PHYSICAL REVIEW D VOLUME 23, NUMBER 2 15 J A N U A R Y 1981

Gravitational/electromagnetic conversion scattering on fixed charges in the Born approximation

R.A. Breuer,* M. Rosenbaum, and M. P. Ryan, Jr.
Centro de Estudios Nucleares, Universidad Nacional Autonoma de Mexico, Mexico

Richard A. Matzner
Centerfor Relativity, University of Texas, Austin, Texas 78712
(Received 30 July 1980)
In the presence of nonzero background electromagnetic and gravitational fields, perturbations in these fields are
coupled, via the Einstein-Maxwell equations. We investigate the differential cross section fbr the conversion
scattering of electromagnetic to gravitational radiation (and vice versa) on fixed charged masses. Our procedure uses
a Born approximation at the level of accuracy in which the horizon structure of putative black holes is lost; our fixed
charges therefore represent charged black holes at this order. We find conversion differential cross sections of the
order of GeZ(G is Newton's gravitational constant, e is the value of the charge), with angular polarization factors,
and with a forward-divergent behavior a ~ i n - ~ ( 0 / 2 )The
. divergence is shown to arise not only in the Born
approximation, but to be implicit in the known behavior of the large-angular-momentum phase shifts for the partial-
wave spherical modes.

only at the lowest nonvanishing order in the grav-

itational constant. This approach, applied here
It was pointed out by Choquet-~ruhat,' G e r l a ~ h , ~ to the classical electromagnetic and gravitational
and Sibgatullin3 that in electrovac solutions (so- wave fields, is the exact computational and con-
lutions to the coupled Maxwell-Einsten equations), ceptual analog of the Born approximation" in the
gravitational and electromagnetic perturbations quantum-mechanical scattering problem,
a r e necessarily coupled. This idea was sub- In addition to the conversion scattering, there
sequently developed by Moncrief4 in the investiga- i s also ordinary, nonconversion scattering, in
tion of perturbations of Reissner-Nordstriim black which, f o r instance, an incident electromagnetic
holes. Chitre, P r i c e , and Sandberg5 developed a wave is scattered ( a s an electromagnetic wave) by
corresponding Newman-Penrose6 formalism and the black hole. Such scattering has been dis-
investigated perturbations due to sources in orbit cussed for various wave fields on uncharged black
around charged black holes. holes in a n angular decomposition by ~ a n c h e z , "
An important aspect to consider in such situa- Matzner,13 Chrzanowski et a1. ,I4 Matzner and
tions is the existence of a conversion c r o s s s e c - Ryan,15 Handler and ~ a t z n e r , " and in the Reiss-
tion. That i s , suppose a n electromagnetic plane ner-Nordstrom case by Gunter.'O Furthermore,
wave is incident on a charged mass. How much Peters,17 Westervelt,la and Sanchezig have cal-
of that flux is converted to outgoing gravitational culatea Born-approximation nonconversion scat-
radiation? The inverse process can also be cal- tering from "Schwarzschild black holes."
culated. Olson and Unruh7 used Moncrief's de- The remarkably wide range of applicability of
composition into modes labeled by angular mo- the Born approximation in such problems a s spin-
mentum I , in the JWKB approximation to calculate l e s s nonrelativistic electron scattering, where it
conversion r a t e s for high-frequency (w >> l/rn )waves reproduces the Rutherford and exact quantum-
scattering in the Reissner-Nordstrom geometry mechanical Coulomb c r o s s sections, has been the
( m a s s m , charge e ) . Matzners numerically ob- subject of much d i s c ~ s s i o n . ~ In
' general situa-
-
tained the w 0, I = 2 conversion c r o s s section,
and Fabbrig obtained the conversion phase shifts
tions, however, it will not give the exact c r o s s
section, although the resulting calculation will
( ~ b )for, o- 0. Recently, Gunterl0 has done a still reflect some of the structure of the exact r e -
numerical study of the conversion process for a sult.
range of w , I , and e / m . In fact, we shall s e e that a consistent Born ap-
In this paper we consider the Reissner-Nord- proximation to scattering from a black hole loses
strijm conversion-scattering problem from a dif- the horizon structure of the black hole, s o it is
ferent viewpoint. Rather than investigating the m o r e accurate to r e f e r to the scattering center a s
partial-wave formulation, a s has been done in a "fixed mass" m . Similarly, rather than referring
previous discussions, we consider the full, to Born-approximation conversion on a "charged
summed conversion differential c r o s s section, but black hole" we use the t e r m "fixed charge" e .

305 O 198 1 The American Physical Society

 BREUER, ROSENBAUM, RYAN, AND MATZNER 23
-
11. CONVERSION SCAlTERING: structure completely disappears from the problem.
ELECTROMAGNETIC TO GRAVITATIONAL The calculation then proceeds in a straightfor-
ward manner. The background electric field i s
In order to display some of the fe&tures of the (using rectangular coordinates in Minkowski
approximation, consider the equation for a metric space)
perturbation propagating on a Reissner-Nordstrijm
0 F o k = e ~ k / ~Y3 ~, = x ~ + Y ~ + ~ ~ . (2.4)
backgroundz0:
- We choose a n incident plane electromagnetic
h,,,, I " = -16nG6TaE, (2.1)
wave of specific polarization moving in the posi-
where gaB= O g a B + h a E i s the perturbed metric tive 2 direction [2i =,;( 3, ;)I :
i s the background metric, all indices a r e raised 62= EgeiLd(t-2)9
and lowered using this metric) and (2.5)
- 6Ei 'F,~
h a5 = ha 6 -'O 2 gaEho0* (2.2)
(there i s also a complementary magnetic field).
A slash denotes a covariant derivative in the back-
The energy flux for this wave i s
ground, and 6T,, i s typically nonzero because
there i s a concomitant electromagnetic perturba-
tion [Eq. (3.1) below]. In Eq. (2.1) the gauge
j i a B , , = O , which i s consistent because of the con-
servation of the s t r e s s tensor, has been assumed.
In the spirit of the Born approximation, we con-
The conversion c r o s s section i s given a s the en-
s i d e r the lowest nonvanishing interaction in cal-
ergy flux of the outgoing gravitational waves p e r
culating conversion scattering. If we momentari-
steradian, divided by the energy flux of the in-
ly ignore the source in Eq. (2.1), we s e e that the
cident electromagnetic wave [Eq. (2.5)].
left-hand side r e f e r s only to the m e t r i c perturba-
Equation (2.3) gives
tion ha,, and is therefore a description of metric
waves moving in a fixed Reissner-Nordstrijm C?izg= - 16nG6Tz,
background. The Born approximation for non- = - 4GEexy 2 e ' ~ ' t - " . (2.7)
conversion scattering of gravitational waves
would be obtained by linearizing this background Furthermore, it is straightforward to show that
to v u v +o ( G ) ( ~ / Y + e2/r2)+ (keeping first-order
6T,, = 6Tz,,
t e r m s in G). This i s the technique used by peters17
and Westervelt18 to calculate the gravitational- 6Tx,i- - bT,, ,
wave-scattering c r o s s section of an uncharged
6Tx,= - (Y/x)~T,,,
mass.
We, however, a r e interested in conversion 6Tx,= - ( z / x ) b ~ , , ,
scattering, which i s due only to the perturbed
6T,,= 0 .
source in Eq. (2.1). This source contains t e r m s
such a s ' F , ~ ' F ~ P , where IFaBi s the electromag- We need not evaluate 6To,because only spatial
netic field tensor perturbation while OF,, i s the components h t j a r e needed to evaluate the trans-
(Coulomb) background electric field of the Reiss- verse-traceless outgoing gravitational wave. The
ner-Nordstrsm solution. Conversion scattering
occurs when an electromagnetic wave i s incident -.-.
Green's function G for Eq. (2.7) satisfies
( v 2 +w 2 ) ~ ( F; I,) = -4r 6(r - r f ), (2.9)
upon the fixed charge. The metric perturbation
which i s generated is proportional to Ge. F o r from which we get
this reason, the deviation of the background metric
from flatness can be ignored in the subsequent
propagation of the gravitational wave generated.
(These metric deviations introduce corrections of where
next order in G t o a t e r m already of f i r s t order.)
r w = w(f
Hence, in the Born approximation one solves the
flat-space propagation of a tensor field z,, with with h u t t h e direction to the observation point, and
source bT,,, where we impose outgoing boundary conditions on
the scattered field. The r sign in Eq. (2.10) means
that we have only taken the lowest-order t e r m of
where = 17"" aa6 and bT,, i s the linearized elec- the Green's function. Equation (2.10) is easily
tromagnetic stress-energy tensor, and the horizon evaluated by aligning the pole with the direction of
23
- G R A V I T A T I O N A L / E L E C T R O M A G N E T I C C O N V E R S I O N ... 307

--t
Aw; one introduces new coordinates x u , y f f , z" s o An interesting alternative evaluation of the inte-
aligned and finds that g r a l s i s given in the Appendix.
F o r the outgoing flux of gravitational radiation
X I = COS? sin3zu + , we use I s a a c ~ o n ' s ' prescription:
~
y f = s i n $ s i n ~ z " + . . ., (2.11)
z f= C O S ~ Z "+ . . , T,":& = ~ g ' ? ~ ~ ~
where the ellipses involve quantities independent - 3271G
-- onmi,ej d,
Pout ( h ~ ~ o h ~ ~ o ) s u m (2.14)
of 2". By symmetry about hw, only t e r m s in x'
proportional to? "-contribute to the integral (2.10).
In Eq. (2.11), (8, @)a r e the angles coordinatizing where h:: is the transverse-traceless part of hi,,
in the original frame; they a r e related by and the average is over several cycles. In our
case one has
&=q3+7r,
(2.12)
e=$(6+n) ( h ~ l o h ~ ~w2o')lhT:h:j'
=- / .
to the angles 0 , cp giving the outgoing direction r.
We obtain Because gauge transformations

- coscp cot-26 - sincp cot-26

1
1 can never introduce nor modify transverse-trace-
l e s s t e r m s , a simple algebraic projection oper-
- sincp cot-2e coscp cot-
2
Q ator based on the direction ,; allows extraction
of the transverse-tracelezs part. Furthermore,
1 0 O
8
coscpcot-
2
. because of the definition h i j in t e r m s of hi, [Eq.
(2.2)], we s e e h i j = h i j when both a r e traceless.
(2.13) The symmetric projector is

P=
I 1 - sin28cos2cp
- sin2$sincp coscp
- sin6 cos6 coscp
- sin2$sincp coscp - sin$ cos6 coscp
1 - sin2$sin2cp - sin8 cos8 sincp
- sine case sincp sin26
1
r - -

and The expression calculated from Eq. (2.18) ap-

i J t r (p~e a~r s~ much
h ~ ~ = ~ ~ ~ = P i l ~ l , ~ , j - $ P(2.17) ) . more complicated than (2.20) a t f i r s t
sight. Only repeated use of trigonometric identi-
We calculate the quantity t i e s gives the final expression.

111. CONVERSION SCATTERING:

GRAVITATIONAL TO ELECTROMAGNETIC
(using the cyclic property of the trace and the
idempotent property of projectors). The c r o s s The production of outgoing electromagnetic radi-
section i s defined a s ation when gravitational radiation i s incident on
a fixed charge i s governed by the Maxwell equa-
tion''
Born
g"B~ ,,,,,-RUaA,=O, (3.1)
Multiplying z,, a s given by Eq. (2.13) by the pro-
where R u e is the Ricci tensor, A , i s the total
jection matrix (P,,) from Eq. (2.16) and calcula-
electromagnetic potential, and the semicolon
ting $ w2 IhTjThTT1 by means of (2.18), we obtain means covariant differentiation using the full
[with T$f given by Eq. (2.6)] metric. In Eq. (3.1) we assume the Lorentz guage
da
- = cos2(i8) (1- cOS'cp) . A';, = 0. We will s e e below that it i s consistent
d n lw t o r w sin2(*$) to make this choice for the perturbed potential.
Born
(2.20) The approach here i s , a s in the preceding sec-
308 BREUER, ROSENBAUM, RYAN, AND MATZNER -
23

tion, to take only the lowest-order t e r m s that ~ A = h o P 0 A , , 4 1~) +O A ~ ~ o P ~ r A ( h T u ,-hoY,

o + h T),P
To,lL
enter conversion scattering and to ignore a l l
+ v ~ ~ v ~ ~ ( ~-hoY,T)OACt.p'
T ~ ~ , ~ + ~ T ~(3.6)
, ~
others. Expand
The background has only a static electrlc fleld
A, =OA, +'A,, (3.2) s o we can take it in a Lorentz gauge where in
where 'A, i s the background potential and 'A, the addition on19 'Ao i s nonzero. Then we derive from
perturbation. The metric and covariant derivatives (3.5) and (3.6) the equations
in (3.1) must also be expanded in t e r m s of the
u1Ao =h(oAo,xr- O ~ o , y y ) e s ~, t - e )
metric perturbation ha,. We will need to examine
each t e r m in the expansion of (3.1) to check the o'A, =iwhOAo,,eiqt-Z),
o r d e r at which it affects the conversion scattering. ulA,= -iwh0A0 ei~w(t-Z) , (3.7)
Y
This i s easily carried out, even though the p e r - I

turbed Eq. (3.1) does not fall a s neatly into con- O'A, = 0 .
version and nonconversion p a r t s a s the equation It i s straightforward to show that the t e r m s on the
for metric perturbations did. The exercise af- right-hand side of (3.7) a r e the components of a
fords an opportunity to apply physical reasoning
conserved current; hence the assumed Lorentz
t o the estimates. gauge is consistent for theperturbation 'A,. As
The 0th-order version of (3.1) i s of course
in Sec. 11, the background electric field i s taken
satisfied. Furthermore, we drop t e r m s of the a s 'Ao = - e/r; then E , = exi/r3.
form Or;,, (background connection coefficients)
The solution of Eqs. (3.7)-again only the spatial
multiplied by h,, o r by 'A,. The reason i s that
O r & - Gm /r2, and h,, and 'A, a r e already small
components A i a r e needed-is obtained by an iden-
tical procedure to the one used to solve Eq. (2.7).
quantities. (Some such t e r m s contribute to non-
We find, with outgoing boundary conditions,
conversion scattering.) On the other hand, prod-
ucts such a s '~;,OA, do survive, since they a r e of
only f i r s t order. Making simplifications of this
type, we a r r i v e a t an intermediate reduced form
for the perturbed Maxwell equations:
O=-hop OA,,o,p+ 0go p CAp.o,p
1
- (1r hpOo
Ah),p
'A,=O.
1 " OA
-p a,o OAP,aI The c r o s s section is obtained by evaluating the
- R,
0 a1
Aa+'gm6 ' R , ~ O A , - ~ ~ ~ (3.3) outgoing (transverse) energy flux in the transverse
waves:
The quantity OR," 'A. i s a nonconversion-scattering
1
t e r m , and will b e dropped. On the other hand, the = q:; = -( E +~~ 2 ) ;
8n
l a s t t e r m , h a 6 OR,,& 'Ab, i s a conversion-scattering
t e r m , but i s of higher o r d e r in G because OR,&
= o ( G ~ ' ) , h a 6 =o(G), 'A& = 0(e) so this t e r m i s a
factor Ge smaller than other t e r m s in (3.3) and
we will drop it. The remaining Ricci tensor t e r m where
0 a61R
g ,B 0Ap will be dropped a s well. In flat space
the vacuum equation for a metric perturbation is
'R,~ -
= O . Here there i s a source, 'R o(G~'A). r
2
cosq cot-(1
e - sin2@cos2q)
Hence this t e r m also has an explicit additional
power of Geand will be dropped.
0
Notice a l s o that the combination =heeiwct-n) - sincp cot-(1+
2
sin2@cos2cp)
2r
o opl a
.-
-10 op 6
g rop-zg (hao,Q+hap,o-hop,a), (3.4) sine cbs0 cot- cos2cp
2
which appears in (Y.3), can be eliminated if a
transverse-traceless gauge i s chosen for the (3.10)
m e t r i c perturbation. Thus, by choosing a s a n in- and, therefore,
cident plane gravitational wave
h = - h =heiw(t-x)
XX YY (3.5)
( a l l other components a r e zero)-which i s trans-
v e r s e and traceless-(3.4) vanishes.
The wave equation for 'A, i s then given by
-
23 GRAVITATIONALIELECTROMAGNETIC CONVERSION ... 309

T h e f l u x due to the incident gravitational plane section.

wave of Eq. (3.5) is The small-angle scattering a r i s e s f r o m distant
GW* - pu$ encounters, and it is f o r these that one expects
TO^ e i - 00 the assumptions of the Born method (planeness of
the incident wave, linearization of the background
m e t r i c ) will b e m o s t accurate. T h e r e i s c o r -
roborating evidence f o r t h i s r e s u l t f r o m the behav-
Hence, the Born-approximation c r o s s section i s
ior of the phase shifts a s determined by FabbrLg
In Moncrief's4 angular decomposition, t h e r e a r e
two independently propagating modes, each of
Born which is a (different) linear combination of the
gravitational and electromagnetic waves. A s
noted by Matzner8 and F a b b ~ - i , it ~ i s the differ-
ence (h6), between the phase shifts f o r these two
modes which d e t e r m i n e s the conversion s c a t t e r -
The similarity between the two c r o s s sections
ing. T h i s m e a n s that the logarithmic phase shift
(3.13) and (2.20) is due t o the fact that the electro-
typical of a long-range (e.g., Coulomb) scattering
magnetic and gravitational potentials a r e almost
f o r c e , and typical of waves in the gravitational
t h e s a m e within the approximation considered.
field of a c e n t r a l m a s s , cancel out in the calcula-
T h e differences in the c r o s s sections a r e due to
tion of the conversion scattering, a fact which
t h e different tensorial nature (spin s) of the fields
lends credence t o our calculation [cf. Eq. (2.5)]
involved. They show up in the azimuthal depen-
dence of the c r o s s section, which is cosz(sro), - which ignores the logarithmic t e r m s entirely.
More interesting is the phase-shift behavior
w h e r e s i s the spin of the incident wave.
It is not obvious that t h i s similarity would exist
,-
( h b ) 2 e w ~ ~ / ~ Although
/l.
-
discovered by Fabbrig f o r 1 m, l/w m. He found -
this quantity d e c r e a s e s
in the full relativistic problem because of the
with I , it does s o sufficiently slowly that the c r o s s
a s y m m e t r y introduced by ingoing-wave boundary
section diverges a s -sin-'(+@) as we now show.
conditions a t the horizon. P e r h a p s a reciprocity
[Matzner8 noted the falloff with I of (h6),, but in-
theorem of the type given by Sanchezlg could s t i l l
c o r r e c t l y concluded that large-1 t e r m s could b e
b e established f o r the full conversion-scattering
ignored.].
problem.
Matzner's8 r e s u l t s can b e used t o show that
~ a b b r i ' s( ~ 6produces
) ~ a conversion amplitude
IV. DISCUSSION: COMPARISON WITH OTHER
CALCULATIONS
-
(gravitational radiation electromagnetic r a d i a -
tion, h e r e with c i r c u l a r incident polarization for
These c r o s s sections w e r e derived f o r plane simplicity):
incident polarization; cp is a n angle f r o m the
direction of polarization. C r o s s sections f o r
c i r c u l a r incident radiation a r e obtained by simply
averaging over p. where
The appearance of the coefficient Ge2 i s con-

- -1
~ Iz ( @ ) ~=z -
iI
@
y;(@, cp)
sistent with our notation which h a s the electrostatic
potential A, e / r ( a s in electrostatic c g s units, is a spin-weighted s p h e r i c a l harmonic. We now
f o r instance). Then, the potential energy is given u s e the relationz2 (cf. Matzner and Ryanz3 f o r a
b y e2/r. Hence e2 h a s dimensions energy x d i s - s i m i l a r calculation)
tance.
Ignoring f a c t o r s of the speed of light, then,
since G x e n e r g y - length, Gez- (length)2. F o r the
charge on a n electron, Ge2- cm2. to write the large-1 relation
P e r h a p s the m o s t s u r p r i s i n g feature of the r e -
s u l t s (2.20) and (3.13) i s the appearance of the
forward-divergent t e r m [ ~ i n ~ ( ; @ ) in ] - ~the c r o s s
section. (This differs f r o m the Coulomb forward
divergence [sin4($@)]-'because the integrand in
Eq. (2.10) i s not the s a m e as the integrand in the Hence, if we concentrate on large-1 t e r m s which
Coulomb problem.) T h i s forward divergence i s dominate the singular behavior of the c r o s s s e c -
one feature we definitely anticipate in any exact tion, we find that the conversion amplitude solves
t r e a t m e n t (summed over I ) of the conversion c r o s s Eq. (4.4) whose s o u r c e i s proportional toz2
310 BREUER, ROSENBAUM, RYAN, AND MATZNER -
23

6(cos9 - 1). The solution of Eq. (4.4) i s Hence the small-angle c r o s s section i s proportion-
a l to sin2(+@),although of course Eq. (4.6), which
was calculated using only the dominant large-2
p a r t s of the phase shift, i s not exact. Nonethe-
l e s s , there i s agreement between this method and
The constant must be determined by demanding the Born approximation on the forward divergence.
expansion coefficients with the form (4.1). Using
the relation between ,Yy and Jacobi polynomials ACKNOWLEDGMENTS
given by Breuer et al.," and evaluating integrals It i s a pleasure to acknowledge helpful discus-
via Sec. 7.39 of the book by Gradshteyn and sions with Dr. S. Hojman and with J. Futterman.
Ryzhikz5 one obtains k = eJ-6, and This work was supported in part by NSF Grants
Nos. PHY77-07619 and INT78-22553, andby a g r a n t
E Ge2-
cos6($e)
w t sin2(p) * from Consejo Nacional de Ciencias y Tecnologia,
large l No. 955.

APPENDIX

The calculation of the integral in Eq. (2.10) may alternately be carried out a s follows. If the outgoing
direction pouti s given by (sine coscp, sine sincp, cose) then

where %'= cp - cp'. Thus we have to solve

= K d d d B f exp[iwrta(8', cos?')] sin81cos8' , (A2c)

v-2n

where K = ( e ~ / . i r r ) e ' ~ ' ~ - ' ' .

If we expand cos(cp - q') and sin(cp - G')
in (A2) we find that all the integrals involving sin?' vanish be-
cause these integrands a r e odd in the regimes [cp - 27r, cp - T ] and [co - T, 91. For the remaining t e r m s ,
the radial integration gives a 6, functionz6:
-
h =--
sincp -
XY COSp h Z Z ,

-
I
The 6 function part in (A3) can only contribute butes nothing in the middle range.27 Then, from
when (Y = 0 which i s the case for 0 ' E l0/2,71 - 9/21. integration over the outer intervals [O, 8/21 and
But in this range, the corresponding integrals [ T - 8/2, T] we obtain (Ref. 25, formula 2.554) Eq.
vanish. Thus only the principal p a r t s in (A3) con- (2.13) in the text.
tribute. The 8' integration of these also contri-

*Permanent address: Max-Planck-Institut f u r Plasma- 'u. Gerlach, Phys. Rev. Lett.s, 1043 (1974); Phys.
physik, D-8046 Garching 6, Munchen, West Germany. Rev. D z , 2762 (1975).
'Y.Choquet-Bruhat, Colloque de Centre National de 3 ~ R.. Sibgatullin, Zh. Eksp. Tear. Fiz. 3 , 1187 (1974)
Recherche Scientifique, 1973 (CNRS,no. 220 - ondes [Sov. Phys.-JETP 2 , 579 (1974)l.
et radiationes gravitationelles, P a r i s , 1974) p. 85. 4 ~ Moncrief,
. Phys. Rev. D 9 , 2707 (1974); 2 , 1526
23
- GRAVITATIONALIELECTROMAGNETIC CONVERSION ... 31 1

(1975). i 8 ~ .J. Westervelt, Phys. Rev. D3, 2319 (1971).

'D. M. Chitre, R. H. P r i c e , and V. D. Sandberg, Phys. "N. G. Sanchez, J. Math. Phys. 17, 688 (1976); Phys.
Rev. Ds, 747 (1975). Rev. D s , 937 (1977).
6 ~ T. . Newman and R. Penrose, J. Math. Phys. 3, 566 2 0 ~ .W. Misner, K. S. Thorne, and J. A. Wheeler,
(1962). Gravitation (Freeman, San Francisco, 1973).
'D. W. Olson and W. G. Unruh, Phys. Rev. Lett. 3 3 , 2 i ~ .Isaacson, Phys. Rev. 166,1263 (1968); 166,1272
1116 (1974). (1968).
'R. A. Matzner, Phys. Rev. D g , 3274 (1976). "5. N. Goldberg, A. J. Macfarlane, E. T. Newman, R.
'R. Fabbri, Nuovo Cimento G , 311 (1977). Rohrlich, and E. C. G. Sudarshan, J. Math. Phys. 3 ,
i o ~ .Gunter, report, 1980 (unpublished). 2155 (1967).
"E. Merzbacher, Quantum Mechanics (Wiley, New York 2 3 ~ A. . Matzner and M. P. Ryan, J r . , Phys. Rev. D g ,
1970), 2nd edition. 1636 (1977).
i 2 ~ .G. Sanchez, Phys. Rev. D 2 , 1030 (1978); 8, "R. A. Breuer, M. P. Ryan, Jr., and S. Waller, Proc.
1798 (1978). R. Soc. Londonm, 71 (1977).
1 3 ~A. . Matzner, J. Math. Phys. 2 , 163 (1968). 2 5 ~ . S. Gradshteyn and I. M. Ryzhik, Table of Integrals,
14p. Chrzanowski, R. A. Matzner, M. P. Ryan, J r . , and Series, andProducts (Academic, New York, 1965).
V. D. Sandberg, Phys. Rev. D z , 318 (1976). Messiah, Quantum Mechanics (North Holland, Am-
"R. A. Matzner and M. P. Ryan, J r . , Astrophys. J. sterdam, 1962).
Suppl. Ser. 36, 451 (1978). "H. Jeffreys and Lady B. Jeffreys, Methods of Math-
. Handler and R. A. Matzner, Phys. Rev. D 22,
i 6 ~ A. ematical Physics (Cambridge University P r e s s , Cam-
2331 (1980). bridge, 1962), cf. Chaps. 12.01 and 12.02.
"P. C. P e t e r s , Phys. Rev. D 13,775 (1976).