Setting the Charm Decay Scale - Texas A&M...
Transcript of Setting the Charm Decay Scale - Texas A&M...
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
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
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
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
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
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
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
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
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
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
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
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
D0/D+ Analysis(PRL 95 121801)
Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 11 / 38
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
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
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
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
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
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
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
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
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
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
Ds Analysis(In Progress)
Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 22 / 38
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
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
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
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
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
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
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
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
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
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
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
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
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
Backup Slides
Peter Onyisi Charm Hadronic BFs Texas A&M, 21 Apr 2006 36 / 38
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
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