Search for τ � e γ / µγ K.Hayasaka(Nagoya U.) Belle Collaboration 1 Sep/15/2004
Introduction � Lepton Flavour Violating ( LFV ) process: forbidden in SM � probe of new physics � SUSY model predicts: ~ ~ B B most sensitive mode in the LFV decays. � We SHOULD search for not only for τ � µγ but also for τ � e γ . 2 Sep/15/2004 Tau04 @ Nara
KEKB accelerator/Belle detector � asymmetric collider •asymmetric detector Peak Lum. 1.4x10 34 (cm -2 s -1 ) Total Logged Lum. 288fb -1 � µ− ID:eff. 87.5% � cross section@ s~110GeV 2 _ σ (bb): σ ( ττ )=1.05:0.912 � e -ID:eff. 92.4% � A B-factory is also a τ -factory! 3 Sep/15/2004 Tau04 @ Nara
Signatures of the Signal and of the BGs � Signal Event � Expected BG Events / – ( e + γ )+( e + n γ ) – Bhabha process ττ process – ν γ e γ e e ν γ τ τ e e e e e e τ τ e generic decay generic decay 4 Sep/15/2004 Tau04 @ Nara
Selection Criteria � 2 charged tracks + more than 1 γ e – signal side : 1 charged+1 photon – tag side : 1 charged+ n photons γ � e-ID>0.9 & e-ID<0.1 τ 1-1 prong � restrict kinematical variables e e – momentum e ,tag, γ ,missing – polar angle e ,tag, γ ,missing τ – opening angle � e -tag, e - γ , tag-missing generic decay – missing mass vs missing mom. � analysis for 87fb -1 data sample tag missing 5 Sep/15/2004 Tau04 @ Nara
Nagoya Cut � P miss vs m 2miss ττ events have large missing mass ν ττ MC γ signal MC e ν τ e e τ ν charged 98% of the τ+τ− bkg is removed. particle 6 Sep/15/2004 Tau04 @ Nara
M e γ and ∆ E resolutions � All selections applied for signal MC events • Asymmetric Gaussian 6.5% # of events. # of events. M e γ ∆ E +1 σ - +1 σ - 25.7/14.3MeV/c 2 84.8/36.0MeV ∆ E=E e +E γ -E beam @CM e γ M e γ = (P e +P γ ) 2 - +5 σ region 7 Sep/15/2004 Tau04 @ Nara
Blind Analysis � Signal dominant region is masked. ττ BG : ττ MC Masked! Bhabha BG : data 0 Shape of BG distribution is evaluated. ∆ E(GeV) Its height is decided with data -0.25 distribution of the side-band region. -0.5 1.6 1.8 2 M e γ (GeV/c 2 ) side-band region 8 Sep/15/2004 Tau04 @ Nara
BG distribution (MC) curve � Estimate BG distribution (Landau+Gauss) from ττ MC M e γ (GeV/c 2 ) 10 # of ev. 1.9 - 2.0 0.5 5 0 10 # of ev. 1.8 - 1.9 5 0 10 # of ev. 1.7 - 1.8 ∆ E(GeV) 5 0 0 10 # of ev. 1.6 - 1.7 5 0 10 # of ev. 1.5 - 1.6 5 reproduce BG 0 -0.5 -1 -0.5 0 0.5 distribution 1.6 1.8 2 ∆ E (GeV) 9 M e γ (GeV/c 2 ) Sep/15/2004 by function Tau04 @ Nara
BG distribution (data) � Estimate from MC and side-band of data data(side band) ττ MC ee ~ 64 = 61 3 ττ MC (Landau+Gauss) Masked! profile plot curve 10 Sep/15/2004 Tau04 @ Nara
Final Candidates ( τ � e γ ) � 60 events found. (# of estimated BG:64) - � 20 events survived in +5 σ region signal MC - signal events +5 σ dominate surviving data 11 Sep/15/2004 Tau04 @ Nara
Evaluation of signal events � fit by unbinned expanded maximum likelihood with signal and BG shape � s 0 =0, b 0 =20 � Estimation for U.L. of s 90 @ 90%CL � by Toy MC: generate 10000 events Signal yield : s 0 result e-ID ineff. 0.01/0.02 ev. BG function 0.13 ev. Efficiency & Luminosity : – s 90 =3.8 events 2 ε N ττ Track rec. eff. 2.0% � Branching fraction Photon rec. eff. 2.8% Selection criteria 2.5% Br=s 0 /2 ε N ττ <3.8x10 -7 Luminosity 1.4% Trigger eff. 5.0% MC statistics 0.2% ε :detection efficiency Total 6.8% N ττ :total event number 12 Sep/15/2004 Tau04 @ Nara
Search for τ � µγ � Almost same selection criteria as τ � e γ – for tag side track, require not to be µ � Main BG : ττγ & µµγ ( � from µ -ID ineff.) τ pair MC ( µµ ) Data ( µµ ) 20 M inv (GeV) Data 10 10 2.1~2.2 (x ineff.) ∆ E (GeV) 0 0 10 10 0.2 2.0~2.1 0 0 10 10 1.9~2.0 0 0 0 10 10 1.8~1.9 0 0 10 10 1.7~1.8 0 0 -0.2 10 10 1.6~1.7 0 0 10 10 1.5~1.6 -0.4 0 0 10 10 1.4~1.5 1.6 1.8 2 0 0 -0.5 0.0 -0.5 0.0 0.5 M inv (GeV/c 2 ) ∆ E (GeV) ∆ E (GeV) 13 Sep/15/2004 Tau04 @ Nara
Final Candidates ( τ � µγ ) � 54 events survived in +5 σ region - 1.71<M inv <1.82GeV/c 2 ∆ E (GeV) signal MC +5 σ 11% - 0.2 10 Data BG function # of events Signal MC 0 5 -0.2 0 -1.0 -0.5 0.0 ∆ E (GeV) • fitting result(UEML) -0.4 • s=0 1.6 1.8 2 Evaluation for U.L. M inv (GeV/c 2 ) including • s=5.1ev. @90%C.L. systemaric • Br<3.1x10 -7 @90%C.L. uncertaunties 14 Sep/15/2004 Tau04 @ Nara
Conclusion & Summary � Obtain BR UL’s with blind analyses. � BG distributions are modeled well. � Results are 10 times more sensitive than CLEO’s. τ � e γ (86.7fb -1 ) τ � µγ (86.3fb -1 ) • N ττ =7.90x10 7 • N ττ =7.87x10 7 • 20 observed events • 54 observed events • ε =6.5% • ε =11% • s=3.8ev. @90%C.L. • s=5.1ev. @90%C.L. • Br<3.8x10 -7 @90%C.L. • Br<3.1x10 -7 @90%C.L. Phys. Rev. Lett. 92(2004) 171802 cf. Br<2.7x10 -6 @90%C.L. (CLEO) cf. Br<1.1x10 -6 @90%C.L. (CLEO) 15 Sep/15/2004 Tau04 @ Nara
Recommend
More recommend