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 Indonesian Computational Society
http://www.opi.lipi.go.id/situs/gfti/ 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) J.M. Tuwankotta (ITB)
A. S. Nugroho (BPPT) L.T. Handoko (LIPI)
A. Sopaheluwakan (LabMath) M. Nurhuda (UNIBRAW)
A. Sulaksono (UI) M. Satriawan (UGM)
B. E. Gunara (ITB) P. Nurwantoro (UGM)
B. Tambunan (BPPT) P. W. Premadi (ITB)
F.P. Zen (ITB) R.K. Lestari (ITB)
H. Alatas (IPB) T. Mart (UI)
I.A. Dharmawan (UNPAD) Y. Susilowati (LIPI)
I. Fachrudin (UI) Z. Su’ud (ITB)
Honorary Editors
B.S. Brotosiswojo (ITB) M.O. Tjia (ITB)
M. Barmawi (ITB) P. Anggraita (BATAN)
M.S. Ardisasmita (BATAN) T.H. Liong (ITB)
Guest Editors
H. Zainuddin (UPM) K. Yamamoto (Hiroshima)
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. Regular : an article contains an original work.
2. Comment : an article responding to another one which has been published before.
3. Review : an article which is a compilation of recent knowledges in a particular topic. This type of article is only by invitation.
4. Proceedings : an article which has been presented in a scientific meeting.
Paper Submission
The submitted paper should be written in English using the LATEX 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 substantially new information on the Y N interaction,

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 manip-
ulate the formalism of the physics problem to pro-
duce an analytic formula, which should become as
simplest and efficient as possible from the computa-
tional 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 cpu-
time is always related to a simple formula. A simple
formula could be obtained from an extreme approxi-
mation. 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 nu-
merical accuracy up to a certain level, should be ob-
tained to reconcile these problems.
In nuclear and particle physics the problems of nu-
merical cpu-time and accuracy are also abundant.
The electroproduction of the hypertriton,
e + 3 He → e′ + K + + 3Λ H ,
provides an example.
Hypernucleus has been very interesting since a few
decades ago. The hypertriton 3Λ H is the lightest (and,
incidentally, a very loosely bound) hypernucleus. It
is believed that studies of the hypertriton can provide
since the available Y N scattering data are still ex-
tremely poor. Furthermore, since the hypertriton is
the lightest hypernucleus, it is obviously the first sys-
tem in which the Y N potential, including the inter-
esting Λ-Σ conversion, can be tested in the nuclear
medium. This is also supported by the fact that nei-
ther the ΛN nor the ΣN interactions possess suffi-
cient strength to produce a bound two-body system.
Therefore the hypertriton is expected to play an im-
portant 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 el-
ementary process operator between the initial and fi-
nal nuclear wave functions. The wave functions are
obtained from the solutions of Faddeev equations us-
ing modern nucleon-nucleon and hyperon-nucleon po-
tentials [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 cal-
culated cross sections.
Recently, we have calculated the hypertriton pho-
toprodutcion 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 re-
quired, whereas on the other hand the accuracy of the

⃝
c 2008 GFTI & MKI 0403-1
2 Numerical problem in computation of the ...

calculation should be maintained up to the present

Table 1: The orbital angular momentum, spin, total an-
experimental error-bars.
gular momentum, for the three-body and two body quan-
This paper has been greatly motivated by the facts tum numbers along with their probabilities (in %, the last
discussed above. The emphasize of this paper is, how- two columns) of the 3 He and the hypertriton wave func-
ever, different from the previous ones. Here, we shall tions [5, 6].
discuss more on the numerical problems of the calcu- 3 3
lation. This paper is organized as follows. In Section 2 α L S J l 2j 2T He ΛH
we shall explain why the cpu-time is critical in our cal- 1 0 0 0 0 1 1 44.58 -
culation. Section 3 will briefly present the theoretical 2 0 1 1 0 1 0 44.90 93.49
formalism of our approach. In Section 4 we discuss 3 2 1 1 0 1 0 2.85 5.79
the result of our investigation. Comparison with ex- 4 0 1 1 2 3 0 0.96 0.03
perimental data will be given in Section 5. Finally, we 5 2 1 1 2 3 0 0.19 0.03
shall conclude our findings in Section 6. 6 1 0 1 1 1 0 0.09 0.00
7 1 0 1 1 3 0 0.20 0.01
2 WHY DOES THE CPU-TIME BECOME 8 1 1 0 1 1 1 1.11 -
CRITICAL? 9 1 1 1 1 1 1 1.11 -
10 1 1 1 1 3 1 0.44 -
In the numerical computation we encounter the multi- 11 1 1 2 1 3 1 0.06 -
dimensional numerical integrations, which are handled 12 3 1 2 1 3 1 0.31 -
by using the Gaussian formula. The basic formula for 13 1 1 2 3 5 1 1.02 -
the Gaussian integration is given by 14 3 1 2 3 5 1 0.02 -
∫ 15 2 0 2 2 3 1 0.27 -
∑
N
16 2 0 2 2 5 1 0.43 -
f (x) dx = f (xi ) wi , (1)
17 2 1 2 2 3 0 0.12 0.02
i=1
18 2 1 2 2 5 0 0.10 0.02
where wi denotes the Gauss weight, whoose total 19 2 1 3 2 5 0 0.21 0.05
number N is determined by the available supporting 20 4 1 3 2 5 0 0.05 0.01
points, in the case of the wave function momenta, or 21 2 1 3 4 7 0 0.13 0.01
by the integration stability in the case of angular in- 22 4 1 3 4 7 0 0.04 0.01
tegrations. 23 3 0 3 3 5 0 0.01 0.00
In the present work the numbers of partial waves 24 3 0 3 3 7 0 0.01 0.00
for the 3 He and 3Λ H wave functions are 34 and 16, re- 25 3 1 3 3 5 1 0.05 -
spectively. These partial waves are shown in Table 1, 26 3 1 3 3 7 1 0.05 -
along with the corresponding probabilities. In both 27 3 1 4 3 7 1 0.01 -
wave functions the numbers of supporting points for 28 5 1 4 3 7 1 0.07 -
the two-body momentum (represented by p) and the 29 3 1 4 5 9 1 0.18 -
spectator momentum (represented by q) are 34 and 30 5 1 4 5 9 1 0.01 -
20, respectively. As can be seen in the next Section, 31 4 0 4 4 7 1 0.05 -
calculation of one point of the corresponding differen- 32 4 0 4 4 9 1 0.06 -
tial cross section is equivalent to an integration with 33 4 1 4 4 7 0 0.01 0.00
almost two billions grid points. 34 4 1 4 4 9 0 0.01 0.00
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 approx-
imate the full calculation, whereas the use of α ≤ 5
kaon angles, where we noticed that the two recent ex-
would lead to a perfect result [2]. It has been also em-
perimental data sets from SAPHIR [7] and CLAS [8]
phasized that careful investigations in a wide range
collaborations show a problem of lack of mutual con-
of kinematics should be performed, before we can ap-
sistency [9]. Thus, the production of hypertriton at
ply this approximation to the hypertriton photo- and
this kinematics could also expected to solve this dis-
electroproduction [2].
crepancy problem. However, it should be realized that
Furthermore, it has been also known that the hy-
the extraction of the information on the elementary
pernucleus production cross section is quite sensitive
amplitude from the nuclear cross sections requires a
to the elementary amplitude, especially at the forward
massive fitting process, which would become definitely
1 The explanation of α is also given in Table 1. impossible if the computation time required to calcu-

0403-2
3 Numerical problem in computation of the ...

late these cross sections were extremely long. K (qK ) Λ N N

These facts indicate that the cpu-time required to Λ
calculate the cross section becomes very critical in the 3H
present work. In this work we shall quantitatively in-
k’1 k2 k3
vestigate the effects of the omission of the higher par-
tial 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. Jµ
Thus, our motivation in this work is obvious, i.e. γv ( k )
to shorten the cpu-time for which the deviation of the
calculated cross section from the full calculation is still
e’
controllable.
k1 k2 k3
3He
3 THEORETICAL FORMALISM

The complete formalism of the hypertriton electropro- N N N

duction off 3 He in an impulse approximation can be e
found, e.g., in Ref. [2]. Up to the difference between
the initial and final nuclear masses, as well as between Figure 1: Feynman diagram describing the process of the
hypertriton electroproduction on a 3 He target in an im-
the initial proton and the final hyperon masses in the
pulse approximation.
elementary operator, the formulas are similar to those
used in pion electroproduction off 3 He [11]. To facili-
tate the reader, here we will only present and briefly
discuss the most important part of them. We start we can reformulate the transition matrix element in
with the corresponding nuclear transition matrix ele- Eq. (2) into the form
ment, which can be written as [2] √ ∑∑
Mµ = 3 (LmL SmS |JmJ )
Mµ ≡ ⟨ 3Λ H | J µ | 3 He ⟩ = α,α′ mm′
√ ∫ 3 ( )
3 d p d3 q Ψ∗3 H (p, q ′ ) × (LmL SmS |J ′ mJ ′ ) lml 12 ms |jmj
Λ ( )( )
× l′ ml′ 12 ms′ |j ′ mj ′ JmJ jmj | 21 Mi
× J µ (k, k1 , k′1 ) Ψ3 He (p, q) , (2) ( )
√ × J ′ mJ ′ j ′ mj ′ | 12 Mf δLL′ δmL mL′ δSS ′
where the factor of 3 on the right hand side of Eq. (2) ∫
comes from the anti-symmetry of the initial state, J µ × δmS mS′ δT 0 p2 dp d3 q ϕα′ (p, q ′ )
represents the elementary operator, while the integra- ′
tions are taken over the three-body momentum coor- × ϕα (p, q) Yml l′ (q̂ ′ ) Yml l (q̂)
dinates2 × ⟨ 21 , ms′ | J µ | 21 , ms ⟩ , (6)
p = 1
(k2 − k3 ) , q = k1 , (3)
2 where α and α′ indicate the partial waves
and the hyperon momentum in the hypertriton is of the initial and final nuclear wave functions
given by given in Table 1, respectively, while the nota-
tions of m = (mL mS ml ms mJ mj ) and m′ =
q′ = k1 + 2
3 (k − q K ) . (4) (mL′ mS ′ ml′ ms′ mJ ′ mj ′ ) are introduced just for con-
By explicitely writing the three-body wave func- veniences.
tions in Eq. (2) in terms of their orbital momenta, Now let us calculate the Gauss supporting points
spins, and isospins, required in the computational procedure. From Ta-
∑ ble 1 we can comprehend that a full calculation of the
Ψ(p, q) = ϕα (p, q)(LmL SmS |JmJ ) four-dimensional integrals in Eq. (6) using all partial
α,m waves could involve integrations with 34 × 16 × 34 ×
×(lml 12 ms |jmj ) (JmJ jmj | 12 Mi ) 20 × 30 × 10 = 110, 976, 000 grid points, where the two
1 last numbers in the multiplications (30 × 10) come
× YmLL (p̂)Yml l (q̂)χSmS χm
2
from the minimum Gauss supporting points for the
¯ 1 1 ⟩
s

¯
× (T ) Mt , (5) numerically-stable angular integrations3 .
2 2
2 See Fig. 1 for the definition of the involved momenta. 3 Integrations 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 4 RESULTS AND DISCUSSION

are performed over all components of the transition
matrix element in the form of 4×4 matrix [ jµν ], where
The cross section obtained from the full calculation
Jµ = (1, σx , σy , σz ) using all partial waves and the deviations from this
  result if we use α ≤ 5, α ≤ 4, and s-waves (α = 2, 4)
j00 jx0 jy0 jz0
 j0x jxx jyx jzx  are displayed in Fig. 2. We note that computation
×  j0y jxy
 . (7)
jyy jzy  time of the numerical data required by the plot of the
j0z jxz jyz jzz differential cross section shown in the upper panel of
Fig. 2 (consisting of 31 × 31 = 961 points) on a PC
Obviously, computation of the cross section becomes with a single processor 3 GHz Pentium-4 is about 11
numerically more challenging, since it is equivalent days (15,344 min).
to the problem of integration with 1,775,616,000 grid
points. Furthermore, the result of this integration From the second panel of Fig. 2 it is obvious that
must be summed over angular-momentum and spin limiting the partial waves up to α = 5 yields an ac-
projections mJ , mJ ′ , mS , and ms [indicated by m and curate approximation, since in general it just slightly
m′ in Eq. (6)]. The selection rule represented by the underestimates the full calculation. The largest dis-
three Kronecker delta functions in Eq. (6) along with crepancies are found at the two cross section peaks
current conservation fortunately reduce this num- at W ≈ 3.75 GeV and 4.10 GeV close to the forward
ber to about 156 millions grid points. Nevertheless, angle, i.e., about 0.15 nb/sr (less than 3%). The av-
this number still indicates a time-consuming numeri- erage deviation over these 961 points is only 0.019
cal computation, since the summations over angular- nb/sr. A similar behavior is also found if we use
momentum and spin projections mJ , mJ ′ , mS , and ms α ≤ 4, although in this case the calculated cross sec-
in Eq. (6) is equivalen to a multiplication of this num- tion slightly overeshoots the cross section of the full
ber with a factor of more than 10. calculation. Here, the average and largest deviations
The differential cross section of the hypertriton elec- are about 0.038 nb/sr and 0.17 nb/sr (less than 4%),
troproduction in the c.m. system is similar to that of respectively. Finally, the largest deviation, almost 1
the meson electroproduction of the nucleon. It can be nb/sr at the top of the highest cross section peak, is
written as obtained if we use only s-waves. In view of the exper-
imental data, it is obvious that the latter provides a
dσ dσT dσL dσTT
= + ϵL +ϵ cos 2ϕ relatively poor approximation method for the hyper-
dΩ dΩ dΩ dΩ triton electroproduction, since the largest differential
√ dσLT
+ 2ϵL (1 + ϵ) cos ϕ , (8) cross section is around 5 nb/sr.
dΩ
Let us consider the cpu-times required to make the
where ϵ indicates the virtual photon polarization, ϵL =
plots just shown. The required cpu-times to calculate
−(k 2 /k2 ) ϵ and the individual cross sections can be
the 961 points of the cross section for the full cal-
written as
culation and the three approximation using a single
dσi e2 qK M3 He M3ΛH processor 3 GHz Pentium-4 PC are shown in Fig. 3.
= Wi , (9)
dΩ 4π (W 2 − M32He )W
Clearly, the cpu-time is significantly reduced by a
with i = T, L, TT, and LT. The structure functions factor of 30 if we limit the partial waves up to α = 5,
Wi are given by while the accuracy is still maintained up to about 0.15
1 nb/sr. As a consequence, the required cpu-time to ob-
WT = (W xx + W yy ) , (10) tain the plot shown at the top panel of Fig. 2 becomes
4π
1 less than 9 hours. If we used the partial waves with
WL = W 00 , (11) α ≤ 4, the cpu-time is reduced by a factor of about
4π
1 60, whereas the maximum deviation slightly increases
WTT = (W xx − W yy ) , (12) to 0.17 nb/sr. The use of only s-waves substantially
4π
1 reduces the cpu-time, i.e., by a factor of 300. The av-
WLT = (W 0x + W x0 ) , (13) erage deviations displayed in Fig. 3 show in general
4π
the same behavior.
where the spin averaged Lorentz tensor W µν is re-
lated to the nuclear transition matrix element given We have also investigated the effects of excluding
in Eq. (2) by the higher partial waves at k 2 = −1.0 GeV2 . However,
∑ since the obtained cross sections are quite small (0.32
W µν = 21 Mµ Mν∗ . (14) nb/sr, at most), it is very hard to draw a quantitative
si sf conclusion at this kinematics.

0403-4
5 Numerical problem in computation of the ...

6
10

6
5
dσ / dΩ (nb/sr)

5
10

t (s)
4
3
2
1
4
0 10
-1 0
5
10
15 3
θ (deg) 20 3.6 10
25 4.0 3.8
4.6 4.4 4.2
30
5.0 4.8 W (GeV)
0.20

1.0 0.15
0.8
∆5 (nb/sr)

∆av. (nb)
0.6
0.4 0.10
0.2
0.0
-0.2 0.05
-0.4 0
5
10
15
0.00
θ (deg) 20 3.8 3.6
25 4.2 4.0
30 4.8 4.6 4.4 -0.05
5.0 W (GeV)

1.5
1.0
0.8
∆4 (nb/sr)

0.6
1.0
∆max. (nb)

0.4
0.2
0.0
-0.2
-0.4 0
0.5
5
10
15
θ (deg) 20 3.6 0.0
25 4.0 3.8
4.6 4.4 4.2
30
5.0 4.8 W (GeV)
-0.5
(1) (2) (3) (4)
1.0
0.8
0.6
∆s (nb/sr)

0.4 Figure 3: (Color online) The cpu-time t required to make

0.2 the three-dimensional plot shown in the top panel of Fig. 2
0.0
-0.2 along with the average and maximum deviations from the
-0.4 0 full calculation (∆av. and ∆max. , respectively) for different
approximations, i.e., using all partial waves (1), α ≤ 5 (2),
5
10
θ (deg)
15
20 3.6
α ≤ 4 (3), and only s-waves (4).
25 4.0 3.8
4.6 4.4 4.2
30
5.0 4.8 W (GeV)

5 COMPARISON WITH EXPERIMENTAL

Figure 2: Effects of the higher partial waves on the dif-
ferential 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 pan-
els display the differences between the full calculation and
the calculation by using α ≤ 5 (∆5 ), α ≤ 4 (∆4 ), and only
s-waves (∆s ).
DATA
The main purpose of this investigation is to quanti-
tatively study the effects of excluding higher partial
waves in the hypertriton electroproduction. Never-
theless, 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 experimen-
tal error bars, the use of the s-wave approximation

0403-5
6 Numerical problem in computation of the ...
3He (e, e’ K +) 3ΛH
12
All
10 W = 4.10 GeV
α≤5
dσ/ dΩ (nb/sr)
8
α≤4
6 s-waves
4 has been shown that an accurate calculation, with a
2
0 isospin zero (i.e., using α ≤ 4, since the selection rule
0 5 10 15
θc.m.
K (deg)
Figure 4: (Color online) Comparison between experimen-
tal data [12] and the calculation using all and specific num-
bers of partial waves.
is clearly still amenable. However, we note that the
largest deviation of using only s-waves does not ap-
pear at W = 4.10 GeV. Instead, at the forward direc-
tions 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 com-
putaional 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 dis-
crepancy 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 ob-
tained 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 . Further-
more, the angular distribution of the transverse cross
section (dσT /dΩ) seems to be more flat than the lon-
gitudinal one. Therefore, we may conclude that such
behavior should originate from the elementary ampli-
tude and not from the properties of the nuclear wave
functions. Thus, if we thought that this experimental
data point were correct, then a substantial improve-
ment of the elementary operator should be in order.
This means that the extraction of the elementary in-
formation from the hypertriton production cross sec-
tion would be mandatory in the future works. Other-
wise, new measurements at this kinematics would be
required.
6 CONCLUSION
The effects of higher partial waves on the accuracy
of the calculated differential cross sections of the hy-
pertriton electroproduction have been investigated. It
maximum deviation of less than 4%, could still be ob-
tained if we used the three lowest partial waves with
20 25 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 consid-
eration the use of partial waves with α ≤ 4 is recom-
mended. 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
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.uni-
mainz.de/MAID/kaon/kaonmaid.html.
[5] V. G. J. Stoks, R. A. M. Klomp, C. P. F. Ter-
heggen, and J. J. de Swart, Phys. Rev. C 49
(1994) 2950.
[6] K. Miyagawa and W. Glöckle, Phys. Rev. C 48
(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. Bydžovský and T. Mart, Phys. Rev. C 76
(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 & Lep-
tons: An Introductory Course in Modern Particle
Physics (John Wiley & Sons, New York, 1984) p.
106.

0403-7