Search for doubly charmed baryons and studyof charmed strange baryons at Belle

Y. Kato,36 T. Iijima,37,36 I. Adachi,11 H. Aihara,62 D. M. Asner,47 T. Aushev,20 A. M. Bakich,56 A. Bala,48 Y. Ban,49

V. Bhardwaj,38 B. Bhuyan,14 A. Bobrov,3 G. Bonvicini,68 A. Bozek,42 M. Bračko,31,21 T. E. Browder,10 D. Červenkov,4

V. Chekelian,32 A. Chen,39 B. G. Cheon,9 K. Chilikin,20 R. Chistov,20 K. Cho,24 V. Chobanova,32 Y. Choi,55 D. Cinabro,68

J. Dalseno,32,58 M. Danilov,20,34 Z. Doležal,4 Z. Drásal,4 A. Drutskoy,20,34 D. Dutta,14 K. Dutta,14 S. Eidelman,3 H. Farhat,68

J. E. Fast,47 T. Ferber,6 V. Gaur,57 N. Gabyshev,3 S. Ganguly,68 A. Garmash,3 R. Gillard,68 Y. M. Goh,9 B. Golob,29,21

J. Haba,11 K. Hayasaka,37 H. Hayashii,38 X. H. He,49 Y. Horii,37 Y. Hoshi,60 W.-S. Hou,41 Y. B. Hsiung,41 K. Inami,36

A. Ishikawa,61 Y. Iwasaki,11 T. Iwashita,38 I. Jaegle,10 T. Julius,33 J. H. Kang,70 E. Kato,61 T. Kawasaki,44 C. Kiesling,32

D. Y. Kim,54 H. J. Kim,27 J. B. Kim,25 J. H. Kim,24 M. J. Kim,27 Y. J. Kim,24 J. Klucar,21 B. R. Ko,25 P. Kodyš,4 S. Korpar,31,21

P. Krokovny,3 T. Kuhr,23 A. Kuzmin,3 Y.-J. Kwon,70 S.-H. Lee,25 J. Li,53 Y. Li,67 L. Li Gioi,32 J. Libby,15 Y. Liu,5

D. Liventsev,11 D. Matvienko,3 K. Miyabayashi,38 H. Miyata,44 R. Mizuk,20,34 A. Moll,32,58 N. Muramatsu,51 R. Mussa,19

Y. Nagasaka,12 E. Nakano,46 M. Nakao,11 H. Nakazawa,39 M. Nayak,15 E. Nedelkovska,32 C. Ng,62 M. Niiyama,26

N. K. Nisar,57 S. Nishida,11 O. Nitoh,65 S. Ogawa,59 S. Okuno,22 P. Pakhlov,20,34 G. Pakhlova,20 C.W. Park,55 H. Park,27

H. K. Park,27 T. K. Pedlar,30 T. Peng,52 R. Pestotnik,21 M. Petrič,21 L. E. Piilonen,67 M. Ritter,32 M. Röhrken,23

A. Rostomyan,6 H. Sahoo,10 T. Saito,61 Y. Sakai,11 S. Sandilya,57 L. Santelj,21 T. Sanuki,61 V. Savinov,50 O. Schneider,28

G. Schnell,1,13 C. Schwanda,17 D. Semmler,7 K. Senyo,69 O. Seon,36 M. Shapkin,18 C. P. Shen,2 T.-A. Shibata,63

J.-G. Shiu,41 B. Shwartz,3 A. Sibidanov,56 Y.-S. Sohn,70 A. Sokolov,18 E. Solovieva,20 S. Stanič,45 M. Starič,21 M. Steder,6

M. Sumihama,8 T. Sumiyoshi,64 U. Tamponi,19,66 K. Tanida,53 G. Tatishvili,47 Y. Teramoto,46 M. Uchida,63 S. Uehara,11

T. Uglov,20,35 Y. Unno,9 S. Uno,11 C. Van Hulse,1 P. Vanhoefer,32 G. Varner,10 A. Vinokurova,3 V. Vorobyev,3

M. N. Wagner,7 C. H. Wang,40 M.-Z. Wang,41 P. Wang,16 M. Watanabe,44 Y. Watanabe,22 K. M. Williams,67 E. Won,25

Y. Yamashita,43 S. Yashchenko,6 Z. P. Zhang,52 V. Zhilich,3 V. Zhulanov,3 and A. Zupanc23

(Belle Collaboration)

1University of the Basque Country UPV/EHU, 48080 Bilbao2Beihang University, Beijing 100191

3Budker Institute of Nuclear Physics SB RAS and Novosibirsk State University, Novosibirsk 6300904Faculty of Mathematics and Physics, Charles University, 121 16 Prague

5University of Cincinnati, Cincinnati, Ohio 452216Deutsches Elektronen–Synchrotron, 22607 Hamburg

7Justus-Liebig-Universität Gießen, 35392 Gießen8Gifu University, Gifu 501-1193

9Hanyang University, Seoul 133-79110University of Hawaii, Honolulu, Hawaii 96822

11High Energy Accelerator Research Organization (KEK), Tsukuba 305-080112Hiroshima Institute of Technology, Hiroshima 731-5193

13IKERBASQUE, Basque Foundation for Science, 48011 Bilbao14Indian Institute of Technology Guwahati, Assam 78103915Indian Institute of Technology Madras, Chennai 600036

16Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 10004917Institute of High Energy Physics, Vienna 1050

18Institute for High Energy Physics, Protvino 14228119INFN - Sezione di Torino, 10125 Torino

20Institute for Theoretical and Experimental Physics, Moscow 11721821J. Stefan Institute, 1000 Ljubljana

22Kanagawa University, Yokohama 221-868623Institut für Experimentelle Kernphysik, Karlsruher Institut für Technologie, 76131 Karlsruhe

24Korea Institute of Science and Technology Information, Daejeon 305-80625Korea University, Seoul 136-71326Kyoto University, Kyoto 606-8502

27Kyungpook National University, Daegu 702-70128École Polytechnique Fédérale de Lausanne (EPFL), Lausanne 1015

29Faculty of Mathematics and Physics, University of Ljubljana, 1000 Ljubljana30Luther College, Decorah, Iowa 5210131University of Maribor, 2000 Maribor

32Max-Planck-Institut für Physik, 80805 München

PHYSICAL REVIEW D 89, 052003 (2014)

1550-7998=2014=89(5)=052003(14) 052003-1 © 2014 American Physical Society

33School of Physics, University of Melbourne, Victoria 301034Moscow Physical Engineering Institute, Moscow 115409

35Moscow Institute of Physics and Technology, Moscow Region 14170036Graduate School of Science, Nagoya University, Nagoya 464-860237Kobayashi-Maskawa Institute, Nagoya University, Nagoya 464-8602

38Nara Women’s University, Nara 630-850639National Central University, Chung-li 3205440National United University, Miao Li 36003

41Department of Physics, National Taiwan University, Taipei 1061742H. Niewodniczanski Institute of Nuclear Physics, Krakow 31-342

43Nippon Dental University, Niigata 951-858044Niigata University, Niigata 950-2181

45University of Nova Gorica, 5000 Nova Gorica46Osaka City University, Osaka 558-8585

47Pacific Northwest National Laboratory, Richland, Washington 9935248Panjab University, Chandigarh 160014

49Peking University, Beijing 10087150University of Pittsburgh, Pittsburgh, Pennsylvania 15260

51Research Center for Electron Photon Science, Tohoku University, Sendai 980-857852University of Science and Technology of China, Hefei 230026

53Seoul National University, Seoul 151-74254Soongsil University, Seoul 156-743

55Sungkyunkwan University, Suwon 440-74656School of Physics, University of Sydney, NSW 2006

57Tata Institute of Fundamental Research, Mumbai 40000558Excellence Cluster Universe, Technische Universität München, 85748 Garching

59Toho University, Funabashi 274-851060Tohoku Gakuin University, Tagajo 985-8537

61Tohoku University, Sendai 980-857862Department of Physics, University of Tokyo, Tokyo 113-0033

63Tokyo Institute of Technology, Tokyo 152-855064Tokyo Metropolitan University, Tokyo 192-0397

65Tokyo University of Agriculture and Technology, Tokyo 184-858866University of Torino, 10124 Torino

67CNP, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 2406168Wayne State University, Detroit, Michigan 48202

69Yamagata University, Yamagata 990-856070Yonsei University, Seoul 120-749

(Received 3 December 2013; published 17 March 2014)

We report results of a study of doubly charmed baryons and charmed strange baryons. The analysis isperformed using a 980 fb−1 data sample collected with the Belle detector at the KEKB asymmetric-energyeþe− collider. We search for doubly charmed baryons ΞþðþÞ

cc with the Λþc K−πþðπþÞ and Ξ0

cπþðπþÞ final

states. No significant signal is observed. We also search for two excited charmed strange baryons,Ξcð3055Þþ and Ξcð3123Þþ with the Σþþ

c ð2455ÞK− and Σþþc ð2520ÞK− final states. The Ξcð3055Þþ signal is

observed with a significance of 6.6 standard deviations including systematic uncertainty, while no signatureof the Ξcð3123Þþ is seen. We also study properties of the Ξcð2645Þþ and measure a width of2.6� 0.2ðstatÞ � 0.4ðsystÞ MeV=c2, which is the first significant determination.

DOI: 10.1103/PhysRevD.89.052003 PACS numbers: 14.20.Lq, 14.20.-c, 14.20.Gk

I. INTRODUCTION

In recent years, there has been significant progress in

charmed baryon spectroscopy, mainly by the Belle and

BABAR experiments [1–8]. In particular, all the ground statesof the single-charmed baryons predicted by the constituent

quarkmodel andseveral excited stateshavebeenobserved [9].

However, there are no experimentally established doublycharmed baryons. The lightest doubly charmed baryoncontains two charm quarks and one up or down quark(Ξþ

cc ¼ ccd, Ξþþcc ¼ ccu), and the spin-parity of the ground

state is expected to be 12þ. The mass of the Ξcc has been

extensively studied theoretically, and the prediction of thequark model ranges from 3.48 GeV=c2 to 3.74 GeV=c2

KATO et al. PHYSICAL REVIEW D 89, 052003 (2014)

052003-2

[10–22], whereas the mass predicted by lattice QCD rangesfrom 3.51 GeV=c2 to 3.67 GeV=c2 [23–27]. The crosssections of the Ξcc production in the process eþe− → ΞccXat

ffiffiffis

p ¼ 10.58 GeV, where X denotes the remainingparticles produced in the fragmentation, is predicted tobe 70 fb in Ref. [28] and 230 fb in Ref. [29]. The crosssection of the pair production of the cc and c̄ c̄ diquarks ispredicted to be 7 fb [30].There have been several experimental studies to

search for the Ξcc. The SELEX collaboration reportedevidence for the Ξþ

cc in the Λþc K−πþ [31] and pDþK−

[32] final states with a mass of about 3.52 GeV=c2

using a 600 GeV=c charged hyperon beam. However,the results have not been supported by FOCUS [33],BABAR [34], Belle [2] nor LHCb [35]. The BABARcollaboration searched for the ΞþðþÞ

cc in the Λþc K−πþðπþÞ

and Ξ0cπ

þðπþÞ decay modes with a 232 fb−1 data sampleof eþe− collisions at or near the Υð4SÞ. They found noevidence for the ΞþðþÞ

cc and set an upper limit on theproduct of the production cross section and branchingfractions of Ξcc and Λþ

c or Ξ0c to be a few fb, depending

on the decay mode. In our search for the Ξþcc in the

Λþc K−πþ final state with a 462 fb−1 data sample of Belle

at or near the Υð4SÞ [2], Belle also found no evidencefor the Ξþ

cc and set an upper limit on σðeþe− → ΞþccXÞ ×

BðΞþcc → Λþ

c K−πþÞ=σðeþe− → Λþc XÞ of 1.5 × 10−4 with

a p�ðΛþc Þ > 2.5 GeV=c requirement. Here, p�ðΛþ

c Þis the momentum of the Λþ

c in the center-of-mass(CM) frame.In this paper, we report on an improved search for the Ξcc

in its weak decays to the Λþc K−πþðπþÞ and Ξ0

cπþðπþÞ final

states. The Belle collaboration has collected a data samplewith a total integrated luminosity of 980 fb−1, which isaround two (four) times the statistics of the previous Ξccsearch by Belle [2] (BABAR [34]) and supersedes the resultsin Ref. [2]. Furthermore, in the previous studies, the Λþ

cand the Ξ0

c states have been reconstructed only fromdecay modes of pK−πþ and Ξ−πþ, respectively. Weincorporate additional decay modes to improve the stat-istical sensitivity.The same data sample can be used to study charmed

strange baryons, as the Λþc K−πþ and the Ξ0

cπþ final states

are strong decay modes of excited Ξþc (Ξ�þ

c ) states. TheBABAR collaboration found two Ξ�þ

c states, Ξcð3055Þþ andΞcð3123Þþ, decaying to the Λþ

c K−πþ final state throughintermediate Σcð2455Þþþ or Σcð2520Þþþ states using adata sample of 384 fb−1 [6]. Their statistical significancewas 6.4 standard deviations (σ) and 3.6σ, respectively. Aconfirmation of these states in other experiments is

necessary. The Ξ0cπ

þ is a strong decay mode of theΞcð2645Þþ. Currently, only the upper limit of3.1 MeV=c2 exists for its width [36]. In this paper, wealso report on a search for the Ξcð3055Þþ and Ξcð3123Þþ inthe Λþ

c K−πþ final state, and the measurement of the widthof the Ξcð2645Þþ.The remaining sections of the paper are organized as

follows. In Sec. II, the data samples and the Belle detectorare described. In Sec. III, a study of the final states with Λþ

c ,i.e., the Ξcc search and the Ξcð3055Þþ and Ξcð3123Þþsearch, are reported. In Sec. IV, a study of the final statewith Ξ0

c i.e., the Ξcc search and measurement of the width ofthe Ξcð2645Þþ, are described. Finally, conclusions aregiven in Sec. V.

II. DATA SAMPLES AND THE BELLE DETECTOR

We use a data sample with a total integrated luminosityof 980 fb−1 recorded with the Belle detector at the KEKBasymmetric-energy eþe− collider [37]. The data sampleswith different beam energies at or near the Υð1SÞ to Υð5SÞare combined in this study. The beam energies andintegrated luminosities are summarized in Table I. Theluminosity-weighted average of

ffiffiffis

pis 10.59 GeV.

The Belle detector is a large-solid-angle magneticspectrometer that consists of a silicon vertex detector(SVD), a 50-layer central drift chamber (CDC), an arrayof aerogel threshold Cherenkov counters (ACC), a barrel-like arrangement of time-of-flight scintillation counters(TOF), and an electromagnetic calorimeter comprised ofCsI(Tl) crystals (ECL) located inside a superconductingsolenoid coil that provides a 1.5 T magnetic field. An ironflux return located outside of the coil is instrumented todetect K0

L mesons and to identify muons (KLM). Thedetector is described in detail elsewhere [38]. Two innerdetector configurations were used. A 2.0 cm radius beam-pipe and a 3-layer silicon vertex detector were used for thefirst sample of 156 fb−1, while a 1.5 cm radius beampipe, a4-layer silicon detector and a small-cell inner drift chamberwere used to record the remaining 824 fb−1 [39].The selection of charged hadrons is based on information

from the tracking system (SVD and CDC) and hadronidentification system (CDC, ACC, and TOF). The chargedproton, kaon, and pion that are not associated with long-lived particles like K0

S, Λ and Ξ−, are required to have apoint of closest approach to the interaction point that iswithin 0.2 cm in the transverse (r-ϕ) direction and within2 cm along the z-axis. (The z-axis is opposite the positronbeam direction.) For each track, the likelihood values Lp,

TABLE I. Summary of the integrated luminosities and beam energies.ffiffiffis

pΥð5SÞ/near it Υð4SÞ/near it Υð3SÞ/near it Υð2SÞ/near it Υð1SÞ/near it

Integrated luminosity (fb−1) 121.0=29.3 702.6=89.5 2.9=0.3 24.9=1.8 5.7=1.8

SEARCH FOR DOUBLY CHARMED BARYONS AND STUDY … PHYSICAL REVIEW D 89, 052003 (2014)

052003-3

LK , and Lπ are provided for the assumption of proton,kaon, and pion, respectively, from the hadron identificationsystem, based on the ionization energy loss in the CDC, thenumber of detected Cherenkov photons in the ACC, and thetime of flight measured by the TOF. The likelihood ratio isdefined as Lði∶jÞ ¼ Li=ðLi þ LjÞ and a track is identifiedas a proton if the likelihood ratios Lðp∶πÞ and Lðp∶KÞ aregreater than 0.6. A track is identified as a kaon if thelikelihood ratios LðK∶πÞ and LðK∶pÞ are greater than 0.6.A track is identified as a pion if the likelihood ratiosLðπ∶KÞ and Lðπ∶pÞ are greater than 0.6. In addition,electron (Le) likelihood is provided based on informationfrom the ECL, ACC, and CDC [40]. A track with anelectron likelihood greater than 0.95 is rejected.The momentum averaged efficiencies of hadron identi-

fication are about 90%, 90%, and 93% for pions, kaons, andprotons, respectively. The momentum averaged probabilityto misidentify a pion (kaon) track as a kaon (pion) track isabout 9 (10)%, and the momentum averaged probability tomisidentify a pion or kaon track as a proton track isabout 5%.We use a Monte-Carlo (MC) simulation events generated

with EVTGEN [41], JETSET [42] with final QED finalstate radiation by PHOTOS [43] and then processed by aGEANT3 [44] based detector simulation to obtain thereconstruction efficiency and the mass resolution.

III. FINAL STATE WITH THE ΛþC

In this section, the analysis using the final states with theΛþc baryon is described. Reconstruction of the Λþ

c candi-date is explained first, followed by the description ofthe ΞþðþÞ

cc search in its decay into Λþc K−πþðπþÞ and the

study of two charmed strange baryons, Ξcð3055Þþ andΞcð3123Þþ. Throughout this paper, the selection criteria aredetermined to maximize the figure of merit (FOM), definedas ε=

ffiffiffiffiffiffiffiffiNbg

p, where ε is the Ξcc efficiency for the selection

criteria and Nbg is the number of background events underthe signal peak except for the scaled momentum selectionfor the Ξcð3055Þþ and Ξcð3123Þþ search, which followedBABAR’s analysis. The distribution of background events isestimated based on data. When the selection criteria aredetermined, we hide the possible signal peak by smearingthe invariant mass of the Ξcc candidates event by event witha Gaussian having a 50 MeV=c2 width in order to avoidany biases.

A. Reconstruction of the Λþc

The Λþc candidates are reconstructed in the pK−πþ and

pK0S decay modes [45]. The K0

S candidate is reconstructedfrom its decay into πþπ−. A pair of oppositely chargedpions that have an invariant mass within 8 MeV=c2 of thenominal K0

S mass, which corresponds to approximately3.5σ of the mass resolution, is used. Two pion tracks arefitted to a common vertex. The result of the fit is used to

suppress misreconstructed K0S candidates and to perform

further vertex fit of the Λþc → pK0

S. The vertex of the twopions for the K0

S is required to be displaced from theinteraction point (IP) in the direction of the pion pairmomentum [46]. The daughters of the Λþ

c are fitted to acommon vertex; the invariant mass of the daughters must bewithin 5ð6Þ MeV=c2, or 1.5σ, nominal Λþ

c mass for thepK−πþ (pK0

S) decay mode. The χ2 value of the commonvertex fit of the Λþ

c is required to be less than 50. For theremaining candidate, a mass constraint fit to the Λþ

c mass isperformed to improve the momentum resolution. As thesignal-to-background ratio for the Λþ

c candidates is similarfor the pK−πþ and pK0

S decay modes, they are combined inthe following analysis. By including the pK0

S mode inaddition to the pK−πþ mode, the yield of the Λþ

c isincreased by about 20%.

B. Search for doubly charmed baryons in Λþc K−πþðπþÞ

We search for the ΞþðþÞcc in its decay intoΛþ

c K−πþðπþÞ inthe mass range of 3.2–4.0 GeV=c2. The expected mass re-solution of the Ξcc estimated from MC is 2.0–3.5MeV=c2,depending on the mass of the Ξcc (degrading with in-creasing mass). In order to reduce the combinatorialbackground, a selection on the scaled momentum xp ¼p� ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

s=4 −m2p

is used, where p� is the CM momentum of aΞcc candidate and s is CM energy squared andm is mass ofthe Ξcc candidate. As there is no measurement of the xpspectrum for Ξcc production, we assume it to be the same asthat of the Λþ

c , which has been precisely measured [47].The xp spectrum is represented by a smooth polynomialfunction and is used to generate a MC sample for the Ξccsignal. Decays of the Ξcc and Λþ

c are generated accordingto the available phase space distribution. The number ofbackground events as a function of the xp cut is estimatedbased on smeared data. The FOM as a function of the xp cutis surveyed in the search region. The optimization pro-cedure yields 0.5 < xp < 1.0 regardless of the Ξcc mass.To check the validity of our analysis, we independentlyexamine the xp spectrum of the Λþ

c and confirm that it isconsistent with that presented in Ref. [47].Figure 1(a) and 1(b) show the MðΛþ

c K−πþÞ andMðΛþ

c K−πþπþÞ distributions, respectively, for data afterall the event selections are applied. No significant signal isseen in the data for either Ξþ

cc or Ξþþcc . The statistical

significance for a given mass is evaluated with an unbinnedextended maximum likelihood (UML) fit. The probabilitydensity function (PDF) for the signal is describedwith signalMCgenerated for eachgivenΞccmass,whereas a third-orderpolynomial function is used as the background PDF. Thestatistical significance is defined as

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi−2 ln ðL0=LÞ

p, where

L0 is the likelihood for the fit without the signal componentand L is the likelihood for the fit with the signal componentincluded. The significance is evaluated for the Ξcc mass

KATO et al. PHYSICAL REVIEW D 89, 052003 (2014)

052003-4

scanned with a 1 MeV=c2 step in the search region. Noneof the mass points give a significance exceeding the3σ level.The 95% confidence level (C.L.) upper limit for the

product of the cross section and branching fraction to theΛcK−πþðπþÞ state produced with the 0.5 < xp < 1.0condition,

σB ≡ σðeþe− → ΞþðþÞcc XÞ

is evaluated. Here, L is the total integrated luminosity, Nsig

is the Ξcc signal yield, BpK−πþ is the branching fraction oftheΛþ

c → pK−πþ (which amounts to 0.050� 0.013), BpKS

is the branching fraction of Λþc → pK0

S measured relativeto the pK−πþ mode (BpKS

=BpK−πþ ¼ 0.24� 0.02), andεpK−πþðpKSÞ is the reconstruction efficiency for the Λþ

c →pK−πþ (Λþ

c → pK0S) decay mode evaluated as a function

of the Ξcc mass. The efficiencies for the ΞþðþÞcc as a function

of their masses are shown in Fig. 2. The factor of two in thedenominator comes from inclusion of the charge-conjugatemode. By including this factor, our measurement can becompared with the theoretical predictions [28,29]; while tocompare with the prediction in Ref. [30], it is necessary tomultiply our σB measurement by 2 because they predictedthe cross section of the pair production of the cc and c̄ c̄diquarks. In BABAR’s measurement [34], they do notintroduce the factor of two, i.e., they report an upper limit

for the sum of the σðeþe− → ΞþðþÞcc XÞ and its charge-

conjugate mode). Therefore, our measurement should bedoubled when comparing with BABAR’s result. We notethat the cross section reported here and elsewhere in thispaper is a visible cross section (i.e., a radiative correction isnot applied.The upper limit is evaluated following the Bayesian

approach. First, we scan the likelihood profile by determin-ing the likelihood values as a function of the σB (LðσBÞ),

M(Λc+K-π+) (GeV/c2)

0

200

400

600

800

1000

1200 (a) (b)

(c) (d)

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

Eve

nts/

(0.0

01G

eV/c

2 )

M(Λc+K-π+π+) (GeV/c2)

0

200

400

600

800

1000

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

Eve

nts/

(0.0

01G

eV/c

2 )

0

5

10

15

20

25

30

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

Mass (GeV/c2)

σ B (

fb)

0

5

10

15

20

25

30

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

Mass (GeV/c2)

σ B (

fb)

FIG. 1 (color online). Invariant mass distribution of the Ξcc candidates for (a)MðΛþc K−πþÞ, (b)MðΛþ

c K−πþπþÞ; the vertical error barsare from data and the dashed histogram are from signal MC for the Ξcc signal generated with a mass of 3.60 GeV=c2 and a productioncross section σðeþe− → ΞþðþÞ

cc XÞ of 500 fb and BðΞþðþÞcc → Λþ

c K−πþðπþÞÞ of 5%. 95% C.L. upper limit of σB as a function of the masswith a 1 MeV=c2 step for (c) Ξþ

cc and (d) Ξþþcc .

SEARCH FOR DOUBLY CHARMED BARYONS AND STUDY … PHYSICAL REVIEW D 89, 052003 (2014)

052003-5

varying Nsig from zero up to the Nsig value for which thelikelihood drops to zero. Then, LðσBÞ is convolved with aGaussian whose width equals the systematic uncertaintiesof σB. The σB value for which the integral (starting fromσB ¼ 0) becomes 95% of the entire area is regarded as the95% C.L. upper limit.We consider the following systematic uncertainties in the

Ξcc search. The systematic uncertainty due to the efficiencyof pion and kaon identification is estimated from the ratioof the yield of the D�þ → D0πþ, D0 → K−πþ with andwithout the pion/kaon identification requirements for dataandMC. The difference of the ratio between data andMC iscorrected and the statistical error of the ratio is regarded asthe systematic uncertainty. The systematic uncertainty dueto the efficiency of proton identification is estimated usingthe ratio of the yield of the Λ → pπ− with and without theproton identification requirement. The difference of theratio between data and MC is corrected and the statisticalerror of the ratio is regarded as the systematic uncertainty.The systematic uncertainty due to the charged trackreconstruction efficiency is estimated using the decay chainD�þ → πþD0, D0 → πþπ−K0

S, and K0S → πþπ−, where

K0S → πþπ− is either partially or fully reconstructed. The

ratio of the yields for partially and fully reconstructed

signals in data and MC is compared, and the difference istaken as the systematic uncertainty. This amounts to 0.35%per track. The systematic uncertainty of the total integratedluminosity is 1.4%. To check the systematic error due to thesignal PDF, we compare the mass resolution of the Λþ

c indata and MC. We find that the resolution for data is 5%larger than in MC. To monitor the effect of this discrepancy,we perform a pseudoexperiment test in which we extractthe signal yield with correct PDF and one that is narrowerby 5%. The largest difference of 3% measured in this test isregarded as the systematic uncertainty. The systematicuncertainty related to the Λþ

c branching fraction is propa-gated from the errors taken from the PDG [9]. To estimatethe systematic uncertainty of the reconstruction efficiencydue to the possible difference of xp spectrum between ourassumption (the same as that of Λþ

c ) and the actual one, weexamine the xp dependence of the reconstruction effi-ciency. The root mean square of the reconstruction effi-ciency in the region of 0.5 < xp < 1.0 is regarded as thesystematic uncertainty. The elements of the systematicuncertainty for the measurement of the σB are enumeratedin the first and second columns of Table II.Figures 1(c) and 1(d) show the 95% C.L. upper limit on

σB for Ξþcc and Ξþþ

cc , respectively, as a function of the mass

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2(a) (b)

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9

Mass (GeV/c2)

Effi

cien

cy

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9

Mass (GeV/c2)

Effi

cien

cy

FIG. 2. Reconstruction efficiency as a function of the Ξcc mass, for (a) Ξþcc, (b) Ξþþ

cc . Circular points are for Λþc → pK−πþ and square

points are for Λþc → pK0

S. The lines show the result of the fit with a linear function.

TABLE II. Summary of the systematic uncertainties in the σB and σB2 measurement (%).

Source Ξþcc with Λþ

c Ξþþcc with Λþ

c Ξcð3123Þþ Ξþcc with Ξ0

c Ξþþcc with Ξ0

c

Particle ID 2.0 2.4 2.0 3.5 3.8Tracking 1.8 2.1 1.8 1.8 2.1Signal PDF 3.5 3.5 28.0 3.5 3.5Luminosity 1.4 1.4 1.4 1.4 1.4B 26.0 26.0 1.6 0.7 0.7xp 2.1 2.3 1.2 6.0 5.7Ni

Ξcð2645Þþ … … … 4.3 4.3total 26.5 26.6 28.2 9.2 9.2

KATO et al. PHYSICAL REVIEW D 89, 052003 (2014)

052003-6

with a 1 MeV=c2 step. The upper limit is in the range of4.1–25.0 fb for the Ξþ

cc and 2.5–26.5 fb for the Ξþþcc .

C. Search for the Ξþc ð3055Þ and Ξþ

c ð3123ÞIn this section, a search for the Ξþ

c ð3055Þ and Ξþc ð3123Þ

is described. Here, we require xp to be greater than 0.7. Inthe analysis by BABAR [6], they required p�ðΛþ

c K−πþÞ >2.9 GeV=c, which is similar to our xp cut as illustrated bythe p�ðΛþ

c K−πþÞ distribution, with the xp cut and2.9 GeV=c2 < MðΛþ

c K−πþÞ < 3.2 GeV=c2 required asshown in Fig. 3(a). Figure 3(b) shows the MðΛþ

c πþÞ

distribution, where contributions from the Σcð2455Þþþand the Σcð2520Þþþ baryons are clearly visible. We selectthe Σcð2455Þþþ [Σcð2520Þþþ] region by requiring

jMðΛþc π

þÞ −mΣþþcj < 5ð18Þ MeV=c2, where mΣþþ

cis the

nominal mass of the Σcð2455Þþþ or Σcð2520Þþþ.Figure 3(c) shows the MðΛþ

c K−πþÞ distribution forthe Σcð2455Þþþ signal region together with the sameplot for the Σcð2455Þþþ sideband region, defined asjMðΛþ

c πþÞ − ðmΣcð2455Þþþ � 15 MeV=c2Þj < 5 MeV=c2.

Clear peaks corresponding to the Ξcð2980Þþ, Ξcð3055Þþ,and Ξcð3080Þþ are seen. To obtain the statistical signifi-cance of the Ξcð3055Þþ, an UML fit is applied. PDFs forthe Ξ�þ

c components are represented by a Breit-Wignerline shape convolved with a Gaussian to account forthe invariant-mass resolution (σres). Using the signal MCevents, we estimate σres to vary from 1.2 to 1.8 MeV=c2,depending on the masses of the Ξ�þ

c states. The width andmean of the Breit-Wigner functions are treated as free

p*(Λc+K-π+)(GeV/c)

0

100

200

300

400

500

600

2.5 3 3.5 4 4.5 5

Eve

nts/

(0.0

25 G

eV/c

)

M(Λc+π+) (GeV/c2)

0

100

200

300

400

500

600

700

800

2.42 2.44 2.46 2.48 2.5 2.52 2.54 2.56 2.58 2.6

500E

vens

/(0.

001G

eV/c

2 )

(a) (b)

)2) (GeV/c+π-K+cΛM(

2.95 3 3.05 3.1 3.15 3.2

)2E

vent

s / (

0.0

04 G

eV/c

0

20

40

60

80

100

120

140 (c)

)2) (GeV/c+π-K+cΛM(

2.95 3 3.05 3.1 3.15 3.2

)2E

vent

s / (

0.0

04 G

eV/c

0

20

40

60

80

100

120

140

160

180

200

220

(d)

FIG. 3 (color online). (a) The p�ðΛþc K−πþÞ distribution from data. (b) TheMðΛþ

c πþÞ distribution. The vertical lines show the selected

regions of the Σcð2455Þþþ and Σcð2520Þþþ. (c) The MðΛþc K−πþÞ distribution with Σcð2455Þþþ selection. The dots with error bars

show the distribution for the Σcð2455Þþþ selected region whereas the rectangles show the distribution for the Σcð2455Þþþ sidebandregion. The solid line shows the fit result. The dashed, dotted, and dash-dotted lines show the contributions from the background,Ξcð3055Þþ, and Ξcð2980Þþ, or Ξcð3080Þþ, respectively. (d) The MðΛþ

c K−πþÞ distribution with Σcð2520Þþþ selection. The dots witherror bars show the distribution for Σcð2520Þþþ selected region whereas the rectangles show the distribution for the Σcð2520Þþþsideband region. The solid line shows the fit result. The dashed, dotted, and dash-dotted lines show the contributions from thebackground, Ξcð3123Þþ, and Ξcð3080Þþ, respectively.

SEARCH FOR DOUBLY CHARMED BARYONS AND STUDY … PHYSICAL REVIEW D 89, 052003 (2014)

052003-7

parameters. The background PDF, f1ðxÞ, is modeled with athreshold function:

f1ðxÞ ¼ 1 − expððx − x0Þ=δmÞðx=x0Þaþbðx=x0 − 1Þ ðif x > x0Þ

f1ðxÞ ¼ 0 ðif x < x0Þ; (1)

where a, b, x0, and δm are free parameters in the fit.The fit result is shown in Fig. 3(c). To estimate the

statistical significance of the Ξcð3055Þþ, we comparethe likelihood values for the fits with and without theΞcð3055Þþ component. The obtained −2 ln ðL0=LÞ value is54.7. By taking into account the change of the numberof degrees of freedom (ndf) by the inclusion of theΞcð3055Þþ component, the statistical significance of theΞcð3055Þþ becomes 6.8σ. The χ2=ndf of the fit withthe Ξcð3055Þþ component, for the binning of Fig. 3(c), is54.8/61.Figure 3(d) shows the MðΛþ

c K−πþÞ distribution forthe Σcð2520Þþþ selected region together with the sameplot for the Σcð2520Þþþ sideband region, defined asjMðΛþ

c πþÞ − ðmΣcð2520Þþþ � 27 MeV=c2Þj < 12 MeV=c2.

A clear peak corresponding to the Ξcð3080Þþ is seen, whileno peak structure is seen in the mass near 3.123 GeV=c2.An UML fit is applied to extract the signal yield. Again,the Ξ�þ

c components are represented by a Breit-Wignerfunction convolved with a Gaussian. For the Ξcð3080Þþcomponent, the mass and width of the Breit-Wignerare treated as free parameters; while for the Ξcð3123Þþcomponent, the mass and width are fixed to the valuesobtained in Ref.[6]. The background shape, f2ðxÞ, isassumed to be:

f2ðxÞ ¼ ð1 − expððx − x0Þ2=δmÞÞðx=x0Þaþbðx=x0 − 1Þ þ cððx=x0Þ2 − 1Þ ðif x > x0Þ

f2ðxÞ ¼ 0 ðif x < x0Þ; (2)

where a, b, c, x0, and δm are fit parameters. The χ2=ndfof the fit with the Ξcð3123Þþ component for the binningof Fig. 3(d) is 28.6=42. The yield of the Ξcð3123Þþ is8� 22 events, which is consistent with zero. Hence, a95% C.L. upper limit for the production cross section isevaluated with the method described in the previoussection. To directly compare with the BABAR result inRef. [6], the upper limit for the product of the crosssection and branching fraction ofΛþ

c producedwith xp > 0.7condition,

σBΛþc≡ σðeþe− → Ξcð3123ÞþXÞ × BðΛþ

c → pK−πþÞ

is evaluated. As in Ref. [6], we assume BðΞcð3123Þþ →Σcð2520ÞþþK−Þ is equal to 1.

To take the uncertainty of the Ξcð3123Þþ mass andwidth from Ref. [6] into account, we perform a pseu-doexperiment test. The background and Ξcð3080Þþ con-tributions are generated with statistics similar to dataand based on the fit result. The Ξcð3123Þþ component isgenerated with mass and width changed by �1σ, corre-sponding to their measured uncertainties (3122.9� 1.3�0.4 MeV=c2 for the mass and 4.4� 3.4� 1.7 MeV=c2

for the width). The yield of Ξcð3123Þþ is extracted byfitting pseudoexperiment data with the procedure usedfor data. The ratio of the generated and extracted yield isregarded as a systematic uncertainty. Because the errorof the width is relatively large, its systematic uncertaintycontribution is dominant (28%). All of the systematicerrors are summarized in the third column of Table II.The 95% C.L. upper limit on σBΛþ

cis 0.17 fb. As in the

case of the Ξcc cross section, the measurement in Ref. [6]does not introduce a factor of two for the σBΛþ

ccalcu-

lation. Therefore, we should double our measurement,which results in 0.34 fb, when comparing with BABAR’sresult. The value is much smaller than that quoted inRef. [6] (1.6� 0.6� 0.2 fb).The systematic uncertainties of the masses and widths

of the Ξ�þc and stability of the statistical significance of

the Ξcð3055Þþ are studied by the following fittingconfigurations. The systematic uncertainties due to thesignal PDF are studied by varying σres by 5%. Thesystematic uncertainties due to possible interferencebetween the Ξcð3055Þþ and Ξcð3080Þþ are studied byfitting the distribution with an additional phase parameterbetween the two Breit-Wigner amplitudes. The system-atic uncertainty due to the background shape is studiedby fitting the mass spectra with a second-order poly-nomial as a background PDF in the range of3.005–3.200 GeV=c2. In none of these fitting configu-rations does the statistical significance of the Ξcð3055Þþfall below 6.6σ. We apply cut conditions of xp > 0.6 andxp > 0.8 instead of xp > 0.7 and reextract the masses andwidths of the Ξ�þ

c states. The differences from the defaultcut condition are regarded as systematic uncertainties.The measured masses, widths, and yields of the three Ξ�þ

cstates are summarized in Table III. All of these mea-surements are consistent with previous Belle measure-ment [2] within 2.5σ and with the BABAR measurement[6] within 2.0σ.

TABLE III. The measured masses and widths of the three Ξ�þc

states. The first error is statistical and second is systematic.

Particle Mass (MeV=c2) Width (MeV=c2) Yield

Ξcð2980Þþ 2974.9� 1.5� 2.1 14.8� 2.5� 4.1 244� 39Ξcð3055Þþ 3058.1� 1.0� 2.1 9.7� 3.4� 3.3 199� 46Ξcð3080Þþ (Σc) 3077.9� 0.4� 0.7 3.2� 1.3� 1.3 185� 31Ξcð3080Þþ (Σ�

c) 3076.9� 0.3� 0.2 2.4� 0.9� 1.6 210� 30

KATO et al. PHYSICAL REVIEW D 89, 052003 (2014)

052003-8

IV. FINAL STATE WITH Ξ0c

In this section, the analysis of the final state with the Ξ0c

is described. The reconstruction of the Ξ0c is presented

first, followed by the analysis of the Ξcð2645Þþ. Finally,a search for ΞþðþÞ

cc decaying into the Ξ0cπ

þðπþÞ final state isdescribed.

A. Reconstruction of Ξ0c

The Ξ0c is reconstructed in three decay modes: Ξ−πþ,

ΛK−πþ, and pK−K−πþ. The Λ is reconstructed from itsdecay intopπ−. The proton andπ− tracks forΛ candidates arefitted to a commonvertex.The fitting result is used to suppressmisreconstructedΛ candidates and to perform the subsequentvertex fit for the Ξ− → Λπ− or Ξ0

c → ΛK−πþ. The invariantmass of theΛ candidate is required to bewithin 3 MeV=c2 ofthe nominal Λmass, which corresponds to approximately 3σof the mass resolution. The selection based on their decayvertex information is also applied [48]. The Ξ− is recon-structed from its decay into Λπ−. The Λ and π− tracks for Ξ−

candidates are fitted to a common vertex. The fitting result isused to clean up the Ξ− candidates and in the commonvertexfit for the Ξ0

c → Ξ−πþ. The closest distance of the Λ and π−

along the z-direction is required to be less than 3 mm. Werequire cos θ > 0.95, where θ is the angle between themomentum vector of the Ξ− and the vector between the IPand the Ξ− decay vertex. The χ2 of the common-vertex fit oftheΛπ− is required to be less than 50. The invariant mass of aΞ− candidate is required to be within 4 MeV=c2 of thenominal Ξ− mass, which corresponds to approximately 3σ ofthemass resolution. The daughter particles of theΞ0

c are fittedto a common vertex. The Ξ0

c candidates are selected byrequiring invariant masses of the daughter particles withcommon vertex fit to be within 12, 7, and 7 MeV=c2 ofthe nominal Ξ0

c mass for the Ξ−πþ, ΛK−πþ, and pK−K−πþdecay modes, respectively, which correspond to approxi-mately 1.5σ of the mass resolution. The χ2 value of the

commonvertex fit for theΞ0c is required to be less than 50.The

mass constraint fit to the Ξ0c mass is performed.

We optimize the selection criteria for xp in the Ξ0cπ

þðπþÞsystem with the method described in Sec. III B, againassuming that the xp spectrum for the Ξcc is the same as thatfor the Λþ

c . We require 0.45 < xp < 1.0 independent of theΞcc mass and the Ξ0

c decay mode. The same cut is appliedfor the analysis of the Ξcð2645Þþ.

B. Study of the Ξþc ð2645Þ

Unlike the Λþc study, the signal-to-background ratio of

the Ξ0c largely depends on the decay modes. Therefore, to

improve our sensitivity for the Ξcc, we perform a simulta-neous fit to the mass spectra for the three Ξ0

c decays withfixed relative signal ratios. We use the relative yields of theΞcð2645Þþ → Ξ0

cπþ measured for the Ξ0

c decay modes toestimate a relative signal yield of the Ξcc. The relativesignal yields of the Ξcc (Ni

Ξcc) in a given Ξ0

c decay channelcan be written as

NiΞcc

¼ NiΞcð2645Þþ

ϵiΞcc

ϵiΞcð2645Þþ; (3)

where NiΞcð2645Þþ is the Ξcð2645Þþ yield, ϵiΞcc

is the recon-struction efficiency of the Ξcc, and ϵiΞcð2645Þþ is thereconstruction efficiency of theΞcð2645Þþ. Both efficienciesinclude the secondary branching fractions for Λ → pπ− ofð63.9� 0.5Þ% and Ξ− → Λπ− of ð99.887� 0.035Þ%. Theindex i denotes the decay mode of the Ξ0

c.Figure 4 shows the MðΞ0

cπþ) distribution for each Ξ0

cdecay mode below the Ξcc search region. Clear peakscorresponding to the Ξcð2645Þþ are seen in all decaymodes. The bump structures near 2.68 GeV=c2 originatefrom the process Ξcð2790Þþ → Ξc

00πþ → Ξ0cγπ

þ with a γmissing in the reconstruction. The simultaneous UML fit isapplied to extract the relative yields and the width of theΞcð2645Þþ. The Ξcð2645Þþ signal is represented by a

)2) (GeV/c+π0cΞM(

2.6 2.62 2.64 2.66 2.68 2.7 2.72 2.74

)2E

vent

s / (

0.0

015

GeV

/c

0

50

100

150

200

250

300

350

400 (a)

)2) (GeV/c+π0cΞM(

2.6 2.62 2.64 2.66 2.68 2.7 2.72 2.74

)2E

vent

s / (

0.0

015

GeV

/c

0

100

200

300

400

500 (b)

)2) (GeV/c+π0cΞM(

2.6 2.62 2.64 2.66 2.68 2.7 2.72 2.74

)2E

vent

s / (

0.0

015

GeV

/c

0

50

100

150

200

250

300

350

400 (c)

FIG. 4 (color online). MðΞ0cπ

þÞ distributions below the Ξcc search region for (a) Ξ0c → Ξ−πþ, (b) Ξ0

c → ΛK−πþ, (c) Ξ0c → pK−K−πþ.

The solid lines show the fit result. The dashed, dotted, and dash-dotted lines show the contributions from background, Ξcð2645Þþ, andΞcð2790Þþ, respectively.

SEARCH FOR DOUBLY CHARMED BARYONS AND STUDY … PHYSICAL REVIEW D 89, 052003 (2014)

052003-9

Breit-Wigner function convolved with a Gaussian whosewidth corresponds to the mass resolution σres. The value ofσres is 1.05 MeV=c2, independent of the decay modes ofthe Ξ0

c. The PDF of the Ξcð2790Þþ reflection is modeledusing MC. f2ðxÞ in Eq. (2) is used as the background PDFfor the Ξ0

c → Ξ−πþ decay mode whereas f1ðxÞ in Eq. (1) isused for the Ξ0

c → ΛK−πþ and Ξ0c → pK−K−πþ decay

modes. The width and mass of the Ξcð2645Þþ are con-strained to be the same for the three decay modes. The yieldof the Ξcð2645Þþ is 1298� 51, 1444� 58, and 974� 47for the Ξ−πþ, ΛK−πþ, and pK−K−πþ decay mode,respectively. The mass and width are obtained to be2645.4� 0.1 MeV=c2 and 2.6� 0.2 MeV=c2, respec-tively. The χ2=ndf of the fit for the binning in Fig. 4 is296/276.To check the consistency of the width measurement

between the Ξ0c decay modes, we fit the three mass spectra

separately. The measured widths are found to be consistentbetween the three decay modes: 2.9� 0.3 MeV=c2,2.6�0.3MeV=c2, and 2.5�0.3MeV=c2 for Ξ0

c→Ξ−πþ,ΛK−πþ, and pK−K−πþ, respectively. The measured widthis found to be consistent for the three decay modes. Thesystematic uncertainty of the width measurement due to thefit procedure is studied with pseudoexperiment eventssamples: the Ξcð2645Þþ component is generated accordingto the signal MC sample with the natural width of2.6 MeV=c2 by signal MC, while contributions from back-ground and Ξcð2790Þþ reflection are generated based on thefit result with the real data. The statistics of the pseudoexperi-ment samples are the same as for those of data. The widthof the Ξcð2645Þþ is extracted from simultaneous fits topseudoexperiment samples, and its mean value is obtained tobe 2.75� 0.03 MeV=c2, which is higher than the inputvalue by 0.15 MeV=c2. The difference is included as thesystematic uncertainty from the fit procedure. The systematicuncertainty due to the background shape is studied by fittingthe data with a second-order polynomial function forthe alternative background shape. The fit region is restrictedto 2.62–2.75 GeV=c2. The width is obtained to be2.9� 0.2 MeV=c2, which is 0.3 MeV=c2 higher than thedefault measurement. This deviation is included as asystematic uncertainty. To check the systematic uncertaintydue to the mass resolution, we evaluate the ratio of theresolution of the Ξ0

c in the data and MC, σdataσmc, where σdata is

the resolution of Ξ0c for data and σmc is that for MC. An

additional cut of 2.64GeV=c2<MðΞ0cπ

þÞ<2.65GeV=c2

is applied to select the Ξcð2645Þþ region only. σdataσmc

is0.96� 0.03, 1.03� 0.07, and 1.07� 0.03 for the Ξ−πþ,ΛK−πþ, and pK−K−πþ decay modes, respectively. Weassign the systematic uncertainty conservatively by increas-ing the mass resolution by 7%, which is the largest deviationof the σdata

σmcfrom unity, for all the decay modes. The resulting

width is 2.5� 0.2 MeV=c2. The difference of 0.1 MeV=c2

with respect to the default measurement is included as a

systematic uncertainty. We use alternative xp ranges of0.35 < xp < 1.0 and 0.55 < xp < 1.0 and extract the widthof 2.6� 0.2 MeV=c2 for the former case and 2.8�0.2 MeV=c2 for the latter. The largest difference of0.2 MeV=c2 with respect to the default measurement isincluded as a systematic uncertainty. By adding all thesystematic uncertainties in quadrature, the total systematicuncertainty for the width measurement is estimated tobe 0.4 MeV=c2.

C. Search for doubly charmed baryonsin the Ξ0

cπþðπþÞ final stateTo obtain the relative yields of the Ξcc, εiΞcc

and εiΞcð2645þÞare evaluated using MC, and the efficiency εiΞcc

is obtainedas a function of the Ξcc mass. Ni

Ξcð2645Þþ and εiΞcð2645þÞ aresummarized in Table IV, while εiΞþ

ccas a function of Ξcc

mass is shown in Fig. 5. As an example, the relativeyield ratio of the Ξþ

cc with a mass of 3.6 GeV=c2

is NΞ−πþΞcc

∶NΛK−πþΞcc

∶NpK−K−πþΞcc

¼ 1∶1.15∶0.84.Figures 6 and 7(a)–(c) show theMðΞ0

cπþðπþÞÞ distribution

in the Ξcc search region with all the selection cuts applied,overlaid with the MC expectation for the Ξcc at the mass of

3.60GeV=c2 with σðeþe−→ΞþðþÞcc XÞ of 500 fb and both

BðΞþðþÞcc →Ξ0

cπþðπþÞÞ and BðΞ0

c→Ξ−πþÞ of 5%. The rela-tive yields of signal MC for each decay mode are basedon Ni

Ξcc.

A simultaneous UML fit, with the relative Ξcc yieldsconstrained as discussed earlier, is applied to evaluate thestatistical significance of the Ξcc. The signal PDF isdescribed with MC events generated for each decay modeand with the Ξcc mass generated in the search region with a1 MeV=c2 step. The mass resolution is 2.7–4.2 MeV=c2,depending on the mass of the Ξcc. The background PDF ismodeled as a third-order polynomial. The highest signifi-cance is 3.2σ for the mass around 3.553 GeV=c2 in theMðΞ0

cπþÞ. We perform a pseudoexperiment test to evaluate

the probability of observing a peak with such a statisticalsignificance. A smooth mass distribution based on data isgenerated and the significance is evaluated in the entiresearch region. The probability to observe a peak with asignificance higher than 3.2σ in one pseudoexperiment is26%. Therefore, the statistical significance of 3.2σ isinsufficient to claim evidence of the Ξþ

cc.

TABLE IV. Ξcð2645Þþ yields and efficiencies used for esti-mation of the relative Ξcc yields.

Decay mode NΞcð2645Þþ εΞcð2645Þþ

Ξ−πþ 1298� 51 0.0748� 0.0002ΛK−πþ 1444� 58 0.0977� 0.0003pK−K−πþ 974� 47 0.1920� 0.0005

KATO et al. PHYSICAL REVIEW D 89, 052003 (2014)

052003-10

0

10

20

30

40

50

60

70

80 (a) (b)

(c) (d)

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

M(Ξc0π+) (GeV/c2)

Eve

nts/

(0.0

02G

eV/c

2 )

0

25

50

75

100

125

150

175

200

225

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

M(Ξc0π+) (GeV/c2)

Eve

nts/

(0.0

02G

eV/c

2 )

0

20

40

60

80

100

120

140

160

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

M(Ξc0π+) (GeV/c2)

Eve

nts/

(0.0

02G

eV/c

2 )

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

Mass (GeV/c2)

σ B (

fb)

2

FIG. 6 (color online). (a)–(c): MðΞ0cπ

þÞ distribution in the Ξþcc search region for Ξ0

c → (a) Ξ−πþ, (b) ΛK−πþ, (c) pK−K−πþ. Thevertical error bars are from data. The dashed histograms are from signal MC. (d): 95% C.L. upper limit of the σB2 for Ξþ

cc as a function ofthe mass with a 1 MeV=c2 step.

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18(a) (b)

0.2

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

Mass (GeV/c2)

Effi

cien

cy

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

Mass (GeV/c2)

Effi

cien

cy

FIG. 5. Reconstruction efficiencies for the Ξcc as a function of the Ξcc mass for (a) Ξþcc, (b) Ξþþ

cc . Circles, square, and triangle points arefor Ξ0

c → Ξ−πþ, ΛK−πþ, and pK−K−πþ, respectively. The lines are the result of the fit with a linear function.

SEARCH FOR DOUBLY CHARMED BARYONS AND STUDY … PHYSICAL REVIEW D 89, 052003 (2014)

052003-11

The 95% C.L. upper limit for the product of the cross section and branching fractions produced with 0.45 < xp < 1.0condition,

σB2 ≡ σðeþe− → ΞþðþÞcc XÞ × BðΞþðþÞ

cc → Ξ0cπ

þðπþÞÞ × BðΞ0c → Ξ−πþÞ ¼ Nsig

2L × εΞ−πþ

Ξcc

�1þ NΛK−πþ

Ξcc

NΞ−πþΞcc

þ NpK−K−πþΞcc

NΞ−πþΞcc

� ;

is evaluated with the same method as in Sec. III B. Inaddition to the sources from the study with the Λþ

c , twoothers are included here. The systematic uncertainty fromthe Λ reconstruction efficiency is estimated to be 3% usingthe yield of the Bþ → ΛΛ̄Kþ with and without the require-ment using decay vertex information. The systematicuncertainties related to Ni

Ξcð2645Þþ are taken from theirstatistical errors. The systematic uncertainties are summa-rized in the fourth column in Table II. Figures 6 and 7(d)show σB2 for the ΞþðþÞ

cc as a function of the mass with a1 MeV=c2 step. The 95% C.L. upper limit on σB2 is 0.076–0.35 fb for the Ξþ

cc and 0.082–0.40 fb for the Ξþþcc .

V. CONCLUSION

We have presented a search for doubly charmed baryonsand a study of the charmed strange baryons Ξcð3055Þþ,Ξcð3123Þþ and Ξcð2645Þþ using the full data sample(980 fb−1) collected with the Belle detector. The searchfor doubly charmed baryons is an improved study of ourprevious work [2]. We use about two times statistics ofprevious work and several additional decay modes thatwere not studied in the previous work.We search for the Ξcc in the Λþ

c K−πþðπþÞ andΞ0cπ

þðπþÞ final states. The Λþc is reconstructed from the

pK−πþ and pK0S decay modes. We do not find any

significant Ξcc signal and set a 95% C.L. upper limit on

0

10

20

30

40

50

60

70

80

90(a) (b)

(c) (d)

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

M(Ξc0π+π+) (GeV/c2)

Eve

nts/

(0.0

02G

eV/c

2 )

0

50

100

150

200

250

300

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

M(Ξc0π+π+) (GeV/c2)

Eve

nts/

(0.0

02G

eV/c

2 )

0

20

40

60

80

100

120

140

3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4

M(Ξc0π+π+) (GeV/c2)

Eve

nts/

(0.0

02G

eV/c

2 )

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

Mass (GeV/c2)

σ B2 (

fb)

FIG. 7 (color online). (a)–(c): MðΞ0cπ

þπþÞ distribution in the Ξþþcc search region for Ξ0

c → (a) Ξ−πþ, (b) ΛK−πþ, (c) pK−K−πþ. Thevertical error bars are from data. The dashed histograms are from signal MC. (d): 95% C.L. upper limit of the σB2 for Ξþþ

cc as a functionof the mass with a 1 MeV=c2 step.

KATO et al. PHYSICAL REVIEW D 89, 052003 (2014)

052003-12

σðeþe− → ΞþðþÞcc XÞ × BðΞþðþÞ

cc → Λþc K−πþðπþÞÞ with the

scaled momentum 0.5 < xp < 1.0: 4.1–25.0 fb for Ξþcc and

2.5–26.5 fb for Ξþþcc . We also search for the Ξcc in the

Ξ0cπ

þðπþÞ final state. The Ξ0c is reconstructed from the

Ξ−πþ, ΛK−πþ, and pK−K−πþ decay modes. We do notfind any significant Ξcc signal and set a 95% C.L. upperlimit on σðeþe− → ΞþðþÞ

cc XÞ × BðΞþðþÞcc → Ξ0

cπþðπþÞÞ×

BðΞ0c → Ξ−πþÞ with the scaled momentum 0.45 <

xp < 1.0: 0.076–0.35 fb for the Ξþcc and 0.082–0.40 fb

for the Ξþþcc . When we compare these values with the

measurements by BABAR, we should note several things.We should multiply our result by two as written in Sec. III.For the final states with Λþ

c , their upper limit is for theproduct of the cross section, BðΞþðþÞ

cc → Λþc K−πþðπþÞÞ

and BðΛþc → pK−πþÞ. The values presented in Ref. [34]

are the highest upper limits in the search region, which canbe compared with our highest values. After taking intoaccount these points, we find our limits represent improve-ments by about a factor two for the final states with Λþ

c anda factor of four for the final states with Ξ0

c.If we assume BðΞþðþÞ

cc → Λþc K−πþðπþÞÞ, BðΞþðþÞ

cc →Ξ0cπ

þðπþÞÞ, and BðΞ0c → Ξ−πþÞ to be 5%, which is equal

to the BðΛþc → pK−πþÞ, the upper limits on the σðeþe− →

ΞþccXÞ are 82–500 fb (Ξþ

cc) and 50–530 fb (Ξþþcc ) for

the decay mode with the Λþc and 30–140 fb (Ξþ

cc) and33–160 fb (Ξþþ

cc ) for the decay mode with the Ξ0c. These

values are comparable to some of the theoretical predic-tions [28,29].We have searched for the Ξcð3055Þþ and Ξcð3123Þþ

in the Λþc K−πþ decays through intermediate Σcð2455Þþþ

or Σcð2520Þþþ states. We observe the Ξcð3055Þþ with asignificance of 6.6σ, including systematic uncertainty. Themass and width are measured to be 3058.1� 1.0ðstatÞ �2.1ðsysÞ MeV=c2 and 9.7� 3.4ðstatÞ � 3.3ðsysÞ MeV=c2,respectively. We do not observe any significant signalcorresponding to the Ξcð3123Þþ.The first measurement of the width of the Ξcð2645Þþ

has been also performed, yielding 2.6� 0.2ðstatÞ�0.4ðsysÞ MeV=c2.

ACKNOWLEDGMENTS

We thank the KEKB group for the excellent operation ofthe accelerator; the KEK cryogenics group for the efficient

operation of the solenoid; and the KEK computer group,the National Institute of Informatics, and the PNNL/EMSLcomputing group for valuable computing and SINET4network support. We acknowledge support from theMinistry of Education, Culture, Sports, Science, andTechnology (MEXT) of Japan, the Japan Society for thePromotion of Science (JSPS), and the Tau-Lepton PhysicsResearch Center of Nagoya University; the AustralianResearch Council and the Australian Department ofIndustry, Innovation, Science and Research; AustrianScience Fund under Grant No. P 22742-N16; theNational Natural Science Foundation of China underContracts No. 10575109, No. 10775142, No. 10825524,No. 10875115, No. 10935008, and No. 11175187; theMinistry of Education, Youth and Sports of the CzechRepublic under Contract No. MSM0021620859; the CarlZeiss Foundation, the Deutsche Forschungsgemeinschaftand the VolkswagenStiftung; the Department of Scienceand Technology of India; the Istituto Nazionale diFisica Nucleare of Italy; The WCU program of theMinistry Education Science and Technology, NationalResearch Foundation of Korea Grants No. 2011-0029457, No. 2012-0008143, No. 2012R1A1A2008330,No. 2013R1A1A3007772, BRL program under NRF GrantNo. KRF-2011-0020333, BK21 Plus program, and GSDCof the Korea Institute of Science and TechnologyInformation; the Polish Ministry of Science and HigherEducation and the National Science Center; the Ministryof Education and Science of the Russian Federation andthe Russian Federal Agency for Atomic Energy; theSlovenian Research Agency; the Basque Foundation forScience (IKERBASQUE) and the UPV/EHU under pro-gram UFI 11/55; the Swiss National Science Foundation;the National Science Council and the Ministry ofEducation of Taiwan; and the U.S. Department ofEnergy and the National Science Foundation. This workis supported by a Grant-in-Aid from MEXT for ScienceResearch in a Priority Area (“New Development ofFlavor Physics”), from JSPS for Creative ScientificResearch (“Evolution of Tau-lepton Physics”), andGrant-in-Aid for Scientific Research on InnovativeAreas “Elucidation of New Hadrons with a Variety ofFlavors”.

[1] R. Mizuk et al. (Belle Collaboration), Phys. Rev. Lett. 94,122002 (2005).

[2] R. Chistov et al. (Belle Collaboration), Phys. Rev. Lett. 97,162001 (2006).

[3] B. Aubert et al. (BABAR Collaboration), Phys. Rev. Lett. 97,232001 (2006).

[4] B. Aubert et al. (BABAR Collaboration), Phys. Rev. Lett. 98,012001 (2007).

[5] K. Abe et al. (Belle Collaboration), Phys. Rev. Lett. 98,262001 (2007).

[6] B. Aubert et al. (BABAR Collaboration), Phys. Rev. D 77,012002 (2008).

SEARCH FOR DOUBLY CHARMED BARYONS AND STUDY … PHYSICAL REVIEW D 89, 052003 (2014)

052003-13

[7] E. Solovieva, R. Chistov, I. Adachi, H. Aihara, K. Arinstein,T. Aushev, A. M. Bakich, and V. Balagura et al., Phys. Lett.B 672, 1 (2009).

[8] T. Lesiak et al. (Belle Collaboration), Phys. Lett. B 665, 9(2008).

[9] J. Beringer et al. (Particle Data Group Collaboration), Phys.Rev. D 86, 010001 (2012).

[10] R. Roncaglia, D. B. Lichtenberg, and E. Predazzi, Phys.Rev. D 52, 1722 (1995).

[11] D. Ebert, R. N. Faustov, V. O. Galkin, A. P. Martynenko,and V. A. Saleev, Z. Phys. C 76, 111 (1997).

[12] B. Silvestre-Brac, Prog. Part. Nucl. Phys. 36, 263 (1996).[13] S.-P. Tong, Y.-B. Ding, X.-H. Guo, H.-Y. Jin, X.-Q. Li, P.-N.

Shen, and R. Zhang, Phys. Rev. D 62, 054024 (2000).[14] S. M. Gerasyuta and D. V. Ivanov, Nuovo Cimento A 112,

261 (1999).[15] C. Itoh, T. Minamikawa, K. Miura, and T. Watanabe, Phys.

Rev. D 61, 057502 (2000).[16] V. V. Kiselev and A. K. Likhoded, Usp. Fiz. Nauk 172, 497

(2002) [Phys. Usp. 45, 455 (2002)].[17] I. M. Narodetskii and M. A. Trusov, arXiv:hep-ph/0204320.[18] D. Ebert, R. N. Faustov, V. O. Galkin, and A. P. Martynenko,

Phys. Rev. D 66, 014008 (2002).[19] J. Vijande, H. Garcilazo, A. Valcarce, and F. Fernandez,

Phys. Rev. D 70, 054022 (2004).[20] S. Migura, D. Merten, B. Metsch, and H.-R. Petry, Eur.

Phys. J. A 28, 41 (2006).[21] C. Albertus, E. Hernandez, J. Nieves, and J. M. Verde-

Velasco, Eur. Phys. J. A 32, 183 (2007); C. Albertus, E.Hernandez, J. Nieves, and J. M. Verde-Velasco, Eur. Phys. J.A 36, 119(E) (2008).

[22] W. Roberts and M. Pervin, Int. J. Mod. Phys. A 23, 2817(2008).

[23] R. Lewis, N. Mathur, and R. M. Woloshyn, Phys. Rev. D 64,094509 (2001).

[24] H. Na and S. Gottlieb, Proc. Sci., LATTICE (2008) 119[arXiv:0812.1235].

[25] L. Liu, H.-W. Lin, K. Orginos, and A. Walker-Loud, Phys.Rev. D 81, 094505 (2010).

[26] Y. Namekawa, (PACS-CS Collaboration) Proc. Sci., LAT-TICE (2012) 139 [arXiv:1212.0073].

[27] C. Alexandrou, J. Carbonell, D. Christaras, V. Drach,M. Gravina, and M. Papinutto, Phys. Rev. D 86, 114501(2012).

[28] V. V. Kiselev, A. K. Likhoded, and M. V. Shevlyagin, Phys.Lett. B 332, 411 (1994).

[29] J. P. Ma and Z. G. Si, Phys. Lett. B 568, 135 (2003).[30] V. V. Braguta, V. V. Kiselev, and A. E. Chalov, Yad. Fiz. 65,

1575 (2002) [Phys. At. Nucl. 65 1537 (2002)].[31] M. Mattson et al. (SELEX Collaboration), Phys. Rev. Lett.

89, 112001 (2002).[32] A. Ocherashvili et al. (SELEX Collaboration), Phys. Lett. B

628, 18 (2005).[33] S. P. Ratti, Nucl. Phys. B, Proc. Suppl. 115, 33 (2003).[34] B. Aubert et al. (BABAR Collaboration), Phys. Rev. D 74,

011103 (2006).[35] R. Aaij et al. (LHCb Collaboration), J. High Energy Phys.

12 (2013) 090.[36] L. Gibbons et al. (CLEO Collaboration), Phys. Rev. Lett.

77, 810 (1996).[37] S. Kurokawa and E. Kikutani, Nucl. Instrum. Methods

Phys. Res., Sect. A 499, 1 (2003), and other papers includedin this volume; T. Abe et al., Prog. Theor. Exp. Phys. (2013)03A001 and following articles up to 03A011.

[38] A. Abashian et al. (Belle Collaboration), Nucl. Instrum.Methods Phys. Res., Sect. A 479, 117 (2002); also seedetector section in J. Brodzicka et al., Prog. Theor. Exp.Phys. (2012), 04D001.

[39] Z.Natkaniec et al. (Belle SVD2 Group), Nucl. Instrum.Methods Phys. Res., Sect. A 560, 1 (2006).

[40] K. Hanagaki, H. Kakuno, H. Ikeda, T. Iijima, and T.Tsukamoto, Nucl. Instrum. Methods Phys. Res., Sect. A485, 490 (2002).

[41] D. J. Lange, Nucl. Instrum. Methods Phys. Res., Sect. A462, 152 (2001).

[42] T. Sjostrand, Comput. Phys. Commun. 82, 74 (1994).[43] E. Barberio and Z. Was, Comput. Phys. Commun. 79, 291

(1994).[44] R. Brun et al., GEANT3.21, CERN Report No. DD/EE/84-1,

1984.[45] Throughout this paper, the inclusion of the charge-conjugate

decay mode is implied unless otherwise stated.[46] K. Sumisawa et al. (Belle Collaboration), Phys. Rev. Lett.

95, 061801 (2005).[47] B. Aubert et al. (BABAR Collaboration), Phys. Rev. D 75,

012003 (2007).[48] K. Abe et al. (Belle Collaboration), Phys. Rev. D 65,

091103 (2002).

KATO et al. PHYSICAL REVIEW D 89, 052003 (2014)

052003-14