SNSN-323-63 January 4, 2011 B → K(∗)(cid:96)+(cid:96)− from B-factories and Tevatron 1 1 0 2 Gerald Eigen n a representing the BABAR collaboration 1 J 3 Department of Physics and Technology ] University of Bergen, 5007 Bergen, NORWAY x e - p e h [ 1 v BABAR and Belle measurements of branching fractions, rate asymme- 0 tries and angular observables in the decay modes B → K(∗)(cid:96)+(cid:96)− are re- 7 viewed and new results from CDF on B → K(∗)µ+µ− branching fractions 4 0 and angular observables are discussed. A first search for B+ → K+τ+τ− . 1 is presented. 0 1 1 : v i X r a PRESENTED AT CKM workshop 2010 Warwick, UK, September 06–10, 2010 1Work supported by the Norwegian Research Council. 1 Introduction Thedecaysb → s(cid:96)+(cid:96)−, where(cid:96)+(cid:96)− isane+e−,µ+µ− orτ+τ− pair, areflavor-changing neutral current (FCNC) processes, which are forbidden in the Standard Model (SM) at tree level but are allowed to proceed via electroweak loops and weak box diagrams. An effective Hamiltonian is used to calculate decay amplitudes [1], which depend on three effective Wilson coefficients, Ceff, Ceff, and Ceff. The first is extracted from 7 9 10 the B → X γ branching fraction, the latter two respectively represent the vector and s axial vector part of the weak penguin and box diagrams. New Physics effects involve new loops that interfere with the SM processes modifying the measured values of Ceff, Ceff, and Ceff with respect to the SM predictions [2]. In addition, scalar and 7 9 10 pseudoscalar processes may contribute that introduce new Wilson coefficients C and s C thatareforbiddenintheSM.Thus, itisimportanttomeasuremanyobservablesin p order to overconstrain the complex Wilson coefficients [3]. These electroweak penguin modes contribute in probing New Physics at a scale of a few TeV [4]. In this review, wefocusonexclusivedecayspresentingresultsfromBABAR, BelleandCDF.Thedata samples are based on luminosities of 349 fb−1, 605 fb−1 and 4.4 fb−1 corresponding to 384 million BB events, 656 Million BB events and 2×1010 bb events, respectively. 2 Selection of B → K(∗)e+e− and B → K(∗)µ+µ− Events BABAR and Belle fully reconstruct ten B → K(∗)e+e− and B → K(∗)µ+µ− final states, in which a K+,K0,K+π−,K+π0 or K0π+ recoils against the lepton pair∗, S S while CDF reconstructs K+µ+µ− and K+π−µ+µ− final states. BABAR (Belle) selects lepton candidates with momenta p > 0.3(0.4) GeV/c and p > 0.7(0.7) GeV/c. e µ BABAR and Belle require good particle identification (PID) for e,µ,K, and π, and select K0 in the π+π− channel. CDF requires muons with p (µ) > 0.4 GeV/c, kaons S T and pions with p (K,π) > 1 GeV/c and B-mesons with p (B) > 6 GeV/c. Both, T T muons and hadrons must have good PID and the muon pair must originate from a secondary vertex. All three experiments suppress combinatorial BB and qq con- tinuum backgrounds (q = u,d,s,c). Here, the leptons dominantly originate from semileptonic b and c decays. BABAR trains neural networks (NN) using event shape variables, vertex information, missing energy, and lepton separation near the inter- action region (IR) optimized in each mode and each q2 bin†. Belle trains a Fisher discriminant using event shape variables, missing mass, B flavor tagging, and lepton separation in z near the IR. CDF trains NNs using vertex information, the angle between the signed vertex displacement with respect to the B momentum, and the µ ∗Charge conjugation is implied unless otherwise stated. †This is the squared momentum transfer into the dilepton system. 1 Experiment Mode B [10−6)] A R CP K(∗) BABAR [6] K(cid:96)+(cid:96)− 0.394+0.073 ±0.02 −0.18+0.18 ±0.01 0.96+0.44 ±0.05 −0.069 −0.18 −0.34 BABAR [6] K∗(cid:96)+(cid:96)− 1.11+0.19 ±0.07 −0.01+0.16 ±0.01 1.10+0.42 ±0.07 −0.18 −0.15 −0.32 Belle [7] K(cid:96)+(cid:96)− 0.48+0.05 ±0.03 0.04±0.1±0.02 1.03±0.19±0.06 −0.04 Belle [7] K∗(cid:96)+(cid:96)− 1.07+0.11 ±0.09 −0.10±0.1±0.01 0.83±0.17±0.8 −0.10 CDF [8] Kµ+µ− 0.38+0.05 ±0.03 −0.05 CDF [8] K∗µ+µ− 1.06+0.14 ±0.09 −0.14 Table 1: Branching fractions, CP asymmetries and lepton flavor ratios for B → K(∗)(cid:96)+(cid:96)− modes in the entire q2 region from BABAR, Belle, and CDF. Uncertainties are statistical and systematic, respectively. separation. BABAR and Belle select signal candidates using the beam-energy substi- (cid:113) tuted mass m = E∗2 −p∗2 and the energy difference ∆E = E∗ −E∗ , where ES beam B B beam E∗ ,E∗ and p∗ are the beam energy, B-meson energy and B-meson momentum in beam B B the Υ(4S) center-of-mass frame, respectively. BABAR extracts the signal yield from a one-dimensional unbinned extended maximum log-likelihood fit in m , while Belle ES performs a one (two) dimensional unbinned extended maximum log-likelihood fit in m (andm )forK(∗)(cid:96)+(cid:96)− modes. CDFselectssignalcandidatesfromanunbinned ES Kπ maximum log-likelihood fit in the B invariant-mass distribution. All experiments re- ject events in the J/ψ and ψ(2S) mass regions and require that Kµ and Kπµ masses are not consistent with a D mass to reject background from B → DX decays. The rejected charmonium events are used as control samples for various cross checks. 3 Results for B → K(∗)e+e− and B → K(∗)µ+µ− Modes Figure 1 (left) shows total branching fractions for B → K(∗)(cid:96)+(cid:96)− (e+e− and µ+µ− modes combined) [6, 7, 8] and B → X (cid:96)+(cid:96)−[9, 7] in comparison to the SM predictions s [5]. The individual exclusive measurements are summarized in Table 1. The Belle inclusive measurement is a recent update based on a luminosity of 605 fb−1, yielding B(B → X (cid:96)+(cid:96)−) = 3.33 ± 0.8+0.19) × 10−6 [10]. The partial branching fractions s −0.24 measured in the three experiments are also consistent with the SM predictions. Rate asymmetries are more precisely measured than branching fractions, since many uncertainties cancel [11]. The isospin asymmetry [12] dB(B0 → K(∗)0(cid:96)+(cid:96)−)/dq2 −(τ /τ )dB(B+ → K(∗)+(cid:96)+(cid:96)−)/dq2 A (q2) = B0 B+ , (1) I dB(B0 → K(∗)0(cid:96)+(cid:96)−)/dq2 +(τ /τ )dB(B+ → K(∗)+(cid:96)+(cid:96)−)/dq2 B0 B+ corrected for the different B0 and B+ lifetimes (τ /τ ), is expected to be small B0 B+ in the SM (A. (q2)/dq2 is −0.01 for q2 = 2.7 − 6 GeV2/c4 after dropping from (cid:39) I 2 BBaaBBaarr 38429 f bfb-1,- 1,’ 0’049 sl+l- CDF 4.4 fb-1, ’10 prelim. Belle 605 fb-1, ’09 Belle 605 fb-1, ’10 prelim. Ali ’02 K*l+l- Kl+l- 10-7 10-6 10-5 Branching Fraction Figure 1: (Left) Total branching fractions measurements of B → K(∗)(cid:96)+(cid:96)− and B → X (cid:96)+(cid:96)− modes from BABAR (red dots), Belle (blue triangles) and CDF (magenta s squares) in comparison to the SM prediction (grey-shaded region). For BABAR and Belle, (cid:96)+(cid:96)− is a combination of e+e− and µ+µ− modes, for CDF it is µ+µ−. (Right) Isospin asymmetry measurements for B → K(∗)(cid:96)+(cid:96)− versus q2 from BABAR (black squares, blue dots) and Belle (red triangles, green triangles). 0.075 at q2 = 0.1 GeV2/c4 and crossing zero near q2 = 1.7 GeV2/c4) [12]. Figure 1 (right) shows the BABAR and Belle A measurements for different q2 regions. The I q2 integrated isospin asymmetry and A for q2 values above the J/ψ are consistent I with the SM prediction. Below the J/ψ, however, BABAR observes a negative A that I deviates significantly from the SM prediction (3.9σ from A = 0) . For models in I which the sign in Ceff is flipped with respect to the value in the SM, a small negative 7 A is expected [12, 13], but it is too small to explain the BABAR measurement. For I low q2, the Belle results are consistent with both BABAR and the SM. In the SM, the direct CP asymmetry B(B → K(∗)(cid:96)+(cid:96)−)−B(B → K(∗)(cid:96)+(cid:96)−) A = . (2) CP B(B → K(∗)(cid:96)+(cid:96)−)+B(B → K(∗)+(cid:96)+(cid:96)−) is expected to be O(10−3), and new physics at the electroweak scale may provide significant enhancements [14]. BABAR performs a simultaneous fit to B+ → K+(cid:96)+(cid:96)− and B → K∗(cid:96)+(cid:96)− modes. The results summarized in Table 1 together with Belle’s measurements are consistent with the SM expectations. In the SM, the lepton flavor ratios R = B(B → Kµ+µ−)/B(B → Ke+e−) and K R = B(B → K∗µ+µ−)/B(B → K∗e+e−) integrated over all q2 are predicted to K∗ be one and 0.75, respectively. The theoretical uncertainties are just a few percent. For example, in two-Higgs-doublet models the presence of a SUSY Higgs might give ∼ 10% corrections to R for large tanβ [13].The BABAR and Belle measurements K(∗) summarized in Table 1 are consistent with the SM expectations. The B → K∗(cid:96)+(cid:96)− angular distribution depends on three angles: θ , the angle K between the K momentum and the B momentum in the K∗ rest frame, θ , the angle (cid:96) 3 Experiment q2 bin [GeV2/c4] F A L FB BABAR [17] 0.1-6.25 0.35±0.16±0.04 0.24+0.18 ±0.05 −0.23 Belle [7] 1-6 0.67±0.23±0.04 0.26+0.27 ±0.07 −0.30 CDF [8] 1-6 0.5+0.27 ±0.04 0.43+0.36 ±0.06 −0.30 −0.37 SM [24] 1-6 0.73+0.13 −0.05+0.03 −0.23 −0.04 Table 2: BABAR, Belle, and CDF measurements of F and A from B → K∗(cid:96)+(cid:96)− L FB modes in the low q2 region. between the (cid:96)+((cid:96)−) momentum and the B(B) momentum in the (cid:96)+(cid:96)− rest frame, and φ, the angle between the two decay planes. The angular distribution involves 12 q2-dependent coefficients J [15, 16] that can be extracted from a full angular fit in i individual bins of q2. Since large data samples are necessary for this study, BABAR , Belle and CDF have analyzed only the one-dimensional angular distributions 3 3 W(cosθ ) = F cos2θ + (1−F )sin2θ , (3) K L K L K 2 4 3 3 W(cosθ ) = F sin2θ + (1−F )(1+cos2θ )+A cosθ , (4) (cid:96) L (cid:96) L (cid:96) FB (cid:96) 4 8 where F is the K∗ longitudinal polarization and A is the lepton forward-backward L FB asymmetry. While Belle and CDF measure F and A in six q2 bins, BABAR mea- L FB sured F and A in two q2 bins due to the limited data sample. An update with the L FB full BABAR data set in six q2 bins is in progress. The measured m and angular dis- ES tributions are fitted with signal, combinatorial background and peaking background components. After determining the signal yield from the m spectrum, F is ex- ES L tracted from a fit to the cosθ distribution for fixed signal yield. Finally, A is K FB extracted from the cosθ distribution for fixed signal yield and fixed F . (cid:96) L Figure 2 shows the BABAR, Belle, and CDF results for F (left) and A (right) L FB in comparison to the SM prediction (lower red curve) [18] and for flipped-sign Ceff 7 models (upper blue curve) [20, 23]. In the SM, A is negative for small q2, crosses FB zero at q2 = (4.2 ± 0.6) GeV2/c4 and is positive for large q2, while for flipped- 0 sign Ceff models A is positive for all q2. Table 2 summarized the F and A 7 FB L FB measurements from B → K∗(cid:96)+(cid:96)− in the low q2 region in comparison to the SM prediction. For F , the three measurements are consistent with each other and the L SM prediction. For A , the three measurements are in good agreement. Though FB they are in better agreement with the flipped-sign Ceff model, they are consistent 7 with the SM prediction. For B → K(cid:96)+(cid:96)−, A is consistent with zero as expected in FB the SM. 4 Figure 2: (left) Measurements of F and (right) Measurements of A in B → L FB K(∗)(cid:96)+(cid:96)− modes by BABAR (black squares), Belle ( brown dots) and CDF (green triangles). The SM prediction (flipped-sign Ceff model) is shown by the upper red 7 (lower blue) curve for F and the lower red (upper blue) curve for A . L FB 4 Search for B+ → K+τ+τ− In the SM, the q2 dependence of the B → X τ+τ− decay rate has a shape similar s to that of B → X µ+µ− in the high q2 region. The B+ → K+τ+τ− branching s fraction is predicted to be ∼ 2 × 10−7 in the SM, which is 50 − 60% of the total inclusive branching fraction [21]. Enhancements are predicted in models beyond the SM. In the next-to-minimal supersymmetric models (NMSSM), for example, the rate may be enhanced by the squared tau-to-muon mass ratio (m /m )2 ∼ 280. Since τ µ signal final states contain 2-4ν, a different analysis strategy is needed here to control backgrounds. BABAR has performed the first search for B+ → K+τ+τ− using an integrated luminosity of 423 fb−1 which corresponds to 465 BB events. The recoiling (”tag”) B is reconstructed in many hadronic final states, B− → D(∗)0,+X, where X represents up to six hadrons (π±,π0,K±,K0). Using m and ∆E the tag is selected with an S ES efficiency of ∼ 0.2%. The single-prong τ decays τ → eνν,τ → µνν and τ → πν are selected as signal modes. Thus, signal candidates are required to have only three charged particles of which one is an identified kaon with charge opposite to the tag B and 0.44 < p < 1.4 GeV/c in the center-of-mass frame. The two remaining K particles must have opposite charge, be consistent with the signal τ decays, have p < 1.59 GeV/c and a mass M < 2.89 GeV/c2. Further requirements are q2 = pair (p(cid:126) − p(cid:126) − p(cid:126) )2/c2 > 14.23 GeV2/c4, a missing energy (i.e. the energy carried Υ(4S) tag K off by neutrinos estimated as the difference between Υ(4S) energy and that of all observed particles) of 1.39 < E < 3.38 GeV, and neutral energy deposited in the miss electromagnetic calorimeter E < 0.74 GeV. Continuum background is suppressed extra by |cosθ | < 0.8, where θ is the opening angle between the thrust axis of the T T tag and that of the rest of the event. The largest remaining background originates 5 from B+ → D0X+, which is suppressed by combining the signal K+ with the τ daughter of opposite charge assigned the π mass hypothesis and requiring a mass M > 1.96 GeV/c2. Kπ BABAR observes 47 events with an expected background of 64.7±7.3 events. In- cludingsystematicuncertaintiesabranchingfractionupperlimitofB(B → K+τ+τ−) < 3.3×10−3 is set at 90% confidence level (CL). 5 Conclusion BABAR and Belle have measured branching fractions, rate asymmetries and angular observables in B → K(∗)(cid:96)+(cid:96)− final states. Recently, CDF contributed new measure- ments on branching fractions and angular observables in B → K(∗)µ+µ−. Except for the isospin asymmetry at low values of q2 all other measurements are consistent with the SM, though F and A agree also with the flipped-sign Ceff model. BABAR L FB 7 has performed the first search for B+ → K+τ+τ− setting a branching fraction upper limit of B(B+ → K+τ+τ−) < 3.3 × 10−3 at 90% CL. Although all experiments are expected to update results with the final data sets, significant improvement in precision will come from LHCb and the Super B-factories. In these new experiments, sufficiently large data samples will be collected to measure the full angular distribu- tion from which the 12 observables J [15] can be measured with high precision in i different bins of q2. In turn, the Wilson coefficients can be determined with high precision to reveal small discrepancies with respect to the SM predictions [3, 23]. ACKNOWLEDGEMENTS I would like to thank my BABAR colleague K. Flood for useful discussions. This work has been supported by the Norwegian Research Council. References [1] G. Buchalla, A. J. Buras and M. E. Lautenbacher, Rev. Mod. Phys. 68, 1125 (1996); C. Bobeth, M. Misiak and J. Urban, Nucl. Phys. B574, 291 (2000); H.H Asatryan et al., Phys. Rev. D65, 034009 (2002); Phys. Lett. B507, 162, (2001); G. Hiller and F.Kru¨ger, Phys.Rev. D69, 074020 (2004); M. Beneke, Th. Feldmann, andD.Seidel; Nucl.Phys.B612, 25(2001);M.Beneke, Th.Feldmann, and D. Seidel; Eur.Phys.J. C41, 173 (2005). [2] G. Burdman, Phys. Rev. D52, 6400 (1995); J. L. Hewett and J. D. Wells, Phys. Rev. D55, 5549 (1997); W. J. Li, Y. B. Dai and C. S. Huang, Eur. Phys. J. C40, 6 565 (2005); Y. G. Xu, R. M. Wang and Y. D. Yang, Phys. Rev. D74, 114019 (2006); P. Colangelo et al., Phys. Rev. D73, 115006 (2006); C.-H. Chen and C.Q. Geng, Phys. Rev. D 66 094018 (2002);C. Bobeth et al, Phys. Rev. D64 074014 (2001). [3] K.S.M. Lee et al., Phys. Rev. D75, 034016 (2007). [4] G. Isidori, Y. Nir, G. Prerez, arXiv:1002.0900 (2010). [5] A. Ali, E. Lunghi, C. Greub and G. Hiller, Phys. Rev. D 66, 034002 (2002). [6] B. Aubert et al. (BABAR collaboration), Phys. Rev. Lett.102, 091803 (2009). [7] J.T. Wei et al. (Belle collaboration), Phys. Rev. Lett.103, 171801 (2009). [8] T. Aaltonen et al. (CDF collaboration), CDF note 10047 (2010). [9] B. Aubert et al. (BABAR collaboration), Phys. Rev. Lett.93, 081862 (2004). [10] C.C.Chiang (Belle collaboration), talk at ICHEP10 (2010). [11] F. Kru¨ger, L. M. Sehgal, N. Sinha and R. Sinha, Phys. Rev. D61, 114028 (2000), [Erratum-ibid. D63, 019901 (2001)]. [12] T. Feldmann and J. Matias, JHEP 0301, 074 (2003). [13] Q. S. Yan, C. S. Huang, W. Liao and S. H. Zhu, Phys. Rev. D 62, 094023 (2000). [14] C. Bobeth, G. Hiller and G. Piranishvili, JHEP 0807, 106 (2008). [15] F. Kru¨ger et al., Phys. Rev. D61, 114028 (2000); Erratum-ibid D63, 019901 (2001). [16] C.S. Kim et al., Phys. Rev. D 62, 034013 (2000). [17] B. Aubert et al. (BABAR collaboration), Phys. Rev. D79, 031102 (2009). [18] G. Buchalla et al., Phys. Rev. D63, 014015 (2001). [19] G. Buchalla et al., Phys. Rev. D63, 014015 (2001). [20] A. Hovhannisyan, W. S. Hou and N. Mahajan, Phys. Rev. D 77, 014016 (2008). [21] J.L. Hewett, Phys. Rev. D53, 4964 (1995). [22] K. Flood, talk at the Int. Conf. on HEP, Paris July 22-28 (2010). [23] F. Kru¨ger and J. Matias, Phys. Rev. D71, 094009 (2005). [24] C. Bobeth, G. Hiller and D. van Dyk, JHEP 1007, 098 (2010). 7