EFERENC

ICTP Publication Fo. 12 of 1967

IKTEMATIOUAL ATOMIC ENERGY ACHUCY

IFTERNATIONAL CENTRE FOR THEORETICAL PHYSICS

SUM RULES IN QUANTUM MECHANICS E M* F e r r o i r a * and J . Sesma **

TRIESTE May 1967

* On leave of abeejnae from Cen-fcro Braeileiro de Pesquiaas Fisioae, Rio de Janeiro, Brazil (permanent address). ** Esouela de Fisioa* Univexsidad Central de Venezuela, Caraoas,

ABSTRACT The sum rules usually derived for the squares of the matrix elements of the co-ordinate operator between eigsnstates of a Schrodinger Hamiltonian are here extended to the matrix elements of more general operators taken between the states of any complete set of states. For that purpose extensive use is made of the closure relation for the set of eigenstates of a hermitian operator. applications of the sum rules obtained are made. Several

-1-

SUM RULES Hf QUANTUM MECHANICS 1. INTRODUCTION The purpose if this paper is to review the sum rules of quantum mechanics (of the type of the Thomas-Reiche-Kuhn sum rule), to identify their limitations, and to generalize them in several respects. It will be shown how a wide family of related sum rules for can be written by direct application of the closure relation

the eigenstates of hermitian operators. The generalization will include the use of operators more general than a Sohrodinger Hamiltonian to form the set of baBe states, and the freedom to sum matrix elements of arbitrary operators. The well-known THOMAS-REICHE-KUHN sum rule1^, states that for a one-particle Hamiltonian of the form

H - |_

v(r) ,

(l)

where V(r) is a statio potential (i.e., V(r) commutes with 1?) the following sum rule for the matrix element of the co-ordinate operator x is valid;

2_(Ek - Ej ) j < k | x ! j > j

= h / 2m

(2)

Here | j>and |k> are the eigenstates of the operator H with eigenvalues Ej and E L respectively. The summation symbol means the sum over the In disorete set of states and integration over the continuum spectrum. the literature this rule is derived by notioing that

(3)

so that

(E. - E k )

-2-

M u l t i p l y i n g b o t h members b y < k i x 1 j > * (E - E, ) | <k \x I j

= < j l x l k >

, t h i s becomes

Interchanging the indices j and k and subtracting, we obtain

Summing over k, using the fact that the states | k"> form a oomplete set and the fundamental relation

x p

x "pxx "

the sum rule (2) is obtained.

In thiB derivation of eq. (2) the

essential points are only the form of the commutator in eq. (3), the fact that the eigenstates of H form a complete set, and the fundamental relation (4) If the commutatorQx:, H) fails to be of the form (3), the sum rule (2) may be not valid, i.e., the second member of (2) may be different from the constant t)/2m. This will ocour, for example, in the case of the particle being subject to a velocity-dependent potential, or in the presence of exchange forces 2) in systems of several partioles , We shall see hov the necessary modification in the form of the sum rule appears in a straightforward way if an appropriate form of the sum rule is -written which does not make use of the explioit form of the commutator (3*). The possible usefulness of a sum rule like this derives from the following. Suppose that the sum in the left-hand side has some physioal meaning (for example, an integrated cross-section, or a total transition probability), or enters as a term in an experimentally measurable quantity (such as an average value of some physical quantity) Then by comparing this value with the second member of the sum rule we can tell whether or not the Hamiltonian governing the system is of the

-3-

form (l), or whether the matrix element relevant to the process is aotually of the assumed form. In the oases where a nonstatic interaction has to be introduced, the use of the sum rule allows us to obtain information on the strength, and form of these new terms. This procedure has actually been used to get information on the amount of exohange and velocity-dependent forces present in the nucleon-nucleon interaction . The application of the sum rule (2) is, however, strongly limited by the condition that the operator in the matrix element relevant for the process being studied be of the form of a single co-ordinate x This is of course a strong restriction and has implied that the sum rule could be used for the photo-nuclear processes only in the BO-oalled dipole approximation. Sameeffort was made , We shall show how more to extend the sum rule to matrix elements of higher multipole moments but only approximated forms were obtained. can easily be

general sum rules, referring to matrix elements of more general operators, written (and perhaps applied successfully to problems of physical interest). The sum rule (2) is often called the sum rule for the oscillator strengths, because oscillator strength f, . is defined as the quantity

which appears under the summation symbol. Other sura r u l e s have been written down, with different powers of the difference (E, - E . ) multiplying the square of the matrix elemer That i s , sums of the type

""*

x I i*> I 2

(5)

with p = 0,1,2,3,4 have been explicitly computed for Hamiltonians of the type (l)**'. property For p 0 the sum rule is of course only the closure

-4-

k
of the set of states \ k"> and no assumption of the form of H is implied. For p = 1 we have the sum rule for the osoillator strengths. The derivation of each of these sum rules was made in a rather special way, using repeatedly expressions like (3) and (4). We shall show how a more straightforward derivation of them which has several advantages can " b e presented. of any operator. First, we do not have to be restricted to the coSecond, the form of the hermitian operator H is ordinate operator * but we oan write sum rules for the matrix element left unspecified so that we can easily perform the sum for any given H. This will be shown in some examples Let H be any hermitian operator (not necessarily an operator representing a Hamiltonian, but any one which represents a constant of motion of our system). We indicate by |i> and E . its eigenstates Let A and B be any two operators acting We shall be concerned with and corresponding eigenvalues*

on the Hilbert space of the eigenstates of H. obtaining closed forms for sums of the type

S p (A , B ) m 2 _ (E k - E . ) P < j | A J k > < k | B | j >

(6)

where the sum is extended to all eigenstates of H (whether in the discrete or in the oontinuum spectrum, in this last case the sum means also the appropriate integration over continuum variables). Nothing physically or mathematically new is introduced or obtained when a sum rule of the kind discussed here is derived. The only faot that is used and is responsible for the closed form written for the infinite Bum of matrix elements is the olosure property of the eigenstates of a hermitian operator. If the closure property is applied in a straightforward manner to the proper expression, the derivation of euoh sum rules appears trivial. However, calling attention to this point seems to be necessary since in the literature we often find derivations, using speoifio models or long calculations, of

-5-

expressions which are nothing more than a particular example of the completeness property of a set of states. The closure property of the set of states k implies directly in the sum rule 3 (A , B) -

A I k>

| A B [ J>

(7)

Here we have a sum of the type given in eq. (6) with p 0. 2.

GENERALIZATION OF THE S U M RULE FOR THE OSCILLATOR STRENGTHS. W e now c o n s i d e r the oase p 1 . Let us c a l c u l a t e the average . W e obtain, E, | k ^ , using

v a l u e of the q u a n t i t y

[A,H] . B

i n a s t a t e | ;}>

the c l o s u r e p r o p e r t y and t h e f a c t t h a t H | k ^ -

[A

H] B I

a> - E. <j I A I k>

ZI <J U | k > E < k

k
Thus we have a sum r u l e

B Id >

S1(A,B)

0 I A l k > <fc I B

[A

H] B I j >

(8)

relating a sum of a oertain combination of matrix elements of arbitrary operators over a complete set of states with the average value of a related operator in a single state. Applications of this formula can easily be made to reproduce the results for special cases reported in the literature and to obtain new formulas. For A = B x and for H being the Schrodinger Hamiltonian we obtain in the left-hand side the sum of the oscillator strengths that appear in the Thomas-ReicheKuhn formula (2). A similar sum rulej but not oompletely equivalent to (8), can be obtained by considering the double commutator l[A , H] , Bj .

-6-

Taking the expectation value of this quantity in the state ( j > and introducing in the appropriate places the complete set of intermediate states we obtain

S^A , B) s

Y (\

~ E j ) ( < J U lk> < k l B | j > + ^ f B I k X k U \ i

, H] , B] | 3 > ,

(9)

which is formally more complicated than eq. (9), "but may result in being simpler in applications. This is due to the fact that the double commutator F[A , H] , B J may have a form which is actually simpler than that of the product [A , H) B . If A and B are hermitian operators, and B s A , eq. (9) becomes

- E .) i <j 1 Al k> | 2 - i <j | [j> , H] , A] I o > .

W e consider a few examples. a) One-particle Schrodinger Hamiltonian Let H be the Schrodinger

Hamiltonian for one-partiole subjected to a static potential, as in eq. (l). Let operators A and B be functions of the co-ordinates only, A - f(?) and B - g(^). Then

[ f ( ? ) f H] -

i (Vf ? + ? . Vf)
2 m

(10)

and

H] , g(r)] - A y J . vg

,*

(11)

so that the sum rule beoomes

( \ - E ) ( < j I f 1k > < k l g | j

> + < j | g | k > < k 0> -

| f |d > ) (12)

-7-

In the particular case of f - g = x w e re-obtain eq. ( 2 ) . obtain the eleotrio and magnetio multipole matrix elements. b)

Particular

cases of interest are those in which these functions are such that we

Many-particle systems -Let H be the Schrodinger Hamiltonian

for a system with N particles

S
and let
A a f

<*;> + * E
w "

be suras of functions of the co-ordinates of each of the particles. We then have

0 ' H] - 9~
and

where Vf. r..

and Vg. mean gradients with respect to the co-ordinates

The sum rule then becomes 1

H

s;(f .

B)

- ~r 2 1 O I ^
i

v2i I d > .

(15)

This sum rule i s relevant for the study of the interaction of electro magnetic radiation with the E T electrons of an atom. approximation i t i s taken In the dipole

-8-

L
and then N

which, is a well-known result, c) Non-static potentials A we11-appreciated characteristic

of the Thomas-Reiche-Kuhn sum rule, which ia also a property of the more general form (12), is that the potential itself does not appear in the second member of the sum. electromagnetic field This fact is a consequence of the If interactions with an particular forni of the Hamiltonian H.

with a velocity-dependent potential or exchange

forces, are introduced, the sum may take a different form from those given in eq, (12) or eq,. (2), Let us see what happens if we include in the Sohrodinger Hamiltonian a velocity-dependent potential of the form often used in nuclear physics '

V(r\ ?) - - A ? . j<?)-?
2m
and an eleotromagnetic field. Then

(16)

and
[f

/ ^ ^ ^ n ( ?

l r ) ) ( i s )

-9-

Then the sum rule "becomes

V
W e Bee that in this sum rule the eleotromagnetic field does not appear explicitly in the second member; i t only affects the determination of the eigenstates of H. the summed quantity. d) Dirao Hamiltonian - Let f(r*) and g(~r*) again be two arbitrary For the Dirac Hamiltonian
2

This i s not true of the velocity-dependent potential

(16) which appears in the operator whose expectation value represents

funotions of the ooordinates.

H= > o , . ( ? - el/c)+/3mc + V(r*) we have [f , H] - i h e oL . V f and jf H) , g.j 0. Then the tvo forms of the sum rule give

g
(20)

< j | g | k > < k | f | o >

and

, g) (21)

Interchanging f with g in eq. (2l) and adding to eq. (21) we obtain the sum (20), which is zero* so that we have a general relation

(22)

valid for every eigenstate | j ^ of the Dirac Hamiltonian H, being an arbitrary function of the co-ordinates. As a particular oase of (22) w e have that the current J - c V: et ^ ; in a stationary

-10-

state vanishes when integrated over a l l spaoe.

Eq, (22) can also

be obtained " b y calculating the expectation value of the commutator [P , H] in an eigenstate of H. If f and g are real funotions, the

sum in eq. (2l) is a pure imaginary quantity. The zero obtained in performing the sum in eq. (20) over a l l states is due to the compensating contributions from positive and negative energy s t a t e s . By inserting in eq. (21) projection operators for positive or negative energy states Tt and adding to the B a m e expression with f and g exchanged we get

\-E.)(<o

|f P+|k> < k | g | j >

+ < j

|gP+!k><kif (23)

ittC O

| ( 3 . 7 f ) P+ g

(5-

Vg) P+ f I 3 > .

In the non-relativistic limit P sum becomes

- (mo _ + H)/mo

so that the above

which should be compared to the expression given in eq. (12). e) Integrals involving eigenfunctions - Applications of

mathematical or physical interest can be found for the sum rules, consisting in writing directly the values of certain integrals involving eigenfunctions of a hermitian operator. Let us, for example, consider eq. (12) where the states are the eigenstates of a Schrodinger Hamiltonian with static potential and in the presence of the electromagnetio field. For simplicity, let all eigenfunctions be written and B so as to be real functions. Let us choose f = ^p/^and g a ^0-/^7 where * f p and ( J L are eigenfunctions of H with eigenvalues E respectively. to the case in which I j " > is the eigenstate To avoid zeros in the denominator, we restrict ourselves of H corresponding to

the lowest energy (so that *K has no zeros and f and g are well behaved); we indicate the eigenfunction corresponding to this ground state by *f0 . The formulas obtained can be extended to other states if oare is taken in the definition or interpretation of the expressions written. The

-11-

sum mile eq. (12) then gives

Using the orthogonality property of the wave functions, this becomes

(ty't.)

P t

(24)

If Vjj and V^ are eigenfunctions corresponding to the continuum spectrum with delta function normalization, 6pa, in the above expression must be understood as 6 (~p - "q) where " p * and " q are the momenta of the particle in the states ^ and ' f i . are eigenfunctions

This formula is valid whenever V^ , H'^. and ^ of the same Schrodinger Hamiltonian.

It can be useful in evaluating certain

integrals once the eigenvalues of H are known. Eq. (24) is a consequence of the completeness and orthogonality of the set of eigenfunctions of H. For spherically symmetric states or for one-dimensional problems, it takes particularly simple forms. Physical applications for relations such as eq.. (24) can be of the following kind Suppose the fundamental state and the corresponding eigenvalue are known. Then, given empirically an funotion ^ extracted from the above expression. approximate wave for an excited state, the corresponding eigenvalue can be Conversely, if the eigenvalue is

experimentally or empirically given, and the wave function is parametrized in some way, we oan use eq, (24) to obtain information on the parameters determining the wave function. Of course, local relations involving two wave funotions and the corresponding eigenvalues but not the potential can be easily written} integral forms duch as the above are, however, of much advantage when the knowledge or the description of the wave funotion is only approximate, since then an average over small looal fluctuations is already performed by the integration process. Another formula whioh can be useful when the wave functions are not real oan be obtained in the following way. Putting g f in eq, (12), and ohoosing f ^f'k
we

-12-

jrus

Using the property of orthogonality of the wave functions in the seoond integral of the first member, and performing the sum over k, we get

Again in t h i s formula the meaning of the i n t e g r a l s should be studied in each p a r t i c u l a r case in which the wave function tyi has zeros. 3. S U M RULES W I T H HIGHER P O W E R S OF (Efc - E.) Again l e t H be a hermitian operator eigenstates | j > and eigenvalues E . , and l e t A and B be any operators. to see t h a t Then i t i s easy

I f [ A , H] , H ) B I j > - Y l < 2 I A I k > < k l B | j

* - 2E . so that we obtain a sum rule

S2(A , B) *. ( V ^ A j M l ^ ^ 1 5 ^ = ^irr^' H J' H J B li-> . (26)

2

It is interesting to remark that the second member can also be written

- < 3 I [A , H] . [B , H] l j >

(27)

since t h i s form may be more convenient when doing actual computations

-13-

Let us look at some particular examples. For the Schrodinger Hamiltonian with a static potential and in the presenoe of the electromagnetic field w e obtain, f and g being functions of the oo-ordinates only, that

s?(f , e) - Z ( E k - B f < j | f | k > < k | g l o >

a

T^<j\(v*f *? + ^ ^f)(^g ? + ? Ve)U > (28)

We should note that the eleotromagnetio field appears

where Y - " p * - eA/o.

explicitly in the second member of the above expression. For f - x we obtain a well-known result. By adding the

corresponding expressions for f y and f z we obtain

-B )2f|<oix|k>|

It is interesting to remark that in this last form the electromagnetic) field does not appear explicitly in the second member of the sum. If a velocity-dependent potential is included eo that the Hamiltonian is of the form (16) we obtain

S (f , g) - JL <)\ (9f.(4-AJ) P +

(30)

->

_ , > .

where P - p - A/o. The differences between this expression and eq. (23) must be noted.

-14-

For the Dirac Hamiltonian in the presenoe of an electromagnetic field and with a static potential we obtain

s2(f ,
For f - g m x we get

(3D

(32)

which is different from the result obtained with the Sohrbdinger Hamiltonian. Here again, projection operators can " b e introduced to separate the contributions coming from positive and negative energy states. Now we consider sum rules where higher powers of {E, - E .) appear, that is, sums of the form (6), with p > 2. It ia easy to verify that

S3(A ,

(33)

,
3,(A , B)

H]

H]

, Hj B |
H|
t

[[A , H] , H] ,

B I0> ,

and thus successively. For a power p we have p commutatorss

U , B) -

< 0 I A | lc> < k l B | a >

"
(34)

< aI [

[[A

H)

, II]

H] B I

e thus have a straightforward way to compute a set of sum rules whioh may find several applications. We have only to evaluate the multiple commutator whioh appears in the seoond member and calculate

-15-

its average value in the state f j J> . The sum rules imply limitations in the magnitude of the matrix elements of an operator and may be used to obtain information in their rate of convergence to zero as E , * oo As p increases the sum tends to increase and may diverge. evaluate the expectation value Suppose we want to obtain indication on the asymptotio behaviour of < k I A I j > for very large E, . " H e have to

<0I [

[[A , H] ,

H]

, Hj A | 3 >

using p - 1,2,3,... commutators.

Suppose that for p - s the expectation Since for high

value is infinite, while for p s - 1 it is finite. to k oik , we have that J k expect that f k E,

energies the number of states in an interval of momenta is proportional | < k | A | j >| dk is finite, while we diverges. Then we may say that Efc | <.k 1 A 1 j > | dk
2

< k I A | o> I
(k E, k ) ~ E, k

(k2 E ^ k ) " 1 ^ * with 5j _ > 0, and that |<k U i O > ( 4 ~

, with6j,>0. The quantities 8* and 6 have upper bounds

determined by the functional dependence between E, and k, with k * ^ " E ^ * For a relativistio kinenatics, 6^+ 6% - 1, and the asymptotic behaviour of the matrix elements is determined up to a unit power of the energy. If negative energy states are included in the sum, our argument should be restricted to even values of p, and the indetermination in the dependence with the energy is of two units in the exponent. If the expectation value appearing1 in the second member of eq. (34) is finite for all numbers of commutators (all values of p ) , then the product <j|A|k> < k | B |. j > tends exponentially (or faster) to zero as E, < By multiplying sum rule for power p byotV'p. and summing over p from zero to infinity we get

-16-

< o I [A + fr [A , H] + f f [[A , H] ,
+ + fr[
[[A , H]

H]

t [[[A , H| ,
, H ] +

H] H]

, H]

.] B I o>
(35)

The term within the brackets is the formal development of e Performing in the left-hand side the sum over p first w e get
01

nt

e
This formula is in fact trivially the closure relation

k>

<k| G | 0>

"

<j |F G

k
with F - e"rtH A and G The several sum rules for different powers p are then the terms of the expansion of eq.. (36) in powers of a, . This makes i t clear how the closure property, and only that, generates a l l this family of r u l e s . However they can be useful if the sum concerned has a physical meaning and can be directly measured . Then one can obtain from the sum rule information such as on the structure of the Hamiltonian governing the system, or on the form of the matrix elements responsible for the process being studied, or on the asymptotic behaviour of these matrix elements, and so on. W e oan also write sum rules of the type S (A , B) for non-integer values of p . oc iA.
u

Let us s t a r t from eq. (36), and just for convenience call . A and integrating

Multiplying both Bides of eq., (36) by e"

over A from zero to infinity, we obtain

-17-

Interchanging the order

of the summation and the integration prooesses

in the left-hand side, and performing the integration over A , we get (if the real part of p i s not positive the integral e defined as a distribution)

|k><k|B

The limit K~* 0 can be taken when eo wished. expectation value .

W e thus obtain a

sum S (A , B) for any complex p, expressed in closed form as a single The expression i s only formal unless w e are able to
a

evaluate explicitly the quantity A(A) by the Harailtonian H.

<?1

A e ~

- This means knowing

the Heisenberg equation of motion for the operator A in a system governed

Sums with functions of E, and E . which are not just powers of (E, - E .) can also be performed. Thus we have

U|k>

<k|B | j>

|[

J|
(38)

where ^ ( H ) is any function of the operator H. matrix elements.

Relations of this kind

can also be used to obtain information on the asymptotic behaviour of

Other olasses of sum rules, obtained by differentiation of an eigenvalue equation with respect to parameters have been obtained 7) recently ; -18-

4.

APPLICATION OF THE S U M RULSS : THE INTEGRATED CROSS-SECTIOIT FOR THE PHOTO-NUCLEAR EFFECT

The Thomas-Reiche-Kuhn sum rule ( 2 ) f o r the matrix elements of a co-ordinate x has found a p p l i c a t i o n i n the c a l c u l a t i o n of t h e c r o s s - s e c t i o n f o r the photo-nuclear effeot i n t e g r a t e d over a l l energies 3) 8") of the incident photon 'and in other problems ' . In the nuclear photoeffeot, a photon inoident on a nuoleus is absorbed, and a proton is ejected. It is the nuclear analogue of the atomic photo-electric effect. In the usual treatment of the photo-nuclear effect the matrix element responsible for the procesB in first-order perturbation theory is subjected to the so-called dipole approximation, which consists in considering the extension of the nuclear wave-functions small compared with the wavelength of the incident photon. The purpose of this approximation is to obtain that only matrix elements of a single coordinate, such as those appearing in the Thomas-Reiche-Kuhn sum rule, appear in the expressions to he calculated. This approximation can be questioned, however, since in the integrated cross-section we use matrix elements corresponding to photons of high energy. Since w e now know how to write sum rules for more general matrix elements, we are able to avoid the dipole approximation in this calculation. In what follows w e show how this can be done. For simplicity and to stress the essential points w e consider the case of a single charged particle bound to a centre of static (non-velocity dependent) forces. The partiole is considered as heavy, so that nonrelativistic kinematics is used. W e assume that the centre of forces admits only one bound state so that we do not have to consider transitions without emission of the charged particle. A photon of frequency CJ , propagation vector k along the z-axis, and polarization vector along the x-axis, collides with the heavy partiole bound by the centre of forces in a state described by the wave funotion |p . . In the final state the partiole in a positive energy state r f whiph asymptotically has a plane wave part normalized in a volume V.

-19-

In first-order perturbation theory the differential crosssection for the process is

(39)
where v is the velocity of the emitted particle For unpolarized photons we have to average this expression over the x and y directions of polarization. This differential cross-seotion is in general diffioult to evaluate due to the lack of knowledge about the wave functions in the final state. For our normalization of the final wave functions the number of states is
V J SI d. (~t i*J )

(IV)' t*

(IT, ft

where p m v i s the momentum of the emitted particle . The total oross-section integrated over the photon energy is then

(40)
_ <?

where ^

means sum over a l l possible final s t a t e s . u>= E^ - E. is the difference between the energies of the

f
final and initial states. Since l/(E - E.) cannot be expanded in

powers of (E - E , ) , we do not have a sum rule to be applied directly to the above expression. However we can rely on properties of the matrix element to perform the sum in an approximate way. The ground state wave function %' occupies a limited region of space, so that the integral over d& can be restricted to a finite region of radius R. On the other hand, the matrix elements are likely to tend rapidly to zero as k inoreases.

-20-

I t may then be that we can neglect, both in the sum over states
and in the integration in tlie space volume, the contributions coming from the regions -which do not satisfy the condition k r << 1. few terms, The dipole approximation consists in putting simply exp(i Then the matrix element can be approximated by In these conditions w e can expand exp(i k z) in power series and keep only a

Js . a J */ [,] % It = - =
In this approximation the integrated oross-section becomes

. a J */

(42)

The sum here indicated oan be performed.

For pure s t a t i c potentials

in the Hamiltonian i t gives the olassical Thomas-Reiohe-Kuhn r e s u l t . If exchange or velocity-dependent forces are present, the s u m will be modified accordingly, and this has in fact been used for a test 2) i) of the presence of these terms in the Hamiltonian ' ' ' . The dipole approximation was the only contribution to crlrtt considered in previous calculations, because of the limitations in the sum rules available, W e now show how other terms of the expansion of exp( i k z) can be inoluded . The next term is usually referred to as the quadrupole The contribution of the interference terms between contribution.

dipole and quadrupole matrix elements to the integrated cross-section i s

-21-

< '

(43)
We should notioe that a factor u > - E . - E . coming from the dipole We

term and another from the k = w/c appearing in the quadrupole matrix element were able to eliminate (B - E.) from the denominator. thus obtained an expression which fits into our sum rule (8) and can be put in closed form. For every other term in the expansion we have that the matrix elements will bring a factor OJ raised to a power larger than two, and all sums to be effected will fall into the general form (34). For the contribution arising purely from the quadrupole matrix element we have

(44)

The sum can again be performed by using eq. (8). Up to quadrupole terms we then have for the integrated crosssection, using the sum rules

fT-

/ <

f T

\ I

-1*2-

f"

-m.

*+*

"7

^'

TJ

-22-

Knowledge of the Hamiltonian and the ground state wave function is neoessary to evaluate this expression-

ACKNOWLEDGMENTS One of the authors (E.P.) is grateful to Professors ATxLus Salam and P . Budini and to the IAEA for the hospitality extended to him at the International Centre for Theoretical Physics, Trieste. Useful discussions with Dr. C O . Bollini and Dr. J.J, Giambiagi are acknowledged with pleasureo

-23-

REFERENCES 1) W. THOMAS, tfaturwiss. _13_, 627 (1925)} W. KUHN, Z. Physik 3 3 , 408 (1925); P . EEICHE and V. THOMAS, Z. Physik 1&, 510 (1925). V. FOCK, Z. Fhysik _89, 744 (1934); E . FEENBERG, P h y s . Rev. 4 , 328 (1936); K. WAT, P h y s . Rev. Jjl, 552 (1937 >5 A . J . F . SIEGERT, Phye. Rev. 2 , 787 (1937). 3) J . S . LEVINGER and H.A. BETHE, Phys . Rev. 2 , 115 (1950); 0 . ROJO and L . S . LEVINGER, Phye. Rev. l ^ , 2177 (1961), R.G. SACHS and N. AUSTERN, Phys Rev. 8l_, 705 (1951 )j J J I . BLATT and V. UEISSKOPF, " T h e o r e t i c a l ITuclear P h y s i o s " , p . 643, John Wiley & Sons (1952). 5) H.A. BETHE, "Intermediate Quantum Mechanics", p . 147, W.A. Benjamin I n c . , (1964)6) M. RAZAVY, G. FIELD and J . S . LEVItfGER, Phys. Rev. 125> 269 (1962). D.J. MORGAN and P.T. LAHDSBERG, Proo. Phys. Soc. jS6, 261 (1965); J. Math. Phys. 1, 2271 (1966). 8) E . MERZBACHER, "Quantum Mechanics", p . 443, John Wiley & Sons, (1961),

2)

4)

7)

-24-