August 12, 2000DPF Search for B +  K + l + l - and B 0  K* 0 l + l - Theoretical predictions and experimental status Analysis methods Signal properties Main background categories –J/  [  (2S)] K and J/  [  (2S)] K* events Results Natalia Kuznetsova, UC Santa Barbara On behalf of the BaBar Collaboration

August 12, 2000DPF Theoretical issues bs l+l+ l-l- bs t W t WW Mode Product Br. Fr. # produced events / 10 fb -1 B +  K + l + l x x B 0  K * 0 e + e - (K * 0  K +  - ) TOTAL events produced 41.4 B 0  K * 0  +  - (K * 0  K +  - ) 5.7 x x x x Predicted Br. Fr.* Predicted Br. Fr.* (~35% uncertainty) (~35% uncertainty) *A.Ali et al., Phys. Rev. D. 61, (2000), (form factors from Light Cone QCD Sum Rules). l-l- l+l+ ,Z,Z b s l-l- l+l+ , Z W t For now, we are only looking at 3.15 fb -1 of data this is the first stage of a blind analysis this is the first stage of a blind analysis we expect to have 20 fb -1 by the end of the year we expect to have 20 fb -1 by the end of the year

August 12, 2000DPF  90% CL limits from CLEO (3.33 x 10 6 BB data set) and CDF (88 pb -1 ): Experimental status Mode CLEO* K+K+ < 9.7 x K0K0 < 31.0 x K* +     < 33.0 x K* 0     < 9.5 x K+e+e-K+e+e- < 11.0 x K0e+e-K0e+e- < 17.0 x K* + e + e - < 38.0 x K* 0 e + e - < 14.0 x CDF** < 5.2 x < 4.0 x *CLEO CONF 98-22, ICHEP (1998) (unpublished). **T. Affolder et al., Phys. Rev. Lett. 83, (1999), 3378.

August 12, 2000DPF Analysis Methods (1) Basic ideas: –Perform a blind analysis –Select charged particle modes only: B + -> K + e + e -, B + -> K +  +  -, B 0 -> K* 0 e + e -, B 0 -> K* 0  +  - (where K* 0 -> K +  - ) –use  E vs. m ES plane to select signal region E B *(p B *) is the B candidate energy (momentum) in the CM frame = center-of-mass energy d nal events lie outside the signal box (signal MC) MC truth tagged signal MC truth tagged signal B +  K + e + e - B +  K + e + e -  E (GeV) M ES (GeV/c 2 )

August 12, 2000DPF Analysis Methods (2) p LAB e > 0.5 GeV/c p LAB  > 1.0 GeV/cSelect electrons with p LAB e > 0.5 GeV/c and muons with p LAB  > 1.0 GeV/c Require high multiplicity events (with # of tracks > 4); veto  conversions Veto J/ ,  (2S) resonance regions Suppress continuum background using Fisher discriminant Particle ID: electron ID based on energy deposition in the CsI calorimeter muon ID based on the penetration length in the Instrumented Flux Return hadron ID based on combined drift chamber dE/dx and Cherenkov angle information K+e+e-K+e+e-K+e+e-K+e+e- K* 0 e + e K+K+K+K+ K* 0     Mode Efficiency, %

August 12, 2000DPF Predicted distributions for q 2 = M 2 l+l- B  K* 0  +  - (pole at q 2 = 0): B  K  +  - : Solid line + blue bands: SM range ( 35%); Ali et al. form factors Dotted line: SUGRA model (R 7 = -1.2, R 9 = 1.03, R 10 = 1) Long-short dashed line: SUSY model (R 7 = -0.83, R 9 = 0.92, R 10 = 1.61)

August 12, 2000DPF We have implemented event generators using the model of Ali et al. Generated distributions for q 2 = M 2 l+l- generated generated reconstructed reconstructed B  K* 0  +  - (pole at q 2 = 0): B  K  +  - :

August 12, 2000DPF Background categories  B  J/  K(K *),  (2S) K(K *) is the most serious background –this background can be controlled by a cut in  E vs. m l+l- plane –also possible to have J/  (  +  - )  K with K and  swapped –re-assign particle masses and cut on the J/  mass  B +  D 0 (  K +   )    with   misidentified as   and K + as   –re-assign particle masses and veto the D 0  Continuum (e + e -  qq) –suppressed by using a 4-variable Fisher discriminant  Combinatorial from B B events –suppressed by using vertexing

August 12, 2000DPF B  J/  [  (2S)] K ( * ) events (1) The most serious background for J/  [  (2S)] K ( * ) this analysis, J/  [  (2S)] K ( * ) events, is suppressed by using a correlated selection in the  E vs. m l+l- plane  This is needed to account for bremsstrahlung and track mismeasurement signal region J/  [  (2S)]  e + e - M e+e- (GeV/c 2 )  E (GeV) Signal B  J/  K and B  (2S) K Monte Carlo J/  (2S)

August 12, 2000DPF B  J/  [  (2S)] K ( * ) events (2) J/  (2S)  K ( * )  J/  (2S)  K ( * )  events are also used as a control sample to verify the analysis efficiency J/  (  e + e - ) K* J  (  e + e - ) K J  (     ) K J/  (     ) K* Entries/(0.005 GeV/c 2 ) M ES (GeV/c 2 ) Entries/(0.005 GeV/c 2 ) M ES (GeV/c 2 ) K ( * ) B  J/  [  (2S)] K ( * ) control sample (8 fb -1 of data)

August 12, 2000DPF Results (1) B 0  K* 0  +  - B +  K +  +  - M ES (GeV/c 2 )  E (GeV) M ES (GeV/c 2 )  E (GeV) M ES (GeV/c 2 )  E (GeV) For the purpose of setting the limit, we assume that all events in the signal region could be due to signal processes  Background is not subtracted B +  K + e + e - B 0  K* 0 e + e -

August 12, 2000DPF Results (2) K+ e+e-K+ e+e-K+ e+e-K+ e+e- K* 0 e + e K+K+K+K+ K* 0     1 0 Mode # obs. evts Total 3 # bkg. evts Preliminary Preliminary 90% C.L. limit < 12.5 x < 24.1 x < 24.5 x < 8.3 x < 8.3 x background The number of background events is extracted from sideband in data 1 signal  Expect 1 signal event based on Geant MC (Ali et al. predictions).

August 12, 2000DPF A candidate B +  K + e + e - event e+ e- K+K+K+K+

August 12, 2000DPF Summary B +  K + l + l -We have searched for the rare decays B +  K + l + l - and B 0  K* 0 l + l - B 0  K* 0 l + l - using a sample of 3.67 x 10 6 B B events –We are using this sample to better understand our backgrounds 3We have found 3 candidate events total and set preliminary 90% C.L. limits B  K l + l - –The limits for the B  K l + l - modes are comparable to those set by other experiments B  K* 0 l + l - –The limits for the B  K* 0 l + l - modes are less sensitive with this data sample We are planning to analyze substantially more data in the near future –BaBar will have 20 fb -1 by the end of the year and an extra 30 fb -1 by next year