Table Of Contentc
6 Upsilon and analyses at CLEO
b
0
0
2
n
a
J
0
2
JeanDuboscq∗for the CLEOCollaboration†
1
v CornellUniversity,USA
6 E-mail: [email protected]
3
0
1
I detail recentwork done by the CLEO collaborationusing data gatheredat the CESR acceler-
0
6 ator concerningthe ¡ system. Results includeB productionat the ¡ (5S), decaysof the ¡ (nS)
S
0 (n=1,2,3)tot pairs,the¡ di-electronwidth,decaysofthe¡ totwohadronsandaphoton,andthe
/
x observationofthefirstdi-piontransitionbetweenc states
e B
-
p
e
h
:
v
i
X
r
a
InternationalEurophysicsConferenceonHighEnergyPhysics
July21st-27th2005
Lisboa,Portugal
∗Speaker.
†IamthankfultotheNationalScienceFoundationforitsessentialsupportofthiswork,aswellastothestaffof
CESRforitsexcellentacceleratorwork.
(cid:13)c Copyrightownedbytheauthor(s)underthetermsoftheCreativeCommonsAttribution-NonCommercial-ShareAlikeLicence. http://pos.sissa.it/
Upsilonandc analysesatCLEO JeanDuboscq
b
1. The ¡ System
The ¡ (nS) system, a bound state of a b and b quark, is produced in the CLEO detector in
collisions ofelectrons andpositrons fromtheCESRaccelerator. Whenn≥4thestatedecays toB
mesons. Forn<4,belowBmesonthreshold,thetwoquarksmustannihilatetoproducenon-Bfinal
states. Theseannihilations aremediated eitherbyavirtualphoton, theexchange oftwogluons, or
the radiation of one photon and emission of a photon. These states can also transition to other
n2s+1L states via emission of, eg, a photon or two pions. These other states are the QCD analog
J
of positronium. Since the b quark is quite heavy, non-relativistic quantum mechanics can be used
to understand the spectroscopy and dynamics. It is interesting to note that the catalogue of final
statesisincompletely known-forinstanceinthecaseofthe¡ (1S)lessthan10%ofallfinalstates
are measured [1]. The ¡ system also supplies a laboratory to study b quarks in a setting different
from B meson decays. For instance, one’s confidence in Lattice QCD (LQCD) as applied to the
extractionofCKMmatrixelementsinBmesondecays,wouldbeboostedifLQCD’spredictionof
the¡ system wereinlinewithexperimental results.
2. The CLEO IIIDetector and Data Sample
The CLEO III detector[2] provides particle measurements over a large solid angle. This de-
tector, in a 1.5 T magnetic field, includes a silicon tracker, and a drift chamber for charged par-
ticle detection, as well as dE/dx particle identification. Outside this detector is a Ring Imaging
Cerenkov (RICH) detector, providing excellent separation of p , K, p. Muon chambers round out
the subsystems, providing m detection above momenta around 1.5 GeV. This detector sits on the
CESRaccelerator, whichprovideselectronpositroncollisionsnearE ≈10GeV,withsymmetric
CM
beam energies. CLEOIIIhastheworld’s largest datasampleforthe¡ resonances below Bmeson
threshold. Thereareapproximately20million¡ (1S),10million¡ (2S)and5million¡ (3S)decays
inthedataset. CLEOIIIhasalsocollected0.42fb−1 ofdatainthe¡ (5S)region.
3. B Production atthe ¡ (5S)
S
The ¡ (5S) is above the threshold for producing B mesons, and should decay, among other
S
(∗) (∗)
things, to B B states. This would be an interesting place for B factories to run to investigate
S S
B mixing. Uptonow,no experiment hasreported the observation ofB production in¡ decays -
S S
observation andmeasurementofthiswouldbeaninvaluableinputtoplanningforB production at
S
theBfactories.
Anelementaryestimationleadstothehypothesis thattheB mesondecaystofinalstateswith
S
a D with a branching fraction of 92±11%. Non strange B meson decays contain far fewer D
S S
mesons. Thus CLEO maps out the production of D mesons in the fp final states as a function
S
of D meson beam energy scaled momentum using off resonance and on resonance ¡ (4S) data,
S
and onresonance ¡ (5S) data. The¡ (4S) off and on resonance yields are then scaled to the ¡ (5S)
energies, to account for continuum and non strange B meson production of the D , which is then
S
subtracted from the observed spectrum. The remainder is attributed to the decay of B mesons to
S
the D , allowing us to observe for the first time the production of the B at the ¡ (5S), and quote:
S S
Br(¡ (5S)→B(∗)B(∗))=16.0±2.6±6.3%. Thisworkisnowpublished in[3].
S S
2
Upsilonandc analysesatCLEO JeanDuboscq
b
AsecondsearchattemptstoidentifyexclusiveB finalstate. CLEOIIIsearchedfortheB final
S S
states yf , fh , yh ′ and D(∗)p −, D(∗)r −. In a plane defined by the difference in energy between
S S
the beam energy and the final state particle energy versus the beam constrained mass of the final
states, kinematics dictate a separation of the B B , B B∗, and B∗B∗ states. CLEO III observes a
S S S S S S
preponderance of events in the y X and D(∗)X modes in the B∗B∗ region. This indicates that the
S S S
B∗B∗ dominates the decay of the ¡ (5S) as expected in the Unitarized Quark Model [5]. Results
S S
fromthisworkaredetailed in[4].
4. ¡ Decaysto Two Leptons
The decay of the ¡ (nS) (n=1,2,3) to two leptons is interesting because it is a probe of the
coupling oftwobquarkstoavirtualphoton. AssuchitcanbeusedasatestbedforLQCD[6]. In
addition, thissimple decay isagood place totestlepton universality bycomparing thefinalstates
ee, mm , tt . Phase space corrections should be very small, and thus branchings fractions to these
states should beequal. Adeviation from this expectation could beamanifestation ofnew physics
[7]. Thedecaysofthe¡ (1S)toleptonpairsarecurrentlymeasuredatthe2to5%level. The¡ (2S)
decaystotheeeandmm finalstatesarealsomeasuredtoapproximately8%,whilethett finalstate
is known only to within 100%. For the ¡ (3S), only the mm final state is well know to an error of
10%, while the ee mode is known to exist (through the production of the 3S at colliders) and the
tt modehasnotbeenobserved.
5. Analysisof¡ →tt
The analysis of the decay ¡ → tt follows the method used in the analysis of the ¡ → mm
decay[8]. The data on and off resonance for the n=1,2,3 ¡ (nS) states are skimmed for two track
events and selection criteria are tuned using data from the ¡ (4S) to isolate both the mm and tt
finalstates (t leptons decay tofinalstates containing onecharged track approximately 75%ofthe
time.) Theoffresonance data arescaled according toluminosity and beam energy (to account for
background evolution) and subtracted from the on resonance data. The excess is attributed to ¡
decays. The ratio of the tt and mm sample, in which interference between on and off resonance
production cancels, provides, after efficiency correction, for adirect test of lepton universality. In
addition, this method allows as a cross check the verification that the decay of the ¡ (4S) to these
finalstatesdoesnotoccur(atourlevelofsensitivity.) Clearexcessesofeventsareobservedinboth
mm and tt modes at the ¡ (nS), n=1,2,3, and no excess is observed at the ¡ (4S). The observed
eventyieldsinbothfinalstatesarecorrected forefficiency, and,forthe2Sand3S,cascade decays
tolower¡ states, withreasonable assumptions forunmeasured decays. Theinterference corrected
numberof mm eventsisinperfectagreement withourprevious measurementofBr(¡ →mm ),and
in addition, the off resonance production of mm and tt final states is in complete agreement with
the Standard Model expectation. The result of the analysis is that R=Br(¡ →tt )/Br(¡ →mm )
ismeasuredasfollows:
R(1S) = 1.06±0.02±0.00±0.03 (5.1)
R(2S) = 1.00±0.03±0.12±0.03 (5.2)
R(3S) = 1.05±0.07±0.05±0.03 (5.3)
3
Upsilonandc analysesatCLEO JeanDuboscq
b
wherethequotederrorsareduetostatistics, cascade modelingandothersystematics, respectively.
The systematic errors for this preliminary result are quite conservative and should decrease in the
course of the analysis. This is a first observation of the decay ¡ (3S)→tt at the 10 s statistical
level,andisamarkedimprovementintheerroronthebranchingfractionfor¡ (2s)→tt . Notethat
althoughthisresultdoesseemtoagreewiththeexpectationsfromleptonuniversality,thedeviation
in amodel such asthat proposed by [7]inwhich the ¡ mightoccasionally radiatively decay toan
h ,whichthendecaysviaa(non-standard)Higgs,willbesensitivetotheenergyofthephoton. The
b
efficiency for seeing the final tt state will be lower in this scenario because of the extra photon,
andwilldilutethepoweroftheresultreported here,depending onthephoton energy.
6. Dielectron Widthofthe ¡
The dielectron width of the ¡ probes the coupling of the bb system to the electron-positron
system. Ratiosofthesewidthamongthedifferent¡ statesareofinteresttothosewhostudyLQCD
asaprobe ofthequark anti-quark wavefunction overlap asthese should becalculable. Thedirect
measurement of the production of e+e− in ¡ decay is hindered by the large background due to
Bhabhascatteringthatmustbesubstractedinadirectanalysis. AtCLEO,thetimeinvertedprocess
can be measured by measuring the line shape for the production of ¡ as a function of the center
of mass energy of electron positron collisions. The total area of this line shape is the desired
width. The observed line shape is a convolution of several components, including the inherent
width G (≈ 1keV) , the accelerator beam energy spread (≈ 4MeV), and the effects of initial
ee
state radiation, which also provides a non Gaussian smearing on the order of MeVs. The initial
state radiation function smears the observed line shape assymetrically, giving a large tail on the
high end of s1/2 =E . The method used in this scan over the ¡ (nS), n=1,2,3 involved weekly
CM
scans over the resonances, including off resonance points 20 MeV below the resonance peak to
gaugecontinuum backgrounds. Inordertoprobethestability ofthebeamenergymeasurement, in
each scan, the point with the largest slope with respect to s1/2 was revisited. The selected events
were hadronic, as these provided large well understood triggering and reconstruction efficiencies.
The line shape was fit to a 1/s shape to account for continuum hadron production, as well as
the convoluted ¡ line shape. Included in this fit was an estimation of the contribution due to
the interference of the continuum hadron production with ¡ production. The number of events
observedwascorrectedforefficiency,t pairfeedthrough,aswellasforthefarsmallercontribution
due to cosmic rays and beam gas collisions. The efficiency correction included a correction for
the effective invisible width due to both trigger inefficiency and real physics, as estimated using
the cascade ¡ (2S) → ¡ (1S)p +p −, and Monte Carlo simulations. The resulting preliminary fits
measurethewidthstobe:
G (¡ (1S)) = 1.336±0.009±0.019keV (6.1)
ee
G (¡ (2S)) = 0.616±0.010±0.009keV (6.2)
ee
G (¡ (3S)) = 0.425±0.009±0.006keV (6.3)
ee
The errors in the above are statistical and systematic. The systematic error includes a 1.3% error
4
Upsilonandc analysesatCLEO JeanDuboscq
b
duetotheknowledgeoftheluminosity. Theratiosthatareofdirectrelevance toLQCDare:
G (¡ (2S))
ee = 0.461±0.008±0.003 (6.4)
G (¡ (1S))
ee
G (¡ (3S))
ee = 0.318±0.007±0.002 (6.5)
G (¡ (1S))
ee
G (¡ (3S))
ee = 0.690±0.019±0.006 (6.6)
G (¡ (2S))
ee
7. Direct Photons in¡ Decays
Below B meson threshold, two important channels in ¡ decay are the three gluon ggg and
the g gg exchange diagrams. The photon in the g gg process escapes into the detector, and thus is
a direct witness of the physics of the bb overlap. The ratio of the number of events observed in
thesetwoprocessesRg =N(g gg)/N(ggg)isafunctionofthestrongandelectromagneticcoupling,
as well as the charge of the b quark, and can be used to gauge the relative importance of these
two processes. The search proceeds by tabulating hadronic events with an isolated high energy
photon reconstructed inCLEO.Theenergyspectrum ofthephoton isusedtountangle thephysics
from the backgrounds. Continuum processes are subtracted by an off resonance subtraction. The
two main backgrounds are photons that come from ISR, which are the dominant background at
the highest photon energy, while at lower energies the dominant background comes from photons
resulting from p 0 decays. The ISR background is expected to be reasonably well simulated by
the event simulation. The background due to p 0’s comes from complicated physics processes,
and one might question the veracity of a Monte Carlo model. To circumvent this, CLEO uses a
data driven Monte Carlo : using the reconstructed spectrum of charged pions, we can predict the
spectrum of real p 0, and then throw Monte Carlo p 0’s that reproduce this spectrum. These p 0’s
arethenallowedtosimulatetherealbackground process. TheISRandp 0 derivedbackground are
subtracted from the continuum subtracted on resonance data, and the resulting distribution is fit
to a sum of two contributions. The first represents the mismatch of our subtraction at low photon
energies, while the second component is the distribution expected for the direct photons. We fit
totwomodels[9][10],both ofwhich expect apredominance ofphotons athigher energies. These
modelsarederivedforthe¡ (1S)butlackingothersuitablemodels,wealsoapplythemtothe2Sand
3S states. The fits for the higher resonances include the effects of cascades to lower resonances.
In the ratio for Rg , the denominator is determined from the known number of ¡ in the CLEO
dataset, as well input from the PDG[1] and Monte Carlo simulations. The preliminary results
are: Rg (1S)=2.90±0.007±0.22±0.15%, Rg(2S)=3.49±0.03±0.58±0.18%, and Rg(3S)=
2.88±0.03±0.38±0.12% where the errors are statistical, systematic and model dependence.
The systematic error is dominated by the p 0 photon feedthrough model. The value of Rg (1S) is
consistent with previous values, and has similar systematic errors, but smaller statistical errors.
Thisisthefirstmeasurement ofRg (2S)andRg(3S).
8. The Decay ¡ (1S)→h+h−g
The radiative decays of the ¡ (1S) can be used to probe the 2 gluon structure. There are
5
Upsilonandc analysesatCLEO JeanDuboscq
b
many interesting results from J/y decays, including the observation of tensor states f (1270)
2
[11, 12, 13, 14] and f′(1525) [15, 16], the claim of a glueball candidate known as the f (2220)
2 j
[17], and the observation of a pp near threshold enhancement termed the X(1860) [18]. In the
¡ (1S) system, one mighthope toobserve similar effects suppressed byafactor due totherelative
quarkchargeandthequarkmassforthepropagator, squared, whichroughlypredictsratesreduced
by a factor of 1/40. In this analysis, CLEO searches for the final state h+h−g in ¡ (1S) decays,
by requiring a bachelor photon of energy greater than 4 GeV. The h± are identified as p , K or p
by combining dEdx and RICH information. In addition, the knowledge of the total beam energy
is used as an extra input to test for PID hypothesis consistency. Continuum decays are accounted
for by subtracting suitably scaled off resonance data. The resulting 2 body mass spectra reveal
a clear f (1270) in the pp spectrum, as well as a clear f′(1525) in the KK spectrum. No clear
2 2
structure is observed in the pp spectrum. No evidence for the f (2220) is observed in any of the
j
spectra. Thereisanunexplained broadexcess ofeventsintheKK massdistribution above2GeV.
In addition, there is no evidence of the X(1860) in the pp spectrum. The decay angles are fitand
itisconfirmedthatthe f (1270) andthe f′(1525) arespin 2objects decaying predominantly with
2 2
helicity 0. Preliminary branching ratios are: Br(¡ (1S)→g f (1270)) =10.2±0.8±0.7×10−5
2
and Br(¡ (1S)→g f′(1525))=3.7+0.9±0.8×10−5. Theexcess inthe KK spectrum for2GeV <
2 −0.7
m <3GeV isBr(¡ (1S)→g K+K−)=1.14±0.08±0.10×10−5. Upperlimitsat90%C.L.are
KK
set at: Br(¡ →g f (980))<3×10−5, Br(¡ →g f (2050)) <0.6×10−5, Br(¡ →g f (1710)) <
2 2 0
0.7×10−5. Forthemassregion2GeV <m <3GeV,Br(¡ (1S)→g pp)<0.6×10−5. Additional
pp
limits on the branching fraction to the f (2220) and the X(1860) are set at the 10−6 level. The
j
branchingfractionsforthe f (1270)and f (1270)areconsistentwiththenaiveexpectationderived
2 2
from J/y decays. It is not clear that any of the other results are in conflict with this scaling. The
resultsofthisworkcanbefoundin[19]. Anadditionalanalysisisunderwaylookingforfinalstates
including aphotonandp 0p 0,hh ,andp 0h .
9. Observationof c (2P)→c (1P)p +p −
b b
Thetransitionofc (2P)→c (1P)p +p −isobservedinanelectromagneticdecayofthe¡ (3S)
b b
tothe c (2P)followed bythedipion transition downtothe c (1P)followed byaelectromagnetic
b b
transition downtothe¡ (1S). The¡ (1S)isthenobserved byitsdecaytoeeormm . Thefinalstate,
excludingtheleptons,consistsof2chargedpionsand2photons,whichisalsothefinalstateofthe
chain¡ (3S)→¡ (2S)p +p −,¡ (2S)→c g ,c →¡ (1S)g . Thismeasuredbackgroundwithslightly
b b
different photon energies and dipion mass can be used to gauge the correctness of the analysis of
the sought after transition. Other backgrounds exist but require the loss of at least 1 photon, and
are thus somewhat suppressed. The analysis plots the energy of largest energy photon against the
inferred mass of the dipion system. Two analyses were pursued. The cascade pions tend to have
lowmomentum,andthus,reconstruction ofbothofthemtendstohavelowefficiency. TheDi-pion
analysis attempts to reconstruct both pions. In this analysis, the ¡ (2S) contamination is found to
be consistent with expectation, and 7 events are observed in the signal region over a background
of 1.2expected events. Theefficiency forthis analysis ise =4.5%. Thesingle pion analysis sees
17 events, over an expected background of 3.3 events, with an efficiency of e =8.5%. The two
analyses see consistent results and can be combined to give a 6 s significant observation. CLEO
6
Upsilonandc analysesatCLEO JeanDuboscq
b
thus reports the first observation of a dipion transition outside of a 3S system, and a preliminary
1
widthofG (c (2P)→c (1P)p +p −)≈0.9keV. Furtherdetailscanbefoundin[20].
b b
10. Conclusions
This talk has highlighted results from the ¡ analyses at CLEOIII. CLEO has made the first
observation of B production in ¡ (5S) decays. CLEO has also for the first time observed the
S
decay ¡ (3S) →tt , and obtained precision results on the ratio of the tt width to the mm width
in ¡ (nS) decay, n=1,2,3. A measurement of the dielectron width of the ¡ (nS) n=1,2,3 was also
presented. Direct photon measurements in hadronic ¡ decays presented here shed light on the
relativeimportanceofg ggandgggintermediatestatesin¡ decays. Thesubstructureofthehadrons
in the ¡ (1S)→g h+h− decays was determined to be consistent with the extrapolated expectation
from Jy decays. A broad enhancement in the KK channel was observed, and no evidence was
seen for the f (2220) or the X(1600). The first observation of the dipion transition between the
j
c (2P)andthe c (1P)wasalsopresented. Allresults presented inthistalkarepreliminary unless
b b
otherwise noted.
References
[1] ParticleDataGroup,S.Eidelmanetal.,Phys.Lett.B592,1(2004).
[2] G.Viehhauser,Nucl.Instrum.MethodsA462,146(2001);M.Artusoetal.,physics/0506132;
D.Petersonetal.,Nucl.Instrum.MethodsPhys.Res.,Sect.A478,142(2002);A.Warburtonetal.,
Nucl.Instrum.andMethodsPhys.Res.,Sect.A488,451(2002);M.A.Selen,R.M.HansandM.J.
Haney,IEEETrans.Nucl.Sci.48,562(2001);R.M.Hans,C.L.Plager,M.A.SelenandM.J.Haney,
IEEETrans.Nucl.Sci.48,552(2001).
[3] M.Artusoetal.(CLEOCollab.),hep-ex/0508047,toappearinPhys.Rev.Lett.
[4] G.Bonvicinietal.(CLEOCollab.),hep-ex/0510034,submittedtoPhys.Rev.Lett.
[5] S.OnoN.Törnqvist,Phys.Rev.Lett.53,878(1984);S.OnoandN.Törnqvist,J.Lee-FranziniandA.
Sanda,Phys.Rev.Lett.55,2938(1985);S.Ono,A.SandaandN.Törnqvist,Phys.Rev.D34,168
(1986).
[6] seeC.Davies,“Non-perturbativeFieldTheory(LQCD)”,thisconference.
[7] M.A.Sanchis-Lozano,arXiv:hep-ph/0503266;M.A.Sanchis-Lozano,arXiv:hep-ph/0510374.
[8] G.S.Adamsetal.(CLEOCollab.),Phys.Rev.Lett94:012001,2005.
[9] R.D.Field,Phys.Lett.B133,248(1983).
[10] X.GarciaandJ.Soto,Phys.Rev.D69,114006(2004),andhep-ph/0507107.
[11] G.Alexanderetal.(PLUTOCollab.),Phys.Lett.B76,652(1978).
[12] D.L. Sharre,“10thInternationalSymposiumonLeptonandPhotonInteractionsatHighEnergy”,
Bonn(1981).
[13] C.Edwardsetal.(CrystalBallCollab.),Phys.Rev.D25,3065(1982).
[14] J.E.Augustinetal.(DM2Collab.),Z.Phys.C36,369(1987).
7
Upsilonandc analysesatCLEO JeanDuboscq
b
[15] R.M.Baltrusaitisetal.(MARKIIICollab.),Phys.Rev.D35,2077(1987).
[16] J.Z.Baietal.(BESCollab.),Phys.Rev.D68,052003(2003).
[17] J.Z.Baietal.(BESCollab.),Phys.Rev.Lett.76,3502(1996).
[18] J.Z.Baietal.(BESCollab.),Phys.Rev.Lett.91,022001(2003).
[19] S.B.Atharetal.(CLEOCollab.),hep-ex/0510015,submittedtoPhys.Rev.D
[20] C.Cawlfieldetal.(CLEOCollab.),hep-ex/0511019,submittedtoPhys.Rev.D.
8
