PHYSICAL REVIEW D 107, 016003 (2023)

Production rates of Ds+ Ds− and DD̄ molecules in B decays

Jia-Ming Xie,1 Ming-Zhu Liu ,2,1,* and Li-Sheng Geng3,1,4,5,†
1
School of Physics, Beihang University, Beijing 102206, China
2
School of Space and Environment, Beihang University, Beijing 102206, China
3
Peng Huanwu Collaborative Center for Research and Education, Beihang University,
Beijing 100191, China
4
Beijing Key Laboratory of Advanced Nuclear Materials and Physics, Beihang University,
Beijing 102206, China
5
School of Physics and Microelectronics, Zhengzhou University, Zhengzhou, Henan 450001, China

(Received 11 August 2022; accepted 11 December 2022; published 5 January 2023)

Motivated by the recent discovery of a charmonium Xð3960Þ in B decays by the LHCb Collaboration,
the likely existence of two bound/virtual states (denoted by X ss̄ and X qq̄ ) below the Dþ −
s Ds and D̄D mass
thresholds has been reexamined recently. In this work, we employ the effective Lagrangian approach
to calculate their production rates in B decays utilizing triangle diagrams. Our results show that the
production yields of Bþ → X ss̄ K þ and Bþ → X qq̄ K þ are of the order of 10−4 , in agreement with the
relevant experimental data, which indicates that, if the Dþ −
s Ds and D̄D bound states indeed exist, they
can be detected in B decays. Moreover, we calculate the production rate of Bþ → Xð3960ÞK þ assuming
that Xð3960Þ is a resonant state of Dþ −
s Ds and find that it is also of the order of 10
−4
but a bit smaller than
þ −
that as a Ds Ds bound state.

DOI: 10.1103/PhysRevD.107.016003

I. INTRODUCTION Given that the mass of Xð3960Þ is very different from

In 2004, the Belle Collaboration observed a state around
3940 MeV in the J=ψω invariant mass distribution of the
B → J=ψωK decay [1], which was later confirmed by the
BABAR Collaboration in the same process but the mass was
determined to be 3915 MeV [2]. In 2009, the Belle
Collaboration observed a state near 3915 MeV in the γγ →
J=ψω reaction [3]. Later the BABAR Collaboration deter-
mined the quantum number of this state to be J PC ¼ 0þþ
that of χ c0 ð3915Þ, they are not likely to be the same state. It
is also difficult to interpret Xð3960Þ as a conventional
charmonium because the mass of χ c0 ð3PÞ (in the quark
model) is around 4200 MeV [11,12]. In Ref. [13], M. Bayar
et al. argued that in the chiral unitary approach there exist
two states below the DD̄ and Dþ −
s Ds thresholds, respec-
tively, and the new Xð3960Þ can be understood as an
enhancement in the Dþ −
s Ds mass distribution. In Ref. [14],

[4]. In 2020, the LHCb Collaboration observed a similar Xin et al. interpreted Xð3960Þ as a JPC ¼ 0þþ Dþ s Ds
−

state χ c0 ð3930Þ in the Dþ D− mass distribution of the Bþ → molecule in the QCD sum rules approach. With a leading
Dþ D− K þ decay [5,6]. In the Review of Particle Physics order contact range effective field theory Ji et al. showed
(RPG) [7], all these states are referred to as χ c0 ð3915Þ and that either a bound state or a virtual state below the Dþ s Ds
−

mass threshold is needed to describe the Ds Ds mass þ −

viewed as a candidate for the χ c0 ð2PÞ charmonium [8,9].
Very recently, the LHCb Collaboration reported a charmo-
nium state named as Xð3960Þ with JPC ¼ 0þþ in the
Dþ − þ
s Ds mass distribution of the B → Ds Ds K
þ − þ decay.
Its mass and width are determined to be m ¼ 3955  6 
11 MeV and Γ ¼ 48  17  10 MeV [10].
*
zhengmz11@buaa.edu.cn
†
lisheng.geng@buaa.edu.cn
Published by the American Physical Society under the terms of
the Creative Commons Attribution 4.0 International license.
Further distribution of this work must maintain attribution to
the author(s) and the published article’s title, journal citation,
and DOI. Funded by SCOAP3.
distribution of the Bþ → Dþ − þ
s Ds K decay [15].
The likely existence of molecules near the DD̄
and Dþ −
s Ds mass thresholds has been studied in several
approaches before the experimental discoveries. In
Ref. [16], Prelovsek et al. performed lattice QCD simu-
lations of coupled-channel DD̄ and Dþ −
s Ds interactions and
found the existence of two bound states near the DD̄ and
Dþ −
s Ds mass thresholds. In Ref. [17], Gamermann et al.
obtained a IðJPC Þ ¼ 0ð0þþ Þ narrow resonance with a mass
of 3719 MeV, which mainly couples to the DD̄ and Ds D̄s
channels. In our previous works [18,19], we investigated
the DD̄ interaction in the one boson exchange (OBE)
model and found that a large cutoff of Λ ¼ 1.415 GeV is
needed to generate a DD̄ bound state. On the other hand, in

2470-0010=2023=107(1)=016003(7) 016003-1 Published by the American Physical Society

XIE, LIU, and GENG PHYS. REV. D 107, 016003 (2023)

order to reproduce Xð3872Þ as a D D̄ bound state in the

OBE model, one only needs a cutoff of Λ ¼ 1.01 GeV.
From this, one concludes that the DD̄ interaction is less
attractive than the D D̄ interaction, in agreement with
several recent studies [20–22].
In Refs. [23,24], the authors described the DD̄ mass
distribution of γγ → DD̄ and demonstrated the existence of
FIG. 2. Triangle diagrams accounting for the two B decays:
a bound state near the DD̄ mass threshold. In Ref. [25], (a) Bþ → Dþ 0 þ þ þ 0 þ
s D̄ → X ss̄ K and (b) B → Ds D̄ → X qq̄ K .
considering the state Xð3720Þ mainly coupled to DD̄, Dai
et al. predicted the DD̄ mass distribution of the B− → than those of the B meson decaying into a charmonium and
D0 D̄0 K − process. In Ref. [26], assuming χ c0 ð3915Þ as a a K þ . Then, taking into account the interaction vertices of
Dþ −
s Ds bound state, Li et al. estimated the branching ratio D̄0 → D−s K þ and Dþ 0 þ þ −
s → D K , the Ds Ds and D̄ D
0 0
of the Bþ → χ c0 ð3915ÞK þ decay to be 6 × 10−4 . In this molecules can be dynamically generated. We illustrate the
work, we assume that there exist two molecules near the decays of Bþ → Xss̄ K þ and Bþ → Xqq̄ K þ at the hadron
DD̄ and Dþ −
s Ds mass thresholds, denoted as X qq̄ and X ss̄ , level via the triangle diagrams shown in Fig. 2. In the
and employ the effective Lagrangian approach to inves- following, we introduce the effective Lagrangians relevant
tigate their production rates in B decays via the triangle to the computation of the Feynman diagrams shown
mechanism. Such an approach has been applied to study the in Fig. 2.
production of Pc and Pcs in the Λb → J=ψpK [27] and The effective Hamiltonian describing the weak decays of
Ξb → J=ψΛK decays [28]. The production rates of DD̄ and Bþ → Dþ s D̄
0
and Bþ → Dþ 0
s D̄ has the following form
Dþ −
s Ds molecules in B decays are helpful to probe the
nature of Xð3960Þ as well as to understand the DD̄ and G
Heff ¼ pFffiffiffi V cb V cs ½ceff
1 O1 þ c2 O2  þ H:c:;
eff ð1Þ
Dþ −
s Ds interactions. 2
This work is organized as follows. We introduce the
triangle mechanism for the decays of Bþ → Xss̄ K þ and where GF is the Fermi constant, V bc and V cs are the Cabibbo-
Bþ → Xqq̄ K þ and the effective Lagrangian approach in Kobayashi-Maskawa (CKM) matrix elements, ceff 1;2 are the
Sec. II. Results and discussions are given in Sec. III, effective Wilson coefficients, and O1 and O2 are the four-
followed by a short summary in the last section. fermion operators of ðsc̄ÞV−A ðcb̄ÞV−A and ðc̄cÞV−A ðsb̄ÞV−A
with ðq̄qÞV−A standing for q̄γ μ ð1 − γ 5 Þq [33–35].
II. THEORETICAL FORMALISM The amplitudes of Bþ → Dþ 0 þ
s D̄ and B → Ds D̄ can
þ 0

be written as the products of two hadronic matrix

The B decays are a good platform to study the weak elements [36,37]
interaction and exotic states [29–32]. In this work, we
employ the triangle diagram to investigate the weak decays GF
of Bþ → Xqq̄ K þ and Bþ → Xss̄ K þ . At the quark level, AðBþ → Dþ 0 þ
s D̄ Þ ¼ pffiffiffi V cb V cs a1 hDs jðsc̄Þj0i
2
the decays of Bþ → Dþ s D̄
0
and Bþ → Dþ 0
s D̄ can both
proceed through the external W-emission mechanism × hD̄0 jðcb̄ÞjBþ i ð2Þ
shown in Fig. 1. Referring to the Review of Particle GF
Physics [7], the absolute branching fractions of the decay AðBþ → Dþ 0 0 þ
s D̄ Þ ¼ pffiffiffi V cb V cs a1 hDs jðsc̄Þj0i
modes Bþ → Dþ 0
and Bþ → Dþ 0 2
s D̄ s D̄ are ð8.2  1.7Þ ×
−3 −3
10 and ð7.6  1.6Þ × 10 , respectively, which are larger × hD̄0 jðcb̄ÞjBþ i ð3Þ
ð0Þ
where a1 ¼ ceff 1 þ c2 =N c with N c ¼ 3 the number of
eff

colors. It should be noted that a1 and a01 can be obtained in

the factorization approach [38].
The current matrix elements between a pseudoscalar
meson or vector meson and the vacuum have the following
form:

hDþ
s jðsc̄Þj0i ¼ gμν f Dþ
s
pνDþ ;
s

hDþ
s jðsc̄Þj0i ¼ mDþ
s
f Dþ
s
ϵμ ; ð4Þ

where f Ds and f Dþ

s
are the decay constants for Ds and Dþ
s ,
External W-emission for Bþ → Dþ 0 þ þ 0  þ
FIG. 1. s D̄ and B → Ds D̄ . respectively, and ϵμ is the polarization vector of Ds .

016003-2
PRODUCTION RATES OF Dþ −
s Ds AND … PHYS. REV. D 107, 016003 (2023)

The hadronic matrix elements can be parametrized in terms of a few form factors [39]

0 þ A2 ðq2 Þ
hD̄ jðcb̄ÞjB i ¼ ϵα −gμα ðmD̄0 þ mBþ ÞA1 ðq2 Þ þ Pμ Pα
mD̄0 þ mBþ
 
μαβγ Vðq2 Þ μ α mD̄0 þ mB
þ
2 mBþ − mD̄0 2 2mD̄0 2
þ iε Pβ qγ þq P A1 ðq Þ − A2 ðq Þ − 2 A0 ðq Þ ; ð5Þ
mD̄0 þ mBþ q2 q2 q

hD̄0 jðcb̄ÞjBþ i ¼ ðpBþ þ pD̄0 Þμ F1D ðq02 Þ þ q0μ F2D ðq02 Þ; ð6Þ As Xss̄ and Xqq̄ are bound states of Dþ −
s Ds and D̄D, we
adopt the compositeness condition to estimate the cou-
where q and q0 represent the momentum of Dþ þ
s and Ds ,
plings of gXss̄ Dþs D−s and gXqq̄ D̄D . The correlation function
respectively, and P ¼ ðpBþ þ pD̄0 Þ. The form factors of Φðy2 Þ is introduced to reflect the distribution of the two
F1;2D ðtÞ, A0 ðtÞ, A1 ðtÞ, A2 ðtÞ, and VðtÞ with t ≡ qð0Þ2 are constituent hadrons in the molecule, which also renders the
parametrized as1 Feynman diagrams ultraviolet finite. Here we choose the
Fourier transformation of the correlation function in form
Xð0Þ of a Gaussian function
XðtÞ ¼ ; ð8Þ
1 − aðt=m2B Þ þ bðt2 =m4B Þ

which could well fit the transition form factors of B → D̄ðÞ Φðp2 Þ ≐ Expð−p2E =Λ2 Þ; ð12Þ
as shown in Ref. [39].
The Lagrangian describing the interaction between the
charmed mesons Ds , D and a kaon has the following form where Λ is a size parameter, which is expected to be around
1 GeV [44,45], and pE is the Euclidean momentum. The
LDs D K ¼ −igDs D K ðDs ∂μ KD†  μ †
μ − Dμ ∂ KDs Þ; ð9Þ
couplings, gXss̄ Dþs D−s and gXqq̄ D̄D , can be estimated by
where gDs D K is the coupling constant. reproducing the binding energies of the Xss̄ and Xqq̄ states
Assuming that Xss̄ and Xqq̄ are dynamically generated by via the compositeness condition [46–48]. The condition
the Dþ −
s Ds and D̄D interactions, respectively, the relevant dictates that the coupling constant can be determined from
Lagrangians can be written as the fact that the renormalization constant of the wave
Z function of a composite particle should be zero. The
compositeness condition can be estimated from the self
LXss̄ Dþs D−s ðxÞ ¼ gXss̄ Dþs D−s Xss̄ ðxÞ dyΦðy2 Þ
energy
   
þ 1 − 1
× Ds x þ y Ds x − y ; ð10Þ
2 2
dΣðk20 Þ
Z Z ¼1− ¼ 0: ð13Þ
dk20 k0 ¼mX
LXqq̄ D̄D ðxÞ ¼ gXqq̄ D̄D Xqq̄ ðxÞ dyΦðy2 Þ
   
1 1
× D̄ x þ y D x − y : ð11Þ With the above relevant Lagrangians, one can easily
2 2 compute the corresponding amplitudes of Fig. 2,

Z
d4 q3 iAðBþ → Dþ 0 0
s D̄ ÞAðD̄ → Ds K Þ
− þ
Aa ¼ g X ss̄ Dþ − ; ð14Þ
s Ds
ð2πÞ4 ðq21 − m2D̄ þ imD̄ ΓD̄ Þðq22 − m2Dþ Þðq23 − m2D−s Þ
s

1
The electric and magnetic distributions of hadrons in the low energy region, such as those of the nucleons, are often parametrized by
dipole form factors of the following form:

GE;M ð0Þ
GE;M ðq2 Þ ¼ ; ð7Þ
ð1 þ q2 =m2 Þ2
which, however, need to be revised in the high energy region [40,41]. We note that the dipole form factors have also been adopted to
describe the internal structure of baryons in lattice QCD simulations [42,43].

016003-3
XIE, LIU, and GENG PHYS. REV. D 107, 016003 (2023)

Z
d4 q3 iAðBþ → Dþ 0 þ
s D̄ ÞAðDs → D K Þ
0 þ
Ab ¼ gXqq̄ D̄0 D0 ; ð15Þ
ð2πÞ4 ðq21 − m2Dþ þ imDþ
s
ΓDþ
s
Þðq22 − m2D̄0 Þðq23 − m2D0 Þ
s

where q1 , q2 , and q3 denote the momenta of D̄0 , Dþ s , and respectively. The D̄0 → D−s K þ transition is expressed as
D−s for Fig. 2(a) and Dþ s , D̄ 0
, and D 0
for Fig. 2(b), AðD̄0 → D−s K þ Þ ¼ gD̄0 D−s Kþ p1 · εðq1 Þ, and the Dþs →
and p1 and p2 represent the momenta of K þ and D0 K þ transition is expressed as AðDþ s → D 0 þ
K Þ¼
Xss̄ ðXqq̄ Þ. The vertices of X ss̄ Dþ − 0 0
s Ds and X qq̄ D̄ D are gDþ 0 þ p1 · εðq1 Þ. The weak decay amplitudes of
s D K
parametrized as the couplings gXss̄ Dþs D−s and gXqq̄ D̄0 D0 , þ þ 0 þ þ 0
B → Ds D̄ and B → Ds D̄ are written as


G A2 ðq22 Þ
AðB → Ds D̄ Þ ¼ pFffiffiffi V cb V cs a1 f Ds −q2 · εðq1 ÞðmD̄0 þ mBþ ÞA1 ðq22 Þ þ ðk0 þ q1 Þ · εðq1 Þq2 · ðk0 þ q1 Þ
2 mD̄0 þ mBþ

þ ðk0 þ q1 Þ · εðq1 Þ½ðmD̄0 þ mBþ ÞA1 ðq22 Þ − ðmBþ − mD̄0 ÞA2 ðq22 Þ − 2mD̄0 A0 ðq22 Þ ;

G
AðB → Ds D̄Þ ¼ pFffiffiffi V cb V cs a01 f Ds mDs εðq1 Þ · ðk0 þ q2 ÞF1D ðq21 Þ; ð16Þ
2

With the Bþ → Xss̄ ðXqq̄ ÞK þ amplitudes determined
above, the corresponding partial decay widths can be
finally written as
1 1 j ⃗pj
Γ¼ jMj2 ;
2J þ 1 8π m2B
where J is the total angular momentum of the initial B
meson, the overline indicates the sum over the polarization
vectors of final states, and j ⃗pj is the momentum of either
final state in the rest frame of the B meson.
III. RESULTS AND DISCUSSIONS
The amplitudes of Bþ → Dþ s D̄
0 and Bþ → Dþ D̄0 are
s ðÞ
obtained by the naive factorization approach as shown in
Eq. (13). In this work, we take GF ¼ 1.166 × 10−5 GeV−2 ,
V cb ¼ 0.041, V cs ¼ 0.987, f Ds ¼ 250 MeV, and f Ds ¼
272 MeV as in Refs. [7,39,49–51]. For the form
and gXqq̄ D̄0 D0 depend on whether the Xss̄ and Xqq̄ states are
below or above the mass thresholds of Dþ −
s Ds and D̄D,
which will be specified below.
It is necessary to analyze the uncertainties of our results,
ð17Þ which mainly come from the coupling constants needed to
evaluate the triangle diagrams. In the weak interaction
vertex, the uncertainties for the parameters Xð0Þ, a and b of
the transition form factors estimated in Ref. [39] are quite
small and can be neglected. On the other hand, the
uncertainties in the experimental branching ratios of B →
D̄0 Dþ 0 þ 0
s =D̄ Ds can propagate to the parameters a1 ða1 Þ and
lead to about 10% uncertainties for them, i.e., a1 ¼
0.93þ0.09 0 þ0.08
−0.10 and a1 ¼ 0.81−0.09 . For the scattering vertices
D̄ðÞ Ds K, the couplings gD̄ðÞ DðÞ K are derived via SU(3)-
s
flavor symmetry. Since SU(3)-flavor symmetry is broken at
the level of 19% [52,53], we attribute an uncertainty of 19%
to the gD̄ðÞ DðÞ K couplings. For the couplings gXqq̄ DD̄ and
s

factors, we adopt those of the covariant light-front gXss̄ Dþs D−s , we take a 10% uncertainty.2 In terms of the
þ 0
quark model, e.g., ðF1D ð0Þ;a;bÞB →D̄ ¼ð0.67;1.22;0.36Þ, average of the three uncertainties mentioned above, we
0 0 arrive at an uncertainty of δ ¼ 13% for the couplings
ðA0 ð0Þ;a;bÞB→D̄ ¼ð0.68;1.21;0.36Þ, ðA1 ð0Þ; a; bÞB→D̄ ¼
0 characterizing the triangle diagrams. Following Ref. [56],
ð0.65; 0.60; 0.00Þ, and ðA2 ð0Þ; a; bÞB→D̄ ¼ ð0.61; 1.12; we estimate the uncertainty for the branching ratios in the
0.31Þ [39]. Note that the terms containing Vðq2 Þ and
following way Γ ¼ Γð1 þ δÞ2. As a result, the calculated
F2D ðq02 Þ do not contribute to the processes we study here. branching ratios have an uncertainty of 28%.
We tabulate the masses and quantum numbers of relevant
particles in Table I. In terms of the branching ratios of
Bþ → Dþ s D̄
0 and Bþ → Dþ D̄0 we determine a ¼ 0.93
s 1
2
Following Refs. [54,55], we vary the cutoff from 1 GeV to
0
and a1 ¼ 0.81, consistent with the estimates of Ref. [36]. 2 GeV, and the couplings gXqq̄ DD̄ and gXss̄ Dþs D−s decrease by
The couplings of gDþs D̄0 Kþ and gDþ 0 þ are determined by approximately 10%. Similarly, with the cutoff Λ ¼ 2 GeV and
s D̄ K
the SU(3)-flavor symmetry, e.g., gDþs D̄0 Kþ ¼ gDþ 0 þ ¼
the contact potentials Ca ¼ −5.25 GeV−2 and Cb ¼ −42.05 GeV−4 ,
pffiffiffi s D̄ K we obtain the coupling gXss̄ Dþs D−s ¼ 8.69 GeV. Compared with the
2gD̄0 D̄0 π0 , where gD̄0 D̄0 π0 ¼ 11.7 is obtained from the couplings obtained with the cutoff Λ ¼ 1 GeV (as shown later), we
decay width of D0 → D0 π 0 [7]. The couplings of gXss̄ Dþs D−s find that the coupling gXss̄ Dþs D−s increases by about 10%.

016003-4
PRODUCTION RATES OF Dþ −
s Ds AND … PHYS. REV. D 107, 016003 (2023)

TABLE I. Masses and quantum numbers of mesons relevant to the branching ratio of Bþ → Xss̄ K þ is of the order of 10−4
the present work [7]. if Xss̄ is a bound state of Dþ −
s Ds .
For the D̄D bound state, we determine the coupling
Meson IðJ P Þ M (MeV) Meson IðJ P Þ M (MeV)
of gXqq̄ D̄D as 9.62–19.38 GeV. The coupling of gXqq̄ D̄0 D0
þ 1 − þ 1 −
2 ð0 Þ 2 ð0 Þ
B 5279.34 K 493.677 is further determined by the isospin symmetry, i.e.,
D0 1 −
2 ð0 Þ
1864.84 Dþ 1 −
2 ð0 Þ
1869.66 gXqq̄ D̄0 D0 ¼ p1ffiffi2 gXqq̄ D̄D . With the couplings so obtained we
D0 1 −
2 ð1 Þ
2006.85 Dþ 1 −
2 ð1 Þ
2010.26
− − calculate the branching ratio of Bþ → Xqq̄ K þ , which turns
Dþs 0ð0 Þ 1968.34 Dþ
s 0ð1 Þ 2112.2
out to be in the range of ð1.3  0.4 ∼ 5.2  1.5Þ × 10−4 as
shown in Fig. 3. We note that the only allowed strong decay
If Xss̄ and Xqq̄ are bound states, the coupling of gXss̄ Dþs D−s mode of a D̄D isoscalar molecule is ηc η, which implies that
and gXqq̄ D̄0 D0 can be estimated by the compositeness the branching ratio of Bþ → ðXqq̄ → ηc ηÞK þ is 10−4 ,
condition [45]. According to Refs. [13,15,16,25,26], we which agrees well with the upper limit of the branching
assume that Xss̄ and Xqq̄ are located below the mass ratio of Bþ → ηc ηK þ , 2.2 × 10−4 [7]. Therefore, our result
thresholds of Dþ − indicates that the D̄D bound state can be detected in the ηc η
s Ds and D̄D respectively by 4 MeV to
30 MeV. For the Dþ − mass distribution of the Bþ → ηc ηK þ decay.
s Ds state, the coupling of gX ss̄ Dþ − is
s Ds
At last, we study the scenario where Xss̄ is a resonant
found to range from 9.41 GeV to 20.09 GeV, and the
state of Dþ −
s Ds , which can be identified as the Xð3960Þ state
corresponding branching ratios of Bþ → Xss̄ K þ varies
recently discovered by the LHCb Collaboration. Here we
from ð2.9  0.8Þ × 10−4 to ð13.3  3.7Þ × 10−4 as shown
assume that Xð3960Þ is dynamically generated by the
in Fig. 3. In Ref. [26], the authors assumed χ c0 ð3915Þ as a
Dþ −
s Ds interaction. With a contact potential of the form
Dþ −
s Ds bound state and estimated the branching ratio of
Ca þ Cb q2 , one can reproduce the mass and width of
B → χ c0 ð3915ÞK þ ¼ 6 × 10−4 , which agrees with our
þ
Xð3960Þ by solving the Lippmann-Schwinger equation
result [ð9.0  2.5Þ × 10−4 ] for a binding energy of B ¼ [57], and then obtain the Xð3960Þ coupling to Dþ −
s Ds from
20 MeV approximately. Referring to the Review of Particle the residues of the pole [58], where Ca and Cb are two low
Physics [7], the upper limit of the branching ratio of energy constants, and q is the momentum of D s in the
Bþ → χ c0 ð3915ÞK þ is 2.8 × 10−4 , which is smaller than center-of-mass system of the Dþ D−
pair. With the exper-
s s
(but consistent with) our result. One should note that a imental mass and width of Xð3960Þ as input, we obtain
shallow bound state below the Dþ −
s Ds mass threshold is Ca ¼ −5.15 GeV−2 and Cb ¼ −148.55 GeV−4 for a cutoff
predicted in two recent works with a binding energy of only of Λ ¼ 1 GeV, and then determine the coupling
several MeV [13,15]. Obviously, the production ratio of gXss̄ Dþs D−s ¼ 7.87 GeV. With the so-determined coupling
such a shallow state in the Bþ → Xss̄ K þ decay should be
we calculate the branching ratio of Bþ → Xss̄ K þ and
smaller than that of a deeply bound state, e.g., χ c0 ð3915Þ
obtain ð1.9  0.5Þ × 10−4 .
treated as a Dþ −
s Ds bound state, but should be of the same
order as shown in Fig. 3. Therefore, our results indicate that
IV. SUMMARY AND DISCUSSION
The recently discovered Xð3960Þ by the LHCb
Collaboration motivated us to study the Dþ −
s Ds and D̄D
molecules predicted in a number of recent works. In this
work, we assumed that there exist two bound states below
the mass thresholds of Dþ −
s Ds and D̄D, respectively, and
studied their production in the B decays via the triangle
mechanism. In such a mechanism, Bþ first weakly decays
into Dþ 0 þ 0 0 þ
s D̄ and Ds D̄ , the D̄ =Ds decays into Ds =D
− 0
þ − 0 0
plus a kaon, and then the final state Ds Ds and D̄ D
interactions dynamically generate the Xss̄ and Xqq̄ mole-
cules. As for the bound states, we employed the compos-
iteness condition approach to determine their couplings to
their constituents. The resonant state of Dþ −
s Ds is dynami-
cally generated in the single-channel approximation, and
the corresponding coupling is determined from the residues
of the pole.
We employed the effective Lagrangian approach to
FIG. 3. Branching ratios of Bþ → X ss̄ K þ and Bþ → X qq̄ K þ as calculate the branching ratios of Bþ → Xss̄ K þ and Bþ →
functions of the binding energies of Dþ −
s Ds and D̄D bound states. Xqq̄ K þ assuming that Xss̄ and Xqq̄ are bound states and

016003-5
XIE, LIU, and GENG
TABLE II. Branching ratios of B → X ss̄=qq̄ =Xð3960ÞK þ where X ss̄=qq̄ is a bound state of Dþ −
Decay modes Our results
þ þ
B → X ss̄ K ð2.1 ∼ 17.0Þ × 10
Bþ → ðX qq̄ → ηc ηÞK þ ð0.9 ∼ 6.7Þ × 10−4
Bþ → Xð3960ÞK þ <2.4 × 10−4
found that both of them are of the order of 10−4 . Our results
indicate that such bound states of D̄D and Dþ −
s Ds (if exist)
have large production rates in the B decays since they
account for a a large portion of the relevant experimental
data as shown in Table II. We note that the D̄D bound
state is likely to be detected in the ηc η mass distribution
of the Bþ → ηc ηK þ decay since the D̄D bound state
only decays into ηc η, while the Dþ −
s Ds bound state has
more decay modes, e.g., B → D̄DK þ , Bþ → ηc ηK þ ,
þ edges support from the National Natural Science
and Bþ → J=ψωK þ . At last, assuming that the Xð3960Þ
state is dynamically generated by the Dþ −
s Ds single-
channel interaction, we obtained the branching ratio of
[1] K. Abe et al. (Belle Collaboration), Phys. Rev. Lett. 94,
182002 (2005).
[2] B. Aubert et al. (BABAR Collaboration), Phys. Rev. Lett.
101, 082001 (2008).
[3] S. Uehara et al. (Belle Collaboration), Phys. Rev. Lett. 104,
092001 (2010).
[4] J. P. Lees et al. (BABAR Collaboration), Phys. Rev. D 86,
072002 (2012).
[5] R. Aaij et al. (LHCb Collaboration), Phys. Rev. D 102,
112003 (2020).
[6] R. Aaij et al. (LHCb Collaboration), Phys. Rev. Lett. 125,
242001 (2020).
[7] P. A. Zyla et al. (Particle Data Group), Prog. Theor. Exp.
Phys. 2020, 083C01 (2020).
[8] M.-X. Duan, S.-Q. Luo, X. Liu, and T. Matsuki, Phys. Rev.
D 101, 054029 (2020).
[9] M.-X. Duan, J.-Z. Wang, Y.-S. Li, and X. Liu, Phys. Rev. D
104, 034035 (2021).
[10] LHCb Collaboration, arXiv:2211.05034; arXiv:2210.15153.
[11] T. Barnes, S. Godfrey, and E. S. Swanson, Phys. Rev. D 72,
054026 (2005).
[12] B.-Q. Li and K.-T. Chao, Phys. Rev. D 79, 094004 (2009).
[13] M. Bayar, A. Feijoo, and E. Oset, arXiv:2207.08490.
[14] Q. Xin, Z.-G. Wang, and X.-S. Yang, AAPPS Bull. 32, 37
(2022).
[15] T. Ji, X.-K. Dong, M. Albaladejo, M.-L. Du, F.-K. Guo, and
J. Nieves, Phys. Rev. D 106, 094002 (2022).
[16] S. Prelovsek, S. Collins, D. Mohler, M. Padmanath, and
S. Piemonte, J. High Energy Phys. 06 (2021) 035.
[17] D. Gamermann, E. Oset, D. Strottman, and M. J. Vicente
Vacas, Phys. Rev. D 76, 074016 (2007).
016003-6
PHYS. REV. D 107, 016003 (2023)
s Ds or D̄D.
Exp [7]
−4
BrðB → χ c0 ð3915ÞK þ Þ < 2.8 × 10−4
þ
BrðBþ → ηc ηK þ Þ < 2.2 × 10−4

Bþ → Xð3960ÞK þ ¼ ð1.9  0.5Þ × 10−4 , which can help
elucidate the nature of Xð3960Þ.
ACKNOWLEDGMENTS
This work is supported in part by the National Natural
Science Foundation of China under Grants No. 11975041,
No. 11735003, and No. 11961141004. M.-Z. L. acknowl-
Foundation of China under Grant No. 12105007 and
China Postdoctoral Science Foundation under Grants
No. 2022M710317, and No. 2022T150036.
[18] M.-Z. Liu, T.-W. Wu, M. Pavon Valderrama, J.-J. Xie, and
L.-S. Geng, Phys. Rev. D 99, 094018 (2019).
[19] T.-W. Wu, M.-Z. Liu, and L.-S. Geng, Phys. Rev. D 103,
L031501 (2021).
[20] M.-Z. Liu and L.-S. Geng, Eur. Phys. J. C 81, 179 (2021).
[21] X.-K. Dong, F.-K. Guo, and B.-S. Zou, Prog. Phys. 41, 65
(2021).
[22] F.-Z. Peng, M. Sánchez Sánchez, M.-J. Yan, and M. Pavon
Valderrama, Phys. Rev. D 105, 034028 (2022).
[23] E. Wang, H.-S. Li, W.-H. Liang, and E. Oset, Phys. Rev. D
103, 054008 (2021).
[24] O. Deineka, I. Danilkin, and M. Vanderhaeghen, Phys. Lett.
B 827, 136982 (2022).
[25] L. R. Dai, J.-J. Xie, and E. Oset, Eur. Phys. J. C 76, 121
(2016).
[26] X. Li and M. B. Voloshin, Phys. Rev. D 91, 114014
(2015).
[27] Q. Wu and D.-Y. Chen, Phys. Rev. D 100, 114002 (2019).
[28] J.-X. Lu, M.-Z. Liu, R.-X. Shi, and L.-S. Geng, Phys. Rev.
D 104, 034022 (2021).
[29] H.-Y. Cheng, C.-K. Chua, and A. Soni, Phys. Rev. D 71,
014030 (2005).
[30] H.-Y. Cheng and K.-C. Yang, Phys. Rev. D 78, 094001
(2008); 79, 039903(E) (2009).
[31] H.-X. Chen, Phys. Rev. D 105, 094003 (2022).
[32] F.-L. Wang, X.-D. Yang, R. Chen, and X. Liu, Phys. Rev. D
104, 094010 (2021).
[33] A. J. Buras, Weak Hamiltonian, CP violation and rare
decays, arXiv:hep-ph/9806471.
[34] C. Q. Geng, Y. K. Hsiao, Y.-H. Lin, and L.-L. Liu,
Phys. Lett. B 776, 265 (2018).
PRODUCTION RATES OF Dþ −
s Ds AND …
[35] J.-J. Han, H.-Y. Jiang, W. Liu, Z.-J. Xiao, and F.-S. Yu,
Chin. Phys. C 45, 053105 (2021).
[36] A. Ali, G. Kramer, and C.-D. Lu, Phys. Rev. D 58, 094009
(1998).
[37] Q. Qin, H.-n. Li, C.-D. Lü, and F.-S. Yu, Phys. Rev. D 89,
054006 (2014).
[38] M. Bauer, B. Stech, and M. Wirbel, Z. Phys. C 34, 103 (1987).
[39] R. C. Verma, J. Phys. G 39, 025005 (2012).
[40] J. Arrington, C. D. Roberts, and J. M. Zanotti, J. Phys. G 34,
S23 (2007).
[41] V. Punjabi, C. F. Perdrisat, M. K. Jones, E. J. Brash, and
C. E. Carlson, Eur. Phys. J. A 51, 79 (2015).
[42] S. Collins et al., Phys. Rev. D 84, 074507 (2011).
[43] K. U. Can, G. Erkol, B. Isildak, M. Oka, and T. T.
Takahashi, J. High Energy Phys. 05 (2014) 125.
[44] X.-Z. Ling, J.-X. Lu, M.-Z. Liu, and L.-S. Geng, Phys. Rev.
D 104, 074022 (2021).
[45] X.-Z. Ling, M.-Z. Liu, L.-S. Geng, E. Wang, and J.-J. Xie,
Phys. Lett. B 826, 136897 (2022).
[46] S. Weinberg, Phys. Rev. 130, 776 (1963).
[47] A. Salam, Nuovo Cimento 25, 224 (1962).
016003-7
PHYS. REV. D 107, 016003 (2023)
[48] K. Hayashi, M. Hirayama, T. Muta, N. Seto, and T.
Shirafuji, Fortschr. Phys. 15, 625 (1967).
[49] S. Aoki et al. (Flavour Lattice Averaging Group), Eur. Phys.
J. C 80, 113 (2020).
[50] G. C. Donald, C. T. H. Davies, R. J. Dowdall, E. Follana, K.
Hornbostel, J. Koponen, G. P. Lepage, and C. McNeile,
Phys. Rev. D 86, 094501 (2012).
[51] Y. Li, P. Maris, and J. P. Vary, Phys. Rev. D 96, 016022
(2017).
[52] S. Dürr et al., Phys. Rev. D 95, 054513 (2017).
[53] N. Miller et al., Phys. Rev. D 102, 034507 (2020).
[54] A. Faessler, T. Gutsche, V. E. Lyubovitskij, and Y.-L. Ma,
Phys. Rev. D 76, 014005 (2007).
[55] A. Faessler, T. Gutsche, S. Kovalenko, and V. E.
Lyubovitskij, Phys. Rev. D 76, 014003 (2007).
[56] X.-Z. Ling, M.-Z. Liu, and L.-S. Geng, Eur. Phys. J. C 81,
1090 (2021).
[57] Q.-Y. Zhai, M.-Z. Liu, J.-X. Lu, and L.-S. Geng, Phys. Rev.
D 106, 034026 (2022).
[58] J.-M. Xie, X.-Z. Ling, M.-Z. Liu, and L.-S. Geng, Eur. Phys.
J. C 82, 1061 (2022).