(βγc). An unbinned maximum-likelihood fit to the observed t distribution is performed to determine CP-violation parameters. Figures shows the t distributions and asymmetry. sin φ is measured to be.64 ±.3 ±.7 by the results of the fits to t distributions. Here, the first error is statistical and the second error is systematic. sin φ will be updated soon at Belle using the full data set (77 million BB pairs). The magnitude of the statistical error would be comparable with that of the systematic error, where the statistical error is expected to be around..

Journal of Physics: Conference Series 335 () 37 doi:.88/74-6596/335//37 Entries /.5 ps Asymmetry 4 3.5 -.5 (d) B J/ψK q=+ q= -7.5-5 -.5.5 5 7.5 -ξ f Δt(ps) Figure. Background subtracted t distributions and asymmetries for events with good tags for B J/ψK. 3. φ b uud quark transitions such as B ππ, ρπ,ρρ are sensitive to φ. The CP-violation parameters are expected to be S CP =sinφ and A CP =bythesm with the tree diagram only, however, direct CP violation is possible, A CP,duetoalarge contribution from the electroweak penguin-mediated decay processes. φ can be determined by S CP = A CP sin(φ + κ), where κ is the shift due to the penguin contribution and determined by the SU() isospin analysis [4]. For the φ measurements through B π + π and B ρ + ρ modes, it is possible to determine κ by considering the set of three B hh decays where hh is either two pions or two longitudinally polarized ρ mesons. The B hh amplitudes obey the complex triangle relations, A + = A + + A, A = A + + A. The electroweak penguins can be neglected in the B + h + h,andthusthesetrianglesshare the same base, A + = A. κ can be determined from the difference between the two triangles. This method has an 8-fold discrete ambiguity in the determination of φ. 3.. φ from B π + π φ measurement is performed using B π + π mode with 535 million BB pairs [5]. The signal yield is estimated by a fit to data, making use of E, M bc and the kaon identification probability. B π + π signal yield is obtained to be 464 ± 65 events. In the time-dependent analysis, direct CP violation is observed, A CP =+.55 ±.8(stat) ±.5(syst) with a significance of 5.5σ, while mixing-induced CP violation is also observed, S CP =.6 ±.(stat) ±.4(syst) with a significance greater than 5.3σ for any A CP value. To constrain φ,theisospinrelationsareusedwithmeasuredvaluesofs CP and A CP,the world average (W.A.) branching ratios of B π + π, π π,andb + π + π,andthew.a. direct CP-asymmetry for B π π [6]. χ is constructed using the measured values and the six parameters, one of which is φ.fourdifferentsolutionsarefoundaftertheminimizationof the χ.withtheexpectationfromotherckmmeasurements,(97±) is found. The φ range <φ < 79 is excluded at the 95% confidence level (C.L.). The resulting two-dimensional confidence regions in the S CP and A CP plane and the difference C.L. plotted for a range of φ values are shown in Fig. 3. 3

Journal of Physics: Conference Series 335 () 37 doi:.88/74-6596/335//37 Aππ.5 -.5 - C.L..8.6.4. Belle - - -.5.5 Sππ 3 6 9 5 8 φ (degrees) Figure 3. The left plot shows confidence regions for S CP and A CP. The curves show the contour from σ to 6σ, frominsidetooutside. ThepointwitherrorbarsistheS CP and A CP measurement. The right plot shows difference C.L. plotted for a range of φ values obtained with an isospin analysis. The solid and dashed lines indicatec.l. =68.3% and 95%, respectively. 3.. φ from B ρ + ρ Possible relative angular momentum for the B ρ + ρ decay is L =,,. The CPviolating parameters receive contributions from a longitudinally polarized state (CP-even) and two transversely polarized states (an admixture of CP-even and CP-odd states). However, measurements of the polarization fraction by Belle and BaBar show that the longitudinal polarization fraction is near unity. There is no need to elaborate the angular analysis, and pure CP-even state (% longitudinal polarization) is assumed in this analysis. Time-dependent analysis is performed for B ρ + ρ mode with 535 million BB pairs [7]. E, M bc and event shape variables for e + e qq (q = u, d, s, c) continuumandbb events are used in a fit to estimate the signal yield. B ρ + ρ signal yield is obtained to be 576 ± 53 events. Direct CP violation and mixing-induced CP violation are obtained to be A CP =+.6 ±.(stat) ±.8(syst), S CP =+.9 ±.3(stat) ±.8(syst), respectively. In the isospin relations to constrain φ, the W.A. branching ratios of B ρ + ρ and B + ρ + π are used [6], while the branching ratio of B ρ ρ is obtained from the Belle measurement [8] to constrain φ. φ is determined to be (9.7 ± 4.9). Figure 4 shows the difference C.L. plotted for φ. Figure 4. C.L. vs. φ obtained from the isospin analysis of B ρρ decays. 4

Journal of Physics: Conference Series 335 () 37 doi:.88/74-6596/335//37 3.3. φ from B ρπ B ρπ mode is also sensitive to φ. Since this mode can be treated as an interference of B ρ + π, ρ π + and ρ π modes, a time-dependent Dalitz plot analysis (TDPA) is employed, where measurements of interfering parameters must be included. Measurements for relative sizes and complex phases of the amplitudes are required to extract φ,however,thoseamplitudesare related to φ through an isospin relation. In this point, B ρπ channel has an advantage which can reduce discrete ambiguities, in contrast to the previous quasi-two-body modes for the φ measurements. With 449 million BB pairs, Belle performed a measurement of φ using B ρπ [9]. The signal yield is determined from a four-dimensional fit to the E-M bc and Dalitz plot distribution. The fit yields 97±4 B π + π π events in the signal region. To constrain φ,measurements with results on the branching fractions for B ρ ± π and B + ρ + π, ρ π +,andflavor asymmetries of the latter two [6] are all combined. χ scans are shown in Fig. 5 where φ can be constrained to be 68 <φ < 95 at 68.3% C.L..8 C.L..6.4. C.L.=68.3% 3 6 9 5 8 φ (degrees) Figure 5. C.L. vs φ. Dotted and solid curves correspond to the result from the TDPA only and that from the TDPA and an isospin combined analysis, respectively. 4. φ 3 B D ( ) K modes provide a good probe for the φ 3 measurement, where the interference between B D ( ) K and B D ( ) K decays is utilized. The magnitude of the amplitude ratio of these decays (r B )canbeevaluatedtobe., using the CKM matrix elements with an additional color suppression factor. The weak phase φ 3 arises from the complex phase of V ub included in the B + D ( ) K + decay. There is no contribution from the penguin diagram in these decays. Among the several approaches proposed for the φ 3 measurement, the current most precise measurement of φ 3 is based on the so-called GGSZ method []. In this method, amplitude analysis of the multibody decay of the neutral D meson is required. However, complex phase variations and imperfect assumption of the D-decay model potentially includes the irreducible uncertainties for the φ 3 measurement. In other approaches, the so-called ADS method [] is also considered to be promising. The ADS method can be free from the model uncertainty, while the measurements based on this method have been statistically limited. 4.. φ 3 from the GGSZ method In the B ± DK ± interference, D can be either D or D,andtheamplitudeofD can be expressed as D = r B e ±iφ 3+iδ B D + D. 5

Journal of Physics: Conference Series 335 () 37 doi:.88/74-6596/335//37 Here δ B is the strong phase difference of the B decays. It is possible to distinguish D from D if the multibody D decays are used. With 657 million BB pairs, Belle updated the φ 3 measurement using the GGSZ method with B + D ( ) K +, D K S π + π decays []. This measurement is based on the model-dependent approach, including the model determination of D-meson decay using the e + e qq continuum processes. The signal yield is estimated by a two-dimensional fit for the E and M bc distribution for each decay mode. Signal yields of B + DK +, B + D K + followed by D Dγ or D Dπ are obtained to be 756, 49 and 4 events in each signal region, respectively. To obtain φ 3,DalitzdistributionsoftheB + and B are fitted separately, using Cartesian parameters x ± = r ± cos(±φ 3 + δ) andy ± = r ± sin(±φ 3 + δ), where the indices + and correspond to B + and B decays, respectively. Here, r + and r are not constrained to be equal. The advantage of this approach is for the low bias and simple distribution of the fitted parameters. Several kinematic variables are used in a fit; Dalitz plot distribution, E, M bc and event shape variables that separates the B decays from the e + e qq continuum processes. The results of the separate B + and B data fits are shown in Fig. 6. The separation of the B and B + positions in the (x, y) planeindicatesdirectcp violation and is measured to be r B sin φ 3. y.4 B DK y.4 B [Dπ ] D* K B [Dγ] D* K with Δδ=8.. B + D * K + -. B + DK + -.4 -.4 -...4 x -. B D * K -.4 -.4 -...4 x Figure 6. Results of signal fits with free parameters x = r cos θ and y = r sin θ for B + DK + (a) and B + D K + (b) samples, separately for B and B + data. Contours indicate,, and 3(forB ± DK ± )andstandarddeviationregions(forb ± D K ± ). Afrequentisttechniqueisusedtoevaluatethestatisticalsignificance of the measurements. In order to improve the sensitivity to φ 3,eventsamplesofB + DK + and B + D K + are combined. φ 3 for the combination of these modes is obtained to be φ 3 =78.4 +.8.6 (stat) ± 3.6 (syst) ± 8.9 (model). CP conservation in this process is ruled out at the confidence level ( C.L.) =5 4,or3.5 standarddeviations,whichleadstotheevidencefordirect CP violation for the B + D ( ) K + process. Model uncertainty is dominated by the uncertainty due to the complex phase. To reduce this model uncertainty, model-independent approach is proposed [, 3, 4]. Based on this method, CLEO recently updated the results of relative phase between D and D by studying the ψ(377) D D decays [5]. The model error in the φ 3 measurement will be replaced by a statistical error of about a few degrees due to the finite ψ(377) D D sample, while the statistical error associated with the B data sample should increase by %-% due to the binned fit procedure. 6

Journal of Physics: Conference Series 335 () 37 doi:.88/74-6596/335//37 4.. φ 3 from the ADS method In the ADS method, measurements of CP asymmetry (A)andpartialrate(R)playanimportant role for the φ 3 measurement. In the decay chain B ± DK ±, D K ± π,whered can be either D or D, A and R are described by A B(B [K + π ] D K ) B(B + [K π + ] D K + ) B(B [K + π ] D K )+B(B + [K π + ] D K + ) = r B r D sin(δ B + δ D )sin(φ 3 )/R, R B(B [K + π ] D K )+B(B + [K π + ] D K + ) B(B [K π + ] D K )+B(B + [K + π ] D K + ) = r B + r D +r Br D cos(δ B + δ D )cos(φ 3 ). Here δ D is the strong phase difference of the D-meson decays, and r D is the amplitude ratio of D K + π,doublycabibbosuppressedmode,tod K + π,cabibbofavored mode. In the B [K + π ] D K interference, either B or D decay is suppressed, while the B [K π + ] D K decay process is dominated by the favored modes of B and D decays. With 77 million BB pairs, Belle updated the measurement for the ADS method using the above decay chain, where preliminary results are obtained on A and R. Inthismeasurement, a tool for multi-variate analysis named NeuroBayes is employed. In order to separate the B DK signal from backgrounds, ten kinematic variables are used as inputsforneurobayes,suchasa likelihood ratio of Fisher discriminant of modified Fox-Wolfram moments, the distance of vertices between the reconstructed B and the other B, andthedecayangleofd Kπ. The signal yield is extracted by a two dimensional fit on E and NeuroBayes output NB, and determined to be 56. 4. +5. for B [K + π ] D K,wherethecharge-conjugatemodeis included. Figure 7 shows the results of the fit to the reconstructed candidates in the fit region. The E distributions for [K + π ] D K and [K π + ] D K + in the NB signal region are shown. 6 4 6 4 Events / MeV 8 6 4 Events / MeV 8 6 4 -....3 ΔE (GeV) -....3 ΔE (GeV) Figure 7. E distributions for [K + π ] D K (left) and [K π + ] D K + (right). Curves represent the various components from the fit: DK signal (thick dashed line), Dπ backgrounds (dashed line), BB backgrounds (dash-dotted line), continuum (dotted line) and sum of all components (solid line). The firstevidence of the DK signal is obtained with a significance of 4.σ including systematic uncertainties. The obtained results for the CP asymmetry and the ratio of the suppressed to favored decay rate are, A =.39 +.6.8 (stat)+.4.3 (syst), and R =[.63+.44.4 (stat)+.7.3 (syst)].toconstrainφ 3, r D, δ D and other relevant observables will be used. 5. φ eff Precise measurement of φ also provides a useful comparison between decays mediated by b sqq and b ccs transitions. A new physics may enter into the loop diagram, and to search 7

Journal of Physics: Conference Series 335 () 37 doi:.88/74-6596/335//37 for the new physics effects, various approaches have been tested by investigating deviations of CP violating parameter φ eff from φ determined by b ccs decays. With 657 million BB pairs, Belle performed a measurement of φ eff using B K + K K S decay with a time-dependent Dalitz approach [6]. Several decay modes are included in the signal model: φ()k S, f (98)K S, f X K S, χ c K S and nonresonant K + K K S. The signal yield is estimated by a fit to data, making use of E, M bc and the flavor-tag quality interval. B K + K K S signal yield is obtained to be 76 ± 5 events. In the timedependent analysis, four solutions are found from a fit. A single solution is preferred from all the solutions using the branching fractions of f (98) ππ and f (5) ππ as external information. Figure 8 shows the t distributions and asymmetry of the preferred solution. Entries /.5ps 4 3 (a) q=+ q= - Raw Asymmetry.5 -.5 (b) -7.5-5 -.5.5 5 Δt (ps) 7.5 - -7.5-5 -.5.5 5 7.5 Δt (ps) Figure 8. (a) t distribution and (b) raw asymmetry for the B K + K K S candidates in the φ()ks region, with good tags. In (a), the dotted black curve shows the background component. In (b), the solid curve shows the fit projection, while the dashed curve shows the SM expectation from the time-dependent CP asymmetry measurement in b ccs decays. In the preferred solution, φ eff is measured to be (3. ± 9. ±.6 ±.4) for φ()k S mode, and (3.3 ± 9. ± 3.4 ± 4.) for f (98)K S mode. The results are still consistent with the SM expectation like other results from Belle, where measurements of φ eff are all consistent with the value of φ obtained from the decay B J/ψK within the two standard deviations. 6. Summary Various studies have been performed to investigate CP violation in the B decays at Belle. The upgraded Belle II experiment will allow us to constrain the CKM angles with much higher statistics, and also help us to explore and clarify the nature ofcp violation. References [] N. Cabibbo, Phys. Rev. Lett., 53(963). [] M. Kobayashi and T. Maskawa, Prog. Theor. Phys. 49, 65(973). [3] K.-F. Chen et al. (Belle Collaboration), Phys. Rev. Lett. 98, 38(7). [4] M. Gronau and D. London, Phys. Rev. Lett. 65, 338(99). [5] H. Ishino et al. (Belle Collaboration), Phys. Rev. Lett. 98, 8(7). [6] Heavy Flavor Averaging Group, http://www.slac.stanford.edu/xorg/hfag. [7] A. Somov et al. (Belle collaboration), Phys. Rev. D 76, 4(7). [8] C.-C. Chiang et al. (Belle collaboration), Phys. Rev. D 78, (R)(8). [9] A. Kusaka et al. (Belle collaboration), Phys. Rev. Lett. 98, 6(7). [] A. Giri, Yu. Grossman, A. Soffer and J. Zupan, Phys, Rev. D 68, 548(3). [] D. Atwood, I. Dunietz and A. Soni, Phys. Rev. Lett. 78, 3357(997). [] A. Poluektov et al. (Belle collaboration), Phys, Rev. D 8, (). [3] A. Bondar and A. Poluektov, Eur. Phys. J. C 47, 347(6). [4] A. Bondar and A. Poluektov, Eur. Phys. J. C 55, 5(8). [5] J. Libby et al. (CLEO collaboration), Phys. Rev. D 8, 6(). [6] Y. Nakahama et al. (Belle collaboration), Phys. Rev. D 8, 73(). 8