MIT OpenCourseWare http://ocw.mit.

edu

Electromagnetic Field Theory: A Problem Solving Approach

For any use or distribution of this textbook, please cite as follows:

Markus Zahn, Electromagnetic Field Theory: A Problem Solving

Approach. (Massachusetts Institute of Technology: MIT
OpenCourseWare). http://ocw.mit.edu (accessed MM DD, YYYY).
License: Creative Commons Attribution-NonCommercial-Share Alike.

For more information about citing these materials or our Terms of

Use, visit: http://ocw.mit.edu/terms.
chapter 1

review of vector analysis

 2 Review of Vector Analysis

Electromagnetic field theory is the study of forces between

charged particles resulting in energy conversion or signal transmis-
sion and reception. These forces vary in magnitude and direction
with time and throughout space so that the theory is a heavy user
of vector, differential, and integral calculus. This chapter presents
a brief review that highlights the essential mathematical tools
needed throughout the text. We isolate the mathematical details
here so that in later chapters most of our attention can be devoted
to the applications of the mathematics rather than to its
development. Additional mathematical material will be presented
as needed throughout the text.

1-1 COORDINATE SYSTEMS

A coordinate system is a way of uniquely specifying the

location of any position in space with respect to a reference
origin. Any point is defined by the intersection of three
mutually perpendicular surfaces. The coordinate axes are
then defined by the normals to these surfaces at the point. Of
course the solution to any problem is always independent of
the choice of coordinate system used, but by taking advantage
of symmetry, computation can often be simplified by proper
choice of coordinate description. In this text we only use the
familiar rectangular (Cartesian), circular cylindrical, and
spherical coordinate systems.

1-1-1 Rectangular (Cartesian) Coordinates

The most common and often preferred coordinate system

is defined by the intersection of three mutually perpendicular
planes as shown in Figure 1-la. Lines parallel to the lines of
intersection between planes define the coordinate axes
(x, y, z), where the x axis lies perpendicular to the plane of
constant x, the y axis is perpendicular to the plane of constant
y, and the z axis is perpendicular to the plane of constant z.
Once an origin is selected with coordinate (0, 0, 0), any other
point in the plane is found by specifying its x-directed, y-
directed, and z-directed distances from this origin as shown
for the coordinate points located in Figure 1-lb.

I
 CoordinateSystems 3

T(-2,2,3)

-3 I
-3 -2 -1
2, 2)? I I I
I . i .
2 3 4
-
(b1

(b)

xdz

dS, =

Figure 1-1 Cartesian coordinate system. (a) Intersection of three mutually perpen-
dicular planes defines the Cartesian coordinates (x,y, z). (b)A point is located in space
by specifying its x-, y- and z-directed distances from the origin. (c) Differential volume
and surface area elements.

By convention, a right-handed coordinate system is always

used whereby one curls the fingers of his or her right hand in
the direction from x to y so that the forefinger is in the x
direction and the middle finger is in the y direction. The
thumb then points in the z direction. This convention is
necessary to remove directional ambiguities in theorems to be
derived later.
Coordinate directions are represented by unit vectors i., i,
and i2 , each of which has a unit length and points in the
direction along one of the coordinate axes. Rectangular
coordinates are often the simplest to use because the unit
vectors always point in the same direction and do not change
direction from point to point.
A rectangular differential volume is formed when one
moves from a point (x, y, z) by an incremental distance dx, dy,
and dz in each of the three coordinate directions as shown in
4 Review of VectorAnalysis

Figure 1-Ic. To distinguish surface elements we subscript the

area element of each face with the coordinate perpendicular
to the surface.

1-1-2 CircularCylindrical Coordinates

The cylindrical coordinate system is convenient to use

when there is a line of symmetry that is defined as the z axis.
As shown in Figure 1-2a, any point in space is defined by the
intersection of the three perpendicular surfaces of a circular
cylinder of radius r, a plane at constant z, and a plane at
constant angle 4 from the x axis.
The unit vectors i,, i6 and iz are perpendicular to each of
these surfaces. The direction of iz is independent of position,
but unlike the rectangular unit vectors the direction of i, and i6
change with the angle 0 as illustrated in Figure 1-2b. For
instance, when 0 = 0 then i, = i, and i# = i,, while if = ir/2,
then i, = i, and i# = -ix.
By convention, the triplet (r, 4, z) must form a right-
handed coordinate system so that curling the fingers of the
right hand from i, to i4 puts the thumb in the z direction.
A section of differential size cylindrical volume, shown in
Figure 1-2c, is formed when one moves from a point at
coordinate (r, 0, z) by an incremental distance dr, r d4, and dz
in each of the three coordinate directions. The differential
volume and surface areas now depend on the coordinate r as
summarized in Table 1-1.

Table 1-1 Differential lengths, surface area, and volume elements for
each geometry. The surface element is subscripted by the coordinate
perpendicular to the surface

CARTESIAN CYLINDRICAL SPHERICAL

dl=dx i+dy i,+dz i, dl=dri,+r d0 i#+dz i, dl=dri,+rdOis

+ r sin 0 do i,
dS. = dy dz dSr = r dO dz dS, = r 9 sin 0 dO d4
dS, = dx dz dS, = drdz dS@ = r sin Odr d4
dS, = dx dy dS, = r dr do dS, = rdrdO
dV=dxdydz dV= r dr d4 dz dV=r 2 sin drdO d

1-1-3 Spherical Coordinates

A spherical coordinate system is useful when there is a

point of symmetry that is taken as the origin. In Figure 1-3a
we see that the spherical coordinate (r,0, 0) is obtained by the
intersection of a sphere with radius r, a plane at constant
 CoordinateSystems 5

(b)
V= rdrdodz

(c)
Figure 1-2 Circular cylindrical coordinate system. (a) Intersection of planes of
constant z and 4 with a cylinder of constant radius r defines the coordinates (r, 4, z).
(b) The direction of the unit vectors i, and i, vary with the angle 46. (c) Differential
volume and surface area elements.

angle 4 from the x axis as defined for the cylindrical coor-

dinate system, and a cone at angle 0 from the z axis. The unit
vectors i,, is and i# are perpendicular to each of these sur-
faces and change direction from point to point. The triplet
(r, 0, 4) must form a right-handed set of coordinates.
The differential-size spherical volume element formed by
considering incremental displacements dr, rdO, r sin 0 d4
 6 Review of Vector Analysis

Al = r sin dr do
I llU
I V I

Figure 1-3 Spherical coordinate system. (a) Intersection of plane of constant angle 0
with cone of constant angle 0 and sphere of constant radius r defines the coordinates
(r, 0, 4). (b) Differential volume and surface area elements.

i
 Vector Algebra 7

Table 1-2 Geometric relations between coordinates and unit vectors for
Cartesian, cylindrical, and spherical coordinate systems*

CAR' FESIAN CYLINDRICAL SPHERICAL

x = r cos 4 = r sin 0 cos 4
y = r sin 4 = r sin 0 sin4
z = z = r cos 0
i. = coso i- sin oi& = sin 0 cos i, +cos 0 cos Oio -sin )i,#
= sin i,+cos
r 4 i, = sin 0 sin 4i, +cos Osin 4 i, +cos 4 it
= cos Oir- sin Oie
CYLINDRICAL CARTESIAN SPHERICAL
r = F+Yl = r sin 0

= tan'
x
Z z = r cos 0
ir = cos 4i. +sin 0i, = sin Oi, + cos Oi,
i* = -sin i,. +cos 0i, = i"
ii = i =- cos 0i, - sin Oi,
SPHERICAL CARTESIAN CYLINDRICAL
r = (x +y + 2 =
cos-1 z -1 os
8 cos /x +y +z - cOs

cot- -
y
sin 0 cos Ai. + sin 0 sin 4)i, + cos Oi, = sin 0i,+cos 0i,
cos 0 cos 4i. + cos 0 sin $i, - sin Oi, = cos Oir- sin Oi,
-sin 4i, +cos Oi, = i,

* Note that throughout this text a lower case roman r is used for the cylindrical radial coordinate
while an italicized r is used for the spherical radial coordinate.
from the coordinate (r, 0, 46) now depends on the angle 0 and
the radial position r as shown in Figure 1-3b and summarized
in Table 1-1. Table 1-2 summarizes the geometric relations
between coordinates and unit vectors for the three coordinate
systems considered. Using this table, it is possible to convert
coordinate positions and unit vectors from one system to
another.

1-2 VECTOR ALGEBRA

1-2-1 Scalars and Vectors

A scalar quantity is a number completely determined by its
magnitude, such as temperature, mass, and charge, the last
8 Review of Vector Analysis

being especially important in our future study. Vectors, such

as velocity and force, must also have their direction specified
and in this text are printed in boldface type. They are
completely described by their components along three coor-
dinate directions as shown for rectangular coordinates in
Figure 1-4. A vector is represented by a directed line segment
in the direction of the vector with its length proportional to its
magnitude. The vector
A = A~i. +A,i, +Ai,
in Figure 1-4 has magnitude
2
A=IA =[A. +A +AI
Note that each of the components in (1) (A., A,, and A,) are
themselves scalars. The direction of each of the components
is given by the unit vectors. We could describe a vector in any
of the coordinate systems replacing the subscripts (x, y, z) by
(r, 0, z) or (r, 0, 4); however, for conciseness we often use
rectangular coordinates for general discussion.

1-2-2 Multiplication of a Vector by a Scalar

If a vector is multiplied by a positive scalar, its direction

remains unchanged but its magnitude is multiplied by the

_AY

A = A I + Ayly+ AI1,

IAI=A =(A +A 2 +A.]'

Figure 1-4 A vector is described by its components along the three coordinate
directions.
 Vector Algebra 9

scalar. If the scalar is negative, the direction of the vector is

reversed:
aA = aAi. + aA,i, + aA i,

1-2-3 Addition and Subtraction

The sum of two vectors is obtained by adding their

components while their difference is obtained by subtracting
their components. If the vector B
B = B. i. +B,i, +B, i,
is added or subtracted to the vector A of (1), the result is a
new vector C:
C= A+ B = (A. B, )i, + (A,+B,)i, + (A, B,)i, (5)
Geometrically, the vector sum is obtained from the
diagonal of the resulting parallelogram formed from A and B
as shown in Figure 1-5a. The difference is found by first

Y
AY + 8, A+B
A,
A I'1
I / I
I/ I
I
IB I

B- A, + B,

-B A+B
A
A
A-K

(b)
Figure 1-5 The sum and difference of two vectors (a) by finding the diagonal of the
parallelogram formed by the two vectors, and (b) by placing the tail of a vector at the
head of the other.
10 Review of Vector Analysis

drawing -B and then finding the diagonal of the paral-

lelogram formed from the sum of A and -B. The sum of the
two vectors is equivalently found by placing the tail of a vector
at the head of the other as in Figure 1-5b.
Subtraction is the same as addition of the negative of a
vector.

EXAMPLE 1-1 VECTOR ADDITION AND SUBTRACTION

Given the vectors

A=4i +4i,, B=i. + 8i,
find the vectors B*A and their magnitudes. For the
geometric solution, see Figure 1-6.

S=A+B
= 5i x + 12i,

= 4(i0, + y)

Figure 1-6 The sum and difference of vectors A and B given in Example 1-1.
 Vector Algebra 11

SOLUTION

Sum
S = A + B = (4 + 1)i, + (4 + 8)i, = 5i, + 12i,
S= [52+ 122] 2 = 13
Difference
D = B - A = (1 - 4)ix +(8- 4)i, = -3i, + 4i,
1/ 2
D = [(-3)2+42] =5

1-2-4 The Dot (Scalar) Product

The dot product between two vectors results in a scalar and

is defined as
A .B= AB cos 0 (6)

where 0 is the smaller angle between the two vectors. The

term A cos 0 is the component of the vector A in the direction
of B shown in Figure 1-7. One application of the dot product
arises in computing the incremental work dW necessary to
move an object a differential vector distance dl by a force F.
Only the component of force in the direction of displacement
contributes to the work
dW=F dl (7)

The dot product has maximum value when the two vectors
are colinear (0 = 0) so that the dot product of a vector with
itself is just the square of its magnitude. The dot product is
zero if the vectors are perpendicular (0 = ir/2). These prop-
erties mean that the dot product between different orthog-
onal unit vectors at the same point is zero, while the dot

Y A

Figure 1-7 The dot product between two vectors.

12 Review of Vector Analysis

product between a unit vector and itself is unity

i. *i. = 1, i. - i, = 0
i, i, 1, i i, = 0 (8)
i, i, 1, i, i, =0

Then the dot product can also be written as

A B = (A,i, +A,i, +A,i,) (B.i, +B,i, + Bi,)

= A.B. + A,B, + A,B. (9)

From (6) and (9) we see that the dot product does not
depend on the order of the vectors

A-B=B-A (10)

By equating (6) to (9) we can find the angle between vectors as

B. + AB, + A,B,
ACos (11)
cos 0=
8= (11)
AB

Similar relations to (8) also hold in cylindrical and spherical

coordinates if we replace (x, y, z) by (r, 4, z) or (r, 0, 4). Then
(9) to (11) are also true with these coordinate substitutions.

EXAMPLE 1-2 DOT PRODUCT

Find the angle between the vectors shown in Figure 1-8,

A = r3 i. +i,, B = 2i.

A B = 2V,

Figure 1-8 The angle between the two vectors A and B in Example 1-2 can be found
using the dot product.
 Vector Algebra

SOLUTION

From (11)
A,B, 4
cos 0 = 2 22 =--
[A +A,] B. 2

0 = cos -= 300
2

1-2-5 The Cross (Vector) Product

The cross product between two vectors A x B is defined as a

vector perpendicular to both A and B, which is in the direc-
tion of the thumb when using the right-hand rule of curling
the fingers of the right hand from A to B as shown in Figure
1-9. The magnitude of the cross product is

JAxBI =AB sin 0

where 0 is the enclosed angle between A and B. Geometric-
ally, (12) gives the area of the parallelogram formed with A
and B as adjacent sides. Interchanging the order of A and B
reverses the sign of the cross product:

AxB= -BxA

AxB

BxA=-AxB

Figure 1-9 (a) The cross product between two vectors results in a vector perpendic-
ular to both vectors in the direction given by the right-hand rule. (b) Changing the
order of vectors in the cross product reverses the direction of the resultant vector.
14 Review of Vector Analysis

The cross product is zero for colinear vectors (0 = 0) so that

the cross product between a vector and itself is zero and is
maximum for perpendicular vectors (0 = ir/2). For rectan-
gular unit vectors we have
i. x i. = 0, i. x i, = i', iv x i = - i
i, x x i, =
i=, i, xi, = -i. (14)
i, x i, = 0, i"Xi. = i,
i, Xi"= -iy
These relations allow us to simply define a right-handed
coordinate system as one where
ix i,=(15)
Similarly, for cylindrical and spherical coordinates, right-
handed coordinate systems have
irXi =i,, i, xiO = i (16)
The relations of (14) allow us to write the cross product
between A and B as
Ax B = (A.i. +A,i, +Ai,)x (B,i1 + B,i, + Bi , )
= i,(AB, - AB,) + i,(A.B. - A.B.) + i,(AxB, - AB,)
(17)
which can be compactly expressed as the determinantal
expansion
i, i, iz
AxB=det A. A, A,
B. B, B.
= i, (A,B - AB,) + i,(A,B, - AB,) + i, (AB, - AB.)
(18)
The cyclical and orderly permutation of (x, y, z) allows easy
recall of (17) and (18). If we think of xyz as a three-day week
where the last day z is followed by the first day x, the days
progress as
xyzxyzxyzxyz . (19)

where the three possible positive permutations are under-

lined. Such permutations of xyz in the subscripts of (18) have
positive coefficients while the odd permutations, where xyz do
not follow sequentially
xzy, yxz, zyx (20)
have negative coefficients in the cross product.
In (14)-(20) we used Cartesian coordinates, but the results
remain unchanged if we sequentially replace (x, y, z) by the
 Vector Algebra 15

cylindrical coordinates (r, 0, z) or the spherical coordinates

(r, 0, 0).

EXAMPLE 1-3 CROSS PRODUCT

Find the unit vector i,, perpendicular in the right-hand

sense to the vectors shown in Figure 1-10.

A= -i.+i,+i,, B=i.-i,+i

What is the angle between A and B?

SOLUTION

The cross product Ax B is perpendicular to both A and B

i, i, i,
AxB=det-1 1 1 =2(i,+i,)
1 -1 1

The unit vector in is in this direction but it must have a

magnitude of unity

i AxB 1
A- = (ix+i,)

B= i,

Figure .1-10 The cross product between the two vectors in Example 1-3.
16 Review of Vector Analysis

The angle between A and B is found using (12) as

sin
=AxBI 21
sin 0
AB %/
=-22 0 = 70.5* or 109.50
The ambiguity in solutions can be resolved by using the dot
product of (11)
A'B -1
cos 0 -B~0
- = 109.5
AB 3

1-3 THE GRADIENT AND THE DEL OPERATOR

1-3-1 The Gradient

Often we are concerned with the properties of a scalar field

f(x, y, z) around a particular point. The chain rule of differ-
entiation then gives us the incremental change df in f for a
small change in position from (x, y, z) to (x + dx, y + dy, z +dz):
af af af
df = -dx + dy + dz (1)
ax Oy Oz

If the general differential distance vector dl is defined as

dl=dx i.+dy i,+dz i, (2)

(1) can be written as the dot product:

df = a-f i, + -- i, + - i)dl
ax ay az
= grad f - dl (3)
where the spatial derivative terms in brackets are defined as
the gradient of f:

grad f = Vf=V i + fi,+ fi (4)

Ox ay az

The symbol V with the gradient term is introduced as a

general vector operator, termed the del operator:

V = ix-+i,-7 +is (5)

ax ay Oz
By itself the del operator is meaningless, but when it premul-
tiplies a scalar function, the gradient operation is defined. We
will soon see that the dot and cross products between the del
operator and a vector also define useful operations.
 The Gradientand the Del Operator

With these definitions, the change in f of (3) can be written

as
df = Vf - dl = IVf dl cos 0 (6)
where 0 is the angle between Vf and the position vector dl.
The direction that maximizes the change in the function f is
when dl is colinear with Vf(O = 0). The gradient thus has the
direction of maximum change in f. Motions in the direction
along lines of constant f have 0 = ir/2 and thus by definition
df = 0.

1-3-2 Curvilinear Coordinates

(a) Cylindrical

The gradient of a scalar function is defined for any coor-

dinate system as that vector function that when dotted with dl
gives df. In cylindrical coordinates the differential change in
f(r, , z) is
af af af
df= dr+ d+-dz (7)
or 0o az

The differential distance vector is

dl= dri,+r do i6 +dz i. (8)

so that the gradient in cylindrical coordinates is

df = Vf dl=>Vf=- af laf af
+i+ i + (9)
Or r 4d az

(b) Spherical
Similarly in spherical coordinates the distance vector is

dl=dri,+rdOi,+rsin Odo i6 (10)

with the differential change of f(r, 0, 46) as

af af af
df = -dr+-dO+- d4 = Vf -dl (11)
Or 0o d4

Using (10) in (11) gives the gradient in spherical coordinates

as
Af Iaf 1 af
Vf=-T+Af
IIi + Ia' i4 +
+ I
iII
i
Ua I U -
r r 80 r s n 0 ad
18 Review of Vector Analysis

EXAMPLE 1-4 GRADIENT

Find the gradient of each of the following functions where

a and b are constants:
(a) f = ax2 y +byst
SOLUTION

af.
- af*4f
Vf = -a
x
i,+-ay i, + -8z i.
8
= 2axyi. + (ax + 3byz)i, + bysi,

(b) f= ar2 sin q +brz cos 24

SOLUTION

S+af . af
Vf = -a ir+ If i +-f i,
ar r a4 az
= (2ar sin 4 + bz cos 20)i,
+ (arcos 4 - 2bz sin 20)i, + br cos 20i,

(c) f =a +br sin 0 cos 4

r
SOLUTION

af. 1 af 1 af
ar rO
r r sin 0 a8

= --+bsin 0 cos 4)i,+bcos 0 cos - sini

1-3-3 The Line Integral

In Section 1-2-4 we motivated the use of the dot product

through the definition of incremental work as depending
only on the component pf force F in the direction of an
object's differential displacement dl. If the object moves along
a path, the total work is obtained by adding up the incremen-
tal works along each small displacement on the path as in
Figure 1-11. If we break the path into N small displacements

I~
 The Gradientand the Del Operator

dl 6

Li wl ai, N N
W dW, F - di,
n 1 n= 1
lim
=
din 0 W F dl
N - fJ
L

Figure 1-11 The total work in moving a body over a path is approximately equal to
the sum of incremental works in moving the body each small incremental distance dl.
As the differential distances approach zero length, the summation becomes a line
integral and the result is exact.

dl, d1 2, . ., dN, the work performed is approximately

W -F *dl +F
1 2 'dl 2 + F 3 di+ +FN dlN
N
Y F, *dl, (13)
n=l

The result becomes exact in the limit as N becomes large with

each displacement dl, becoming infinitesimally small:

N
W= lim Y Fn.dl,= F dl (14)
N-o n=1
dl,,--O

In particular, let us integrate (3) over a path between the

two points a and b in Figure 1-12a:

Sdf = fi,-fl.= Vf dl (15)

Because df is an exact differential, its line integral depends

only on the end points and not on the shape of the contour
itself. Thus, all of the paths between a and b in Figure 1-12a
have the same line integral of Vf, no matter what the function
f may be. If the contour is a closed path so that a = b, as in
20 Review of Vector Analysis

IJ
0
Y
2 b
I
4
2 31. Vf" di = f(b) - f(a)

a
1

(a)
Figure 1-12 The component of the gradient of a function integrated along a line
contour depends only on the end points and not on the contour itself. (a) Each of the
contours have the same starting and ending points at a and b so that they all have the
same line integral of Vf. (b) When all the contours are closed with the same beginning
and ending point at a, the line integral of Vf is zero. (c) The line integral of the
gradient of the function in Example (1-5) from the origin to the point P is the same for
all paths.

Figure 1-12b, then (15) is zero:

Lvf dl=L.-f.=0 (16)

where we indicate that the path is closed by the small circle in

the integral sign f. The line integral of the gradient of a
function around a closed path is zero.

EXAMPLE 1-5 LINE INTEGRAL

For f=x 2y, verify (15) for the paths shown in Figure 1-12c
between the origin and the point P at (xo, yo).

SOLUTION

The total change in f between 0 and P is

df f -flo = xoYo

From the line integral along path 1 we find

Yo 0o xo 2
Vf -dl= f dy+f= xoyo
Y_0 Ir X=0 ax
 Flux and Divergence 21

Similarly, along path 2 we also obtain

Vf dl= o &o,+ o dy_ xoyo

while along path 3 we must relate x and y along the straight

line as
y = Yo x dy = Yodx
xo xo

to yield

Jo \ax dx +.ay dy) = Jx.=o Xo

vf *dl= I(- o dx = xOyo

1-4 FLUX AND DIVERGENCE

If we measure the total mass of fluid entering the volume in

Figure 1-13 and find it to be less than the mass leaving, we
know that there must be an additional source of fluid within
the pipe. If the mass leaving is less than that entering, then

Flux in = Flux out

Flux in < Flux out

"----/ -, -- -----

Source

Flux in > Flux out

Sink

Figure 1-13 The net flux through a closed surface tells us whether there is a source or
sink within an enclosed volume.
22 Review of Vector Analysis

there is a sink (or drain) within the volume. In the absence of

sources or sinks, the mass of fluid leaving equals that entering
so the flow lines are continuous. Flow lines originate at a
source and terminate at a sink.

1-4.1 Flux

We are illustrating with a fluid analogy what is called the

flux 4 of a vector A through a closed surface:

D= f AdS (1)

The differential surface element dS is a vector that has

magnitude equal to an incremental area on the surface but
points in the direction of the outgoing unit normal n to the
surface S, as in Figure 1-14. Only the component of A
perpendicular to the surface contributes to the flux, as the
tangential component only results in flow of the vector A
along the surface and not through it. A positive contribution
to the flux occurs if A has a component in the direction of dS
out from the surface. If the normal component of A points
into the volume, we have a negative contribution to the flux.
If there is no source for A within the volume V enclosed by
the surface S, all the flux entering the volume equals that
leaving and the net flux is zero. A source of A within the
volume generates more flux leaving than entering so that the
flux is positive (D > 0) while a sink has more flux entering than
leaving so that D< 0.
dS n dS

Figure 1-14 The flux of a vector A through the closed surface S is given by the
surface integral of the component of A perpendicular to the surface S. The differential
vector surface area element dS is in the direction of the unit normal n.
 Flux and Divergence 23

Thus we see that the sign and magnitude of the net flux
relates the quantity of a field through a surface to the sources
or sinks of the vector field within the enclosed volume.

1-4-2 Divergence

We can be more explicit about the relationship between the

rate of change of a vector field and its sources by applying (1)
to a volume of differential size, which for simplicity we take to
be rectangular in Figure 1-15. There are three pairs of plane
parallel surfaces perpendicular to the coordinate axes so that
(1) gives the flux as

()= f A.(x) dy dz - A. (x -Ax) dy dz

+ JA,(y + Ay) dx dz - A, (y) dx dz

+j A(z +Az) dxdy- A,(z)dxdy (2)

where the primed surfaces are differential distances behind

the corresponding unprimed surfaces. The minus signs arise
because the outgoing normals on the primed surfaces point in
the negative coordinate directions.
Because the surfaces are of differential size, the
components of A are approximately constant along each
surface so that the surface integrals in (2) become pure

dS, = Ax Ay

Ay Aze

Figure 1-15 Infinitesimal rectangular volume used to define the divergence of a

vector.
24 Review of Vector Analysis

multiplications of the component of A perpendicular to the

surface and the surface area. The flux then reduces to the form

( [Ax(x)-Ax(x -Ax)] + [A,(y +Ay)-A,(y)]

Ax Ay

+ Ax Ay Az (3)

We have written (3) in this form so that in the limit as the

volume becomes infinitesimally small, each of the bracketed
terms defines a partial derivative
aA 3A AMz
lim (= + + AV (4)
ax-o \ax ay az

where AV = Ax Ay Az is the volume enclosed by the surface S.

The coefficient of AV in (4) is a scalar and is called the
divergence of A. It can be recognized as the dot product
between the vector del operator of Section 1-3-1 and the
vector A:

div A = V - A = -+
aAx aA, + aA, (5)
ax ay az

1-4-3 Curvilinear Coordinates

In cylindrical and spherical coordinates, the divergence

operation is not simply the dot product between a vector and
the del operator because the directions of the unit vectors are
a function of the coordinates. Thus, derivatives of the unit
vectors have nonzero contributions. It is easiest to use the
generalized definition of the divergence independent of the
coordinate system, obtained from (1)-(5) as

V A= lim 'A.dS (6)

v-,o AV

(a) Cylindrical Coordinates

In cylindrical coordinates we use the small volume shown in
Figure 1-16a to evaluate the net flux as

D= sA dS =f (r+Ar)Ar,,, , dk dz - rArr,, ddz

+ I A4,. dr dz - A6, dr dz

r
+ i rA,,+. dr db - I rA,,= dr db
.J3 J13
 Flux and Divergence

2
dS, = (r + Ar) sin0dO do

/ dS = r dr dO

Figure 1-16 Infinitesimal volumes used to define the divergence of a vector in

(a) cylindrical and (b) spherical geometries.

Again, because the volume is small, we can treat it as approx-

imately rectangular with the components of A approximately
constant along each face. Then factoring out the volume
A V= rAr A4 Az in (7),

([(r + Ar)A ,.l+A,- rA r,]

r\ Ar
[A. -A.1] [A
+ A.
AAA + .A, .... ) rAr AO Az
r oyr u*

M M
26 Review of Vector Analysis

lets each of the bracketed terms become a partial derivative as

the differential lengths approach zero and (8) becomes an
exact relation. The divergence is then
"A*dS la 1AA + A,
V A=lim =--(rAr)+-- + (9)
ar&o AV r r r 84 8z
Az-.0

(b) Spherical Coordinates

Similar operations on the spherical volume element AV=
r 9 sin 0 Ar AO A4 in Figure 1-16b defines the net flux through
the surfaces:

4= A. dS

[( r)2 A,,+A, - r2 AA,j]

([(r + r2 Ar

+ [AA.A, sin (0 +AO)-A e, sin 8]

r sin 0 AO

+ [AA+. - A ] r2 sin 0 Ar AO A (10)

r sin 0 A4
The divergence in spherical coordinates is then

sA *dS
V. A= lim
a-.o AV

1Ir a (r'A,)+ 1 8 (As sin 0)+ BA,

A (
(11)
r ar r sin 0 80 r sin 0 a

1-4-4 The Divergence Theorem

If we now take many adjoining incremental volumes of any

shape, we form a macroscopic volume V with enclosing sur-
face S as shown in Figure 1-17a. However, each interior
common surface between incremental volumes has the flux
leaving one volume (positive flux contribution) just entering
the adjacent volume (negative flux contribution) as in Figure
1-17b. The net contribution to the flux for the surface integral
of (1) is zero for all interior surfaces. Nonzero contributions
to the flux are obtained only for those surfaces which bound
the outer surface S of V. Although the surface contributions
to the flux using (1) cancel for all interior volumes, the flux
obtained from (4) in terms of the divergence operation for
 Flux and Divergence 27

n1 -n 2

Figure 1-17 Nonzero contributions to the flux of a vector are only obtained across
those surfaces that bound the outside of a volume. (a) Within the volume the flux
leaving one incremental volume just enters the adjacent volume where (b) the out-
going normals to the common surface separating the volumes are in opposite direc-
tions.

each incremental volume add. By adding all contributions

from each differential volume, we obtain the divergence
theorem:

QO=cIAdS= lim (V*A)AV.=f VAdV (12)

A V.-.O

where the volume V may be of macroscopic size and is

enclosed by the outer surface S. This powerful theorem con-
verts a surface integral into an equivalent volume integral and
will be used many times in our development of electromag-
netic field theory.

EXAMPLE 1-6 THE DIVERGENCE THEOREM

Verify the divergence theorem for the vector

A = xi. +yi, +zi, = ri,
by evaluating both sides of (12) for the rectangular volume
shown in Figure 1-18.

SOLUTION

The volume integral is easier to evaluate as the divergence

of A is a constant
V A = A.+ aA, A=
ax ay az
28 Review of Vector Analysis

Figure 1-18 The divergence theorem is verified in Example 1-6 for the radial vector
through a rectangular volume.

(In spherical coordinates V A= (1/r )(8ar)(/r)(r)=3) so that

the volume integral in (12) is

vV A dV= 3abc
The flux passes through the six plane surfaces shown:

q=fA-dS= jj(a dydz- AJO) dydz

a 0

+A ,(b)dx dz- A, dx dz
/O

c 0

which verifies the divergence theorem.

1.5 THE CURL AND STOKES' THEOREM

1-5-1 Curl

We have used the example of work a few times previously

to motivate particular vector and integral relations. Let us do
so once again by considering the line integral of a vector
 The Curl and Stokes' Theorem 29

around a closed path called the circulation:

C= A dl (1)

where if C is the work, A would be the force. We evaluate (1)

for the infinitesimal rectangular contour in Figure 1-19a:

C= A,(y)dx+ A,(x+Ax)dy+ Ax(y+Ay)dx

I 3

+ A,(x) dy (2)

The components of A are approximately constant over each

differential sized contour leg so that (2) is approximated as

S([A(y)- (y +Ay)] + [A,(x + Ax)- A,(x)] (3)

j x y

x. y)

(a)

Figure 1-19 (a) Infinitesimal rectangular contour used to define the circulation.
(b)The right-hand rule determines the positive direction perpendicular to a contour.
30 Review of Vector Analysis

where terms are factored so that in the limit as Ax and Ay

become infinitesimally small, (3) becomes exact and the
bracketed terms define partial derivatives:

lim C= aA ) AS (4)
ax-o ax ay
AS.-AxAy

The contour in Figure 1-19a could just have as easily been

in the xz or yz planes where (4) would equivalently become

C = \( 'AS.
z" (yz plane)
ay a.,
C= \ AS, (xz plane) (5)

by simple positive permutations of x, y, and z.

The partial derivatives in (4) and (5) are just components of
the cross product between the vector del operator of Section
1-3-1 and the vector A. This operation is called the curl of A
and it is also a vector:
i, i, i,

curl A= Vx A= det
ax ay az
A. A, A,
a. _A,
A *(a aAM\
tA az )
ay az ax
aAzx
+aA, aA) (6)

The cyclical permutation of (x, y, z) allows easy recall of (6) as

described in Section 1-2-5.
In terms of the curl operation, the circulation for any
differential sized contour can be compactly written as

C= (VxA)- dS (7)
where dS = n dS is the area element in the direction of the
normal vector n perpendicular to the plane of the contour in
the sense given by the right-hand rule in traversing the
contour, illustrated in Figure 1-19b. Curling the fingers on
the right hand in the direction of traversal around the
contour puts the thumb in the direction of the normal n.
For a physical interpretation of the curl it is convenient to
continue to use a fluid velocity field as a model although the
general results and theorems are valid for any vector field. If
 The Curl and Stokes' Theorem
Theorem
Stokes'
and
Curl
The

- --

-- - - - - - - - - -
- --
*- =-------
---.----- -_---_*-_---------
- - -- - - - - -

No circulation Nonzero circulation

Figure 1-20 A fluid with a velocity field that has a curl tends to turn the paddle wheel.
The curl component found is in the same direction as the thumb when the fingers of
the right hand are curled in the direction of rotation.

a small paddle wheel is imagined to be placed without dis-

turbance in a fluid flow, the velocity field is said to have
circulation, that is, a nonzero curl, if the paddle wheel rotates
as illustrated in Figure 1-20. The curl component found is in
the direction of the axis of the paddle wheel.

1-5-2 The Curl for Curvilinear Coordinates

A coordinate independent definition of the curl is obtained

using (7) in (1) as

fA~dl
(V x A), = lim (8)
dS.-. dS.

where the subscript n indicates the component of the curl

perpendicular to the contour. The derivation of the curl
operation (8) in cylindrical and spherical. coordinates is
straightforward but lengthy.

(a) Cylindrical Coordinates

To express each of the components of the curl in cylindrical
coordinates, we use the three orthogonal contours in Figure
1-21. We evaluate the line integral around contour a:

A dl= A,(4) dz + A,(z -&Az) r d4

+ A( +A)r dz + A(z) r d4

([A,( +A4)-A,(,)] _[A#(z)-A#(z -Az)] r Az

rA AZA jr

M M
 32 Review of Vector Analysis

(r-Ar,

(rO + AO, -AZ)

(V x A),

Figure 1-21 Incremental contours along cylindrical surface area elements used to
calculate each component of the curl of a vector in cylindrical coordinates.

to find the radial component of the curl as

l
aA aAA
(V x A)r = lim (10)
a,-o ,&
r Az r a4 az
Az l around contour b:

We evaluate the line integral around contour b:

r z -As r-Ar

A* dl = Ar A,(z)dr+ zA
dz
A,(r)dz + j Ar(z - Az)dr
+ -Ar -,r)
+j Az(r-Ar) dz

[A,(z)-A,(z -Az)] [A(r)-A(r- Ar)] Ar

Az Ar (11)

I
 The Curl and Stokes' Theorem 33

to find the 4 component of the curl,

A dl OAr aa
(V x A) = li = (12)
Ar-.o Ar Az az ar
Az 0

The z component of the curl is found using contour c:

r
S rrr--d "+A4 r
A -dl= Arlo dr+ rAld4+ A,,,,. dr

+ A(r - r)A,-. d b

[rAp,-(r-Ar)A4,_-,] [Arl4$]-Arlb]rAr
rAr rA ]
(13)
to yield

A - dl
______1 8 aA\
(Vx A), = lim =C - (rA) - ) (14)
Ar-O rAr AO rr 84

The curl of a vector in cylindrical coordinates is thus

IM(A, aA aA aA
VxA= =
)ir+(A A.
r a4 az az r

+ (rA) a)i, (15)

r ar 84

(b) Spherical Coordinates

Similar operations on the three incremental contours for
the spherical element in Figure 1-22 give the curl in spherical
coordinates. We use contour a for the radial component of
the curl:

A dl=J+A A4,r sin 0 d + - rAI,., dO

+i r sin ( -AO)A ,. 4 d + rA, dO

.+A -AO

[Ad. sin 0 - A4._.. sin (0- AO)]

rsin 0 AO
-[A,.,-A_+
r2 sin 0A A4 (16)
r sin 0 AO
 34 Review of Vector Analysis

r sin (0 - AO)

(r,0- AO,

Figure 1-22 Incremental contours along spherical surface area elements used to
calculate each component of the curl of a vector in spherical coordinates.

to obtain

I A dl(Ain
(V x A), = lim r(A.I sin )
a:o r sin 0 AO AO r sin 0 O
(17)
The 0 component is found using contour b:

J
- A
r r 4+A
A dl= A,0.dr+ (r - Ar)A%,_, sin 0 d

+ A.... dr + rA$, sin 0 d4

-A +A4

\ r sin 0 Ab

[rA4-(r-Ar)A_.]~ r sin 0 Ar A4
r Ar

I
 The Curl and Stokes' Theorem 35

as

1 1 aA, a (rA
(Vx A)@ = lim
r o)
a,-o r sin 0 Ar A r sin 08 a r
(19)
The 4 component of the curl is found using contour c:

fA dl= e-A, rA + r
+dO A,,,dr
-Ao r
+ (r -Ar)AoA dO + A,,,_ ,dr

([rA,, - (r-Ar)Ae.al ] [A,1 -Ar,_-,]) r Ar AO

\ r Ar r AO
(20)
as

Ad 1 a BA
(Vx A), = lim -(rAe) - (21)
Ar-.o r Ar AO r aOr 801)

The curl of a vector in spherical coordinates is thus given

from (17), (19), and (21) as
1 aA
VxA= I (A.sin 0)- i,

+- -r(rA)- i (22)

1-5-3 Stokes' Theorem

We now piece together many incremental line contours of

the type used in Figures 1-19-1-21 to form a macroscopic
surface S like those shown in Figure 1-23. Then each small
contour generates a contribution to the circulation
dC = (V x A) *dS (23)
so that the total circulation is obtained by the sum of all the
small surface elements
C
C= I(VxA)'dS
J's
36 Review of Vector Analysis

Figure 1-23 Many incremental line contours distributed over any surface, have
nonzero contribution to the circulation only along those parts of the surface on the
boundary contour L.

Each of the terms of (23) are equivalent to the line integral

around each small contour. However, all interior contours
share common sides with adjacent contours but which are
twice traversed in opposite directions yielding no net line
integral contribution, as illustrated in Figure 1-23. Only those
contours with a side on the open boundary L have a nonzero
contribution. The total result of adding the contributions for
all the contours is Stokes' theorem, which converts the line
integral over the bounding contour L of the outer edge to a
surface integral over any area S bounded by the contour

A
*dl= J(Vx A)* dS (25)

Note that there are an infinite number of surfaces that are

bounded by the same contour L. Stokes' theorem of (25) is
satisfied for all these surfaces.

EXAMPLE 1-7 STOKES' THEOREM

Verify Stokes' theorem of (25) for the circular bounding

contour in the xy plane shown in Figure 1-24 with a vector
 The Curl and Stokes' Theorem 37

zi -1 ri - i

Figure 1-24 Stokes' theorem for the vector given in Example 1-7 can be applied to
any surface that is bounded by the same contour L.

field
A = -yi, +xi, -zi, = ri6 -zi,
Check the result for the (a) flat circular surface in the xy
plane, (b) for the hemispherical surface bounded by the
contour, and (c) for the cylindrical surface bounded by the
contour.

SOLUTION

For the contour shown

dl= R do i"
so that
A * dl= R 2 d4
where on L, r = R. Then the circulation is
2sr

C= A dl= o R2do=2n-rR 2

The z component of A had no contribution because dl was

entirely in the xy plane.
The curl of A is

VxA=i aA, =y2i,

ax ay
 38 Review of Vector Analysis

(a) For the circular area in the plane of the contour, we

have that

S(Vx A) *dS = 2 dS.= 2R

which agrees with the line integral result.

(b) For the hemispherical surface
v/2 2w

(Vx A) dS = 0 2i, iR sin 0 dOd

From Table 1-2 we use the dot product relation

i=i,= cos 0

which again gives the circulation as

/2 COs20 w/2
C= 2 sin 20 dO dO = -2wR
'2 20 /=2nR2
=1=1=o -o2 0e-o

(c) Similarly, for th'e cylindrical surface, we only obtain

nonzero contributions to the surface integral at the upper
circular area that is perpendicular to Vx A. The integral is
then the same as part (a) as V XA is independent of z.

1-5-4 Some Useful Vector Identities

The curl, divergence, and gradient operations have some

simple but useful properties that are used throughout the
text.

(a) The Curl of the Gradient is Zero [V x (Vf)= 0]

We integrate the normal component of the vector V x (Vf)
over a surface and use Stokes' theorem

JVx (Vf).dS= Vf dl= (26)

where the zero result is obtained from Section 1-3-3, that the
line integral of the gradient of a function around a closed
path is zero. Since the equality is true for any surface, the
vector coefficient of dS in (26) must be zero

vx(Vf)=O

The identity is also easily proved by direct computation

using the determinantal relation in Section 1-5-1 defining the

I
 Problems 39

curl operation:
i. i, i,

V x (Vf)= det a a
ax ay az
af af af
ax ay az
.=(-
L f. +,( af a/2f) + i(.I2- a2f).0
ayaz azay azax axaz axay ayax)
(28)
Each bracketed term in (28) is zero because the order of
differentiation does not matter.

(b) The Divergence of the Curl of a Vector is Zero

[V* (V x A)= 0]
One might be tempted to apply the divergence theorem to
the surface integral in Stokes' theorem of (25). However, the
divergence theorem requires a closed surface while Stokes'
theorem is true in general for an open surface. Stokes'
theorem for a closed surface requires the contour L to shrink
to zero giving a zero result for the line integral. The diver-
gence theorem applied to the closed surface with vector V X A
is then

VxA dS= IV> (VxA) dV= O V - (VxA) = 0

(29)
which proves the identity because the volume is arbitrary.
More directly we can perform the required differentiations
V. (VxA)
a,aA, aA,\ a aA aA) a aA, aA\
-ax\y az / y az ax I az\ ax ay /
2
xa A, a2 A , a2 A, A 2 a2 A, a2 Ax (
-\xay avax \ayaz azay) \azax ax(z
where again the order of differentiation does not matter.

PROBLEMS

Section 1-1
1. Find the area of a circle in the xy plane centered at the
origin using:
(a) rectangular coordinates x 2 +y = a2 (Hint:
J -x dx = A[x a2-x a 2 sin-'(x/a)l)