Search for Neutral B Meson Decays to Two Charged Leptons

Abstract The decays Bd0, Bs0 → e+e−, μ+μ−, e±μ∓ are searched for in 3.5 million hadronic Z events, which constitute the full LEP I data sample collected by the L3 detector. No signals are observed, therefore upper limits at the 90% (95%) confidence levels are set on the following branching fractions: Br(B 0 d →e + e − ) −5 ; Br(B 0 s →e + e − ) −5; Br(B 0 d →μ + μ − ) −5 ; Br(B 0 s →μ + μ − ) −5; Br(B 0 d →e ± μ −∓ ) −5 ; Br(B 0 s →e ± e −∓ ) −5; The results for Bs0 → e+e− and Bs0 → e±μ∓ are the first limits set on these decay modes.


L Introduction
Measurements of rare B hadron decay rates provide clean tests of the Standard Model and are in general sensitive to new physics. B° -> decays 7 are par ticularly clean both theoretically and experimentally. The flavour-changing neutral current decays, B° -> e+e~ and B° -> ¡ x~, are forbidden at tree level. They occur in the Standard Model due to higher-order processes with branching fractions < 0 ( 10-9 ) [1], which is beyond the sensitivity of LEP» and are sen sitive to the Cabibbo-Kobayashi-Maskawa matrix el ements, the top quark mass, and the B meson decay constants. B° -* decays violate lepton number conservation and are forbidden in the Standard Model.
Observation of B° -> decays at LEP would be a clear indication for physics beyond the Standard Model. For example, two Higgs doublet models pre 1 Supported by the German Bundesministerium für Bildung, Wis senschaft, Forschung und Technologie. 2 Supported by the Hungarian OTKA fund under contract number T14459.
Supported also by the Comisión Interministerial de Ciencia y Technología. 4 Also supported by CONICET and Universidad Nacional de La Plata, CC 67, 1900 La Plata, Argentina. 5 Also supported by Panjab University, Chandigarh-160014, India. 6 Supported by the National Natural Science Foundation of China. 7 Throughout this paper we use B° to denote either B¡¡ or Bj?, and £+£-to denote e+e"1 or /x+/a-or e^/x^. The latter case is a sum of e+/¿" and e" /¿+ . diet significant enhancements to the B° -» 1~ de cay rates [2][3][4][5].
This analysis is performed on data recorded during 1991-1995 running at the Z, corresponding to a sam ple of approximately 3.5 million hadronic Z decays. The mixed sample of B hadrons produced in Z decays provides an opportunity to study B, meson decays which are not accessible at the center-of-mass energy of the Y(4S).

The L3 detector
The L3 detector is described in detail else where [6,7]. The central tracking chamber is a Time Expansion Chamber (TEC) consisting of two coaxial cylindrical drift chambers with 12 inner and 24 outer sectors. The Z-chamber surrounding the TEC consists of two coaxial proportional cham bers with cathode strip readout. The electromag netic calorimeter is composed of bismuth germanate (BGO) crystals. Hadronic energy depositions are measured by a uranium-proportional wire chamber sampling calorimeter surrounding the BGO. Scintil lator time-of-flight counters are located between the electromagnetic and hadronic calorimeters. The muon spectrometer, located outside the hadron calorimeter, consists of three layers of drift chambers measuring the muon trajectory in both the bending and the non bending planes. All subdetectors are installed inside a solenoidal magnet which provides a uniform field 478

M, Acciarri et a l . / Physics Letters B 391 (1997) 474-480
The invariant mass resolution of pairs of electrons measured in the BGO calorimeter is approximately 70M eV for the typical kinematics of B° e+e~ decays. Similarly, the di-muon mass resolution is 180MeV and the e^/x^ mass resolution is 140 MeV.

£+£ decays and
The JETSET Monte Carlo program [8,9] is used to simulate hadronic Z decays* To model b quark frag mentation the Peterson function [10] is used as a function of xe = 2£hadron/v^> with a mean value of (xe) = 0.703. The masses of the B[j and Bi? mesons are assumed to be 5279 MeV and 5373 MeV respectively, which are consistent with the most recent world aver age values [11]. The events produced by JETSET are passed through the GEANT-based L3 detector simula tion program [12] which allows for the effects of en ergy loss, multiple scattering, decays and interactions in the detector material, as well as time-dependent de tector effects. These events are then reconstructed us ing the same algorithms as for the data, To study the sensitivity of the L3 detector to events containing B° -► £+£~ decays, we first simulate e 1 e bb events. Bjjs mesons are forced to decay via the chain B[}s -(£ = e, /¿). If, however, there is more than one B § s meson in the event, only one is required to decay in this way. For the back ground studies a sample of hadronic events is used which does not include the B° -» £+£~ decay modes.
plane transverse to the beam direction. Electron can didates are required to have an energy of more than 2GeV. Muon candidate tracks in the muon spectrometer are required to be within | cos 0\ < 0.8 and to have hits in at least two of the three r< j> layers and at least one of the two z layers. Backgrounds from punchthrough hadrons, decays in flight, and cosmic rays are sup pressed by requiring the muon chamber track to point towards the primary vertex. Residual contamination due to cosmic rays which coincide with a genuine hadronic event is reduced to a negligible level by scin tillator timing cuts. Muon candidates are required to have a momentum of more than 2GeV.
Due to the hard fragmentation of the b quark, the energy carried by the two leptons from a B° decay is large. Pairs of oppositely charged candidate leptons are therefore required to have a combined energy of more than 20GeV. The opening angle of the candi date lepton pair is required to be less than 90°, to en sure that both lepton candidates originated from the decay chain of the same primary quark. After these cuts the background consists predominantly of fake leptons with a small irreducible contribution from the decay chain b -► c£~P; c sZ+v.

Event selection
Hadronic events are selected by making use of their characteristic energy distributions and high multiplic ity [13]. A total of 3453780 events from the 1991-1995 data samples are selected.
Candidate electrons are selected in the barrel and endcap BGO calorimeters within |cos#| < 0.97, where 6 is the polar angle. An electron is character ized by an isolated energy cluster in the BGO with a shower shape consistent with that of electromagnetic particles. To reject photons, the cluster is required to match with a charged track to within 5mrad in the for Bd,s e+e T>0 » dfs the data (solid line) for d) e+e~\ e) and f) e V * -The dashed line represents the background contribution which is estimated from the Monte Carlo sample of hadronic Z decays for which the same se lection used for the data is applied. No signal is seen /.i+f£~, or B°i s decays.

Determination of branching fractions
Upper limits on the branching fractions for these processes are obtained from binned maximumlikelihood fits to the Z+i~~ invariant mass distributions. For each di-lepton sample, the signal comprises two components, corresponding to B[| and Bj? decays. The likelihood function is given by: where Nj is the number of observed data events in mass bin and /¿d,i, ¿¿s,i» and /¿b.j denote the expected numbers of events for B §, B^, and background respec-tively.
The results for B® -+ e^e" and B|? -> are the first limits set on these decay modes. The results for Bg -> e+e" , Bg -► / ¿ V . Bg -> e*/**, and B° -» ¡xv a r e consistent with the slightly more stringent limits from CLEO [ 15] and CDF [ 16].