This action might not be possible to undo. Are you sure you want to continue?

# Volume 8 Number 0403

ISSN 1979-3898

Journal of Theoretical and Computational Studies

**Numerical problem in computation of the hypertriton production cross section
**

T. Mart J. Theor. Comput. Stud. 8 (2008) 0403 Received: December 4th , 2008; Accepted for publication: December 6th , 2008

Published by

Indonesian Theoretical Physicist Group http://www.opi.lipi.go.id/situs/gfti/ Indonesian Computational Society http://www.opi.lipi.go.id/situs/mki/

**Journal of Theoretical and Computational Studies
**

Journal devoted to theoretical study, computational science and its cross-disciplinary studies URL : http://www.jurnal.lipi.go.id/situs/jtcs/ Editors A. Purwanto (ITS) A. S. Nugroho (BPPT) A. Sopaheluwakan (LabMath) A. Sulaksono (UI) B. E. Gunara (ITB) B. Tambunan (BPPT) F.P. Zen (ITB) H. Alatas (IPB) I.A. Dharmawan (UNPAD) I. Fachrudin (UI) Honorary Editors B.S. Brotosiswojo (ITB) M. Barmawi (ITB) M.S. Ardisasmita (BATAN) Guest Editors H. Zainuddin (UPM) T. Morozumi (Hiroshima)

Coverage area 1. Theoretical study : employing mathematical models and abstractions of a particular ﬁeld in an attempt to explain known or predicted phenomenon. E.g. : theoretical physics, mathematical physics, biomatter modeling, etc. 2. Computational science : constructing mathematical models, numerical solution techniques and using computers to analyze and solve natural science, social science and engineering problems. E.g. : numerical simulations, model ﬁtting and data analysis, optimization, etc. 3. Cross-disciplinary studies : inter-disciplinary studies between theoretical study and computational science, including the development of computing tools and apparatus. E.g. : macro development of Matlab, paralel processing, grid infrastructure, etc. Types of paper 1. 2. 3. 4. Regular : an article contains an original work. Comment : an article responding to another one which has been published before. Review : an article which is a compilation of recent knowledges in a particular topic. This type of article is only by invitation. Proceedings : an article which has been presented in a scientiﬁc meeting.

J.M. Tuwankotta (ITB) L.T. Handoko (LIPI) M. Nurhuda (UNIBRAW) M. Satriawan (UGM) P. Nurwantoro (UGM) P. W. Premadi (ITB) R.K. Lestari (ITB) T. Mart (UI) Y. Susilowati (LIPI) Z. Su’ud (ITB) M.O. Tjia (ITB) P. Anggraita (BATAN) T.H. Liong (ITB) K. Yamamoto (Hiroshima)

Paper Submission

A The submitted paper should be written in English using the L TEX template provided in the web. All communication thereafter should be done only through the online submission page of each paper.

Referees All submitted papers are subject to a refereeing process by an appropriate referee. The editor has an absolute right to make the ﬁnal decision on the paper. Reprints Electronic reprints including covers for each article and content pages of the volume are available from the journal site for free.

Indonesian Theoretical Physicist Group Indonesian Computational Society Secretariat Oﬃce : c/o Group for Theoretical and Computational Physics, Research Center for Physics LIPI, Kompleks Puspiptek Serpong, Tangerang 15310, Indonesia http://www.opi.lipi.go.id/situs/gfti/ http://www.opi.lipi.go.id/situs/mki/ c ⃝ 2008 GFTI & MKI ISSN 1979-3898

J. Theor. Comput. Stud. Volume 8 (2008) 0403

**Numerical problem in computation of the hypertriton production cross section
**

T. Mart Departemen Fisika, FMIPA, Universitas Indonesia, Depok 16424, Indonesia

**Abstract : We discuss the numerical problems in the calculation of the hypertriton electroproduction cross section.
**

We found that the accuracy of the calculation could still be maintained if we used at least four lowest partial waves of the nuclear wave functions. We also present the cpu-times required to calculate the cross sections. We found that the cpu-time could be reduced by a factor of 60 if we used these speciﬁc partial waves, whereas the deviation of calculated cross section from the full calculation is less than 4%. We conclude that in view of the present experimental error bars future investigations could just use these lowest partial waves.

Keywords : Numerical accuracy, cpu-time, partial waves, meson electroproduction, hypernuclei E-mail : tmart@ﬁsika.ui.ac.id Received: December 4th , 2008; Accepted for publication: December 6th , 2008

1

INTRODUCTION

It has been widely accepted that two main problems in the computational physics are the cpu-time and the accuracy of a calculation. Ideally, one could manipulate the formalism of the physics problem to produce an analytic formula, which should become as simplest and eﬃcient as possible from the computational physics point of view. However, in the reality this is very diﬃcult to achieve, since by nature the real physics problems are quite complicated. A short cputime is always related to a simple formula. A simple formula could be obtained from an extreme approximation. On the other hand, an accurate calculation does not allow for such an extreme approximation. Therefore, an optimal approximation, which produces a relatively simple formula and still maintains the numerical accuracy up to a certain level, should be obtained to reconcile these problems. In nuclear and particle physics the problems of numerical cpu-time and accuracy are also abundant. The electroproduction of the hypertriton, e + 3 He → e′ + K + + 3 H , Λ provides an example. Hypernucleus has been very interesting since a few decades ago. The hypertriton 3 H is the lightest (and, Λ incidentally, a very loosely bound) hypernucleus. It is believed that studies of the hypertriton can provide c ⃝ 2008 GFTI & MKI

substantially new information on the Y N interaction, since the available Y N scattering data are still extremely poor. Furthermore, since the hypertriton is the lightest hypernucleus, it is obviously the ﬁrst system in which the Y N potential, including the interesting Λ-Σ conversion, can be tested in the nuclear medium. This is also supported by the fact that neither the ΛN nor the ΣN interactions possess suﬃcient strength to produce a bound two-body system. Therefore the hypertriton is expected to play an important role in hypernuclear physics, similar to that of the deuteron in conventional nuclear physics. In the hypertriton electroproduction process the cross section is obtained from the nuclear transition amplitude, which is calculated by sandwiching the elementary process operator between the initial and ﬁnal nuclear wave functions. The wave functions are obtained from the solutions of Faddeev equations using modern nucleon-nucleon and hyperon-nucleon potentials [5, 6]. These wave functions are expanded in terms of the partial waves with diﬀerent angular momentum, spin, and isospin states. The number of partial waves used determines the accuracy of the calculated cross sections. Recently, we have calculated the hypertriton photoprodutcion and electroproduction oﬀ a 3 He [1, 2, 3]. Our ﬁnding shows that the limitation of the number of partial waves used in the calculation is urgently required, whereas on the other hand the accuracy of the 0403-1

2

Numerical problem in computation of the ...

calculation should be maintained up to the present experimental error-bars. This paper has been greatly motivated by the facts discussed above. The emphasize of this paper is, however, diﬀerent from the previous ones. Here, we shall discuss more on the numerical problems of the calculation. This paper is organized as follows. In Section 2 we shall explain why the cpu-time is critical in our calculation. Section 3 will brieﬂy present the theoretical formalism of our approach. In Section 4 we discuss the result of our investigation. Comparison with experimental data will be given in Section 5. Finally, we shall conclude our ﬁndings in Section 6. 2 WHY DOES THE CPU-TIME BECOME CRITICAL?

Table 1: The orbital angular momentum, spin, total angular momentum, for the three-body and two body quantum numbers along with their probabilities (in %, the last two columns) of the 3 He and the hypertriton wave functions [5, 6].

α 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

L 0 0 2 0 2 1 1 1 1 1 1 3 1 3 2 2 2 2 2 4 2 4 3 3 3 3 3 5 3 5 4 4 4 4

S 0 1 1 1 1 0 0 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 0 0 1 1 1 1 1 1 0 0 1 1

J 0 1 1 1 1 1 1 0 1 1 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4

l 0 0 0 2 2 1 1 1 1 1 1 1 3 3 2 2 2 2 2 2 4 4 3 3 3 3 3 3 5 5 4 4 4 4

2j 1 1 1 3 3 1 3 1 1 3 3 3 5 5 3 5 3 5 5 5 7 7 5 7 5 7 7 7 9 9 7 9 7 9

2T 1 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0

3

He

3 ΛH

In the numerical computation we encounter the multidimensional numerical integrations, which are handled by using the Gaussian formula. The basic formula for the Gaussian integration is given by ∫ f (x) dx =

N ∑ i=1

f (xi ) wi ,

(1)

where wi denotes the Gauss weight, whoose total number N is determined by the available supporting points, in the case of the wave function momenta, or by the integration stability in the case of angular integrations. In the present work the numbers of partial waves for the 3 He and 3 H wave functions are 34 and 16, reΛ spectively. These partial waves are shown in Table 1, along with the corresponding probabilities. In both wave functions the numbers of supporting points for the two-body momentum (represented by p) and the spectator momentum (represented by q) are 34 and 20, respectively. As can be seen in the next Section, calculation of one point of the corresponding diﬀerential cross section is equivalent to an integration with almost two billions grid points. In the hypertriton photoproduction it has been shown that at W = 4.04 GeV the use of four lowest partial waves (i.e. with α ≤ 4)1 would nicely approximate the full calculation, whereas the use of α ≤ 5 would lead to a perfect result [2]. It has been also emphasized that careful investigations in a wide range of kinematics should be performed, before we can apply this approximation to the hypertriton photo- and electroproduction [2]. Furthermore, it has been also known that the hypernucleus production cross section is quite sensitive to the elementary amplitude, especially at the forward

1 The

44.58 44.90 2.85 0.96 0.19 0.09 0.20 1.11 1.11 0.44 0.06 0.31 1.02 0.02 0.27 0.43 0.12 0.10 0.21 0.05 0.13 0.04 0.01 0.01 0.05 0.05 0.01 0.07 0.18 0.01 0.05 0.06 0.01 0.01

93.49 5.79 0.03 0.03 0.00 0.01 0.02 0.02 0.05 0.01 0.01 0.01 0.00 0.00 0.00 0.00

explanation of α is also given in Table 1.

kaon angles, where we noticed that the two recent experimental data sets from SAPHIR [7] and CLAS [8] collaborations show a problem of lack of mutual consistency [9]. Thus, the production of hypertriton at this kinematics could also expected to solve this discrepancy problem. However, it should be realized that the extraction of the information on the elementary amplitude from the nuclear cross sections requires a massive ﬁtting process, which would become deﬁnitely impossible if the computation time required to calcu-

0403-2

3

Numerical problem in computation of the ...

late these cross sections were extremely long. These facts indicate that the cpu-time required to calculate the cross section becomes very critical in the present work. In this work we shall quantitatively investigate the eﬀects of the omission of the higher partial waves of nuclear wave functions on the accuracy of the calculation. For this purpose we make use of the electroproduction process, since photoproduction is only a special case of electroproduction. Thus, our motivation in this work is obvious, i.e. to shorten the cpu-time for which the deviation of the calculated cross section from the full calculation is still controllable. 3 THEORETICAL FORMALISM

K (qK )

Λ N N

Λ 3H

k’ 1

k2 k3

γv ( k ) e’

Jµ

k1 k2 k3

3He

The complete formalism of the hypertriton electroproduction oﬀ 3 He in an impulse approximation can be found, e.g., in Ref. [2]. Up to the diﬀerence between the initial and ﬁnal nuclear masses, as well as between the initial proton and the ﬁnal hyperon masses in the elementary operator, the formulas are similar to those used in pion electroproduction oﬀ 3 He [11]. To facilitate the reader, here we will only present and brieﬂy discuss the most important part of them. We start with the corresponding nuclear transition matrix element, which can be written as [2] Mµ ≡ ⟨ 3 H | J µ | 3 He ⟩ = Λ √ ∫ 3 3 d p d3 q Ψ∗ H (p, q ′ ) 3

Λ

e

N

N N

Figure 1: Feynman diagram describing the process of the hypertriton electroproduction on a 3 He target in an impulse approximation.

we can reformulate the transition matrix element in Eq. (2) into the form √ ∑∑ Mµ = 3 (LmL SmS |JmJ ) ( ) × (LmL SmS |J ′ mJ ′ ) lml 1 ms |jmj 2 ( )( ) 1 × l′ ml′ 1 ms′ |j ′ mj ′ JmJ jmj | 2 Mi 2 ) ( × J ′ mJ ′ j ′ mj ′ | 1 Mf δLL′ δmL mL′ δSS ′ 2 ∫ × δmS mS′ δT 0 p2 dp d3 q ϕα′ (p, q ′ )

l l q q × ϕα (p, q) Yml′ (ˆ′ ) Yml (ˆ) 1 1 × ⟨ 2 , ms′ | J µ | 2 , ms ⟩ ,

′

α,α′ mm′

× J µ (k, k1 , k′ ) Ψ3 He (p, q) , (2) 1 √ where the factor of 3 on the right hand side of Eq. (2) comes from the anti-symmetry of the initial state, J µ represents the elementary operator, while the integrations are taken over the three-body momentum coordinates2 p =

1 2

(6)

(k2 − k3 ) , q = k1 ,

(3)

**and the hyperon momentum in the hypertriton is given by q′ = k1 +
**

2 3

(k − q K ) .

(4)

By explicitely writing the three-body wave functions in Eq. (2) in terms of their orbital momenta, spins, and isospins, ∑ ϕα (p, q)(LmL SmS |JmJ ) Ψ(p, q) =

α,m

**×(lml 1 ms |jmj ) (JmJ jmj | 1 Mi ) 2 2
**

L × YmL (p)Yml (ˆ)χS S χms ˆ l q m 2 ⟩ 1 × (T 1 ) 2 Mt , 2

2 See

1

(5)

where α and α′ indicate the partial waves of the initial and ﬁnal nuclear wave functions given in Table 1, respectively, while the notations of m = (mL mS ml ms mJ mj ) and m′ = (mL′ mS ′ ml′ ms′ mJ ′ mj ′ ) are introduced just for conveniences. Now let us calculate the Gauss supporting points required in the computational procedure. From Table 1 we can comprehend that a full calculation of the four-dimensional integrals in Eq. (6) using all partial waves could involve integrations with 34 × 16 × 34 × 20 × 30 × 10 = 110, 976, 000 grid points, where the two last numbers in the multiplications (30 × 10) come from the minimum Gauss supporting points for the numerically-stable angular integrations3 .

3 Integrations

Fig. 1 for the deﬁnition of the involved momenta.

over the nucleon angles θ and ϕ [2].

0403-3

4

Numerical problem in computation of the ...

It is also important to note that these integrations are performed over all components of the transition matrix element in the form of 4×4 matrix [ jµν ], where Jµ = (1, σx , σy , σz ) j00 jx0 j0x jxx × j0y jxy j0z jxz jy0 jyx jyy jyz jz0 jzx . jzy jzz

4

RESULTS AND DISCUSSION

(7)

Obviously, computation of the cross section becomes numerically more challenging, since it is equivalent to the problem of integration with 1,775,616,000 grid points. Furthermore, the result of this integration must be summed over angular-momentum and spin projections mJ , mJ ′ , mS , and ms [indicated by m and m′ in Eq. (6)]. The selection rule represented by the three Kronecker delta functions in Eq. (6) along with current conservation fortunately reduce this number to about 156 millions grid points. Nevertheless, this number still indicates a time-consuming numerical computation, since the summations over angularmomentum and spin projections mJ , mJ ′ , mS , and ms in Eq. (6) is equivalen to a multiplication of this number with a factor of more than 10. The diﬀerential cross section of the hypertriton electroproduction in the c.m. system is similar to that of the meson electroproduction of the nucleon. It can be written as dσ dσT dσL dσTT = + ϵL +ϵ cos 2ϕ dΩ dΩ dΩ dΩ √ dσLT + 2ϵL (1 + ϵ) cos ϕ , (8) dΩ where ϵ indicates the virtual photon polarization, ϵL = −(k 2 /k2 ) ϵ and the individual cross sections can be written as e2 qK M3 He M3H dσi Λ = Wi , (9) 2 dΩ 4π (W 2 − M3 He )W with i = T, L, TT, and LT. The structure functions Wi are given by 1 (W xx + W yy ) , 4π 1 W 00 , WL = 4π 1 WTT = (W xx − W yy ) , 4π 1 WLT = (W 0x + W x0 ) , 4π where the spin averaged Lorentz tensor W µν lated to the nuclear transition matrix element in Eq. (2) by ∑ 1 Mµ Mν∗ . W µν = 2 WT =

si sf

The cross section obtained from the full calculation using all partial waves and the deviations from this result if we use α ≤ 5, α ≤ 4, and s-waves (α = 2, 4) are displayed in Fig. 2. We note that computation time of the numerical data required by the plot of the diﬀerential cross section shown in the upper panel of Fig. 2 (consisting of 31 × 31 = 961 points) on a PC with a single processor 3 GHz Pentium-4 is about 11 days (15,344 min). From the second panel of Fig. 2 it is obvious that limiting the partial waves up to α = 5 yields an accurate approximation, since in general it just slightly underestimates the full calculation. The largest discrepancies are found at the two cross section peaks at W ≈ 3.75 GeV and 4.10 GeV close to the forward angle, i.e., about 0.15 nb/sr (less than 3%). The average deviation over these 961 points is only 0.019 nb/sr. A similar behavior is also found if we use α ≤ 4, although in this case the calculated cross section slightly overeshoots the cross section of the full calculation. Here, the average and largest deviations are about 0.038 nb/sr and 0.17 nb/sr (less than 4%), respectively. Finally, the largest deviation, almost 1 nb/sr at the top of the highest cross section peak, is obtained if we use only s-waves. In view of the experimental data, it is obvious that the latter provides a relatively poor approximation method for the hypertriton electroproduction, since the largest diﬀerential cross section is around 5 nb/sr. Let us consider the cpu-times required to make the plots just shown. The required cpu-times to calculate the 961 points of the cross section for the full calculation and the three approximation using a single processor 3 GHz Pentium-4 PC are shown in Fig. 3. Clearly, the cpu-time is signiﬁcantly reduced by a factor of 30 if we limit the partial waves up to α = 5, while the accuracy is still maintained up to about 0.15 nb/sr. As a consequence, the required cpu-time to obtain the plot shown at the top panel of Fig. 2 becomes less than 9 hours. If we used the partial waves with α ≤ 4, the cpu-time is reduced by a factor of about 60, whereas the maximum deviation slightly increases to 0.17 nb/sr. The use of only s-waves substantially reduces the cpu-time, i.e., by a factor of 300. The average deviations displayed in Fig. 3 show in general the same behavior. We have also investigated the eﬀects of excluding the higher partial waves at k 2 = −1.0 GeV2 . However, since the obtained cross sections are quite small (0.32 nb/sr, at most), it is very hard to draw a quantitative conclusion at this kinematics.

(10) (11) (12) (13) is regiven (14)

0403-4

5

Numerical problem in computation of the ...

10

6 5 4 3 2 1 0 -1 0 5 10

6

dσ / dΩ (nb/sr)

t (s)

15 20 25 3.8 3.6 4.2 4.0 4.6 4.4 30 5.0 4.8 W (GeV)

10

5

10

4

θ (deg)

10

3

0.20 0.15

∆av. (nb)

5 10 15 20 3.6 4.0 3.8 25 4.4 4.2 30 4.8 4.6 5.0 W (GeV) 5 10 15 20 25 30 3.8 3.6 4.2 4.0 4.6 4.4 5.0 4.8 W (GeV)

1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 0

∆5 (nb/sr)

0.10 0.05 0.00

θ (deg)

-0.05 1.5

∆max. (nb)

1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 0

∆4 (nb/sr)

1.0 0.5 0.0 -0.5 (1) (2) (3) (4)

θ (deg)

1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 0 5 10

∆s (nb/sr)

θ (deg)

15 20 25 3.8 3.6 4.2 4.0 4.6 4.4 30 5.0 4.8 W (GeV)

Figure 3: (Color online) The cpu-time t required to make the three-dimensional plot shown in the top panel of Fig. 2 along with the average and maximum deviations from the full calculation (∆av. and ∆max. , respectively) for diﬀerent approximations, i.e., using all partial waves (1), α ≤ 5 (2), α ≤ 4 (3), and only s-waves (4).

5

Figure 2: Eﬀects of the higher partial waves on the differential cross section of the hypertriton electroproduction oﬀ 3 He. The upper panel shows the result obtained from the full calculation by using all partial waves. Other panels display the diﬀerences between the full calculation and the calculation by using α ≤ 5 (∆5 ), α ≤ 4 (∆4 ), and only s-waves (∆s ).

COMPARISON WITH EXPERIMENTAL DATA

The main purpose of this investigation is to quantitatively study the eﬀects of excluding higher partial waves in the hypertriton electroproduction. Nevertheless, it is also imperative to compare the results with the available experimental data given in Ref. [12]. This is shown in Fig. 4. Due to the large experimental error bars, the use of the s-wave approximation

0403-5

6

3He

Numerical problem in computation of the ...

12 10 dσ/ dΩ (nb/sr) 8 6 4 2 0 0 5

(e, e’ K +) 3 H Λ All α≤5 α≤4 s-waves

wise, new measurements at this kinematics would be required. 6 CONCLUSION

W = 4.10 GeV

10 15 θc.m. (deg) K

20

25

Figure 4: (Color online) Comparison between experimental data [12] and the calculation using all and speciﬁc numbers of partial waves.

The eﬀects of higher partial waves on the accuracy of the calculated diﬀerential cross sections of the hypertriton electroproduction have been investigated. It has been shown that an accurate calculation, with a maximum deviation of less than 4%, could still be obtained if we used the three lowest partial waves with isospin zero (i.e., using α ≤ 4, since the selection rule excludes the α = 1 component). In this case, the time for calculating diﬀerential cross sections can be reduced by a factor of about 60. For future consideration the use of partial waves with α ≤ 4 is recommended. Our ﬁnding is also supported by the available experimental data. New experimental measurement with 10% accuracy would be very useful to clarify the validity of the approximations discussed in this paper. 7 ACKNOWLEDGEMENTS

is clearly still amenable. However, we note that the largest deviation of using only s-waves does not appear at W = 4.10 GeV. Instead, at the forward directions the largest deviation (approximately 0.99 µb/sr, see Fig. 2) is found with W = 4.20 GeV. Therefore, we propose to measure the hypertrion electroproduction with the same experimental quality, but at W = 4.20 GeV. This would be very useful to check the validity of the s-waves approximation. For the sake of computaional accuracy and eﬃcient cpu-time we would, however, recommend the use of partial waves with α ≤ 4. It is importan to make a remark here that the discrepancy between the calculated cross sections and the experimental data point found at θK = 18.9◦ requires a special explanation. At this kinematics we note that the our calculations are much smaller than those at the forward direction. This behavior does not depend on the total c.m. energy. However, we have also obtained that the longitudinal part of the cross section (dσL /dΩ) dominates other parts of the cross section [see Eq. (8)] in the whole kinematics displayed in Fig. 2 and decreases quickly as a function of the θK . Furthermore, the angular distribution of the transverse cross section (dσT /dΩ) seems to be more ﬂat than the longitudinal one. Therefore, we may conclude that such behavior should originate from the elementary amplitude and not from the properties of the nuclear wave functions. Thus, if we thought that this experimental data point were correct, then a substantial improvement of the elementary operator should be in order. This means that the extraction of the elementary information from the hypertriton production cross section would be mandatory in the future works. Other-

**This work has been partially supported by University of Indonesia.
**

JTCS

REFERENCES [1] T. Mart, L. Tiator, D. Drechsel, and C. Bennhold, Nucl. Phys. A640 (1998) 235. [2] T. Mart and B. I. S. van der Ventel, Phys. Rev. C 78 (2008) 014004. [3] T. Mart, Nucl. Phys. A 2009 (in press). [4] T. Mart and C. Bennhold, Phys. Rev. C 61 (1999) 012201(R); T. Mart, Phys. Rev. C 62 (2000) 038201; C. Bennhold, H. Haberzettl and T. Mart, arXiv:nucl-th/9909022; T. Mart, C. Bennhold, H. Haberzettl, and L. Tiator, http://www.kph.unimainz.de/MAID/kaon/kaonmaid.html. [5] V. G. J. Stoks, R. A. M. Klomp, C. P. F. Terheggen, and J. J. de Swart, Phys. Rev. C 49 (1994) 2950. [6] K. Miyagawa and W. Gl¨ckle, Phys. Rev. C 48 o (1993) 2576. [7] K.-H. Glander et al., Eur. Phys. J. A 19 (2004) 251. [8] R. Bradford et al., Phys. Rev. C 73 (2006) 035202. [9] P. Bydˇovsk´ and T. Mart, Phys. Rev. C 76 z y (2007) 065202. [10] R. A. Adelseck, C. Bennhold and L. E. Wright, Phys. Rev. C 32 (1985) 1681.

0403-6

7

Numerical problem in computation of the ...

[11] L. Tiator and D. Drechsel, Nucl. Phys. A 360, 208 (1981). [12] F. Dohrmann et al., Phys. Rev. Lett. 93 (2004) 242501. [13] F. Halzen and A. D. Martin, Quarks & Leptons: An Introductory Course in Modern Particle Physics (John Wiley & Sons, New York, 1984) p. 106.

0403-7