Semileptonic B physics at CLEO

Ron PolingUniversity of Minnesota

CLEO CollaborationCLEO-c Collaboration

Ron Poling WIN03 October 7, 2003 2

CESR and CLEO: 23 years of B Physics

Three generations of detectors, ~170 Ph.D.’s, ~360 papers

Ron Poling WIN03 October 7, 2003 3

The Global Program of CKM Tests

• Goal: redundant determinations of sides and angles of the Unitarity Triangle(s)

• CLEO’s Role:– Measure UT sides through |Vcb|

and |Vub| in semileptonic B decays.– Detailed studies to guide

interpretation and help clear the “fog” of nonperturbative effects.

• Challenging experimentally and theoretically.

bW-

q c or uq

e, µ, ν

τVcb or Vub

Ron Poling WIN03 October 7, 2003 4

B Semileptonic Branching Fraction

• CLEO II: 10 fb-1 at ϒ(4S)– Mature, well-understood detector,

data, Monte Carlo, generators, etc.• Lepton-Tagged Analysis

– ptag>1.4 GeV/c plus accompanying electron with pe>0.6 GeV/c.

– Charge, angular correlations to separate primary (B→Xeν) from secondary (B→D →Yeν).

• Refined electron ID, background and efficiency determinations.– Maximize understanding and

minimize momentum dependence.

CLEO - Preliminary

Preliminary: B(B→Xeν)=(10.88 ± 0.08 ± 0.33)%

Ron Poling WIN03 October 7, 2003 5

BSL – Still small-ish, but consistent

• Good agreement among different techniques, experiments.

• Measurements at ϒ(4S) have come up and LEP Z0 average has come down.– Most recent LEP fit

result is (10.59±0.22)%

Ron Poling WIN03 October 7, 2003 6

The Reality of Semileptonic B DecaysThe Reality of Semileptonic B Decays

• We idealize them as…

• The real picture is…

• View of b decay obscured by…– Hadronic effects describable by

both perturbative and nonperturbative QCD.

– Uncertainty about our assumption of quark-hadron duality.

• Available tools include Heavy Quark Effective Theory and Lattice QCD. Must be validated by experiment.

bW-

q c or uq

e, µ, ν

τVcb or Vub

Ron Poling WIN03 October 7, 2003 7

HQET+OPE

• Expresses observables of inclusive b decay as expansions in inverse powers of MB (Voloshin +Shifman, …).

• One parameter at order ΛQCD/MB:– Λ~ MB – mb is the energy of light-quark and gluon degrees of freedom

in the B meson.• Two parameters at order Λ2

QCD/M2B :

– λ1 – negative of the average b-quark momentum squared.– λ2 – energy of the hyperfine interaction of the b spin with the light

degrees of freedom (from B*-B splitting: λ2 = (0.128 ± 0.010) GeV).• Six more parameters at order Λ3

QCD/M3B (Gremm and Kapustin):

– ρ1, ρ2, T1, T2, T3, T4 (not all independent).• Meaning of parameters depends on choices: order of perturbative

(β0(αs/π)2 ) and nonperturbative (1/M3B) expansions and mass

scheme. Consistency required. |Vcb| is what counts!

Ron Poling WIN03 October 7, 2003 8

MOMENTS, INC. – CLEO Branch

• B→Xsγ Photon-Energy Moments• B→Xlν Lepton-Energy Moments• B→Xlν Hadronic Recoil Mass Moments

First and second moments, different lepton-energy ranges…

Lots of complementary information.

Data from different processes will tell us if this approach provides a coherent picture of inclusive B decays.

Ron Poling WIN03 October 7, 2003 9

Measuring B→Xsγ

• Use all available tools!– Traditional shape variables.– Pseudo-reconstruction: demand

consistency with B→K(n)πγ– Use neural net to compute signal

probability and measure spectrum of “weights.”

– Subtract continuum with below-ϒ(4S) data.

– Subtract backgrounds with Monte Carlo, informed by data wherever possible.

Photon-energy spectrum at ϒ(4S)

⇒Continuum must be suppressed.

Ron Poling WIN03 October 7, 2003 10

Measuring B→Xsγ• CLEO II – 10 fb-1

• PRL 87, 251807 (2001)• Hard measurement, simple

interpretation:– Almost two-body, with

broadening by Fermi motion, gluon bremsstrahlung, boost of B’s at ϒ(4S)

– <Eγ> gives nearly direct access to .

Next, measure the moments of the Eγ spectrum by fitting to model of Ali-Greub or HQT calculation of Kagan-Neubert:

Ron Poling WIN03 October 7, 2003 11

Hadronic Mass Moments in B→Xclν

• Select events with leptons with– 1.5 < pl < 2.5 GeV/c

• “ν reconstruction” to measure pν.• Calculate approximate hadronic

recoil mass using lν (take advantage of pB small):

• Construct continuum-subtracted spectrum and fit to mixture:– B→Dlν (CLEO FF’s)– B→D*lν (CLEO FF’s)– High-mass, nonresonant by

models (ISGW2, Goity/Roberts)

• Use fit components to compute moments:

= 0.251±0.023 ±0.062 GeV2

= 0.639±0.056 ±0.178 GeV2

PRL 88, 251808 (2001)

Ron Poling WIN03 October 7, 2003 12

Interpreting the Moments

• First moments <Eγ> and <MX

2- MD2 > fitted to extract

Λ and λ1.• Combined solution is

Λ = 0.35±0.070±0.10 GeV λ1 = -0.236±0.071±0.078 GeV2

• Precise meaning of parameters is scheme- and order-dependent; these are to order 1/MB

3, order β0αs2 in MS

scheme.

Ron Poling WIN03 October 7, 2003 13

Extracting |Vcb| from B(B→Xclν)

• Ingredients:B(B→Xclν) = (10.8±0.3)% (CLEO)τB0 and τB± (PDG), f+−/ f00 (CLEO)

⇒ ΓSL = (0.44 ± 0.02) × 10-10 MeV

• Result:|Vcb| = 0.041 ± 0.0005 ± 0.0007 ± 0.0009

Overall precision: ~3% + quark-hadron duality.

Ron Poling WIN03 October 7, 2003 14

|Vcb| from B→D*lν dΓ/dw

0.0388 ±0.0011

|Vcb|=(42.6±1.2±2.3)×10-3

PRL 88, 251808 (2001)

Ron Poling WIN03 October 7, 2003 15

New and Improved Measurement of theHadronic Mass Moments in B→Xclν

hep-ex/0307081 - Contributed to Lepton-Photon 2003

• Compute recoiling hadronic mass from charged lepton and neutrino kinematics - neutrino “detection.”

• Near hermeticity of CLEO II ⇒Neutrino=“What’s missing”

Fit 3-dimensional differential decay rate, extract hadronic mass squared as a function of lepton-energy cut (pl > 1 GeV/c).

Ron Poling WIN03 October 7, 2003 16

Fit Projections• Selection criteria:

– Cuts to enhance ν reconstruction– Continuum suppression– Efficiency ~2% for B→Xclν

• Sample to fit: 122K events• Components of fit:

– B→Dlν– B→D*lν– B→D**lν ISGW2

– B→(Xc)NRlν Goity/Roberts– B→Xulν ISGW2+NR– Secondaries CLEO MC– Fake Leptons, Continuum fixed

with data

}HQET+measured FFs

Ron Poling WIN03 October 7, 2003 17

Fits →Mode-by-Mode BRs

+Generator-Level Info

(Fraction above pl cut, moment value for that cut.)

⇓Moment Value for B→Xclν

Results

• Consistent with previous CLEO measurements, BaBar summer ’03

• Interpretation continues

PRELIMINARY

DELPHI: energy and hadronic mass moments for full lepton-energy spectrum. hep-ph/0210319 PRELIMINARY

Ron Poling WIN03 October 7, 2003 18

B→Xclν Energy Moments PRD 67, 072001 (2003)

Pending…El moments down to 0.6 GeV

eµ

Ron Poling WIN03 October 7, 2003 19

Improved Inclusive Measurement of |Vub|

• End-point technique:– Isolate B→Xulν by looking above

(and just below) the B→Xclνkinematic limit.

– Used for b→u discovery (CLEO, ARGUS) and CLEO 93: |Vub|/|Vcb|= 0.08±0.02Significant model dependence.

• Can now do better:– Better knowledge of B→Xclν.– B→Xsγ helps determine fraction

of spectrum in end point. B→Xsγ spectrum ⇒ shape function ⇒ fu

Ron Poling WIN03 October 7, 2003 20

Measuring B→Xulν

2.0 < pl <2.6 GeV/cSuppress continuum with neural net (energy flow w.r.t. lepton). Less q2- and M(Xu)-dependent.Estimate B→Xclν by fitting inclusive spectrum. Components reflect state-of-the-art on form factors, b.r.’s, etc.Trade-off: B→Xclν uncertainty vs. spectral-fraction uncertainty.

b→ulν excess

ON

OFF

b →clν

PRL 88, 231803 (2002)

Ron Poling WIN03 October 7, 2003 21

• Hoang, Ligeti, Manohar; Uraltsev:

• Determine from B→ Xsγ shape-function fits (Kagan-Neubert).

Extracting |Vub|

In preparation: Paper with details of fits for shape-function parameters and fuvalues for phase-space regions other than just end point.

Ron Poling WIN03 October 7, 2003 22

Branching Fractions and q2 Dependenceof B→πlν and B→ρ(ω)l ν

and Extraction of |Vub|

• Exclusive reconstruction of B→(π/η/ρ/ω)l ν through neutrino “detection.”

• Near hermeticity of CLEO II ⇒ Neutrino=“What’s missing”

hep-ex/0304019 to be published

in PRD

Signal defined by…

Ron Poling WIN03 October 7, 2003 23

B→πlν (pl > 1.0 GeV/c) B→ρlν (pl > 1.5 GeV/c)

• CLEO II 10 fb-1

• Event selection to enhance ν reconstruction, suppress continuum

• Fit Mmlν and ∆E in three bins in q2 to find branching fractions

Ron Poling WIN03 October 7, 2003 24

• Extracting |Vub|

– Fit dΓ/dq2

– Discriminates among FFs• B→πlν

– FF dependence is small– Disfavors ISGW2

• B→ρlν– Larger FF

dependence⇒greater model uncertainty in |Vub|

B→πlν

B→ρlν

Ron Poling WIN03 October 7, 2003 25

RESULTS

Ron Poling WIN03 October 7, 2003 26

Comparisons and CKM Impact

|Vub|/10-3

0 2 4 6

CLEO lepton endpoint

BABAR endpoint

LEP M

BABAR M

BELLE D*�ν (prelim)

CLEO exclusives

BABAR ρ�ν

BELLE M & q (prelim)

x

x

x 2

Mean with 25% error

V x 10ub-3

-1

-0.5

0

0.5

1

-1 -0.5 0 0.5 1 1.5 2

∆md

∆ms & ∆md

εK

εK

|Vub/Vcb|

ρ

ηCK M

f i t t e rpackage

βγ

α

90% c. l. contours

Ron Poling WIN03 October 7, 2003 27

• CLEO is still contributing on |Vub| and |Vcb|.– Sample smaller than B factories, but well understood with

mature tools.– We’re all limited by theoretical uncertainties.

• Continuing and forthcoming CLEO CKM-related measurements:– Completion of moment measurements, coordinated fits of all

available inputs.– Inclusive b→ulν for all CLEO II and CLEO III via

multidimensional fit and bigger phase-space acceptance.• CESR-c/CLEO-c will have significant impact on CKM.

See the upcoming of Karl Ecklund.

Perspectives and ProspectsPerspectives and Prospects