The Classical Electromagnetic Field

(Gaussian).

1

CONCEPTS OF A FIELD THEORY

This is a book on the electromagnetic field, one of the many examples of fields that are important in physics and natural science. In the words of Morse and Feshbach (B), Practically all of modern physics deals with fields: potential fields, probability fields, electromagnetic fields, tensor fields, and spinor fields. Since the concept of field is applied so widely, it is perhaps natural that it has taken on somewhat different contexts with, of course, an underlying common denominator. In this first brief chapter, we shall try to analyze the concept and its variants and so highlight the essential aspects of the electromagnetic field.

One mathematical common denominator is easy to isolate. Mathematically, a field is a function, or a set of functions considered as an entity, of the coordinates of a point in space (and possibly of time). For example, if the temperature is defined at every point in some volume, we say there is a scalar temperature field throughout that volume. If, in a moving fluid, the three vector components of velocity are known as a function of position in the fluid, they constitute as a whole a vector velocity field. In the theory of elasticity, the relative vector displacements of points of an elastic solid from their unstrained positions are described in terms of a double vectorial or, as it is more usually called, a tensor field. In modern physics, the Schrödinger probability amplitude ψ or the generalized Dirac spinor amplitudes are examples of fields.

It is worth noting that in this above list, there are really two kinds of fields. The first kind, exemplified by the temperature or velocity fields, is an idealization that is really defined only in a certain approximation of coarseness or fineness. For example, the velocity field of hydrodynamics is meaningful only in an average or continuum approximation in which the atomic and grainy structure of the fluid is not considered. This point is discussed in some detail in Morse and Feshbach (B) and we shall not elaborate on it here. By contrast, the Schrödinger or Dirac fields are not approximations to an underlying discontinuous physical model but must be assumed to exist no matter how finely space is divided.

To describe the evolution of the electromagnetic field concept, we recall some history that begins with Newton (1642-1726). One of the great laws of physics is Newton’s universal law of gravitation. This law embodies the concept of action at a distance, according to which gravitational masses exert the forces they do on each other by virtue of their positions in space, the intervening space playing no active role. This is meant to contrast with forces which work via contiguous action whereby two masses at a distance exert forces which are transmitted by the intervening medium. For example, if several billiard balls are in contact in a row on a table and the first one is struck, the last one will move. Thus, the one billiard ball exerts a force on the other distant one, but by a mechanism which involves successive actions of the intermediate balls, the one moving the next, moving the next, etc. The concept of contiguous action is then quite different from that of action at a distance where no intermediate mechanism or medium is considered.

The work of Newton is relevant in a second way. His laws of the motion of point particles and rigid bodies paved the way for the development of the continuum mechanics of fluids and later of elastic bodies. Some tentative beginnings on the subject of fluid flow were made by Newton himself, but the real groundwork was later laid by John Bernoulli (1667-1748) and Euler (1707 – 1783). They bypassed the problem of the actual microscopic structure of fluids by adopting a continuum model and then applied Newton’s laws of point mechanics to small elements of the continuum. The same idea was later applied to elastic solids, and the vibrations of these solids was discussed by applying Newton’s laws to a small element of the solid, assuming that the forces acting on it were the stresses due to the rest of the solid, plus any external forces. Hydrodynamics was therefore formulated in terms of the velocity and acceleration of the moving fluid at every point, i.e., in terms of velocity and acceleration vector fields. The theory of elastic solids was similarly formulated in terms of stress and strain tensorial fields.

So much for mechanics; we turn now to electromagnetism. A basic law of electrostatics is Coulomb’s law for the force between two charged particles. Except for the fact that the electric force can be either attractive or repulsive, whereas the gravitational force is always attractive, this law is obviously similar to Newton’s law of gravitation. It was then considered from the time of its discovery as an example of action at a distance, in which two charges act on each other in a way that has nothing to do with the intervening medium. But this view began to be questioned, at least in the mind of Faraday (1791-1867), by his work on dielectric polarization. This phenomenon led him to attribute more and more importance to the intervening medium. We cannot go into the details of Faraday’s results and reasoning but, for illustration, shall concentrate on one of his findings. This has to do with the effect of insulators or dielectrics on the capacitance of condensers. Consider a parallel-plate capacitor with air between its plates; it has a certain capacitance. If the air is replaced by a dielectric medium, the capacitance will be increased. Faraday viewed the phenomenon of enhanced capacitance as somehow due to the fact that the electric force generated by the charges on the plates was weakened by the dielectric medium. But if changing the medium that intervened between the charges changed the force, then somehow the forces must depend on, or be transmitted by, the medium. As a corollary of this view, Faraday considered that the essential feature of the interaction between charged particles was the lines of force that carried the stresses of the medium from one charge to another. These lines of force that extend from charge to charge through the medium were considered primary; the charges merely happen to be the places where the lines of force start and stop. Although these views were mainly derived from experiments on dielectric polarization with strongly dielectric substances, they seemed to be equally valid for those whose dielectric constant was close to unity. By extension, then, Faraday considered them valid for that insulator whose dielectric constant is exactly unity, i.e., free space. In short, in Faraday’s view, free space was a substance, qualitatively like all other insulating substances, that contained charges which could be separated and displaced, as charges were separated and displaced in the material insulators he used in his capacitor experiments.

The next great name in the history of the electromagnetic field is that of Maxwell (1831-1879). He took up Faraday’s ideas on the nature of the force between charges and the importance of the intervening medium. He succeeded in showing that the forces that charge complexes exert on each other and their energies could be expressed not only in terms of the magnitudes of the charges and their positions but in terms of a stress energy tensor that was defined throughout the medium (even if that medium was free space) and that had as components certain functions of the field strengths. For example, the force that one point charge exerts on another could be calculated by either using Coulomb’s law, or surrounding the charge by an imaginary surface and integrating over that surface the total electrical stress as given by the stress energy tensor. This concept of stresses in the medium gave no new result, but it did at least show that the Faraday-Maxwell conception of the state of the medium was consistent with the results, if not the concept, of action at a distance.

The next step in the development of the field concept was also due to Maxwell. This was, of course, the discovery of the equations that bear his name and, as a corollary, the discovery of electromagnetic waves and their identity with light waves. Two basic guides in this work were Faraday’s theory of charge polarization or displacement, and the theory, well developed by Maxwell’s time, of the vibrations of an elastic solid. By generalizing Faraday’s idea of displacement to the time-varying case and by introducing the so-called displacement current D, Maxwell found that electromagnetic phenomena could be described in terms of four field vectors, E, B, D, H. But like the velocity fields of hydrodynamics or the stress or strain fields of elasticity, these fields were not considered to exist by themselves but were somehow considered to be vibrations or displacements of an underlying luminiferous ether whose properties were those of a somewhat special kind of elastic solid. Electromagnetic waves were then, so to speak, secondary: the ether could exist without electromagnetic waves but electromagnetic waves could not exist without the ether.

Maxwell’s theory was a great triumph. But as the years after its discovery passed, it was accompanied by an increasing perplexity as to the nature of the hypothetical ether that underlay it. The history of the researches and speculations on the nature of the ether is beyond us here. It is well described in Whittaker (B) and some detail is given in Chapter 12 of this book. Suffice it to say here that a famous experiment by Michelson and Morley showed that the ether did not exist. Nonetheless, electromagnetic waves continued doggedly to be generated and propagated. The Michelson-Morley experiment thus served, so to speak, to emancipate the electromagnetic field. After it, the fields could not be considered as merely vibrations of an underlying medium. From the time of the disproof of the existence of the ether, the electromagnetic field had to be looked on as an entity in its own right, as real as matter and everywhere on a par with it.

In fact, the essence of the electromagnetic field theory is that the field does have properties that we usually associate with matter. It can possess energy, momentum, and angular momentum. The field is thus a dynamical concept and is not merely a mathematical function of the space coordinates and time. We shall highlight this essential aspect of the electromagnetic field by a simple example. In it, we presuppose a small amount of elementary knowledge on the reader’s part; namely, that accelerated charged particles radiate and that this radiation is propagated with a finite velocity c.

Suppose two charged particles q1 and q2 are separated by a distance d. Imagine that we suddenly move particle 1, say, and then quickly bring it to rest again. Having been accelerated, the particle will emit a pulse of radiation which travels at velocity c and hence would make itself felt on particle 2 at a time t = d/c. If we look at this system of two charged particles at some time after the first particle is brought to rest but at a time which is less than d/c, we would see simply two charged particles, each at rest; they would constitute an isolated system with kinetic energy zero. Suddenly, however, at a time t = d/c, we would find that the second charged particle began to move. Superficially, the energy (and momentum) of this isolated system would appear to change even though there were no external forces acting on it. How can we reconcile this with the conservation laws of energy and momentum? The field theory gives one possible answer to this question. According to it, we have simply overlooked the fact that there are forms of energy other than kinetic; there is in fact another physical entity that we have not mentioned, the electromagnetic field, and this entity contains energy and momentum. The energy and momentum that begin to be transmitted to particle 2 at time t = d/c were, in fact, contained in the electromagnetic field for earlier times.

We shall spell out these conservation laws involving the electromagnetic field in Chapter 11. For the moment, however, we make another point: the conceptual difficulties with energy and momentum conservation come about primarily because of the finite velocity c of electromagnetic propagation. For if c were infinite, we could not set up, even in thought, the above experiment and the concomitant difficulties would not arise. For static fields, then, in which propagation velocities are not involved, we shall see that the field concept does not so inevitably impose itself. Many of the results of electrostatics and magnetostatics can be formulated in terms either of action at a distance or of a field.

In summary, the basic idea of electromagnetism as a field theory is that charges and currents produce at each point of space a field that has a reality of its own, that can contain and propagate energy, momentum, and angular momentum, and that acts on other charges. The field is produced by, and acts on, charges. Correspondingly, there are two sets of equations. These are, first, Maxwell’s equations which describe the field produced by a given set of charges and currents. Second, there is the Lorentz force equation which shows how a given field acts on charges.

We should not leave the impression that classical field theory itself is without its difficulties. It has these too. There is a conceptual difficulty in describing the self-action of the field on the charge that produces it, which is not soluble in classical field theory. This is outlined in Section 14.9, and an extensive discussion is given in Rohrlich (B). For practical computational purposes, these difficulties are resolved in part by quantum electrodynamics, which is the extension of the classical field theory of this book to incorporate quantum concepts; there remain however difficulties in principle connected with the occurrence of infinite quantities. There have been attempts over the years to bypass the difficulties with field theory, by reviving a sophisticated version of the action-at-a-distance theory. Examples are the theories of Wheeler and Feynman (R) and a recent paper by Kennedy (R).

REFERENCES

Hesse, Mary B., Forces and Fields, Philosophical Library, New York, 1962.

This book is subtitled, The concept of action at a distance in the history of physics. The development of this concept is studied from primitive times to the present.

Kennedy, Frederick James, Instantaneous Action at a Distance Formulation of Classical Electrodynamics, J. Math. Phys., 10, 1349 (1969).

A recent attempt to provide an alternative to the field description of the interaction of charges. There is a useful set of references.

Wheeler, J. A., and R. P. Feynman, Interaction with the Absorber as the Mechanism of Radiation, Rev. Mod. Phys., 17, 157 (1945); Classical Electrodynamics in Terms of Direct Interparticle Action, Rev. Mod. Phys., 21, 425 (1949).

An interesting attempt to alter the standard framework of electromagnetic theory.

Williams, L. Pearce, The Origins of Field Theory, Random House, New York, 1966.

A nonmathematical, but interesting and detailed account of the history of the field concept.

2

THE ELECTROSTATIC FIELD

It has been known for a long time that there are materials which, having been rubbed, will attract or repel other small pieces of matter. In particular, the Greek philosophers knew that amber possessed this property; it is from the Greek word for amber, ηλεκτρον, that the word electricity is derived. There appears to have been nothing extensive made of this knowledge, however, and the beginning of electrostatics as a science must be dated, along with many other beginnings, from the Renaissance. The detailed history of this is complicated and fascinating, but we can mention only some highlights. Gilbert (1540 – 1605) found that materials could be roughly divided into those that could be electrified by friction (electrics) and those that could not (nonelectrics). Today we call such materials insulators and conductors, respectively. DuFay (1698-1739) discovered that the electrical fluid was of two types, vitreous and resinous; this was the origin of the concept of positive and negative electrical charges. Moreover, it was found that vitreous and resinous electricity could not be produced from nothing; a change in the amount of one was accompanied by a like change in the amount of the other. This observation was the primitive basis of the idea of the conservation of charge; in effect, it justified our treating amounts of positive and negative electricity according to the algebraic laws for combining positive and negative quantities.

All the early experiments were done with frictional electricity, i.e., with electrical forces generated by rubbing bodies together. It is amusing to note that even today we still do not have a clear understanding of the detailed mechanism of this phenomenon, as Harper (R) discusses.

Be that as it may, the developments over the years and centuries have been distilled into a system of concepts and nomenclature that we have inherited. In brief, a body which is in a state to exhibit the kind of electrical phenomena described here is said to be electrically charged or to possess charge. Somewhat tautologically, then, charge is a name for that quantity which a body possesses when it exhibits electrical forces. This charge is of two kinds, negative and positive, and the amounts of these can, when appropriate, be added algebraically. Charges of like sign tend to repel each other and of unlike to attract. From modern experiments we know even more exactly that charge is invariable in the sense that the total charge of an isolated body cannot be changed. This most important fact is the Law of Charge Conservation. We know moreover that charge is quantized, i.e., the total charge on a body is an integral multiple of a certain unit of charge. This unit of charge is that which appears on the electron, which elementary particle is, in general, the mobile particle whose mass motion constitutes the subtile electrical fluid of the earlier days.

All this is qualitative and descriptive, however. The science of electrostatics begins with a quantitative version of the above facts. This is embodied in Coulomb’s law which is described below. But first we discuss the question of how charge may be distributed in space, and how its distribution is described mathematically.

2.1 THE DISTRIBUTION OF CHARGE

Charge can manifest itself in either discrete or continuous distributions; here we discuss the relation between these two concepts. Historically the basic charge element was considered to be the point charge, with vanishingly small dimensions. Starting from this, one conceives of a continuous distribution of charge as one in which very many point charges are distributed in close proximity so that on some spatial or temporal average they are distributed continuously. In much the same way, one often describes a gas or liquid by a smooth mass-density function even though it is known that on an ultimate small scale it is grainy and molecular. For many purposes the graininess is not important, however, and equally, a complex of point charges can be considered for many purposes as a continuous smooth charge density. This may be so, for example, if the observation is made at distances which are large compared to the scale of graininess.

Quantum mechanics on the other hand, yields an opposite point of view; this discipline teaches that it is more natural to think of charge as distributed continuously. Thus in an atom, molecule, or solid, the charge is conceived as being spread out in a continuous smooth cloud, whose density is governed by quantum-mechanical laws. The idea of a continuous charge distribution is then basic. This does not, however, destroy the usefulness of the concept of a point charge. For we shall see later that any charge distribution of one sign whose linear dimensions are all small compared with the distance at which its properties (field and potential) are measured will act like a point charge. The precise meaning of this statement will emerge when we discuss multipoles. For the moment, we simply say that the properties of such a distribution are given by an expansion whose first term is the point charge term, and whose succeeding terms diminish as powers of the reciprocal of the distance to the point of observation. But before doing that, we should emphasize that a point charge is not one which is small in any absolute sense; there is no such absolute sense! A point charge is one that is relatively small, and the scale is set by the distance to the point of observation. Thus the sun, if charged, would be to a good approximation a point charge from the earth as observation point, whereas the proton is not necessarily a point charge if it is probed at distances of the order 10 – 13 cm or less.

We shall have to deal with some charge distributions that are relatively diffuse, and with others that are very sharply localized and concentrated. The various laws and formulas of electrostatics frequently apply equally well to either type of distribution. It is very convenient to have a single mathematical description that is applicable to either case. We shall accomplish this by introducing a density function ρ(x, y, z), where x, y, z are space coordinates (not necessarily rectangular ones). We shall often write this more briefly as ρ(r), where r is the position vector associated with the point x, y, z. Then for the continuous case, the quantity of charge in an infinitesimal volume element dv is defined as

(2.1)

Since charge in the Gaussian system is measured in electrostatic units (esu or statcoulombs), p has dimensions esu x (cm) – 3. A point charge is included in this general description by letting ρ(r) be a δ-function. This kind of function is discussed in Appendix B; briefly, it can be considered as the limit of a sequence of functions which are increasingly singular, but which always have unit integral. In terms of the δ-function, a point charge q at r1 is described by

and more than one point charge, with charge qi at ri, by

(2.2)

2.2 COULOMB’S LAW AND SUPERPOSITION

There are two fundamental laws of electrostatics. The first is Coulomb’s law, which describes the force between two point charges; the second is the law of superposition, which extends this to more than two charges.

For Coulomb’s law, consider two point charges of magnitudes q and q′, a distance r apart. Then the law states that the strength F of the force between them is

where C is a constant of proportionality that depends on the system of units. The direction of the force is along the line joining the two charges and is a repulsion if the charges are like in sign, an attraction if unlike. Both aspects of this law can be stated in vector notation. We refer to Fig. 2.1 in which r is the vector (magnitude r) from the charge q′ to the charge q is the unit vector² in that direction. In cgs units C is unity, and the vector force on q due to q′ can be written

(2.3)

Fig. 2.1 Illustration for Coulomb’s law.

The second law of electrostatics is that of superposition; it has to do with the force exerted on a given charge by a distribution of charges. The law is this: the force on the given charge due to the distribution is the vector sum of the forces that each charge of the distribution would produce, if it were alone. In other words, Coulomb’s law between any two point charges is not affected by the presence of other point charges. This may sound banal to the point of being trivial, but it is far from that; without superposition, electrostatics would be a quite different (and more difficult) subject.

2.3 THE ELECTRIC FIELD

Although Coulomb’s law and superposition are the theoretical bases of electrostatics, it is useful for both practical and conceptual reasons to introduce two secondary ideas, those of electric field and potential. As we indicated in Chapter 1, the concept of field is not essential in electrostatics, but it almost inevitably enters later in the discussion of time-varying phenomena; it then makes for a nicer continuity to deal with it from the beginning. The concept of potential is likewise not so basic as Coulomb’s law, but it has many virtues, not the least of which is that it enables one to reduce vectorial problems to scalar ones.

We begin with electric field. From Fig. 2.1, the electric field E at the position of the charge q is defined as the vectorial force exerted on that charge, divided by the magnitude q,

(2.4)

Comparing Eqs. (2.3) and (2.4), we see that for this case E is also defined by

(2.5)

The split of Coulomb’s law, (2.3), into the two equations (2.4) and (2.5) is, as we have said, somewhat arbitrary. Nonetheless, these equations do point up an essential aspect of the field concept: the field is produced (Eq. 2.5) by the charge q′ and then acts to exert a force (Eq. 2.4) on the charge q. It follows from the law of superposition for forces that the field E obeys the superposition law as well. For an arbitrary distribution of charge, the field is then given by vectorially summing the field due to all the charge elements. For example, for point charges q′, the field at a point is, with ri the to the field point,

(2.6)

and if the charge distribution is continuous, there is an analogous integral formula. From (2.4) we see that the electric field is alternatively defined as the force on a unit positive test charge at a point. It is assumed, of course, that the presence of the test charge does not perturb the charge distribution whose field it measures.

Fig. 2.2 Lines of force and equipotential surfaces for some charge distributions.

The electric field is frequently characterized pictorially by its lines of force. These are closed lines whose directed tangent at any point of space gives the direction of the field at that point. Since the field can be defined as the force on a unit positive charge, it points away from the immediate vicinity of positive charges and toward the immediate vicinity of negative ones. One says that the field lines start on positive charges and stop on negative ones. Figure 2.2 gives examples of the lines of force produced by various charge distributions.

The electric field of any charge distribution is a vector field defined at all points in space. What are its general properties? Now it is known from vector analysis that a vector field is uniquely determined if its divergence and curl are known throughout space; this is sometimes called Helmholtz’ Theorem and is discussed in Appendix C. We might then begin to answer this question by investigating the divergence and curl of E. If we do this, we ask about the differential properties of the field. Alternatively, we might look at integral properties, i.e., properties the field has when it is integrated in some specified way. The differential and integral properties are, of course, not independent; the one can be derived from the other, and we can then start with whichever is most convenient.

We shall begin with an integral property that is known as Gauss’ Theorem. To derive it, consider a point charge q enclosed in some volume bounded by a surface as shown in Fig. 2.3. We want to consider the total flux of E through this surface. Recall that the flux of E through any small surface element is defined as the product of the normal component of E times the area of the element. If we let ds be an outward pointing vector element of surface, the flux of E through the whole surface S s E · ds. Using the expression (2.5) for E and recognizing that the element of solid angle d, we have

Fig. 2.3 Gauss’ theorem.

This result for a single point charge in the volume is obviously independent of where in the volume the point charge is. It follows then from the law of superposition that for a distribution of charges in the volume a similar result would hold except that instead of q on the right-hand side there would appear the total charge, call it Q, in the volume. We can then write Gauss’ Theorem as

(2.7)

If one does a calculation analogous to the above for the case that the point charge is outside the surface, then for the simple surface shown, a line of force will enter the surface at one point and exit on the opposite side. Remembering that ds is the outwardly drawn vector surface element, we readily see that the net entering flux of a narrow cone of lines of force will just cancel the corresponding exiting flux on the opposite surface. Explicitly we may say what is implied by Eq. (2.7): the net flux through a closed surface that does not enclose any charge is zero.

To get the differential counterpart of Eq. (2.7), we can proceed here in either of two ways, and it is instructive to mention both of them. The first relies on the divergence theorem which is discussed in texts on vector analysis and stated in Appendix C. For any vector field A(r) which is defined through a volume V, that is bounded by a simply-connected surface S, as in Fig. 2.3, the divergence theorem is the following relation between integrals:

(2.8)

Now imagine a continuous distribution of charge with a density ρ(r). Apply Gauss’ Theorem to an arbitrary volume V0 in it. Then the Q that appears on the right-hand side of (2.7) can be written as Q vo ρ dv, and the left-hand side can be rewritten via the divergence theorem to yield

This equation can hold for an arbitrary volume only if the integrand vanishes, and so we have

(2.9)

This is the differential expression of Gauss’ Law. Another way of getting the same result is to apply Eq. (2.7) to an infinitesimal volume ΔV with bounding surface ΔS. Then the right-hand side of Eq. (2.7) can be approximated by 4π is the average value of ρ through the volume. Equation (2.7) then reads

In the limit that the volume ΔV shrinks to a point, the mean value p simply becomes p, the value at that point. Moreover, the left-hand side of the last equation then defines ∇ · E at that point, and so we again get Eq. (2.9).

So much for ∇ · E. To finish characterizing E as a vector field we now calculate V x E. Here it will be easier to begin with the curl considered explicitly as a differential operator. Although we want to calculate V × E for the field of an arbitrary charge distribution, we shall begin by calculating for a single point charge and later invoking superposition. Without loss of generality we can imagine that this point charge q is at the origin of a system of coordinates and we call its field Eq• Then Eq. (2.5) gives for the components Eqx, etc.,

In rectangular coordinates, the x-component of V x Eq is

With a similar result for the other components we have for this example

∇ × Eq = 0.

This result is rather more general than might appear at first sight. For, given an arbitrary charge distribution, then at any point where the charge density is zero, the field can be considered as the superposition of the fields due to more or less distant charge elements. From the above result, each of these elements produces a field whose curl is zero; the equation V × E = 0 is thus valid at any point outside an arbitrary charge distribution.

For a point inside the distribution, where ρ is finite, we can use an argument whose general form we shall have occasion to repeat several times later in this book in other contexts. At such a point we imagine a small sphere, centered at the point. The field there is then the superposition of the fields due to those charges inside the sphere and to those charges outside. These latter, as we have just seen, produce a field whose curl is zero. We then consider the field at some point in the sphere due to the charge in the sphere itself, and for this we must be more explicit about its size. If the function p(r) is continuous, which is assumed, we can always take the radius of the sphere small enough that the charge density inside is essentially constant to whatever accuracy we may wish to assign. Call that radius r0, and the density ρ0. We then have the problem of calculating the field inside a uniformly charged sphere. This field will be radial, by symmetry, and it can be calculated by Gauss’ Theorem. Thus the flux of E through some spherical surface of radius r < r0 is just 4πr² E and the total charge within the surface is ρ. From Eq. (2.7)

The curl of this field is zero everywhere for r < r0 and in particular at r = 0. Thus the charges outside the sphere and those inside produce fields whose curl is zero. We can therefore state that everywhere inside or outside of nonsingular charge distributions,

∇ × Ε = 0.

An important corollary of the above discussion is that the field is finite inside or outside any continuous charge distribution for which ρ is finite everywhere.

From this differential characterization of V x E we get the integral counterpart by invoking Stokes’ Theorem. This theorem is discussed in Appendix C. In brief, it states that given a closed curve C (element of length dl) spanned by a surface S (element of area ds), then for a vector field A(r) one has, with a sign convention discussed in Appendix C:

If we apply this result to the vector field E, we have for any closed path, ,

(2.11)

2.4 THE POTENTIAL

The fact that the curl of E is zero enables a simplification. For we know from vector analysis³ that this implies that E can be calculated as the gradient of a single scalar, the potential function, which we call Φ:

(2.12)

One virtue of Φ is that it reduces a vectorial problem to a scalar one. Instead of calculating each component of E separately, we can calculate the single function Φ and then use (2.12). Also, Φ has a useful physical interpretation: the work done in taking a unit charge between two points can be simply expressed in terms of it. We will discuss this now.

Note first that the work the field E does in moving a unit positive charge from one point to another along some infinitesimal length dl is

dW = E dl.

The work done in moving the particle a distance dl against the force E is just the negative of this. The total work W12 required to move a particle a finite distance between two points 1 and 2 in a field E is then

(2.13)

Using Eq. (2.12), we have

(2.14)

Thus the difference of potential between two points is the work that must be done to move a unit charge between those points. Equations (2.13) and (2.14) also show that the line integral of the tangential component of E along any path joining two points is independent of the path and depends only on the endpoints. This is an important property of conservative fields. The equation, Φ(r) = constant, defines a surface called an equipotential surface; such surfaces play an important role in electrostatics. Some equipotential surfaces for simple charge distributions are sketched in Fig. 2.2.

In the spirit of the discussion of E, consider the properties of Φ. The differential property is obtained by combining (2.9) and (2.12), and is

(2.15)

This is Poisson’s equation. In a region where ρ is zero it becomes Laplace’s equation

(2.16)

We can get the integral counterpart of Poisson’s equation from the potential for a point charge and the principle of superposition. To do this we first observe that the potential which a single point charge produces at a given distance is the magnitude of the charge divided by that distance. The potential Φq of a point charge of magnitude q′ at the origin should then be Φq = q′/r. This is easily confirmed. Taking the negative gradient of Φq, we are led to the expression (2.5) for E. Now suppose we have a charge distribution ρ(r′) as shown in Fig. 2.4. (It is described in terms of a primed variable to save r for the vector to the field point.) The small element of charge p(r′) dv′ produces at r a contribution to the potential which equals the magnitude of this charge element divided by its distance from the field point. This contribution is ρ dv′/R = ρ dυ′/|r – r′|. The total potential at r is, by superposition,

(2.17)

Fig. 2.4 Coordinate system for calculating Φ at the field point r due to the charge distribution ρ(r′).

This is a very important equation, since it permits us to calculate Φ, and hence E, for an arbitrary charge distribution, continuous or discrete. The function 1/|r – r′| that appears in (2.17) is sometimes called a unit source or Green function.⁴ It is called a unit source function because it gives the effect at r due to a unit source at r′, and is called a Green function for historical reasons that we shall see later. It is one of the simplest examples of a large class of Green functions that we shall be concerned with.

In summary, the differential and integral properties of E and Φ that have been found are:

PROBLEMS

Use Gauss’ Theorem to calculate the field inside and outside an infinitely long cylinder of circular cross section which contains a uniform density of charge. Can you calculate the field similarly for a cylinder of finite length?

Find the distribution of charge giving rise to an electric field whose potential is

where x and y are Cartesian coordinates. Such a distribution is called a two-dimensional one since it does not depend on the third coordinate z.

The potential of a certain charge distribution is

where r is a radial coordinate and q and x are parameters. Find the distribution of charge, continuous and discrete, that would produce such a potential.

Suppose the force F between two like particles is governed by a modified Coulomb’s law

If a potential exists for the force field F(r), find it. Investigate ∇ F and V x F.

The potential of a certain idealized charge distribution in spherical coordinates r, θ, is

Φ = m cosθ/r²

Find the components Er and Eθ of the field, and sketch the lines of force.

REFERENCES

Canby, Edward Tatnall, A History of Electricity, Hawthorn Books, Inc., New York, 1963.

Semipopular, elementary, but accurate, and the illustrations are delightful.

Harper, W. R., How Do Solid Surfaces Become Charged? Static Electrification: 1967

Conference Proceedings, The Institute of Physics, London, 1967.

Roller, Duane and Roller, Duane H. D., The Development of the Concept of the Electric

Charge, Harvard University Press, Cambridge, Mass., 1954.

Scholarly, and devoted to a rather specialized topic, but one which has many ramifications.

3

THE SUMMATION PROBLEM FOR CHARGES

In this chapter we turn to the problem of evaluating the expression (2.17) for Φ(r) assuming ρ(r′) is given. Following Sommerfeld, we call this the summation problem. By this is meant the problem of explicitly working out, in a convenient and practical way, the integral or sum in (2.17). This kind of problem is to be contrasted with the so-called boundary value problem that we shall treat later; this latter type is much more difficult, essentially because ρ(r′) is not known in advance.

To keep perspective here, we remark that although (2.17) represents a general solution to the problem of calculating the potential, and hence the field, of any distribution, there may be other, more convenient, methods of solution. For example, it may sometimes be easier to use directly Poisson’s equation, the differential counterpart of (2.17), than to try to evaluate the integral. Or, with charge configurations of considerable symmetry, Gauss’ Law may give a quicker and easier result. And for certain idealized distributions that extend to infinity (two-dimensional or one-dimensional distributions), there may be convergence problems with the integral. But with these minor qualifications, Eq. (2.17) stands as a most useful and general expression.

We shall discuss several kinds of distribution, without being able to treat all the interesting distributions that occur in the various branches of physics. As a small preview, we mention some examples. A charge may be paired with an equal and opposite one at a small distance, forming in the limit of zero distance a so-called point dipole. A point dipole may then be paired with another of the opposite sense to produce a point quadrupole, which may be paired to form point octupoles; the complete generalization of such concepts is embodied in an important general theory of the multipole expansion. Charge may be distributed on or throughout long slender cylinders in two-dimensional distribution, where quantities