charm–anticharm baryon production asymmetries in photon–nucleon interactions

b

Physics Letters B 581 (2004) 39–48

www.elsevier.com/locate/physlet

Charm–anticharm baryon production asymmetriesin photon–nucleon interactions

FOCUS Collaboration1

J.M. Linka, P.M. Yagera, J.C. Anjosb, I. Bediagab, C. Göbelb, A.A. Machadob,J. Magninb, A. Massafferrib, J.M. de Mirandab, I.M. Pepeb, E. Polycarpob,

A.C. dos Reisb, S. Carrilloc, E. Casimiroc, E. Cuautlec, A. Sánchez-Hernándezc,F. Vázquezc, C. Uribed, L. Agostinoe, L. Cinquinie, J.P. Cumalate, B. O’Reilly e,I. Segonie, M. Wahle, J.N. Butlerf, H.W.K. Cheungf, G. Chiodinif, I. Gainesf,P.H. Garbinciusf, L.A. Garrenf, E. Gottschalkf, P.H. Kasperf, A.E. Kreymerf,R. Kutschkef, M. Wangf, L. Benussig, M. Bertanig, S. Biancog, F.L. Fabbrig,

A. Zallo g, M. Reyesh, C. Cawlfieldi, D.Y. Kim i, A. Rahimii, J. Wissi, R. Gardnerj,A. Kryemadhij, Y.S. Chungk, J.S. Kangk, B.R. Kok, J.W. Kwakk, K.B. Leek, K. Chol,

H. Parkl, G. Alimontim, S. Barberism, M. Boschinim, A. Ceruttim, P. D’Angelom,M. DiCoratom, P. Dinim, L. Ederam, S. Erbam, M. Giammarchim, P. Inzanim,

F. Leverarom, S. Malvezzim, D. Menascem, M. Mezzadrim, L. Moronim, D. Pedrinim,C. Pontogliom, F. Prelzm, M. Roverem, S. Salam, T.F. Davenport IIIn, V. Arenao,

G. Bocao, G. Bonomio, G. Gianinio, G. Liguorio, M.M. Merlo o, D. Panteao,D. Lopes Pegnao, S.P. Rattio, C. Riccardio, P. Vituloo, H. Hernandezp, A.M. Lopezp,

E. Luiggip, H. Mendezp, A. Parisp, J.E. Ramirezp, Y. Zhangp, J.R. Wilsonq,T. Handlerr, R. Mitchellr, D. Enghs, M. Hosacks, W.E. Johnss, M. Nehrings,

P.D. Sheldons, K. Stensons, E.W. Vaanderings, M. Websters, M. Sheafft

a University of California, Davis, CA 95616, USAb Centro Brasileiro de Pesquisas Físicas, Rio de Janeiro, RJ, Brazil

c CINVESTAV, 07000 México City, DF, Mexicod IFUAP, 72570 Puebla, Mexico

e University of Colorado, Boulder, CO 80309, USAf Fermi National Accelerator Laboratory, Batavia, IL 60510, USA

g Laboratori Nazionali di Frascati dell’INFN, I-00044 Frascati, Italyh University of Guanajuato, 37150 Leon, Guanajuato, Mexico

i University of Illinois, Urbana-Champaign, IL 61801, USAj Indiana University, Bloomington, IN 47405, USA

k Korea University, Seoul 136-701, South Koreal Kyungpook National University, Taegu 702-701, South Korea

m INFN and University of Milano, Milano, Italyn University of North Carolina, Asheville, NC 28804, USA

o Dipartimento di Fisica Nucleare e Teorica and INFN, Pavia, Italy

0370-2693/$ – see front matter 2004 Published by Elsevier B.V.doi:10.1016/j.physletb.2003.12.011

http://www.elsevier.com/locate/physletb

40 FOCUS Collaboration / Physics Letters B 581 (2004) 39–48

he regiontries of the

the

p University of Puerto Rico, Mayaguez, PR 00681, USAq University of South Carolina, Columbia, SC 29208, USA

r University of Tennessee, Knoxville, TN 37996, USAs Vanderbilt University, Nashville, TN 37235, USA

t University of Wisconsin, Madison, WI 53706, USA

Received 26 November 2003; accepted 3 December 2003

Editor: M. Doser

Abstract

We report measurements of the charm–anticharm production asymmetries forΛ+c , Σ++

c , Σ0c , Σ++∗

c ,Σ0∗c , andΛ+

c (2625)baryons from the Fermilab photoproduction experiment FOCUS (E831). These asymmetries are integrated over twhere the spectrometer has good acceptance. In addition, we have obtained results for the photoproduction asymmeΛc baryons as functions ofpL, p2

T, andxF . The integrated asymmetry forΛ+

c production,(σΛ+c

− σΛ−c)/(σ

Λ+c

+ σΛ−c), is

0.111± 0.018± 0.012, significantly different from zero. The asymmetries of the excited states are consistent withΛcasymmetry. 2004 Published by Elsevier B.V.

ones.of

ntlyeredib-or-armre-ea-ics.and

canon

ted-

s toese

pleark

ntse,

.re-

n-et-ym-

ons

de-edrgy

mryl-edeVus-ntsr-ck-five

woreeldandom-

The FOCUS experiment uses a photon beama beryllium oxide target to produce charm particlIn high energy photon–hadron interactions, pairscharm–anticharm quarks are produced predominathrough photon–gluon fusion [1]. At leading ordin quantum chromodynamics (QCD), the produccharm and anticharm particles are identically distruted in the kinematic variables. At next-to-leadingder, small asymmetries are expected between chand anticharm production [2–4]. However, these pdicted asymmetries would be too small to be msured at the current level of experimental statistInteractions between the produced charm quarksthose in the struck nucleon during hadronizationalso induce production asymmetries. One commmodel is that of string fragmentation as implemenin PYTHIA [5]. In this model, the charm and anticharm quarks are connected through color stringthe quarks in the struck nucleon. The energy in thstrings is converted to particles by “popping”qq pairsout of the vacuum. The simplest asymmetry examoccurs when a charm quark is connected to the diquin the nucleon by a low energy string with insufficieenergy to produce any additional particles. In this cathe charm quark can combine with valenceu and d

1 Seehttp://www-focus.fnal.gov/authors.htmlfor additional au-thor information.

quarks to form aΛc while a c quark can only formmesons when it combines with the valence quarks

In this Letter we present a high-statistics measument of the production asymmetry ofΛc baryons fromphoton–nucleon interactions, providing the first covincing evidence for a non-zero asymmetry. This Lter also contains the first measurements of this asmetry as functions ofpL, p2

T , andxF . In addition, theproduction asymmetries of the excited charm baryΣ++c , Σ0

c , Σ++∗c , Σ0∗

c , andΛ+c (2625) are presented

for the first time. All of these states decay to aΛc plusone or two charged pions.

The FOCUS (Fermilab E831) experiment wassigned to study charm particle physics. Charmhadrons are produced by the interaction of high enephotons (〈E〉 ≈ 175 GeV for events in which a chardecay was reconstructed) with a segmented belium oxide (BeO) target. The photons are producby bremsstrahlung in a lead target with a 300 Ge+/e− beam. Vertex reconstruction is performeding four silicon strip planes interleaved with segmeof the target followed by a 12 plane silicon strip vetex detector. Downstream of the vertex detector, traing and momentum measurements are made usingstations of multiwire proportional chambers and tlarge aperture magnets with opposite polarity. Thmulticell Cerenkov counters operating in threshomode are used to identify electrons, pions, kaons,protons over a wide range of momenta. The spectr

http://www-focus.fnal.gov/authors.html

FOCUS Collaboration / Physics Letters B 581 (2004) 39–48 41

tro-ata

un.

ofandckireated

esses,beby

re

s,

tonsese

iredchsis

logisnsTo

me8

oveen

e0

n

ise.

ed.drors

theonm

hisnfit

in-

geif-the

amthe

ofseith

ofrmdnts

welland

wero-so,and

e, ificlered,m-

me-at

as a

eter also contains a hadron calorimeter, two elecmagnetic calorimeters, and two muon detectors. Dwere collected during the 1996–1997 fixed-target r

TheΛ+c particles are reconstructed using thepK−

π+ decay mode.2 The decay vertex is formed fromthree charged tracks in the event. The vector sumtheir momenta is projected back toward the targetused as aseed to intersect with at least one other train the event to form a production vertex. We requthe production and decay vertices to be separby at least 5.5σ�, whereσ� is the uncertainty in themeasured vertex separation. We also place goodnof-fit criteria on the primary and secondary verticrequiring that the confidence level for each vertexgreater than 1%. Identification of the decay tracksparticle type is performed using theCerenkov detectosystem [6]. Aχ2-like variable is formed using thon/off status of all cells within a particle’sCerenkovcone (β = 1). For each of the four possibilitieelectron, pion, kaon, and proton, we calculateWi =−2Σcells

j logPj , wherei is the particle type andj thecell number.Pj is the probability that thej th cellwill yield the observed response given particle typei.Identification is then based on differences in theWi .The proton candidate is required to have the prohypothesis favored over the kaon and pion hypotheby 1 and 4 units of log likelihood, respectively. Thkaon hypothesis for the kaon candidate is requto be favored over the pion hypothesis by 3 suunits. For the pion candidate, the pion hypothecannot be disfavored by more than 6 units oflikelihood relative to the most likely hypothesis. Thvery loose requirement is also applied to the piofrom the excited charm decays, described below.remove longer lived charm backgrounds, the lifetiof the Λc in its rest frame must be shorter thantimes the world average lifetime [7].Λc candidates areidentified as all those events which satisfy the abcriteria and which fall into the mass range betwe2.10 and 2.45 GeV/c2. Finally, we restrict our samplto be in the (large) phase space region given by 4<pL < 200 GeV/c andp2

T < 6.0 (GeV/c)2.The Σ++

c , Σ0c , Σ++∗

c , and Σ0∗c candidates are

reconstructed by combining theΛ+c candidates within

2σ of the meanΛ+c mass with a single charged pio

2 Charge conjugate states are implied, unless stated otherw

-

track. All possible combinations in the event are triThe confidence level of theΣc decay vertex is requireto be greater than 1%. To reduce systematic ercoming from the reconstruction of theΛ+

c and obtainbetter signal-to-noise, we study theΣc states usingthe difference in the invariant mass of eachΣc stateand the invariant mass of theΛ+

c , ∆M. Because thepion is typically of low momentum, it suffers froma considerable amount of multiple scattering, anduncertainty on its momentum dominates the errorthe Σc invariant mass. To improve the momentumeasurement, the primary vertex is refit without tsoft pion track (if possible), and the pion directiois recomputed, forcing it to come from the reprimary.Λ+c (2625) candidates are reconstructed by comb

ing theΛ+c candidates within 2σ of the meanΛ+

c masswith all combinations of two pions of opposite charin the event. As for theΣc states, we use the mass dference plots and force the two pions to come fromprimary.

FOCUS data were taken with three different beenergies and with two different radiators. Sinceasymmetries can depend on the energy spectrumthe incident beam photons, we include only thoevents in our sample that come from data taken wa 300 GeVe−/e+ beam on a lead radiator of 20%a radiation length. This selection results in a chabaryon yield which is about 75% of the yield founusing all of the data. For reconstructed charm evepassing the trigger, the beam energy is reasonablydescribed by a Gaussian with mean of 175 GeVwidth of 45 GeV.

From studies of high statistics meson decaysfind no evidence that there is any asymmetry intduced by charge bias in the spectrometer. Evensince the acceptance depends on the longitudinaltransverse momenta of the produced baryon statthese spectra are different for particle and antipartover the range for which the asymmetry is measuthere will be an acceptance difference for the two saples which can be corrected. This corrected asymtry gives the asymmetry for an experiment with flacceptance inpL andp2

T over the range 40< pL <200 GeV/c andp2

T < 6.0 (GeV/c)2.To compensate for this acceptance difference

function ofp2T andpL, the production asymmetryA

is determined by:


entum on
Fig. 1.pK−π+ weighted invariant mass distribution split by charge. The weights are used to account for efficiency loss versus moman event-by-event basis. The average weight for the whole histogram is given in the figure asw. The unweighted yields are also shown.
ed

ef-Sm-USsin

dis-atate

thethe

nsed

ht

gtri-ndidth

lu-tly

theis

theto

de-thethemall

o-are–6.

(1)A= N/ε − N/εN/ε + N/ε ,

whereN and N are the numbers of reconstructbaryons and antibaryons, respectively, andε (ε) is thebaryon (antibaryon) reconstruction efficiency. Theficienciesε and ε are calculated using the FOCUMonte Carlo simulation program. The detector siulation uses a detailed description of the FOCdetector. The physics processes are generated uPYTHIA (version 6.127), with many PYTHIA parame-ters tuned to match the observed FOCUS physicstributions. The remaining discrepancy between dand Monte Carlo is removed by weighting MonCarlo events to exactly match the observed datapLandp2

T distributions. For the excited charm states,particle/antiparticle efficiencies are used to correctasymmetry. For the higher statisticsΛc decays a fur-ther step is taken. Since very littlep2

T dependence oefficiency is observed, the Monte Carlo events are uto obtain the efficiency variation versuspL. This effi-ciency is fit to a function which is then used to weigeach event.

g

The globalΛc asymmetry is obtained by fittineach of the two weighted invariant mass data disbutions shown in Fig. 1 with a Gaussian signal aquadratic background. The Gaussian mean and ware allowed to float separately forΛ+

c andΛ−c . The

unweighted yields forΛ+c andΛ−

c are 5427±120 and4242± 108, respectively.

To account for natural widths and changing resotions, the excited charm baryon states are fit slighdifferently. The Monte Carlo is used to generatesignal shape which is fit to a spline function. Thspline function, properly normalized, is used assignal shape for the data, with the mass allowedfloat. The advantage in the case of theΣc, andΣ∗

c

is that the spline function is able to account for thetector resolution and natural width of the state. ForΛ∗c state, the spline function is a better model of

detector resolution than a Gaussian due to the sphase space. The background shapes forΣc , Σ∗

c andΛc(2625) are a threshold functionN(1 + α(∆M −mπ)∆M

β), quadratic polynomial, and linear polynmial, respectively. The fits to each data sampleshown on the corresponding data plots in Figs. 2


Fig. 2.Λ+c π

+ invariant mass distribution split by charge.

re-

ch

assandFors

rtain

eingof

.ndughtheby

aly-theine-

shthelsoed.tic

fornd.

raleoraand

esforn.

theofn.an

The global asymmetries are calculated from theturned yields.

TheΛc data samples are divided into bins of eaof the kinematic variablespL, p2

T , andxF , integratingover the full range of the other variables. The mplots for each bin are fit using a Gaussian signalquadratic background, as for the full data set.these fits, theΛ+

c (Λ−c ) mass is fixed to the mas

obtained by fitting the fullΛ+c (Λ−

c ) sample. Thewidths are fixed to follow the Monte Carlo widths foeach bin. The yields from these fits are used to obthe production asymmetry versuspL, p2

T , andxF . Theefficiency corrected asymmetry distributions forΛc vspL, p2

T , andxF are shown in Figs. 7, 8, and 9. ThxF measurement requires knowledge of the incombeam energy which is only available in about 30%the data sample and is therefore of lower statistics

In our first systematic error check, backgroustudies were performed to assure that feedthrofrom other charm states does not contribute toasymmetry. The systematic errors are obtainedmaking the same measurements under different ansis conditions. We performed these studies forglobal asymmetry and the asymmetry versus the kmatic variables reported.

For theΛc and the three excited states (Σc, Σ∗c ,

andΛ∗c ), fits with different bin widths and with bin

shifted by 1/2 bin were performed. A variation whicfixed the mass for baryons and antibaryons tovalue from the total sample was also studied. We ainvestigated differences due to the fit functions usFor theΣc states a Gaussian function with a quadrabackground was utilized to fit the signals. For theΣ∗

c

mass distributions we used a linear and cubic fitthe background instead of the quadratic backgrouFor Σc andΣ∗

c , fits with spline functions from twoadditional Monte Carlo samples where the natuwidths were varied by±1σ were used to estimate thuncertainty in the knowledge of the natural widths. FtheΛ∗

c we used two different fit functions to providesystematic check: a Gaussian for the signal shapea linear background and theΛ∗

c spline function with aquadratic background. Additionally, for all the statwe used the respective raw asymmetries to checksystematic problems with the efficiency correctioFor theΛc we also calculated the asymmetry usingefficiencies directly from the Monte Carlo insteadusing the Monte Carlo to obtain an efficiency functio

The r.m.s. of all of these variations is used asestimate of the systematic error.


Fig. 3.Λ+c π

− invariant mass distribution split by charge.

Fig. 4.Λ+c π

+ invariant mass distribution split by charge.


Fig. 5.Λ+c π

− invariant mass distribution split by charge.

Fig. 6.Λ+c π

+π− invariant mass distribution split by charge.


re

re

Fig. 7.pL asymmetry distribution forΛc along with statistical and systematic errors and the PYTHIA prediction—the systematic errors asuperposed on the statistical errors and are smaller than the statistical errors in every bin.

Fig. 8.p2T asymmetry distribution forΛc along with statistical and systematic errors and the PYTHIA prediction—the systematic errors a

superposed on the statistical errors and are smaller than the statistical errors in every bin.

nshe

sys-

nts.n

ng.en-

m-

e-nts

.ex-for-asealsoles

try

of

The global asymmetries for all of the charm baryostudied in this analysis are shown in Table 1. Tfirst errors are statistical and the second aretematic. Table 1 also shows comparisons to PYTHIA

asymmetries calculated with unreconstructed eveThe PYTHIA predictions come from running versio6.203 with all parameters left at the default settiThe events are generated with the correct beamergy distribution and only candidates within the noinal phase space, 40< pL < 200 GeV/c andp2

T <

6.0 (GeV/c)2, are used in determining the asymmtry. The E691 and E687 photoproduction experime

have previously reported global asymmetries forΛc of0.110± 0.089 [8] and 0.035± 0.076 [9], respectivelyOur results are similar to those obtained by theseperiments although comparisons are not straightward since all three experiments have different phspace and beam energy distributions. Table 1shows the efficiency ratios of particles to antiparticfor the charm baryons studied.

We have studied the photoproduction asymmeof Λ+

c versusΛ−c using the decay channelpK−π+.

From∼ 10 000Λc events we present the first resultsthis asymmetry as functions ofpL, p2

T , andxF . These


re
Fig. 9. xF asymmetry distribution forΛc along with statistical and systematic errors and the PYTHIA prediction—the systematic errors asuperposed on the statistical errors and are smaller than the statistical errors in every bin.
Table 1Raw and corrected global production asymmetry for the charm baryons compared to the predictions of default PYTHIA version 6.203. Theefficiency ratio of particles to antiparticles is also shown

Baryon Raw asymmetry Corrected asymmetry PYTHIA Efficiency ratio

Λ+c 0.123± 0.017 0.111± 0.018± 0.012 0.073 1.023± 0.005

Σ++c 0.147± 0.073 0.136± 0.073± 0.036 0.126 1.024± 0.010

Σ0c 0.005± 0.089 0.005± 0.089± 0.024 0.128 1.000± 0.009

Σ++∗c 0.188± 0.105 0.181± 0.105± 0.033 0.133 1.016± 0.006

Σ0∗c 0.299± 0.165 0.298± 0.165± 0.023 0.132 1.002± 0.006

Λ+c (2625) −0.066± 0.086 −0.075± 0.087± 0.021 N/A 1.019± 0.017

bal

tes

stitiver

with

m-e

s no

ofof

in-rra-detedUSdi

ni-er-

,ntou-un-

results show a clear positive asymmetry. The gloΛ+c asymmetry is measured to be 0.111± 0.018±

0.012.The production asymmetry of excited charm sta

which decay toΛ+c , including theΣ++

c , Σ0c , Σ++∗

c ,Σ0∗c , andΛ+

c (2625) was also measured for the firtime. The measurements generally indicate a posasymmetry similar to theΛc. Because of the smallesample size, however, they are also consistentzero.

We find that the string fragmentation model as iplemented in PYTHIA version 6.203 does not describtheΛ+

c asymmetry dependence onpL, p2T , orxF . Our

measurements indicate that the asymmetry showsignificant dependence in these variables.

Acknowledgements

We acknowledge the assistance of the staffsFermi National Accelerator Laboratory, the INFN

Italy, the physics departments of the collaboratingstitutions and the Instituto de Física “Luis Rivera Tezas” de la Benemérita Universidad AutónomaPuebla (IFUAP), México. This research was supporin part by the US National Science Foundation, theDepartment of Energy, the Italian Istituto NazionaleFisica Nucleare and Ministero della Istruzione, Uversità e Ricerca, Organización de los Estados Amicanos (OEA), IFUAP-México, CONACyT-Méxicothe Brazilian Conselho Nacional de DesenvolvimeCientífico e Tecnológico, the Korean Ministry of Edcation, and the Korean Science and Engineering Fodation.

References

[1] L.M. Jones, H.W. Wyld, Phys. Rev. D 17 (1978) 759.[2] R.K. Ellis, P. Nason, Nucl. Phys. B 312 (1989) 551.[3] J. Smith, W.L. van Neerven, Nucl. Phys. B 374 (1992) 36.


s.

8.ds

02)

89)

70

[4] S. Frixione, M. Mangano, P. Nason, G. Ridolfi, Nucl. PhyB 412 (1994) 225.

[5] T. Sjöstrand, et al., Comput. Phys. Commun. 135 (2001) 23[6] FOCUS Collaboration, J.M. Link, et al., Nucl. Instrum. Metho

A 484 (2002) 270.[7] Particle Data Group, K. Hagiwara, et al., Phys. Rev. D 66 (20

010001.

[8] E691 Collaboration, J.C. Anjos, et al., Phys. Rev. Lett. 62 (19513.

[9] E687 Collaboration, P.L. Frabetti, et al., Phys. Lett. B 3(1996) 222.

charm–anticharm baryon production asymmetries in photon–nucleon interactions

Documents