Physics Letters B 693 (2010) 88–94

Contents lists available at ScienceDirect

Physics Letters B

www.elsevier.com/locate/physletb

A study of charged κ in J/ψ → K ±K Sπ∓π0

BES Collaboration

M. Ablikim a, J.Z. Bai a, Y. Bai a, Y. Ban k, X. Cai a, H.F. Chen p, H.S. Chen a, H.X. Chen a, J.C. Chen a,Jin Chen a, X.D. Chen e, Y.B. Chen a, Y.P. Chu a, Y.S. Dai r, Z.Y. Deng a, S.X. Du a,1, J. Fang a, C.D. Fu a,C.S. Gao a, Y.N. Gao n, S.D. Gu a, Y.T. Gu d, Y.N. Guo a, Z.J. Guo o,2, F.A. Harris o, K.L. He a, M. He l, Y.K. Heng a,H.M. Hu a, T. Hu a, G.S. Huang a,3, X.T. Huang l, Y.P. Huang a, X.B. Ji a, X.S. Jiang a, J.B. Jiao l, D.P. Jin a, S. Jin a,G. Li a, H.B. Li a, J. Li a, L. Li a, R.Y. Li a, W.D. Li a, W.G. Li a, X.L. Li a, X.N. Li a, X.Q. Li j, Y.F. Liang m, B.J. Liu a,4,C.X. Liu a, Fang Liu a, Feng Liu f, H.M. Liu a, J.P. Liu q, H.B. Liu d,5, J. Liu a, Q. Liu o, R.G. Liu a, S. Liu h,Z.A. Liu a, F. Lu a, G.R. Lu e, J.G. Lu a, C.L. Luo i, F.C. Ma h, H.L. Ma b, Q.M. Ma a, M.Q.A. Malik a, Z.P. Mao a,X.H. Mo a, J. Nie a, S.L. Olsen o, R.G. Ping a, N.D. Qi a, J.F. Qiu a, G. Rong a, X.D. Ruan d, L.Y. Shan a, L. Shang a,C.P. Shen o, X.Y. Shen a, H.Y. Sheng a, H.S. Sun a, S.S. Sun a, Y.Z. Sun a, Z.J. Sun a, X. Tang a, J.P. Tian n,G.L. Tong a, G.S. Varner o, X. Wan a, L. Wang a, L.L. Wang a, L.S. Wang a, P. Wang a, P.L. Wang a, Y.F. Wang a,Z. Wang a, Z.Y. Wang a, C.L. Wei a, D.H. Wei c, N. Wu a,∗, X.M. Xia a, G.F. Xu a, X.P. Xu f, Y. Xu j, M.L. Yan p,H.X. Yang a, M. Yang a, Y.X. Yang c, M.H. Ye b, Y.X. Ye p, C.X. Yu j, C.Z. Yuan a, Y. Yuan a, Y. Zeng g,B.X. Zhang a, B.Y. Zhang a, C.C. Zhang a, D.H. Zhang a, F. Zhang n,6, H.Q. Zhang a, H.Y. Zhang a, J.W. Zhang a,J.Y. Zhang a, X.Y. Zhang l, Y.Y. Zhang m, Z.X. Zhang k, Z.P. Zhang p, D.X. Zhao a, J.W. Zhao a, M.G. Zhao a,P.P. Zhao a, Z.G. Zhao p, B. Zheng a, H.Q. Zheng k, J.P. Zheng a, Z.P. Zheng a, B. Zhong i, L. Zhou a, K.J. Zhu a,Q.M. Zhu a, X.W. Zhu a, Y.S. Zhu a, Z.A. Zhu a, Z.L. Zhu c, B.A. Zhuang a, B.S. Zou a

a Institute of High Energy Physics, Beijing 100049, People’s Republic of Chinab China Center for Advanced Science and Technology (CCAST), Beijing 100080, People’s Republic of Chinac Guangxi Normal University, Guilin 541004, People’s Republic of Chinad Guangxi University, Nanning 530004, People’s Republic of Chinae Henan Normal University, Xinxiang 453002, People’s Republic of Chinaf Huazhong Normal University, Wuhan 430079, People’s Republic of Chinag Hunan University, Changsha 410082, People’s Republic of Chinah Liaoning University, Shenyang 110036, People’s Republic of Chinai Nanjing Normal University, Nanjing 210097, People’s Republic of Chinaj Nankai University, Tianjin 300071, People’s Republic of Chinak Peking University, Beijing 100871, People’s Republic of Chinal Shandong University, Jinan 250100, People’s Republic of Chinam Sichuan University, Chengdu 610064, People’s Republic of Chinan Tsinghua University, Beijing 100084, People’s Republic of Chinao University of Hawaii, Honolulu, HI 96822, USAp University of Science and Technology of China, Hefei 230026, People’s Republic of Chinaq Wuhan University, Wuhan 430072, People’s Republic of Chinar Zhejiang University, Hangzhou 310028, People’s Republic of China

* Corresponding author.E-mail address: wuning@ihep.ac.cn (N. Wu).

1 Currently at: Zhengzhou University, Zhengzhou 450001, People’s Republic of China.2 Currently at: Johns Hopkins University, Baltimore, MD 21218, USA.3 Currently at: University of Oklahoma, Norman, OK 73019, USA.4 Currently at: University of Hong Kong, Pok Fu Lam Road, Hong Kong.5 Currently at: Graduate University of Chinese Academy of Sciences, Beijing 100049, People’s Republic of China.6 Currently at: Harbin Institute of Technology, Harbin 150001, People’s Republic of China.

0370-2693/$ – see front matter © 2010 Elsevier B.V. All rights reserved.doi:10.1016/j.physletb.2010.08.015

BES Collaboration / Physics Letters B 693 (2010) 88–94 89

a r t i c l e i n f o a b s t r a c t

Article history:Received 1 July 2010Accepted 9 August 2010Available online 14 August 2010Editor: M. Doser

Keywords:Charged κLow mass scalarSU(3) symmetry breakingJ/ψ decays

Based on 58 × 106 J/ψ events collected by BESII, the decay J/ψ → K ± K Sπ∓π0 is studied. In the

invariant mass spectrum recoiling against the charged K ∗(892)±, the charged κ particle is found asa low mass enhancement. If a Breit–Wigner function of constant width is used to parameterize theκ , its pole locates at (849 ± 77+18

−14) − i(256 ± 40+46−22) MeV/c2. Also in this channel, the decay J/ψ →

K ∗(892)+ K ∗(892)− is observed for the first time. Its branching ratio is (1.00 ± 0.19+0.11−0.32) × 10−3.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

The σ and κ are controversial particles in hadron spectroscopy.They were first found in the analysis of ππ and π K scatteringdata. Because the total phase shifts in the lower mass region aremuch less than 180◦ and they do not fit into ordinary qq̄ mesonnonets, they have been the subject of violent debates. Refs. [1–12]are some recent analyses that support their existence.

Evidence for κ particles comes from the study of produc-tion processes and the re-analysis of Kπ scattering data. Evi-dence for the κ has been found in D+ → K −π+π− [8], J/ψ →K̄ ∗(892)0 K +π− [2,6], and D+ → K −π+μ+ν [13]. A Kπ s-wavecomponent is found in D0 → K −K +π0 [14], D+ → K −π+e+νe

[15], and τ → K Sπ−ντ [16]. But no evidence for the κ in D0 →

K −π+π0 [17] or for the charged κ in D0 → K −K +π0 [18] is seen.The κ is found in the phenomenological analysis of Kπ scatter-ing phase shift data [12,19–26]. However, some theorists are notconvinced by the evidence [27–30]. The present status of the κ issummarized by the Particle Data Group PDG [31].

The σ and κ particles have been studied with BES data, whereevidence for σ and κ particles is quite clear [1–6]. A neutral κ wasfound in the decay J/ψ → K̄ ∗(892)0κ0 → K̄ ∗(892)0 K +π− [2,6].Because of isospin symmetry, if a neutral κ exists, a charged κshould exist and could be produced in J/ψ → K̄ ∗(892)±κ∓ .

In this study, we search for and study the charged κ in J/ψ →K ±K Sπ

∓π0. Our analysis is based on 58 million J/ψ decays col-lected by BESII at the BEPC (Beijing Electron Positron Collider). BE-SII is a large solid-angle magnetic spectrometer which is describedin detail in Ref. [32]. The momentum of charged particles is deter-mined by a 40-layer cylindrical main drift chamber (MDC). Parti-cle identification is accomplished using specific ionization (dE/dx)measurements in the MDC and time-of-flight (TOF) information ina barrel-like array of 48 scintillation counters. Outside of the TOFis a barrel shower counter (BSC) which measures the energy anddirection of photons.

2. Event selection

In the event selection, candidate tracks are required to have agood track fit with the point of closest approach of the track tothe beam axis being within the interaction region of 2 cm in rxy

and ±20 cm in z (the beam direction), polar angles θ satisfying| cos θ | < 0.80, and transverse momenta Pt > 50 MeV/c. Photonsare required to be isolated from charged tracks, to come from theinteraction region, and have deposited energy in the BSC greaterthan 40 MeV. Events are required to have four good charged trackswith total charge zero and at least two good photons.

For the K S reconstruction, we loop over all oppositely chargedpairs of tracks, assuming them to be pions, and fit them toK S → π+π− , which determines vertices and π+π− invariant

masses, Mππ , for the four possible combinations. The combina-tion with Mππ closest to MK S is selected and is required to sat-isfy |Mπ+π− − MK S | < 20 MeV/c2 and have its decay vertex inthe xy-plane satisfy rxy > 0.008 m. After K S selection, the particletype of the remaining two tracks, that is whether they are K +π−or K −π+ , is decided by selecting the combination with smallestχ2

TOF + χ2DEDX .

A four constraint (4C) kinematic fit is applied under theK ±π∓π+π−γ γ hypothesis, and χ2

4C < 20 is required. Events witha γ γ invariant mass satisfying |Mγ γ − Mπ0 | < 40 MeV/c2 arefitted with a 5C kinematic fit to K ±π∓π+π−π0 with the twophotons constrained to the π0 mass, and events with χ2

5C < 50are selected. The π∓π0 mass distribution is shown in Fig. 1(a),where the ρ is clearly seen. Decays with an intermediate ρ arebackground, and the requirement |Mπ∓π0 − Mρ | > 100 MeV/c2

is applied to remove them. The requirements |MK Sπ0 − 0.897| >

40 MeV/c2 and |MK ±π∓ − 0.897| > 40 MeV/c2 are used to re-move backgrounds from J/ψ → K ∗(892)0 K ±π∓ and J/ψ →K ∗(892)0 K Sπ

0. After these requirements, the combined K ±π0

and K Sπ∓ mass distribution is shown in Fig. 1(b); the high-

est narrow peak is charged K ∗(892). The MK ±π0 versus MK Sπ∓scatter plot is shown in Fig. 1(c). There are two clear bands,a vertical and horizontal band, which correspond to J/ψ →K ∗(892)±K Sπ

∓ and J/ψ → K ∗(892)± K ∓π0, respectively. The re-quirements |MK ±π0 − 0.892| < 80 MeV/c2 and |MK Sπ± − 0.892| <

80 MeV/c2 are imposed to select J/ψ → K ∗(892)±K Sπ∓ and

J/ψ → K ∗(892)± K ∓π0 events, respectively.After the above selection, the combined K Sπ

∓ and K ∓π0 in-variant mass distribution recoiling against the K ∗(892)± is shownin Fig. 1(d). It is the sum of the four decays listed in Table 1with about 1000 events for each, as listed in the table, and a totalof 4121 events. In Fig. 1(d), a clear narrow peak at 892 MeV/c2

and a wider peak at about 1430 MeV/c2 are seen. In addition,there is a broad low mass enhancement just above threshold. Thespectrum is quite similar to the spectrum of K +π− in the decayJ/ψ → K̄ ∗(892)0 K +π− [6]. The biggest difference between themis that the charged K ∗(892) peak is much larger than the neutralK ∗(892) peak. The K ∗(892)π invariant mass distribution is shownin Fig. 1(e), where there are indications of peaks at 1270 MeV/c2

and 1400 MeV/c2. The resulting Dalitz plot is shown in Fig. 1(f).The two diagonal bands correspond to the low mass enhancementcombined with the 892 MeV/c2 peak and the peak around 1430MeV/c2 in the Kπ spectrum.

3. Background studies

Possible sources of background are studied. First, sidebandbackgrounds are studied. The π+π− mass distribution is shown inFig. 2(a), where a clear K S signal can be seen and the background

90 BES Collaboration / Physics Letters B 693 (2010) 88–94

Fig. 1. (a) π∓π0 mass distribution. (b) Combined K ±π0 and K Sπ∓ mass distribution after final cuts except the K ∗(892) cuts. (c) Scatter plot of MK ±π0 versus MK S π∓ .

(d) Invariant mass distribution of K ±π0 and K Sπ± recoiling against K ∗(892)∓ . (e) Invariant mass distribution of K ∗(892)π . (f) Dalitz plot.

level is quite low. The Kπ spectrum from K S side-band events isshown in Fig. 2(b). (The K S side-band is defined by 0.02 GeV/c2 <

|Mππ − 0.497| < 0.04 GeV/c2.) There are no clear structures. Theγ γ spectrum is shown in Fig. 2(c), and the Kπ spectrum of the284 π0 side-band events is shown in Fig. 2(d). (The π0 side-band is defined by 0.04 GeV/c2 < |mγ γ − 0.135| < 0.08 GeV/c2.)The structures in Fig. 2(d) are similar to those in the signal re-gion and come from signal events in the π0 tails. The K ∗(892)

side-band background is shown by the dark shaded histogram inFig. 2(e). (The K ∗(892) side-band is defined by 0.08 GeV/c2 <

|MKπ − 0.892| < 0.16 GeV/c2.) The clear K ∗(892) in the Kπ massdistribution of K ∗(892) side-band events mainly comes from thecross channel. (There are two bands in the scatter plot in Fig. 1(c).When we select one band, we will also select some events fromthe other band where it crosses the first band. These events corre-spond to cross channel background.) Fig. 2(f) shows the Kπ spec-trum after side-band subtraction. The low mass enhancement andK ∗(892) peak survive after side-band subtraction.

Next, we perform Monte Carlo simulation to study the mainphysics background processes, including J/ψ → γ ηc → γ K ∗ K̄ ∗ →γ K ±K Sπ

∓π0, J/ψ → γ ηc → γ K ∗(892)±K ∓π0 → γ K ±K Sπ∓π0,

Table 1Four signal channels and number of events from each.

Channel Events

J/ψ → K ∗(892)+ K Sπ− → K +π0 K Sπ

− 1023J/ψ → K ∗(892)− K Sπ

+ → K −π0 K Sπ+ 946

J/ψ → K ∗(892)+ K −π0 → K Sπ+ K −π0 1055

J/ψ → K ∗(892)− K +π0 → K Sπ− K +π0 1097

J/ψ → γ ηc → γ K ∗(892)± K Sπ∓ → γ K ±K Sπ

∓π0, J/ψ →π0π+π−π+π− , and J/ψ → π0 K +K −π+π− . The selection ef-ficiencies are much lower than 1%, and the largest number ofbackground events contributed is about 6. Therefore, the physicsbackground is quite low.

4. Partial wave analysis

A partial wave analysis (PWA), which is based on the co-variant helicity amplitude analysis [33–37], is performed for thecharged κ . We add the likelihoods of all four channels together,and find the minimum of the sum. This is the same method as

BES Collaboration / Physics Letters B 693 (2010) 88–94 91

Fig. 2. (a) mπ+π− mass distribution after final event selection except for the K S requirement. (b) K S side-band structure. (c) mγ γ mass distribution after final data selectionexcept the π0 requirement. (d) MKπ distribution from π0 side-band events. (e) Kπ spectrum recoiling against K ∗(892). The dark shaded histogram is from K ∗ side-bandevents. (f) The Kπ spectrum after side-band subtraction.

Table 2Resonances included in the fit of this channel. Masses and widths of various resonances are determined by mass and width scans. κ (1), (2) and (3) are results given byfits using Breit–Wigner functions (1), (2) and (3) to fit κ respectively. IPS refers to the broad 0+ structure which interferes with κ . K ∗ BG refers to the K ∗(892) backgroundcoming from the cross channel. PS BG refers to the background with no interference with resonances.

Resonance Spin-parity Decay mode Mass(MeV/c2)

Width(MeV/c2)

Significance

κ (1) 0+ Kπ 810 ± 68+15−24 536 ± 87+106

−47 > 6σ

κ (2) 0+ Kπ 884 ± 40+11−22 478 ± 77+71

−41 > 6σ

κ (3) 0+ Kπ 1165 ± 58+120−41 1349 ± 500+472

−176 > 6σ

K ∗0 (1430) 0+ Kπ 1400 ± 86 325 ± 200 0.6σ

IPS 0+ Kπ – – > 6σK ∗

2 (1430) 2+ Kπ 1411 ± 30 111 ± 46 > 6σK ∗

2 (1920) 2+ Kπ 2020 ± 140 705 ± 160 > 6σK ∗(1410) 1− Kπ 1420 ± 14 130 ± 28 > 6σK ∗(892) 1− Kπ 896 ± 8 57 ± 12 > 6σK1(1270) 1+ K ∗(892)π 1254 ± 14 60 ± 28 > 6σK1(1400) 1+ K ∗(892)π 1390 ± 30 146 ± 44 > 6σb1(1235) 1+ K ∗(892)K 1230 ± 52 142 ± 38 4.5σK ∗(892) BG 1− Kπ – – > 6σPS BG – – – – –

92 BES Collaboration / Physics Letters B 693 (2010) 88–94

Table 3Masses, widths and pole positions of the charged κ . In the table, the first errors arestatistical, and the second are systematic. BW (1) means Eq. (1) is used to fit the κ .BW (2) and BW (3) have similar meanings.

BW (1) BW (2) BW (3)

Mass (MeV/c2) 810 ± 68+15−24 884 ± 40+11

−22 1165 ± 58+120−41

Width (MeV/c2) 536 ± 87+106−47 478 ± 77+71

−41 1349 ± 500+472−176

Pole (MeV/c2) (849 ± 77+18−14)

− i(256 ± 40+46−22)

(849 ± 51+14−28)

− i(288 ± 101+64−30)

(839 ± 145+24−7 )

− i(297 ± 51+50−18)

used to fit J/ψ → K ∗(892)0 K +π− [6]. The main difference here isthat the decay J/ψ → K ∗(892)± K ∗(892)∓ is included in the fit.

In the PWA analysis, ten resonances, κ , K ∗0 (1430), IPS, K ∗

2 (1430),K ∗

2 (1920), K ∗(1410) and K ∗(892) in the Kπ spectrum, K1(1270)

and K1(1400) in the K ∗(892)π spectrum, and b1(1235) in theK ∗(892)K spectrum, and two backgrounds, listed as the last itemsin Table 2, are considered in the fit. In Table 2, IPS (InterferencePhase Space) refers to a broad 0+ structure with a Kπ invariantmass spectrum the same as phase space, that interferes with κ .K ∗ BG refers to the K ∗(892) background coming from the crosschannel. PS (Phase Space) refers to the background with no inter-ference with resonances and with a shape almost the same as thatof phase space.

Three different parameterizations are used to fit the κ . They are

BWκ = 1

m2κ − s − imκΓκ

, Γκ = const, (1)

BWκ = 1

m2κ − s − i

√sΓκ(s)

, Γκ(s) = g2κ · kκ

8π s, (2)

BWκ = 1

m2κ − s − i

√sΓκ(s)

, Γκ(s) = α · kκ , (3)

where kκ is the magnitude of the K momentum in the Kπ , orthe κ , center of mass system [38], and α is a constant which willbe determined by the fit. Parameters in the Breit–Wigner functionare determined by mass and width scans. The minima of the scancurves give the central values of mass and width parameters. Fromthese, the corresponding Breit–Wigner pole positions can be di-rectly calculated from Eqs. (1), (2) and (3). Our final results arelisted in Table 3, where the first errors are statistical, and the sec-ond are systematic. The mass and width parameters obtained bydifferent parameterizations are quite different, but their poles arealmost the same, which is quite similar to what was found in thestudy of the neutral κ . The results for the neutral κ [6] are shownin Table 4 and are consistent with the charged κ .

Table 4Masses, widths and pole positions of the neutral κ [6]. BW (1) means Eq. (1) isused to fit the κ . BW (2) and BW (3) have similar meaning.

BW (1) BW (2) BW (3)

Mass (MeV/c2) 745 ± 26+14−91 874 ± 25+12

−55 1140 ± 39+47−80

Width (MeV/c2) 622 ± 77+61−78 518 ± 65+27

−87 1370 ± 156+406−148

Pole (MeV/c2) (799 ± 37+16−90)

− i(290 ± 33+25−38)

(836 ± 38+18−87)

− i(329 ± 66+28−46)

(811 ± 74+17−83)

− i(285 ± 20+18−42)

Our final results correspond to the solution with the mini-mum least likelihood. Differences among solutions with similarlikelihood values are included as systematic uncertainties. Also in-cluded are the effect of removing K0(1430), IPS, b1(1235) andK ∗(892)K̄ ∗(892), the result of a fit with the K ∗(892) backgroundlevel floating, and the result from a fit using direct side-band sub-traction.

The masses and widths of all resonances obtained by mass andwidth scans are shown in Table 2. In the fit, the contribution fromK ∗

0 (1430) is small; its statistical significance is only 0.6σ . Becauseit is expected in this channel, it is included in the final solution.The Kπ mass distribution is shown in Fig. 3(a), where points witherror bars are data, and the light shaded histogram is the final fit.In the figure, the dark shaded histogram shows the contribution ofthe charged κ . Fig. 3(b) shows the fit for the K ∗(892)π spectrum,and Fig. 4 shows the angular distributions.

5. Branching ratio measurements

The decay J/ψ → K ∗(892)±κ∓ contributes 655 events. MonteCarlo simulation of J/ψ → K ∗(892)±κ∓ → K ±K Sπ

∓π0 deter-mines an efficiency of 2.33%, and the branching ratio of J/ψ →K ∗(892)+κ− or J/ψ → K ∗(892)−κ+ is

BR = 655/4

2.33% × 5.8 × 107 × 1/9

= (1.09 ± 0.18+0.94

−0.54

) × 10−3, (4)

where, the first error is statistical, the second error is systematic,the factor 1

4 is because the events are the sum of four channels,and 1

9 is the isospin factor. In the previous study of the decayJ/ψ → K̄ ∗(892)0 K +π− , the corresponding branching ratio for theneutral κ7 is (0.52–0.97) × 10−3. The two results are consistent

7 The branching ratio of the neutral κ was not reported in Ref. [6]. In thestudy of the neutral κ , the number of κ events was in the range 1891–3516,

Fig. 3. (a) Final fit results for the Kπ spectrum. Points with error bars are data, the light shaded histogram is the final global fit, and the dark shaded histogram is thecontribution of the κ . (b) Final fit of K ∗(892)π spectrum. Dots with error bars are data, and the histogram is the final global fit. There are two peaks in the lower massregion. The lower one is fit by the K1(1270), and the higher one is by the K1(1400).

BES Collaboration / Physics Letters B 693 (2010) 88–94 93

Fig. 4. Final fit result for the angular distributions. Dots with error bars are data, and the histogram is the final global fit. The upper left figure shows the final fit of the θ1

distribution, the upper right shows the final fit of the φ1 distribution, the lower left shows the final fit of the θ2 distribution, the lower right shows the final fit of the φ2

distribution. θ1 and φ1 are the polar angle and azimuthal angle of the Kπ system in the J/ψ center of mass system. θ2 and φ2 are the polar angle and azimuthal angle ofthe K meson in the Kπ center of mass system.

with isospin symmetry. The systematic error includes uncertain-ties from multi-solutions, from different background fit methods,and from removing some components from the fit (K ∗

0 (1430), IPS,b1(1235), and K ∗(892)± K ∗(892)∓).

Also interesting is the existence of the J/ψ electromagneticdecay J/ψ → K ∗(892)± K ∗(892)∓ . The peak at 892 MeV/c2 inthe Kπ invariant mass spectrum (see Fig. 1(d)) comes from bothcross channel background and from J/ψ → K ∗(892)± K ∗(892)∓ .The number of events from the cross channel can be calculatedapproximately from the K ∗(892) side-band structure, shown inFig. 2(e), where the narrow peak at 892 MeV/c2 is clear. Inthe final fit, the decay J/ψ → K ∗(892)± K ∗(892)∓ contributes323 events. The Monte Carlo simulation of the decay J/ψ →K ∗(892)±K ∗(892)∓ → K ±K Sπ

∓π0 yields an efficiency of 1.25%,and the branching ratio of J/ψ → K ∗(892)+K ∗(892)− is

BR = 323/2

1.25% × 5.8 × 107 × 1/9 × 2

= (1.00 ± 0.19+0.11

−0.32

) × 10−3, (5)

where, the first error is statistical, the second error is systematic,the factor 1

2 is because events are counted twice, 19 is the isospin

factor, and 2 is because the data comes from two decay chan-nels: J/ψ → K ∗(892)+ K ∗(892)− → (K +π0)(K Sπ

−) and J/ψ →K ∗(892)−K ∗(892)+ → (K −π0)(K Sπ

+). The systematic error in-cludes uncertainties from multi-solutions, from different back-ground fit methods, and from removing some components fromfit (K ∗

0 (1430), IPS, and b1(1235)).

and the selection efficiency was 14.2%. Its branching ratio is 1891–351614.2%×5.8×107×4/9

=(0.52–0.97) × 10−3.

6. Summary

In conclusion, the charged κ is observed and studied in thedecay J/ψ → K ∗(892)±κ∓ → K ±K Sπ

∓π0. The low mass en-hancement in the Kπ spectrum cannot be fit well unless acharged κ is added into the solution. If we use a Breit–Wignerfunction of constant width to parameterize the κ , its pole lo-cates at (849 ± 77+18

−14) − i(256 ± 40+46−22) MeV/c2. In our analysis,

three different κ parameterizations are tried in the fit, and fi-nal results are shown in Table 3 and are consistent with thoseof the neutral κ and are also in good agreement with those ob-tained in the analysis of Kπ scattering phase shifts. Also, the decayJ/ψ → K ∗(892)± K ∗(892)∓ is observed for the first time with thebranching ratio (1.00 ± 0.19+0.11

−0.32) × 10−3. The corresponding de-cay mode is not observed in J/ψ → K̄ ∗(892)0 K +π− . The decaysJ/ψ → K ∗(892)±K ∗(892)∓ can be produced through J/ψ elec-tromagnetic decays, while J/ψ → K ∗(892)0 K̄ ∗(892)0 can only beproduced through J/ψ hadronic decays, which would be SU(3)

symmetry breaking decays and are suppressed.

Acknowledgements

The BES Collaboration thanks the staff of BEPC and computingcenter for their hard efforts. This work is supported in part by theNational Natural Science Foundation of China under contract Nos.10491300, 10225524, 10225525, 10425523, 10625524, 10521003,10821063, 10825524, the Chinese Academy of Sciences under con-tract No. KJ 95T-03, the 100 Talents Program of CAS under contractNos. U-11, U-24, U-25, and the Knowledge Innovation Project ofCAS under contract Nos. U-602, U-34 (IHEP), the National Nat-ural Science Foundation of China under contract Nos. 10775077,10225522 (Tsinghua University), and the Department of Energy un-der contract No. DE-FG02-04ER41291 (U. Hawaii).

94 BES Collaboration / Physics Letters B 693 (2010) 88–94

References

[1] Ning Wu, BES Collaboration, BES R measurements and J/ψ decays, in: J. TranThanh Van (Ed.), QCD and High Energy Hadronic Interactions, Proceedings ofthe XXXVIth Rencontres de Moriond, Les Arcs, France, March 17–24, 2001,pp. 3–6.

[2] J.Z. Bai, BES Collaboration, hep-ex/0304001.[3] J.Z. Bai, et al., BES Collaboration, High Energy Phys. Nucl. Phys. 28 (2004)

215.[4] M. Ablikim, et al., BES Collaboration, Phys. Lett. B 598 (2004) 149.[5] M. Ablikim, et al., BES Collaboration, Phys. Lett. B 645 (2007) 19.[6] M. Ablikim, et al., BES Collaboration, Phys. Lett. B 633 (2006) 681.[7] E.M. Aitala, et al., Fermilab E791 Collaboration, Phys. Rev. Lett. 86 (2001) 770.[8] E.M. Aitala, et al., Fermilab E791 Collaboration, Phys. Rev. Lett. 89 (2002)

121801.[9] D. Alde, et al., Phys. Lett. B 397 (1997) 350.

[10] T. Ishida, et al., in: Proceedings of International Conference Hadron ’95, WorldScientific, Manchester, UK, 1995.

[11] V.E. Markushin, M.P. Locher, Frascati Phys. Ser. 15 (1999) 229.[12] Z. Xiao, H.Q. Zheng, Nucl. Phys. A 695 (2001) 273.[13] J.M. Link, et al., FNAL FOCUS Collaboration, Phys. Lett. B 535 (2002) 430.[14] C. Cawlfield, et al., CLEO Collaboration, Phys. Rev. D 74 (2006) 031108R.[15] M.R. Shepherd, et al., CLEO Collaboration, Phys. Rev. D 74 (2006) 052001.[16] D. Epifanov, et al., BELLE Collaboration, Phys. Lett. B 654 (2007) 65.[17] S. Anderson, et al., Phys. Rev. D 63 (2001) 09001.

[18] B. Aubert, et al., BABAR Collaboration, Phys. Rev. D 76 (2007) 011102R.[19] E. van Beveren, et al., Z. Phys. C 30 (1986) 651.[20] D. Aston, et al., Nucl. Phys. B 296 (1988) 253.[21] S. Ishida, et al., Progr. Theor. Phys. 98 (1997) 621.[22] D. Black, et al., Phys. Rev. D 58 (1998) 054012.[23] J.A. Oller, E. Oset, Phys. Rev. D 60 (1999) 074023.[24] M.J. Jamin, et al., Nucl. Phys. B 587 (2000) 331.[25] D. Bugg, Phys. Lett. B 572 (2003) 1.[26] H.Q. Zheng, et al., Nucl. Phys. A 733 (2004) 235.[27] D. Lohs, Phys. Lett. B 234 (1990) 235.[28] N.A. Tornqvist, Z. Phys. C 68 (1995) 467.[29] A.V. Anisovich, A.V. Sarantsev, Phys. Lett. B 413 (1997) 137.[30] S.N. Cherry, M.R. Pennington, Nucl. Phys. A 688 (2001) 823.[31] Particle Data Group, Phys. Lett. B 667 (2008) 1.[32] J.Z. Bai, et al., BES Collaboration, Nucl. Instrum. Methods A 344 (1994) 319;

J.Z. Bai, et al., BES Collaboration, Nucl. Instrum. Methods A 458 (2001) 627.[33] M. Jacob, G.C. Wick, Ann. Phys. (N.Y.) 7 (1959) 404.[34] S.U. Chung, Phys. Rev. D 57 (1998) 431.[35] Ning Wu, Tu-Nan Ruan, Commun. Theor. Phys. (Beijing, China) 35 (2001) 547.[36] Ning Wu, Tu-Nan Ruan, Commun. Theor. Phys. (Beijing, China) 35 (2001) 693.[37] Ning Wu, Tu-Nan Ruan, Commun. Theor. Phys. (Beijing, China) 37 (2002) 309.[38] H.Q. Zheng, How to parameterize a resonance with finite width, Talk given at

International Symposium on Hadron Spectroscopy, Chiral Symmetry and Rel-ativistic Description of Bound Systems, Tokyo, Japan, 24–26 February 2003,hep-ph/0304173.