1. Measurements of Leptonic B Decays from BaBar
Gregory Dubois-Felsmann Caltech for the BaBar collaboration ICHEP 2004 Heavy Quark Mesons and Baryons session

2. All results are preliminary unless otherwise noted
Overview
Motivations for studying leptonic B decays
SM and beyond-SM considerations
B+  l+ nl
B0  l+ l–, l+ l'– , B0  n n, n n g
B+  K+ n n
Analyses
Decays to nX states: tag-recoil analyses
general methods
specific results
B0  l+ l– analysis
Conclusions and prospects
All results are preliminary unless otherwise noted
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

3. Leptonic B decays to t+ n, l+ l–, n n
Leptonic decays of heavy-quark mesons provide a laboratory
For testing straightfoward SM predictions:
For searching for non-SM effects in highly suppressed processes. Some new-physics in loops (e.g., SUSY) can enhance these by orders of magnitude. Also LFV?
Note helicity suppression b u l+ n W+ B+
Calculable in Lattice QCD
In SM: B(B+  t+ nt ) ~ (9.3±3.9)×10-5 (PDG’04 Vub, fB) (pred.) B(B+  m+ nm ) ~ (4.2±1.8)×10-7 b d
l+,n W+ W–
l–,l’–,n b
l+,n
g,Z B0 qu d W–
l–,l’–,n b qu
g,Z
l+,n
In SM: B(B0  m+m–) ~ 8×10-11 (pred.) B(B0  n n ) ~ zero W qu d
l–,l’–,n
(nn via Z only)
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

4. B decays to K/p+ n n Motivation
The flavor-changing neutral current decays B  K/p+nn occur in the Standard Model via one-loop radiative penguin and box diagrams:
SM expectation: B(B+K+nn) ~ ( ) × 10-6, B(B+p+nn ) ~ 2.8 × 10-7
Their analysis is theoretically very clean; observation of these processes would be complementary to the observation of B  K(*) l+l– and will help in understanding the basic Standard Model physics of such diagrams.
These also present another opportunity for the observation of new-physics effects in the loops. n l
– 0.6
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

5. Tag-recoil analyses: B1  tag, B2  t+n, K+nn, nn
Neutrino analyses require extra kinematic and/or particle selection constraints.
Use BB initial state to achieve this.
Explicitly identify the decay of one “tag” B
Fully reconstructed B  D +X hadronic decays full knowledge of kinematics
b  c semileptonic decays some ambiguity, more tags
Study the recoil for the decay of interest
Typically we require the recoil to have
Exact charged-particle content expected for signal
Number and total energy Eextra of neutrals observed less than an analysis-dependent threshold
Tagging efficiencies can be checked by “double-tagging”
All analyses apply anti-continuum shape cuts
n•p OR l + n•p
Btag e-
D(*)
Brecoil e+ ν
visible charged and neutral particles
Eextra = S EEMC neutrals
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

6. Semileptonic (b  c + l– + X) tags
Selection varies somewhat with analysis
The lepton must be either an electron or muon with positive PID and p*>~1GeV/c
Kaons are identified by the DIRC
D mesons are reconstructed in several K +Np final states, identified by inv. mass ±3s
D* mesons are identified by the mass difference Dm=mD*–mD
The semileptonic decay kinematics are enforced by looking at the variable cos qB,D(*)l , defined as and requiring it to be in the physical region. Some of the analyses are more tolerant of missing/extra particles (higher D*(*) states), trading purity for statistics.
Analysis
t+nt (I)
t+nt (II)
K+ n n
nn, nn g
D selection
D0Kp, K3p, Kpp0, KS0pp
D0Kp, K3p, Kpp0
D+Kpp D0Kp, K3p, Kpp0
Tag selection
D0l–nlX X: g, p0, nothing
D*0l–nl D*0D0p0,D0g
D0l–nlX X unseen
D*+l–nlX, D+l–nlX D*+D0 p+
cos qB,D(*)l requirement
-2.5 < cosq < 1.1
-1.1 < cosq < 1.1
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

7. Hadronic ( B  D + X ) tags
An additional B–  D(*)0 + X– hadronic tag sample is used for the K/p+nn and t+nt (I) analyses:
D mesons are reconstructed in D0K–p+, K–p+p+p–, K–p+p0, KS0p+p–
In the t+nt (I) analysis, D*0 Dp0 candidates are formed as well. The X– system is composed of 1-5 p±, 0-2 K±, 0-2 p0, and 0-1 KS0.
In the K+nn analysis, only D0 are considered, and the X– system is composed of 1-5 mesons, including p±, K±, and p0, with no more than two p0 used.
B mesons are identified with the kinematic variables DE=EB–Ebeam and the beam-energy substituted mass
sideband
signal
t+nt (I)
BaBar preliminary
K+ n n
BaBar preliminary
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

8. hep-ex/0407038 submitted to PRL
B+  t+ nt analysis I
Semileptonic tags: ML fit to Eextra
Consider t+  e+nent and m+nmnt
Signal: ± events Background: ± 11.8 events
Significance as a signal: 2.3s (stat. only)
Limit (CLs method, incl. syst.): B(B+  t+ nt) < 6.7×10-4 (90%CL)
Hadronic tags: event counts
t+  p+nt, p+p0nt, p+p–p+nt, e+nent, m+nmnt
Signal in kin./Eextra regions: events Background: ± 2.1(stat) ± 1.3 (syst) events
Limit (L-ratio w/ null hypothesis): B(B+  t+ nt) < 4.2×10-4 (90%CL)
Combined limit: B(B+  t+ nt) < 4.2×10-4 (90%CL)
BaBar preliminary
t branching fraction
eff’c’y given tag
backgr’nd estimate
Nobs
BB pairs used: (88.9±1.0)×106
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

9. B+  t+ nt analysis II – D*l–n tags
Conference paper: ICHEP: CONF-04/026
B+  t+ nt analysis II – D*l–n tags
Refinement of semileptonic part of analysis I
Tighten tag selection to require clean D*
Consider p+nt, p+p0nt, p+p–p+nt, e+nent, and m+nmnt
Mode-dependent kinematic selections
Determine signals by counting events in Eextra regions
Using (124.1±1.4)×106 BB pairs
Validations
Signal efficiency, Eextra:
Double tags
Backgrounds:
Control samples with extra tracks
Mode
e+nent
m+nmnt
p+nt
p+p0nt
p+p–p+nt
BG incl. systematics
15.2±3.1
8.1±2.1
55.3 ±7.4
29.8 ±5.1
25.1 ±3.4
Observed 13 10 72 30 26
(Signal MC scaled to BF=10-3)
Result:
B(B+  t+ nt) < 4.3×10-4 (90%CL)
Combined with hadronic tags:
(independent, 88.9M BB)
B(B+  t+ nt) < 3.3×10-4 (90%CL)
Also recall (PRL 92, , ’04) BaBar B(B+  m+ nm) < 6.6×10-6
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

10. B+  K/p+ n n analysis Eextra pK hadronic semileptonic
hep-ex/ ICHEP: CONF-04/050
B+  K/p+ n n analysis
Eextra
hadronic semileptonic
Uses tags of both types
Signal obtained from Eextra/pK region
Combined limit:
B(B+  K+nn) < 5.2×10-5 (90%CL)
Also possible to set limit on B+  p+nn
Change PID requirement; use only hadronic analysis: B(B+  p+nn) < 1.0 ×10-4 (90%CL)
BaBar preliminary
Tags
Hadronic
Semileptonic
Background events (incl. systematics)
3.9±1.1
3.4±1.2 (non-peaking)
Signal events 3 6
Efficiency (%)
0.055±0.005
0.115±0.009
Limit (90%CL) on B(B+  K+nn)
6.2×10-5
7.0×10-5
BaBar preliminary pK
Points: data; solid: background MC; dashed: signal MC (arbitrary scale)
BaBar preliminary
BaBar preliminary
BB pairs used: (88.9±1.0)×106
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

11. B0  invisible (n n ), n n g analysis
hep-ex/ accepted by PRL
B0  invisible (n n ), n n g analysis
Semileptonic tags only: B0  D(*)+l–n (D*+  D0 p+)
Require “nothing” in recoil: no charged tracks, limited neutrals
Signal obtained from ML fit to Eextra (KEYS shapes from MC):
n n :
Signal: 17 ± 9, background events
n n g:
Signal: – , background
Upper limits:
Systematics for n n ( n n g ):
Additive: 7.4 ( 4.3 ) events
Multiplicative: 10.9% ( 11.1% )
Frequentist limit-setting procedure
Systematics taken as Gaussian
B(B0  invisible) < × 10-5 (90%CL) B(B0  n n g ) < × 10-5 (90%CL)*
BB pairs used: (88.5±1.0)×106
– 8
n n
n n g
– 1.9
– 5
* Depends on constituent quark model for Dalitz plot shape (Lu & Zhang, Phys. Lett. B 381, 349 (1996))
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

12. B0  l+ l– (e+e–, m+m–, e+m–) analysis overview
Very straightforward mES, DE reconstruction
Extremely clean monochromatic kinematics
Key features: Lepton identification, rejection of QED and qq backgrounds
Bremsstrahlung recovery applied to e’s
Strategy: define cuts for ID and purity, and sideband region, blind signal box in mES, DE
Selection variables
| cos qT |: qT is angle between thrust axes of l+l– candidate and rest of event
mROE: invariant mass of rest of event
R2: norm’d 2nd Fox-Wolfram moment
Ntrk + ½Ng: measure of multiplicity
EEMC < 11 GeV, rejects QED
accept
accept
accept
accept
Signal MC (blue) vs. sideband data
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

13. B0  l+ l– (e+e–, m+m–, e+m–) results
In prep. for PRL: ICHEP: PUB-04/032
B0  l+ l– (e+e–, m+m–, e+m–) results
Unblinded data for (122.5±1.2)×106 BB:
Signal regions
Events observed
e+e–
m+m–
Systematics:
Dominated by multiplicative uncertainties on the efficiency, primarily from tracking and m ID: 5.7%, 7.1%, 6.8% for e+e–, m+m–, e+m–
Limits:
Using the Barlow method:
B(B0  e+e–) < 6.1 × 10-8 (90%CL)
B(B0  m+m–) < 8.3 × 10-8 (90%CL)
B(B0  e+ m–) < 18 × 10-8 (90%CL)
A substantial improvement on previous measurements!
e±m+
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

14. Finding additional details
B+ m+nm
Published in PRL 92, (2004); hep-ex/
B+ t+nt (88.9M BB)
Submitted to PRL: hep-ex/ , BABAR-PUB-04/021
B+ t+nt (124.1M BB, new semileptonic analysis with D*0l–n tags)
ICHEP04 conference paper, abstract # , BABAR-CONF-04/026
B+ K+ n n
ICHEP04 conference paper, abstract # , BABAR-CONF-04/050, hep-ex/
B0 nn, nng
Accepted for publication in PRL: hep-ex/ , BABAR-PUB-04/014; ICHEP abs. #
B0 l+l–
About to be submitted to PRL; BABAR-PUB-04/032; ICHEP abstract #
See also D0 l+l–: hep-ex/ , BABAR-PUB-04/027, submitted to PRL; ICHEP #
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

15. Summary
We have set upper limits on branching fractions of several rare leptonic and/or FCNC decays of B mesons useful as …
checks of basic predictions for weak decays of heavy-quark mesons (e.g., fB), and
probes for new physics beyond the Standard Model
These decays have not yet come into sight but some are now within an order of magnitude or less of the SM expectations…
We have 2-3 times the data available already (>240M BB) and being analyzed, with another factor of two to come in the next few years…
Some of these will very likely be observed one or two ICHEPs from now…
nn g
(absent in SM) nn
p+nn
K+nn
t+n
m+m–
m+n
m+m– (e+e–, e+m– similar)
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

16. Additional slides

17. The BaBar detector and dataset
Asymmetric C.M. for Y(4S)  BB, bg ~ 0.56
Work presented today
B+  t+ nt (D*ln tags)
B0  l+ l–
B+  t+ nt, m+ nm
B+  K+ n n
B0  n n, n n g
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

18. B+  t+ nt analysis I details
For semileptonic tags D0l–nlX, t+ decays e+nent and m+nmnt are considered
Exactly one charged track originating near the interaction point,
Positively identified as an electron or muon, with p* < 1.2 GeV/c.
Continuum quark and t+t– events are rejected by means of event shape cuts and a minimum invariant-mass requirement on triplets of charged tracks.
Require -2.5 < cos qB,Dl < 1.1; admits excited D states with some unobserved particles
Signal determined from fit to shape of distribution of remaining neutral energy (Eextra)
Peaks near zero and at low values (~0.15 GeV) for signal, from X=(nil) and X=g, p0, respectively; rising at larger Eextra for principal backgrounds.
For hadronic tags, we consider p+nt, p+p0nt, p+p–p+nt, e+nent, and m+nmnt
Total charge 0. Positive PID is required for charged tracks.
Extra p0 or KS0 or high energy photons are rejected.
Characteristic kinematical and dynamical features of the t decay modes are required
Minimum missing momentum (calculable using the fully reconstructed tag) req’d per mode
Signal determined by counting events in the signal region for each t decay mode mES > 5.27 GeV/c2, < DE < 0.06 GeV
Background extrapolated from sideband 5.21 < mES < 5.26 GeV/c2, checked against simulation
168k tags
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

19. B+  t+ nt analysis II – D*0l–n tags – details
Refinement of semileptonic analysis, tightening tag selection
Require a D*0, in either of its decays (D0p0,D0g)
Tight kinematic cut rejects other DX: –1.1 < cos qB,D*l < 1.1
Select best D, D*
Tag efficiency: (1.818±0.074)10-3
Consider p+nt, p+p0nt, p+p–p+nt, e+nent, and m+nmnt
Require 1 or 3 charged tracks of good quality & matching PID & a good p0 (Eg > 50 MeV) for the p+p0nt mode (no p0 for p+nt)
Mode-dependent cuts on missing mass, energies of “t” daughters, n-p sub-resonance masses (r, a1)
Eextra from remaining neutrals identifies signal
Signal: Eextra < 300 MeV
Background: 350 < Eextra < 1000 MeV
D0p0en
D0gen
DM (Gev/c2)
Entries/ GeV
Eextra validation from double tags
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

20. B+  t+ nt analysis II – event count breakdown
Expectations and results for (124.1±1.4)×106 BB:
BG incl. systematic unc’y
15.2±3.1
8.1±2.1
55.3 ±7.4
29.8 ±5.1
25.1 ±3.4
Observed 13 10 72 30 26
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

21. B+  K+ n n analysis details
Semileptonic tags
D0 mesons are reconstructed in D0K–p+, K–p+p+p–, K–p+p0
Leptons are identified as e, m
Require -2.5 < cosqB,Dl < 1.1 (accepts D*(*)0); reject tags with extra p± consistent w/D*+
Hadronic tags
D mesons and D0 X– combinations are reconstructed as stated earlier, selecting on mES and DE.
Signal determination
Require exactly one additional charged track, a well-identified kaon. Reject p0’s, >6 neutrals.
Compute Eextra as above
Count events:
Signal: Eextra < 250 MeV, pK* > 1.25 GeV/c
Continuum and combinatoric B background determined from sidebands (of D0 and B);
“Peaking” B background determined from simuation for hadronic analysis, not subtracted in semileptonic
BaBar preliminary
“D0” candidate mass (GeV/c2)
continuum background subtracted
BaBar preliminary
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004

22. B0  invisible (n n ), n n g analysis overview
Semileptonic tags only: B0  D(*)+l–n (D*+  D0 p+) with D+  K–p+p+ and D0K–p+, K–p+p+p–, K–p+p0
Require -2.5 < cosqB,D(*)l < 1.1 (accepts some higher D*(*)’s)
Result: tags, purity 66%
Reconstruct “nothing”
No additional charged tracks
n n: N(“g-like” clusters in EMC) < 3, N(“K0L-like” clusters) < 3
n n g: Exactly 1 EMC cluster with Eg* > 1.2 GeV
Again look at Eextra
Determine signals:
Construct PDFs in Eextra for signal, background
Non-parametric KEYS PDFs used
Perform unbinned ML fit to distribution in data
Validations:
Efficiency checked against double tags
Analyses repeated in “B+  invisible” control sample
n n
n n g
Eextra
Leptonic B Decays - BaBar preliminary
Gregory Dubois-Felsmann – 5 August 2004