The Standard Model predicts that CP-violating phenomena are governed solely by the single imaginary parameter of the Cabibbo-Kobayashi-Maskawa matrix of complex quark couplings. The first observations of CP violation in the neutral B system were recently reported, and they have been interpreted widely as induced by B0-B0bar mixing. To date, direct CP violation has only been observed in the neutral kaon system. Direct CP violation in a given decay requires contributions from two or more amplitudes which differ in both CP-violating (weak) and CP-conserving (strong) phases. In the B system, these conditions are expected to be met in some charmless hadronic decays, and direct CP violation can occur at sizeable levels, depending on the magnitude of the strong phase difference or on the presence of new physics. Previous analyses, mainly focusing on two-pseudoscalar final states, have not observed direct CP violation in these decays. We present a search for direct CP violation in the vector-pseudoscalar decay B -> K*(892)+- pi-+. We express the difference between the decay rates for B0bar -> K*(892)- pi+ and B0 -> K*(892)+ pi- in terms of an asymmetry: A_CP = ( Br( B0bar -> K*(892)- pi+ ) - Br( B0 -> K*(892)+ pi- ) ) / ( Br( B0bar -> K*(892)- pi+ ) + Br( B0 -> K*(892)+ pi- ) ). We consider both K*(892)+- submodes, K*(892)+- -> K0S pi+- and K*(892)+- -> K+- pi0 by analyzing the final states K0S pi+- h-+ and K+- h-+ pi0, where h-+ denotes a charged pion or kaon. We perform a maximum likelihood fit in the K0S pi+- h-+ and K+- h-+ pi0 Dalitz plots to distinguish B -> K*(892)+- pi-+ from other intermediate resonances or non-resonant three-body decays. The CP-averaged branching fraction for B -> K*(892)+- pi-+ has been measured by the Belle and CLEO Collaborations, and the work described here is an extension of that previous CLEO analysis.

We perform a maximum likelihood fit to the two topologies mentioned above. We allow for twelve signal components describing three-body non-resonant decay and quasi-two-body decay with intermediate resonances (K*(892), K*0(1430), rho(770), and f0(980)), including K*(892)+- pi-+. We neglect interference among these signal processes and assign a systematic uncertainty estimated from Monte Carlo simulation. The fit also has four background components: pion and kaon hypotheses for h+- for continuum e+e- -> q qbar (with q = u, d, s, c) and b -> c decays. We fit for the charge asymmetry between h+ and h- in each signal mode. In both topologies, the h+- is chosen to be the faster of the two tracks. In B -> K*(892)+- pi-+ decays, the h+- is the track that recoils from the K*(892)+- more than 99.99% of the time, so its charge uniquely distinguishes B0bar decays from B0 decays. Thus, the charge asymmetry is essentially the same as the CP asymmetry in this mode. We do not fit for the charge asymmetry in the background components, but we measure them to be consistent with zero. We fit both topologies simultaneously, constraining the branching fractions and charge asymmetries in each B -> K*(892)+- h-+ or B -> K*0(1430)+- h-+ mode to be equal in its two K* submodes, which are related by isospin.

In the previous CLEO analysis, the B -> K*(892)+- pi-+ event yields were measured to be 12.6+4.6-3.9 for K*(892)+- -> K0S pi+- and 6.1+2.2-1.9 for K*(892)+- -> K+- pi0, with a combined statistical significance of 4.6 sigma. In this analysis, we measure the CP asymmetry in B -> K*(892)+- pi-+ to be 0.26 +0.33-0.34(stat.) +0.10-0.08(syst.). The dominant contributions to the systematic uncertainty are statistical uncertainties in the probability density functions used in the likelihood fit and variations in the fitting method. Using both statistical and systematic uncertainties, we establish a 90% confidence level interval for the CP asymmetry of [-0.31,0.78].

*Measurement of the Charge Asymmetry in B --> K ^{*}(892)^{+-} pi^{-+}*

B. I. Eisenstein et al.

Phys.Rev.D68:017101,2003

CLNS 03/1823, CLEO 03-07 - SPIRES Search - hep-ex/0304036

Figure 1

cleoac@mail.lepp.cornell.edu Last modified: Wed Jan 28 18:36:25 EST 2004