Observation of the decay Bs0 → K 0 K 0

B. Pal,7 A. J. Schwartz,7 A. Abdesselam,64 I. Adachi,15, 12 H. Aihara,71 D. M. Asner,54 T. Aushev,42, 25

R. Ayad,64 T. Aziz,65 V. Babu,65 I. Badhrees,64, 30 S. Bahinipati,17 A. M. Bakich,63 E. Barberio,40 P. Behera,19
V. Bhardwaj,61 B. Bhuyan,18 J. Biswal,26 A. Bobrov,4, 52 A. Bozek,49 M. Bračko,38, 26 T. E. Browder,14
D. Červenkov,5 V. Chekelian,39 A. Chen,46 B. G. Cheon,13 R. Chistov,25 K. Cho,31 V. Chobanova,39 Y. Choi,62
D. Cinabro,77 J. Dalseno,39, 66 N. Dash,17 Z. Doležal,5 Z. Drásal,5 A. Drutskoy,25, 41 D. Dutta,65 S. Eidelman,4, 52
H. Farhat,77 J. E. Fast,54 B. G. Fulsom,54 V. Gaur,65 A. Garmash,4, 52 R. Gillard,77 Y. M. Goh,13
P. Goldenzweig,28 D. Greenwald,67 O. Grzymkowska,49 J. Haba,15, 12 T. Hara,15, 12 K. Hayasaka,44
arXiv:1512.02145v3 [hep-ex] 20 Apr 2016

H. Hayashii,45 X. H. He,55 W.-S. Hou,48 K. Inami,43 A. Ishikawa,69 Y. Iwasaki,15 W. W. Jacobs,20 I. Jaegle,14

H. B. Jeon,33 D. Joffe,29 K. K. Joo,6 T. Julius,40 K. H. Kang,33 E. Kato,69 T. Kawasaki,50 C. Kiesling,39
D. Y. Kim,60 H. J. Kim,33 K. T. Kim,32 M. J. Kim,33 S. H. Kim,13 K. Kinoshita,7 P. Kodyš,5 S. Korpar,38, 26
P. Križan,35, 26 P. Krokovny,4, 52 T. Kuhr,36 R. Kumar,57 T. Kumita,73 A. Kuzmin,4, 52 Y.-J. Kwon,79
I. S. Lee,13 C. H. Li,40 H. Li,20 L. Li,58 L. Li Gioi,39 J. Libby,19 D. Liventsev,76, 15 P. Lukin,4, 52 T. Luo,56
M. Masuda,70 D. Matvienko,4, 52 K. Miyabayashi,45 H. Miyata,50 R. Mizuk,25, 41 G. B. Mohanty,65
S. Mohanty,65, 75 A. Moll,39, 66 H. K. Moon,32 T. Mori,43 R. Mussa,24 E. Nakano,53 M. Nakao,15, 12 T. Nanut,26
Z. Natkaniec,49 M. Nayak,19 N. K. Nisar,65, 1 S. Nishida,15, 12 S. Ogawa,68 S. Okuno,27 P. Pakhlov,25, 41
G. Pakhlova,42, 25 C. W. Park,62 H. Park,33 S. Paul,67 T. K. Pedlar,37 L. Pesántez,3 R. Pestotnik,26
M. Petrič,26 L. E. Piilonen,76 C. Pulvermacher,28 J. Rauch,67 E. Ribežl,26 M. Ritter,36 A. Rostomyan,8
S. Ryu,59 H. Sahoo,14 Y. Sakai,15, 12 S. Sandilya,65 T. Sanuki,69 Y. Sato,43 V. Savinov,56 T. Schlüter,36
O. Schneider,34 G. Schnell,2, 16 C. Schwanda,22 Y. Seino,50 K. Senyo,78 O. Seon,43 I. S. Seong,14 V. Shebalin,4, 52
T.-A. Shibata,72 J.-G. Shiu,48 B. Shwartz,4, 52 F. Simon,39, 66 Y.-S. Sohn,79 A. Sokolov,23 E. Solovieva,25
S. Stanič,51 M. Starič,26 J. Stypula,49 M. Sumihama,11 T. Sumiyoshi,73 U. Tamponi,24, 74 Y. Teramoto,53
K. Trabelsi,15, 12 M. Uchida,72 S. Uehara,15, 12 T. Uglov,25, 42 S. Uno,15, 12 P. Urquijo,40 Y. Usov,4, 52
C. Van Hulse,2 P. Vanhoefer,39 G. Varner,14 A. Vinokurova,4, 52 A. Vossen,20 M. N. Wagner,10 C. H. Wang,47
M.-Z. Wang,48 X. L. Wang,76 M. Watanabe,50 Y. Watanabe,27 K. M. Williams,76 E. Won,32 J. Yamaoka,54
J. Yelton,9 C. Z. Yuan,21 Y. Yusa,50 Z. P. Zhang,58 V. Zhilich,4, 52 V. Zhulanov,4, 52 and A. Zupanc35, 26
(The Belle Collaboration)
1
Aligarh Muslim University, Aligarh 202002
2
University of the Basque Country UPV/EHU, 48080 Bilbao
3
University of Bonn, 53115 Bonn
4
Budker Institute of Nuclear Physics SB RAS, Novosibirsk 630090
5
Faculty of Mathematics and Physics, Charles University, 121 16 Prague
6
Chonnam National University, Kwangju 660-701
7
University of Cincinnati, Cincinnati, Ohio 45221
8
Deutsches Elektronen–Synchrotron, 22607 Hamburg
9
University of Florida, Gainesville, Florida 32611
10
Justus-Liebig-Universität Gießen, 35392 Gießen
11
Gifu University, Gifu 501-1193
12
SOKENDAI (The Graduate University for Advanced Studies), Hayama 240-0193
13
Hanyang University, Seoul 133-791
14
University of Hawaii, Honolulu, Hawaii 96822
15
High Energy Accelerator Research Organization (KEK), Tsukuba 305-0801
16
IKERBASQUE, Basque Foundation for Science, 48013 Bilbao
17
Indian Institute of Technology Bhubaneswar, Satya Nagar 751007
18
Indian Institute of Technology Guwahati, Assam 781039
19
Indian Institute of Technology Madras, Chennai 600036
20
Indiana University, Bloomington, Indiana 47408
21
Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049
22
Institute of High Energy Physics, Vienna 1050
23
Institute for High Energy Physics, Protvino 142281
24
INFN - Sezione di Torino, 10125 Torino
25
Institute for Theoretical and Experimental Physics, Moscow 117218
 2

26
J. Stefan Institute, 1000 Ljubljana
27
Kanagawa University, Yokohama 221-8686
28
Institut für Experimentelle Kernphysik, Karlsruher Institut für Technologie, 76131 Karlsruhe
29
Kennesaw State University, Kennesaw GA 30144
30
King Abdulaziz City for Science and Technology, Riyadh 11442
31
Korea Institute of Science and Technology Information, Daejeon 305-806
32
Korea University, Seoul 136-713
33
Kyungpook National University, Daegu 702-701
34
École Polytechnique Fédérale de Lausanne (EPFL), Lausanne 1015
35
Faculty of Mathematics and Physics, University of Ljubljana, 1000 Ljubljana
36
Ludwig Maximilians University, 80539 Munich
37
Luther College, Decorah, Iowa 52101
38
University of Maribor, 2000 Maribor
39
Max-Planck-Institut für Physik, 80805 München
40
School of Physics, University of Melbourne, Victoria 3010
41
Moscow Physical Engineering Institute, Moscow 115409
42
Moscow Institute of Physics and Technology, Moscow Region 141700
43
Graduate School of Science, Nagoya University, Nagoya 464-8602
44
Kobayashi-Maskawa Institute, Nagoya University, Nagoya 464-8602
45
Nara Women’s University, Nara 630-8506
46
National Central University, Chung-li 32054
47
National United University, Miao Li 36003
48
Department of Physics, National Taiwan University, Taipei 10617
49
H. Niewodniczanski Institute of Nuclear Physics, Krakow 31-342
50
Niigata University, Niigata 950-2181
51
University of Nova Gorica, 5000 Nova Gorica
52
Novosibirsk State University, Novosibirsk 630090
53
Osaka City University, Osaka 558-8585
54
Pacific Northwest National Laboratory, Richland, Washington 99352
55
Peking University, Beijing 100871
56
University of Pittsburgh, Pittsburgh, Pennsylvania 15260
57
Punjab Agricultural University, Ludhiana 141004
58
University of Science and Technology of China, Hefei 230026
59
Seoul National University, Seoul 151-742
60
Soongsil University, Seoul 156-743
61
University of South Carolina, Columbia, South Carolina 29208
62
Sungkyunkwan University, Suwon 440-746
63
School of Physics, University of Sydney, NSW 2006
64
Department of Physics, Faculty of Science, University of Tabuk, Tabuk 71451
65
Tata Institute of Fundamental Research, Mumbai 400005
66
Excellence Cluster Universe, Technische Universität München, 85748 Garching
67
Department of Physics, Technische Universität München, 85748 Garching
68
Toho University, Funabashi 274-8510
69
Department of Physics, Tohoku University, Sendai 980-8578
70
Earthquake Research Institute, University of Tokyo, Tokyo 113-0032
71
Department of Physics, University of Tokyo, Tokyo 113-0033
72
Tokyo Institute of Technology, Tokyo 152-8550
73
Tokyo Metropolitan University, Tokyo 192-0397
74
University of Torino, 10124 Torino
75
Utkal University, Bhubaneswar 751004
76
CNP, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061
77
Wayne State University, Detroit, Michigan 48202
78
Yamagata University, Yamagata 990-8560
79
Yonsei University, Seoul 120-749
We measure the decay Bs0 → K 0 K 0 using data collected at the Υ(5S) resonance with the Belle
detector at the KEKB e+ e− collider. The data sample used corresponds to an integrated luminosity
of 121.4 fb−1 . We measure a branching fraction B(Bs0 → K 0 K 0 ) = [19.6 +5.8
−5.1 (stat.) ± 1.0(sys.) ±
2.0(NB 0 B 0 )] × 10−6 with a significance of 5.1 standard deviations. This measurement constitutes
s s
the first observation of this decay.

PACS numbers: 13.25.Hw, 14.40.Nd

The two-body decays Bs0 → h+ h0− , where h(0) is either a pion or kaon, have now all been observed [1]. In con-
 3

trast, the neutral-daughter decays Bs0 → h0 h00 have yet tween the two track helices at their closest approach; the
to be observed. The decay Bs0 → K 0 K 0 [2] is of partic- flight length in the x-y plane; the angle between the KS0
ular interest because the branching fraction is predicted momentum and the vector joining the KS0 decay vertex
to be relatively large. In the standard model (SM), the to the IP; the angle between the pion momentum and
decay proceeds mainly via a b → s loop (or “penguin”) the laboratory-frame direction in the KS0 rest frame; the
transition as shown in Fig. 1, and the branching fraction distance-of-closest-approach in the x-y plane between the
is predicted to be in the range (16 − 27) × 10−6 [3]. The IP and the two pion helices; and the pion hit information
presence of non-SM particles or couplings could enhance in the SVD and CDC. The selection efficiency is 87% over
this value [4]. It has been pointed out that CP asymme- the momentum range of interest. We also require that the
tries in Bs0 → K 0 K 0 decays are promising observables in π + π − invariant mass be within 12 MeV/c2 (about 3.5σ
which to search for new physics [5]. in resolution) of the nominal KS0 mass [1].
To identify Bs0 → KS0 KS0 candidates, we define two
variables:
p the beam-energy-constrained mass Mbc =
2
Ebeam pB |2 c2 /c2 ; and the energy difference ∆E =
− |~
EB −Ebeam , where Ebeam is the beam energy and EB and
p~B are the energy and momentum, respectively, of the Bs0
candidate. These quantities are evaluated in the e+ e−
center-of-mass frame. We require that events satisfy
Mbc > 5.34 GeV/c2 and −0.20 GeV < ∆E < 0.10 GeV.
To suppress background arising from continuum
FIG. 1. Loop diagram for Bs0 → K 0 K 0 decays. e+ e− → qq (q = u, d, s, c) production, we use a sec-
ond NN [11] that distinguishes jetlike continuum events
(∗)0 (∗)0
The current upper limit on the branching fraction, from more spherical Bs B s events. This NN uses the
B(Bs0 → K 0 K 0 ) < 6.6 × 10−5 at 90% confidence level, following input variables, which characterize the event
was set by the Belle Collaboration using 23.6 fb−1 of topology: the cosine of the angle between the thrust
data recorded at the Υ(5S) resonance [6]. Here, we axis [12] of the Bs0 candidate and the thrust axis of the
update this result using the full data set of 121.4 fb−1 rest of the event; the cosine of the angle between the Bs0
recorded at the Υ(5S). The analysis presented here uses thrust axis and the +z axis; a set of 16 modified Fox-
improved tracking, K 0 reconstruction, and continuum Wolfram moments [13]; and the ratio of the second to
suppression algorithms. The data set corresponds to zeroth (unmodified) Fox-Wolfram moments. All quanti-
(6.53 ± 0.66) × 106 Bs0 B s0 pairs [7] produced in three ties are evaluated in the center-of-mass frame. The NN is
Υ(5S) decay channels: Bs0 B s0 , Bs∗0 B s0 or Bs0 B s∗0 , and trained using Monte Carlo (MC) simulated signal events
Bs∗0 B s∗0 . The latter two channels dominate, with pro- and qq background events. The MC samples are obtained
duction fractions of fB ∗0 B 0 = (7.3±1.4)% and fB ∗0 B ∗0 = using EvtGen [14] for event generation and Geant3 [15]
s s s s
(87.0 ± 1.7)% [8]. for modeling the detector response. The NN has a sin-
The Belle detector is a large-solid-angle magnetic spec- gle output variable (CNN ) that ranges from −1 for back-
trometer consisting of a silicon vertex detector (SVD), a groundlike events to +1 for signal-like events. We re-
50-layer central drift chamber (CDC), an array of aerogel quire CNN > −0.1, which rejects approximately 85% of
threshold Cherenkov counters, a barrel-like arrangement qq background while retaining 83% of signal decays. We
of time-of-flight scintillation counters, and an electromag- subsequently translate CNN to a new variable
netic calorimeter comprising CsI(Tl) crystals. These de- 
CNN − CNNmin

0
tector components are located inside a superconducting CNN = ln max − C , (1)
CNN NN
solenoid coil that provides a 1.5 T magnetic field. An
min max
iron flux-return located outside the coil is instrumented where CNN = −0.1 and CNN is the maximum value of
to detect KL0 mesons and to identify muons. The detec- CNN obtained from a large sample of signal MC decays.
0
tor is described in detail elsewhere [9, 10]. The origin of The distribution of CNN is well modeled by a Gaussian
the coordinate system is defined as the position of the function.
nominal interaction point (IP). The +z axis is aligned After applying all selection criteria, approximately
with the direction opposite the e+ beam and is parallel 1.0% of events have multiple Bs0 candidates. For these
to the direction of the magnetic field within the solenoid. events, we retain the candidate having the smallest value
The +x axis is horizontal and points towards the outside of χ2 obtained from the deviations of the reconstructed
of the storage ring; the +y axis points vertically upward. KS0 masses from their nominal values [1]. According to
Candidate K 0 mesons are reconstructed via the decay MC simulation, this criterion selects the correct Bs0 can-
KS → π + π − using a neural network (NN) technique [11].
0
didate > 99% of the time.
The NN uses the following information: the KS0 momen- We measure the signal yield by performing an un-
tum in the laboratory frame; the distance along z be- binned extended maximum likelihood fit to the variables
 4

0
Mbc , ∆E, and CNN . The likelihood function is defined estimated from a high statistics sample of D0 → KS0 π 0
as decays. The factor 0.50 accounts for the 50% probability
  for K 0 K 0 → KS0 KS0 (since K 0 K 0 is CP even). Inserting
N
P
− j Yj
Y X
i 0i  these values gives B(Bs0 → K 0 K 0 ) = (19.6 +5.8
−5.1 ) × 10
−6
,
L=e  Yj Pj (Mbc , ∆E i , CNN ) , (2)
where the error is statistical.
i j
The systematic uncertainty on B(Bs0 → K 0 K 0 ) arises
where Yj is the yield of component j; Pj (Mbc i 0i
, ∆E i , CNN ) from several sources, as listed in Table I. The uncertain-
is the probability density function (PDF) of component ties due to the fixed parameters in the PDF shape are
j for event i; j runs over the two event categories (signal estimated by varying the parameters individually accord-
and qq background); and i runs over all events in the sam- ing to their statistical uncertainties. For each variation
ple (N ). Backgrounds arising from other Bs0 and non-Bs0 the branching fraction is recalculated, and the difference
decays were studied using MC simulation and found to be with the nominal branching fraction is taken as the sys-
negligible. As correlations among the variables Mbc , ∆E, tematic uncertainty associated with that parameter. We
and CNN0
are found to be small, the three-dimensional add together all uncertainties in quadrature to obtain the
PDFs Pj (Mbc i 0i
, ∆E i , CNN ) are factorized into the prod- overall uncertainty due to fixed parameters. The uncer-
uct of separate one-dimensional PDFs. tainties due to errors in the calibration factors and the
The signal PDF is defined as fractions f (∗) (∗) are evaluated in a similar manner. To
Bs B s
test the stability of our fitting procedure, we generate
Psig = fB ∗0 B ∗0 PB ∗0 B ∗0 + fB ∗0 B 0 PB ∗0 B 0 (3) and fit a large ensemble of MC pseudoexperiments. By
s s s s s s s s

+(1 − fB ∗0 B ∗0 − fB ∗0 B 0 )PB 0 B 0 , comparing the mean of the fitted yields with the input
s s s s s s
value, a bias of −2.6% is found. We attribute this bias
where PB ∗0 B ∗0 , PB ∗0 B 0 , and PB 0 B 0 are the PDFs for sig- to our neglecting small correlations among the fitted ob-
s s s s s s
nal arising from Υ(5S) → Bs∗0 B s∗0 , (Bs∗0 B s0 + B s∗0 Bs0 ), servables. An 0.9% systematic uncertainty is assigned
0
and Bs0 B s0 decays. The Mbc and CNN PDFs are mod- due to the CNN selection; this is obtained by comparing
eled with Gaussian functions, and the ∆E PDFs are the selection efficiencies in MC simulationand data for
each modeled with a sum of two Gaussian functions the B 0 → D− (→ K + π − π − )π + control sample. We as-
having a common mean. All parameters of the signal sign a 2.0% systematic uncertainty for each reconstructed
PDF are fixed to the corresponding MC values. The KS0 → π + π − ; this is determined using a D0 → KS0 π 0
peak positions for Mbc and ∆E are adjusted according sample. The uncertainty on ε due to the MC sample size
to small data-MC differences observed in a control sam- is 0.2%. The total of the above systematic uncertainties
ple of Bs0 → Ds− π + decays [8]. As this control sample is calculated as their sum in quadrature. In addition,
has only modest statistics, the resolutions for Mbc , ∆E, there is a 10.1% uncertainty due to the number of Bs0 B s0
0 0
and CNN , and the peak position for CNN , are adjusted pairs. As this large uncertainty does not arise from our
for data-MC differences using a high statistics sample of analysis, we quote it separately.
B 0 → D− (→ K + π − π − )π + decays. For qq background,
0
the Mbc , ∆E, and CNN PDFs are modeled with an AR-
GUS function [16], a first-order Chebyshev polynomial, TABLE I. Systematic uncertainties on B(Bs0 → K 0 K 0 ).
and a Gaussian function, respectively. All parameters of Those listed in the upper section are associated with fitting
the qq background PDFs except for the end point of the for the signal yields and are included in the signal significance.
ARGUS function are floated in the fit. Source Uncertainty (%)
The results of the fit are 29.0 +8.5 −7.6 signal events and PDF parametrization 0.2
1095.0 +33.9
−33.4 continuum background events. Projections Calibration factor +0.9
−0.8
of the fit are shown in Fig. 2. The branching fraction is f (∗) (∗) +1.2
−1.1
Bs Bs
calculated via Fit bias +0.0
−2.6
Ys KS0 → π + π − reconstruction 4.0
B(Bs0 → K 0 K 0 ) = 2 ε, (4)
2NB 0 B 0 (0.50)BK 0 CNN selection 0.9
s s
MC sample size 0.2
where Ys is the fitted signal yield; NB 0 B 0 = (6.53 ± B(KS0 → π + π − ) 0.1
s s
0.66)×106 is the number of Bs0 B s0 events; BK 0 = (69.20± Total (without NB 0 B 0 ) +4.4
−5.1
s s
0.05)% is the branching fraction for KS0 → π + π − [1]; NB 0 B 0 10.1
and ε = (46.3 ± 0.1)% is the signal efficiency as deter- s s

mined from MC simulation. The efficiency is corrected

by a factor 1.01 ± 0.02 for each reconstructed KS0 , to pThe signal significance is calculated as
account for a small difference in KS0 reconstruction effi- −2 ln(L0 /Lmax ), where L0 is the likelihood value
ciency between data and simulation. This correction is when the signal yield is fixed to zero, and Lmax is
 5

16
Events / ( 0.01 GeV/c2 )

Events / ( 0.8 )
Events / ( 0.02 GeV )
(a) (b) 25
(c)
14
30
12 20
10
20 15
8

6 10
10 4
5
2

0 0 0
5.34 5.36 5.38 5.4 5.42 5.44 -0.2 -0.1 0 0.1 -10 -5 0 5 10
Mbc (GeV/c2) ΔE (GeV) /
CNN

0
FIG. 2. Projections of the 3D fit to the real data: (a) Mbc in −0.11 GeV < ∆E < 0.02 GeV and CNN > 0.5; (b) ∆E
0 0
2 2
in 5.405 GeV/c < Mbc < 5.427 GeV/c and CNN > 0.5; and (c) CNN in 5.405 GeV/c < Mbc < 5.427 GeV/c2 and
2

−0.11 GeV < ∆E < 0.02 GeV. The points with error bars are data, the (green) dashed curves show the signal, (magenta)
dotted curves show the continuum background, and (blue) solid curves show the total. The χ2 /(number of bins) values of these
fit projections are 0.30, 0.43, and 0.26, respectively, which indicate that the fit gives a good description of the data. The three
peaks in Mbc arise from Υ(5S) → Bs0 B s0 , Bs∗0 B s0 + Bs0 B s∗0 , and Bs∗0 B s∗0 decays.

the likelihood value of the nominal fit. We include KS0 π + π − contribution is observed. We check this quanti-
systematic uncertainties in the significance by convolv- tatively by performing our signal fit for events in the mass
ing the likelihood function with a Gaussian function sidebands of each KS0 [M (π + π − ) ∈ (0.460, 0.485) GeV/c2
whose width is equal to that part of the systematic and M (π + π − ) ∈ (0.510, 0.530) GeV/c2 ]. The extracted
uncertainty that affects the signal yield. We obtain a signal yields, −0.7 +2.9 +2.2
−2.1 and 1.6 −1.2 for the higher mo-
signal significance of 5.1 standard deviations; thus, our mentum KS and lower momentum KS0 , respectively, are
0

measurement constitutes the first observation of this consistent with zero. We calculate the expected num-
decay. ber of Bs0 → KS0 π + π − events in our signal sample us-
ing MC simulation and the measured branching fraction,
18 B(Bs0 → K 0 π + π − ) = 15.0×10−6 [18]; the result is 0.001.
Events / ( 0.003 GeV/c2 )

16
(a)
In summary, we report the first observation of the de-
14
cay Bs0 → K 0 K 0 . The branching fraction is measured to
12
10
be
8 −6
6
B(Bs0 → K 0 K 0 ) = (19.6 +5.8
−5.1 ± 1.0 ± 2.0) × 10 ,
4
where the first uncertainty is statistical, the second is
2
0
systematic, and the third reflects the uncertainty due to
the total number of Bs0 B s0 pairs. This value is in good
0.49 0.5 0.51
M1(π+π-) (GeV/c2)
agreement with the SM predictions [3], and it implies
that the Belle II experiment [19] will reconstruct over
Events / ( 0.003 GeV/c2 )

18 (b) 1000 of these decays. Such a sample would allow for

16 a much higher sensitivity search for new physics in this
14
b → s penguin-dominated decay.
12
10
ACKNOWLEDGMENTS
8
6
We thank the KEKB group for the excellent opera-
4
2
tion of the accelerator; the KEK cryogenics group for
0 the efficient operation of the solenoid; and the KEK
0.49 0.5 0.51 computer group, the National Institute of Informatics,
M2(π+π-) (GeV/c2) and the PNNL/EMSL computing group for valuable
computing and SINET4 network support. We acknowl-
FIG. 3. Background subtracted sPlot distributions of edge support from the Ministry of Education, Culture,
M (π + π − ) for the (a) higher momentum and (b) lower mo- Sports, Science, and Technology (MEXT) of Japan, the
mentum KS0 candidates. Japan Society for the Promotion of Science (JSPS),
and the Tau-Lepton Physics Research Center of Nagoya
Figure 3 shows the background-subtracted sPlot [17] University; the Australian Research Council; Austrian
distributions of M (π + π − ), where the KS0 selection is Science Fund under Grants No. P 22742-N16 and P
removed for the π + π − pair being plotted. No Bs0 → 26794-N20; the National Natural Science Foundation
 6

of China under Contracts No. 10575109, No. 10775142, bined fit on the annihilation corrections in Bu,d,s → P P
No. 10875115, No. 11175187, and No. 11475187; the decays within QCDF,” Phys. Lett. B 740, 56 (2015);
Chinese Academy of Science Center for Excellence in H. Y. Cheng, C. W. Chiang and A. L. Kuo, “Updating
Particle Physics; the Ministry of Education, Youth B → P P, V P decays in the framework of flavor symme-
and Sports of the Czech Republic under Contract try,” Phys. Rev. D 91, 014011 (2015).
[4] Q. Chang, X. Q. Li and Y. D. Yang, “A comprehensive
No. LG14034; the Carl Zeiss Foundation, the Deutsche analysis of hadronic b → s transitions in a family non-
Forschungsgemeinschaft and the VolkswagenStiftung; universal Z 0 model,” J. Phys. G 41, 105002 (2014).
the Department of Science and Technology of India; [5] S. Baek, D. London, J. Matias and J. Virto, “Bs0 →
the Istituto Nazionale di Fisica Nucleare of Italy; the K + K − and Bs0 → K 0 K̄ 0 decays within supersymme-
WCU program of the Ministry of Education, National try,” J. High Energy Phys. 12, (2006) 019; A. Hayakawa,
Research Foundation (NRF) of Korea Grants No. 2011- Y. Shimizu, M. Tanimoto and K. Yamamoto, “Search-
0029457, No. 2012-0008143, No. 2012R1A1A2008330, ing for the squark flavor mixing in CP violations of
Bs → K + K − and K 0 K̄ 0 decays,” Prog. Theor. Exp.
No. 2013R1A1A3007772, No. 2014R1A2A2A01005286, Phys. 2014, 023B04 (2014).
No. 2014R1A2A2A01002734, [6] C.-C. Peng et al. (Belle Collaboration), “Search for Bs0 →
No. 2015R1A2A2A01003280 , No. 2015H1A2A1033649; hh decays at the Υ(5S) resonance,” Phys. Rev. D 82,
the Basic Research Lab program under NRF Grant 072007 (2010).
No. KRF-2011-0020333, Center for Korean J-PARC [7] C. Oswald et al. (Belle Collaboration), “Semi-inclusive
Users, No. NRF-2013K1A3A7A06056592; the Brain studies of semileptonic Bs decays at Belle,” Phys. Rev.
Korea 21-Plus program and Radiation Science Research D 92, 072013 (2015)
[8] S. Esen et al. (Belle Collaboration), “Precise measure-
Institute; the Polish Ministry of Science and Higher Ed- (∗)+ (∗)−
ment of the branching fractions for Bs → Ds Ds
ucation and the National Science Center; the Ministry of ∗+ ∗−
and first measurement of the Ds Ds polarization using
Education and Science of the Russian Federation and the e+ e− collisions,” Phys. Rev. D 87, 031101(R) (2013).
Russian Foundation for Basic Research; the Slovenian [9] A. Abashian et al. (Belle Collaboration), “The Belle de-
Research Agency; the Basque Foundation for Science tector,” Nucl. Instrum. Methods Phys. Res., Sect. A 479,
(IKERBASQUE) and the Euskal Herriko Unibertsitatea 117 (2002); also see the detector section in J.Brodzicka
(UPV/EHU) under program UFI 11/55 (Spain); the et al., “Physics achievements from the Belle experiment,”
Swiss National Science Foundation; the National Science Prog. Theor. Exp. Phys. 2012, 04D001 (2012).
[10] Z. Natkaniec et al. (Belle SVD2 Group), “Status of the
Council and the Ministry of Education of Taiwan; and Belle silicon vertex detector,” Nucl. Instrum. Methods
the U.S. Department of Energy and the National Science Phys. Res., Sect. A 560, 1 (2006).
Foundation. This work is supported by a Grant-in-Aid [11] M. Feindt and U. Kerzel, “The NEUROBAYES neural
from MEXT for Science Research in a Priority Area network package,” Nucl. Instrum. Methods Phys. Res.,
(“New Development of Flavor Physics”) and from Sect. A 559, 190 (2006).
JSPS for Creative Scientific Research (“Evolution of [12] S. Brandt, C. Peyrou, R. Sosnowski and A. Wroblewski,
Tau-lepton Physics”). “The principal axis of jets. An attempt to analyze high-
energy collisions as two-body processes,” Phys. Lett. 12,
57 (1964).
[13] G. C. Fox and S. Wolfram, “Observables for the Analysis
of Event Shapes in e+ e− Annihilation and Other Pro-
cesses,” Phys. Rev. Lett. 41, 1581 (1978); The modified
[1] K. A. Olive et al. (Particle Data Group), “Review of moments used in this Letter are described in S. H. Lee
particle physics,” Chin. Phys. C 38, 090001 (2014). et al. (Belle Collaboration), “Evidence for B 0 → π 0 π 0 ,”
[2] Unless stated otherwise, charge-conjugate modes are im- Phys. Rev. Lett. 91, 261801 (2003).
plicitly included. [14] D. J. Lange, “The EvtGen particle decay simulation
[3] C. H. Chen, “Analysis of Bs → KK decays in the package,” Nucl. Instrum. Methods Phys. Res., Sect. A
PQCD,” Phys. Lett. B 520, 33 (2001); A. R. Williamson 462, 152 (2001).
and J. Zupan, “Two body B decays with isosinglet fi- [15] R. Brun et al., GEANT 3.21, CERN Report DD/EE/84-
nal states in SCET,” Phys. Rev. D 74, 014003 (2006); 1, 1984.
A. Ali, G. Kramer, Y. Li, C. D. Lu, Y. L. Shen, W. Wang [16] H. Albrecht et al. (ARGUS Collaboration), “Search for
and Y. M. Wang, “Charmless non-leptonic Bs decays to hadronic b → u decays,” Phys. Lett. B 241, 278 (1990).
P P , P V and V V final states in the pQCD approach,” [17] M. Pivk and F. R. Le Diberder, “sPlot: A statistical tool
Phys. Rev. D 76, 074018 (2007); C. K. Chua, “Rescat- to unfold data distributions”, Nucl. Instrum. Methods
tering effects in charmless B̄u,d,s → P P decays,” Phys. Phys. Res., Sect. A 555, 356 (2005).
Rev. D 78, 076002 (2008); K. Wang and G. Zhu, “Flavor [18] R. Aaij et al. (LHCb Collaboration), “Study of B(s) 0
→
dependence of annihilation parameters in QCD factor- 0 + 0− 0
KS h h decays with first observation of Bs →
ization,” Phys. Rev. D 88, 014043 (2013); J. J. Wang,
KS0 K ± π ∓ and Bs0 → KS0 π + π − ,” J. High Energy Phys.
D. T. Lin, W. Sun, Z. J. Ji, S. Cheng and Z. J. Xiao,
10, (2013) 143.
“B̄s0 → Kπ, KK decays and effects of the next-to-
[19] T. Abe et al. (Belle II Collaboration), “Belle II Technical
leading order contribution,” Phys. Rev. D 89, 074046
Design Report,” arXiv:1011.0352.
(2014); Q. Chang, J. Sun, Y. Yang and X. Li, “A com-