Hodgson
Hodgson
Hodgson
INTRODUCTORY OVERVIEW
Abstract
The formalism of the optical model is brie
y reviewed, together with the ways it can be used to
analyse experimental data. The usefulness and limitations of global potentials is assessed. The justi-
cation for using optical potentials to give distorted waves for nuclear reaction calculations is examined
and some remaining problems discussed. The nal section is devoted to the application of optical model
potentials to calculate reaction cross-sections at energies up to 200 MeV.
1. Introduction
During the last forty years the nuclear optical model has been extensively applied to analyse the
elastic scattering of pions, nucleons and heavier particles by nuclei over a wide range of energies [1]. It
has been extended to include inelastic scattering by the coupled-channels formalism [2] and consideration
of dispersion eects enables both bound and scattering states to be described by the same mean eld [3].
The interaction of a nucleon with a nucleus is inherently complicated, and the optical model
represents it by a phenomenological potential of very simple form, with parameters that are adjusted to
t the experimental data. This potential has an imaginary part that takes into account the absorption of
the reaction
ux from the elastic channel to the non-elastic reaction channels. This is analogous to the
scattering and absorption of light by a medium of complex refractive index, which is why it is called the
optical model.
Elastic scattering cross-sections are aected by the excited states of the compound system and
by the residual states in the non-elastic channels. This causes the cross-section to
uctuate, and the
uctuations may be seen and analysed when measurements of high precision are made on light nuclei.
Usually the experimental resolution of the detector is sucient to average over these
uctuations, but it
must be emphasised that the optical model applies only to energy-averaged cross-sections.
The optical model is sometimes criticised for its many parameters: with so many you could t
anything! However the
exibility of the model is limited, and the form and the magnitude of the real
part of the potential, and to a lesser extent the imaginary part, can be estimated quite well from simple
physical considerations. Furthermore, the strength of its physical basis is conrmed by its ability to t
with very similar parameters the scattering from many nuclei over a large energy range including the
bound states of the simple shell model. To achieve this, the depths of the potential vary smoothly with
energy, and this can be attributed in the case of the real part of the potential to its non-local nature.
Optical potentials obtained by analysis of elastic scattering are widely used to generate the distorted
waves used to analyse the cross-sections of many reactions, and these analyses have proved to be a powerful
tool in determining nuclear structures.
In this review, the forms of the optical potential are described in Section 2, and the methods
of optical model analysis in Section 3. In spite of its many successes, the optical model is still being
developed and there remain some problems to study. Among these are the validity of the distorted
waves generated by optical potentials and the usefulness of reaction cross-sections, and this is discussed
in Section 4. The concluding Section 5 is devoted to reactions up to 200 MeV.
1
2. Optical Model Potentials
Over the years optical model potentials have become increasingly sophisticated to enable them to
account for more accurate and extensive experimental data.
The early potentials had the form
V = (U + iW )f (r) (2:1)
where U and W are the real and imaginary potential depths and f (r) is the radial form factor. This
usually has the Saxon-Woods form
f (r) = 1, (2:2)
1 + exp r,aR
where R and a are the radius and surface diuseness parameters. This is the four-parameter optical
potential. When inserted into the quantum mechanical formalism it gives the dierential cross-section for
elastic scattering and reaction (absorption) cross-sections for charged particles, and for neutral particles
the total elastic cross-section and the total cross-section as well.
Analyses of more accurate data showed that a better t could be obtained by allowing the form
factor parameters to dier for the real and imaginary potentials; this gives the six-parameter potential.
At low energies the interaction takes place predominantly in the surface region of the nucleus, so
some analyses use a surface-peaked form factor for the imaginary part of the potential having the form
g(r) = ,4a d fdr(r) = f1 +4 exp[(r , R)=a] (2:3)
exp[(r , R)=a]g2
where the factor ,4a is introduced so that g(R) = 1. The imaginary potentials with f (r) and g(r)
are referred to as volume and surface absorption respectively. It is possible to include both surface and
volume absorption terms and this is necessary for the dispersion relation analyses described below. Since
the cross-sections are rather insensitive to the form of the potential, very accurate data are necessary for
these analyses. In some early analyses a Gaussian form was used for g(r).
The addition of a spin-orbit term enables the polarisation of the scattered particle to be calculated.
This term has the form 2
Vso (r) = m c (Us + iWs ) 1r d fdrs (r) L:
h
(2:4)
The square of the pion Compton wavelength is a relic of the derivation of this term from the meson theory
of nuclear forces, and its numerical value is close to 2 fm2 . The form factor fs (r) has the Saxon-Woods
form (2.2) and usually a rather smaller value of the radius parameter. There is very little evidence
favouring the inclusion of the imaginary spin-orbit potential, so it is usually omitted. This gives the
nine-parameter optical potential.
Since the optical potential represents all the nucleon-nucleon interactions between the incident
particle and the target nucleus it is possible to estimate nucleon optical potentials by summing the
nucleon-nucleon interactions over the nuclear density distribution (r), giving
A
X Z
V (r) = v(jr , ri j) (r0 )v(jr , r0 j)dr0 (2:5)
i=1
where v(jr , r0 j) is the eective interaction between the incident nucleon and a nucleon in the nucleus.
This eective interaction can be represented by a Gaussian or a Yukawa form, with adjustable range and
strength. This folded potential is useful only for the real part of the potential, so when it is used it is
supplemented by a phenomenological imaginary part as before. Folded potentials are more frequently
used for alpha-particle and heavy ion scattering, since they often give a better representation of the
surface part of the potential than the phenomenological forms (2.1) to (2.4). An additional renement is
the inclusion of a density-dependent term in the expression for the eective interaction. Double folded
potentials can similarly be obtained for composite projectiles.
These optical potentials treat explicitly only the elastic scattering channel, and all other non-elastic
channels are taken into account in a global way by the imaginary potential that represents the total
ux
2
removed from the elastic channel by all other reactions. This is referred to as the simple optical model
(SOM).
It is also possible to treat explicitly one or more inelastic scattering channels by the coupled-
channels formalism [2]. This consists of a set of coupled equations that can be solved to give both the
elastic and the inelastic scattering in as many channels as required. In addition to the optical potential,
such calculations also require a model of the nuclear excitation, so that the coupling terms can be
evaluated. A coupled-channels analysis is made when there is strong coupling between the elastic channel
and the inelastic channels that lead to the excitation
P of low-lying collective states. Since the computing
time increases approximately as the cube of J (2J + 1), where J is the total angular momentum of the
target states, it is practicable to couple only a small number of channels. The inelastic scattering with
excitation of the higher states can be found using the distorted wave Born approximation.
The optical potential can be regarded as the positive energy part of the mean eld that is expe-
rienced by a nucleon in the region of a nucleus. The negative part is the potential of the simple shell
model that gives the allowed bound states and hence the magic numbers. Examination of this mean eld
over the whole energy range shows that it is essentially continuous as a function of energy. The real part
decreases monotonically with increasing energy, and the imaginary part has a parabolic variation around
the Fermi surface energy. For negative energies, the imaginary potential is connected to the fragmentation
width of the bound states.
In the region of the Fermi surface the real potential depth deviates from the regular behaviour
due to dispersion eects. The real and imaginary pats of the potential are connected by the dispersion
relations that follow from the requirement of causality namely that the scattering wave is not emitted
before the incident wave arrives [4]. Thus, since the optical potential V (K; E ) is a complex analytic
function I
V (K; E ) = 21i VE(K; E 0 ) dE 0 (2:6)
c 0 ,E
where the contour encloses the singularity at E = E 0 .
Putting
V (K; E ) = U (K; E ) + iW (K; E ) (2:7)
and separating into real and imaginary parts gives the dispersion relations
Z 1
U (K; E ) = P W (K; E 0 ) dE 0 (2:8)
,1 E 0 , E
and Z 1 U (K; E 0 )
W (K; E ) = , P dE 0 (2:9)
,1 E 0 , E
where P indicates the principal part of the integral.
These dispersion relations further unify the optical potential by relating the real and imaginary
parts, but there is the practical diculty that the potentials are required over the whole energy range
from minus to plus innity. Several methods have been developed to overcome this diculty. Using the
symmetric behaviour of the imaginary potential about the Fermi surface gives
Z 1
2
U (E; r) = UHF (E; r) + (E , EF ) W (E 0 ; r) dE 0 (2:10)
EF (E 0 , EF )2 + (E , EF )2
where EF is the Fermi energy. This integral has satisfactory convergence. The Hartree-Fock poten-
tial UHF (E; r) has a volume form and falls smoothly and almost linearly with increasing energy. The
imaginary potential W (E 0 ; r) has both volume and surface parts; the volume part is absorbed into the
Hartree-Fock term, while the surface part modulates the potential around the Fermi energy. To illustrate
this, the energy variations of the volume (WV ) and surface (WD ) imaginary potentials for neutrons on
40 Ca, together with the corresponding dispersive additions to the real potential, are shown in Fig.1 [5].
The energy dependence of the volume integrals of the real potential for neutrons and protons on 208 Pb
are shown in Fig.2 [6]. It is apparent that the modulation due to the dispersion eect is appreciable only
at low energies. It is, however, essential to include it in this energy region, as shown in Fig.3 [7].
3
Fig.
40
1. Energy variations of the volume (WV ) and surface (WD ) imaginary potentials for neutrons on
Ca, together with the corresponding dispersive additions to the real potential (Delaroche and Tornow,
[5]).
Fig.
208
2. Energy dependence of the volume integral of the real potential for neutrons and protons on
Pb showing the Fermi surface anomaly (Finlay et al, [6]).
4
Fig. 3. Total cross-section for the interaction of neutrons with 48Ca compared with calculations using
a global optical potential with (full line) and without (dashed line) the dispersion term [7].
3. Optical Model Analyses
In order to use the optical model to calculate the dierential cross-section for elastic scattering,
and if necessary also the polarisation, it is necessary to specify the potentials numerically. In the case of
nucleon scattering, approximate values of the parameters of the potential can be obtained from elementary
physical considerations. The depth of the real potential is similar to that of the simple shell model, which
is known from studies of bound single-particle states to be about 50 MeV. The splitting of these states
gives 4 MeV for the spin-orbit potential. The depth of the imaginary part of the potential can be
estimated semi-classically from the total absorption cross-section to be about 8 MeV. The form factor
parameters can be obtained from the folding model potential (2.5). Since the eective nucleon-nucleon
interaction is short range, it can be approximated by a delta function, and then (2.5) becomes
V (r) / (r): (3:1)
The nuclear density (r) can be represented quite well by the Saxon-Woods form (2.2), with R = 1:25A1=3
and a = 0:6fm.
To analyse a set of experimental data, which is usually a dierential elastic scattering cross-section
and possibly also polarisations, a reaction cross-section and (for neutrons) a total cross-section, these
quantities are calculated from a potential with initial parameter values such as the above, and then
compared with the data. An automatic search routine then varies the parameters systematically until
the optimum t is obtained. Many analyses have been made in this way, and the resulting potentials are
available in the literature or in tabular form [8], so quite accurate starting values can be readily obtained.
One of the great strengths of the optical model is that it is possible to t the elastic scattering by
many nuclei over a range of energies with very similar parameters. This makes it possible to obtain global
potentials that are very useful in reaction analysis, since we often require distorted waves for energies
and target nuclei for which no elastic scattering measurements are available.
To obtain a global potential it is usual to analyse many sets of elastic scattering data, using form
factor parameter xed to average values determined from preliminary analyses where they were allowed
to vary. The resulting potential depths usually have a smooth variation with energy and from nucleus
to nucleus, and these can be parameterised. The dependence on nuclear asymmetry can be included for
nucleons. Since global potentials are seldom required at low energies it is usually not necessary to include
dispersion eects, but these can be included if desired.
Many global parameter sets have been obtained in this way, and in order to select the best it is
very useful to make comparisons with a wide range of data, as has been done for neutrons by Young [9].
Global potentials take no account of the very dierent structures of particular nuclei, and so we
cannot expect them to t the data perfectly; indeed it is rather surprising that they usually t as well as
5
they do. It is always possible to improve the t substantially to any particular set of data by varying the
parameters, and such potentials may be better than global ones in reaction analyses. There are however
serious problems with such analyses due to ambiguities that are discussed in section 4.
The optical potential is applicable when the cross-sections are energy-averaged to remove the
uctuations due to individual states in the compound system. It therefore cannot be applied at low
energies. At high energies relativistic eects must be included. Global optical potentials thus apply over
a limited energy range.
There are quite satisfactory global potentials for neutrons [9] and protons [10], and some for
deuterons [11]. Despite sustained eorts there are none for helions and alpha-particles. This may be due
to the greater sensitivity of the scattering to nuclear structure, particularly in the surface region.
An essential preliminary to any optical model analysis is to verify that the computer program is
correct. This may be done by comparing some trial calculations with the results published in program
intercomparisons [12].
4. The Use of Optical Model Potentials in Reaction Analyses
Most quantum-mechanical analyses of nuclear reaction data use distorted waves generated by opti-
cal potentials tted to the appropriate elastic scattering data or obtained from a global parametrisation.
It is however possible to nd many optical potentials that t the elastic scattering data equally well,
and these frequently give substantially dierent distorted waves and hence dierent nuclear reaction
cross-sections. This is a serious problem that deserves more attention than it has so far received.
The source of the diculty is that the optical potential describes with rather few parameters the
elastic scattering data that is exactly specied by a much larger number of S -matrix elements. It is
adjusted to give the best overall t to elastic scattering, but it may not be the best for reaction analyses.
It may well be that the elastic scattering is more sensitive to one set of matrix elements and the reaction
to another set.
One way to overcome this diculty is to require the optical potential to t selected reaction data
as well as elastic scattering data. Another possibility is to extract the S -matrix elements from the
experimental data using inversion techniques, but these are not without their diculties.
6
Fig. 4. The potentials obtained by inversion and by optical model analysis of the elastic scattering of
334 MeV tritons and 72 MeV 3 He by 14 C. The full lines show the real parts and the dashed lines the
imaginary parts of the central potential V0 and the isospin potential Viso [16].
As an example, Fig. 4 shows two quite dierent potentials obtained by optical model analysis and
by inversion that both give excellent ts to the elastic scattering of 334 MeV tritons by 14 C and 72 MeV,
3 He by 14 C.
Another diculty concerning the distorted waves is that the optical model is determined by t-
ting the asymptotic values of the partial waves whereas the cross-section of a reaction depends on the
wavefunction throughout the nucleus. In addition the eect of the non-locality of the potential should
be considered.
Since the reaction cross-sections are the sum of the cross-sections in all the reaction channels, they
may provide useful additional information. Many measurements of reaction cross-sections have been
made, and the results usually agree reasonably well with those calculated from optical potentials tted to
the elastic scattering cross-sections. Reaction cross-sections are however rather dicult to measure, and
until recently the values obtained were not accurate enough to restrict the choice of optical potential.
Recent work has however given more accurate values that may enable this to be done. Thus
Auce et al [13] have measured the cross-sections of the scattering of 75{190 MeV alpha-particles by
nuclei from 12 C to 208 Pb. The experimental reaction cross-sections were found to be 10{20% less than
those calculated from optical potentials tted to the elastic scattering cross-sections. This also could be
attributed to dierent sensitivities to dierent sets of S -matrix elements. However there is apparently
some diculty with these measurements that has yet to be resolved.
Very recently, some accurate measurements by Yamaya et al [14] of the forward glory scattering
of 12 C, 13 C, 15 N and 16 O by 28 Si at energies about twice the Coulomb barrier have been analysed to
give values of the reaction cross-section that are more accurate than those obtained from optical model
analyses, but consistent with them.
7
An ingenious method to extract reaction cross-sections from elastic scattering data has been pro-
posed by Masaki and Aoki [15], but this has not yet been suciently tested.
It may well be that the only way to obtain potentials free from these ambiguities is to calculate
them from the nucleon-nucleon potential. As an example of such a calculation, Karantaghitic et al [16]
have used the Paris potential to calculate the G-matrix elements and hence, using the local density
approximation, the dierential cross-sections and analysing powers. These calculations, however, are
inherently complicated, and the potentials obtained are very non-local. It is thus not possible to relate
them to the corresponding optical potentials.
5. Reactions at energies up to 200 MeV
The optical model can be used up to about 200 MeV without appreciable diculties due to rela-
tivistic eects. There is some evidence that the form factors used at lower energies are no longer adequate,
and there are few global potentials that extend to such high energies.
Fig. 5. The energy spectrum of neutrons inelastically scattered from 90Zr at 14.1 MeV compared with
MSD (thin solid curve), MSC (thin dashed curve), CN (thin dotted curve), secondary compound nucleon
emission (CN2) (thin long-dashed curve) and collective cross-sections calculated as described in the text.
The collective excitations contribute to the continuum and include the 2+ (thick dotted curve), 3, (thick
long-dashed curve), 4+ (thick dot-dashed curve, 5, (thin dot-dashed curve) states and the LEOR (thick
long dashed curve). The thick full curve is the incoherent sum of all these cross-sections [17].
In many nuclear reactions there are many reaction channels and multistep processes contribute
strongly to the measured cross-sections. As an illustration of this complexity, Fig.5 shows the angle-
integrated cross-section for the inelastic scattering of 14 MeV neutrons by 93 Nb. At the higher outgoing
energies, corresponding to the lower excitation energies, there are resolved peaks corresponding to the low-
lying collective states; these cross-sections may also be calculated using the coupled-channels formalism.
At the lower outgoing energies there are contributions from the higher multipole resonances, including the
octupole resonance, and from the multistep compound [18] and multistep direct [19] reactions. Finally
at the lowest outgoing energies there are the compound nucleus reactions. It used to be thought that
reactions could take place by just two processes, direct and compound nucleus, and that the experimental
8
data could be reproduced by evaluating these separately and then adding them incoherently. In some
cases this is quite a good approximation, but in others it is seriously inadequate.
Fig. 6. The eective interaction strength V0 as a function of incident energy for nucleons with the
Yukawa form factor with range 1 fm. The line V0 = 31 exp(,0:15E=31) is obtained from the energy
dependence of the optical model [20].
As the incident energy increases, the contributions of the compound nucleus, multistep compound
and multipole resonances fall, until by about 50 MeV only the multistep direct remains, except perhaps at
small outgoing energies. The cross-sections for these reactions may be calculated using the multistep direct
theory with standard optical potentials. The only parameter in this theory is the eective interaction
strength V0 which may be derived from the folding model expression for the real part of the optical
potentials. A comparison between values of V0 calculated in this way with those obtained by analysing
experimental (p; p0 ) cross-sections is given in Fig.6. Many double-dierential (p; p0 ) cross-sections have
been well tted with the multistep direct theory [20], and recently it has been extended to analyse (p; )
and (p, 3 He) cross-sections from 20 to 200 MeV [21].
I thank my colleagues who have kindly permitted me to quote their work.
References
1. P.E. Hodgson, The Optical Model of Elastic Scattering, Clarendon Press, Oxford, 1963; Nuclear Re-
actions and Nuclear Structure, Oxford University Press, 1971; The Nucleon Optical Potential, World
Scientic, 1995.
2. T. Tamura, Rev. Mod. Phys. 37, 241, 1965.
B. Buck, Phys. Rev. 130, 712, 1963.
B. Buck, A.P. Stamp and P.E. Hodgson, Phil. Mag. 8, 1805, 1963.
3. P.E. Hodgson, Contemporary Phys. 31, 295, 1990.
4. C. Mahaux, H. Ng^o and G.R. Satchler, Nucl. Phys. A449, 354, 1986.
5. J.P. Delaroche and W. Tornow, Phys. Lett. B203, 4, 1988.
6. R.W. Finlay, J.R.M. Annand, J.R. Petler and F.S. Dietrich, Phys. Lett. B155, 313, 1985.
7. Su Zong Di and P.E. Hodgson, J. Phys. G14, 1485, 1988.
9
8. C.M. Perey and F.G. Perey, Atom. Data and Nucl. Data Tables 13, 293, 1974; 17, 1, 1976.
9. P.G. Young, Proceedings of the Specialists' Meeting on the Use of the Optical Model for the Calculation
of Neutron Cross-Sections below 20 MeV. OECD, Paris, 127, 1986.
10. F.G. Perey, Phys. Rev. 131, 745, 1963.
G.W. Greenlees, G.J. Pyle and Y.C. Tang, Phys. Rev. 171, 1115, 1968.
11. C.M. Perey and F.G. Perey, Phys. Rev. 132, 755, 1963.
12. P.E. Hodgson and E. Sartori, NEA Data Bank Report{198{U.INDC (NEA), 5, 1985.
13. A. Auce, R.F. Carlson, A.J. Cox, A. Ingemarsson, R. Johansson, P.U. Renberg, U. Sundberg, G. Tibell
and R. Zorro, Phys. Rev. C50, 871, 1994.
14. T. Yamaya. Private communication, 1996.
15. M. Masaki and Y. Aoki, Annual Report of Tandem Accelerator Centre, University of Tsukuba, 20,
1992.
16. K. Amos. Private communication, 1996.
17. P. Demetriou, A. Marcinkowski and P.E. Hodgson, Nucl. Phys. A596, 67, 1996.
18. R. Bonetti, M.B. Chadwick, B.V. Carlson, M.S. Hussein and P.E. Hodgson, Phys. Rep. 202, 171,
1991.
19. R. Bonetti, A.J. Koning, J.M. Akkermans and P.E. Hodgson, Phys. Rep. 247, 1, 1994.
20. E. Gadioli and P.E. Hodgson, Pre-Equilibrium Nuclear Reactions, 1992.
21. H.B. Olaniyi, P. Demetriou and P.E. Hodgson, J. Phys. G21, 361, 1995.
P. Demetriou and P.E. Hodgson, J. Phys. G22, 1811, 1996.
A.A. Cowley, G.J. Arendse, J.W. Koen, W.A. Richter, J.A. Stander, G.F. Steyn, P. Demetriou, P.E.
Hodgson and Y. Watanabe, Phys. Rev. C54, 778, 1996.
A.A. Cowley, G.J. Arendse, G.F. Steyn, J.A. Stander, W.A. Richter, S.S. Dimitrova, P. Demetriou and
P.E. Hodgson. In preparation.
10