




Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
magnetostatic multipole moments and related electromagnetic problems. 2. Review of scalar spherical harmonics. This review is intended primarily to define ...
Typology: Exercises
1 / 8
This page cannot be seen from the preview
Don't miss anything!





Eur J. Ph\s 6 I 19'351 287-294 Pnnted in Northern Ireland
287
f Institute of Physics, University of Mexico, Mexico 20, DF, Mexico $ Department of Mathematics, Chemistry, Physics and Computer Science, Inter American University, San German, PR 00753, USA S; Departamento de Fisica. Universidad Uacional de Colombia, Bogota, DE, Colombia
Received 29 June 1984, infinal form 27 March 1985
Abstract An alternativeandsomewhatsystematic definition of the vector spherical harmonics, in analogy with the commonly used scalar spherical harmonics, is presented. The new set of vector spherical harmonics satisfies the properties of orthogonality and complete- ness. and is compared with other existing definitions of vector spherical harmonics. Some applications to prob- lems in magnetostatics are illustrated.
Vectorsphericalharmonics (VSH) havebeenused in theexpansion of plane waves tostudytheab- sorption and scattering of light by a sphere (see, for example.BohrenandHuffman1983).Theyhave also been widely used in nuclear and atomic physics (see, for example, Blatt and Weisskopf 1978). The definitions of the various existing sets of VSH in different fields of physics areoftendictated by convenience. For example, one method of defining such sets makes use of an operator which is propor- tional t o theusualorbitalangularmomentum operator of quantummechanics.Whenthis operatoractsuponthescalarsphericalharmonics (SSH) function, it generates one (out of a triad of) VSH. Thepurpose of thisnote is todevelopan alternate set of VSH which is particularlyuseful in classical electrodynamics. An alternative,simple treatmentbased on SSH usesthescalarDebye potentials(Gray1978a,GrayandNickel1978). Thelayout of thispaper is as follows. In 52 we present a brief review of SSH, 63 is devoted to the definition and formal properties of VSH and finally,
1 1 Present address: Institute of Theoretical Physics, Chal- mers University of Technology, Goteborg, Sweden.
Zusammenfassung Einealternativeundmehr sys- tematische Definition vektorieller Kugelfunktionen, analog zu den gewohnlich verwendeten skalaren Kugel- funktionen wird vorgectellt. Der neue Satz vektorieller Kugelfunktionen erfullt die Orthogonalitats- und Vollstandigkeitsbedingungen: er wird verglichen mit vorhandenen anderen Definitionen vektorieller Kugel- funktionen und zur Veranschaulichung auf Probleme der Magnetostatik angewendet.
magnetostaticmultipolemomentsandrelated electromagnetic problems.
2. Review of scalarsphericalharmonics This review is intended primarily to define notation and underscore the parallel between the use of SSH andthe VSH tobeintroduced in 83. Familiarity with the properties and uses of the SSH at the level of developmentpresented in thestandardelec- tromagnetismtext of Jackson(19751 will beas- sumed. Where possible we will follow the notation of Jackson.
completenessorclosurerelation.i.e.anyarbitrary function g of 8. 6 can be expanded as
g ( & + ) = c Ao,,I'~,,l(8.+1. (2.1)
Further. if g is a function of other variables, e.g. S and r. then the expansion coefficients A,,,, are func- tions of theseadditionalvariables.Thecomputa- tion of the expansion coefficients is made relatively simple by the orthogonality condition
f d.RY?,,(B. 4)I',,,J& 6 ) = 6 , d ,,,,, ( 2. 2 )
I = ( l P > > = ~ I
cients are given by
lifiedusing the symmetries of the SSH, namely,
Y[,,,(~-e. 4 ) = (-i)l*mYl,,,(e, 4 ) (2.4) Y l m ( n - ~. ~ + n ) = ( - l 1 1 Y l ~ , ( 8 , 4 ~. One of thereasons why the SSH are usefulin physics is thattheybehave in anexemplary way when operated upon by the Laplacian V-:
The simplifications that can be achieved with spher- ical harmonic expansions are evident in the Poisson equation V'QE= -457~. Expanding both @E and p,
%(rl %41= c 1 C,,,(r)Ylm(0, 4 ) (2.6) I m
the Poisson equation becomes
must match the coefficient for p we are left with
Theangulardependenceshave ineffect cancelled outandweneeddealonly with anordinary differential equation. To achieve a general solution of the differential equation given by equation (2.9) we start by con- sidering the very special case pin? = 6 ( r - r') and the differential equation
(2.10) vanishes and the solution is simple: f =r < r ' C r '
Notice that we have chosen only solutions that are well behavedattheoriginandat infinity. Con-
obtain the remaining coefficient, C, the differential equationmustbeintegratedoveran infinitesimal
C = [ 4 ~ / ( 2 1 +l)]r"".Details of the procedure to obtain C are to be found in Jackson (1975. $3.9). Introducing the usual notation
thesolution,i.e.theGreenfunction,canbe written
To obtainthesolutiontoequation (2.9) weneed onlyrecognisethatany pi,,, canbewrittenasa superposition of delta functions. As is well known
Therefore the solution to equation (2.9) must be a similarsuperposition of f s as given by equation (2.131:
If we now require that the charge distribution p be bounded in extent, and we ask explicitly for the value of C,, at a value of r outside the source, we have r = r, in the integrand and
charge distribution are called its multipole moments, andtheexpressionthatresultsfromsubstituting equations (2.161 and (2.17) backin equation (2.6),
= - 4 ~ 8 ( r ' - r ). (2.10)
where V;,,,. Vi: and Vi:: aretheexpansion coeffi- cients analogous to fit,, in equation (3.3). In physics theoverwhelminglyimportantprop-
operator. Specifically, if we haveanyequationin- volving the C operator (as gradient, Laplacian, curl, etc)and if all functionsareexpanded in spherical harmonics(scalars in scalarsphericalharmonics, vectors in vectorsphericalharmonics),thenthe angular dependence will ‘cancel out’. Towards this endonecanreadilyconfirmtherelationsamong scalar and vector spherical harmonics given in the following compendium:
A fewexplicit values of thevectorspherical
definitions it is aneasymatterto verify thatthe following relations are satisfied:
Yk,, = (- l)’ny:,,, W k n , = (-l)t”W:,n (3.15) @l,-,,, = (- 1 )”‘@? 111. Further,fromthedefiningequations(3.5)and (3.8) we have
which greatly simplify the tabulation of the explicit values. T o gain confidence that our mathematical arsenal does fill our needs we turn back to the problem at thebeginning of this section.Utilisingequations (3.31. (3.10) and (3.11) we have
therefore
Noticethattheangulardependencehascancelled out as expected; one is thusleftwithanordinary differential equation.
Vector spherical harmonicsand their application to magnetostatics 29 1
Otherdefinitions of VSH havebeengiven,for example, by Hill(1954)whodefinesseta [V,,,,, W,,, X,,,] and by Blatt and Weisskopf (1978) who define a set [ Y;yh.I 1 where (Y - I = O , + 1. The connectionbetweentheseandthepresentvector harmonics is as follows:
Konopinski (1981) and Morse and Feshbach (1953) present in their popular books two more definitions
We remark that with the useof group theory, the discussion by Blatt and Weisskopf and the explicit equivalences(3.19)betweentheirsphericalhar- monicsandthepresentonesshouldconvincethe readerthatthevectorsphericalharmonicsdoin- deed form a complete set. A simple proof of com- pletenessbased on the Debye potentials has been recently reported (Gray and Nickel 1978). We will now show that the VSH are orthogonal in the same sense as the SSH, and that the VSH have useful symmetry properties. First of all we note that at a point, for the same
Of VSH.
Yl,,, * W[,,, = 0 YIP,, * ah,,, = 0 (3.20) W[pma,, = 0.
Morerelevantaletheorthogonalityproperties
m. m'.
= dRW,,,, a,:, = 0.
Although equations (3.21) are not obvious, they are easily verifiable from equation (2.2) and the defini- tions of the SSH. Wemention in passingthatthe
make the vector spherical harmonics orthonormal. Using the above, the expansion coefficients in a vector spherical harmonic expansion are relatively straightforward, namely,
with the coefficients V;,,,, Vi::, V;:: being
V;,,, = d R V * Y;",, 5
4. Magnetostatic multipole moments Our goal here is to develop a formalism for multi- poles of the magnetic induction field, B. similar to that (see $2) for those of the E field.This will be possible becausefor no current ( J = currentden- sity) at the field point
C
andthereforethereexistsafunction QM, the magnetic scalar potential, from which we can find the magnetic induction field as B = -V@.,. (4.2) This scalar :unction can be expanded in a manner preciselyan .logous totheexpansion in equation (2.18)
sincethis is thegeneral well behavedsolution of C B = -T2QM = 0 ; the quantities MI,,, are the mag- netostatic multipole moments for the field. Notice that in general we cannot use Q,. directly
being that 0 , has meaning only where J = 0. Thevectorpotential is relatedtothecurrent sources through the equation
Fromequation(4.4)thedifferentialequationfor A:? is seen to be:
4 a C
therefore, a Equation (4.17) can be written in the same form as that of an electrostatic dipole
m e r r 3
where m denotesthemagneticdipolemoment which characterisesthe field. Thecorrespondence of m and Mlo could have been written down intui- tively.electrostatic Andipole coefficientis as- Figure 1 sociated with a dipole component (see, e.g. Jackson 1975 p 137) p ‘ = ($.x)”2q10. (4.19) l = 2 yields atonce thattherefore follows It M2.o = M2.*z= 0
c Considerasasecondexampleacurrentpath
as shown in figure 1. It is ‘intuitively obvious’ that
moment of thetwo Ds cancel), so thedistant
magneticinduction field will beatmostaquad- rupole field. We follow the same procedure to find M z , as we did to find the dipole coefficients of the ring. The current distribution is Theleadingmagneticinduction field atlargedis- tances then can be found from the potential It is possible to describe this field with a symmet- ric quadrupoletensor Q, fromtherelations(see, e.g. Jackson 1975 p 137). 4ia31 ( ) 1 ; li2 - 1 ( 1 5 ) M,,=-? S ”_ - (Qxy - iQ,, Since e, * @L= 0, the third part of the current gives the other contributions. The integration J J @Ld 4 3 8i-r no contribution to $2. Consider the integration for M20=
will haveformtheJzexp(-im4) d 4 - J:nexp(-imq5)d4. Weconcludethat m mustbe (4.25) oddforanonvanishing $2 andthat,forodd m, Forthis field, then,theonlynonvanishingquad- { exp(-imd) d 4 canbereplaced by -4i/m.rupolecomponent is 4.14),dingquationatarriveto We now take precisely the same steps as we did QY Z =--, (4.26) S ( r - a ) - Y ~ , J ~ = , , ~ .( 4. 2 2 ) Themagneticinduction field, of course, will be 3 a m ae samepreciselythe as an E field described by equation (4.24).A simpler example where the lead- Substitution of thisresultintoequation(4.11)with ing multipole is thequadrupole, is twoloopsof C p , = - - 2iI a
294 R G Barrera, G A Esttvzz and J Giraldo
equal and opposite dipoles (Gray 1979). Introduc- toryaccounts of multipoleexpansions, using both sphericalandCartesiantensorsandtherelations between them are found in the literature (see, for example,GrayandGubbins 19831.
5. Concluding remarks The new results of this report are the definitions of vector spherical harmonics through equations ( 3. 5 ) , (3.8) and (3.9). We emphasise that their develop- ment assumes nothing more than a familiarity with thepropertiesandusage of thescalarspherical harmonics. Indeed. the vector spherical harmonics areformulated in analogytothescalarspherical harmonics.Weremarkthatprovidedone is not going to do transformations of axes, one can have scalars and vectors as they are in the present paper.
description giveninthis paper is not adequate. Several convenient approaches for treating radia- tion fields using vectorsphericalharmonicshave been recently reported (Gray 1978a. b, 1979, Gray and Nickel 1978. Lambert 19781. We have seen by way of illustrativeexamplesthatcertainproblems in magnetostatics offer straightforward solutions in terms of vectorsphericalharmonics.Furthermore. combining the approach by which the vector spher- ical harmonicsareintroduced in thisarticle with theusualexpressionsforthemultipoleexpansion of a plane wave. standard electromagnetic scatter- ing problems can be easily workedout.These featuresshouldrender this article useful tosenior undergraduate and graduate students.
Acknowledgments Dedicated to Professor J D Jackson in admiration and affection on his sixtieth birthday. We wish to express our sincere thanks to L B Bhuiyan, Ismael Cervantes, Alan Lev Estevez and Albert0 Jose Moreno-Bernal for their kind encouragement in so many ways. Financial support for this work was provided in part bv a grant to GAE from Universidad Pedagogica y Tecnologca de Colombia. Tunja. Colombia.
References Blatt J M and Weisskopf V F 1978 Theoretical Nuclear Physics 2nd edn (New York: Wileyi Bohren C F and Huffman D R 1983 Absorptionand Scatrering of Light by Small Particles (New York: Wiley~ch4. Bronzan JB1971 A m. J. Phys. 39 1357 __ 1982 Electromagnetism.Path to Researched D
GraqC G 1978a Am. J. Phys. 46 169
Teplitz (New York: Plenum)
~ 1978b A m. J. Phys. 46 582
Gray C G andNickelBG1978 A ~ I .1. Phys. 46 735 Hill E L 1954 Ant. J. Phys. 22 211 JacksonJ D 1975ClassicalElectrodynamics2ndedn (New York: Wilevi ch 3. 5. 16 Konopinski E 1981 Electronlagnetrc Fields and Rela- ticistic Particles (New York: McGraw-Hill) ch 9 Lambert R H 1978 A m. J. Phys. 46 849 Morse P M and Feshbach H 1953 Methods of Theoreri-
Fluids (Oxford:OxfordUniversityPress)
cal Physics 2 v01 INea York: McGraw-Hill)