Arxiv:1706.02588V2 [Hep-Ph] 29 Apr 2019 D D O Oeua Tts Ntehde Hr Etrw Have We Sector Candidates Charm Good Hidden the Particularly the in As Are States

Home , C parity, Isospin multiplet

arXiv:1706.02588v2 [hep-ph] 29 Apr 2019 D D o oeua tts ntehde hr etrw have we sector candidates charm good hidden the particularly the In as are states. to few molecular referred a for usually and different charmonium are states in They the XYZ in found then. fit since been experiments not have (hid- do charm spectrum that hidden (bottomonium) puzzling states similarly bottom) of den series a alone: P oeua tt.The for state. candidate molecular paradigmatic most a the provides probably oet 1–8.I h idnbto etrw aethe have we sector bottom Z hidden the In [12–18]. ponents ocsta ih esrn nuht idte [1–5]. them bind to enough the of heavy strong discovery be The nucleons, might analogy the exchange that generating in bind mesons, forces light that the- robust: exchange The forces can is hadrons nuclear hadron. existence the exotic their of with for type basis a are oretical – hadrons heavy etr the sector, B icvrdbfr the before discovered ohv large a have to ∗ † ¯ c b ∗ [email protected] [email protected] D Σ ev arnmlcls–budsae opsdof composed states bound – molecules hadron Heavy (4450) ¯ 160 and (10610) B ¯ c ∗ ∗ Z , ∗ , c D (3900), oeue 2,2] fw osdrteoe charm open the consider we If 22]. [21, molecules ¯ D ∗ ∗ + Σ D ¯ c D etqaksae 1] hc ih contain might which [11], states pentaquark ∗ , n niayn hsfaeoki aito fteoe alr ones the of the variation describe a is to framework states This antimeson antibaryon. and xhne,i atclrtetno oc.W n hti gen in determin that We find an We Σ the system. S-wave force. for this tensor in perturbative the of is particular dynamics in pair energy exchanges, quark low pote light baryon-antibaryon the heavy the mediates the which constrains symmetry for spin baryons heavy of ntehde-hr n idnbto etr.Terdoubl Their sectors. hidden-bottom (Λ and hidden-charm the in rbbet n ev aynu tt.Tems rbbeca probable most The state. Λ f baryonium isoscalar estimate heavy can the a we find mesons, c to vector contact-range probable of the exchange that the assume by we saturated If perturbative. are they ASnmes 03.65.Ge,13.75.Lb,14.40.Lb,14.40.Nd,14. numbers: PACS s D 0 ¯ oeue 9 0,adthe and 10], [9, molecules Z 21)and (2317) Q ∗ Z DK edsusteeetv edter ecito fbudsta bound of description theory field effective the discuss We ev aynAtbro oeue nEetv il Theor Field Effective in Molecules Baryon-Antibaryon Heavy c Σ .INTRODUCTION I. Λ b 42)[,8 hc r upce obe to suspected are which 8] [7, (4020) 160 1,2] hc ih be might which 20], [19, (10650) Q c ∗ Λ , / n even and X D Q X ∗ nentoa eerhCne o uliadPrilsi t in Particles and Nuclei for Center Research International 37)mr hnadcd g [6] ago decade a than more (3872) u-uLu, Jun-Xu Σ K 37)adhv entheorized been have and (3872) Q ( ∗ D ejn e aoaoyo dacdNcerMtrasadPh and Materials Nuclear Advanced of Laboratory Key Beijing X Q ) oeua opnn [25–28]. component molecular Σ , 2 s Λ 37)trsotntt be to not out turns (3872) ¯ ntttd hsqeNcár,CR-NP,Ui.Paris- Univ. Nucléaire, CNRS-IN2P3, Physique de Institut 1 Q D 26)msn 2,2]were 24] [23, mesons (2460) ¯ Q ( Σ , ∗ Λ ) nvri´ ai-aly -10 ra ee,France Cedex, Orsay F-91406 Paris-Saclay, Université Σ Q c 1 Q 29)mlclrcom- molecular (2590) Q ( colo hsc n ula nryEngineering, Energy Nuclear and Physics of School Σ ¯ ∗ Σ ,2 1, ¯ ) Q r logo addtsfrbinding. for candidates good also are ) Q Σ , ehn nvriy ejn 011 China 100191, Beijing University, Beihang Σ , iSegGeng, Li-Sheng Q ∗ Q ∗ P Σ ¯ Σ ¯ c Q (4380) Q X n Σ and n Σ and 37)o the or (3872) Dtd pi 0 2019) 30, April (Dated: + Q ∗ B Q ∗ Σ ¯ and B Σ ¯ Q ∗ ¯ 1, ∗ Q ∗ ytm,wiefrteΞ the for while systems, , ∗ n h svco Λ isovector the and n aulPvnValderrama Pavon Manuel and ttshv ece h ev eo-eo n heavy and (3 meson-meson discovered threshold heavy experimentally the meson-baryon the For reached of have threshold. mass states charm the open moment the the above happening states ih epsil opeittelcto fhaybary- heavy Λ of the location with the states, it predict instances onium to few possible a In be heavy might states. rel- of these the behind explore descriptions dynamics and evant available them straight- to the fairly molecules is meson-antimeson extend it yet, to found been forward not candi- Though have states state. excellent these bound baryon-antibaryon an heavy a be for date would threshold baryon-baryon Λ aynbro hehl (4 threshold baryon-baryon fteXZsae.Tems la xml rbbyis probably example clear few most a The of the states. nature XYZ the the uncover of to exploited can successfully spectrum of molecular be the mixture in Yet a regularities nature: potential be components. in these exotic will realized other they and fully charmonium shallow molecule, be 30– are 21, not states these [10, will unless spectrum spectrum regular This that fairly symmetry 35]. a of into degree translate high can a and symmetry, chiral quark SU(3)-flavour, heavy isospin, simul- to are subjected they content, taneously physics. quark heavy-light hadronic their of to Owing objects theoretical interesting most [29]. example lustrative ydffrn elztoso ev ur ymty[10]. symmetry quark heavy of realizations different by D Z 40Pq,14.40Rt c c D eepc oeua ttst erltvl arwfor narrow relatively be to states molecular expect We ev arnathdo oeue r mn the among are molecules hadron-antihadron Heavy ¯ Σ ¯ and Z c n Λ and c n Σ and sand ’s ulnso h ffciefil hoyare theory field effective the of ouplings Z ta.Teoepo xhnepotential exchange pion one The ntial. rlpo xhne r rbbynon- probably are exchanges pion eral rwihqatmnmesi smore is it numbers quantum which or b -hre n bto counterparts -bottom and y-charmed c eoacs ecnie h case the consider We resonances. D aydvlpdfrhaymeson- heavy for developed eady ddtst ombudsae are states bound form to ndidates c ¯ h eaieiprac fpion of importance relative the e e opsdo ev baryon heavy a of composed tes Σ Z ¯ /Σ eepoehwhayquark heavy how explore we d Q c b .Anro eoac erteheavy the near resonance narrow A ). srsnne,wihse ob related be to seem which resonances, ’s Σ ¯ c Q ′ D Q ¯ Ξ ¯ eCso and Cosmos he Q ′ n Λ and epciey,btbrl h heavy the barely but respectively), Ξ , Q ∗ Sud, Ξ c ysics, ¯ Q (2590) Q ′ Σ ¯ . n 4 and 7 . Q ∗ n Ξ and ,4 5, 1, ytm,both systems, † Σ ¯ . ,4 7, Q ∗ c Ξ ¯ ytm en nil- an being systems . Q ∗ GeV/4 1 . y cases e o Λ for GeV 9 . e for GeV 3 c Λ ¯ c , 2

Heavy hadron molecules possess another interesting case their interactions can be described in terms of well- quality: they show a separation of scales. On the one known physics such as pion exchanges. In the following hand we have the size of the hadrons, which is of the order lines we will present a brief introduction to the applica- of 0.5 fm, while on the other we have the size of the bound tion of the EFT framework to heavy baryon-antibaryon state, which should be bigger than the individual hadrons systems. within it. As a consequence heavy hadron molecules are amenable to an effective field theory (EFT) treatment, where all quantities can be expressed as an expansion of A. The Effective Field Theory Expansion a light over a heavy energy scale. EFT descriptions of heavy hadron molecules have been exploited successfully EFTs rely on the existence of a separation of scales, in the past specifically to systems composed of heavy where a distincion is made between low and high en- mesons and antimesons [31, 36–38]. In this manuscript ergy physics and their respetive characteristic momentum we extend the heavy hadron EFT formulated in Ref. [32] scales M and M , which are sometimes called the and put in use in Refs. [10, 33, 34] to the case of the heavy soft hard soft (or light) and hard (or heavy) scales. The separation baryon-antibaryon molecules. As commented, these type of scales can be used to express physical quantities at low of molecules might very well be discovered in the next energies as expansions in terms of the small parameter few years. The purpose of this work is to explore the M /M . If we consider a system of heavy baryons symmetry constrains and the kind of EFT that is to be soft hard for concreteness, there are two possible EFT expansions expected in these systems, rather than to make concrete depending on which type of low energy symmetry we are predictions of the possible location of these states. Yet considering: we will speculate a bit about this later issue on the basis of the relative strength of the long range pion exchange (i) heavy quark spin symmetry (HQSS) and and the saturation of the EFT low energy constants by σ, ρ, ω and φ meson exchange. (ii) chiral symmetry. The manuscript is structured as follows: in Section II we make a brief introduction to the EFT formalism. In For HQSS the soft and hard scales are ΛQCD 200 MeV ∼ Section III we present the leading order EFT potential for and the mass of the heavy quark mQ, which is either heavy baryon-antibaryon states, which consists of a series mc 1.5GeV or mb 4.5 GeV. For chiral symmetry ∼ ∼ of contact four-baryon vertices plus the time-honoured we call the soft and hard scales Q and M, where if there one pion exchange potential. In Section IV we explore are no baryons we have Q mπ q 100 200 MeV ∼ ∼ ∼ − the question of whether pions are perturbative or not (with mπ the pion mass and q the momenta of the pions) for this type of hadron molecule. In Section V we discuss and M 2πfπ 1GeV. If there are baryons Q in- ∼ ∼ the possible power countings to describe molecular states. cludes the soft momenta of the baryons, while practical In Section VI we speculate about which heavy baryon- calculations in the two-baryon sector suggest a more con- antibaryon molecules might be more probable. Finally servative value of M 0.5 1.0 GeV for the hard scale. ∼ − in Section VII we present our conclusions. In Appendix We advance that the scale Q can contain more than the A we present the complete derivation of the one pion pion mass and the momenta of the pions and baryons, as exchange potential, in Appendix B we briefly explain the we will discuss in Sect. II B. one eta and one kaon exchange potential and in Appendix Heavy baryons are non-relativistic at the soft scales of C we derive the heavy quark symmetry constrains for the either of the two previous symmetries. This implies that four-baryon contact vertices. their interactions can be described in terms of an effective potential VEFT, which admits the double expansion

Λ µ Q ν II. EFFECTIVE FIELD THEORY FOR HEAVY V = Vˆ (µ,ν) QCD , (1) EFT m M BARYON MOLECULES µ,ν Q X Effective field theories (EFTs) are generic and system- where the indexes µ and ν indicate the order in the heavy atic descriptions of low energy processes. They can be quark and chiral expansion, respectively, with µ 0 and ν 1 (this second point will be explained in Sect.≥ II B). applied to physical systems in which there is a distinct ≥− separation of scales, but where the underlying high en- The HQSS expansion converges remarkably faster than ergy theory for that system is unknown or unsolvable. the chiral expansion, owing to the sizes of the soft and Hadronic molecules are a good candidate for the EFT hard scales involved in each of these expansions. For this reason from now on we will work in the m limit treatment: the separation among the hadrons forming Q → ∞ a hadronic molecule is expected to be larger than the and ignore any HQSS breaking effect. With this in mind, size of the hadrons. When the hadrons are close to each the expansion of the EFT potential simplifies to other they overlap and the ensuing description in terms of ν max Q νmax+1 quantum chromodynamics (QCD) is unsolvable. But this V = V (ν) + , (2) EFT O M is not the case when the hadrons are far away, in which ν=ν X0 3

(0) 2 which converges for Q

1 from which V Q− is required. To explain why the po- D. Kaon/Eta Exchanges and SU(3) Symmetry tential is counted∼ this way we have to revisit the estima- (0) (0) tions of the size of the couplings C and F contained If we want to preserve SU(3) flavour symmetry, the in Eq. (8). If any of these two couplings contains a light exchange of kaons and eta mesons should be treated on scale equal footing as the exchange of pions, at least in prin- ( 1) 1 ( 1) 1 ciple. But the mass of the kaon and the eta meson is of C − and/or F − , (12) ∼ MQ ∼ MQ the order of 0.5 GeV, which is comparable to the hard scale M. We have two choices: (i) ignore kaon and eta 0 1 the LO potential will be promoted from Q to Q− , al- exchanges or, (ii) include them explicitly. lowing for the existence of bound states 2. The light ( 1) In the first option, the contribution from kaon and eta momentum scale that appears in C − can be identi- exchange are implicitly included in the contact-range po- fied with the inverse scattering length of the two-body ( 1) tential. There is a disadvantage, though: the contact- system [39–42], while the light scale in F − is related range potential breaks SU(3)-flavour symmetry in this with the strength of the OPE potential [45]. We stress case. We expect the size of this breaking to be paramet- that it is enough to promote one of the two couplings (0) (0) 0 1 rically small for heavy baryon-baryon and heavy baryon- C and F from Q to Q− , where for a more detailed ¯ antibaryon systems of the type ΣQΣQ, ΣQΣQ, ΞQ′ ΞQ′ , discussion we refer to Sect. V. ¯ ΞQ′ ΞQ′ , etc., i.e. systems containing only one species of baryon. Besides the pion, this type of system only ex- C. Coupled Channels change eta mesons, where their coupling to the heavy baryons is considerably weaker than that of the pions. Regarding the kaons, they are relevant for heavy baryon- Now we consider the power counting of coupled channel baryon systems that involve different species: Σ Ξ , effects. Heavy baryons can come in HQSS multiplets Q Q′ Ξ Ω , etc. But if we consider instead heavy baryon- which are degenerate in the heavy quark limit, e.g. the Q′ Q antibaryon system the exchange of a single kaon im- Σ and the Σ in the charm sector or the Σ and the Σ in c c∗ b b∗ plies a transition between two-baryon states with differ- the bottom sector. If we take the Σ and the Σ heavy c c∗ ent thresholds, which involves coupled channel effects. baryons as an example, the two heavy baryon system ¯ ¯ For example in the Ξc′ Σc ΩcΞc′ transition mediated can have transitions of the type Σ Σ Σ Σ∗, Σ Σ c c → c c c c → by the exchange of a kaon/antikaon,→ the energy gap Σc∗Σc∗, etc. In EFT these transitions have a characteristic momentum scale is ∆CC = 240MeV and the coupled channel scale is ΛCC = 790 MeV, which is certainly hard. ΛCC = 2µ∆CC , (13) From these reasons we expect the exchange of kaon and eta mesons to have a small impact on the description of p with µ the reduced mass of the system and ∆CC the en- heavy baryon-antibaryon systems in general. This sug- ergy difference of the transition. Coupled channel effects gests that ignoring explicit kaon and eta exchange, which can be argued to be suppressed by a factor of amounts to include them implicitly in the contact-range Q 2 couplings, is not likely to generate a sizable breakdown , (14) of SU(3)-flavour in the contact-range potential. The im- Λ CC plicit inclusion of eta and kaon effects in the contact- with ΛCC the coupled channel scale. For the ΣcΣc family range couplings will change their values from the ones of systems we have that ΛCC = 400/564 MeV depending expected from exact SU(3)-flavour symmetry, but proba- on the transition, while for the ΣbΣb case we have ΛCC = bly this change will be numerically smaller than the usual 350/495 MeV. This scale is softer than M, but not much 20% uncertainty associated with SU(3)-flavour symmetry softer: we can effectively ignore the coupled channels at relations. This point is supported by an analysis of the the price of reducing the range of applicability of the strength of the eta and kaon exchanges in Appendix B, EFT. where we also present the kaon and eta exchange poten- That is, there are two choices for constructing the EFT tials in case one wants to include them explicitly in the in this case: (i) consider the coupled channel effects to EFT. be subleading (at the price of reducing the convergence radius of the EFT), (ii) include them at leading order. Here we will opt for the first option, owing to its simplic- III. THE LEADING ORDER POTENTIAL ity. Be it as it may, most of the results of this manuscript can be easily extended to the coupled channel case. In this section we write down the heavy baryon- antibaryon potential at LO within the EFT expansion. This potential can contain a contact- and a finite-range 2 We mention in passing that the promotion can also be under- piece stood in terms of the anomalous dimension of the coupling C0, i.e. to its scaling with respect to the cut-off [43, 44]. (0) (0) (0) VEFT = VC + VF , (15) 5 where we are assuming NDA for the purpose of fix- which belong to the 6 representation of SU(3). ing the notation and simplifying the discussion. If If the heavy quark within the heavy baryons is a charm there are bound states the actual power counting of the quark, Q = c, the flavour components of the B3¯ field are heavy baryon-antibaryon system will differ from NDA, 0 see Sect. II B. But we will address this problem later in Ξc + Sects. IV and V. B3¯ = Ξ , (20) − c  The LO contact-range potential is a momentum- and Λ+ energy-independent potential in momentum space (or a  c    Dirac-delta in coordinate space). The LO finite-range where we follow the convention of Cho [46]. For the B6 potential is the OPE potential 3. For the contact-range field the flavour components are component we cannot determine if it is perturbative or ′ not a priori without resorting to experimental or phe- Σ++ 1 Σ+ 1 Ξ+ c √2 c √2 c nomenological input. For the finite-range components, ′ B = 1 Σ+ Σ0 1 Ξ0 . (21) 6  √2 c c √2 c  i.e. the pion exchanges, the situation is different and ′ ′ 1 + 1 0 0 we can in fact determine if they are perturbative, see  √ Ξc √ Ξc Ωc   2 2  Sect. IV. The discussion about the possible power count-   ings that arise depending on which pieces of the EFT For the B6∗ baryons we have exactly the same compo- potential are perturbative and non-perturbative will be nents as for B6, but with a star to indicate that they presented later in Sect. V. are spin-3/2 baryons. Depending on the case, it can be This section is organized as follows: we begin by ex- practical to simply consider the SU(2)-isospin structure plaining the details of how the heavy baryons are or- rather than the complete SU(3)-flavour one. ganized in superfields that are well-behaved according The fields B3¯, B6 and B6∗ can be organized into the to HQSS in Sect. IIIA. Next we will consider the C- superfields T and S, which have good transformation and G-parity properties of the heavy baryon-antibaryon properties under rotation of the heavy quark. For non- system in Sect. III B. After this, we will first introduce relativistic heavy baryons we write the superfields as [46] the general form of the contact-range potential and the 1 constrains imposed on it by heavy-quark spin symmetry, T = B¯ , S~ = ~σB + B~ ∗ , (22) 3 √ 6 6 SU(2)-isospin and SU(3)-flavour symmetry in Sect. III C. 3 Last, we will present the general form of the OPE poten- where the letters T and S stands for (anti-)triplet and tial and its partial wave projection in Sect. III D. Owing sextet. The definition of the non-relativistic superfield T to the scope of the discussion, the notation will be com- is redundant — it acts merely as a second name for B3¯ — plex. We overview the most used notation in this section but we include it for completeness. Notice that we have in Table I. ~ written the spin-3/2 heavy baryon field as a vector: B6∗. The reason is that this is a Rarita-Schwinger field, where the spin-3/2 nature of this field is taken into account by A. The Heavy Baryon Superfields coupling a spatial vector with a Dirac spinor and then projecting to the spin-3/2 channel with the condition ~σ Heavy baryons have the structure ~ · B6∗ = 0. Under rotations of the heavy quark spin the Q(qq) , (16) superfields behave as | i i ~ǫ ~σ i ~ǫ ~σ where Q is the heavy quark and qq the light quark pair, T e 2 · Q T, S~ e 2 · Q S.~ (23) which is in S-wave. The light quarks can couple their → → spin to SL = 0, 1. If the light spin is SL = 0, we have a For a more complete account of the heavy baryon fields P 1 + and superfields we refer to Appendix A. J = 2 heavy baryon

B¯ = Q(qq) . (17) 3 | SL=0i This type of heavy baryon belongs to the 3¯ representation B. C- and G-Parity of the SU(3) ﬂavour group. If the light spin is SL = 1, + + we have instead a J P = 1 or a J P = 3 baryon We are considering heavy baryon-antibaryon states. 2 2 If a state is electrically neutral and does not have strangeness, then C-parity will be a well-deﬁned quan- B6 = Q(qq)SL=1 , (18) | i J=1/2 tum number. If the heavy baryon and antibaryon have P identical SL and J , i.e. if they have the structure B6∗ = Q(qq)SL=1 , (19) | i J=3/2 ¯ ¯ ¯ B3¯B3¯ , B6B6 , B6∗B6∗ , (24) | i | i | i then the C-parity of the system is 3 Regarding the exchange of the other SU(3) Nambu-Goldstone bosons, we refer to the discussion in Section II D. C = ( 1)L+S , (25) − 6

Symbol Meaning

B3¯ Antitriplet heavy baryon field 1 B6 Sextet heavy baryon field, ground state (J = 2 ) 3 B6∗ Sextet heavy baryon field, excited state (J = 2 ) M3¯ Antitriplet heavy baryon mass 1 M6 Sextet heavy baryon mass, ground state (J = 2 ) 3 M6∗ Sextet heavy baryon mass, excited state (J = 2 ) T Antitriplet heavy baryon superfield S Sextet heavy baryon superfield C (i) C-parity (ii) Coupling of the momentum and energy independent contact-interaction G G-parity A(M) Antitriplet-antitriplet (A), antitriplet-sextet (B) and sextet-sextet (C) contact interactions SL (M) (M) B , B M refers to the SU(3)-flavour representation, SL to total light-quark spin SLD SLD C(M) D and E to whether it is a direct or exchange term SL I~i Generic isospin operator for vertex i = 1, 2 of the two-body potential

~ti Isosin-0 to isospin-1 transition matrices 1 ~τi Isospin- 2 Pauli matrices T~i Isospin-1 matrices

~ai Generic spin operator for vertex i = 1, 2 of the two-body potential 1 ~σi Spin- 2 Pauli matrices ~ ~+ 1 3 Si, Si Spin- 2 to spin- 2 transition matrices ~ 3 Σi Spin- 2 angular momentum matrices C12 Spin-spin operator (~a1 ~a2) · S12 Tensor operator (3~a1 rˆ ~a2 rˆ ~a1 ~a2) 2S+1 · · − · LJ Spectroscopic notation for partial waves with S the spin, L the orbital angular momentum and J the total angular momentum

C12 Matrix elements of the spin-spin operator in the partial wave basis

S12 Matrix elements of the tensor operator in the partial wave basis

TABLE I. List of the most used symbols in Sect. III. with L and S the orbital angular momentum and spin 4. fields. With this convention we define the states Examples of this type of heavy baryon-antibaryon system ¯ 1 ¯ ¯ + 0 ¯ 0 + B3¯B6(η) = B3¯B6 + η B6B3¯ , (29) are Λc Λc−, ΣcΣc and Σc∗ Σc∗−. | i √2 | i | i P If the light-quark spin SL and spin-parity J of the ¯ 1 ¯ ¯ heavy baryon and antibaryon are not identical, we first B3¯B6∗(η) = B3¯B6∗ + η B6∗B3¯ , (30) | i √2 | i | i have to choose a C-parity convention, for instance: 1 B6B¯∗(η) = B6B¯∗ + η B∗B¯6 , (31) | 6 i √2 | 6 i | 6 i C B¯ =+ B¯¯ , (26) 3 3 | i | ¯ i where η = 1, for which the C-parity is C B6 =+ B6 , (27) ± | i | i C = η ( 1)L+S , (32) C B∗ = B¯∗ , (28) | 6 i −| 6 i − where L (S) is the total orbital angular momentum (spin) 5 where there is a relative minus sign for the C-parity trans- of the heavy baryon-antibaryon pair′ . Examples′ of this 3 1 + 0 0 + type of molecule include Λ Σ−, Ξ Ξ¯ and Ξ Ξ∗−. formation of the spin- 2 fields with respect to the spin- 2 c c c c c c

5 The C-parity is the product of the intrinsic C-parity and the symmetry factor of exchanging the particles, which now includes 4 This comes from multiplying the intrinsic C-parity of a fermion- a contribution from η, i.e. C = (+η) × (−1) × (−1)L × (−1)S+1 L S antifermion system with the symmetry factors of exchanging the for B3¯B6 and C = (−η) × (−1) × (−1) × (−1) for B3¯B6∗ / L S+1 particles, i.e. C = (−1) × (−1) × (−1) . B6B6∗. 7

For a heavy baryon-antibaryon state that is not elec- where SL is the total light-quark spin of the heavy trically neutral but has no strangeness and belongs to baryon-antibaryon system. The dSL ’s are coefficients the same SU(2) isospin representation as a neutral state, that depend on the heavy- and light-quark decomposi- C-parity is not a well-defined quantum number but there tion of the specific heavy baryon-antibaryon molecule, exists an extension that includes isospin. This extension see Appendix C for details. is G-parity [47], which can be defined as follows The contact-range couplings of the T T¯, ST¯ and SS¯ molecules are independent and we will use a different iπI2 G = Ce , (33) notation for each case: A, B and C respectively. For the ¯ that is, a C-parity transformation combined with a rota- T T system we write the contact-range potential as tion in isospin space 6. For a electrically charged state, (0) T T¯ V T T¯ = A0 , (37) the G-parity is well-defined and its eigenvalues are h | C | i where we are already taking into account that the total G = C ( 1)I , (34) − light quark spin is always SL = 0 (we also ignore the coefficient because there is actually no decomposition). where I is the isospin of the electrically charged state For the ST¯ system we write and C is the C-parity of the electrically neutral compo- (0) nent of the isospin multiplet. For example if we consider ST¯ V ST¯ = d1DB1D , (38) ++ ¯ 0 h | C | i Σc Σc, its isospin is I = 2 and the neutral component (0) 0 ¯ 0 ST¯ V T S¯ = d1EB1E , (39) of its isospin multiplet is a linear combination of ΣcΣc, h | C | i + ++ ++ ¯ 0 Σc Σc− and Σc Σc−−: the G-parity of Σc Σc is then where the total light spin is always SL = 1, but where G = ( 1)I C = ( 1)L+S. − − we have to make the distinction between a diagonal and non-diagonal potential. For the SS¯ system we have

C. The Contact-Range Potential SS¯ V (0) SS¯ = d C + d C + d C , (40) h | C | i 0 0 1 1 2 2 where the total light spin is SL = 0, 1, 2. We list The LO contact-range potential takes the generic form the contact-range potential for heavy baryon-antibaryon (0) molecules with well-defined C-parity in Table II, which p′ V p = C, (35) h | C | i also applies by extension to the molecules with well- with C a coupling constant and where p (p′) is the center- defined G-parity. of-mass momentum of the incoming (outgoing) heavy If the molecules do not have well-defined C- or G-parity baryon-antibaryon pair. In principle there should be one (i.e. molecules with strangeness), the form of the contact- independent coupling C for each J PC quantum num- range potential depends on the particular case. For the ber and type of heavy baryon-antibaryon molecule. But family of molecules the contact-range potential is constrained by HQSS and ¯ ¯ ¯ B3¯B3¯ , B6B6 , B∗B∗ , (41) SU(3)-flavour symmetry, which greatly reduces the num- | i | i | 6 6 i ¯ ¯ ′ ¯ ber of possible couplings. We first consider the HQSS (e.g. ΛcΞc, ΣcΞc, Σc∗Ξc∗) the contact-range couplings are structure of the contact-range potential and then the exactly as shown in Table II for the case in which C- SU(3) flavour one. parity is well-defined. For the molecules involving different types of heavy hadrons the contact-range potentials are defined in coupled channels. If we consider the bases 1. HQSS structure ¯ ¯ 1 = B3¯B6 , B6B3¯ , (42) B {| ¯ i | ¯ i} The application of HQSS to the heavy baryon- 2 = B3¯B6∗ , B6∗B3¯ , (43) B {| ¯ i | ¯ i} antibaryon system implies that the contact-range cou- 3 = B6B∗ , B∗B6 , (44) B {| 6 i | 6 i} pling does not depend on the heavy quark spin, only on we end up with the following contact-range potentials the light quark spin. This means that the coupling C can be decomposed in terms of light-quark components B1D B1E V 1 (0−)= − , (45) B B1E B1D ! C = dSL CSL , (36) − 1 X B1D 3 B1E V 1 (1−)= , (46) B 1 3 B1E B1D ! 6 Notice that the G-parity transformation is sometimes defined as iπI2 1 G = Ce− . with a minus sign. For particles with integer B1D 3 B1E isospin this is equivalent to the definition with a plus sign, G = V 2 (1−)= 1 − , (47) +iπI B B B Ce 2 . For particles with half-integer isospin each of these − 3 1E 1D ! conventions generate anti-particle states that differ by a sign. This has no observable consequence, as it amounts to a global B1D B1E V 2 (2−)= − . (48) redefinition of the amplitudes by a phase. B B B − 1E 1D ! 8

P PC System J /J VC (HQSS) System Type Isospin VC ¯ + ¯ ¯ 1 (1) 2 (8) B3¯B3¯ 0− A0 ΛcΛc T T 0 3 A + 3 A ¯ ¯ ¯ 2 (1) 1 (8) B3¯B3¯ 1−− A0 ΞcΞc T T 0 3 A + 3 A + (8) B¯B¯6 0− B1D B1E ΞcΞ¯ c T T¯ 1 A 3 − ¯ ¯ ¯ (8) B3¯B6 0−− B1D + B1E Ξc′ Ξc ST 0 B + 1 1 (8) 2 (10) B¯B¯6 1− B1D B1E Ξ′ Ξ¯ c ST¯ 1 B + B 3 − 3 c 3 3 ¯ 1 ¯ ¯ 2 (8) 1 (10) B3¯B6 1−− B1D + 3 B1E ΣcΛc ST 1 3 B + 3 B ¯ + 1 ¯ ¯ 1 (1) 8 (8) 3 (27) B3¯B6∗ 1− B1D + 3 B1E ΩcΩc SS 0 6 C + 15 C + 10 C ¯ 1 ¯ ¯ 1 (1) 1 (8) 3 (27) B3¯B6∗ 1−− B1D 3 B1E Ξc′ Ξc′ SS 0 3 C + 15 C + 5 C + − 1 (1) 2 (8) 1 (27) B¯B¯∗ 2− B1D B1E ΣcΣ¯ c SS¯ 0 C + C + C 3 6 − 2 5 10 ¯ ¯ ¯ 1 (8) 4 (27) B3¯B6∗ 2−− B1D + B1E Ξc′ Ξc′ SS 1 5 C + 5 C ¯ + 1 2 ¯ ¯ 4 (8) 1 (27) B6B6 0− 3 C0 + 3 C1 ΣcΣc SS 1 5 C + 5 C ¯ 1 6 20 ¯ ¯ (27) B6B6 1−− 27 C0 + 27 C1 + 27 C2 ΣcΣc SS 2 C ¯ + B6B6∗ 1− C1 ¯ 16 6 5 B6B6∗ 1−− 27 C0 + 27 C1 + 27 C2 + 1 2 TABLE III. SU(3)-ﬂavour decomposition of the contact- B B¯∗ 2− C + C 6 6 3 1 3 2 range couplings depending on the type heavy baryon- ¯ B6B6∗ 2−− C2 antibaryon molecule. Notice that the HQSS structure of the ¯ + 2 1 B6∗B6∗ 0− 3 C0 + 3 C1 couplings is independent of the SU(3) one, which is why we ¯ 10 15 2 do not show the light-spin indices for the couplings. For the B6∗B6∗ 1−− 27 C0 + 27 C1 + 27 C2 + 2 1 S heavy baryons we show the decomposition for the lightest B∗B¯∗ 2− C + C 6 6 3 1 3 2 member of the HQSS multiplet only. ¯ B6∗B6∗ 3−− C2

TABLE II. HQSS decomposition of the contact-range cou- For the ST¯ case we have 6 3=8 10: plings for heavy baryon-antibaryon molecule with well-deﬁned ⊗ ⊕ C-parity. A, B and C refers to the coupling of the T T¯, ST¯ ¯ and SS system respectively, with the subscript indicating the B B(8),B(10) , (52) light-quark spin decomposition. For the cases where C-parity 1D →{ 1D 1D } is not well-deﬁne, we refer to the explanations in the main (8) (10) B1E B1E ,B1E . (53) text. →{ }

Finally for the SS¯ case we have 6 6=1¯ 8 27: 16 C0+33 C1+5 C2 16 C0 21 C1+5 C2 ⊗ ⊕ ⊕ 54 − 54 V 3 (1−)= ,(49) B 16 C0 21 C1+5 C2 16 C0+33 C1+5 C2 − ! (1) (8) (27) 54 54 C C , C , C . (54) SL SL SL SL 1 5 1 1 →{ } 6 C1 + 6 C2 6 C1 6 C2 V 3 (2−)= 1 1 1 − 5 . (50) B C1 C2 C1 + C2! 6 − 6 6 6 The decomposition for a specific molecule can be consulted in Table III, which have been obtained from the ¯ ¯ ¯ Examples of bases 1, 2 and 3 are the ΞcΣc-Ξc′ Λc, ΞcΣc∗- SU(3) Clebsch-Gordan coefficients for 3¯ 3, 6 3 and ¯ ¯ ¯ ⊗ ⊗ Ξc∗Λc and Ξc′ Σc∗-Ξc∗Σc systems. 6 6¯ of Ref. [48]. Notice that we are not explicitly considering⊗ the SU(2)-isospin structure as it is a subgroup of SU(3)-flavour.

2. SU(3)-flavour structure Finally, we remind that the SU(3)-flavour structure of the contact-range potential can be broken if the finite- range potential is not SU(3)-flavour symmetric. Whether Besides HQSS, heavy baryon-antibaryon systems also this happens depends on two factors. The first is the have SU(3)-flavour symmetry. In SU(3)-flavour the T particular power counting we are using and the order we and S heavy baryons belong to the antitriplet and sextet are considering within the EFT expansion, e.g. if the ¯ ¯ representation (3 and 6), respectively. For the T T the contact-range interaction is leading, but the exchange HQSS coupling is further divided into the SU(3)-flavour of pions, kaons and etas is subleading, the violations of representations 3¯ 3=1 8, a singlet and an octet. ⊗ ⊕ SU(3)-flavour symmetry if we ignore kaon and eta ex- That is, there are two independent SU(3)-flavour contact changes will be subleading. The second factor is that interactions kaon and eta exchanges are parametrically small, as was explained in Sect. II D and where a more detailed deriva- A A(1), A(8) . (51) tion can be found in Appendix B. 0 →{ 0 0 } 9

Vertex Ri R¯i I~i ~ai D. The One Pion Exchange Potential 2 2 ~ Λc Σc q 3 -q 3 ti ~σi → The OPE potential in momentum space reads Λc Σ∗ √2 √2 ~ti S~† → c i 2 2 1 (0) Ξc Ξ′ - ~τi ~σi ¯ ¯ → c q 3 q 3 2 T1T2 VF T1′T2′ =0 , (55) 1 h | | i Ξc Ξ∗ √2 √2 ~τi S~† 2 → c 2 i (0) g3 ~a1 ~q~a2 ~q 2 2 T1S¯2 V S′ T¯′ = R1 R¯2 I~1 I~2 · · , Σc Σc - T~i ~σi F 1 2 2 2 2 3 3 h | | i − 2fπ · q + µπ → 1 1 Σ∗ Σc T~i S~i c → √3 √3 (56) 1 1 ~ ~ Σc Σc∗ Ti Si† 2 → √3 √3 (0) g2 ~a1 ~q~a2 ~q 2 2 ~ ~ S1S¯2 V S′ S¯′ = R1 R¯2 I~1 I~2 · · , Σc∗ Σc∗ - Ti Σi F 1 2 2 2 2 3 3 h | | i − 2fπ · q + µπ → 2 2 1 Ξ′ Ξ′ - ~τ ~σ c c 3 3 2 i i (57) → 1 1 1 Ξ∗ Ξ′ ~τi S~i c → c √3 √3 2 1 1 1 Ξ′ Ξ∗ ~τi S~† where we have chosen the specific notation above to cover c → c √3 √3 2 i 2 2 1 all the possible combinations. The subscripts 1 and 2 are Ξ∗ Ξ∗ - ~τi Σ~ i c → c 3 3 2 used to denote the vertices 1 and 2 in the diagrams of Fig. 1. In the equation above R1 and R¯2 are numeri- TABLE IV. Numerical, isospin and spin factor associated cal factors which depend on the transition we are con- with each vertex in the ST¯ V T S¯ and SS¯ V SS¯ heavy sidering, see Table IV (the bar indicates an antibaryon h | | i h | | i baryon-antibaryon potential. The arrows are used to indicate to antibaryon transition). I~1 and I~2 are isospin matri- ~ the final baryon state in the vertex. The symbols ~τi and Ti ces, while ~a1 and ~a2 spin matrices. For the couplings we represent the Pauli matrices (in isospin space) and the isospin have that g2 is the axial coupling for the SL = 1 heavy I = 1 matrices respectively, while t is a special isospin matrix i baryon, g3 the coupling involved in T Sπ transitions for connecting the isoscalar Λc with the isovector Σc and the → ~ and fπ the pion decay constant. µπ is the effective pion pion. The symbols ~σi and Σi are the Pauli matrices and the mass for the vertices involved in the particular channel spin S = 3/2 matrices, while S~i is a 2 4 matrix that is used × considered. Finally we notice that OPE vanishes for the for the transitions involving a spin 1/2 and spin 3/2 baryon. ¯ Notice that this table can also be used to compute the heavy T T molecules, which can be described solely in terms of baryon-baryon potential. contact-interactions at lowest order. Regarding the isospin structure of the OPE potential, we have that I~1 and I~2 are the isospin matrices corresponding to vertex 1 and 2. If we have a heavy baryon with isospin 1/2 at vertex i we can simply make the sub- ~ ~τi stitution Ii = 2 , where ~τ are the Pauli matrices. If the heavy baryon at vertex i has isospin 1 we use the J =1

S′ T ′ S′ S′ angular momentum matrices in isospin space, for which 1 2 1 2 we use the notation T~i, i.e. we make the substitution I~i = T~i. The exact isospin factor for each type of vertex can be consulted in Table IV. Regarding the spin structure, we note that the spin 1 2 1 2 operators a1 and a2 depend on which is the initial and final spin of the heavy baryons at vertex 1 and 2. If the initial and final heavy baryons at vertex 1(2) have spin 1/2 we have ~a1 = ~σ1 (~a2 = ~σ2). If the initial and final heavy baryons at vertex 1(2) have spin 3/2 we have ~a1 = T S S S 1 2 1 2 Σ~ 1 (~a2 = Σ~ 2), where Σ~ are the S =3/2 spin matrices. If the initial and final heavy baryons at vertex 1(2) switch from spin 1/2 to spin 3/2 (or viceversa), then ~a = S~ FIG. 1. Diagrams for OPE potential between two heavy 1 1 ~ ~ hadrons. In the left we show the TS potential and in the (~a2 = S2), where S are special 2x4 spin matrices that right the SS potential, where T (S) is the heavy baryon with describe the transition from a different initial to final spin light spin SL =0(SL = 1). (see their definition in Appendix A). As with isospin, the exact spin matrix to use in each type of vertex can be checked in Table IV. Regarding the axial couplings g2 and g3, we notice ++ that the value of g3 can be extracted from the Σc + + → Λc π decay measured by Belle in Ref. [49], yielding +0.019 g3 = 0.973 0.042 [50] (notice that the previous reference originally uses− the convention of Yan et al. [51] to define 10 the axial couplings, instead of the one by Cho [46, 52] 1. The OPE Potential in Coordinate Space that we employ here and we have consequently adapted the numbers of Ref. [50] to our convention). In con- If we Fourier-transform the potential into coordinate trast g2 is experimentally unavailable, but on the ba- space we obtain sis of quark model relations one can estimate it to be (0) g2 = √2g3 = 1.38. If we consider the values of g2 T S¯ V (~r) S′ T¯′ = − − h 1 2| F | 1 2i and g3 from the lattice QCD calculation of Ref. [53], we 2 ¯ ~ ~ g3 3 obtain instead g2 = 0.84 0.20 and g3 = 0.71 0.13, R1 R2 I1 I2 2 C12δ (~r) where it is important− to mention± that they are calculated± − · 6 fπ in the mQ = limit (notice again that our convention +R1 R¯2 I~1 I~2 C12 WC (r) for g differs∞ from the definition used in Ref. [53] by a · 2 h sign, which has been taken into account). Had we ap- +S12(ˆr) WT (r) , (60) plied the quark model relations to the lattice QCD value (0) i of g , we would have obtained g = 1.00, which is con- S1S¯2 V (~r) S′ S¯′ = 3 2 h | F | 1 2i siderably larger than the lattice QCD− determination but 2 ¯ ~ ~ g2 3 yet within its error bar. R1 R2 I1 I2 2 C12δ (~r) − · 6 fπ For the effective pion mass, we have that if the particles +R R¯ I~ I~ C W (r) 1 2 1 · 2 12 C in the vertex 1 and 2 have the same mass, then µπ = mπ. h On the other hand if they have different masses (e.g. +S12(ˆr) WT (r) , (61) S1 = B, S1′ = B′) and the mass splitting is given by ∆, i 2 2 2 then we have that µπ = mπ ∆ (a relation that assumes where C12 and S12 are the spin-spin and tensor operators, heavy, non-relativistic baryons).− defined as C = ~a ~a , (62) Finally we notice that we can also compute the heavy 12 1 · 2 baryon-baryon potential potential by making the change S12(ˆr)=3 ~a1 r~aˆ 2 rˆ ~a1 ~a2 . (63) · · − · The OPE potential contains a Dirac-delta contribution which can be reabsorbed into the contact-range potential if one wishes too. The spin-spin and tensor pieces of the R1R¯2 R1R2 , (58) potential W and W can be written as → C T 2 3 µπr gi µπ e− WC = 2 , (64) 24πfπ µπr g2µ3 e µπr 3 3 in Eqs. (56) and (57) and consulting the proper values in W = i π − 1+ + , (65) T 24πf 2 µ r µ r (µ r)2 Table IV. π π π π where g = g or g depending on the case and µ is The most explicit way to construct the potential for i 2 3 π the effective pion mass for the channel under considera- one particular channel is to make use of Table IV, where tion. The spin-spin piece of the OPE potential is often all the factors are listed. For instance if we are con- referred to as central OPE, a naming convention that of- sidering the Ξ Σ¯ Ξ Σ¯ potential, we can see that c∗ c c′ c∗ ten appears in nuclear physics for historical reasons and it contains a Ξ → Ξ transition in vertex 1 and a c∗ c′ which permeates the notation W and W . Central is Σ¯ Σ¯ transition→ in vertex 2. If we use Table IV we C T c c∗ used in opposition to tensor to convey the idea that the → √ ~ find R1 = 1/ 3, I1 = ~τ1/2, ~a1 = S1 for vertex 1 and central piece carries no orbital angular momentum (while ¯ ~ ~ R2 = 1/√3, I2 = T1, ~a2 = S2† for vertex 2. Putting the the tensor piece carries two units of orbital angular mo- pieces together the potential reads mentum). The term central OPE is indeed convenient and we will use it in what follows (instead of the more accurate spin-spin OPE). We notice that the OPE potential also contains a contact-range contribution, which 2 ~ ~ ~ is mostly harmless: it can be reabsorbed into the EFT ¯ (0) ¯ 1 g3 ~τ1 T2 S1 ~q S2† ~q Ξc∗Σc VF Ξc′ Σc∗ = 2 · 2· 2· ,(59) contact-range contribution to the potential by a redefini- h | | i −3 2fπ 2 q + µπ tion of the couplings. Hence it can be simply ignored.

2. Partial Wave Projection of the OPE Potential 2 2 2 where µ = m ∆ , with ∆ (m(Ξ∗) m(Ξ′ )) π π − ≃ c − c ≃ (m(Σc∗) m(Σc)) 70 MeV and µπ 120MeV. The We consider now the projection of the coordinate space other cases− can be≃ obtained analogously.∼ potential into the partial wave basis. For that we work 11 with baryon-antibaryon states with well-defined total an- the lowest order two pion exchange irreducible diagrams gular momentum and parity J P . If the total strangeness enter at order Q2 in the chiral expansion. of the baryon-antibaryon state is zero, we will consider states with well-defined C-parity (for neutral systems) or G-parity (if the system is not electrically neutral). Be- A. The Central Potential sides we will only consider states that contain an S-wave, as they are the more likely to form a bound state. If The perturbative nature of the central piece of OPE 2S+1 we use the spectroscopic notation LJ to denote the can be determined from the comparison of tree-level ver- partial waves, we have the following combinations sus once-iterated central OPE, i.e. V and V G0V in oper- 1 ator form. This type of comparison was made in Ref. [54] B B¯ (0−) = S , (66) | Q 6 i { 0} in the context of nucleon-nucleon scattering 7. Here we ¯ 3 3 BQB6(1−) = S1, D1 , (67) are merely adapting it to the particular case of the heavy | i { } baryon-antibaryon system. The ratio of iterated vs tree- 3 3 5 level OPE can be expressed as a ratio of scales B B¯∗(1−) = S , D , D , (68) | Q 6 i { 1 1 1} 3 5 5 5 B B¯ (2 ) = D , S , D , G , (69) p V G0V p Q Q 6∗ − 2 2 2 2 h | | i = , (75) | i { } p V p Λ h | | i C ¯ 1 5 B6∗B6∗(0−) = S0, D0 , (70) where Q is a light scale (either the external momentum p | i {3 3 }7 7 B∗B¯∗(1−) = S , D , D , G , (71) or the pion mass mπ) and ΛC is a scale that characterizes | 6 6 i { 1 1 1 1} ¯ 1 5 5 5 central OPE. The evaluation of this ratio at p = 0 leaves B6∗B6∗(2−) = D2, S2, D2, G2 , (72) | i {3 3 7 7 }7 7 the pion mass as the only light scale left, in which case B∗B¯∗(3−) = D , G , S , D , G , I , (73) | 6 6 i { 3 3 3 3 3 3} we obtain the following value for the central scale with BQ = B3¯ or B6. The calculation of the matrix el- 1 24πf 2 ements is in general straightforward, where we refer to Λ = π , (76) C σ τ µ R R¯ g2 Appendix A for the details. The result of these calcu- | | | 1 2| i lations is that the C12 and S12(ˆr) operators can be ex- with µ the reduced mass of the system, σ and τ the eval- C pressed as matrices, which we denote with the 12 and uation of the spin and isospin operators corresponding S 12 notation. With this in mind for r > 0 we write the to the particular case under consideration and where R1 OPE potential as and R¯2 can be consulted on Table IV. For the charm and (0) ¯ ~ ~ C S bottom sectors the value is respectively VF (r)= R1R2 I1 I2 [ 12WC (r)+ 12WT (r)] , · 1060 MeV (74) Λ (Q = c)= , (77) C στ R R¯ g2 | | | 1 2| i where the dimension of the matrices is set by the number 450 MeV of partial waves. The explicit matrices that apply in each Λ (Q = b)= , (78) C στ R R¯ g2 case can be consulted in Appendix A, where it is also | | | 1 2| i explained how they are calculated. which depend on the value of the couplings g2 and g3. The discussion about the values of the axial couplings, and in particular g2, is important because it can change IV. HOW TO COUNT THE ONE PION the value of ΛC by a large factor. For the T S¯ and SS¯ EXCHANGE POTENTIAL molecules the value of ΛC in the charm sector is +100 The LO EFT heavy baryon-antibaryon potential can T S¯ 1120 40 MeV ΛC (Q = c) − , (79) in principle contain a contact- and a finite-range piece, ∼ στ R1R¯2 where the latter is the well-known OPE potential. While | | | | ¯ (560 1500) MeV there is no a priori way to determine if the contact- ΛSS(Q = c) − , (80) C ∼ στ R R¯ range potential is perturbative, this is not the case for | | | 1 2| the OPE potential where there exist a series of theoret- where ΛSS¯ can change almost by a a factor of 3 owing ical developments to evaluate its strength. In this sec- C to the uncertainty of g2 (notice that instead of a num- tion we will check how these ideas apply to the central ber with an error, we have simply indicated a range of and tensor pieces of the OPE potential. Before start- possible values). In the bottom sector it is instead more ing the discussion it is important to stress that we make a very explicit distinction between iterated OPE (or re- ducible multi-pion exchange) and irreducible multi-pion exchanges. The former is merely the outcome of iterating 7 Recently a more sophisticated method for determining the per- the EFT potential within the Schrödinger or Lippmann- turbativeness of OPE has been developed in Ref. [55] for periph- Schwinger equations while the latter is a genuine contri- eral waves with L ≥ 1. Unfortunately it has not been extended bution to the EFT potential, though a subleading one: yet to S-waves. 12 advisable to use the lattice QCD determination for g2 der to obtain a finite result again. For the perturbative 3 and g3, leading to series of the 1/r potentials the quantity we renormalize +440 is the angular momentum. The zero-th order term in the T S¯ 900 260 MeV ΛC (Q = b) − , (81) perturbative expansion of the wave function is now ∼ στ R1R¯2 | | | | J (kr) 660+440 MeV (0) ν(ka3) SS¯ 220 Ψl (r; k)= , (83) ΛC (Q = b) − . (82) √r ∼ στ R1R¯2 | | | | Jl+1/2(kr) The previous values have to be combined with the instead of the standard , where Jν (z) refers to ¯ √r στ R1R2 factor. The maximum value of this factor the Bessel function of order ν. In the expression above ν | | | | happens for the channels with lowest spin and isospin. is the renormalized angular momentum, which happens In Table V we list the scale ΛC for a few represen- to be a function of the momentum k and a length scale tative heavy baryon-antibaryon molecules. In general a3 that is related to the strength of the potential (it will ΛC 0.5GeV (if not harder) in most cases, which means be defined later). The secular series is built not only ∼ that we expect central OPE to be perturbative. The by adding higher order terms but also by making ν(x) + ¯ + exceptions are the isoscalar 0− ΣQΣQ and 0− , 1−− to depend on κ = ka3. If we switch off the potential 1 Σ∗ Σ¯ ∗ molecules, at least in the bottom sector. In the Q Q and take ka3 = 0, we have ν(0) = l + 2 and we recover charm sector the scale ΛC varies considerably as a con- the free wave function. For small enough values of κ we sequence of the uncertainty of the axial coupling g2. In expect ν(κ) to be expansible in powers of x, i.e. to be particular if the absolute value of the axial coupling g2 perturbative. By reexpanding the secular series and the | | is on the high end, i.e. the value g2 = 1.38 deduced renormalized angular momentum we can recover the orig- − from the quark model, central OPE will be important for inal perturbative series. However the interesting feature certain molecules in the hidden charm sector. In the bot- of the series above is that we can determine the values of tom sector the situation is more clear: central OPE will κ for which ν(κ) is analytic. When ν(κ) is not analytic, ¯ be non-perturbative for the aforementioned ΣQΣQ and it does not admit a power series in κ anymore. This in ¯ ΣQ∗ ΣQ∗ molecules. Finally for having a comparison with turn means that there is no way to rearrange the secular a well-known state, we mention that the central scale for series into the standard perturbative series, leading to its OPE in the two-nucleon system is ΛC 590 MeV. failure. ≃ For the 1/r3 potential, which is equivalent to the tensor force for distances mπr < 1, the secular series has been B. The Tensor Potential studied in detail by Gao [57] for the uncoupled channel case. Birse [45] extended the previous techniques for The tensor piece of the OPE potential requires a more the coupled channel case and particularized them for the involved analysis. A direct comparison of V and V G0V is nucleon-nucleon system. In a previous publication by one not possible. The reason is that the iteration of the tensor of the authors [32] the analysis of Birse was applied to the piece of OPE diverges, see for instance Refs. [32, 55] for a heavy meson-antimeson system. In this work we extend detailed explanation. Thus we must resort to a method it to the heavy baryon-antibaryon system. that does not involve the direct evaluation of iterated We will consider the tensor force in the limit mπ 0, tensor OPE. for which the OPE potential can be written as → The type of power-law behaviour of the tensor OPE a3 potential is analogous to the behaviour of a few phys- 2µ V(r)= Sj , (84) r3 ical systems studied in atomic physics. The potential between two dipoles is of the 1/r3 type, just like the where the potential is a matrix in the coupled channel S tensor force. The failure of standard perturbation the- space and j is the tensor operator (in matrix form, ory for these systems is well-known in atomic physics, as written in Sect. III D), with j referring to the to- where alternative techniques have been developed to deal tal angular momentum. We have that µ is the reduced with this type of potentials. The work of Cavagnero [56] mass of the heavy baryon-antibaryon system and a3 is explains that the divergences of perturbation theory in the length scale that determines the strength of the ten- these type of systems is similar to the role of secular sor piece of the potential. The potential in this limit is perturbations in classical mechanics, i.e. a type of per- amenable to the secular perturbative series developped turbation that is small at short time scales but ends up in Refs. [32, 45, 57]. The corrections stemming from the diverging at large time scales. The solution is to redefine finite value of mπ where considered in Ref. [32] and will (or, loosely speaking, renormalize 8) some quantity in or- be discussed later on in this section.

C. The Renormalized Angular Momentum 8 We simply adopt the terminology in use in the ﬁeld of atomic physics for these redeﬁnitions in secular perturbation theory, though it does not exactly corresponds with the standard mean- Now we explain how the secular perturbation theory ing of renormalization. looks like and most importantly, how to calculate the 13

Channel I Sign ΛC Channel I Sign ΛC ¯ +70 ¯ +290 ΞcΞc′ (0−±) 0 720 30 ΞbΞb′ (0−±) 0 580 170 ∓ − ∓ − ¯ +50 ¯ +220 ΛcΣc(0−±) 1 580 20 ΛbΣb(0−±) 1 450 130 ± − ± − ¯ + ¯ + +450 Ξc′ Ξc′ (0− ) 0 530 1420 Ξb′ Ξb′ (0− ) 0 620 210 − − − − ¯ + ¯ + +170 ΣcΣc(0− ) 0 210 560 ΣbΣb(0− ) 0 240 80 − − − − ¯ + ¯ + +490 Ξc∗Ξc′ (1− ) 0 580 1560 Ξb∗Ξb′ (1− ) 0 680 230 − − − − ¯ + ¯ + +190 Σc∗Σc(1− ) 0 230 610 Σb∗Σb(1− ) 0 260 90 − − − − ¯ ¯ +230 Σc∗Σc(1−−) 0 280 750 Σb∗Σb(1−−) 0 320 110 − − − − ¯ ¯ +230 Σc∗Σc(2−−) 0 + 280 750 Σb∗Σb(2−−) 0 + 320 110 − − ¯ + ¯ + +350 Ξc∗Ξc∗(0− ) 0 410 1100 Ξb∗Ξb∗(0− ) 0 490 170 − − − − ¯ + ¯ + +140 Σc∗Σc∗(0− ) 0 160 440 Σb∗Σb∗(0− ) 0 190 70 ¯ − − ¯ − +190− Σc∗Σc∗(1−−) 0 220 590 Σb∗Σb∗(1−−) 0 260 90 − − − − ¯ ¯ +230 Σc∗Σc∗(3−−) 0 + 270 740 Σb∗Σb∗(3−−) 0 + 320 110 − −

TABLE V. The central scale ΛC (in units of MeV) for a few selected heavy baryon-antibaryon molecules. For momenta above this scale, p > ΛC , the central force becomes non-perturbative. The scale ΛC is inversely proportional to στ, reaching its minimum for the channels with the lowest total spin and isospin and growing quickly for other conﬁgurations. The selection ( ) ¯ ( ) includes the heavy baryon-antibaryon states for which ΛC < 0.5 GeV, most of which are of the ΣQ∗ ΣQ∗ type. In addition we ′ ′ ( / ) ¯ ( / ) include the ΞQ ∗ ΞQ ∗ molecule for which the central force is strongest. and the antitriplet-sextet molecules. The uncertainty +0.019 in the antitriplet-sextet case comes from the axial coupling, which is taken to be g3 = 0.973 0.042 in the charm sector [50] and − g3 = 0.71 0.13 in the bottom one [53]. For the sextet-sextet case in the charm sector the axial coupling is not well determined ± and we take it in the range g2 0.84 1.38, while in the bottom sector we have g2 = 0.84 0.20 [53]. For comparison in | | ∼ − − ± the deuteron channel of the two-nucleon system the central scale is ΛC 590 MeV. ≃

S Channel κc survives, see Eq. (84). In the equation above j is the tensor operator matrix, while Lj is a diagonal matrix B3B¯6(0−)/B6B¯6(0−) - representing the angular momentum operator L~ 2 B3B¯6(1−)/B6B¯6(1−) 0.6835 ¯ B3B6∗(1−) 1.412 L2 = diag(l (l + 1),...,l (l + 1)) . (86) ¯ j 1 1 N N B3B6∗(2−) 1.934 ¯ + u B6B6∗(1− ) 0.8533 The reduced wave function k,j is an N N matrix, ¯ × B6B6∗(1−−) 0.7264 where column j represents a solution that behaves as ¯ + a free wave with angular momentum l when we take B6B6∗(2− ) 0.5784 j ¯ a3 0. The solution of the Schrödinger equation is a B6B6∗(2−−) 0.6998 → ¯ + linear combination of the functions ξ and η: B6∗B6∗(0− ) 0.6378 ¯ B6∗B6∗(1−−) 0.6674 u (r)= α ξ (r; k)+ β η (r; k) , (87) + k,j lj lj lj lj B∗B¯∗(2− ) 0.6833 6 6 l ¯ X{ j } B6∗B6∗(3−−) 0.5922 where we sum over the possible values of the angular momenta and with ξ and η N-component vectors that TABLE VI. Reduced critical momenta κc for the different can be written as sums of Bessel functions: S-wave heavy baryon-antibaryon systems B′B¯. ∞ ξ (r; k)= b (ν ) √rJ (kr) , (88) lj m lj m+νlj m= renormalized angular momenta ν. We begin by writ- X−∞ ing the reduced Schrödinger equation in coupled channels ∞ m √ (the uncoupled channel can be consulted in Ref. [32]) for ηlj (r; k)= ( 1) b m(νlj ) rJ m νl (kr) , − − − − j 3 m= the particular case of a pure 1/r potential X−∞ (89) a L2 u + S 3 + j u (r)= k2 u (r) , (85) k,j′′ j 3 2 k,j k,j where the ν ’s are functions of κ = ka , i.e. ν = ν (κ). − " r r # lj 3 lj lj The νlj ’s are the renormalized angular momenta that we where we are considering N angular momentum channels. previously introduced in Eq. (83). In turn the expansion Notice that we have taken the chiral limit mπ 0, which of the reduced wave functions ξ and η in Eqs. (88) and means that only the tensor piece of the OPE→ potential (89) is simply the extension of Eq. (83) to arbitrary or- 14

9 ders . We have N different solutions for ξ and η that matrix Sj , which can be found in Eq. (A138) and we have labeled with the subscript lj to indicate that for Eqs. (A147-A150) of Appendix A, then plug this matrix κ = 0 they behave as a free wave of angular momentum into Eq. (92), from which finally we solve Eq. (93) to ob- ¯ lj . The recursive relation from which one can compute tain κc. The B3B6∗ molecules the procedure is the same, bm(νlj ) can be consulted in Ref. [45] but are of no con- with the tensor force matrices defined in Eqs.(A167) ¯ cern if we are only interested in the νli ’s. For κ = 0 only and (A168). The B6B6∗ molecules are the most compli- the bm(νlj ) coefficient for m = 0 survives. cated because they contain a direct and exchange ten- ¯ ¯ The renormalized angular momenta νli = νli (κ) (with sor operator, which mediate the B6B6∗ B6B6∗ and ¯ ¯ → κ = ka3) can be calculated as follows. First we define B6B6∗ B6∗B6 potential, respectively. In addition the the following N N matrix effective→ pion masses are different for the direct and ex- × change tensor operators. Here we ignore this effect: in κ2 Fj (ν,κ) fj (ν) [R1(ν) R1( ν)] , (90) the present calculation we are making the approximation ≡ − ν − − that HQSS is exact and therefore there is no mass split- which depends on two other matrices, fj (ν) and R1(ν); ting between the B6 and B6∗ heavy baryons. However the fj (ν) is a diagonal matrix defined as length scale a3 is different for the direct and exchange operators, i.e. f j (ν) 2 1 2 2 1 2 D E = diag(ν (l1 + ) ,...,ν (lN + ) ) , a3 D a3 E 2 ν 2 2 2µ V ¯∗ (r)= S + S . (95) − − B6B6 3 j 3 j (91) r r It happens that both scales are proportional to each other while R1(ν) is given by the recursive relation E 3 D a3 = a3 , (96) 2 1 −4 Rn(ν)= fj (n + ν) κ Sj Rn+1(ν) Sj − , − as can be checked by inspecting the coupled-channel form (92) of the potential in Eq. (A106) from Appendix A. Thus in ¯ which can be accurately solved with between 20 and the B6B6∗ system we will be computing the critical values 30 iterations [45] (that is, one takes RN = 0 for large of the matrix enough N, e.g. 20 30, and solves the recursion rela- S SD 3 SE − j = j j . (97) tion backwards). Once we have the matrix Fj , we obtain − 4

νli = νli (κ) by ﬁnding the zeros of

det (Fj (ν,κ))=0 , (93) D. Critical Momenta This equation admits N solutions, one for each value of the angular momentum. For κ = 0 these solutions behave The critical κc for which the convergence criterion fails as are listed in Table VI for the different possible heavy baryon-antibaryon states that contain an S-wave. The 1 previous values have been obtained under the assump- νl (κ =0)= li + , (94) i 2 tion that the effective pion mass can be taken to be zero. The effect of finite pion mass was considered in Ref. [32], with i = 1,...,N. As κ increases νl (κ) moves slowly i where it was found that it increases the range of mo- downwards. Once we reach νl = li at the critical value i menta where the tensor part of OPE is perturbative by κ = κc, we have that νl splits into the complex conjugate i the following factor solutions νli (κ) = li iρli (κ). This is a non-analyticity ± +mπRc which marks the point above which νli (κ) cannot be ex- κc(mπ)= κc e , (98) pressed as a perturbative series. This in turn defines κc, the critical value of κ for which there is a ν (κ) that be- where Rc is the radius below which we do not expect li the OPE potential to be valid. The value of this radius comes non-analytic in κ. Usually the first νlj to split is the one that corresponds to the smallest angular momen- is rather ambiguous. In Ref. [32] the estimation Rc = 0.5 0.8 fm was proposed, yielding tum and also the one that determines the breakdown of − the perturbative series. κc(mπ) 1.5 κc . (99) For computing the critical momenta we need first the ≃ Higher values might be more appropriate indeed, but here matrix elements of the tensor operator in the channel we will stick to this value. under consideration. For the B¯B¯¯, B B¯ and B B¯ 3 3 6 6 6∗ 6∗ To obtain the critical momenta we multiply κ (m ) molecules this is trivial: first we take the tensor force c π by the relation k = κ (m )/ a , where a is the tensor c c π | 3| 3 length scale. If we match the mπr 0 limit of the OPE 3 → potential to the 1/r form we have used to derive κc, we find that 9 We note that Eq. (83) is written in terms of the standard wave 2 function, while Eqs. (88) and (89) use the reduced wave function ¯ µgi instead. a3 = R1R2 τ 2 , (100) | | | | 4πfπ 15

1 where τ = T~1 T~2 and gi = g2, g3 depending on whether the potential is promoted to Q− , see the discussion in we are considering· the SS¯ or the T S¯ potential. The Sect. IIB for details. There are different choices depend- factors R1 and R¯2 and the proper isospin operator to ing on which piece of the interaction is promoted. We ¯ use can be checked in Table IV. For the B6B6∗ case we will consider three scenarios: will use the factors corresponding to the direct channels, (a) promotion of the contact terms, i.e. B B and B∗ B∗, in agreement with the 6 → 6 6 → 6 convention that we have used in Eq. (97) for writing their (b) promotion of central OPE and, tensor matrices. From the previous, we can define the tensor scale as (c) promotion of tensor OPE. κ(m ) κ(m ) 4πf 2 Λ (m )= π = π π , (101) Each scenario represents a different binding mechanism: T π ¯ 2 a3 R1R2τ µgi | | | | (a) short-range, which is useful because it allows a direct comparison with the central scale ΛC that we defined in Eq. (76). If we (b) long-range and particularize for the T S¯ and SS¯ in the charm sector (c) a combination of both, +18 T S¯ 186 8 MeV Λ (m ,Q = c) κ (m ) − , (102) where we notice that (c) is not obvious but a consequence T π ≃ c π R R¯ τ | 1 2 | of the technicalities of power counting, as we will explain. SS¯ 90 250 MeV We present an overview of these scenarios in Table VIII. Λ3 (mπ,Q = c) κc(mπ) − , (103) ≃ R1R¯2 τ But we stress that the discussion here will be theoretical: | | in the absence of experimental data it is not particu- while for the bottom sector we obtain larly useful to consider the subleading orders of the EFT +80 expansion. The exploration in this section provides in- T S¯ 140 40 MeV ΛT (mπ,Q = b) κc(mπ) − , (104) formation about the theoretical uncertainties that are to ≃ R1R¯2 τ | +80 | be expected from a LO calculation in each scenario, SS¯ 100 40 MeV Λ (m ,Q = b) κ (m ) − , (105) T π ≃ c π R R¯ τ (a) (Q/M), | 1 2 | where we have used the lattice QCD values of g2 and (b) (Q/M), g3 [53]. These scales look rather soft at first sight but the 5 2 factors R1, R¯2 and τ will increase the values of ΛT consid- (c) (Q/M) , erably in most cases. A few representative values of Λ T where we will explain in detail how we obtain these un- are compiled in Table VII both for the chiral limit and the certainties and also what is the general form of the first physical pion mass. In general we find that tensor OPE subleading corrections, which can actually be consulted is considerably stronger than central OPE, For the Ξ Ξ¯ Q Q′ in Table VIII. For a more in depth discussion of the ¯ and ΛQΣQ molecules we find that ΛT is markedly softer power counting of heavy meson-antimeson in particular than ΛC, and in the bottom sector the tensor force prob- and two-body systems in general we refer the reader to ably requires a non-perturbative treatment. For the iso- Refs. [32, 44]. ¯ ¯ ¯ scalar ΣQΣQ, ΣQ∗ ΣQ, ΣQ∗ ΣQ∗ molecules the tensor scale ΛT is moderately soft, particularly in the bottom sector. We notice that the same comments are also valid in the A. Counting with Perturbative Pions two-nucleon system, in which ΛT (0) = 66MeV in the chiral limit [45] and Λ(m ) = 99MeV for the physical π The first possibility — scenario (a) — is that the bind- pion mass. ing mechanism for heavy baryon-antibaryon molecules is of a short-range nature. Within the EFT language this amounts to the promotion of the contact-range potential V. POWER COUNTING FOR HEAVY BARYON 0 1 MOLECULES from Q to Q− . Within this power counting the leading 1 order (LO Q− in this case) potential will be composed of contact≡ terms, while the next-to-leading order poten- In this section we discuss the different possible power tial (NLO Q0) will contain the OPE potential plus a counting rules for the heavy baryon-antibaryon states. few additional≡ contact interactions We are interested in the case where there are bound states. This excludes NDA, for which LO ( 1) V = VC− , (107) LO 0 LO 0 NLO (0) V (~q) Q , V Q , (106) V = V + VOPE . (108) C ∼ F ∼ C as this counting leads to purely perturbative heavy We do not have to promote all the possible contact in- baryon-antibaryon interactions. The existence of bound teractions that we obtain from the heavy-light spin de- states requires that at least one of the components of composition: in general a subset of it will be enough. 16

Channel I ΛT (0) ΛT (mπ) Channel I ΛT (0) ΛT (mπ) ¯ +23 +34 ¯ +100 +150 ΞcΞc′ (1−±) 0 246 10 368 14 ΞbΞb′ (1−±) 0 200 60 300 90 ¯ +18− +27− ¯ −+80 +110− ΛcΣc(1−±) 1 196 7 295 11 ΛbΣb(1−±) 1 150 40 230 70 ¯ − − ¯ +160− +230− Ξc′ Ξc′ (1−−) 0 180 490 270 740 Ξb′ Ξb′ (1−−) 0 210 70 320 110 ¯ − − ¯ +60− −+90 ΣcΣc(1−−) 0 70 190 110 290 ΣbΣb(1−−) 0 80 30 120 40 − − − − ¯ + ¯ + +130 +200 Ξc∗Ξc′ (2− ) 0 150 410 230 610 Ξb∗Ξb′ (2− ) 0 180 60 270 90 − − − − ¯ + ¯ + +80 +110 Σc∗Σc(1− ) 0 90 240 130 350 Σb∗Σb(1− ) 0 100 30 150 50 − − − − ¯ ¯ +60 +90 Σc∗Σc(1−−) 0 70 200 110 300 Σb∗Σb(1−−) 0 90 30 130 40 − − − − ¯ + ¯ + +50 +80 Σc∗Σc(2− ) 0 60 160 90 240 Σb∗Σb(2− ) 0 70 20 100 30 − − − − ¯ ¯ +60 +90 Σc∗Σc(2−−) 0 70 190 110 290 Σb∗Σb(2−−) 0 80 30 130 50 − − − − ¯ ¯ +130 +200 Ξc∗Ξc′ (3−−) 0 160 420 230 630 Ξb∗Ξb′ (3−−) 0 190 70 280 100 ¯ + − − ¯ + +50− −+80 Σc∗Σc∗(0− ) 0 60 170 100 260 Σb∗Σb∗(0− ) 0 80 30 110 40 − − − − ¯ ¯ +60 +90 Σc∗Σc∗(1−−) 0 70 180 100 270 Σb∗Σb∗(1−−) 0 80 30 120 40 − − − − ¯ + ¯ + +60 +90 Σc∗Σc∗(2− ) 0 70 190 100 280 Σb∗Σb∗(2− ) 0 80 30 120 40 − − − − ¯ ¯ +50 +80 Σc∗Σc∗(3−−) 0 60 160 90 240 Σb∗Σb∗(3−−) 0 70 20 100 40 − − − −

TABLE VII. The tensor scale ΛT (in units of MeV) for a series of heavy baryon-antibaryon molecules. For momenta above this scale, p> ΛT , the tensor force becomes non-perturbative. The tensor scale depends on the pion mass: ΛT (0) and ΛT (mπ) represent the value in the chiral limit and physical pion mass, respectively. The values of ΛT are shown for the antitriplet-sextet ( ) ¯ ( ) and sextet-sextet molecules. In the latter case we concentrate on the isoscalar ΣQ∗ ΣQ∗ molecules, for which pion exchanges ′ ′ ( / ) ¯( / ) are stronger (for the isovector case the value of ΛT is twice that of the isoscalar case). For the ΞQ ∗ ΞQ ∗ molecules we only show the channel in which the tensor force is strongest. For comparison purposes, in the deuteron channel of the two-nucleon system the tensor scale is ΛT (0) = 66 MeV and ΛT (mπ) = 99 MeV, respectively.

Power Counting Q LO V LO NLO V NLO 0 2 2 2 NDA p, mπ Q C, VOPE Q D (p + p′ ), VTPE 1 0 2 2 (a) p, mπ, √2µB2 Q− C Q D (p + p′ ), VOPE 1 0 (b) p, mπ, ΛC Q− VOPE(C) Q C, VOPE(T) 1 3/2 2 2 (c) p, mπ, ΛT Q− C, VOPE Q D (p + p′ )

TABLE VIII. Possible power countings for the heavy baryon-antibaryon system: NDA (naive dimensional analysis) and the three scenarios (a), (b), (c) that we consider in Sect. V. Q refers to the soft scales in each power counting, which include the momenta p of the heavy baryons and pions, the mass mπ of the pions, the binding momentum √2µB2 of a heavy baryon- antibaryon bound state (with µ the reduced mass and B2 the binding energy) and the central and tensor scales ΛC and ΛT defined in Eqs. (76) and (101) and listed in Tables V and VII for a series of heavy baryon-antibaryon systems. The LO and NLO columns indicate the counting Qν of the leading order and first subleading correction. In the V LO and V NLO columns we write 2 2 what are the contributions to the EFT potential in each case. C and D (p + p′ ) refer to a contact-range potential without derivatives and with two derivatives of the heavy baryon field, respectively. VOPE is the OPE potential, VOPE(C) its central piece and VOPE(T) its tensor piece, Finally VTPE refers to two pion exchange potential (i.e. irreducible diagrams containing two pions), which we have not considered in this work.

There is one important detail with this counting. If we D couplings is consider the S-wave contact-range interactions in EFT, they admit the momentum expansion 1 1 C ,D . (110) ∼ M 2 ∼ M 4 2 2 p′ V p = C + D(p′ + p )+ .... (109) h | C | i But if we promote the coupling C to LO, the coupling D must also be promoted [39–41] where the dots denote couplings involving more deriva- 1 1 tives of the baryon fields. Here we use C and D as a C ,D . (111) generic notation for the couplings of a contact-range po- ∼ MQ ∼ M 2Q2 tential with no derivatives (C) or with two derivatives (D). The naive expectation for the scaling of the C and As a consequence the ordering of the contact-range po- 17 tential will be potentials in general lead to non-trivial consequences in LO EFT [58–63]. The tensor force is not only singular, but p′ V p = C, (112) h | C | i also attractive for the case at hand: for an S-wave heavy NLO 2 2 p′ V p = D(p′ + p ) . (113) baryon-antibaryon state that mixes with a D-wave, there h | C | i That is, the NLO potential will contain a contact-range is always a configuration for which the tensor force is at- 10 interaction with two-derivatives on the baryon fields. As tractive . For attractive singular potentials short-range a consequence if we promote a particular C coupling to physics is enhanced: the non-perturbative treatment of 1 attractive singular potentials requires the inclusion of a Q− , the corresponding derivative coupling with D will be promoted to Q0. We notice that in this work we have contact-range interaction at LO [60, 61]. not explicitly considered a contact-range potential with The application of these ideas for a heavy baryon- derivatives. The take-home message is that in this sce- antibaryon S-wave molecule implies that a non- nario the theoretical uncertainty of the calculations is perturbative tensor force requires a non-perturbative Q/M because the first correction to a LO calculation is contact potential. As a consequence, the LO potential suppressed by one order in the EFT expansion. will be LO 1 V = VOPE + VC− , (119)

B. Counting with Non-Perturbative Central OPE ( 1) 11 with VC− the lowest order contact-range potential . The counting of the contacts will be modified as fol- The second possibility is that the binding mechanism lows [44, 45] depends also on the attraction provided by central OPE. 1 1 We can distinguish two cases: (i) the binding depends on C ,D . (120) central OPE alone, i.e. scenario (b), and (ii) the binding ∼ MQ ∼ M 7/2Q1/2 depends on the interplay of the contact terms and central or equivalently we can write OPE, i.e. scenario (a+b). ( 1) In the first case — scenario (a) — we have a relatively p′ V − p = C, (121) h | C | i simple power counting in which (3/2) 2 2 p′ V p = D(p′ + p ) , (122) LO h | C | i V = VOPE(C) , (114) where the contacts with derivatives get promoted by half where by OPE(C) it is meant the central piece of OPE. an order. Thus the theoretical uncertainty of a LO cal- The NLO potential contains tensor OPE and the contact culation is (Q/M)5/2 (the first subleading correction, a interactions derivative contact interaction, enters Q5/2 orders after NLO (0) V = VOPE(T) + VC , (115) LO), which is considerably better than for the other scenarios. where The previous analysis is a simplification though: ten- (0) p′ V p = C0 . (116) sor forces mix channels with different orbital angular h | C | i Contacts with 2n derivatives on the baryon fields will en- momentum, which might lead to complications in cer- ter at order Q2n. In this scenario the relative uncertainty tain cases (in particular the power counting of contact- of a LO calculation is Q/M because the first correction interactions mixing partial waves). We have not ad- to the EFT potential enters at NLO. dressed these problems here: they depend on the particu- The second case — scenario (a+b) — is identical to the lar system under consideration and the aim of the present power counting of scenario (a) except for the fact that we discussion is to provide an overview of power counting include OPE in the LO: rather than a detailed account. LO LO V = VC + VOPE(C) , (117) V NLO = V NLO + V , (118) VI. PREDICTING HEAVY BARYON C OPE(T) MOLECULES LO where VOPE(T) is the tensor piece of OPE, while VC and NLO VC are the contact-range potentials of Eqs. (112) and In this section we investigate the question of whether (113). The uncertainty of the LO calculation is Q/M. we can predict heavy baryon molecules. The answer to

C. Counting with Non-Perturbative Tensor OPE 10 This can be seen by inspecting the partial wave projection of The third possibility — scenario (c) — arises when the tensor operator S12, which can be consulted in Appendix A. It happens that for this matrices there is always at least one tensor OPE is non-perturbative. This is the most in- positive and one negative eigenvalue. volved of the three power countings considered. Ten- 11 We have simply included the full OPE potential in LO because sor OPE is a singular potential, which means that it di- the addition of central OPE does not further modify the power verges as fast as (or faster than) 1/r2 for r 0. Singular counting induced by tensor OPE. → 18 this question depends on which is the binding mechanism. out making explicit assumptions about the binding mech- If the binding mechanism is of a short-range nature, the anism. prediction of bound states will rely on phenomenology. In this section we will examine the short- and long- Within the EFT framework this is illustrated by the range binding mechanisms for heavy baryon-antibaryon fact that the contact-range couplings are free parame- molecules. The most obvious short-range mechanism is ters. If there is no preexisting experimental information the saturation of the EFT contact-range couplings from about the heavy baryon-antibaryon system, we will have scalar and vector meson exchange, while the most impor- to determine the contact-range couplings by matching tant long-range mechanisms is the OPE potential. Now to a phenomenological model. Conversely if the binding we will explain these binding mechanisms in detail. mechanism is of a long-range nature, the prediction of bound states is possible within EFT. Examples are the ¯ Λc1Σc [29] and DDs∗0 / D∗Ds∗1 [64] systems, which in- A. Short-Range Binding Dynamics teract via a long-range Yukawa potential that is strong enough to bind. This is not the standard situation though First we explore the short-range dynamics, in particu- and more often than not we will need phenomenological lar scalar and vector meson exchange. For taking this ef- input. fect into account we saturate the contact-range couplings At this point it is interesting to notice the relation of the EFT with the exchange of a meson with mass mS between power counting and the predictability of heavy of the order of the hard scale of the EFT (mS M), baryon-antibaryon molecules. In Sect. V we proposed see Ref. [70] for a detailed exposition of this idea.∼ For three power counting scenarios: (a), (b) and (c). Sce- this we expand the exchange potential VS for momenta 2 nario (a) corresponds to a short-range binding mech- ~q mS and match it with the expansion of the EFT anisms, which requires phenomenological input. Sce- contact-range≪ potential nario (b) corresponds to a long-range binding mechanism, 2 which allows for EFT predictions. Finally scenario (c) is VS(~q)= V0 + V2 ~q + ..., (123) 2 a mixture of short- and long-range binding, which in a VC (~q)= C + D ~q + ..., (124) few cases will lead to predictions. The heavy baryon- antibaryon system belongs to scenario (a) or (c) depend- from which we arrive to ing on the particular state and quantum numbers con- C(Λ m ) V = V ( ~q = 0) , (125) sidered. ∼ S ∼ 0 S | | Theoretical studies of hadronic molecules have at- where Λ is the cutoff. Notice that we take Λ mS: tributed the binding mechanism to either short- or long- this is because the saturation hypothesis is only expected∼ range causes. In the pioneering work of Voloshin and to work if the cutoff is of the order of the mass of the Okun [1] it is the exchange of light mesons (π, σ, ρ exchange meson [70]. For a Yukawa-like meson exchange and ω) which generates heavy hadron molecules, i.e. a potential mixture of short- and long-range physics. Early specu- 2 lations [2–5] often predicted binding from the OPE po- gS VS(~q)= 2 2 , (126) tential (long-range physics) alone. It is notable to men- ~q + mS tion that Ericson and Karl [4] indicated that hadronic the saturated contact-range coupling is proportional to molecules should be possible in the charm sector and that Törnqvist [5] predicted the existence of a isoscalar 1++ g2 ¯ C S , (127) D∗D bound state. The experimental discovery of the ∝ m2 X(3872) a decade after [6] suggests that these theoret- S ical speculations were on the right track. At this point where the proportionality constant will depend on the we find it interesting to notice that a molecular X(3872) details of the regularization process. This argument is also arises naturally from short-range physics [65]. Be- independent of the nature of the exchanged meson, it fore the discovery of the Pc(4450) by the LHCb [11], only matters that the mass of this meson is of the order which is suspected (but not confirmed) to be a D¯ ∗Σc of the hard scale. molecule [14, 16–18], there were theoretical predictions Next we calculate the scalar and vector meson ex- of its existence too. The work or Refs. [66, 67] used change contribution to the saturation of the EFT cou- contact-range interactions derived from vector meson ex- pling. We begin by considering scalar meson exchange. change saturation to make quantitative predictions of The sigma meson exchange potential is an I = 1 , J P = 3 − D¯ Σ molecule (among others). 2 2 ∗ c g2 Meanwhile the work of Ref. [68] used the OPE potential σ Vσ(~q)= 2 2 , (128) instead to make qualitative predictions about probable −~q + mσ 1 P 3 − hadronic molecules, including a possible I = 2 , J = 2 with gσ the sigma coupling. We can determine gσ from D¯ ∗Σc molecule. Finally EFT and EFT-inspired works the quark model, in which gσ is simply proportional to explain the properties of shallow molecular states solely the number of u and d quarks in the hadron. Here we take on the basis of short-range interactions [31, 36, 69], with- the sigma-nucleon-nucleon coupling as input, which in 19

the non-linear sigma model [71] is g = √2 M /f with m = 770 MeV, the universality of the ρ coupling σNN N π ∼ ρ 10.2, where MN is the nucleon mass and fπ the pion implies that gρ = mρ/2fπ 2.9. If we match to the decay constant. From this we have g = g /3 3.4 potentials in Eqs. (132) to (135),≃ we ﬁnd σ σNN ∼ for ΞQ, ΞQ′ and ΞQ∗ and gσ = 2gσNN /3 6.8 for ΛQ, ∼ λ = √2 g and λ =2√2 g . (137) ΣQ and ΣQ∗ . The contributions of the scalar meson to T ρ S ρ the saturation of the contact-range couplings are listed The saturation of the EFT contact couplings by the vec- in Table IX. tor mesons is easy to obtain and can be consulted in We continue with the vector meson exchange poten- Table IX. We mention that is is possible to consider the tial, for which the starting point is the heavy baryon - contributions of the diﬀerent vector mesons separately vector meson Lagrangian for the SSV and T T V vertices

(where V represents the vector meson). If we consider CV = Cρ + Cω + Cφ . (138) interactions with no derivatives, which allow for saturation of the lowest order EFT couplings, we can write the This form is interesting because it makes it easy to de- following Lagrangians duce the strength of the heavy baryon-baryon short-range interaction from the heavy baryon-antibaryon one. This = λ ǫ ǫ T¯l v (V µ)i T , (129) LT T V T ikl jkm Q µ j Qm merely involves changing the sign of the contributions = λ S¯ v (V µ)i Sνjk , (130) from the negative G-parity mesons, the ω and the φ, LSSV S Qνik µ j Q yielding CV′ = Cρ Cω Cφ. Though the vector me- where the latin indices indicate the sum over the SU(3) son saturation of the− heavy− baryon-baryon system is not components 12. The vector meson nonet ﬁeld is given by listed here, they can be obtained from Table IX where 0 0 the Cρ, Cω and Cφ contributions are listed. ρ + ω ρ+ K + √2 √2 ∗ Finally we add the contribution to the EFT contact 0 ρ ω0 0 V = ρ− + K∗ , (131) couplings from scalar and vector meson exchange satura-  − √2 √2  ¯ 0 tion, that is  K∗− K∗ φ    C(Λ m ,m ) C + C , (139) where the Lorentz index µ is implicitly understood. The ∼ V σ ∼ S V vector meson exchange contribution to the potential can where CS and CV are the scalar and vector meson con- be worked out along the lines of Appendix A. A few rep- tributions. At this point it is interesting to compare resentative examples are saturation in the heavy baryon-antibaryon system with λ2 (~τ ~τ 3) the heavy meson-antimeson and heavy meson-antibaryon Ξ Ξ¯ V Ξ Ξ¯ = T 1 · 2 − , (132) ¯ ¯ h Q Q| | Q Qi ~q2 + m2 2 cases. The X(3872) and Pc(4450) are D∗D and D∗Σc V molecular candidates for which we can apply the sat- ¯ ¯ λS λT (~τ1 ~τ2 3) uration argument as well, as can be seen in Table IX. ΞQ′ ΞQ V ΞQ′ ΞQ = 2 2 · − , (133) h | | i ~q + mV 4 If we compare the saturated contact-range couplings 2 λS (~τ1 ~τ2 3) of the X(3872) and Pc(4450) with the ones for the Ξ′ Ξ¯′ V Ξ′ Ξ¯′ = · − , (134) h Q Q| | Q Qi ~q2 + m2 8 heavy baryon-antibaryon system, we can identify the V most promising molecular candidates. Heavy baryon- 2 T~1 T~2 1 antibaryon systems for which the short-range interaction ¯ ¯ λS · − ΣQΣQ V ΣQΣQ = , (135) is expected to be more attractive than the X(3872) in- h | | i ~q2 + m2 2 V clude which have been calculated in the SU(3) limit. We have Λ Λ¯ , Ξ Ξ¯ (I = 0) , Ξ′ Ξ¯ (I = 0) , taken mρ = mω = mφ = mV with mV the vector meson Q Q Q Q Q Q ¯ ¯ ¯ mass. Notice that we do not have to write explicitly the ΣQΛQ , ΞQ′ ΞQ′ (I = 0) , ΣQΣQ(I =0, 1) , potential for the sextet spin-3/2 heavy baryons: the vec- (140) tor meson potentials for the ΣQ∗ , ΞQ∗ and ΩQ∗ are identical to the ones for ΣQ, ΞQ′ and ΩQ. The couplings λS and λT to which we have to add the molecules containing the are not arbitrary: they can be determined from the uni- excited sextet baryons, i.e. the molecules we obtain from versality of the ρ coupling constant [72]. If we consider the substitutions ΞQ Ξ∗ and ΣQ Σ∗ . The systems → Q → Q the ρ-meson exchange potential between two isospin 1/2 for which there is more short-range attraction than for baryons the Pc(4450) include g2 Λ Λ¯ , Σ Λ¯ , Σ Σ¯ (I =0, 1) , (141) V (~q)= ρ ~τ ~τ , (136) Q Q Q Q Q Q ρ ~q2 + m2 1 · 2 ρ where we notice that they are a subset of Eq. (140). The obvious conclusion is that the heavy baryon-antibaryon pairs listed in Eq. (141) are the strongest candidates to 12 Notice that we are not considering T SV vertices because they bind. The particular case of ΛcΛ¯ c has been recently stud- involve derivatives and do not saturate the LO contact-range ied in Ref. [73], leading to binding in agreement with our couplings. conclusions. 20

If we consider the heavy baryon-baryon system instead, System Isospin CS CV Cρ Cω Cφ the contribution of the ω and φ mesons is repulsive and in NN 0 9 +6 3 +9 0 general there is less attraction that in the heavy baryon- − − NN 1 9 +10 +1 +9 0 antibaryon case. Yet for the following heavy baryon- − DD¯ 0 1 4 3 1 0 baryon system − − − − DD¯ 1 1 0 +1 1 0 − − Σ Σ (I = 0) , (142) ¯ 1 Q Q ΣcD 2 2 2 4 +2 0 3 − − − ΣcD¯ 2 +4 +2 +2 0 there is more short-range attraction than in the X(3872) 2 − ΛQΛ¯ Q 0 4 4 0 4 0 and the Pc(4450). Further candidates for binding can be − − − ΞQΞ¯ Q 0 1 6 3 1 2 inferred from a comparison with the deuteron, for which − − − − − ΞQΞ¯ Q 1 1 2 +1 1 2 the short-range interaction is repulsive. In Table IX we − − − − see that for the nucleon-nucleon system there is a strong Ξ′ Ξ¯ Q 0 1 6 3 1 2 Q − − − − − short-range repulsion from the exchange of the ω meson Ξ′ Ξ¯ Q 1 1 2 +1 1 2 Q − − − − but also a strong attraction coming from the exchange of ΣcΛ¯ Q 1 4 4 0 4 0 the σ meson. The existence of the deuteron indicates that − − − ΩQΩ¯ Q 0 0 8 0 0 8 attraction wins in this case. This is not surprising if we − − ΞQΞ¯ Q 0 1 6 3 1 2 notice that mσ gρ, which suggests − − − − − ΞQΞ¯ Q 1 1 2 +1 1 2 that σ meson saturation overcomes ω meson saturation − − − − ΣQΣ¯ Q 0 4 12 8 4 0 ( CS > Cω ). Here it is worth noticing that binding in − − − − | | | | ΣQΣ¯ Q 1 4 8 4 4 0 non-relativistic systems depends on the reduced poten- − − − − tial, the product of the potential by twice the reduced ΣQΣ¯ Q 2 4 0 +4 4 0 − − mass of the system. This in turn implies that for the following systems TABLE IX. Contact-range coupling from saturation of vector 2 2 ΛQΛQ , ΣQΛQ , ΣQΣQ(I =0, 1) , (143) and scalar meson exchange in units proportional to gρ/mV 2 and gσqq/mσ, with gρ 2.9 and gσqq 3.4. Scalar and the net eﬀect of the short-range attraction from the σ ≃ ≃ 3 vector meson exchange saturation of the spin- 2 sextet heavy meson will be larger than in the two-nucleon system, i.e. 1 baryons (ΣQ∗ , ΞQ∗ , ΩQ∗ ) is identical to their spin- 2 partners (ΣQ, ΞQ, ΩQ) and are not listed independently. The contri- 2µ C > 2µ CNN , (144) | S| NN| S | butions to vector meson exchange are also listed separately as CV = Cρ + Cω + Cφ. Though not listed, the saturation with µ and µNN the reduced masses of the systems listed for the heavy baryon-baryon system can be obtained from the in Eq. (143) and the two-nucleon system respectively, and heavy baryon-antibaryon case by changing the sign of the ω NN CS and CS their σ-saturated couplings. But we warn and φ contributions. For comparison purposes we also in- that this argument is incomplete: common-wisdom in clude the NN (two-nucleon), D∗D¯ and D¯ ∗Σc systems, which nuclear physics attributes binding in the deuteron to the are related to the deuteron, the X(3872) and the Pc(4450), interplay of short- and long-range physics, in particular respectively. the short-range repulsion from the ω meson, the attraction from the σ meson and the tensor force from OPE. This suggests that, with the exception of the isoscalar meson-antibaryon molecules leads to the prediction of ¯ ¯ ¯ ¯ ΣcΣc system, the other molecular candidates listed in seven DΣc, DΣc∗, D∗Σc and D∗Σc∗ molecules but only Eq. (143) require a more thorough theoretical exploration one experimental candidate, the Pc(4450). This situa- to determine if there is binding. tion also happens in other theoretical approaches that The saturation argument probably provides incom- derive heavy hadron interactions from vector meson sat- plete information about the LO contact-range couplings. uration [65–67]. The probable conclusion is that we are The saturated couplings are independent of the total probably missing something in the resonance saturation light spin of the heavy hadron-antihadron system. This arguments we are using to derive the LO couplings. Be it is compatible with HQSS — it represents a subset of as it may, for the set of molecules in Eq. (141) the short- the possible interactions that respect HQSS — but not range attraction is expected to be remarkably stronger necessarily with experiments. If we review the heavy than in the X(3872) and Pc(4450). meson-antimeson system, to which the X(3872) is suspected to belong, scalar and vector meson exchange saturation predicts exactly the same potential for DD¯, D∗D¯, B. Long-Range Binding Dynamics DD¯ ∗ and D∗D¯ ∗ irrespectively of the spin and C-parity quantum numbers. That is, the saturation argument leads to the prediction of six isoscalar heavy meson- The long range dynamics of the heavy baryon- antimeson molecules. This is to be compared with only antibaryon system is driven by OPE. We assess the rel- one obvious molecular candidate, the X(3872). Anal- ative strength of the OPE potential for each channel in ogously, the application of this argument to the heavy the following way: ﬁrst we modify the OPE potential by 21

Channel I = 0 I = 1 I = 2 Channel I = 0 I = 1 I = 2 + + +0.23 ΣcΣ¯ c(0− ) ΣbΣ¯ b(0− ) 0 − − − − − ¯ ¯ +0.30 +0.21 +0.21 ΣcΣc(1−−) 0.42 0.94 0.22 0.55 0.19 0.48 ΣbΣb(1−−) 0.86 0.31 0.49 0.19 0.43 0.18 − − − − − − ¯ + ¯ + +0.38 +0.23 +0.13 ΣcΣc∗(1− ) 0.39 1.00 0.19 0.53 0.15 0.36 ΣbΣb∗(1− ) 0.88 0.35 0.46 0.20 0.32 0.12 ¯ − − − ¯ +0− .40 +0− .26 +0− .16 ΣcΣc∗(1−−) 0.44 1.11 0.22 0.59 0.18 0.44 ΣbΣb∗(1−−) 0.99 0.39 0.52 0.22 0.39 0.15 − − − − − − ¯ + ¯ + +0.31 +0.21 +0.24 ΣcΣc∗(2− ) 0.46 1.01 0.24 0.59 0.25 0.64 ΣbΣb∗(2− ) 0.91 0.32 0.51 0.21 0.56 0.22 − − − − − − ¯ ¯ +0.27 +0.18 +0.23 ΣcΣc∗(2−−) 0.39 0.85 0.21 0.50 0.20 0.55 ΣbΣb∗(2−−) 0.77 0.27 0.45 0.17 0.48 0.21 − − − − − − ¯ + ¯ + +0.46 +0.32 +0.14 Σc∗Σc∗(0− ) 0.56 1.36 0.27 0.75 0.18 0.41 Σb∗Σb∗(0− ) 1.20 0.46 0.64 0.27 0.36 0.30 − − − − − − ¯ ¯ +0.42 +0.29 +0.15 Σc∗Σc∗(1−−) 0.52 1.24 0.26 0.68 0.18 0.42 Σb∗Σb∗(1−−) 1.09 0.41 0.59 0.25 0.37 0.14 − − − − − − ¯ + ¯ + +0.34 +0.23 +0.18 Σc∗Σc∗(2− ) 0.44 1.02 0.22 0.57 0.20 0.50 Σb∗Σb∗(2− ) 0.91 0.34 0.50 0.20 0.43 0.17 ¯ − − − ¯ +0− .27 +0− .20 +0− .28 Σc∗Σc∗(3−−) 0.45 0.91 0.25 0.55 0.28 0.69 Σb∗Σb∗(3−−) 0.82 0.28 0.48 0.18 0.60 0.25 − − − − − −

TABLE X. Critical radius Rc for which OPE is able to bind certain heavy baryon-antibaryon molecules, where I refers to + +0.23 the isospin of the system. For the ΣQΣ¯ C (0− ), where there is no tensor force, the notation 0 indicate that there is only binding for Rc 0.23 fm if the coupling g2 lies on the high end of the lattice calculations. For comparison purposes, the critical radius for the deuteron≤ and the P +(4450) pentaquark is 1.00 fm and 0.30 0.49 fm, respectively. c −

including a cut-off tainty is again a consequence of g2). In comparison for the heavy meson-antimeson molecular candidates the V (r; r )= V (r) θ(r r ) , (145) OPE c OPE − c radii are 0.30 fm, 0.10fm and 0.26 fm for the X(3872), where rc is the cut-off. Then we calculate the largest rc Zc(3900) and Zb(10610), respectively. The rather small for which OPE alone is able to bind a molecule. We call critical radii of the X, the Zc and the Zb suggest that this radius rc = Rc the critical radius. Notice that we are these hadron molecules depend on the short-range at- in fact assuming that (i) OPE is valid from infinity till traction (instead of OPE) to bind. For the deuteron the critical radius and (ii) there is no short-range physics. the critical radius is significantly larger, indicating that If this critical radius turns out to be large enough we OPE is an important component of the binding mecha- + will consider that the system is likely to bind. By large nism. Last the situation for the Pc(4450) pentaquark enough we mean for instance that the critical radius is seems to be in the middle. From Table X it is appar- larger than the size of the hadrons or the range of other ent that for the heavy baryon-antibaryon system OPE contributions to the hadron-hadron potential that have can provide as much attraction as in the deuteron. If not been taken into account (e.g. two-pion exchange). we combine this observation with what we know about It is important to notice that most heavy baryon- short-range physics according to Table IX, the conclusion ( ) ¯ ( ) antibaryon molecules bind if rc is sufficiently small be- is that there will be a rich ΣQ∗ ΣQ∗ molecular spectrum, cause of the tensor force. Thus the crucial factor is not particularly in the I =0, 1 configurations. whether there is a critical radius for which the molecule binds, but whether the critical radius Rc is reasonable or not. The reasons why the tensor force is able to bind VII. CONCLUSIONS in most cases is because for S-wave molecules it behaves as an attractive singular potential, see the discussion in In this work we have presented a general EFT frame- Sect. V C. This is why it is important to consider whether work for the heavy baryon-antibaryon system. EFTs ex- the distance at which OPE binds is reasonable or not. ploit the existence of a separation of scales to express ( ) ¯ ( ) We list the critical radii for the ΣQ∗ ΣQ∗ molecules the observable quantities of a low energy system as a ( ) ¯ ( ) in Table X. We have chosen the ΣQ∗ ΣQ∗ system be- power series. In the case at hand the size of a hadron cause this is the case in which the OPE potential is molecule is expected to be larger than the hadrons form- expected to be stronger owing to the higher isospin of ing it. As a consequence this type of system is amenable the ΣQ’s. Besides, from scalar and vector meson ex- to an EFT description. Besides, heavy hadron molecules change saturation we expect a very strong short range are constrained by chiral, SU(2)-isospin, SU(3)-flavour attraction. The isoscalar molecules are the ones show- and HQSS symmetries. This degree of symmetry trans- ing more attraction and higher critical radii, reaching in lates into a few interesting regularities in their spectrum. a few cases 1 fm. For the hidden charm molecules the EFT explains the heavy baryon-antibaryon interac- uncertainty is really big as the value of g2 is not exper- tion in terms of contact-range interactions and pion ex- imentally known. To give a sense of scale we mention changes. The relative importance of these two contribu- that for the deuteron the critical radius is 1.00 fm. For tions changes from system to system. In general the LO the Pc(4450) pentaquark-like state as a ΣcD¯ ∗ molecule, EFT description involves four-baryon contact-range in- the critical radius is 0.30 0.49 fm (where the uncer- teractions and pion exchanges (OPE), but this depends − 22 on the molecule. Pion exchanges are expected to be par- relations explains why the Zc’s and Zb’s resonances ap- ¯ ¯ ticularly important in the isoscalar ΣQΣQ, ΣQ∗ ΣQ and pear in pairs [31, 74], while the latter predicts that the ¯ X(3872) should have a 2++ partner, the X(4012) [32, 33]. ΣQ∗ ΣQ∗ molecules (but less important for other configu- In the heavy meson-antimeson system there are a series rations). In contrast OPE vanishes in the ΞQΞ¯Q and of dynamical effects (besides the aforementioned tensor ΛQΛ¯ Q molecules, which can be described in terms of a ¯ ¯ ¯ OPE) that might break these patterns, which include contact-theory at LO. For the ΞQ′ ΞQ′ , ΞQ∗ ΞQ′ and ΞQ∗ ΞQ∗ molecules, particularly in the hidden charm sector, OPE decays into nearby channels [75], coupled channel dy- is probably a NLO effect. We warn that the conclusions namics [76], the existence of nearby quarkonia [77] and about the relevance of the OPE potential are only well- annihilation [78]. Though they have not been studied in established for the bottom sector. In the charm sector the heavy baryon-antibaryon case, these effects could be relevant. the value of the g2 axial coupling that appears in the Σc Σcπ amplitude is not known experimentally, and Finally there is the important question of whether the → a determination either in a future experiment or in the existence of heavy baryon molecules can be predicted. lattice will be welcomed. Particle coupled channels, i.e. EFTs are generic frameworks that usually require pre- transitions in which a heavy baryon changes from the existing experimental input to make predictions. The ground to the excited state (B6 B6∗), are subleading if EFT potential is composed of a long-range and short- → the molecules are not too tightly bound, i.e. for binding range piece. The short-range piece involves unknown momenta γ = M6 EB 350 400MeV. The previous couplings, which have to be determined from external | |≤ − findings regarding pion exchanges and coupled channels information. In the absence of experimental data, there p are analogous to the ones in the heavy meson-antimeson is the possibility of using phenomenological arguments molecules [32]. It is also worth mentioning that right now to estimate the contact-range couplings. If we assume the LO EFT is more than enough for the description of the saturation of these couplings from ρ-, ω-, φ- and heavy hadron-antihadron molecules, where the scarcity σ-meson exchange, the most probable candidates for a of experimental data makes it superfluous to calculate heavy baryon-antibaryon bound state are the isoscalar ¯ ¯ ¯ ¯ subleading orders. ΛcΛc, ΣcΣc, Σc∗Σc and Σc∗Σc∗ molecules, located at 4573, The EFT potential is constrained by HQSS. This is 4906, 4970 and 5035MeV and the isovector ΛcΣ¯ c and ¯ particularly evident for S-wave interactions, such as the ΛcΣc∗ molecules at 4740 and 4805MeV. If we consider LO contact-range potential and central OPE. Symme- the heavy baryon-baryon system instead, saturation in- tries in the S-wave interaction are likely to translate into dicates that the isoscalar, doubly-charmed ΣcΣc, Σ∗Σc ¯ c symmetries in the spectrum. For the T T case the LO and Σc∗Σc∗ molecules are good candidates for binding, fol- EFT potential does not depend on the total spin of the lowed by their isovector counterparts, the isoscalar ΛcΛc system: and isovector ΛcΣc and ΛcΣc∗ systems. For the heavy baryon-antibaryon system we supplement the saturation T T¯ V (0−) T T¯ = T T¯ V (1−) T T¯ , (146) h | s | i h | s | i argument with an estimate of the relative strength of the OPE potential, which we assess by calculating the radius where the subscript s is used to indicate S-wave. That for which OPE would be able to bind the system by itself. is, the T T¯ heavy baryon molecules are expected to come This second argument also points to isoscalar molecules in pairs. For the T S¯/ST¯ molecules we have the following as the most likely to bind. It might be possible to observe two relations: these heavy baryon-antibaryon molecules in experiments such as LHCb and PANDA, which is expected to be par- T S¯ Vs(0−±) T S¯ = T S¯ Vs(2−±) T S¯ , (147) h | | i h | | i ticularly suited for the precision study of hidden charm T B¯ V (1−±) T B¯ = T B¯∗ V (1−∓) T B¯∗ , (148) h 6| s | 6i h 6 | s | 6 i exotic states [79], or alternatively in the lattice. where in the second line we have explicitly indicated whether the sextet heavy baryon is in the ground or excited state. The conclusion is again that T S¯ molecules appear in pairs. For SS¯ molecules the contacts have a far richer structure, with only one obvious symmetry re- ACKNOWLEDGMENTS lation:

SS¯ Vs(2−−) SS¯ = SS¯ Vs(3−−) SS¯ . (149) M.P.V thanks Johann Haidenbauer for discussions and h | | i h | | i the Institute de Physique Nucl´eaire d’Orsay, where part Tensor OPE mixes partial waves and will induce devi- of this work was carried out, for its hospitality. We ations from the previous relations, which will be mod- also thank Chu-Wen Xiao for his comments on the erate if the bound states are shallow. At this point it manuscript. This work is partly supported by the Na- is worth noticing the analogy with the heavy meson- tional Natural Science Foundation of China under Grants antimeson case, where this type of twin structure also No. 11375024, No.11522539, No. 11735003, the Funda- + + happens for (i) the 1 − D∗D¯ and 1 − D∗D¯ ∗ and (ii) the mental Research Funds for the Central Universities and ++ ++ 1 D∗D¯ and 2 D∗D¯ ∗ molecules. The ﬁrst of these and the Thousand Talents Plan for Young Professionals. 23

Appendix A: The One Pion Exchange Potential in The superfield TQ(v) transforms as Heavy Hadron Chiral Perturbation Theory i~ǫ S~ T (v) e− · v T (v) , (A8) Q → Q The OPE potential is the LO piece of the finite-range ~ EFT potential. In this appendix we explain how to com- where Sv is related to SU(2)v, the SU(2) spin group of pute it. The idea is to obtain non-relativistic amplitudes the heavy quark Q moving at velocity v. For the Ψ6Q for processes involving an incoming and outgoing heavy and Ψ6∗Qµ sextet heavy baryons the definition is baryon and a pion, which we write as 1 1+ v/ (B B′π; ~q) , (A1) SQµ(v)= (γµ + vµ) γ5 Ψ6Q A → √3 2 where B(B′) are the initial/final baryon and ~q the mo- 1+ v/ + Ψ∗ , (A9) mentum of the pion if outgoing (if incoming we change 2 6Qµ the momentum to ~q). If written in a suitable normal- − ¯ 1 ¯ 1+ v/ ization these amplitudes can be combined to compute SQµ(v)= Ψ6Q γ5 (γµ + vµ) −√3 2 the OPE potential (or for that matter any one boson 1+ v/ exchange potential) as follows + Ψ¯ ∗ , (A10) 6Qµ 2 1(~q) 2( ~q) B′ B′ V B B = A A − , (A2) h 1 2| | 1 2i ~q2 + µ2 where γµ, γ5 are the Dirac matrices. The SQµ(v) su- π perfield contains a Lorentz index that comes from the where 1 and 2 refer to the pion vertices 1 and 2 and µπ Rarita-Schwinger field Ψ6∗Qµ. It obeys a contrain analo- is the effective pion mass for this particular transition gous to Eq. (A5) (which is not necessarily the physical pion mass because ′ µ mB mB1 = 0 and gives the pion a non-vanishing zero- v S (v) = 0 and vS/ (v)= S (v) . 1 − 6 Qµ Qµ Qµ th component to its 4-momentum). The amplitudes 1 (A11) and may refer to baryons or antibaryons indistinc-A A2 tively. In the following lines we will explain how to do The SQµ(v) superfield transforms as the derivation in detail. i~ǫ S~ S (v) e− · v S (v) . (A12) Qµ → Qµ 1. The Heavy Baryon Field In general we take the velocity parameter to be v = (1,~0). We are interested in heavy baryon-antibaryon Heavy baryons contain a heavy quark and two light molecules, i.e. we need the antibaryon fields. Here it quarks, i.e. they have the structure Q qq . The total is important to notice that | i spin of the light quark pair is SL = 0, 1. If the light ¯ quark spin is SL = 0, we have an antitriplet spin-1/2 TQ(v)= TQ† (v) γ0 , (A13) heavy baryons, which we denote by the Dirac field ¯ SQµ(v)= SQµ† (v) γ0 , (A14) Ψ3¯Q . (A3) are the operators for creating heavy baryons, which are If the light quark spin is SL = 1, we have spin-1/2 and unrelated to the antibaryon fields. Heavy antibaryons spin-3/2 heavy baryons which we denote by the Dirac ¯ require the definition of new TQ¯ (v), TQ¯ (v), SQ¯ (v) and and Rarita-Schwinger fields ¯ SQ¯ (v) fields. We will not need to define them explicitly Ψ6Q , Ψ6∗Qµ . (A4) though. Instead we will use C- and G-parity transforma- tions to deduce the interactions of the heavy antibaryons Notice that the Rarita-Schwinger field contains a Lorenzt with the pions. index: it is the external product of a Minkowsky vector The heavy baryon fields have SU(2)-isospin and SU(3)- and a Dirac spinor. The product contains a spurious spin- flavour structure. If we add SU(3)-flavour indices for the 1 component that can be removed with the condition 2 heavy baryons with Q = b and SL = 0, we have µ γ Ψ6∗Qµ =0 . (A5) Ξb− Within heavy hadron EFT it is customary to use the 0 Ψ3¯b = Ξ (A15) fields T (v) and Sµ(v) instead, which have good trans- − b  formation properties under rotations of the spin of the Λ0  b  heavy quark (where v refers to the velocity of the heavy   while for Q = b, SL = 1 we have quark). For the Ψ3¯Q heavy baryon the definition is ′ 1+ v/ Σ+ 1 Σ0 1 Ξ0 b √2 b √2 b TQ(v)= Ψ3¯Q , (A6) ′ 2 1 0 1 Ψ6¯b =  √ Σb Σb− √ Ξb−  , (A16) 2 ′ 2 1+ v/ 1 ′ 1 T¯ (v)= Ψ¯ ¯ . (A7) 0 Q 3Q  √ Ξb √ Ξb− Ωb−  2  2 2    24

3 where the corresponding expressions for the spin- 2 heavy 3. The Non-Relativistic Limit baryons Ψ6∗c are identical. For the SU(3)-flavour structure of the Q = c charmed baryons we refer to Eqs. (20) The potential is well-defined in the non-relativistic and (21). In the following we will mostly consider the limit, where the heavy baryon fields reduce to SU(2)-isospin structure: when we talk about the TQ we could either be referring to ΛQ (isoscalar) or ΞQ χs′ Ψ3¯Q 2M3¯ , (A25) (isospinor), while when we talk about SQ it could either → 0 ! be ΣQ (isovector), ΞQ′ (isospinor) or ΩQ (isoscalar). p Notice that we are interested in the OPE potential: χs Ψ6Q 2M6 , (A26) the isoscalar ΩQ cannot exchange a single pion and will → 0 ! not be further considered here. The isoscalar ΛQ and the p isospinor Ξ can only exchange pions in vertices involv- 0 ~χs Q Ψ∗ = Ψ∗ , Ψ~ ∗ 2M ∗ , , ( ) 6Qµ 6Q0 6Q → 6 0 0 ing a ΞQ′ and a ΣQ∗ respectively, i.e. there is no ΞQΞQπ ( ! !) ′ n o p ( ) ( ) (A27) or ΛQΛQπ vertex but there are ΞQΞQ∗ π and ΛQΣQ∗ π vertices. where χs′ , χs are standard spinors, while ~χs = (χs1,χs2,χs3) is a vector in which each component is a spinor. The vector ~χs fulfills the condition 2. The Heavy Baryon Chiral Lagrangian at LO ~σ ~χ =0 , (A28) · s The interaction of heavy baryons and pions can be i.e. the non-relativistic version of Eq. (A5), which ensures 3 written as [46, 52] that Ψ6∗Qµ is a genuine spin- 2 field. We have that M3¯ is the SL = 0 heavy baryon mass, while M6, M6∗ are the =0 , (A17) LTTπ SL = 1 heavy baryon masses. In the heavy quark limit = g ǫ T¯i (Aµ)j Skl + the SL = 1 baryon masses are identical: LSTπ 3 ijk Q l Qµ M6 = M6∗ for mQ . (A29) hijk ¯µ µ l → ∞ ǫ SQkl (A )j TQi , (A18) However this does not happen with the M3¯ mass, which ¯µ ν σi i λ jk remains different from M6 and M ∗ in the heavy quark SSπ = ig2 ǫµνσλ S v (A )j (S ) , (A19) 6 L Qik Q limit. Putting all the pieces together, in the non- where the latin indices i,j,k,l indicate either the SU(2)- relativistic limit the SL = 0 heavy field reduces to isospin or the SU(3)-flavour components, g2, g3 are cou- 1 B3¯ pling constants and ǫµνσλ is the 4-dimensional Levi- TQ(v) , (A30) µ Civita symbol. In the equation above A is the pseudo √2M3¯ → 0 ! Goldstone-boson field, 1 T¯ (v) † , (A31) Q B3¯ 0 µ i µ µ √2M3¯ → A = ξ†∂ ξ ξ∂ ξ† , (A20) 2 − while the SL = 1 heavy fields read where ξ is defined as 1 ~ 1 3 ~σB6 + B6∗ i S~Q(v) , (A32) f M ξ = e π , (A21) √2M6 → q 0 ! with the matrix M 1 ¯ 1 S~ (v) † ~ ∗† , (A33) 0 Q 3 ~σB6 + B6 0 π + η π+ K+ √2M6 → √2 √6 0 q π η 0 with B3¯, B6 and B6∗ the non-relativistic heavy baryon M = π− + K , (A22)  − √2 √6  fields. The notation can be further simplified by notic- ¯ 0 2 ¯ ¯  K− K η ing that (i) there is no difference between the TQ / SQ  − 3    and T † / S† fields in the non-relativistic limit, (ii) by which entails that we are taking the normalizationq choice Q Q ignoring the antibaryon components and (iii) by absorb- f 132 MeV. π √ √ If≃ we consider the SU(2) subgroup of SU(3), the Aµ ing the normalization factors 2M3¯ and 2M6 in a field redefinition. In this case we end up with field reduces to the following expansion in the pion field 1 1 TQ = B3¯ , (A34) Aµ = ∂µπ + [π, [π, ∂µπ]] + ..., (A23) 3 T † = B† , (A35) −fπ 6fπ Q 3¯ where π refers to the SU(2) submatrix in the equation above (after removing the contribution from the η), i.e. ~ 1 ~ SQ = ~σB6 + B6∗ , (A36) 0 3 π π+ r √2 π = 0 . (A24) 1 π S~† = B† ~σ + B~ ∗† . (A37) π− ! Q 6 6 − √2 r3 25

The pion field Aµ reduces in the heavy baryon non- for the Λc, Ξc, Ξ′ and Σc baryons respectively. The rela- c ′ relativistic limit to tions for the excited Ξ ∗ and Σ∗ sextet baryons are iden- ′c c 3 tical to those of the Ξ and Σ baryons. Alternatively if ~ 1 ~ π c c A = ~π + ( 3 ) , (A38) we consider isospin vectors we can write −fπ ∇ O fπ where we ignore the zero-th component of A because + µ Λc = 00 I , (A47) it couples to the zero-th component of SQµ(v), which | i + 0 1 1 1 1 vanishes in the non-relativistic limit. From this we can Ξc , Ξc = , , , , (A48) |2 2iI |2 −2iI rewrite the Lagrangian as ′ ′ + 0 1 1 1 1 i ~ j ~kl Ξc , Ξc = , , , , (A49) STπ = g3 ǫijkT † (A) SQ + |2 2iI |2 −2iI L − Q l · ++n + 0o ijkh ~ ~ l Σc , Σc , Σc = 1, 1 I , 1, 0 I , 1, 1 I . ǫ SQkl† (A)j TQi , (A39) {| i | i | − i } · (A50) i = ig Tr S~† A~ S~ , (A40) LSSπ − 2 Q · × Q h i The isospin factors can be extracted by first expanding where the trace is over isospin space. Alternatively we the isospin / flavour indices in the particle basis and can expand the Lagrangian in terms of the fields B3¯, B6 later reinterpreting the result in terms of matrices in the and B6∗: isospin space. We begin with the ST π Lagrangian, for

g3 which † ~ STπ = B3¯ ~σ πB6 L √3fπ · ∇ a a Λ†π Σc = t , (A51) g3 ~ c + B† ~σ πB6′ a √3f 6 · ∇ a τ π Ξc†π Ξc′ = , (A52) g3 2 + B† ~ π B~ ∗ f 3¯ ∇ · 6 π where πa is the pion ﬁeld in the Cartesian basis, τ a are g3 ~ ~ a + B6∗† πB3¯ (A41) the Pauli matrices and t are given by fπ · ∇

1 g2 ~ √2 SSπ = i B6† ~σ ~π σ B6 1 L 3fπ · ∇ × t =  0  , (A53) 1 g2 ~ ~ ~ √ + i B6∗† ~π B6∗ − 2  fπ · ∇ ×  i  g2 √2 + i B† ~σ ~ ~π B~ 6 6∗ t2 = , (A54) √3 fπ · ∇ ×  0  i g2 ~ ~ √2 + i B6∗† ~π ~σ B6 , (A42)   √3 fπ · ∇ ×   0 where we have removed the isospin / ﬂavour indices to 3 t = 1 . (A55) make the expressions shorter. 0     4. The Spin and Isospin Factors In the SSπ case we have

′ a a τ Now we calculate the matrix elements of the diﬀerent Ξc†π Ξc′ = , (A56) 2 √2 vertices, which depend on a series of spin and isospin fac- a a T tors. In the charm sector (Q = c) the relations between Σc†π Σc = , (A57) the isospin and particle basis are √2

+ B3¯(Λc)=Λc , (A43) where Ta are the J = 1 angular momentum matrices ′ + in isospin space. The isospin factors for the Ξc∗ and Σc∗ Ξc B¯(Ξc)= ′ , (A44) baryons are identical to those of the Ξc and Σc baryons. 3 Ξ0 c ! Next we factor out the spin in terms of angular mo- ′ + mentum matrices or equivalent expressions. For the ST π Ξc ′ B6(Ξc′ )= 0 , (A45) vertices the factors are Ξc ! 1 Σ++ Σ+ B† ~σ ~qB = ~σ ~q, (A58) c √2 c 3¯ · 6 · B6(Σc)= 1 + 0 , (A46) Σc Σc ! B† ~q B~ ∗ = S~ ~q, (A59) √2 3¯ · 6 · 26 while for the SSπ vertices we have Finally for the transitions with the Σc and Σc∗ we have the following amplitudes B6† ~σ (~q ~σ) B6 = i 2 ~σ ~q, (A60) · × − · 2g T a ~ ~ 2 ~ (Σ Σ πa)= 2 ~σ ~q, (A79) B∗† ~q B6∗ = i Σ ~q, (A61) c c 6 · × − 3 · A → 3fπ √2 · a 2g2 T B† ~σ ~q B~ ∗ = i S~ ~q, (A62) a ~ 6 6 (Σc∗ Σc∗π )= Σ ~q, (A80) · × − · A → 3fπ √2 · B~ ∗† (~q ~σ) B = i S~† ~q, (A63) a 6 6 a g2 T ~ · × − · (Σc∗ Σcπ )= S ~q, (A81) where ~σ are the Pauli matrices, Σ~ are the J = 3 angular A → √3fπ √2 · 2 a ~ a g2 T momentum matrices and S are 2 4 matrices that con- (Σc Σ∗π )= S~† ~q. (A82) 1 3 × ~ A → c √3f √2 · nect the spin- 2 and spin- 2 baryons. These S matrices π read 1 0 1 0 √2 √6 5. G-parity and Heavy Antibaryons S1 = − , (A64) 0 1 0 1 √6 √2 ! − The amplitudes in Eqs. (A69-A82) are for the heavy i 0 i 0 √2 √6 baryons. Here we deduce the amplitudes for the heavy S2 = , (A65) 0 i 0 i antibaryons by working in the isospin basis and applying √6 √2 ! a G-parity transformation, which is a combination of a C- 2 0 3 0 0 parity transformation and a rotation in isospin space [47] S3 = − , (A66) 0q 0 2 0 G = CeiπI2 , (A83) − 3  q  which are normalized as follows with I2 the second Cartesian component of the isospin matrix. Now we will determine how G operates on the 2 δij iǫijkσk SiS† = − . (A67) different fields we consider here. For instance, pions have j 3 well-defined G-parity Now we define the non-relativistic amplitudes as G π = π . (A84) (B Bπa)= i Bπa B , (A68) | i −| i A → − h |L| i We can write it in terms of the components of the pion with B = B3¯, B6, B6∗. For the transitions involving Λc field for completeness we have + + a g3 a π π (Λc Σcπ )= t ~σ ~q, (A69) A → √ · | 0 i | 0 i 3fπ G  π  =  π  . (A85) g3 | i − | i a a ~ π− π− (Λc Σc∗π )= t S† ~q, (A70) | i | i A → fπ ·     a g3 a ~ If we consider baryons instead, G will transform a baryon (Σc∗ Λcπ )= t S ~q. (A71) A → fπ · into an antibaryon in the same isospin state. If we con- 1 sider nucleons or other isospin 2 baryons, the G-parity For the transitions involving Ξc we have transformation works as follows a a g3 τ (Ξc′ Ξcπ )= ~σ ~q, (A72) p n¯ A → √3fπ 2 · G | i = | i , (A86) a n ! p¯ ! a g3 τ ~ | i −| i (Ξc′ Ξc∗π )= S† ~q, (A73) A → fπ 2 · and now we can identify antiparticle states with a a g3 τ ~ isospinors as follows (Ξc∗ Ξc′ π )= S ~q. (A74) A → fπ 2 · 1 1 n¯ = + , (A87) For the ones with the Ξc′ and Ξc∗, the amplitudes read | i |2 2iI a 1 1 a 2g2 τ p¯ = , (A88) (Ξc′ Ξc′ π )= ~σ ~q, (A75) | i −|2 − 2iI A → 3fπ 2√2 · ′ ′ a where we use the subscript I to indicate that we a 2g2 τ ~ (Ξc∗ Ξc∗π )= Σ ~q, (A76) are indeed referring to isospinors. From the nu- A → 3fπ 2√2 · cleon/antinucleon example we can appreciate that the ′ a a g2 τ ~ (Ξc∗ Ξc′ π )= S ~q, (A77) idea of a G-parity transformation is to have a good map- A → √3fπ 2√2 · ping between the isospin and the particle/antiparticle ba- ′ a a g2 τ ~ sis. The crucial point in the G-parity transformation for (Ξc Ξc∗π )= S† ~q, (A78) A → √3fπ 2√2 · the baryons is the relative minus sign between the isospin 27 vectors of the antineutron and antiproton, which in turn know for the baryons allows for the use of the same SU(2) Clebsch-Gordan co- (B¯ B¯ πa)= (B B πa) , (A94) efficients in the baryon and antibaryon cases. A → −A → ¯ ¯ a a For the heavy baryons the idea is the same as for the (B B6∗ π )=+ (B B6∗ π ) , (A95) A → a A → a nucleons. But there is a subtlety: we are considering (B¯∗ B¯ π )=+ (B∗ B π ) , (A96) 1 3 6 6 different C-parity conventions for the spin- and spin- A → a A → a 2 2 (B¯∗ B¯∗ π )= (B∗ B∗ π ) , (A97) fields A 6 → 6 −A 6 → 6 ¯ where B = B3¯,B6. The signs simply reflect the sign for C B = B , (A89) the G-parity transformation of the pion, plus the extra | i | ¯i C B6∗ = B6∗ , (A90) sign involved in the B¯/B B∗ and B∗ B¯/B | i −| i 3 6 → 6 6 → 3 6 1 transitions. For a detailed example we consider where B = B3¯,B6. For the antitriplet and sextet spin- 2 a a heavy cascades the transformation works exactly as in (B¯6 B¯6 π )= i B¯6π B¯6 nucleons A → h |L|a i = i G(B6π ) GB6 ′ ′ − h |L| i +( ) 0( ) a Ξc Ξ¯c = i B6π G† G B6 ′ ′ − h | L | i G | 0( ) i = | ( )i , (A91) a Ξc ! Ξc− ! = i B6π B6 | i −| i − h |L| ia = (B6 B6 π ) , (A98) while for the sextet spin-3/2 heavy cascades we have −A → 2 where we have used that G B6 = 1 ( 1 for ΞQ′ , +1 + ¯0 2 | i ± − Ξc ∗ Ξc∗ for ΣQ), G π = +1 and G† G = . G | 0 i = | i . (A92) | i L L Ξ ∗ − Ξ−∗ 6. The OPE Potential | c i ! −| c i! ++ + 0 For the isotriplet heavy baryons Σc , Σc , Σc we have With the amplitudes of Eqs. (A69-A82) we can derive { } instead the potential by using Eq. (A2). For simplicity we will ++ ¯ 0 consider the heavy quark limit, in which the B6 and B6∗ Σc Σc heavy baryons are degenerate. For the T S¯ =ΛcΣ¯ c, ΛcΣ¯ ∗ | + i | i c G Σ = Σ− , (A93) ¯ ¯ ¯  c   c  and T S = ΞcΞc′ , ΞcΞc∗ potentials we write the potential | 0 i −| i Σ Σ−− in the bases | ci   | c i     plus the transformation for the excited isotriplet heavy ¯ = Λ Σ¯ , Σ Λ¯ , Λ Σ¯ ∗, Σ∗Λ¯ , (A99) BΛcΣc { c c c c c c c c} baryons, which will carry an extra minus sign. ¯ ¯ ¯ ¯ Ξ′ Ξ¯ = ΞcΞ′ , Ξ′ Ξc, ΞcΞ∗, Ξ∗Ξc . (A100) With the G-parity transformation we can deduce the B c c { c c c c } amplitudes for the antibaryons from the ones we already in which the potential reads as

0 1 ~σ ~q~σ ~q 0 1 S~ ~q~σ ~q 3 1 · 2 · √3 1 · 2 · 2 1 1 g 1 ~σ1 ~q~σ2 ~q 0 ~σ1 ~q S~2 ~q 0 ST¯ 3  3 · · √3 · ·  VOPE(~q)= 2 τ 2 2 1 f ~q + µ 0 ~σ ~q S~† ~q 0 S~† ~q S~† ~q π π  √3 1 · 2 · 1 · 2 ·   1 ~ ~ ~   S† ~q~σ2 ~q 0 S† ~q S2 ~q 0   √3 1 · · 1 · ·  1   + , (A101) O m Q

where the isospin factor τ is The eﬀective pion mass in the heavy quark limit is given 2 2 2 2 2 2 by µ = m (mΣ mΛ ) or µ = m (mΞ′ mΞ ) π π − c − c π π − c − c respectively. For the SS¯ potential we use the bases

′ ¯′ = Ξ′ Ξ¯′ , Ξ′ Ξ¯∗, Ξ∗Ξ¯′ , Ξ∗Ξ¯∗ , (A104) τ(Λ Σ¯ )=1 , (A102) BΞcΞc { c c c c c c c c } c c ¯ ¯ ¯ ¯ Σ Σ¯ = ΣcΣc, ΣcΣc∗, Σc∗Σc, Σc∗Σc∗ , (A105) ~τ1 ~τ2 B c c { } τ(Ξ Ξ¯′ )= · . (A103) in which the potential reads c c 4 28

2 + ~σ1 ~q~σ2 ~q λ ~σ1 ~q S~2 ~q +λ S~1 ~q~σ2 ~q λ S~1 ~q S~2 ~q 2 · · − · · 2 · · − · · ¯ 2g 1 λ ~σ ~q S~ ~q + ~σ ~q Σ~ ~q λ S~ ~q S~† ~q +λ S~ ~q Σ~ ~q V SS (~q)= 2 τ  1 2 1 2 1 2 1 2  OPE 2 2 2 − ~ · · 2 ~ · ~ · − ~ · · ~ · ~ · 9fπ ~q + mπ +λ S1† ~q~σ2 ~q λ S1† ~q S2 ~q + Σ1 ~q~σ2 ~q λ Σ1 ~q S2 ~q  2 · · − · · · · − · ·   λ S~† ~q S~† ~q +λ S~† ~q Σ~ ~q λ Σ~ ~q S~† ~q + Σ~ ~q Σ~ ~q  − 1 · 2 · 1 · 2 · − 1 · 2 · 1 · 2 ·  1   + , (A106) O m Q

√3 where λ = 2 and where the isospin factor is where gi = g2 or g3 depending on the case. From this we obtain the expressions we already wrote in Eqs. (61), ~τ1 ~τ2 (62), (63), (64) and (65). Here we will write the coordi- τ (Ξ′ Ξ¯′ )= · , (A107) c c 4 nate space potential in coupled channels, in which case τ (Σ Σ¯ )= T~ T~ . (A108) we obtain c c 1 · 2

7. Coordinate Space

The general form of the ST¯ and SS¯ potential in momentum space can be written as

2 T S¯ ¯ g3 ~ ~ ~a1 ~q~a2 ~q VOBE = R1 R2 2 I1 I2 2· 2· , (A109) − 2fπ · q + µπ 2 SS¯ ¯ g2 ~ ~ ~a1 ~q~a2 ~q VOPE = R1 R2 2 I1 I2 2· 2· , (A110) − 2fπ · q + µπ where R1 and R¯2 are numerical factors, µπ is the effective pion mass at the vertices (as explained in the previous 2 (0) g3 ST¯ 3 section), I1, I2 the appropriate isospin matrices and ~a1, V (~r)= τ C δ (~r) ST¯ 3 f 2 12 ~a1 are the spin matrices acting on vertex 1 and 2. The π specific factors can be worked out easily from Eqs. (A101) CST¯ SST¯ τ 12 WC (r)+ 12 (ˆr) WT (r) , and (A106) to obtain the results of Table IV. The coordi- − nate space potential is obtained by Fourier transforming h i (A112) 2 the momentum space potential (0) 2g3 CSS¯ 3 VSS¯ (~r)= τ 2 12 δ (~r) 3 27 fπ (0) d q (0) i~q ~r V (~r)= V (~q) e− · 2 ¯ ¯ (2π)3 τ CSS W (r)+ SSS(ˆr) W (r) , Z − 9 12 C 12 T 2 µπr ¯ gi ~ ~ e− h i(A113) = R1 R2 2 I1 I2 (~a1 ) (~a2 ) , 2fπ · · ∇ · ∇ 4πr where WC and WT are defined in Eqs. (64) and (65), and (A111) with the central and tensor matrices given by

0 1 ~σ ~σ 0 1 S~ ~σ 3 1 · 2 √3 1 · 2 1 ~σ ~σ 0 1 ~σ S~ 0 CST¯  3 1 · 2 √3 1 · 2  12 = 1 , (A114) 0 ~σ S~† 0 S~† S~†  √3 1 · 2 1 · 2   1 ~ ~ ~   S† ~σ2 0 S† S2 0   √3 1 · 1 ·   2  + ~σ1 ~σ2 λ ~σ1 S~2 +λ S~1 ~σ2 λ S~1 S~2 · − · 2 · − · ¯ λ ~σ S~ + ~σ Σ~ λ S~ S~† +λ S~ Σ~ CSS =  1 2 1 2 1 2 1 2  , (A115) 12 − ~ · 2 ~ · ~ − ~ · ~ · ~ +λ S1† ~σ2 λ S1† S2 + Σ1 ~σ2 λ Σ1 S2  2 · − · · − ·   λ S~† S~† +λ S~† Σ~ λ Σ~ S~† + Σ~ Σ~  − 1 · 2 1 · 2 − 1 · 2 1 · 2    29

0 1 S (~σ , ~σ , rˆ) 0 1 S (S~ , ~σ , rˆ) 3 12 1 2 √3 12 1 2 1 S (~σ , ~σ , rˆ) 0 1 S (~σ , S~ , rˆ) 0 SST¯  3 12 1 2 √3 12 1 2  12 (ˆr)= 1 , (A116) 0 S (~σ , S~†, rˆ) 0 S (S~†, S~†, rˆ)  √3 12 1 2 12 1 2   1 ~ ~ ~   S12(S†, ~σ2, rˆ) 0 S12(S†, S2, rˆ) 0   √3 1 1   2  + S12(~σ1, ~σ2, rˆ) λS12(~σ1, S~2, rˆ) +λS12(S~1, ~σ2, rˆ) λ S12(S~1, S~2, rˆ) − 2 − ¯ λS (~σ , S~ , rˆ) + S (~σ , Σ~ , rˆ) λ S (S~ , S~†, rˆ) +λS (S~ , Σ~ , rˆ) SSS(ˆr)=  12 1 2 12 1 2 12 1 2 12 1 2  . (A117) 12 − ~ 2 ~ ~ − ~ ~ ~ +λS12(S1†, ~σ2, rˆ) λ S12(S1†, S2, rˆ) + S12(Σ1, ~σ2, rˆ) λS12(Σ1, S2, rˆ)  2 − −   λ S (S~†, S~†, rˆ) +λS (S~†, Σ~ , rˆ) λS (Σ~ , S~†, rˆ) + S (Σ~ , Σ~ , rˆ)  − 12 1 2 12 1 2 − 12 1 2 12 1 2   

8. The Partial Wave Projection The tensor force is more involved. Owing to conservation of parity l l′ must be an even number. The spin transi- | − | Heavy baryon-antibaryon bound states have well de- tions are a more complicated because in the general case ¯ fined J P quantum numbers. Hence we can simplify the s s′ can be even or odd. The exception are the B3¯B3¯, | −¯ | ¯ OPE potential by projecting it into partial waves with B6B6 and B6∗B6∗ systems: these systems usually have well defined C- or G-parity, which implies that s s′ is well-defined parity and angular momentum. For this we | − | define the states even. Even when C-/G-parity is not well defined, like in ¯ the Ξc∗Σc∗ system, the tensor operator involves identical ¯ spin-matrices and is symmetrical under the exchange of ST (jm) = Ylml (ˆr) sms lmlsms jm , | i | i h | i particles 1 and 2: lmXlsms (A118) B3¯B¯3¯ B6B¯6 S12 = S12 =3 ~σ1 rˆ ~σ2 rˆ ~σ1 ~σ2 , (A123) ¯ ∗ ¯∗ · · − · SS(jm) = Ylml (ˆr) sms lmlsms jm , B6 B6 ~ ~ ~ ~ | i | i h | i S12 =3 Σ1 rˆΣ2 rˆ Σ1 Σ2 . (A124) lmXlsms · · − · (A119) As a consequence the matrix elements for the tensor oper- ¯ ator vanishes for odd s s′ . But if we consider the B3¯B6∗ ¯ | − | ¯ ¯ where j, m is the total angular momentum and its third and B6B6∗ systems (for example the ΞcΞc∗ and Σc′ Σc∗ sys- component for the heavy baryon-antibaryon pair, while tems), it is perfectly possible to have a mix of even and l,ml and s,ms refer to the angular momentum and spin odd spin. The odd s s′ transitions have a particularity | − | of the pair. The product lmlsms jm is the Clebsch- that it is worth mentioning: the matrix element for the h | i Gordan coefficient for that particular combination of to- tensor operator for odd s s′ changes sign as follows tal, orbital and spin angular momentum (notice that for | − | ¯ Sj ¯ angular momentum the Clebsch-Gordan are independent BB6∗ s(s 1)ll′ BB6∗ = h | ± | i on whether we have particles or antiparticles). The spin ¯ Sj ¯ B6∗B s(s 1)ll′ B6∗B , (A125) wave function can be further decomposed as −h | ± | i ¯ Sj ¯ BB6∗ s(s 1)ll′ B6∗B = h | ± | i sms = s1m1 s2m2 s1m1s2m2 sms , j | i | i| i h | i B∗B¯ S ′ BB¯∗ , (A126) m1m2 6 s(s 1)ll 6 X −h | ± | i (A120) with B = B3¯ or B6. In fact the previous two equations indicate that it is actually a good idea to take explicitly with s and s the spin of the heavy baryon 1 and 2 1 2 into account the particle coupled channel structure. Now (either 1 or 3 ). 2 2 if we consider the bases In this basis we can compute the partial wave projec- C S ¯ ¯ tion of 12 and 12 as ′ = B3¯B6∗,B6∗B3¯ , (A127) B { ¯ ¯ } = B6B6∗,B6∗B6 , (A128) (s′l′)j′m′ O (sl)jm = B { } h | 12| i 2 then for ′ we can write the central and tensor operators d rˆ (s′l′)j′m′ O (ˆr) (sl)jm = B h | 12 | i as Z j δjj′ δmm′ O ′ ′ , (A121) ~ ~ ss ll 0 S1† S2 C12′ = · , (A129) S~ S~† 0 where the total angular momentum and its third compo- 1 · 2 ! nent are conserved. The central force C12 conserves in S (S~†, S~ , rˆ) S (ˆr)= 12 1 2 . addition the orbital angular momentum and the spin 12′ ~ ~ S12(S1, S2†, rˆ) 0 ! j j C ′ ′ C ss′ll′ = δss δll ls . (A122) (A130) 30 while for they read 15 0 0 0 B − 4 ∗ ¯∗ 3 B6 B6 0 0 0 ~ ~ ~ C (2−)=  − 4  , ~σ1 Σ2 S1† S2 12 3 C12 = · · , (A131) 0 0 4 0 S~ S~† Σ~ ~σ  −  1 2 1 2!  0 0 0 3  · ·  − 4  S (~σ , Σ~ , rˆ) S (S~†, S~ , rˆ)   (A145) S (ˆr)= 12 1 2 12 1 2 . 12 ~ ~ ~ 11 0 0000 S12(S1, S2†, rˆ) S12(Σ1, ~σ2, rˆ)! − 4 · 0 11 0 0 0 0 (A132)  − 4  ∗ ¯∗ 9 CB6 B6 0 0+ 4 0 0 0 12 (3−)=  9  , This increases the complexity of the partial wave projec-  0 0 0+ 0 0   4  tion. However in most cases we can define states with  0 0 00+ 9 0  good C- or G-parity, which effectively amounts to re-  4   0 0 000+ 9  ducing the previous coupled channel systems to a single  4    channel problem. (A146) The calculation of the matrix elements is straightforward in most cases. We begin with the BB¯ system (with B = B3¯ or B6), for which we have the partial waves 1 BB¯(0−) = S0 , (A133) | i {3 }3 BB¯(1−) = S , D , (A134) | i { 1 1} which lead to the matrix elements

∗ ∗ BB¯ B B¯ 0 3 C (0−)= 3 , (A135) S 6 6 12 12 (0−)= − , (A147) − 3 3! BB¯ 1 0 − − C (1 )= , (A136) √ 12 − 0 17 3 7 0 0 1! 5 √2 − 5 17 17 3 2 9 6 ¯ ∗ ¯∗   SBB B6 B6 5 √2 10 5 7 5 7 12 (0−)=0 , (A137) S (1−)= − − , 12  3√7 3 2 q108 9√q3   5 5 7 35 35  SBB¯ 0 2 √2 − −  12 (1−)= . (A138)  9q 6 9√3 45  2 √2 2  0 5 7 35 14  − !  − −   q  ¯ (A148) In terms of complexity the next case is the B6∗B6∗ system, for which the partial waves are ¯ 1 5 B6∗B6∗(0−) = S0, D0 , (A139) | i {3 3 }7 7 B∗B¯∗(1−) = S , D , D , G , (A140) | 6 6 i { 1 1 1 1} ¯ 1 5 5 5 B6∗B6∗(2−) = D2, S2, D2, G2 , (A141) | i {3 3 7 7 }7 7 B∗B¯∗(3−) = D , G , S , D , G , I , | 6 6 i { 3 3 3 3 3 3} (A142) 3 2 2 0 3 7 9 35 which translate into the following matrices − √5 − 3 7 ∗ ¯∗  0 3 q q0  B6 B6 √5 10 ∗ ∗ 15 S (2−)= − , B B¯ 0 12 C 6 6 4  3 2 3 7 q9 18  12 (0−)= − 3 , (A143)  7 10 14 7 √5  0 !   − 4  q 2 q 18 15   9 35 0 7  11 0 0 0 − 7 √5 −  − 4  q  ∗ ¯∗ 11 (A149) B6 B6 0 0 0 C (1−)=  − 4  , 12 0 0+ 9 0  4   0 0 0+ 9   4    (A144) 31

17 51 √3 3√3 36 3√66 35 35 5 35 35 0 −√ − √ −  51 3 17 0 3 3 9 2 3 3  35 − 14 − 35 7 11 − 11 3√3 9√3 ∗ ¯∗ 0 0 q0q 0 B6 B6  5 5  S (3−)=  − √ √ √  , (A150) 12  36 3 3 9 3 99 9 66 0   35 − 35 5 70 35   3√66 9 2 9√66 27 9 3   35 7 11 0 35 77 11 2  − −   q 3 9 3 q63   0 3 11 0 0 11 2 22   − −   q q 

¯ L+S We will continue with the B3¯B6 system, where we have where C = η ( 1) , we ﬁnd the combinations the partial waves − ¯ + B3¯B6∗(1− ) = | 3 i3 5 ¯ 1 S1( ), D1( ), D1(+) , (A161) B3¯B6(0−) = S0 , (A151) { − − } | i { } ¯ + ¯ 3 3 B3¯B6∗(2− ) = B3¯B6(1−) = S1, D1 . (A152) | i | i { } 3D ( ), 5S (+), 5D (+), 5G (+) , (A162) { 2 − 2 2 2 }

For these systems OPE is non-diagonal: the B3¯B3¯π ver- ¯ ¯ B¯B¯∗(1−−) = tex is zero and OPE involves the B3¯B6 B6B3¯ transi- 3 6 → | 3 i3 5 tion. In fact there is only one case, the ΞcΞc′ system, in S1(+), D1(+), D1( ) , (A163) which OPE does not cancel. For this system either C- or {¯ − } B3¯B∗(2−−) = G-parity is well-defined, for which we define the standard | 6 i 3D (+), 5S ( ), 5D ( ), 5G ( ) , (A164) states { 2 2 − 2 − 2 − } where the number in parentheses is the value of η = 1. ± ¯ 1 ¯ ¯ That is, there is the complication that each partial wave B3¯B6(η) = B3¯B6 + η B6B3¯ , | i √2 | i | i in a particular channel can have a different η. Concrete (A153) calculations yield the following matrices 1 0 0 L+S 3 with C = η ( 1) . With this definition in mind, the E − 1 C (1−±)= 0 0 , (A165) projection of− the central and tensor operators read 12 ±  − 3  0 01    1  ¯ 0 0 0 CB3¯B6 3 12 (0−)= 3 η , (A154) − − CE  0 100 ¯ 12(2−±)= , (A166) CB3¯B6 1 0 ± 0 010 12 (1−)= η , (A155)   0 1!  0 001   ¯   SB3¯B6 12 (0−)=0 , (A156) 0 5 1 3√2 √2 B¯B¯ 0 2 √2 S 3 6 E 5 5 1 12 (1−)= η . (A157) S (1−±)=   , (A167) 2 √2 2 ! 12 ± 3√2 − 6 2 − 1 1 1  √ 2 2   2   5 3 1 3 3 The next case is the B¯B¯ system. As happened with 2 3 6∗ 6 − 10 2 7 35 the B¯B¯ case involves only non-diagonal OPE transi- 3 6  3 q0 q 7 q0  ¯ ¯ E 10 10 tions, i.e. B3¯B6∗ B3¯B6. We have the partial waves S (2−±)= − − . → 12 ±  1q 3 7 q3 6   2 7 10 14   − − − 7√5  ¯ 3 3 5 3 6 5 B3¯B6∗(1−) = S1, D1, D1 , (A158)  2 q q0   35 7√5 7  | i {3 5 5 }5  −  B¯B¯∗(2−) = D , S , D , G . (A159) | 3 6 i { 1 2 2 2}  q (A168)

If we are considering states with well-deﬁned C-parity Here we have departed from the previous notation of indicating the channel with a superindex. Instead, we use the superindex E to indicate that it is exchange or non- 1 B¯B¯∗(η) = B¯B¯∗ + η B∗B¯¯ , diagonal OPE. | 3 6 i √ | 3 6 i | 6 3i ¯ 2 Last, the most complex case are the B6B6∗ system. (A160) The partial wave and C-parity structure is identical to 32

¯ ¯ the B3¯B6∗ case, but now there is also a diagonal or di- Now there is a B6B6∗ system for which good G-parity rect piece of OPE (besides the non-diagonal piece). The states cannot be defined, which is vertex factors R1 and R¯2 are different for the direct and exchange pieces, which has to be taken into account when ′′ = Ξ′ Σ¯ ∗, Ξ∗Σ¯ . (A178) writing down the potential B { c c c} V = V D + V E OPE OPE OPE This system is more involved than usual because we have D ¯D ~ ~ CD SD ¯ = R1 R2 I1 I2 12WC + 12WT to consider the two particle channels α = ΞcΣc∗ and β = · ¯ E ¯E ~ ~ CE SE Ξc∗Σc separately. We can define the central and tensor + R1 R2 I1 I2 12WC + 12WT . · matrices in the particle basis as (A169)

If we ignore the fact that the direct and exchange pieces Cαα Cαβ Sαα Sαβ 12 12 12 12 can have a diﬀerent eﬀective pion mass µπ, we can merge C = and S = . 12 Cβα Cββ 12 Sβα Sββ together the direct and exchange central and tensor ma- 12 12 ! 12 12 ! trices. For this we notice the relation (A179) E ¯E 3 D ¯D R1 R2 = R1 R2 . (A170) −4 In this particle basis, the diagonal central matrices are This implies that we can reexpress the OPE potential as D E Cαα Cββ CD VOPE = VOPE + VOPE 12 (1−)= 12 (1−)= 12(1−±) , (A180) D D αα ββ D = R R¯ I~ I~ [C W + S W ] , C (2−)= C (2−)= C (2−±) , (A181) 1 2 1 · 2 12 C 12 T 12 12 12 (A171) while the non-diagonal are given by where the central and tensor matrices C12 and S12 are a sum of the direct and exchange pieces 3 1 0 0 C PC CD PC CE PC 3 12(J )= 12(J ) 12(J ) , (A172) αβ βα 1 C (1−)= C (1−)= , (A182) − 4 12 12 0 3 0 S PC SD PC 3 SE PC 0 0 1 12(J )= 12(J ) 12(J ) . (A173)   − 4   1 CD SE 3 0 0 0 The 12 and 12 are identical to the ones we discussed 0 100 in the B¯B¯∗ case, see Eqs. (A165), (A166), (A167) and Cαβ Cβα   3 6 12 (2−)= 12 (2−)= . (A183) (A168). The direct matrices are given by 0 010   0 001 5   2 0 0   D − 5 (A184) C (1−±)= 0 0 , (A174) 12  − 2  0 0 3  2  For the tensor matrices we have  5  2 0 0 0 − 3 D 0 + 0 0 1 3 C (2−±)=  2  , (A175) 0 12 0 0+ 3 0 − √2 √2 2 αα 1 1 3  3  S (1−)=   , (A185)  0 0 0+  12 − √2 2 2  2  3 3 3    √ 2 2   2 −  1 3   0 0 5 1 − √2 √2 − 3√2 √2 SD 1 1 3 αβ 5 5 1 12(1−±)=   , (A176) S (1−)=   , (A186) − √2 2 2 12 − 3√2 6 2 3 3 3 1 1 1  √2 2 2   √ 2 2   −   − 2 −      0 5 1 − 3√2 − √2 1 3 3 3 3 6 3 βα 5 5 1 2 10 2 7 35 S (1−)=   , (A187) − − 12 − 3√2 6 − 2  3 3 q0 3 q7 q0  1 1 1 SD 10 10  √2 2 2  12(2−±)= − .    3 q3 7 q9 18   1 3   2 7 3 10 14 √  0  7 5  − √2 − √2  q3 q 18 15  Sββ 1 1 3  6 0  12 (1−)=   , (A188)  35 7√5 − 7  − √2 2 − 2   3 3 3 q (A177)  √ 2 2  − 2 − −    33

1 3 3 3 3 2 2 2 3 6 P , which is µP = mP ∆ , where mP is the mass of the − 2 − 10 2 7 35 meson. Notice the analogy− between Eqs. (B1) and (B2)  3 3 q0 3 q7 q0  αα 10 10 and Eqs. (A109) and (A110) of Appendix A. S (2−)= − , 12  3 q3 7 q9 18  The numerical, flavour and isospin factors of the one  2 7 3 10 14   7√5  eta exchange potential are listed in Table XI for all ver-  6 q3 q0 18 15  tices in which eta exchange is allowed. The format is  35 7√5 7   −  similar to Table IV, which was dedicated to OPE. There  q (A189) is a difference worth mentioning: the G-parity of the η 5 3 1 3 3 meson is positive (in contrast to the negative G-parity of 6 10 2 7 2 35 − − the pion), which implies that the signs of the numerical  3 q0 q 7 q0  η ¯η Sαβ 10 − 10 factors Ri and Ri is not the same as for the pion factors 12 (2−)= , ¯  q1 3 7 q3 6  Ri and Ri of Table IV. In particular we have that  − 2 7 − 10 − 14 − 7√5    RηR¯η > 0 R R¯ < 0 , (B3)  2 q3 q0 6 5  i i i i  35 7√5 7  ⇒ − −  and vice versa. Besides this, it is easy to see that the  q (A190) η η strength of the flavour factors F1 F2 is in general much weaker than the corresponding isospin factors for OPE. 5 3 1 3 3 ¯ η η 2 For example, in the ΣcΣc system we have that F1 F2 = − 6 10 − 2 7 − 35 1/3 for eta exchange in contrast to T~1 T~2 = 2, 1 and  3 q0 q7 q0  · − − βα 10 10 1 for I =0, 1, 2 with the OPE potential. S (2−)= − − . 12  1q 3 7 q3 6  For the one kaon exchange potential we list the dif-  2 7 10 14   − − − 7√5  ferent factors in Table XII for half the non-vanishing  q3 q 6 5   2 0  vertices. We write the vertices in order of decreasing  35 − 7√5 7   q (A191) strangeness for the heavy baryons, with a final kaon. The vertices with an initial antikaon are indeed identical: 1 3 3 3 3 6 3 − 2 10 − 2 7 − 35 (B B′K)= (KB¯ B′) , (B4) A → A →  3 3 q0 3 q7 q0  ββ 10 10 where B and B′ denote the initial and final heavy S (2−)= . 12  3q 3 7 q9 18  baryons. The flavour factors in Table XII are listed as  2 7 3 10 14   − 7√5  numbers, or as the symbol θa, which is a matrix in isospin  q3 q 18 15  i  6 0  space. When the flavour factor is a number, it is implic- − 35 7√5 − 7   q (A192) itly understood that they are multiplied by the identity in isospin space (all vertices conserve isospin). The ma- a 1 trices θi , where a = 2 refers to the isospin state of ± 1 Appendix B: The One Pseudoscalar Meson the kaon, mediate the transitions from isospin-1 to - 2 Exchange Potential in Heavy Hadron Chiral baryons (e.g. Σc Ξc′ K). Their explicit expressions are Perturbation Theory → 1 +1/2 1 0 0 1/2 0 0 θ = , θ− = √2 . (B5) i 0 1 0 i 1. General Form of the Potential √2 ! 0 0 1! a The evaluation of the product of two θi matrices can be The potential generated from the exchange of the kaon done in terms of isospin matrices and the eta can be derived from the same rules we have a a 1 used to calculate the OPE potential. We will not present θ †θ = [~τ T~ + 1] , (B6) 1 2 2 1 · 2 a complete derivation here, but simply a quick overview. a X The one pseudoscalar meson exchange (OME) potentials ~ are formally identical to the OPE potential. For a general where ~τ1 and T2 refer to the isospin matrices as evaluated pseudo Nambu-Goldstone boson P we can write for the initial or final state.

2 ¯ g ~a ~q~a ~q V T S = RP R¯P 3 F P F P 1 · 2 · , (B1) OME − 1 2 2f 2 1 2 q2 + µ2 2. Strenth of the Eta and Kaon Exchange P P Potentials 2 SS¯ P ¯P g2 P P ~a1 ~q~a2 ~q VOME = R1 R2 2 F1 F2 2· 2· , (B2) − 2fP q + µP The strength of eta and kaon exchange can be com- P pared to that of OPE using the techniques of Sect. IV. with fP the weak decay constant for the meson P , R1 , P ¯P P The central scale ΛC for the OME potential can be de- R2 numerical factors that depend on the meson P , F1 P fined as follows and F2 flavour factors (which for P = π are the isospin 2 factors we have previously used for OPE), ~a1 and ~a1 P 1 24πfP ΛC = P P P P 2 , (B7) spin operators and µP the effective mass of the meson σ F F µ R R¯ g | 1 2 | | 1 2 | i 34

P ¯P ~ P ¯P P Vertex P Ri Ri Ii ~ai Vertex P Ri Ri Fi ~ai 2 2 1 2 2 Ξc Ξ′ η √3 ~σi Λc Ξ′ K 1 ~σi → c q 3 q 3 2 → c q 3 q 3 1 Ξc Ξ∗ η √2 -√2 √3 S~† Λc Ξ∗ K √2 -√2 1 S~† → c 2 i → c i 2 2 1 2 2 Σc Σc η ~σi Σc Ξc K θi ~σi → 3 3 √3 → q 3 q 3 − 1 1 1 ~ Σc∗ Σc η - Si Σ∗ Ξc K √2 -√2 θi S~i → √3 √3 √3 c → − 1 1 1 ~ 2 2 Σc Σc∗ η - Si† Ξ Ω K 1 ~σ → √3 √3 √3 c c q 3 q 3 i 2 2 1 ~ → Σc∗ Σc∗ η Σi √ √ ~ → 3 3 √3 Ξc Ωc K 2 - 2 1 Si† 2 2 1 → 2 2 Ξc′ Ξc′ η ~σi Σ Ξ′ K θ ~σ → 3 3 2 √3 c c 3 3 i i 1 1 1 → 1 1 Ξ∗ Ξ′ η - S~i Σ∗ Ξ′ K - θi S~i c → c √3 √3 2 √3 c → c √3 √3 1 1 1 1 1 Ξ′ Ξ∗ η - S~† Σc Ξ∗ K - θi S~† c → c √3 √3 2 √3 i → c √3 √3 i 2 2 1 2 2 Ξ∗ Ξ∗ η Σ~ i Σ∗ Ξ∗ K θi Σ~ i c → c 3 3 2 √3 c → c 3 3 2 2 2 2 2 Ω Ω η ~σ Ξc′ Ωc K 1 ~σi c c 3 3 √3 i 3 3 → → 1 1 1 1 2 Ξ∗ Ωc K - 1 S~i Ω∗ Ωc η - S~i c √3 √3 c √3 √3 √3 → → 1 1 1 1 2 Ξ′ Ω∗ K - 1 S~† Ωc Ω∗ η - S~† c c √3 √3 i c √3 √3 √3 i → → 2 2 2 2 2 Ξ∗ Ω∗ K 1 Σ~ i Ω∗ Ω∗ η Σ~ i c c 3 3 c → c 3 3 √3 →

TABLE XII. Numerical, isospin and spin factor associated TABLE XI. Numerical, isospin and spin factor associated with each vertex of the one kaon exchagne potential for the with each vertex of the one eta exchange potential for the ST¯ and SS¯ systems. The arrows are used to indicate the final ST¯ and SS¯ systems. The arrows are used to indicate the baryon state in the vertex and P indicates the pseudo Gold- final baryon state in the vertex and P indicates the pseudo stone meson in that vertex. The numerical and spin factors Goldstone meson in that vertex. The numerical factors RP i — RP , R¯P and ~σ , Σ~ , S~ — are as in Table XI. The flavour and R¯P are for baryons and antibaryons respectively, while i i i i i i factors F P are is most cases a number (implicitly multiplied F P are the flavour factors. ~σ are the Pauli matrices, Σ~ are i i i i by the identity in isospin space to guarantee isospin conser- the spin S = 3/2 matrices and S~i are the 2 4 matrices that × varion). The exception are the Σc Ξc family of transitions are used for the transitions from spin-1/2 to spin-3/2 baryon. → a in which the flavour factors are 2 3 matrices θi , with a de- noting which of the two isospin states× of the kaon have been exchanged. which is analogous to Eq. (76), expect for the changes . f f and τ F P F P , (B8) π → P → 1 2 ¯ P ¯P mπRc while R1R2 = R1 R2 . The comparison is in fact direct and e that are included to take into account that if we write| it| as| | the OME and OPE potentials cease to be valid below a certain distance, see of Eq. (98). If we add these factors, τ f 2 P P we end up with ΛC = P| | P 2 ΛC , (B9) F1 F2 × fπ × | | 2 mP Rc P τ fP e which merely involves the evaluation of a few numerical ΛT (mP )= | | ΛT (mp) . F P F P × f 2 × emπRc × factors. For the eta meson we have | 1 2 | π 2 (B12) τ fη | | =3, 6, 9 and 1.5 , (B10) (mP mπ)Rc F ηF η f 2 ∼ where e − 3 5 and 2.5 4 for the eta and | 1 2 | π kaon, respectively. The∼ addition− of this− factor means that where we have taken fη fK 160 MeV. By multiply- the tensor scale ΛP is in general considerably larger than ≃ ∼ η T ing these factors it is apparent that ΛC is considerably the one for OPE. In particular it can be regarded as a harder than ΛC. The conclusion is that one eta exchange hard scale. is very suppressed with respect to OPE. For one kaon exchange the factors are Appendix C: The Contact-Range Potential τ f 2 | | =1, 2 and K 1.5 , (B11) F K F K f 2 ∼ | 1 2 | π The calculation of the contact-range potential will use which are not particularly large. The flavour factor sup- a different set of techniques that the one of the OPE po- presion is larger for the lower isospin states, for which tential. While we derived the OPE potential from the the OPE is stronger. The outcome is that the one kaon lowest-dimensional Lagrangian compatible with HQSS exchange potential is perturbative. and chiral symmetry, for the contact-range potential we P The comparison of the tensor scales ΛT and ΛT is sim- will use HQSS without any explicit reference to a La- ilar except for the existence of factors of the type emP Rc grangian. We will first take into account that the lowest 35 order contact-range potential is simply a constant in mo- function into heavy and light components as follows mentum space ( ) ( ) (0) ∗ ¯ ∗ p~ V ~p = C, (C1) B6 B6 (J −) = DSH ,SL (J) SH SL ,(C6) h ′| C | i | i ⊗ J SH ,SL X where ~p and ~p′ are the initial and final center-of-mass momentum of the heavy baryon-antibaryon pair and C a where D are the coefficients for this change of basis. They coupling. In principle the coupling C depends on spe- fulfill the condition cific baryon-antibaryon system and its quantum numbers. But HQSS precludes precludes C to depend on D (J) 2 =1 . (C7) the heavy quark spin, which will translate into a reduc- | SH ,SL | tion of the number of couplings. We will explain in the SXH ,SL following lines how to do that. Heavy baryons are Qqq states with the structure From this decomposition we can calculate the light-spin | i components of the contact-range potential: 1 B3¯ = Q (qq)sL=0 = , (C2) | ij 2 1 ′ ′ B6 = Q (qq)sL=1 = , (C3) SH′ SL′ V SH SL = δSH S δSLS VSL , (C8) | ij 2 h ⊗ | | ⊗ i H L 3 B6∗ = Q (qq)sL=1 j= , (C4) | i 2 where the light and heavy spin decouple. with sL = 0, 1 the light-quark pair spin and with j the The general way to carry on the heavy-light spin de- total spi The application of HQSS to the contact-range composition is to consider the spin wave function of the couplings C implies that they depend on the total light heavy hadrons spin S~L = ~sL1 +~sL2, but not on the total spin J~ = ~j1 +~j2 or the total heavy spin S~H = ~sH1+~sH2. To determine the H1 = sH sL j1 , H2 = sH sL j2 , (C9) exact structure we have to study the coupling of the light- | i | 1 1 i | i | 2 2 i where sH , sH is the heavy spin, sL , sL the light quark spin for different types of heavy baryon-antibaryon 1 2 1 2 spin and j1, j2 the angular momenta of the two hadrons. molecules. In the following lines we will explain how to When we couple the two hadrons together we have ( ) ¯( ) ¯( ) do this for the B6∗ B6∗ , B3¯B6∗ systems, ordered by decreasing degree of complexity. We did not list the system H1H2 = sH1 sL1 j1 sH2 sL2 j2 ¯ | i | i | i B3¯B3¯ SL =0 , (C5) = D (J) (s s )S (s s )S (j j )J , ⇒ SH ,SL | H1 H2 H L1 L2 L 1 2 i SH ,SL for which the light spin structure is trivial because sL1 = X sL2 = 0. (C10)

which is merely a detailed version of Eq. (C6), where the 1. The SS¯ Contact Potential notation indicates that the heavy spins coupled to SH , the light spins to SL and the angular momenta to J. The ¯ For a heavy SS baryon-antibaryon system (i.e. SL = coeﬃcients DSH ,SL (J) can in fact be expressed in terms 1 for both baryons) we can decompose the spin wave of 9-J symbols

D (J)= s s j s s j (s s )S (s s )S (j j )J SH ,SL h H1 L1 1 H2 L2 2|| H1 H2 H L1 L2 L 1 2 i

sH1 sL1 j1 = (2j + 1)(2j + 1)(2S + 1)(2S + 1) . (C11) 1 2 H L sH2 sL2 j2   p SH SL J   

Finally if we are considering antihadrons, we should con- tem, for the B6B¯6 case we find the following sider their behaviors under C-parity to define their spin wave functions consistently: they might differ by a sign 1 2 B6B¯6(0−) = 0H 0L + 1H 1L , from the ansatz sH sLj . | i √3 ⊗ 3 ⊗ J=0 | i r (C12)

√2 1 B6B¯6(1−) = 0H 1L 1H 0L | i 3 ⊗ − 3√3 ⊗ If we go back to the SS¯ heavy baryon-antibaryon sys- 2 5 + 1H 2L , (C13) 3 3 ⊗ J=1 r

¯ ¯ For the B6B6∗ and B6∗B6 cases, we include a minus sign cases where it applies ¯ in front of the states containing a B6∗ to highlight the ¯ + C-parity convention that we employ here: B6∗B6(1− ) =1H 1L , (C22) | i ⊗ J=1

√2 4 B∗B¯6(1−−) = 0H 1L 1H 0L | 6 i 3 ⊗ − √ ⊗ 1 2 2 3 3 B B¯∗(1−) =+ 0 1 1 0 −| 6 6 i 3 H ⊗ L − 3 3 H ⊗ L 1 5 r 1H 2L (C23) − 3 3 ⊗ J=1 1 1 5 r + 1H 1L 1H 2L , 2 1 √2 ⊗ J=1 − 3 6 ⊗ J=1 B B¯ (2 +) = 0 2 1 1 . r 6∗ 6 − H L H L (C14) | i r3 ⊗ − √3 ⊗ J=2

¯ 1 2 2 B6∗B6(2−−) =1H 2L , (C24) + B∗B¯ (1−) = 0 1 + 1 0 | i ⊗ J=2 | 6 6 i −3 H ⊗ L 3 3 H ⊗ L r From the previous decomposition and applying Eq. (C8) 1 1 5 + 1H 1L + 1H 2L , we obtain the contact-range potentials of Section III. √2 ⊗ J=1 3 6 ⊗ J=1 r (C15) 2. The T S¯/ST¯ Contact Potential

For a heavy T S¯/ST¯ baryon-antibaryon system we have to pay attention to the fact that one baryon has SL =0 and the other SL = 1. The expectation is that there will be a direct (exchange) contact term for the transition ¯ 1 1 B6B6∗(2−) =+ 0H 2L 1H 1L T S¯ T S¯ (T S¯ ST¯). The heavy-light decomposition −| i √3 ⊗ − √6 ⊗ J=2 → → of the potential is in this case 1 + 1H 2L , (C16) ′ √2 ⊗ J=2 SH′ (SL′ 1 SL′ 2 )S V SH (SL1 SL2 )SL = h ⊗ ⊗ L | | ⊗ ⊗ i 1 1 ′ ′ ¯ δSH S δSLS SL′ 1 SL′ 2 VSL SL1 SL2 , + B6∗B6(2−) = 0H 2 L + 1H 1L H L h | | i | i −√3 ⊗ √6 ⊗ J=2 (C25)

1 + 1H 2L , (C17) where now we have take into account that the light quark √2 ⊗ J=2 spin of particles 1 and 2 is diﬀerent. If we have a particle-

antiparticle system, C-parity implies ¯ Finally, for the B∗B∗ case we have S′ S′ V S S = S′ S′ V S S . 6 6 h L1 L2 | SL | L1 L2 i h L2 L1 | SL | L2 L1 i (C26) As a consequence, for the T S¯/ST¯ case there are two con- ¯ 2 1 B6∗B6∗(0−) = 0H 0L 1H 1L , tact couplings corresponding to −| i 3 ⊗ − √3 ⊗ J=0 r (C18) 01 V1 01 = 10 V1 10 , (C27) h | | i h | | i √ 01 V1 10 = 10 V1 01 . (C28) ¯ 5 1 10 h | | i h | | i B6∗B6∗(1−) = 0H 1L + 1H 0L −| i 3 ⊗ 3 r 3 ⊗ That is, a contact term that conserves the spin of par- 1 2 ticles 1 and 2 and a contact that exchanges it. For the 1H 2L , (C19) ¯ ¯ − 3 3 ⊗ J=1 B3¯B6 and B6B3¯ the heavy-light spin decomposition reads r 1 2 ¯ B∗B¯∗(2−) = 0H 2L + 1H 1L , B3¯B6(0−) = +1H 1q¯q¯ , (C29) 6 6 | i ⊗ J=0 −| i √3 ⊗ r3 ⊗ J=2 ¯ (C20) B6B3¯(0−) = 1H 1qq , (C30) | i − ⊗ J=0 ¯ B6∗B6∗(3−) =1H 2L , (C21) −| i ⊗ J=3 ¯ 1 2 B3¯B6(1−) = 0H 1q¯q¯ + 1H 1q¯q¯ , | i −√3 ⊗ 3 ⊗ J=1 r (C31) where we have included the minus sign to stress the convention. ¯ 1 2 B6B3¯(1−) =+ 0H 1qq + 1H 1qq , Finally for the B B¯ we can also write the decompo- | i √3 ⊗ 3 ⊗ J=1 6 6∗ r sition in the basis with well-deﬁned C-parity for those (C32)

¯ ¯ while for the B3¯B6∗ and B6∗B3¯ we include the minus sign manifest in front of the states to make the C-parity convention ¯ 2 1 B3¯B6∗(1−) = 0H 1q¯q¯ + 1H 1q¯q¯ , −| i 3 ⊗ √3 ⊗ J=1 r (C33)

¯ 2 1 + B6∗B3¯(1−) = 0H 1qq 1H 1qq , | i 3 ⊗ − √3 ⊗ J=1 r (C34)

¯ B3¯B6∗(2−) =1H 1q¯q¯ , (C35) −| i ⊗ J=2 ¯ + B6∗B3¯(2−) =1H 1qq . (C36) | i ⊗ J=2

In the decomposition above only the quark pair with SL = 1 is written. The other quark pair is implicitly understood, i.e. 1 =1 0 , (C37) qq qq ⊗ q¯q¯ 1 =0 1 . (C38) q¯q¯ qq ⊗ q¯q¯ From the decomposition and the deﬁnitions B = 1 V 1 = 1 V 1 , (C39) D1 h qq| | qqi h q¯q¯| | q¯q¯i B = 1 V 1 = 1 V 1 , (C40) E1 h qq| | q¯q¯i h q¯q¯| | qqi we obtain the contact-range potentials of Section III.

[1] M. Voloshin and L. Okun, JETP Lett. 23, 333 (1976). arXiv:1507.03704 [hep-ph]. [2] N. A. Tornqvist, Phys.Rev.Lett. 67, 556 (1991), revised [13] H.-X. Chen, W. Chen, X. Liu, T. G. Steele, version. and S.-L. Zhu, Phys. Rev. Lett. 115, 172001 (2015), [3] A. V. Manohar and M. B. arXiv:1507.03717 [hep-ph]. Wise, Nucl.Phys. B399, 17 (1993), [14] L. Roca, J. Nieves, and E. Oset, arXiv:hep-ph/9212236 [hep-ph]. Phys. Rev. D92, 094003 (2015), [4] T. E. O. Ericson and G. Karl, arXiv:1507.04249 [hep-ph]. Phys.Lett. B309, 426 (1993). [15] J. He, Phys. Lett. B753, 547 (2016), [5] N. A. Tornqvist, Z.Phys. C61, 525 (1994), arXiv:1507.05200 [hep-ph]. arXiv:hep-ph/9310247 [hep-ph]. [16] C. W. Xiao and U. G. Meißner, [6] S. K. Choi et al. (Belle), Phys. Rev. D92, 114002 (2015), Phys. Rev. Lett. 91, 262001 (2003), arXiv:1508.00924 [hep-ph]. arXiv:hep-ex/0309032. [17] T. J. Burns, Eur. Phys. J. A51, 152 (2015), [7] M. Ablikim et al. (BESIII), arXiv:1509.02460 [hep-ph]. Phys. Rev. Lett. 110, 252001 (2013), [18] L. Geng, J. Lu, and M. P. Valder- arXiv:1303.5949 [hep-ex]. rama, Phys. Rev. D97, 094036 (2018), [8]Z. Q. Liu et al. (Belle), arXiv:1704.06123 [hep-ph]. Phys. Rev. Lett. 110, 252002 (2013), [19] A. Bondar et al. (Belle), arXiv:1304.0121 [hep-ex]. Phys. Rev. Lett. 108, 122001 (2012), [9] Q. Wang, C. Hanhart, and Q. Zhao, arXiv:1110.2251 [hep-ex]. Phys. Rev. Lett. 111, 132003 (2013), [20] I. Adachi et al. (Belle), (2012), arXiv:1207.4345 [hep-ex]. arXiv:1303.6355 [hep-ph]. [21] M. Voloshin, Phys.Rev. D84, 031502 (2011), [10] F.-K. Guo, C. Hidalgo-Duque, J. Nieves, and arXiv:1105.5829 [hep-ph]. M. Pavon Valderrama, Phys. Rev. D88, 054007 (2013), [22] M. Cleven, F.-K. Guo, C. Hanhart, and U.- arXiv:1303.6608 [hep-ph]. G. Meissner, Eur. Phys. J. A47, 120 (2011), [11] R. Aaij et al. (LHCb), arXiv:1107.0254 [hep-ph]. Phys. Rev. Lett. 115, 072001 (2015), [23] B. Aubert et al. (BaBar), arXiv:1507.03414 [hep-ex]. Phys. Rev. Lett. 90, 242001 (2003), [12]R. Chen, X. Liu, X.-Q. Li, and S.- arXiv:hep-ex/0304021 [hep-ex]. L. Zhu, Phys. Rev. Lett. 115, 132002 (2015), 38

[24] D. Besson et al. (CLEO), [48] T. A. Kaeding, Atom. Data Nucl. Data Tabl. 61, 233 (1995), Phys. Rev. D68, 032002 (2003), [Erratum: Phys. arXiv:nucl-th/9502037 [nucl-th]. Rev.D75,119908(2007)], arXiv:hep-ex/0305100 [hep-ex]. [49] S. H. Lee et al. (Belle), Phys. Rev. D89, 091102 (2014), [25] F.-K. Guo, P.-N. Shen, H.-C. Chiang, R.-G. Ping, arXiv:1404.5389 [hep-ex]. and B.-S. Zou, Phys. Lett. B641, 278 (2006), [50] H.-Y. Cheng and C.-K. Chua, arXiv:hep-ph/0603072 [hep-ph]. Phys. Rev. D92, 074014 (2015), [26] F.-K. Guo, P.-N. Shen, and H.- arXiv:1508.05653 [hep-ph]. C. Chiang, Phys. Lett. B647, 133 (2007), [51] T.-M. Yan, H.-Y. Cheng, C.-Y. Cheung, G.-L. Lin, Y. C. arXiv:hep-ph/0610008 [hep-ph]. Lin, and H.-L. Yu, Phys. Rev. D46, 1148 (1992), [Erra- [27] F.-K. Guo and U.-G. Meiss- tum: Phys. Rev.D55,5851(1997)]. ner, Phys. Rev. D84, 014013 (2011), [52] P. L. Cho, Phys. Lett. B285, 145 (1992), arXiv:1102.3536 [hep-ph]. arXiv:hep-ph/9203225 [hep-ph]. [28] M. Altenbuchinger, L. S. Geng, and [53] W. Detmold, C. J. D. Lin, and W. Weise, Phys. Rev. D89, 014026 (2014), S. Meinel, Phys. Rev. D85, 114508 (2012), arXiv:1309.4743 [hep-ph]. arXiv:1203.3378 [hep-lat]. [29] L.-S. Geng, J.-X. Lu, M. P. Valderrama, and [54] S. Fleming, T. Mehen, and I. W. X.-L. Ren, Phys. Rev. D97, 091501 (2018), Stewart, Nucl. Phys. A677, 313 (2000), arXiv:1705.00516 [hep-ph]. arXiv:nucl-th/9911001 [nucl-th]. [30] M. T. AlFiky, F. Gabbiani, and A. A. Petrov, [55] M. Pavón Valderrama, M. Sánchez Sánchez, Phys. Lett. B640, 238 (2006), arXiv:hep-ph/0506141. C. J. Yang, B. Long, J. Carbonell, and [31] T. Mehen and J. W. Pow- U. van Kolck, Phys. Rev. C95, 054001 (2017), ell, Phys.Rev. D84, 114013 (2011), arXiv:1611.10175 [nucl-th]. arXiv:1109.3479 [hep-ph]. [56] M. Cavagnero, Phys.Rev. A50, 2841 (1994). [32] M. Pavon Valderrama, Phys. Rev. D85, 114037 (2012), [57] B. Gao, Phys.Rev. A59, 2778 (1999). arXiv:1204.2400 [hep-ph]. [58] S. R. Beane, P. F. Bedaque, L. Chil- [33] J. Nieves and M. Pavon Valder- dress, A. Kryjevski, J. McGuire, and rama, Phys. Rev. D86, 056004 (2012), U. van Kolck, Phys. Rev. A64, 042103 (2001), arXiv:1204.2790 [hep-ph]. arXiv:quant-ph/0010073 [quant-ph]. [34] C. Hidalgo-Duque, J. Nieves, and M. Pavon Valder- [59] M. Pavon Valderrama and E. Ruiz Arriola, rama, Phys. Rev. D87, 076006 (2013), Phys. Rev. C72, 054002 (2005), arXiv:nucl-th/0504067. arXiv:1210.5431 [hep-ph]. [60] M. Pavon Valderrama and E. R. Arriola, [35] F.-K. Guo, C. Hidalgo-Duque, J. Nieves, and Phys. Rev. C74, 054001 (2006), arXiv:nucl-th/0506047. M. Pavon Valderrama, Phys. Rev. D88, 054014 (2013), [61] M. Pavon Valderrama and E. Ruiz Arriola, arXiv:1305.4052 [hep-ph]. Phys. Rev. C74, 064004 (2006), arXiv:nucl-th/0507075. [36] E. Braaten and M. Kusunoki, [62] M. Pavon Valderrama and E. Ruiz Ar- Phys.Rev. D69, 074005 (2004), riola, Phys. Rev. C79, 044001 (2009), arXiv:hep-ph/0311147 [hep-ph]. arXiv:0809.3186 [nucl-th]. [37] S. Fleming, M. Kusunoki, T. Mehen, and U. van Kolck, [63] M. Pavon Valderrama and E. Ruiz Ar- Phys. Rev. D76, 034006 (2007), arXiv:hep-ph/0703168. riola, Phys. Rev. C83, 044002 (2011), [38] D. L. Canham, H. W. Hammer, and arXiv:1005.0744 [nucl-th]. R. P. Springer, Phys. Rev. D80, 014009 (2009), [64] M. Sanchez Sanchez, L.-S. Geng, J.-X. Lu, T. Hyodo, arXiv:0906.1263 [hep-ph]. and M. P. Valderrama, Phys. Rev. D98, 054001 (2018), [39] D. B. Kaplan, M. J. Savage, and M. B. Wise, arXiv:1707.03802 [hep-ph]. Phys. Lett. B424, 390 (1998), arXiv:nucl-th/9801034. [65] D. Gamermann and E. Oset, [40] D. B. Kaplan, M. J. Savage, and M. B. Wise, Eur. Phys. J. A33, 119 (2007), Nucl. Phys. B534, 329 (1998), arXiv:nucl-th/9802075. arXiv:0704.2314 [hep-ph]. [41] U. van Kolck, Nucl. Phys. A645, 273 (1999), [66] J.-J. Wu, R. Molina, E. Oset, and B. S. arXiv:nucl-th/9808007 [nucl-th]. Zou, Phys. Rev. Lett. 105, 232001 (2010), [42] M. C. Birse, J. A. McGovern, and K. G. arXiv:1007.0573 [nucl-th]. Richardson, Phys. Lett. B464, 169 (1999), [67] C. W. Xiao, J. Nieves, and arXiv:hep-ph/9807302 [hep-ph]. E. Oset, Phys. Rev. D88, 056012 (2013), [43] M. Pavón Valderrama and D. R. arXiv:1304.5368 [hep-ph]. Phillips, Phys. Rev. Lett. 114, 082502 (2015), [68] M. Karliner and J. L. Rosner, arXiv:1407.0437 [nucl-th]. Phys. Rev. Lett. 115, 122001 (2015), [44] M. Pavon Valderrama, Int. J. Mod. Phys. E25, 1641007 (2016), arXiv:1506.06386 [hep-ph]. arXiv:1604.01332 [nucl-th]. [69] M. Voloshin, Int.J.Mod.Phys. A21, 1239 (2006), [45] M. C. Birse, Phys.Rev. C74, 014003 (2006), arXiv:hep-ph/0509192 [hep-ph]. arXiv:nucl-th/0507077 [nucl-th]. [70] E. Epelbaum, U. G. Meissner, W. Gloeckle, [46] P. L. Cho, Nucl. Phys. B396, 183 (1993), and C. Elster, Phys.Rev. C65, 044001 (2002), [Erratum: Nucl. Phys.B421,683(1994)], arXiv:nucl-th/0106007 [nucl-th]. arXiv:hep-ph/9208244 [hep-ph]. [71] M. Gell-Mann and M. Levy, Nuovo Cim. 16, 705 (1960). [47] T. D. Lee and C.-N. Yang, Nuovo Cim. 10, 749 (1956), [72] J. J. Sakurai, Annals Phys. 11, 1 (1960). [,233(1956)]. [73] R. Chen, A. Hosaka, and X. Liu, Phys. Rev. D96, 116012 (2017), 39

arXiv:1707.08306 [hep-ph]. Nefediev, Phys. Lett. B763, 20 (2016), [74] A. E. Bondar, A. Garmash, A. I. Milstein, R. Mizuk, arXiv:1605.09649 [hep-ph]. and M. B. Voloshin, Phys. Rev. D84, 054010 (2011), [77] E. Cincioglu, J. Nieves, A. Ozpineci, and arXiv:1105.4473 [hep-ph]. A. U. Yilmazer, Eur. Phys. J. C76, 576 (2016), [75] M. Albaladejo, F. K. Guo, C. Hidalgo- arXiv:1606.03239 [hep-ph]. Duque, J. Nieves, and M. Pavon Valder- [78] L.-Y. Dai, J. Haidenbauer, and U.-G. Meißner, rama, Eur. Phys. J. C75, 547 (2015), JHEP 07, 078 (2017), arXiv:1702.02065 [nucl-th]. arXiv:1504.00861 [hep-ph]. [79] G. Barucca et al. (PANDA), (2018), [76] V. Baru, E. Epelbaum, A. A. Filin, arXiv:1812.05132 [hep-ex]. C. Hanhart, U.-G. Meißner, and A. V.