Rare B Decays : B?ln, ll, llg

1
Rare B Decays B?ln, ll, llg
Nagoya University Koji Ikado The 11th
International Conference on B-Physics at Hadron
Machines (Beauty 2006) Sep. 28, 2006
2
Introduction
B?ln

• Leptonic decay proceeding through W boson
  annihilation in the Standard Model
• Decay rate simply related to B meson decay
  constant fB and Vub
• Helicity suppressed by lepton mass tn is
  favored over en and mn
• Clean experimental method of measuring B meson
  decay constant fB
• Physics beyond the SM could enhance the branching
  fraction through the introduction of a charged
  Higgs boson

3
B0?l l-
Introduction (cont.)

• Neutral B mesons decay to ll- via box or penguin
  annihilation
• Branching fractions are suppressed by lepton mass
• Flavor violating channel (B0 ? e m, etc.) are
  forbidden in SM
• Some new-physics can enhance the branching
  fractions by orders of magnitude

B(B0 ? ee) 10-15 B(B0 ? mm) 10-10 B(B0
? n n ) zero
4
B0?l l-g
Introduction (cont.)

• Radiative dilepton decays
• Photon emission from the initial state relaxes
  the helicity suppression
• SM predictions
• SM branching fraction much below experimental
  sensitivity
• ? Search for new physics

B(B0 ? eeg) 10-10B(B0 ? mmg) 10-10
B(B0 ? nng) 9 10-10
5
B?tn
6
B?tn Analysis Concepts_at_Belle

• B decays with missing neutrinos lack the
  kinematic constraints which are used to separate
  signal events from backgrounds (Mbc and DE)
• Reconstruct the decay of the non-signal B
  (tagging), then look for the signal decay in
  whatever is left over

More than 2 neutrinos appear in B ?tn decay
Tagging side Fully reconstruct hadronic modes
Signal side Reconstruct particles from t decay
7
Fully Reconstructed Bs with 449 M BB _at_ Belle
7 modes
6 modes

• 180 channels used

2 modes

• Beam constrained mass

N 680k eff. 0.29 purity 57
N 412 k eff. 0.19 purity 52

• m 5.28 GeV/c2
• s 3 MeV/c2 due to s(Ebeam)

10 for feed-across between B and
B0
Neutral B
Charged B
Signal region -0.08 lt DE lt 0.06 GeV, Mbc gt 5.27
GeV/c2
8
Signal Selection

• t lepton is identified in the 5 decay modes

81 of all t decay modes
Total efficiency with t decay branching fraction
15.81 ?0.05
All the selection criteria have been optimized to
achieve the highest sensitivity
9
Signal Selection

• Extra neutral energy in calorimeter EECL
• - Most powerful variable for separating signal
  and background
• -Total calorimeter(ECL) energy from the neutral
  clusters which are not associated with the tag B

• Minimum energy threshold
• Barrel 50 MeV
• Forward(Backward) endcap 100(150) MeV

Zero or small value of EECL arising only
from beam background
Higher EECL due to additional neutral clusters

• BB- is dominant in background

10
Extra Calorimeter Energy in Data

• Observed events compared with background expected
• 
  

414 fb-1
After finalizing the signal selection criteria,
the signal region is examined
Number of data in signal region

Observe excess in signal region
11

Fit Results

• Unbinned Likelihood fit to the obtained EECL
  distributions

Signal background
S Significance with systematics
Background
Background yield is consistent with the
expectation from the MC simulation
(peaking background is included)
B?tn Signal
Observe 17.2 events with a significance of
3.5s
5.3 - 4.7
Signal shape Gauss exponential Background
shape Gauss second-order
12
Branching Fraction

• All t decay modes combined
• Product of B meson decay constant fB and CKM
  matrix element Vub
• Using Vub (4.38 ? 0.33)10-3 from HFAG

• B(B?tn) (1.59 ? 0.40)10-4 (SM prediction)
• (Vub (4.38 ? 0.33)10-3 fB
  0.216 ? 0.022 GeV )

fB 0.216 ? 0.022 GeV (HPQCD)
Phys. Rev. Lett. 95, 212001 (2005)
13
Constraints on Physics Parameters

• Constraint on Charged Higgs
  

Phys. Rev. D48, 2342 (1993)
rH
2s
95.5C.L. exclusion boundaries
14
B?tn Search _at_ Babar

• Babar searches for in a sample of 324x106 BB
  events
• Reconstruct one B in a semileptonic final state
  B?DlnX
• D?K p, K p p p, K p p, Ks p p (Xg, p from D0
  is not explicitly reconstructed)
• Require lepton CM momentum gt 0.8 GeV
• Require that -2 lt cosqB-D0l lt 1
• Parent B energy and momentum are determined from
  the beam energy
• Tagged B reconstruction efficiency 0.7
• Discriminate signal from
• background using Eextra

• t lepton is identified in the
• 4 decay modes

15

B?tn Search _at_ Babar (cont.)

• Results
• Observe a result consistent with zero signal at
  1.3s, set a limit on the branching fraction and
  quote a central value
• Calculate the product of fB and Vub

Babar preliminary
16
B?en B?mn
17
B?en B?mn_at_Babar

• Search for B?ln in a sample of 229x106 BB events
• Reconstruct one B in a fully hadronic final state
  
• Reconstruct B? D()0 X

Identify a monoenergetic electron or muon
recoiling against Btag
18
B?en B?mn_at_Babar

• Observe 0 (0) events in the signal box in
  electron (muon) events
• Set upper limits on the branching fractions

208.7fb-1
Babar preliminary
19
B?en B?mn_at_Belle

• Highly energetic lepton
• Companion B reconstructed with the remaining
  particles

Belle preliminary
PlB rest (GeV)
Belle preliminary
Update is coming soon
20
B0 ? l l
21
B0 ? l l (ee, mm, em) _at_Babar

• Very straightforward reconstruct mES and DE
• Extremely clean monochromatic kinematics
• Strategy define selection for sideband
• and blind signal box in mES, DE
• Selection variables
• cos qT qT is angle between thrust
• axes of ll candidate and rest of event
• mROE invariant mass of rest of event
• R2 normd 2nd Fox-Wolfram moment
• Ntrk ½Ng measure of multiplicity
• EEMC lt 11 GeV, rejects QED

accept
accept
accept
accept
Signal MC (blue) vs. sideband data
22
B0 ? l l (ee, mm, em) results
BB pairs used (122.51.0)106
Signal regions
Events observed
ee
mm

• Limits

B(B0 ? ee) lt 6.1 10-8 (90CL) B(B0 ?
mm) lt 8.3 10-8 (90CL) B(B0 ? e m) lt 18
10-8 (90CL)
e-m
111 fb-1
Phys. Rev. Lett. 94, 221803 (2005)
B(B0 ? ee) lt 1.9 10-7 (90CL) B(B0 ?
mm) lt 1.6 10-7 (90CL) B(B0 ? e m) lt
1.7 10-7 (90CL)
78 fb-1
Phys. Rev. D 68, 111101 (2003)
B(B0d?mm-) lt 2.310-8 (90 CL)
780 pb-1
23
B0 ? tt _at_Babar

• First limit on this channel
• Experimentally very challenging 2-4 neutrinos
  in signal
• Constrain in particular leptoquark coupling and
  tanb in SUSY
• Analysis
• Reconstruct one B in a fully
• hadronic final state B? D() X
• gt280k events
• In the event remainder, look for
• two t decays (t?lnn, pn, rn)
• Kinematics of charged partilce
• momenta and residual energy
• are fed into a neutral network
• to separate signal and BG

Control sample
Data
B(B0 ? tt) lt 4.1 10-3 (90CL)
Observed events 263?19 Expected events 281?48
Phys. Rev. Lett. 96, 241802 (2006)
210 fb-1
24
B0 ? n n (invisible) _at_Babar
B pairs used (88.51.0)106

• Semileptonic tags B0?D()-ln (D- ?D0 p-)
• Require nothing in recoil no charged tracks,
  limited neutrals
• Signal obtained from ML fit to Eextra
• Signal 17 9,
• background 19 10
• Upper limits
• Systematics
• Additive 7.4 events
• Multiplicative 10.9
• Frequentist limit-setting procedure

nn
8
nng
B(B0 ? invisible) lt 22 10-5 (90CL)
Phys. Rev. Lett. 93, 091802 (2004)
25
B0?l l-g
26
B0?l l-g _at_Babar

• Search for B?llg in a sample of 324x106 BB events
• Reconstruct B candidates from two leptons and a
  photon
• Leptons required to be 0.3 lt mll lt 4.9 (4.7) GeV
  for electrons (muons)
• Reject backgrounds from J/y, y (2S) decay
  (leptons) or p0 decay(photon)
• Reject qq background using signal B kinematics
  and event shape in a Fisher discriminant
• Background is determined by extrapolating the mES
  sideband into the signal box

eesignal 6.07 ?0.14 emsignal 4.93 ?0.12
Nebkg 1.28 ?0.80 Nmbkg 1.40 ?0.42
ee-g
mm-g
27
B0?l l-g _at_Babar
e e-g
m m-g

• Observe 0 (3) events in the signal box in
  electron (muon) events
• Set upper limits on the branching fractions

B(B0 ? ee-g) lt 0.7 10-7 (90CL)B(B0 ?
mm-g ) lt 3.4 10-7 (90CL)
Babar preliminary
B(B0 ? n n g ) lt 4.7 10-5 (90CL)
Phys. Rev. Lett. 93, 091802 (2004)
28
Summary

• Performed searches for rare leptonic B decays
  B?ln, ll, llg
• Belle found first evidence for B ?tn
• - First direct measurement of the B meson decay
  constant fB

4
29
Backup Slides
30
KEKB Belle
500 fb-1

• Asymmetric-energy ee- collider
• 8GeV 3.5 GeV
• High Luminosity
• L 1.5 x 1034

Collected 560 fb-1
31
Features with Fully Reconstructed B Tag

• Merit Offline B meson Beam
• B momentum is available
• -Resolution of Mmiss2 can be significantly
  improved.
• -separate similar semileptonic decays
• ? reduce background significantly
• B-flavor is also available
• -We can treat charged neutral
• B separately
• Demerit Low statistics
• Efficiency 0.2 - 0.3
• However, we can rely on KEKB, providing World
  record luminosity

Mmiss2 for B-?D0 m- n (MC)
w/o B momentum
with B momentum
S/N
Full recon.
D() l n tag
Advanced n recon.
Lum.
trad. n recon.
Eff.
32
Signal Selection

• Extra neutral energy EECL Validation
• -Double tagged sample, Btag is fully
  reconstructed and Bsig is semileptonic mode

Bg D()0 X (fully reconstruction) B- g D0
l-n D0 p0 K- p
K- p p- p
BB- 494 ? 18
B0B0 7.9 ? 2.2
Total 502 ? 18
Data 458
Purity 90
Validate with double tagged events
33
Background Estimation
MC 41.3 ? 6.2 Data 43
MC 89.6 ? 8.0 Data 93
MC 94.2 ? 8.0 Data 96
MC 18.5 ? 4.1 Data 21
Sideband Total MC 267 ? 14 Data 274
MC 23.3 ? 4.7 Data 21
Large MC samples for ee- ?BB, qq, Xuln, Xu tn,
t t- , and rare B decays are used (including
beam-background)
Good agreement between data and MC in sideband
region ? Validity of background MC simulation
34
B?tn Candidate Event
B g D0 p K p- p p- B- g t -
n e-nn
35
B?tn and Decay Constant fB

• Expected branching fraction
• Currently, our best knowledge of fB comes from
  lattice QCD calculations with uncertainty of 10

Vub (4.38 ? 0.33)10-3 from HFAG
(hep-ex/0603003) fB 0.216 ? 0.022 GeV from
lattice QCD
HPQCD result, Phys. Rev. Lett. 95, 212001 (2005)
36
Systematic Uncertainty

• Signal selection efficiencies
• Tag reconstruction efficiency 10.5
• Difference of yields between data and MC in
  the B-? D0l-n control sample
• Number of BB 1
• Signal yield
• signal shape ambiguity estimated by varying
  the signal PDF parameters
• BG shape changing PDF
• Total systematic uncertainty

22.5 - 25.7
25.5 - 28.4
37
Constraints on Physics Parameters

• CKM parameters
• -Constraint in the (r,h) plane from the B?tn
  branching fraction and Dmd

38

Fit Results

• Mbc and Pmis distributions

39
Constraints at Super-B

• Br(B?t n) measurement
• Further accumulation of luminosity helps to
  reduce both statistical and systematic errors
• - Some of the major systematic errors come from
  limited statistics of the control sample
• Vub measurement
• lt 5 in future is an realistic goal
• fB from theory
• 10 now ? 5 (?)

Lum. DB(B?tn) exp DVub
414 fb-1 36 7.5
5 ab-1 10 5.8
50 ab-1 3 4.4
Assumption in the following plots
40
Constraints at Super-B (cont.)
95.5C.L. exclusion boundaries
DfB(LQCD) 5
5ab -1
rH
2s
50ab -1
rH