Searches for new physics in B decays at BaBar Richard Kass for the BaBar Collaboration Outline of Talk Searches for: B0invisible (γ) Charged Lepton Flavor Violation Measurement of BD(*)τν Richard Kass CIPANP 1 Search for Binvisible (g) Search for B decays with neutrinos or other non-interacting (e.g. SUSY) particles in the final state. In SM B vv is highly suppressed by ~(mv/mb)2 In SM BF(B vvγ) is ~10-9 In SUSY models BF(B invisible) can be enhanced to ~(0.1-1)x10-6 due to neutrino + neutralino production A. Dedes, H. Dreiner, and P. Richardson, Phys. Rev.D65, 015001 (2001) Previous Babar results with 88.5 fb-1 (B. Aubert et al., Phys.Rev.Lett.93:091802,2004 ): BF(B invisible)<22x10-5, BF(B vvγ) <4.7x10-5 Richard Kass CIPANP 2 Binvisible (γ) analysis details Use BaBar data sample: 471x106 BB pairs taken at U(4S) Divide the event into a signal B and a tag B Use B0D(*)-l+v as the tag B. l=electron/muon Semi-leptonic B tag has higher efficiency than fully reconstructed B tag B0 → D-lν D- → K+π-π-/KSπB0 → D*-lν D*- → D0π-/D-π0 D0 → K-π+/K-π+π0/K-π+π+πUse cosine between B & D(*)l assuming only 1 v is missing common vertex for D(*)l system Signal side of the event: no charged tracks, require photon with ECM>1.2 GeV for Binvisible γ Use Neural Net to discriminate signal from background (B’s, qq, etc) B → Invisible BaBar Preliminary Dtag : 9 variables D*tag : 6 variables B → Inv.+γ Dtag : 6 variables D*tag : 4 variables Richard Kass CIPANP 3 Binvisible (γ) analysis details Extract event yields using unbinned Max. Likelihood fit Discriminating variable is the extra energy in the EM calorimeter, Eextra For B → invisible+γ the high energy γ is removed from Eextra. Signal region has Eextra <1.2 GeV Signal efficiencies: B0invisible =17.8x10-4, B0invisible+γ= 16.0x10-4 Preliminary B0invisible Preliminary B0invisible+γ Signal and background yields determined from ML fit Richard Kass CIPANP 4 Binv (γ) systematic errors & results Use “unphysical” B+inv (γ) as a control sample to check the analysis Bayesian approach used to calculate 90% CL upper limits UL 0 P( B)dB / P( B)dB 0.9 0 P(B) assumes flat priors, systematics modeled as gaussians BR(B0invisible) <2.4x10-5 BR(B0invisible + γ) <1.7x10-5 Richard Kass CIPANP 5 Search for Charged Lepton Flavor Violation in B decay submitted to PRD arXiv:1204.2852v1 Look for charged B decays B±h±τl with h=π/K and l=e/μ LFV highly suppressed in SM ~ (mv/mW)2 Possible to greatly enhance the rate in extensions to SM 2nd & 3rd generations favored if coupling ~ mass NP SM+v mixing φs is a scalar Sher & Yuan, Phys.Rev.D44:1461-1472,1991 consider φs and estimate: B±h±τe < 3x10-3 B±h±τ μ < 3x10-4 BaBar: B±K±τ μ < 7.7x10-5 @90% CL, PRL 99, 201801 (2007) Richard Kass CIPANP 6 B±h±τl Analysis Details Use BaBar data sample: 472x106 BB pairs taken at Y(4S) Divide the event into a signal B and a tag B Tag B reconstructed using hadronic decays B±D(*)0XX-=combination of ps & Ks Signal B reconstructed from B±h± tl 3 charged tracks h±= π± or K±, l=e or m indirectly reconstruct the tau (use 1-prong taus: tm, te, , tnp0)p-): pt=-ptag- ph- pl Et=Ebeam– Eh- El m2t=E2t- |pt|2 |mt-mt,PDG|<60 MeV/c2 Use BD(*)0lv, D0Kπ as a control sample for BF normalization NDlv /ND*lv /ND**lv determined from ML fit to data BD(*)0 μv BaBar Preliminary Richard Kass D0 D*0 D**0 CIPANP 7 B±h±τl Analysis Details Reject continuum (e+e-qq) background using a likelihood ratio BaBar Preliminary Variables include: cosqthrust neutral energy in calorimeter muon quality (primary, tau daughter) Extract event yields using an unbinned Max. Likelihood fit Use poisson PDFs for the three tau channels (m, e, np0)p-) Assume uniform 3-body phase space for the signal decays Set 90% UL CL using Feldman-Cousins procedure Richard Kass CIPANP 8 B±h±τl Results Look at 8 channels, NO signals, Only ULs BaBar Preliminary BaBar Preliminary Combine modes assuming B(B+h+τl )= B(B-h-τl ) 5X better model independent bounds on NP scale in μτ flavor changing operators. Λbd>11 TeV, Λbs>15 TeV D. Black, et al, PRD 66, 053002 (2002) more details in “extra slides” Richard Kass CIPANP 9 BD(*)τν D(*)tn H-,W- t- arXiv:1205.5442 submitted to PRL Large mass of tau adds sensitivity to additional helicity amplitude. For BD(*)τν we have: We can compare our rate measurements with the SM predictions for: B( B Dtn ) R ( D) 0.293 0.017 B( B Dn ) B( B D*tn ) R( D ) 0.252 0.003 B( B D*n ) * S. Fajfer, J. F. Kamenik, I. Nisandzic, arXiv:1203.2654, , J. F. Kamenik, F. Mescia, PRD 78 014003 (2008) These ratios are sensitive to physics beyond the SM Richard Kass CIPANP 10 Analysis Details Use BaBar data sample: 471x106 BB pairs taken at Y(4S) Divide the event into a signal B and a tag B Fully reconstructed tag B using hadronic decays Signal B reconstructed from a lepton (te/m) & D(*) No additional charged particles allowed in event Require q2 =(pB-pD(*))2 > 4 GeV2 Backgrounds are suppressed using Boosted Decision Trees Extract event yields using an unbinned 2-D Max. Likelihood fit Perform fit using lepton momentum in B rest frame (pl*) & (missing mass)2 mm2=(pe+e--ptag-pD(*) -pl)2 Signal Samples: D0l, D*0l, D+l, D*+l with l= e or m Fixed backgrounds: B0-B+ cross feed, combinatorial, continuum Use data control samples for corrections & validation: BD**lν, D(*)lν control samples BaBar Preliminary BaBar Preliminary Richard Kass CIPANP 11 BD*τν Yields D*0tn D*+tn D*tn Nsig 693±62 245±27 888±63 Stat. sig. (s) 11.3 11.6 16.4 R(D*) 0.322±0.032 0.355±0.039 0.332±0.024 1.74±0.19 1.76±0.13 B(BD*τν ) % 1.71±0.17 fit assumes R(D*0)=R(D*+) statistical errors only D*0 D*+ m2miss (GeV2) Richard Kass D*0 Events/100MeV MeV Events/100 Events/25 MeV BaBar Preliminary Free yields D*+ Fixed yield pl* (GeV) CIPANP 12 BDτν Yields D0tn D+tn Dtn Nsig 314±60 177±31 489±63 Stat. sig. (s) 5.5 6.1 8.4 R(D) 0.429±0.082 0.469±0.084 0.440±0.058 B(BDτν ) % 0.99±0.19 1.01±0.18 1.02±0.13 fit assumes R(D0)=R(D+) statistical errors only D0 D0 D+ Events/100 MeV Events/25 MeV BaBar Preliminary Free yields D+ Fixed yield M2miss (GeV2) Richard Kass pl* (GeV) CIPANP 13 Systematic Errors & Results Source R(D) % R(D*) % r D**lv bkg 5.8 3.7 0.62 MC statistics 5.0 2.5 -0.48 Cont. & BB bkg 4.9 2.7 -0.30 esig/enorm 2.6 1.6 0.22 System. Uncert. 9.5 5.3 0.05 Statist. Uncert. 13.1 7.1 -0.45 Total Uncertainty 16.2 9.0 -0.27 r=correlation coeff. R(D)=0.440±0.058±0.042 R(D*)=0.332±0.024±0.018 SM Average does not include this measurement Richard Kass CIPANP 14 Comparison with SM BABAR SM D R(D) 0.440 +/- 0.071 0.293 +/- 0.017 2.0 s R(D*) 0.332 +/- 0.029 0.252 +/- 0.003 2.7 s The combination of the two measurements (-0.27 correlation) yields 2=14.6 for 2 DOF and p-value = 6.9x10-4 Data differ from SM rate by 3.4 s Richard Kass CIPANP 15 Charged Higgs? Compare our R(D) & R(D*) measurements with Type II 2HDM red=theory ±1s blue=exp ± 1s SM:tanb/mH=0 PDF shapes and efficiencies are recalculated vs tanb/mH Data match the 2HDM model for: R(D): tanb/mH=0.44 ±0.02 & R(D*): tanb/mH=0.75 ±0.04 Exclude Type II 2HDM at 99.8% CL Richard Kass CIPANP 16 Summary & Conclusions *Search for B0invisible (γ): improved upper limits: BR(B0invisible) <2.4x10-5 BR(B0invisible + γ) <1.7x10-5 *Search for Charged Lepton Flavor Violation in B decay BR(B+K+ τμ) <4.8x10-5 BR(B+K+ τe) <3.0x10-5 BR(B+π+ τμ) <7.2x10-5 BR(B+π+ τe) <7.5x10-5 improved new submitted to PRD arXiv:1204.2852v1 *Measurement of BD(*)τν: Larger than SM predictions R(D) R(D*) BABAR 0.440 +/- 0.071 0.332 +/- 0.029 SM 0.293 +/- 0.017 0.252 +/- 0.003 2.0 s 2.7 s D submitted to PRL arXiv:1205.5442 Exclude Type II 2HDM at 99.8% CL Richard Kass CIPANP 17 Extra slides Richard Kass CIPANP 18 B±h±τl Analysis Details Use BaBar data sample: 472x106 BB pairs taken at Y(4S) Divide the event into a signal B and a tag B The tag B is fully reconstructed using hadronic decays B±D(*)0Xwith X-=n1π, n2K, n3Ks, n4π0, n1+n2 ≤ 5, n3 ≤2, n4 ≤2 The signal B is reconstructed from B±h± tl 3 charged tracks h= π or K l=e or m indirectly reconstruct the tau (use 1-prong taus: tm, te, , tnp0)p-): pt=-ptag- ph- pl Et=Ebeam– Eh- El m2t=E2t- |pt|2 |mt-mt,PDG|<60 MeV/c2 Use BD(*)0lv, D0Kπ as a control sample for BF normalization NDlv /ND*lv /ND**lv determined from ML fit to data BD(*)0 μv BaBar Preliminary Richard Kass D0 D*0 D**0 N htl N Dln Bhtl e htl BDln e Dln CIPANP e tag Dln e tag htl Bhtl e htl S 0 ≈1 19 B±h±τl Results submitted to PRD arXiv:1204.2852v1 Richard Kass CIPANP 20 BD(*)τν Yields arXiv:1205.5442 Richard Kass CIPANP 21 PEP-II at SLAC asymmetric e+e− collider: 9 GeV (e-)/3.1 GeV (e+) PEP-II Peak Luminosity 1.2 x 1034 cm-2s-1 BaBar recorded 429 fb-1 at Y(4S) 4.7 x 108 U(4S)→BB events Richard Kass CIPANP 22 BaBar Detector Electromagnetic Calorimeter (EMC) 1.5 T Solenoid Detector of Internally Recflected Cherenkov Light (DIRC) Drift Chamber (DCH) Instrumented Flux Return (IFR) Silicon Vertex Tracker (SVT) SVT, DCH: charged particle tracking: vertex & mom. resolution, K0s/Λ EMC: electromagnetic calorimeter: g/e/π0/η DIRC, IFR, DCH: charged particle ID: π/μ/K/p Highly efficient trigger for B mesons Richard Kass CIPANP 23 Analysis Technique Threshold kinematics: we know the initial energy (E*beam) of the Y(4S) system Therefore we know the energy & magnitude of momentum of each B meson *2 mES Ebeam - p*B2 Signal * DE E B* - E beam Event topology Signal (spherical) Background Background (jet-structure) Also, use neural networks + unbinned maximum likelihood fits Richard Kass CIPANP 24 BaBar K/p ID D*+ → D0p+ D0→ K+ p- BaBar DIRC Richard Kass CIPANP 25