Observation of the decay B̄0→Λc+p̄π0

In a sample of 467 (cid:1) 10 6 B (cid:1) B pairs collected with the BABAR detector at the PEP-II collider at SLAC we have observed the decay (cid:1) B 0 ! (cid:1) þ c (cid:1) p(cid:2) 0 and measured the branching fraction to be ð 1 : 94 (cid:2) 0 : 17 (cid:2) 0 : 14 (cid:2) 0 : 50 Þ (cid:1) 10 (cid:3) 4 , where the uncertainties are statistical, systematic, and the uncertainty on the (cid:1) þ c ! pK (cid:3) (cid:2) þ branching fraction, respectively. We determine an upper limit of 1 : 5 (cid:1) 10 (cid:3) 6 at 90% C.L. for the product branching fraction B ð (cid:1) B 0 ! (cid:3) þ c ð 2455 Þ (cid:1) p Þ (cid:1) B ð (cid:1) þ c ! pK (cid:3) (cid:2) þ Þ . Furthermore, we observe an enhancement at the threshold of the invariant mass of the baryon-antibaryon pair.

In a sample of 467 Â 10 6 B " B pairs collected with the BABAR detector at the PEP-II collider at SLAC we have observed the decay " B 0 ! Ã þ c " p% 0 and measured the branching fraction to be ð1:94 AE 0:17 AE 0:14 AE 0:50Þ Â 10 À4 , where the uncertainties are statistical, systematic, and the uncertainty on the Ã þ c ! pK À % þ branching fraction, respectively. We determine an upper limit of 1:5 Â 10 À6 at 90% C.L. for the product branching fraction Bð " B 0 ! AE þ c ð2455Þ " pÞ Â BðÃ þ c ! pK À % þ Þ. Furthermore, we observe an enhancement at the threshold of the invariant mass of the baryon-antibaryon pair. DOI Although approximately 7% of B-meson decays have baryons in the final state, presently the sum of all measured branching fractions of exclusive baryonic B decays is only about 1% [1]. B mesons decay dominantly via b ! c transitions, hence decays to baryons should be dominated by charm baryon production or a charmed meson accompanied by noncharmed baryons. Both types of decays have been observed [2,3], and are found to have comparable branching fractions for decays to final states with the same multiplicity.
In baryonic B decays and in baryon production in general, enhancements at the threshold for the baryonantibaryon invariant mass have been observed [3,4]. This may indicate resonances near threshold or another mechanism for enhanced production of baryon-antibaryon pairs. This threshold enhancement may also explain the increase in branching fraction with final-state multiplicity and the apparent suppression of two-body decays to baryons [1,5].
The mechanisms of baryon production in heavy meson decays are poorly understood, and studies of exclusive decays may provide insight into different decay mechanisms. As will be discussed below, isospin relations will also help distinguish different primary processes.
While the B À ! Ã þ c " p% À final-state can only have an isospin I of 3=2, " B 0 ! Ã þ c " p% 0 can also have I ¼ 1=2. If both decays proceed via the same weak decay mechanism, I ¼ 3=2, the ratio of the partial decay widths of " B 0 to B À should be 2=3. However, it is also possible that the decay mechanisms are different. Thus a deviation of the ratio of partial decay widths from 2=3 would suggest a contribution from the I ¼ 1=2 final-state to the " B 0 ! Ã þ c " p% 0 decay or a contribution from the decay process where the % À is coming from the W in the B À ! Ã þ c " p% À decay. This analysis is based on a data set of about 426 fb À1 corresponding to 467 Â 10 6 B " B pairs. These data were collected with the BABAR detector at the PEP-II asymmetric-energy e þ e À collider with a center-of-mass energy, ffiffi ffi s p , at the Çð4SÞ resonance mass. An additional sample of 44:5 fb À1 , collected 40 MeV below the mass of the Çð4SÞ resonance, are used to study the continuum background e þ e À ! q " q, where q ¼ u, d, s, or c. The signal efficiency is determined using a detailed GEANT4 [11] Monte Carlo (MC) simulation of the BABAR detector that generates MC events uniformly in the Ã þ c " p% 0 phase space. MC events are also used to study the background contributions.
The BABAR detector is described in detail elsewhere [12]. Charged particles are distinguished and their momenta measured in the tracking system consisting of a five-layer double-sided silicon vertex tracker (SVT) and a 40-layer drift chamber (DCH). An internally reflecting ring imaging Cherenkov detector (DIRC) is also used to distinguish charged particles and a CsI(Tl) electromagnetic calorimeter (EMC) is used to detect photons.
Likelihood ratios based on information from SVT, DCH and DIRC are used to identify protons and kaons. The efficiency for the kaon selection is around 90% while the rate for misidentifying pions and protons as kaons varies between 5% and 10%, depending on track momentum. The identification efficiency for the proton selection is greater than 90% while the misidentification rate of identifying kaons and pions as protons varies between 3% and 15%, depending on track momentum.
Two photons are selected as electromagnetic showers in the EMC with the expected shape and are combined to form a % 0 candidate, where the photon with the lower energy must have an energy greater than 60 MeV, while the second photon must have an energy greater than 100 MeV. The invariant mass of the combination is required to be between 120 MeV=c 2 and 145 MeV=c 2 .
The Ã þ c candidates are reconstructed in the decay mode Ã þ c ! pK À % þ , and a fit with geometric constraint applied to the common vertex must have a 1 2 probability greater than 0.1%. The invariant pK À % þ mass must be within 2:5' of the fitted peak of the mass distribution, 2:276 < mðpK À % þ Þ < 2:296 GeV=c 2 . The Ã þ c and % 0 candidates are then combined with a " p candidate in a fit using kinematic constraints to form a " B 0 candidate. In the fit, the mass of the pK À % þ candidate is constrained to the mass of the Ã þ c and the mass of the combination to the mass of the % 0 [1]. The 1 2 probability of this fit must be greater than 0.1%.
The analysis makes use of two almost independent kinematic variables, ÁE and m ES , where where ðE 0 ; p 0 Þ is the four momentum of the e þ e À system and p B is the B candidate momentum, both measured in the laboratory frame. The m ES distribution for signal events peaks at the B mass and the distribution of ÁE for signal events is centered around zero. Candidates arising from other B decays, with more final-state particles, such as " A considerable background comes from B À ! Ã þ c " p% À decays, and, in particular, from the B À ! AE 0 c ð2455Þ " p, AE 0 c ð2455Þ ! Ã þ c % À decays, in which the Ã þ c " p pair from B À decay is combined with a % 0 from the decay of the B þ meson. To suppress this background, we reconstruct B À ! Ã þ c " p% À , and reject the event if jÁEj < 50 MeV and m ES > 5:27 GeV=c 2 for such a B À candidate, or if the condition 2400 MeV=c 2 < mðÃ þ c % À Þ < 2465 MeV=c 2 is satisfied (veto cuts). These two requirements keep 98% of the signal, while they remove 85% of B À ! AE 0 c ð2455Þ " p events. The remaining 15% of the background events do not peak in the signal ÁE À m ES region.
The continuum background is reduced by a requirement on the thrust value of the event T < 0:75, where we include both charged particles and photons in this calculation. The thrust is defined as whereT is the thrust axis defined as the direction which maximizes the sum of the longitudinal momenta of the particles, and p i the momentum vector of the i-th particle in the CMS. This selection keeps 83% of the signal but only 25% of the continuum background, as determined from MC simulation and continuum data collected 40 MeV below the Çð4SÞ energy.
To further reduce the background from continuum and B " B events, mainly coming from combinations of low energy, only one " B 0 candidate per event is selected. In events with more than one candidate (about 10% of the events), first the candidate(s) with the invariant mass mðÞ closest to the % 0 nominal mass are selected. For events with multiple candidates containing the same % 0 , the candidate with the pK À % þ mass closest to the nominal Ã c mass is retained. If there are still multiple B candidates, the candidate with the highest probability of the kinematic vertex fit is used. Figure 1 shows a comparison between the ÁE distribution of candidates reconstructed in data and in signal MC events, in which signal events are obtained by a fit to the m ES distribution in every ÁE bin, as described below.
The number of reconstructed signal candidates is determined from a binned 1 2 fit to the observed m ES distribution shown in Fig. 2. The sum of two Gaussian distributions with different means is used to describe the signal. The parameters of the two Gaussians are fixed to the values obtained from a fit to signal MC events. The background is described by the function [13] where m 0 ¼ 5:289 GeV=c 2 is the kinematic end-point value, c a shape parameter left free in the fit, and n is the normalization. There are 273 AE 23 signal candidates seen in data and the significance of this observation is more than 10'.
The number of produced signal events used to measure the branching fraction is determined by a fit to the efficiency-corrected m ES distribution using the same parametrization as before. The events are weighted with the inverse of the efficiency as a function of the invariant mass mðÃ þ c % 0 Þ. To compute the efficiency, the signal MC sample is divided in 10 intervals of mðÃ þ c % 0 Þ. For each interval the m ES distribution is fitted to extract the signal MC yield. The efficiency for each interval is computed dividing the yield by the number of events generated in this interval. The resulting efficiency distribution is then fitted by a 4th order polynomial. The averaged signal efficiency is 6.0%.
To check for peaking background from other B decays and random combinations, the analysis is repeated for selected samples without mass constraints on the % 0 and Ã þ c mass. The signal yields after subtraction of background obtained from the invariant mass distributions of the % 0 and Ã þ c are found to be consistent with the default analysis. The systematic uncertainties are mainly derived from studies of data control samples and by comparison of data and MC events. The main systematic uncertainty arises from differences between data and MC events in the ÁE distribution seen in Fig. 1. The difference between the cut efficiency in MC and data, relative to the MC one, is used as the systematic uncertainty (4.6%). Other systematic uncertainties arise from the veto cuts (3.4%), the % 0 reconstruction efficiency (3.0%), the particle identification (1.2%), the number of B 0 " B 0 pairs (1.1%) and the reconstruction efficiency of charged tracks (0.9%). To determine the uncertainty from the MC model we use to generate signal events, these signal events are reweighted depending on mð " p% 0 Þ, and a new efficiency function is calculated. The data m ES distribution is then corrected for reconstruction efficiencies with this function and fitted as before. The difference in the number of signal events we use as the systematic uncertainty of the specific MC model (2.2%). The systematic uncertainty due to the fit is determined by changing the cut-off value of the background function by 1 MeV=c 2 (0.50%). The individual contributions to the systematic uncertainty are added in quadrature, resulting in the total of 7.1%.
In Fig. 4, the measured mðÃ þ c " pÞ distribution is compared with a MC simulated one, generated with a phase space distribution for the decay to Ã þ c " p% 0 and normalized to the number of data events. To extract the signal distribution events, the m ES distribution is fitted in every bin of mðÃ þ c " pÞ. There is a clear difference in shape between data and simulation, with a clear enhancement at low mass, with