Search for lepton-flavor-violating $τ\to\ell V^0$ decays at Belle

0.28 MB·English

by  Y. Nishio

Search for lepton-flavor-violating τ → ℓV0 decays at Belle
Y. Nishio,20 K. Inami,20 T. Ohshima,20 I. Adachi,7 H. Aihara,41 K. Arinstein,1 V. Aulchenko,1 T. Aushev,16,11
T. Aziz,37 A. M. Bakich,36 V. Balagura,11 E. Barberio,19 I. Bedny,1 K. Belous,9 U. Bitenc,12 S. Blyth,23
A. Bondar,1 M. Braˇcko,7,18,12 T. E. Browder,6 A. Chen,22 W. T. Chen,22 B. G. Cheon,5 I.-S. Cho,45 Y. Choi,35
J. Dalseno,19 M. Danilov,11 M. Dash,44 A. Drutskoy,3 S. Eidelman,1 D. Epifanov,1 N. Gabyshev,1 B. Golob,17,12
H. Ha,14 J. Haba,7 K. Hara,20 K. Hayasaka,20 H. Hayashii,21 M. Hazumi,7 Y. Horii,40 Y. Hoshi,39
W.-S. Hou,24 Y. B. Hsiung,24 T. Iijima,20 A. Ishikawa,32 R. Itoh,7 M. Iwasaki,41 Y. Iwasaki,7 N. J. Joshi,37
D. H. Kah,15 H. Kaji,20 J. H. Kang,45 N. Katayama,7 H. Kawai,2 T. Kawasaki,27 H. Kichimi,7 Y. J. Kim,4
8 S. Korpar,18,12 Y. Kozakai,20 P. Kriˇzan,17,12 P. Krokovny,7 R. Kumar,31 C. C. Kuo,22 Y. Kuroki,30 A. Kuzmin,1
0 Y.-J. Kwon,45 J. S. Lee,35 M. J. Lee,34 S. E. Lee,34 T. Lesiak,25 A. Limosani,19 Y. Liu,4 D. Liventsev,11
0 F. Mandl,10 S. McOnie,36 H. Miyake,30 H. Miyata,27 Y. Miyazaki,20 R. Mizuk,11 G. R. Moloney,19 T. Mori,20
2
E. Nakano,29 M. Nakao,7 H. Nakazawa,22 S. Nishida,7 O. Nitoh,43 S. Ogawa,38 S. Okuno,13 H. Ozaki,7 P. Pakhlov,11
n
G. Pakhlova,11 C. W. Park,35 H. Park,15 R. Pestotnik,12 L. E. Piilonen,44 A. Poluektov,1 Y. Sakai,7
a
J O. Schneider,16 K. Senyo,20 M. E. Sevior,19 M. Shapkin,9 V. Shebalin,1 H. Shibuya,38 J.-G. Shiu,24 B. Shwartz,1
6 J. B. Singh,31 A. Sokolov,9 A. Somov,3 S. Staniˇc,28 M. Stariˇc,12 T. Sumiyoshi,42 F. Takasaki,7 M. Tanaka,7
1 G. N. Taylor,19 Y. Teramoto,29 I. Tikhomirov,11 S. Uehara,7 K. Ueno,24 T. Uglov,11 Y. Unno,5 S. Uno,7

Y. Usov,1 G. Varner,6 S. Villa,16 A. Vinokurova,1 C. H. Wang,23 P. Wang,8 X. L. Wang,8 Y. Watanabe,13
]
x E. Won,14 Y. Yamashita,26 Z. P. Zhang,33 V. Zhilich,1 V. Zhulanov,1 A. Zupanc,12 and O. Zyukova1
e
- (The Belle Collaboration)
p
1Budker Institute of Nuclear Physics, Novosibirsk, Russia
e
2Chiba University, Chiba, Japan
h
[ 3University of Cincinnati, Cincinnati, OH, USA
4The Graduate University for Advanced Studies, Hayama, Japan

1 5Hanyang University, Seoul, South Korea
v 6University of Hawaii, Honolulu, HI, USA
5 7High Energy Accelerator Research Organization (KEK), Tsukuba, Japan
7 8Institute of High Energy Physics, Chinese Academy of Sciences, Beijing, PR China
4 9Institute for High Energy Physics, Protvino, Russia
2
10Institute of High Energy Physics, Vienna, Austria
.
1 11Institute for Theoretical and Experimental Physics, Moscow, Russia
0 12J. Stefan Institute, Ljubljana, Slovenia
8 13Kanagawa University, Yokohama, Japan
0 14Korea University, Seoul, South Korea
v: 15Kyungpook National University, Taegu, South Korea
i 16E´cole Polytechnique F´ed´erale de Lausanne, EPFL, Lausanne, Switzerland
X 17Faculty of Mathematics and Physics, University of Ljubljana, Ljubljana, Slovenia
r 18University of Maribor, Maribor, Slovenia
a 19University of Melbourne, Victoria, Australia
20Nagoya University, Nagoya, Japan
21Nara Women’s University, Nara, Japan
22National Central University, Chung-li, Taiwan
23National United University, Miao Li, Taiwan
24Department of Physics, National Taiwan University, Taipei, Taiwan
25H. Niewodniczanski Institute of Nuclear Physics, Krakow, Poland
26Nippon Dental University, Niigata, Japan
27Niigata University, Niigata, Japan
28University of Nova Gorica, Nova Gorica, Slovenia
29Osaka City University, Osaka, Japan
30Osaka University, Osaka, Japan
31Panjab University, Chandigarh, India
32Saga University, Saga, Japan
33University of Science and Technology of China, Hefei, PR China
34Seoul National University, Seoul, South Korea
35Sungkyunkwan University, Suwon, South Korea
36University of Sydney, Sydney, NSW, Australia
2
37Tata Institute of Fundamental Research, Mumbai, India
38Toho University, Funabashi, Japan
39Tohoku Gakuin University, Tagajo, Japan
40Tohoku University, Sendai, Japan
41Department of Physics, University of Tokyo, Tokyo, Japan
42Tokyo Metropolitan University, Tokyo, Japan
43Tokyo University of Agriculture and Technology, Tokyo, Japan
44Virginia Polytechnic Institute and State University, Blacksburg, VA, USA
45Yonsei University, Seoul, South Korea
Wehavesearchedforneutrinolessτ leptondecaysintoℓandV0,whereℓstandsforanelectronor
muon,and V0 for avectormeson (φ,ω,K∗0,K¯∗0 orρ0),using543 fb−1 of datacollected with the
Belle detector at the KEKB asymmetric-energy e+e− collider. No excess of signal events over the
expected background has been observed, and we set upperlimits on the branching fractions in the
range(5.9−18)×10−8 at the90% confidencelevel. These upperlimits includethefirst results for
the ℓω mode as well as new limits that are significantly more restrictive than our previous results
for theℓφ, ℓK∗0, ℓK¯∗0 and ℓρ0 modes.
PACSnumbers: 11.30.Fs,13.35.Dx,14.60.Fg
INTRODUCTION calorimeter comprised of CsI(Tl) crystals located inside
a superconducting solenoid coil that provides a 1.5 T
magnetic field. An iron flux-return located outside the
In the Standard Model (SM), lepton-flavor-violating
coilis instrumentedto detect K0 mesons andto identify
(LFV) decays of charged leptons are forbidden; even if L
muons. The detector is described in detail elsewhere [7].
neutrino mixing is taken into account, they are highly
Two inner detector configurations were used. A 2.0 cm
suppressed. However, LFV is expected to appear in
radius beam-pipe and a 3-layer silicon vertex detector
many extensions of the SM. Some such models pre-
wereusedforthefirstsampleof158fb−1,whilea1.5cm
dict branching fractions for τ LFV decays in the range
radius beam-pipe, a 4-layer silicon detector and a small-
10−8−10−7[1,2,3],whichcanbereachedatthepresent
cellinnerdriftchamberwereusedtorecordtheremaining
B-factories. AnobservationofLFVwouldprovideunam-
385 fb−1 [10].
biguous evidence for new physics beyond the SM.
A search for LFV τ− decays into neutrinoless final
stateswithonechargedleptonℓ−(e− orµ−)andavector EVENT SELECTION
mesonwasfirstperformedbytheCLEOcollaborationin
theℓ−φ,ℓ−K∗0,ℓ−K¯∗0andℓ−ρ0finalstatesin1998;this
We search for events, in which one τ decays to a
search set upper limits in the range (2.0−7.5)×10−6
charged lepton and two charged hadrons (3-prong de-
[4]. Later, Belle obtained upper limits in the range
cay) while the other τ decays into one charged particle
(2.0−7.7)×10−7 using 158 fb−1 of data. In this paper,
(1-prong decay) and missing neutral particle(s). We re-
wereportanimprovedsearchforLFVτ− decays[5]into
construct φ candidates from K+K−, ω from π+π−π0,
achargedleptonandaneutralvectormeson(V0),where
K∗0 from K+π−, K¯∗0 from K−π+ and ρ0 from π+π−.
V0 includes ω in addition to the φ, K∗0, K¯∗0 and ρ0 [†]
The selection criteria described below are optimized
. The analysis is based on a data sample of 543 fb−1,
from studies of Monte Carlo (MC) simulated events and
corresponding to 4.99 × 108 τ-pairs collected with the
the experimental data distributions. The background
Belle detector[7]atthe KEKBasymmetric-energye+e−
estimation is based on MC simulations of the reaction
collider [8] taken at the Υ(4S) resonance and 60 MeV
e+e− → τ+τ− as well as qq¯continuum and two-photon
below it. These resultssupersede our previouspublished
processes. The τ+τ− samplecorrespondingto1524fb−1
results [9].
is generated using the KKMC code [11]. The MC sam-
TheBelledetectorisalarge-solid-anglemagneticspec-
ples of qq¯and two-photon processes are produced using
trometer that consists of a silicon vertex detector, a 50-
EvtGen [12] and AAFH [13], respectively, in amounts
layercentraldriftchamber,anarrayofaerogelthreshold
corresponding to the luminosity of the experiment. The
Cherenkov counters, a barrel-like arrangement of time-
signal MC events are generated by KKMC assuming a
of-flight scintillation counters, and an electromagnetic
phase-space distribution for τ decay. The detector re-
sponse is simulated by a GEANT3 [14] based program.
The transverse momentum for each charged track is
requiredtobelargerthan0.06GeV/cinthebarrelregion
[†]While preparing this paper, we became aware that the BaBar
group had also reported on a search for τ− → ℓ−ω in a (−0.6235 < cosθ < 0.8332, where θ is the polar angle
preprint[6]. relative to the direction opposite to that of the incident
3
e+ beam in the laboratory frame) and 0.1 GeV/c in the Mπ+π− < 1.074 GeV/c2 (±4σ), requiring that the
endcap region(−0.8660<cosθ <−0.6235and 0.8332< daughter pions have P < 0.1, P < 0.1 and momenta
K e
cosθ < 0.9563). The energies of photon candidates are greater than 0.5 GeV/c. In addition, for the τ− →e−ρ0
required to be larger than 0.1 GeV in both regions. mode,werequireP <0.5fordaughterpionsinorderto
µ
To select the signal topology, we require four charged reduce the two-photon background from ee→eeµµ.
tracksinaneventwithzeronetcharge,andthetotalen- Figures1(a)-(d)showtheinvariantmassdistributions
ergyofchargedtracksandphotons inthe center-of-mass of the φ, ω, K∗0 and ρ0 candidates for the τ− → µ−φ,
(CM) frame to be less than 11 GeV. We also require the τ− → µ−ω, τ− → µ−K∗0 and τ− → µ−ρ0 modes, re-
missing momentum in the laboratory frame to exceed spectively. Theestimatedbackgrounddistributionsagree
0.6 GeV/c, and to point into the detector acceptance withthedata. Themainbackgroundcontributionforthe
(−0.8660 < cosθ < 0.9563). Here the missing momen- τ− →ℓ−φ mode is due to qq¯events involving φ mesons.
tum is defined as the difference between the momentum Fortheτ− →ℓ−ωmodethedominantbackgroundcomes
of the initial e+e− system, and the sum of the observed from τ− → π−ων decay with the pion misidentified as
τ
momentum vectors. An eventis subdivided into 3-prong alepton. Theτ− →π−π+π−ν decayisoneofthe main
τ
and 1-prong hemispheres with respect to the thrust axis background sources for the τ− → ℓ−K∗0, ℓ−K¯∗0 and
calculated from the momenta of all charged tracks and τ− → ℓ−ρ0 modes. In this background source one pion
photonsintheCMframe. Thesehemispheresarereferred is misidentified as a lepton for all modes while for the
to as the signal and tag sides, respectively. We allow at τ− → ℓ−K∗0 and ℓ−K¯∗0 modes one additional pion is
most two photons on the tag side to take into account misidentified as a kaon.
initial state radiation. To reduce the qq¯background,not
more than one photon on the signal side is allowed for
theℓ−φ,ℓ−K∗0,ℓ−K¯∗0,ℓ−ρ0modeswhilenotmorethan 2V/c 10 (a) 2V/c220205 (b)
twophotonsinadditiontoπ0daughtersarepermittedfor Me 8 Me175
eetht,eeKr,ℓ−Poωrx.πm)Tohidsiesisdi.senAdteicfifihenaderdguesadisnpgPaxtrht=eiclLleikxoe/fl(ihPthoexodtLyxrpa)e,tiwxoh(pxearre=aLmµx-, vents / 2.0 46 events / 1311170255050
risestphoenlsikeseloihfotohde rfoerlepvaanrtticdleetteycptoersx,[1d5e]t.eFrmorinmeduofnrocmantdhie- # of e 2 # of 2550
0 0
dates on the signalside we require P > 0.95 while their 0.99 1 1.011.021.031.041.05 0.6 0.7 0.8 0.9 1
µ M (Gev/c2) M 0(Gev/c2)
momentum should be greater than 1.0 GeV/c. The effi- KK ppp
ciency formuonidentificationis 92%witha 1.2%proba- 2c (c) 2c160 (d)
bility to misidentify a pion as a muon. Electrons on the V/ 30 V/140
e e
swgiadirgipeetCnnahioaattninelfiridsncaiitavdsdhateaaiarotninaenarne0eφilts.e5rc9metmG4qre%uoaesnisoVrswnei/dsshci.ni0atl.eor1Tet%thhhhsa.eeeevlepeerfficraPotcnebeidgea>enbfcirly01oit..m9y0fo1atrKontGdh+meeKmiVsei−o/dlemcec2pnetatrniiotfr<nyas # of events / 23M 112250505 # of events / 43 M11246802000000
0 0
MK+K− <1.03GeV/c2 (±4σ). Forbothkaondaughters 0.60.70.80M.9 1 (1G.1e1v./2c21).3 0.4 0.6 0.8M1 1(.G2e1v.4/c21).6
werequireP >0.8. Toreducethebackgroundfromthe Kp pp
K
γ →e+e− conversions the cut P <0.1 is applied.
e FIG. 1: Invariant mass distributions of (a) φ → K+K− for
Candidate ω mesons are reconstructed from π+π−π0 τ− → µ−φ, (b) ω → π+π−π0 for τ− → µ−ω, (c) K∗0 →
with the invariant mass requirement 0.757 GeV/c2 < K+π− for τ− →µ−K∗0 and (d) ρ0 →π+π− for τ− →µ−ρ0
Mπ+π−π0 < 0.808 GeV/c2 (±3σ). A π0 candidate is in the region 1.5 GeV/c2 < MℓV0 < 1.95 GeV/c2 and −0.5
selected from γ pairs with invariant mass in the range GeV< ∆E < 0.5 GeV. The points with error bars are data.
0.11GeV/c2 <M <0.15GeV/c2. Inordertoimprove The open histogram shows the expected τ+τ− background
γγ
the ω mass resolution, the π0 mass is constrained to its MCwhilethefilledhistogramisthesumofqq¯andtwo-photon
world average value of 134.9766 GeV/c2 for the ω mass MCs. The regions between thevertical lines are selected.
reconstruction.
Candidate K∗0 and K¯∗0 mesons are selected from To reduce the remaining background from τ+τ− and
K±π∓ pairs with invariant mass in the range qq¯, the events from the triangular area defined by the
0.827 GeV/c2 < MKπ < 0.968 GeV/c2 (±3σ), which missing momentum, pmiss (GeV/c), and missing mass
satisfy the condition P > 0.8 for the kaon candidate squared, m2 ((GeV/c2)2) are selected for further con-
K miss
and Pe <0.1 for both daughters. sideration. These requirements are summarized in Ta-
Candidate ρ0 mesons are selected from π+π− pairs ble I and illustrated in Fig. 2 by the two-dimensional
with invariant mass in the range 0.478 GeV/c2 < plots of p (GeV/c) versus m2 ((GeV/c2)2) for the
miss miss
4
direction measured by the calorimeter.
TABLE I: Selection criteria using pmiss (GeV/c) and m2miss The resolutions in ∆E and M , evaluated using the
((GeV/c2)2) where pmiss is missing momentum and m2miss is signal MC, are summarized in TℓVab0le II. We define the
missing mass squared.
signal region in the ∆E −M plane as a ±3σ ellipse.
ℓV0
In order to avoid biases in the event selection, we blind
Mode Selection criteria
τ− → the signal region until the analysis is finalized.
ℓ−φ pmiss> 89m2miss and m2miss>−0.5
ℓµ−−ωK∗0 ppmmiissss>> 8348.m5m2mi2mssis−s−8389 aanndd pmm2misisss>>8−m02m.5iss TTAheBsLuEpeIrIs:cRriepstoslulotiwonasnidnhMigℓhV0inidnicMateeVt/hc2ealonwde∆r aEndinhMigehVer.
e−K∗0 pmiss> 58.5m2miss− 181 and m2miss>0 sides of thepeak, respectively.
µ−K¯∗0 pmiss> 68.5m2miss and m2miss>−0.5
eµ−−Kρ¯0∗0 ppmmiissss>>−65m8m2m2mississ−4 aanndd ppmmiissss>>−2m18.2m4imss2miss Mτ−od→e σMhigℓVh0 σMlowℓV0 σ∆hiEgh σ∆lowE
e−ρ0 pmiss>−8m2miss−4 and pmiss>1.6m2miss µ−φ 3.4±0.2 3.4±0.2 13.2±0.4 14.0±0.5
e−φ 3.7±0.1 3.6±0.1 13.3±0.7 15.4±0.7
µ−ω 5.9±0.1 6.2±0.1 19.3±0.6 30.3±0.8
V/c) 8 signal V/c) 8 data e−ω 6.1±0.1 6.5±0.1 20.4±0.7 32.5±1.3
Ge 7 Ge 7 µ−K∗0 4.5±0.4 4.5±0.4 13.8±0.3 14.4±0.4
(miss6 (miss6 e−K∗0 4.3±0.1 5.1±0.1 12.9±0.3 18.0±0.4
p 5 p 5 µ−K¯∗0 4.7±0.1 4.4±0.1 14.0±0.3 15.0±0.3
4 4 e−K¯∗0 4.6±0.1 4.9±0.1 12.6±0.6 17.8±0.5
3 3 µ−ρ0 5.6±0.1 5.0±0.1 13.9±0.4 15.9±0.4
2 2 e−ρ0 4.7±0.1 6.3±0.1 14.5±0.4 17.4±0.5
1 1
0-10 -7.5 -5 -2.5 0 2.5 5 7.5 10 0-10 -7.5 -5 -2.5 0 2.5 5 7.5 10
m2miss ((GeV/c2)2) m2miss ((GeV/c2)2)
BACKGROUND ESTIMATION
FIG. 2: pmiss vs. m2miss plots for signal MC and data for the
τ → µρ0 mode. The regions between the vertical lines are
After all selections, a few events remain in the region
selected.
−10σlow < M < 10σhigh and −10σlow < ∆E <
MℓV0 ℓV0 MℓV0 ∆E
3σhigh, which we define as a background region. This
τ− →µ−ρ0 mode. ∆E
regionwill be used to estimate the expected background
For the ℓ−ω (ℓ−K∗0 and ℓ−K¯∗0) mode, we require
and is shown in Fig. 3. Table III lists the numbers of
that the opening angle between the lepton and ω (K∗0)
events in the background region excluding the signal re-
on the signal side in the CM frame, θCM (θCM ), sat-
ℓω ℓK∗0 gion. The background is efficiently suppressed by the
isfy cosθCM < 0.88 (cosθCM < 0.93). To remove two-
ℓω ℓK∗0 event selection. The comparison between the data and
photonbackgroundfortheeV0modes,wefurtherrequire
MCshowsreasonableagreementforallmodes,exceptfor
that the opening angle, α, between the direction of the
total momentum of charged tracks and γ’s on the sig-
nal side and that on the tag side satisfy the condition
TABLE III: Number of events in the background region ex-
cosα>−0.999 for the e−φ, cosα>−0.996 for the e−ω,
cluding thesignal region for data and MC.
e−K∗0(K¯∗0) and cosα>−0.990 for the e−ρ0 mode.
To identify signalτ decays,we reconstructthe ℓV0 in-
Mode
variantmass,MℓV0,andthe energydifferenceintheCM τ− → Data MC
frame, ∆E, between the sum of energies on the signal µ−φ 2 1.72±1.23
side and the beam-energy, Ebeam. Signal events should e−φ 2 0±1.19
0
concentrate around MℓV0 = mτ and ∆E = 0, where µ−ω 7 10.46±1.91
mτ is the nominal τ mass. For the ℓω modes, we used e−ω 0 1.07±0.62
the beam-energy constrained mass, Mbc, instead of the µ−K∗0 3 3.78±1.72
invariant mass Minv, where Mbc = pEb2eam−(p~τ)2, in e−K∗0 1 2.26±1.26
order to improve the mass resolution, which is smeared µ−K¯∗0 2 1.21±0.92
due to the γ energy resolution. In calculating the τ mo- e−K¯∗0 0 0.31±0.31
mentum p~ , we replacethe magnitude of the π0 momen- µ−ρ0 12 4.82±1.25
τ
tum with the value obtained from the beam energy, the e−ρ0 0 0.36±0.36
energies of charged tracks on the signal side and the π0
5
E(GeV) 0.1 t - → m - f E(GeV) 0.1 t - → e- f tnhaentτb−ac→kgµro−uρn0dmsooudrec.esFiosrqtq¯h,iswhmicohdeatonloewomf tuhlteipdloicmitiy-
D0.05 D0.05 is poorly described by MC. For all other modes we es-
timate the number of background events in the signal
0 0 ellipse from the data in the backgroundregion using the
-0.05 following method.
-0.05
Sinceformostofthemodestherearesingleeventsonly
-0.1
-0.1 in the ∆E − M plane both in the data and MC, we
lV0
-0.15 first study the backgrounddistribution in a sideband re-
1.751.761.771.781.79 1.8 1.81 1.751.761.771.781.79 1.8 1.81
Mmf (GeV/c2) Mef(GeV/c2) gion larger than the background region (1.5 GeV/c2 <
GeV)0.15 t - → m - w GeV) 0.2 t - → e- w MℓV0 < 1.95 GeV/c2 and −0.5 GeV< ∆E < 0.5 GeV)
E( 0.1 E( 0.1 and find that the distribution of events in it is approxi-
D D
0.05 mately flat in the background region. Therefore we as-
0 0 sume that this distribution is flat inside the background
-0.05 regionand estimate the expected number of background
-0.1 -0.1
eventsinthesignalregionfromthenumberofdataevents
-0.15
-0.2 -0.2 in the background region times the ratio of the areas of
-0.25 the signal ellipse and background region. If the number
-0.3
1.72 1.74 1.76 1.78 1.8 1.82 1.72 1.74 1.76 1.78 1.8 1.82 ofdataeventsinthebackgroundregioniszero,weassign
Mmw (GeV/c2) Mew(GeV/c2) anupperlimitof2.44eventsatthe90%confidencelevel.
E(GeV) 0.1 t - → m - K*0 E(GeV) 0.1 t - → e- K*0 TthhisemexepthecotdedisnsuhmowbenrionftbhaeckthgirroduncodluemvenntosfoTbatbalieneIdV.by
D D0.05
0.05 For the µ−ω mode, where the number of events is
0 larger, we estimate the background contribution in the
0
signalregionusing the shape of the backgroundMC dis-
-0.05
-0.05 tribution normalized to the data yield in the sideband
-0.1
region of the backgroundregion.
-0.1
-0.15 For the µ−ρ0 mode, the background events in Fig.3
1.74 1.76 1.78 1.8 1.82 1.74 1.76 1.78 1.8 1.82 come mostly from τ− → π−π+π−ν decay and qq¯ when
MmK*(GeV/c2) MeK*(GeV/c2) one of the pions is misidentified as a muon. To estimate
eV) t - → m - K–*0 eV) 0.1 t - → e- K–*0 the background contribution we select a special event
E(G 0.1 E(G samplerequiringP <0.1formuoncandidatesinsteadof
D D0.05 µ
0.05 P > 0.95. The number of expected background events
µ
0
is then calculated from the product of the number of
0
-0.05 events with Pµ < 0.1 in the signal region and the muon
-0.05 fake rate.
-0.1
-0.1
-0.15
-0.15 1.74 1.76 1.78 1.8 1.82 1.74 1.76 1.78 1.8 1.82 RESULTS
MmK–*(GeV/c2) MeK–*(GeV/c2)
eV) t - → m - r 0 eV) t - → e- r 0
E(G 0.1 E(G 0.1 Afterunblindingthesignalregionsingleeventsonlyre-
D D main in some modes, see Fig.3. The observednumber of
0.05 0.05
eventsinthesignalregionisconsistentwiththeexpected
0 0 background. Fromthenumbersofobservedeventsinthe
-0.05 signal region and the numbers of expected background
-0.05
events, listed in the second and third columns of Ta-
-0.1
-0.1 ble IV, respectively, we evaluate the upper limit on the
-0.15
-0.15 number of signal events at the 90% CL, s90, with sys-
1.74 1.76 1.78 1.8 1.82 1.72 1.74 1.76 1.78 1.8 1.82
Mmr (GeV/c2) Mer(GeV/c2) tematic uncertainties included in the Feldman-Cousins
method [16] using the POLE code [17]. In the cases
FIG. 3: Distributions of ∆E - MℓV0 in the ±10σ box for (a) whenwe giveanupper limit ofthe expectedbackground
τ− → µ−φ, (b) τ− → e−φ, (c) τ− → µ−ω, (d) τ− → e−ω, (τ →e−ω, e−K¯∗ and e−ρ0 modes), the number of back-
(e) τ− → µ−K∗0, (f) τ− → e−K∗0, (g) τ− → µ−K¯∗0, (h)
ground events is taken to be zero. This results in con-
τ− → e−K¯∗0, (i) τ− → µ−ρ0 and (j) τ− → e−ρ0 after all
servative upper limits. The main systematic uncertain-
selections. Dots are data and filled boxes show the signal
ties on the detection efficiency come from track recon-
MC. The elliptical area is the3σ signal region.
struction(1.0%pertrack),electronidentification(2.2%),
6
muon identification (2.0%), kaon/pion separation (1.4% efficiency is improved, e.g. it increases by a factor of 2.8
for φ reconstruction, 1.1% for K∗0 and 1.5% for ρ0), π0 for the µ−φ and 2.5 for the e−φ mode. The new upper
reconstruction (4.0%), statistics of the signal MC (1.3% limits can be used to constrain the parameter space of
for ℓ−φ, 0.7%forℓ−ω,0.6%for ℓ−K∗0 andℓ−K¯∗0, 0.5% various scenarios beyond the SM.
for µ−ρ0 and 0.6% for e−ρ0) and uncertainties in the
branching fractions for φ → K+K− and ω → π+π−π0
(1.2% and 0.8%). The uncertainty in the number of τ- ACKNOWLEDGEMENTS
pair events mainly comes from the luminosity measure-
ment (1.6%).
We thank the KEKB group for the excellent opera-
Theupperlimitsonthebranchingfractions,B,arecal- tionoftheaccelerator,theKEKcryogenicsgroupforthe
culatedasB < 2Ns9τ0τǫ,whereNττ =4.99×108,isthetotal efficient operation of the solenoid, and the KEK com-
number of the τ-pairs produced and ǫ is the signal effi- puter group and the National Institute of Informatics
ciency including the branching fractions of φ→K+K−, for valuable computing and Super-SINET network sup-
ω →π+π−π0, K∗0 → K+π− and ρ0 →π+π− [18]. The port. We acknowledge support from the Ministry of
resultingupperlimitsonthebranchingfractionsaresum- Education, Culture, Sports, Science, and Technology of
marized in Table IV. Japan and the Japan Society for the Promotion of Sci-
ence;theAustralianResearchCouncilandtheAustralian
DepartmentofEducation,Science andTraining;the Na-
TABLEIV:SummaryofthenumberofobservedeventsN ,
obs
the number of expected background events Nexp, detection tional Natural Science Foundation of China under con-
efficiency ǫ,total systematic error ∆ǫ/ǫ, 90% CL upperlimit tractNo.10575109and10775142;theDepartmentofSci-
ofthenumberofsignaleventss90 and90%CLupperlimitof ence and Technology of India; the BK21 program of the
thebranching fractions B. MinistryofEducationofKorea,theCHEPSRCprogram
and Basic Research program (grant No. R01-2005-000-
Mode Nobs Nexp ǫ ∆ǫ/ǫ s90 UL on B 10089-0) of the Korea Science and Engineering Founda-
τ− → (%) (%) (90% CL)
tion,andthe PureBasicResearchGroupprogramofthe
µ−φ 1 0.17±0.12 3.14 5.2 4.17 1.3×10−7
KoreaResearchFoundation;thePolishStateCommittee
e−φ 0 0.18±0.12 3.10 5.3 2.27 7.3×10−8
for Scientific Research; the Ministry of Education and
µ−ω 0 0.19±0.20 2.51 6.3 2.22 8.9×10−8
Science of the Russian Federation and the Russian Fed-
e−ω 1 <0.24 2.46 6.3 4.34 1.8×10−7
eral Agency for Atomic Energy; the Slovenian Research
µ−K∗0 0 0.26±0.15 3.71 4.8 2.20 5.9×10−8
Agency; the Swiss National Science Foundation; the Na-
e−K∗0 0 0.08±0.08 3.04 4.9 2.35 7.8×10−8
µ−K¯∗0 1 0.17±0.12 4.02 4.8 4.14 1.0×10−7 tional Science Council and the Ministry of Education of
e−K¯∗0 0 <0.17 3.21 4.9 2.45 7.7×10−8 Taiwan; and the U.S. Department of Energy.
µ−ρ0 1 1.04±0.28 4.89 4.9 3.34 6.8×10−8
e−ρ0 0 <0.17 3.94 5.1 2.46 6.3×10−8
[1] A. Ilakovac, Phys.Rev.D 62, 036010 (2000).
[2] A. Brignole and A.Rossi, Nucl. Phys.B 701, 3 (2004).
[3] C.-H. Chen and C.-Q. Geng, Phys. Rev. D 74, 035010
SUMMARY (2006).
[4] D.W. Bliss et al. (CLEO Collaboration), Phys. Rev. D
57, 5903 (1998).
Wehavesearchedforthe LFVdecaysτ− →ℓ−φ,ℓ−ω, [5] Throughout this paper, the inclusion of the charge-
ℓ−K∗0, ℓ−K¯∗0 andℓ−ρ0 using543fb−1 ofdataobtained conjugatedecaymodesisimpliedunlessotherwisestated.
in the Belle experiment. No evidence for a signal is ob- [6] B.Aubertetal.(BaBarCollaboration), arXiv:0711.0980
[hep-ex].
served, and the upper limits on the branching fractions
[7] A.Abashianetal.(BelleCollaboration),Nucl.Instr.and
are set in the range (5.9−18)×10−8 at the 90% CL.
Meth. A 479, 117 (2002).
This analysis is the first search for the τ− →ℓ−ω mode.
[8] S. Kurokawa and E. Kikutani, Nucl. Instr. and Meth. A
The results for the τ− → ℓ−φ, ℓ−K∗0, ℓ−K¯∗0 and ℓ−ρ0 499,1(2003), andotherpapersincludedinthisvolume.
modes are 3−10 times more restrictive than our previ- [9] Y. Yusa et al. (Belle Collaboration), Phys. Lett. B 640,
ous results obtained using 158 fb−1 of data. The sen- 138 (2006).
sitivity improvement comes from a factor of 3.4-times [10] Z.Natkaniecet al.(BelleSVD2Group),Nucl.Instr.and
Meth. A 560, 1 (2006).
larger statistics and an optimized analysis. In particu-
[11] S.Jadach,B.F.L.Ward,Z.Wa¸s,Comp.Phys.Commun.
lar, we have improved the conditions on p and m2
miss miss 130, 260 (2000).
to reduce ττ and qq¯background as well as those on the [12] D. J. Lange, Nucl. Instr.and Meth. A 462, 152 (2001).
opening angleα to reduce two-photonbackground. As a [13] F.A. Berends, P.H. Daverveldt, R. Kleiss, Comp. Phys.
result,betterbackgroundsuppressionisachievedandthe Commun. 40, 285 (1986).
7
[14] CERNProgramLibraryLongWriteupNo.W5013,1993. [17] TheHighland-Cousinsmethodforcalculatingupperlim-
[15] K. Hanagaki et al., Nucl. Instr. and Meth. A 485, 490 itswithsystematicerrorsisdescribedinJ.Conradetal.,
(2002);A.Abashianetal.,Nucl.Instr.andMeth.A491, Phys. Rev.D 67, 012002 (2003).
69(2002); E.Nakano,Nucl.Instr.andMeth. A494,117 [18] W.-M. Yao et al. (Particle Data Group), J. Phys. G 33,
(2002). 1 (2006).
[16] G.J.FeldmanandR.D.Cousins,Phys.Rev.D57,3873
(1998).