Setting the Charm Decay Scale - Texas A&M...

40
Setting the Charm Decay Scale e + e - D * s D s D + s D - s γ π + K + K - π - K + K - γ Peter Onyisi Cornell University CLEO Collaboration Texas A&M, 21 Apr 2006 Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 1 / 38

Transcript of Setting the Charm Decay Scale - Texas A&M...

Page 1: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Setting the Charm Decay Scale

e+e− → D∗s Ds → D+s D−

s γ

π+

K+K−

π−

K+

K−

γ

Peter Onyisi

Cornell UniversityCLEO Collaboration

Texas A&M, 21 Apr 2006

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 1 / 38

Page 2: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Outline

I Intro to CLEO-c

I Open charm physics at CLEO-cI Hadronic branching fraction analyses

I D0/D+

I Ds

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 2 / 38

Page 3: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

CESR

I CESR is a 768 m circumferencee+e− storage ring

I Provides collisions for CLEOand beams for the Cornell HighEnergy Synchrotron Source

I Originally designed to operateat Ecm = 9–12 GeV, ran mostlyat Υ resonances with (at thetime) world-record luminosities(1.25 × 1033)

I Upgraded to provide collisionsdown to Ecm = 3 GeV

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 3 / 38

Page 4: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

CESR-c Upgrades

0140902-005Beam

Trajectory

BeamTrajectory

I Old CESR could not cool the beams atlow energy enough for good luminosity(synchrotron power ∝ E 4)

I CLEO can no longer runsimultaneously with CHESS

I Solution: 12 wiggler magnets installed

I “Anti-solenoids” cancelenergy-dependent effects of CLEOfield on beam

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 4 / 38

Page 5: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

CLEO-c Upgrade

Solenoid Coil Barrel Calorimeter

Drift Chamber

Inner Drift Chamber /Beampipe

EndcapCalorimeter

IronPolepiece

Barrel MuonChambers

MagnetIron

Rare EarthQuadrupole

SC Quadrupoles

SC QuadrupolePylon

2230104-002

CLEO-c

Ring Imaging CherenkovDetector

I CLEO-III silicon vertex detectorreplaced with (all stereo) driftchamber (typical z0-resolution 700µm)

I Solenoid magnetic field changedfrom 1.5T to 1.0T to compensatefor lower-momentum tracks

I DAQ, trigger, software, etc. fromCLEO-III with only minor changes

I Particle ID (from dE/dx ,Cerenkov) better due to lower ptracks

I Muon system now only useful forhigh momentum (e.g.J/ψ→ µ+µ−),

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 5 / 38

Page 6: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

CLEO-c Missions

I Enable measurements at other experimentsI Find branching fractions of reference modesI Test theoretical input to B-factoriesI Measure strong phases in D system

I Explore QCD dynamics in the charm region and belowI CharmoniumI D Dalitz studiesI Light mesons in radiative decays

I Search for new physics

I D0–D0

mixingI CP violationI Rare decays

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 6 / 38

Page 7: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

CLEO-c Missions

I Enable measurements at other experimentsI Find branching fractions of reference modesI Test theoretical input to B-factoriesI Measure strong phases in D system

I Explore QCD dynamics in the charm region and belowI CharmoniumI D Dalitz studiesI Light mesons in radiative decays

I Search for new physics

I D0–D0

mixingI CP violationI Rare decays

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 6 / 38

Page 8: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

CLEO-c Missions

I Enable measurements at other experimentsI Find branching fractions of reference modesI Test theoretical input to B-factoriesI Measure strong phases in D system

I Explore QCD dynamics in the charm region and belowI CharmoniumI D Dalitz studiesI Light mesons in radiative decays

I Search for new physics

I D0–D0

mixingI CP violationI Rare decays

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 6 / 38

Page 9: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Open Charm Decays In A Nutshell

c

d ν

W+

D+

ℓ+

Leptonic decays:

I Probe wavefunction at origin (decay constants fD ,

fDs )

I Test lattice QCD and meson structure

models

s

u

ℓ+

νW

+

u

c

K0

D0

Semileptonic decays:

I Probe overlap of initial and final hadron states

(form factors f+(q2) . . . )

I Test LQCD and decay models

s

s

d

u

s

c

W+

K0

K+

D+s

Hadronic decays:

I Reference modes normalize decays

I QCD dynamics

I Search for new phenomena: CP violation, D0–D0

mixing

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 7 / 38

Page 10: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Absolute Charm Branching Fractions

What is the problem?

I Measurements of decays to c quarks depend on reconstructingdecays of light charmed mesons and baryons

I Branching fraction measurements can be limiting systematics

I Since b → c is a dominant decay mode, B measurementsoften rely on knowing various D(s) BFs

I Affects precision measurements of Z → cc (H → cc . . . )

I Reference modes (D0 → K−π+, D+ → K−π+π+, D+s → φπ+)

normalize virtually all other branching fractions

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 8 / 38

Page 11: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

General Method

(Follows pioneering analysis by Mark III. . . )

I Exploit low-energy production processes:

I At 3.77 GeV, open charm only produced as D0D0

and D+D−

I At 4.17 GeV, Ds produced almost entirely as D∗s Ds

I In the events we use, a meson of interest is always producedalong with its antiparticle

I We choose a set of modes and reconstruct single tags (wereconstruct a decay) and double tags (we reconstruct bothdecays)

I A double tag can count as single tagsI For N decay modes we have 2N single tags (separated by

charge) and N2 double tags

I Since the single tag yield in mode i is proportional to Bi , andthe double tag yield for i , j is proportional to BiBj , we candetermine each of the branching fractions

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 9 / 38

Page 12: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Method Numerics

Single tag yields: Ni = NDDBiεi

Double tag yields: Nij = NDDBiBjεij

⇒ Branching fractions: Bj =Nij

Ni

εiεij

I In practice, we fit all the yields simultaneouslyI Maximizes power: limiting statistical uncertainty is√

total double tags in every modeI Bad χ2 → something wrong. . .

I Can correlate systematics

I Obtain cross-section as well

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 10 / 38

Page 13: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

D0/D+ Analysis(PRL 95 121801)

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 11 / 38

Page 14: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Data Used

I L = (55.8± 0.6) pb−1 at Ecm ≈ 3.773 GeV, at the peak of theψ(3770) resonance

I CLEO-c dataset as of end of 2004

I We are updating to summer 2005 (281 pb−1)

I Also exploit ≈ 3 pb−1 of CLEO-c ψ ′ data for systematics studies

DD threshold

operating point

Ecm

σ(hadrons)σ(µ+µ−)

PDG

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 12 / 38

Page 15: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

D Hadronic Decay Overview

I Reference decay modes are D0→ K−π+ and D+→ K−π+π+

I 18 single tag, 45 double tag modes

Decay PDG 2004 fit Rel uncert

D0→ K−π+ 3.80% 2.4%

D0→ K−π+π0 13.0% 6.2%

D0→ K−π+π+π− 7.46% 4.2%

D+→ K−π+π+ 9.2% 6.5%

D+→ K−π+π+π0 6.5% 17%

D+→ KS π+ 1.41% 6.7%

D+→ KS π+π0 4.85% 31%

D+→ KS π+π+π− 3.55% 14%

D+→ K−K+π+ 0.89% 9.0%

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 13 / 38

Page 16: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

The CLEO-c D-Hunter’s Guide

I Charged K , π distinguished using dE/dx (all momenta) andCerenkov (for high momentum)

I Find π0’s by combining pairs of isolated showers in theCsI calorimeter, requiring 3σ consistency with π0 mass (σ ∼ 6MeV)

I Find KS ’s by combining pairs of tracks that lie within a masswindow

I Two crucial kinematic variables:I “Beam-constrained mass” MBC =

√E 2

beam − ~p2D tests that

total momentum of candidate is rightI ∆E = ED − Ebeam tests particle ID, is sensitive to missing

particles

“D Tagging will solve the world’s problems and make it sunny in Ithaca” - Anon.

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 14 / 38

Page 17: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Yield extraction

10

210

310

410

10

210

310

410+π- K→0D +π+π- K→+D +π0

S K→+D

210

310

410

210

310

4100π+π- K→0D 0π+π+π- K→+D 0π+π0

S K→+D

1.84 1.86 1.88

210

310

410

1.84 1.86 1.88

210

310

410+π-π+π- K→0D

1.84 1.86 1.881.84 1.86 1.88

-π+π+π0S K→+D

1.84 1.86 1.881.84 1.86 1.88

+π+K- K→+D

)2c (GeV/M

)2 cE

ven

ts /

(0.0

02 G

eV/

DATA: Single tags

1.84 1.86 1.880

100

200

300

400

500

600

700

800

1.84 1.86 1.880

100

200

300

400

500

600

700

800 0 D0D

)2c (GeV/M

)2 cE

ven

ts /

(0.0

012

GeV

/

1.84 1.86 1.880

100

200

300

400

500

600

1.84 1.86 1.880

100

200

300

400

500

600-D+D

)2c (GeV/M

)2 cE

ven

ts /

(0.0

012

GeV

/

DATA: Double tag projections

I Fit signal with a priori function of

physical parameters (detector momentum

resolution, beam energy spread, ψ(3770)

lineshape, ISR spectrum)

I Smooth backgrounds fit as combinatoric

phase space (“ARGUS function”)

I Peaking backgrounds estimated from

known BFs and subtracted

I In double tags, fit 2D plane of MBC (1)

vs. MBC (2)

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 15 / 38

Page 18: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Systematics studies using ψ ′

I Clean decays ψ ′ → J/ψπ+π−

and J/ψπ0π0 used to comparetracking and π0 efficiencies inMC and data

I Reconstruct J/ψ and one pion;compute recoil mass: peaks atpion mass

I Find fraction of such events withother pion reconstructed

I Right: Plots for J/ψπ+π−,0.15 < cos θπ < 0.55

I ε = (95.89± 0.20)%; agreeswith MC within statistics

M2miss (GeV)-0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1

Eve

nts

/ ( 0

.001

5 G

eV )

0

100

200

300

400

500

psiprime_data_new_costheta_p_psiprime.evt

22±NBkgdExp = 98 98±Npeak = 9230 0.0033±fract = 0.5000

0.00022±mpisq1 = 0.01744 0.00032±mpisq2 = 0.02214 0.00015±sigma1 = 0.00836 0.00025±sigma2 = 0.01612

Fit parameters

M2miss (GeV)-0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1

Eve

nts

/ ( 0

.001

5 G

eV )

0

100

200

300

400

500

psiprime_data_new_costheta_p_psiprime.evt

DATA2nd πfound

M2miss (GeV)-0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1

Eve

nts

/ ( 0

.001

5 G

eV )

0

5

10

15

20

25

30

35

40

psiprime_data_new_costheta_p_nopsiprime.evt

30±NBkgd = 457

12±NBkgdExp = -100.0 24±Npeak = 396

1.1±p2 = 14.8

Fit parameters

M2miss (GeV)-0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1

Eve

nts

/ ( 0

.001

5 G

eV )

0

5

10

15

20

25

30

35

40

psiprime_data_new_costheta_p_nopsiprime.evt

DATA2nd π

not found

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 16 / 38

Page 19: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Double DCSD: Quantum Effects

Consider double tags with D0 → Kπ:

I Dominant decay is D0 → K−π+ (D0 → K+π−). However there are

doubly Cabibbo-suppressed decays D0 → K+π− and vice versa.

I The ratio of amplitudes is the complex number A, which encodesboth the suppression and a relative strong interaction phase δ.

I The direct contribution of the double DCSD amplitude in doubletags is negligible.

I But you have interference between D0D0

going Cabibbo-allowed toboth modes and double DCSD:

I = |1 + A2|2 = 1 + 2|A|2 cos 2δ+ O(|A|4)

I |A|2 ∼ 0.4%; we don’t know cos 2δ: 0.8% uncertainty on D0 doubletag rates!

(We expect cosδ ∼ 1; we’re doing analyses to test this. . . )

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 17 / 38

Page 20: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Systematic uncertainties

Source Fractional uncertainty (%)

Tracking/KS /π0 0.7/3.0/2.0 per particleParticle ID 0.3 per π, 1.3 per K

Trigger efficiency < 0.2∆E cut 1.0–2.5 per D

FSR modeling 0.5 per single tagψ ′′ width 0.6

Resonant substructure 0.4–1.5Event environment 0.0–1.3Yield fit functions 0.5

Misc. event selection 0.3Double DCSD interference 0.8 in neutral double tags

D0→K−π+ uncert 2.3%, D+→K−π+π+ uncert 2.8%For these, largest contributors are kaon PID and ∆E cut

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 18 / 38

Page 21: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Results

Branching fractions ...

Mode Value

B(D0 → K−π+) (3.91± 0.08± 0.09)%

B(D0 → K−π+π0) (14.9± 0.3± 0.5)%

B(D0 → K−π+π+π−) (8.3± 0.2± 0.3)%

B(D+ → K−π+π+) (9.5± 0.2± 0.3)%

B(D+ → K−π+π+π0) (6.0± 0.2± 0.2)%

B(D+ → KS π+) (1.55± 0.05± 0.06)%

B(D+ → KS π+π0) (7.2± 0.2± 0.4)%

B(D+ → KS π+π+π−) (3.2± 0.1± 0.2)%

B(D+ → K+K−π+) (0.97± 0.04± 0.04)%

σD+D− (nb) σD0D

0 (nb) σDD (nb) σD+D−/σD0D

0

2.79± 0.07+0.10−0.04 3.60± 0.07+0.07

−0.05 6.39± 0.10+0.17−0.08 0.776± 0.024+0.014

−0.006

... and cross sections from 55.8 pb−1

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 19 / 38

Page 22: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Result comparison

0.4 0.6 0.8 1.0 1.2 1.4 1.6

PDGCLEO-c

B(CLEO-c)B(PDG)

0.4 0.6 0.8 1.0 1.2 1.4 1.6

PDGCLEO-c

Br. Ratio(CLEO-c)Br. Ratio(PDG)

0.010 0.020 0.030 0.040 0.050 0.060

PDG

B(D0 → K−π+)

previous absolute

measurements and

PDG fit

0.045 0.065 0.085 0.105

PDG

B(D+ →K−π+π+)

previous absolute

measurements and

PDG fit

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 20 / 38

Page 23: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

D0/D+ Summary and Outlook

I Branching fractions from 56 pb−1 have precision comparableto world averages

I Updating to 281 pb−1: we will be systematics-limited

I Aiming for < 1.5% uncertainty on reference modes

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 21 / 38

Page 24: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Ds Analysis(In Progress)

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 22 / 38

Page 25: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Ds Hadronic Decay Overview

I The classic reference decay hasbeen the exclusive modeD+

s → φπ+ → K−K+π+

I Essentially all otherdecays have branchingratios to this mode

I This causes problems since φsignal is ambiguous given theprecision we will soon achieve

I We instead measureinclusive branching fractions.

Modes used

Decay PDG 2004 BF (%)

D+s → KS K+ 1.8± 0.55

D+s → K−K+π+ 4.3± 1.2

D+s → K−K+π+π0 —

D+s → π+π+π− 1.00± 0.28

Relative uncertainties are roughly 25–30%,

limited by the φπ+ BF in PDG 2004.

BaBar has a 2005 φπ+ measurement,

34% higher than the PDG, with 13%

errors.

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 23 / 38

Page 26: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

The φπ+ problem

I Expect (f0(980) → K−K+)π+ to contributeto any φ mass region, with badly controlledparameters

I Correction might be on the order of 5% ormore — but depends on experiment’s masswindow, resolution, angular distributionrequirements!

Looking at low-mass KK pairs(m(KK ) < 1.005 GeV) we see evidence forscalar production by looking at helicity angle

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts /

( 0.0

0325

GeV

)

0

20

40

60

80

100

120

140

160

180

+π φ → +s, Dinvm

bkg = 177.21 +/- (-15.0, 15.8)

bkgbar = 156.36 +/- (-14.0, 14.8)

c1 = -0.217 +/- (-0.1, 0.1)

yield = 588.76 +/- (-25.1, 25.8)

yieldbar = 575.62 +/- (-24.6, 25.4)

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts /

( 0.0

0325

GeV

)

0

20

40

60

80

100

120

140

160

180

+π φ → +s, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts /

( 0.0

0325

GeV

)

0

20

40

60

80

100

120

140

160

180

-π φ → -s

, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts /

( 0.0

0325

GeV

)

0

20

40

60

80

100

120

140

160

180

-π φ → -s

, Dinvm

khlangphiEntries 2384Mean 1.005e-16RMS 0.7427

-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 10

100

200

300

400

500

khlangphiEntries 2384Mean 1.005e-16RMS 0.7427

K helicity angle, phi band

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts /

( 0.0

0325

GeV

)

0

2

4

6

8

10

12

14

16

18

SIDEBAND+π φ → +s, Dinvm

bkg = 50.73 +/- (-7.5, 8.3)

bkgbar = 38.88 +/- (-6.6, 7.4)

c1 = -0.154 +/- (-0.2, 0.2)

yield = 36.24 +/- (-6.5, 7.2)

yieldbar = 47.11 +/- (-7.2, 7.9)

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts /

( 0.0

0325

GeV

)

0

2

4

6

8

10

12

14

16

18

SIDEBAND+π φ → +s, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts /

( 0.0

0325

GeV

)

0

2

4

6

8

10

12

14

16

18

SIDEBAND-π φ → -s

, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts /

( 0.0

0325

GeV

)

0

2

4

6

8

10

12

14

16

18

SIDEBAND-π φ → -s

, Dinvm

-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 102

4

6

8

1012

14

16

1820

22

K helicity angle, low KK mass

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 24 / 38

Page 27: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Dataset and Landscape

I Use 76 pb−1 of data collected at Ecm ∼ 4170 MeV as part ofthe recent (Aug.–Jan.) Ds energy scan and data run

I Chose running point for maximal Ds production

I Dominant Ds production channel in this region is D∗s Ds , ≈ 1nb

I D0, D+ events produced at ≈ 7 nb, are the major background

D∗s Ds threshold

operating point

Ecm

PDG

CLEO-c ScanPreliminary

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 25 / 38

Page 28: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Production Channel

I We use events with the topologye+e− → D∗±s D∓s → D+

s D−s (γ,π0).

I We do not reconstruct the γ or π0.

I We use the momentum of the Ds candidates to select forevents with an intermediate D∗s . (The quantity

mBC =√

E 2beam − ~p2

D is a proxy for momentum.)

I We can use a loose cut to include the daughters of D∗s , or atight cut for the directly produced Ds

1.94 1.96 1.98 2 2.02 2.04 2.06 2.08 2.10

200

400

600

800

1000

1200

1400

1600

+ KSK→+s

, DBCm

1.94 1.96 1.98 2 2.02 2.04 2.06 2.08 2.10

200

400

600

800

1000

1200

1400

1600

+ KSK→+s

, DBCm

MC

Direct Ds

Ds from D∗s

1.94 1.96 1.98 2 2.02 2.04 2.06 2.08 2.10

20

40

60

80

100

120

140

160

180

+ KSK→+s

, DBCm mbcplts_400Entries 2178Mean 2.02RMS 0.03742

+ KSK→+s

, DBCm

Data

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 26 / 38

Page 29: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Kinematic Separation

(GeV)BCm1.85 1.9 1.95 2 2.05

(G

eV)

inv

m

1.8

1.85

1.9

1.95

2

2.05

2.1

candidates+π + K- for KBC vs. minvm MC

√E 2

beam − ~p2cand

DD D∗D D∗D∗

DsDs D∗s Ds

D∗ refl

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 27 / 38

Page 30: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

D∗s Ds Wrinkles

I We do a binned maximum likelihood fit for all the observedyields (utilizing Poisson statistics for double tags)

I Maximizing statistical power important given relatively low Ds

cross-sectionI The KKπ mode is 62% of all single tags, but KKπ/KKπ is

only 26% of the double tag yield

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 28 / 38

Page 31: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Backgrounds and Reflections

I Crossfeed between Ds modes from KS ↔ π+π− isCabibbo-suppressed

I Use vetoes and sidebands

I Peaking structure can arise from reflections: for example, thedecay chain D∗+ → D0π+ → K−K+π+ has correct mBC ,peaks at 2.06 GeV.

I We veto certain mass regions; e.g. for KKπ we reject eventsconsistent with D0 → KK .

I This doesn’t affect signal but makes the background easier tomodel

kkEntries 20094Mean 1.403RMS 0.2443

0.8 1 1.2 1.4 1.6 1.8 20

50

100

150

200

250

300

350

400

450

kkEntries 20094Mean 1.403RMS 0.2443

candidates+π + K- mass, K+K-K

Data KKπ

D0 → K−K+

kkpiEntries 34331Mean 1.665RMS 0.1476

1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 20

50

100

150

200

250

kkpiEntries 34331Mean 1.665RMS 0.1476

candidates0π +π+K- mass in K+π+K-K

Data KKππ0

D+ → KKπ

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 29 / 38

Page 32: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Yield extraction

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ (

0.00

325

GeV

)

0

100

200

300

400

500

600

+π + K-

K→+s, Dinvm

bkg = 6406.98 +/- (-87.0, 87.7)

bkgbar = 6319.17 +/- (-86.6, 87.3)

c1 = -0.098 +/- (-0.0, 0.0)

yield = 1607.02 +/- (-52.9, 53.6)

yieldbar = 1736.83 +/- (-54.1, 54.9)

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ (

0.00

325

GeV

)

0

100

200

300

400

500

600

+π + K-

K→+s, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ (

0.00

325

GeV

)

0

100

200

300

400

500

600

-π - K+K→-s

, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ (

0.00

325

GeV

)

0

100

200

300

400

500

600

-π - K+K→-s

, Dinvm

DATA: KKπ Single Tags

D+s D−

s

)2) (GeV/c-

sm(D

1.88 1.9 1.92 1.94 1.96 1.98 2 2.02 2.04 2.06

)2)

(GeV

/c+ s

m(D

1.88

1.9

1.92

1.94

1.96

1.98

2

2.02

2.04

2.06)-π - K+K→-

s / D+π + K-K→+s

(D-sD vs. m+

sDm

DATA: KKπ+/KKπ− Double Tag

I Fit single tag signals with doubleGaussian or Crystal Ball function(parameters fixed from MonteCarlo) plus a linear background

I Each charge done separately

I In double tags, count events insignal and sideband boxes

I Combinatoric background isflat in m(D+

s ) − m(D−s ), has

structure in m(D+s ) + m(D−

s )

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 30 / 38

Page 33: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Data Results

) (GeV)s

m(D1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ 3

MeV

0

20

40

60

80

100

120

140

160

180

200

+ KSK→+s

), Ds

m(D

) (GeV)s

m(D1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ 3

MeV

0

20

40

60

80

100

120

140

160

180

200

+ KSK→+s

), Ds

m(D 788± 34

) (GeV)s

m(D1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ 3

MeV

0

200

400

600

800

1000

1200

+π + K-K→+s

), Ds

m(D

) (GeV)s

m(D1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ 3

MeV

0

200

400

600

800

1000

1200

+π + K-K→+s

), Ds

m(D 3344± 77

) (GeV)s

m(D1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ 3

MeV

0

50

100

150

200

250

300

0π +π + K-K→+s

), Ds

m(D

) (GeV)s

m(D1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ 3

MeV

0

50

100

150

200

250

300

0π +π + K-K→+s

), Ds

m(D 709± 54

) (GeV)s

m(D1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ 3

MeV

0

50

100

150

200

250

-π +π +π→+s

), Ds

m(D

) (GeV)s

m(D1.92 1.94 1.96 1.98 2 2.02

Eve

nts

/ 3

MeV

0

50

100

150

200

250

-π +π +π→+s

), Ds

m(D 539± 41)2) (GeV/c-

sm(D

1.88 1.9 1.92 1.94 1.96 1.98 2 2.02 2.04 2.06

)2)

(GeV

/c+ s

m(D

1.88

1.9

1.92

1.94

1.96

1.98

2

2.02

2.04

2.06

All double tagsAll double tags

) (GeV)-

s) - m(D+

sm(D

-0.1 -0.05 0 0.05 0.1

Eve

nts

/ 8

MeV

0

10

20

30

40

50

60

All double tagsAll double tags

136 signal28 sideband

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 31 / 38

Page 34: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Systematic uncertainties

Source Fractional uncertainty (%)

Tracking/KS /π0 0.35/1.1/5.0 per particleParticle ID 0.3–1.4 correlated by decay

Resonant substructure 0–6.0 correlated by decayFit procedure 3.5 in fit result

Event environment 3.5 in KKππ0

Initial state radiation correction 0–5 per single tagB(D∗+s → π0D+

s ) 0.7 in KKππ0, πππ

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 32 / 38

Page 35: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Preliminary Results for Ds

Mode CLEO-c (%) PDG 2004 fit (%)

B(KS K+) 1.28+0.13−0.12 ± 0.07 1.8± 0.55

B(K−K+π+) 4.54+0.44−0.42 ± 0.25 4.3± 1.2

B(K−K+π+π0) 4.83+0.49−0.47 ± 0.46 —

B(π+π+π−) 1.02+0.11−0.10 ± 0.05 1.00± 0.28

Branching Fraction (%)0 1 2 3 4 5 6 7

-π +π +π

0π +π - K+K

+π - K+K

+ KSK

-π +π +π

0π +π - K+K

+π - K+K

+ KSK

PDG 2004 fit-1CLEO Preliminary, 76 pb

BF/PDG 2004 fit0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4

-π +π +π

+π - K+K

+ KSK

-π +π +π

+π - K+K

+ KSK

PDG 2004 fitPDG 2004 fit, BR error only

-1CLEO Preliminary, 76 pb

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 33 / 38

Page 36: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Comparison with BaBar φπ+

Can we compare with the BaBar B(D+s → φπ+) result?

I We can use the PDG fit branching ratios...

Branching Fraction (%)0 1 2 3 4 5 6 7

-π +π +π

0π +π - K+K

+π - K+K

+ KSK

-π +π +π

0π +π - K+K

+π - K+K

+ KSK

PDG 2004 fit-1CLEO Preliminary, 76 pb

PDG 04: 3.6%

Branching Fraction (%)0 1 2 3 4 5 6 7

-π +π +π

0π +π - K+K

+π - K+K

+ KSK

-π +π +π

0π +π - K+K

+π - K+K

+ KSK

+πφPDG 2004 BRs + BaBar -1CLEO Preliminary, 76 pb

BaBar 05: 4.81%

I We are more consistent with 3.6% than 4.8%

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 34 / 38

Page 37: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Ds Summary and Outlook

I We have preliminary absolute branching fractions for four Ds

decay modes from 76 pb−1 of dataI Precision about 11% for all-charged modesI Inclusive K−K+π+π0 is a first measurement

I The measured BFs are consistent with the PDG 2004 fit

I We are actively working on adding more modes (especiallydecays with η, η ′)

I We are aiming for < 4% uncertainties with full CLEO-cdataset

I Already have more than 100 pb−1 additional data on tape!

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 35 / 38

Page 38: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Backup Slides

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 36 / 38

Page 39: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Backgrounds

I Non-peaking backgrounds removed in the yield fitI Peaking backgrounds are from crossfeed between modes we

consider, and contamination from other modesI Latter dominated by Cabibbo-suppressed decays in KS modes,

e.g. prompt D+→ 5π fakes D+→KS 3π; in some modes up to3% correction

I Estimate backgrounds to single and double tags with PDGbranching fractions and efficiencies from MC, subtract frommeasured yields

Beam constrained mass (GeV)1.83 1.84 1.85 1.86 1.87 1.88 1.89

Eve

nts

/ ( 0

.000

6 G

eV )

0

5

10

15

20

25

30

35

40

45

Fake type kpi

15±yield = 228

Beam constrained mass (GeV)1.83 1.84 1.85 1.86 1.87 1.88 1.89

Eve

nts

/ ( 0

.000

6 G

eV )

0

5

10

15

20

25

30

35

40

45

Fake type kpi

MC30x data DCSD decay D

0→ K−π+ faking D0→K−π+ in

30x MC sample. In data, contributes ≈ 0.15% of

observed peak.

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 37 / 38

Page 40: Setting the Charm Decay Scale - Texas A&M Universitypeople.physics.tamu.edu/kamon/research/talk/ppc_lunch/...Setting the Charm Decay Scale e+e− → D∗ s D s → D+sD s −γ π+

Resonant Substructure

I Our Monte Carlo has some reasonablemixture of intermediate resonances

I Our efficiencies depend on theintermediate state

I We reweight the expected efficienciesby comparing data yields with MCexpectations

I Size of correction is largestsystematic for K−K+π+π0

I The correction for a given modeaffects that mode’s BF only

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

20

40

60

80

100

120

140

160

180

+π φ → +s, Dinvm

bkg = 177.21 +/- (-15.0, 15.8)

bkgbar = 156.36 +/- (-14.0, 14.8)

c1 = -0.217 +/- (-0.1, 0.1)

yield = 588.76 +/- (-25.1, 25.8)

yieldbar = 575.62 +/- (-24.6, 25.4)

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

20

40

60

80

100

120

140

160

180

+π φ → +s, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

20

40

60

80

100

120

140

160

180

-π φ → -s

, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

20

40

60

80

100

120

140

160

180

-π φ → -s

, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

20

40

60

80

100

120

140

160

180

200

+ K*0

K → +s, Dinvm

bkg = 1038.45 +/- (-35.2, 35.9)

bkgbar = 1033.49 +/- (-35.2, 36.0)

c1 = -0.175 +/- (-0.0, 0.0)

yield = 524.55 +/- (-27.0, 27.7)

yieldbar = 608.52 +/- (-28.7, 29.4)

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

20

40

60

80

100

120

140

160

180

200

+ K*0

K → +s, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

20

40

60

80

100

120

140

160

180

200

- K*0 K→ -s

, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

20

40

60

80

100

120

140

160

180

200

- K*0 K→ -s

, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

5

10

15

20

25

30

35

40

0π +π φ → +s, Dinvm

bkg = 285.40 +/- (-19.6, 20.3)

bkgbar = 276.75 +/- (-19.2, 20.1)

c1 = -0.101 +/- (-0.1, 0.1)

yield = 117.62 +/- (-14.8, 15.5)

yieldbar = 131.20 +/- (-15.2, 15.9)

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

5

10

15

20

25

30

35

40

0π +π φ → +s, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

5

10

15

20

25

30

35

40

0π -π φ → -s

, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

5

10

15

20

25

30

35

40

0π -π φ → -s

, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

10

20

30

40

50

0π + K*0

K → +s, Dinvm

bkg = 827.61 +/- (-32.8, 33.7)

bkgbar = 864.18 +/- (-33.4, 34.2)

c1 = -0.125 +/- (-0.0, 0.0)

yield = 111.42 +/- (-19.4, 20.0)

yieldbar = 141.03 +/- (-20.3, 20.6)

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

10

20

30

40

50

0π + K*0

K → +s, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

10

20

30

40

50

0π - K*0 K→ -s

, Dinvm

Invariant mass (GeV)1.9 1.92 1.94 1.96 1.98 2 2.02

Even

ts / (

0.003

25 G

eV )

0

10

20

30

40

50

0π - K*0 K→ -s

, Dinvm

Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 38 / 38