MITRES 6 002S08 Part2 PDF

chapter 4
electricfield boundary
value problems
258 Electric Field Boundary Value Problems
The electric field distribution due to external sources is

disturbed by the addition of a conducting or dielectric body
because the resulting induced charges also contribute to the
field. The complete solution must now also satisfy boundary
conditions imposed by the materials.
4-1 THE UNIQUENESS THEOREM
Consider a linear dielectric material where the permittivity

may vary with position:
D=e (r)E = -e (r)VV (1)
The special case of different constant permittivity media

separated by an interface has e (r) as a step function. Using (1)
in Gauss's law yields
V - [e(r)V V]= -pf (2)
which reduces to Poisson's equation in regions where e (r) is a

constant. Let us call V, a solution to (2).
The solution VL to the homogeneous equation
V - [e(r)VV= 0 (3)
which reduces to Laplace's equation when e(r) is constant,
can be added to V, and still satisfy (2) because (2) is linear in
the potential:
V - [e (r)V(V, + VL)] = V - [e (r)V V.]+V - [e (r)V V] = -pf
0 (4)
Any linear physical problem must only have one solution
yet (3) and thus (2) have many solutions. We need to find
what boundary conditions are necessary to uniquely specify
this solution. Our method is to consider two different solu
tions V, and V 2 for the same charge distribution
V (eVVi)=-p, V - (eVV2 )=-Pf (5)
so that we can determine what boundary conditions force

these solutions to be identical, V, = V 2
.
Boundary Value Problems in CartesianGeometries 259
The difference of these two solutions VT = V, - V 2 obeys

the homogeneous equation
V - (EVVT)=0 (6)
We examine the vector expansion
2
V-(eVTVVT)= VTV - (EVVT)+eVVT . VVT=6eVVT (7)
0
noting that the first term in the expansion is zero from (6) and
that the second term is never negative.
We now integrate (7) over the volume of interest V, which
may be of infinite extent and thus include all space
.V.-(eVVVT)dV= eVTVVT-dS= JIVVTI2dV (8)
The volume integral is converted to a surface integral over

the surface bounding the region using the divergence
theorem. Since the integrand in the last volume integral of (8)
is never negative, the integral itself can only be zero if VT is
zero at every point in the volume making the solution unique
(VT =0> V = V2). To force the volume integral to be zero,
the surface integral term in (8) must be zero. This requires
that on the surface S the two solutions must have the same
value (VI = V2) or their normal derivatives must be equal
[V V - n = V V 2 -n]. This last condition is equivalent to
requiring that the normal components of the electric fields be
equal (E = -V V).
Thus, a problem is uniquely posed when in addition to
giving the charge distribution, the potential or the normal
component of the electric field on the bounding surface sur
rounding the volume is specified. The bounding surface can
be taken in sections with some sections having the potential
specified and. other sections having the normal field
component specified.
If a particular solution satisfies (2) but it does not satisfy
the boundary conditions, additional homogeneous solutions
where pf =0, must be added so that the boundary conditions
are met. No matter how a solution is obtained, even if
guessed, if it satisfies (2) and all the boundary conditions, it is
the only solution.
4-2 BOUNDARY VALUE PROBLEMS IN CARTESIAN GEOMETRIES
For most of the problems treated in Chapters 2 and 3 we

restricted ourselves to one-dimensional problems where the
electric field points in a single direction and only depends on
that coordinate. For many cases, the volume is free of charge
so that the system is described by Laplace's equation. Surface
260 Electric FieldBoundary Value Problems
charge is present only on interfacial boundaries separating

dissimilar conducting materials. We now consider such
volume charge-free problems with two- and three dimen
sional variations.
4-2-1 Separation of Variables
Let us assume that within a region of space of constant

permittivity with no volume charge, that solutions do not
depend on the z coordinate. Then Laplace's equation reduces
to
a2V a2V
8x y -- y+ 2=0(1)
We try a solution that is a product of a function only of the x

coordinate and a function only of y:
V(x, y) = X(x) Y(y) (2)
This assumed solution is often convenient to use if the system

boundaries lay in constant x or constant y planes. Then along
a boundary, one of the functions in (2) is constant. When (2) is
substituted into (1) we have
Yd 2X d2Yy d 2X I d2'y
Y- +X =0,+2 = 0=>_ 0 (3)
S dy" X Yy
where the partial derivatives become total derivatives because

each function only depends on a single coordinate. The
second relation is obtained by dividing through by XY so that
the first term is only a function of x while the second is only a
function of y.
The only way the sum of these two terms can be zero for all
values of x and y is if each term is separately equal to a
constant so that (3) separates into two equations,
1 d 2X 2 1 d 2 Y_2
=k =-kY (4)
where k2 is called the separation constant and in general can

be a complex number. These equations can then be rewritten
as the ordinary differential equations:
d2 X dY 2
(5)
4-2-2 Zero Separation Constant Solutions
When the separation constant is zero (k2 =0) the solutions

to (5) are
X = alx +bl, Y=cly+dl (6)
where a,, b1 , cl, and dl are constants. The potential is given by

the product of these terms which is of the form
V= a2 +b 2x+c 2y+d 2xy (7)

The linear and constant terms we have seen before, as the
potential distribution within a parallel plate capacitor with no
fringing, so that the electric field is uniform. The last term we
have not seen previously.
(a) Hyperbolic Electrodes

A hyperbolically shaped electrode whose surface shape
obeys the equation xy = ab is at potential Vo and is placed
above a grounded right-angle corner as in Figure 4-1. The
.U.!
0.2
4~0 125*
Equipotential lines -
-
VO ab
Field lines
x
-
-r y2 -X2 = const.
Figure 4-1 The equipotential and field lines for a hyperbolically shaped electrode at
potential Vo above a right-angle conducting corner are orthogonal hyperbolas.
boundary conditions are

V(x =0)= 0, V(y = 0)=0, V(xy = ab)= Vo (8)
so that the solution can be obtained from (7) as
V(x, y) Voxy/(ab) (9)
The electric field is then
V
E=-VV= L 0 [yi.+xi,] (10)
ab
The field lines drawn in Figure 4-1 are the perpendicular
family of hyperbolas to the equipotential hyperbolas in (9):
dy
-= E. x
-=--> -x 22 =const (11)
dx E. y
(b) Resistor in an Open Box

A resistive medium is contained between two electrodes,
one of which extends above and is bent through a right-angle
corner as in Figure 4-2. We try zero separation constant
S -0.0 -

0.1 - - -- EO, a 0
03 - dx E s-y
0.4 _S
2
(X 1)2 = const.
0.5
0.6 \I
0.7
0.8
0.9 \ \
d
V++=++V
--------- pt - -

~-~---
~~--~----
--

Figure--- 4-- A- reitv eimpatal i= n pncnutn ox
solutions given by (7) in each region enclosed by the elec

trodes:
V={al+bix+cy+dixy, o!ysd (12)

a2 +b 2 x +c 2y +d 2xy, d:5 y 5 s
With the potential constrained on the electrodes and being

continuous across the interface, the boundary conditions are
V(x =0)= Vo= a 1+c 1y > a1 = Vo, c 1 =0 (0!5 y S d)
+b11+c y+djly>bj=-Vo/I, d 1 =O
V(x=l)= vo (O:sy d)
a2+bul+c2y+d2ly =>a 2 +b 2 l=0, c2 +d 2 l=0
(dSy ss)
V(y=s)=O=a2 +b 2x+c2s+d2xs =>a2 +c 2s=O, b 2 +d 2 s=O
70 70
V(y=d,)=V(y =d-)=a1+bx+1 d+,d1 xd
=a 2 +b 2 x +c 2 d +d 2xd (13)
=>a =Vo=a2 +c2d, b 1 = -Vo/l =b2 +d 2 d
so that the constants in (12) are

a 1 = V0 , bi=-V0 /l, c 1 =0, d1 =0
V0 V0
a2 = , b 2 =- (14)
(I - d/s) l(I - d/s)'
V0 V0
C2 =- d2 =
s(1 - d/s)' Is(1 - d/s)
The potential of (12) is then
VO( - x/1), 0:5y!5d (5

V - I- xy+-), d:yss
s -d ( s
SIs/
with associated electric field

V0
.
VOix, 0:5 y:5 d

E = -- V V = (16)
ss) +!- -x , d<y<s
s -d I s s 1)3
Note that in the dc steady state, the conservation of charge

boundary condition of Section 3-3-5 requires that no current
cross the interfaces at y = 0 and y = d because of the surround
ing zero conductivity regions. The current and, thus, the
electric field within the resistive medium must be purely

tangential to the interfaces, E,(y=d..)=E,(y=0+)=0. The
surface charge density on the interface at y = d is then due only
to the normal electric field above, as below, the field is purely
tangential:
(y=d_)=E V
Of(y=d)=EoE,(y=d)--E, /s-d - (17)
1)
The interfacial shear force is then
1 2
Fx= oyE.(y-d)wdx= w (18)

0 2(s-d)
If the resistive material is liquid, this shear force can be used

to pump the fluid.*
4-2-3 Nonzero Separation Constant Solutions
Further solutions to (5) with nonzero separation constant

(k 2 0 0) are
X = A sinhkx +A 2 cosh kx = B1 e*+ B 2 e
Y= C1 sin ky+C2 cos ky =D1 eiky+D 2 e-ky
When k is real, the solutions of X are hyperbolic or
equivalently exponential, as drawn in Figure 4-3, while those
of Y are trigonometric. If k is pure imaginary, then X
becomes trigonometric and Y is hyperbolic (or exponential).
The solution to the potential is then given by the product
of X and Y:
V = E1 sin ky sinh kx + E 2 sin ky cosh kx
+E3 cos ky sinh kx + E 4 cos ky cosh kx
or equivalently
V = F1 sin ky e"* + F2 sin ky e~k' + F3 COS ky e"* + F4 COS ky e *x

(21)
We can always add the solutions of (7) or any other
Laplacian solutions to (20) and (21) to obtain a more general
* SeeJ. R. Melcher and G. I. Taylor, Electrohydrodynamics: A Review of the Role of

InterfacialShear Stresses, Annual Rev. Fluid Mech., Vol. 1, Annual Reviews, Inc., Palo
Alto, Calif., 1969, ed. by Searsand Van Dyke, pp. 111-146. See also J. R. Melcher, "Electric
Fields and Moving Media", film produced for the National Committee on Electrical
EngineeringFilms by the EducationalDevelopment Center, 39 Chapel St., Newton, Mass.
02160. This film is describedin IEEE Trans. Education E-17, (1974) pp. 100-110.
cosh x lim sinh x cosh x e

3
2
J ,~~inh x = _
2
lim sinh x x 2
hx
2
Ix I< 1 cosh x 1 + 2
-3 -2 -1 1 2 3
sinh x
Figure 4-3 The exponential and hyperbolic functions for positive and negative
arguments.
solution because Laplace's equation is linear. The values of

the coefficients and of k are determined by boundary condi
tions.
When regions of space are of infinite extent in the x direc
tion, it is often convenient to use the exponential solutions in
(21) as it is obvious which solutions decay as x approaches co.
For regions of finite extent, it is usually more convenient to
use the hyperbolic expressions of (20). A general property of
Laplace solutions are that they are oscillatory in one direction
and decay in the perpendicular direction.
4-2-4 Spatialy Periodic Excitation
A sheet in the x =0 plane has the imposed periodic poten

tial, V = Vo sin ay shown in Figure 4-4. In order to meet this
boundary condition we use the solution of (21) with k = a.
The potential must remain finite far away from the source so
266 ElectricField Boundary Value Problems
V = VO sin aye ax V Osin aye-O
E =- Voaeax[cosayi, + sin ayi,] E = Vo aeax (cos ayi y sin ayixl
N> N
a , ,Il
cosaye =const
/ ,,N/'NX4.YY\Y~
I ~
>N/ / / / / field lines

- - - Cos ayea = const
Nz Z
I N
N N N \
N N
N N N N
N -. N
N N N
Figure 4-4 The potential and electric field decay away from an infinite sheet with
imposed spatially periodic voltage. The field lines emanate from positive surface
charge on the sheet and terminate on negative surface charge.
we write the solution separately for positive and negative x as
V=sinaye ", x-0 (22)

I Vo sin aye , xS0
where we picked the amplitude coefficients to be continuous
and match the excitation at x = 0. The electric field is then
E=VV - Voae"[cosayi,-sinayix] x >0

-Voae-[cos ayi +sin ayi], x<0
(
The surface charge density on the sheet is given by the dis

continuity in normal component of D across the sheet:
o(x =0) = e[E.(x = 0+) - E.(x = 0)]
= 2E Voa sin ay (24)
The field lines drawn in Figure 4-4 obey the equation

dy-y= E,,= -F cot ay -:>cos ay e "=const x>O 25
>0 (25)
& E. x0
4-2-5 Rectangular Harmonics
When excitations are not sinusoidally periodic in space,

they can be made so by expressing them in terms of a trig
onometric Fourier series. Any periodic function of y can be
expressed as an infinite sum of sinusoidal terms as
f(y)= -bo + . a.sin +bn cos ) (26)
where A is the fundamental period of f(y).

The Fourier coefficients an are obtained by multiplying
both sides of the equation by sin (2piry/A ) and integrating over
a period. Since the parameter p is independent of the index n,
we may bring the term inside the summation on the right
hand side. Because the trigonometric functions are orthog
onal to one another, they integrate to zero except when the
function multiplies itself:
pin si 2ntry dy=0, p +n

fA i A snA d=A/2,~~2
P=n (27)
(7
sin 2p[ cos 2niy dy=0
J" A A
Every term in the series for n # p integrates to zero. Only the
term for n = p is nonzero so that
a,= f(y) sin 2pjdy (28)

A .A
To obtain the coefficients b., we similarly multiply by
cos (2piry/A) and integrate over a period:
2 2pr
b,=- f(y)cos-p! dy (29)
A 0A
Consider the conducting rectangular box of infinite extent

in the x and z directions and of width d in the y direction
shown in Figure 4-5. The potential along the x = 0 edge is Vo
while all other surfaces are grounded at zero potential. Any
periodic function can be used for f(y) if over the interval
0:s y5 d, f(y) has the properties
f(y)=Vo,0<y<d;f(y=0)=f(y=d)= (30)
(
y/d vo
=
i.e --- - --
+
-01
+ . w/V . .250
0
-- o,= ( s p
-
In particular, we choose the periodic square wave function

with A==2d shown in Figure 4-6 so that performing the
integrations in (28) and (29) yields
S2Vo
pir
0, P even (1
4 VO/pir, p odd
Thus the constant potential at x =0 can be written as the

Fourier sine series
V(x 4 VO 0" sin (ntry/d)
= 0)= VOw= - (32)
1r n..1 n
n odd
In Figure 4-6 we plot various partial sums of the Fourier
series to show that as the number of terms taken becomes
large, the series approaches the constant value VO except for
the Gibbs overshoot of about 18% at y = 0 and y = d where the
function is discontinuous.
The advantage in writing VO in a Fourier sine series is that
each term in the series has a similar solution as found in (22)
where the separation constant for each term is k,, = nir/d with
associated amplitude 4 Vo/(nir).
The solution is only nonzero for x > 0 so we immediately
write down the total potential solution as
s e)v=
n " (31)
V(x,
Ir n=i n ed33
, o
100 terms
10 terms 3 terms 1 term Gibbs overshoot
10------. . . .
~ j.. uVVVY
.
-V I
.
0.75
Y
-N
V
V0
0.50
0.25
-Vo I
2d
d2d y
2d---
I I I I
0.0 0 0.25 0.50 0.75 1.0
y/d
Figure 4-6 Fourier series expansion of the imposed constant potential along the x = 0
edge in Figure 4-5 for various partial sums. As the number of terms increases, the
series approaches a constant except at the boundaries where the discontinuity in
potential gives rise to the Gibbs phenomenon of an 18% overshoot with narrow width.
The electric field is then
E=-VV= -V- (-sin i o (34)

d dd

The field and equipotential lines are sketched in Figure 4-5.

Note that for x d, the solution is dominated by the first
harmonic. Far from a source, Laplacian solutions are insensi
tive to the details of the source geometry.
4-2-6 Three-Dimensional Solutions
If the potential depends on the three coordinates (x, y, z),

we generalize our approach by trying a product solution of
the form
V(x, y, z) = X(x) Y(y) Z(z) (35)
which, when substituted into Laplace's equation, yields after

division through by XYZ
I d 2X
1 d 2Y l d2 Z
Xd2+-
X x +- 2=0 (36)
Y dy Zd
three terms each wholly a function of a single coordinate so

that each term again must separately equal a constant:
Id k2X2 1 d2Y 2 1 d 2Z 2
Xd Ydy Zdzkk+k (37)
We change the sign of the separation constant for the z

dependence as the sum of separation constants must be zero.
The solutions for nonzero separation constants are
X=A 1 sin kx+A2 coskx

Y=B1 sin ky+B 2 cos k~y (38)
Z= C1 sinh k~z+C 2 cosh kaz =D1 ekz+D2 ek
The solutions are written as if k, k,, and k. are real so that
the x and y dependence is trigonometric while the z depen
dence is hyperbolic or equivalently exponential. However, k.,
k,, or k. may be imaginary converting hyperbolic functions to
trigonometric and vice versa. Because the squares of the
separation constants must sum to zero at least one of the
solutions in (38) must be trigonometric and one must be
hyperbolic. The remaining solution may be either trigono
metric or hyperbolic depending on the boundary conditions.
If the separation constants are all zero, in addition to the
solutions of (6) we have the similar addition
Z = e z +f3) (39)
Separationof Variables in CylindricalGeometry 271
4-3 SEPARATION OF VARIABLES IN CYLINDRICAL GEOMETRY
Product solutions to Laplace's equation in cylindrical

coordinates
18 (rV 1 a2 V a2 V
r r r )r r2O4V 8z
also separate into solvable ordinary differential equations.
4-3-1 Polar Solutions
If the system geometry does not vary with z, we try a

solution that is a product of functions which only depend on
the radius r and angle 4:
V(r, 4) = R(r)b(4) (2)
which when substituted into (1) yields
4) d dR Rd 2.0
r + 2=0 (3)
r dr dr r* do
This assumed solution is convenient when boundaries lay at a

constant angle of 46 or have a constant radius, as one of the
functions in (2) is then constant along the boundary.
For (3) to separate, each term must only be a function of a
single variable, so we multiply through by r2 /R and set each
term equal to a constant, which we write as n2:
r d d 2 1 d4
- r - n - - = -_n2 (4)
R dr ( dr ' (D do2
The solution for 4) is easily solved as
( A I sin n4+A 2 cos n46, n00 5)

B 14+B2 , n=0
The solution for the radial dependence is not as obvious.

However, if we can find two independent solutions by any
means, including guessing, the total solution is uniquely given
as a linear combination of the two solutions. So, let us try a
power-law solution of the form
R= Ar (6)
which when substituted into (4) yields
p2= n 2 ->p = n 7
(7)
For n # 0, (7) gives us two independent solutions. When n =0

we refer back to (4) to solve
dR
r -= const* R= D ln r+D2 (8)
dr
so that the solutions are
RCjr"+C2 r-", n
Dllnr+D, n=0
We recognize the n =0 solution for the radial dependence

as the potential due to a line charge. The n =0 solution for
the 46 dependence shows that the potential increases linearly
with angle. Generally n can be any complex number,
although in usual situations where the domain is periodic and
extends over the whole range 0 = 4 t 2wr, the potential at
4 = 21r must equal that at 4 = 0 since they are the same point.
This requires that n be an integer.
EXAMPLE 4-1 SLANTED CONDUCTING PLANES
Two planes of infinite extent in the z direction at an angle a

to one another, as shown in Figure 4-7, are at a potential
difference v. The planes do not intersect but come sufficiently
close to one another that fringing fields at the electrode ends
may be neglected. The electrodes extend from r = a to r = b.
What is the approximate capacitance per unit length of the
structure?
+ o
0 ab
Figure 4-7 Two conducting planes at angle a stressed by a voltage v have a

4-directed electric field.
Separation of Variables in CylindricalGeometry 273
SOLUTION
We try the n =0 solution of (5) with no radial dependence

as
V=B 1 4+B2
The boundary conditions impose the constraints
V('0=0)=0, V(4=a)=v =v a
The electric field is
l dV v
E= = -=--
r do ra
The surface charge density on the upper electrode is then
sv
oy(f a) = -- E.*(-0 = a)=--
ra
with total charge per unit length
E ev b
A (,= a)= of(4 =a) dr =-ln-
J. c a
so that the capacitance per unit length is
A e In (b/a)
V a
4-3-2 Cylinder in a Uniform Electric Field
(a) Field Solutions

An infinitely long cylinder of radius a, permittivity E2, and
Ohmic conductivity 0-2 is placed within an infinite medium of
permittivity E1 and conductivity o,-. A uniform electric field at
infinity E = Eoi, is suddenly turned on at t =0. This problem
is analogous to the series lossy capacitor treated in Section
3-6-3. As there, we will similarly find that:
(i) At t = 0 the solution is the same as for two lossless

dielectrics, independent of the conductivities, with no
interfacial surface charge, described by the boundary
condition
o f(r = a) = Dr(r= a+)- Dr(r= a-) =0

=eiEr(r=a,)=8
2 Er(r=a-) (10)
(ii) As t ->o , the steady-state solution depends only on

the conductivities, with continuity of normal current
at the cylinder interface,

erE,(r = a.) = c-2Er(r = a-) Jr(r= a,) = J,(r = a-)
(11)
(iii) The time constant describing the transition from the
initial to steady-state solutions will depend on some
weighted average of the ratio of permittivities to
conductivities.
To solve the general transient problem we must find the
potential both inside and outside the cylinder, joining the
solutions in each region via the boundary conditions at r = a.
Trying the nonzero n solutions of (5) and (9), n must be an
integer as the potential at 4 =0 and 4 = 21r must be equal,
since they are the same point. For the most general case, an
infinite series of terms is necessary, superposing solutions
with n = 1, 2, 3, 4, - - - . However, because of the form of the
uniform electric field applied at infinity, expressed in cylin
drical coordinates as
E(r- o)= Eoi.= E 0 [i, cos 4-# sin 4] (12)
we can meet all the boundary conditions using only the n = 1
solution.
Keeping the solution finite at r =0, we try solutions of the
form
V(r, 4)= A MrCOS4,r (13)

I[B(t)r+C(t)/r] cos 4, r-I-a
-A (t)[cos 4i, - sin 4i]= -A(t)i, r <a

E= -VV= -[B(t)-Ct)/r2] cos Oir (14)
2
1 +[B(()+C(t)r1 sin 46id,, r>a
We do not consider the sin 4 solution of (5) in (13) because at

infinity the electric field would have to be y directed:
V= Drsin 4 > E = -V V= -D[i, sin 4+i cos 4] = -Di,

(15)
The electric field within the cylinder is x directed. The
solution outside is in part due to the imposed x-directed
uniform field, so that as r - co the field of (14) must approach
(12), requiring that B(t) = -Eo. The remaining contribution
to the external field is equivalent to a two-dimensional line
dipole (see Problem 3.1), with dipole moment per unit length:
p.= Ad = 2'rsC(t) (16)

The other time-dependent amplitudes A (t) and C(t) are

found from the following additional boundary conditions:
(i) the potential is continuous at r= a, which is the same

as requiring continuity of the tangential component of
E:
V(r= a.)= V(r = a-)>E6(r= a-)=E#(r = a+)
> Aa = Ba + C/a (17)
(ii) charge must be conserved on the interface:
Jr(r = a+) -J,(r = a) + = 0

at
> a-1Er(r = a,) - 0-2Er(r = a-)
+a [e 1 E,(r = a+) - e 2E,(r=a)] = 0

at
(18)
In the steady state, (18) reduces to (11) for the continuity of
normal current, while for t =0 the time derivative must be
noninfinite so o is continuous and thus zero as given by (10).
Using (17) in (18) we obtain a single equation in C(t):
dC (0-1+0-2)= -a2 dEo)
d + + 2 C= -a (Eo(o l--0- 2 )+(I-E 2 )-
dt 1+E2
61e dt
(19)
Since EO is a step function in time, the last term on the
right-hand side is an impulse function, which imposes the
initial condition
C(t = 0) = -a 2 ( 2) Eo (20)
so that the total solution to (19) is
C(t) = aEo +
(2 2(o62-0261) 61+E2
\0.-+02 (0-j+02)(81+82) / 0-3+O-2
(21)
The interfacial surface charge is
of(r = a, t) = e E,(r=a+) - E 2E,(r a)
= -e B--)+e 2 A] cos q
2(021-2)Eo+(ei+E2) Cos
=2(o-2- 0-162)Eo[1-e-I]cosk (22)

0-1 + 02
The upper part of the cylinder (-r/25 0 7r/2) is charged of

one sign while the lower half (7r/2:5 46 i ir) is charged with
the opposite sign, the net charge on the cylinder being zero.
The cylinder is uncharged at each point on its surface if the
relaxation times in each medium are the same, E/o-1= e2/0-2
The solution for the electric field at t =0 is
2______ 2e1 E 0
[eos
[CS:r -sm die0]=26Eo., r< a
61+62 E1+62
Et=0 Eo[ 1+- 61+62 COS ,ir (23)

[( r 61+62)
-- ) sin Oi.], r>a
The field inside the cylinder is in the same direction as the

applied field, and is reduced in amplitude if 62>El and
increased in amplitude if E2 <6 1, up to a limiting factor of two
as e1 becomes large compared to E2. If E2 = E1, the solution
reduces to the uniform applied field everywhere.
The dc steady-state solution is identical in form to (23) if we
replace the permittivities in each region by their conduc
tivities;
[cos 4O,- sin 40ij= i., r<a

a-1+0-2 471+0-2
[( 2
02-2~ i
E(t -+ co)=< Eo 1+- 0 cos Oi (24)
r -I+E-2)
- 1--,2,rl smin.0 , r>a

r O-l+O-2
(b) Field Line Plotting

Because the region outside the cylinder is charge free, we
know that V- E =0. From the identity derived in Section
1-5-4b, that the divergence of the curl of a vector is zero, we
thus know that the polar electric field with no z component
can be expressed in the form
E(r, 4)= VX X(r, O)i.

I aY. aY.
-,--14 (25)
r a4 ar
where x is called the stream function. Note that the stream

function vector is in the direction perpendicular to the elec
tric field so that its curl has components in the same direction
as the field.
Along a field line, which is always perpendicular to the

equipotential lines,
dr = Er I o84 (26)
r d4 Es rX/8r
By cross multiplying and grouping terms on one side of the
equation, (26) reduces to
a1 al
d =-dr+-dO =0>Y=const (27)
ar 84
Field lines are thus lines of constant 1.
For the steady-state solution of (24), outside the cylinder
=Er=Eo 1+ 2 cos
r or+o2 (28)
- =E4s=-Eo ( -a 2 2-O sin4
ar r 2 Ol+0-2/
we find by integration that
I= Eo( r+ ai2 ii sin# (29)
The steady-state'field and equipotential lines are drawn in

Figure 4-8 when the cylinder is perfectly conducting (- 2 -> x)
or perfectly insulating (o-2 = 0).
If the cylinder is highly conducting, the internal electric
field is zero with the external electric field incident radially, as
drawn in Figure 4-8a. In contrast, when the cylinder is per
fectly insulating, the external field lines must be purely
tangential to the cylinder as the incident normal current is
zero, and the internal electric field has double the strength of
the applied field, as drawn in Figure 4-8b.
4-3-3 Three-Dimensional Solutions
If the electric potential depends on all three coordinates,

we try a product solution of the form
V(r, 4-, z) = R(r)4(.)Z(z) (30)
which when substituted into Laplace's equation yields
ZF d dR RZd2 + d 2Z (31)
rr-ddr\r + r2 dY2+R
r) 0dZ dz-2 = 0
(
We now have a difficulty, as we cannot divide through by a

factor to make each term a function only of a single variable.
2
r<a
+ )
dr - E - r- AV ao ar
= COSO 0 r>a
rdO E, 2) r a
(i-'-) r<a
r0
-
E=-V V= + 2 2
a
-
+ [(1 + _a- ) COS Ir (1 a ()
sin ia r> a
- ----------- - ---- 4.25
---- - - --- -- 2.75
~ .~~~~~~ - ~- --- - _---__..

- 0.5
5
V/(Eoa)
- -- - 02--
- - ------ --- 0.165
0.16
---- 0.5
0.0
- -------- 2.75
a--------------------------------4.25
Eoi =E-
Figure 4-8 Steady-state field and equipotential lines about a (a) perfectly conducting
or (b) perfectly insulating cylinder in a uniform electric field.
However, by dividing through by V = R(bZ,
I d dR I d 2 4 1 d 2Z
do2+Z dz2 0 (32)
Rr dr (r +r)2
-k 2 k2
we see that the first two terms are functions of r and 4 while
the last term is only a function of z. This last term must
therefore equal a constant:
2 Alsinhkz+A 2 coshkz, k O
Id Z (33)
Z dz LZ+A,
k =0
Separationof Variables in Cylindrical Geometry 279
=-2Eorcos$
r<a
-Eoa(a + )cosO r l a
sinOiO)=2EOi,
E = r<a
s
V [ sa2 (+ sa2
r Sir -0 r2)ioI r>a
V
Ea
- -- -- -- -4.25
- - -------- 3.33
---------- - -- -- -- 2.5
-- - - ------- -2.0
- C2 , 2 -0 0.5
------
--- -- --- - - --- 0.0
--------- 0.5
-- a
-
E1, 01
- - ------ 2.0
-- - - -- - 2.0
- --- - - - 2.5
- 3.33
4.25 2
a
dr Er ___Coto
rdo ~ E 2 cot$
Eoi= E6 (i, coso - i, sin$) ( r
(b) =>(- _ a)sin$ const
Figure 4-8b
2
The first two terms in (32) must now sum to -k so that after
multiplying through by r 2 we have
r d dR 2 2 1 d2 D
R r--rdr +k r +- =0 (34)
Now again the first two terms are only a function of r, while
the last term is only a function of 0 so that (34) again
separates:
r d dR 2 2
-- n 2
rr- +k
2 1 d 23
r =n (5
dr dr ' D2
where n 2 is the second separation constant. The angular

dependence thus has the same solutions as for the two-
dimensional case
{Bsin n +B 2 cosnO, n=O (36)
B30 + B4, n =O
The resulting differential equation for the radial dependence

d dR 22
r- (r- +(k2r2-n2)R=O (37)
dr dr)
is Bessel's equation and for nonzero k has solutions in terms
6.

5. 1, (x) =j- J, (jx)
4.
3.
2.
1 2 (x)
Io (x)
1.
2 2x
-1.
(a)
Figure 4-9 The Bessel functions (a) J,(x) and I,(x), and (b) Y,(x) and K,,(x).
of tabulated functions:
C1.J(kr)+C2 Y(kr), k *0
R= C 3 r"+C4 r-', k=0, n 0 (38)
C 5 In r+C6 , k=0, n=O
where J. is called a Bessel function of the first kind of order n
and Y, is called the nth-order Bessel function of the second
kind. When n = 0, the Bessel functions are of zero order while
if k =0 the solutions reduce to the two-dimensional solutions
of (9).
Some of the properties and limiting values of the Bessel
functions are illustrated in Figure 4-9. Remember that k
2.5
Ko x)
2.0
K, (x) K,(x) 7r/2)j J, (jx) + Y (jx)]
K 2 (x)
1.5
1.0
Yo(x)
0.5 Y (x) Y2
(X)
4 6 8 10
0.5

7 22
-1.0
(b)
Figure 4-9b
can also be purely imaginary as well as real. When k is real so

that the z dependence is hyperbolic or equivalently exponen
tial, the Bessel functions are oscillatory while if k is imaginary
so that the axial dependence on z is trigonometric, it is con
venient to define the nonoscillatory modified Bessel functions
as
I.(kr)=j "J.(jkr)
K.(kr)j= u+j1 J(jkr)+jY.(jkr)]

As in rectangular coordinates, if the solution to Laplace's
equation decays in one direction, it is oscillatory in the
perpendicular direction.
4-3-4 High Voltage Insulator Bushing
The high voltage insulator shown in Figure 4-10 consists

of a cylindrical disk with Ohmic conductivity o- supported
by a perfectly conducting cylindrical post above a ground
plane.*
The plane at z = 0 and the post at r = a are at zero potential,
while a constant potential is imposed along the circumference
of the disk at r = b. The region below the disk is free space so
that no current can cross the surfaces at z = L and z = L - d.
Because the boundaries lie along surfaces at constant z or
constant r we try the simple zero separation constant solutions
in (33) and (38), which are independent of angle 4:
, L-d<z<L
V(rz) =Az+Blz lnr+C1 lnr+D1
' A 2 z+B 2 zlnr+C2 lnr+D2, 0tz!L-d (40)
Applying the boundary conditions we relate the coefficients
as
V(z =0)=0>C 2 =D 2 =0
In a=0
[A 2 +B 2
V(r=a)=0> A 1 +B 1lna=0
IC1 Ina+D,=0
V(r=b,z>L-d)=Vo> (41)
IC1 Inb+D = Vo
V(z=(L-d).)=V(z=(L-d)+)='(L-d)(A 2 +B 2 lnr)
=(L-d)(A1 +Bllnr)+Cilnr+Dj
* M. N. Horenstein, "Particle Contaminationof High Voltage DC Insulators," PhD thesis,

MassachusettsInstitute of Technology, 1978.
b
a
L E.
V=VO@r= b
E= 0
cv- -c
(a) I
L
V = VO V= VO
a- |
~.++ +++ + +'+: d + + :+ I+ + I+ +1+ +

+
V0
+
Vo
0.9
0.8
- 0.7
,
- 0.6
-
-
0.5
-
-
0.4
/
0.3
- ~/
0.2
0.1
Field lines a--

*
2
Z2 = r [In(r/a)
2
] + const
- - - - Equipotential V _ Vozln(r/a)
lines (L -d)In(b/a)
(b)
Figure 4-10 (a) A finitely conducting disk is mounted upon a perfectly conducting
cylindrical post and is placed on a perfectly conducting ground plane. (b) Field and
equipotential lines.
which yields the values
Cl= 0
,V Vo In a
A, = B, = 0,
In (b/a)' In (b/a) (42)
B 2 =(L- (/ C 2 = D 2 =0
(L -d) In (b/a)' (L - d) In (b/a)'
The potential of (40) is then
V(r, z)=
J
Vo In (r/a)
n(b/a)
Voz In (r/a)
L-dszsL
(43)
0_zSL-d
(L - d) In (b/a)
V0 i L-d<z<L
E=-VV= rIn(b/a)
(44)
0, 0<z<L-d
(- ) n- b/V) In r+
The field lines in the free space region are
dr r r/z)Z2= r2 In r- +const (45)

dzE,,r n (ra) I.a 2J
and are plotted with the equipotential lines in Figure 4-10b.
4-4 PRODUCT SOLUTIONS IN SPHERICAL GEOMETRY
In spherical coordinates, Laplace's equation is

Iar 2 V 2s a (sin1 + 1 V
r~r\8/ r inG ao r2 sin
(1)
4-4-1 One-Dimensional Solutions
If the solution only depends on a single spatial coordinate,

the governing equations and solutions for each of the three
coordinates are
d r2 dV(r)= A(
(i) (r r 1=> V(r) =- +A 2 (2)
ProductSolutions in Spherical Geometry 285
(ii) (sin 6 dVG) = 0 =V(0)=B1 In (tan +B 2
(3)
d2 V(O)
(iii) d => V(O) = CIO4+ C2 (4)
We recognize the radially dependent solution as the poten

tial due to a point charge. The new solutions are those which
only depend on 0 or 4.
EXAMPLE 4-2
TWO CONES
Two identical cones with surfaces at angles 0 = a and 0=

ir -a and with vertices meeting at the origin, are at a poten
tial difference v, as shown in Figure 4-11. Find the potential
and electric field.
SIn (tan k)
.. (.)..=. -

2 In(tan
)
E 2rsinO ln(tan
)
.. .2 . .. .. . .
Figure 4-11 Two cones with vertices meeting at the origin are at a potential
difference v.
286 Electric FieldBQundary Value Problems
SOLUTION
Because the boundaries are at constant values of 0, we try

(3) as a solution:
V(O)= B In [tan (0/2)1+ B 2
From the boundary conditions we have
V(O = a) =v
2
-v v
V(O = r - a) = -=> Bl = B2=0
2 2 In [tan (&/2)]
so that the potential is
(0/2)]
V(O)= v In [tan
2 In [tan (a/2)]
with electric field
-v
E = -v V=i
2r sin 9 In [tan (a/2)]
4-4-2 Axisymmetric Solutions
If the solution has no dependence on the coordinate 4, we

try a product solution
V(r, 9) = R(r)9(0) (5)
which when substituted into (1), after multiplying through by

r2 IRO, yields
/dR d .dO I d 2
- r -+ W- sin -- =0 (6)
R dr( dr - 9n d9
Because each term is again only a function of a single vari

able, each term is equal to a constant. Anticipating the form
of the solution, we choose the separation constant as n(n + 1)
so that (6) separates to
r2 -n(n + 1)R =0 (7)
d(snd9+<n(n+1)esine=0 (8)
d9' dM
I
Product Solutions in SphericalGeometry 287
For the radial dependence we try a power-law solution
R=Arp (9)
which when substituted back into (7) requires
p(p +1)=n(n +1) (10)
which has the two solutions
p =n, p =-(n +1) (1
When n = 0 we re-obtain the I/r dependence due to a point

charge.
To solve (8) for the 9 dependence it is convenient to intro
duce the change of variable
i=cos9 (12)
so that
dO dO dp dO
---- =-si - -= _( ) -P)/ 2d (13) sd
.
dO d16 dO dp dp
Then (8) becomes
- (I-p2)- +n(n+1)0=0 (14)

dp ( dp
which is known as Legendre's equation. When n is an integer,
the solutions are written in terms of new functions:
e= B.P.(P)+ C.Q.(P) (15)
where the P.(P) are called Legendre polynomials of the first

kind and are tabulated in Table 4-1. The Q. solutions are
called the Legendre functions of the second kind for which
the first few are also tabulated in Table 4-1. Since all the Q.
are singular at 9=0 and 9= r, where P = * 1, for all problems
which include these values of angle, the coefficients C. in (15)
must be zero, so that many problems only involve the Legen
dre polynomials of first kind, P.(cos 0). Then using (9)-(11)
and (15) in (5), the general solution for the potential with no
* dependence can be written as
V(r,o)= Y (A.r"+Br~"+I))P.(cos0) (16)

Table 4-1 Legendre polynomials of first and second kind
n P.(f= Cos 6) Q.(P =cos 6)

0 1 -2LIn(jj)
1 is =cos 6
1 -2
2 2(3#2-1) -4(3 _ 1) In
+ 13) -p2
='(3 cos 2 0 _ 1) -4(53 - 3) In 1+# 3
3 2(505-39)
-1 (5 cos 6-3 cos 6)
m Im(p2_
)
2'm! dpm
4-4-3 Conducting Sphere in a Uniform Field
(a) Field Solution

A sphere of radius R, permittivity s2 , and Ohmic conduc
tivity a2 is placed within a medium of permittivity El and
conductivity o1. A uniform dc electric field Eoi. is applied at
infinity. Although the general solution of (16) requires an
infinite number of terms, the form of the uniform field at
infinity in spherical coordinates,
E(r - 00) = Eoi. = Eo(i. cos 6 -ie sin 6) (17)
suggests that all the boundary conditions can be met with just
the n = 1 solution:
r:s R
V(r, 0).= (Ar cos 0, 2 (18)
V (Br + C/r ) cos 0, r R
We do not include the I/r2 solution within the sphere (r< R)

as the potential must remain finite at r =0. The associated
I
Product Solutions in Spherical Geometry 289
electric field is
sin 0)=-Ai,, r<R

E=-VV= -A(ircos 0-iO
-(B -2Cr 3 ) cos Oit+(B +C/r3)sin i., r>R
(19)
The electric field within the sphere is uniform and z direct
ed while the solution outside is composed of the uniform
z-directed field, for as r->o the field must approach (17) so
that B = -Eo, plus the field due to a point dipole at the origin,
with dipole moment
p =41re 1 C (20)
Additional steady-state boundary conditions are the
continuity of the potential at r = R [equivalent to continuity of
tangential E(r =R)], and continuity of normal current at
r= R,
V(r = R)= V(r = R-)=>Ee(r= R,)= E(r = R-)
> AR=BR+C/R 2
J,(r = R+)= J,(r = R-):>o-E,(r= R+) = o-2E,(r = R-) (21)
=c> 1 (B -2C/R)= Or 2A
for which solutions are
A =- 3- 1 Eo, B = -Eo, C = (c 2 -co 1 )R3 Eo (22)

2o-1 + 0-2 2o-+-2
The electric field of (19) is then
3c-1E0 3c-1E0
(i cos 6 -ie sin 6)= .i, r<R
20-1 + -2 2a- + -2
E=I Eo 1+2 R3-2 ) cos 6i, (23)

r 3(2o- + -2))
( 3(0-2 -a1) s ]i, r>R

r 3(2cr-1+ 0-2))
The interfacial surface charge is
orf(r = R)= eiE,(r = R+)-e 2E,(r = R-)

3(- 2s1 -1 0-1E2)Eo cos 0
(24)
2r, + 0r2
which is of one sign on the upper part of the sphere and of

opposite sign on the lower half of the sphere. The total
charge on the entire sphere is zero. The charge is zero at
every point on the sphere if the relaxation times in each

region are equal:
=9 2(25)
O*l 02
The solution if both regions were lossless dielectrics with

no interfacial surface charge, is similar in form to (23) if we
replace the conductivities by their respective permittivities.
(b) Field Line Plotting

As we saw in Section 4-3-2b for a cylindrical geometry, the
electric field in a volume charge-free region has no diver
gence, so that it can be expressed as the curl of a vector. For
an axisymmetric field in spherical coordinates we write the
electric field as
E'8'-Vx (X(r, 0).
sr,,- r sin 0/
1 81, 1 a1,
= 2 -IT- . i (26)
r sin 08o r sin0 ar
Note again, that for a two-dimensional electric field, the

stream function vector points in the direction orthogonal to
both field components so that its curl has components in the
same direction as the field. The stream function I is divided
by r sin 0 so that the partial derivatives in (26) only operate on
1.
The field lines are tangent to the electric field
dr = E, . 1 81180 (27)
r d6 Es r81,1r
which after cross multiplication yields
d1=-dr+-d =0::>I=const (28)

8r 80
so that again I is constant along a field line.

For the solution of (23) outside the sphere, we relate the
field components to the stream function using (26) as
1 81, 2R_____-_
E,.= -=E= 1+ 3, cos8
r2 sin 0 aO r (2o-
1+ 2 ))
1 81, / R 3 (o,2 - 1 ) sine
r sin 6 ar r3(2oi + 02))
ProductSolutions in Soherical Geometry 291
so that by integration the stream function is
I= Eo 2+ sin2 8 (30)
2 r(2a + a2))
The steady-state field and equipotential lines are drawn in

Figure 4-12 when the sphere is perfectly insulating (Or 2 = 0) or
perfectly conducting (o2-00).
v 2
EoR[ + R2 I cos r> R
{EO(i, Cos 0- i.sinO)= EOi, r<R

i, } r >R
E3 [ r ) Cos-i- - (1 + 2r ) sin
(1
r3Coto 9
rd6 E
+ R
2-R I sin 20 const To R

R r
4.0 -
------ -3.1
-------------- -2.1
-- -- -- - 1.
- - - - -- .- - ----
- -- -- - - 0.45
- - - - - - - - -*-
.-- -- - - - - - -- 0. 4
-
-0.75
-- Eoi4 - -- ------ - 1.4
-
--------- ------ -- 0o.75
- - - - --- - - - -----
--.-..--.-1.1
----- 1.3
-
_ 2.1
-
---- ..
-
EO i, = E O(ir cos - i. sin 0)

(a)
Figure 4-12 Steady-state field and equipotential lines about a (a) perfectly insulating
or (b) perfectly conducting sphere in a uniform electric field.
r<R
0
-EOR(r - )COS6 r>R
t
r<R
E=-V V= ~
2R
3
R3
- V oi,' 1 _ ) sinioI r>R
Z ~
Eo[(1 + r 3
r
)
(1+ 2R3
dr 3
E, r coto
rdO E0 (0 Re3
r
( )2]sin 2 6 const
r 2 -R
- -2.75
-1.75
------- 1.0
02 ----- 0.25
---------------
--~-- - - V
0
0.25 EoR
0.6
------- 1.0
1.75
- - '~- 02
-
E1 , Oi
2.75
Eoi ,= Eo(icos -i sinG)
(b)
Figure 4-12b
If the conductivity of the sphere is less than that of the

surrounding medium (0-2<-1), the electric field within the
sphere is larger than the applied field. The opposite is true
for (U2 >o1 ). For the insulating sphere in Figure 4-12a, the
field lines go around the sphere as no current can pass
through.
For the conducting sphere in Figure 4-12b, the electric field
lines must be incident perpendicularly. This case is used as a
polarization model, for as we see from (23) with 0-2 - O, the
external field is the imposed field plus the field of a point
Product Solutions in Spherical Geometry 293
dipole with moment,

p, = 47r 1R3 Eo (31)
If a dielectric is modeled as a dilute suspension of nonin
teracting, perfectly conducting spheres in free space with
number density N, the dielectric constant is
eoEo-+-P eoEo+ Np,
E= = = Eo(1+4TR 3N) (32)
4-4-4 Charged Particle Precipitation Onto a Sphere
The solution for a perfectly conducting sphere surrounded

by free space in a uniform electric field has been used as a
model for the charging of rain drops.* This same model has
also been applied to a new type of electrostatic precipitator
where small charged particulates are collected on larger
spheres.t
Then, in addition to the uniform field Eoi, applied at
infinity, a uniform flux of charged particulate with charge
density po, which we take to be positive, is also injected, which
travels along the field lines with mobility A. Those field lines
that start at infinity where the charge is injected and that
approach the sphere with negative radial electric field,
deposit charged particulate, as in Figure 4-13. The charge
then redistributes itself uniformly on the equipotential sur
face so that the total charge on the sphere increases with time.
Those field lines that do not intersect the sphere or those that
start on the sphere do not deposit any charge.
We assume that the self-field due to the injected charge is
very much less than the applied field E0 . Then the solution of
(23) with O-2 = C is correct here, with the addition of the radial
field of a uniformly charged sphere with total charge Q(t):
S 2R 3 R3
E [EO
r
)+3)
cosO+ Q
4 7r--r
iEO(1 ) sin io,
)
r>R (33)
Charge only impacts the sphere where E,(r=R) is nega
tive:
E,(r = R)= 3EO cos 9+

Q 2<0 (34)
4ireR
* See: F. J. W. Whipple and J. A. Chalmers, On Wilson's Theory of the Collection of Charge

by FallingDrops, Quart. J. Roy. Met. Soc. 70, (1944), p. 103.
t See: H. J. White, Industrial Electrostatic Precipitation Addison-Wesley, Reading. Mass.
1963, pp. 126-137.

11Z
ft.
+
+ :0
+
Injected charge E0 i
with dnity pEoi
Q La I
and mobilityM 1.0 9 =-.7071
1.. .7
=Q 1 .7071 Q =+1.
Q.aQ
(a) (b) (c) (d) (e)
2R l 2[ 2 20- oS
E0 R + (-) 1 sin _ C = constant
Figure 4-13 Electric field lines around a uniformly charged perfectly conducting sphere in a uniform electric field with continuous
positive charge injection from z = -oo. Only those field lines that impact on the sphere with the electric field radially inward [E,(R) <0]
deposit charge. (a) If the total charge on the sphere starts out as negative charge with magnitude greater or equal to the critical charge,
the field lines within the distance y. of the z axis impact over the entire sphere. (b)-(d) As the sphere charges up it tends to repel some of
the incident charge and only part of the sphere collects charge. With increasing charge the angular window for charge collection
decreases as does y.,. (e) For Q - Q, no further charge collects on the sphere so that the charge remains constant thereafter. The angular
window and y, have shrunk to zero.
Product Solutions in Spherical Geometr, 295
which gives us a window for charge collection over the range

of angle, where
-2 (35) cosO 2 Q
121reER
Since the magnitude of the cosine must be less than unity, the
maximum amount of charge that can be collected on the
sphere is
Q = 127reEoR 2 (36)
As soon as this saturation charge is reached, all field lines

emanate radially outward from the sphere so that no more
charge can be collected. We define the critical angle 6. as the
angle where the radial electric field is zero, defined when (35)
is an equality cos 0, = -Q/Q,. The current density charging
the sphere is
J, = popE,(r = R)
=3pogEo(cosO+QQ,), 0,<0< (37)
The total charging current is then
dQ
-=-
r 2
J,2R2 sin 0dO
= -6rpoAEoR 2
(cos 0 + Q/Q3 ) sin 0 dB
= -6rposEoR 2 (-- cos 20 - (Q/Q,) cos 6) |=..

= -6irpop.EoR (- (1-cos 20,) + (Q/Q,) (1 + cos 0.))
(38)
As long as IQ1 <Q, 0, is defined by the equality in (35). If Q

exceeds Q,, which can only occur if the sphere is intentionally
overcharged, then 0, = 7r and no further charging can occur
as dQldt in (38) is zero. If Q is negative and exceeds Q, in
magnitude, Q < -Q, then the whole sphere collects charge as
0, =0. Then for these conditions we have
-I , Q>Q,
cos0,= -QIQ, -Q,<Q<Q, (39)
1, Q<-Q,
cos20,=2cos 2 -C1={_1 )2 Q1> Q (40)

12(Q/Q,)I'IQ<2
so that (38) becomes
0, Q>Q.
1 , Q,<Q<Q, (41)
dtj4s i,2
d PoI Q
6 Q,
with integrated solutions
Qo
Q1,
Qo+ (0t){
Q Q. 4 \ Q,/
Q= , -Q,<Q<Q, (42)
1+ 1 _o
4r Q,
Qo Q <-Q
Q
,
where Qo is the initial charge at t=0 and the characteristic

charging time is
r = E/(Po) (43)
If the initial charge Qo is less than -Q,, the charge magni

tude decreases with the exponential law in (42) until the total
charge reaches -Q, at t = to. Then the charging law switches
to the next regime with Qo = -Q., where the charge passes
through zero and asymptotically slowly approaches Q = Q,.
The charge can never exceed Q, unless externally charged. It
then remains constant at this value repelling any additional
charge. If the initial charge Qo has magnitude less than Q.,
then to=0. The time dependence of the charge is plotted in
Figure 4-14 for various initial charge values Qo. No matter
the initial value of Qo for Q < Q,, it takes many time constants
for the charge to closely approach the saturation value Q,.
The force of repulsion on the injected charge increases as the
charge on the sphere increases so that the charging current
decreases.
The field lines sketched in Figure 4-13 show how the fields
change as the sphere charges up. The window for charge
collection decreases with increasing charge. The field lines
are found by adding the stream function of a uniformly
charged sphere with total charge Q to the solution of (30)
A Numerical Method-Successive Relaxation 297
2.0
1.0
_____
Q0
1 ----
= 15 1 _Q _ Qo
QS Qs
.0
- -- -- ---
- - - - - - - -
- - - -
ko+ ( - to
Q
6 8 10 Q. +( - to
( 2.0 eo T (p
)
-1.0
--
0
3. -Q QO

a. =QD,
-2.0
-3.0
Figure 4-14 There are three regimes describing the charge build-up on the sphere. It
takes many time constants ['r = e/(pos)] for the charge to approach the saturation value
Q, because as the sphere charges up the Coulombic repulsive force increases so that
most of the charge goes around the sphere. If the sphere is externally charged to a
value in excess of the saturation charge, it remains constant as all additional charge is
completely repelled.
with 0-2->00:
I= EoR 2 [!!+I sin 2 _Q coS

r 2 R 47re
(44)
The streamline intersecting the sphere at r = R, 0 = 0,
separates those streamlines that deposit charge onto the
sphere from those that travel past.
4-5 A NUMERICAL METHOD-SUCCESSIVE RELAXATION
In many cases, the geometry and boundary conditions are

irregular so that closed form solutions are not possible. It
then becomes necessary to solve Poisson's equation by a
computational procedure. In this section we limit ourselves to
dependence on only two Cartesian coordinates.
4-5-1 Finite Difference Expansions
The Taylor series expansion to second order of the poten

tial V, at points a distance Ax on either side of the coordinate
(x, y), is
2(1
aV A. I 2V
V(x+-Ax, V(x, y)+-
-V(x'y)+-
V(x+-AX, y)y)' Ax+---2-
x+- (,&X)22
(Ax)
ax 2Ox X.y
If we add these two equations and solve for the second

derivative, we have
2
a V V(x+Ax, y)+ V(x -Ax,
2
y)-2V(x, y) (2)
ixT (Ax)
Performing similar operations for small variations from y

yields
a9V V(x, y+Ay)+ V(x, y -Ay)-2V(x, y)
(Ay) 2 (3)
y*
If we add (2) and (3) and furthermore let Ax = Ay, Poisson's
equation can be approximated as
a2 V 02 V 1
-y -' _iY _2 [V(x +Ax, y)+ V(x - Ax, y)
+V(x, y +Ay)+ V(x, y - Ay)-4V(x, y)] =

(4)
so that the potential at (x, y) is equal to the average potential
of its four nearest neighbors plus a contribution due to any
volume charge located at (x, y):
V(x, y)= 4[V(x +Ax, y)+ V(x - Ax, y)
pj(x y) (Ax)2 5
+ V(x, y+ Ay)+ V(x, y- Ay)]+ 4 (5)
4e
The components of the electric field are obtained by taking
the difference of the two expressions in (1)
E.(x,y)=- ax- 2 AxL~xY

[V(x+Ax, y)-V(x-Ax,y)]
(6)
OV 1
E,(x, y) = -- aV - -- ,[V(x, y + AY)- V(x, y - Ay)
ay MAy
4-5-2 Potential Inside a Square Box
Consider the square conducting box whose sides are con

strained to different potentials, as shown in Figure (4-15). We
discretize the system by drawing a square grid with four
A Numerical Method-Successive Relaxation 299
V2 = 2
4
V(3, 2) V(3, 3)
3V V3 = 3 d
2 - V(2, 2) V(2, 3)
1 2 3 4
Figure 4-15 The potentials at the four interior points of a square conducting box
with imposed potentials on its surfaces are found by successive numerical relaxation.
The potential at any charge free interior grid point is equal to the average potential of
the four adjacent points.
interior points. We must supply the potentials along the

boundaries as proved in Section 4-1:
4 4
Vl= Y V(I, J= 1)= 1, V3 = Y V(I, J= 4)= 3
I=1 1=1
4 4 (7)
V2= Y V(I=4,J)=2, V4 = Y V(I=1,J)=4
J=1 J=1
Note the discontinuity in the potential at Che corners.

We can write the charge-free discretized version of (5) as
V(I, J) = 4[ V(I + 1, J) + V(I - 1, J) + V(I, J+ 1) + V(I, J - 1)]

(8)
We then guess any initial value of potential for all interior

grid points not on the boundary. The boundary potentials
must remain unchanged. Taking the interior points one at a
time, we then improve our initial guess by computing the
average potential of the four surrounding points.
We take our initial guess for all interior points to be zero
inside the box:
V(2, 2) = 0, V(3, 3) = 0
V(3, 2) = 0, V(2, 3) = 0
Then our first improved estimate for V(2, 2) is
V(2, 2)= [ V(2, 1)+ V(2, 3)+ V(1, 2)+ V(3, 2)]
= [1+0+4+0]= 1.25 (10)
Using this value of V(2, 2) we improve our estimate for

V(3, 2) as
V(3, 2)= [V(2, 2)+ V(4, 2)+ V(3, 1)+ V(3, 3)]
A1.25+2+1+0]= 1.0625
=4 (11)
Similarly for V(3, 3),
V(3, 3)= [ V(3, 2)+ V(3, 4)+ V(2, 3)+ V(4, 3)]
=;[1.0625+3+0+2]= 1.5156 (12)
and V(2, 3)
V(2, 3)=1[ V(2, 2)+ V(2, 4)+ V(1, 3)+ V(3, 3)]
=f[1.25+3+4+1.5156]=2.4414 (13)
We then continue and repeat the procedure for the four
interior points, always using the latest values of potential. As
the number of iterations increase, the interior potential
values approach the correct solutions. Table 4-2 shows the
first ten iterations and should be compared to the exact solu
tion to four decimal places, obtained by superposition of the
rectangular harmonic solution in Section 4-2-5 (see problem
4-4):
VOx, y)= I . sin !!f'y(Vs sinh

.. n smh n s d d
n odd
- V, sinh nr(x-d))
+sin n( V 2 sinh V4sinh n(y - d) (14)
where Vi, V2, Vs and V 4 are the boundary potentials that for
this case are
V 1=1, V2=2, Vs=3, V4=4 (15)
To four decimal places the numerical solutions remain
unchanged for further iterations past ten.
Table 4-2 Potential values for the four interior points in

Figure 4-15 obtained by successive relaxation for the first
ten iterations
0 1 2 3 4 5
V1 0 1.2500 2.1260 2.3777 2.4670 2.4911

V2 0 1.0625 1.6604 1.9133 1.9770 1.9935
V3 0 1.5156 2.2755 2.4409 2.4829 2.4952
V4 0 2.4414 2.8504 2.9546 2.9875 2.9966
Problems 301
6 7 8 9 10 Exact
V1 2.4975 2.4993 2.4998 2.4999 2.5000 2.5000

V 2 1.9982 1.9995 1.9999 2.0000 2.0000 1.9771
V3 2.4987 2.4996 2.4999 2.5000 2.5000 2.5000
V 4 2.9991 2.9997 2.9999 3.0000 3.0000 3.0229
The results are surprisingly good considering the coarse

grid of only four interior points. This relaxation procedure
can be used for any values of boundary potentials, for any
number of interior grid points, and can be applied to other
boundary shapes. The more points used, the greater the
accuracy. The method is easily implemented as a computer
algorithm to do the repetitive operations.
PROBLEMS
Section 4.2
1. The hyperbolic electrode system of Section 4-2-2a only
extends over the range 0 : x : xo, 0 ! y t yo and has a depth D.
(a) Neglecting fringing field effects what is the approxi
mate capacitance?
(b) A small positive test charge q (image charge effects are
negligible) with mass m is released from rest from the surface
of the hyperbolic electrode at x = xo, y = ab/xo. What is the
velocity of the charge as a function of its position?
(c) What is the velocity of the charge when it hits the
opposite electrode?
2. A sheet of free surface charge at x = 0 has charge dis
tribution
of = oo cos ay
o =o cos ay
)x
(a) What are the potential and electric field distributions?

(b) What is the equation of the field lines?
3. Two sheets of opposite polarity with their potential dis
tributions constrained are a distance d apart.
VO cos dy - VO cos ay
C -* X
(a) What are the potential and electric field distributions

everywhere?
(b) What are the surface charge distributions on each
sheet?
4. A conducting rectangular box of width d and length 1 is of'
infinite extent in the z direction. The potential along the x =0
edge is V, while all other surfaces are grounded (V2= V 3
V4=0).
3,
V4
d
V1 V3
V2

(b) The potential at y = 0 is now raised to V2 while the
surface at x =0 remains at potential VI. The other two sur
faces remain at zero potential (V3= V4=0). What are the
potential and electric field distributions? (Hint: Use super
position.)
(c) What is the potential distribution if each side is respec
tively at nonzero potentials V1, V2, Vs, and V4?
.
I
Problems 303
5. A sheet with potential distribution

V = Vo sin ax cos bz
is placed parallel and between two parallel grounded
conductors a distance d apart. It is a distance s above the
lower plane.
d
-
V = VO si nax cos bz
s
-
0
-

(Hint: You can write the potential distribution by inspection
using a spatially shifted hyperbolic function sinh c(y -d).)
(b) What is the surface charge distribution on each plane at
y=0,y=s, and y=d?
6. A uniformly distributed surface charge o-o of width d and
of infinite extent in the z direction is placed at x = 0 perpen
dicular to two parallel grounded planes of spacing d.
-x

(Hint: Write o-o as a Fourier series.)
(b) What is the induced surface charge distribution on each
plane?
(c) What is the total induced charge per unit length on
each plane? Hint:
S r2
.=, n 8
n odd
7. A slab of volume charge of thickness d with volume charge

density p1= po sin ax is placed upon a conducting ground
plane.
(a) Find a particular solution to Poisson's equation. Are the

boundary conditions satisfied?
(b) If the solution to (a) does not satisfy all the boundary
conditions, add a Laplacian solution which does.
(c) What is the electric field distribution everywhere and
the surface charge distribution on the ground plane?
(d) What is the force per unit length on the volume charge
and on the ground plane for a section of width 2 r/a? Are
these forces equal?
(e) Repeat (a)-(c), if rather than free charge, the slab is a
permanently polarized medium with polarization
P= Po sin axi,
8. Consider the Cartesian coordinates (x, y) and define the

complex quantity
z=x+jy, j=/
where z is not to be confused with the Cartesian coordinate.
Any function of z also has real and imaginary parts
w(z) = u(x, y)+jv(x, y)
(a) Find u and v for the following functions:
(i) z2
(ii) sin z
(iii) cos z
(iv) e'
(v) In z
(b) Realizing that the partial derivatives of w are
aw = dw az = dw au .av
ax dz ax dz ax ax
aw= dw az .dw .u .av
1
ay dz ay dz -~ y
show that u and v must be related as
au av au Ov
Ox ay' ay ax
Problems 305
These relations are known as the Cauchy-Riemann equations

and u and v are called conjugate functions.
(c) Show that both u and v obey Laplace's equation.
(d) Show that lines of constant u and v are perpendicular
to each other in the xy plane. (Hint: Are Vu and Vv perpen
dicular vectors?)
Section 4.3
9. A half cylindrical shell of length I having inner radius a
and outer radius b is composed of two different lossy dielec
tric materials (e1, o-1) for 0<4 <a and (e2, 0-2) for a <4< r.
A step voltage Vo is applied at t =0. Neglect variations with z.
Depth I
VO
within the shell at times t =0, t = co, and during the transient
interval? (Hint: Assume potentials of the form V(O)=A(t)o
+B(t) and neglect effects of the region outside the half
cylindrical shell.)
(b) What is the time dependence of the surface charge at
O= a?
(c) What is the resistance and capacitance?
10. The potential on an infinitely long cylinder is constrained
to be
V(r = a)= Vo sin n46
V V/2
(bi)
(a) Find the potential and electric field everywhere.

(b) The potential is now changed so that it is constant on
each half of the cylinder:

Vo/2, O<4<nr
=V~. 0r)= ')=
V(r=a.
- Vo/2 , vr < -0 < 2 7r
Write this square wave of potential in a Fourier series.

(c) Use the results of (a) and (b) to find the potential and
electric field due to this square wave of potential.
11. A cylindrical dielectric shell of inner radius a and outer
radius b is placed in frce space within a uniform electric field
Eoi.. What are the potential and electric field distributions
everywhere?
60
4, = EO[i, cos 0-iosin 01
12. A permanently polarized cylinder P2i. of radius a is

placed within a polarized medium Pli, of infinite extent. A
uniform electric field Eoi. is applied at infinity. There is no
free charge on the cylinder. What are the potential and elec
tric field distributions?
P i
tix
Problems 307
13. One type of electrostatic precipitator has a perfectly

conducting cylinder of radius a placed within a uniform
electric field Eoi.. A uniform flux of positive ions with charge
qo and number density no are injected at infinity and travel
along the field lines with mobility s. Those field lines that
approach the cylinder with Er<0 deposit ions, which redis
tribute themselves uniformly on the surface of the cylinder.
The self-field due to the injected charge is negligible
compared to E0
.
x
Uniform flux of ions with mobility
E , number density no, and charge q0
(a) If the uniformly distributed charge per unit length on

the cylinder is A(t), what is the field distribution? Where is
the electric field zero? This point is called a critical point
because ions flowing past one side of this point miss the
cylinder while those on the other side are collected. What
equation do the field lines obey? (Hint: To the field solution
of Section 4-3-2a, add the field due to a line charge A.)
(b) Over what range of angle 0, 0, <4 <27r - k ,, is there
a window (shaded region in figure) for charge collection as a
function of A (t)? (Hint: Er < 0 for charge collection.)
(c) What is the maximum amount of charge per unit
length that can be collected on the cylinder?
(d) What is the cylinder charging current per unit length?
(Hint: dI = -qonopEra do)
(e) Over what range of y = y* at r = 0s,= -r7 do the
injected ions impact on the cylinder as a function of A(t)?
What is this charging current per unit length? Compare to
(d).
14. The cylinder of Section 4-3,2 placed within a lossy
medium is allowed to reach the steady state.
(a) At t = 0 the imposed electric field at infinity is suddenly
set to zero. What is the time dependence of the surface charge

distribution at r = a?
(b) Find the surface charge distribution if the field at
infinity is a sinusoidal function of time E0 cos wt.
15. A perfectly conducting cylindrical can of radius c open at
one end has its inside surface coated with a resistive layer.
The bottom at z = 0 and a perfectly conducting center post of
radius a are at zero potential, while a constant potential Vo is
imposed at the top of the can.
_ V0
+
-Oo
0-
00

within the structure (a<r<c, 0<z <l)? (Hint: Try the zero
separation constant solutions n = 0, k = 0.)
(b) What is the surface charge distribution and the total
charge at r=a,r=b,and z =0?
(c) What is the equation of the field lines in the free space
region?
16. An Ohmic conducting cylinder of radius a is surrounded
by a grounded perfectly conducting cylindrical can of radius b
open at one end. A voltage Vo is applied at the top of the
resistive cylinder. Neglect variations with o.
within the structure, 0<z< 1, 0<r<b? (Hint: Try the
zero separation constant solutions n =0, k =0 in each region.)
(b) What is the surface charge distribution and total charge
on the interface at r = a?
(c) What is the equation or the field lines in the free space
region?
Problems 309
VO
XII
S b
o--,
Section 4.4
17. A perfectly conducting hemisphere of radius R is placed
upon a ground plane of infinite extent. A uniform field Eoi. is
applied at infinity.
(a) How much more charge is on the hemisphere than

would be on the plane over the area occupied by the hemi
sphere.
(b) If the hemisphere has mass density pm and is in a
gravity field -gi., how large must E0 be to lift the hemi
sphere? Hint:
cosm+i e
sin 0 cos' 0 dO = -____0
m+1
18. A sphere of radius R, permittivity e2, and Ohmic
conductivity a-2 is placed within a medium of permittivity ei
and conductivity o-,. A uniform electric field Eoi, is suddenly
turned on at t = 0.
(a) What are the necessary boundary and initial condi
tions?
~2.0
(b) What are the potential and electric field distributions as

a function of time?
(c) What is the surface charge at r = R?
(d) Repeat (b) and (c) if the applied field varies sinusoidally
with time as EO cos ot and has been on a long time.
19. The surface charge distribution on a dielectric sphere
with permittivity 62 and radius R is
crf = o-0( 3 cos2 0-1)
The surrounding medium has permittivity eI. What are the

potential and electric field distributions? (Hint: Try the n =
2 solutions.)
20. A permanently polarized sphere P2i, of radius R is
placed within a polarized medium Pii.. A uniform electric
field Eoi, is applied at infinity. There is no free charge at
r R. What are the potential and electric field distributions?
P2'Z
P~i
EO iz
21. A point dipole p= pi, is placed at the center of a dielec

tric sphere that is surrounded by a different dielectric
medium. There is no free surface charge on the interface.
Problems 311
What are the potential and electric field distributions? Hint:

p cos 9
lim V(r, 0) = Cos22
r-0 47Er
R
P
E2
Section 4.5
22. The conducting box with sides of length d in Section
4-5-2 is filled with a uniform distribution of volume charge
with density
Po 7-d22 [coul-m ]
What are the potentials at the four interior points when the
outside of the box is grounded?
23. Repeat the relaxation procedure of Section 4-5-2 if the
boundary potentials are:
V2 =-2 V 2 =-2
V1 = 1
V 1 =1 V 3 =3 V3 =-3
V4 =-4 V4=4
(a) (b)
(a) V, = 1, V 2 = -2, V 3 = 3, V 4 = -4
(b) V,= 1, V 2 = -2, V 3 = -3, V4 = 4
(c) Compare to four decimal places with the exact solution.
chapter S
the magnetic
field
314 The Magnetic Field
The ancient Chinese knew that the iron oxide magnetite

(FesO 4 ) attracted small pieces of iron. The first application of
this effect was the navigation compass, which was not
developed until the thirteenth century. No major advances
were made again until the early nineteenth century when
precise experiments discovered the properties of the
magnetic field.
5-1 FORCES ON MOVING CHARGES
5-1-1 The Lorentz Force Law
It was well known that magnets exert forces on each other,

but in 1820 Oersted discovered that a magnet placed near a
current carrying wire will align itself perpendicular to the
wire. Each charge q in the wire, moving with velocity v in the
magnetic field B [teslas, (kg-s 2 -A-')], felt the empirically
determined Lorentz force perpendicular to both v and B
f =q(vx B) (1)
as illustrated in Figure 5-1. A distribution of charge feels a
differential force df on each moving incremental charge
element dq:
df = dq(vx B) (2)
B q
f q(v x B)
Figure 5-1 A charge moving through a magnetic field experiences the Lorentz force
perpendicular to both its motion and the magnetic field.
Forces on Moving Charges 315
Moving charges over a line, surface, or volume, respectively

constitute line, surface, and volume currents, as in Figure 5-2,
where (2) becomes
pfv x B dV= Jx B dV (J = pfv, volume current density)

dS
df= a-vxB dS=KXB
(K = orfv, surface current density) (3)
AfvxB dl =IxB dl (I=Afv, line current)
v : -- I dl =--ev
di
df = Idl x B
(a)
B
dS
K dS
di >
d1 KdSx B
(b)
dV
1K----------+-. JdV
df JdVx B
(c)
Figure 5-2 Moving line, surface, and volume charge distributions constitute currents.
(a) In metallic wires the net charge is zero since there are equal amounts of negative
and positive charges so that the Coulombic force is zero. Since the positive charge is
essentially stationary, only the moving electrons contribute to the line current in the
direction opposite to their motion. (b) Surface current. (c) Volume current.
316 T& Magewic Field
The total magnetic force on a current distribution is then

obtained by integrating (3) over the total volume, surface, or
contour containing the current. If there is a net charge with
its associated electric field E, the total force densities include
the Coulombic contribution:
f=q(E+vxB) Newton
FL=Af(E+vxB)=AfE+IXB N/m
Fs=a'(E+vxB)=o-rE+KXB N/M 2
Fv=pf(E+vxB)=pfE+JXB N/M 3
In many cases the net charge in a system is very small so that
the Coulombic force is negligible. This is often true for
conduction in metal wires. A net current still flows because of
the difference in velocities of each charge carrier.
Unlike the electric field, the magnetic field cannot change
the kinetic energy of a moving charge as the force is perpen
dicular to the velocity. It can alter the charge's trajectory but
not its velocity magnitude.
5-1-2 Charge Motions in a Uniform Magnetic Field
The three components of Newton's law for a charge q of

mass m moving through a uniform magnetic field Bi, are
dv.
m -d = qv,B,
di
dv dv,
m-=qvxB4' m--=-qv.B. (5)
dt dt
dv.
m =0 * v, = const
The velocity component along the magnetic field is

unaffected. Solving the first equation for v, and substituting
the result into the siecond equation gives us a single equation
in v.:
d v +2 1 dv. qB.
S+WoV. = 0, V, =- - , w=-m (6)
where Wo is called the Larmor angular velocity or the cyclo

tron frequency (see Section 5-1-4). The solutions to (6) are
v. =A sin wot + A 2 COS (7)
1 dv,
v, -- =A1 cos wot-A 2 sin coot
wo dt
where A I and A 2 are found from initial conditions. If at t =0,
v(t = 0) = voi (8)
then (7) and Figure 5-3a show that the particle travels in a
circle, with constant speed vo in the xy plane:
v = vo(cos aoti. -sin woti,) (9)
with radius
R = volwo (10)
If the particle also has a velocity component along the

magnetic field in the z direction, the charge trajectory
becomes a helix, as shown in Figure 5-3b.
y
2-ir gB,
Vo iY q
V0 V 0 ix
t ( 2n + 1) t =--(2n
wo 2
+
WO 2 r
_x
Bzis/ -V01 WO (2n +- 1)
(a)
- B,
00 UUUU.MUUU
()
Figure 5-3 (a) A positive charge q, initially moving perpendicular to a magnetic field,
feels an orthogonal force putting the charge into a circular motion about the magnetic
field where the Lorentz force is balanced by the centrifugal force. Note that the charge
travels in the direction (in this case clockwise) so that its self-field through the loop [see
Section 5-2-1] is opposite in direction to the applied field. (b) A velocity component in
the direction of the magnetic field is unaffected resulting in a helical trajectory.
5-1-3 The Mass Spectrograph
The mass spectrograph uses the circular motion derived in

Section 5-1-2 to determine the masses of ions and to measure
the relative proportions of isotopes, as shown in Figure 5-4.
Charges enter between parallel plate electrodes with a y-
directed velocity distribution. To pick out those charges with
a particular magnitude of velocity, perpendicular electric and
magnetic fields are imposed so that the net force on a charge
is
q(E.
q. +vB.) (11.)
For charges to pass through the narrow slit at the end of the
channel, they must not be deflected by the fields so that the
force in (11) is zero. For a selected velocity v, = vo this
requires a negatively x directed electric field
V
E. =- = - voBo (12)
S
which is adjusted by fixing the applied voltage V. Once the
charge passes through the slit, it no longer feels the electric
field and is only under the influence of the magnetic field. It
thus travels in a circle of radius
r= = m (13)
wo qBo
+ v
B0 i,
Photographic
plate
Iq
insulator qBo -Ex
Figure 5-4 The mass spectrograph measures the mass of an ion by the radius of its
trajectory when moving perpendicular to a magnetic field. The crossed uniform
electric field selects the ion velocity that can pass through the slit.
which is directly proportional to the mass of the ion. By

measuring the position of the charge when it hits the photo
graphic plate, the mass of the ion can be calculated. Different
isotopes that have the same number of protons but different
amounts of neutrons will hit the plate at different positions.
For example, if the mass spectrograph has an applied
voltage of V= -100 V across a 1-cm gap (E. = -- 10 V/m) with
a magnetic field of 1 tesla, only ions with velocity
v,= -EIBo= 104 m/sec (14)
will pass through. The three isotopes of magnesium, 12 Mg24
25 26
12 Mg , 12 Mg , each deficient of one electron, will hit the
photographic plate at respective positions:
2 x 10 4N(1.67 x 10- 2 7
d=2r= 1.610-'(1) 2X10 N
)
= 0.48, 0.50, 0.52cm (15)
where N is the number of protons and neutrons (m = 1.67 x
10-27 kg) in the nucleus.
5-1-4 The Cyclotron
A cyclotron brings charged particles to very high speeds by

many small repeated accelerations. Basically it is composed of
a split hollow cylinder, as shown in Figure 5-5, where each
half is called a "dee" because their shape is similar to the
_ -- D2
Figure 5-5 The cyclotron brings ions to high speed by many small repeated accelera
tions by the electric field in the gap between dees. Within the dees the electric field is
negligible so that the ions move in increasingly larger circular orbits due to an applied
magnetic field perpendicular to their motion.
fourth letter of the alphabet. The dees are put at a sinusoi

dally varying potential difference. A uniform magnetic field
Boi, is applied along the axis of the cylinder. The electric field
is essentially zero within the cylindrical volume and assumed
uniform E, = v(t)/s in the small gap between dees. A charge
source at the center of D, emits a charge q of mass m with zero
velocity at the peak of the applied voltage at t = 0. The electric
field in the gap accelerates the charge towards D 2 . Because the
gap is so small the voltage remains approximately constant at
VO while the charge is traveling between dees so that its
displacement and velocity are
dv, q Vo qVO
dt s Sm
dy qVot 2 (16)
dt 2ms
The charge thus enters D 2 at time t = [2ms 2 /qV 0]" 2 later with
velocity v, = -,12qVo/m. Within D 2 the electric field is negligible
so that the charge travels in a circular orbit of radius r =
v,/co = mvIqBo due to the magnetic field alone. The
frequency of the voltage is adjusted to just equal the angular
velocity wo = qBo/m of the charge, so that when the charge
re-enters the gap between dees the polarity has reversed
accelerating- the charge towards D, with increased
velocity. This process is continually repeated, since every time
the charge enters the gap the voltage polarity accelerates the
charge towards the opposite dee, resulting in a larger radius
of travel. Each time the charge crosses the gap its velocity is
increased by the same amount so that after n gap traversals its
velocity and orbit radius are
V = , R1 = = (2nm Vo) 1/2 (17)

M - 0o qBO
If the outer radius of the dees is R, the maximum speed of

the charge
Vma. =oR = -R (18)
is reached after 2n = qB2R 2/mVo round trips when R. = R.

For a hydrogen ion (q = 1.6x 10-' 9 coul, m = 1.67X 10-27 kg),
within a magnetic field of 1 tesla (wo= 9.6 X 107 radian/sec)
and peak voltage of 100 volts with a cyclotron radius of one
meter, we reach vma,= 9 .6 x 10 7 m/s (which is about 30% of
the speed of light) in about 2n -9.6 x 105 round-trips, which
takes a time r=4nir/w, 27r/100-0.06 sec. To reach this
Forceson Moving Charges
speed with an electrostatic accelerator would require

2
b mv2 =qV4 Vmvma 48 x 106 Volts (19)
2q
The cyclotron works at much lower voltages because the
angular velocity of the ions remains constant for fixed qBo/m
and thus arrives at the gap in phase with the peak of the
applied voltage so that it is sequentially accelerated towards
the opposite dee. It is not used with electrons because their
small mass allows them to reach relativistic velocities close to
the speed of light, which then greatly increases their mass,
decreasing their angular velocity wo, putting them out of
phase with the voltage.
5-1-5 HaDl Effect
When charges flow perpendicular to a magnetic field, the

transverse displacement due to the Lorentz force can give rise
to an electric field. The geometry in Figure 5-6 has a uniform
magnetic field Boi, applied to a material carrying a current in
the y direction. For positive charges as for holes in a p-type
semiconductor, the charge velocity is also in the positive y
direction, while for negative charges as occur in metals or in
n-type semiconductors, the charge velocity is in the negative y
direction. In the steady state where the charge velocity does
not vary with time, the net force on the charges must be zero,
BO i,
Figure 5-6 A magnetic field perpendicular to a current flow deflects the charges
transversely giving rise to an electric field and the Hall voltage. The polarity of the
voltage is the same as the sign of tbe charge carriers.
which requires the presence of an x-directed electric field

E+vx B=0->Ex = -v,Bo (20)
A transverse potential difference then develops across the
material called the Hall voltage:
Vh=- Exdx =vBod (21)
The Hall voltage has its polarity given by the sign of v,;
positive voltage for positive charge carriers and negative
voltage for negative charges. This measurement provides an
easy way to determine the sign of the predominant charge
carrier for conduction.
5-2 MAGNETIC FIELD DUE TO CURRENTS
Once it was demonstrated that electric currents exert forces

on magnets, Ampere immediately showed that electric cur
rents also exert forces on each other and that a magnet could
be replaced by an equivalent current with the same result.
Now magnetic fields could be turned on and off at will with
their strength easily controlled.
5-2-1 The Biot-Savart Law
Biot and Savart quantified Ampere's measurements by

showing that the magnetic field B at a distance r from a
moving charge is
goqv X2 i-
B= 47r teslas (kg-s -A') (1)
as in Figure 5-7a, where go is a constant called the permeabil

ity of free space and in SI units is defined as having the exact
numerical value
o= 47T X 10-7 henry/m (kg-m-A-2s 2 ) (2)
The 47r is introduced in (1) for the same reason it was intro
duced in Coulomb's law in Section 2-2-1. It will cancel out a
47r contribution in frequently used laws that we will soon
derive from (1). As for Coulomb's law, the magnetic field
drops off inversely as the square of the distance, but its direc
tion is now perpendicular both to the direction of charge flow
and to the line joining the charge to the field point.
In the experiments of Ampere and those of Biot and
Savart, the charge flow was constrained as a line current
within a wire. If the charge is distributed over a line with
Magnetic Field Due to Currents 323
V
r
QP
IdI
>
B W
rQP
1
QP
K dS
e (c)
rgp
JdV
8 (d)
Figure 5-7 The magnetic field generated by a current is perpendicular to the current
and the unit vector joining the current element to the field point; (a) point charge; (b)
line current; (c) surface current; (d) volume current.
current I, or a surface with current per unit length K, or over

a volume with current per unit area J, we use the differential-
sized current elements, as in Figures 5-7b-5-7d:
I dl (line current)
dq v = K dS (surface current) (3)
I jdV (volume current)

The total magnetic field for a current distribution is then
obtained by integrating the contributions from all the incre
mental elements:
I dl x io
__
-AO JL 2 (line current)
41r QP
so KdSXIQP
B-- -u--- (surface current) (4)
41r is rQP
AoJJdVxiQP
__
i- (volume current)
41r f rQp
The direction of the magnetic field due to a current element

is found by the right-hand rule, where if the forefinger of the
right hand points in the direction of current and the middle
finger in the direction of the field point, then the thumb
points in the direction of the magnetic field. This magnetic
field B can then exert a force on other currents, as given in
Section 5-1-1.
5-2-2 Line Currents
A constant current I, flows in the z direction along a wire of

infinite extent, as in Figure 5-8a. Equivalently, the right-hand
rule allows us to put our thumb in the direction of current.
Then the fingers on the right hand curl in the direction of B,
as shown in Figure 5-8a. The unit vector in the direction of
the line joining an incremental current element I, dz at z to a
field point P is
r z
iQp = i,.cos 0 -i, sin 0=.- (5)
rQP rQP
2 +
-p [r r2 1/2
dB, = p r Ii2d
V 41Fr Qp
r p
r P iQP
12
1 =L B1' 0
BO 2Ira
go i12 L
r 2ira
/1
I(b
CJ~BO
(a) (b)
Figure 5-8 (a) The magnetic field due to an infinitely long z-directed line current is
in the 0 direction. (b) Two parallel line currents attract each other if flowing in the
same direction and repel if oppositely directed.
Magnetic Field Due to Currents 325
with distance
r 2=(z2+r2)1/2 (6)
The magnetic field due to this current element is given by (4)
as
dB=. I dz(i XiQp) - AoIr dz

47r rQp 41r(z 2 +r 2
(
The total magnetic field from the line current is obtained by
integrating the contributions from all elements:
BI AoIir
[ dz
B,-= 4r .Lc (z 2 +r)2 3/2
pjz1r z
-2(Z2 2 1/2
47r r (z+r ) 10
= 'o'i
(8)
21rr
If a second line current 12 of finite length L is placed at a
distance a and parallel to I, as in Figure 5-8b, the force on 12
due to the magnetic field of I, is
+L/2
f=J 12 dzi.xB
-L/2
+4L/2
lpoIi
= I 2 dz (iXi)
L/2 2ara
_4 _1 1oi2L
.
2ra ir (9)
If both currents flow in the same direction (1112>0), the

force is attractive, while if they flow in opposite directions
(1112<0), the force is repulsive. This is opposite in sense to
the Coulombic force where opposite charges attract and like
charges repel.
5-2-3 Current Sheets
(a) Single Sheet of Surface Current

A constant current Koi, flows in the y =0 plane, as in
Figure 5-9a. We break the sheet into incremental line cur
rents Ko dx, each of which gives rise to a magnetic field as
given by (8). From Table 1-2, the unit vector is is equivalent
to the Cartesian components
i' = -sin Oi. 2+cos 4i, (10)

dx
12= KO dx Bz
Ko
-Ko
/2
11 =Kodx 2
x K dB dB dB, +dB2
T-, dB 2
J r (X2+y P
/
(a)
- JdB = - poJo dy' B,

Jo' 2 Lp"d
dB. = ma*1 dy' 7
-y y
2r -d/2 2
t dK=Jody'
-- yojod
IT/
dy'
(b)
Figure 5-9 (a) A uniform surface current of infinite extent generates a uniform
magnetic field oppositely directed on each side of the sheet. The magnetic field is
perpendicular to the surface current but parallel to the plane of the sheet. (b) The
magnetic field due to a slab of volume current is found by superimposing the fields
due to incremental surface currents. (c) Two parallel but oppositely directed surface
current sheets have fields that add in the region between the sheets but cancel outside
the sheet. (d) The force on a current sheet is due to the average field on each side of
the sheet as found by modeling the sheet as a uniform volume current distributed over
an infinitesimal thickness A.
Magnetic FieldDue to Currents 327
z
K, = Koi ,K2 = -Koi

Ao lim JO A = KO
J, o
A-O
!;,OK. - d
.
"I I
-
-~ AO
BI
A2 ADK 0
- o Ko =
(d)
+
B=B +B
1 2
(c)
Figure 5-9
The symmetrically located line charge elements a distance x

on either side of a point P have y magnetic field components
that cancel but x components that add. The total magnetic
field is then
Bx +0 AoKo sin iod

B=--. 2 r (x2 + 2 )1/2
-oKoy +0 dx
21w . (x2 +y 2
- oKo tanIx
21r y -cc
_ -joKo/2, y> 0 (11)

l.oKo/2, y <0
The field is constant and oppositely directed on each side of

the sheet.
(b) Slab of Volume Current

If the z-directed current Joi, is uniform over a thickness d,
as in Figure 5-9b, we break the slab into incremental current
sheets Jo dy'. The magnetic field from each current sheet is
given by (11). When adding the contributions of all the
differential-sized sheets, those to the left of a field point give a

negatively x directed magnetic field while those to the right
contribute a positively x-directed field:
. +d/2 -ojody' -oJod d

Id/2 2 2 ' 2
B==< -- jOd y <-- (12)
J-d 2 2' Ajy 2
p oo dy'+ yfd/ oodyp
2
2
yfdy' d poJo dy' d d
,/2 2 , 2 2 2
The total force per unit area on the slab is zero:
+d/2 +d/2
Fs,=[ JoB dy=-, Jof y dy
2 +d/2
(13)
=
221fld =0
-d/2
A current distribution cannot exert a net force on itself.
(c) Two Parallel Current Sheets

If a second current sheet with current flowing in the
opposite direction - Koi, is placed at y = d parallel to a cur
rent sheet Koi, at y = 0, as in Figure 5-9c, the magnetic field
due to each sheet alone is
-oKo .oKo.
2 2 X, y >O 2 12o, y> d
Bj= B2=' (14)

_oKo
2 12, y'<0
<
Thus in the region outside the sheets, the fields cancel while
-/oKo.
2
2 ', y<d
they add in the region between:

B =, +2 -/yLKoi , 0<y d (5
B0, y<, (15)
The force on a surface current element on the second sheet
is
df = -Koi dSxB (16)
However, since the magnetic field is discontinuous at the
current sheet, it is not clear which value of magnetic field to
use. Tp take the limit properly, we model the current sheet at
y = d as a thin volume current with density Jo and thickness A,
as in Figure 5-9d, where KO = JoA.
Magnetic Field Due to Cu4rents 329
The results of (12) show that in a slab of uniform volume

current, the magnetic field changes linearly to its values at the
surfaces
B.(y = d -A)= -soKo (17)
B.(y =d)=0
so that the magnetic field within the slab is
B. =loKo(y d) (18)
The force per unit area on the slab is then
Fs A Jo(y - d)i, dy
-poKojo(y-d)2 . d
A 2 Id-A
joKoJoA. poKo
.
2 2 (19)
The force acts to separate the sheets because the currents are
in opposite directions and thus repel one another.
Just as we found for the electric field on either side of a
sheet of surface charge in Section 3-9-1, when the magnetic
field is discontinuous on either side of a current sheet K,
being B, on one side and B 2 on the other, the average
magnetic field is used to compute the force on the sheet:
df= KdS x (Bi+ B 2) (20)

(20
2
In our case
Bi =- LoKoi., B 2 =0 (21)
5-2-4 Hoops of Line Current
(a) Single hoop

A circular hoop of radius a centered about the origin in the
xy plane carries a constant current I, as in Figure 5-1Oa. The
distance from any point on the hoop to a point at z along the z
axis is
r 2(Z2+a2 1/2 (22)
in the direction
(-ai,+zi.)
Q (z2+ 2) 2 (23)
2B,a
Helmholtz coil
with d=a
a
-2 -1 0 1 2 3
Highly uniform magnetic
dB = dB, + dB 2 d field in central region
dB 2 > dB, A around a= d
a
iQP [-ai,+ aI 2
( 2 + a2Y1 f a
di =KodA'
'o.i
y
adO
(a) (b) (c)

Figure 5-10 (a) The magnetic field due to a circular current loop is z directed along
the axis of the hoop. (b) A Helmholtz coil, formed by two such hoops at a distance
apart d equal to their radius, has an essentially uniform field 'iear the center at z = d/2.
(c) The magnetic field on the axis of a cylinder with a 45-directed surface current is
found by integrating the fields due to incremental current loops.
so that the incremental magnetic field due to a current ele

ment of differential size is
dB= ""Ia d4a i)xiQP

4,rrto,
= oIad (+
2 E s/2(aiz+zir) (24)
47r(z +a )
The radial unit vector changes direction as a function of 4,
being oppositely directed at -0, so that the total magnetic
field due to the whole hoop is purely z directed:
B ola 2 2d
B 41r(z2 +a) 2
2
poIa
(25)
2(z 2 +a 2)512
The direction of the magnetic field can be checked using
the right-hand rule. Curling the fingers on the right hand in
the direction of.the current puts the thumb in the direction of
Magnetic Field Due to Cunents 331
the magnetic field. Note that the magnetic field along the z axis
is positively z directed both above and below the hoop.
(b) Two Hoops (Hehnholtz Coil)

Often it is desired to have an accessible region in space with
an essentially uniform magnetic field. This can be arianged
by placing another coil at z = d, as in Figure 5-1 Ob. Then the
total magnetic field along the z axis is found by superposing
the field of (25) for each hoop:
I 0 Ia2 1 1
B.= 2 \(z2+a2)/2+ ((z - d)2+a )S/2) (26)
We see then that the slope of B.,
aB. 3 oIa2 ( -z (z -d) \

az 2 \(z 2 +a) 5 2
((z -d) 2 +a 2)5/2) (27)
is zero at z = d/2. The second derivative,
a2B. 3 poIa 2 ( 5z 2
az 2 (z 2+a )7/ (z 2 +a 2 )5 /2
5(z-d)2 1
((z - d) +a 2) ((z - d)2+ a 2)/2
can also be set to zero at z = d/2, if d = a, giving a highly (

uniform field around the center of the system, as plotted in
Figure 5-10b. Such a configuration is called a Helmholtz coil.
(c) Hollow Cylinder of Surface Current

A hollow cylinder of length L and radius a has a uniform
surface current K 0 i* as in Figure 5-10c. Such a configuration
is arranged in practice by tightly winding N turns of a wire
around a cylinder and imposing a current I through the wire.
Then the current per unit length is
Ko= NIIL (29)
The magnetic field along the z axis at the position z due to

each incremental hoop at z' is found from (25) by replacing z
by (z - z') and I by Ko dz':
B.t a2Ko dz'

dB. 2[(z - Z')2+ a 2/
332
The Magnetic FieUd
The total axial magnetic field is then

B 12 Oa2 dz'
B.=-. 2 2 [(z - z')+a'
Jsoa2 Ko (z'-z) +02
2 a _[(z,) +a2]I .'---/2
2+ z+L/2 3 _ioKo( -z+L/2 +
2 \[(z - L/2)2 + a2 ]m [(z+L/2)2 +a2]" 2 (31)
As the cylinder becomes very long, the magnetic field far
from the ends becomes approximately constant
lim B.=p, K0 (32)
5-3 DIVERGENCE AND CURL OF THE MAGNETIC FIELD
Because of our success in examining various vector opera

tions on the electric field, it is worthwhile to perform similar
operations on the magnetic field. We will need to use the
following vector identities from Section 1-5-4, Problem 1-24
and Sections 2-4-1 and 2-4-2:
V - (V XA)=0 (1)
Vx(Vf)=O (2)
(3)
rQP) rop
rQp=O (4)
V 2(r V= 0,
dV= - 41, rap= 0(4
V (A x B)= B - (V x A)- A - V x B (5)

V x (A XB)= (B - V)A -(A - V)B+(V - B)A -(V - A)B (6)
V(A - B)= (A - V)B+(B - V)A+ A x (V X B)+B x (V x A)
(7)
5-3-1 Gauss's Law for the Magnetic Field
Using (3) the magnetic field due to a volume distribution of

current J is rewritten as
B=E2 Jx()dV
=2 JA xV( dV (8)
4r Jv rP
Divergence and Curl of the Magnetic Field 333
If we take the divergence of the magnetic field with respect to

field coordinates, the del operator can be brought inside the
integral as the integral is only over the source coordinates:
V-B -. Jxv dV (9)

v L rQp
The integrand can be expanded using (5)
v-Jxv-)]=v(-,)- -(VXJ)-J-Vx[V(-)=0
0
0 (10)
The first term on the right-hand side in (10) is zero because j

is not a function of field coordinates, while the second term is
zero from (2), the curl of the gradient is always zero. Then (9)
reduces to
V-B=0 (11)
This contrasts with Gauss's law for the displacement field
where the right-hand side is equal to the electric charge
density. Since nobody has yet discovered any net magnetic
charge, there is no source term on the right-hand side of (11).
The divergence theorem gives us the equivalent integral
representation
B-dS=0 (12)
tV-BdV=
which tells us that the net magnetic flux through a closed
surface is always zero. As much flux enters a surface as leaves
it. Since there are no magnetic charges to terminate the
magnetic field, the field lines are always closed.
5-3-2 Ampere's Circuital Law
We similarly take the curl of (8) to obtain
VxB=- VxJxVkI dV (13)

47r v I rQP
where again the del operator can be brought inside the
integral and only operates on rQp.
We expand the integrand using (6):
Vx JxV )= v
rP rQP) ~-1(i-VV rQP)
0
rQC)
where two terms on the right-hand side are zero because J is

not a function of the field coordinates. Using the identity of
(7),
V J-v = [V(-V J+( -V)V(

rQP) rQP) IrQP)
+ V x )+Jx [x I (15)
0
the second term on the right-hand side of (14) can be related
to a pure gradient of a quantity because the first and third
terms on the right of (15) are zero since J is not a function of
field coordinates. The last term in (15) is zero because the curl
of a gradient is always zero. Using (14) and (15), (13) can be
rewritten as
VxB=- 4 I J - V( I )-JVk dV (16)

1 v L rQp/J rQp/
Using the gradient theorem, a corollary to the divergence

theorem, (see Problem 1-15a), the first volume integral is
converted to a surface integral
VdS .
I JVB (17)
41r s
rar/ , vrQ7
This surface completely surrounds the current distribution so

that S is outside in a zero current region where J =0 so that
the surface integral is zero. The remaining volume integral is
nonzero only when rQp =0, so that using (4) we finally obtain
V x B= goJ (18)
which is known as Ampere's law.
Stokes' theorem applied to (18) results in Ampere's circuital
law:
Vx--. dS= -- dl= J-dS (19)

s Lo S
Like Gauss's law, choosing the right contour based on sym

metry arguments often allows easy solutions for B.
If we take the divergence of both sides of (18), the left-hand
side is zero because the divergence of the curl of a vector is
always zero. This requires that magnetic field systems have
divergence-free currents so that charge cannot accumulate.
Currents must always flow in closed loops.
Divergence and Curl of the Magnetic Field 335
5-3-3 Currents With Cylindrical Symmetry
(a) Surface Current

A surface current Koi, flows on the surface of an infinitely
long hollow cylinder of radius a. Consider the two sym
metrically located line charge elements dI = Ko ad4 and their
effective fields at a point P in Figure 5-1 la. The magnetic
field due to both current elements cancel in the radial direc
tion but add in the 4 direction. The total magnetic field can
be found by doing a difficult integration over 4. However,
dB = dB1 + dB 2
dB,
dl= Koado . dB2\
C5 p= Ia2+ r2 _-2 ar COS n1
(rP - c O )r, + a sin i,

IQP rQ P fraction of the current
crosses this surface
No current
crosses this
a surface
All the cu rrent
I III
crosses this urface
r i
K =Ko i, Koi
2 B
0 r <a 2r
- rd$=
r 2 r <a
f B0 rd$= 27rKoa r>a
f
0 Po Joira2
(a) (b) (c)
Figure 5-11 (a) The magnetic field of an infinitely long cylinder carrying a surface
current parallel to its axis can be found using the Biot-Savart law for each incremental
line current element. Symmetrically located elements have radial field components
that cancel but 4 field components that add. (b) Now that we know that the field is
purely 4 directed, it is easier to use Ampere's circuital law for a circular contour
concentric with the cylinder. For r <a no current passes through the contour while for
r>a all the current passes through the contour. (c) If the current is uniformly
distributed over the cylinder the smaller contour now encloses a fraction of the
current.
using Ampere's circuital law of (19) is much easier. Since we

know the magnetic field is 4 directed and by symmetry can
only depend on r and not 4 or z, we pick a circular contour of
constant radius r as in Figure 5-11 b. Since dl= r d4 i# is in the
same direction as B, the dot product between the magnetic
field and dl becomes a pure multiplication. For r <a no cur
rent passes through the surface enclosed by the contour,
while for r>a all the current is purely perpendicular to the
normal to the surface of the contour:
B w"B* 2vrrB_ Ko21ra=I, r>a
-t.do 0, r<a
(20)
where I is the total current on the cylinder.
The magnetic field is thus
IoKoa/r = joI/(2arr), r> a
0, r<a
Outside the cylinder, the magnetic field is the same as if all
the current was concentrated along the axis as a line current.
(b) Volume Current

If the cylinder has the current uniformly distributed over
the volume as Joi , the contour surrounding the whole cylin
der still has the total current I = Joira2 passing through it. If
the contour has a radius smaller than that of the cylinder,
only the fraction of current proportional to the enclosed area
passes through the surface as shown in Figure 5-1 1c:
B4 2rrB Jora =I r>a
- r d= 2rrB,= fJO (22)
L go go Jo7r=Ir/a2, r<a
so that the magnetic field is
gojoa 2 =~oo
oI r>a
2r 2vrr'
B, = (3
B ojor soIr (23)
2 2
i,2, r<a
5-4 THE VECTOR POTENTIAL
5-4-1 Uniqueness
Since the divergence of the magnetic field is zero, we may

write the magnetic field as the curl of a vector,
V -B =0=>B= V x A (1)
The Vector Potential 337
where A is called the vector potential, as the divergence of the

curl of any vector is always zero. Often it is easier to calculate
A and then obtain the magnetic field from (1).
From Ampere's law, the vector potential is related to the
current density as
V x B=V x (V x A)=V(V -A)-V 2A = poJ (2)
We see that (1) does not uniquely define A, as we can add

the gradient of any term to A and not change the value of the
magnetic field, since the curl of the gradient of any function is
always zero:
A-+A+Vf>B=Vx(A+Vf)=VxA (3)
Helmholtz's theorem states that to uniquely specify a

vector, both its curl and divergence must be specified and that
far from the sources, the fields must approach zero. To prove
this theorem, let's say that we are given, the curl and diver
gence of A and we are to determine what A is. Is there any
other vector C, different from A that has the same curl and
divergence? We try C of the form
C=A+a (4)
and we will prove that a is zero.

By definition, the curl of C must equal the curl of A so that
the curl of a must be zero:
VxC=Vx(A+a)=VxA=Vxa=0 (5)
This requires that a be derivable from the gradient of a scalar

function f:
V x a= 0>a=Vf (6)
Similarly, the divergence condition requires that the diver

gence of a be zero,
V - C=V - (A+a)=V - A>V - a=0 (7)
so that the Laplacian of f must be zero,
V-a=V 2f=0 (8)
In Chapter 2 we obtained a similar equation and solution for

the electric potential that goes to zero far from the charge
distribution:
V2= _ _ pdV (9)

E Jv4rerr(9)
If we equate f to V, then p must be zero giving us that the

scalar function f is also zero. That is, the solution to Laplace's
equation of (8) for zero sources everywhere is zero, even
though Laplace's equation in a region does have nonzero
solutions if there are sources in other regions of space. With f
zero, from (6) we have that the vector a is also zero and then
C = A, thereby proving Helmholtz's theorem.
5-4-2 The Vector Potential of a Current Distribution
Since we are free to specify the divergence of the vector

potential, we take the simplest case and set it to zero:
V A=0 (10)
Then (2) reduces to
V2A= -oJ(11)
Each vector component of (11) is just Poisson's equation so

that the solution is also analogous to (9)
o J dV
A - --d (12)
41r fv rQp
The vector potential is often easier to use since it is in the

same direction as the current, and we can avoid the often
complicated cross product in the Biot-Savart law. For moving
point charges, as well as for surface and line currents, we use
(12) with the appropriate current elements:
J dV-+K dS-+I dL -+qv (13)
5-4-3 The Vector Potential and Magnetic Flux
Using Stokes' theorem, the magnetic flux through a surface

can be expressed in terms of a line integral of the vector
potential:
<D B - dS Vx A - dS A - dl (14)
(a) Finite Length Line Current

The problem of a line current I of length L, as in Figure
5-12a appears to be nonphysical as the current must be
continuous. However, we can imagine this line current to be
part of a closed loop and we calculate the vector potential and
magnetic field from this part of the loop.
The distance rQp from the current element I dz' to the field
point at coordinate (r, <5, z) is
r2 = [(Z 1Z /)2+ r2]"2 (15)
The vector potential is then
_1_I 1/2 dz'

2
4A r u [(z -z') +r11/2
2
tpoI -z + L/2+[(z - L/2)2+r 12
2 1 1 2
47r -(z+ L/2)+[(z+L/2) 2 +r
pOI( -z+L/2 +sinh z+L/21 (16)

47r r r
2
P(r, 0, z)
Sr '
+r21/2
(zZ')2
SIdz'
L
x
(a)
Figure 5-12 (a) The magnetic field due to a finite length line current is most easily
found using the vector potential, which is in the direction of the current. This problem
is physical only if the line current is considered to be part of a closed loop. (b) The
magnetic field from a length w of surface current is found by superposing the vector
potential of (a) with L - oo. The field lines are lines of constant A. (c) The magnetic
flux through a square current loop is in the -x direction by the right-hand rule.
di = KodxKo i
f (X - x'f +Y 2 112
dx' .- (x,y)
Magnetic field lines (lines of constant A,)

y 2
-x) In [(x -- +y ] + (2 +x)n[(x+2)2 +y 2
)
+2y tan W =Const
x 2 2
X2YI
7;
27)2
(b)
=f B - dS = #A - dl
S
2a
j -11 y
x
L
-D
(c)
Figure 5-12
with associated magnetic field
B=VxA
(1 aA. A aA aA 1 a aAr
" i,+ -'i +- (- (rA,) i
r 8o az az r r \r 84
8A,,
ar
-poIr
4
,7T \[(z - L/2)2 + r 2] 2
_ z + L/2 + [(z - L/2) 2 + r 2] 2
[(z + L/2) 2 + r 2 ] 112 _ (z + L2)+ [(z + L/2) 2 + r] 1/2)
po1I -z + L/2 z + L/2 .(

4rr \[r 2+(Z - L/2)21/2+[r2+(Z + L/2()2)2
+
For large L, (17) approaches the field of an infinitely long
line current as given in Section 5-2-2:
A,= _ Inr+const
27T
lim (18)
ar 27rr
Note that the vector potential constant in (18) is infinite, but

this is unimportant as this constant has no contribution to the
magnetic field.
(b) Finite Width Surface Current

If a surface current Koi,, of width w, is formed by laying
together many line current elements, as in Figure 5-12b, the
vector potential at (x, y) from the line current element KO dx' at
position x' is given by (18):
dA, =-oKo
4 dx' In [(x - x') 2 +y 2
] (19)
7r
The total vector potential is found by integrating over all

elements:
A.=- -LoK +w/2 In [(x -x') 2 +y ] dx'

41rI w12
-LoKo (x'- x) In [(x-x')2 +y]2(x'-x)

4
1r
+2y tan-
K2 - x In x +y = r 1
+2 +x ) n 2)x+ 2 2
-2w +2y tan' g2 +X2 W / (20)*

-W /41)
y _+
The magnetic field is then
ax ay _
= OKO 2 tanx+ln i +w/2)2 +Y2

Y2
47r -w2/4 (x-w/2)2+
(21)
The vector potential in two-dimensional geometries is also
useful in plotting field lines,
dy = B, -- 8A/x (22)
dx B. aA./ay
for if we cross multiply (22),
-'dx+ -'dy=dA=0->A.=const (23)

ax ay
we see that it is constant on a field line. The field lines in

Figure 5-12b are just lines of constant A,. The vector poten
tial thus plays the same role as the electric stream function in
Sections 4.3.2b and 4.4.3b.
(c) Flux Through a Square Loop

The vector potential for the square loop in Figure 5-12c with
very small radius a is found by superposing (16) for each side
with each component of A in the same direction as the current
in each leg. The resulting magnetic field is then given by four
*tan (a - b)+ tan-' (a + b)= tan-' 1 -a'2

Magnetization 343
terms like that in (17) so that the flux can be directly computed
by integrating the normal component of B over the loop area.
This method is straightforward but the algebra is cumber
some.
An easier method is to use (14) since we already know the
vector potential along each leg. We pick a contour that runs
along the inside wire boundary at small radius a. Since each
leg is identical, we only have to integrate over one leg, then
multiply the result by 4:
-a+D/2
4)=4 A, dz
ra-D/2
_ ,O -a+D/2 sinh +D/2 _ z+D/2

irf.- D/2 aa
)
= o - _-z) sinh 1 -D/+ --Z) +a 21/
V H 2 a [2
+ D+Z sinW z+D/2

D +z2 + 21/21 -a+D/2
2 a 2 a-D/2
=2 oI -- a sinh- I +al+(D -a) sinhV D-a
-[(D - a) 2 + a21/2) (24)

As a becomes very small, (24) reduces to
lim4D=2 LD sinh D 1) (25)

a-0 7r \a/
We see that the flux through the loop is proportional to the

current. This proportionality constant is called the self-
inductance and is only a function of the geometry:
1
L = =2 - sinh-' - )- 1 (26)
I 7T\(
(
Inductance is more fully developed in Chapter 6.
5-5 MAGNETIZATION
Our development thus far has been restricted to magnetic

fields in free space arising from imposed current dis
tributions. Just as small charge displacements in dielectric
materials contributed to the electric field, atomic motions
constitute microscopic currents, which also contribute to the
magnetic field. There is a direct analogy between polarization
and magnetization, so our development will parallel that of
Section 3-1.
5-5-1 The Magnetic Dipole
Classical atomic models describe an atom as orbiting elec

trons about a positively charged nucleus, as in Figure 5-13.
Figure 5-13 Atomic currents arise from orbiting electrons in addition to the spin
contributions from the electron and nucleus.
The nucleus and electron can also be imagined to be spin

ning. The simplest model for these atomic currents is analo
gous to the electric dipole and consists of a small current loop
of area dS carrying a current I, as in Figure 5-14. Because
atomic dimensions are so small, we are only interested in the
magnetic field far from this magnetic dipole. Then the shape
of the loop is not important, thus for simplicity we take it to be
rectangular.
The vector potential for this loop is then
I I I I
A=-.i r3 r
dx idy i, r4 rs (1)
41r
where we assume that the distance from any point on each

side of the loop to the field point P is approximately constant.
m = Idxdyi! m =IdS
1P
r3'
r2 r1
dSdxdyi,
dy X
,
dx r4 X
dS
S ir iV COSX,
A dy ,-(- )=COSX2
Figure 5-14 A magnetic dipole consists of a small circulating current loop. The
magnetic moment is in the direction normal to the loop by the right-hand rule.
Magnetization 345
Using the law of cosines, these distances are related as
r=r2+ -Yrdycosi; r=r2+ ( Xdxc

r) =- r dy2COS x1, r2 =r+ - +rdx COS X2() (2)
2 2(dy2 2 =2 d2()
r3 =r +) + rdycos X, r4 r + -+rdx COS X2
where the angles X, and X2 are related to the spherical coor

dinates from Table 1-2 as
i, i,=cosX =sin 6 sin (3)
-i,-ix= cos X2= -sin 6 cos k
In the far field limit (1) becomes
lim A = /.0I [dx( I

>>dx 41r [r dy dy 1/2
r>>dy 1 + -+2 cOSX
S2r 2r
1 1/2
1+-2 cosi
r\ dx dx 1 112
r + -(-+2 cos K2)
2r 2r
dx 1 1/2)]
1+- (--2 cos X2
2r 2r
~ 2 dxdy[cos Xiii +cos X i,]

11r (4)
2
Using (3), (4) further reduces to
MoI dS
2
A = 47Tr sin [ - sin i + cos 0i,]
MoIdS
= 4Tr
2
sin Oi4, (5)
where we again used Table 1-2 to write the bracketed

Cartesian unit vector term as is. The magnetic dipole
moment m is defined as the vector in the direction perpen
dicular to the loop (in this case i,) by the right-hand rule with
magnitude equal to the product of the current and loop area:
m= I dS i =I dS (6)
Then the vector potential can be more generally written as
A=
4 lrr7sinNO=7x, 4 7rr2 (7)
with associated magnetic field
a(A,6sin O)i, - (rAs)ie

1 a ia
B=VxA=
r sin 0ea r ar
p.om
- M [2 cos Oir+ sin i]
4,7rr'
(8)
This field is identical in form to the electric dipole field of

Section 3-1-1 if we replace p/Eo by Mom.
5-5-2 Magnetization Currents
Ampere modeled magnetic materials as having the volume

filled with such infinitesimal circulating current loops with
number density N, as illustrated in Figure 5-15. The
magnetization vector M is then defined as the magnetic dipole
density:
M= Nm= NI dS amp/m (9)
For the differential sized contour in the xy plane shown in
Figure 5-15, only those dipoles with moments in the x or y
directions (thus z components of currents) will give rise to
currents crossing perpendicularly through the surface
bounded by the contour. Those dipoles completely within the
contour give no net current as the current passes through the
contour twice, once in the positive z direction and on its
return in the negative z direction. Only those dipoles on
either side of the edges-so that the current only passes
through the contour once, with the return outside the
contour-give a net current through the loop.
Because the length of the contour sides Ax and Ay are of
differential size, we assume that the dipoles along each edge
do not change magnitude or direction. Then the net total
current linked by the contour near each side is equal to the
pioduct of the current per dipole I and the number of
dipoles that just pass through the contour once. If the normal
vector to the dipole loop (in the direction of m) makes an
angle 0 with respect to the direction of the contour side at
position x, the net current linked along the line at x is
- INdS Ay cos 01,= -M,(x) Ay (10)
The minus sign arises because the current within the contour
adjacent to the line at coordinate x flows in the - z direction.
Magnetization 347
O Oo
001~00h
O 010 U/9
009
0 0 0
70
IY
Ae
M,
(X, Y)
/0 Oor--- -7
, , 9 - 0
/Ax /0000
Ay-
1.M
(x. y)
A Cos
dS
Figure 5-15 Many such magnetic dipoles within a material linking a closed contour
gives rise to an effective magnetization current that is also a source of the magnetic
field.
Similarly, near the edge at coordinate x +Ax, the net current
linked perpendicular to the contour is
IN dSAy cos 01.+ =M,(x+Ax) Ay (11)

Along the edges at y and y + Ay, the current contributions are

INdS Ax cos 01,= M,(y) Ax
-INdS Ax cos 61,,A, = -M (y +Ay) Ax (12)
The total current in the z direction linked by this contour is
thus the sum of contributions in (10)-(12):
M,(x)_ Mx(Y +AY)- M.(Y)

~A (x+AX)-
Itot=I,AX AY(M,
AY=AxA
(13)
If the magnetization is uniform, the net total current is zero
as the current passing through the loop at one side is canceled
by the current flowing in the opposite direction at the other
side. Only if the magnetization changes with position can
there be a net current through the loop's surface. This can be
accomplished if either the current per dipole, area per dipole,
density of dipoles, of angle of orientation of the dipoles is a
function of position.
In the limit as Ax and Ay become small, terms on the
right-hand side in (13) define partial derivatives so that the
current per unit area in the z direction is
Iz ., / M, OM,\
lim Im m= (V X M), (14)
A,-0
Ay-0
Ax Ay ax ay
which we recognize as the z component of the curl of the

magnetization. If we had orientated our loop in the xz or yz
planes, the current density components would similarly obey
the relations
j,= az , ax) = (V xM)
(15)
= (V x M)(
jx = (aM. amy)
ay az
so that in general
Jm=VxM (16)
where we subscript the current density with an m to represent
the magnetization current density, often called the Amperian
current density.
These currents are also sources of the magnetic field and
can be used in Ampere's law as
B
V x -= J_+J= J+V xM (17)
where Jf is the free current due to the motion of free charges

as contrasted to the magnetization current J_, which is due to
the motion of bound charges in materials.
Magnetization 349
As we can only impose free currents, it is convenient to

define the vector H as the magnetic field intensity to be
distinguished from B, which we will now call the magnetic
flux density:
B
H =--M=> B= o(H + M) (18)
Ao
Then (17) can be recast as
Vx - M) =V x H=J, (19)
Ao/
The divergence and flux relations of Section 5-3-1 are
unchanged and are in terms of the magnetic flux density B.
In free space, where M = 0, the relation of (19) between B and
H reduces to
B= pOH (20)
This is analogous to the development of the polarization
with the relationships of D, E, and P. Note that in (18), the
constant parameter uo multiplies both H and M, unlike the
permittivity eo which only multiplies E.
Equation (19) can be put into an equivalent integral form
using Stokes' theorem:
I (VH)-dS= H-dl= J,-dS (21)
The free current density J1 is the source of the H field, the

magnetization current density J. is the source of the M field,
while the total current, Jf+J., is the source of the B field.
5-5-3 Magnetic Materials
There are direct analogies between the polarization pro

cesses found in dielectrics and magnetic effects. The consti
tutive law relating the magnetization M to an applied
magnetic field H is found by applying the Lorentz force to
our atomic models.
(a) Diamagnetism
The orbiting electrons as atomic current loops is analogous
to electronic polarization, with the current in the direction
opposite to their velocity. If the electron (e = 1.6x 10- 9coul)
rotates at angular speed w at radius R, as in Figure 5-16, the
current and dipole moment are
I=- m=IrR2= R (22)

2,fr 2
L=m, wR 2i=- m
-eB
ID
2
2'2
m =-IrR20 ~w 2
Figure 5-16 The orbiting electron has its magnetic moment m in the direction
opposite to its angular momentum L because the current is opposite to the electron's
velocity.
Note that the angular momentum L and magnetic moment m

are oppositely directed and are related as
L= mRi, x v= moR 2 i.= -2m, (23)

e
where m, = 9.1 X 10-3' kg is the electron mass.

Since quantum theory requires the angular momentum to
be quantized in units of h/2w, where Planck's constant is
h=6.62xi0 4'joule-sec, the smallest unit of magnetic
moment, known as the Bohr magneton, is
mB = A ~9.3 x 10 24 amp-m2 (24)

41rm,
Within a homogeneous material these dipoles are

randomly distributed so that for every electron orbiting in
one direction, another electron nearby is orbiting in the
opposite direction so that in the absence of an applied
magnetic field there is no net magnetization.
The Coulombic attractive force on the orbiting electron
towards the nucleus with atomic number Z is balanced by the
centrifugal force:
Ze 2 2
m.10 2R = 4eo2 (25)
41reoR
Since the left-hand side is just proportional to the square of

the quantized angular momentum, the orbit radius R is also
quantized for which the smallest value is
47T- h 2 5X10
R = M,-e02
"
m (26)
Magnetization 351
with resulting angular speed
C = Z2 e sM. 1 3 X 10' 6 Z2 (27)

(4re)2(h/2Fr
When a magnetic field Hoi, is applied, as in Figure 5-17,

electron loops with magnetic moment opposite to the field
feel an additional radial force inwards, while loops with
colinear moment and field feel a radial force outwards. Since
the orbital radius R cannot change because it is quantized,
this magnetic force results in a change of orbital speed Aw:
e +(w +AwI)RIoHo) m,(w +Awl) 2 R= e

(41reoR
2
m,(W +AW 2) R=e Ze 2-(W +AW 2)RyoHo (28)
(47rsoR
where the first electron speeds up while the second one slows
down.
Because the change in speed Aw is much less than the
natural speed w, we solve (28) approximately as
Awl = ewApoHo
2ma - ejoHo (29)
- epi oHo
2mpw + eyoHo
where we neglect quantities of order (AW)2 . However, even

with very high magnetic field strengths of Ho= 106 amp/m we
see that usually
eI.oHo< 2mwo
(1.6 X 10~ 19)(41r X 101).106 < 2(9.1 X 10-3")(1.3 X I 0'r)(30)
Hoi, Hoi,
-e v xB
evxR
+
Figure 5-17 Diamagnetic effects, although usually small, arise in all materials because
dipoles with moments parallel to the magnetic field have an increase in the orbiting
electron speed while those dipoles with moments opposite to the field have a decrease
in speed. The loop radius remains constant because it is quantized.
so that (29) further reduces to
Aw - Aw 2 " eyoHo _1. 1 X I 5 Ho (31)

2m,
The net magnetic moment for this pair of loops,
eR 2 _ 2p oR 2
M 2 (w2 -oi)=-eR2 Aw e Ho (32)
2 2m,
is opposite in direction to the applied magnetic field.
If we have N such loop pairs per unit volume, the
magnetization field is
NesioR2
M=Nm= - Hoi. (33)
2m,
which is also oppositely directed to the applied magnetic field.

Since the magnetization is linearly related to the field, we
define the magnetic susceptibility Xm as
M= XmH, X- = -2 0R (34)
2m,
where X, is negative. The magnetic flux density is then

B = AO(H +M)= po(1+Xm)H = Aog H = yH (35)
where , = 1 +Xm is called the relative permeability and A is
the permeability. In free space Xm = 0 so that ji, = 1 and
A = yLo. The last relation in (35) is usually convenient to use, as
all the results in free space are still correct within linear
permeable material if we replace ylo by 1L. In diamagnetic
materials, where the susceptibility is negative, we have that
y, < 1, y < jO. However, substituting in our typical values
Ne 2oR 4.4 X 10 3 5
Xm = - 2
2m z2 N (36)
we see that even with Nz 1030 atoms/M3 , Xy is much less than

unity so that diamagnetic effects are very small.
(b) Paramagnetism
As for orientation polarization, an applied magnetic field
exerts a torque on each dipole tending to align its moment
with the field, as illustrated for the rectangular magnetic
dipole with moment at an angle 0 to a uniform magnetic field
B in Figure 5-18a. The force on each leg is
dfI = - df2 = I Ax i. X B = I Ax[Bi, - Bzi,]
df3 = -df 4 = I Ay i, B= I Ay(- B.+Bj+,j)
In a uniform magnetic field, the forces on opposite legs are
equal in magnitude but opposite in direction so that the net
Magnetization 353
df4 =-Ii xBAy=-IAy(-Bxi, +Bi 2
)
df1 =Iix BAx = IAx [By i, -B , i _
AxB 2 B >
df2 -Iii x BAx =- Ax(B i, -B, i

d =Iiy x BAy = Ay-B, i + B, ii
B
BB
Ayx
Figure 5-18 (a) A torque is exerted on a magnetic dipole with moment at an angle 9
to an applied magnetic field. (b) From Boltzmann statistics, thermal agitation opposes
the alignment of magnetic dipoles. All the dipoles at an angle 0, together have a net
magnetization in the direction of the applied field.
force on the loop is zero. However, there is a torque:

4
T= l',r xdf.
(-i,xdf 1 +,,xd 2 )+ 2 (i~xdfi- df 4

)
= I Ax Ay(B.i,-B,i.)=:mXB (38)
The incremental amount of work necessary to turn the

dipole by a small angle dO is
dW = Td = myzoHo sin 0 dO (39)
so that the total amount of work necessa'ry to turn the dipole
from 0 =0 to any value of 0 is
W= TdO= -myoH cos 0I = mj.oHo(1-cos 0)

(40)
This work is stored as potential energy, for if the dipole is
released it will try to orient itself with its moment parallel to
the field. Thermal agitation opposes this alignment where
Boltzmann statistics describes the number density of dipoles
having energy W as
n = ne -WAT = n I mpoHO(I-cos 0)/kT = noemoLOHo cos 0/hT
(41)
where we lump the constant energy contribution in (40)
within the amplitude no, which is found by specifying the
average number density of dipoles N within a sphere of
radius R:
2w R
1 r
N=i- I noe's*0r2sin 0 drdOdo
rR 9
o .o f-0
=no sin 0e' "**dO (42)
where we let
a = myoHo/kT (43)
With the change of variable
u =acos 0, du = -a sin 9 dO (44)
the integration in (42) becomes
N= s sinh a
e' du =- (45)
2a a
so that (41) becomes
Na e (46)
sinh a
From Figure 5-18b we see that all the dipoles in the shell
over the interval 0 to 0 + dO contribute to a net magnetization.
which is in the direction of the applied magnetic field:
dM= cos 0 r2 sin dr dO do (47)

3irR
Magnetization 355
so that the total magnetization due to all the dipoles within

the sphere is
maN a
M=I sin 0 cos Oec de (48)
2 sinh a J.=
Again using the change of variable in (44), (48) integrates
to
-- nN C"
M= ue" du
2a sinh aI
-mN u _-"
2a sinh a
= m [e -( -- )a
-inN
2a siha[e-a(-a-1)-ea(a -l)I
-ainN
= [-a cosh a+sinh a]
a sinh a
= mN[coth a - 1/a] (49)
which is known as the Langevin equation and is plotted as a

function of reciprocal temperature in Figure 5-19. At low
temperatures (high a) the magnetization saturates at M = mN
as all the dipoles have their moments aligned with the field.
At room temperature, a is typically very small. Using the
parameters in (26) and (27) in a strong magnetic field of
Ho= 106 amps/m, a is much less than unity:
a= mkoHo R2 OHO=8x 10-4 (50)

kT 2 kT
M
IMmNa M
I3
M = mN(cotha--a)
5 10 15
a- kT
Figure 5-19 The Langevin equation describes the net magnetization. At low
temperatures (high a) all the dipoles align with the field causing saturation. At high
temperatures (a << 1) the magnetization increases linearly with field.
In this limit, Langevin's equation simplifies to

lim M- M 1+a2/2 1
aI La+a3/6 a]
MN((I+a 2/2)(1-a/6) 1
a a]
mNa ptom2 N
3 3 3hT
kTH 0 (51)
In this limit the magnetic susceptibility Xm is positive:

o2N
M=X.H, X.= 3T (52)
3
but even with N 1030 atoms/M , it is still very small:
X-7 X 10-4 (53)
(c) Ferromagnetism
As for ferroelectrics (see Section 3-1-5), sufficiently high
coupling between adjacent magnetic dipoles in some iron
alloys causes them to spontaneously align even in the absence
of an applied magnetic field. Each of these microscopic
domains act like a permanent magnet, but they are randomly
distributed throughout the material so that the macroscopic
magnetization is zero. When a magnetic field is applied, the
dipoles tend to align with the field so that domains with a
magnetization along the field grow at the expense of non
aligned domains.
The friction-like behavior of domain wall motion is a lossy
process so that the magnetization varies with the magnetic
field in a nonlinear way, as described by the hysteresis loop in
Figure 5-20. A strong field aligns all the domains to satura
tion. Upon decreasing H, the magnetization lags behind so
that a remanent magnetization M, exists even with zero field.
In this condition we have a permanent magnet. To bring the
magnetization to zero requires a negative coercive field - H,.
Although nonlinear, the main engineering importance of
ferromagnetic materials is that the relative permeability s,. is
often in the thousands:
IL= IIO= B/H (54)
This value is often so high that in engineering applications we
idealize it to be infinity. In this limit
lim B=tyH=>H=0, B finite (55)
the H field becomes zero to keep the B field finite.

Magnetization 357
M
- l H
/H,
Figure 5-20 Ferromagnetic materials exhibit hysteresis where the magnetization

saturates at high field strengths and retains a net remanent magnetization M, even
when H is zero. A coercive field -H, is required to bring the magnetization back to
zero.
EXAMPLE 5-1 INFINITE LINE CURRENT WITHIN A

MAGNETICALLY PERMEABLE CYLINDER
A line current I of infinite extent is within a cylinder of

radius a that has permeability 1L, as in Figure 5-21. The
cylinder is surrounded by free space. What are the B, H, and
M fields everywhere? What is the magnetization current?
ti
BO
I-,
2rr
t Line current
( -1)1
2rr
Sr r
Surface current
K. =-(A -- 1)
Figure 5-21 A free line current of infinite extent placed within a permeable cylinder
gives rise to a line magnetization current along the axis and an oppositely directed
surface magnetization current on the cylinder surface.
SOLUTION
Pick a circular contour of radius r around the current.

Using the integral form of Ampere's law, (21), the H field is
of the same form whether inside or outside the cylinder:
H -d1= H,27rr= I=>H =-

L 27rr
The magnetic flux density differs in each region because the

permeability differs:
pgH=-, 0<r<a
2wrr
Bo H 4= 27rr'
A r> a
The magnetization is obtained from the relation

B
M=--_ H
go
as
I) H, - O<r<a
Mo = (Ao / o 21rr'
10, r>a
The volume magnetization current can be found using
(16):
J =VM = i, +-(rMs)i,=0, 0<r<a

az r ar
There is no bulk magnetization current because there are no
bulk free currents. However, there is a line magnetization
current at r =0 and a surface magnetization current at r = a.
They are easily found using the integral form of (16) from
Stokes' theorem:
JsVxM-dS= LM-dl= J.dS
Pick a cortour around the center of the cylinder with r <a:
M4s27rr = ( oI= IM
where I. is the magnetization line current. The result

remains unchanged for any radius r <a as no more current is
enclosed since J.=0 for 0<r<a., As .soon as r>a, Mo
becomes zero so that the total magnetization current becomes
BoundaryConditions 359
zero. Therefore, at r = a a surface magnetization current

must flow whose total current is equal in magnitude but
opposite in sign to the line magnetization current:
K.,. = --- = - (Z-u)

2ra so2ira
5-6 BOUNDARY CONDITIONS
At interfacial boundaries separating materials of differing

properties, the magnetic fields on either side of the boundary
must obey certain conditions. The procedure is to use the
integral form of the field laws for differential sized contours,
surfaces, and volumes in the same way as was performed for
electric fields in Section 3-3.
To summarize our development thus far, the field laws
for magnetic fields in differential and integral form are
VxH=Jf, fH-di= Jf-dS (1)
VxM=J,., fM-dl= J,.-dS (2)
V-B=0, JB-dS=0 (3)
5-6-1 Tangential Component of H
We apply Ampere's circuital law of (1) to the contour of

differential size enclosing the interface, as shown in Figure
5-22a. Because the interface is assumed to be infinitely thin,
the short sides labelled c and d are of zero length and so offer
B2
Free surface current K 1 n
H2 perpendicular to contour L Area dS
up out of the page.
d 2
'P L --H,) K,
H2 n - (81 - B2) = 0
2
C
H, (a) (b)
Figure 5-22 (a) The tangential component of H can be discontinuous in a free
surface current across a boundary. (b) The normal component of B is always continu
ous across an interface.
no contribution to the line integral. The remaining two sides

yield
fH -dl=(H,- H2,) dl = KA.dl (4)
where KA. is the component of free surface current perpen

dicular to the contour by the right-hand rule in this case up out
of the page. Thus, the tangential component of magnetic field
can be discontinuous by a free surface current,
(HI - H2,)= Kf.>n X(H2 - H)= Kf (5)
where the unit normal points from region 1 towards region 2.

If there is no surface current, the tangential component of H
is continuous.
5-6-2 Tangential Component of M
Equation (2) is of the same form as (6) so we may use the

results of (5) replacing H by M and Kf by K,,, the surface
magnetization current:
(Mi- M 2,)= K,., nX(M2 -Ml)=K. (6)
This boundary condition confirms the result for surface

magnetization current found in Example 5-1.
5-6-3 Normal Component of B
Figure 5-22b shows a small volume whose upper and lower

surfaces are parallel and are on either side of the interface.
The short cylindrical side, being of zero length, offers no
contribution to (3), which thus reduces to
B-dS= (B2 , 1-B 1 ) dS=0 (7)
yielding the boundary condition that the component of B

normal to an interface of discontinuity is always continuous:
B1. - B 2 .=0>n - (BI - B2 )= 0 (8)
EXAMPLE 5-2 MAGNETIC SLAB WITHIN A UNIFORM MAGNETIC

FIELD
A slab of infinite extent in the x and y directions is placed

within a uniform magnetic field Hoi, as shown in Figure 5-23.
Magnetic Field Boundary Value Problems 361
I
i
i,
t H oi 0 tHo
Mo Mo
Mai , H -- (Ho -Mo)
Mo Mo
Hoia Ho(i
(a) (b)
Figure 5-23 A (a) permanently magnetized or (b) linear magnetizable material is

placed within a uniform magnetic field.
Find the H field within the slab when it is

(a) permanently magnetized with magnetization Moi,
(b) a linear permeable material with permeability A.
SOLUTION
For both cases, (8) requires that the B field across the
boundaries be continuous as it is normally incident.
(a) For the permanently magnetized slab, this requires that
y 0 H 0 = po(H + Mo)>H = Ho-Mo

Note that when there is no externally applied field (Ho = 0),
the resulting field within the slab is oppositely directed to the
magnetization so that B = 0.
(b) For a linear permeable medium (8) requires
yxoHo = IH=>H = A0Ho
For p. >po the internal magnetic field is reduced. If H0 is set

to zero, the magnetic field within the slab is also zero.
5-7 MAGNETIC FIELD BOUNDARY VALUE PROBLEMS
5-7-1 The Method of Images
A line current I of infinite extent in the z direction is a

distance d above a plane that is either perfectly conducting or
infinitely permeable, as shown in Figure 5-24. For both cases
2
x + - d)
2
= Const
110 Y d x 2 +(y +d)2
2 2
[x 2 + (y -d) 1 [X + (y+ d = Const
0o (y d
A. t
JA* d
(b)
Figure 5-24 (a) A line current above a perfect conductor induces an oppositely
directed surface current that is equivalent to a symmetrically located image line
current. (b) The field due to a line current above an infinitely permeable medium is the
same as if the medium were replaced by an image current now in the same direction as
the original line current.
the H field within the material must be zero but the boundary
conditions at the interface are different. In the perfect
conductor both B and H must be zero, so that at the interface
the normal component of B and thus H must be continuous
and thus zero. The tangential component of H is dis
continuous in a surface current.
In the infinitely permeable material H is zero but B is finite.
No surface current can flow because the material is not a
conductor, so the tangential component of H is continuous
and thus zero. The B field must be normally incident.
Both sets of boundary conditions can be met by placing an
image current I at y = - d flowing in the opposite direction
for the conductor and in the same direction for the perme
able material.
Using the upper sign for the conductor and the lower sign
for the infinitely permeable material, the vector potential due
to both currents is found by superposing the vector potential
found in Section 5-4-3a, Eq. (18), for each infinitely long line
current:
-IO
A.= - {ln [x2 +(y -d) 2
1" 2 FIn [x2 +(y+d) 211 2 1
21r
=4 {ln [x 2+(y -d)2] F In [x +(y +d) 2 ]} (1)

1w
with resultant magnetic field
1
H =-IVxA=-- 1~.A.
I(X a
aAz)
Ao pAo ()y i,--A
8x
- II (y -d)i. -xi,, (y +d)i. -xA, (2

21r[x 2 +(y-d) 2] [X2+(y+d)2 (2)
The surface current distribution for the conducting case is
given by the discontinuity in tangential H,
Id
K. =-H.(y=)= [2 2 (3)
which has total current
I + Id + dx
I=K~dx=J(
2
1r L (x2 +d
)
-- tan - (4)
ir d d I-
just equal to the image current.
The force per unit length on the current for each case is
just due to the magnetic field from its image:
2
f *.OI (5)
47rd
being repulsive for the conductor and attractive for the

permeable material.
The magnetic field lines plotted in Figure 5-24 are just lines
of constant A, as derived in Section 5-4-3b. Right next to the
line current the self-field term dominates and the field lines
are circles. The far field in Figure 5-24b, when the line and
image current are in the same direction, is the same as if we
had a single line current of 21.
5-7-2 Sphere in a Uniform Magnetic Field
A sphere of radius R is placed within a uniform magnetic

field Hoi.. The sphere and surrounding medium may have
any of the following properties illustrated in Figure 5-25:
(i) Sphere has permeability /12 and surrounding medium

has permeability pr.
(ii) Perfectly conducting sphere in free space.
(iii) Uniformly magnetized sphere M 2i, in a uniformly
magnetized medium Mli..
For each of these three cases, there are no free currents in

either region so that the governing equations in each region
are
V-B=0
VxH=O (5)
2 2
(_) sin o = Const
[.f
r +_2 R
(a) Hoi = Ho(i, cosO - i0 sin6)
Figure 5-25 Magnetic field lines about an (a) infinitely permeable and (b) perfectly
conducting sphere in a uniform magnetic field.
MagneticField Boundary Value Problems 365
2
[- r + (1)2]
R sin 0 = Const
I)
Ho i , =Ho(icos - i s inel y
(b)
Figure 5-25
Because the curl of H is zero, we can define a scalar magnetic

potential
H =VX (6)
where we avoid the use of a negative sign as is used with the

electric field since the potential x is only introduced as a
mathematical convenience and has no physical significance.
With B proportional to H or for uniform magnetization, the
divergence of H is also zero so that the scalar magnetic
potential obeys Laplace's equation in each region:
v2x =0 (7)
We can then use the same techniques developed for the

electric field in Section 4-4 by trying a scalar potential in each
region as
{ Ar cos 0,
2
r<R
(8)
(Dr+C/r ) cos 0 r>R
366 me Magetic FieLd
The associated magnetic field is then
H=V XIi,+ -i ,+ 1 IX4

Or r 80 r sin084
r<R
{A(i,cos 0-io sin
0)= Ai,
(D-2C/r3)cosTi,-(D+ C/r)sin Oig, r>R (9)
For the three cases, the magnetic field far from the sphere
must approach the uniform applied field:
H(r=c0)=Hoi.=Ho(i,.cose-io sin 0)>D=Ho (10)
The other constants, A and C, are found from the boundary
conditions at r = R. The field within the sphere is uniform, in
the same direction as the applied field. The solution outside
the sphere is the imposed field plus a contribution as if there
were a magnetic dipole at the center of the sphere with
moment m, - 4rC.
(i) If the sphere has a different permeability from the sur

rounding region, both the tangential components of H and
the normal components of B are continuous across the
spherical surface:
3
He(r=R)= H,(r= R_)' A = D +C/R
B,(r= R+) =B,(r = R-)=iH,(r = R,) = y 2 H,(r =R-)
which yields solutions
A = pI___, C - A2-A
2
R 5Ho (12)
p2+2pi 2+ MI
The magnetic field distribution is then
3pi1Ho .3. SpHoi.

2 (i, Cos 0 - , sin )= , r <R
92+ 1L p2+2pJ
I]C i, (13) H=- Ho 1+2 (.2- s

1 r pI2+2AI J
-/ NJ sin }, r>R
r3 p2+ 21A
)
The magnetic field lines are plotted in Figure 5-25a when

IA2-*O. In this limit, H within the sphere is zero, so that the
field lines incident on the sphere are purely radial. The field
lines plotted are just lines of constant stream function 1,
found in the same way as for the analogous electric field
problem in Section 4-4-3b.
.(ii) If the sphere is perfectly conducting, the internal

magnetic field is zero so that A = 0. The normal component
of B right outside the sphere is then also zero:
3
H,(r = R) = 0> C = HOR /2 (14)
yielding the solution
H=Ho 1 3 cosir-
R)2r
r
I+3) sin ioj, r>R
(15)
The interfacial surface current at r = R is obtained from the
discontinuity in the tangential component of H:
K5= He(r=R)=-2H sin 6 (16)
The current flows in the .negative < direction around the
sphere. The right-hand rule, illustrated in Figure 5-25b,
shows that the resulting field from the induced current acts in
the direction opposite to the imposed field. This opposition
results in the zero magnetic field inside the sphere.
The field lines plotted in Figure 5-25b are purely tangential
to the perfectly conducting sphere as required by (14).
(iii) If both regions are uniformly magnetized, the bound
ary conditions are
Ho(r = R,)= Ho(r=R_)4A = D+C/R 3
B,(r = R ) = B,(r R)) H,(r = R+) + M 1 cos 0
=H,(r=R_)+M2 cos9 (17)
with solutions
A = H,+A(Mi - M 2
(18)
)
R3
C=- (MI - M2
)
so that the magnetic field is
[Ho+ - (M1 - M2 )][cos Oi, - sin io]

3
=[H0 +-(M 1-M 2 )]i. r<R

3
2R
3 (19)
H=
(H0 - 3r 3 3 (M1 -M 2 )lcos~i,
3
R
r3HOMl +-M2))sin Oi 0 r>R
,
Because the magnetization is uniform in each region, the

curl of M is zero everywhere but at the surface of the sphere,
so that the volume magnetization current is zero with a sur

face magnetization current at r = R given by
Km = n x (Mi - M 2
)
= i, x (MI - M 2)i.
= i, x (MI - M 2 )(i, cos 0 - sin Oio)
= - (MI - M 2 ) sin Oik (20)
5-8 MAGNETIC FIELDS AND FORCES
5-8-1 Magnetizable Media
A magnetizable medium carrying a free current J1 is placed

within a magnetic field B, which is a function of position. In
addition to the Lorentz force, the medium feels the forces on
all its magnetic dipoles. Focus attention on the rectangular
magnetic dipole shown in Figure 5-26. The force on each
current carrying leg is
f = i dl x (Bi + Byi, + Bi
)
> f(x)= -i Ay[-Bji + Bri]

f(x + Ax) =i Ay[ - Bi, + Bix]j \.
f(y) = i Ax[Byi, -Bi,]j
,
f(y +Ay) = - i Ax[Byi, - Bzi,]l ,A, (1)

so that the total force on the dipole is
f = f(x)+f(x+Ax)+f(y)+f(y +Ay)
.
, Ax AY 1
B,(x +Ax)-B,(x ) .Bx(x+ Ax )-B (x ).
Ax Ax
B.(y+Ay)-Bz(y) . B,(y+Ay)--B,(y).
+ Ay 1i (2)
/t
Sx
Figure 5-26 A magnetic dipole in a magnetic field B.

Magnetic Fields and Forces 369
In the limit of infinitesimal Ax and Ay the bracketed terms

define partial derivatives while the coefficient is just the
magnetic dipole moment m = i Ax Ay i :
lim f= M.--LB.-_8.a
+- ,+---,, (3)
Ay-0 ax ax ay ay
Ampere's and Gauss's law for the magnetic field relate the
field components as
V - B =0 =- ( + (4)
az \Ox ay
VxB=pto(Jf+VxM)= OJT -- = AOJT.

ay az
aB. aB.
-----
Oz Ox
= AoJT,
- = AoJ. (5)
Ox Oy
which puts (3) in the form
f m, -- I
.Oz z i I I
= + - O(JTi, -JT.i,)
z
=(m - V)B+pomX JT (6)
where JT is the sum of free and magnetization currents.
If there are N such dipoles per unit volume, the force
density on the dipoles and on the free current is
F=Nf= (M- V)B+iLoMXJT+J!XB

= lo(M - V)(H+M)+AoM x (Jf+V X M)+poJJ X (H+M)
= po(M - V)(H+M)+ oM x (V x M) +IJ xH (7)
Using the vector identity
M x (V x M)= -(M - V)M+1V(M - M) (8)
(7) can be reduced to
F= Lo(M - V)H +lojf XH +V (M -M) (9)
The total force on the body is just the volume integral of F:
f = fv F dV (10)
In particular, the last contribution in (9) can be converted

to a surface integral using the gradient theorem, a corollary
to the divergence theorem (see Problem 1-15a):
V( "M - M dV=f M - MdS (11)
Since this surface S surrounds the magnetizable medium, it

is in a region where M = 0 so that the integrals in (11) are
zero. For this reason the force density of (9) is written as
F= io(M - V)H + oJfX H (12)
It is the first term on the right-hand side in (12) that accounts

for an iron object to be drawn towards a magnet. Magnetiz
able materials are attracted towards regions of higher H.
5-8-2 Force on a Current Loop
(a) Lorentz Force Only

Two parallel wires are connected together by a wire that is
free to move, as shown in Figure 5-27a. A current I is
imposed and the whole loop is placed in a uniform magnetic
field Boi.. The Lorentz force on the moveable wire is
f, = IBol (13)
where we neglect the magnetic field generated by the current,

assuming it to be much smaller than the imposed field B 0
.
(b) Magnetization Force Only

The sliding wire is now surrounded by an infinitely
permeable hollow cylinder of iliner radius a and outer radius
b, both being small compared to the wire's length 1, as in
Figure 5-27b. For distances near the cylinder, the solution is
approximately the same as if the wire were infinitely long. For
r>0 there is no current, thus the magnetic field is curl and
divergence free within each medium so that the magnetic
scalar potential obeys Laplace's equation as in Section 5-7-2.
In cylindrical geometry we use the results of Section 4-3 and
try a scalar potential of the form
x=(Ar+ )cos (14)

x
B
f
'Y
(a)
B
oo
y f = IBoliy t I Pa
(b) BOix = Bo(i, coso io sin
p )
hi,
B IA
Srb2
rbI
(c)
Figure 5-27 (a) The Lorentz-force on a current carrying wire in a magnetic field. (b)
If the current-carrying wire is surrounded by an infinitely permeable hollow cylinder,
there is no Lorentz force as the imposed magnetic field is zero where the current is.
However, the magnetization force on the cylinder is the same as in (a). (c) The total
force on a current-carrying magnetically permeable wire is also unchanged.
in each region, where B = VX because V x B = 0. The

constants are evaluated by requiring that the magnetic field
approach the imposed field Boi. at r = 0 and be normally
incident onto the infinitely permeable cylinder at r =a and
r = b. In addition, we must add the magnetic field generated
by the line current. The magnetic field in each region is then
The Magnetic Field
(see Problem 32a):
sII 4, O<r<a
2Bob2 as222IL
b 2-a 2 [ r rr +- sin
Is-
n2arr
0i,
B=
a<r<b (15)
B 21+yCos # I sin di, +.E- i',

R r r 27rr
r>b
Note the infinite flux density in the iron (A - oo) due to the
line current that sets up the finite H field. However, we see
that none of the imposed magnetic field is incident upon the
current carrying wire because it is shielded by the infinitely
permeable cylindrical shell so that the Lorentz force contri
bution on the wire is zero. There is, however, a magnetization
force on the cylindrical shell where the internal magnetic field
H is entirely due to the line current, H, = I/27rr because with
i - oo, the contribution due to BO is negligibly small:
F = o(M - V)H
( aM,
=
(3r
(A) (16)
r aw
Within the infinitely permeable shell the magnetization and
H fields are
H#21rr
AoMr=Br-oflrH=b2- a ) cos4 (17)
2Bob2 / \ ( - o)I
M~oM = Bo - &oH= 2- 1+- sin 4+
(b 2 - r/ 2irr
Although Hs only depends on r, the unit vector i, depends on
i,=(-sini.+cosi,) (18)
so that the force density of (16) becomes
F= - Bj I6+ (B.0 - ~oH#)I d00

22? 21rr2 d4
= [- B(- sin Oi.+cos Oi,)
+ (B,* - p~oH#)(- cos 46i. - sin Oi,)]

I 2Bobe2t 2
irr=2 21r2
-bs -as r
-L2r) cos 0(-sin Oi. +cos Oi,)
--(1 +- sin O(cos Oi. + sin i,)
+ (/ - AO) (cos ) ijI +sin Oil)
21rrI
I (2Bob 2 2a 2
=Fr2L b-2 -2 sin 0 cos 0 i. r 2 i')
+ O -I 1 (cos
i. +sin Oi,)] (19)
27rr
The total force on the cylinder is obtained by integrating

(19) over r and 4:
b
J
2w
f= Flrdrdo (20)
-=0 r=a
All the trigonometric terms in (19) integrate to zero over 4 so

that the total force is
2Bob 2Il ar2
A, 2 2 -3 dr
(b -a ) .. ar
Bob 2I1 a22

(b -a) r 2a
=IB01 (21)
The force on the cylinder is the same as that of an unshield

ed current-carrying wire given by (13). If the iron core has a
finite permeability, the total force on the wire (Lorentz force)
and on the cylinder (magnetization force) is again equal to
(13). This fact is used in rotating machinery where current-
carrying wires are placed in slots surrounded by highly
permeable iron material. Most of the force on the whole
assembly is on the iron and not on the wire so that very little
restraining force is necessary to hold the wire in place. The
force on a current-carrying wire surrounded by iron is often
calculated using only the Lorentz force, neglecting the
presence of the iron. The correct answer is obtained but for
the wrong reasons. Actually there is very little B field near the
wire as it is almost surrounded by the high permeability iron
so that the Lorentz force on the wire is very small. The force
is actually on the iron core.
(c) Lorentz and Magnetization Forces

If the wire itself is highly permeable with a uniformly
distributed current, as in Figure 5-27c, the magnetic field is
(see Problem 32a)
2B 0 (, OrCOS Ir
4 - is sin 0) +Ir2 4
A +jo l2rb
2B0 I
= i.+ 2(-yi.+xi,), r<b
H= /+ IL rb (22)
B0 + bo2i
A0 R r2 + OS
- A -1- sin Oi,6 + I i, r> b

r2 A +A0o) 2r
It is convenient to write the fields within the cylinder in

Cartesian coordinates using (18) as then the force density
given by (12) is
F = pO(M -V)H +Aojf X H
= (A - A0)(H - V)H + iL x H
= (A - IAO) (H. ax+ H, T~y)(H.i. + Hi,) + ;b2(H.i, - Hi.)

(23)
Since within the cylinder (r < b) the partial derivatives of H
are
,3H. aH,
ax ay(24)
aH. aH, I
ay - ax -ib
(23) reduces to
H +A
F=(IA-o)(H. i,+H,
lxay rbT (H.,-Hi.)
=2 (IA+ho)(H.i,-Hi.)
I(p+po) 2B, -y \. Ix 1
2'rb 2 +po
/A+A0 (25)
Realizing from Table 1-2 that

yi, +xi = r[sin Oi, +cos 4i.] = r26) (26)
Problems 375
the force density can be written as
F=IB, I 2(st+y0)
2 r (27)
Frb
rb 2 - 2
(2 Irb (sin ki, +cos 0i )
)
The total force on the permeable wire is
2,r b
f= Flr dr do (28)
We see that the trigonometric terms in (27) integrate to zero

so that only the first term contributes:
IB 01 2, b
f,= 2 rdrdo
=IB1 (29)
The total force on the wire is independent of its magnetic

permeability.
PROBLEMS
Section 5-1
1. A charge q of mass m moves through a uniform magnetic
field Boi,. At t =0 its velocity and displacement are
v(t = 0) = vooi + VYoi, + v~o0z
r(t = 0) = xoi. + yoi, + zoi
(a) What is the subsequent velocity and displacement?
(b) Show that its motion projected onto the xy plane is a
circle. What is the radius of this circle and where is its center?
(c) What is the time dependence of the kinetic energy of
the charge 2mlvI 2?
2. A magnetron is essentially a parallel plate capacitor

stressed by constant voltage Vo where electrons of charge -e
are emitted at x = 0, y = 0 with zero initial velocity. A trans
verse magnetic field Boi, is applied. Neglect the electric and
magnetic fields due to the electrons in comparison to the
applied field.
(a) What is the velocity and displacement of an electron,
injected with zero initial velocity at t = 0?
(b) What value of magnetic field will just prevent the elec
trons from reaching the other electrode? This is the cut-off
magnetic field.
V0
Y -111
VO
y a
Sx
b
GBoi
-- S-

(a) (c)
(c) A magnetron is built with coaxial electrodes where

electrons are injected from r = a, 4 =0 with zero initial veloc
ity. Using the relations from Table 1-2,
i, = cos 4i. + sin Oi,

i,= -sin 4i. +cos Oi,
show that
di, . d46 v4
.
dt di r
di, . d4 v,.
dt dt r
What is the acceleration of a charge with velocity
V Vri, +Vqi?
(d) Find the velocity of the electrons as a function of radial

position.
Hint:
dv, dVrdr dVr d 1 2
dt dr dt Vr dr dr
dv, dvodr dvo
dt dr dt V dr
(e) What is the cutoff magnetic field? Check your answer

with (b) in the limit b = a +s where s < a.
3. A charge q of mass m within a gravity field -gi, has an
initial velocity voi.. A magnetic field Boi is applied. What
Problems 377
i
q V
Bo i,
value of BO will keep the particle moving at constant speed in

mg the x direction?
4. The charge to mass ratio of an electron e/m was first
measured by Sir J. J. Thomson in 1897 by the cathode-ray
tube device shown. Electrons emitted by the cathode pass
through a slit in the anode into a region with crossed electric
and magnetic fields, both being perpendicular to the elec
trons velocity. The end of the tube is coated with a fluorescent
material that produces a bright spot where the electron beam
impacts.
V1
V2
Screen
(a) What is the velocity of the electrons when passing

through the slit if their initial cathode velocity is vo?
(b) The electric field E and magnetic field B are adjusted so
that the vertical deflection of the beam is zero. What is the
initial electron velocity? (Neglect gravity.)
(c) The voltage V2 is now set to zero. What is the radius R
of the electrons motion about the magnetic field?
(d) What is e/m in terms of E, B, and R?
5. A charge q of mass m at t = 0 crosses the origin with
velocity vo = v.oi. + v,oi,. For each of the following applied
magnetic fields, where and when does the charge again cross
Y the y =0 plane?
(a) B= Boi.
(b) B= Boi,
(c) B=Boi.
So = voti, cos6 + iv sin6]
x
6. In 1896 Zeeman observed that an atom in a magnetic field

had a fine splitting of its spectral lines. A classical theory of
the Zeeman effect, developed by Lorentz, modeled the elec
tron with mass m as being bound to the nucleus by a spring-
like force with spring constant k so that in the absence of a
magnetic field its natural frequency was wk =
.
(a) A magnetic field Boi, is applied. Write Newton's law for
the x, y, and z displacements of the electron including the
spring and Lorentz forces.
(b) Because these equations are linear, guess exponential
solutions of the form e". What are the natural frequencies?
(c) Because wh is typically in the optical range (wh

10 5 radian/sec), show that the frequency splitting is small
compared to wk even for a strong field of B 0 = 1 tesla. In this
limit, find approximate expressions for the natural frequen
cies of (b).
7. A charge q moves through a region where there is an

electric field E and magnetic field B. The medium is very
viscous so that inertial effects are negligible,
Pv=q(E+vxB)
where 6 is the viscous drag coefficient. What is the velocity of

the charge? (Hint: (vxB)XB= -v(B-B)+B(v-B) and
v - B= (q/#)E - B.)
8. Charges of mass m, charge q, and number density n move

through a conducting material and collide with the host
medium with a collision frequency P in the presence of an
electric field E and magnetic field B.
(a) Write Newton's first law for the charge carriers, along
the same lines as developed in Section 3-2-2, with the addition
of the Lorentz force.
(b) Neglecting particle inertia and diffusion, solve for the
particle velocity v.
(c) What is the constitutive law relating the current density
J = qnv to E and B. This is the generalized Ohm's law in the
presence of a magnetic field.
(d) What is the Ohmic conductivity or? A current i is passed
through this material in the presence of a perpendicular
magnetic field. A resistor RL is connected across the
terminals. What is the Hall voltage? (See top of page 379).
(e) What value of RL maximizes the power dissipated in the
load?
Problems 379
- RL Vh
IdBoiz
Section 5.2
9. A point charge q is traveling within the magnetic field of
an infinitely long line current I. At r = ro its velocity is
v(t = 0) = Vri, + Vb 0i, + Vzoi
Its subsequent velocity is only a function of r.

(a) What is the velocity of the charge as a function of
position? Hint: See Problem 2c and 2d,
dx = (lnx)2
(b) What is the kinetic energy of the charge?

(c) What is the closest distance that the charge can
approach the line current if v6 0 = 0?
10. Find the magnetic field at the point P shown for the
following line currents:
n.sided
a Pregula regular
equilateral
polygon
(a) (b) (c)
(d) (e) (f)
11. Two long parallel line currents of mass per unit length
m in a gravity field g each carry a current I in opposite
The Magnetic Field
directions. They are suspended by cords of length L What is

the angle 6 between the cords?
12. A constant current KOi6 flows on the surface of a sphere

of radius R.
Koi#
(a) What is the magnetic field at the center of the sphere?

(HINT: i, x i, = ie= cos 0 cos Oi. +cos 0 sin Oi, -sin Oi,.)
(b) Use the results of (a) to find the magnetic field at the
center of a spherical shell of inner radius R1 and outer radius
R 2 carrying a uniformly distributed volume current jois.
13. A line current I of length 2L flows along the z axis.
-Y K = Koi
2Lx
(a) Y (a)y(b)
Problems 381
(a) What is the magnetic field everywhere in the z =0

plane?
(b) Use the results of (a) to find the magnetic field in the
z = 0 plane due to an infinitely long current sheet of height 2L
and uniform current density Koi. Hint: Let u= x2+y2
du 1 . _ bu+2a
fu(u2+bu-a)12 [ ubsi+4a
14. Closely spaced wires are wound about an infinitely long
cylindrical core at pitch angle 0o. A current flowing in the
wires then approximates a surface current
K = KO(i. sin 8 0+i, cos 00
)
K = Ko(ij sin 0 + io cosOO)
00
- .2a
What is the magnetic field everywhere?
15. An infinite slab carries a uniform current Joi, except

within a cylindrical hole of radius a centered within the slab.
Y
SJO 12
(a) (b)
(a) Find the magnetic field everywhere? (Hint: Use

superposition replacing the hole by two oppositely directed
currents.)
(b) An infinitely long cylinder of radius a carrying a uni
form current Joi, has an off-axis hole of radius b with center a
distance d from the center of the cylinder. What is the
magnetic field within the hole? (Hint: Convert to Cartesian
coordinates ri = xi-yix.)
Section 5.3
16. Which of the following vectors can be a magnetic field B?
If so, what is the current density J?
(a) B=ari,
(b) B= a(xi,--yi.)
(c) B=a(xi.-yi,)
(d) B= ario
17. Find the magnetic field everywhere for each of the
following current distributions:
..............
(a) (c)
(a) J joi, -a<y<O
(b) j y i., -a<y<a

a
joi , O<r<a
(c) j= -joi,,, a<r <b
for.i, ~
(d) j= a
0, r>a
Problems 383
Section 5.4
18. Two parallel semi-infinite current sheets a distance d
apart have their currents flowing in opposite directions and
extend over the interval -00 < x <0.
y
2d -Koi, x
- K0 i
(a) What is the vector potential? (Hint: Use superposition
of the results in Section 5-3-4b.)
(b) What is the magnetic field everywhere?
(c) How much magnetic flux per unit length emanates
through the open face at x = 0? How much magnetic flux per
unit length passes through each current sheet?
(d) A magnetic field line emanates at the position yo(O <
yo < d) in the x = 0 plane. At what value of y is this field line at
x =-00?
19. (a) Show that V - A 0 for the finite length line current
in Section 5-4-3a. Why is this so?
In 3 y
(b) Find the vector potential for a square loop.

(c) What is V - A now?
20. Find the magnetic vector potential and ma netic field for
the following current distributions: (Hint: VfA = V(V - A)

V x (V x A))
(i) Infinitely long cylinder of radius a carrying a
(a) surface current KOi5
(b) surface current, Koi,
(c) volume current Joi,
(a) (d)
(ii) Infinitely long slab of thickness d carrying a

(d) volume cur'ent Joi
lox.
(e) volume current x
Section 5.5
21. A general definition for the magnetic dipole moment for
any shaped current loop is
m=- rxI dl
2
If the current is distributed over a surface or volume or is due
to a moving point charge we use
Idi -qv-*KdS-+JdV
What is the magnetic dipole moment for the following cur

rent distributions:
(a) a point charge q rotated at constant angular speed w at
radius a;
(b) a circular current loop of radius a carrying a current I;
(c) a disk of radius a with surface current Koi4;
(d) a uniformly distributed sphere of surface or volume
charge with total charge Q and radius R rotating in the 4
direction at constant angular speed w. (Hint: i, X i# = -i =
-[cos 0 cos #i.+cos 6 sin Oi, - sin Gi])
22. Two identical point magnetic dipoles m with magnetic

polarizability a (m =aH) are a distance a apart along the z
axis. A macroscopic field Hoi, is applied.
(a) What is the local magnetic field acting on each dipole?
(b) What is the force on each dipole?
Problems 385
m aH
HO i,
(c) Repeat (a) and (b) if we have an infinite array of such

dipoles. Hint:
- 1.2
n=1 n
(d) If we assume that there is one such dipole within each

volume of a , what is the permeability of the medium?
23. An orbiting electron with magnetic moment mai, is in a
uniform magnetic field Boi. when at t = 0 it is slightly dis
placed so that its angular momentum L = -( 2 m,/e)m now also
has x and y components.
(a) Show that the torque equation can be put in terms of
the magnetic moment
dm
=t-ymxB
where y is called the gyromagnetic ratio. What is y?

(b) Write out the three components of (a) and solve for the
magnetic moment if at t =0 the moment is initially
m(t = 0)= m Oi + mOi, + mOi
(c) Show that the magnetic moment precesses about the

applied magneticfield. What is the precessional frequency?
24. What are the B, H, and M fields and the resulting
magnetization currents for the following cases:
(a) A uniformly distributed volume current Joi, through a
cylinder of radius a and permeability A surrounded by
free space.
(b) A current sheet Koi, centered within a permeable slab
of thickness d surrounded by free space.
386 Th' Magnetic Field
Joiz
d :
(a) (b)
Ij 0MM.gIgg
I IMER
Section 5.6
25. A magnetic field with magnitude H, is incident upon the
flat interface separating two different linearly permeable
materials at an angle 01 from the normal. There is no surface
N2
H,
current on the interface. What is the magnitude and angle of

the magnetic field in region 2?
26. A cylinder of radius a and length L is permanently
magnetized as Moi.
(a) What are the B and H fields everywhere along its axis?
(b) What are the fields far from the magnet (r w> a, r >>L)?
(c) Use the results of (a) to find the B and H fields every
where due to a permanently magnetized slab Moi, of infinite
xy extent and thickness L.
(d) Repeat (a) and (b) if the. cylinder has magnetization
MO(I -r/a)i,. Hint:
=) (rr
Problems 387
2
M0 1
L ... :y
Section 5.7
27. A z-directed line current I is a distance d above the
interface separating two different magnetic materials with
permeabilities 11 and 142
...........
(a) Find the image currents I' at position x =-d and I" at
x=d that satisfy all the boundary conditions. The field in
region 1 is due to I and I' while the field in region 2 is due to
I". (Hint: See the analogous dielectric problem in Section
3-3-3.)
(b) What is the force per unit length on the line current I?
28. An infinitely long line current I is parallel to and a
distance D from the axis of a perfectly conducting cylinder of
radius a carrying a total surface current 1o.
(a) Find suitable image currents and verify that the bound
ary conditions are satisfied. (Hint: xi,-vi.=ri#; i,=
sin Oir+cos 414; x = r cos 4.)
388 Te Magnetic Field
KO 0 2va
@1 x
(a)
t Ir
(d)
(b) What is the surface current distribution on the cylin

der? What total current flows on the cylinder? Hint:
f d4 2 _an([a 2 - b 2]"/ 2 tan (t4)

2
Ja + b cos 4 [a - b" \ (a+ b)
(c) What is the force per unit length on the cylinder?

(d) A perfectly conducting cylinder of radius a carrying a
total current I has its center a distance d above a perfectly
conducting plane. What image currents satisfy the boundary
conditions?
(e) What is the force per unit length on the cylinder?
29. A current sheet KO cos ayi is placed at x =0. Because
there are no volume currents for x #0, a scalar magnetic
potential can be defined H = VX.
Problems 389
KO cosayi, (a) What is the general form of solution for x? (Hint: See
Section 4-2-3.)
(b) What boundary conditions must be satisfied?
(c) What is the magnetic field and vector potential every
-Z-
where?
(d) What is the equation of the magnetic field lines?
30. A slab of thickness d carries a volume current distribution

Jo sin axi, and is placed upon a perfectly conducting ground
plane.
MO
d ".4.
(a) Find a particular solution for the vector potential. Are

all the boundary conditions satisfied?
(b) Show that additional solutions to Laplace's equations
can be added to the vector potential to satisfy the boundary
conditions. What is the magnetic field everywhere?
(c) What is the surface current distribution on the ground
plane?
(d) What is the force per unit length on a section of ground
plane of width 21r/a? What is the body force per unit length
on a section of the current carrying slab of width 2ir/a?
(e) What is the magnetic field if the slab carries no current
but is permanently magnetized as MO sin axiy Repeat (c) and
(d).
31. A line current of length L stands perpendicularly upon a
perfectly conducting ground plane.
It ..IZ
L
(a) Find a suitable image current that is equivalent to the

induced current on the z =0 plane. Does the direction of the
image current surprise you?
(b) What is the magnetic field everywhere? (Hint: See
Section 5-4-3a.)
(c) What is the surface current distribution on the
conducting plane?
32. A cylinder of radius a is placed within a uniform
magnetic field Hoi.. Find the magnetic field for each of the
following cases:
.I
(a)
(a) Cylinder has permeability s42 and surrounding medium

has permeability jp
.
(b) Perfectly conducting cylinder in free space.

(c) Uniformly magnetized cylinder M2 i, in a uniformly
magnetized medium Mji..
33. A current sheet Koi, is placed along the y axis. at x =0
between two parallel perfectly conducting planes a distance d
apart.
d
Kois
........
(a) Write the constant current at x =0 as an infinite Fourier
series of fundamental period 2d. (Hint: See Section 4-2-5.)
(b) What general form of a scalar potential x, where H=
Vx, will satisfy the boundary conditions?
(c) What is the magnetic field everywhere?
Problems 391
(d) What is the surface current distribution and the total

current on the conducting planes? Hint:
1 r2
n=1 n 8
(n odd)
Section 5.8
34. An infinitely long cylinder of radius a is permanently mag
netized as Mi.
Moxa
0 Y
(a) Find the magnetic field everywhere.

(b) An infinitely long line current I is placed either at
y = -b or at x = b (b > a). For each of these cases, what is
the force per unit length on the line current? (Hint: See
problem 32c.)
35. Parallel plate electrodes are separated by a rectangular
conducting slab that has a permeability A. The system is
driven by a dc current source.
Depth D
/0
0 :x
d
(a) Neglecting fringing field effects assume the magnetic

field is H,(x)i,. If the current is uniformly distributed
throughout the slab, find the magnetic field everywhere.
(b) What is the total force on the slab? Does the force
change with different slab permeability? Why not?
36. A permeable slab is partially inserted into the air gap

of a magnetic circuit with uniform field Ho. There is a
nonuniform fringing field right outside the magnetic circuit
near the edges.
Depth D y
(a) What is the total force on the slab in the x direction?

(b) Repeat (a) if the slab is permanently magnetized M =
Moi,. (Hint: What is H,(x = -no)? See Example 5-2a.)
chapter 6
electromagnetic
induction
394 ELeciromagneticInduction
In our development thus far, we have found the electric

and magnetic fields to be uncoupled. A net charge generates
an electric field while a current is the source of a magnetic
field. In 1831 Michael Faraday experimentally discovered
that a time varying magnetic flux through a conducting loop
also generated a voltage and thus an electric field, proving
that electric and magnetic fields are coupled.
6-1 FARADAY'S LAW OF INDUCTION
6-1-1 The Electromotive Force (EMF)
Faraday's original experiments consisted of a conducting

loop through which he could impose a dc current via a switch.
Another short circuited loop with no source attached was
nearby, as shown in Figure 6-1. When a dc current flowed in
loop 1, no current flowed in loop 2. However, when the
voltage was first applied to loop 1 by closing the switch, a
transient current flowed in the opposite direction in loop 2.
~ji1 (t)
+ A Ammeter i2(t) Dies off because of Ohmic

V losses with time constant
" = L/R
Positive current is induced

to try to keep magnetic flux
equal to a non-zero constant
Negative current is induced
to try to keep magnetic flux
equal to zero
Figure 6-1 Faraday's experiments showed that a time varying magnetic flux through
a closed conducting loop induced a current in the direction so as to keep the flux
through the loop constant.
Faraday'sLaw of Induction 395
When the switch was later opened, another transient current

flowed in loop 2, this time in the same direction as the original
current in loop 1. Currents are induced in loop 2 whenever a
time varying magnetic flux due to loop I passes through it.
In general, a time varying magnetic flux can pass through a
circuit due to its own or nearby time varying current or by the
motion of the circuit through a magnetic field. For any loop,
as in Figure 6-2, Faraday's law is
dlb d
EMF= E-dl=---=_ -- B-dS (1)
fL dt dt
where EMF is the electromotive force defined as the line
integral of the electric field. The minus sign is introduced on
the right-hand side of (1) as we take the convention that
positive flux flows in the direction perpendicular to the direc
tion of the contour by the right-hand rule.
6-1-2 Lenz's Law
The direction of induced currents is always such as to

oppose any changes in the magnetic flux already present.
Thus in Faraday's experiment, illustrated in Figure 6-1, when
the switch in loop 1 is first closed there is no magnetic flux in
loop 2 so that the induced current flows in the opposite
direction with its self-magnetic field opposite to the imposed
field. The induced current tries to keep a zero flux through
- dS
'b=fB
s
ndS dS
f E - di = -fB - dS
L
Figure 6-2 Faraday's law states that the line integral of the electric field around a
closed loop equals the time rate of change of magnetic flux through the loop. The
positive convention for flux is determined by the right-hand rule of curling the fingers
on the right hand in the direction of traversal around the loop. The thumb then points
in the direction of positive magnetic flux.
396 Eectromagnetic Induction
loop 2. If the loop is perfectly conducting, the induced cur

rent flows as long as current flows in loop 1, with zero net flux
through the loop. However, in a real loop, resistive losses
cause the current to exponentially decay with an LIR time
constant, where L is the self-inductance of the loop and R is
its resistance. Thus, in the dc steady state the induced current
has decayed to zero so that a constant magnetic flux passes
through loop 2 due to the current in loop 1.
When the switch is later opened so that the current in loop
1 goes to zero, the second loop tries to maintain the constant
flux already present by inducing a current flow in the same
direction as the original current in loop 1. Ohmic losses again
make this induced current die off with time.
If a circuit or any part of a circuit is made to move through
a magnetic field, currents will be induced in the direction
such as to try to keep the magnetic flux through the loop
constant. The force on the moving current will always be
opposite to the direction of motion.
Lenz's law is clearly demonstrated by the experiments
shown in Figure 6-3. When a conducting ax is moved into a
magnetic field, eddy currents are induced in the direction
where their self-flux is opposite to the applied magnetic field.
The Lorentz force is then in the direction opposite to the
motion of the ax. This force decreases with time as the cur
rents decay with time due to Ohmic dissipation. If the ax was
slotted, effectively creating a very high resistance to the eddy
currents, the reaction force becomes very small as the
induced current is small.
f, = 2rR IB,
t = JJ x BdV
V
R
Opposition magnetic
field due to
induced current Insulating i(t)
)
N B
VO
V
Conducting
(a) (b)
Figure 6-3 Lenz's law. (a) Currents induced in a conductor moving into a magnetic
field exert a force opposite to the motion. The induced currents can be made small by
slotting the ax. (b) A conducting ring on top of a coll is flipped off when a current is
suddenly applied, as the induced currents try to keep a zero flux through the ring.
When the current is first turned on in the coil in Figure 6-3b,

the conducting ring that sits on top has zero flux through it.
Lenz's law requires that a current be induced opposite to that
in the coil. Instantaneously there is no z component of
magnetic field through the ring so the flux must return radi
ally. This creates an upwards force:
f = 27RI X B= 2rRI4Bri. (2)

which flips the ring off the coil. If the ring is cut radially so
that no circulating current can flow, the force is zero and the
ring does not move.
(a) Short Circuited Loop

To be quantitative, consider the infinitely long time varying
line current I(t) in Figure 6-4, a distance r from a rectangular
loop of wire with Ohmic conductivity o-, cross-sectional area
A, and total length I = 2(D+ d). The magnetic flux through
the loop due to I(t) is
rD2 r+d
cDb,. = LoH,(r') dr' dz
z--D/2 r
jtoID r+dr' joID r+d (3)

2
1r r r' 2v r
H,(r')= 21rr'
,
cross sectional area A

Ohmic conductivity a
Par
pr
-Ed-
Figure 6-4 A rectangular loop near a time varying line current. When the terminals
are short circuited the electromotive force induces a current due to the time varying
mutual flux and/or because of the motion of the circuit through the imposed nonuni
form magnetic field of the line current. If the loop terminals are open circuited there is
no induced current but a voltage develops.
398 Eleciromagnetic Induction
The mutual inductance M is defined as the flux to current

ratio where the flux through the loop is due to an external
current. Then (3) becomes
D,. = M(r)I, M(r) = In r+d (4)

21r r
When the loop is short circuited (v = 0), the induced Ohmic
current i gives rise to an electric field [E = J/o = i/(Ao)] so that
Faraday's law applied to a contour within the wire yields an
electromotive force just equal to the Ohmic voltage drop:
il d(D
E - dI=-= iR = --- (5)
fL oA dt
where R = L/(crA) is the resistance, of the loop. By convention,
the current is taken as positive in the direction of the line
integral.
The flux in (5) has contributions both from the imposed
current as given in (3) and from the induced current pro
portional to the loop's self-inductance L, which for example is
given in Section 5-4-3c for a square loop (D = d):
(D = M(r)I + Li (6)
If the loop is also moving radially outward with velocity
vr = dr/dt, the electromotively induced Ohmic voltage is
-iR =
dt
dI dM(r) di
d= di d
dI dMdr di
= M(r)-+I +L-d(7)
dt dr dt dt
where L is not a function of the loop's radial position.
If the loop is stationary, only the first and third terms on
the right-hand side contribute. They are nonzero only if the
currents change with time. The second term is due to the
motion and it has a contribution even for dc currents.
Turn-on Transient. If the loop is stationary (drldt =0) at
r = ro, (7) reduces to
di dl 8
L-+ iR = -M(ro) (8)
di dt
If the applied current I is a dc step turned on at t =0, the
solution to (8) is
i(t) =
M(ro)I Lt
(/~.t>O (9)
where the impulse term on the right-hand side of (8) imposes

the initial condition i(t=0)=-M(ro)I/L. The current is
negative, as Lenz's law requires the self-flux to oppose the
applied flux.
Turn-off Transient. If after a long time T the current I is
instantaneously turned off, the solution is
i(S)= M(ro)IeLaT t> T (10)

L
where now the step decrease in current I at t = T reverses the
direction of the initial current.
Motion with a dc Current. With a dc current, the first term
on the right-hand side in (7) is zero yielding
di 0oIDd dr 1)
L-+iR (_____)
dt 27rr(r+d) dt
To continue, we must specify the motion so that we know how
r changes with time. Let's consider the simplest case when the
loop has no resistance (R = 0). Then (11) can be directly
integrated as
Li L 'oIDIn
2=- l+d/r
21r lIn r
I+ d/ro (12)
where we specify that the current is zero when r =ro. This

solution for a lossless loop only requires that the total flux of
(6) remain constant. The current is positive when r > ro as the
self-flux must aid the decreasing imposed flux. The current is
similarly negative when r < ro as the self-flux must cancel the
increasing imposed flux.
The force on the loop for all these cases is only due to the
force on the z-directed current legs at r and r+d:
SAODiI I
21 r+d r)
yLoDiId
21rr(r+d)
being attractive if iI> 0 and repulsive if iI <0.
(b) Open Circuited Loop

If the loop is open circuited, no induced current can flow
and thus the electric field within the wire is zero (J = rE =0).
The electromotive force then only has a contribution from
the gap between terminals equal to the negative of the
voltage:
d4 d4
fE-dl= E-dl=-v -- :v- (14)

JL bdt d
400 Electromagnetic Induction
Note in Figure 6-4 that our convention is such that the cur
rent i is always defined positive flowing out of the positive
voltage terminal into the loop. The flux (D in (14) is now only
due to the mutual flux given by (3), as with i =0 there is no
self-flux. The voltage on the moving open circuited loop is
then
dl dMdr
v=M(r) + I (15)
dt dr dt
(c) Reaction Force

The magnetic force on a short circuited moving loop is
always in the direction opposite to its motion. Consider the
short circuited loop in Figure 6-5, where one side of the loop
moves with velocity v,. With a uniform magnetic field applied
normal to the loop pointing out of the page, an expansion of
the loop tends to' link more magnetic flux requiring the
induced current to flow clockwise so that its self-flux is in the
direction given by the right-hand rule, opposite to the applied
field. From (1) we have
il dCI dx
E - dl=-= iR -- BoD- = BoDv. (16)
L o-A ' dt dt
where I 2(D+x) also changes with time. The current is then
.BoDv.
Z=B (17)
R
y
D '' B : eB B= Boi.
F p
0 Expanding loop
A I F eB
Contracting loop
Figure 6-5 If a conductor moves perpendicular to a magnetic field a current is

induced in the direction to cause the Lorentz force to be opposite to the motion. The
total flux through the closed loop, due to both the imposed field and the self-field
generated by the induced current, tries to remain constant.
where we neglected the self-flux generated by i, assuming it to

be much smaller than the applied flux due to Bo. Note also
that the applied flux is negative, as the right-hand rule
applied to the direction of the current defines positive flux
into the page, while the applied flux points outwards.
The force on the moving side is then to the left,
2 2
f = -iDi, x Boi. = -iDB i= = - i (18)

R
opposite to the velocity.
However if the side moves to the left (v, <0), decreasing
the loop's area thereby linking less flux, the current reverses
direction as does the force.
6-1-3 Laminations
The induced eddy currents in Ohmic conductors results in

Ohmic heating. This is useful in induction furnaces that
melt metals, but is undesired in many iron core devices. To
reduce this power loss, the cores are often sliced into many
thin sheets electrically insulated from each other by thin oxide
coatings. The current flow is then confined to lie within a thin
sheet and cannot cross over between sheets. The insulating
laminations serve the same purpose as the cuts in the slotted
ax in Figure 6-3a.
The rectangular conductor in Figure 6-6a has a time vary
ing magnetic field B(t) passing through it. We approximate
the current path as following the rectangular shape so that
dx w
- -dy L
L dy 3 x L
B() (a) B(t) (b)
Figure 6-6 (a) A time varying magnetic field through a conductor induces eddy
currents that cause Ohmic heating. (b) If the conductor is laminated so that the
induced currents are confined to thin strips, the dissipated power decreases.
402 ELectromagneticInduction
the flux through the loop of incremental width dx and dy of

area 4xy is
D= -4xyB(t) (19)
where we neglect the reaction field of the induced current
assuming it to be much smaller than the imposed field. The
minus sign arises because, by the right-hand rule illustrated in
Figure 6-2, positive flux flows in the direction opposite to
B(t). The resistance of the loop is
4I x 4 LXF+ IE\ 2 1
RL= + = (20)
The electromotive force around the loop then just results in

an Ohmic current:
-d(D dB 4L 2 dB
Edl= iR== =4xy-=-x -- (21)
ftd dt w dt
with dissipated power
i 2R. = 4Dx 3 oL(dB/dt) 2 dx

w[1+(w/L)(]
The total power dissipated over the whole sheet is then
found by adding the powers dissipated in each incremental
loop:
w/2
P= dp
4D(dB/dt)2 o-L w12 x 3

dx
w[ l+(w/L)2 o
LDw~o-(dB/dt)2
2
16[1+(w/L) ]
If the sheet is laminated into N smaller ones, as in Figure
6-6b, each section has the same solution as (23) if we replace w
by wIN. The total power dissipated is then N times the power
dissipated in a single section:
_ LD(wIN)3r(dB/dt)2 N crLDw3(dBldt) 2
16[1+(w/NL) 2] 16N 2[l+(wINL)2]
As N becomes large so that w/NL < 1, the dissipated power
decreases as 1/N2
6-1-4 Betatron
The cyclotron, discussed in Section 5-1-4, is not used to

accelerate electrons because their small mass allows them to
reach relativistic speeds, thereby increasing their mass and

decreasing their angular speed. This puts them out of phase
with the applied voltage. The betatron in Figure 6-7 uses the
transformer principle where the electrons circulating about
the evacuated toroid act like a secondary winding. The
imposed time varying magnetic flux generates an electric field
that accelerates the electrons.
Faraday's law applied to a contour following the charge's
trajectory at radius R yields
fE - dl= E02rR = -- (25)

di
which accelerates the electrons as
dv, e dcl e (6
m-ds= -eEs = e Q=v#= e ( (26)
dt 27rR dt 27rmR
The electrons move in the direction so that their self-
magnetic flux is opposite to the applied flux. The resulting
Lorentz force is radially inward. A stable orbit of constant
radius R is achieved if this force balances the centrifugal
force:
2
dv, my,
M-=--ev.B(R) =0 (27)
di R
which from (26) requires the flux and magnetic field to be

related as
0=21rR2B,(R) (28)
This condition cannot be met by a uniform field (as then
0 = 1rR 2B ) so in practice the imposed field is made to
approximately vary with radial position as
B.(r)=Bo(- Q>Q=21rJ B,(r)rdr=2,rR2 Bo (29)
Figure 6-7 "thebetatron accelerates electrons to high speeds using the electric field
generated by a time varying magnetic field.
where R is the equilibrium orbit radius, so that (28) is

satisfied.
The magnetic field must remain curl free where there is no
current so that the spatial variation in (29) requires a radial
magnetic field component:
(aB, _ B) BoR
VxB= 14=0>B= 2z (30)
( z ar r
Then any z-directed perturbation displacements
d2 z ev!t, eB\ 2
-, = - - (R)= - - z
t _M In
z =A 1 sin wt+A 2 cos w0 t, wo= (31)
have sinusoidal solutions at the cyclotron frequency wo=

eBo/m, known as betatron oscillations.
6-1-5 Faraday's Law and Stokes' Theorem
The integral form of Faraday's law in (1) shows that with

magnetic induction the electric field is no longer conservative
as its line integral around a closed path is non-zero. We may
convert (1) to its equivalent differential form by considering a
stationary contour whose shape does not vary with time.
Because the area for the surface integral does not change
with time, the time derivative on the right-hand side in (1)
may be brought inside the integral but becomes a partial
derivative because B is also a function of position:
E - dl=- -B- dS (32)

at
Using Stokes' theorem, the left-hand side of (32) can be
converted to a surface integral,
E.dl= VxE.dS- -dS (33)

, fS s at
which is equivalent to
B
j(VxE+- -dS=0 (34)
Since this last relation is true for any surface, the integrand
itself must be zero, which yields Faraday's law of induction in
differential form as
BB
VxE- (35)
at
Magnetic Circuits 405
6-2 MAGNETIC CIRCUITS
Various alloys of iron having very high values of relative

permeability are typically used in relays and machines to
constrain the magnetic flux to mostly lie within the permeable
material.
6-2-1 Self-Inductance
The simple magnetic circuit in Figure 6-8 has an N turn

coil wrapped around a core with very high relative
permeability idealized to be infinite. There is a small air gap
of length s in the core. In the core, the magnetic flux density
B is proportional to the magnetic field intensity H by an
infinite permeability g. The B field must remain finite to keep
the flux and coil voltage finite so that the H field in the core
must be zero:
H=O
lim B=AH-> (1)
B finite
Contour of integration of
Ampere's law
PO
p * 00
Nturns _-------
di
+ H = -S s
Flux leaving
Closed surface S
has zero net flux
through it
Flux entering SL_

Current i passes perpendicularly
through contour N times a
a
v
u
u
'
Faraday s law evaluated for dashed contour o ow ngN
turn coil in the direction of the current
Figure 6-8 The magnetic field is zero within an infinitely permeable magnetic core
and is constant in the air gap if we neglect fringing. The flux through the air gap is
constant at every cross section of the magnetic circuit and links the N turn coil N times.
406 Eleciromagaetic Induction
The H field can then only be nonzero in the air gap. This
field emanates perpendicularly from the pole faces as no
surface currents are present so that the tangential component
of H is continuous and thus zero. If we neglect fringing field
effects, assuming the gap s to be much smaller than the width
d or depth D, the H field is uniform throughout the gap.
Using Ampere's circuital law with the contour shown, the
only nonzero contribution is in the air gap,
H - d= Hs = Itotal encosed = Ni (2)
where we realize that the coil current crosses perpendicularly

through our contour N times. The total flux in the air gap is
then
Db= ioHDd= uoNDd (3)

S
Because the total flux through any closed surface is zero,
5SB - dS =0 (4)
all the flux leaving S in Figure 6-8 on the air gap side enters
the surface through the iron core, as we neglect leakage flux
in the fringing field. The flux at any cross section in the iron
core is thus constant, given by (3).
If the coil current i varies with time, the flux in (3) also
varies with time so that a voltage is induced across the coil. We
use the integral form of Faraday's law for a contour that lies
within the winding with Ohmic conductivity o-, cross sectional
area A, and total length 1. Then the current density and
electric field within the wire is
11 J
J=-, E=-=-- (5)
A a- oA
so that the electromotive force has an Ohmic part as well as a
contribution due to the voltage-across the terminals:
C di"
E-d= -- d1+J E-dI=---- fB-dS (6)
-L f A bdt
iR across
in wire terminals
The surface S on the right-hand side is quite complicated

because of the spiral nature of the contour. If the coil only
had one turn, the right-hand side of (6) would just be the time
derivative of the flux of (3). For two turns, as in Figure 6-9,
the flux links the coil twice, while for N turns the total flux
- dS
c = fB
Spiral surface S
Flux 0 through
t 0 9 Flux linked by a
N turn coil is
a single loop
Flux linked
4
by a two turn
loop is 20 4
Figure 6-9 The complicated spiral surface for computation of the linked flux by an N
turn coil can be considered as N single loops each linking the same flux 4,.
linked by the coil is NM. Then (6) reduces to

di
v=iR+L (7)
dt
where the self-inductance is defined as
B dS = 1N2Dd henry [kg-m 2-A -s2] (8) L =N= N =

S fL H - dl s
For linearly permeable materials, the inductance is always
independent of the excitations and only depends on the
geometry. Because of the fixed geometry, the inductance is a
constant and thus was taken outside the time derivative in (7).
In geometries that change with time, the inductance will also
be a function of time and must remain under the derivative.
The inductance is always proportional to the square of the
number of coil turns. This is because the flux 4, in the air gap
is itself proportional to N and it links the coil N times.
EXAMPLE 6-1 SELF-INDUCTANCES
Find the self-inductances for the coils shown in Figure

6-10.
(a) Solenoid
An N turn coil is tightly wound upon a cylindrical core of
radius a, length 1, and permeability A.
Contour o f integration No net current cuts

of Am pere's law contour (equal but opposite
(magnetic field negligible Contributions from upward
flutsi de coil) and downward currents)
outs
= Ni
a t .-- -No current

-cuts contour
4
4
/
4
4
4
4
4
N turns
qN turns
Net current cutting
contour = Ni
(a) Wb
Figure 6-10 Inductances. (a) Solenoidal coil; (b) toroidal coil.
SOLUTION
A current i flowing in the wire approximates a surface

current
K6 = Ni/l
If the length I is much larger than the radius a, we can neglect
fringing field effects at the ends and the internal magnetic
field is approximately uniform and equal to the surface cur
rent,
Ni
H.= K0=
as we assume the exterior magnetic field is negligible. The

same result is obtained using Ampere's circuital law for the
contour shown in Figure 6-10a. The flux links the coil N
times:
NID NAH. ra2 N2 1A'ra
2
L=
(b) Toroid
AnN turn coil is tightly wound around a donut-shaped core
of permeability 1A with a rectangular cross section and inner
and outer radii R, and R 2
.
SOLUTION
Applying Ampere's circuital law to the three contours

shown in Figure 6-10b, only the contour within the core has a
net current passing through it:
0, r<R1
fH-dl=H027rr= Ni, R><r<R2
0, r>R2
The inner contour has no current through it while the outer
contour enclosing the whole toroid has equal but opposite
contributions from upward and downward currents.
The flux through any single loop is
R2
(1) yD Ho Hdr
piDNi R2 dr
21r fRr r
yDNi R 2
= In-
27r R1
so that the self-inductance is
N'F pgDN 2R 2
L= = -D In-
i 27r R1
6-2-2 Reluctance
Magnetic circuits are analogous to resistive electronic

circuits if we define the magnetomotive force (MMF) 9
analogous to the voltage (EMF) as
Ni (9)
The flux then plays the same role as the current in electronic
circuits so that we define the magnetic analog to resistance as
the reluctance:
9 N2 (length)
(D L (permeability)(cross-sectional area)
which is proportional to the reciprocal of the inductance.

The advantage to this analogy is that the rules of adding
reluctances in series and parallel obey the same rules as resist
ances.
410 ElectromagneticInduction
(a) Reluctances in Series

For the iron core of infinite permeability in Figure 6-11 a,
with two finitely permeable gaps the reluctance of each gap is
found from (8) and (10) as
iai = s2
22
2 1= (11)
pjajD' ApaD
so that the flux is

_ Ni NO N2
(D= = - =>L=-= (12)
The iron core with infinite permeability has zero reluctance.

If the permeable gaps were also iron with infinite permeabil
ity, the reluctances of (11) would also be zero so that the flux
p-+4 o
i
+
Nturns
U1a,D
L----a2
-
. S2 evaluating Ampere's law
Depth D (a)
)- - M
i
C
C
P N turns-
r_-_-_-_-_-_-_-_-P-
S
: _
Jr= Ni
02 =afaq
C
pA a,D
L- Paths for evaluation

of Ampere's circuital
Depth D law which give us
that H, = H2 = Ni/s
(b)
Figure 6-11 Magnetic circuits are most easily analyzed from a circuit approach where
(a) reluctances in series add and (b) permeances in parallel add.
in (12) becomes infinite. This is analogous to applying a

voltage across a short circuit resulting in an infinite current.
Then the small resistance in the wires determines the large
but finite current. Similarly, in magnetic circuits the small
reluctance of a closed iron core of high permeability with no
gaps limits the large but finite flux determined by the satura
tion value of magnetization.
The H field is nonzero only in the permeable gaps so that
Ampere's law yields
Hls 1+H2s2 =Ni (13)
Since the flux must be continuous at every cross section,
= s 1iHia 1D = A2 Ha2D (14)
we solve for the H fields as
.=lals2a2Ni H 2= .LaiNi
H1 a1s2+A2a2S iaIs2 +A 2a2sI
(b) Reluctances in Parallel

If a gap in the iron core is filled with two permeable materials,
as in Figure 6-1 lb, the reluctance of each material is still given
by (11). Since each material sees the same magnetomotive
force, as shown by applying Ampere's circuital law to
contours passing through each material,
Ni
His = H 2s = Ni ->Hi= H 2 =- (16)
s
the H fields in each material are equal. The flux is then
Ni(Mt 1 +22)
0 = (giHlaa +A2H2a2)D = Ni91+R)= Ni (91 +92)
(17)
where the permeances 01 and -2 are just the reciprocal
reluctances analogous to conductance.
6-2-3 Transformer Action
(a) Voltages are not Unique

Consider two small resistors R, and R 2 forming a loop
enclosing one leg of a closed magnetic circuit with permeabil
ity A, as in Figure 6-12. An N turn coil excited on one leg with
a time varying current generates a time varying flux that is
approximately
'D(t)=pNAiI/ (18)
where I is the average length around the core.
Cross sectional R2 d4
area A V 2 =-iR 2 RI + R 2 dt
R2
II
-N turns
Li
R1 df
V, = iRl = R 1 + R2 dt
del
V1 - V2 = dt
Figure 6-12 Voltages are not unique in the presence of a time varying magnetic field.
A resistive loop encircling a magnetic circuit has different neasured voltages across the
same node pair. The voltage difference is equal to the time rate of magnetic flux
through the loop.
Applying Faraday's law to the resistive loop we have
d4(Q) =>i= 1 dQi

f E-dl=i(Rs+R 2 )=+ (19)
-
L dt R 1 +R 2 dt
where we neglect the self-flux produced by the induced cur

rent i and reverse the sign on the magnetic flux term because
D penetrates the loop in Figure 6-12 in the direction opposite
to the positive convention given by the right-hand rule illus
trated in Figure 6-2.
If we now measured the voltage across each resistor, we
would find different values and opposite polarities even
though our voltmeter was connected to the same nodes:
v, =iR, =+ R
R,1 dot
-(
R 1 +R2 di
-R 2 d1 (20)
V= -iR 2
R 1 +R 2 dt

This nonuniqueness of the voltage arises because the elec

tric field is no longer curl free. The voltage difference
between two points depends on the path of the connecting
wires. If any time varying magnetic flux passes through the
contour defined by the measurement, an additional contri
bution results.
(b) Ideal Transformers

Two coils tightly wound on a highly permeable core, so that
all the flux of one coil links the other, forms an ideal trans
former, as in Figure 6-13. Because the iron core has an
infinite permeability, all the flux is confined within the core.
The currents flowing in each coil, it and i2 , are defined so
that when they are positive the fluxes generated by each coil
are in the opposite direction. The total flux in the core is then
Nji, -N 2 i I
@= (21)
where 92 is the reluctance of the core and I is the average

length of the core.
The flux linked by each coil is then
yEA 2i
A
=N 1=-(Niii-NN2 i2
)
(22)
yA
A 2 =N2)= -(NN 2 ii-Nii2
I
)
Cross sectional
area A
7
+
i 2
N1 turns
N2
V1 (1) turns _
onV2R L
+
-F.
Secondary winding
Primary ci,
winding
I,
vrg core eng

v1 N,
v2 N2
ii N2 N =V2
I2 N,
(a)
Figure 6-13 (a) An ideal transformer relates primary and secondary voltages by the
ratio of turns while the currents are in the inverse ratio so that the input power equals
the output power. The H field is zero within the infinitely permeable core. (b) In a real
transformer the nonlinear B-H hysteresis loop causes a nonlinear primary current i
with an open circuited secondary (i 2 =0) even though the imposed sinusoidal voltage
v, = VO cos wt fixes the flux to be sinusoidal. (c) A more complete transformer equivalent
circuit.
414 Eectromagnetic Induction
li ur,m dB =;J
B =0 sin W
I i II I KIH I
0 B field time scale r I

IIi
Iii I
III I
H=
(b)
. 1I R1 N, N2'~~~ R2 X2 -2
+
+
L _ .....
Ideal transformer
(c)
Figure 6.13.
which can be written as
AI Llij-Mi2
(23)
A 2 =Mil-L 2i 2
where LI and L 2 are the self-inductances of each coil alone
and M is the mutual inductance between coils:
LI=NIL, L 2 = N2L 0 , M =N,N2 Lo, Lo =pAl
(24)
In general, the mutual inductance obeys the equality:

2
M=k(LIL 2) , 0 ks1 (25)
where k is called the coefficient of coupling. For a noninfinite

core permeability, k is less than unity because some of the flux
of each coil goes into the free space region and does not link
the other coil. In an ideal transformer, where the permeabil
ity is infinite, there is no leakage flux so that k = 1.
From (23), the voltage across each coil is
dA, di, di2
=-~=L1 -M
di dt dt
(26)
dA 2 di, di2
v= -- =M--L2
di di di
Because with no leakage, the mutual inductance is related
to the self-inductances as
N2 N
M=-L =-L1
2 (27)
N1 N2
the ratio of coil voltages is the same as the turns ratio:
v, dA,/dt N1
v2 dA 2/dt N2
In the ideal transformer of infinite core permeability, the
inductances of (24) are also infinite. To keep the voltages and
fluxes in (26) finite, the currents must be in the inverse turns
ratio
-=- N(29)
i2
N,
The electrical power delivered by the source to coil 1, called
the primary winding, just equals the power delivered to the
load across coil 2, called the secondary winding:
v~iI=V2 i2 (30)
If N 2 >Nl, the voltage on winding 2 is greater than the

voltage on winding I but current i2 is less than iI keeping the
powers equal.
If primary winding 1 is excited by a time varying voltage
vI(t) with secondary winding 2 loaded by a resistor RL so that
V2= i2 RL (31)
the effective resistance seen by the primary winding is
R= v= -2 .V - RL (32)
ii N 2 (N2/Ni)i 2 N2
A transformer is used in this way as an impedance trans

former where the effective resistance seen at the primary
winding is increased by the square of the turns ratio.
(c) Real Transformers

When the secondary is open circuited (i 2 = 0), (29) shows
that the primary current of an ideal transformer is also zero.
In practice, applying a primary sinusoidal voltage VO cos wt
will result in a small current due to the finite self-inductance
of the primary coil. Even though this self-inductance is large
if the core permeability y is large, we must consider its effect
because there is no opposing flux as a result of the open
circuited secondary coil. Furthermore, the nonlinear
hysteresis curve of the iron as discussed in Section 5-5-3c will
result in a nonsinusoidal current even though the voltage is
sinusoidal. In the magnetic circuit of Figure 6.13a with i 2 =0,
the magnetic field is
H= (33)
while the imposed sinusoidal voltage also fixes the magnetic

flux to be sinusoidal
d'F 0 V
v1=---= Vocos wt |> b= BA =--sin wt (34)
.
dtW
Thus the upper half of the nonlinear B-H magnetization

characteristic in Figure 6-13b has the same shape as the flux-
current characteristic with proportionality factors related to
the geometry. Note that in saturation the B-H curve
approaches a straight line with slope .Lo. For a half-cycle of
flux given by (34), the nonlinear open circuit magnetizing
current is found graphically as a function of time in Figure
6-13b. The current is symmetric over the negative half of the
flux cycle. Fourier analysis shows that this nonlinear current is
composed of all the odd harmonics of the driving frequency
dominated by the third and fifth harmonics. This causes
problems in power systems and requires extra transformer
windings to trap the higher harmonic currents, thus prevent
ing their transmission.
A more realistic transformer equivalent circuit is shown in
Figure 6-13c where the leakage reactances X, and X 2
represent the fact that all the flux produced by one coil does
not link the other. Some small amount of flux is in the free
space region surrounding the windings. The nonlinear
inductive reactance X, represents the nonlinear magnetiza
tion characteristic illustrated in Figure 6-13b, while R,
represents the power dissipated in traversing the hysteresis
Faraday'sLaw for Moving Media 41 7
loop over a cycle. This dissipated power per cycle equals the
area enclosed by the hysteresis loop. The winding resistances
are R, and R 2
.
6-3 FARADAY'S LAW FOR MOVING MEDIA
6-3-1 The Electric Field Transformation
If a point charge q travels with a velocity v through a region

with electric field E and magnetic field B, it experiences the
combined Coulomb-Lorentz force
F= q(E+vX B) (1)
Now consider another observer who is travelling at the same
velocity v as the charge carrier so that their relative velocity is
zero. This moving observer will then say that there is no
Lorentz force, only a'Coulombic force
F'= qE' (2)
where we indicate quantities measured by the moving obser

ver with a prime. A fundamental postulate of mechanics is
that all physical laws are the same in every inertial coordinate
system (systems that travel at constant relative velocity). This
requires that the force measured by two inertial observers be
the same so that F'= F:
E'= E+vX B (3)
The electric field measured by the two observers in relative

motion will be different. This result is correct for material
velocities much less than the speed of light and is called a
Galilean field transformation. The complete relativistically
correct transformation slightly modifies (3) and is called a
Lorentzian transformation but will not be considered here.
In using Faraday's law of Section 6-1-1, the question
remains as to which electric field should be used if the
contour L and surface S are moving. One uses the electric
field that is measured by an observer moving at the same
velocity as the convecting contour. The time derivative of the
flux term cannot be brought inside the integral if the surface
S is itself a function of time.
6-3-2 Ohm's Law for Moving Conductors
The electric field transformation of (3) is especially

important in modifying Ohm's law for moving conductors.
For nonrelativistic velocities, an observer moving along at the
418 ELectromagnetic Induction
same velocity as an Ohmic conductor measures the usual

Ohm's law in his reference frame,
Jf= a-E' (4)
where we assume the conduction process is unaffected by the
motion. Then in Galilean relativity for systems with no free
charge, the current density in all inertial frames is the same so
that (3) in (4) gives us the generalized Ohm's law as
J'=J1 = o-(E+vx B) (5)
where v is the velocity of the conductor.
The effects of material motion are illustrated by the parallel
plate geometry shown in Figure 6-14. A current source is
applied at the left-hand side that distributes itself uniformly
as a surface current K. = *I/D on the planes. The electrodes
are connected by a conducting slab that moves to the right with
constant velocity U. The voltage across the current source can
be compu-ted using Faraday's law with the contour shown. Let
us have the contour continually expanding with the 2-3 leg
moving with the conductor. Applying Faraday's law we have
4 0 1
f2/ P 3
E' - dl dl+ E'- dl+ -dl+ E- di
iR -V
=--d B-dS (6)

dt
Surface current
K.=I I
II D
it
4' _ _.3.
---- d+
H. =-Return surface
D current K, -1
Figure 6-14 A moving, current-carrying Ohmic conductor generates a speed voltage

as well as the usual resistive voltage drop.
Faraday's Law for Moving Media 419
where the electric field used along each leg is that measured
by an observer in the frame of reference of the contour.
Along the 1-2 and 3-4 legs, the electric field is zero within the
stationary perfect conductors. The second integral within the
moving Ohmic conductor uses the electric field E', as
measured by a moving observer because the contour is also
expanding at the same velocity, and from (4) and (5) is related
to the terminal current as
E'=
J' I
i (7)
o- o-Dd
(
In (6), the last line integral across the terminals defines the
voltage.
Is d Cd
is -v=--- B-dS=- d(oHexs) (8)
o-Dd dt sdt
The first term is just the resistive voltage drop across the
conductor, present even if there is no motion. The term on
the right-hand side in (8) only has a contribution due to the
linearly increasing area (dxldt = U) in the free space region
with constant magnetic field,
H,= I/D (9)

The terminal voltage is then
v = IfR +AoUsI, R= (10)

\ DI oDd
We see that the speed voltage contribution arose from the

flux term in Faraday's law. We can obtain the same solution
using a contour that is stationary and does not expand with
the conductor. We pick the contour to just lie within the
conductor at the time of interest. Because the contour does
not expand with time so that both the magnetic field and the
contour area does not change with time, the right-hand side
of (6) is zero. The only difference now is that along the 2-3 leg
we use the electric field as measured by a stationary observer,
E=E'-vxB (11)
so that (6) becomes
IR+ IS- V=0 (12)

D
which agrees with (10) but with the speed voltage term now
arising from the electric field side of Faraday's law.
This speed voltage contribution is the principle of electric
generators converting mechanical work to electric power
when moving a current-carrying conductor through a

magnetic field. The resistance term accounts for the electric
power dissipated. Note in (10) that the speed voltage contri
bution just adds with the conductor's resistance so that the
effective terminal resistance is v/I = R +(yt.Us/D). If the slab
moves in the opposite direction such that U is negative, the
terminal resistance can also become negative for sufficiently
large U (U<-RD/os). Such systems are unstable where the
natural modes grow rather than decay with time with any
small perturbation, as illustrated in Section 6-3-3b.
6-3-3 Faraday's Disk (Homopolar Generator)*
(a) Imposed Magnetic Field

A disk of conductivity o- rotating at angular velocity w
transverse to a uniform magnetic field Boi,, illustrates the
basic principles of electromechanical energy conversion. In
Figure 6-15a we assume that the magnetic field is generated
by an N turn coil wound on the surrounding magnetic circuit,
BO=ILoNif (13)
The disk and shaft have a permeability of free space yo, so

that the applied field is not disturbed by the assembly. The
shaft and outside surface at r = RO are highly conducting and
make electrical connection to the terminals via sliding
contacts.
We evaluate Faraday's law using the contour shown in
Figure 6-15a where the 1-2 leg within the disk is stationary so
the appropriate electric field to be used is given by (11):
E,=r wrBo -wrBo (14)

" 2 7rodr
where the electric field and current density are radial and i, is
the total rotor terminal current. For the stationary contour
with a constant magnetic field, there is no time varying flux
through the contour:
24
E - dl= E, dr + E - dl= 0 (15)
* Some of the treatment in this section is similarto that developed in: H. H. Woodson andJ. R.
Melcher, Electromechanical Dynamics, Part I, Wiley, N. Y., 1968, Ch. 6.
Faraday'sLaw for Moving Media 421
IR
I
VjN turns 2 i
-O poNf - 4+7
Stationary contour of
integration of Faraday's law
(a)
Rr Lr
fg Lf
V/ Rt Vr Gwil
(b)
Figure 6-15 (a) A conducting disk rotating in an axial magnetic field is called a
homopolar generator. (b) In addition to Ohmic and inductive voltages there is a speed
voltage contribution proportional to the speed of the disk and the magnetic field.
Using (14) in (15) yields the terminal voltage as
vr= RO -wrBo dr
Ri 27rTo-d
= Ro wBo(R -R
)
2,ro-d Ri 2
=irRr-GWif (16)
where R, is the internal rotor resistance of the disk and G is
called the speed coefficient:
In (Ro/Ri) ~ -~ 2 2
R,= ( G= ON(R 0 -R,) (17)
27ro-d 2s
We neglected the self-magnetic field due to the rotor current,
assuming it to be much smaller than the applied field Bo, but
it is represented in the equivalent rotor circuit in Figure 6-15b

as the self-inductance L, in series with a resistor and a speed
voltage source linearly dependent on the field current. The
stationary field coil is represented by its self-inductance and
resistance.
For a copper disk (o= 6 x 107 siemen/m) of thickness 1 mm
rotating at 3600 rpm (w = 1207r radian/sec) with outer and
inner radii RO = 10 cm and Ri = 1 cm in a magnetic field of
Bo = 1 tesla, the open circuit voltage is
vc= >Bo(R2- R) - -1.9 V (18)

2
while the short circuit current is
i.= VoC 2Iro-d - 3 X 105 amp (19)

In (RolRi)
Homopolar generators are typically high current, low voltage

devices. The electromagnetic torque on the disk due to the
Lorentz force is
T=f
2vr
-
Lo
d
= f-
Ro
riX(JXB)rdrd46 dz
RO
=-i,Boi. rdr
=--- (R2 -R )i,

2
=-Gii,i. (20)
The negative sign indicates that the Lorentz force acts on

the disk in the direction opposite to the motion. An external
torque equal in magnitude but opposite in direction to (20) is
necessary to turn the shaft.
This device can also be operated as a motor if a rotor
current into the disk (i, <0) is imposed. Then the electrical
torque causes the disk to turn.
(b) Self-Excited Generator

For generator operation it is necessary to turn the shaft and
supply a field current to generate the magnetic field.
However, if the field coil is connected to the rotor terminals,
as in Figure 6-16a, the generator can supply its own field
current. The equivalent circuit for self-excited operation is
shown in Figure 6-16b where the series connection has i,= ir.
N turnsW
if BO
vItR
L =L, + L,
i,= R=R 1 +Rr
ir Rr
L,
Vr Gwif
Figure 6-16 A homopolar generator can be self-excited where the generated rotor
current is fed back to the field winding to generate its own magnetic field.
Kirchoff's voltage law around the loop is

di
L +i(R -G>)=0, R = R,+ Rf, L = L,+ Lf
dt
(21)
where R and L are the series resistance and inductance of the
coil and disk. The solution to (21) is
i = Io e -[(R-C ILl (22

where Io is the initial current at t = 0. If the exponential factor
is positive
Gw>R (23)
the current grows with time no matter how small Io is. In

practice, Io is generated by random fluctuations (noise) due to
residual magnetism in the iron core. The exponential growth
is limited by magnetic core saturation so that the current
reaches a steady-state value. If the disk is rotating in the
opposite direction (w <0), the condition of (23) cannot be
satisfied. It is then necessary for the field coil connection to be
reversed so that i,= -if. Such a dynamo model has been used
as a model of the origin of the earth's magnetic field.
(c) Self-Excited ac Operation

Two such coupled generators can spontaneously generate
two phase ac power if two independent field windings are
connected, as in Figure 6-17. The field windings are con
nected so that if the flux through the two windings on one
machine add, they subtract on the other machine. This
accounts for the sign difference in the speed voltages in the
equivalent circuits,
di1
L--+ (R - GOw)ii + G0i2 = 0
dt
di2 (24)
L-+(R-GW)i-Gwi1= 0
dt
where L and R are the total series inductance and resistance.

The disks are each turned at the same angular speed w.
Since (24) are linear with constant coefficients, solutions are
of the form
i = Ie"', i2 =12 e" (25)
which when substituted back into (24) yields
(Ls + R - Gw)I + GwI2 =0

-GfoIi+(Ls + R -Go)I 2 =0 (26)
For nontrivial solutions, the determinant of the coefficients of
I, and I2 must be zero,
(Ls + R -GW)2 =-(GW)2 (27)
which when solved for s yields the complex conjugate natural

frequencies,
(R-Gw) .Gw
L L
I112=*Ej (28)
where the currents are 90* out of phase. If the real part of s is
positive, the system is self-excited so that any perturbation
/~~~
0
7
R1 Lj R1 Ll Rr Lr
L=-2L + L,
Rf L1 L, R Lr R=2R1 +Rr
;2
+
Figure 6-17 Cross-connecting two homopolar generators can result in self-excited

two-phase alternating currents. Two independent field windings are required where
on one machine the fluxes add while on the other they subtract.
grows at an exponential rate:

Gw>R (29)
The imaginary part of s yields the oscillation frequency
Co= I m (s)=GwIL (30)
Again, core saturation limits the exponential growth so that
two-phase power results. Such a model may help explain the
periodic reversals in the earth's magnetic field every few
hundred thousand years.
(d) Periodic Motor Speed Reversals

If the field winding of a motor is excited by a dc current, as
in Figure 6-18, with the rotor terminals connected to a
generator whose field and rotor terminals are in series, the
circuit equation is
di (R - Gio,) . & if (31)
-- + =- If (1
dt L L
where L and R are the total series inductances and resis

tances. The angular speed of the generator w is externally
Generator
Motor
RI-
If= RV
- Motor
Rrm Lrm
G WmIf L=Lrm +Lrg+Lfg

R =Rrm + Rig + Rrg
R 1g
Lfg
Generator
g Lrg
Ggwgi
Figure 6-18 Cross connecting a homopolar generator and motor can result in spon
taneous periodic speed reversals of the motor's shaft.
constrained to be a constant. The angular acceleration of the

motor's shaft is equal to the torque of (20),
dom
J--= -- GIfi (32)
dt
where J is the moment of inertia of the shaft and If = Vf/Rm is
the constant motor field current.
Solutions of these coupled, linear constant coefficient
differential equations are of the form
i = Ie"
t (33)
& = Wes
which when substituted back into (31) and (32) yield
+ R -G' G I
I + Ws =0 (34)
Again, for nontrivial solutions the determinant of coefficients

of I and W must be zero,
s(s+RZ#JG +(G1) =0 (35)

L e!L L
which when solved for s yields
(R-Ggw,) [(R -Gp !) 2 (G.nif) 2 11 (36)

2L 2L JLI
For self-excitation the real part of s must be positive,
G,> R (37)
while oscillations will occur if s has an imaginary part,
(G.f) 2> R - GgW 2 (8

JL 2L)
Now, both the current and shaft's angular velocity spon
taneously oscillate with time.
6--4 Basic Motors and Generators
(a) ac Machines
Alternating voltages are generated from a dc magnetic field
by rotating a coil, as in Figure 6-19. An output voltage is
measured via slip rings through carbon brushes. If the loop
of area A is vertical at t = 0 linking zero flux, the imposed flux
428 Electromagnetic induction
+t
.NArea A dt
Wt
NIOA
2
'1PO W COSWt
Figure 6-19 A coil rotated within a constant magnetic field generates a sinusoidal
voltage.
through the loop at any time, varies sinusoidally with time

due to the rotation as
Di = (o sin wt (39)
Faraday's law applied to a stationary contour instantaneously
passing through the wire then gives the terminal voltage as
dQ di
v = iR +--= iR +L -i+(Dow cos wt (40)
dt dt
where R and L are the resistance and inductance of the wire.

The total flux is equal to the imposed flux of (39) as well as
self-flux (accounted for by L) generated by the current i. The
equivalent circuit is then similar to that of the homopolar
generator, but the speed voltage term is now sinusoidal in
time.
(b) dc Machines
DC machines have a similar configuration except that the
slip ring is split into two sections, as in Figure 6-20a. Then
whenever the output voltage tends to change sign, the
terminals are also reversed yielding the waveform shown,
which is of one polarity with periodic variations from zero to a
peak value.
(V
wr 3irW
2 2
V V V A wt
VT 7r 31r
2 2
Winding 1
----.. . . .... Winding 2
(b) Commutated volta le
Figure 6-20 (a) If the slip rings are split so that when the voltage tends to change sign
the terminals are also reversed, the resulting voltage is of one polarity. (b) The voltage
waveform can be smoothed out by placing a second coil at right angles to the first and
using a four-section commutator.
The voltage waveform can be smoothed out by using a

four-section commutator and placing a second coil perpen
dicular to the first, as in Figure 6-20b. This second coil now
generates its peak voltage when the first coil generates zero
voltage. With more commutator sections and more coils, the
dc voltage can be made as smooth as desired.
6-3-5 MHD Machines
Magnetohydrodynamic machines are based on the same

principles as rotating machines, replacing the rigid rotor by a
conducting fluid. For the linear machine in Figure 6-21, a
fluid with Ohmic conductivity a- flowing with velocity v,
moves perpendicularly to an applied magnetic field Boi.. The
terminal voltage V is related to the electric field and current
as
E=L -, J= o-(E+v XB)= -+VBo ix=-ix

s 4s Dd
(41)
which can be rewritten as
V = iR -vBos (42)
which has a similar equivalent circuit as for the homopolar
generator.
The force on the channel is then
f=tJxBdV
= -iBosi, (43)
again opposite to the fluid motion.
6-3-6 Paradoxes
Faraday's law is prone to misuse, which has led to

numerous paradoxes. The confusion arises because the same
B0
R = d
I -Dd
V
- ~ .E
V vBos
+
D
y d
Figure 6-21 An MHD (magnetohydrodynamic) machine replaces a rotating conduc

tor by a moving fluid.
contribution can arise from either the electromotive force

side of the law, as a speed voltage when a conductor moves
orthogonal to a magnetic field, or as a time rate of change of
flux through the contour. This flux term itself has two
contributions due to a time varying magnetic field or due to a
contour that changes its shape, size, or orientation. With all
these potential contributions it is often easy to miss a term or
to double count.
(a) A Commutatorless de Machine*

Many persons have tried to make a commutatorless dc
machine but to no avail. One novel unsuccessful attempt is
illustrated in Figure 6-22, where a highly conducting wire is
vibrated within the gap of a magnetic circuit with sinusoidal
velocity:
v.=vo sin (at (44)
Faraday's law applied to a

stonary contour (dashed)
instantaneously within
wire.
vibrating
1+ 1-4
4 -N 3
Eoint -, vx=o
sinwt
sinwt
'2
-L
Fcc 6-22 It is impossible to design a commutatorless dc machine. Although the speed

voltage alone can have a dc average, it will be canceled by the transformer elec
tromotive force due to the time rate of change of magnetic flux through the loop. The
total terminal voltage will always have a zero time average.
* H. Sohon, ElectricalEssays for Recreation. Electrical Engineering, May (1946), p. 294.

The sinusoidal current imposes the air gap flux density at the
same frequency w:
B. = Bo sin wt, Bo = goNIols (45)
Applying Faraday's law to a stationary contour instan

taneously within the open circuited wire yields
i0 j 2 3
4
E-dl= ,E-dl+ E -dI + E-dl+ E-dl
E=-vxB -v
- B -dS (46)
dts
where the electric field within the highly conducting wire as
measured by an observer moving with the wire is zero. The
electric field on the 2-3 leg within the air gap is given by (11),
where E'= 0, while the 4-1 leg defines the terminal voltage. If
we erroneously argue that the flux term on the right-hand side
is zero because the magnetic field B is perpendicular to dS, the
terminal voltage is
v =v.B.l =voBol sin2 a2t (47)
which has a dc time-average value. Unfortunately, this result

is not complete because we forgot to include the flux that
turns the corner in the magnetic core and passes perpen
dicularly through our contour. Only the flux to the right of
the wire passes through our contour, which is the fraction
(L - x)/L of the total flux. Then the correct evaluation of (46) is
d
-v +v.Bl =+ [(L -x)B1l) (48)
where x is treated as a constant because the contour is sta

tionary. The change in sign on the right-hand side arises
because the flux passes through the contour in the direction
opposite to its normal defined by the right-hand rule. The
voltage is then
dB
v = v,,B.l -(L - dx)lE (49)
dt
where the wire position is obtained by integrating (44),
x = v.dt = -V(cos wt - 1)+xo (50)

and xo is the wire's position at t = 0. Then (49) becomes
v =-(xB,)- Ld
dt dt
=Bolvo [(+ I)cos wt -cos 2wt - LBow cos wt (51)

VO
which has a zero time average.
(b) Changes in Magnetic Flux Due to Switching

Changing the configuration of a circuit using a switch does
not result in an electromotive force unless the magnetic flux
itself changes.
In Figure 6-23a, the magnetic field through the loop is
externally imposed and is independent of the switch position.
Moving the switch does not induce an EMF because the
magnetic flux through any surface remains unchanged.
In Figure 6-23b, a dc current source is connected to a
circuit through a switch S. If the switch is instantaneously
moved from contact 1 to contact 2, the magnetic field due to
the source current I changes. The flux through any fixed area
has thus changed resulting in an EMF.
(c) Time Varying Number of Turns on a Coil*

If the number of turns on a coil is changing with time, as in
Figure 6-24, the voltage is equal to the time rate of change of
flux through the coil. Is the voltage then
d(D
v1 N-- (52)
or
d d4 dN
dt dt d(
No current is induced
by switching.
jS IB E 7
1 2 1 2
(a) (b)
Figure 6-23 (a) Changes in a circuit through the use of a switch does not by itself
generate an EMF. (b) However, an EMF can be generated if the switch changes the
magnetic field.
* L. V. Bewley. Flux Linkages and Electromagnetic Induction. Macmillan, New York,

1952.
V 1(t), Nit)
Nit)
to, No
-.-
V
Moving
contact lt) N(t)
i794, To NO
(a)
p 0 N(t)1(t)A p'ON(t)I()A
Loo l(t)
Area A poN10 (t)
i v
t 0~
(
lit)
I
v = Nit)$ = MoN(t)A d[N(t)l]

t v=N(t) dd
dt I dt dt
poNoAN(t) di(t)
(b) (c) 10 dt
Figure 6-24 (a) If the number of turns on a coil is changing with time, the induced
voltage is v = N(t) d4'/dt. A constant flux does not generate any voltage. (b) If the flux
itself is proportional to the number of turns, a dc current can generate a voltage. (c)
With the tap changing coil, the number of turns per unit length remains constant so
that a dc current generates no voltage because the flux does not change with time.
For the first case a dc flux generates no voltage while the

second does.
We use Faraday's law with a stationary contour instan
taneously within the wire. Because the contour is stationary,
its area of NA is not changing with time and so can be taken
outside the time derivative in the flux term of Faraday's law so
that the voltage is given by (52) and (53) is wrong. Note that
there is no speed voltage contribution in the electromotive
force because the velocity of the wire is in the same direction
as the contour (v x B - dl = 0).
Magnetic Diffusion into an Ohmic Conductor 435
If the flux (D itself depends on the number of turns, as in

Figure 6-24b, there may be a contribution to the voltage even
if the exciting current is dc. This is true for the turns being
wound onto the cylinder in Figure 6-24b. For the tap changing
configuration in Figure 6-24c, with uniformly wound turns,
the ratio of turns to effective length is constant so that a dc
current will still not generate a voltage.
6-4 MAGNETIC DIFFUSION INTO AN OHMIC CONDUCTOR*
If the current distribution is known, the magnetic field can

be directly found from the Biot-Savart or Ampere's laws.
However, when the magnetic field varies with time, the
generated electric field within an Ohmic conductor induces
further currents that also contribute to the magnetic field.
6-4-1 Resistor-Inductor Model
A thin conducting shell of radius Ri, thickness A, and depth

I is placed within a larger conducting cylinder, as shown in
Figure 6-25. A step current Io is applied at t =0 to the larger
cylinder, generating a surface current K =(Io/1)i4. If the
length I is much greater than the outer radius R 0 , the
magnetic field is zero outside the cylinder and uniform inside
for Ri <r < Ro. Then from the boundary condition on the
discontinuity of tangential H given in Section 5-6-1, we have
'0
Ho=-Loi., Ri<r<Ro (1)
The magnetic field is different inside the conducting shell

because of the induced current, which from Lenz's law, flows
in the opposite direction to the applied current. Because the
shell is assumed to be very thin (A< R1 ), this induced current
can be considered a surface current related to the volume
current and electric field in the conductor as
K. = JA = (o-A)E, (2)
The product (a-A) is called the surface conductivity. Then the
magnetic fields on either side of the thin shell are also related
by the boundary condition of Section 5-6-1:
Hi - Ho= K. = (o-A)E4 (3)
of the treatment of this section is similar to that of H. H. Woodson and J. R. Melcher,

* Much
Electromechanical Dynamics, PartII, Wiley, N. Y., 1968, Ch. 7.
Kx
OH,
Ro
Induced
I~)Hi current
10
Depth I
Faraday's law
applied to contour
within cylindrical shell
of Ohmic conductivity a.
Figure 6-25 A step change in magnetic field causes the induced current within an
Ohmic conductor to flow in the direction where its self-flux opposes the externally
imposed flux. Ohmic dissipation causes the induced current to exponentially decay
with time with a LIR time constant.
Applying Faraday's law to a contour within the conducting

shell yields
E - di=
d B - dS=> E#21rRi = -povrR. 2
dH.
(4)
where only the magnetic flux due to H passes through the

contour. Then using (1)-(3) in (4) yields a single equation in
Hi:
dHi+Hi I(t) AoRicrA (5)
dt 7 IT' 2
where we recognize the time constant r as just being the ratio

of the shell's self-inductance to resistance:
4) AoirR? 2rRi L AoRio-A
L= -- R=-, =- (6
K,0 ' R I-1A R 2 (6)
The solution to (5) for a step current with zero initial
magnetic field is
Hi=L(1-e-1T) (7)
Initially, the magnetic field is excluded from inside the

conducting shell by the induced current. However, Ohmic
dissipation causes the induced current to decay with time so

that the magnetic field may penetrate through the shell with
characteristic time constant r.
64-2 The Magnetic Diffusion Equation
The transient solution for a thin conducting shell could be

solved using the integral laws because the geometry con
strained the induced current to flow azimuthally with no
radial variations. If the current density is not directly known,
it becomes necessary to self-consistently solve for the current
density with the electric and magnetic fields:
B
V xE = a (Faraday's law) (8)
at
V x H = Jf (Ampere's law) (9)
V - B=0 (Gauss's law) (10)
For linear magnetic materials with constant permeability si

and constant Ohmic conductivity o moving with velocity U,
the constitutive laws are
B =H, J=cr(E+Ux LH) (11)
We can reduce (8)-(11) to a single equation in the magnetic

field by taking the curl of (9), using (8) and (11) as
V x (V x H) = V xJ
= c[V X E+tV x (U X H)]
= +Vx(Ux H)) (12)
The double cross product of H can be simplified using the

vector identity
~0
Vx(VxH)=V(V H)-V 2 H
- -V2H =--H Vx(UxH) (13)

Af at
where H has no divergence from (10). Remember that the

Laplacian operator on the left-hand side of (13) also
differentiates the directionally dependent unit vectors in
cylindrical (i, and i#) and spherical (i,,i#, and i,) coordinates.
6-4.3 Transient Solution with Jo Motion (U =0)
A step current is turned on at t =0, in the parallel plate

geometry shown in Figure 6-26. By the right-hand rule and
with the neglect of fringing, the magnetic field is in the z
direction and only depends on the x coordinate, B.(x, t), so
that (13) reduces to
2H7 aH.
(14)
t
ax
which is similar in form to the diffusion equation of a dis
tributed resistive-capacitive cable developed in Section 3-6-4.
In the dc steady state, the second term is zero so that the
solution in each region is of the form
a H.
2 =0->H,=ax+b (15)
aX
K 0= I/D DepthD
1(t) I() H y s
=-1D
YI
101Kx
-x
0 d
(a)
t = 0. (Surface current= I/D)
1.0 2.0
-
Jv (x,t)
H, (x, t)
-
lD 0.5 t- 1/(Dd)
2.0
.50 0.1 0.25
0. .25 1.0
0.
T i 0.5
0 0.5 1.0 0.5 1.0
x/d x/d
(b)
Figure 6-26 (a) A current source is instantaneously turned on at t =0. The resulting
magnetic field within the Ohmic conductor remains continuous and is thus zero at t = 0
requiring a surface current at x =0. (b) For later times the magnetic field and current
diffuse into the conductor with longest time constant r = o-d2 /ir2 towards a steady
state of uniform current with a linear magnetic field.
where a and b are found from the boundary conditions. The

current on the electrodes immediately spreads out to a uni
form surface distribution E (IID)i, traveling from the upper
to lower electrode uniformly through the Ohmic conductor.
Then, the magnetic field is uniform in the free space region,
decreasing linearly to zero within the Ohmic conductor being
continuous across the interface at x =0:
- -I S x 0
D
lim H.(x) = (16)
sooII
t-0--I(d -x), 0:5 x - d
Dd
In the free space region where o-=0, the magnetic field
remains constant for all time. Within the conducting slab,
there is an initial charging transient as the magnetic field
builds up to the linear steady-state distribution in (16).
Because (14) is a linear equation, for the total solution of the
magnetic field as a function of time and space, we use super
position and guess a solution that is the sum of the steady-
state solution in (16) and a transient solution which dies off
with time:
IA
H.(x, t)= -(d - x) + (x) e" (17)
We follow the same procedures as for the lossy cable in

Section 3-6-4. At this point we do not know the function H(x)
or the parameter a. Substituting the assumed solution of (17)
back into (14) yields the ordinary differential equation
d2 H(x)
dx 2 +0-jkaH(x)=0 (18)
which has the trigonometric solutions

H(x)=Aisin Vlx+A 2 cosloAax (19)
Since the time-independent part in (17) already meets the
boundary conditions of
H(x = 0) = ID
H,(x=d)= 0
the transient part of the solution must be zero at the ends
H(x =0)=0=>A 2 =0
H(x = d)=0=A i sin a d =0 (21)
which yields the allowed values of a as
-lo,a d = nir =>a. = I (nsr)2 , n =1, 2,3,. ... (22)

s~o- d/
Since there are an infinite number of allowed values of a, the

most general solution is the superposition of all allowed solu
tions:
I narx
H.(x, 0)=--(d-x)+ Y An sin e -a (23)
Dd n=1 d
This relation satisfies the boundary conditions but not the
initial conditions at t = 0 when the current is first turned on.
Before the current takes its step at t =0, the magnetic field is
zero in the slab. Right after the current is turned on, the
magnetic field must remain zero. Faraday's law would
otherwise make the electric field and thus the current density
infinite within the slab, which is nonphysical. Thus we impose
the initial condition
I flnTx
H,(x,t = 0)=0=-(d-x)+ An sin (24)
Dd 71.1 d
which will allow us to solve for the amplitudes An by multi
plying (24) through by sin (m'wx/d) and then integrating over
x from 0 to d:
I d Mm CO d .nirx .mr
O=j (d-x) sin-mdx +_ An sin - sin -dx
d d n.1 0 d d
(25)
The first term on the right-hand side is easily integrable*
while the product of sine terms integrates to zero unless
m = n, yielding
2I
A. - (26)
mirD
The total solution is thus
I x * sin (nixld) _ni,.

H (x, t)= D- - 2 E - e(7 (27)
d .=1 n*1
where we define the fundamental continuum magnetic

diffusion time constant T as
1 ILc-d 2
r = -- = 2 (28)
a, ir
analogous to the lumped parameter time constant of (5) and
(6).
(d -x) sin mx dx =d
d MIT
The magnetic field approaches the steady state in times

long compared to r. For a perfect conductor (o - >), this time
is infinite and the magnetic field is forever excluded from the
slab. The current then flows only along the x = 0 surface.
However, even for copper (o--6x 10 7 siemens/m) 10-cm
thick, the time constant is r=80 msec, which is fast for many
applications. The current then diffuses into the conductor
where the current density is easily obtained from Ampere's law
as
Jf=VXH - _
,
ax
1+2 nirx
S + Cos e"i, (29)
The diffusion of the magnetic field and current density are

plotted in Figure 6-26b for various times
The force on the conducting slab is due to the Lorentz
force tending to expand the loop and a magnetization force
due to the difference of permeability of the slab and the
surrounding free space as derived in Section 5-8-1:
F= sO(M - V)H +poJf x H

= (A - o)(H - V)H + oJfX H (30)
For our case with H = H,(x)i,, the magnetization force density
has no contribution so that (30) reduces to
F= AoJf X H
= Ao(V x H) x H
=so(H -V)H -V(2 0 oH - H)
d1 2.
dx (bxoH )i (31)
Integrating (31) over the slab volume with the magnetic

field independent of y and z,
d d 1
f- sD-(WoH ) dx
dx
= -ioHsD
2 s (32)
D
gives us a constant force with time that is independent of the

permeability. Note that our approach of expressing the cur
rent density in terms of the magnetic field in (31) was easier
than multiplying the infinite series of (27) and (29), as the
result then only depended on the magnetic field at the

boundaries that are known from the boundary conditions of
(20). The resulting integration in (32) was easy because the
force density in (31) was expressed as a pure derivative of x.
6-4-4 The Sinusoidal Steady State (Skin Depth)
We now place an infinitely thick conducting slab a distance

d above a sinusoidally varying current sheet Ko cos ti,, which
lies on top of a perfect conductor, as in Figure 6-27a. The
X
-yMo - H.= -Ko coswt
d -n. O
YA
Kocaswti,
(a)
1Wt
0
1.0 - Wt H. (x, t
KO
12
-2 3
44
K0
/
-1.0 -0
(b)
Figure 6-27 (a) A stationary conductor lies above a sinusoidal surface current placed
upon a perfect conductor so that H =0 for x < - d. (b) The magnetic field and current
density propagates and decays into the conductor with the same characteristic length
given by the skin depth 8= 12/(wiA-). The phase speed of the wave is wo.
magnetic field within the conductor is then also sinusoidally

varying with time:
H.(x, t)= Re [A.(x) e*l] (33)
Substituting (33) into (14) yields
d2 A'.
jOs
.2- =0 (34)
with solution
I.(x)= A, e(+i)x O+A2 e-(l+)xa (35)
where the skin depth 8 is defined as
8 = l2/(a(w-) (36)
Since the magnetic field must remain finite far from the
current sheet, A I must be zero.. The magnetic field is also
continuous across the x =0 boundary because there is no
surface current, so that the solution is
H,(x, t) = Re [-Ko e-(I+i)xa elo ]

= -Ko cos (wt -x/8) e 1, x-0 (37)
where the magnetic field in the gap is uniform, determined

by the discontinuity in tangential H at x = -d to be H, = -K,
for -d <x - 0 since within the perfect conductor (x <-d)H =
0. The magnetic field diffuses into the conductor as a strongly
damped propagating wave with characteristic penetration
depth 8. The skin depth 8 is also equal to the propagating
wavelength, as drawn in Figure 6-27b. The current density
within the conductor
JfX=VxH=
ax
i,
= +K " [sin (wt -- cos (Wt -)i, (38)
is also drawn in Figure 6-27b at various times in the cycle,

being confined near the interface to a depth on the order of 8.
For a perfect conductor, 8 -*0, and the volume current
becomes a surface current.
Seawater has a conductivity of -4 siemens/m so that at a
frequency of f =1 MHz (w = 2-rf) the skin depth is 8
0.25 m. This is why radio communications to submarines are
difficult. The conductivity of copper is cr 6 x 107 siemens/m
so that at 60 Hz the skin depth is8 - 8 mm. Power cables with
larger radii have most of the current confined near the sur
face so that the center core carries very little current. This
reduces the cross-sectional area through which the current

flows, raising the cable resistance leading to larger power
dissipation.
Again, the magnetization force density has no contribution
to the force density since H only depends on x:
F = so(M - V)H + lioJf x H

= o(V XH) x H
= -V(-2stoH - H) (39)
The total force per unit area on the slab obtained by

integrating (39) over x depends only on the magnetic field at
x = 0:
f(L= -- H ) dx
1 2 2
= ijoK0 cos Wt (40)
because again H is independent of y and z and the x

component of the force density of (39) was written as a pure
derivative with respect to x. Note that this approach was easier
than integrating the cross product of (38) with (37).
This force can be used to levitate the conductor. Note that
the region for x >8 is dead weight, as it contributes very little
to the magnetic force.
6-4-5 Effects of Convection
A distributed dc surface current -Koi, at x =0 flows along

parallel electrodes and returns via a conducting fluid moving
to the right with constant velocity voi., as shown in Figure
6-28a. The flow is not impeded by the current source at x = 0.
With the neglect of fringing, the magnetic field is purely z
directed and only depends on the x coordinate, so that (13) in
the dc steady state, with U = voi. being a constant, becomes*
d 2 H. dH(.
T -Avo---= 0 (41)
Solutions of the form

Hx(x) =A eo" (42)
) - H(V/)+
x =(V / - (U - V)H = dH
*V x (U x H) = U (V/)- /1) - (]V ) ( ) =-od
----------
-
D
-0-- --------- -------
0 I
x
(a)
H, W) Ko (eRm xI-eRM) (x) Ko Rm Rmxl

A I
1-e R 1-e R.
5
1.0 2.01
5
H, (x) ill (X)
KOl
0.5- 1.0 0
R. =0 Rm
0.5 1.0 0.5 1.0

x/lI x/I
(b)
Figure 6-28 (a) A conducting material moving through a magnetic field tends to pull
the magnetic field and current density with it. (b) The magnetic field and current
density are greatly disturbed by the flow when the magnetic Reynolds number is large,
R. = oyUI > 1.
when substituted back into (41) yield two allowed values of p,

P-- 2rvop=0:P=0, P= pOvo (43)
Since (41) is linear, the most general solution is just the sum
of the two allowed solutions,
H,(x )= A I e R-X +A2 (44)
where the magnetic Reynold's number is defined as
R.= ovol 2(45)
and represents the ratio of a representative magnetic

diffusion time given by (28) to a fluid transport time (1/vo).
The boundary conditions are
H,(x =0)=Ko, H.(x=L)=0 (46)

so that the solution is
H.x=Ko
H.() KeR
0
R_1
(e~n/ eR ~ (47)
1-e
The associated current distribution is then
Jt=VxH= H.i,
KoR - .,
_e
(48)
1-e - eI (48
The field and current distributions plotted in Figure 6-28b
for various R., show that the magnetic field and current are
pulled along in the direction of flow. For small R_ the
magnetic field is hardly disturbed from the zero flow solution
of a linear field and constant current distribution. For very
large R. >> 1, the magnetic field approaches a uniform dis
tribution while the current density approaches a surface cur
rent at x = 1.
The force on the moving fluid is independent of the flow
velocity:
f JJx pHsDdx
K 1,
R____ R.XR~x11 m~
laiIAOIsDfeR-xl R d_ i
(1-e _)2
K20osD R_cIL (eR32 10
-1 2
.
= ioKosDi. (49)
6-4-6 A Linear Induction Machine
The induced currents in a conductor due to a time varying

magnetic field give rise to a force that can cause the conductor
to move. This describes a motor. The inverse effect is when
we cause a conductor to move through a time varying
magnetic field generating a current, which describes a

generator.
The linear induction machine shown in Figure 6-29a
assumes a conductor moves to the right at constant velocity
Ui. Directly below the conductor with no gap is a surface
current placed on top of an infinitely permeable medium
K(t) = -K 0 cos (wt - kz)i, = Re [-K0 e !(Wk)i,] (50)
which is a traveling wave moving to the right at speed wk.
For x >0, the magnetic field will then have x and z components
of the form
H.(x, z, t) = Re [A. (x) e'("" )]
H.(x, z, t) = Re [A.(x) e'(" t k)]
H,
a, 4: U
x
Y ' kz)
(a)
1.01
<fz>
0.5
j- oKo2
I

3 6
S -o"- 2
(w -kU)
k
(b)
Figure 6-29 (a) A traveling wave of surface current induces currents in a conductor
that is moving at a velocity U different from the wave speed a>k. (b) The resulting
forces can levitate and propel the conductor as a function of the slip S, which measures
the difference in speeds of the conductor and traveling wave.
448 Electromagneticinduction
where (10) (V - B= 0) requires these components to be related

as
--- jk,=0 (52)

dx
The z component of the magnetic diffusion equation of
(13) is
-k 2 =jo-(w -AkU)H (53)
which can also be written as
- . 2fz = 0 (54)
where
2
= k 2 (l+jS), S (W-kU) (55)
and S is known as the slip. Solutions of (54) are again

exponential but complex because y is complex:
A = A i e'+A 2 e~' (56)
Because H. must remain finite far from the current sheet,

A1 =0, so that using (52) the magnetic field is of the form
H=Koe-" -k (57)
where we use the fact that the tangential component of H is discon

tinuous in the surface current, with H = 0 for x<0.
The current density in the conductor is
=V x H= ( 8 ) =-jkH
2 2
(Y -k
)
-Kok2e-7
Kokej--w (58)
If the conductor and current wave travel at the same speed

(w/k = U), no current is induced as the slip is zero. Currents
are only induced if the conductor and wave travel at different
velocities. This is the principle of all induction machines.
The force per unit area on the-conductor then has x and z

components:
f= IJ oHdx
= jtoJ,(H.i.- H.i.) dx (59)
These integrations are straightforward but lengthy because

first the instantaneous field and current density must be
found from (51) by taking the real parts. More important is
the time-average force per unit area over a period of excita
tion:
2w/w
<f>=f- fdt (60)
2v o
Since the real part of a complex quantity is equal to half the
sum of the quantity and its complex conjugate,
A = Re [A e "'] = 2(A e "+
t
A* ei"') (61)
B = e[ -]=-L, -+A e-i")
the time-average product of two quantities is
ABdte
2wr 4 21r
+A*A* e-1ivt) dt
=:(A *+ A*)
= IRe (AE*) (62)
which is a formula often used for the time-average power in
circuits where A and B are the voltage and current.
Then using (62) in (59), the x component of the time-
average force per unit area is
<f.>=4Re fLoffk*dx)
=2 KokS Re ( e+* dx
- 2Kok S R
\Y(Y +-y*)/
22
2 -1+S 2 -I\
=-I soK
zoosS =;1 2 ,so 1 (63)
4 [1 +S2+(l +S2)12 I1S
where the last equalities were evaluated in terms of the slip S

from (55).
We similarly compute the time-average shear force per unit

area as
< f,>= -- Re 00yojH* dx)
5 e
=po K k
2 yy* (I
2=- y Re )cd)
-2 * *Re (Re *)
2oKS
= jiT? (64)
4M +S2 Re (,[I+ jS)
When the wave speed exceeds the conductor's speed (w/k >
U), the force is positive as S >0 so that the wave pulls the
conductor along. When S <0, the slow wave tends to pull the
conductor back as <f.> <0. The forces of (63) and (64),
plotted in Figure 6-29b, can be used to simultaneously lift and
propel a conducting material. There is no force when the
wave and conductor travel at the same speed (w/k = U) as the
slip is zero (S = 0). For large S, the levitating force <f.>
approaches the constant value IiyoKo while the shear force
approaches zero. There is an optimum value of S that maxi
mizes <f,>. For smaller S, less current is induced while for
larger S the phase difference between the imposed and
induced currents tend to decrease the time-average force.
6-4-7 Superconductors
In the limit of infinite Ohmic conductivity (o-->-oo), the

diffusion time constant of (28) becomes infinite while the skin
depth of (36) becomes zero. The magnetic field cannot
penetrate a perfect conductor and currents are completely
confined to the surface.
However, in this limit the Ohmic conduction law is no
longer valid and we should use the superconducting consti
tutive law developed in Section 3-2-2d for a single charge
carrier:
aj= w 2E (65)
at
Then for a stationary medium, following the same pro
cedure as in (12) and (13) with the constitutive law of (65),
(8)-(1 1) reduce to
V2a 2 H =0=>V2(-Ho) Ey(-Ho)=O

t 6)t
(66)
Energy Stored in the Magnetic Field 451
where Ho is the instantaneous magnetic field at t =0. If the

superconducting material has no initial magnetic field when
an excitation is first turned on, then Ho =0.
If the conducting slab in Figure 6-27a becomes super-
conducting, (66) becomes
d 2H. w2
-- H =0, c = (67)
where c is the speed of light in the medium.

The solution to (67) is
H = A 1 e'c +A 2 e'**
=-Ko cos wt e ,"'c (68)
where we use the boundary condition of continuity of
tangential H at x = 0.
The current density is then
ax
= Kow, cos wt e
(69)
For any frequency w, including dc (w = 0), the field and

current decay with characteristic length:
Ic= c/W, (70)
Since the plasma frequency w, is typically on the order of
1 0 15 radian/sec, this characteristic length is very small, 1,

3 x 10/10' 5 -3x 10- m. Except for this thin sheath, the
magnetic field is excluded from the superconductor while the
volume current is confined to this region near the interface.
There is one experimental exception to the governing
equation in (66), known as the Meissner effect. If an ordinary
conductor is placed within a dc magnetic field Ho and then
cooled through the transition temperature for superconduc
tivity, the magnetic flux is pushed out except for a thin sheath
of width given by (70). This is contrary to (66), which allows
the time-independent solution H = Ho, where the magnetic
field remains trapped within the superconductor. Although
the reason is not well understood, superconductors behave as
if Ho =0 no matter what the initial value of magnetic field.
6-5 ENERGY STORED IN THE MAGNETIC FIELD

6-5-1 A Single Current Loop
The differential amount of work necessary to overcome
the electric and magnetic forces on a charge q moving an
incremental distance ds at velocity v is

dW, = -q(E+v X B) -ds (1)
(a) Electrical Work

If the charge moves solely under the action of the electrical
and magnetic forces with no other forces of mechanical ori
gin, the incremental displacement in a small time dt is related
to its velocity as
ds = v dt (2)
Then the magnetic field cannot contribute to any work on the
charge because the magnetic force is perpendicular to the
charge's displacement:
dW, = -qv - E dt (3)
and the work required is entirely due to the electric field.

Within a charge neutral wire, the electric field is not due to
Coulombic forces but rather arises from Faraday's law. The
moving charge constitutes an incremental current element,
qv= idl>dW,=-iE -d dt (4)
so that the total work necessary to move all the charges in the
closed wire is just the sum of the work done on each current
element,
dw= fdW, =-i dt fE -dl
d
=idt- B-dS
dt s
= i dt
dt
=idD (5)
which through Faraday's law is proportional to the change of
flux through the current loop. This flux may be due to other
currents and magnets (mutual flux) as well as the self-flux due
to the current i. Note that the third relation in (5) is just
equivalent to the circuit definition of electrical power
delivered to the loop:
P dW d=V (6)
dt dt
All of this energy supplied to accelerate the charges in the
wire is stored as no energy is dissipated in the lossless loop
and no mechanical work is performed if the loop is held
stationary.
(b) Mechanical Work

The magnetic field contributed no work in accelerating the
charges. This is not true when the current-carrying wire is
itself moved a small vector displacement ds requiring us to
perform mechanical work,
dW= -(idXB) -ds=i(BXdl) -ds
= iB - (dxds) (7)
where we were able to interchange the dot and the cross using
the scalar triple product identity proved in Problem 1-10a.
We define S, as the area originally bounding the loop and S 2
as the bounding area after the loop has moved the distance
ds, as shown in Figure 6-30. The incremental area dS3 is then
the strip joining the two positions of the loop defined by the
bracketed quantity in (7):
dS3 = dl x ds (8)
The flux through each of the contours is
I= IB - dS, ( 2 =1 B -dS (9)
where their difference is just the flux that passes outward

through dSs:
dF=4 1-( 2 =B-dS3 (10)
The incremental mechanical work of (7)-necessary to move
the loop is then identical to (5):
dW= iB - dS 3 =id4 (11)
Here there was no change of electrical energy input, with
the increase of stored energy due entirely to mechanical work
in moving the current loop.
S2
ds
dS3 di x ds
d \
idi
fB
Figure 6-30 The mechanical work necessary to move a current-carrying loop is
stored as potential energy in the magnetic field.
6-5-2 Energy and Inductance
If the loop is isolated and is within a linear permeable

material, the flux is due entirely to the current, related
through the self-inductance of the loop as
D= Li (12)
so that (5) or (11) can be integrated to find the total energy in
a loop with final values of current I and flux (D:
W= j i d4
= -dQ
1 D2 1
=---=-_L12I-I (13)
2L 2 2
6.5-3 Current Distributions
The results of (13) are only true for a single current loop.
For many interacting current loops or for current dis
tributions, it is convenient to write the flux in terms of the
vector potential using Stokes' theorem:
= sB -dS= s(V x A) - dS A - dl (14)
Then each incremental-sized current element carrying a

current I with flux d(l has stored energy given by (13):
dW =2I d(l=I - A di (15)
For N current elements, (15) generalizes to
W=2(I -A 1 dl 1+I2 -A 2 dl 2 +- +IN *AN d N)
N
I.- A.
din (16) = n=1 ,
If the current is distributed over a line, surface, or volume,
the summation is replaced by integration:
j If -A dl (line current)
W=- sKf A dS (surface current) (17)
J, A dV (volume current)
Remember that in (16) and (17) the currents and vector

potentials are all evaluated at their final values as opposed to
(11), where the current must be expressed as a function of
flux.
6-5-4 Magnetic Energy Density
This stored energy can be thought of as being stored in the

magnetic field. Assuming that we have a free volume dis
tribution of current Jf, we use (17) with Ampere's law to
express Jf in terms of H,
W=f Jf-AdV=f (VXH)-AdV (18)
where the volume V is just the volume occupied by the

current. Larger volumes (including all space) can be used in
(18), for the region outside the current has Jf = 0 so that no
additional contributions arise.
Using the vector identity
V - (A x H)= H - (V x A) -A - (V x H)
=H - B-A - (V X H) (19)
we rewrite (18) as
W=I v[H - B-V - (A x H)] dV (20)
The second term on the right-hand side can be converted

to a surface integral using the divergence theorem:
(A xH) -dS (21)

V - (A x H) dV =
It now becomes convenient to let the volume extend over all
space so that the surface is .at infinity. If the current dis
tribution does not extend to infinity the vector potential dies
off at least as I/r and the magnetic field as I/r Then, even
though the area increases as r2 , the surface integral in (21)
decreases at least as I/r and thus is zero when S is at infinity.
Then (20) becomes simply
W=1jH -BdV=Ij LH2dV= -dV (22)
where the volume V now extends over all space. The

magnetic energy density is thus
1B2
2 - (23)
w=HB=iH A
These results are only true for linear materials where yA

does not depend on the magnetic field, although it can
depend on position.
For a single coil, the total energy in (22) must be identical to
(13), which gives us an alternate method to calculating the
self-inductance from the magnetic field.
6-5-5 The Coaxial Cable
(a) External Inductance

A typical cable geometry consists of two perfectly conduct
ing cylindrical shells of radii a and b and length 1, as shown in
Figure 6-31. An imposed current I flows axially as a surface
current in opposite directions on each cylinder. We neglect
fringing field effects near the ends so that the magnetic field is
the same as if the cylinder were infinitely long. Using
Ampere's law we find that
I
H4=--, a<r<b (24)
2ir
The total magnetic flux between the two conductors is
(D f goH.ldr
I.oIL b
= In a (25)
2w a
HO 0
2irb
KK, a "H,(r)
Depth I H =
a b r
Figure 6-31 The magnetic field between two current-carrying cylindrical shells
forming a coaxial cable is confined to the region between cylinders.
giving the self-inductance as

) A1o b
L =- =- In - (26)
I 2,7r a
The same result can just as easily be found by computing
the energy stored in the magnetic field
W LI2 =jio H21Trldr
p4 ln b 2W A1 In (b/a)
= AOI n-=>L- 2 -(27)
47r a I 27r
(b) Internal Inductance

If the inner cylinder is now solid, as in Figure 6-32, the
current at low enough frequencies where the skin depth is
much larger than the radius, is uniformly distributed with
density
J= - (28)
Ta
so that a linearly increasing magnetic field is present within
the inner cylinder while the outside magnetic field is
* Depth
X Xdi . 21r dr
ra2r
** . 7aj/ * J
>
2HO
ab
_.1dr H, 21rr
. K a 0
Figure 6-32 At low frequencies the current in a coaxial cable is uniformly distributed
over the solid center conductor so that the internal magnetic field increases linearly with
radius. The external magnetic field remains unchanged. The inner cylinder can be
thought of as many incremental cylindrical shells of thickness dr carrying a fraction of
the total current. Each shell links its own self-flux-as well as the mutual flux of the other
shells of smaller radius. The additional flux within the current-carrying conductor
results in the internal inductance of the cable.
unchanged from (24):
Ir
2
27ra , O<r<a
H, = (29)
r',O a<r<b
2,7r r
The self-inductance cannot be found using the flux per

unit current definition for a current loop since the current is
not restricted to a thin filament. The inner cylinder can be
thought of as many incremental cylindrical shells, as in Figure
6-32, each linking its own self-flux as well as the mutual flux
of the other shells of smaller radius. Note that each shell is at
a different voltage due to the differences in enclosed flux,
although the terminal wires that are in a region where the
magnetic field is negligible have a well-defined unique voltage
difference.
The easiest way to compute the self-inductance as seen by
the terminal wires is to use the energy definition of (22):
W= jyo I H2rlrdr
= rl 0[ 2 1r dr+ I r dr
2o(7ra2 a 271rr
A'jI
2
1 b
= -+ln - (30)
4,7r 4 a)
which gives the self-inductance as
2W Ayol b
L= -= -+ln- (31)
I 2ir\4 aI
The additional contribution of g~ol/87r is called the internal

inductance and is due to the flux within the current-carrying
conductor.
6-5-6 Self-Inductance, Capacitance, and Resistance
We can often save ourselves further calculations for the

external self-inductance if we already know the capacitance or
resistance for the same two-dimensional geometry composed
of highly conducting electrodes with no internal inductance
contribution. For the arbitrary geometry shown in Figure
6-33 of depth d, the capacitance, resistance, and inductance
are defined as the ratios of line and surface integrals:
ed is E -n, ds
IL E - dl
R j1 E -di (2
~ fs E -nd(32)
cR
o-d n, ds
L d JL H - ni di
Ls H - ds
Because the homogeneous region between electrodes is
charge and current free, both the electric and magnetic fields
can be derived from a scalar potential that satisfies Laplace's
equation. However, the electric field must be incident
normally onto the electrodes while the magnetic field is
incident tangentially so that E and H are perpendicular
everywhere, each being along the potential lines of the other.
This is accounted for in (32) and Figure 6-33 by having n, ds
perpendicular to ds and ni dl perpendicular to dl. Then since
C, R, and L are independent of the field strengths, we can
take E and H to both have unit magnitude so that in the
products of LC and LIR the line and surface integrals cancel:
LC etd2 = d 2/c 2 , c 1(E3
2 2 (33)
LIR=wo-d , RC=elo-
These products are then independent of the electrode
geometry and depend only on the material parameters and
the depth of the electrodes.
We recognize the LIR ratio to be proportional to the
magnetic diffusion time of Section 6-4-3 while RC is just the
charge relaxation time of Section 3-6-1. In Chapter 8 we see
that the NIC product is just equal to the time it takes an
electromagnetic wave to propagate a distance d at the speed
of light c in the medium.
E /
-----
L---.
L dl Q
-Q \
-ds
ns Depth d
Figure C-33 The electric and magnetic fields in the two-dimensional homogeneous
charge and current-free region between hollow electrodes can be derived from a scalar
potential that obeys Laplace's equation. The electric field lines are along the magnetic
potential lines and vice versa so E and H are perpendicular. The inductance-capaci
tance product is then a constant.
460 ElecromagneticInduction
6-6 THE ENERGY METHOD FOR FORCES
6-6-1 The Principle of Virtual Work
In Section 6-5-1 we calculated the energy stored in a

current-carrying loop by two methods. First we calculated the
electric energy input to a loop with no mechanical work done.
We then obtained the same answer by computing the
mechanical work necessary to move a current-carrying loop
in an external field with no further electrical inputs. In the
most general case, an input of electrical energy can result in
stored energy dW and mechanical work by the action of a
force f. causing a small displacement dx:
idO=dW+f.dx (1)
If we knew the total energy stored in the magnetic field as a
function of flux and position, the force is simply found as
8W
f.= a (2)
We can easily compute the stored energy by realizing that

no matter by what process or order the system is assembled, if
the final position x and flux 0 are the same, the energy is the
same. Since the energy stored is independent of the order
that we apply mechanical and electrical inputs, we choose to
mechanically assemble a system first to its final position x with
no electrical excitations so that 0 =0. This takes no work as
with zero flux there is no force of electrical origin. Once the
system is mechanically assembled so that its position remains
constant, we apply the electrical excitation to bring the system
to its final flux value. The electrical energy required is
W=J idO (3)

x const
For linear materials, the flux and current are linearly
related through the inductance that can now be a function of
x because the inductance depends on the geometry:
-i =/L(x) (4)
Using (4) in (3) allows us to take the inductance outside the
integral because x is held constant so that the inductance is
also constant:
W= 0 dO
L((x)
= =c L(x)i)
2L(x)2(5
The Energy Method for Forces 461
The stored energy is the same as found in Section 6-5-2 even

when mechanical work is included and the inductance varies
with position.
To find the force on the moveable member, we use (2) with
the energy expression in (5), which depends only on flux and
position:
aW
8x
.
,t2 d [ IL (x)]
2 dx
1 dL(x) D2
2
2 L (x) dx
1i 2 dL(x)
dx
6-6-2 Circuit Viewpoint
This result can also be obtained using a circuit description

with the linear flux-current relation of (4):
di
V=
dt
di .dL(x)
=x+ dt
di .dL(x)dx (7)
=x+ dx d7
The last term, proportional to the speed of the moveable
member, just adds to the usual inductive voltage term. If the
geometry is fixed and does not change with time, there is no
electromechanical coupling term.
The power delivered to the system is
d
which can be expanded as

d 1 .2 2 dL(x)dx
d (9) P=-(iL(x)i
di )+Ai
)+id di
This is in the form
dW dx W=2L(x)i 2 (10)
dt dt f.=-1
dL(x)
p= +2
which states that the power delivered to the inductor is equal

to the sum of the time rate of energy. stored and mechanical
power performed on the inductor. This agrees with the
energy method approach. If the inductance does not change
with time because the geometry is fixed, all the input power is
stored. as potential energy W.
Example 6-2 MAGNETIC FIELDS AND FORCES
(a) Relay
Find the force on the moveable slug in the magnetic circuit
shown in Figure 6-34.
SOLUTION
It is necessary to find the inductance of the system as a

function of the slug's position so'that we can use (6). Because
of the infinitely permeable core and slug, the H field is non
zero only in the air gap of length x. We use Ampere's law to
obtain
H = NI/x
The flux through the gap
(= yoNIA/x
is equal to the flux through each turn of the coil yielding the
inductance as
NO yAoN 2 A
L(x)=-=
Ix
N turns H=N
Mo Cross-sectional area A
-*-I
Figure 6-34 The magnetic field exerts a force on the moveable member in the relay
pulling it into the magnetic circuit.
The Energy Method for Forces 463
The force is then

f 2 dL(x)
dx2
2C
i.L0N 2A12
2
2
The minus sign means that the force is opposite to the direc
tion of increasing x, so that the moveable piece is attracted to
the coil.
(b) One Turn Loop

Find the force on the moveable upper plate in the one turn
loop shown in Figure 6-35.
SOLUTION
The current distributes itself uniformly as a surface current

K = I/D on the moveable plate. If we neglect nonuniform
field effects near the corners, the H field being tangent to the
conductors just equals K:
H = IID
The total flux linked by the current source is then
(D= oHxl
= I
D
which gives the inductance as
x pDoxl
L(x)=-=
I D
K IID
Figure 6-35 the magnetic force on a current-carrying ioop tends to expand the ioop.
The force is then constant
f=,l1dL(x)
dx
1 olI 2
2 D
6-6-3 Magnetization Force
A material with permeability A is partially inserted into the

magnetic circuit shown in Figure 6-36. With no free current
in the moveable material, the x-directed force density from
Section 5-8-1 is
F. = Mo(M - V)H.
= (; - iMo)(H - V)H.
aHH
-'UO) H. -+ H,
=(;~IL~o)\axY ay11(11)
) 11
where we neglect variations with z. This force arises in the

fringing field because within the gap the magnetic field is
essentially uniform:
H, = NI/s (12)
Because the magnetic field in the permeable block is curl free,
V xH =>O--- 8 (13)
0
p.y x
-L a
t= NI N y
xo
Depth D
(a)
Figure 6-36 - A permeable material tends to be pulled into regions of higher magnetic
field.
Probtems 465
(11) can be rewritten as
F.= - o)(H+ H') (14)

2 8x
The total force is then
f =sD F.dx
=(I - I)
1
N I2D
&)NID(15)
2 S
where the fields at x = -oo are zero and the field at x= x 0 is

given by (12). High permeability. material is attracted to
regions of stronger magnetic field. It is this force that causes
iron materials to be attracted towards a magnet. Diamagnetic
materials (A <po) will be repelled.
This same result can more easily be obtained using (6)
where the flux through the gap is
NID
4D= HD[ipx +p o(a -x)] ---- [(p - po)x+ao] (16)
S
so that the inductance is
NO N 2D
L=-= -[( -po)x +aizo] (17)
I S
Then the force obtained using (6) agrees with (15)
f 2 dL(x)
dx
=A2s A N 2,2D (18)

2s
PROBLEMS
Section 6-1
1. A circular loop of radius a with Ohmic conductivity a- and
cross-sectional area A has its center a small distance D away
from an infinitely long time varying current.
I~t) aCross-sectional area A
a i
-- D
(a) Find the mutual inductance M and resistance R of the

loop. Hint:
dx 2 tan.., [VCb" tan (x/ 2
)
Sa+b cosx - L a+b
rdr =
73D-r"
(b) This loop is stationary. and has a self-inductance L.

What is the time dependence of the induced short circuit
current when the line current is instantaneously stepped on
to a dc level I at t = 0?
(c) Repeat (b) when the line current has been on a long
time and is suddenly turned off at t = T.
(d) If the loop has no resistance and is moving with radial
velocity v, = dr/dt, what is the short circuit current and open
circuit voltage for a dc line current?
(e) What is the force on the loop when it carries a current
i? Hint:
[ cos 0 do
J D+acos
= -- sin [cos]
D ._,a +Dcos4
+ smi
aDT-a D+a cos/
2. A rectangular loop at the far left travels with constant

velocity Ui. towards and through a dc surface current sheet
Koi, at x =0. The right-hand edge of the loop first reaches
the current sheet at t =0.
(a) What is the loop's open circuit voltage as a function of
time?
(b) What is the short circuit current if the loop has self-
inductance L and resistance R?
(c) Find the open circuit voltage if the surface current is
replaced by a fluid with uniformly distributed volume cur
rent. The current is undisturbed as the loop passes through.
Problems 467
+V -.
Ut -b
a) (ca
Specifically consider the case when d > b and then sketch the
results when d= b and d <b. The right edge of the current
loop reaches the volume current at t =0.
3. A short circuited rectangular loop of mass m and self-
inductance L is dropped with initial velocity voi. between the
pole faces of a magnet that has a concentrated uniform
magnetic field Boi. Neglect gravity.
0 B0
SN S
X V0
(a) What is the imposed flux through the loop as a function

of the loop's position x (0< x <s) within the magnet?
(b) If the wire has conductivity a- and cross-sectional area
A, what equation relates the induced current i in the loop and
the loop's velocity?
(c) What is the force on the loop in terms of i? Obtain a
single equation for the loop's velocity. (Hint: Let Wo =
Bob 2/mL, a= RIL.)
(d) How does the loop's velocity and induced current vary
with time?
(e) If a--+ c, what minimum initial velocity is necessary for
the loop to pass through the magnetic field?
4. Find the mutual inductance between the following cur
rents:
(a) Toroidal coil of rectangular or circular cross section
''I "I
N turns
S
u rns
R
51IN a-b
cross--section
Toroid
cross-section
(a)
I I
b
-7
-d a

(b)
coaxially centered about an infinitely long line current. Hint:

dx 2 {Ia"-b?
t-a -1 tan(x/2 )}
f a+bcosx a+b'
r dr =--R
-
(b) A very long rectangular current loop, considered as two

infinitely long parallel line currents, a distance D apart, car
rying the same current I in opposite directions near a small
rectangular loop of width a, which is a distance d away from
the left line current. Consider the cases d + a <D, d <D <
d+a, and d>D.
5. A circular loop of radius a is a distance D above a point
magnetic dipole of area dS carrying a current Ii.
C:'2
D
4 m = I 1 dS
I1
Problems 469
(a) What is the vector potential due to the dipole at all

points on the circular loop? (Hint: See Section 5-5-1.)
(b) How much flux of the dipole passes through the circu
lar loop?
(c) What is the mutual inductance between the dipole and
the loop?
(d) If the loop carries a current 12, what is the magnetic
field due to 12 at the position of the point dipole? (Hint: See
Section 5-2-4a.)
(e) How much flux due to I2 passes through the magnetic
dipole?
(f) What is the mutual inductance? Does your result agree
with (c)?
6. A small rectangular loop with self-inductance L, Ohmic
conductivity a-, and cross-sectional area A straddles a current
sheet.
,K(t)
nK(t~i,
T
(a) The current sheet is instantaneously turned on to a dc

level Koi, at t =0. What is the induced loop current?
(b) After a long time T the sheet current is instantaneously
set to zero. What is the induced loop current?
(c) What is the induced loop current if the current sheet
varies sinusoidally with time as Ko cos wti,.
7. A point magnetic dipole with area dS lies a distance d
below a perfectly conducting plane of infinite extent. The
dipole current I is instantaneously turned on at t =0.
(a) Using the method of images, find the magnetic field
everywhere along the conducting plane. (Hint: i, - ir = sin 0,
dA = wa2
470 ELectomagnetic Induction
is - ir = Cos 9.)
(b) What is the surface current distribution?
(c) What is the force on the plane? Hint:
J r3 dr 2
(r + d/4)
2
(r +d 2)5 6(r2 +d 2 )4
(d) If the plane has a mass M in the gravity field g, what

current I is necessarl to just lift the conductor? Evaluate for
M= 10-3 kg, d= 10~ m, and a = 10- m.
8. A thin block with Ohmic conductivity o- and thickness 8
moves with constant velocity Vi, between short circuited
perfectly conducting parallel plates. An initial surface current
Ko is imposed at t =0 when x = xo, but the source is then
removed.
-K(t)
x Depth D
(a) The surface current on the plates K(t) will vary with
time. What is the magnetic field in terms of K(t)? Neglect
fringing effects.
(b) Because the moving block is so thin, the current is
uniformly distributed over the thickness 8. Using Faraday's
law, find K(t) as a function of time.
(c) What value of velocity will just keep the magnetic field
constant with time until the moving block reaches the end?
(d) What happens to the magnetic field for larger and
smaller velocities?
9. A thin circular disk of radius a, thickness d, and conduc
tivity a- is placed in a uniform time varying magnetic field
B(t).
(a) Neglecting the magnetic field of the eddy currents,
what is the current induced in a thin circular filament at
radius r of thickness dr.
Problems 471
B(t)
a..
d(a
(a) (d)
(b) What power is dissipated in this incremental current

loop?
(c) How much power is dissipated in the whole disk?
(d) If the disk is instead cut up into N smaller circular disks
with negligible wastage, what is the approximate radius of
each smaller disk?
(e) If these N smaller disks are laminated together to form
a thin disk of closely packed cylindrical wires, what is the
power dissipated?
Section 6-2
10. Find the self-inductance of an N turn toroidal coil of
circular cross-sectional radius a and mean radius b. Hint:
d 2 tn - b2- r2 tan (0/2)

f b+r cos 0 bT--rT b+r
frdr = - 2
N turns
11. A large solenoidal coil of long length l, radius a,, and

number of turns N, coaxially surrounds a smaller coil of long
length 12, radius a 2 , and turns N 2
.
+ V1
N, turns
_L 21A
a2 J _W~V 2
a2
a1
ZN2 turns
(a) Neglecting fringing field effects find the self-

inductances and mutual inductances of each coil.
(Hint: Assume the magnetic field is essentially uniform
within the cylinders.)
(b) What is the voltage across each coil in terms of iI and i 2 ?
(c) If the coils are connected in series so that ii = i 2 with the
fluxes of each coil in the same direction, what is the total
self-inductance?
(d) Repeat (c) if the series connection is reversed so that
iI= -i 2 and the fluxes due to each coil are in opposite direc
tions.
(e) What is the total self-inductance if the coils are
connected in parallel so that v1 = v 2 or v 1 = -V2?
12. The iron core shown with infinite permeability has three
gaps filled with different permeable materials.
(a) What is the equivalent magnetic circuit?
(b) Find the magnetic flux everywhere in terms of the gap
reluctances.
Nj<
<1 V1
- S3
y 4 a3
Depth D
Problems 473
(c) What is the total magnetic flux through each winding?

(d) What is the self-inductance and mutual inductance of
each winding?
13. A cylindrical shell of infinite permeability, length I and

inner radius b coaxially surrounds a solid cylinder also with
infinite permeability and length I but with smaller radius a so
that there is a small gap g = b - a. An N, turn coil carrying a
current I, is placed within two slots on the inner surface of
the outer cylinder.
Depth I
+-1
(a) What is the magnetic field everywhere? Neglect all

radial variations in the narrow air gap. (Hint: Separately
consider 0< 0 <nr and ir <k <27r.)
(b) What is the self-inductance of the coil?
(c) A second coil with N 2 turns carrying a current 12 is
placed in slots on the inner cylinder that is free to rotate.
When the rotor is at angle 0, what is the total magnetic field
due to currents I, and 12? (Hint: Separately consider 0<
(k <0, 0 < <7r, 7r < 4 <r +6, and ir+0<q<27r.)
(d) What is the self-inductance and mutual inductance of
coil 2 as a function of 6?
(e) What is the torque on the rotor coil?
14. (a) What is the ratio of terminal voltages and currents for
the odd twisted ideal transformer shown?
(b) A resistor RL is placed across the secondary winding
(v 2 , i 2 ). What is the impedance as seen by the primary
winding?
~1 i2
V1 N N2 C . V2
C+
15. An N turn coil is wound onto an infinitely permeable

magnetic core. An autotransformer is formed by connecting a
tap of N' turns.
V1 2N
+
RL V2 N'
(a) What are the terminal voltage (V 2 /Vi) and current (i2 /iI)
ratios?
(b) A load resistor RL is connected across the terminals
of the tap. What is the impedance as seen by the input
terminals?
Section 6-3
16. A conducting material with current density Jji. has two
species of charge carriers with respective mobilities u+ and ;s
and number densities n+ and n... A magnetic field Boi, is
imposed perpendicular to the current flow.
Problems 475
Bo i,
Vh
+-+
(a) What is the open circuit Hall voltage? (Hint: The

transverse current of each carrier must be zero.)
(b) What is the short circuit Hall current?
17. A highly conducting hollow iron cylinder with
permeability A and inner and outer radii R, and R 2 is
concentric to an infinitely long dc line current (adapted from
L. V. Bewley, Flux Linkages and Electromagnetic Induction.
Macmillan, New York, 1952, pp. 71-77).
dt
I
ty b b
a V 0
V0
R2
/10 R1
I d
LE
I
C
I
AO
d
/10
| |
(b) (d)
(a) What is the magnetic flux density everywhere? Find the

electromotive force (EMF) of the loop for each of the follow
ing cases.
476 ELecromagneticInduction
(b) A highly conducting circuit abcd is moving downward

with constant velocity Vo making contact with the surfaces of
the cylinders via sliding brushes. The circuit is completed
from c to d via the iron cylinder.
(c) Now the circuit remains stationary and the iron cylin
der moves upwards at velocity Vo.
(d) Now a thin axial slot is cut in the cylinder so that it can
slip by the complete circuit abcd, which remains stationary as
the cylinder moves upwards at speed Vo. The brushes are
removed and a highly conducting wire completes the c-d
path.
18. A very long permanently magnetized cylinder MoIL rotates on

a shaft at constant angular speed w. The inner and outer surfaces
at r = a and r = b are perfectly conducting, so that brushes can
make electrical contact.
a G
(a) If the cylinder is assumed very long compared to its

radius, what are the approximate values of B and H in the
magnet?
(b) What is the open circuit voltage?
(c) If the magnet has an Ohmic conductivity a-, what is the
equivalent circuit of this generator?
(d) What torque is required to turn the magnet if the
terminals are short circuited?
19. A single spoke wheel has a perfectly conducting cut rim.

The spoke has Ohmic conductivity a- and cross-sectional area
A. The wheel rotates at constant angular speed cvo in a
sinusoidally varying magnetic field B. = Bo cos at.
(a) What is the open circuit voltage and short circuit cur
rent?
(b) What is the equivalent circuit?
Problems 477
B, = Bocoswt
(t)
Ohmic conductivity y
cross-sectional area A
20. An MHD machine is placed within a magnetic circuit.
+a
Depth D
(a) A constant dc current if = Io is applied to the N turn

coil. How much power is delivered to the load resistor RL?
(b) The MHD machine and load resistor RL are now
connected in series with the N turn coil that has a resistance
R. No current is applied. For what minimum flow speed can
the MHD machine be self-excited?
21. The field winding of a homopolar generator is connected

in series with the rotor terminals through a capacitor C. The
rotor is turned at constant speed o.
(a) For what minimum value of rotor speed is the system
self-excited?
(b) For the self-excited condition of (a) what range of
values of C will result in dc self-excitation or in ac self-
excitation?
(c) What is the frequency for ac self-excitation?
II R
L1
C== Rr Lr
Section 6-4
22. An Ohmic block separates two perfectly conducting
parallel plates. A dc current that has been applied for a long
time is instantaneously turned off at t =0.
Depth D
t x
0 d
(a) What are the initial and final magnetic field dis
tributions? What are the boundary conditions?
(b) What are the transient magnetic field and current dis
tributions?
(c) What is the force on the block as a function of time?
23. A block of Ohmic material is placed within a magnetic
circuit. A step current Io is applied at t =0.
(a) What is the dc steady-state solution for the magnetic
field distribution?
(b) What are the boundary and initial conditions for the
magnetic field in the conducting block?
*(c) What are the transient field and current distributions?
(d) What is the time dependence of the force on the
conductor?
(e) The current has been on a long time so that the system
is in the dc steady state found in (a) when at t = T the current
Problems 479
A(t)
i(t) Depth D
o 4Na
T t ----- 40 X
is instantaneously turned off. What are the transient field and

current distributions in the conductor?
(f) If the applied coil current varies sinusoidally with time
as i(t) = Io cos wt, what are the sinusoidal steady-state field
and current distributions? (Hint: Leave your answer in
terms of complex amplitudes.)
(g) What is the force on the block?
24. A - semi-infinite conducting block is placed between
parallel perfect conductors. A sinusoidal current source is
applied.
Io COS~st
Itf PO S
Dph
Depth D
BOX
(a) What are the magnetic field and current distributions

within the conducting block?
(b) What is the total force on the block?
(c) Repeat (a) and (b) if the block has length d.
25. A current sheet that is turned on at t =0 lies a distance d
above a conductor of thickness D and conductivity a-. The
conductor lies on top of a perfectly conducting plane.
(a) What are the initial and steady-state solutions? What
are the boundary conditions?
(b) What are the transient magnetic field and current dis
tributions?
(c) After a long time T, so that the system has reached the
dc steady state, the surface current is set to zero. What are the
subsequent field and current distributions?
K(t)K
d ;O, a = 0
y
I, ~ ~ ~ a y..+:.
(d) What are the field and current distributions if the cur
rent sheet varies as Ko cos ot?
26. Distributed dc currents at x =0 and x = I flow through a

conducting fluid moving with constant velocity vOi,.
Ko- -- ___
X Depth D
(a) What are the magnetic field and current distributions?

(b) What is the force on the fluid?
27. A sinusoidal surface current at x =0 flows along parallel

electrodes and returns through a conducting fluid moving to
the right with constant velocity voi.. The flow is not impeded
by the current source. The system extends to x =o.
K0 -----------------
-
Depth D
x
(a) What are the magnetic field and current density dis
tributions?
(b) What is the time-average force on the fluid?
Problems 481
28. The surface current for the linear induction machine

treated in Section 6-4-6 is now put a distance s below the
conductor.
(a) What are the magnetic field and current distributions in
each region of space? (Hint: Check your answer with
Section 6-4-6 when s = 0.)
(b) Repeat (a) if s is set to zero but the conductor has a
finite thickness d.
29. A superconducting block with plasma frequency (o, is
placed within a magnetic circuit with exciting current
1Cos t.
M 11
N s
7 Depth D
(a) What are the magnetic field and current distributions in

the superconductor?
(b) What is the force on the block?
Section 6.5
30. Find the magnetic energy stored and the self-inductance
for the geometry below where the current in each shell is
uniformly distributed.
sw
d (i: St.n
P2, 0Depth I
31. Find the external self-inductance of the two wire lines

shown. (Hint: See Section 2-6-4c.)
482 ElectroagneticInduction
Depth I
RIR R 2
Depth I
32. A coaxial cable with solid inner conductor is excited by a

sinusoidally varying current Io cos wt at high enough
frequency so that the skin depth is small compared to the
radius a. The current is now nonuniformly distributed over
the inner conductor.
to Cos Wt
(a) Assuming that H = H,(r)is, what is the governing

equation for H#(r) within the inner cylinder. (Hint: V 2 H=
0
V(V, H) -V x (V x H).)
(b) Solve (a) for solutions of the form
H#(r) = Re [HI#(r) e"']
Hint: Bessel's equation is
d y dy 2
x 2 ;+x +(x -_ )y=0
with solutions
y = A iJ,(x)+A 2 Y,(x)
where Y, is singular at x = 0.
(c) What is the current distribution? Hint:
d 1
-[Ui(x)]
dx +-J(x)
x = Io(x)
Section 6-6
33. A reluctance motor is made by placing a high permeabil
ity material, which is free to rotate, in the air gap of a
magnetic circuit excited by a sinusoidal current Io cos Lot.
Problems 483
The inductance of the circuit varies as

L(O)=LO+L1 cos20
where the maximum inductance LO+L, occurs when 6 =0 or
6 = 1r and the minimum inductance LO--L 1 occurs when 6=
:Er/2.
(a) What is the torque on the slab as a function of the angle
6?
(b) The rotor is rotating at constant speed w, where 6=
wt + 8 and 8 is the angle of the rotor at t = 0. At what value of
w does the torque have a nonzero time average. The reluc
tance motor is then a synchronous machine. Hint:
cos2 wot sin 26 = [sin 20+cos 2ot sin 20]
= f{sin 26 + $[sin 2(wot + 6) + sin 2(0 - wot)]}
(c) What is the maximum torque that can be delivered by
the shaft and at what angle 5 does it occur?
34. A system of two coupled coils have the following flux-

current relations:
,i
I,
V 1
l2
(c)
=1=Li(0)ii+M(0)i
2
2 = M(0)ii+L2(0)i2
(a) What is the power p delivered to the coils?

(b) Write this power in the form
dW dO
p=-+T dt
dt
What are W and T?

(c) A small coil is free to rotate in the uniform magnetic
field produced by another coil. The flux-current relation is
(1=Li+MOi 2 sin 0
02 = Moi I sin 0+ L 2 i2
The coils are excited by dc currents I, and 12. What is the
torque on the small coil?
(d) If the small coil has conductivity o-, cross-sectional area
A, total length 1, and is short circuited, what differential
equation must the current il obey if 0 is a function-of time? A
dc current I2 is imposed in coil 2.
(e) The small coil has moment of inertia J. Consider only
small motions around 0 = 0 so that cos 0 1. With the torque
and current equations linearized, try exponential solutions of
the form e" and solve for the natural frequencies.
(f) The coil is released from rest at 0 = Oo. What is 0(t) and
il(t)? Under what conditions are the solutions oscillatory?
Damped?
35. A coaxial cable has its short circuited end free to move.
(a) What is the inductance of the cable as a function of x?

(b) What is the force on the end?
36. For the following geometries, find:
(a) The inductance;
(b) The force on the moveable member.
Problems 485
x1
s-.
Depth D
37. A coaxial cylinder is dipped into a magnetizable fluid with

permeability yA and mass density p.. How high h does the fluid
rise within the cylinder?
~ bt
4~
AO0
hjIA~
MIT OpenCourseWare
http://ocw.mit.edu
Resource: (OHFWURPDJQHWLF)LHOG7KHRU\$3UREOHP6ROYLQJ$SSURDFK
0DUNXV=DKQ
The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
the

MITRES 6 002S08 Part2 PDF

Uploaded by

Copyright:

Available Formats

MITRES 6 002S08 Part2 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

MITRES 6 002S08 Part2 PDF

Uploaded by

Copyright:

Available Formats

chapter 4

The electric field distribution due to external sources is

4-1 THE UNIQUENESS THEOREM

Consider a linear dielectric material where the permittivity

The special case of different constant permittivity media

which reduces to Poisson's equation in regions where e (r) is a

V (eVVi)=-p, V - (eVV2 )=-Pf (5)

so that we can determine what boundary conditions force

The difference of these two solutions VT = V, - V 2 obeys

.V.-(eVVVT)dV= eVTVVT-dS= JIVVTI2dV (8)

The volume integral is converted to a surface integral over

4-2 BOUNDARY VALUE PROBLEMS IN CARTESIAN GEOMETRIES

For most of the problems treated in Chapters 2 and 3 we

charge is present only on interfacial boundaries separating

4-2-1 Separation of Variables

Let us assume that within a region of space of constant

We try a solution that is a product of a function only of the x

V(x, y) = X(x) Y(y) (2)

This assumed solution is often convenient to use if the system

where the partial derivatives become total derivatives because

where k2 is called the separation constant and in general can

4-2-2 Zero Separation Constant Solutions

When the separation constant is zero (k2 =0) the solutions

where a,, b1 , cl, and dl are constants. The potential is given by

V= a2 +b 2x+c 2y+d 2xy (7)

(a) Hyperbolic Electrodes

boundary conditions are

(b) Resistor in an Open Box

Figure--- 4-- A- reitv eimpatal i= n pncnutn ox

Boundary Value Problems in CartesianGeometries 263

solutions given by (7) in each region enclosed by the elec­

V={al+bix+cy+dixy, o!ysd (12)

With the potential constrained on the electrodes and being

so that the constants in (12) are

The potential of (12) is then

VO( - x/1), 0:5y!5d (5

with associated electric field

VOix, 0:5 y:5 d

Note that in the dc steady state, the conservation of charge

electric field within the resistive medium must be purely

Fx= oyE.(y-d)wdx= w (18)

If the resistive material is liquid, this shear force can be used

4-2-3 Nonzero Separation Constant Solutions

Further solutions to (5) with nonzero separation constant

V = F1 sin ky e"* + F2 sin ky e~k' + F3 COS ky e"* + F4 COS ky e *x

* SeeJ. R. Melcher and G. I. Taylor, Electrohydrodynamics: A Review of the Role of

cosh x lim sinh x cosh x e

solution because Laplace's equation is linear. The values of

4-2-4 Spatialy Periodic Excitation

A sheet in the x =0 plane has the imposed periodic poten­

V = VO sin aye ax V Osin aye-O

E =- Voaeax[cosayi, + sin ayi,] E = Vo aeax (cos ayi y sin ayixl

>N/ / / / / field lines

we write the solution separately for positive and negative x as

V=sinaye ", x-0 (22)

E=VV - Voae"[cosayi,-sinayix] x >0

The surface charge density on the sheet is given by the dis­

The field lines drawn in Figure 4-4 obey the equation

4-2-5 Rectangular Harmonics

solutions given by (7) in each region enclosed by the elec

A sheet in the x =0 plane has the imposed periodic poten

The surface charge density on the sheet is given by the dis