Observation of the decay $B^0 \to J/ψη$

0.15 MB·

by  M. -C. Chang

Observation of the decay B0 J/ψ η
→
M.-C. Chang,5 K. Abe,8 K. Abe,43 I. Adachi,8 H. Aihara,45 D. Anipko,1 K. Arinstein,1 T. Aushev,18,13
A. M. Bakich,40 E. Barberio,20 A. Bay,18 I. Bedny,1 K. Belous,12 U. Bitenc,14 I. Bizjak,14 S. Blyth,23 A. Bondar,1
A. Bozek,26 T. E. Browder,7 P. Chang,25 Y. Chao,25 A. Chen,23 K.-F. Chen,25 W. T. Chen,23 B. G. Cheon,3
R.Chistov,13 S.-K.Choi,51 Y. Choi,39 Y.K.Choi,39 S.Cole,40 J.Dalseno,20 M.Danilov,13 M.Dash,49 A.Drutskoy,4
S. Eidelman,1 S. Fratina,14 N. Gabyshev,1 T. Gershon,8 A. Go,23 G. Gokhroo,41 H. Ha,16 J. Haba,8 T. Hara,31
K. Hayasaka,21 M. Hazumi,8 D. Heffernan,31 T. Hokuue,21 Y. Hoshi,43 S. Hou,23 W.-S. Hou,25 Y. B. Hsiung,25
T. Iijima,21 K. Inami,21 A. Ishikawa,45 H. Ishino,46 R. Itoh,8 M. Iwasaki,45 Y. Iwasaki,8 H. Kaji,21 J. H. Kang,50
S. U. Kataoka,22 H. Kawai,2 T. Kawasaki,28 H. R. Khan,46 H. Kichimi,8 S. K. Kim,37 Y. J. Kim,6 K. Kinoshita,4
7
0 S. Korpar,19,14 P. Kriˇzan,52,14 P. Krokovny,8 R. Kulasiri,4 R. Kumar,32 C. C. Kuo,23 A. Kuzmin,1 Y.-J. Kwon,50
0 G. Leder,11 M. J. Lee,37 S. E. Lee,37 T. Lesiak,26 A. Limosani,8 S.-W. Lin,25 G. Majumder,41 F. Mandl,11
2
T. Matsumoto,47 A. Matyja,26 S. McOnie,40 W. Mitaroff,11 K. Miyabayashi,22 H. Miyake,31 H. Miyata,28

n Y. Miyazaki,21 R. Mizuk,13 T. Mori,21 T. Nagamine,44 I. Nakamura,8 E. Nakano,30 M. Nakao,8 S. Nishida,8
a
J O.Nitoh,48 S. Noguchi,22 S.Ogawa,42 T.Ohshima,21 S. Okuno,15 S.L. Olsen,7 Y.Onuki,34 H.Ozaki,8 P.Pakhlov,13
1 G. Pakhlova,13 H. Palka,26 H. Park,17 K. S. Park,39 L. S. Peak,40 R. Pestotnik,14 L. E. Piilonen,49 Y. Sakai,8
N. Satoyama,38 T. Schietinger,18 O. Schneider,18 J. Schu¨mann,24 A. J. Schwartz,4 R. Seidl,9,34 K. Senyo,21

2 M. Shapkin,12 H. Shibuya,42 B. Shwartz,1 J. B. Singh,32 A. Somov,4 N. Soni,32 S. Staniˇc,29 M. Stariˇc,14 H. Stoeck,40
v
K. Sumisawa,8 T. Sumiyoshi,47 S. Suzuki,35 O. Tajima,8 F. Takasaki,8 K. Tamai,8 N. Tamura,28 M. Tanaka,8
7
4 G. N. Taylor,20 Y. Teramoto,30 X. C. Tian,33 I. Tikhomirov,13 K. Trabelsi,7 T. Tsuboyama,8 T. Tsukamoto,8
0 S. Uehara,8 T. Uglov,13 S. Uno,8 P. Urquijo,20 Y. Ushiroda,8 Y. Usov,1 G. Varner,7 K. E. Varvell,40 S. Villa,18
9
C. C. Wang,25 C. H. Wang,24 M.-Z. Wang,25 Y. Watanabe,46 R. Wedd,20 E. Won,16 Q. L. Xie,10 A. Yamaguchi,44
0
6 Y. Yamashita,27 M. Yamauchi,8 Y. Yusa,49 C. C. Zhang,10 L. M. Zhang,36 Z. P. Zhang,36 and A. Zupanc14
0 (The Belle Collaboration)
/
x 1Budker Institute of Nuclear Physics, Novosibirsk
e 2Chiba University, Chiba
p- 3Chonnam National University, Kwangju
e 4University of Cincinnati, Cincinnati, Ohio 45221
h 5Department of Physics, Fu Jen Catholic University, Taipei
: 6The Graduate University for Advanced Studies, Hayama, Japan
v
7University of Hawaii, Honolulu, Hawaii 96822
i
X 8High Energy Accelerator Research Organization (KEK), Tsukuba
9University of Illinois at Urbana-Champaign, Urbana, Illinois 61801
r
a 10Institute of High Energy Physics, Chinese Academy of Sciences, Beijing
11Institute of High Energy Physics, Vienna
12Institute of High Energy Physics, Protvino
13Institute for Theoretical and Experimental Physics, Moscow
14J. Stefan Institute, Ljubljana
15Kanagawa University, Yokohama
16Korea University, Seoul
17Kyungpook National University, Taegu
18Swiss Federal Institute of Technology of Lausanne, EPFL, Lausanne
19University of Maribor, Maribor
20University of Melbourne, Victoria
21Nagoya University, Nagoya
22Nara Women’s University, Nara
23National Central University, Chung-li
24National United University, Miao Li
25Department of Physics, National Taiwan University, Taipei
26H. Niewodniczanski Institute of Nuclear Physics, Krakow
27Nippon Dental University, Niigata
28Niigata University, Niigata
29University of Nova Gorica, Nova Gorica
30Osaka City University, Osaka
2
31Osaka University, Osaka
32Panjab University, Chandigarh
33Peking University, Beijing
34RIKEN BNL Research Center, Upton, New York 11973
35Saga University, Saga
36University of Science and Technology of China, Hefei
37Seoul National University, Seoul
38Shinshu University, Nagano
39Sungkyunkwan University, Suwon
40University of Sydney, Sydney NSW
41Tata Institute of Fundamental Research, Bombay
42Toho University, Funabashi
43Tohoku Gakuin University, Tagajo
44Tohoku University, Sendai
45Department of Physics, University of Tokyo, Tokyo
46Tokyo Institute of Technology, Tokyo
47Tokyo Metropolitan University, Tokyo
48Tokyo University of Agriculture and Technology, Tokyo
49Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061
50Yonsei University, Seoul
51Gyeongsang National University, Chinju
52University of Ljubljana, Ljubljana
We report the first observation of B0 → J/ψ η decay. These results are obtained from a data
sample that contains 449 million BB¯ pairs accumulated at the Υ(4S) resonance with the Belle
detectorat theKEKBasymmetric-energy e+e− collider. Weobserve asignal with asignificance of
8.1σ and obtain a branching fraction of (9.5 ± 1.7 (stat) ± 0.8 (syst))×10−6.
PACSnumbers: 13.20.He,13.25.Hw,14.40.Aq,14.40.Nd
At the quark level, the decay B0 J/ψη proceeds
primarily via a ¯b c¯cd¯transition. As→is apparent from
→ J=
its Feynman diagram (Fig. 1), this is a Cabibbo- and
b
color-suppresseddecay,similartoB0 J/ψπ0 forwhich
→
tbheeenbrmaneachsuinregdfrbaycttihoenBanadBaCrPanvdioBlaetliloencpoallraabmoreatteirosnhsa[v1e, B0 W+ d
2, 3, 4, 5, 6]. The best existing limit for the branching (cid:17)
fraction for B0 J/ψη comes from BaBar [7] and is d d
based on 56 milli→on BB¯ pairs.
FIG.1: Feynmandiagramfortheleadingcontributiontothe
Iffactorizationandflavor-SU(3)symmetriccoefficients
decay B0 →J/ψ η.
forthed¯dcontentareassumed,thebranchingfractionfor
B0 J/ψη decay is expected to be comparable to that
→
for B0 J/ψπ0. If the measured branching fraction is
andanelectromagneticcalorimetercomprisedofCsI(Tl)
→
significantlydifferentfromthis expectation,itwouldim-
crystals (ECL). All these detectors are located inside a
ply the influence of large final-state interactions or non-
superconductingsolenoidcoilthatprovidesa1.5Tmag-
standard contributions. In the latter case, precise mea-
netic field. An iron flux-return located outside of the
surements of CP violation parameters might also reveal coil is instrumented to detect K0 mesons and to iden-
L
unexpected phenomena.
tify muons (KLM). The detector is described in detail
In this letter, we report the first observation of the elsewhere [10].
decay B0 J/ψη [8]. The measurement is based on a The data used in this analysis was collected with two
→
datasamplethatcontains449millionBBpairs,collected detector configurations. A 2.0 cm radius beampipe and
withthe Belledetector atthe KEKBasymmetric-energy a 3-layer silicon vertex detector are used for the first
e+e− (3.5 on 8 GeV) collider [9] operating at the Υ(4S) sample of 152 million BB¯ pairs, while a 1.5 cm radius
resonance. beampipe, a 4-layer silicon detector and a small-cell in-
TheBelledetectorisalarge-solid-anglemagneticspec- ner drift chamber are used to record the remaining 297
trometerthatconsistsofasiliconvertexdetector(SVD), million BB¯ pairs [11]. A GEANT-based Monte Carlo
a50-layercentraldriftchamber(CDC),anarrayofaero- (MC) simulation is used to model the response of the
gel threshold Cˇerenkovcounters (ACC), a barrel-like ar- detector and determine efficiencies [12].
rangementof time-of-flight scintillation counters (TOF), Charged tracks are selected based on the impact pa-
3
rameters relative to the interaction point: dr for the ra- dates, the B candidate with the minimum χ2 value from
dialdirectionand dz for the directionalongthe positron the mass- and vertex-constrainedfit is chosen.
beam. Our requirements are dr < 1 cm and dz < 5 To suppress the combinatorial background dominated
cm. From among the selectedchargedtracks,e+| an|d e− by the two-jet-like e+e− qq¯ (q = u,d,s) contin-
→
candidates are identified by combining information from uum process, we remove events with the ratio of sec-
the ECL,the CDC (dE/dx) andthe ACC.Identification ond to zeroth Fox-Wolfram moments R > 0.4 [13].
2
of µ+ and µ− candidates is based on track penetration This requirement is determined using a figure-of-merit
depth and the hit pattern in the KLM system. Charged N /√N +N ,whereN isthenumberofexpectedsig-
S S B S
pionsareidentifiedusingthecombinedinformationfrom nal events and N is the number of background events.
B
the CDC (dE/dx), the TOF and the ACC. For N , we use a MC simulation with the assumption
S
Photoncandidatesareselectedfromcalorimetershow- that (B0 J/ψ η) is 6 10−6 [14]. For N , we use a
B
B → ×
ers not associated with charged tracks. An energy de- sampleofcontinuumbackgroundMCthatcorrespondsto
position with a photon-like shape and with energy of at anintegratedluminositythatisaboutthreetimesthatof
least50MeVinthebarrelregionor100MeVintheend- the data sample andnormalizedto the number of events
cap region is considered a photon candidate. A pair of in an M ∆E sideband (5.2 GeV/c2 < M < 5.26
bc bc
−
photons with an invariant mass 117.8 MeV/c2 <M < GeV/c2 and 0.1 GeV < ∆E < 0.2 GeV). This require-
γγ
150.2 MeV/c2 is considered as a π0 candidate. This in- mentonR eliminates88%ofthecontinuumbackground
2
variantmassregioncorrespondstoa 3σintervalaround and retains 97% of the signal.
±
the π0 mass, where σ is the mass resolution. After the above continuum suppression, the back-
WereconstructJ/ψ mesonsintheℓ+ℓ− decaychannel ground is dominated by BB¯ events with B decays to
(ℓ=eorµ)andincludeuptoonebremsstrahlungphoton J/ψ. We use a MC sample corresponding to 3.86
thatiswithin50mradofeachofthee+ande−tracks(de- 1010 generic BB¯ decays that includes all known B0 ×
noted as e+e−(γ)). The invariant mass is required to be J/ψX processes to investigate these backgrounds. W→e
within 0.15 GeV/c2 <M m < 0.036 GeV/c2 find that the dominant backgrounds come from B0
and 0.−06GeV/c2<M eme(γ)−<0J./0ψ36GeV/c2,where J/ψK (25.1%), B± J/ψK∗±(892) (19.3%), B0 →
m −denotesthenominµaµl−maJss/,ψM andM arethe J/ψKL∗0(892) (14.5%),→B0 J/ψK (14.1%), B0 →
recJo/ψnstructed invariant mass fromeee(+γe)−(γ) anµµd µ+µ−, J/ψπ0 (8.4%), and a few ot→her excluSsive B J/ψ →X
→
respectively. Asymmetric intervals are used to include decaymodes. ThesebackgroundspeakintheM distri-
bc
part of the radiative tails. butionsbutnotin∆E. Thus,combinatorialbackgrounds
Candidate η mesons are reconstructed in the γγ and from B J/ψ X decays do not affect the signal yield
π+π−π0 finalstates. Werequiretheinvariantmasstobe extracted→from the ∆E distribution.
in the range 500 MeV/c2 < M < 575 MeV/c2 ( 3σ) Signalyields and backgroundlevels are determinedby
γγ
±
and 535 MeV/c2 < Mπ+π−π0 < 560 MeV/c2 (±3σ). In fitting the distributions in∆E forcandidatesintheMbc
theγγ finalstate,weremoveη candidatesifeitherofthe signal region. The ∆E distribution is fitted with a func-
daughter photons form a π0 candidate when combined tion that is the sum of a Crystal Ball line function [15]
with any other photon in the event. Asymmetric decays andtwoGaussiansforsignal,andasecond-orderpolyno-
are removed with the requirement E E /(E + mial for background. The probability density functions
γ1 γ2 γ1
| − |
E ) < 0.8, where E and E are the energies of the (PDF) that describe the signal and background shapes
γ2 γ1 γ2
two photons that form the η candidate. are determined from MC simulations. We use the MC
We combine the J/ψ and η to form B mesons. Sig- sample for this process to determine the PDF for the
nal candidates are identified by two kinematic variables total background.
defined in the Υ(4S) center-of-mass (CM) frame: the We use the decay B+ J/ψK∗+(K∗+ K+π0) as
baneadmth-eeneenrgeyrgycodnisffterareinnecde,m∆aEss=, ME∗bc =EpEb2,ewamhe−repp∗B2∗, aancdonMtrColfosarmthpelefittotedcomrree→actnfaonrddwiffiedrtehncveaslub→eestwofetehned∆atEa
and E∗ are the momentum and eBne−rgyboeafmthe B canB- signal peak. For this selection, we require the helicity
B
didate and E is the run-dependent beam energy. angle in the K∗+ K+π0 decay, i.e. the angle between
To improve thbeeam∆E resolution, the masses of the se- the π0 and the ne→gative of the B+ momenta in the K∗
lected π0, η and J/ψ candidates are constrainedto their rest frame, to be less than 86 degrees. This requirement
nominal masses using mass-constrained kinematic fits. primarily selects events with a high momentum π0 and
In addition, vertex-constrained fits are applied to the producesacontrol-sample∆Edistributionthatissimilar
η π+π−π0 and J/ψ ℓ+ℓ− candidates. We retain to that for our signal decay mode. The signal PDF is
→ →
events with M > 5.2 GeV/c2 and ∆E < 0.2 GeV, modifiedbasedonthe resultsforthe controlsample: the
bc
| |
and define a signal region to be 5.27GeV/c2 < M < mean value of ∆E is shifted by 3.81 0.65 MeV and
bc
− ±
5.29GeV/c2, 0.10GeV<∆E <0.05GeV (η γγ) or the width is scaled by a factor of 0.99 0.04.
∆E < 0.05G−eV (η π+π−π0). In events w→ith more There are 98 and 58 events in the M± signal region
bc
t|han|one B candidate→, usually due to multiple η candi- for the η γγ and η π+π−π0 modes, respectively.
→ →
4
We determine the signal content by performing an un-
binned extended maximum-likelihood fit to the candi-
date data events. The unbinned extended maximum-
likelihood function is 14
e−(NS+NB) N 12
= Y[NSPS(∆Ei)+NBPB(∆Ei)], (1) V
L N! e
i M10
0
1
where N is the total number of candidate events, NS s/
and NB denote the number of signal and BB¯ back- rie 8
ground events, P (∆E ) and P (∆E ) denote the sig- nt
S i B i e
nal and background ∆E PDFs respectively, and i is of 6
the event index. Separate fits to the B0 J/ψη(γγ) o.
and B0 J/ψη(π+π−π0) candidate samp→les give sig- N 4
→
nal yields of 43.1 8.9 and 16.6 5.8 events, respec-
± ± 2
tively. The results of the fits to the data are shown
in Fig. 2. These signal yields correspond to branching
0
fractions of (9.6 1.9) 10−6 and (10.0 3.5) 10−6, -0.2-0.15-0.1-0.05 0 0.05 0.1 0.15 0.2
respectively. We±fitthe×twosamplessimult±aneous×ly,con- D E (GeV)
strainingtheresultstoacommonbranchingfraction,and
obtain59.7 10.5events. Thecombinedbranchingfrac-
tion (9.5 1±.7) 10−6. 10
± ×
We correctfor discrepancies of the signal detection ef- 9
ficiency between data and MC using control samples.
8
The J/ψ µ+µ− and D∗+ D0π+ (D0 K−π+)
→ → → V
events are used as control samples to correct for the e 7
M
muon identification and pion identification, respectively. 0 6
We find that the MC overestimates the efficiencies for 1
J/ψ µ+µ− and η π+π−π0 by 8.5% and 1.8%, re- es/ 5
spect→ively. → ntri
e 4
A 4.8% systematic error associated with uncertainties of
inthesignalandbackgroundPDFsis estimatedbycom- o. 3
N
paring the fit results for the cases when the polynomial
2
parameters are fixed by either MC or data in the M
bc
sidebandregion,andfromchangesthatresultfromvary- 1
ing each parameter by one standard deviation. For the
0
contributionfromthecorrectionstotheleptonidentifica- -0.2-0.15-0.1-0.05 0 0.05 0.1 0.15 0.2
tionefficiencies,we useacontrolsampleofJ/ψ ℓ+ℓ−, D E (GeV)
→
which indicates uncertainties of 2.7% per electron track
and1.2%permuontrack. Byweightingtherelativecon- FIG. 2: ∆E distributions for the decay B0 → J/ψη. The
tributions from J/ψ e+e− and J/ψ µ+µ−, we plot on the top shows B0 → J/ψη(η → γγ) while B0 →
assign a 3.9% system→atic error. The sys→tematic error J/ψη(η → π+π−π0) is on the bottom. The curves show the
signal (dash-dotlines) and background (dashed lines) contri-
from the pion identification correctionis 0.7% per track;
this is determined from a study of the D∗+ D0π+ butionsas well as theoverall fit (solid lines).
(D0 K−π+) control sample. Since this app→lies only
toth→eη π+π−π0 decaymode,theeffectivesystematic
erroris0→.4%. Thesystematicerrorduetothetrackfind- J/ψ ℓ+ℓ−, η γγ,and η π+π−π0 are 1.0%, 0.7%,
→ → →
ing efficiencies for the Belle detector is 1.0%per charged and 1.8%, respectively [16]. These errors contribute 1%
lepton and 1.3% per charged pion. This contributes 2% and 1.1% systematic errors due to (J/ψ ℓ+ℓ−) and
per J/ψ and0.7%perη, andthe totalof2.7%(fromlin- (η γγ and π+π−π0), respectiveBly. We→assign a sys-
B →
early adding) is included. A 4.1% systematic error due tematic errorof1.2%due to the uncertaintyinthe num-
to π0 detection is determined from a comparison of the ber of BB¯ pairs.
data and MC ratios for a large sample of η π+π−π0 The systematic errorsaresummarizedin Table I. The
→
andη 3π0 decays. Sinceη γγ issimilartoπ0 decay, totaluncertaintyisthe quadraticsumofeachterm. The
→ →
we also assign 4.1% as the systematic error for η γγ detection efficiencies and branching fractions are listed
→
reconstruction. Theerrorsonthebranchingfractionsfor in Table II. The statistical significance of the observed
5
signal,definedasp 2ln( 0/ max)where max ( 0)de- the NII for valuable computing and Super-SINET net-
− L L L L
notesthelikelihoodvalueatthemaximum(withthesig- worksupport. WeacknowledgesupportfromMEXTand
nal yield fixed at zero), is 8.1σ. JSPS (Japan); ARC and DEST (Australia); NSFC and
KIP of CAS (contract No. 10575109 and IHEP-U-503,
China); DST (India); the BK21 program of MOEHRD,
TABLEI:Systematicerrorsforthecombinedbranchingfrac-
and the CHEP SRC and BR (grant No. R01-2005-
tion.
000-10089-0) programs of KOSEF (Korea); KBN (con-
Systematic errors(%)
tract No. 2P03B 01324,Poland); MIST (Russia); ARRS
PDFs 4.8
(Slovenia); SNSF (Switzerland); NSC and MOE (Tai-
Lepton identification 3.9
wan); and DOE (USA).
Charged pion identification 0.4
Tracking (lepton and charged pion) 2.7
η→γγ, π0 selection 4.1
B(J/ψ→ℓ+ℓ−) 1.0
B(η→γγ and π+π−π0) 1.1 [1] P.Averyetal.(CLEOCollab.),Phys.Rev.D62,051101
NBB¯ 1.2 (2000).
Total 8.1 [2] B. Aubert et al. (BaBar Collab.), Phys. Rev. D 65,
032001 (2002).
[3] K. Abe et al. (Belle Collab.), Phys. Rev. D 67, 032003
(2003).
TABLE II: Detection efficiencies and branchingfractions. [4] B. Aubert et al. (BaBar Collab.), Phys. Rev. Lett. 91,
061802 (2003).
Mode Eff.(%) BF(10−6) [5] S.U.Kataokaetal.(BelleCollab.),Phys.Rev.Lett.93,
B0→J/ψη(γγ) 21.9 9.6 ± 1.9 261801 (2004).
B0→J/ψη(π+π−π0)1 13.8 10.0 ± 3.5 [6] B. Aubert et al. (BaBar Collab.), Phys. Rev. D 74,
B0→J/ψη (combined) 9.5 ± 1.7 ± 0.8 011101 (2006).
[7] B. Aubert et al. (BaBar Collab.), Phys. Rev. Lett. 91,
071801 (2003).
In summary, the first observation of the decay B0
[8] Inclusion of charge conjugate modes is implied.
→
J/ψη is reported. We observe 59.7 10.5 signal events [9] S.KurokawaandE.Kikutani,Nucl.Instr.and.Meth.A
with8.1σsignificanceina449million±BB¯ pairdatasam- 499,1(2003), andotherpapersincludedinthisvolume.
ple at the Υ(4S) resonance. The measured branching [10] A.Abashianetal.(BelleCollab.),Nucl.Instr.andMeth.
fraction is (B0 J/ψ η) = (9.5 1.7 0.8) 10−6, A 479, 117 (2002).
B → ± ± × [11] Z. Natkaniec (Belle SVD2 Group), Nucl. Instr. and
where the first error is statistical and the second is sys-
Meth.A 560, 1 (2006).
tematic. The measured branching fraction is consis-
[12] R. Brun et al.,GEANT3.21, CERN DD/EE/84-1,1984.
tent with the theoretical expectation [17] and compa-
[13] TheFox-WolframmomentswereintroducedinG.C.Fox
rable to the world-average value for the J/ψπ0 mode: and S.Wolfram, Phys. Rev.Lett. 41, 1581 (1978).
(2.2 0.4) 10−5 [16]. There is no indication of either [14] Forthisoptimization,weassumeabranchingfractionfor
large±final s×tate interactions or non-standardmodel con- J/ψ η that is about one third that for J/ψ π0.
tributions. [15] DESY Internal Report,DESY F31-86-02(1990).
[16] W.-M. Yaoet al., (PDG),J. Phys.G 33, 1 (2006).
We thank the KEKB group for excellent operation of
[17] A. Deandrea et al.,Phys. Lett. B 318, 549 (1993).
the accelerator, the KEK cryogenics group for efficient
solenoid operations, and the KEK computer group and