USC(NT)-95-6, nucl-th/9512023 Chiral Perturbation Approach to the pp ppπ0 Reaction → Near Threshold B.-Y. Park, F.Myhrer, J.R.Morones, T.Meissner and K.Kubodera ∗ 6 Department of Physics and Astronomy, University of South Carolina 9 9 Columbia, SC 29208, U. S. A. 1 n a J 6 Abstract 2 2 v The usual theoretical treatments of the near-threshold pp ppπ0 reac- 3 → 2 tion are based on various phenomenological Lagrangians. In this work we 0 2 1 examine the relationship between these approaches and a systematic chiral 5 9 perturbation method. Our chiral perturbation calculation indicates that the / h t pion rescattering term should be significantly enhanced as compared with - l c u the traditional phenomenological treatment, and that this term should have n v: substantial energy and momentum dependence. An important consequence i X of this energy-momentum dependence is that, for a representative threshold r a kinematics and within the framework of our semiquantitative calculation, the rescatteringterminterferesdestructivelywiththeBorn-terminsharpcontrast to the constructive interference obtained in the conventional treatment. This destructive interference makes theoretical cross sections for pp ppπ0 much → smaller than the experimental values, a feature that suggests the importance On leave of absence from Department of Physics, Chungnam National University, Daejeon 305- ∗ 764, Korea 1 of the heavy-meson exchange contributions to explain the experimental data. 13.75.Cs, 13.75.Gx, 12.39.Fe Typeset using REVTEX 2 I. INTRODUCTION Recently Meyer et al. [1] carried out high-precision measurements of the total cross sections near threshold for the reaction p+p p+p+π0. (1) → These measurements were confirmed by Bondar et al. [2]. The early theoretical calculations [3–5] underestimate these s-wave π0 production cross sections by a factor of 5. The basic ∼ features of these early calculations may be summarized as follows. The pion production reactionsareassumedtobedescribedbythesinglenucleonprocess(theBornterm),Fig.1(a), and the s-wave pion rescattering process, Fig.1(b). The π-N vertex for the Born term is assumed to be given by the pseudovector interaction Hamiltonian g i = A ψ¯ σ ∇(τ π) σ ∇,τ π˙ ψ, (2) 0 H 2f · · − 2m { · · } π (cid:18) N (cid:19) where g is the axial coupling constant, and f = 93 MeV is the pion decay constant. A π The first term represents p-wave pion-nucleon coupling, while the second term accounts for the nucleon recoil effect and makes “Galilean-invariant”. For s-wave pion production 0 H only the second term contributes. Since this second term is smaller than the first term by a factor of m /m , the contribution of the Born term to s-wave pion production is π N ∼ intrinsically suppressed, and as a consequence the process becomes sensitive to two-body contributions, Fig.1(b). The s-wave rescattering vertex in Fig.1(b) is commonly calculated using the phenomenological Hamiltonian [3] λ λ = 4π 1 ψ¯π πψ +4π 2 ψ¯τ π π˙ψ (3) H1 m · m2 · × π π The two coupling constants λ and λ in Eq.(3) were determined from the S and S pion 1 2 11 31 nucleon scattering lengths a and a as 1/2 3/2 m m π π λ = 1+ (a +2a ), (4a) 1 1/2 3/2 6 m (cid:18) N(cid:19) 3 m m π π λ = 1+ (a a ). (4b) 2 1/2 3/2 6 m − (cid:18) N(cid:19) The current algebra prediction [6] for the scattering lengths, a = 2a = m /4πf2 = 1/2 − 3/2 π π 0.175m 1, implies that only chiral symmetry breaking terms will give a non-vanishing value −π of the coupling constant λ in Eq.(3). Therefore λ is expected to be very small. Indeed, the 1 1 empirical values a 0.175m 1 and a 0.100m 1 obtained by H¨ohler et al. [7] lead to 1/2 ≃ −π 3/2 ≃ − −π λ 0.005 and λ 0.05. So the contribution of the λ term in Eq.(3) is significantly sup- 1 2 1 ∼ ∼ pressed. Meanwhile, although λ is much larger than λ , the isospin structure of the λ term 2 1 2 is such that it cannot contribute to the π0 production from two protons at the rescattering vertex in Fig.1(b). Thus, the use of the phenomenological Hamiltonians, Eqs.(2) and (3), to calculate the Born term and the rescattering terms illustrated in Figs.1(a) and (b), gives significantly suppressed cross sections for the pp ppπ0 reaction near threshold. Therefore, → theoretically calculated cross sections can be highly sensitive to any deviations from this conventional treatment. These delicate features should be kept in mind in discussing the large discrepancy (a factor of 5) between the observed cross sections and the predictions ∼ of the earlier calculations. A plausible mechanism to increase the theoretical cross section was suggested by Lee and Riska [8]. They proposed to supplement the contribution of the pion-exchange diagram, Fig.1(b), with the contributions of the short-range axial-charge exchange operators which were directly related to heavy-meson exchanges in the nucleon-nucleon interactions [9]. Ac- cording to Lee and Riska, the shorter-range meson exchanges (scalar and vector exchange contributions) can enhance the cross section by a factor 3–5. Subsequently, Horowitz et al. [10] demonstrated, for the Bonn meson exchange potential, a prominent role of the σ me- son in enhancing the cross section, thereby basically confirming the conclusions of Lee and Riska. The possible importance of heavy-meson exchanges may be inferred from the follow- ing simple argument. Consider Fig.1(b) in the center-of-mass (CM) system with the initial and final interactions turned off and with the exchanged particle allowed to be any particle (not necessarily a pion). At threshold, q = m , q = 0, p = p = 0, so that any exchanged 0 π ′1 ′2 4 particle must have k = m /2 = 70 MeV and k = m m +(m /2)2 370 MeV/c, 0 π π N π | | ∼ q which implies k2 = m m . Thus the rescattering process probes two-nucleon forces at π N − distances 0.5 fm corresponding to a typical effective exchanged mass √m m = 370 π N ∼ MeV. Its sensitivity to the intermediate-range N-N forces indicates the possible importance of the two-body heavy meson axial exchange currents considered by Lee and Riska. The par- ticular kinematical situation we considered here shall be referred to as the typical threshold kinematics. Meanwhile, Hern´andez and Oset [11] considered the off-shell dependence of the πN s- wave isoscalar amplitude featuring in the rescattering process, Fig.1(b). They pointed out that the s-wave amplitude could be appreciably enhanced for off-shell kinematics pertinent to the rescattering process. We have seen above that, for the typical threshold kinemat- ics, the exchanged pion can indeed be far off-shell. The actual kinematics of course may deviate from the typical threshold kinematics rather significantly due to energy-momentum exchanges between the two nucleons in the initial and final states, but the importance of the off-shell kinematics for the exchanged pion is likely to persist. Hern´andez and Oset examined two types of off-shell extrapolation: (i) the Hamilton model for πN isoscalar am- plitude based on σ-exchange plus a short range piece [12], and (ii) an extrapolation based on the current algebra constraints. In either case the enhancement of the total cross sec- tion due to the rescattering process was estimated to be strong enough to reproduce the experimental data. A more detailed momentum-space calculation carried out by Hanhart et al. [13] supports the significant enhancement due to an off-shell effect in the rescattering process, although the enhancement is not large enough to explain the experimental data. It should be emphasized that Hanhart et al.’s calculation eliminates many of the kinematical approximations employed in the previous calculations. Given these developments based on the phenomenological Lagrangians, we consider it important to examine the significance of these phenomenological Lagrangians in chiral per- turbation theory (χPT) [14,15] which in general serves as a guiding principle for low-energy 5 hadron dynamics. In the present work we shall describe an attempt at relating the tradi- tional phenomenological approaches to χPT. The fact that χPT accounts for and improves the results of the current algebra also makes it a natural framework for studying thresh- old pion production. Furthermore, in this low-energy regime, it is natural to employ the heavy-fermion formalism (HFF) [16]. The HFF has an additional advantage of allowing easy comparison with Eqs.(2) and (3). It should be mentioned, however, that the application of χPT to nuclei involves some subtlety. As emphasized by Weinberg [17], naive chiral counting fails for a nucleus, which is a loosely bound many-body system. This is because purely nucleonic intermediate states occurring in a nucleus can have very low excitation energies, which spoils the ordinary chiral counting. To avoid this difficulty, one must first classify diagrams appearing in perturbation series into irreducible and reducible diagrams, according to whether or not a diagram is free from purely nucleonic intermediate states. Thus, in an irreducible diagram, every interme- diate state contains at least one meson. The χPT can be safely applied to the irreducible diagrams. The contribution of all the irreducible diagrams (up to a specified chiral order) is then to be used as an effective operator acting on the nucleonic Hilbert space. This second step allows us to incorporate the contributions of the reducible diagrams. We may refer to this two-step procedure as the nuclear chiral perturbation theory (nuclear χPT). This method was first applied by Weinberg [17] to chiral-perturbation-theoretical derivation of the nucleon-nucleon interactions and subsequently used by van Kolck et al. [18]. Park, Min and Rho (PMR) [19] applied the nuclear χPT to meson exchange currents in nuclei. The success of the nuclear χPT in describing the exchange currents for the electromagnetic and weak interactions is well known [19–21]. The present paper is in the spirit of the work of PMR. This article is organized as follows: In the next section we define our pion field and the chiral counting procedure. Then in section III we present the two lowest order Lagrangians, discuss their connection to the early works on this reaction and determine within certain approximations the numerical values of the effective pion rescattering vertex strength, κ . th 6 InsectionIVwe brieflydiscuss theconnectionbetween thetransitionmatrixforthisreaction and the χPT calculated amplitude. In section V we present necessary loop corrections to the Born term, and in section VI we calculate the cross section and discuss the various approximations and the uncertainties of the low energy constants in χPT. Finally in section VII, after discussing some higher chiral order diagrams, we present our main conclusions. A work very similar in spirit to ours has recently been completed by Cohen et al. [22]. II. CHIRAL PERTURBATION THEORY The effective chiral Lagrangian involves an SU(2) matrix U(x) that is non-linearly ch L related to the pion field and that has standard chiral transformation properties [23]. An example is [24] U(x) = 1 [π(x)/f ]2 +iτ π(x)/f . (5) π π − · q In the meson sector, the sum of chiral-invariant monomials constructed from U(x) and its derivatives constitutes the chiral-symmetric part of . Furthermore, one can construct ch L systematically the symmetry-breaking part of with the use of a mass matrix the ch L M chiral transformation of which is dictated by that of the quark mass term in the QCD Lagrangian. To each term appearing in one can assign a chiral order index ν¯ defined by ch L ν¯ d 2, (6) ≡ − where d is the summed power of the derivative and the pion mass involved in this term. A low energy phenomenon is characterized by a generic pion momentum Q, which is small compared to the chiral scale Λ 1 GeV. It can be shown that the contribution of a term of ∼ chiral order ν¯ carry a factor (Q˜/Λ)ν¯, where Q˜ represents either Q or the pion mass m . This π suggests the possibility of describing low-energy phenomena in terms of that contains ch L only a manageably limited number of terms of low chiral order. This is the basic idea of χPT. 7 The heavy fermion formalism (HFF) [16] allows us to easily extend χPT to the meson- nucleon system. In HFF, the ordinary Dirac field ψ describing the nucleon, is replaced by the heavy nucleon field N(x) and the accompanying “small component field” n(x) through the transformation ψ(x) = exp( im v x)[N(x)+n(x)] (7) N − · with v/N = N, v/n = n, (8) − where the four-velocity v is assumed to be almost static, i.e., v (1,0,0,0) [25]. Elim- µ µ ≈ ination of n(x) in favor of N(x) leads to expansion in ∂ /m . Since m 1 GeV Λ, µ N N ≈ ≈ an expansion in ∂ /m may be treated like an expansion in ∂ /Λ. in HFF consists µ N µ ch L of chiral symmetric monomials constructed from U(x), N(x) and their derivatives and of symmetry-breaking terms involving . The chiral order ν¯ in HFF is defined by M ν¯ d+n/2 2, (9) ≡ − where d is, as before, the summed power of the derivative and the pion mass, while n is the number of nucleon fields involved in a given term. As before, a term in with chiral order ch L ν¯ can be shown to carry a factor (Q˜/Λ)ν¯ 1. In what follows, ν¯ stands for the chiral order ≪ defined in Eq.(9). In addition to the chiral order index ν¯ defined for each term in , we assign a chiral ch L order index ν for each irreducible Feynman diagram appearing in the chiral perturbation series for a multifermion system [17]. Its definition is ν = 4 E 2C +2L+ ν¯, (10) N i − − i X where E is the number of nucleons in the Feynman diagram, L the number of loops, and C N the number of disconnected parts of the diagram. The sum over i runs over all the vertices in the Feynman graph, and ν¯ is the chiral order of each vertex. One can show [17] that an i irreducible diagram of chiral order ν carries a factor (Q˜/Λ)ν 1. ≪ 8 Intheliteraturetheterm“effectiveLagrangian”(or“effectiveHamiltonian”)isoftenused to imply that that Lagrangian (or Hamiltonian) is only meant for calculating tree diagrams. The Hamiltonians given in Eqs.(2) and (3) are regarded as effective Hamiltonians of this type. We must note, however, that the effective Lagrangian in χPT has a different meaning. Notonlycan beused beyondtreeapproximationbut, infact,aconsistent chiralcounting ch L even demands inclusion of every loop diagram whose chiral order ν is lower than or equal to the chiral order of interest. As will be discussed below, for a consistent χPT treatment of the problem at hand, we therefore need to consider loop corrections. However, since the inclusion of the loop corrections is rather technical, we find it useful to first concentrate on thetree-diagramcontributions. This simplification allows us to understand thebasic aspects oftherelationbetween thecontributions fromχPTandthephenomenologicalHamiltonians, Eqs.(2) and (3). Therefore, in the next two sections (III and IV) we limit our discussion to tree diagrams. A more elaborate treatment including loop corrections will be described in section V. III. TREE DIAGRAM CONSIDERATIONS In order to produce the one-body and two-body diagrams depicted in Figs.1(a) and 1(b), we minimally need (see below) terms with ν¯ = 1 and 2 in . We therefore work with ch L = (0) + (1), (11) ch L L L where (ν¯) represents terms of chiral order ν¯. Their explicit forms are [15,26] L (0) = fπ2Tr[∂ U ∂µU +m2(U +U 2)] (12a) L 4 µ † π † − +N¯(iv D +g S u)N (12b) · A · 1 C (N¯Γ N)2 (12c) − 2 A A A X (1) = igA N¯ S D,v u N (12d) L − 2mN { · · } +2c m2N¯NTr(U +U 2) (12e) 1 π † − 9 +(c gA2 )N¯(v u)2N +c N¯u uN (12f) 2−8mN · 3 · c9 (N¯N)(N¯iS uN) (12g) − 2mN · c10 (N¯SµN)(N¯iu N) (12h) − 2mN µ In the above ξ U(x), (13) ≡ q u i(ξ ∂ ξ ξ∂ ξ ), (14) µ † µ µ † ≡ − D N (∂ + 1[ξ ,∂ ξ])N, (15) µ ≡ µ 2 † µ and S is the covariant spin operator defined by µ S 1γ [v/,γ ]. (16) µ ≡ 4 5 µ In (1) abovewehaveretainedonlytermsofdirect relevanceforourdiscussion. Thecoupling L constants c ,c and c can be fixed from phenomenology [15]. They are related to the 1 2 3 ¯ pion-nucleon σ-term, σ (t) p m¯(u¯u + dd) p (m¯ = average mass of the light quarks, πN ′ ∼ h | | i t = (p p)2), the axial polarizability α and the isospin-even πN s-wave scattering length ′ A − a+ 1(a + 2a ) 0.008m 1 [7]. (The explicit expressions will be given below.) It ≡ 3 1/2 3/2 ≈ − −π should be noted that in HFF, a part of the term in (1) with the coefficient (c g2/8m ), L 2− A N namely the g2/8m piece, represents the s-wave π-N scattering contribution, which in a − A N traditional calculation is obtained from the crossed Born-term . The four-Fermi non-derivative contact terms in Eq.(12) were introduced by Weinberg [17] and further investigated in two- and three-nucleon systems by van Kolck et al. [18]. Although these terms are important in the chiral perturbative derivation of the nucleon- nucleon interactions [17,18], they do not play a major role in the following discussion of the threshold pp ppπ0 reaction. We therefore temporarily ignore these four-fermion terms → and come back to a discussion of these terms in the last section. 10