π−d → π−d On rescattering effects in the reaction V.V. Barua,b, A.E. Kudryavtsevb, V.E. Tarasovb aInstitut fu¨r Kernphysik, Forschungszentrum Ju¨lich GmbH, D–52425 Ju¨lich, Germany bInstitute of Theoretical and Experimental Physics, 3 117259, B. Cheremushkinskaya 25, Moscow, Russia 0 0 2 Abstract n We discuss rescattering corrections to the impulse approximation for the processes a J π−d → π−d and γd → π0d. It is shown that the rescattering effects (RE) give non- 8 negligible contribution to the real part of these amplitudes. At the same time the contributions from the imaginary parts of impulse and rescattering corrections dras- 1 v tically cancel each other. This cancellation means that the processes π−d → π0nn 1 and γd → π+nn/π−pp are strongly suppressed near threshold as required by the Pauli 2 principle. 0 1 0 3 1 Introduction 0 / h Thestudyofthereactionsγd → π0dandπ−d → π−dnearthresholdhasattractedcontinuous t - l attention in the past few decades. Moreover the new experimental data appeared due to c u recent success of the accelerator technologies stimulate increasing theoretical interest in this n field. Inthispaperwewouldliketoconcentrateontherescattering effects(RE)andtheirrole : v for these reactions. Indeed, these effects are found to be important in many of theoretical i X investigations of the reaction γd → π0d (cf., e.g. Refs. [1, 2, 3, 4]). However, recently r in Ref. [5] the discussion about the role of these effects was renewed. In particular it was a emphasized in[5]thatthecontributionfromthetwo-stepprocess γd → π−pp → π0d(cf. Fig. 1a) is totally compensated by the loop corrections to the impulse approximation (LCIA) (cf. Fig. 1b). This was argued by the Pauli principle for the intermediate NN states. Thus, the rescatteringeffectsinRef. [5]donotcontributetotheprocessofcoherentπ0 photoproduction on deuteron near threshold. Obviously, this conclusion of Ref. [5] disagrees with the results of other calculations performed, e.g., in Refs. [1, 2, 3, 4]. Let us discuss the arguments of Ref. [5] in more detail: i) The final π0d state has quantum numbers JP = 1− at low energies where pion is in the S wave with respect to the deuteron. However, the only possible state of the system − − ppπ with l = l = 0 is 0 (here l is the orbital angular momentum of the pp system 1 2 1 and l is the orbital momentum of the pion relative to the pp system). Therefore the 2 S-wave intermediate state ppπ− does not contribute to the process γd → π0d. ii) In other words, the contribution of the diagram in Fig. 1a has to be compensated by the loop corrections to the impulse approximation (Fig. 1b) because of antisymmetry of the wave function for the pair of the intermediate nucleons. 1 Note, that the process γd → π0np → π0d is allowed by quantum numbers. However the amplitude γn → π0n which contributes to this reaction is about factor of 20 smaller than the corresponding amplitude for the charged pion production. In this remark we are going to discuss the role of rescattering effects for the process of pion-deuteron elastic scattering at low energies. The diagrams corresponding to RE and LCIA for the πd-scattering are very similar to the ones for the reaction γd → πd (cf. Fig. 1 and Fig 2b and 2c). Therefore we will investigate the relevance of RE and the problem of the cancellation of RE and LCIA performing the calculation of the πd-scattering amplitude. The πd-scattering length was measured with a high accuracy [6, 7] and its value coincides with the theoretical predictions (cf., e.g., Refs. [8, 9, 10, 11]). In all these theoretical calcula- tionsrescatteringeffects(includingthetwo-stepchargeexchangeprocessπ−p → π0n → π−p) give significant contribution to the value of the pion-deuteron scattering length. Inwhatfollows wewilldirectly demonstratethat therealpartoftherescattering diagram (cf. Fig.2c) gives non-negligible contribution to the pion-deuteron scattering length. It is not compensated by the real part of LCIA (cf. Fig.2b). However, the imaginary parts of RE and LCIA cancel each other. This cancellation means that there is no contribution to observables from the πNN states forbidden by the Pauli principle. 2 Calculation of the πd scattering amplitude Below we use a simple potential approachfor the calculation ofthe πN-scattering amplitude. This approach was already applied to the problem of the determination of the πN-scattering length in Ref. [9]. The model utilizes a pion-nucleon potential V (p,q) which is required πN for solving the Lippman-Schwinger equation T = V +VGT . (1) The S-wave πN-lengths b and b are related to the scattering length a by the equation 0 1 πN a = b +b Iτ , (2) πN 0 1 where Iandτ areisospin operatorsforpionandnucleon, b andb areisoscalar andisovector 0 1 scattering lengths. The analyses [9,10] ofthe experimental data [6,7]show that theabsolute values of b and b are small compared to the typical scale of the problem ∼ µ−1 (where 0 1 µ is the pion mass). Note also that b ≪ b . Thus, the amplitude T in eq. (1) may be 0 1 perturbatively expanded in terms of the potential V (p,q). πN Following Ref. [9] we choose V in the S-wave in the separable form: πN (λ +λ Iτ) V (k,q) = − 0 1 g(k)g(q), (3) πN 2m πN where g(k) = (c2 + k2)−1, m = mµ/(m + µ) and m is the nucleon mass. The cut off πN parameter c characterizes the range of the πN-forces, and usually it is varied in the range 2.5µ ≤ c ≤ 5µ [9, 10]. The parameters λ and λ are chosen in such a way to reproduce the 0 1 scattering lengths b and b . In what follows we will calculate the pion-deuteron scattering 0 1 amplitude up to the second order in terms of the potential V . With this accuracy λ and πN 0 λ are equal to 1 2 c4 c λ = b − (b2 +2b2) , 0 2π2 0 2 0 1 (cid:18) (cid:19) c4 c λ = b 1− (2b −b ) . (4) 1 2π2 1 2 0 1 (cid:18) (cid:19) Corrections to these expressions are of the order of ∼ O(b3,b3) which are negligible. 0 1 Let us calculate the pion-deuteron scattering length using the potential V (cf. eq. πN (3)). i) Single scattering amplitude in the Born approximation The diagram corresponding to this amplitude is shown in Fig. 2a. The expression for the 1(V) πd amplitude f corresponding to the sum of two diagrams with the scattering of pion on πd proton and neutron has the form: µ fπ(1d)V = − (2π)(1+µ/m ) dpϕ2d(p)[Vπ−p +Vπ−n]. (5) d Z Here ϕ (p) is the deuteron wave-function in the momentum space with the normalization d condition dpϕ2(p) = (2π)3. Neglecting by the small corrections of order of ∼ µ/m, one d may take out the potential V in Eq. (5) of the integral and then get: R c f1(V) = 2 b − (b2 +2b2) . (6) πd 0 2 0 1 (cid:20) (cid:21) This contribution is real as it should be in the Born approximation. Note also that the 1(V) value f depends on the value of the parameter c. πd ii) Single scattering in the one-loop approximation The diagram for the one-loop correction to the Born approximation is shown in Fig. 2b. We have to calculate the sum of two diagrams with the scattering of pion on proton and neutron taking into account the sum over all intermediate states. The expression for the (1)VGV amplitude f corresponding to this sum has the form: πd 2µ f(1)VGV = (λ2 +λ2)I(∆m=0)+λ2I(∆m) , πd 1+µ/m 0 1 1 d h i dp g2 mmk−+µµp ϕ2d(p) ds g2 mmk−+µµp + mm−+µµs! I(∆m) = (cid:18) (cid:19) (.7) 2π (2m )2 (k+s)2 p2 (p−s)2 k2 Z πN Z + + +ε −∆m− −i0 d " 2µ 2m 2m 2µ # Here k is the 3-momentum of the initial and final pion, ∆m = mπ− +mp −mπ0 −mn = 3.3 MeV is the excess energy for the charge exchange process π−p → π0n in the intermediate state. For the case of the elastic rescattering ∆m = 0. The integral in eq. (7) is calculated numerically for some values of the cut off parameter c. In the limit of large c, i.e. when c ≫ µ and for µ/m ≪ 1 this integral can be calculated analytically: 3 f(1)VGV = c(b2 +2b2)+2i k (∆m=0)(b2 +b2)+k (∆m)b2 , (8) πd 0 1 0 0 1 0 1 h i where we introduced the notation k2 = k2 + 2µ∆m− 2µε . Note that k ≪ c near the 0 d threshold. Thus, in the limit of large c the resulting contribution from the impulse approximation (cf. Fig. 2a and 2b) to the real part of the πd-scattering amplitude is Ref1 = f1(V) +Ref(1)VGV = 2b . (9) πd πd πd 0 This is a naive but expected result for the real part of the amplitude corresponding (1)VGV to the impulse approximation. The values of Ref for the charge exchange process πd π−d → π0nn → π−d are presented in Table 1 for the different values of parameter c. In (1)VGV contrary to the real part of the loop amplitude the imaginary part of f (cf. eq. (8)) πd does not depend on c as required by the unitarity. Nowletusdiscussthecontributiontothepion-deuteronscatteringlengthfromthedouble scattering process. iii) Double scattering contribution Double scattering diagram is shown in Fig. 2c. Performing the calculation we have the (2) following integral for the doublescattering amplitude f (cf. Ref. [9] for details): πd 4c4 f(2) = (b2 −b2) J(∆m=0)−b2 J(∆m) , πd (2π)5 0 1 1 h i dq dq ϕ (q )ϕ (q )g2(k+q −q ) 1 2 d 1 d 2 1 2 J(∆m) = . (10) (k+q −q )2 +(µ/m)(q2 +q2)+2µ(ε −∆m)−k2 −i0 Z 1 2 1 2 d Inthelimit oflargecandforµ/m ≪ 1thisintegralisreduced tothefollowing expression: e−ikr+ik0(∆m=0)r e−ikr+ik0(∆m)r f(2) = 2(b2 −b2) Ψ2(r) dr −2b2 Ψ2(r) dr , (11) πd 0 1 d r 1 d r Z Z where Ψ (r) is the deuteron wave function in the coordinate space. d (2) In the limit of small k and k , i.e. near the threshold for the real part of f we get 0 πd 1 Ref(2) = 2(b2 −2b2) . (12) πd 0 1 r (cid:28) (cid:29)d This expression is well known as a static limit for the doublescattering amplitude, see, e.g. [12] and references therein. (2) The imaginary part of the amplitude f (11) in the same limit is πd Imf(2) = 2k (∆m=0)(b2 −b2) −2k (∆m)b2. (13) πd 0 0 1 0 1 Note that this contribution is negative because b ≫ b . 1 0 4 iv) Total pion-deuteron amplitude Let us discuss the value for total pion-deuteron scattering amplitude in the limit of large c (c ≫ µ) and for µ/m ≪ 1. For the imaginary part of the resulting amplitude in this limit from eqs. (8) and (13) we get Imf ≈ 4k (∆m=0)b2. (14) πd 0 0 Thus, we obtain that the contributions from LCIA and RE to the imaginary part of the pion-deuteron scattering amplitude cancel each other in the leading order (i.e. terms ∼ b2). The non-vanishing part of Imf is proportional to b2 what corresponds to the elastic 1 πd 0 − − − rescatteringprocessπ d → π pn → π d. Note, thattheimaginarypartsofbothexpressions (8) and (13) behave as two-particle phase space, i.e. proportional to k ∼ Q1/2, where Q 0 is the kinetic energy of the intermediate πNN system. However, three-particle πNN phase space should behave as Q2. This paradox can be resolved if we remind the reader that the approximation µ/m ≪ 1, which implies that the kinetic energies of the intermediate nucleons in Eqs. (7) and (10) are neglected, was used to obtain Eqs. (8) and (13). This approximation corresponds to the the rescattering of pion on the fixed centers. That is why the imaginary parts in Eqs. (8) and (13) behave as Q1/2. Of course, one can avoid this unnecessary simplification and calculate Eqs. (7) and (10) with taking into account the terms of order of O(µ). But it should not change the main result that the contributions m from LCIA and RE to the imaginary part of the πd-scattering amplitude with intermediate charge exchange cancel each other. The result (14) means that the only possible final state − which can be formed in the S-wave in the process of deuteron desintegration is pnπ (with S = 1 and I = 0 for pair of nucleons). The virtual charge exchange does not contribute to the imaginary part of the pion-deuteron amplitude. This conclusion is in agreement with the remark of Ref. [5]. At the same time we would like to stress that there is no complete cancellation between (1)VGV (2) real parts of the amplitudes f and f , i.e. the resulting contribution from LCIA and πd πd RE to the real part of the pion-deuteron scattering amplitude is not small. This conclusion, which is also correct for the process γd → π0d, is in contrary to the arguments of Ref. [5]. (1)VGV As can be seen from Eq. (8), the expression for Ref depends linearly on the cut-off πd (2) parameter c for large values of c and µ/m ≪ 1, whereas Ref in the same limit is totally πd determined by the deuteron wave function, i.e. independent of c (c.f. Eq. (12)). Therefore (1)VGV (2) the cancellation of the real parts of the amplitudes f and f can not be achieved in πd πd this limit (the value c = 2 <|1/r|> ≈ 1.2µ is obviousely not realistic). d In Ref. [9] we have calculated the sum of the real parts of the diagrams presented in Fig. 2varying theparametercinthelimits2.5µ ≤ c ≤ 3.5µ. Theresultsofthepresent numerical calculation are presented in Table 1 for the case when c varies in a larger range and the terms ∼ O(µ/m) are taken into account. In the calculation we use the purely hadronic values for b and b presented in Ref. [7], i.e. b = −2.2×10−3 m−1;b = 90.5×10−3 m−1. This Table 0 1 0 π 1 π clearly confirmes the conclusion discussed above that the real parts of the diagrams of Fig. 2b and 2c do not cancel each other. 5 3 Summary We developed a consequent potential approach to the problem of the calculation of the pion-deuteron scattering length. The πd amplitude was calculated including terms of the second order with respect to the pion-nucleon potential V . The proper symmetrization of πN the wave function for the intermediate nucleons is taken into account automatically in our approach. We show that there is a significant cancellation of the contributions from the imaginary parts of LCIA (cf. Fig. 2b) and RE (cf. Fig. 2c). This cancellation is expected. It simply reflects the fact that the process π−d → π0nn is strongly suppressed near threshold as required by the Pauli principle. However, no such cancellations take place between the real parts of these processes. The integrals for the real parts of the amplitudes (7) and (10) are quite different. In particular, they have different dependence on the cut off parameter c in (1)VGV (2) the formfactor. Therefore we see no reasons for the cancellation of Ref and Ref . π−d π−d The situation for the reaction γd → π0d is quite analogous to that discussed for the reaction πd → πd. There is no any reasons for the cancellation of the real parts of the diagrams shown in Fig. 1a and 1b. This conclusion is in agreement with the results of papers [1, 2, 3, 4] where rescattering effects are found to be important for the reaction γd → π0d . We would like to thank A. M. Gasparyan and V. G. Ksenzov for useful discussions. This work was partly supported by RFBR grant N0 02-02-16465, DFG-RFBR grant 436 RUS 113/652/0-1. References [1] J.H. Koch and R.M. Woloshin, Phys. Rev. C 16, 1968 (1977). [2] P. Bosted and J.M. Laget, Nucl. Phys. A 296, 413 (1978). [3] G. F¨aldt, Phys. Scripta 22, 5 (1980). [4] S. R. Beane et al., Nucl. Phys. A618, 381 (1997) [5] M. Rekalo and Egle Tomasi-Gustafsson, Phys. Rev. C66, 015203 (2002); e-print nucl-th/0112063 (2001). [6] D. Chatellard et al., Phys. Rev. Lett., 74, 4157 (1995); Nucl. Phys. A 625, 855 (1997). [7] H.-Ch. Schr¨oder et al., Phys.Lett. B469, 25 (1995), Eur. Phys. J. C 21, 473 (2001). [8] V.M. Kolybasov and A.E. Kudryavtsev, Nucl. Phys B 41, 510 (1972). [9] V.V. Baru and A.E. Kudryavtsev, Phys. Atom. Nucl. 60, 1475 (1997), πN-Newsletter, 12, 64 (1997) [10] T.E.O. Ericson, B. Loiseau and A.W. Thomas, e-print hep-ph/0009312 (2000), submit- ted to Phys. Rev. C. [11] A. Deloff, Phys. Rev. C 64, 065205 (2001); nucl-th/0104067. [12] T. Ericson and W. Weise, Pions and Nuclei. Clarendon Press. Oxford, 1988, p. 126. 6 (1)VGV (2) Ref [fm] Ref [fm] πd πd c = 2µ 0.0280 -0.0084 c = 3µ 0.0443 -0.0098 c = 4µ 0.0608 -0.0104 c = 5µ 0.0774 -0.0107 Table 1: The real parts of the contributions from the diagrams shown in Fig.2b and 2c for the charge exchange process. γ γ d π- π0 p π-- p p p d π0 d d a b Figure 1: Diagrams with intermediate negative pion rescattering contributing to the process γd → π0d. π- π- π- π π- π- d N π N N N d d d d d π- a b c Figure 2: Feynman diagrams contributing to the πd-scattering amplitude: a – diagram of the first order on the πN potential; b,c – diagrams of the second order on the πN potential. 7