CHAPTER 1
1.1. Given the vectors M = 10ax + 4ay 8az and N = 8ax + 7ay
a) a unit vector in the direction of M + 2N.
M + 2N = 10ax
4ay + 8az + 16ax + 14ay
Thus
a=
b) the magnitude of 5ax + N
(5, 0, 0) + (8, 7, 2)
c) |M||2N|(M + N):
2az , find:
4az = (26, 10, 4)
(26, 10, 4)
= (0.92, 0.36, 0.14)
|(26, 10, 4)|
3M:
( 30, 12, 24) = (43, 5, 22), and |(43, 5, 22)| = 48.6.
|( 10, 4, 8)||(16, 14, 4)|( 2, 11, 10) = (13.4)(21.6)( 2, 11, 10)
= ( 580.5, 3193, 2902)
1.2. Vector A extends from the origin to (1,2,3) and vector B from the origin to (2,3,-2).
a) Find the unit vector in the direction of (A
A
B = (ax + 2ay + 3az )
whose
magnitude is |A B| = [( ax
⌥
3 3 = 5.20. The unit vector is therefore
B): First
(2ax + 3ay
2az ) = ( ax
ay + 5az ) · ( ax
aAB = ( ax
ay + 5az )
1/2
ay + 5az )]
=
⌥
1 + 1 + 25 =
ay + 5az )/5.20
b) find the unit vector in the direction of the line extending from the origin to the midpoint of the
line joining the ends of A and B:
The midpoint is located at
Pmp = [1 + (2
1)/2, 2 + (3
2)/2, 3 + ( 2
3)/2)] = (1.5, 2.5, 0.5)
The unit vector is then
(1.5ax + 2.5ay + 0.5az )
amp = ↵
= (1.5ax + 2.5ay + 0.5az )/2.96
(1.5)2 + (2.5)2 + (0.5)2
1.3. The vector from the origin to the point A is given as (6, 2, 4), and the unit vector directed from
the origin toward point B is (2, 2, 1)/3. If points A and B are ten units apart, find the coordinates
of point B.
With A = (6, 2, 4) and B = 13 B(2, 2, 1), we use the fact that |B A| = 10, or
|(6 23 B)ax (2 23 B)ay (4 + 13 B)az | = 10
Expanding, obtain
36 8B + 49 B 2 + 4 83 B + 49 B 2 + 16 + 83 B + 19 B 2 = 100
⇧
176
or B 2 8B 44 = 0. Thus B = 8± 64
= 11.75 (taking positive option) and so
2
B=
2
(11.75)ax
3
2
1
(11.75)ay + (11.75)az = 7.83ax
3
3
1
7.83ay + 3.92az
1.4. A circle, centered at the origin with a radius of 2 units, lies in the xy plane. Determine ⌥
the unit
vector in rectangular components that lies in the xy plane, is tangent to the circle at (
3, 1, 0),
and is in the general direction of increasing values of y:
A unit vector tangent to this circle in the general increasing y direction is t = a⇤ . Its⌥x and
y components are tx = a⇤ · ax = sin ⇧, and ty = a⇤ ·⌥
ay = cos ⇧. At the point (
3, 1),
⇧ = 150⇤ , and so t = sin 150⇤ ax cos 150⇤ ay = 0.5(ax + 3ay ).
1.5. A vector field is specified as G = 24xyax + 12(x2 + 2)ay + 18z 2 az . Given two points, P (1, 2, 1)
and Q( 2, 1, 3), find:
a) G at P : G(1, 2, 1) = (48, 36, 18)
b) a unit vector in the direction of G at Q: G( 2, 1, 3) = ( 48, 72, 162), so
aG =
( 48, 72, 162)
= ( 0.26, 0.39, 0.88)
|( 48, 72, 162)|
c) a unit vector directed from Q toward P :
aQP =
P
|P
Q
(3, 1, 4)
= ⌥
= (0.59, 0.20, 0.78)
Q|
26
d) the equation of the surface on which |G| = 60: We write 60 = |(24xy, 12(x2 + 2), 18z 2 )|, or
10 = |(4xy, 2x2 + 4, 3z 2 )|, so the equation is
100 = 16x2 y 2 + 4x4 + 16x2 + 16 + 9z 4
1.6. Find the acute angle between the two vectors A = 2ax + ay + 3az and B = ax 3ay + 2az by using
the definition of:
⌥
⌥
a) the dot product:⌥First, A · B = ⌥
2 3 + 6 = 5 = AB cos ⇥, where A = 22 + 12 + 32 = 14,
and where B = 12 + 32 + 22 = 14. Therefore cos ⇥ = 5/14, so that ⇥ = 69.1⇤ .
b) the cross product: Begin with
⇧
⇧ ax
⇧
A ⇤ B = ⇧⇧ 2
⇧ 1
⇧
ay az ⇧⇧
1
3 ⇧⇧ = 11ax
3 2 ⇧
ay
7az
⌥
⌥
⌥
and then |A ⇤ B|
= 11⇥2 + 12 + 72 = 171. So now, with |A ⇤ B| = AB sin ⇥ = 171,
⌥
find ⇥ = sin 1
171/14 = 69.1⇤
1.7. Given the vector field E = 4zy 2 cos 2xax + 2zy sin 2xay + y 2 sin 2xaz for the region |x|, |y|, and |z|
less than 2, find:
a) the surfaces on which Ey = 0. With Ey = 2zy sin 2x = 0, the surfaces are 1) the plane z = 0,
with |x| < 2, |y| < 2; 2) the plane y = 0, with |x| < 2, |z| < 2; 3) the plane x = 0, with |y| < 2,
|z| < 2; 4) the plane x = ⇤/2, with |y| < 2, |z| < 2.
b) the region in which Ey = Ez : This occurs when 2zy sin 2x = y 2 sin 2x, or on the plane 2z = y,
with |x| < 2, |y| < 2, |z| < 1.
c) the region in which E = 0: We would have Ex = Ey = Ez = 0, or zy 2 cos 2x = zy sin 2x =
y 2 sin 2x = 0. This condition is met on the plane y = 0, with |x| < 2, |z| < 2.
2
1.8. Demonstrate the ambiguity that results when the cross product is used to find the angle between
two vectors by finding the angle between A = 3ax 2ay + 4az and B = 2ax + ay 2az . Does this
ambiguity exist when the dot product is used?
We use the relation A ⇤ B = |A||B| sin ⇥n. With the given vectors we find
⌥ 2ay + az
⌥
⌥
⌥
A ⇤ B = 14ay + 7az = 7 5
= 9 + 4 + 16 4 + 1 + 4 sin ⇥ n
5
✏
±n
where n is identified as shown; we see that n can be positive or negative, as sin ⇥ can be
positive or negative. This apparent sign ambiguity is not the real problem, however, as we
really want
of the angle anyway. Choosing the positive sign, we are left with
⌥ the
⌥ magnitude
⌥
sin ⇥ = 7 5/( 29 9) = 0.969. Two values of ⇥ (75.7⇤ and 104.3⇤ ) satisfy this equation, and
hence the real ambiguity.
⌥
In using the dot
⌥ product, we find A · B = ⇤6 2 8 = 4 = |A||B| cos ⇥ = 3 29 cos ⇥, or
cos ⇥ = 4/(3 29) = 0.248 ⇧ ⇥ = 75.7 . Again, the minus sign is not important, as we
care only about the angle magnitude. The main point is that only one ⇥ value results when
using the dot product, so no ambiguity.
1.9. A field is given as
G=
(x2
25
(xax + yay )
+ y2 )
Find:
a) a unit vector in the direction of G at P (3, 4, 2): Have Gp = 25/(9 + 16) ⇤ (3, 4, 0) = 3ax + 4ay ,
and |Gp | = 5. Thus aG = (0.6, 0.8, 0).
b) the angle between G and ax at P : The angle is found through aG · ax = cos ⇥. So cos ⇥ =
(0.6, 0.8, 0) · (1, 0, 0) = 0.6. Thus ⇥ = 53⇤ .
c) the value of the following double integral on the plane y = 7:
⌦
0
⌦
0
4
⌦
0
2
4
⌦
2
0
G · ay dzdx
25
(xax + yay ) · ay dzdx =
2
x + y2
⌦
0
4
⌦
0
2
⌃ ⌥
1
4
1
= 350 ⇤
tan
7
7
25
⇤ 7 dzdx =
2
x + 49
⌦
0
4
350
dx
+ 49
x2
0 = 26
1.10. By expressing diagonals as vectors and using the definition of the dot product, find the smaller angle
between any two diagonals of a cube, where each diagonal connects diametrically opposite corners,
and passes through the center of the cube:
Assuming a side length, b, two diagonal vectors would be A = ⌥
b(ax +
⌥ ay + az ) and B =
b(ax ay + az ). Now use A · B = |A||B| cos ⇥, or b2 (1 1 + 1) = ( 3b)( 3b) cos ⇥ ⇧ cos ⇥ =
1/3 ⇧ ⇥ = 70.53⇤ . This result (in magnitude) is the same for any two diagonal vectors.
3
1.11. Given the points M (0.1, 0.2, 0.1), N ( 0.2, 0.1, 0.3), and P (0.4, 0, 0.1), find:
a) the vector RM N : RM N = ( 0.2, 0.1, 0.3)
(0.1, 0.2, 0.1) = ( 0.3, 0.3, 0.4).
b) the dot product RM N · RM P : RM P = (0.4, 0, 0.1) (0.1, 0.2, 0.1) = (0.3, 0.2, 0.2). RM N ·
RM P = ( 0.3, 0.3, 0.4) · (0.3, 0.2, 0.2) = 0.09 + 0.06 + 0.08 = 0.05.
c) the scalar projection of RM N on RM P :
(0.3, 0.2, 0.2)
0.05
RM N · aRM P = ( 0.3, 0.3, 0.4) · ⌥
=⌥
= 0.12
0.09 + 0.04 + 0.04
0.17
d) the angle between RM N and RM P :
⇥M = cos
1
⌃
RM N · RM P
|RM N ||RM P |
⌥
= cos
1
⌃
0.05
⌥
⌥
0.34 0.17
⌥
= 78⇤
1.12. Write an expression in rectangular components for the vector that extends from (x1 , y1 , z1 ) to
(x2 , y2 , z2 ) and determine the magnitude of this vector.
The two points can be written as vectors from the origin:
A1 = x1 ax + y1 ay + z1 az
and A2 = x2 ax + y2 ay + z2 az
The desired vector will now be the di⇥erence:
A12 = A2
A1 = (x2
x1 )ax + (y2
y1 )ay + (z2
z1 )az
y1 )2 + (z2
z1 )2
whose magnitude is
|A12 | =
↵
⇤
A12 · A12 = (x2
x1 )2 + (y2
⌅1/2
1.13. a) Find the vector component of F = (10, 6, 5) that is parallel to G = (0.1, 0.2, 0.3):
F||G =
F·G
(10, 6, 5) · (0.1, 0.2, 0.3)
G=
(0.1, 0.2, 0.3) = (0.93, 1.86, 2.79)
2
|G|
0.01 + 0.04 + 0.09
b) Find the vector component of F that is perpendicular to G:
FpG = F
F||G = (10, 6, 5)
(0.93, 1.86, 2.79) = (9.07, 7.86, 2.21)
c) Find the vector component of G that is perpendicular to F:
GpF = G
G||F = G
G·F
F = (0.1, 0.2, 0.3)
|F|2
4
1.3
(10, 6, 5) = (0.02, 0.25, 0.26)
100 + 36 + 25
1.14. Given that A + B + C = 0, where the three vectors represent line segments and extend from a
common origin,
a) must the three vectors be coplanar?
In terms of the components, the vector sum will be
A + B + C = (Ax + Bx + Cx )ax + (Ay + By + Cy )ay + (Az + Bz + Cz )az
which we require to be zero. Suppose the coordinate system is configured so that vectors A
and B lie in the x-y plane; in this case Az = Bz = 0. Then Cz has to be zero in order for the
three vectors to sum to zero. Therefore, the three vectors must be coplanar.
b) If A + B + C + D = 0, are the four vectors coplanar?
The vector sum is now
A + B + C + D = (Ax + Bx + Cx + Dx )ax + (Ay + By + Cy + Dy )ay + (Az + Bz + Cz + Dz )az
Now, for example, if A and B lie in the x-y plane, C and D need not, as long as Cz + Dz = 0.
So the four vectors need not be coplanar to have a zero sum.
1.15. Three vectors extending from the origin are given as r1 = (7, 3, 2), r2 = ( 2, 7, 3), and r3 =
(0, 2, 3). Find:
a) a unit vector perpendicular to both r1 and r2 :
ap12 =
r1 ⇤ r2
(5, 25, 55)
=
= (0.08, 0.41, 0.91)
|r1 ⇤ r2 |
60.6
b) a unit vector perpendicular to the vectors r1 r2 and r2
( 2, 5, 6). So r1 r2 ⇤ r2 r3 = (19, 52, 32). Then
ap =
r3 : r1
r2 = (9, 4, 1) and r2
(19, 52, 32)
(19, 52, 32)
=
= (0.30, 0.81, 0.50)
|(19, 52, 32)|
63.95
c) the area of the triangle defined by r1 and r2 :
Area =
1
|r1 ⇤ r2 | = 30.3
2
d) the area of the triangle defined by the heads of r1 , r2 , and r3 :
Area =
1
|(r2
2
r1 ) ⇤ (r2
r3 )| =
5
1
|( 9, 4, 1) ⇤ ( 2, 5, 6)| = 32.0
2
r3 =
1.16. If A represents a vector one unit in length directed due east, B represents a vector three units in
length directed due north, and A + B = 2C D and 2A B = C + 2D, determine the length and
direction of C. (di⇧culty 1)
Take north as the positive y direction, and then east as the positive x direction. Then we may write
A + B = ax + 3ay = 2C
and
2A
B = 2ax
D
3ay = C + 2D
Multiplying the first equation by 2, and then adding the result to the second equation eliminates
D, and we get
4
3
4ax + 3ay = 5C ⇧ C = ax + ay
5
5
⇤
⌅
1/2
The length of C is |C| = (4/5)2 + (3/5)2
=1
C lies in the x-y plane at angle from due north (the y axis) given by = tan 1 (4/3) = 53.1⇤ (or
36.9⇤ from the x axis). For those having nautical leanings, this is very close to the compass point
NE 34 E (not required).
1.17. Point A( 4, 2, 5) and the two vectors, RAM = (20, 18, 10) and RAN = ( 10, 8, 15), define a
triangle.
a) Find a unit vector perpendicular to the triangle: Use
ap =
RAM ⇤ RAN
(350, 200, 340)
=
= (0.664, 0.379, 0.645)
|RAM ⇤ RAN |
527.35
The vector in the opposite direction to this one is also a valid answer.
b) Find a unit vector in the plane of the triangle and perpendicular to RAN :
aAN =
( 10, 8, 15)
⌥
= ( 0.507, 0.406, 0.761)
389
Then
apAN = ap ⇤ aAN = (0.664, 0.379, 0.645) ⇤ ( 0.507, 0.406, 0.761) = ( 0.550, 0.832, 0.077)
The vector in the opposite direction to this one is also a valid answer.
c) Find a unit vector in the plane of the triangle that bisects the interior angle at A: A non-unit
vector in the required direction is (1/2)(aAM + aAN ), where
aAM =
Now
(20, 18, 10)
= (0.697, 0.627, 0.348)
|(20, 18, 10)|
1
1
(aAM + aAN ) = [(0.697, 0.627, 0.348) + ( 0.507, 0.406, 0.761)] = (0.095, 0.516, 0.207)
2
2
Finally,
abis =
(0.095, 0.516, 0.207)
= (0.168, 0.915, 0.367)
|(0.095, 0.516, 0.207)|
6
1.18. A certain vector field is given as G = (y + 1)ax + xay . a) Determine G at the point (3,-2,4):
G(3, 2, 4) =
ax + 3ay .
b) obtain a unit vector defining the direction of G at (3,-2,4).
⌥
|G(3, 2, 4)| = [1 + 32 ]1/2 = 10. So the unit vector is
aG (3, 2, 4) =
ax + 3ay
⌥
10
1.19. a) Express the field D = (x2 + y 2 ) 1 (xax + yay ) in cylindrical components and cylindrical variables:
Have x = ⌅ cos ⇧, y = ⌅ sin ⇧, and x2 + y 2 = ⌅2 . Therefore
D=
Then
D⇥ = D · a⇥ =
and
D⇤ = D · a⇤ =
1
(cos ⇧ax + sin ⇧ay )
⌅
⌅ 1
1
1⇤ 2
[cos ⇧(ax · a⇥ ) + sin ⇧(ay · a⇥ )] =
cos ⇧ + sin2 ⇧ =
⌅
⌅
⌅
1
1
[cos ⇧(ax · a⇤ ) + sin ⇧(ay · a⇤ )] = [cos ⇧( sin ⇧) + sin ⇧ cos ⇧] = 0
⌅
⌅
Therefore
D=
1
a⇥
⌅
b) Evaluate D at the point where ⌅ = 2, ⇧ = 0.2⇤, and z = 5, expressing the result in cylindrical
and cartesian coordinates: At the given point, and in cylindrical coordinates, D = 0.5a⇥ . To
express this in cartesian, we use
D = 0.5(a⇥ · ax )ax + 0.5(a⇥ · ay )ay = 0.5 cos 36⇤ ax + 0.5 sin 36⇤ ay = 0.41ax + 0.29ay
1.20. If the three sides of a triangle are represented by the vectors A, B, and C, all directed counterclockwise, show that |C|2 = (A + B) · (A + B) and expand the product to obtain the law of cosines.
With the vectors drawn as described above, we find that C = (A + B) and so |C|2 = C 2 = C · C =
(A + B) · (A + B) So far so good. Now if we expand the product, obtain
(A + B) · (A + B) = A2 + B 2 + 2A · B
where A · B = AB cos(180⇤
) = AB cos where is the interior angle at the junction of A
and B. Using this, we have C 2 = A2 + B 2 2AB cos , which is the law of cosines.
7
1.21. Express in cylindrical components:
a) the vector from C(3, 2, 7) to D( 1, 4, 2):
C(3, 2, 7) ⌅ C(⌅ = 3.61, ⇧ = 33.7⇤ , z = 7) and
D( 1, 4, 2) ⌅ D(⌅ = 4.12, ⇧ = 104.0⇤ , z = 2).
Now RCD = ( 4, 6, 9) and R⇥ = RCD · a⇥ = 4 cos(33.7) 6 sin(33.7) = 6.66. Then
R⇤ = RCD · a⇤ = 4 sin(33.7) 6 cos(33.7) = 2.77. So RCD = 6.66a⇥ 2.77a⇤ + 9az
b) a unit vector at D directed toward C:
RCD = (4, 6, 9) and R⇥ = RDC · a⇥ = 4 cos( 104.0) + 6 sin( 104.0) = 6.79. Then R⇤ =
RDC · a⇤ = 4[ sin( 104.0)] + 6 cos( 104.0) = 2.43. So RDC = 6.79a⇥ + 2.43a⇤ 9az
Thus aDC = 0.59a⇥ + 0.21a⇤ 0.78az
c) a unit vector at D directed toward the origin: Start with rD = ( 1, 4, 2), and so the
vector toward the origin will be rD = (1, 4, 2). Thus in cartesian the unit vector is a =
(0.22, 0.87, 0.44). Convert to cylindrical:
a⇥ = (0.22, 0.87, 0.44) · a⇥ = 0.22 cos( 104.0) + 0.87 sin( 104.0) = 0.90, and
a⇤ = (0.22, 0.87, 0.44) · a⇤ = 0.22[ sin( 104.0)] + 0.87 cos( 104.0) = 0, so that finally,
a = 0.90a⇥ 0.44az .
1.22. A sphere of radius a, centered at the origin, rotates about the z axis at angular velocity
The rotation direction is clockwise when one is looking in the positive z direction.
rad/s.
a) Using spherical components, write an expression for the velocity field, v, which gives the tangential velocity at any point within the sphere:
As in problem 1.20, we find the tangential velocity as the product of the angular velocity and
the perperdicular distance from the rotation axis. With clockwise rotation, we obtain
v(r, ⇥) = r sin ⇥ a⇤ (r < a)
b) Convert to rectangular components:
From here, the problem is the same as part c in Problem 1.20, except the rotation direction is
reversed. The answer is v(x, y) = [ y ax + x ay ], where (x2 + y 2 + z 2 )1/2 < a.
1.23. The surfaces ⌅ = 3, ⌅ = 5, ⇧ = 100⇤ , ⇧ = 130⇤ , z = 3, and z = 4.5 define a closed surface.
a) Find the enclosed volume:
Vol =
⌦
4.5
3
⌦
130
100
⌦
5
⌅ d⌅ d⇧ dz = 6.28
3
NOTE: The limits on the ⇧ integration must be converted to radians (as was done here, but not
shown).
b) Find the total area of the enclosing surface:
Area = 2
+
⌦
3
4.5
⌦
⌦
130
100
130
⌦
5
⌅ d⌅ d⇧ +
3
5 d⇧ dz + 2
100
⌦
3
8
⌦
4.5
⌦
130
3
100
4.5 ⌦ 5
3 d⇧ dz
d⌅ dz = 20.7
3
1.23c) Find the total length of the twelve edges of the surfaces:
Length = 4 ⇤ 1.5 + 4 ⇤ 2 + 2 ⇤
30⇤
30⇤
⇤
2⇤
⇤
3
+
⇤ 2⇤ ⇤ 5 = 22.4
360⇤
360⇤
d) Find the length of the longest straight line that lies entirely within the volume: This will be
between the points A(⌅ = 3, ⇧ = 100⇤ , z = 3) and B(⌅ = 5, ⇧ = 130⇤ , z = 4.5). Performing
point transformations to cartesian coordinates, these become A(x = 0.52, y = 2.95, z = 3)
and B(x = 3.21, y = 3.83, z = 4.5). Taking A and B as vectors directed from the origin, the
requested length is
Length = |B A| = |( 2.69, 0.88, 1.5)| = 3.21
1.24. Two unit vectors, a1 and a2 lie in the xy plane and pass through the origin. They make angles ⇧1
and ⇧2 with the x axis respectively.
a) Express each vector in rectangular components; Have a1 = Ax1 ax + Ay1 ay , so that Ax1 =
a1 · ax = cos ⇧1 . Then, Ay1 = a1 · ay = cos(90 ⇧1 ) = sin ⇧1 . Therefore,
a1 = cos ⇧1 ax + sin ⇧1 ay
and similarly, a2 = cos ⇧2 ax + sin ⇧2 ay
b) take the dot product and verify the trigonometric identity, cos(⇧1
sin ⇧1 sin ⇧2 : From the definition of the dot product,
⇧2 ) = cos ⇧1 cos ⇧2 +
a1 · a2 = (1)(1) cos(⇧1 ⇧2 )
= (cos ⇧1 ax + sin ⇧1 ay ) · (cos ⇧2 ax + sin ⇧2 ay ) = cos ⇧1 cos ⇧2 + sin ⇧1 sin ⇧2
c) take the cross product and verify the trigonometric identity sin(⇧2 ⇧1 ) = sin ⇧2 cos ⇧1
cos ⇧2 sin ⇧1 : From the definition of the cross product, and since a1 and a2 both lie in the x-y
plane,
⇧
⇧
⇧ ax
ay
az ⇧⇧
⇧
a1 ⇤ a2 = (1)(1) sin(⇧1 ⇧2 ) az = ⇧⇧ cos ⇧1 sin ⇧1 0 ⇧⇧
⇧ cos ⇧2 sin ⇧2 0 ⇧
= [sin ⇧2 cos ⇧1
cos ⇧2 sin ⇧1 ] az
thus verified.
1.25. Given point P (r = 0.8, ⇥ = 30⇤ , ⇧ = 45⇤ ), and
E=
1
r2
⌃
⌥
sin ⇧
cos ⇧ ar +
a⇤
sin ⇥
a) Find E at P : E = 1.10a⇥ + 2.21a⇤ .
⌥
b) Find |E| at P : |E| = 1.102 + 2.212 = 2.47.
c) Find a unit vector in the direction of E at P :
aE =
E
= 0.45ar + 0.89a⇤
|E|
9
1.26. Express the uniform vector field, F = 5 ax in
a) cylindrical components: F⇥ = 5 ax · a⇥ = 5 cos ⇧, and F⇤ = 5 ax · a⇤ =
obtain F(⌅, ⇧) = 5(cos ⇧ a⇥ sin ⇧ a⇤ ).
5 sin ⇧. Combining, we
b) spherical components: Fr = 5 ax ·ar = 5 sin ⇥ cos ⇧; F = 5 ax ·a = 5 cos ⇥ cos ⇧; F⇤ = 5 ax ·a⇤ =
5 sin ⇧. Combining, we obtain F(r, ⇥, ⇧) = 5 [sin ⇥ cos ⇧ ar + cos ⇥ cos ⇧ a
sin ⇧ a⇤ ].
1.27. The surfaces r = 2 and 4, ⇥ = 30⇤ and 50⇤ , and ⇧ = 20⇤ and 60⇤ identify a closed surface.
a) Find the enclosed volume: This will be
Vol =
⌦
60
20
⌦
50
30
⌦
4
r2 sin ⇥drd⇥d⇧ = 2.91
2
where degrees have been converted to radians.
b) Find the total area of the enclosing surface:
Area =
⌦
60
20
⌦
50
(4 + 2 ) sin ⇥d⇥d⇧ +
2
2
30
⌦
4
2
+2
⌦
⌦
60
r(sin 30⇤ + sin 50⇤ )drd⇧
20
50 ⌦ 4
30
rdrd⇥ = 12.61
2
c) Find the total length of the twelve edges of the surface:
Length = 4
⌦
4
2
= 17.49
dr + 2
⌦
50
30
(4 + 2)d⇥ +
⌦
60
(4 sin 50⇤ + 4 sin 30⇤ + 2 sin 50⇤ + 2 sin 30⇤ )d⇧
20
d) Find the length of the longest straight line that lies entirely within the surface: This will be
from A(r = 2, ⇥ = 50⇤ , ⇧ = 20⇤ ) to B(r = 4, ⇥ = 30⇤ , ⇧ = 60⇤ ) or
A(x = 2 sin 50⇤ cos 20⇤ , y = 2 sin 50⇤ sin 20⇤ , z = 2 cos 50⇤ )
to
B(x = 4 sin 30⇤ cos 60⇤ , y = 4 sin 30⇤ sin 60⇤ , z = 4 cos 30⇤ )
or finally A(1.44, 0.52, 1.29) to B(1.00, 1.73, 3.46). Thus B
Length = |B
A = ( 0.44, 1.21, 2.18) and
A| = 2.53
1.28. State whether or not A = B and, if not, what conditions are imposed on A and B when
a) A · ax = B · ax : For this to be true, both A and B must be oriented at the same angle, ⇥, from
the x axis. But this would allow either vector to lie anywhere along a conical surface of angle
⇥ about the x axis. Therefore, A can be equal to B, but not necessarily.
b) A ⇤ ax = B ⇤ ax : This is a more restrictive condition because the cross product gives a vector.
For both cross products to lie in the same direction, A, B, and ax must be coplanar. But if A
lies at angle ⇥ to the x axis, B could lie at ⇥ or at 180⇤ ⇥ to give the same cross product. So
again, A can be equal to B, but not necessarily.
10
1.28c) A · ax = B · ax and A ⇤ ax = B ⇤ ax : In this case, we need to satisfy both requirements in parts
a and b – that is, A, B, and ax must be coplanar, and A and B must lie at the same angle, ⇥, to
ax . With coplanar vectors, this latter condition might imply that both +⇥ and ⇥ would therefore
work. But the negative angle reverses the direction of the cross product direction. Therefore both
vectors must lie in the same plane and lie at the same angle to x; i.e., A must be equal to B.
d) A · C = B · C and A ⇤ C = B ⇤ C where C is any vector except C = 0: This is just the general
case of part c. Since we can orient our coordinate system in any manner we choose, we can
arrange it so that the x axis coincides with the direction of vector C. Thus all the arguments
of part c apply, and again we conclude that A must be equal to B.
1.29. Express the unit vector ax in spherical components at the point:
a) r = 2, ⇥ = 1 rad, ⇧ = 0.8 rad: Use
ax = (ax · ar )ar + (ax · a )a + (ax · a⇤ )a⇤ =
sin(1) cos(0.8)ar + cos(1) cos(0.8)a + ( sin(0.8))a⇤ = 0.59ar + 0.38a
0.72a⇤
b) x = 3, y = 2, ⌥z = 1: First, transform the point to spherical coordinates. Have r =
⇥ = cos 1 ( 1/ 14) = 105.5⇤ , and ⇧ = tan 1 (2/3) = 33.7⇤ . Then
⌥
14,
ax = sin(105.5⇤ ) cos(33.7⇤ )ar + cos(105.5⇤ ) cos(33.7⇤ )a + ( sin(33.7⇤ ))a⇤
= 0.80ar 0.22a
0.55a⇤
c) ⌅
⌥= 2.5, ⇧ = 0.71rad, z = 1.5: Again,
⌥convert the point to spherical coordinates. r =
8.5, ⇥ = cos (z/r) = cos 1 (1.5/ 8.5) = 59.0⇤ , and ⇧ = 0.7 rad = 40.1⇤ . Now
ax = sin(59⇤ ) cos(40.1⇤ )ar + cos(59⇤ ) cos(40.1⇤ )a + ( sin(40.1⇤ ))a⇤
= 0.66ar + 0.39a
0.64a⇤
↵
⌅2 + z 2 =
1.30. Consider a problem analogous to the varying wind velocities encountered by transcontinental aircraft.
We assume a constant altitude, a plane earth, a flight along the x axis from 0 to 10 units, no vertical
velocity component, and no change in wind velocity with time. Assume ax to be directed to the
east and ay to the north. The wind velocity at the operating altitude is assumed to be:
v(x, y) =
(0.01x2
0.08x + 0.66)ax (0.05x
1 + 0.5y 2
0.4)ay
a) Determine the location and magnitude of the maximum tailwind encountered: Tailwind would
be x-directed, and so we look at the x component only. Over the flight range, this function
maximizes at a value of 0.86/(1 + 0.5y 2 ) at x = 10 (at the end of the trip). It reaches a local
minimum of 0.50/(1 + 0.5y 2 ) at x = 4, and has another local maximum of 0.66/(1 + 0.5y 2 ) at
the trip start, x = 0.
b) Repeat for headwind: The x component is always positive, and so therefore no headwind exists
over the travel range.
c) Repeat for crosswind: Crosswind will be found from the y component, which is seen to maximize
over the flight range at a value of 0.4/(1 + 0.5y 2 ) at the trip start (x = 0).
d) Would more favorable tailwinds be available at some other latitude? If so, where? Minimizing
the denominator accomplishes this; in particular, the lattitude associated with y = 0 gives the
strongest tailwind.
11
.
CHAPTER 3
3.1. Suppose that the Faraday concentric sphere experiment is performed in free space using a
central charge at the origin, Q1 , and with hemispheres of radius a. A second charge Q2 (this
time a point charge) is located at distance R from Q1 , where R >> a.
a) What is the force on the point charge before the hemispheres are assembled around Q1 ?
This will be simply the force beween two point charges, or
F=
Q1 Q2
ar
4⇧⇥0 R2
b) What is the force on the point charge after the hemispheres are assembled but before
they are discharged? The answer will be the same as in part a because induced charge
Q1 now resides as a surface charge layer on the sphere exterior. This produces the same
electric field at the Q2 location as before, and so the force acting on Q2 is the same.
c) What is the force on the point charge after the hemispheres are assembled and after they
are discharged? Discharging the hemispheres (connecting them to ground) neutralizes the
positive outside surface charge layer, thus zeroing the net field outside the sphere. The
force on Q2 is now zero.
d) Qualitatively, describe what happens as Q2 is moved toward the sphere assembly to the
extent that the condition R >> a is no longer valid. Q2 itself begins to induce negative
surface charge on the sphere. An attractive force thus begins to strengthen as the charge
moves closer. The point charge field approximation used in parts a through c is no longer
valid.
3.2. An electric field in free space is E = (5z 2 /⇥0 ) âz V/m. Find the total charge contained within
a cube, centered at the origin, of 4-m side length, in which all sides are parallel to coordinate
axes (and therefore each side intersects an axis at ±2.
The flux density is D = ⇥0 E = 5z 2 az . As D is z-directed only, it will intersect only the
top and bottom surfaces (both parallel to the x-y plane). From Gauss’ law, the charge in
the cube is equal to the net outward flux of D, which in this case is
Qencl =
D · n da =
2
2
2
2
5(2)2 az · az dx dy +
2
2
2
2
5( 2)2 az · ( az ) dx dy = 0
where the first and second integrals on the far right are over the top and bottom surfaces
respectively.
25
3.3. The cylindrical surface ⌃ = 8 cm contains the surface charge density, ⌃s = 5e 20|z| nC/m2 .
a) What is the total amount of charge present? We integrate over the surface to find:
⌅
Q=2
2⇤
0
5e
20z
(.08)d⌥ dz nC = 20⇧(.08)
0
⌃
1
20
⌥
⇧⌅
⇧
⇧
e 20z ⇧ = 0.25 nC
⇧
0
b) How much flux leaves the surface ⌃ = 8 cm, 1 cm < z < 5cm, 30⇥ < ⌥ < 90⇥ ? We just
integrate the charge density on that surface to find the flux that leaves it.
⇥ = Q⇤ =
.05
90
.01
= 9.45 ⇤ 10
5e
20z
(.08) d⌥ dz nC =
30
3
⌃
90 30
360
⌥
2⇧(5)(.08)
⌃
1
20
⌥
nC = 9.45 pC
⇧.05
⇧
⇧
e 20z ⇧
⇧
.01
3.4. An electric field in free space is E = (5z 3 /⇥0 ) âz V/m. Find the total charge contained within
a sphere of 3-m radius, centered at the origin. Using Gauss’ law, we set up the integral in free
space over the sphere surface, whose outward unit normal is ar :
Q=
⇥0 E · n da =
2⇤
⇤
0
0
5z 3 az · ar (3)2 sin ⇤ d⇤ d⌥
where in this case z = 3 cos ⇤ and (in all cases) az · ar = cos ⇤. These are substituted to yield
⇤
Q = 2⇧
5(3)5 cos4 ⇤ sin ⇤d⇤ =
2⇧(5)(3)5
0
⌃ ⌥
⇧2⇤
1
⇧
cos5 ⇤⇧ = 972⇧
5
0
3.5. Let D = 4xyax + 2(x2 + z 2 )ay + 4yzaz C/m2 and evaluate surface integrals to find the total
charge enclosed in the rectangular parallelepiped 0 < x < 2, 0 < y < 3, 0 < z < 5 m: Of the 6
surfaces to consider, only 2 will contribute to the net outward flux. Why? First consider the
planes at y = 0 and 3. The y component of D will penetrate those surfaces, but will be inward
at y = 0 and outward at y = 3, while having the same magnitude in both cases. These fluxes
will thus cancel. At the x = 0 plane, Dx = 0 and at the z = 0 plane, Dz = 0, so there will be
no flux contributions from these surfaces. This leaves the 2 remaining surfaces at x = 2 and
z = 5. The net outward flux becomes:
5
⇥=
3
0
0
3
=5
0
⇧
D⇧x=2 · ax dy dz +
3
4(2)y dy + 2
0
26
3
0
2
0
⇧
D⇧z=5 · az dx dy
4(5)y dy = 360 C
3.6. In free space, volume charge of constant density ⌃v = ⌃0 exists within the region
< y < , and d/2 < z < d/2. Find D and E everywhere.
<x<
,
From the symmetry of the configuration, we surmise that the field will be everywhere
z-directed, and will be uniform with x and y at fixed z. For finding the field inside the
charge, an appropriate Gaussian surface will be that which encloses a rectangular region
defined by 1 < x < 1, 1 < y < 1, and |z| < d/2. The outward flux from this surface
will be limited to that through the two parallel surfaces at ±z:
⇥in =
D · dS = 2
1
1
1
1
Dz dxdy = Qencl =
z
1
1
z
1
1
⌃0 dxdydz ⇤
where the factor of 2 in the second integral account for the equal fluxes through the
two surfaces. The above readily simplifies, as both Dz and ⌃0 are constants, leading to
Din = ⌃0 z az C/m2 (|z| < d/2), and therefore Ein = (⌃0 z/⇥0 ) az V/m (|z| < d/2).
Outside the charge, the Gaussian surface is the same, except that the parallel boundaries
at ±z occur at |z| > d/2. As a result, the calculation is nearly the same as before, with
the only change being the limits on the total charge integral:
⇥out =
D · dS = 2
1
1
1
1
Dz dxdy = Qencl =
d/2
1
1
d/2
1
1
⌃0 dxdydz ⇤
Solve for Dz to find the constant values:
⌦
⌦
(⌃0 d/2) az (z > d/2)
(⌃0 d/2⇥0 ) az (z > d/2)
2
Dout =
C/m and Eout =
V/m
(⌃0 d/2) az (z < d/2)
(⌃0 d/2⇥0 ) az (z < d/2)
3.7. Volume charge density is located in free space as ⌃v = 2e 1000r nC/m3 for 0 < r < 1 mm, and
⌃v = 0 elsewhere.
a) Find the total charge enclosed by the spherical surface r = 1 mm: To find the charge we
integrate:
2⇤
Q=
0
⇤
0
.001
2e
r sin ⇤ dr d⇤ d⌥
1000r 2
0
Integration over the angles gives a factor of 4⇧. The radial integration we evaluate using
tables; we obtain
Q = 8⇧
r2 e 1000r ⇧⇧.001
2 e 1000r
+
( 1000r
⇧
1000
1000 (1000)2
0
⇧.001
⇧
1)⇧
= 4.0 ⇤ 10
0
9
nC
b) By using Gauss’s law, calculate the value of Dr on the surface r = 1 mm: The gaussian
surface is a spherical shell of radius 1 mm. The enclosed charge is the result of part a.
We thus write 4⇧r2 Dr = Q, or
Dr =
Q
4.0 ⇤ 10 9
=
= 3.2 ⇤ 10
4⇧r2
4⇧(.001)2
27
4
nC/m2
3.8. Use Gauss’s law in integral form to show that an inverse distance field in spherical coordinates,
D = Aar /r, where A is a constant, requires every spherical shell of 1 m thickness to contain
4⇧A coulombs of charge. Does this indicate a continuous charge distribution? If so, find the
charge density variation with r.
The net outward flux of this field through a spherical surface of radius r is
⇥=
D · dS =
2⇤
⇤
0
0
A
ar · ar r2 sin ⇤ d⇤ d⌥ = 4⇧Ar = Qencl
r
We see from this that with every increase in r by one m, the enclosed charge increases
by 4⇧A (done). It is evident that the charge density is continuous, and we can find the
density indirectly by constructing the integral for the enclosed charge, in which we already
found the latter from Gauss’s law:
2⇤
Qencl = 4⇧Ar =
0
⇤
0
r
r
⌃(r⇤ ) (r⇤ )2 sin ⇤ dr⇤ d⇤ d⌥ = 4⇧
0
⌃(r⇤ ) (r⇤ )2 dr⇤
0
To obtain the correct enclosed charge, the integrand must be ⌃(r) = A/r2 .
3.9. A uniform volume charge density of 80 µC/m3 is present throughout the region 8 mm < r <
10 mm. Let ⌃v = 0 for 0 < r < 8 mm.
a) Find the total charge inside the spherical surface r = 10 mm: This will be
2⇤
Q=
0
⇤
0
= 1.64 ⇤ 10
.010
.008
10
(80 ⇤ 10
6
)r2 sin ⇤ dr d⇤ d⌥ = 4⇧ ⇤ (80 ⇤ 10
C = 164 pC
6
)
r3 ⇧⇧.010
⇧
3 .008
b) Find Dr at r = 10 mm: Using a spherical gaussian surface at r = 10, Gauss’ law is
written as 4⇧r2 Dr = Q = 164 ⇤ 10 12 , or
Dr (10 mm) =
164 ⇤ 10 12
= 1.30 ⇤ 10
4⇧(.01)2
7
C/m2 = 130 nC/m2
c) If there is no charge for r > 10 mm, find Dr at r = 20 mm: This will be the same
computation as in part b, except the gaussian surface now lies at 20 mm. Thus
Dr (20 mm) =
164 ⇤ 10 12
= 3.25 ⇤ 10
4⇧(.02)2
28
8
C/m2 = 32.5 nC/m2
3.10. An infinitely long cylindrical dielectric of radius b contains charge within its volume of density
⌃v = a⌃2 , where a is a constant. Find the electric field strength, E, both inside and outside
the cylinder.
Inside, we note from symmetry that D will be radially-directed, in the manner of a line
charge field. So we apply Gauss’ law to a cylindrical surface of radius ⌃, concentric with
the charge distribution, having unit length in z, and where ⌃ < b. The outward normal
to the surface is a⌅ .
D · n da =
1
0
2⇤
0
D⌅ a⌅ · a⌅ ⌃ d⌥ dz = Qencl =
1
0
2⇤
0
⌅
a(⌃⇤ )2 ⌃⇤ d⌃⇤ d⌥ dz
0
in which the dummy variable ⌃⇤ must be used in the far-right integral because the upper
radial limit is ⌃. D⌅ is constant over the surface and can be factored outside the integral.
Evaluating both integrals leads to
⌃ ⌥
1
a⌃3
a⌃3
2⇧(1)⌃D⌅ = 2⇧a
⌃4 ⌥ D⌅ =
or Ein =
a⌅ (⌃ < b)
4
4
4⇥0
To find the field outside the cylinder, we apply Gauss’ law to a cylinder of radius ⌃ > b.
The setup now changes only by the upper radius limit for the charge integral, which is
now the charge radius, b:
D · n da =
1
0
2⇤
0
D⌅ a⌅ · a⌅ ⌃ d⌥ dz = Qencl =
1
0
2⇤
0
b
a⌃2 ⌃ d⌃ d⌥ dz
0
where the dummy variable is no longer needed. Evaluating as before, the result is
D⌅ =
ab4
ab4
or Eout =
a⌅ (⌃ > b)
4⌃
4⇥0 ⌃
3.11. In cylindrical coordinates, let ⌃v = 0 for ⌃ < 1 mm, ⌃v = 2 sin(2000⇧⌃) nC/m3 for 1 mm <
⌃ < 1.5 mm, and ⌃v = 0 for ⌃ > 1.5 mm. Find D everywhere: Since the charge varies only
with radius, and is in the form of a cylinder, symmetry tells us that the flux density will be
radially-directed and will be constant over a cylindrical surface of a fixed radius. Gauss’ law
applied to such a surface of unit length in z gives:
a) for ⌃ < 1 mm, D⌅ = 0, since no charge is enclosed by a cylindrical surface whose radius
lies within this range.
b) for 1 mm < ⌃ < 1.5 mm, we have
2⇧⌃D⌅ = 2⇧
⌅
.001
= 4⇧ ⇤ 10
2 ⇤ 10
9
9
sin(2000⇧⌃⇤ )⌃⇤ d⌃⇤
1
sin(2000⇧⌃)
(2000⇧)2
⌃
cos(2000⇧⌃)
2000⇧
⌅
.001
or finally,
D⌅ =
⇤
10 15
sin(2000⇧⌃) + 2⇧ 1
2
2⇧ ⌃
⌅✏
103 ⌃ cos(2000⇧⌃) C/m2
29
(1 mm < ⌃ < 1.5 mm)
3.11c) for ⌃ > 1.5 mm, the gaussian cylinder now lies at radius ⌃ outside the charge distribution, so
the integral that evaluates the enclosed charge now includes the entire charge distribution. To
accomplish this, we change the upper limit of the integral of part b from ⌃ to 1.5 mm, finally
obtaining:
2.5 ⇤ 10 15
D⌅ =
C/m2 (⌃ > 1.5 mm)
⇧⌃
3.12. The sun radiates a total power of about 3.86 ⇤ 1026 watts (W). If we imagine the sun’s surface
to be marked o⇤ in latitude and longitude and assume uniform radiation,
a) What power is radiated by the region lying between latitude 50⇥ N and 60⇥ N and longitude 12⇥ W and 27⇥ W?
50⇥ N lattitude and 60⇥ N lattitude correspond respectively to ⇤ = 40⇥ and ⇤ = 30⇥ .
12⇥ and 27⇥ correspond directly to the limits on ⌥. Since the sun for our purposes is
spherically-symmetric, the flux density emitted by it is I = 3.86 ⇤ 1026 /(4⇧r2 ) ar W/m2 .
The required power is now found through
P1 =
27
12
3.86 ⇤ 1026
ar · ar r2 sin ⇤ d⇤ d⌥
4⇧r2
40
30
3.86 ⇤ 1026
=
[cos(30⇥ )
4⇧
cos(40 )] (27
⇥
⇥
12 )
⇥
⌃
2⇧
360
⌥
= 8.1 ⇤ 1023 W
b) What is the power density on a spherical surface 93,000,000 miles from the sun in W/m2 ?
First, 93,000,000 miles = 155,000,000 km = 1.55 ⇤ 1011 m. Use this distance in the flux
density expression above to obtain
3.86 ⇤ 1026
ar = 1200 ar W/m2
4⇧(1.55 ⇤ 1011 )2
I=
3.13. Spherical surfaces at r = 2, 4, and 6 m carry uniform surface charge densities of 20 nC/m2 ,
4 nC/m2 , and ⌃s0 , respectively.
a) Find D at r = 1, 3 and 5 m: Noting that the charges are spherically-symmetric, we
ascertain that D will be radially-directed and will vary only with radius. Thus, we apply
Gauss’ law to spherical shells in the following regions: r < 2: Here, no charge is enclosed,
and so Dr = 0.
2 < r < 4 : 4⇧r2 Dr = 4⇧(2)2 (20 ⇤ 10
So Dr (r = 3) = 8.9 ⇤ 10
9
) ⌥ Dr =
80 ⇤ 10
r2
9
C/m2
C/m2 .
4 < r < 6 : 4⇧r2 Dr = 4⇧(2)2 (20 ⇤ 10
So Dr (r = 5) = 6.4 ⇤ 10
9
10
9
) + 4⇧(4)2 ( 4 ⇤ 10
9
) ⌥ Dr =
16 ⇤ 10
r2
9
C/m2 .
b) Determine ⌃s0 such that D = 0 at r = 7 m. Since fields will decrease as 1/r2 , the question
could be re-phrased to ask for ⌃s0 such that D = 0 at all points where r > 6 m. In this
region, the total field will be
16 ⇤ 10 9
⌃s0 (6)2
+
r2
r2
9
= (4/9) ⇤ 10 C/m2 .
Dr (r > 6) =
Requiring this to be zero, we find ⌃s0
30
3.14. A certain light-emitting diode (LED) is centered at the origin with its surface in the xy plane.
At far distances, the LED appears as a point, but the glowing surface geometry produces a
far-field radiation pattern that follows a raised cosine law: That is, the optical power (flux)
density in Watts/m2 is given in spherical coordinates by
Pd = P0
cos2 ⇤
ar
2⇧r2
2
Watts/m
where ⇤ is the angle measured with respect to the normal to the LED surface (in this case,
the z axis), and r is the radial distance from the origin at which the power is detected.
a) Find, in terms of P0 , the total power in Watts emitted in the upper half-space by the
LED: We evaluate the surface integral of the power density over a hemispherical surface
of radius r:
2⇤
Pt =
⇤/2
P0
0
0
cos2 ⇤
ar · ar r2 sin ⇤ d⇤ d⌥ =
2⇧r2
⇧⇤/2 P
P0
⇧
0
cos3 ⇤⇧
=
3
3
0
b) Find the cone angle, ⇤1 , within which half the total power is radiated; i.e., within the
range 0 < ⇤ < ⇤1 : We perform the same integral as in part a except the upper limit for ⇤
is now ⇤1 . The result must be one-half that of part a, so we write:
⌃
⌥
⇧⇥1
⇥
Pt
P0
P0
P0
1
⇧
=
=
cos3 ⇤⇧ =
1 cos3 ⇤1 ⌥ ⇤1 = cos 1
= 37.5⇥
2
6
3
3
0
21/3
c) An optical detector, having a 1 mm2 cross-sectional area, is positioned at r = 1 m and at
⇤ = 45⇥ , such that it faces the LED. If one nanowatt (stated in error as 1mW) is measured
by the detector, what (to a very good estimate) is the value of P0 ? Start with
Pd (45⇥ ) = P0
cos2 (45⇥ )
P0
ar =
ar
2
2⇧r
4⇧r2
Then the detected power in a 1-mm2 area at r = 1 m approximates as
. P0
P [W ] =
⇤ 10
4⇧
6
= 10
.
⌥ P0 = 4⇧ ⇤ 10
9
3
W
If the originally stated 1mW value is used for the detected power, the answer would have
been 4⇧ kW (!).
3.15. Volume charge density is located as follows: ⌃v = 0 for ⌃ < 1 mm and for ⌃ > 2 mm,
⌃v = 4⌃ µC/m3 for 1 < ⌃ < 2 mm.
a) Calculate the total charge in the region 0 < ⌃ < ⌃1 , 0 < z < L, where 1 < ⌃1 < 2 mm:
We find,
L
2⇤
⌅1
8⇧L 3
Q=
4⌃ ⌃ d⌃ d⌥ dz =
[⌃1 10 9 ] µC
3
0
0
.001
b) Use Gauss’ law to determine D⌅ at ⌃ = ⌃1 : Gauss’ law states that 2⇧⌃1 LD⌅ = Q, where
Q is the result of part a. So, with ⌃1 in meters,
D⌅ (⌃1 ) =
4(⌃31
31
10
3⌃1
9
)
µC/m2
3.15c) Evaluate D⌅ at ⌃ = 0.8 mm, 1.6 mm, and 2.4 mm: At ⌃ = 0.8 mm, no charge is enclosed by
a cylindrical gaussian surface of that radius, so D⌅ (0.8mm) = 0. At ⌃ = 1.6 mm, we evaluate
the part b result at ⌃1 = 1.6 to obtain:
D⌅ (1.6mm) =
4[(.0016)3 (.0010)3 ]
= 3.6 ⇤ 10
3(.0016)
6
µC/m2
At ⌃ = 2.4, we evaluate the charge integral of part a from .001 to .002, and Gauss’ law is
written as
8⇧L
2⇧⌃LD⌅ =
[(.002)2 (.001)2 ] µC
3
from which D⌅ (2.4mm) = 3.9 ⇤ 10 6 µC/m2 .
3.16. An electric flux density is given by D = D0 a⌅ , where D0 is a given constant.
a) What charge density generates this field? Charge density is found by taking the divergence: With radial D only, we have
1 d
D0
⌃v = ⌦ · D =
(⌃D0 ) =
C/m3
⌃ d⌃
⌃
b) For the specified field, what total charge is contained within a cylinder of radius a and
height b, where the cylinder axis is the z axis? We can either integrate the charge density
over the specified volume, or integrate D over the surface that contains the specified
volume:
b
Q=
0
2⇤
0
a
0
D0
⌃ d⌃ d⌥ dz =
⌃
b
0
2⇤
0
D0 a⌅ · a⌅ a d⌥ dz = 2⇧abD0 C
3.17. A cube is defined by 1 < x, y, z < 1.2. If D = 2x2 yax + 3x2 y 2 ay C/m2 :
a) apply Gauss’ law to find the total flux leaving the closed surface of the cube. We call the
surfaces at x = 1.2 and x = 1 the front and back surfaces respectively, those at y = 1.2
and y = 1 the right and left surfaces, and those at z = 1.2 and z = 1 the top and bottom
surfaces. To evaluate the total charge, we integrate D · n over all six surfaces and sum
the results. We note that there is no z component of D, so there will be no outward flux
contributions from the top and bottom surfaces. The fluxes through the remaining four
are
⇥=Q=
D · n da =
1.2
+
✓1
1.2
1
1.2
✓1
1.2
1
2(1.2)2 y dy dz +
⌘⇣
◆ ✓1
front
3x2 (1)2 dx dz +
⌘⇣
◆ ✓1
left
1.2
1.2
1
1.2
1.2
1
2(1)2 y dy dz
⌘⇣
◆
back
3x2 (1.2)2 dx dz = 0.1028 C
⌘⇣
◆
right
b) evaluate ⌦ · D at the center of the cube: This is
⇤
⌅
⌦ · D = 4xy + 6x2 y (1.1,1.1) = 4(1.1)2 + 6(1.1)3 = 12.83
c) Estimate the total charge enclosed within the cube by using Eq. (8): This is
⇧
.
Q = ⌦ · D⇧center ⇤ v = 12.83 ⇤ (0.2)3 = 0.1026 Close!
32
3.18. State whether the divergence of the following vector fields is positive, negative, or zero:
a) the thermal energy flow in J/(m2 s) at any point in a freezing ice cube: One way to
visualize this is to consider that heat is escaping through the surface of the ice cube as it
freezes. Therefore the net outward flux of thermal energy through the surface is positive.
Calling the thermal flux density F, the divergence theorem says
s
F · dS =
v
⌦ · F dv
and so if we identify the left integral as positive, the right integral (and its integrand)
must also be positive. Answer: positive.
b) the current density in A/m2 in a bus bar carrying direct current: In this case, we have
no accumulation or dissipation of charge within any small volume, since current is dc;
this also means that the net outward current flux through the surface that surrounds any
small volume is zero. Therefore the divergence must be zero.
c) the mass flow rate in kg/(m2 s) below the surface of water in a basin, in which the
water is circulating clockwise as viewed from above: Here again, taking any small volume
in the water, the net outward flow through the surface that surrounds the small volume
is zero; i.e., there is no accumulation or dissipation of mass that would result in a change
in density at any point. Divergence is therefore zero.
3.19. A spherical surface of radius 3 mm is centered at P (4, 1, 5) in free space. Let D = xax C/m2 .
Use the results of Sec. 3.4 to estimate the net electric flux leaving the spherical surface: We
.
use ⇥ = ⌦ · D v, where in this case ⌦ · D = ( / x)x = 1 C/m3 . Thus
. 4
⇥ = ⇧(.003)3 (1) = 1.13 ⇤ 10
3
7
C = 113 nC
3.20. A radial electric field distribution in free space is given in spherical coordinates as:
r⌃0
ar
(r ⇧ a)
3⇥0
(2a3 r3 )⌃0
=
ar
(a ⇧ r ⇧ b)
3⇥0 r2
(2a3 b3 )⌃0
=
ar
(r ⌃ b)
3⇥0 r2
E1 =
E2
E3
where ⌃0 , a, and b are constants.
a) Determine the volume charge density in the entire region (0 ⇧ r ⇧ ) by appropriate
use of ⌦ · D = ⌃v . We find ⌃v by taking the divergence of D in all three regions, where
D = ⇥0 E. As D has only a radial component, the divergences become:
⌃
⌥
1 d 2 ⇥
1 d r3 ⌃0
⌃v1 = ⌦ · D1 = 2
r D1 = 2
= ⌃0
(r ⇧ a)
r dr
r dr
3
⌃
⌥
1 d 2 ⇥
1 d 1
3
3
⌃v2 = 2
r D2 = 2
(2a
r )⌃0 = ⌃0
(a ⇧ r ⇧ b)
r dr
r dr 3
⌃
⌥
1 d 2 ⇥
1 d 1
3
3
⌃v3 = 2
r D3 = 2
(2a
b )⌃0 = 0
(r ⌃ b)
r dr
r dr 3
33
3.20b) Find, in terms of given parameters, the total charge, Q, within a sphere of radius r where
r > b. We integrate the charge densities (piecewise) over the spherical volume of radius b:
2⇤
⇤
a
2⇤
⇤
b
⇥
4
Q=
⌃0 r2 sin ⇤ dr d⇤ d⌥
⌃0 r2 sin ⇤ dr d⇤ d⌥ = ⇧ 2a3 b3 ⌃0
3
0
0
0
0
0
a
3.21. Calculate the divergence
of D at the point specified if ⌅
⇤
a) D = (1/z 2 ) 10xyz ax + 5x2 z ay + (2z 3 5x2 y) az at P ( 2, 3, 5): We find
10y
10x2 y
+0+2+
z
z3
⌦·D=
= 8.96
( 2,3,5)
b) D = 5z 2 a⌅ + 10⌃z az at P (3, 45⇥ , 5): In cylindrical coordinates, we have
⌦·D=
1
1 D⇧
Dz
5z 2
(⌃D⌅ ) +
+
=
+ 10⌃
⌃ ⌃
⌃ ⌥
z
⌃
= 71.67
(3, 45 ,5)
c) D = 2r sin ⇤ sin ⌥ ar + r cos ⇤ sin ⌥ a⇥ + r cos ⌥ a⇧ at P (3, 45⇥ , 45⇥ ): In spherical coordinates, we have
1
1
1
D⇧
⌦ · D = 2 (r2 Dr ) +
(sin ⇤D⇥ ) +
r r
r sin ⇤ ⇤
r sin ⇤ ⌥
cos 2⇤ sin ⌥ sin ⌥
= 6 sin ⇤ sin ⌥ +
= 2
sin ⇤
sin ⇤ (3,45 , 45 )
3.22. (a) A flux density field is given as F1 = 5az . Evaluate the outward flux of F1 through the
hemispherical surface, r = a, 0 < ⇤ < ⇧/2, 0 < ⌥ < 2⇧.
The flux integral is
⇥1 =
hem.
F1 · dS =
2⇤
0
⇤/2
0
5 az · ar a2 sin ⇤ d⇤ d⌥ =
✓ ⌘⇣ ◆
cos ⇥
2⇧(5)a2
cos2 ⇤ ⇧⇧⇤/2
= 5⇧a2
⇧
2 0
b) What simple observation would have saved a lot of work in part a? The field is constant,
and so the inward flux through the base of the hemisphere (of area ⇧a2 ) would be equal
in magnitude to the outward flux through the upper surface (the flux through the base
is a much easier calculation).
c) Now suppose the field is given by F2 = 5zaz . Using the appropriate surface integrals,
evaluate the net outward flux of F2 through the closed surface consisting of the hemisphere
of part a and its circular base in the xy plane:
Note that the integral over the base is zero, since F2 = 0 there. The remaining flux
integral is that over the hemisphere:
2⇤
⇥2 =
0
⇤/2
0
5z az · ar a2 sin ⇤ d⇤ d⌥ =
⇤/2
= 10⇧a3
cos2 ⇤ sin ⇤ d⇤ d⌥ =
0
2⇤
0
⇤/2
5(a cos ⇤) cos ⇤ a2 sin ⇤ d⇤ d⌥
0
⇧⇤/2 10
10 3
⇧
⇧a cos3 ⇤⇧
=
⇧a3
3
3
0
d) Repeat part c by using the divergence theorem and an appropriate volume integral:
The divergence of F2 is just dF2 /dz = 5. We then integrate this over the hemisphere
volume, which in this case involves just multiplying 5 by (2/3)⇧a3 , giving the same answer
as in part c.
34
3.23. a) A point charge Q lies at the origin. Show that div D is zero everywhere except at the
origin. For a point charge at the origin we know that D = Q/(4⇧r2 ) ar . Using the formula
for divergence in spherical coordinates (see problem 3.21 solution), we find in this case that
⌃
⌥
1 d
2 Q
⌦·D= 2
r
=0
r dr
4⇧r2
The above is true provided r > 0. When r = 0, we have a singularity in D, so its divergence
is not defined.
b) Replace the point charge with a uniform volume charge density ⌃v0 for 0 < r < a. Relate
⌃v0 to Q and a so that the total charge is the same. Find div D everywhere: To achieve
the same net charge, we require that (4/3)⇧a3 ⌃v0 = Q, so ⌃v0 = 3Q/(4⇧a3 ) C/m3 . Gauss’
law tells us that inside the charged sphere
4⇧r2 Dr =
Thus
4 3
Qr3
⇧r ⌃v0 = 3
3
a
Qr
1 d
Dr =
C/m2 and ⌦ · D = 2
3
4⇧a
r dr
⌃
Qr3
4⇧a3
⌥
=
3Q
4⇧a3
as expected. Outside the charged sphere, D = Q/(4⇧r2 ) ar as before, and the divergence
is zero.
3.24. In a region in free space, electric flux density is found to be:
⌦
⌃0 (z + 2d) az C/m2
( 2d ⇧ z ⇧ 0)
D=
2
⌃0 (z 2d) az C/m
(0 ⇧ z ⇧ 2d)
Everywhere else, D = 0.
a) Using ⌦ · D = ⌃v , find the volume charge density as a function of position everywhere:
Use
⌦
dDz
⌃0 ( 2d ⇧ z ⇧ 0)
⌃v = ⌦ · D =
=
dz
⌃0 (0 ⇧ z ⇧ 2d)
b) determine the electric flux that passes through the surface defined by z = 0, a ⇧ x ⇧ a,
b ⇧ y ⇧ b: In the x-y plane, D evaluates as the constant D(0) = 2d⌃0 az . Therefore the
flux passing through the given area will be
⇥=
a
b
a
b
2d⌃0 dxdy = 8abd ⌃0 C
c) determine the total charge contained within the region a ⇧ x ⇧ a,
b ⇧ y ⇧ b,
d ⇧ z ⇧ d: From part a, we have equal and opposite charge densities above and below
the x-y plane. This means that within a region having equal volumes above and below
the plane, the net charge is zero.
d) determine the total charge contained within the region
0 ⇧ z ⇧ 2d. In this case,
Q=
⌃0 (2a) (2b) (2d) =
a ⇧ x ⇧ a,
b ⇧ y ⇧ b,
8abd ⌃0 C
This is equivalent to the net inward flux of D into the volume, as was found in part b.
35
3.25. Within the spherical shell, 3 < r < 4 m, the electric flux density is given as
D = 5(r
3)3 ar C/m2
a) What is the volume charge density at r = 4? In this case we have
⌃v = ⌦ · D =
1 d 2
5
(r Dr ) = (r
r2 dr
r
3)2 (5r
6) C/m3
which we evaluate at r = 4 to find ⌃v (r = 4) = 17.50 C/m3 .
b) What is the electric flux density at r = 4? Substitute r = 4 into the given expression to
find D(4) = 5 ar C/m2
c) How much electric flux leaves the sphere r = 4? Using the result of part b, this will be
⇥ = 4⇧(4)2 (5) = 320⇧ C
d) How much charge is contained within the sphere, r = 4? From Gauss’ law, this will be
the same as the outward flux, or again, Q = 320⇧ C.
3.26. If we have a perfect gas of mass density ⌃m kg/m3 , and assign a velocity U m/s to each
di⇤erential element, then the mass flow rate is ⌃m U kg/(m2 s). Physical reasoning then
leads to the continuity equation, ⌦ · (⌃m U) =
⌃m / t.
a) Explain in words the physical interpretation of this equation: The quantity ⌃m U is the
flow (or flux) density of mass. Then the divergence of ⌃m U is the outward mass flux per
unit volume at a point. This must be equivalent to the rate of depletion of mass per unit
volume at the same point, as the continuity equation states.
↵
b) Show that s ⌃m U · dS = dM/dt, where M is the total mass of the gas within the
constant closed surface, S, and explain the physical significance of the equation.
Applying the divergence theorem, we have
s
⌃m U · dS =
v
⌦ · (⌃m U) dv =
v
⌃m
dv =
t
d
dt
⌃m dv =
v
dM
dt
This states in large-scale form what was already stated in part a. That is – the net
outward mass flow (in kg/s) through a closed surface is equal to the negative time rate
of change in total mass within the enclosed volume.
3.27. Let D = 5.00r2 ar mC/m2 for r ⇧ 0.08 m and D = 0.205 ar /r2 µC/m2 for r ⌃ 0.08 m (note
error in problem statement).
a) Find ⌃v for r = 0.06 m: This radius lies within the first region, and so
⌃v = ⌦ · D =
1 d 2
1 d
(r Dr ) = 2 (5.00r4 ) = 20r mC/m3
r2 dr
r dr
which when evaluated at r = 0.06 yields ⌃v (r = .06) = 1.20 mC/m3 .
b) Find ⌃v for r = 0.1 m: This is in the region where the second field expression is valid.
The 1/r2 dependence of this field yields a zero divergence (shown in Problem 3.23), and
so the volume charge density is zero at 0.1 m.
36
3.27c) What surface charge density could be located at r = 0.08 m to cause D = 0 for r > 0.08 m?
The total surface charge should be equal and opposite to the total volume charge. The
latter is
2⇤
Q=
0
⇤
0
.08
0
20r(mC/m3 ) r2 sin ⇤ dr d⇤ d⌥ = 2.57 ⇤ 10
So now
2.57
=
4⇧(.08)2
⌃s =
3
mC = 2.57 µC
32 µC/m2
3.28. Repeat Problem 3.8, but use ⌦ · D = ⌃v and take an appropriate volume integral.
We begin by finding the charge density directly through
1 d
⌃v = ⌦ · D = 2
r dr
⌃
⌥
A
2A
r
= 2
r
r
Then, within each spherical shell of unit thickness, the contained charge is
r+1
Q(1) = 4⇧
r
A
(r⇤ )2 dr⇤ = 4⇧A(r + 1
(r⇤ )2
r) = 4⇧A
3.29. In the region of free space that includes the volume 2 < x, y, z < 3,
D=
2
(yz ax + xz ay
z2
2xy az ) C/m2
a) Evaluate the volume integral side of the divergence theorem for the volume defined above:
In cartesian, we find ⌦ · D = 8xy/z 3 . The volume integral side is now
vol
⌦ · D dv =
3
2
3
2
3
2
8xy
dxdydz = (9
z3
4)(9
4)
⌃
1
4
1
9
⌥
= 3.47 C
b. Evaluate the surface integral side for the corresponding closed surface: We call the surfaces
at x = 3 and x = 2 the front and back surfaces respectively, those at y = 3 and y = 2
the right and left surfaces, and those at z = 3 and z = 2 the top and bottom surfaces.
To evaluate the surface integral side, we integrate D · n over all six surfaces and sum the
results. Note that since the x component of D does not vary with x, the outward fluxes
from the front and back surfaces will cancel each other. The same is true for the left
and right surfaces, since Dy does not vary with y. This leaves only the top and bottom
surfaces, where the fluxes are:
D · dS =
3
✓
2
3
2
4xy
dxdy
32
⌘⇣
◆
top
3
✓
2
3
2
4xy
dxdy = (9
22
⌘⇣
◆
bottom
37
4)(9
4)
⌃
1
4
1
9
⌥
= 3.47 C
3.30 a) Use Maxwell’s first equation, ⌦ · D = ⌃v , to describe the variation of the electric field
intensity with x in a region in which no charge density exists and in which a non-homogeneous
dielectric has a permittivity that increases exponentially with x. The field has an x component
only: The permittivity can be written as ⇥(x) = ⇥1 exp( 1 x), where ⇥1 and 1 are constants.
Then
⌦ · D = ⌦ · [⇥(x)E(x)] =
This reduces to
dEx
+
dx
d
[⇥1 e
dx
1 Ex
1x
Ex (x)] = ⇥1
1
e
= 0 ⌥ Ex (x) = E0 e
1x
Ex + e
1x
dEx
=0
dx
1x
where E0 is a constant.
b) Repeat part a, but with a radially-directed electric field (spherical coordinates), in which
again ⌃v = 0, but in which the permittivity decreases exponentially with r. In this case,
the permittivity can be written as ⇥(r) = ⇥2 exp( 2 r), where ⇥2 and 2 are constants.
Then
1 d ⇤ 2
r ⇥2 e
r2 dr
⌦·D = ⌦·[⇥(r)E(r)] =
This reduces to
2r
Er (r) = E0 exp
r2 Er + r2
⌃
2
r
2
r
⌥
dr = E0 exp [ 2 ln r +
2
2
⌥
2
dEr
+
dr
whose solution is
⌃
⌅ ⇥2
Er = 2 2rEr
r
dEr
e
dr
2r
=0
Er = 0
2
r] =
E0
e
r2
2r
where E0 is a constant.
3.31. Given the flux density
16
cos(2⇤) a⇥ C/m2 ,
r
use two di⇤erent methods to find the total charge within the region 1 < r < 2 m, 1 < ⇤ < 2
rad, 1 < ⌥ < 2 rad: We use the divergence theorem and first evaluate the surface integral
side. We are evaluating the net outward flux through a curvilinear “cube”, whose boundaries
are defined by the specified ranges. The flux contributions will be only through the surfaces
of constant ⇤, however, since D has only a ⇤ component. On a constant-theta surface, the
di⇤erential area is da = r sin ⇤drd⌥, where ⇤ is fixed at the surface location. Our flux integral
becomes
D=
D · dS =
=
2
✓1
2
1
2
16
cos(2) r sin(1) drd⌥ +
r
⌘⇣
◆ ✓1
⇥=1
16 [cos(2) sin(1)
cos(4) sin(2)] =
2
1
16
cos(4) r sin(2) drd⌥
r
⌘⇣
◆
⇥=2
3.91 C
We next evaluate the volume integral side of the divergence theorem, where in this case,
⌦·D=
1 d
1 d 16
16 cos 2⇤ cos ⇤
(sin ⇤ D⇥ ) =
cos 2⇤ sin ⇤ = 2
r sin ⇤ d⇤
r sin ⇤ d⇤ r
r
sin ⇤
38
2 sin 2⇤
3.31 (continued) We now evaluate:
vol
⌦ · D dv =
2
1
2
1
2
1
16 cos 2⇤ cos ⇤
r2
sin ⇤
2 sin 2⇤ r2 sin ⇤ drd⇤d⌥
The integral simplifies to
2
1
2
1
2
16[cos 2⇤ cos ⇤
2
2 sin 2⇤ sin ⇤] drd⇤d⌥ = 8
1
1
39
[3 cos 3⇤
cos ⇤] d⇤ =
3.91 C
CHAPTER 4
4.1. The value of E at P (⇧ = 2, ⌃ = 40⇥ , z = 3) is given as E = 100a⇤ 200a⌅ + 300az V/m.
Determine the incremental work required to move a 20 µC charge a distance of 6 µm:
a) in the direction of a⇤ : The incremental work is given by dW =
case, dL = d⇧ a⇤ = 6 ⇤ 10 6 a⇤ . Thus
dW =
(20 ⇤ 10
6
C)(100 V/m)(6 ⇤ 10
6
m) =
12 ⇤ 10
b) in the direction of a⌅ : In this case dL = 2 d⌃ a⌅ = 6 ⇤ 10
dW =
(20 ⇤ 10
6
)( 200)(6 ⇤ 10
6
c) in the direction of az : Here, dL = dz az = 6 ⇤ 10
dW =
(20 ⇤ 10
6
)(300)(6 ⇤ 10
d) in the direction of E: Here, dL = 6 ⇤ 10
aE =
6
6
6
9
J=
12 nJ
a⌅ , and so
) = 2.4 ⇤ 10
6
q E · dL, where in this
8
J = 24 nJ
az , and so
)=
3.6 ⇤ 10
8
J=
36 nJ
aE , where
100a⇤ 200a⌅ + 300az
= 0.267 a⇤
[1002 + 2002 + 3002 ]1/2
0.535 a⌅ + 0.802 az
Thus
dW =
=
(20 ⇤ 10
44.9 nJ
6
)[100a⇤
e) In the direction of G = 2 ax
aG =
200a⌅ + 300az ] · [0.267 a⇤
0.535 a⌅ + 0.802 az ](6 ⇤ 10
3 ay + 4 az : In this case, dL = 6 ⇤ 10
2ax 3ay + 4az
= 0.371 ax
[22 + 32 + 42 ]1/2
6
6
)
6
)
aG , where
0.557 ay + 0.743 az
So now
dW =
=
(20 ⇤ 10
(20 ⇤ 10
6
)[100a⇤
6
) [37.1(a⇤ · ax )
+ 222.9] (6 ⇤ 10
6
200a⌅ + 300az ] · [0.371 ax
)
55.7(a⇤ · ay )
0.557 ay + 0.743 az ](6 ⇤ 10
74.2(a⌅ · ax ) + 111.4(a⌅ · ay )
where, at P , (a⇤ · ax ) = (a⌅ · ay ) = cos(40⇥ ) = 0.766, (a⇤ · ay ) = sin(40⇥ ) = 0.643, and
(a⌅ · ax ) = sin(40⇥ ) = 0.643. Substituting these results in
dW =
(20 ⇤ 10
6
)[28.4
35.8 + 47.7 + 85.3 + 222.9](6 ⇤ 10
40
6
)=
41.8 nJ
4.2. A positive point charge of magnitude q1 lies at the origin. Derive an expression for the
incremental work done in moving a second point charge q2 through a distance dx from the
starting position (x, y, z), in the direction of ax : The incremental work is given by
dW =
q2 E12 · dL
where E12 is the electric field arising from q1 evaluated at the location of q2 , and where
dL = dx ax . Taking the location of q2 at spherical coordinates (r, ⇥, ⌃), we write:
dW =
q2 q1
ar · ( dx)ax
4⌅ 0 r2
where r2 = x2 + y 2 + z 2 , and where ar · ax = sin ⇥ cos ⌃. So
dW =
q2 q1
2
4⌅ 0 (x + y 2 + z 2 )
x2 + y 2
⇠
x2
+
⌫
y2
+
x
z2
⇡⇠
sin
x2
⌫
+
cos ⌅
y2
⇡
dx =
q2 q1 x dx
4⌅ 0 (x2 + y 2 + z 2 )3/2
4.3. If E = 120 a⇤ V/m, find the incremental amount of work done in moving a 50 µm charge a
distance of 2 mm from:
a) P (1, 2, 3) toward Q(2, 1, 4): The vector
along this direction will be Q
↵
from which aP Q = [ax ay + az ]/ 3. We now write
dW =
=
P = (1, 1, 1)
⌦
↵
(ax ay + az
↵
qE · dL = (50 ⇤ 10 ) 120a⇤ ·
(2 ⇤ 10
3
1
(50 ⇤ 10 6 )(120) [(a⇤ · ax ) (a⇤ · ay )] ↵ (2 ⇤ 10 3 )
3
6
3
)
At P , ⌃ = tan 1 (2/1) = 63.4⇥ . Thus (a⇤ · ax ) = cos(63.4) = 0.447 and (a⇤ · ay ) =
sin(63.4) = 0.894. Substituting these, we obtain dW = 3.1 µJ.
b) Q(2, 1, 4) toward P (1, 2, 3): A little thought is in order here: Note that the field has only
a radial component
and does not depend on ⌃ or z. Note also that P and Q are at the
↵
same radius ( 5) from the z axis, but have di⇥erent ⌃ and z coordinates. We could just
as well position the two points at the same z location and the problem would not change.
If this were so, then moving along a straight line
↵ between P and Q would thus involve
moving along a chord of a circle whose radius is 5. Halfway along this line is a point of
symmetry in the field (make a sketch to see this). This means that when starting from
either point, the initial force will be the same. Thus the answer is dW = 3.1 µJ as in part
a. This is also found by going through the same procedure as in part a, but with the
direction (roles of P and Q) reversed.
41
4.4. An electric field in free space is given by E = xax + yay + zaz V/m. Find the work done in
moving a 1µC charge through this field
a) from (1,1,1) to (0,0,0): The work will be
⌦✓ 0
✓
✓
6
W = q E · dL = 10
xdx +
1
0
1
ydy +
✓
0
↵
zdz J = 1.5 µJ
1
b) from (⇧ = 2, ⌃ = 0) to (⇧ = 2, ⌃ = 90⇥ ): The path involves changing ⌃ with ⇧ and z
fixed, and therefore dL = ⇧ d⌃ a⌅ . We set up the integral for the work as
W =
10
6
✓
⇥/2
0
(xax + yay + zaz ) · ⇧ d⌃ a⌅
where ⇧ = 2, ax · a⌅ = sin ⌃, ay · a⌅ = cos ⌃, and az · a⌅ = 0. Also, x = 2 cos ⌃ and
y = 2 sin ⌃. Substitute all of these to get
✓ ⇥/2
⇤
⌅
6
W = 10
(2)2 cos ⌃ sin ⌃ + (2)2 cos ⌃ sin ⌃ d⌃ = 0
0
Given that the field is conservative (and so work is path-independent), can you see a much
easier way to obtain this result?
c) from (r = 10, ⇥ = ⇥0 ) to (r = 10, ⇥ = ⇥0 + 180⇥ ): In this case, we are moving only in the
a direction. The work is set up as
✓ 0 +⇥
6
W = 10
(xax + yay + zaz ) · r d⇥ a
0
Now, substitute the following relations: r = 10, x = r sin ⇥ cos ⌃, y = r sin ⇥ sin ⌃, z =
r cos ⇥, ax · a = cos ⇥ cos ⌃, ay · a = cos ⇥ sin ⌃, and az · a = sin ⇥. Obtain
✓ 0 +⇥
⇤
⌅
6
W = 10
(10)2 sin ⇥ cos ⇥ cos2 ⌃ + sin ⇥ cos ⇥ sin2 ⌃ cos ⇥ sin ⇥ d⇥ = 0
0
where we use cos2 ⌃ + sin2 ⌃ = 1.
⇣P
4.5. Compute the value of A G · dL for G = 2yax with A(1, 1, 2) and P (2, 1, 2) using the path:
a) straight-line segments A(1, 1, 2) to B(1, 1, 2) to P (2, 1, 2): In general we would have
✓ P
✓ P
G · dL =
2y dx
A
A
The change in x occurs when moving between B and P , during which y = 1. Thus
✓ P
✓ P
✓ 2
G · dL =
2y dx =
2(1)dx = 2
A
B
1
b) straight-line segments A(1, 1, 2) to C(2, 1, 2) to P (2, 1, 2): In this case the change in
x occurs when moving from A to C, during which y = 1. Thus
✓ P
✓ C
✓ 2
G · dL =
2y dx =
2( 1)dx = 2
A
A
1
42
4.6. An electric field in free space is given as E = x âx + 4z ây + 4y âz . Given V (1, 1, 1) = 10 V.
Determine V (3, 3, 3). The potential di⇥erence is expressed as
V (3, 3, 3)
✓
V (1, 1, 1) =
3,3,3
1,1,1
⌦✓ 3
=
(x âx + 4z ây + 4y âz ) · (dx ax + dy ay + dz az )
x dx +
1
✓
3
4z dy +
1
✓
3
4y dz
1
↵
We choose the following path: 1) move along x from 1 to 3; 2) move along y from 1 to 3,
holding x at 3 and z at 1; 3) move along z from 1 to 3, holding x at 3 and y at 3. The integrals
become:
⌦✓ 3
↵
✓ 3
✓ 3
V (3, 3, 3) V (1, 1, 1) =
x dx +
4(1) dy +
4(3) dz = 36
1
So
V (3, 3, 3) =
1
1
36 + V (1, 1, 1) =
36 + 10 =
26 V
3
4.7. Let
⇣ G = 3xy ax + 2zay . Given an initial point P (2, 1, 1) and a final point Q(4, 3, 1), find
G · dL using the path:
a) straight line: y = x
✓
1, z = 1: We obtain:
✓
G · dL =
4
✓
3xy dx +
2
2
3
✓
2z dy =
1
4
3x(x
1) dx +
2
2
✓
3
2(1) dy = 90
1
b) parabola: 6y = x2 + 2, z = 1: We obtain:
✓
G · dL =
✓
4
3xy 2 dx +
2
✓
3
2z dy =
1
✓
4
2
1
x(x2 + 2)2 dx +
12
✓
3
2(1) dy = 82
1
4.8. Given E = xax +yay , a) find the work involved in moving a unit
↵ positive charge on a circular
arc, the circle centered at the origin, from x = a to x = y = a/ 2.
In moving along the arc, we start at ⌃ = 0 and move to ⌃ = ⌅/4. The setup is
W =
=
q
✓
✓
E · dL =
⇥/4
✓
⇥/4
0
E · ad⌃ a⌅ =
✓
2a2 sin ⌃ cos ⌃ d⌃ =
0
⇥/4
0
where q = 1, x = a cos ⌃, and y = a sin ⌃.
✓
⇥/4
( x ax · a⌅ + y ay · a⌅ )a d⌃
⇠ ⌫ ⇡
⇠ ⌫ ⇡
0
sin ⌅
cos ⌅
⇧⇥/4
a
⇧
a2 sin(2⌃) d⌃ =
cos(2⌃)⇧
=
2
0
2
a2
2
Note that the field is conservative, so we would get the same result by integrating along
a two-segment path over x and y as shown:
W =
✓
E · dL =
✓
⌥
a/ 2
a
( x) dx +
✓
0
43
⌥
a/ 2
y dy =
a2 /2
4.8b) Verify that the work done in moving the charge around the full circle from x = a is zero: In
this case, the setup is the same, but the integration limits change:
✓
W =
2⇥
⇧2⇥
a2
⇧
cos(2⌃)⇧ = 0
2
0
a2 sin(2⌃) d⌃ =
0
4.9. A uniform surface charge density of 20 nC/m2 is present on the spherical surface r = 0.6 cm
in free space.
a) Find the absolute potential at P (r = 1 cm, ⇥ = 25⇥ , ⌃ = 50⇥ ): Since the charge density
is uniform and is spherically-symmetric, the angular coordinates do not matter. The
potential function for r > 0.6 cm will be that of a point charge of Q = 4⌅a2 ⇧s , or
V (r) =
4⌅(0.6 ⇤ 10 2 )2 (20 ⇤ 10
4⌅ 0 r
9
)
=
0.081
V with r in meters
r
At r = 1 cm, this becomes V (r = 1 cm) = 8.14 V
b) Find VAB given points A(r = 2 cm, ⇥ = 30⇥ , ⌃ = 60⇥ ) and B(r = 3 cm, ⇥ = 45⇥ , ⌃ = 90⇥ ):
Again, the angles do not matter because of the spherical symmetry. We use the part a
result to obtain
⌦
↵
1
1
VAB = VA VB = 0.081
= 1.36 V
0.02 0.03
4.10. A sphere of radius a carries a surface charge density of ⇧s0 C/m2 .
a) Find the absolute potential at the sphere surface: The setup for this is
V0 =
✓
a
⌃
where, from Gauss’ law:
E=
So
✓
V0 =
a
⌃
E · dL
a2 ⇧s0
ar
2
0r
V/m
a2 ⇧s0
a2 ⇧s0 ⇧⇧a
a⇧s0
a
·
a
dr
=
V
⇧ =
r
r
2
0r
0r ⌃
0
b) A grounded conducting shell of radius b where b > a is now positioned around the charged
sphere. What is the potential at the inner sphere surface in this case? With the outer
sphere grounded, the field exists only between the surfaces, and is zero for r > b. The
potential is then
V0 =
✓
b
a
⌦
a2 ⇧s0
a2 ⇧s0 ⇧⇧a a2 ⇧s0 1
ar · ar dr =
⇧ =
2
a
0r
0r b
0
44
1
b
↵
V
4.11. Let a uniform surface charge density of 5 nC/m2 be present at the z = 0 plane, a uniform line
charge density of 8 nC/m be located at x = 0, z = 4, and a point charge of 2 µC be present
at P (2, 0, 0). If V = 0 at M (0, 0, 5), find V at N (1, 2, 3): We need to find a potential function
for the combined charges which is zero at M . That for the point charge we know to be
Vp (r) =
Q
4⌅ 0 r
Potential functions for the sheet and line charges can be found by taking indefinite integrals
of the electric fields for those distributions. For the line charge, we have
✓
⇧l
⇧l
Vl (⇧) =
d⇧ + C1 =
ln(⇧) + C1
2⌅ 0 ⇧
2⌅ 0
For the sheet charge, we have
✓
Vs (z) =
⇧s
dz + C2 =
2 0
⇧s
z + C2
2 0
The total potential function will be the sum of the three. Combining the integration constants,
we obtain:
Q
⇧l
⇧s
V =
ln(⇧)
z+C
4⌅ 0 r 2⌅ 0
2 0
The terms in this expression are not referenced to a common origin, since the charges are at
di⇥erent positions. The parameters r, ⇧, and z are scalar distances from the charges, and will
be treated as
↵ such here.↵To evaluate the constant, C, we first look at point M , where VT = 0.
At M , r = 22 + 52 = 29, ⇧ = 1, and z = 5. We thus have
0=
2 ⇤ 10 6
↵
4⌅ 0 29
At point N , r =
VN =
8 ⇤ 10
2⌅ 0
9
ln(1)
5 ⇤ 10
2 0
9
5+C
⌥ C=
1.93 ⇤ 103 V
↵
↵
↵
1 + 4 + 9 = 14, ⇧ = 2, and z = 3. The potential at N is thus
2 ⇤ 10 6
↵
4⌅ 0 14
8 ⇤ 10
2⌅ 0
9
↵
ln( 2)
5 ⇤ 10
2 0
45
9
(3)
1.93 ⇤ 103 = 1.98 ⇤ 103 V = 1.98 kV
4.12. In spherical coordinates, E = 2r/(r2 + a2 )2 ar V/m. Find the potential at any point, using
the reference
a) V = 0 at infinity: We write in general
✓
V (r) =
With a zero reference at r ⌃
2r dr
1
+C = 2
+C
2
2
+a )
r + a2
(r2
, C = 0 and therefore V (r) = 1/(r2 + a2 ).
b) V = 0 at r = 0: Using the general expression, we find
V (0) =
1
+C =0 ⌥ C =
a2
V (r) =
1
2
r + a2
Therefore
1
a2
1
r2
=
a2
a2 (r2 + a2 )
c) V = 100V at r = a: Here, we find
V (a) =
Therefore
V (r) =
r2
1
+ C = 100 ⌥ C = 100
2a2
1
+ a2
1
2a2
1
a2 r2
+ 100 = 2 2
+ 100
2
2a
2a (r + a2 )
4.13. Three identical point charges of 4 pC each are located at the corners of an equilateral triangle
0.5 mm on a side in free space. How much work must be done to move one charge to a point
equidistant from the other two and on the line joining them? This will be the magnitude of
the charge times the potential di⇥erence between the finishing and starting positions, or
⌦
(4 ⇤ 10 12 )2 1
W =
2⌅ 0
2.5
↵
1
⇤ 104 = 5.76 ⇤ 10
5
10
J = 576 pJ
4.14. Given the electric field E = (y + 1)ax + (x 1)ay + 2az , find the potential di⇥erence between
the points
a) (2,-2,-1) and (0,0,0): We choose a path along which motion occurs in one coordinate
direction at a time. Starting at the origin, first move along x from 0 to 2, where y = 0;
then along y from 0 to 2, where x is 2; then along z from 0 to 1. The setup is
Vb
Va =
✓
2
0
⇧
⇧
(y + 1)⇧
y=0
dx
✓
2
(x
0
b) (3,2,-1) and (-2,-3,4): Following similar reasoning,
Vb
Va =
✓
3
2
⇧
⇧
(y + 1)⇧
y= 3
dx
46
✓
2
3
(x
⇧
⇧
1)⇧
x=2
⇧
⇧
1)⇧
x=3
dy
✓
1
2 dz = 2
0
dy
✓
4
1
2 dz = 10
4.15. Two uniform line charges, 8 nC/m each, are located at x = 1, z = 2, and at x = 1, y = 2
in free space. If the potential at the origin is 100 V, find V at P (4, 1, 3): The net potential
function for the two charges would in general be:
V =
At the origin, R1 = R2 =
⇧l
ln(R1 )
2⌅ 0
↵
5, and V = 100 V. Thus, with ⇧l = 8 ⇤ 10
100 =
2
↵
(8 ⇤ 10 9 )
ln( 5) + C
2⌅ 0
At P (4, 1, 3), R1 = |(4, 1, 3) (1, 1, 2)| =
VP =
⇧l
ln(R2 ) + C
2⌅ 0
9
,
⌥ C = 331.6 V
↵
↵
10 and R2 = |(4, 1, 3) ( 1, 2, 3)| = 26. Therefore
↵ 
(8 ⇤ 10 9 ) ◆ ↵
ln( 10) + ln( 26) + 331.6 =
2⌅ 0
68.4 V
4.16. A spherically-symmetric charge distribution in free space (with a < r <
) – note typo
in problem statement, which says (0 < r <
) – is known to have a potential function
V (r) = V0 a2 /r2 , where V0 and a are constants.
a) Find the electric field intensity: This is found through
E=
V =
dV
a2
ar = 2V0 3 ar V/m
dr
r
b) Find the volume charge density: Use Maxwell’s first equation:
⇧v =
·D=
1 d
· ( 0 E) = 2
r dr
⌦
↵
a2
2
r 2 0 V0 3
=
r
2 0 V0
a2
C/m3
r4
c) Find the charge contained inside radius a: Here, we do not know the charge density inside
radius a, but we do know the flux density at that radius. We use Gauss’ law to integrate
D over the spherical surface at r = a to find the charge enclosed:
Qencl =
✏
r=a
D · dS = 4⌅a2 D|r=a = 4⌅a2 2 0 V0
a2
a3
= 8⌅ 0 aV0 C
d) Find the total energy stored in the charge (or equivalently, in its electric field) in the
region (a < r < ). We integrate the energy density in the field over the region:
✓
✓ 2⇥ ✓ ⇥ ✓ ⌃
1
a4
We =
D · E dv =
2 0 V02 6 r2 sin ⇥ dr d⇥ d⌃
r
v 2
0
0
a
⇧⌃
1
⇧
= 8⌅ V02 0 a4 3 ⇧ = 8⌅ 0 aV02 /3 J
3r a
47
4.17. Uniform surface charge densities of 6 and 2 nC/m2 are present at ⇧ = 2 and 6 cm respectively,
in free space. Assume V = 0 at ⇧ = 4 cm, and calculate V at:
a) ⇧ = 5 cm: Since V = 0 at 4 cm, the potential at 5 cm will be the potential di⇥erence
between points 5 and 4:
V5 =
✓
5
4
✓
E · dL =
5
4
(.02)(6 ⇤ 10
a⇧sa
d⇧ =
0⇧
9
)
ln
0
5
4
=
3.026 V
b) ⇧ = 7 cm: Here we integrate piecewise from ⇧ = 4 to ⇧ = 7:
✓
V7 =
6
4
a⇧sa
d⇧
0⇧
✓
7
6
(a⇧sa + b⇧sb )
d⇧
0⇧
With the given values, this becomes
V7 =
=
⌦
(.02)(6 ⇤ 10
0
9.678 V
9
)
↵
⌦
6
4
ln
(.02)(6 ⇤ 10
9
) + (.06)(2 ⇤ 10
9
0
)
↵
ln
7
6
4.18. Find the potential at the origin produced by a line charge ⇧L = kx/(x2 + a2 ) extending along
the x axis from x = a to + , where a > 0. Assume a zero reference at infinity.
Think of the line charge as an array of point charges, each of charge dq = ⇧L dx, and each
having potential at the origin of dV = ⇧L dx/(4⌅ 0 x). The total potential at the origin is
then the sum of all these potentials, or
V =
✓
⌃
a
⇧L dx
=
4⌅ 0 x
✓
⌃
a
k dx
k
=
tan
4⌅ 0 (x2 + a2 )
4⌅ 0 a
1
⌃ x ⌥⌃
a
a
=
k ◆⌅
4⌅ 0 a 2
⌅
k
=
4
16 0 a
4.19. The annular surface, 1 cm < ⇧ < 3 cm, z = 0, carries the nonuniform surface charge density
⇧s = 5⇧ nC/m2 . Find V at P (0, 0, 2 cm) if V = 0 at infinity: We use the superposition integral
form:
✓ ✓
⇧s da
VP =
4⌅ 0 |r r⇧ |
where r = zaz and r⇧ = ⇧a⇤ . We integrate over the surface of the annular region, with
da = ⇧ d⇧ d⌃. Substituting the given values, we find
VP =
✓
2⇥
0
✓
.03
(5 ⇤ 10
.01
4⌅
0
9
)⇧2 d⇧ d⌃
⇧2 + z 2
Substituting z = .02, and using tables, the integral evaluates as
⌦
(5 ⇤ 10
VP =
2 0
9
)
↵⌦
⇧
2
⇧2
+
(.02)2
ln(⇧ +
2
(.02)2
48
⇧2
+
↵.03
(.02)2 )
.01
= .081 V
4.20. In a certain medium, the electric potential is given by
V (x) =
⇧0
1
a 0
e
ax
where ⇧0 and a are constants.
a) Find the electric field intensity, E:
⌦
d ⇧0
E ==
V =
1
dx a 0
ax
e
↵
⇥
⇥
ax =
⇧0
e
ax
0
ax V/m
b) find the potential di⇥erence between the points x = d and x = 0:
Vd0 = V (d)
⇧0
1
a 0
V (0) =
c) if the medium permittivity is given by (x) =
the volume charge density, ⇧v , in the region:
D= E=
Then ⇧v =
⇧0
ax
0e
e
ax
0e ,
ax
0
ax
e
ad
⇥
V
find the electric flux density, D, and
=
⇧0 ax C/m2
· D = 0.
d) Find the stored energy in the region (0 < x < d), (0 < y < 1), (0 < z < 1):
We =
✓
v
1
D · E dv =
2
✓
1
0
✓
1
0
✓
0
d
⇧20
e
2 0
ax
⇧20
dx dy dz =
e
2 0a
⇧d
ax ⇧
⇧ =
0
⇧20
1
2 0a
e
ad
⇥
J
4.21. Let V = 2xy 2 z 3 +3 ln(x2 +2y 2 +3z 2 ) V in free space. Evaluate each of the following quantities
at P (3, 2, 1):
a) V : Substitute P directly to obtain: V = 15.0 V
b) |V |. This will be just 15.0 V.
c) E: We have
⌦
⇧
⇧
6x
12y
⇧
⇧
E⇧ =
V⇧ =
2y 2 z 3 + 2
ax + 4xyz 3 + 2
x + 2y 2 + 3z 2
x + 2y 2 + 3z 2
P
P
↵
18z
2 2
+ 6xy z + 2
az
= 7.1ax + 22.8ay 71.1az V/m
x + 2y 2 + 3z 2
P
d) |E|P : taking the magnitude of the part c result, we find |E|P = 75.0 V/m.
e) aN : By definition, this will be
⇧
⇧
aN ⇧ =
P
⇧
⇧
f) D: This is D⇧ =
P
⇧
⇧
0 E⇧
P
E
=
|E|
0.095 ax
= 62.8 ax + 202 ay
49
0.304 ay + 0.948 az
629 az pC/m2 .
ay
4.22. A line charge of infinite length lies along the z axis, and carries a uniform linear charge density
of ⇧◆ C/m. A perfectly-conducting cylindrical shell, whose axis is the z axis, surrounds the
line charge. The cylinder (of radius b), is at ground potential. Under these conditions, the
potential function inside the cylinder (⇧ < b) is given by
⇧◆
ln(⇧)
2⌅ 0
V (⇧) = k
where k is a constant.
a) Find k in terms of given or known parameters: At radius b,
⇧◆
⇧◆
ln(b) = 0 ⌥ k =
ln(b)
2⌅ 0
2⌅ 0
V (b) = k
b) find the electric field strength, E, for ⇧ < b:
Ein =
⌦
d
⇧◆
ln(b)
d⇧ 2⌅ 0
V =
↵
⇧◆
⇧◆
ln(⇧) a⇤ =
a⇤ V/m
2⌅ 0
2⌅ 0 ⇧
c) find the electric field strength, E, for ⇧ > b: Eout = 0 because the cylinder is at ground
potential.
d) Find the stored energy in the electric field per unit length in the z direction within the
volume defined by ⇧ > a, where a < b:
We =
✓
v
1
D · E dv =
2
✓
0
1
✓
2⇥
0
✓
b
a
⇧2◆
8⌅ 2
0
⇧2
⇧ d⇧ d⌃ dz =
⇧2◆
ln
4⌅ 0
b
a
J
4.23. It is known that the potential is given as V = 80⇧.6 V. Assuming free space conditions, find:
a) E: We find this through
E=
dV
a⇤ =
d⇧
V =
48⇧
b) the volume charge density at ⇧ = .5 m: Using D =
through
⇧
⇧
⇧v ⇧ = [
.5
· D].5 =
1
⇧
⇧
d
⇧
(⇧D⇤ ) ⇧ =
d⇧
.5
.4
V/m
0 E,
28.8 0 ⇧
we find the charge density
⇧
⇧ =
1.4 ⇧
.5
673 pC/m3
c) the total charge lying within the closed surface ⇧ = .6, 0 < z < 1: The easiest way to do
this calculation is to evaluate D⇤ at ⇧ = .6 (noting
⇧ that it is constant), and then multiply
⇧
by the cylinder area: Using part a, we have D⇤ ⇧ = 48 0 (.6) .4 = 521 pC/m2 . Thus
Q=
2⌅(.6)(1)521 ⇤ 10
12
C=
1.96 nC.
50
.6
4.24. A certain spherically-symmetric charge configuration in free space produces an electric field
given in spherical coordinates by:
(⇧0 r2 )/(100 0 ) ar V/m
(100⇧0 )/( 0 r2 ) ar V/m
E(r) =
(r ⌅ 10)
(r ⇧ 10)
where ⇧0 is a constant.
a) Find the charge density as a function of position:
⇧v (r ⌅ 10) =
· ( 0 E1 ) =
⇧v (r ⇧ 10) =
1 d
r2 dr
· ( 0 E2 ) =
r2
1 d
r2 dr
⇧0 r2
100
r2
=
100⇧0
r2
⇧0 r
C/m3
25
=0
b) find the absolute potential as a function of position in the two regions, r ⌅ 10 and r ⇧ 10:
✓
✓ r
100⇧0
⇧0 (r⇧ )2
a
·
a
dr
ar · ar dr⇧
r
r
2
r
100
0
0
⌃
10
⌅
100⇧0 ⇧⇧10 ⇧0 (r⇧ )3 ⇧⇧r
10⇧0 ⇤
=
4 (10 3 r3 ) V
⇧
⇧ =
300 0 10
3 0
0r ⌃
V (r ⌅ 10) =
✓
V (r ⇧ 10) =
10
100⇧0
100⇧0 ⇧⇧r
100⇧0
⇧
a
·
a
dr
=
V
⇧ =
r
r
⇧
2
⇧
⌃
0 (r )
0r
0r
r
⌃
c) check your result of part b by using the gradient:
E1 =
⌦
d 10⇧0 ⇤
4
dr 3 0
V (r ⌅ 10) =
↵
⌅
10⇧0
⇧0 r2
(10 r ) ar =
(3r2 )(10 3 ) ar =
ar
3 0
100 0
⌦
↵
d 100⇧0
100⇧0
V (r ⇧ 10) =
ar =
ar
2
dr
0r
0r
E2 =
3 3
d) find the stored energy in the charge by an integral of the form of Eq. (42) (not Eq. (43)):
1
We =
2
=
✓
⇧v V dv =
v
4⌅⇧20
150
0
✓
0
✓
0
10
2⇥
⌦
40r3
↵
⌅ 2
10⇧0 ⇤
3 3
4 (10 r ) r sin ⇥ dr d⇥ d⌃
3 0
0
0
↵
⌦
↵10
r6
4⌅⇧20
r7
⇧2
4
dr =
10r
= 7.18 ⇤ 103 0
100
150 0
700 0
0
✓
⇥
✓
10
1 ⇧0 r
2 25
⌦
e) Find the stored energy in the field by an integral of the form of Eq. (44) (not Eq. (45)).
✓
✓
1
1
We =
D1 · E1 dv +
D2 · E2 dv
(r⇤10) 2
(r⌅10) 2
✓ 2⇥ ✓ ⇥ ✓ 10
✓ 2⇥ ✓ ⇥ ✓ ⌃ 4 2
⇧20 r4
10 ⇧0 2
2
=
r sin ⇥ dr d⇥ d⌃ +
r sin ⇥ dr d⇥ d⌃
4
4
(2 ⇤ 10 ) 0
0
0
0
0
0
10 2 0 r
⌦
⌦
↵
✓ 10
✓ ⌃ ↵
2
2⌅⇧20
dr
2⌅⇧20 1
4
6
4
3
3
3 ⇧0
=
10
r dr + 10
=
(10
)
+
10
=
7.18
⇤
10
2
7
0
0
0
0
10 r
51
4.25. Within the cylinder ⇧ = 2, 0 < z < 1, the potential is given by V = 100 + 50⇧ + 150⇧ sin ⌃ V.
a) Find V , E, D, and ⇧v at P (1, 60⇥ , 0.5) in free space: First, substituting the given point,
we find VP = 279.9 V. Then,
E=
1 V
a⌅ =
⇧ ⌃
V
a⇤
⇧
V =
Evaluate the above at P to find EP =
Now D =
⇧v =
0 E,
·D =
so DP =
1
⇧
1.59a⇤
179.9a⇤
[150 cos ⌃] a⌅
75.0a⌅ V/m
.664a⌅ nC/m2 . Then
d
1 D⌅
(⇧D⇤ ) +
=
d⇧
⇧ ⌃
At P , this is ⇧vP =
[50 + 150 sin ⌃] a⇤
⌦
↵
1
1
(50 + 150 sin ⌃) + 150 sin ⌃
⇧
⇧
0
=
50
⇧
0
C
443 pC/m3 .
b) How much charge lies within the cylinder? We will integrate ⇧v over the volume to obtain:
Q=
✓
0
1
✓
0
2⇥
✓
2
0
50 0
⇧ d⇧ d⌃ dz =
⇧
2⌅(50) 0 (2) =
5.56 nC
4.26. Let us assume that we have a very thin, square, imperfectly conducting plate 2m on a side,
located in the plane z = 0 with one corner at the origin such that it lies entirely within the
first quadrant. The potential at any point in the plate is given as V = e x sin y.
a) An electron enters the plate at x = 0, y = ⌅/3 with zero initial velocity; in what direction
is its initial movement? We first find the electric field associated with the given potential:
E=
V =
e
x
[sin y ax
cos y ay ]
Since we have an electron, its motion
↵ is opposite that of the field, so the direction on
entry is that of E at (0, ⌅/3), or 3/2 ax 1/2 ay .
b) Because of collisions with the particles in the plate, the electron achieves a relatively low
velocity and little acceleration (the work that the field does on it is converted largely into
heat). The electron therefore moves approximately along a streamline. Where does it
leave the plate and in what direction is it moving at the time? Considering the result
of part a, we would expect the exit to occur along the bottom edge of the plate. The
equation of the streamline is found through
✓
Ey
dy
cos y
=
=
⌥ x=
tan y dy + C = ln(cos y) + C
Ex
dx
sin y
At the entry point (0, ⌅/3), we have 0 = ln[cos(⌅/3)] + C, from which C = 0.69. Now,
along the bottom edge (y = 0), we find x = 0.69, and so the exit point is (0.69, 0). From
the field expression evaluated at the exit point, we find the direction on exit to be ay .
52
4.27. Two point charges, 1 nC at (0, 0, 0.1) and
a) Calculate V at P (0.3, 0, 0.4): Use
VP =
1 nC at (0, 0, 0.1), are in free space.
q
4⌅ 0 |R+ |
q
4⌅ 0 |R |
where R+ = (.3, 0, .3) and R = (.3, 0, .5), so that |R+ | = 0.424 and |R | = 0.583. Thus
10
VP =
4⌅
9
0
⌦
1
.424
↵
1
= 5.78 V
.583
b) Calculate |E| at P : Use
EP =
q(.3ax + .3az )
4⌅ 0 (.424)3
q(.3ax + .5az )
10
=
4⌅ 0 (.583)3
4⌅
9
0
[2.42ax + 1.41az ] V/m
Taking the magnitude of the above, we find |EP | = 25.2 V/m.
c) Now treat the two charges as↵a dipole at the origin and find V at P : In spherical coordinates, P is located at r = .32 + .42 = .5 and ⇥ = sin 1 (.3/.5) = 36.9⇥ . Assuming a
dipole in far-field, we have
VP =
qd cos ⇥
10
=
2
4⌅ 0 r
9
(.2) cos(36.9⇥ )
= 5.76 V
4⌅ 0 (.5)2
4.28. Use the electric field intensity of the dipole (Sec. 4.7, Eq. (36)) to find the di⇥erence in
potential between points at ⇥a and ⇥b , each point having the same r and ⌃ coordinates. Under
what conditions does the answer agree with Eq. (34), for the potential at ⇥a ?
We perform a line integral of Eq. (36) along an arc of constant r and ⌃:
Vab =
✓
a
b
qd
[2 cos ⇥ ar + sin ⇥ a ] · a r d⇥ =
4⌅ 0 r3
qd
=
[cos ⇥a
4⌅ 0 r2
✓
a
b
qd
sin ⇥ d⇥
4⌅ 0 r2
cos ⇥b ]
This result agrees with Eq. (34) if ⇥a (the ending point in the path) is 90⇥ (the xy plane).
Under this condition, we note that if ⇥b > 90⇥ , positive work is done when moving (against
the field) to the xy plane; if ⇥b < 90⇥ , negative work is done since we move with the field.
4.29. A dipole having a moment p = 3ax 5ay + 10az nC · m is located at Q(1, 2, 4) in free space.
Find V at P (2, 3, 4): We use the general expression for the potential in the far field:
V =
where r
r⇧ = P
p · (r
4⌅ 0 |r
r⇧ )
r⇧ |3
Q = (1, 1, 8). So
VP =
(3ax
5ay + 10az ) · (ax + ay + 8az ) ⇤ 10
4⌅ 0 [12 + 12 + 82 ]1.5
53
9
= 1.31 V
4.30. A dipole for which p = 10 0 az C · m is located at the origin. What is the equation of the
surface on which Ez = 0 but E = 0?
First we find the z component:
10
5 ⇤
[2
cos
⇥
(a
·
a
)
+
sin
⇥
(a
·
a
)]
=
2 cos2 ⇥
r
z
z
4⌅r3
2⌅r3
⇤
⌅
This will be zero when 2 cos2 ⇥ sin2 ⇥ = 0. Using identities, we write
Ez = E · az =
2 cos2 ⇥
sin2 ⇥ =
sin2 ⇥
⌅
1
[1 + 3 cos(2⇥)]
2
The above becomes zero on the cone surfaces, ⇥ = 54.7⇥ and ⇥ = 125.3⇥ .
4.31. A potential field in free space is expressed as V = 20/(xyz) V.
a) Find the total energy stored within the cube 1 < x, y, z < 2. We integrate the energy
density over the cube volume, where wE = (1/2) 0 E · E, and where
E=
V = 20
⌦
↵
1
1
1
a
+
a
+
a
x
y
z V/m
x2 yz
xy 2 z
xyz 2
The energy is now
WE = 200
0
✓
2
1
✓
2
1
✓
⌦
2
1
1
1
1
+ 2 4 2+ 2 2 4
4
2
2
x y z
x y z
x y z
↵
dx dy dz
The integral evaluates as follows:
WE = 200
0
✓
2
1
= 200
0
✓
= 200
0
2
2
1
✓
2
1
1
✓
✓
✓
⌦
1
2
⌦
⌦
⌦
1
3
7
24
7
24
1
x3 y 2 z 2
1
+
2
y z2
1
yz 2
7
1
7
= 200 0
+
2
48 z
48
1
⌦ ↵
7
= 200 0 (3)
= 387 pJ
96
2
1
6
1
xy 4 z 2
1
2
1
xy 2 z 4
1
+
4
y z2
1
1
2
y3 z2
1
+
z2
1
4
1
z4
1
2
↵
↵2
dy dz
1
↵
1
dy dz
y2 z4
↵2
1
dz
yz 4 1
dz
b) What value would be obtained by assuming a uniform energy density equal to the value
at the center of the cube? At C(1.5, 1.5, 1.5) the energy density is
⌦
↵
1
wE = 200 0 (3)
= 2.07 ⇤ 10
(1.5)4 (1.5)2 (1.5)2
10
J/m3
This, multiplied by a cube volume of 1, produces an energy value of 207 pJ.
54
4.32. Using Eq. (36), a) find the energy stored in the dipole field in the region r > a:
We start with
qd
[2 cos ⇥ ar + sin ⇥ a ]
4⌅ 0 r3
E(r, ⇥) =
Then the energy will be
We =
✓
vol
1
0 E · E dv =
2
✓
0
2⇥
✓
0
⇥
✓
⌃
a
⌅ 2
(qd)2 ⇤
2
2
4
cos
⇥
+
sin
⇥
r sin ⇥ dr d⇥ d⌃
32⌅ 2 0 r6 ⇠
⌫
⇡
3 cos2 +1
✓
⌅
2⌅(qd) 1 ⇧⇧⌃ ⇥ ⇤
(qd)2 ⇤
2
=
3
cos
⇥
+
1
sin
⇥
d⇥
=
cos3 ⇥
⇧
32⌅ 2 0 3r3 a 0
48⌅ 2 0 a3 ⇠
⌫
2
cos ⇥
4
=
(qd)2
J
12⌅ 0 a3
⌅⇥
0
⇡
b) Why can we not let a approach zero as a limit? From the above result, a singularity in the
energy occurs as a ⌃ 0. More importantly, a cannot be too small, or the original far-field
assumption used to derive Eq. (36) (a >> d) will not hold, and so the field expression
will not be valid.
4.33. A copper sphere of radius 4 cm carries a uniformly-distributed total charge of 5 µC in free
space.
a) Use Gauss’ law to find D external to the sphere: with a spherical Gaussian surface at
radius r, D will be the total charge divided by the area of this sphere, and will be ar directed. Thus
Q
5 ⇤ 10 6
D=
a
=
ar C/m2
r
4⌅r2
4⌅r2
b) Calculate the total energy stored in the electrostatic field: Use
✓ 2⇥ ✓ ⇥ ✓ ⌃
1
1 (5 ⇤ 10 6 )2 2
D · E dv =
r sin ⇥ dr d⇥ d⌃
2
4
vol 2
0
0
.04 2 16⌅ 0 r
✓
1 (5 ⇤ 10 6 )2 ⌃ dr
25 ⇤ 10 12 1
= (4⌅)
=
= 2.81 J
2
2
16⌅ 2 0
8⌅ 0
.04
.04 r
WE =
✓
c) Use WE = Q2 /(2C) to calculate the capacitance of the isolated sphere: We have
C=
Q2
(5 ⇤ 10 6 )2
=
= 4.45 ⇤ 10
2WE
2(2.81)
55
12
F = 4.45 pF
4.34. A sphere of radius a contains volume charge of uniform density ⇧0 C/m3 . Find the total stored
energy by applying
a) Eq. (43): We first need the potential everywhere inside the sphere. The electric field
inside and outside is readily found from Gauss’s law:
E1 =
⇧0 r
⇧0 a3
ar r ⌅ a and E2 =
ar r ⇧ a
3 0
3 0 r2
The potential at position r inside the sphere is now the work done in moving a unit
positive point charge from infinity to position r:
✓
V (r) =
a
E2 · ar dr
⌃
✓
r
a
✓
E1 · ar dr =
⇧
a
⇧0 a3
dr
3 0 r2
⌃
✓
a
r
⇧0 r⇧ ⇧
⇧0
dr =
3a2
3 0
6 0
r2
⇥
Now, using this result in (43) leads to the energy associated with the charge in the sphere:
We =
✓
1
2
✓
2⇥
0
✓
⇥
0
0
a
⇧20
3a2
6 0
⇥
⌅⇧0
r2 r2 sin ⇥ dr d⇥ d⌃ =
3 0
✓
0
a
3a2 r2
⇥
4⌅a5 ⇧20
r4 dr =
15 0
b) Eq. (45): Using the given fields we find the energy densities
we1 =
1
⇧20 r2
1
⇧2 a6
r ⌅ a and we2 = 0 E2 · E2 = 0 4 r ⇧ a
0 E1 · E1 =
2
18 0
2
18 0 r
We now integrate these over their respective volumes to find the total energy:
We =
✓
0
2⇥
✓
0
⇥
✓
0
a
⇧20 r2 2
r sin ⇥ dr d⇥ d⌃ +
18 0
56
✓
0
2⇥
✓
0
⇥
✓
a
⌃
⇧20 a6 2
4⌅a5 ⇧20
r
sin
⇥
dr
d⇥
d⌃
=
18 0 r4
15 0
4.35. Four 0.8 nC point charges are located in free space at the corners of a square 4 cm on a side.
a) Find the total potential energy stored: This will be given by
4
1⌘
WE =
qn Vn
2 n=1
where Vn in this case is the potential at the location of any one of the point charges that
arises from the other three. This will be (for charge 1)
⌦
↵
q
1
1
1
↵
V1 = V21 + V31 + V41 =
+
+
4⌅ 0 .04 .04 .04 2
Taking the summation produces a factor of 4, since the situation is the same at all four
points. Consequently,
⌦
↵
1
(.8 ⇤ 10 9 )2
1
WE = (4)q1 V1 =
2 + ↵ = 7.79 ⇤ 10 7 J = 0.779 µJ
2
2⌅ 0 (.04)
2
b) A fifth 0.8 nC charge is installed at the center of the square. Again find the total stored
energy: This will be the energy found in part a plus the amount of work done in moving
the fifth charge into position from infinity. The latter is just the potential at the square
center arising from the original four charges, times the new charge value, or
WE =
4(.8 ⇤ 10 9 )2
↵
= .813 µJ
4⌅ 0 (.04 2/2)
The total energy is now
WE net = WE (part a) +
WE = .779 + .813 = 1.59 µJ
4.36 Surface charge of uniform density ⇧s lies on a spherical shell of radius b, centered at the origin
in free space.
a) Find the absolute potential everywhere, with zero reference at infinity: First, the electric
field, found from Gauss’ law, is
E=
Then
V (r) =
✓
r
⌃
b2 ⇧s
a V/m
2 r
0r
E · dL =
✓
r
⌃
b2 ⇧s
b2 ⇧s
⇧
dr
=
V
⇧ 2
0 (r )
0r
b) find the stored energy in the sphere by considering the charge density and the potential
in a two-dimensional version of Eq. (42):
✓
✓
✓
1
1 2⇥ ⇥ b2 ⇧s 2
2⌅⇧2s b3
We =
⇧s V (b) da =
⇧s
b sin ⇥ d⇥ d⌃ =
2 S
2 0
0b
0
0
c) find the stored energy in the electric field and show that the results of parts b and c are
identical.
✓
✓ 2⇥ ✓ ⇥ ✓ ⌃ 4 2
1
1 b ⇧s 2
2⌅⇧2s b3
We =
D · E dv =
r
sin
⇥
dr
d⇥
d⌃
=
2 0 r4
0
v 2
0
0
b
57
CHAPTER 5
5.1. Given the current density J =
104 [sin(2x)e
2y
ax + cos(2x)e
2y
ay ] kA/m2 :
a) Find the total current crossing the plane y = 1 in the ay direction in the region 0 < x < 1,
0 < z < 2: This is found through
I=
⌫ ⌫
⌫ 2⌫ 1
⌫ 2⌫ 1
⇧
⇧
⇧
⇧
J · n⇧ da =
J · ay ⇧
dx dz =
104 cos(2x)e
S
y=1
S
0
0
0
0
⇧1
1
⇧
4
2
10 (2) sin(2x)⇧ e = 1.23 MA
2
0
=
2
dx dz
b) Find the total current leaving the region 0 < x, x < 1, 2 < z < 3 by integrating J·dS over
the surface of the cube: Note first that current through the top and bottom surfaces will
not exist, since J has no z component. Also note that there will be no current through the
x = 0 plane, since Jx = 0 there. Current will pass through the three remaining surfaces,
and will be found through
I=
⌫
3
2
⌫
1
0
= 10
4
⌫
3
2
= 104
1
2
⇧
⇧
J · ( ay )⇧
⌫
y=0
1
0
⇤
cos(2x)e
dx dz +
3
2
⌫
0
cos(2x)e
0
⇧1
⇧
sin(2x)⇧ (3
⇤
2) 1
✏Jx
✏Jy
+
=
✏x
✏y
10
0
⌫
e
2
⌅
1
⇧
⇧
J · (ay )⇧
y=1
2
⌅
dx dz
dx dz +
10
4
⌫
3
2
1
2
+ 104
sin(2)e
⌫
⌫
3
2
1
⌫
0
1
⇧
⇧
J · (ax )⇧
x=1
sin(2)e
2y
dy dz
dy dz
0
⇧1
2y ⇧
⇧ (3
0
2) = 0
c) Repeat part b, but use the divergence theorem: We find the net outward current through
the surface of the cube by integrating the divergence of J over the cube volume. We have
·J=
4
⇤
2 cos(2x)e
2y
2 cos(2x)e
2y
⌅
= 0 as expected
5.2. Given J = 10 4 (yax + xay ) A/m2 , find the current crossing the y = 0 plane in the
direction between z = 0 and 1, and x = 0 and 2.
At y = 0, J(x, 0) =
I=
⌫
104 xay , so that the current through the plane becomes
J · dS =
⌫
0
1
⌫
0
2
104 x ay · ( ay ) dx dz = 2 ⇤ 10
58
4
A
ay
5.3. Let
J=
400 sin ⇥
ar A/m2
r2 + 4
a) Find the total current flowing through that portion of the spherical surface r = 0.8,
bounded by 0.1⌅ < ⇥ < 0.3⌅, 0 < ⌥ < 2⌅: This will be
I=
⌫ ⌫
= 346.5
⌫
⇧
⇧
J · n⇧ da =
⌫
S
.3⇤
.1⇤
1
[1
2
0
2⇤
⌫
.3⇤
.1⇤
400 sin ⇥
400(.8)2 2⌅
2
(.8)
sin
⇥
d⇥
d⌥
=
(.8)2 + 4
4.64
⌫
.3⇤
sin2 d⇥
.1⇤
cos(2⇥)] d⇥ = 77.4 A
b) Find the average value of J over the defined area. The area is
Area =
⌫
0
2⇤
⌫
.3⇤
(.8)2 sin ⇥ d⇥ d⌥ = 1.46 m2
.1⇤
The average current density is thus Javg = (77.4/1.46) ar = 53.0 ar A/m2 .
5.4. If volume charge density is given as ⇧v = (cos t)/r2 C/m3 in spherical coordinates, find J. It
is reasonable to assume that J is not a function of ⇥ or ⌥.
We use the continuity equation (5), along with the assumption of no angular variation to
write
⇥
1 ✏
✏⇧v
✏ cos t
sin t
·J= 2
r2 Jr =
=
=
r ✏r
✏t
✏t
r2
r2
So we may now solve
by direct integration to obtain:
⇥
✏
r2 Jr =
✏r
J = Jr ar =
sin t
sin t
ar A/m2
r
where the integration constant is set to zero because a steady current will not be created
by a time-varying charge density.
59
5.5. Let
J=
25
a⌅
⇧
20
az A/m2
⇧2 + 0.01
a) Find the total current crossing the plane z = 0.2 in the az direction for ⇧ < 0.4: Use
I=
⌫ ⌫
S
=
1
2
⇧
⇧
J · n⇧
z=.2
da =
⌫
2⇤
0
⌫
0
.4
⇧2
⇧.4
⇧
20 ln(.01 + ⇧2 )⇧ (2⌅) =
0
20
⇧ d⇧ d⌥
+ .01
20⌅ ln(17) =
178.0 A
b) Calculate ✏⇧v /✏t: This is found using the equation of continuity:
✏⇧v
=
✏t
·J=
1 ✏
✏Jz
1 ✏
✏
(⇧J⌅ ) +
=
(25) +
⇧ ✏⇧
✏z
⇧ ✏⇧
✏z
⇧2
20
+ .01
=0
c) Find the outward current crossing the closed surface defined by ⇧ = 0.01, ⇧ = 0.4, z = 0,
and z = 0.2: This will be
⌫ .2 ⌫ 2⇤
⌫ .2 ⌫ 2⇤
25
25
I=
a⌅ · ( a⌅ )(.01) d⌥ dz +
a⌅ · (a⌅ )(.4) d⌥ dz
.01
.4
0
0
0
0
⌫ 2⇤ ⌫ .4
⌫ 2⇤ ⌫ .4
20
20
+
az · ( az ) ⇧ d⇧ d⌥ +
az · (az ) ⇧ d⇧ d⌥ = 0
2
2
0
0 ⇧ + .01
0
0 ⇧ + .01
since the integrals will cancel each other.
d) Show that the divergence theorem is satisfied for J and the surface specified in part b.
In part c, the net outward flux was found to be zero, and in part b, the divergence of J
was found to be zero (as will be its volume integral). Therefore, the divergence theorem
is satisfied.
5.6. In spherical coordinates, a current density J = k/(r sin ⇥) a⇥ A/m2 exists in a conducting
medium, where k is a constant. Determine the total current in the az direction that crosses a
circular disk of radius R, centered on the z axis and located at a) z = 0; b) z = h.
Integration over a disk means that we use cylindrical coordinates. The general flux integral
assumes the form:
⌫
⌫ 2⇤ ⌫ R
k
I = J · dS =
a⇥ · az ⇧ d⇧ d⌥
r
sin
⇥
s
0
0
sin ⇥
⇢
Then, using r = ⇧2 + z 2 , this becomes
I=
⌫
0
2⇤
⌫
0
R
⇧R
⇠⇢
⇢
k⇧
⇧
⇢
= 2⌅k ⇧2 + z 2 ⇧ = 2⌅k
R2 + z 2
0
⇧2 + z 2
At z = 0 (part a), we have I(0) = 2⌅kR , and at z = h (part b): I(h) = 2⌅k
60
z
⇤
⇡
R2 + h2
⌅
h .
5.7. Assuming that there is no transformation of mass to energy or vice-versa, it is possible to
write a continuity equation for mass.
a) If we use the continuity equation for charge as our model, what quantities correspond to J
and ⇧v ? These would be, respectively, mass flux density in (kg/m2 s) and mass density
in (kg/m3 ).
b) Given a cube 1 cm on a side, experimental data show that the rates at which mass is
leaving each of the six faces are 10.25, -9.85, 1.75, -2.00, -4.05, and 4.45 mg/s. If we
assume that the cube is an incremental volume element, determine an approximate value
for the time rate of change of density at its center. We may write the continuity equation
for mass as follows, also invoking the divergence theorem:
⌫
⌫
✏⇧m
dv =
· Jm dv =
Jm · dS
v ✏t
v
s
where
s
Jm · dS = 10.25
9.85 + 1.75
2.00
4.05 + 4.45 = 0.550 mg/s
Treating our 1 cm3 volume as di⇥erential, we find
✏⇧m .
=
✏t
0.550 ⇤ 10 3 g/s
=
10 6 m3
550 g/m3
s
5.8. A truncated cone has a height of 16 cm. The circular faces on the top and bottom have radii
of 2mm and 0.1mm, respectively. If the material from which this solid cone is constructed
has a conductivity of 2 ⇤ 106 S/m, use some good approximations to determine the resistance
between the two circular faces.
Consider the cone upside down and centered on the positive z axis. The 1-mm radius end
is at distance z = from the x-y plane; the wide end (2-mm radius) lies at z = + 16 cm.
is chosen such that if the cone were not truncated, its vertex would occur at the origin.
The cone surface subtends angle ⇥c from the z axis (in spherical coordinates). Therefore,
we may write
0.1mm
2 mm
=
and tan ⇥c =
tan ⇥c
160 +
Solving these, we find = 8.4 mm, tan ⇥c = 1.19 ⇤ 10 2 , and so ⇥c = 0.68⇤ , which gives
us a very thin cone! With this understanding, we can assume that the current density is
uniform with ⇥ and ⌥ and will vary only with spherical radius, r. So the current density
will be constant over a spherical cap (of constant r) anywhere within the cone. As the
cone is thin, we can also assume constant current density over any flat surface within the
cone at a specifed z. That is, any spherical cap looks flat if the cap radius, r, is large
compared to its radius as measured from the z axis (⇧). This is our primary assumption.
Now, assuming constant current density at constant r, and net current, I, we may write
⌫ 2⇤ ⌫ ⇥c
I=
J(r) ar · ar r2 sin ⇥ d⇥ d⌥ = 2⌅r2 J(r)(1 cos ⇥c )
0
0
or
J(r) =
I
2⌅r2 (1
cos ⇥c )
61
ar =
(1.42 ⇤ 104 )I
ar
2⌅r2
5.8 (continued) The electric field is now
J(r)
(1.42 ⇤ 104 )I
=
ar = (7.1 ⇤ 10
⌃
2⌅r2 (2 ⇤ 106 )
E(r) =
3
)
I
ar V/m
2⌅r2
The voltage between the ends is now
V0 =
⌫
rin
rout
.
where rin = / cos ⇥c =
V0 =
⌫
⇣⇥10
.
and where rout = (160 + )/ cos ⇥c = 160 + . The voltage is
3
3
(160+⇣)⇥10
=
E · ar dr
(7.1 ⇤ 10
3
I
)
ar · ar dr = (7.1 ⇤ 10
2⌅r2
3
0.40
I
⌅
⌦
I
1
)
2⌅ .0084
1
.1684
↵
from which we identify the resistance as R = 0.40/⌅ = 0.128 ohms.
A second method uses the idea that we can construct the cone from a stack of thin circular
plates of linearly-increasing radius, ⇧. Assuming each plate is of di⇥erential thickness, dz,
the di⇥erential resistance of a plate will be
dR =
dz
⌃⌅⇧2
where ⇧ = z tan ⇥c = z(1.19 ⇤ 10 2 ). The cone resistance will be the resistance of the
stack of plates (in series), found through
R=
=
⌫
dR =
⌫
3.53 ⇤ 10
⌅
(160+⇣)⇥10
⇣⇥10
⌦
3
3
1
.0084
3
⌫ (160+⇣)⇥10
dz
=
⌃⌅⇧2
⇣⇥10 3
↵
1
= 0.127 ohms
.1684
3
dz
(2 ⇤
106 )⌅z 2 (1.19
⇤ 10
2 )2
5.9. a) Using data tabulated in Appendix C, calculate the required diameter for a 2-m long nichrome
wire that will dissipate an average power of 450 W when 120 V rms at 60 Hz is applied to it:
The required resistance will be
R=
V2
l
=
P
⌃(⌅a2 )
Thus the diameter will be
d = 2a = 2
⌧
lP
2(450)
=2
= 2.8 ⇤ 10
2
6
⌃⌅V
(10 )⌅(120)2
4
m = 0.28 mm
b) Calculate the rms current density in the wire: The rms current will be I = 450/120 =
3.75 A. Thus
3.75
7
2
J=
2 = 6.0 ⇤ 10 A/m
⌅ (2.8 ⇤ 10 4 /2)
62
5.10. A large brass washer has a 2-cm inside diameter, a 5-cm outside diameter, and is 0.5 cm thick.
Its conductivity is ⌃ = 1.5 ⇤ 107 S/m. The washer is cut in half along a diameter, and a
voltage is applied between the two rectangular faces of one part. The resultant electric field
in the interior of the half-washer is E = (0.5/⇧) a⇧ V/m in cylindrical coordinates, where the
z axis is the axis of the washer.
a) What potential di⇥erence exists between the two rectangular faces? First, we orient the
washer in the x-y plane with the cut faces aligned with the x axis. To find the voltage,
we integrate E over a circular path of radius ⇧ inside the washer, between the two cut
faces:
⌫
⌫ 0
0.5⌅
V0 =
E · dL =
a⇧ · a⇧ ⇧ d⌥ = 0.5⌅ V
⇧
⇤
b) What total current is flowing? First, the current density is J = ⌃E, so
J=
1.5 ⇤ 107 (0.5)
7.5 ⇤ 106
a⇧ =
a⇧ A/m2
⇧
⇧
Current is then found by integrating J over any transverse plane in the washer (the
rectangular cross-section):
I=
⌫
s
J · dS =
= 0.5 ⇤ 10
2
⌫
2
0.5⇥10
0
⌫
2.5⇥10
10
(7.5 ⇤ 106 ) ln
2
2.5
1
2
7.5 ⇤ 106
a⇧ · a⇧ d⇧ dz
⇧
= 3.4 ⇤ 104 A
c) What is the resistance between the two faces?
R=
V0
0.5⌅
=
= 4.6 ⇤ 10
I
3.4 ⇤ 104
5
ohms
5.11. Two perfectly-conducting cylindrical surfaces of length l are located at ⇧ = 3 and ⇧ = 5 cm.
The total current passing radially outward through the medium between the cylinders is 3 A
dc.
a) Find the voltage and resistance between the cylinders, and E in the region between the
cylinders, if a conducting material having ⌃ = 0.05 S/m is present for 3 < ⇧ < 5 cm:
Given the current, and knowing that it is radially-directed, we find the current density
by dividing it by the area of a cylinder of radius ⇧ and length l:
J=
3
a⌅ A/m2
2⌅⇧l
63
5.11a) (continued)
Then the electric field is found by dividing this result by ⌃:
3
9.55
a⌅ =
a⌅ V/m
2⌅⌃⇧l
⇧l
E=
The voltage between cylinders is now:
V =
⌫
3
E · dL =
5
⌫
5
3
Now, the resistance will be
R=
9.55
9.55
a⌅ · a⌅ d⇧ =
ln
⇧l
l
5
3
=
4.88
V
l
V
4.88
1.63
=
=
I
3l
l
b) Show that integrating the power dissipated per unit volume over the volume gives the
total dissipated power: We calculate
P =
⌫
v
E · J dv =
⌫ l⌫
0
0
2⇤
⌫
.05
.03
32
32
⇧
d⇧
d⌥
dz
=
ln
(2⌅)2 ⇧2 (.05)l2
2⌅(.05)l
5
3
=
14.64
W
l
We also find the power by taking the product of voltage and current:
4.88
14.64
(3) =
W
l
l
P =VI =
which is in agreement with the power density integration.
5.12. Two identical conducting plates, each having area A, are located at z = 0 and z = d. The region between plates is filled with a material having z-dependent conductivity, ⌃(z) = ⌃0 e z/d ,
where ⌃0 is a constant. Voltage V0 is applied to the plate at z = d; the plate at z = 0 is at
zero potential. Find, in terms of the given parameters:
a) the resistance of the material: We start with the di⇥erential resistance of a thin slab of
the material of thickness dz, which is
dz
ez/d dz
dR =
=
so that R =
⌃A
⌃0 A
⌫
dR =
⌫
0
d
ez/d dz
d
=
(e
⌃0 A
⌃0 A
1) =
1.72d
⌃0 A
b) the total current flowing between plates: We use
I=
V0
⌃0 AV0
=
R
1.72 d
c) the electric field intensity E within the material: First the current density is
J=
I
⌃0 V0
J
V0 ez/d
az =
az so that E =
=
az V/m
A
1.72 d
⌃(z)
1.72 d
64
5.13. A hollow cylindrical tube with a rectangular cross-section has external dimensions of 0.5 in by
1 in and a wall thickness of 0.05 in. Assume that the material is brass, for which ⌃ = 1.5 ⇤ 107
S/m. A current of 200 A dc is flowing down the tube.
a) What voltage drop is present across a 1m length of the tube? Converting all measurements
to meters, the tube resistance over a 1 m length will be:
R1 =
(1.5 ⇤
107 ) [(2.54)(2.54/2)
= 7.38 ⇤ 10
4
⇤ 10
4
1
2.54(1
The voltage drop is now V = IR1 = 200(7.38 ⇤ 10
4
.2) ⇤ 10
.1)(2.54/2)(1
4]
= 0.147 V.
b) Find the voltage drop if the interior of the tube is filled with a conducting material for
which ⌃ = 1.5 ⇤ 105 S/m: The resistance of the filling will be:
R2 =
1
(1.5 ⇤
105 )(1/2)(2.54)2
⇤ 10
4 (.9)(.8)
= 2.87 ⇤ 10
2
The total resistance is now the parallel combination of R1 and R2 :
RT = R1 R2 /(R1 + R2 ) = 7.19⇤10 4 , and the voltage drop is now V = 200RT = .144 V.
5.14. A rectangular conducting plate lies in the xy plane, occupying the region 0 < x < a, 0 < y < b.
An identical conducting plate is positioned directly above and parallel to the first, at z = d.
The region between plates is filled with material having conductivity ⌃(x) = ⌃0 e x/a , where
⌃0 is a constant. Voltage V0 is applied to the plate at z = d; the plate at z = 0 is at zero
potential. Find, in terms of the given parameters:
a) the electric field intensity E within the material: We know that E will be z-directed,
but the conductivity varies with x. We therefore expect no z variation in E, and also
note that the line integral of E between the bottom and top plates must always give V0 .
Therefore E = V0 /d az V/m.
b) the total current flowing between plates: We have
J = ⌃(x)E =
⌃0 e
x/a
V0
d
az
Using this, we find
I=
⌫
J · dS =
⌫
0
b
⌫
0
a
⌃0 e
x/a
d
V0
az · ( az ) dx dy =
c) the resistance of the material: We use
R=
V0
d
=
I
0.63 ab ⌃0
65
⌃0 abV0
(1
d
e
1
)=
0.63ab⌃0 V0
A
d
5.15. Let V = 10(⇧ + 1)z 2 cos ⌥ V in free space.
a) Let the equipotential surface V = 20 V define a conductor surface. Find the equation of
the conductor surface: Set the given potential function equal to 20, to find:
(⇧ + 1)z 2 cos ⌥ = 2
b) Find ⇧ and E at that point on the conductor surface where ⌥ = 0.2⌅ and z = 1.5: At
the given values of ⌥ and z, we solve the equation of the surface found in part a for ⇧,
obtaining ⇧ = .10. Then
E=
=
Then
1 ✏V
✏V
a⇧
az
⇧ ✏⌥
✏z
⇧+1 2
10z 2 cos ⌥ a⌅ + 10
z sin ⌥ a⇧ 20(⇧ + 1)z cos ⌥ az
⇧
V =
✏V
a⌅
✏⇧
E(.10, .2⌅, 1.5) =
18.2 a⌅ + 145 a⇧
26.7 az V/m
c) Find |⇧s | at that point: Since E is at the perfectly-conducting surface, it will be normal
to the surface, so we may write:
⇧s =
⇧
⇧
E
·
n
⇧
0
surface
=
0
E·E
=
|E|
0
E·E=
0
⇢
(18.2)2 + (145)2 + (26.7)2 = 1.32 nC/m2
5.16. A coaxial transmission line has inner and outer conductor radii a and b. Between conductors
(a < ⇧ < b) lies a conductive medium whose conductivity is ⌃(⇧) = ⌃0 /⇧, where ⌃0 is a
constant. The inner conductor is charged to potential V0 , and the outer conductor is grounded.
a) Assuming dc radial current I per unit length in z, determine the radial current density
field J in A/m2 : This will be the current divided by the cross-sectional area that is normal
to the current direction:
I
J=
a⌅ A/m2
2⌅⇧(1)
b) Determine the electric field intensity E in terms of I and other parameters, given or
known:
J
I⇧
I
E= =
a⌅ =
a⌅ V/m
⌃
2⌅⌃0 ⇧
2⌅⌃0
c) by taking an appropriate line integral of E as found in part b, find an expression that
relates V0 to I:
⌫ a
⌫ a
I
I(b a)
V0 =
E · dL =
a⌅ · a⌅ d⇧ =
V
2⌅⌃
2⌅⌃0
0
b
b
d) find an expression for the conductance of the line per unit length, G:
G=
I
2⌅⌃0
=
S/m
V0
(b a)
66
5.17. Given the potential field V = 100xz/(x2 + 4) V. in free space:
a) Find D at the surface z = 0: Use
E=
V =
100z
✏
✏x
x2
x
+4
At z = 0, we use this to find D(z = 0) =
ax
0 E(z
100x
az V/m
x2 + 4
0 ay
= 0) =
100 0 x/(x2 + 4) az C/m2 .
b) Show that the z = 0 surface is an equipotential surface: There are two reasons for this:
1) E at z = 0 is everywhere z-directed, and so moving a charge around on the surface
involves doing no work; 2) When evaluating the given potential function at z = 0, the
result is 0 for all x and y.
c) Assume that the z = 0 surface is a conductor and find the total charge on that portion
of the conductor defined by 0 < x < 2, 3 < y < 0: We have
⇧
⇧
⇧s = D · az ⇧
So
Q=
⌫
0
3
⌫
0
2
100 0 x
dx dy =
x2 + 4
(3)(100)
100 0 x
C/m2
x2 + 4
=
z=0
0
1
2
⇧2
⇧
ln(x2 + 4)⇧ =
0
150
0
ln 2 =
0.92 nC
5.18. Two parallel circular plates of radius a are located at z = 0 and z = d. The top plate (z = d)
is raised to potential V0 ; the bottom plate is grounded. Between the plates is a conducting
material having radial-dependent conductivity, ⌃(⇧) = ⌃0 ⇧, where ⌃0 is a constant.
a) Find the ⇧-independent electric field strength, E, between plates: The integral of E between plates must give V0 , independent of position on the plates. Therefore, it must be
true that
V0
E=
az V/m (0 < ⇧ < a)
d
b) Find the current density, J between plates:
J = ⌃E =
⌃0 V0 ⇧
az A/m2
d
c) Find the total current, I, in the structure:
I=
⌫
0
2⇤
⌫
0
a
⌃0 V0 ⇧
2⌅a3 ⌃0 V0
az · ( az ) ⇧ d⇧ d⌥ =
A
d
3d
d) Find the resistance between plates:
R=
V0
3d
=
ohms
I
2⌅a3 ⌃0
67
5.19. Let V = 20x2 yz
10z 2 V in free space.
a) Determine the equations of the equipotential surfaces on which V = 0 and 60 V: Setting the
given potential function equal to 0 and 60 and simplifying results in:
At 0 V : 2x2 y
z=0
At 60 V : 2x2 y
z=
6
z
b) Assume these are conducting surfaces and find the surface charge density at that point
on the V = 60 V surface where x = 2 and z = 1. It is known that 0 ⌅ V ⌅ 60 V is the
field-containing region: First, on the 60 V surface, we have
2x2 y
Now
E=
6
= 0 ⌃ 2(2)2 y(1)
z
z
V =
40xyz ax
1
20x2 z ay
6=0 ⌃ y=
[20xy
7
8
20z] az
Then, at the given point, we have
D(2, 7/8, 1) =
0 E(2, 7/8, 1)
=
0 [70 ax
+ 80 ay + 50 az ] C/m2
We know that since this is the higher potential surface, D must be directed away from
it, and so the charge density would be positive. Thus
⇢
⇧s = D · D = 10 0 72 + 82 + 52 = 1.04 nC/m2
c) Give the unit vector at this point that is normal to the conducting surface and directed
toward the V = 0 surface: This will be in the direction of E and D as found in part b, or
⌦
↵
7ax + 8ay + 5az
an =
= [0.60ax + 0.68ay + 0.43az ]
72 + 82 + 52
5.20. Two point charges of 100⌅ µC are located at (2,-1,0) and (2,1,0). The surface x = 0 is a
conducting plane.
a) Determine the surface charge density at the origin. I will solve the general case first, in
which we find the charge density anywhere on the y axis. With the conducting plane in
the yz plane, we will have two image charges, each of +100⌅ µC, located at (-2, -1, 0)
and (-2, 1, 0). The electric flux density on the y axis from these four charges will be
⇣
D(y) =
100⌅
4⌅
 [(y 1) a
2 ax ] [(y + 1) ay 2 ax ]

y
+

2
✓ [(y 1) + 4]3/2
[(y + 1)2 + 4]3/2
given charges
[(y 1) ay + 2 ax ]
[(y 1)2 + 4]3/2
[(y + 1) ay + 2 ax ]
µC/m2
[(y + 1)2 + 4]3/2 ◆
image charges
68
⌘
5.20 a) (continued)
In the expression, all y components cancel, and we are left with
⌦
D(y) = 100
[(y
1
1
+
2
3/2
1) + 4]
[(y + 1)2 + 4]3/2
We now find the charge density at the origin:
⇧
⇧
⇧s (0, 0, 0) = D · ax ⇧
y=0
↵
ax µC/m2
= 17.9 µC/m2
b) Determine ⇧S at P (0, h, 0). This will be
⇧
⇧
⇧s (0, h, 0) = D · ax ⇧
y=h
= 100
⌦
1
1
+
2
3/2
1) + 4]
[(h + 1)2 + 4]3/2
[(h
↵
µC/m2
5.21. Let the surface y = 0 be a perfect conductor in free space. Two uniform infinite line charges
of 30 nC/m each are located at x = 0, y = 1, and x = 0, y = 2.
a) Let V = 0 at the plane y = 0, and find V at P (1, 2, 0): The line charges will image across
the plane, producing image line charges of -30 nC/m each at x = 0, y = 1, and x = 0,
y = 2. We find the potential at P by evaluating the work done in moving a unit positive
charge from the y = 0 plane (we choose the origin) to P : For each line charge, this will
be:
⌦
↵
⇧l
final distance from charge
VP V0,0,0 =
ln
2⌅ 0
initial distance from charge
where V0,0,0 = 0. Considering the four charges, we thus have
VP =
⇧l
ln
2⌅ 0
1
2
+ ln
⇧l
=
ln (2) + ln
2⌅ 0
= 1.20 kV
1
2
2
1
+ ln
ln
⌃
10
1
⌥
10 + ln
ln
17
2
17
2
✏
✏
30 ⇤ 10
=
2⌅ 0
9
ln
10 17
2
b) Find E at P : Use
⌦
⇧l
(1, 2, 0) (0, 1, 0) (1, 2, 0) (0, 2, 0)
+
2⌅ 0
|(1, 1, 0)|2
|(1, 0, 0)|2
↵
(1, 2, 0) (0, 1, 0) (1, 2, 0) (0, 2, 0)
|(1, 3, 0)|2
|(1, 4, 0)|2
⌦
↵
⇧l
(1, 1, 0) (1, 0, 0) (1, 3, 0) (1, 4, 0)
=
+
= 723 ax
2⌅ 0
2
1
10
17
EP =
69
18.9 ay V/m
✏
5.22. The line segment x = 0, 1 ⌅ y ⌅ 1, z = 1, carries a linear charge density ⇧L = ⌅|y| µC/m.
Let z = 0 be a conducting plane and determine the surface charge density at: (a) (0,0,0); (b)
(0,1,0).
We consider the line charge to be made up of a string of di⇥erential segments of length, dy ⌅ ,
and of charge dq = ⇧L dy ⌅ . A given segment at location (0, y ⌅ , 1) will have a corresponding
image charge segment at location (0, y ⌅ , 1). The di⇥erential flux density on the y axis
that is associated with the segment-image pair will be
dD =
⇧L dy ⌅ [(y y ⌅ ) ay az ]
4⌅[(y y ⌅ )2 + 1]3/2
⇧L dy ⌅ [(y y ⌅ ) ay + az ]
⇧L dy ⌅ az
=
4⌅[(y y ⌅ )2 + 1]3/2
2⌅[(y y ⌅ )2 + 1]3/2
In other words, each charge segment and its image produce a net field in which the y
components have cancelled. The total flux density from the line charge and its image is
now
⌫
⌫ 1
⌅|y ⌅ | az dy ⌅
D(y) = dD =
y ⌅ )2 + 1]3/2
1 2⌅[(y
↵
⌫ 1⌦
az
y⌅
y⌅
=
+
dy ⌅
⌅
2
3/2
⌅
2
3/2
2 0 [(y y ) + 1]
[(y + y ) + 1]
⌦
↵1
⌅
az
y(y y ) + 1
y(y + y ⌅ ) + 1
=
+
2 [(y y ⌅ )2 + 1]1/2
[(y + y ⌅ )2 + 1]1/2 0
⌦
↵
az
y(y 1) + 1
y(y + 1) + 1
2
1/2
=
+
2(y + 1)
2 [(y 1)2 + 1]1/2
[(y + 1)2 + 1]1/2
Now, at the origin (part a), we find the charge density through
⌦
↵
⇧
az 1
1
⇧
⇧s (0, 0, 0) = D · az ⇧
=
+
2 = 0.29 µC/m2
2
y=0
2
2
Then, at (0,1,0) (part b), the charge density is
⌦
⇧
az
3
⇧
⇧s (0, 1, 0) = D · az ⇧
=
1+
2
y=1
5
↵
2 =
0.24 µC/m2
5.23. A dipole with p = 0.1az µC · m is located at A(1, 0, 0) in free space, and the x = 0 plane is
perfectly-conducting.
a) Find V at P (2, 0, 1). We use the far-field potential for a z-directed dipole:
V =
p cos ⇥
p
z
=
2
2
2
4⌅ 0 r
4⌅ 0 [x + y + z 2 ]1.5
The dipole at x = 1 will image in the plane to produce a second dipole of the opposite
orientation at x = 1. The potential at any point is now:
⌦
↵
p
z
z
V =
4⌅ 0 [(x 1)2 + y 2 + z 2 ]1.5
[(x + 1)2 + y 2 + z 2 ]1.5
Substituting P (2, 0, 1), we find
⌦
.1 ⇤ 106
1
V =
4⌅ 0
2 2
70
1
10 10
↵
= 289.5 V
5.23 b) Find the equation of the 200-V equipotential surface in cartesian coordinates: We just set the
potential exression of part a equal to 200 V to obtain:
⌦
[(x
↵
z
= 0.222
[(x + 1)2 + y 2 + z 2 ]1.5
z
2
1) + y 2 + z 2 ]1.5
5.24. At a certain temperature, the electron and hole mobilities in intrinsic germanium are given as
0.43 and 0.21 m2 /V · s, respectively. If the electron and hole concentrations are both 2.3⇤1019
m 3 , find the conductivity at this temperature.
With the electron and hole charge magnitude of 1.6 ⇤ 10
case can be written:
⌃ = |⇧e |µe + ⇧h µh = (1.6 ⇤ 10
19
19
C, the conductivity in this
)(2.3 ⇤ 1019 )(0.43 + 0.21) = 2.36 S/m
5.25. Electron and hole concentrations increase with temperature. For pure silicon, suitable expressions are ⇧h = ⇧e = 6200T 1.5 e 7000/T C/m3 . The functional dependence of the mobilities
on temperature is given by µh = 2.3 ⇤ 105 T 2.7 m2 /V · s and µe = 2.1 ⇤ 105 T 2.5 m2 /V · s,
where the temperature, T , is in degrees Kelvin. The conductivity will thus be
⇤
⇧e µe + ⇧h µh = 6200T 1.5 e 7000/T 2.1 ⇤ 105 T
⌅
1.30 ⇤ 109 7000/T ⇤
=
e
1 + 1.095T .2 S/m
T
⌃=
2.5
+ 2.3 ⇤ 105 T
Find ⌃ at:
a) 0⇤ C: With T = 273⇤ K, the expression evaluates as ⌃(0) = 4.7 ⇤ 10
5
b) 40⇤ C: With T = 273 + 40 = 313, we obtain ⌃(40) = 1.1 ⇤ 10
3
S/m.
c) 80⇤ C: With T = 273 + 80 = 353, we obtain ⌃(80) = 1.2 ⇤ 10
2
S/m.
2.7
⌅
S/m.
5.26. A semiconductor sample has a rectangular cross-section 1.5 by 2.0 mm, and a length of 11.0
mm. The material has electron and hole densities of 1.8⇤1018 and 3.0⇤1015 m 3 , respectively.
If µe = 0.082 m2 /V · s and µh = 0.0021 m2 /V · s, find the resistance o⇥ered between the end
faces of the sample.
Using the given values along with the electron charge, the conductivity is
⌃ = (1.6 ⇤ 10
19
The resistance is then
⇤
⌅
) (1.8 ⇤ 1018 )(0.082) + (3.0 ⇤ 1015 )(0.0021) = 0.0236 S/m
R=
⌃A
=
0.011
= 155 k
(0.0236)(0.002)(0.0015)
71
5.27. Atomic hydrogen contains 5.5 ⇤ 1025 atoms/m3 at a certain temperature and pressure. When
an electric field of 4 kV/m is applied, each dipole formed by the electron and positive nucleus
has an e⇥ective length of 7.1 ⇤ 10 19 m.
a) Find P: With all identical dipoles, we have
P = N qd = (5.5 ⇤ 1025 )(1.602 ⇤ 10
b) Find
r:
We use P =
e
Then
r
=1+
e
0 e E,
=
19
)(7.1 ⇤ 10
19
) = 6.26 ⇤ 10
12
C/m2 = 6.26 pC/m2
and so
P
6.26 ⇤ 10 12
=
= 1.76 ⇤ 10
(8.85 ⇤ 10 12 )(4 ⇤ 103 )
0E
4
= 1.000176.
5.28. Find the dielectric constant of a material in which the electric flux density is four times the
polarization.
First we use D =
r = 4/3.
0E
+P =
0E
+ (1/4)D. Therefore D = (4/3) 0 E, so we identify
5.29. A coaxial conductor has radii a = 0.8 mm and b = 3 mm and a polystyrene dielectric for
which r = 2.56. If P = (2/⇧)a⌅ nC/m2 in the dielectric, find:
a) D and E as functions of ⇧: Use
E=
P
0( r
1)
=
(2/⇧) ⇤ 10 9 a⌅
144.9
=
a⌅ V/m
12
(8.85 ⇤ 10 )(1.56)
⇧
Then
2 ⇤ 10
D = 0E + P =
⇧
b) Find Vab and
e:
9
a⌅
⌫
3
=
r
↵
1
3.28 ⇤ 10
+1 =
1.56
⇧
9
a⌅
C/m2 =
3.28a⌅
nC/m2
⇧
Use
Vab =
e
⌦
0.8
144.9
d⇧ = 144.9 ln
⇧
3
0.8
= 192 V
1 = 1.56, as found in part a.
c) If there are 4 ⇤ 1019 molecules per cubic meter in the dielectric, find p(⇧): Use
p=
P
(2 ⇤ 10 9 /⇧)
5.0 ⇤ 10
=
a⌅ =
19
N
4 ⇤ 10
⇧
72
29
a⌅ C · m
5.30. Consider a composite material made up of two species, having number densities N1 and N2
molecules/m3 respectively. The two materials are uniformly mixed, yielding a total number
density of N = N1 + N2 . The presence of an electric field E, induces molecular dipole
moments p1 and p2 within the individual species, whether mixed or not. Show that the
dielectric constant of the composite material is given by r = f r1 + (1 f ) r2 , where f is the
number fraction of species 1 dipoles in the composite, and where r1 and r2 are the dielectric
constants that the unmixed species would have if each had number density N .
We may write the total polarization vector as
Ptot = N1 p1 + N2 p2 = N
N1
N2
p1 +
p2
N
N
= N [f p1 + (1
f )p2 ] = f P1 + (1
In terms of the susceptibilities, this becomes Ptot = 0 [f e1 + (1
and e2 are evaluated at the composite number density, N . Now
D=
r 0E
=
0E
+ Ptot =
0
[1 + f
e1
+ (1
f)
f)
e2 ] E,
f )P2
where
e1
e2 ] E
r
Identifying
r
r
as shown, we may rewrite it by adding and subracting f :
= [1 + f f + f
= [f r1 + (1 f )
+ (1 f )
r2 ] Q.E.D.
e2 ]
e1
= [f (1 +
e1 )
+ (1
5.31. The surface x = 0 separates two perfect dielectrics. For x > 0, let
where x < 0. If E1 = 80ax 60ay 30az V/m, find:
a) EN 1 : This will be E1 · ax = 80 V/m.
r
f )(1 +
=
r1
b) ET 1 . This has components of E1 not normal to the surface, or ET 1 =
⇢
c) ET 1 = (60)2 + (30)2 = 67.1 V/m.
⇢
d) E1 = (80)2 + (60)2 + (30)2 = 104.4 V/m.
e2 )]
= 3, while
60ay
r2
30az V/m.
e) The angle ⇥1 between E1 and a normal to the surface: Use
cos ⇥1 =
f) DN 2 = DN 1 =
g) DT 2 =
h) D2 =
r1 0 EN 1
r2 0 ET 1
= 5(8.85 ⇤ 10
r1 0 EN 1 ax
i) P2 = D2
0 E2
+
r2 0 ET 1
= D2 [1
E1 · ax
80
=
⌃ ⇥1 = 40.0⇤
E1
104.4
= 3(8.85 ⇤ 10
12
12
(1/
)(80) = 2.12 nC/m2 .
)(67.1) = 2.97 nC/m2 .
= 2.12ax
r2 )]
2.66ay
1.33az nC/m2 .
= (4/5)D2 = 1.70ax
2.13ay
1.06az nC/m2 .
j) the angle ⇥2 between E2 and a normal to the surface: Use
cos ⇥2 =
Thus ⇥2 = cos
1
E2 · ax
D2 · ax
2.12
=
=⇢
= .581
E2
D2
(2.12)2 = (2.66)2 + (1.33)2
(.581) = 54.5⇤ .
73
=5
5.32. Two equal but opposite-sign point charges of 3µC are held x meters apart by a spring that
provides a repulsive force given by Fsp = 12(0.5 x) N. Without any force of attraction, the
spring would be fully-extended to 0.5m.
a) Determine the charge separation: The Coulomb and spring forces must be equal in magnitude. We set up
(3 ⇤ 10 6 )2
9 ⇤ 10 12
=
= 12(0.5
4⌅ 0 x2
4⌅(8.85 ⇤ 10 12 )x2
x)
which leads to the cubic equation:
x3
0.5x2 + 6.74 ⇤ 10
3
.
whose solution, found using a calculator, is x = 0.136 m.
b) what is the dipole moment?
Dipole moment magnitude will be p = qd = (3 ⇤ 10
6
)(0.136) = 4.08 ⇤ 10
7
C-m.
5.33. Two perfect dielectrics have relative permittivities r1 = 2 and r2 = 8. The planar interface
between them is the surface x y +2z = 5. The origin lies in region 1. If E1 = 100ax +200ay
50az V/m, find E2 : We need to find the components of E1 that are normal and tangent to
the boundary, and then apply the appropriate boundary conditions. The normal component
will be EN 1 = E1 · n. Taking f = x y + 2z, the unit vector that is normal to the surface is
n=
f
1
=
[ax
| f|
6
ay + 2az ]
This normal will point in the direction of increasing f , which will be away from the origin, or
into region 2 (you can visualize a portion of the surface as a triangle whose vertices are on the
three coordinate axes at x = 5, y = 5, and z = 2.5). So EN 1 = (1/ 6)[100 200 100] =
81.7 V/m. Since the magnitude is negative, the normal component points into region 1 from
the surface. Then
EN 1 =
81.65
1
6
[ax
ay + 2az ] =
33.33ax + 33.33ay
66.67az V/m
Now, the tangential component will be ET 1 = E1 EN 1 = 133.3ax + 166.7ay + 16.67az . Our
boundary conditions state that ET 2 = ET 1 and EN 2 = ( r1 / r2 )EN 1 = (1/4)EN 1 . Thus
1
E2 = ET 2 + EN 2 = ET 1 + EN 1 = 133.3ax + 166.7ay + 16.67az
4
= 125ax + 175ay V/m
74
8.3ax + 8.3ay
16.67az
5.34. Region 1 (x ⇧ 0) is a dielectric with
E1 = 20ax 10ay + 50az V/m.
r1
= 2, while region 2 (x < 0) has
r2
= 5. Let
a) Find D2 : One approach is to first find E2 . This will have the same y and z (tangential)
components as E1 , but the normal component, Ex , will di⇥er by the ratio r1 / r2 ; this
arises from Dx1 = Dx2 (normal component of D is continuous across a non-charged
interface). Therefore E2 = 20( r1 / r2 ) ax 10 ay + 50 az = 8 ax 10 ay + 50 az . The flux
density is then
D2 =
= 40
r2 0 E2
0
ax
50
0
ay + 250
0
az = 0.35 ax
0.44 ay + 2.21 az nC/m2
b) Find the energy density in both regions: These will be
we1 =
1
2
we2 =
1
2
r1 0 E1
· E1 =
1
(2)
2
· E2 =
1
(5)
2
r2 0 E2
0
⇤
⌅
(20)2 + (10)2 + (50)2 = 3000
0
⇤ 2
⌅
(8) + (10)2 + (50)2 = 6660
= 26.6 nJ/m3
0
0
= 59.0 nJ/m3
5.35. Let the cylindrical surfaces ⇧ = 4 cm and ⇧ = 9 cm enclose two wedges of perfect dielectrics,
r1 = 2 for 0 < ⌥ < ⌅/2, and r2 = 5 for ⌅/2 < ⌥ < 2⌅. If E1 = (2000/⇧)a⌅ V/m, find:
a) E2 : The interfaces between the two media will lie on planes of constant ⌥, to which E1
is parallel. Thus the field is the same on either side of the boundaries, and so E2 = E1 .
b) the total electrostatic energy stored in a 1m length of each region: In general we have
wE = (1/2) r 0 E 2 . So in region 1:
WE1 =
⌫
1
0
⌫
⌫
⇤/2
0
4
9
1
(2000)2
⌅
(2) 0
⇧ d⇧ d⌥ dz =
2
⇧2
2
2
0 (2000)
ln
9
4
= 45.1 µJ
In region 2, we have
WE2 =
⌫
0
1
⌫
2⇤
⇤/2
⌫
4
9
1
(2000)2
15⌅
(5) 0
⇧ d⇧ d⌥ dz =
2
⇧2
4
75
2
0 (2000)
ln
9
4
= 338 µJ
CHAPTER 6.
6.1. Consider a coaxial capacitor having inner radius a, outer radius b, unit length, and filled with
a material with dielectric constant, ⇥r . Compare this to a parallel-plate capacitor having plate
width, w, plate separation d, filled with the same dielectric, and having unit length. Express
the ratio b/a in terms of the ratio d/w, such that the two structures will store the same energy
for a given applied voltage.
Storing the same energy for a given applied voltage means that the capacitances will
be equal. With both structures having unit length and containing the same dielectric
(permittivity ⇥), we equate the two capacitances:
2⇧⇥
⇥w
=
ln(b/a)
d
b
d
= exp 2⇧
a
w
6.2. Let S = 100 mm2 , d = 3 mm, and ⇥r = 12 for a parallel-plate capacitor.
a) Calculate the capacitance:
C=
⇥r ⇥0 A
12⇥0 (100 ⇤ 10
=
d
3 ⇤ 10 3
6
)
= 0.4⇥0 = 3.54 pf
b) After connecting a 6 V battery across the capacitor, calculate E, D, Q, and the total
stored electrostatic energy: First,
E = V0 /d = 6/(3 ⇤ 10
3
) = 2000 V/m, then D = ⇥r ⇥0 E = 2.4 ⇤ 104 ⇥0 = 0.21 µC/m2
The charge in this case is
Q = D · n|s = DA = 0.21 ⇤ (100 ⇤ 10
6
) = 0.21 ⇤ 10
4
µC = 21 pC
Finally, We = (1/2)QV0 = 0.5(21)(6) = 63 pJ.
c) With the source still connected, the dielectric is carefully withdrawn from between the
plates. With the dielectric gone, re-calculate E, D, Q, and the energy stored in the
capacitor.
E = V0 /d = 6/(3 ⇤ 10
3
) = 2000 V/m, as before. D = ⇥0 E = 2000⇥0 = 17.7 nC/m2
The charge is now Q = DA = 17.7 ⇤ (100 ⇤ 10
6
) nC = 1.8 pC.
Finally, We = (1/2)QV0 = 0.5(1.8)(6) = 5.4 pJ.
d) If the charge and energy found in (c) are less than that found in (b) (which you should
have discovered), what became of the missing charge and energy? In the absence of
friction in removing the dielectric, the charge and energy have returned to the battery
that gave it.
76
6.3. Capacitors tend to be more expensive as their capacitance and maximum voltage, Vmax ,
increase. The voltage Vmax is limited by the field strength at which the dielectric breaks
down, EBD . Which of these dielectrics will give the largest CVmax product for equal plate
areas: (a) air: ⇥r = 1, EBD = 3 MV/m; (b) barium titanate: ⇥r = 1200, EBD = 3 MV/m;
(c) silicon dioxide: ⇥r = 3.78, EBD = 16 MV/m; (d) polyethylene: ⇥r = 2.26, EBD = 4.7
MV/m? Note that Vmax = EBDd, where d is the plate separation. Also, C = ⇥r ⇥0 A/d, and
so Vmax C = ⇥r ⇥0 AEBD , where A is the plate area. The maximum CVmax product is found
through the maximum ⇥r EBD product. Trying this with the given materials yields the winner,
which is barium titanate.
6.4. An air-filled parallel-plate capacitor with plate separation d and plate area A is connected to
a battery which applies a voltage V0 between plates. With the battery left connected, the
plates are moved apart to a distance of 10d. Determine by what factor each of the following
quantities changes:
a) V0 : Remains the same, since the battery is left connected.
b) C: As C = ⇥0 A/d, increasing d by a factor of ten decreases C by a factor of 0.1.
c) E: We require E ⇤ d = V0 , where V0 has not changed. Therefore, E has decreased by a
factor of 0.1.
d) D: As D = ⇥0 E, and since E has decreased by 0.1, D decreases by 0.1.
e) Q: Since Q = CV0 , and as C is down by 0.1, Q also decreases by 0.1.
f) ⌃S : As Q is reduced by 0.1, ⌃S reduces by 0.1. This is also consistent with D having been
reduced by 0.1.
g) We : Use We = 1/2 CV02 , to observe its reduction by 0.1, since C is reduced by that factor.
6.5. A parallel plate capacitor is filled with a nonuniform dielectric characterized by ⇥r = 2 + 2 ⇤
106 x2 , where x is the distance from one plate. If S = 0.02 m2 , and d = 1 mm, find C: Start by
assuming charge density ⌃s on the top plate. D will, as usual, be x-directed, originating at the
top plate and terminating on the bottom plate. The key here is that D will be constant over
the distance between plates. This can be understood by considering the x-varying dielectric as
constructed of many thin layers, each having constant permittivity. The permittivity changes
from layer to layer to approximate the given function of x. The approximation becomes exact
as the layer thicknesses approach zero. We know that D, which is normal to the layers, will
be continuous across each boundary, and so D is constant over the plate separation distance,
and will be given in magnitude by ⌃s . The electric field magnitude is now
E=
D
⌃s
=
⇥0 ⇥r
⇥0 (2 + 2 ⇤ 106 x2 )
The voltage beween plates is then
V0 =
◆
0
10
3
⌃s dx
⌃s
1
↵
=
tan
6
2
⇥0 (2 + 2 ⇤ 10 x )
⇥0 4 ⇤ 106
1
↵
x 4 ⇤ 106
2
Now Q = ⌃s (.02), and so
C=
Q
⌃s (.02)⇥0 (2 ⇤ 103 )(4)
=
= 4.51 ⇤ 10
V0
⌃s ⇧
77
10
⇧10
⇧
⇧
0
3
=
F = 451 pF
⌃⇧ ⌥
⌃s
1
⇥0 2 ⇤ 103 4
6.6. Repeat Problem 6.4 assuming the battery is disconnected before the plate separation is increased: The ordering of parameters is changed over that in Problem 6.4, as the progression
of thought on the matter is di⌅erent.
a) Q: Remains the same, since with the battery disconnected, the charge has nowhere to go.
b) ⌃S : As Q is unchanged, ⌃S is also unchanged, since the plate area is the same.
c) D: As D = ⌃S , it will remain the same also.
d) E: Since E = D/⇥0 , and as D is not changed, E will also remain the same.
e) V0 : We require E ⇤ d = V0 , where E has not changed. Therefore, V0 has increased by a
factor of 10.
f) C: As C = ⇥0 A/d, increasing d by a factor of ten decreases C by a factor of 0.1. The
same result occurs because C = Q/V0 , where V0 is increased by 10, whereas Q has not
changed.
g) We : Use We = 1/2 CV02 = 1/2 QV0 , to observe its increase by a factor of 10.
6.7. Let ⇥r1 = 2.5 for 0 < y < 1 mm, ⇥r2 = 4 for 1 < y < 3 mm, and ⇥r3 for 3 < y < 5 mm.
Conducting surfaces are present at y = 0 and y = 5 mm. Calculate the capacitance per square
meter of surface area if: a) ⇥r3 is that of air; b) ⇥r3 = ⇥r1 ; c) ⇥r3 = ⇥r2 ; d) region 3 is silver:
The combination will be three capacitors in series, for which
⌦
↵
1
1
1
1
d1
d2
d3
10 3 1
2
2
=
+
+
=
+
+
=
+ +
C
C1
C2
C3
⇥r1 ⇥0 (1) ⇥r2 ⇥0 (1) ⇥r3 ⇥0 (1)
⇥0
2.5 4 ⇥r3
So that
C=
(5 ⇤ 10 3 )⇥0 ⇥r3
10 + 4.5⇥r3
Evaluating this for the four cases, we find a) C = 3.05 nF for ⇥r3 = 1, b) C = 5.21 nF for
⇥r3 = 2.5, c) C = 6.32 nF for ⇥r3 = 4, and d) C = 9.83 nF if silver (taken as a perfect
conductor) forms region 3; this has the e⌅ect of removing the term involving ⇥r3 from the
original formula (first equation line), or equivalently, allowing ⇥r3 to approach infinity.
78
6.8. A parallel-plate capacitor is made using two circular plates of radius a, with the bottom plate
on the xy plane, centered at the origin. The top plate is located at z = d, with its center on
the z axis. Potential V0 is on the top plate; the bottom plate is grounded. Dielectric having
radially-dependent permittivity fills the region between plates. The permittivity is given by
⇥(⌃) = ⇥0 (1 + ⌃2 /a2 ). Find:
a) E: Since ⇥ does not vary in the z direction, and since we must always obtain V0 when
integrating E between plates, it must follow that E = V0 /d az V/m.
b) D: D = ⇥E =
[⇥0 (1 + ⌃2 /a2 )V0 /d] az C/m2 .
c) Q: Here we find the integral of the surface charge density over the top plate:
◆
◆
◆
⇥0 (1 + ⌃2 /a2 )V0
D · dS =
az · ( az ) ⌃ d⌃ d
d
S
0
0
◆ a
2⇧⇥0 V0
3⇧⇥0 a2
=
(⌃ + ⌃3 /a2 ) d⌃ =
V0
d
2d
0
Q=
2⇤
a
d) C: We use C = Q/V0 and our previous result to find C = 3⇥0 (⇧a2 )/(2d) F.
6.9. Two coaxial conducting cylinders of radius 2 cm and 4 cm have a length of 1m. The region
between the cylinders contains a layer of dielectric from ⌃ = c to ⌃ = d with ⇥r = 4. Find the
capacitance if
a) c = 2 cm, d = 3 cm: This is two capacitors in series, and so
⌦
1
1
1
1
1
=
+
=
ln
C
C1
C2
2⇧⇥0 4
3
2
+ ln
4
3
↵
C = 143 pF
b) d = 4 cm, and the volume of the dielectric is the same as in part a: Having equal volumes
requires that 32 22 = 42 c2 , from which c = 3.32 cm. Now
⌦
1
1
1
1
=
+
=
ln
C
C1
C2
2⇧⇥0
3.32
2
79
1
+ ln
4
4
3.32
↵
C = 101 pF
6.10. A coaxial cable has conductor dimensions of a = 1.0 mm and b = 2.7 mm. The inner conductor
is supported by dielectric spacers (⇥r = 5) in the form of washers with a hole radius of 1 mm
and an outer radius of 2.7 mm, and with a thickness of 3.0 mm. The spacers are located every
2 cm down the cable.
a) By what factor do the spacers increase the capacitance per unit length? The net capacitance can be constructed as a composite quantity, composed of weighted contributions
from the air-filled and dielectric-filled regions:
Cnet =
2⇧⇥0
2⇧⇥r ⇥0
f1 +
f2
ln(b/a)
ln(b/a)
where f1 = (2 0.3)/2 and f2 = 0.3/2 are the filling factors for air and dielectric.
Substituting these gives
⌦
↵
2⇧⇥0
1.7
Cnet =
+ 0.15⇥r
ln(b/a) 2
⇡
⇠⌫
⇢
f
where the bracketed term, f , is the capactance increase factor that we seek. Substituting
⇥r = 5 gives f = 1.6.
b) If 100V is maintained across the cable, find E at all points:
Method 1: We recall the expression for electric field in a coaxial line from Gauss’ Law:
E=
a⌃s
a⌅
⇥⌃
where ⌃s is the surface charge density on the inner conductor. We also note that electric
field will be the same both inside and outside the dielectric rings because the integral of
E between conductors must always give 100 V. Another way to justify this is through
the dielectric boundary condition that requires tangential electric field to be continuous
across an interface between two dielectrics. This again means that E in our case will be
the same inside and outside the rings. We can therefore set up the integral for the voltage:
◆ a
◆ a
a⌃s
a⌃s
V0 =
E · dL =
d⌃ =
ln(b/a) = 100
⇥⌃
⇥
b
b
Then
a⌃s
100
=
= 101
⇥
ln(2.7)
E=
101
a⌅ V/m
⌃
Method 2: Solve Laplace’s equation. This was done for the cylindrical geometry in
Example 6.3, which gave the potential distribution between conductors, Eq. (35):
V (⌃) = V0
ln(b/⌃)
ln(b/a)
from which
E=
V =
dV
V0
100
101
a⌅ =
a⌅ =
a⌅ =
a⌅ V/m
d⌃
⌃ ln(b/a)
⌃ ln(2.7)
⌃
as before.
80
6.11. Two conducting spherical shells have radii a = 3 cm and b = 6 cm. The interior is a perfect
dielectric for which ⇥r = 8.
a) Find C: For a spherical capacitor, we know that:
C=
4⇧⇥r ⇥0
1
a
=
1
b
1
3
4⇧(8)⇥0
⇥
= 1.92⇧⇥0 = 53.3 pF
1
6 (100)
b) A portion of the dielectric is now removed so that ⇥r = 1.0, 0 < < ⇧/2, and ⇥r = 8,
⇧/2 <
< 2⇧. Again, find C: We recognize here that removing that portion leaves
us with two capacitors in parallel (whose C’s will add). We use the fact that with the
dielectric completely removed, the capacitance would be C(⇥r = 1) = 53.3/8 = 6.67 pF.
With one-fourth the dielectric removed, the total capacitance will be
C=
1
3
(6.67) + (53.4) = 41.7 pF
4
4
6.12. a) Determine the capacitance of an isolated conducting sphere of radius a in free space (consider
an outer conductor existing at r ⌥ ). If we assume charge Q on the sphere, the electric
field will be
Q
E=
ar V/m (a < r < )
4⇧⇥0 r2
The potential on the sphere surface will then be
V0 =
◆
a
⌅
E · dL =
Capacitance is then
C=
◆
a
⌅
Q dr
Q
=
2
4⇧⇥0 r
4⇧⇥0 a
Q
= 4⇧⇥0 a F
V0
b) The sphere is to be covered with a dielectric layer of thickness d and dielectric contant
⇥r . If ⇥r = 3, find d in terms of a such that the capacitance is twice that of part a: Let
the dielectric radius be b, where b = d + a. The potential at the conductor surface is then
V0 =
◆
b
⌅
Q
dr
4⇧⇥0 r2
◆
a
b
⌦
Q
Q
1
1
dr =
+
2
4⇧⇥r ⇥0 r
4⇧⇥0 b ⇥r
1
a
1
b
↵
Then, substituting capacitance, C = Q/V0 , and solving for b, we find, after a little algebra:
b=
a(⇥r 1)
a⇥r (4⇧⇥0 /C)
1
Then, substitute ⇥r = 3 and C = 8⇧⇥0 a (twice the part a result) to obtain:
b = 4a
d=b
81
a = 3a
6.13. With reference to Fig. 6.5, let b = 6 m, h = 15 m, and the conductor potential be 250 V.
Take ⇥ = ⇥0 . Find values for K1 , ⌃L , a, and C: We have
K1 =
✏
h+
⇣2 ✏
↵
15 +
h2 + b2
=
b
We then have
⌃L =
Next, a =
↵
h2
b2 =
(15)2
(15)2 + (6)2
6
⇣2
= 23.0
4⇧⇥0 V0
4⇧⇥0 (250)
=
= 8.87 nC/m
ln K1
ln(23)
(6)2 = 13.8 m. Finally,
2⇧⇥
2⇧⇥0
=
= 35.5 pF
1
cosh (h/b)
cosh 1 (15/6)
C=
6.14. Two #16 copper conductors (1.29-mm diameter) are parallel with a separation d between
axes. Determine d so that the capacitance between wires in air is 30 pF/m.
We use
C
2⇧⇥0
= 60 pF/m =
L
cosh 1 (h/b)
The above expression evaluates the capacitance of one of the wires suspended over a plane
at mid-span, h = d/2. Therefore the capacitance of that structure is doubled over that
required (from 30 to 60 pF/m). Using this,
h
= cosh
b
2⇧⇥0
C/L
= cosh
2⇧ ⇤ 8.854
60
= 1.46
Therefore, d = 2h = 2b(1.46) = 2(1.29/2)(1.46) = 1.88 mm.
6.15. A 2 cm diameter conductor is suspended in air with its axis 5 cm from a conducting plane.
Let the potential of the cylinder be 100 V and that of the plane be 0 V. Find the surface
charge density on the:
a) cylinder at a point nearest the plane: The cylinder will image across the plane, producing
an equivalent two-cylinder problem, with the second one at location 5 cm below the
plane. We will take the plane as the zy plane, with the cylinder
positions at x = ±5.
↵
2
Now b = 1 cm, h = 5 cm, and V0 = 100 V. Thus a = h
b2 = 4.90 cm. Then
2
K1 = [(h + a)/b] = 98.0, and ⌃L = (4⇧⇥0 V0 )/ ln K1 = 2.43 nC/m. Now
D = ⇥0 E =
⌦
⌃L (x + a)ax + yay
2⇧
(x + a)2 + y 2
(x a)ax + yay
(x a)2 + y 2
↵
and
⌃s, max
⇧
⇧
= D · ( ax )⇧
x=h b,y=0
⌦
⌃L
h b+a
=
2⇧ (h b + a)2
82
↵
h b a
= 473 nC/m2
(h b a)2
6.15b) plane at a point nearest the cylinder: At x = y = 0,
D(0, 0) =
from which
⌦
⌃L aax
2⇧ a2
⌃s = D(0, 0) · ax =
↵
aax
=
a2
⌃L
=
⇧a
⌃L 2
ax
2⇧ a
15.8 nC/m2
c) Find the capacitance per unit length. This will be C = ⌃L /V0 = 2.43 [nC/m]/100 =
24.3 pF/m.
6.16. Consider an arrangement of two isolated conducting surfaces of any shape that form a capacitor. Use the definitions of capacitance (Eq. (2) in this chapter) and resistance (Eq. (14)
in Chapter 5) to show that when the region between the conductors is filled with either conductive material (conductivity ⌥) or with a perfect dielectric (permittivity ⇥), the resulting
resistance and capacitance of the structures are related through the simple formula RC = ⇥/⌥.
What basic properties must be true about both the dielectric and the conducting medium for
this condition to hold for certain?
Considering the two surfaces, with location a one surface, and location b on the other,
the definitions are written as:
⌘
✓a
⇥E · dS
E · dL
s✓
C=
and R = ⌘ b
a
⌥E · dS
E · dL
s
b
Note that the integration surfaces in the two definitions are (in this case) closed, because
each must completely surround one of the two conductors. The surface integrals would
thus yield the total charge on – or the total current flowing out of – the conductor inside.
The two formulas are muliplied together to give
⌘
⇥E · dS
⇥
RC = ⌘ s
=
⌥
⌥E · dS
s
The far-right result is valid provided that ⇥ and ⌥ are constant-valued over the integration
surfaces. Since in prinicple any surface can be chosen over which to integrate, the safest
(and correct) conclusion is that ⇥ and ⌥ must be constants over the capacitor (or resistor)
volume; i.e., the medium must be homogeneous. A little more subtle points are that ⇥
and ⌥ generally cannot vary with field orientation (isotropic medium), and cannot vary
with field intensity (linear medium), for the simple relation RC = ⇥/⌥ to work.
83
6.17 Construct a curvilinear square map for a coaxial capacitor of 3-cm inner radius and 8-cm outer
radius. These dimensions are suitable for the drawing.
a) Use your sketch to calculate the capacitance per meter length, assuming ⇥R = 1: The
sketch is shown below. Note that only a 9⇥ sector was drawn, since this would then be
duplicated 40 times around the circumference to complete the drawing. The capacitance
is thus
NQ
40
.
C = ⇥0
= ⇥0
= 59 pF/m
NV
6
b) Calculate an exact value for the capacitance per unit length: This will be
C=
2⇧⇥0
= 57 pF/m
ln(8/3)
84
6.18 Construct a curvilinear-square map of the potential field about two parallel circular cylinders,
each of 2.5 cm radius, separated by a center-to-center distance of 13cm. These dimensions are
suitable for the actual sketch if symmetry is considered. As a check, compute the capacitance
per meter both from your sketch and from the exact formula. Assume ⇥R = 1.
Symmetry allows us to plot the field lines and equipotentials over just the first quadrant, as
is done in the sketch below (shown to one-half scale). The capacitance is found from the
formula C = (NQ /NV )⇥0 , where NQ is twice the number of squares around the perimeter
of the half-circle and NV is twice the number of squares between the half-circle and the left
vertical plane. The result is
C=
NQ
32
⇥0 =
⇥0 = 2⇥0 = 17.7 pF/m
NV
16
We check this result with that using the exact formula:
C=
⇧⇥0
⇧⇥0
=
= 1.95⇥0 = 17.3 pF/m
1
cosh (d/2a)
cosh 1 (13/5)
85
6.19. Construct a curvilinear square map of the potential field between two parallel circular cylinders, one of 4-cm radius inside one of 8-cm radius. The two axes are displaced by 2.5 cm.
These dimensions are suitable for the drawing. As a check on the accuracy, compute the
capacitance per meter from the sketch and from the exact expression:
C=
cosh
1
[(a2
2⇧⇥
+ b2 D2 )/(2ab)]
where a and b are the conductor radii and D is the axis separation.
The drawing is shown below. Use of the exact expression above yields a capacitance value of
C = 11.5⇥0 F/m. Use of the drawing produces:
. 22 ⇤ 2
C=
⇥0 = 11⇥0 F/m
4
86
6.20. A solid conducting cylinder of 4-cm radius is centered within a rectangular conducting cylinder
with a 12-cm by 20-cm cross-section.
a) Make a full-size sketch of one quadrant of this configuration and construct a curvilinearsquare map for its interior: The result below could still be improved a little, but is
nevertheless su⌥cient for a reasonable capacitance estimate. Note that the five-sided
region in the upper right corner has been partially subdivided (dashed line) in anticipation
of how it would look when the next-level subdivision is done (doubling the number of field
lines and equipotentials).
b) Assume ⇥ = ⇥0 and estimate C per meter length: In this case NQ is the number of squares
around the full perimeter of the circular conductor, or four times the number of squares
shown in the drawing. NV is the number of squares between the circle and the rectangle,
or 5. The capacitance is estimated to be
C=
NQ
4 ⇤ 13
.
⇥0 =
⇥0 = 10.4⇥0 = 90 pF/m
NV
5
87
6.21. The inner conductor of the transmission line shown in Fig. 6.14 has a square cross-section
2a ⇤ 2a, while the outer square is 5a ⇤ 5a. The axes are displaced as shown. (a) Construct
a good-sized drawing of the transmission line, say with a = 2.5 cm, and then prepare a
curvilinear-square plot of the electrostatic field between the conductors. (b) Use the map to
calculate the capacitance per meter length if ⇥ = 1.6⇥0 . (c) How would your result to part b
change if a = 0.6 cm?
a) The plot is shown below. Some improvement is possible, depending on how much time
one wishes to spend.
b) From the plot, the capacitance is found to be
. 16 ⇤ 2
.
C=
(1.6)⇥0 = 12.8⇥0 = 110 pF/m
4
c) If a is changed, the result of part b would not change, since all dimensions retain the same
relative scale.
88
6.22. Two conducting plates, each 3 by 6 cm, and three slabs of dielectric, each 1 by 3 by 6 cm,
and having dielectric constants of 1, 2, and 3 are assembled into a capacitor with d = 3 cm.
Determine the two values of capacitance obtained by the two possible methods of assembling
the capactor.
The two possible configurations are 1) all slabs positioned vertically, side-by-side; 2) all
slabs positioned horizontally, stacked on top of one another. For vertical positioning, the
1x3 surfaces of each slab are in contact with the plates, and we have three capacitors in
parallel. The individual capacitances will thus add to give:
Cvert = C1 + C2 + C3 =
⇥0 (1 ⇤ 3)
(1 + 2 + 3) = 6⇥0
3
With the slabs positioned horizontally, the configuration becomes three capacitors in
series, with the total capacitance found through:
1
Choriz
=
1
1
1
3
+
+
=
C1
C2
C3
⇥0 (3)
1+
1 1
+
2 3
=
11
6⇥0
Choriz =
6⇥0
11
6.23. A two-wire transmission line consists of two parallel perfectly-conducting cylinders, each having a radius of 0.2 mm, separated by center-to-center distance of 2 mm. The medium surrounding the wires has ⇥r = 3 and ⌥ = 1.5 mS/m. A 100-V battery is connected between the
wires. Calculate:
a) the magnitude of the charge per meter length on each wire: Use
C=
⇧⇥
⇧ ⇤ 3 ⇤ 8.85 ⇤ 10
=
1
cosh (h/b)
cosh 1 (1/0.2)
12
= 3.64 ⇤ 10
9
C/m
Then the charge per unit length will be
Q = CV0 = (3.64 ⇤ 10
11
)(100) = 3.64 ⇤ 10
9
C/m = 3.64 nC/m
b) the battery current: Use
RC =
⇥
⌥
R=
Then
I=
3 ⇤ 8.85 ⇤ 10 12
(1.5 ⇤ 10 3 )(3.64 ⇤ 10
11 )
V0
100
=
= 0.206 A = 206 mA
R
486
89
= 486 ⇤
6.24. A potential field in free space is given in spherical coordinates as:
V (r) =
[⌃0 /(6⇥0 )] [3a2 r2 ] (r ⇧ a)
(a3 ⌃0 )/(3⇥0 r) (r ⌃ a)
where ⌃0 and a are constants.
a) Use Poisson’s equation to find the volume charge density everywhere: Inside r = a, we
apply Poisson’s equation to the potential there:
2
V1 =
⌃v
1 d
= 2
⇥0
r dr
from which we identify ⌃v = ⌃0
Outside r = a, we use
2
V2 =
r2
dV1
dr
⇥
⌃0 1 d 2
r ( 2r) =
2
6⇥0 r dr
=
(r ⇧ a).
⌃v
1 d
= 2
⇥0
r dr
r2
dV2
dr
=
1 d
r2 dr
r2
a3 ⌃0
3⇥0 r2
⌃0
⇥0
=0
from which ⌃v = 0 (r ⌃ a).
b) find the total charge present: We have a constant charge density confined within a spherical volume of radius a. The total charge is therefore Q = (4/3)⇧a3 ⌃0 C.
6.25. Let V = 2xy 2 z 3 and ⇥ = ⇥0 . Given point P (1, 2, 1), find:
a) V at P : Substituting the coordinates into V , find VP =
b) E at P : We use E =
V = 2y z ax 4xyz ay
at P , becomes EP = 8ax + 8ay 24az V/m
2 3
c) ⌃v at P : This is ⌃v =
·D=
⇥0
2
3
V =
8 V.
6xy 2 z 2 az , which, when evaluated
4xz(z 2 + 3y 2 ) C/m3
d) the equation of the equipotential surface passing through P : At P , we know V =
so the equation will be xy 2 z 3 = 4.
8 V,
e) the equation of the streamline passing through P : First,
Ey
dy
4xyz 3
2x
=
= 2 3 =
Ex
dx
2y z
y
Thus
ydy = 2xdx, and so
1 2
y = x2 + C1
2
Evaluating at P , we find C1 = 1. Next,
Ez
dz
6xy 2 z 2
3x
=
=
=
2
3
Ex
dx
2y z
z
Thus
3 2 1 2
x = z + C2
2
2
Evaluating at P , we find C2 = 1. The streamline is now specified by the equations:
y 2 2x2 = 2 and 3x2 z 2 = 2.
3xdx = zdz, and so
f) Does V satisfy Laplace’s equation? No, since the charge density is not zero.
90
6.26. Given the spherically-symmetric potential field in free space, V = V0 e r/a , find:
a) ⌃v at r = a; Use Poisson’s equation, 2 V = ⌃v /⇥, which in this case becomes
⌃v
1 d
= 2
⇥0
r dr
r2
dV
dr
V0 d ⌃ 2
r e
ar2 dr
=
r/a
⌥
V0 ⌃
2
ar
=
r⌥
e
a
r/a
from which
⌃v (r) =
⇥0 V0 ⌃
2
ar
r⌥
e
a
⌃v (a) =
r/a
⇥0 V0
e
a2
1
C/m3
b) the electric field at r = a; this we find through the negative gradient:
E(r) =
V =
dV
V0
ar =
e
dr
a
r/a
E(a) =
ar
V0
e
a
1
ar V/m
c) the total charge: The easiest way is to first find the electric flux density, which from part
b is D = ⇥0 E = (⇥0 V0 /a)e r/a ar . Then the net outward flux of D through a sphere of
radius r would be
⇥(r) = Qencl (r) = 4⇧r2 D = 4⇧⇥0 V0 r2 e
As r ⌥
r/a
C
, this result approaches zero, so the total charge is therefore Qnet = 0.
6.27. Let V (x, y) = 4e2x + f (x) 3y 2 in a region of free space where ⌃v = 0. It is known that both
Ex and V are zero at the origin. Find f (x) and V (x, y): Since ⌃v = 0, we know that 2 V = 0,
and so
✏2V
✏2V
d2 f
2
2x
V =
+
=
16e
+
6=0
✏x2
✏y 2
dx2
Therefore
d2 f
=
dx2
df
=
dx
16e2x + 6
Now
Ex =
8e2x + 6x + C1
✏V
df
= 8e2x +
✏x
dx
and at the origin, this becomes
Ex (0) = 8 +
Thus df /dx |x=0 =
df ⇧⇧
= 0(as given)
⇧
dx x=0
8, and so it follows that C1 = 0. Integrating again, we find
f (x, y) =
4e2x + 3x2 + C2
which at the origin becomes f (0, 0) = 4 + C2 . However, V (0, 0) = 0 = 4 + f (0, 0). So
f (0, 0) = 4 and C2 = 0. Finally, f (x, y) = 4e2x + 3x2 , and V (x, y) = 4e2x 4e2x + 3x2
3y 2 = 3(x2 y 2 ).
91
6.28. Show that in a homogeneous medium of conductivity ⌥, the potential field V satisfies Laplace’s
equation if any volume charge density present does not vary with time: We begin with the
continuity equation, Eq. (5), Chapter 5:
·J=
✏⌃v
✏t
where J = ⌥E, and where, in our homogeneous medium, ⌥ is constant with position. Now
write
✏⌃v
· J = · (⌥E) = ⌥ · (
V ) = ⌥ 2V =
✏t
This becomes
2
V = 0 (Laplace⇤ s equation)
when ⌃v is time-independent. Q.E.D.
6.29. Given the potential field V = (A⌃4 + B⌃ 4 ) sin 4 :
a) Show that 2 V = 0: In cylindrical coordinates,
2
1 ✏
✏V
1 ✏2V
⌃
+ 2
⌃ ✏⌃
✏⌃
⌃ ✏ 2
⇥
1 ✏
1
=
⌃(4A⌃3 4B⌃ 5 ) sin 4
16(A⌃4 + B⌃ 4 ) sin 4
⌃ ✏⌃
⌃2
16
16
=
(A⌃3 + B⌃ 5 ) sin 4
(A⌃4 + B⌃ 4 ) sin 4 = 0
⌃
⌃2
V =
b) Select A and B so that V = 100 V and |E| = 500 V/m at P (⌃ = 1, = 22.5⇥ , z = 2):
First,
✏V
1 ✏V
E=
V =
a⌅
a⇧
✏⌃
⌃✏
⇤
⌅
= 4 (A⌃3 B⌃ 5 ) sin 4 a⌅ + (A⌃3 + B⌃ 5 ) cos 4 a⇧
and at P , EP =
equations are:
4(A
B) a⌅ . Thus |EP | = ±4(A
4(A
and
B). Also, VP = A + B. Our two
B) = ±500
A + B = 100
We thus have two pairs of values for A and B:
A = 112.5, B =
12.5 or A =
92
12.5, B = 112.5
6.30. A parallel-plate capacitor has plates located at z = 0 and z = d. The region between plates
is filled with a material containing volume charge of uniform density ⌃0 C/m3 , and which has
permittivity ⇥. Both plates are held at ground potential.
a) Determine the potential field between plates: We solve Poisson’s equation, under the
assumption that V varies only with z:
2
V =
d2 V
=
dz 2
⌃0
⇥
V =
⌃0 z 2
+ C1 z + C2
2⇥
At z = 0, V = 0, and so C2 = 0. Then, at z = d, V = 0 as well, so we find C1 = ⌃0 d/2⇥.
Finally, V (z) = (⌃0 z/2⇥)[d z].
b) Determine the electric field intensity, E between plates: Taking the answer to part a, we
find E through
E=
V =
d  ⌃0 z
(d
dz 2⇥
dV
az =
dz
z) =
⌃0
(2z
2⇥
d) az V/m
c) Repeat a and b for the case of the plate at z = d raised to potential V0 , with the z = 0
plate grounded: Begin with
⌃0 z 2
+ C1 z + C2
2⇥
V (z) =
with C2 = 0 as before, since V (z = 0) = 0. Then
V (z = d) = V0 =
⌃0 d2
+ C1 d
2⇥
So that
C1 =
V0
⌃0 d
+
d
2⇥
V0
⌃0 z
z+
(d z)
d
2⇥
We recognize this as the simple superposition of the voltage as found in part a and the
voltage of a capacitor carrying voltage V0 , but without the charged dielectric. The electric
field is now
dV
V0
⌃0
E=
az =
az + (2z d) az V/m
dz
d
2⇥
V (z) =
93
6.31. Let V = (cos 2 )/⌃ in free space.
a) Find the volume charge density at point A(0.5, 60⇥ , 1): Use Poisson’s equation:
⌃v =
⇥0
=
⇥0
2
So at A we find:
⌃vA =
V =
1 ✏
⌃ ✏⌃
1 ✏
⌃ ✏⌃
cos 2
⌃
⇥0
3⇥0 cos(120⇥ )
=
0.53
⌃
✏V
1 ✏2V
+ 2
✏⌃
⌃ ✏ 2
4 cos 2
3⇥0 cos 2
=
2
⌃
⌃
⌃3
12⇥0 =
106 pC/m3
b) Find the surface charge density on a conductor surface passing through B(2, 30⇥ , 1): First,
we find E:
✏V
1 ✏V
E=
V =
a⌅
a⇧
✏⌃
⌃✏
cos 2
2 sin 2
=
a⌅ +
a⇧
⌃2
⌃2
At point B the field becomes
EB =
cos 60⇥
2 sin 60⇥
a⌅ +
a⇧ = 0.125 a⌅ + 0.433 a⇧
4
4
The surface charge density will now be
⌃sB = ±|DB | = ±⇥0 |EB | = ±0.451⇥0 = ±0.399 pC/m2
The charge is positive or negative depending on which side of the surface we are considering. The problem did not provide information necessary to determine this.
94
6.32. A uniform volume charge has constant density ⌃v = ⌃0 C/m3 , and fills the region r < a, in
which permittivity ⇥ as assumed. A conducting spherical shell is located at r = a, and is held
at ground potential. Find:
a) the potential everywhere: Inside the sphere, we solve Poisson’s equation, assuming radial
variation only:
2
V =
1 d
r2 dr
r2
dV
dr
⌃0
⇥
=
V (r) =
⌃0 r2
C1
+
+ C2
6⇥0
r
We require that V is finite at the orgin (or as r ⌥ 0), and so therefore C1 = 0. Next,
V = 0 at r = a, which gives C2 = ⌃0 a2 /6⇥. Outside, r > a, we know the potential must
be zero, since the sphere is grounded. To show this, solve Laplace’s equation:
2
V =
1 d
r2 dr
r2
dV
dr
=0
V (r) =
C1
+ C2
r
Requiring V = 0 at both r = a and at infinity leads to C1 = C2 = 0. To summarize
V (r) =
⌅0
2
6 (a
r2 ) r < a
0 r>a
b) the electric field intensity, E, everywhere: Use
E=
V =
dV
⌃0 r
ar =
ar r < a
dr
3⇥
Outside (r > a), the potential is zero, and so E = 0 there as well.
6.33. The functions V1 (⌃, , z) and V2 (⌃, , z) both satisfy Laplace’s equation in the region a < ⌃ < b,
0 ⇧ < 2⇧, L < z < L; each is zero on the surfaces ⌃ = b for L < z < L; z = L for
a < ⌃ < b; and z = L for a < ⌃ < b; and each is 100 V on the surface ⌃ = a for L < z < L.
a) In the region specified above, is Laplace’s equation satisfied by the functions V1 + V2 ,
V1 V2 , V1 + 3, and V1 V2 ? Yes for the first three, since Laplace’s equation is linear. No
for V1 V2 .
b) On the boundary surfaces specified, are the potential values given above obtained from
the functions V1 + V2 , V1 V2 , V1 + 3, and V1 V2 ? At the 100 V surface (⌃ = a), No for all.
At the 0 V surfaces, yes, except for V1 + 3.
c) Are the functions V1 + V2 , V1 V2 , V1 + 3, and V1 V2 identical with V1 ? Only V2 is,
since it is given as satisfying all the boundary conditions that V1 does. Therefore, by the
uniqueness theorem, V2 = V1 . The others, not satisfying the boundary conditions, are
not the same as V1 .
95
6.34. Consider the parallel-plate capacitor of Problem 6.30, but this time the charged dielectric
exists only between z = 0 and z = b, where b < d. Free space fills the region b < z < d. Both
plates are at ground potential. No surface charge exists at z = b, so that both V and D are
continuous there. By solving Laplace’s and Poisson’s equations, find:
a) V (z) for 0 < z < d: In Region 1 (z < b), we solve Poisson’s equation, assuming z variation
only:
d2 V1
⌃0
dV1
⌃0 z
=
=
+ C1 (z < b)
2
dz
⇥
dz
⇥
In Region 2 (z > b), we solve Laplace’s equation, assuming z variation only:
d2 V2
=0
dz 2
dV2
= C1⇤ (z > b)
dz
At this stage we apply the first boundary condition, which is continuity of D across the
interface at z = b. Knowing that the electric field magnitude is given by dV /dz, we write
⇥
dV1 ⇧⇧
dV2 ⇧⇧
=
⇥
⇧
⇧
0
dz z=b
dz z=b
⌃0 b + ⇥C1 = ⇥0 C1⇤
C1⇤ =
⌃0 b
⇥
+ C1
⇥0
⇥0
Substituting the above expression for C1⇤ , and performing a second integration on the
Poisson and Laplace equations, we find
V1 (z) =
and
V2 (z) =
⌃0 z 2
+ C1 z + C2 (z < b)
2⇥
⌃0 bz
⇥
+ C1 z + C2⇤ (z > b)
2⇥0
⇥0
Next, requiring V1 = 0 at z = 0 leads to C2 = 0. Then, the requirement that V2 = 0 at
z = d leads to
0=
⌃0 bd
⇥
+ C1 d + C2⇤
⇥0
⇥0
C2⇤ =
⌃0 bd
⇥0
⇥
C1 d
⇥0
⌃0 z 2
⌃0 b
+ C1 z and V2 (z) =
(d
2⇥
⇥0
z)
⇥
C1 (d
⇥0
With C2 and C2⇤ known, the voltages now become
V1 (z) =
z)
Finally, to evaluate C1 , we equate the two voltage expressions at z = b:
⌦
↵
⌃0 b b + 2⇥r (d b)
V1 |z=b = V2 |z=b
C1 =
2⇥ b + ⇥r (d b)
where ⇥r = ⇥/⇥0 . Substituting C1 as found above into V1 and V2 leads to the final
expressions for the voltages:
⌦
↵
⌃0 bz
b + 2⇥r (d b)
z
V1 (z) =
(z < b)
2⇥
b + ⇥r (d b)
b
⌦
↵
⌃0 b2
d z
V2 (z) =
(z > b)
2⇥0 b + ⇥r (d b)
96
6.34b) the electric field intensity for 0 < z < d: This involves taking the negative gradient of the final
voltage expressions of part a. We find
E1 =
⌦
dV1
⌃0
az =
z
dz
⇥
E2 =
b
2
b + 2⇥r (d b)
b + ⇥r (d b)
⌦
dV2
⌃0 b2
1
az =
dz
2⇥0 b + ⇥r (d
b)
↵
↵
az V/m (z < b)
az V/m (z > b)
6.35. The conducting planes 2x + 3y = 12 and 2x + 3y = 18 are at potentials of 100 V and 0,
respectively. Let ⇥ = ⇥0 and find:
a) V at P (5, 2, 6): The planes are parallel, and so we expect variation in potential in the
direction normal to them. Using the two boundary condtions, our general potential
function can be written:
V (x, y) = A(2x + 3y
and so A =
18) + 0
100/6. We then write
V (x, y) =
and VP =
12) + 100 = A(2x + 3y
100
3 (5)
100
(2x + 3y
6
18) =
100
x
3
50y + 300
100 + 300 = 33.33 V.
b) Find E at P : Use
E=
V =
100
ax + 50 ay V/m
3
6.36. The derivation of Laplace’s and Poisson’s equations assumed constant permittivity, but there
are cases of spatially-varying permittivity in which the equations will still apply. Consider the
vector identity, · ( G) = G ·
+
· G, where and G are scalar and vector functions,
respectively. Determine a general rule on the allowed directions in which ⇥ may vary with
respect to the electric field.
In the original derivation of Poisson’s equation, we started with ·D = ⌃v , where D = ⇥E.
Therefore
· D = · (⇥E) =
· (⇥ V ) =
V · ⇥ ⇥ 2 V = ⌃v
We see from this that Poisson’s equation, 2 V = ⌃v /⇥, results when V · ⇥ = 0. In
words, ⇥ is allowed to vary, provided it does so in directions that are normal to the local
electric field.
97
6.37. Coaxial conducting cylinders are located at ⌃ = 0.5 cm and ⌃ = 1.2 cm. The region between
the cylinders is filled with a homogeneous perfect dielectric. If the inner cylinder is at 100V
and the outer at 0V, find:
a) the location of the 20V equipotential surface: From Eq. (35) we have
V (⌃) = 100
ln(.012/⌃)
V
ln(.012/.005)
We seek ⌃ at which V = 20 V, and thus we need to solve:
20 = 100
ln(.012/⌃)
ln(2.4)
b) E⌅ max : We have
E⌅ =
✏V
=
✏⌃
⌃=
.012
= 1.01 cm
(2.4)0.2
dV
100
=
d⌃
⌃ ln(2.4)
whose maximum value will occur at the inner cylinder, or at ⌃ = .5 cm:
E⌅ max =
100
= 2.28 ⇤ 104 V/m = 22.8 kV/m
.005 ln(2.4)
c) ⇥r if the charge per meter length on the inner cylinder is 20 nC/m: The capacitance per
meter length is
2⇧⇥0 ⇥r
Q
C=
=
ln(2.4)
V0
We solve for ⇥r :
⇥r =
(20 ⇤ 10 9 ) ln(2.4)
= 3.15
2⇧⇥0 (100)
6.38. Repeat Problem 6.37, but with the dielectric only partially filling the volume, within 0 <
⇧, and with free space in the remaining volume.
<
We note that the dielectric changes with , and not with ⌃. Also, since E is radiallydirected and varies only with radius, Laplace’s equation for this case is valid (see Problem
6.36) and is the same as that which led to the potential and field in Problem 6.37.
Therefore, the solutions to parts a and b are unchanged from Problem 6.37. Part c,
however, is di⌅erent. We write the charge per unit length as the sum of the charges along
each half of the center conductor (of radius a)
Q = ⇥r ⇥0 E⌅,max (⇧a) + ⇥0 E⌅,max (⇧a) = ⇥0 E⌅,max (⇧a)(1 + ⇥r ) C/m
Using the numbers given or found in Problem 6.37, we obtain
1 + ⇥r =
20 ⇤ 10 9 C/m
(8.852 ⇤ 10 12 )(22.8 ⇤ 103 V/m)(0.5 ⇤ 10
2
m)⇧
= 6.31
⇥r = 5.31
We may also note that the average dielectric constant in this problem, (⇥r + 1)/2, is the
same as that of the uniform dielectric constant found in Problem 6.37.
98
6.39. The two conducting planes illustrated in Fig. 6.14 are defined by 0.001 < ⌃ < 0.120 m,
0 < z < 0.1 m, = 0.179 and 0.188 rad. The medium surrounding the planes is air. For
region 1, 0.179 < < 0.188, neglect fringing and find:
a) V ( ): The general solution to Laplace’s equation will be V = C1 + C2 , and so
20 = C1 (.188) + C2 and 200 = C1 (.179) + C2
Subtracting one equation from the other, we find
180 = C1 (.188
Then
20 =
.179)
C1 =
2.00 ⇤ 104 (.188) + C2
C2 = 3.78 ⇤ 103
Finally, V ( ) = ( 2.00 ⇤ 104 ) + 3.78 ⇤ 103 V.
b) E(⌃): Use
E(⌃) =
2.00 ⇤ 104
1 dV
2.00 ⇤ 104
=
a⇧ V/m
⌃d
⌃
V =
c) D(⌃) = ⇥0 E(⌃) = (2.00 ⇤ 104 ⇥0 /⌃) a⇧ C/m2 .
d) ⌃s on the upper surface of the lower plane: We use
⇧
⇧
⌃s = D · n⇧
surf ace
=
2.00 ⇤ 104
2.00 ⇤ 104
a⇧ · a⇧ =
C/m2
⌃
⌃
e) Q on the upper surface of the lower plane: This will be
Qt =
◆
0
.1
◆
.120
.001
2.00 ⇤ 104 ⇥0
d⌃ dz = 2.00 ⇤ 104 ⇥0 (.1) ln(120) = 8.47 ⇤ 10
⌃
8
C = 84.7 nC
f) Repeat a) to c) for region 2 by letting the location of the upper plane be = .188 2⇧,
and then find ⌃s and Q on the lower surface of the lower plane. Back to the beginning,
we use
20 = C1⇤ (.188 2⇧) + C2⇤ and 200 = C1⇤ (.179) + C2⇤
Subtracting one from the other, we find
180 = C1⇤ (.009
2⇧)
C1⇤ = 28.7
Then 200 = 28.7(.179) + C2⇤
C2⇤ = 194.9. Thus V ( ) = 28.7 + 194.9 in region 2.
Then
28.7
28.7⇥0
E=
a⇧ V/m and D =
a⇧ C/m2
⌃
⌃
⌃s on the lower surface of the lower plane will now be
⌃s =
28.7⇥0
28.7⇥0
a⇧ · ( a⇧ ) =
C/m2
⌃
⌃
The charge on that surface will then be Qb = 28.7⇥0 (.1) ln(120) = 122 pC.
99
6.39g) Find the total charge on the lower plane and the capacitance between the planes: Total charge
will be Qnet = Qt + Qb = 84.7 nC + 0.122 nC = 84.8 nC. The capacitance will be
C=
Qnet
84.8
=
= 0.471 nF = 471 pF
V
200 20
6.40. A parallel-plate capacitor is made using two circular plates of radius a, with the bottom plate
on the xy plane, centered at the origin. The top plate is located at z = d, with its center on
the z axis. Potential V0 is on the top plate; the bottom plate is grounded. Dielectric having
radially-dependent permittivity fills the region between plates. The permittivity is given by
⇥(⌃) = ⇥0 (1 + ⌃2 /a2 ). Find:
a) V (z): Since ⇥ varies in the direction normal to E, Laplace’s equation applies, and we
write
d2 V
2
V =
=0
V (z) = C1 z + C2
dz 2
With the given boundary conditions, C2 = 0, and C1 = V0 /d. Therefore V (z) = V0 z/d V.
b) E: This will be E =
V =
dV /dz az =
(V0 /d) az V/m.
c) Q: First we find the electric flux density: D = ⇥E = ⇥0 (1 + ⌃2 /a2 )(V0 /d) az C/m2 . The
charge density on the top plate is then ⌃s = D · az = ⇥0 (1 + ⌃2 /a2 )(V0 /d) C/m2 . From
this we find the charge on the top plate:
Q=
◆
2⇤
0
◆
a
⇥0 (1 + ⌃2 /a2 )(V0 /d) ⌃ d⌃ d =
0
3⇧a2 ⇥0 V0
C
2d
d) C. The capacitance is C = Q/V0 = 3⇧a2 ⇥0 /(2d) F.
6.41. Concentric conducting spheres are located at r = 5 mm and r = 20 mm. The region between
the spheres is filled with a perfect dielectric. If the inner sphere is at 100 V and the outer
sphere at 0 V:
a) Find the location of the 20 V equipotential surface: Solving Laplace’s equation gives us
V (r) = V0
1
r
1
a
1
b
1
b
where V0 = 100, a = 5 and b = 20. Setting V (r) = 20, and solving for r produces
r = 12.5 mm.
b) Find Er,max : Use
E=
Er,max = E(r = a) =
V =
dV
V0 ar
ar = 2 1 1 ⇥
dr
r a b
V0
100
=
= 26.7 V/mm = 26.7 kV/m
a(1 (a/b))
5(1 (5/20))
c) Find ⇥r if the surface charge density on the inner sphere is 1.0 µC/m2 : ⌃s will be equal
in magnitude to the electric flux density at r = a. So ⌃s = (2.67 ⇤ 104 V/m)⇥r ⇥0 =
10 6 C/m2 . Thus ⇥r = 4.23. Note, in the first printing, the given charge density was
100 µC/m2 , leading to a ridiculous answer of ⇥r = 423.
100
6.42. The hemisphere 0 < r < a, 0 < ⇤ < ⇧/2, is composed of homogeneous conducting material of
conductivity ⌥. The flat side of the hemisphere rests on a perfectly-conducting plane. Now,
the material within the conical region 0 < ⇤ < , 0 < r < a, is drilled out, and replaced with
material that is perfectly-conducting. An air gap is maintained between the r = 0 tip of this
new material and the plane. What resistance is measured between the two perfect conductors?
Neglect fringing fields.
With no fringing fields, we have ⇤-variation only in the potential. Laplace’s equation is
therefore:
1
d
dV
2
V = 2
sin ⇤
=0
r sin ⇤ d⇤
d⇤
This reduces to
dV
C1
=
d⇤
sin ⇤
V (⇤) = C1 ln tan (⇤/2) + C2
We assume zero potential on the plane (at ⇤ = ⇧/2), which means that C2 = 0. On the
cone (at ⇤ = ), we assume potential V0 , and so V0 = C1 ln tan( /2)
C1 = V0 / ln tan( /2) The potential function is now
V (⇤) = V0
ln tan(⇤/2)
ln tan( /2)
< ⇤ < ⇧/2
The electric field is then
E=
V =
1 dV
a⇥ =
r d⇤
V0
a⇥ V/m
r sin ⇤ ln tan( /2)
The total current can now be found by integrating the current density, J = ⌥E, over any
cross-section. Choosing the lower plane at ⇤ = ⇧/2, this becomes
I=
◆
0
2⇤
◆
a
0
⌥V0
a⇥ · a⇥ r dr d =
r sin(⇧/2) ln tan( /2)
The resistance is finally
R=
V0
=
I
2⇧a⌥V0
A
ln tan( /2)
ln tan( /2)
ohms
2⇧a⌥
Note that R is in fact positive (despite the minus sign) since ln tan( /2) is negative when
< ⇧/2 (which it must be).
101
6.43. Two coaxial conducting cones have their vertices at the origin and the z axis as their axis.
Cone A has the point A(1, 0, 2) on its surface, while cone B has the point B(0, 3, 2) on its
surface. Let VA = 100 V and VB = 20 V. Find:
a)
for each cone: Have
A
= tan
1
(1/2) = 26.57⇥ and
B
= tan
1
(3/2) = 56.31⇥ .
b) V at P (1, 1, 1): The potential function between cones can be written as
V (⇤) = C1 ln tan(⇤/2) + C2
Then
20 = C1 ln tan(56.31/2) + C2 and 100 = C1 ln tan(26.57/2) + C2
Solving↵these two equations, we find C1 =
tan 1 ( 2) = 54.7⇥ . Thus
VP =
97.7 and C2 =
97.7 ln tan(54.7/2)
41.1. Now at P , ⇤ =
41.1 = 23.3 V
6.44. A potential field in free space is given as V = 100 ln tan(⇤/2) + 50 V.
a) Find the maximum value of |E⇥ | on the surface ⇤ = 40⇥ for 0.1 < r < 0.8 m, 60⇥ <
90⇥ . First
E=
1 dV
a⇥ =
r d⇤
100
a⇥ =
2r tan(⇤/2) cos2 (⇤/2)
100
a⇥ =
2r sin(⇤/2) cos(⇤/2)
<
100
a⇥
r sin ⇤
This will maximize at the smallest value of r, or 0.1:
Emax (⇤ = 40⇥ ) = E(r = 0.1, ⇤ = 40⇥ ) =
100
a⇥ = 1.56 a⇥ kV/m
0.1 sin(40)
b) Describe the surface V = 80 V: Set 100 ln tan ⇤/2 + 50 = 80 and solve for ⇤: Obtain
ln tan ⇤/2 = 0.3
tan ⇤/2 = e.3 = 1.35
⇤ = 107⇥ (the cone surface at ⇤ = 107
degrees).
102
6.45. In free space, let ⌃v = 200⇥0 /r2.4 .
a) Use Poisson’s equation to find V (r) if it is assumed that r2 Er ⌥ 0 when r ⌥ 0, and also
that V ⌥ 0 as r ⌥ : With r variation only, we have
2
V =
or
1 d
r2 dr
or
dV
=
dr
dV
dr
dV
dr
⌃v
=
⇥
r2
=
200 .6
r + C1 =
.6
333.3r
1.4
+
dV
dr
=
d
dr
Integrate once:
r2
r2
=
C1
=
r2
200r
200r
2.4
.4
333.3r.6 + C1
V (in this case) =
Er
Our first boundary condition states that r2 Er ⌥ 0 when r ⌥ 0 Therefore C1 = 0.
Integrate again to find:
333.3 .4
V (r) =
r
+ C2
.4
From our second boundary condition, V ⌥ 0 as r ⌥ , we see that C2 = 0. Finally,
V (r) = 833.3r
.4
V
b) Now find V (r) by using Gauss’ Law and a line integral: Gauss’ law applied to a spherical
surface of radius r gives:
4⇧r2 Dr = 4⇧
◆
r
0
Thus
Er =
Now
V (r) =
200⇥0 ⇤ 2
r.6
(r ) dr = 800⇧⇥0
⇤
2.4
(r )
.6
Dr
800⇧⇥0 r.6
=
= 333.3r
⇥0
.6(4⇧)⇥0 r2
◆
r
333.3(r⇤ )
⌅
103
1.4
1.4
V/m
dr⇤ = 833.3r
.4
V
6.46. By appropriate solution of Laplace’s and Poisson’s equations, determine the absolute potential
at the center of a sphere of radius a, containing uniform volume charge of density ⌃0 . Assume
permittivity ⇥0 everywhere. HINT: What must be true about the potential and the electric
field at r = 0 and at r = a?
With radial dependence only, Poisson’s equation (applicable to r ⇧ a) becomes
2
V1 =
1 d
r2 dr
r2
dV1
dr
=
⌃0
⇥0
V1 (r) =
⌃0 r2
C1
+
+ C2
6⇥0
r
(r ⇧ a)
For region 2 (r ⌃ a) there is no charge and so Laplace’s equation becomes
2
V2 =
1 d
r2 dr
r2
dV2
dr
=0
V2 (r) =
C3
+ C4
r
(r ⌃ a)
Now, as r ⌥ , V2 ⌥ 0, so therefore C4 = 0. Also, as r ⌥ 0, V1 must be finite, so
therefore C1 = 0. Then, V must be continuous across the boundary, r = a:
So now
⇧
⇧
V1 ⇧r=a = V2 ⇧r=a
V1 (r) =
⌃0 a2
C3
+ C2 =
6⇥0
a
⌃0 2
(a
6⇥0
r2 ) +
C2 =
C3
⌃0 a2
+
a
6⇥0
C3
C3
and V2 (r) =
a
r
Finally, since the permittivity is ⇥0 everywhere, the electric field will be continuous at
r = a. This is equivalent to the continuity of the voltage derivatives:
dV1 ⇧⇧
dV2 ⇧⇧
=
⇧
⇧
dr r=a
dr r=a
⌃0 a
=
3⇥0
C3
a2
C3 =
⌃0 a3
3⇥0
So the potentials in their final forms are
V1 (r) =
⌃0
(3a2
6⇥0
r2 ) and V2 (r) =
⌃0 a3
3⇥0 r
The requested absolute potential at the origin is now V1 (r = 0) = ⌃0 a2 /(2⇥0 ) V.
104
CHAPTER 7
7.1a. Find H in cartesian components at P (2, 3, 4) if there is a current filament on the z axis carrying
8 mA in the az direction:
Applying the Biot-Savart Law, we obtain
⇠ ⌃
⇠ ⌃
⇠ ⌃
IdL ⇤ aR
Idz az ⇤ [2ax + 3ay + (4 z)az ]
Idz[2ay 3ax ]
Ha =
=
=
2
2
3/2
2
4⌅R
4⌅(z
8z + 29)
8z + 29)3/2
⌃
⌃
⌃ 4⌅(z
Using integral tables, this evaluates as
⌦
↵⌃
I 2(2z 8)(2ay 3ax )
I
Ha =
=
(2ay
4⌅ 52(z 2 8z + 29)1/2
26⌅
⌃
Then with I = 8 mA, we finally obtain Ha =
3ax )
294ax + 196ay µA/m
b. Repeat if the filament is located at x = 1, y = 2: In this case the Biot-Savart integral
becomes
⇠ ⌃
⇠ ⌃
Idz az ⇤ [(2 + 1)ax + (3 2)ay + (4 z)az ]
Idz[3ay ax ]
Hb =
=
2
3/2
2
4⌅(z
8z
+
26)
4⌅(z
8z + 26)3/2
⌃
⌃
Evaluating as before, we obtain with I = 8 mA:
⌦
↵⌃
I 2(2z 8)(3ay ax )
I
Hb =
=
(3ay
2
1/2
4⌅ 40(z
20⌅
8z + 26)
⌃
ax ) =
127ax + 382ay µA/m
c. Find H if both filaments are present: This will be just the sum of the results of parts a and
b, or
HT = Ha + Hb = 421ax + 578ay µA/m
This problem can also be done (somewhat more simply) by using the known result for H from
an infinitely-long wire in cylindrical components, and transforming to cartesian components.
The Biot-Savart method was used here for the sake of illustration.
7.2. A filamentary conductor is formed into an equilateral triangle with sides of length
current I. Find the magnetic field intensity at the center of the triangle.
carrying
I will work this one from scratch, using the Biot-Savart law. Consider one side of the
triangle, oriented along the z axis, with its end points at z = ± /2. Then consider a
point, x0 , on the x axis, which would correspond to the center of the triangle, and at
which we want to find H associated with the wire segment. We thus have IdL = Idz az ,
R = z 2 + x20 , and aR = [x0 ax z az ]/R. The di⇤erential magnetic field at x0 is now
dH =
IdL ⇤ aR
Idz az ⇤ (x0 ax z az )
I dz x0 ay
=
=
2
2
2
3/2
4⌅R
4⌅(x0 + z )
4⌅(x20 + z 2 )3/2
where ay would be normal to the plane of the triangle. The magnetic field at x0 is then
H=
⇠
✓/2
✓/2
⇧✓/2
I dz x0 ay
I z ay
I ay
⇧
=
=
⇧
2
2
3/2
2
2
✓/2
4⌅(x0 + z )
4⌅x0 x0 + z
2⌅x0 2 + 4x20
105
7.2. (continued). Now, x0 lies at the center of the equilateral
↵ triangle, and from the geometry of
⇤
the triangle, we find that x0 = ( /2) tan(30 ) = /(2 3). Substituting this result into the
just-found expression for H leads to H = 3I/(2⌅ ) ay . The contributions from the other two
sides of the triangle e⇤ectively multiply the above result by three. The final answer is therefore
Hnet = 9I/(2⌅ ) ay A/m. It is also possible to work this problem (somewhat more easily) by
using Eq. (9), applied to the triangle geometry.
7.3. Two semi-infinite filaments on the z axis lie in the regions
< z < a (note typographical
error in problem statement) and a < z < . Each carries a current I in the az direction.
a) Calculate H as a function of ⇧ and ⌥ at z = 0: One way to do this is to use the field from
an infinite line and subtract from it that portion of the field that would arise from the
current segment at a < z < a, found from the Biot-Savart law. Thus,
⇠ a
I
I dz az ⇤ [⇧ a⌅ z az ]
H=
a⇧
2⌅⇧
4⌅[⇧2 + z 2 ]3/2
a
The integral part simplifies and is evaluated:
⇠ a
⇧a
I dz ⇧ a⇧
I⇧
z
Ia
⇧
=
a⇧
a⇧
⇧ =
2 + z 2 ]3/2
2
2
2
4⌅
a
4⌅[⇧
⇧ ⇧ +z
2⌅⇧ ⇧2 + a2
a
Finally,
I
H=
1
2⌅⇧
a
⇧2 + a2
✏
a⇧ A/m
b) What value of a will cause the magnitude of H at ⇧ = 1, z = 0, to be one-half the value
obtained for an infinite filament? We require
✏
↵
a
1
a
1
↵
1
=
=
a = 1/ 3
2
1 + a2
⇧2 + a2 ⌅=1 2
7.4. Two circular current loops are centered on the z axis at z = ±h. Each loop has radius a and
carries current I in the a⇧ direction.
a) Find H on the z axis over the range h < z < h: As a first step, we find the magnetic
field on the z axis arising from a current loop of radius a, centered at the origin in the
plane z = 0. It carries↵a current I in the a⇧ direction. Using
↵ the Biot-Savart law, we have
IdL = Iad⌥ a⇧ , R = a2 + z 2 , and aR = (zaz aa⌅ )/ a2 + z 2 . The field on the z axis
is then
⇠ 2⇤
⇠ 2⇤
Iad⌥ a⇧ ⇤ (zaz aa⌅ )
Ia2 d⌥ az
a2 I
H=
=
=
az A/m
4⌅(a2 + z 2 )3/2
4⌅(a2 + z 2 )3/2
2(a2 + z 2 )3/2
0
0
In obtaining this result, the term involving a⇧ ⇤ zaz = za⌅ has integrated to zero, when
taken over the range 0 < ⌥ < 2⌅. Substitute a⌅ = ax cos ⌥ + ay sin ⌥ to show this.
We now have two loops, displaced from the x-y plane to z = ±h. The field is now the
superposition of the two loop fields, which we can construct using displaced versions of
the H field we just found:
✏
a2 I
1
1
H=
+
az A/m
3/2
2 [(z h)2 + a2 ]3/2
[(z + h)2 + a2 ]
106
7.4 (continued) We can rewrite this in terms of normalized distances, z/a, and h/a:
⇣
I✓
aH =
2
z
a
h
a
3/2
2
+1
+
z
h
+
a a
3/2
2
+1
⌘
◆ az A
Take I = 1 A and plot |H| as a function of z/a if:
b) h = a/4,
c) h = a/2:
d) h = a.
Which choice for h gives the most uniform field? From the results, h = a/2 is evidently the
best. This is the Helmholtz coil configuration – in which the spacing is equal to the coil radius.
107
7.5. The parallel filamentary conductors shown in Fig. 8.21 lie in free space. Plot |H| versus
y, 4 < y < 4, along the line x = 0, z = 2: We need an expression for H in cartesian
coordinates. We can start with the known H in cylindrical for an infinite filament along the
z axis: H = I/(2⌅⇧) a⇧ , which we transform to cartesian to obtain:
H=
Iy
Ix
ax +
ay
2
+y )
2⌅(x2 + y 2 )
2⌅(x2
If we now rotate the filament so that it lies along the x axis, with current flowing in positive
x, we obtain the field from the above expression by replacing x with y and y with z:
H=
Iz
Iy
ay +
az
2
+z )
2⌅(y 2 + z 2 )
2⌅(y 2
Now, with two filaments, displaced from the x axis to lie at y = ±1, and with the current
directions as shown in the figure, we use the previous expression to write
⌦
↵
⌦
↵
Iz
Iz
I(y 1)
I(y + 1)
H=
a
+
az
y
2⌅[(y + 1)2 + z 2 ] 2⌅[(y 1)2 + z 2 ]
2⌅[(y 1)2 + z 2 ] 2⌅[(y + 1)2 + z 2 ]
↵
We now evaluate this at z = 2, and find the magnitude ( H · H), resulting in
I
|H| =
2⌅
2
2
y + 2y + 5
y2
2
2y + 5
2
+
(y
y2
1)
2y + 5
(y + 1)
2
y + 2y + 5
✏
2 1/2
This function is plotted below
7.6. A disk of radius a lies in the xy plane, with the z axis through its center. Surface charge of
uniform density ⇧s lies on the disk, which rotates about the z axis at angular velocity ⇥ rad/s.
Find H at any point on the z axis.
We use the Biot-Savart law in the form of Eq. (6), with the following parameters: K =
⇧s v = ⇧s ⇧⇥ a⇧ , R = z 2 + ⇧2 , and aR = (z az ⇧ a⌅ )/R. The di⇤erential field at point
z is
dH =
Kda ⇤ aR
⇧s ⇧ ⇥ a⇧ ⇤ (z az ⇧ a⌅ )
⇧s ⇧ ⇥ (z a⌅ + ⇧ az )
=
⇧ d⇧ d⌥ =
⇧ d⇧ d⌥
2
2
2
3/2
4⌅R
4⌅(z + ⇧ )
4⌅(z 2 + ⇧2 )3/2
108
7.6. (continued). On integrating the above over ⌥ around a complete circle, the a⌅ components
cancel from symmetry, leaving us with
H(z) =
⇠
2⇤
0
=
⇠
a
0
⇧s ⇥
2
⇠ a
⇧s ⇧ ⇥ ⇧ az
⇧s ⇥ ⇧3 az
⇧
d⇧
d⌥
=
d⇧
2
2 3/2
4⌅(z 2 + ⇧2 )3/2
0 2(z + ⇧ )
⌃
⌥⌘
⇣
✏a
2
2
2 /z 2
a
+
2z
1
1
+
a
2
z
⇧s ⇥ ✓
◆ az A/m
z 2 + ⇧2 +
az =
2
2
2z
1 + a2 /z 2
z +⇧ 0
7.7. A filamentary conductor carrying current I in the az direction extends along the entire negative
z axis. At z = 0 it connects to a copper sheet that fills the x > 0, y > 0 quadrant of the xy
plane.
a) Set up the Biot-Savart law and find H everywhere on the z axis (Hint: express a⇧ in terms
of ax and ay and angle ⌥ in the integral): First, the contribution to the field at z from
the current on the negative z axis will be zero, because the cross product, IdL ⇤ aR = 0
for all current elements on the z axis. This leaves the contribution of the current sheet in
the first quadrant. On exiting the origin, current fans out over the first quadrant in the
a⌅ direction and is uniform at a given radius. The surface current density can therefore
be written as K(⇧) = 2I/(⌅⇧) a⌅ A/m2 over the region (0 < ⌥ < ⌅/2). The Biot-Savart
law applicable to surface current is written as
⇠
K ⇤ aR
H=
dA
2
s 4⌅R
where R = z 2 + ⇧2 and aR = (zaz ⇧a⌅ )/ z 2 + ⇧2 Substituting these and integrating
over the first quadrant yields the setup:
H=
⇠
0
⇤/2
⇠
0
⌃
2I a⌅ ⇤ (zaz ⇧a⌅ )
⇧ d⇧ d⌥ =
4⌅ 2 ⇧(z 2 + ⇧2 )3/2
⇠
Following the hint, we now substitute a⇧ = ay cos ⌥
0
⇤/2
⇠
0
⌃
Iz a⇧
d⇧ d⌥
+ ⇧2 )3/2
2⌅ 2 (z 2
ax sin ⌥, and write:
⇠
⇠
⇠ ⌃
Iz ⇤/2 ⌃ (ay cos ⌥ ax sin ⌥)
Iz
d⇧
H=
d⇧ d⌥ = 2 (ax ay )
2
2
2
3/2
2
2⌅ 0
2⌅
(z + ⇧ )
(z + ⇧2 )3/2
0
0
⇧
Iz
⇧
I
⇧⌃
= 2 (ax ay )
⇧ = 2 (ax ay ) A/m
2
2
2
2⌅
2⌅
z
0
z z +⇧
b) repeat part a, but with the copper sheet occupying the entire xy plane. In this case, the
⌥ limits are (0 < ⌥ < 2⌅). The cos ⌥ and sin ⌥ terms would then integrate to zero, so the
answer is just that: H = 0.
109
7.8. For the finite-length current element on the z axis, as shown in Fig. 8.5, use the Biot-Savart
law to derive Eq. (9) of Sec. 8.1: The Biot-Savart law reads:
H=
⇠
z2
z1
IdL ⇤ aR
=
4⌅R2
⇠
⌅ tan
⌅ tan
Idzaz ⇤ (⇧a⌅ zaz )
=
4⌅(⇧2 + z 2 )3/2
2
1
⇠
⌅ tan
⌅ tan
2
1
I⇧a⇧ dz
4⌅(⇧2 + z 2 )3/2
The integral is evaluated (using tables) and gives the desired result:
Iza⇧
H=
⇧⌅ tan
⇧
⇧
tan
I
=
4⌅⇧
2
4⌅⇧ ⇧2 + z 2 ⌅ tan 1
I
=
(sin 2 sin 1 )a⇧
4⌅⇧
tan
2
1 + tan2
1
1 + tan2
2
1
✏
a⇧
7.9. A current sheet K = 8ax A/m flows in the region 2 < y < 2 in the plane z = 0. Calculate
H at P (0, 0, 3): Using the Biot-Savart law, we write
HP =
⇠ ⇠
K ⇤ aR dx dy
=
4⌅R2
⇠
2
2
⇠
⌃
⌃
8ax ⇤ ( xax yay + 3az )
dx dy
4⌅(x2 + y 2 + 9)3/2
Taking the cross product gives:
HP =
⇠
2
2
⇠
⌃
⌃
8( yaz 3ay ) dx dy
4⌅(x2 + y 2 + 9)3/2
We note that the z component is anti-symmetric in y about the origin (odd parity). Since the
limits are symmetric, the integral of the z component over y is zero. We are left with
HP =
=
⇠
2
2
⇠
6
ay
⌅
⌃
⌃
⇠
24 ay dx dy
=
2
4⌅(x + y 2 + 9)3/2
2
2
2
dy =
y2 + 9
6
ay
⌅
⇠
2
x
⇧⌃
⇧
dy
⇧
2
2
2
⌃
2 (y + 9) x + y + 9
⇧
⌃
⌥
12 1
y ⇧2
4
ay tan 1
(2)(0.59) ay = 1.50 ay A/m
⇧ =
⌅
3
3
⌅
2
7.10. A hollow spherical conducting shell of radius a has filamentary connections made at the top
(r = a, ⇥ = 0) and bottom (r = a, ⇥ = ⌅). A direct current I flows down the upper filament,
down the spherical surface, and out the lower filament. Find H in spherical coordinates (a)
inside and (b) outside the sphere.
Applying Ampere’s circuital law, we use a circular contour, centered on the z axis, and
find that within the sphere, no current is enclosed, and so H = 0 when r < a. The same
contour drawn outside the sphere at any z position will always enclose I amps, flowing
in the negative z direction, and so
H=
I
a⇧ =
2⌅⇧
I
a⇧ A/m (r > a)
2⌅r sin ⇥
110
7.11. An infinite filament on the z axis carries 20⌅ mA in the az direction. Three uniform cylindrical
current sheets are also present: 400 mA/m at ⇧ = 1 cm, 250 mA/m at ⇧ = 2 cm, and 300
mA/m at ⇧ = 3 cm. Calculate H⇧ at ⇧ = 0.5, 1.5, 2.5, and 3.5 cm: We find H⇧ at each of the
required radii by applying Ampere’s circuital law to circular paths of those radii; the paths
are centered on the z axis. So, at ⇧1 = 0.5 cm:
H · dL = 2⌅⇧1 H⇧1 = Iencl = 20⌅ ⇤ 10
Thus
H⇧1 =
10 ⇤ 10
⇧1
3
=
10 ⇤ 10
0.5 ⇤ 10
3
2
3
A
= 2.0 A/m
At ⇧ = ⇧2 = 1.5 cm, we enclose the first of the current cylinders at ⇧ = 1 cm. Ampere’s law
becomes:
2⌅⇧2 H⇧2 = 20⌅ + 2⌅(10
2
)(400) mA
H⇧2 =
Following this method, at 2.5 cm:
H⇧3 =
and at 3.5 cm,
10 + 4.00 (2 ⇤ 10
2.5 ⇤ 10 2
H⇧4 =
10 + 4.00
2
)(250)
5.00 (3 ⇤ 10
3.5 ⇤ 10 2
10 + 4.00
= 933 mA/m
1.5 ⇤ 10 2
= 360 mA/m
2
)(300)
=0
7.12. In Fig. 8.22, let the regions 0 < z < 0.3 m and 0.7 < z < 1.0 m be conducting slabs carrying
uniform current densities of 10 A/m2 in opposite directions as shown. The problem asks you
to find H at various positions. Before continuing, we need to know how to find H for this type
of current configuration. The sketch below shows one of the slabs (of thickness D) oriented
with the current coming out of the page. The problem statement implies that both slabs are of
infinite length and width. To find the magnetic field inside a slab, we apply Ampere’s circuital
law to the rectangular path of height d and width w, as shown, since by symmetry, H should
be oriented horizontally. For example, if the sketch below shows the upper slab in Fig. 8.22,
current will be in the positive y direction. Thus H will be in the positive x direction above
the slab midpoint, and will be in the negative x direction below the midpoint.
H out
H out
111
7.12. (continued). In taking the line integral in Ampere’s law, the two vertical path segments will
cancel each other. Ampere’s circuital law for the interior loop becomes
H · dL = 2Hin ⇤ w = Iencl = J ⇤ w ⇤ d
Hin =
Jd
2
The field outside the slab is found similarly, but with the enclosed current now bounded by
the slab thickness, rather than the integration path height:
2Hout ⇤ w = J ⇤ w ⇤ D
Hout =
JD
2
where Hout is directed from right to left below the slab and from left to right above the slab
(right hand rule). Reverse the current, and the fields, of course, reverse direction. We are now
in a position to solve the problem. Find H at:
a) z = 0.2m: Here the fields from the top and bottom slabs (carrying opposite currents)
will cancel, and so H = 0.
b) z = 0.2m. This point lies within the lower slab above its midpoint. Thus the field will
be oriented in the negative x direction. Referring to Fig. 8.22 and to the sketch on the
previous page, we find that d = 0.1. The total field will be this field plus the contribution
from the upper slab current:
H=
10(0.1)
ax
2⌧
10(0.3)
ax =
2⌧
lower slab
2ax A/m
upper slab
c) z = 0.4m: Here the fields from both slabs will add constructively in the negative x
direction:
10(0.3)
H= 2
ax = 3ax A/m
2
d) z = 0.75m: This is in the interior of the upper slab, whose midpoint lies at z = 0.85.
Therefore d = 0.2. Since 0.75 lies below the midpoint, magnetic field from the upper
slab will lie in the negative x direction. The field from the lower slab will be negative
x-directed as well, leading to:
H=
10(0.2)
ax
2⌧
upper slab
10(0.3)
ax =
2⌧
2.5ax A/m
lower slab
e) z = 1.2m: This point lies above both slabs, where again fields cancel completely: Thus
H = 0.
112
7.13. A hollow cylindrical shell of radius a is centered on the z axis and carries a uniform surface
current density of Ka a⇧ .
a) Show that H is not a function of ⌥ or z: Consider this situation as illustrated in Fig.
8.11. There (sec. 8.2) it was stated that the field will be entirely z-directed. We can see
this by applying Ampere’s circuital law to a closed loop path whose orientation we choose
such that current is enclosed by the path. The only way to enclose current is to set up
the loop (which we choose to be rectangular) such that it is oriented with two parallel
opposing segments lying in the z direction; one of these lies inside the cylinder, the other
outside. The other two parallel segments lie in the ⇧ direction. The loop is now cut by the
current sheet, and if we assume a length of the loop in z of d, then the enclosed current
will be given by Kd A. There will be no ⌥ variation in the field because where we position
the loop around the circumference of the cylinder does not a⇤ect the result of Ampere’s
law. If we assume an infinite cylinder length, there will be no z dependence in the field,
since as we lengthen the loop in the z direction, the path length (over which the integral
is taken) increases, but then so does the enclosed current – by the same factor. Thus H
would not change with z. There would also be no change if the loop was simply moved
along the z direction.
b) Show that H⇧ and H⌅ are everywhere zero. First, if H⇧ were to exist, then we should be
able to find a closed loop path that encloses current, in which all or or portion of the path
lies in the ⌥ direction. This we cannot do, and so H⇧ must be zero. Another argument is
that when applying the Biot-Savart law, there is no current element that would produce
a ⌥ component. Again, using the Biot-Savart law, we note that radial field components
will be produced by individual current elements, but such components will cancel from
two elements that lie at symmetric distances in z on either side of the observation point.
c) Show that Hz = 0 for ⇧ > a: Suppose the rectangular loop was drawn such that the
outside z-directed segment is moved further and further away from the cylinder. We
would expect Hz outside to decrease (as the Biot-Savart law would imply) but the same
amount of current is always enclosed no matter how far away the outer segment is. We
therefore must conclude that the field outside is zero.
d) Show that Hz = Ka for ⇧ < a: With our rectangular path set up as in part a, we have no
path integral contributions from the two radial segments, and no contribution from the
outside z-directed segment. Therefore, Ampere’s circuital law would state that
H · dL = Hz d = Iencl = Ka d
Hz = Ka
where d is the length of the loop in the z direction.
e) A second shell, ⇧ = b, carries a current Kb a⇧ . Find H everywhere: For ⇧ < a we would
have both cylinders contributing, or Hz (⇧ < a) = Ka + Kb . Between the cylinders, we are
outside the inner one, so its field will not contribute. Thus Hz (a < ⇧ < b) = Kb . Outside
(⇧ > b) the field will be zero.
113
7.14. A toroid having a cross section of rectangular shape is defined by the following surfaces: the
cylinders ⇧ = 2 and ⇧ = 3 cm, and the planes z = 1 and z = 2.5 cm. The toroid carries a
surface current density of 50az A/m on the surface ⇧ = 3 cm. Find H at the point P (⇧, ⌥, z):
The construction is similar to that of the toroid of round cross section as done on p.239. Again,
magnetic field exists only inside the toroid cross section, and is given by
H=
Iencl
a⇧
2⌅⇧
(2 < ⇧ < 3) cm, (1 < z < 2.5) cm
where Iencl is found from the given current density: On the outer radius, the current is
Iouter =
50(2⌅ ⇤ 3 ⇤ 10
2
)=
3⌅ A
This current is directed along negative z, which means that the current on the inner radius
(⇧ = 2) is directed along positive z. Inner and outer currents have the same magnitude. It is
the inner current that is enclosed by the circular integration path in a⇧ within the toroid that
is used in Ampere’s law. So Iencl = +3⌅ A. We can now proceed with what is requested:
a) PA (1.5cm, 0, 2cm): The radius, ⇧ = 1.5 cm, lies outside the cross section, and so HA = 0.
b) PB (2.1cm, 0, 2cm): This point does lie inside the cross section, and the ⌥ and z values do
not matter. We find
HB =
Iencl
3a⇧
a⇧ =
2⌅⇧
2(2.1 ⇤ 10
2)
= 71.4 a⇧ A/m
c) PC (2.7cm, ⌅/2, 2cm): again, ⌥ and z values make no di⇤erence, so
HC =
3a⇧
2(2.7 ⇤ 10
2)
= 55.6 a⇧ A/m
d) PD (3.5cm, ⌅/2, 2cm). This point lies outside the cross section, and so HD = 0.
7.15. Assume that there is a region with cylindrical symmetry in which the conductivity is given by
⌃ = 1.5e 150⌅ kS/m. An electric field of 30 az V/m is present.
a) Find J: Use
J = ⌃E = 45e 150⌅ az kA/m2
b) Find the total current crossing the surface ⇧ < ⇧0 , z = 0, all ⌥:
⇠ ⇠
⇠ 2⇤ ⇠ ⌅0
2⌅(45) 150⌅
I=
J · dS =
45e 150⌅ ⇧ d⇧ d⌥ =
e
[ 150⇧
(150)2
0
0
⇤
⌅
= 12.6 1 (1 + 150⇧0 )e 150⌅0 A
⇧⌅0
⇧
1] ⇧ kA
0
c) Make use of Ampere’s circuital law to find H: Symmetry suggests that H will be ⌥directed only, and so we consider a circular path of integration, centered on and perpendicular to the z axis. Ampere’s law becomes: 2⌅⇧H⇧ = Iencl , where Iencl is the current
found in part b, except with ⇧0 replaced by the variable, ⇧. We obtain
H⇧ =
2.00 ⇤
1
⇧
(1 + 150⇧)e
114
150⌅
⌅
A/m
7.16. A current filament carrying I in the az direction lies along the entire positive z axis. At the
origin, it connects to a conducting sheet that forms the xy plane.
a) Find K in the conducting sheet: The current fans outward radially with uniform surface
current density at a fixed radius. The current density at radius ⇧ will be the total current,
I, divided by the circumference at radius ⇧:
K=
I
a⌅ A/m
2⌅⇧
b) Use Ampere’s circuital law to find H everywhere for z > 0: Circular lines of H are
expected, centered on the z axis – in the a⇧ direction. Ampere’s law is set up by
considering a circular path integral taken around the wire at fixed z. The enclosed current
is that which passes through any surface that is bounded by the line integration path:
H · dL = 2⌅⇧H⇧ = Iencl
If the surface is that of the disk whose perimeter is the integration path, then the enclosed
current is just I, and the magnetic field becomes
H⇧ =
I
2⌅⇧
H=
I
a⇧ A/m
2⌅⇧
But the disk surface can be “stretched” so that it forms a balloon shape. Suppose the
“balloon” is a right circular cylinder, with its open top circumference at the path integral
location. The cylinder extends downward, intersecting the surface current in the x-y
plane, with the bottom of the cylinder below the x-y plane. Now, the path integral is
unchanged from before, and the enclosed current is the radial current in the x-y plane
that passes through the side of the cylinder. This current will be I = 2⌅⇧[I/(2⌅⇧)] = I,
as before. So the answer given above for H applies to anywhere in the region z > 0.
c) Find H for z < 0: Consider the same cylinder as in part b, except take the path integral of H around the bottom circumference (below the x-y plane). The enclosed current
now consists of the filament current that enters through the top, plus the radial current
that exits though the side. The two currents are equal magnitude but opposite in sign.
Therefore, the net enclosed current is zero, and thus H = 0 (z < 0).
115
7.17. A current filament on the z axis carries a current of 7 mA in the az direction, and current
sheets of 0.5 az A/m and 0.2 az A/m are located at ⇧ = 1 cm and ⇧ = 0.5 cm, respectively.
Calculate H at:
a) ⇧ = 0.5 cm: Here, we are either just inside or just outside the first current sheet, so
both we will calculate H for both cases. Just inside, applying Ampere’s circuital law to
a circular path centered on the z axis produces:
2⌅⇧H⇧ = 7 ⇤ 10
H(just inside) =
3
7 ⇤ 10 3
2⌅(0.5 ⇤ 10
2
a⇧ = 2.2 ⇤ 10
1
a⇧ A/m
Just outside the current sheet at .5 cm, Ampere’s law becomes
2⌅⇧H⇧ = 7 ⇤ 10
2⌅(0.5 ⇤ 10
3
H(just outside) =
2
)(0.2)
7.2 ⇤ 10 4
a⇧ = 2.3 ⇤ 10
2⌅(0.5 ⇤ 10 2 )
2
a⇧ A/m
b) ⇧ = 1.5 cm: Here, all three currents are enclosed, so Ampere’s law becomes
2⌅(1.5 ⇤ 10
2
)H⇧ = 7 ⇤ 10
3
6.28 ⇤ 10
3
H(⇧ = 1.5) = 3.4 ⇤ 10
+ 2⌅(10
1
2
)(0.5)
a⇧ A/m
c) ⇧ = 4 cm: Ampere’s law as used in part b applies here, except we replace ⇧ = 1.5 cm with
⇧ = 4 cm on the left hand side. The result is H(⇧ = 4) = 1.3 ⇤ 10 1 a⇧ A/m.
d) What current sheet should be located at ⇧ = 4 cm so that H = 0 for all ⇧ > 4 cm? We
require that the total enclosed current be zero, and so the net current in the proposed
cylinder at 4 cm must be negative the right hand side of the first equation in part b. This
will be 3.2 ⇤ 10 2 , so that the surface current density at 4 cm must be
K=
3.2 ⇤ 10 2
az =
2⌅(4 ⇤ 10 2 )
1.3 ⇤ 10
1
az A/m
7.18. A wire of 3-mm radius is made up of an inner material (0 < ⇧ < 2 mm) for which ⌃ = 107
S/m, and an outer material (2mm < ⇧ < 3mm) for which ⌃ = 4 ⇤ 107 S/m. If the wire carries
a total current of 100 mA dc, determine H everywhere as a function of ⇧.
Since the materials have di⇤erent conductivities, the current densities within them will
di⇤er. Electric field, however is constant throughout. The current can be expressed as
⇤
⌅
I = ⌅(.002)2 J1 + ⌅[(.003)2 (.002)2 ]J2 = ⌅ (.002)2 ⌃1 + [(.003)2 (.002)2 ]⌃2 E
Solve for E to obtain
E=
⌅[(4 ⇤ 10
6 )(107 )
0.1
+ (9 ⇤ 10
6
4 ⇤ 10
6 )(4
⇤ 107 )]
= 1.33 ⇤ 10
4
V/m
We next apply Ampere’s circuital law to a circular path of radius ⇧, where ⇧ < 2mm:
2⌅⇧H⇧1 = ⌅⇧2 J1 = ⌅⇧2 ⌃1 E
116
H⇧1 =
⌃1 E⇧
= 663 A/m
2
7.18 (continued) . Next, for the region 2mm < ⇧ < 3mm, Ampere’s law becomes
2⌅⇧H⇧2 = ⌅[(4 ⇤ 10
6
)(107 ) + (⇧2
4 ⇤ 10
6
8.0 ⇤ 10
⇧
H⇧2 = 2.7 ⇤ 103 ⇧
3
)(4 ⇤ 107 )]E
A/m
Finally, for ⇧ > 3mm, the field outside is that for a long wire:
I
0.1
1.6 ⇤ 10
=
=
2⌅⇧
2⌅⇧
⇧
H⇧3 =
2
A/m
7.19. In spherical coordinates, the surface of a solid conducting cone is described by ⇥ = ⌅/4 and a
conducting plane by ⇥ = ⌅/2. Each carries a total current I. The current flows as a surface
current radially inward on the plane to the vertex of the cone, and then flows radially-outward
throughout the cross-section of the conical conductor.
a) Express the surface current density as a function of r: This will be the total current
divided by the circumference of a circle of radius r in the plane, directed toward the
origin:
I
K(r) =
ar A/m2 (⇥ = ⌅/2)
2⌅r
b) Express the volume current density inside the cone as a function of r: This will be the
total current divided by the area of the spherical cap subtending angle ⇥ = ⌅/4:
J(r) = I
⇠
0
2⇤
⇠
⇤/4
✏
1
r sin ⇥ d⇥ d⌥
⇧
2
⇧
0
ar =
I ar
↵ A/m2
1/ 2)
2⌅r2 (1
(0 < ⇥ < ⌅/4)
c) Determine H as a function of r and ⇥ in the region between the cone and the plane: From
symmetry, we expect H to be ⌥-directed and uniform at constant r and ⇥. Ampere’s
circuital law can therefore be stated as:
H · dL = 2⌅r sin ⇥ H⇧ = I
H=
I
a⇧ A/m (⌅/4 < ⇥ < ⌅/2)
2⌅r sin ⇥
d) Determine H as a function of r and ⇥ inside the cone: Again, ⌥-directed H is anticipated,
so we apply Ampere’s law in the following way:
H · dL = 2⌅r sin ⇥ H⇧ =
⇠
s
J · dS =
⇠
2⇤
0
⇠
This becomes
2⌅r sin ⇥ H⇧ =
or
H=
I
2⌅r(1
↵
1/ 2)
⌦
(1
2⌅
⇥
0
I
2⌅(1
I ar
↵ · ar r2 sin ⇥⇧ d⇥⇧ d⌥
1/ 2)
2⌅r2 (1
⇧⇥
⇧⇧
↵ cos ⇥ ⇧
0
1/ 2)
↵
cos ⇥)
a⇧ A/m (0 < ⇥ < ⌅/4)
sin ⇥
As a test of this, note that the inside and outside fields (results of parts c and d) are equal
at the cone surface (⇥ = ⌅/4) as they must be.
117
7.20. A solid conductor of circular cross-section with a radius of 5 mm has a conductivity that varies
with radius. The conductor is 20 m long and there is a potential di⇤erence of 0.1 V dc between
its two ends. Within the conductor, H = 105 ⇧2 a⇧ A/m.
a) Find ⌃ as a function of ⇧: Start by finding J from H by taking the curl. With H
⌥-directed, and varying with radius only, the curl becomes:
J=
⇤H=
⇥
1 d
1 d
(⇧H⇧ ) az =
105 ⇧3 az = 3 ⇤ 105 ⇧ az A/m2
⇧ d⇧
⇧ d⇧
Then E = 0.1/20 = 0.005 az V/m, which we then use with J = ⌃E to find
⌃=
J
3 ⇤ 105 ⇧
=
= 6 ⇤ 107 ⇧ S/m
E
0.005
b) What is the resistance between the two ends? The current in the wire is
I=
⇠
s
J · dS = 2⌅
⇠
0
a
(3 ⇤ 105 ⇧) ⇧ d⇧ = 6⌅ ⇤ 105
Finally, R = V0 /I = 0.1/0.079 = 1.3 ⇥
118
1 3
a
3
= 2⌅ ⇤ 105 (0.005)3 = 0.079 A
7.21. A cylindrical wire of radius a is oriented with the z axis down its center line. The wire carries
a non-uniform current down its length of density J = b⇧ az A/m2 , where b is a constant.
a) What total current flows in the wire? We integrate the current density over the wire
cross-section:
⇠
⇠ 2⇤ ⇠ a
2⌅ba3
Itot = J · dS =
b⇧ az · az ⇧ d⇧ d⌥ =
A
3
s
0
0
b) find Hin (0 < ⇧ < a), as a function of ⇧: From the symmetry, ⌥-directed H (= H⇧ a⇧ )
is expected in the interior; this will be constant at a fixed radius, ⇧. Apply Ampere’s
circuital law to a circular path of radius ⇧ inside:
Hin · dL = 2⌅⇧H⇧,in =
⇠
So that
Hin =
s
J · dS =
⇠
2⇤
0
⇠
⌅
0
b⇧⇧ az · az ⇧⇧ d⇧⇧ d⌥ =
2⌅b⇧3
3
b⇧2
a⇧ A/m (0 < ⇧ < a)
3
c) find Hout (⇧ > a), as a function of ⇧ Same as part b, except the path integral is taken at
a radius outside the wire:
⇠
⇠ 2⇤ ⇠ a
2⌅ba3
Hout · dL = 2⌅⇧H⇧,out = J · dS =
b⇧ az · az ⇧ d⇧ d⌥ =
3
s
0
0
So that
Hout =
ba3
a⇧ A/m (⇧ > a)
3⇧
d) verify your results of parts b and c by using ⇤ H = J: With a ⌥ component of H only,
varying only with ⇧, the curl in cylindrical coordinates reduces to
J=
⇤H=
1 d
(⇧H⇧ ) az
⇧ d⇧
Apply this to the inside field to get
Jin =
1 d
⇧ d⇧
b⇧3
3
az = b⇧ az
For the outside field, we find
Jout =
1 d
⇧ d⇧
as expected.
119
⇧ba3
3⇧
az = 0
7.22. A solid cylinder of radius a and length L, where L >> a, contains volume charge of uniform
density ⇧0 C/m3 . The cylinder rotates about its axis (the z axis) at angular velocity ⇥ rad/s.
a) Determine the current density J, as a function of position within the rotating cylinder:
Use J = ⇧0 v = ⇧0 ⇧⇥ a⇧ A/m2 .
b) Determine the magnetic field intensity H inside and outside: It helps initially to obtain
the field on-axis. To do this, we use the result of Problem 8.6, but give the rotating
charged disk in that problem a di⇤erential thickness, dz. We can then evaluate the onaxis field in the rotating cylinder as the superposition of fields from a stack of disks which
exist between ±L/2. Here, we make the problem easier by letting L ⌥
(since L >> a)
thereby specializing our evaluation to positions near the half-length. The on-axis field is
therefore:
⌃
⌥⌘
⇣
2
2
⇠ ⌃
2 /z 2
a
+
2z
1
1
+
a
⇧0 ⇥ ✓
◆ dz
Hz (⇧ = 0) =
1 + a2 /z 2
⌃ 2z
⌦
↵
⇠ ⌃
⇧0 ⇥
a2
2z 2
↵
=2
+↵
2z dz
2
z 2 + a2
z 2 + a2
0
⌦ 2
↵⌃
a
z
a2
z2
2
2
2
2
2
2
= 2⇧0 ⇥
ln(z + z + a ) +
z +a
ln(z + z + a )
2
2
2
2 0
⇡
⇢⌃
⇡
⇢
= ⇧0 ⇥ z z 2 + a2 z 2
= ⇧0 ⇥ z z 2 + a2 z 2
0
z⌅⌃
Using the large z approximation in the radical, we obtain
⌦
↵
a2
⇧0 ⇥a2
2
2
Hz (⇧ = 0) = ⇧0 ⇥ z 1 + 2
z =
2z
2
To find the field as a function of radius, we apply Ampere’s circuital law to a rectangular
loop, drawn in two locations described as follows: First, construct the rectangle with one
side along the z axis, and with the opposite side lying at any radius outside the cylinder.
In taking the line integral of H around the rectangle, we note that the two segments that
are perpendicular to the cylinder axis will have their path integrals exactly cancel, since
the two path segments are oppositely-directed, while from symmetry the field should not
be di⇤erent along each segment. This leaves only the path segment that coindides with
the axis, and that lying parallel to the axis, but outside. Choosing the length of these
segments to be , Ampere’s circuital law becomes:
⇠
⇠ ✓⇠ a
H · dL = Hz (⇧ = 0) + Hz (⇧ > a) = Iencl = J · dS =
⇧0 ⇧⇥ a⇧ · a⇧ d⇧ dz
s
0
0
⇧0 ⇥a2
=
2
But we found earlier that Hz (⇧ = 0) = ⇧0 ⇥a2 /2. Therefore, we identify the outside field,
Hz (⇧ > a) = 0. Next, change the rectangular path only by displacing the central path
component o⇤-axis by distance ⇧, but still lying within the cylinder. The enclosed current
is now somewhat less, and Ampere’s law becomes
⇠
⇠ ✓⇠ a
H · dL = Hz (⇧) + Hz (⇧ > a) = Iencl = J · dS =
⇧0 ⇧⇧ ⇥ a⇧ · a⇧ d⇧ dz
s
⇧0 ⇥ 2
=
(a
2
⇧2 )
⇧0 ⇥ 2
H(⇧) =
(a
2
120
0
⌅
⇧2 ) az A/m
7.22c) Check your result of part b by taking the curl of H. With H z-directed, and varying only with
⇧, the curl in cylindrical coordinates becomes
dHz
a⇧ = ⇧0 ⇥⇧ a⇧ A/m2 = J
d⇧
⇤H=
as expected.
7.23. Given the field H = 20⇧2 a⇧ A/m:
a) Determine the current density J: This is found through the curl of H, which simplifies
to a single term, since H varies only with ⇧ and has only a ⌥ component:
J=
⇤H=
⇥
1 d(⇧H⇧ )
1 d
az =
20⇧3 az = 60⇧ az A/m2
⇧ d⇧
⇧ d⇧
b) Integrate J over the circular surface ⇧ = 1, 0 < ⌥ < 2⌅, z = 0, to determine the total
current passing through that surface in the az direction: The integral is:
I=
⇠ ⇠
J · dS =
⇠
2⇤
0
⇠
0
1
60⇧az · ⇧ d⇧ d⌥az = 40⌅ A
c) Find the total current once more, this time by a line integral around the circular path
⇧ = 1, 0 < ⌥ < 2⌅, z = 0:
I=
H · dL =
⇠
0
2⇤
⇧
20⇧ a⇧ ⇧⌅=1 · (1)d⌥a⇧ =
2
121
⇠
0
2⇤
20 d⌥ = 40⌅ A
7.24. Infinitely-long filamentary conductors are located in the y = 0 plane at x = n meters where
n = 0, ±1, ±2, . . . Each carries 1 A in the az direction.
a) Find H on the y axis. As a help,
⌃
⌫
1
⌅
+
2y e2⇤y
y
⌅
=
2 + n2
y
2
n=1
1
We can begin by determining the field on the y axis arising from two wires only, located
at x = ±n. We start from basics, using the Biot-Savart law:
⇠
IdL ⇤ aR
H=
4⌅R2
where R = n2 + y 2 + z 2
a+
R =
⇥2
for both wires, and where, for the wire at x = +n
nax + yay + zaz
(n2 + y 2 +
and
1/2
z2)
aR =
nax + yay + zaz
1/2
(n2 + y 2 + z 2 )
for the wire located at x = n. Then with IdL = dz az (I = 1) for both wires, the
Biot-Savart construction becomes
H=
⇠
⌃
dz az ⇤ [( nax + yay + zaz ) + (nax + yay + zaz )]
⌃
4⌅ (n2
+
y2
+
3/2
z2)
This evaluates as
H=
1
⌅
n2
y
+ y2
=
yax
2⌅
⇠
⌃
⌃ (n2
dz
+
y2
3/2
+ z2)
ax A/m
Now if we include all wire pairs, the result is the superposition of an infinite number of
fields of the above form. Specifically,
Hnet =
⌃
1⌫
⌅ n=1
y
n2 + y 2
ax A/m
Using the given closed form of the series expansion, our final answer is
⌦
↵
1
1
1
Hnet =
+
ax A/m
2 2⌅y e2⇤y 1
b) Compare your result of part a to that obtained if the filaments are replaced by a current
sheet in the y = 0 plane that carries surface current density K = 1az A/m: This is found
through Eq. (11):
1
1
1
H = K ⇤ aN = az ⇤ ay =
ax
2
2
2
Our answer of part a approaches this value as y ⌥
, demonstrating that at large
distances, the parallel wires act like a current sheet. Interestingly, at close-in locations,
.
such that 2⌅y << 1, we may expand e2⇤y = 1 + 2⌅y, leading to the cancellation of the
.
last two terms in the part a result, and again, Hnet = 1/2 ax .
122
7.25. When x, y, and z are positive and less than 5, a certain magnetic field intensity may be
expressed as H = [x2 yz/(y + 1)]ax + 3x2 z 2 ay [xyz 2 /(y + 1)]az . Find the total current in the
ax direction that crosses the strip, x = 2, 1 ⌃ y ⌃ 4, 3 ⌃ z ⌃ 4, by a method utilizing:
a) a surface integral: We need to find the current density by taking the curl of the given H.
Actually, since the strip lies parallel to the yz plane, we need only find the x component
of the current density, as only this component will contribute to the requested current.
This is
Hz
Hy
xz 2
Jx = ( ⇤ H)x =
=
+ 6x2 z ax
y
z
(y + 1)2
The current through the strip is then
I=
⇠
s
=
J · ax da =
⇠
3
4
⇠
3
3 2
z + 72z
5
4
⇠
4
1
2z 2
+ 24z
(y + 1)2
1 3
z + 36z 2
5
dz =
dy dz =
⇠
3
4
=
4
2z 2
+ 24zy
(y + 1)
4
dz
1
259
3
b) a closed line integral: We integrate counter-clockwise around the strip boundary (using
the right-hand convention), where the path normal is positive ax . The current is then
I=
H · dL =
= 108(3)
⇠
4
3(2)2 (3)2 dy +
1
8 3
(4
15
33 ) + 192(1
⇠
⇠ 1
⇠ 3
2(4)z 2
dz +
3(2)2 (4)2 dy +
(4 + 1)
3
4
4
1 3
3
4)
(3
4 ) = 259
3
4
2(1)z 2
dz
(1 + 1)
7.26. Consider a sphere of radius r = 4 centered at (0,0,3). Let S1 be that portion of the spherical
surface that lies above the xy plane. Find S1 ( ⇤ H) · dS if H = 3⇧ a⇧ in cylindrical
coordinates: ↵
First, the intersection
of the sphere with the x-y plane is a disk in the plane of
↵
2
2
radius ⇧d = 4
3 = 7. The curl of the given field (having a ⌥ component that varies
only with ⇧) is
⇥
1 d
1 d
⇤H=
(⇧H⇧ ) az =
3⇧2 az = 6 az
⇧ d⇧
⇧ d⇧
Since this is a constant field, its flux through S1 will simply be the flux through the disk,
which in turn is simply the product of the field with the disk area:
↵
= 6 ⇤ ⌅( 7)2 = 42⌅
Another way to solve the problem is to use Stokes’ theorem and write the flux as
=
H · dL
where the path integral is taken around the disk perimeter. Doing this gives:
↵
↵
= 2⌅⇧d H⇧ |⌅d = 2⌅ 7 ⇤ 3 7 = 42⌅
123
7.27. The magnetic field intensity is given in a certain region of space as
H=
a) Find
to
x + 2y
2
ay + az A/m
2
z
z
⇤H: For this field, the general curl expression in rectangular coordinates simplifies
⇤H=
Hy
Hy
2(x + 2y)
1
ax +
az =
ax + 2 az A/m
3
z
x
z
z
b) Find J: This will be the answer of part a, since
⇤ H = J.
c) Use J to find the total current passing through the surface z = 4, 1 < x < 2, 3 < y < 5,
in the az direction: This will be
I=
⇠ ⇠
⇧
J⇧z=4 · az dx dy =
⇠
5
3
⇠
1
2
1
dx dy = 1/8 A
42
d) Show
that the same result is obtained using the other side of Stokes’ theorem: We take

H · dL over the square path at z = 4 as defined in part c. This involves two integrals of
the y component of H over the range 3 < y < 5. Integrals over x, to complete the loop,
do not exist since there is no x component of H. We have
I=
⇧
H⇧z=4 · dL =
⇠
3
5
2 + 2y
dy +
16
124
⇠
5
3
1 + 2y
1
dy = (2)
16
8
1
(2) = 1/8 A
16
7.28. Given H = (3r2 / sin ⇥)a⇥ + 54r cos ⇥a⇧ A/m in free space:
a) find the total current in the a⇥ direction through the conical surface ⇥ = 20⇤ , 0 ⌃ ⌥ ⌃ 2⌅,
0 ⌃ r ⌃ 5, by whatever side of Stokes’ theorem you like best. I chose the line integral
side, where the integration path is the circular path in ⌥ around the top edge of the cone,
at r = 5. The path direction is chosen to be clockwise looking down on the xy plane.
This, by convention, leads to the normal from the cone surface that points in the positive
a⇥ direction (right hand rule). We find
H · dL =
=
⇠
2⇤
0
⇤ 2
⌅
(3r / sin ⇥)a⇥ + 54r cos ⇥a⇧ r=5,⇥=20 · 5 sin(20⇤ ) d⌥ ( a⇧ )
2⌅(54)(25) cos(20⇤ ) sin(20⇤ ) =
2.73 ⇤ 103 A
This result means that there is a component of current that enters the cone surface in
the a⇥ direction, to which is associated a component of H in the positive a⇧ direction.
b) Check the result by using the other side of Stokes’ theorem: We first find the current
density through the curl of the magnetic field, where three of the six terms in the spherical
coordinate formula survive:
⇤H=
1
(54r cos ⇥ sin ⇥)) ar
r sin ⇥ ⇥
⇥
1
1
54r2 cos ⇥ a⇥ +
r r
r r
Thus
3r3
sin ⇥
a⇧ = J
9r
a⇧
sin ⇥
The calculation of the other side of Stokes’ theorem now involves integrating J over the
surface of the cone, where the outward normal is positive a⇥ , as defined in part a:
J = 54 cot ⇥ ar
⇠
S
(
⇤ H) · dS =
=
⇠
2⇤
0
⇠
0
=
⇠
5
⌦
54 cot ⇥ ar
0
2⇤ ⇠ 5
108 cos ⇥ a⇥ +
9r
108 cos ⇥ a⇥ +
a⇧
sin ⇥
108 cos(20⇤ ) sin(20⇤ )rdrd⌥ =
0
2.73 ⇤ 103 A
125
↵
20⇥
· a⇥ r sin(20⇤ ) dr d⌥
2⌅(54)(25) cos(20⇤ ) sin(20⇤ )
7.29. A long straight non-magnetic conductor of 0.2 mm radius carries a uniformly-distributed
current of 2 A dc.
a) Find J within the conductor: Assuming the current is +z directed,
J=
2
⌅(0.2 ⇤ 10
3 )2
az = 1.59 ⇤ 107 az A/m2
b) Use Ampere’s circuital law to find H and B within the conductor: Inside, at radius ⇧, we
have
⇧J
2⌅⇧H⇧ = ⌅⇧2 J
H=
a⇧ = 7.96 ⇤ 106 ⇧ a⇧ A/m
2
Then B = µ0 H = (4⌅ ⇤ 10 7 )(7.96 ⇤ 106 )⇧a⇧ = 10⇧ a⇧ Wb/m2 .
c) Show that
⇤ H = J within the conductor: Using the result of part b, we find,
⇤H=
1 d
1 d
(⇧H⇧ ) az =
⇧ d⇧
⇧ d⇧
1.59 ⇤ 107 ⇧2
2
az = 1.59 ⇤ 107 az A/m2 = J
d) Find H and B outside the conductor (note typo in book): Outside, the entire current is
enclosed by a closed path at radius ⇧, and so
H=
I
1
a⇧ =
a⇧ A/m
2⌅⇧
⌅⇧
Now B = µ0 H = µ0 /(⌅⇧) a⇧ Wb/m2 .
e) Show that
write:
⇤ H = J outside the conductor: Here we use H outside the conductor and
⇤H=
1 d
1 d
(⇧H⇧ ) az =
⇧ d⇧
⇧ d⇧
⇧
1
⌅⇧
az = 0 (as expected)
7.30. (an inversion of Problem 8.20). A solid nonmagnetic conductor of circular cross-section has
a radius of 2mm. The conductor is inhomogeneous, with ⌃ = 106 (1 + 106 ⇧2 ) S/m. If the
conductor is 1m in length and has a voltage of 1mV between its ends, find:
a) H inside: With current along the cylinder length (along az , and with ⌥ symmetry, H
will be ⌥-directed only. We find E = (V0 /d)az = 10 3 az V/m. Then J = ⌃E =
103 (1 + 106 ⇧2 )az A/m2 . Next we apply Ampere’s circuital law to a circular path of radius
⇧, centered on the z axis and normal to the axis:
⇠ ⇠
⇠ 2⇤ ⇠ ⌅
H · dL = 2⌅⇧H⇧ =
J · dS =
103 (1 + 106 (⇧⇧ )2 )az · az ⇧⇧ d⇧⇧ d⌥
S
Thus
H⇧ =
103
⇧
⇠
⌅
0
0
⇧⇧ + 106 (⇧⇧ )3 d⇧⇧ =
0
⌦
↵
103 ⇧2
106 4
+
⇧
⇧
2
4
Finally, H = 500⇧(1 + 5 ⇤ 10 ⇧ )a⇧ A/m (0 < ⇧ < 2mm).
5 3
b) the total magnetic flux inside the conductor: With field in the ⌥ direction, a plane normal
to B will be that in the region 0 < ⇧ < 2 mm, 0 < z < 1 m. The flux will be
⇠ ⇠
⇠ 1 ⇠ 2⇥10 3
⇥
=
B·dS = µ0
500⇧ + 2.5 ⇤ 108 ⇧3 d⇧dz = 8⌅⇤10 10 Wb = 2.5 nWb
S
0
0
126
7.31. The cylindrical shell defined by 1 cm < ⇧ < 1.4 cm consists of a non-magnetic conducting
material and carries a total current of 50 A in the az direction. Find the total magnetic flux
crossing the plane ⌥ = 0, 0 < z < 1:
a) 0 < ⇧ < 1.2 cm: We first need to find J, H, and B: The current density will be:
50
J=
2 )2
⌅[(1.4 ⇤ 10
(1.0 ⇤ 10
az = 1.66 ⇤ 105 az A/m2
2 )2 ]
Next we find H⇧ at radius ⇧ between 1.0 and 1.4 cm, by applying Ampere’s circuital law,
and noting that the current density is zero at radii less than 1 cm:
2⌅⇧H⇧ = Iencl =
⇠
2⇤
0
⇠
⌅
10
2
1.66 ⇤ 105 ⇧⇧ d⇧⇧ d⌥
H⇧ = 8.30 ⇤ 104
(⇧2
Then B = µ0 H, or
B = 0.104
10
⇧
(⇧2
4
)
A/m (10
10
⇧
4
)
2
m < ⇧ < 1.4 ⇤ 10
2
m)
a⇧ Wb/m2
Now,
a
⌦
↵
⇠ 1 ⇠ 1.2⇥10 2
10 4
B · dS =
0.104 ⇧
d⇧ dz
⇧
0
10 2
⌦
↵
(1.2 ⇤ 10 2 )2 10 4
1.2
4
= 0.104
10 ln
= 3.92 ⇤ 10
2
1.0
=
⇠ ⇠
7
Wb = 0.392 µWb
b) 1.0 cm < ⇧ < 1.4 cm (note typo in book): This is part a over again, except we change the
upper limit of the radial integration:
b
=
⇠ ⇠
⇠
1
⇠
1.4⇥10
B · dS =
0
10 2
⌦
(1.4 ⇤ 10 2 )2 10
= 0.104
2
2
4
⌦
0.104 ⇧
↵
10 4
d⇧ dz
⇧
↵
1.4
10 4 ln
= 1.49 ⇤ 10
1.0
6
Wb = 1.49 µWb
c) 1.4 cm < ⇧ < 20 cm: This is entirely outside the current distribution, so we need B there:
We modify the Ampere’s circuital law result of part a to find:
Bout = 0.104
[(1.4 ⇤ 10
)
2 2
10
4
ln
20
1.4
⇧
]
a⇧ =
10
⇧
5
a⇧ Wb/m2
We now find
c
=
⇠
0
1
⇠
20⇥10
1.4⇥10
2
2
10
⇧
5
d⇧ dz = 10
127
5
= 2.7 ⇤ 10
5
Wb = 27 µWb
7.32. The free space region defined by 1 < z < 4 cm and 2 < ⇧ < 3 cm is a toroid of rectangular
cross-section. Let the surface at ⇧ = 3 cm carry a surface current K = 2az kA/m.
a) Specify the current densities on the surfaces at ⇧ = 2 cm, z = 1cm, and z = 4cm.
All surfaces must carry equal currents. With this requirement, we find: K(⇧ = 2) =
3 az kA/m. Next, the current densities on the z = 1 and z = 4 surfaces must transistion
between the current density values at ⇧ = 2 and ⇧ = 3. Knowing the the radial current
density will vary as 1/⇧, we find K(z = 1) = (60/⇧)a⌅ A/m with ⇧ in meters. Similarly,
K(z = 4) = (60/⇧)a⌅ A/m.
b) Find H everywhere: Outside the toroid, H = 0. Inside, we apply Ampere’s circuital law
in the manner of Problem 8.14:
⇠ 2⇤
H · dL = 2⌅⇧H⇧ =
K(⇧ = 2) · az (2 ⇤ 10 2 ) d⌥
0
H=
2⌅(3000)(.02)
a⇧ =
⇧
60/⇧ a⇧ A/m (inside)
c) Calculate the total flux within the toriod: We have B =
=
⇠
.04
.01
⇠
.03
.02
(60µ0 /⇧)a⇧ Wb/m2 . Then
60µ0
a⇧ · ( a⇧ ) d⇧ dz = (.03)(60)µ0 ln
⇧
3
2
= 0.92 µWb
7.33. Use an expansion in rectangular coordinates to show that the curl of the gradient of any scalar
field G is identically equal to zero. We begin with
G
G
G
ax +
ay +
az
x
y
z
G=
and
⇤
G=
+
⌦
⌦
y
x
G
z
G
y
z
y
↵
⌦
G
G
ax +
y
z
x
↵
G
az = 0 for any G
x
128
x
G
z
↵
ay
7.34. A filamentary conductor on the z axis carries a current of 16A in the az direction, a conducting
shell at ⇧ = 6 carries a total current of 12A in the az direction, and another shell at ⇧ = 10
carries a total current of 4A in the az direction.

a) Find H for 0 < ⇧ < 12: Ampere’s circuital law states that H · dL = Iencl , where the
line integral and current direction are related in the usual way through the right hand
rule. Therefore, if I is in the positive z direction, H is in the a⇧ direction. We proceed
as follows:
0 < ⇧ < 6 : 2⌅⇧H⇧ = 16
H = 16/(2⌅⇧)a⇧
6 < ⇧ < 10 : 2⌅⇧H⇧ = 16
⇧ > 10 : 2⌅⇧H⇧ = 16
12
12
H = 4/(2⌅⇧)a⇧
4=0
H=0
b) Plot H⇧ vs. ⇧:
c) Find the total flux
=
⇠
0
1
⇠
1
6
crossing the surface 1 < ⇧ < 7, 0 < z < 1: This will be
16µ0
d⇧ dz +
2⌅⇧
⇠
0
1
⇠
6
7
4µ0
2µ0
d⇧ dz =
[4 ln 6 + ln(7/6)] = 5.9 µWb
2⌅⇧
⌅
129
7.35. A current sheet, K = 20 az A/m, is located at ⇧ = 2, and a second sheet, K = 10 az A/m is
located at ⇧ = 4.
a.) Let Vm = 0 at P (⇧ = 3, ⌥ = 0, z = 5) and place a barrier at ⌥ = ⌅. Find Vm (⇧, ⌥, z) for
⌅ < ⌥ < ⌅: Since the current is cylindrically-symmetric, we know that H = I/(2⌅⇧) a⇧ ,
where I is the current enclosed, equal in this case to 2⌅(2)K = 80⌅ A. Thus, using the
result of Section 8.6, we find
Vm =
80⌅
⌥=
2⌅
I
⌥=
2⌅
40⌥ A
which is valid over the region 2 < ⇧ < 4, ⌅ < ⌥ < ⌅, and
<z<
outer current contributes, leading to a total enclosed current of
Inet = 2⌅(2)(20)
. For ⇧ > 4, the
2⌅(4)(10) = 0
With zero enclosed current, H⇧ = 0, and the magnetic potential is zero as well.
b) Let A = 0 at P and find A(⇧, ⌥, z) for 2 < ⇧ < 4: Again, we know that H = H⇧ (⇧),
since the current is cylindrically symmetric. With the current only in the z direction, and
again using symmmetry, we expect only a z component of A which varies only with ⇧.
We can then write:
dAz
µ0 I
⇤A=
a⇧ = B =
a⇧
d⇧
2⌅⇧
Thus
dAz
=
d⇧
µ0 I
2⌅⇧
Az =
µ0 I
ln(⇧) + C
2⌅
We require that Az = 0 at ⇧ = 3. Therefore C = [(µ0 I)/(2⌅)] ln(3), Then, with I = 80⌅,
we finally obtain
A=
µ0 (80⌅)
[ln(⇧)
2⌅
ln(3)] az = 40µ0 ln
3
⇧
az Wb/m
7.36. Let A = (3y z)ax + 2xzay Wb/m in a certain region of free space.
a) Show that · A = 0:
·A=
x
(3y
z) +
y
b) At P (2, 1, 3), find A, B, H, and J: First AP =
formula in cartesian coordinates,
B=
⇤A=
2xax
ay + (2z
3)az
2xz = 0
6ax + 12ay . Then, using the curl
BP =
4ax
ay + 3az Wb/m2
Now
HP = (1/µ0 )BP =
Then J =
⇤ H = (1/µ0 )
3.2 ⇤ 106 ax
8.0 ⇤ 105 ay + 2.4 ⇤ 106 az A/m
⇤ B = 0, as the curl formula in cartesian coordinates shows.
130
7.37. Let N = 1000, I = 0.8 A, ⇧0 = 2 cm, and a = 0.8 cm for the toroid shown in Fig. 8.12b. Find
Vm in the interior of the toroid if Vm = 0 at ⇧ = 2.5 cm, ⌥ = 0.3⌅. Keep ⌥ within the range
0 < ⌥ < 2⌅: Well-within the toroid, we have
H=
NI
a⇧ =
2⌅⇧
Thus
Vm =
Then,
0=
or C = 120. Finally
Vm
⌦
= 120
Vm =
1 dVm
a⇧
⇧ d⌥
N I⌥
+C
2⌅
1000(0.8)(0.3⌅)
+C
2⌅
↵
400
⌥ A (0 < ⌥ < 2⌅)
⌅
131
7.38. A square filamentary di⇤erential current loop, dL on a side, is centered at the origin in the
z = 0 plane in free space. The current I flows generally in the a⇧ direction.
a) Assuming that r >> dL, and following a method similar to that in Sec. 4.7, show that
dA =
µ0 I(dL)2 sin ⇥
a⇧
4⌅r2
We begin with the expression for the di⇤erential vector potential, Eq. (48):
dA =
µ0 IdL
4⌅R
where in our case, we have four di⇤erential elements. The net vector potential at some
distant point will consist of the vector sum of the four individual potentials. Referring to
the figures below, the net potential at the indicated point is initially constructed as:
⌦
µ0 IdL
dA =
ax
4⌅
1
R2
1
R1
+ ay
1
R3
1
R4
↵
The challenge is to determine the four distances, R1 through R4 , in terms of spherical
coordinates, r, ⇥, and ⌥, thus referencing the four potentials to a common origin. This is
where the treatment in Sec. 4.7 is useful, although it is more complicated here because
the problem is three-dimensional.
The diagram for the y-displaced elements is shown in the left-hand figure. The three
distance lines, r, R1 , and R2 , are approximately parallel because the observation point is
in the far zone. Therefore, beyond the blue line segment that crosses the three lines, the
lengths are essentially equal. The di⇤erence in lengths between r and R1 , for example,
is the length of the line segment along r, from the origin to the blue line. This length
will be the projection of the distance vector dL/2 ay along ar , or dL/2 ay · ar . As a look
ahead, this principle is discussed with illustrations in Sec. 14.5, which handles antenna
arrays of two elements.
132
7.38a) (continued). We now may write:
⌦
↵
dL
.
R1 = r
ay · ar
2
⌦
↵
dL
.
R2 = r +
ay · ar
2
and
Referring to the right-hand figure and applying similar reasoning there leads to
⌦
↵
⌦
↵
dL
dL
.
.
R3 = r
ax · ar
and
R4 = r +
ax · ar
2
2
where we know that ax · ar = sin ⇥ cos ⌥ and ay · ar = sin ⇥ sin ⌥ We now substitute all
these relations into the original expression for dA:
µ0 IdL
dA =
4⌅
dL
r+
sin ⇥ sin ⌥
2
+
r
dL
sin ⇥ cos ⌥
2
=
µ0 IdL
4⌅r
1+
1
dL
sin ⇥ cos ⌥
2r
+
1
dL
sin ⇥ sin ⌥
2r
1
1
r
dL
sin ⇥ sin ⌥
2
dL
r+
sin ⇥ cos ⌥
2
1
1
1
ay
dL
sin ⇥ sin ⌥
2r
dL
1+
sin ⇥ cos ⌥
2r
1
ay
Now, since dL/r << 1, this simplifies to
⌦
dL
dL
. µ0 IdL
dA =
1
sin ⇥ sin ⌥
1+
sin ⇥ sin ⌥
4⌅r
2r
2r
↵
dL
dL
+
1+
sin ⇥ cos ⌥
1
sin ⇥ cos ⌥
ay
2r
2r
µ0 I(dL)2 sin ⇥
=
[ sin ⌥ ax + cos ⌥ ay ]
4⌅r2
µ0 I(dL)2 sin ⇥
=
a⇧
4⌅r2
1
ax
✏
1
ax
✏
ax
b) Show that
I(dL)2
(2 cos ⇥ ar + sin ⇥ a⇥ )
4⌅r3
Using the part a expression, we construct dH = (1/µ0 ) ⇤dA, where in this case, we have
a ⌥-directed dA that varies with r and ⇥. The curl expression in spherical coordinates
reduces to:
1
(dA sin ⇥)
1 (rdA)
⇤ dA =
ar
a⇥
r sin ⇥
⇥
r
r
or
1
I(dL)2 sin2 ⇥
1
I(dL)2 sin ⇥
⇤ dA =
a
a⇥
r
r sin ⇥ ⇥
4⌅r2
r r
4⌅r
I(dL)2
=
(2 cos ⇥ ar + sin ⇥ a⇥ )
4⌅r3
dH =
133
7.39. Planar current sheets of K = 30az A/m and 30az A/m are located in free space at x = 0.2
and x = 0.2 respectively. For the region 0.2 < x < 0.2:
a) Find H: Since we have parallel current sheets carrying equal and opposite currents, we
use Eq. (12), H = K ⇤ aN , where aN is the unit normal directed into the region between
currents, and where either one of the two currents are used. Choosing the sheet at x = 0.2,
we find
H = 30az ⇤ ax = 30ay A/m
b) Obtain and expression for Vm if Vm = 0 at P (0.1, 0.2, 0.3): Use
H=
So
Then
30ay =
dVm
= 30
dy
dVm
ay
dy
Vm = 30y + C1
0 = 30(0.2) + C1
c) Find B: B = µ0 H =
Vm =
C1 =
6
Vm = 30y
6A
30µ0 ay Wb/m2 .
d) Obtain an expression for A if A = 0 at P : We expect A to be z-directed (with the
current), and so from ⇤ A = B, where B is y-directed, we set up
dAz
=
dx
Then 0 = 30µ0 (0.1) + C2
30µ0
C2 =
Az = 30µ0 x + C2
3µ0 . So finally A = µ0 (30x
3)az Wb/m.
7.40. Show that the line integral of the vector
 potential A about any closed path is equal to the
magnetic flux enclosed by the path, or A · dL = B · dS.
We use the fact that B =
⇤ A, and substitute this into the desired relation to find
⇠
A · dL =
⇤ A · dS
This is just a statement of Stokes’ theorem (already proved), so we are done.
134
7.41. Assume that A = 50⇧2 az Wb/m in a certain region of free space.
a) Find H and B: Use
B=
Then H = B/µ0 =
Az
a⇧ =
⇧
⇤A=
100⇧ a⇧ Wb/m2
100⇧/µ0 a⇧ A/m.
b) Find J: Use
J=
⇤H=
1
1
(⇧H⇧ )az =
⇧ ⇧
⇧ ⇧
100⇧2
µ0
az =
200
az A/m2
µ0
c) Use J to find the total current crossing the surface 0 ⌃ ⇧ ⌃ 1, 0 ⌃ ⌥ < 2⌅, z = 0: The
current is
⇠ ⇠
⇠ 2⇤ ⇠ 1
200
200⌅
I=
J · dS =
az · az ⇧ d⇧ d⌥ =
A = 500 MA
µ0
µ0
0
0
d) Use the value of H⇧ at ⇧ = 1 to calculate
H · dL = I =
7.42. Show that
2
2 (1/R12 )
1
R12
=
=
Also note that
=
⇠
2⇤
0
1 (1/R12 )

H · dL for ⇧ = 1, z = 0: Have
100
a⇧ · a⇧ (1)d⌥ =
µ0
200⌅
A=
µ0
500 MA
3
= R21 /R12
. First
⇤
⌅ 1/2
(x2 x1 )2 + (y2 y1 )2 + (z2 z1 )2
⌦
↵
1 2(x2 x1 )ax + 2(y2 y1 )ay + 2(z2 z1 )az
R12
R21
=
= 3
3
2
2
2
3/2
2
R12
R12
[(x2 x1 ) + (y2 y1 ) + (z2 z1 ) ]
2
1 (1/R12 )
would give the same result, but of opposite sign.
135
7.43. Compute the vector magnetic potential within the outer conductor for the coaxial line whose
vector magnetic potential is shown in Fig. 8.20 if the outer radius of the outer conductor
is 7a. Select the proper zero reference and sketch the results on the figure: We do this by
first finding B within the outer conductor and then “uncurling” the result to find A. With
z-directed current I in the outer conductor, the current density is
I
Jout =
⌅(7a)2
⌅(5a)2
az =
I
az
24⌅a2
Since current I flows in both conductors, but in opposite directions, Ampere’s circuital law
inside the outer conductor gives:
2⌅⇧H⇧ = I
⇠
0
2⇤
⇠
⌅
5a
I
⇧⇧ d⇧⇧ d⌥
24⌅a2
⌦
↵
I
49a2 ⇧2
H⇧ =
2⌅⇧
24a2
Now, with B = µ0 H, we note that ⇤A will have a ⌥ component only, and from the direction
and symmetry of the current, we expect A to be z-directed, and to vary only with ⇧. Therefore
⇤A=
and so
dAz
=
d⇧
Then by direct integration,
Az =
⇠
µ0 I(49)
d⇧ +
48⌅⇧
⇠
dAz
a⇧ = µ0 H
d⇧
⌦
↵
µ0 I 49a2 ⇧2
2⌅⇧
24a2
⌦
µ0 I⇧
µ0 I ⇧2
d⇧
+
C
=
48⌅a2
96⌅ a2
↵
98 ln ⇧ + C
As per Fig. 8.20, we establish a zero reference at ⇧ = 5a, enabling the evaluation of the
integration constant:
µ0 I
C=
[25 98 ln(5a)]
96⌅
Finally,
⌦ 2
↵
µ0 I
⇧
5a
Az =
25 + 98 ln
Wb/m
96⌅
a2
⇧
A plot of this continues the plot of Fig. 8.20, in which the curve goes negative at ⇧ = 5a, and
then approaches a minimum of .09µ0 I/⌅ at ⇧ = 7a, at which point the slope becomes zero.
136
7.44. By expanding Eq.(58), Sec. 8.7 in cartesian coordinates, show that (59) is correct. Eq. (58)
can be rewritten as
2
A = ( · A)
⇤ ⇤A
We begin with
Ax
Ay
Az
+
+
x
y
z
·A=
Then the x component of
(
· A) is
[ (
Now
Az
y
⇤A=
and the x component of
[
· A)]x =
Ay
z
⇤
⇤
2
2
2
Ax
Ay
Az
+
+
2
x
x y
x z
Ax
z
ax +
Az
x
Ay
x
ay +
Ax
y
az
⇤ A is
⇤ A]x =
2
2
Ay
x y
2
2
Ax
Az
+
2
z
z y
Ax
y2
Then, using the underlined results
[ (
· A)
⇤
⇤ A]x =
2
Ax
+
x2
2
Ax
+
y2
2
Ax
=
z2
2
Ax
Similar results will be found for the other two components, leading to
(
· A)
⇤
⇤A=
2
Ax ax +
137
2
Ay ay +
2
Az az ⇧
2
A
QED
CHAPTER 8
8.1. A point charge, Q = 0.3 µC and m = 3 ⇤ 10 16 kg, is moving through the field E = 30 az V/m.
Use Eq. (1) and Newton’s laws to develop the appropriate di⌅erential equations and solve them,
subject to the initial conditions at t = 0: v = 3 ⇤ 105 ax m/s at the origin. At t = 3 µs, find:
a) the position P (x, y, z) of the charge: The force on the charge is given by F = qE, and Newton’s
second law becomes:
F = ma = m
d2 z
= qE = ( 0.3 ⇤ 10
dt2
6
)(30 az )
describing motion of the charge in the z direction. The initial velocity in x is constant, and
so no force is applied in that direction. We integrate once:
dz
qE
= vz =
t + C1
dt
m
The initial velocity along z, vz (0) is zero, and so C1 = 0. Integrating a second time yields the
z coordinate:
qE 2
z=
t + C2
2m
The charge lies at the origin at t = 0, and so C2 = 0. Introducing the given values, we find
z=
( 0.3 ⇤ 10 6 )(30) 2
t =
2 ⇤ 3 ⇤ 10 16
1.5 ⇤ 1010 t2 m
At t = 3 µs, z = (1.5 ⇤ 1010 )(3 ⇤ 10 6 )2 = .135 cm. Now, considering the initial constant
velocity in x, the charge in 3 µs attains an x coordinate of x = vt = (3⇤105 )(3⇤10 6 ) = .90 m.
In summary, at t = 3 µs we have P (x, y, z) = (.90, 0, .135).
b) the velocity, v: After the first integration in part a, we find
vz =
qE
t=
m
(3 ⇤ 1010 )(3 ⇤ 10
6
)=
9 ⇤ 104 m/s
Including the intial x-directed velocity, we finally obtain v = 3 ⇤ 105 ax
9 ⇤ 104 az m/s.
c) the kinetic energy of the charge: Have
K.E. =
1
1
m|v|2 = (3 ⇤ 10
2
2
138
16
)(1.13 ⇤ 105 )2 = 1.5 ⇤ 10
5
J
8.2. Compare the magnitudes of the electric and magnetic forces on an electron that has attained a
velocity of 107 m/s. Assume an electric field intensity of 105 V/m, and a magnetic flux density
associated with that of the Earth’s magnetic field in temperate latitudes, 0.5 gauss. We use the
Lorentz Law, F = Fe + Fm = q(E + v ⇤ B), where |B| = 0.5 G = 5.0 ⇤ 10 5 T. We find
|Fe | = (1.6 ⇤ 10
|Fm | = (1.6 ⇤ 10
19
19
C)(105 V/m) = 1.6 ⇤ 10
C)(107 m/s)(5.0 ⇤ 10
5
14
T) = 8.0 ⇤ 10
N
17
N = 0.005|Fe |
8.3. A point charge for which Q = 2 ⇤ 10 16 C and m = 5 ⇤ 10 26 kg is moving in the combined fields
E = 100ax 200ay + 300az V/m and B = 3ax + 2ay az mT. If the charge velocity at t = 0 is
v(0) = (2ax 3ay 4az ) ⇤ 105 m/s:
a) give the unit vector showing the direction in which the charge is accelerating at t = 0: Use
F(t = 0) = q[E + (v(0) ⇤ B)], where
v(0) ⇤ B = (2ax
3ay
4az )105 ⇤ ( 3ax + 2ay
az )10
3
= 1100ax + 1400ay
500az
So the force in newtons becomes
F(0) = (2⇤10
16
)[(100+1100)ax +(1400 200)ay +(300 500)az ] = 4⇤10
14
[6ax +6ay
az ]
The unit vector that gives the acceleration direction is found from the force to be
aF =
6ax + 6ay
↵
73
az
= .70ax + .70ay
.12az
b) find the kinetic energy of the charge at t = 0:
K.E. =
1
1
m|v(0)|2 = (5 ⇤ 10
2
2
26
kg)(5.39 ⇤ 105 m/s)2 = 7.25 ⇤ 10
15
J = 7.25 fJ
8.4. Show that a charged particle in a uniform magnetic field describes a circular orbit with an orbital
period that is independent of the radius. Find the relationship between the angular velocity and
magnetic flux density for an electron (the cyclotron frequency).
A circular orbit can be established if the magnetic force on the particle is balanced by the
centripital force associated with the circular path. We assume a circular path of radius R, in
which B = B0 az is normal to the plane of the path. Then, with particle angular velocity ⇤, the
velocity is v = R⇤ a⇤ . The magnetic force is then Fm = qv⇤B = qR⇤ a⇤ ⇤B0 az = qR⇤B0 a⇥ .
This force will be negative (pulling the particle toward the center of the path) if the charge
is positive and motion is in the a⇤ direction, or if the charge is negative, and motion is in
positive a⇤ . In either case, the centripital force must counteract the magnetic force. Assuming
particle mass m, the force balance equation is qR⇤B0 = m⇤2 R, from which ⇤ = qB0 /m. The
revolution period is T = 2⌅/⇤ = 2⌅m/(qB0 ), which is independent of R. For an electron, we
have q = 1.6 ⇤ 10 9 C, and m = 9.1 ⇤ 1031 kg. The cyclotron frequency is therefore
⇤c =
q
B0 = 1.76 ⇤ 1011 B0 s
m
139
1
8.5. A rectangular loop of wire in free space joins points A(1, 0, 1) to B(3, 0, 1) to C(3, 0, 4) to D(1, 0, 4)
to A. The wire carries a current of 6 mA, flowing in the az direction from B to C. A filamentary
current of 15 A flows along the entire z axis in the az direction.
a) Find F on side BC:
✓ C
FBC =
Iloop dL ⇤ Bfrom wire at BC
B
Thus
FBC =
✓
1
4
(6 ⇤ 10
3
) dz az ⇤
15µ0
ay =
2⌅(3)
1.8 ⇤ 10
8
ax N =
18ax nN
b) Find F on side AB: The field from the long wire now varies with position along the loop
segment. We include that dependence and write
✓ 3
15µ0
45 ⇤ 10 3
FAB =
(6 ⇤ 10 3 ) dx ax ⇤
ay =
µ0 ln 3 az = 19.8az nN
2⌅x
⌅
1
c) Find Ftotal on the loop: This will be the vector sum of the forces on the four sides. Note that
by symmetry, the forces on sides AB and CD will be equal and opposite, and so will cancel.
This leaves the sum of forces on sides BC (part a) and DA, where
FDA =
✓
1
4
(6 ⇤ 10
3
) dz az ⇤
The total force is then Ftotal = FDA + FBC = (54
15µ0
ay = 54ax nN
2⌅(1)
18)ax = 36 ax nN
8.6. Show that the di⌅erential work in moving a current element IdL through a distance dl in a magnetic
field B is the negative of that done in moving the element Idl through a distance dL in the same
field: The two di⌅erential work quantities are written as:
dW = (IdL ⇤ B) · dl and dW ⇤ = (Idl ⇤ B) · dL
We now apply the vector identity, Eq.(A.6), Appendix A: (A ⇤ B) · C = (B ⇤ C) · A, and write:
(IdL ⇤ B) · dl = (B ⇤ dl) · IdL =
(Idl ⇤ B) · dL
QED
8.7. Uniform current sheets are located in free space as follows: 8az A/m at y = 0, 4az A/m at y = 1,
and 4az A/m at y = 1. Find the vector force per meter length exerted on a current filament
carrying 7 mA in the aL direction if the filament is located at:
a) x = 0, y = 0.5, and aL = az : We first note that within the region 1 < y < 1, the magnetic
fields from the two outer sheets (carrying 4az A/m) cancel, leaving only the field from the
center sheet. Therefore, H = 4ax A/m (0 < y < 1) and H = 4ax A/m ( 1 < y < 0).
Outside (y > 1 and y < 1) the fields from all three sheets cancel, leaving H = 0 (y > 1,
y < 1). So at x = 0, y = .5, the force per meter length will be
F/m = Iaz ⇤ B = (7 ⇤ 10
3
)az ⇤ 4µ0 ax =
b.) y = 0.5, z = 0, and aL = ax : F/m = Iax ⇤ 4µ0 ax = 0.
35.2ay nN/m
c) x = 0, y = 1.5, aL = az : Since y = 1.5, we are in the region in which B = 0, and so the force
is zero.
140
8.8. Two conducting strips, having infinite length in the z direction, lie in the xz plane. One occupies
the region d/2 < x < b + d/2 and carries surface current density K = K0 az ; the other is situated
at (b + d/2) < x < d/2 and carries surface current density K0 az .
a) Find the force per unit length in z that tends to separate the two strips:
We begin by evaluating the magnetic field arising from the left-hand strip (in the region x < 0)
at any location on the x axis. Because the source strip is infinite in z, this field will not depend
on z and will be valid at any location in the x-z plane. We use the Biot-Savart law and find
the field at a fixed point x0 on the x axis. The Biot-Savart law reads:
✓
K ⇤ aR
H(x0 ) =
da
2
s 4⌅R
where the integral is taken over the left strip area, and where R is the distance from point
(x, z) on the strip to the fixed observation point, x0 . Thus R = (x x0 )2 + z 2 , and
(x x0 ) ax z az
(x x0 )2 + z 2
aR =
so that
H(x0 ) =
✓
⌅
⌅
✓
d
2
K0 az ⇤ [ (x
4⌅ [(x
(d
2 +b)
x0 ) ax
x0 )2 +
z az ]
3/2
z2]
dx dz
Taking the cross product leaves only a y component:
H(x0 ) =
✓
⌅
⌅
✓
d
2
K0 ay (x
4⌅ [(x
(d
2 +b)
x0 )
3/2
x0 )2 + z 2 ]
dx dz
It is easiest to evaluate the z integral first, leading to
K0
H0 =
ay
4⌅
✓
d
2
(d
2 +b)
(x
z
x0 ) (x
x0 )2 + z 2
Evaluate the x integral to find:
K0
H0 =
ay ln(x
2⌅
⇧
⇧
x0 )⇧
d
2
(d
2 +b)
=
⇧⌅
⇧
⇧
⌅
K0
ln
2⌅
K0
dx =
ay
2⌅
d
2
+ b + x0
d
2 + x0
✏
✓
d
2
(d
2 +b)
dx
(x x0 )
ay A/m
Now the force acting on the right-hand strip per unit length is
✓
F = K ⇤ B da
s
where K is the surface current density in the right-hand strip, and B is the magnetic flux
density (µ0 H) arising from the left strip, evaluated within the right strip area (over which the
integral is taken). Over a unit lengh in z, the force integral is written:
F=
✓
0
1
✓
(b+ d
2)
d
2
k0 az ⇤
µ0 K0
ln
2⌅
141
d
2
✏
+ b + x0
ay dx0 dz
d
2 + x0
8.8a (continued)
The z integration yields a factor of 1, the cross product gives an x-directed force, and we can
rewrite the expression as:
µ0 K02
F=
ax
2⌅
µ0 K02
=
ax
2⌅
✓
⌦
(b+ d
2)
d
2
⌦
ln
d
+ b + x0
2
d
+ b + x0 ln
2
↵
d
+ x0
2
ln
d
+ b + x0
2
dx0
d
+ x0 ln
2
x0
d
+ x0
2
+ x0
↵(b+ d2 )
d
2
Evaluating this result over the integration limits and then simplifying results in the following
expression, which is one of many ways of writing the result:
2b
1+
d
µ0 dK02
F=
ax
2⌅
ln
1 + 2b
d
1 + db
b
ln 1 +
d
✏
b) let b approach zero while maintaining constant current, I = K0 b, and show that the force per
unit length approaches µ0 I 2 /(2⌅d) N/m.
As b gets small, so does the ratio b/d. We may then write:
1
1+
b
d
.
= 1
b
+
d
b
d
2
✏
The force expression now becomes:
2b
1+
d
µ0 dK02
F=
ax
2⌅
2b
1+
d
ln
b
+
d
1
b
d
2
✏
⌦
↵✏
b
ln 1 +
d
↵
⌦
↵↵
b
ln 1 +
d
The product in the natural log function is expanded:
µ0 dK02
F=
ax
2⌅
⌦
2b
1+
d
⌦
ln 1
b
b2
2b
+ 2+
d d
d
2b2
2b2
+ 2
2
d
d
All terms in the natural log functions involve 1 + f (b/d) where f (b/d) << 1. Therefore, we
.
may expand the log functions in power series, keeping only the first terms: i.e., ln(1 + f ) = f ,
if f << 1. With this simplification, the force becomes:
. µ0 dK02
F=
ax
2⌅
⌦
2b
1+
d
b
d
b2
2b3
+
d2
d3
Expanding and simplifying, the final result is
. µ0 (bK0 )2
F=
ax N/m
2⌅d
where the term, 4b4 /d4 , has been neglected.
142
b
<< 1
d
b
d
↵
8.9. A current of 100az A/m flows on the conducting cylinder ⇧ = 5 mm and +500az A/m is present
on the conducting cylinder ⇧ = 1 mm. Find the magnitude of the total force acting to split the
outer cylinder apart along its length: The di⌅erential force acting on the outer cylinder arising
from the field of the inner cylinder is dF = Kouter ⇤ B, where B is the field from the inner cylinder,
evaluated at the outer cylinder location:
2⌅(1)(500)µ0
a⇤ = 100µ0 a⇤ T
2⌅(5)
B=
Thus dF = 100az ⇤ 100µ0 a⇤ = 104 µ0 a⇥ N/m2 . We wish to find the force acting to split the outer
cylinder, which means we need to evaluate the net force in one cartesian direction on one half of
the cylinder. We choose the “upper” half (0 < ⌃ < ⌅), and integrate the y component of dF over
this range, and over a unit length in the z direction:
Fy =
✓
0
1
✓
0
10 µ0 a⇥ · ay (5 ⇤ 10
4
3
✓
) d⌃ dz =
0
50µ0 sin ⌃ d⌃ = 100µ0 = 4⌅ ⇤ 10
5
N/m
Note that we did not include the “self force” arising from the outer cylinder’s B field on itself.
Since the outer cylinder is a two-dimensional current sheet, its field exists only just outside the
cylinder, and so no force exists. If this cylinder possessed a finite thickness, then we would need
to include its self-force, since there would be an interior field and a volume current density that
would spatially overlap.
8.10. A planar transmission line consists of two conducting planes of width b separated d m in air,
carrying equal and opposite currents of I A. If b >> d, find the force of repulsion per meter of
length between the two conductors.
Take the current in the top plate in the positive z direction, and so the bottom plate current
is directed along negative z. Furthermore, the bottom plate is at y = 0, and the top plate is
at y = d. The magnetic field stength at the bottom plate arising from the current in the top
plate is H = K/2 ax A/m, where the top plate surface current density is K = I/b az A/m.
Now the force per unit length on the bottom plate is
F=
✓
0
1
✓
0
b
Kb ⇤ Bb dS
where Kb is the surface current density on the bottom plate, and Bb is the magnetic flux
density arising from the top plate current, evaluated at the bottom plate location. We obtain
F=
✓
0
1
✓
0
b
I
µ0 I
az ⇤
ax dS =
b
2b
143
µ0 I 2
ay N/m
2b
8.11. a) Use Eq. (14), Sec. 8.3, to show that the force of attraction per unit length between two
filamentary conductors in free space with currents I1 az at x = 0, y = d/2, and I2 az at x = 0,
y = d/2, is µ0 I1 I2 /(2⌅d): The force on I2 is given by
↵
⌘ ⌦⌘
I1 I2
aR12 ⇤ dL1
F2 = µ0
⇤ dL2
2
4⌅
R12
Let z1 indicate the z coordinate along I1 , and z2 indicate the z coordinate along I2 . We then have
R12 = (z2 z1 )2 + d2 and
(z2 z1 )az day
aR12 =
(z2 z1 )2 + d2
Also, dL1 = dz1 az and dL2 = dz2 az The “inside” integral becomes:
⌘
⌘
✓ ⌅
aR12 ⇤ dL1
[(z2 z1 )az day ] ⇤ dz1 az
=
=
2
R12
[(z2 z1 )2 + d2 ]1.5
⌅ [(z2
The force expression now becomes
↵
⌘ ⌦✓ ⌅
✓ ✓
I1 I2
d dz1 ax
I1 I2 1
F2 = µ0
⇤ dz2 az = µ0
4⌅
z1 )2 + d2 ]1.5
4⌅ 0
⌅ [(z2
d dz1 ax
z1 )2 + d2 ]1.5
⌅
⌅
d dz1 dz2 ay
[(z2 z1 )2 + d2 ]1.5
Note that the “outside” integral is taken over a unit length of current I2 . Evaluating, obtain,
✓ 1
I1 I2 d ay
µ0 I1 I2
F2 = µ0
(2)
dz2 =
ay N/m
2
4⌅d
2⌅d
0
as expected.
b) Show how a simpler method can be used to check your result: We use dF2 = I2 dL2 ⇤ B12 ,
where the field from current 1 at the location of current 2 is
µ0 I1
B12 =
ax T
2⌅d
so over a unit length of I2 , we obtain
µ0 I1
I1 I2
ax = µ0
ay N/m
2⌅d
2⌅d
This second method is really just the first over again, since we recognize the inside integral of
the first method as the Biot-Savart law, used to find the field from current 1 at the current 2
location.
F2 = I2 az ⇤
8.12. Two circular wire rings are parallel to each other, share the same axis, are of radius a, and are
separated by distance d, where d << a. Each ring carries current I. Find the approximate force
of attraction and indicate the relative orientations of the currents.
With the loops very close to each other, the primary force on each “segment” of current in
either loop can be assumed to arise from the local B field from the immediately adjacent
segment in the other loop. So the problem becomes essentially that of straightening out both
loops and considering them as two parallel wires of length L = 2⌅a. The magnetic induction
at points very close to either current loop is approximately that of an infinite straight wire,
.
or B = µ0 I/(2⌅d). The force is therefore found using Eq. (11) to be
F = ILB = I(2⌅a)
µ0 I
µ0 aI 2
=
2⌅d
d
If the force is attractive, then the currents must be in the same direction.
144
8.13. A current of 6A flows from M (2, 0, 5) to N (5, 0, 5) in a straight solid conductor in free space. An
infinite current filament lies along the z axis and carries 50A in the az direction. Compute the
vector torque on the wire segment using:
a) an origin at (0, 0, 5): The B field from the long wire at the short wire is B = (µ0 Iz ay )/(2⌅x) T.
Then the force acting on a di⌅erential length of the wire segment is
dF = Iw dL ⇤ B = Iw dx ax ⇤
µ0 Iz
µ0 Iw Iz
ay =
dx az N
2⌅x
2⌅x
Now the di⌅erential torque about (0, 0, 5) will be
dT = RT ⇤ dF = xax ⇤
µ0 Iw Iz
dx az =
2⌅x
µ0 Iw Iz
dx ay
2⌅
The net torque is now found by integrating the di⌅erential torque over the length of the wire
segment:
T=
✓
5
2
µ0 Iw Iz
dx ay =
2⌅
3µ0 (6)(50)
ay =
2⌅
1.8 ⇤ 10
4
ay N · m
b) an origin at (0, 0, 0): Here, the only modification is in RT , which is now RT = x ax + 5 az So
now
µ0 Iw Iz
µ0 Iw Iz
dT = RT ⇤ dF = [xax + 5az ] ⇤
dx az =
dx ay
2⌅x
2⌅
Everything from here is the same as in part a, so again, T =
c) an origin at (3, 0, 0): In this case, RT = (x
dT = [(x
3)ax + 5az ] ⇤
1.8 ⇤ 10
4
ay N · m.
3)ax + 5az , and the di⌅erential torque is
µ0 Iw Iz
dx az =
2⌅x
µ0 Iw Iz (x
2⌅x
3)
dx ay
Thus
T=
✓
2
5
µ0 Iw Iz (x
2⌅x
3)
dx ay =
6.0 ⇤ 10
145
5
⌦
3
3 ln
5
2
↵
ay =
1.5 ⇤ 10
5
ay N · m
8.14. A solenoid is 25cm long, 3cm in diameter, and carries 4 A dc in its 400 turns. Its axis is perpendicular to a uniform magnetic field of 0.8 Wb/m2 in air. Using an origin at the center of the solenoid,
calculate the torque acting on it.
First, we consider the torque, referenced to the origin, of a single wire loop of radius a in the
plane z = z0 . This is one loop of the solenoid. We will take the applied magnetic flux density
as in the ax direction and write it as B = B0 ax . The di⌅erential torque associated with a
di⌅erential current element on this loop, referenced to the origin, will be:
dT = R ⇤ dF
where R is the vector directed from the origin to the current element, and is given by R =
z0 az + aa⇥ , and where
dF = IdL ⇤ B = Ia d⌃ a⇤ ⇤ B0 ax =
So now
IaB0 cos ⌃ d⌃ az
dT = (z0 az + aa⇥ ) ⇤ ( IaB0 cos ⌃ d⌃ az ) = a2 IB0 cos ⌃ d⌃ a⇤
Using a⇤ = cos ⌃ ay
sin ⌃ ax , the di⌅erential torque becomes
dT = a2 IB0 cos ⌃ (cos ⌃ ay
sin ⌃ ax ) d⌃
Note that there is no dependence on z0 . The net torque on the loop is now
✓
✓ 2
⇥
T = dT =
a2 IB0 cos2 ⌃ ay cos ⌃ sin ⌃ ax d⌃ = ⌅a2 IB0 ay
0
But we have 400 identical turns at di⌅erent z locations (which don’t matter), so the total
torque will be just the above result times 400, or:
T = 400⌅a2 IB0 ay = 400⌅(1.5 ⇤ 10
) (4)(0.8)ay = 0.91ay N · m
2 2
8.15. A solid conducting filament extends from x = b to x = b along the line y = 2, z = 0. This
filament carries a current of 3 A in the ax direction. An infinite filament on the z axis carries 5
A in the az direction. Obtain an expression for the torque exerted on the finite conductor about
an origin located at (0, 2, 0): The di⌅erential force on the wire segment arising from the field from
the infinite wire is
dF = 3 dx ax ⇤
5µ0
a⇤ =
2⌅⇧
15µ0 cos ⌃ dx
↵
az =
2⌅ x2 + 4
15µ0 x dx
az
2⌅(x2 + 4)
So now the di⌅erential torque about the (0, 2, 0) origin is
dT = RT ⇤ dF = x ax ⇤
15µ0 x dx
15µ0 x2 dx
a
=
ay
z
2⌅(x2 + 4)
2⌅(x2 + 4)
The torque is then
✓
◆
15µ0 x2 dx
15µ0
a
=
a
x 2 tan
y
y
2
2⌅
b 2⌅(x + 4)
⌦
↵
b
6
1
= (6 ⇤ 10 ) b 2 tan
ay N · m
2
T=
b
146
1
⌃ x ⌥b
2
b
8.16. Assume that an electron is describing a circular orbit of radius a about a positively-charged nucleus.
a) By selecting an appropriate current and area, show that the equivalent orbital dipole moment
is ea2 /2, where is the electron’s angular velocity: The current magnitude will be I = Te ,
where e is the electron charge and T is the orbital period. The latter is T = 2⌅/ , and so
I = e /(2⌅). Now the dipole moment magnitude will be m = IA, where A is the loop area.
Thus
e
1
m=
⌅a2 = ea2 //
2⌅
2
b) Show that the torque produced by a magnetic field parallel to the plane of the orbit is ea2 B/2:
With B assumed constant over the loop area, we would have T = m ⇤ B. With B parallel to
the loop plane, m and B are orthogonal, and so T = mB. So, using part a, T = ea2 B/2.
c) by equating the Coulomb and centrifugal forces, show that
is the electron mass: The force balance is written as
e2
= me
4⌅ 0 a2
2
=
a ⌃
is (4⌅ 0 me a3 /e2 )
4⌅ 0 me a3
e2
1/2
, where me
1/2
//
d) Find values for the angular velocity, torque, and the orbital magnetic moment for a hydrogen
atom, where a is about 6 ⇤ 10 11 m; let B = 0.5 T: First
⌦
(1.60 ⇤ 10 19 )2
=
4⌅(8.85 ⇤ 10 12 )(9.1 ⇤ 10 31 )(6 ⇤ 10
T =
1
(3.42 ⇤ 1016 )(1.60 ⇤ 10
2
Finally,
m=
19
11 )3
)(0.5)(6 ⇤ 10
T
= 9.86 ⇤ 10
B
147
24
↵1/2
= 3.42 ⇤ 1016 rad/s
) = 4.93 ⇤ 10
11 2
A · m2
24
N·m
8.17. The hydrogen atom described in Problem 16 is now subjected to a magnetic field having the same
direction as that of the atom. Show that the forces caused by B result in a decrease of the angular
velocity by eB/(2me ) and a decrease in the orbital moment by e2 a2 B/(4me ). What are these
decreases for the hydrogen atom in parts per million for an external magnetic flux density of 0.5
T? We first write down all forces on the electron, in which we equate its coulomb force toward the
nucleus to the sum of the centrifugal force and the force associated with the applied B field. With
the field applied in the same direction as that of the atom, this would yield a Lorentz force that is
radially outward – in the same direction as the centrifugal force.
Fe = Fcent + FB
With B = 0, we solve for
⌃
e2
= me
4⌅ 0 a2
2
QvB
to find:
=
0
a + e⇡ ⇠⌫
aB⇢
e2
4⌅ 0 me a3
=
Then with B present, we find
2
e2
4⌅ 0 me a3
=
e B
=
me
2
0
e B
me
Therefore
=
But
.
=
0,
0
1
e B .
=
2
0 me
1
eB
2 0 me
0
1
e B
2 02 me
=
0
eB
//
2me
and so
.
=
0
As for the magnetic moment, we have
m = IS =
Finally, for a = 6 ⇤ 10
=
where
0
e
1
. 1
⌅a2 =
ea2 = ea2
2⌅
2
2
11
0
eB
2me
=
1
2
2
0 ea
1 e2 a2 B
//
4 me
m, B = 0.5 T, we have
eB 1 . eB 1
1.60 ⇤ 10 19 ⇤ 0.5
=
=
= 1.3 ⇤ 10
2me
2me 0
2 ⇤ 9.1 ⇤ 10 31 ⇤ 3.4 ⇤ 1016
= 3.4 ⇤ 1016 sec
1
is found from Problem 16. Finally,
m
e2 a2 B
2 . eB
=
⇤
=
m
4me
ea2
2me
148
0
= 1.3 ⇤ 10
6
6
8.18. Calculate the vector torque on the square loop shown in Fig. 8.16 about an origin at A in the field
B, given:
a) A(0, 0, 0) and B = 100ay mT: The field is uniform and so does not produce any translation
of the loop. Therefore, we may use T = IS ⇤ B about any origin, where I = 0.6 A and
S = 16az m2 . We find T = 0.6(16)az ⇤ 0.100ay = 0.96 ax N m.
b) A(0, 0, 0) and B = 200ax + 100ay mT: Using the same reasoning as in part a, we find
T = 0.6(16)az ⇤ (0.200ax + 0.100ay ) =
0.96ax + 1.92ay N m
c) A(1, 2, 3) and B = 200ax + 100ay 300az mT: We observe two things here: 1) The field is
again uniform and so again the torque is independent of the origin chosen, and 2) The field
di⌅ers from that of part b only by the addition of a z component. With S in the z direction,
this new component of B will produce no torque, so the answer is the same as part b, or
T = 0.96ax + 1.92ay N m.
d) A(1, 2, 3) and B = 200ax + 100ay 300az mT for x ⇧ 2 and B = 0 elsewhere: Now, force is
acting only on the y-directed segment at x = +2, so we need to be careful, since translation
will occur. So we must use the given origin. The di⌅erential torque acting on the di⌅erential
wire segment at location (2,y) is dT = R(y) ⇤ dF, where
dF = IdL ⇤ B = 0.6 dy ay ⇤ [0.2ax + 0.1ay
and R(y) = (2, y, 0)
(1, 2, 3) = ax + (y
2)ay
0.3az ] = [ 0.18ax
0.12az ] dy
3az . We thus find
dT = R(y) ⇤ dF = [ax + (y 2)ay 3az ] ⇤ [ 0.18ax
= [ 0.12(y 2)ax + 0.66ay + 0.18(y 2)az ] dy
0.12az ] dy
The net torque is now
T=
✓
2
[ 0.12(y
2)ax + 0.66ay + 0.18(y
2)az ] dy = 0.96ax + 2.64ay
2
8.19. Given a material for which ⌥m = 3.1 and within which B = 0.4yaz T, find:
a) H: We use B = µ0 (1 + ⌥m )H, or
H=
b) µ = (1 + 3.1)µ0 = 5.15 ⇤ 10
6
0.4yay
= 77.6yaz kA/m
(1 + 3.1)µ0
H/m.
c) µr = (1 + 3.1) = 4.1.
d) M = ⌥m H = (3.1)(77.6yay ) = 241yaz kA/m
e) J =
⇤ H = (dHz )/(dy) ax = 77.6 ax kA/m2 .
f) Jb =
⇤ M = (dMz )/(dy) ax = 241 ax kA/m2 .
g) JT =
⇤ B/µ0 = 318ax kA/m2 .
149
1.44az N m
8.20. Find H in a material where:
a) µr = 4.2, there are 2.7 ⇤ 1029 atoms/m3 , and each atom has a dipole moment of 2.6 ⇤ 10 30 ay
A · m2 . Since all dipoles are identical, we may write M = N m = (2.7 ⇤ 1029 )(2.6 ⇤ 10 30 ay ) =
0.70ay A/m. Then
M
0.70 ay
H=
=
= 0.22 ay A/m
µr 1
4.2 1
b) M = 270 az A/m and µ = 2 µH/m: Have µr = µ/µ0 = (2 ⇤ 10
H = 270az /(1.59 1) = 456 az A/m.
6
)/(4⌅ ⇤ 10
7
) = 1.59. Then
c) ⌥m = 0.7 and B = 2az T: Use
H=
B
2az
=
= 936 az kA/m
µ0 (1 + ⌥m )
(4⌅ ⇤ 10 7 )(1.7)
d) Find M in a material where bound surface current densities
of 12 az A/m and 9 az A/m
⇣
exist at ⇧ = 0.3 m and ⇧ = 0.4 m, respectively: We use M · dL = Ib , where, since currents
are in the z direction and are symmetric about the z axis, we chose the path integrals to be
circular loops centered on and normal to z. From the symmetry, M will be ⌃-directed and
will vary only with radius. Note first that for ⇧ < 0.3 m, no bound current will be enclosed
by a path integral, so we conclude that M = 0 for ⇧ < 0.3m. At radii between the currents
the path integral will enclose only the inner current so,
⌘
M · dL = 2⌅⇧M⇤ = 2⌅(0.3)12 ⌃ M =
3.6
a⇤ A/m (0.3 < ⇧ < 0.4m)
⇧
Finally, for ⇧ > 0.4 m, the total enclosed bound current is Ib,tot = 2⌅(0.3)(12) 2⌅(0.4)(9) = 0,
so therefore M = 0 (⇧ > 0.4m).
8.21. Find the magnitude of the magnetization in a material for which:
a) the magnetic flux density is 0.02 Wb/m2 and the magnetic susceptibility is 0.003 (note that
this latter quantity is missing in the original problem statement): From B = µ0 (H + M) and
from M = ⌥m H, we write
M=
B
µ0
1
+1
⌥m
1
=
B
0.02
=
= 47.7 A/m
µ0 (334)
(4⌅ ⇤ 10 7 )(334)
b) the magnetic field intensity is 1200 A/m and the relative permeability is 1.005: From B =
µ0 (H + M) = µ0 µr H, we write
M = (µr
1)H = (.005)(1200) = 6.0 A/m
c) there are 7.2 ⇤ 1028 atoms per cubic meter, each having a dipole moment of 4 ⇤ 10 30 A · m2
in the same direction, and the magnetic susceptibility is 0.0003: With all dipoles identical the
dipole moment density becomes
M = n m = (7.2 ⇤ 1028 )(4 ⇤ 10
150
30
) = 0.288 A/m
8.22. Under some conditions, it is possible to approximate the e⌅ects of ferromagnetic materials by
assuming linearity in the relationship of B and H. Let µr = 1000 for a certain material of which
a cylindrical wire of radius 1mm is made. If I = 1 A and the current distribution is uniform, find
a) B: We apply Ampere’s circuital law to a circular path of radius ⇧ around the wire axis, and
where ⇧ < a:
⌅⇧2
I⇧
I ⌃ H=
2
⌅a
2⌅a2
2
= 200⇧ a⇤ Wb/m
2⌅⇧H =
⌃B=
1000µ0 I⇧
(103 )4⌅ ⇤ 10
a
=
⇤
2⌅a2
2⌅ ⇤ 10
7
6
(1)⇧
a⇤
b) H: Using part a, H = B/µr µ0 = ⇧/(2⌅) ⇤ 106 a⇤ A/m.
c) M:
M = B/µ0
H=
(2000 2)⇧
⇤ 106 a⇤ = 1.59 ⇤ 108 ⇧ a⇤ A/m
4⌅
d) J:
J=
⇤H=
1 d(⇧H⇤ )
az = 3.18 ⇤ 105 az A/m
⇧ d⇧
Jb =
⇤M=
1 d(⇧M⇤ )
az = 3.18 ⇤ 108 az A/m2
⇧ d⇧
e) Jb within the wire:
8.23. Calculate values for H⇤ , B⇤ , and M⇤ at ⇧ = c for a coaxial cable with a = 2.5 mm and b = 6 mm
if it carries current I = 12 A in the center conductor, and µ = 3 µH/m for 2.5 < ⇧ < 3.5 mm,
µ = 5 µH/m for 3.5 < ⇧ < 4.5 mm, and µ = 10 µH/m for 4.5 < ⇧ < 6 mm. Compute for:
a) c = 3 mm: Have
I
12
H⇤ =
=
= 637 A/m
2⌅⇧
2⌅(3 ⇤ 10 3 )
Then B⇤ = µH⇤ = (3 ⇤ 10
Finally, M⇤ = (1/µ0 )B⇤
b. c = 4 mm: Have
6
)(637) = 1.91 ⇤ 10
Finally, M⇤ = (1/µ0 )B⇤
c) c = 5 mm: Have
Wb/m2 .
H⇤ = 884 A/m.
H⇤ =
Then B⇤ = µH⇤ = (5 ⇤ 10
3
6
I
12
=
2⌅⇧
2⌅(4 ⇤ 10
)(478) = 2.39 ⇤ 10
3
3)
= 478 A/m
Wb/m2 .
H⇤ = 1.42 ⇤ 103 A/m.
H⇤ =
I
12
=
2⌅⇧
2⌅(5 ⇤ 10
3)
= 382 A/m
Then B⇤ = µH⇤ = (10 ⇤ 10 6 )(382) = 3.82 ⇤ 10 3 Wb/m2 .
Finally, M⇤ = (1/µ0 )B⇤ H⇤ = 2.66 ⇤ 103 A/m.
151
8.24. Two current sheets, K0 ay A/m at z = 0 and K0 ay A/m at z = d, are separated by an inhomogeneous material for which µr = az + 1, where a is a constant.
a) Find expressions for H and B in the material: The z variation in the permeability leaves the
H field una⌅ected, and so we may find this using Ampere’s circuital law. This is done in
Chapter 7, culminating in Eq. (12) there. Applying this to the conductor in the z = 0 plane,
we find
H = K ⇤ an = K0 ay ⇤ az = K0 ax A/m
b) find the total flux that crosses a 1m2 area on the yz plane: Because the permeability varies
with z, the flux will depend on the location and dimensions of the 1m2 area. Choose a rectangle
located in the range 0 < y < y1 , and z1 < z < z2 , where we require that (z2 z1 )y1 = 1.
Therefore, y1 = 1/(z2 z1 ). The flux through this area is now
✓
✓ z2 ✓ 1/(z2 z1 )
✓ z2
µ0 K0
⇥m = µH · dS =
µ0 K0 (az + 1) ax · ax dy dz =
(az + 1) dz
(z2 z1 ) z1
s
z1
0

◆a

µ0 K0 ◆ a 2
=
(z2 z12 ) + (z2 z1 ) = µ0 K0 (z2 + z1 ) + 1 Wb/m2
(z2 z1 ) 2
2
8.25. A conducting filament at z = 0 carries 12 A in the az direction. Let µr = 1 for ⇧ < 1 cm, µr = 6
for 1 < ⇧ < 2 cm, and µr = 1 for ⇧ > 2 cm. Find
a) H everywhere: This result will depend on the current and not the materials, and is:
H=
I
1.91
a⇤ =
A/m (0 < ⇧ < ⌥)
2⌅⇧
⇧
b) B everywhere: We use B = µr µ0 H to find:
B(⇧ < 1 cm) = (1)µ0 (1.91/⇧) = (2.4 ⇤ 10 6 /⇧)a⇤ T
B(1 < ⇧ < 2 cm) = (6)µ0 (1.91/⇧) = (1.4 ⇤ 10 5 /⇧)a⇤ T
B(⇧ > 2 cm) = (1)µ0 (1.91/⇧) = (2.4 ⇤ 10 6 /⇧)a⇤ T where ⇧ is in meters.
8.26. A long solenoid has a radius of 3cm, 5,000 turns/m, and carries current I = 0.25 A. The region
0 < ⇧ < a within the solenoid has µr = 5, while µr = 1 for a < ⇧ < 3 cm. Determine a so that
a) a total flux of 10 µWb is present: First, the magnetic flux density in the coil is written in
general as B = µnI az Wb/m2 . Using b = 0.03m as the outer radius, the total flux in the coil
becomes
✓
✓
✓
✓
✓
2
⇥m =
B · dS =
s
0
⇤
= µ0 nI 5a2 + (b2
a
2
b
5µ0 nI⇧ d⇧ d⌃ +
0
0
⌅
⇤
⌅
a2 ) = µ0 nI 4a2 b2
µ0 nI ⇧ d⇧ d⌃
a
Substituting the given numbers, we have
⇥m = (4⌅ ⇤ 10
7
⇤
)(5000)(0.25) 4a2
Solve for a to find a = 2.7 cm.
⌅
0.032 = 10
5
Wb (as required)
b) Find a so that the flux is equally-divided between the regions 0 < ⇧ < a and a < ⇧ < 3 cm:
Using the expression for the flux in part a, we set
5a2 = b2
b
3
a2 ⌃ a = ↵ = ↵ = 1.22 cm
6
6
152
8.27. Let µr1 = 2 in region 1, defined by 2x+3y 4z > 1, while µr2 = 5 in region 2 where 2x+3y 4z < 1.
In region 1, H1 = 50ax 30ay + 20az A/m. Find:
a) HN 1 (normal component of H1 at the boundary): We first need a unit vector normal to the
surface, found through
aN =
(2x + 3y
| (2x + 3y
4z)
2ax + 3ay
↵
=
4z)|
29
4az
= .37ax + .56ay
.74az
Since this vector is found through the gradient, it will point in the direction of increasing
values of 2x + 3y 4z, and so will be directed into region 1. Thus we write aN = aN 21 . The
normal component of H1 will now be:
HN 1 = (H1 · aN 21 )aN 21
= [(50ax 30ay + 20az ) · (.37ax + .56ay
= 4.83ax 7.24ay + 9.66az A/m
.74az )] (.37ax + .56ay
.74az )
b) HT 1 (tangential component of H1 at the boundary):
HT 1 = H1 HN 1
= (50ax 30ay + 20az ) ( 4.83ax 7.24ay + 9.66az )
= 54.83ax 22.76ay + 10.34az A/m
c) HT 2 (tangential component of H2 at the boundary): Since tangential components of H are
continuous across a boundary between two media of di⌅erent permeabilities, we have
HT 2 = HT 1 = 54.83ax
22.76ay + 10.34az A/m
d) HN 2 (normal component of H2 at the boundary): Since normal components of B are continuous across a boundary between media of di⌅erent permeabilities, we write µ1 HN 1 = µ2 HN 2
or
HN 2 =
µr1
2
HN 1 = ( 4.83ax
µR 2
5
7.24ay + 9.66az ) =
1.93ax
2.90ay + 3.86az A/m
e) ⇥1 , the angle between H1 and aN 21 : This will be
⌦
↵
H1
50ax 30ay + 20az
cos ⇥1 =
· aN 21 =
· (.37ax + .56ay
|H1 |
(502 + 302 + 202 )1/2
Therefore ⇥1 = cos
1
.74az ) =
0.21
( .21) = 102⇥ .
f) ⇥2 , the angle between H2 and aN 21 : First,
H2 = HT 2 + HN 2 = (54.83ax 22.76ay + 10.34az ) + ( 1.93ax
= 52.90ax 25.66ay + 14.20az A/m
⌦
H2
52.90ax
cos ⇥2 =
· aN 21 =
|H2 |
Therefore ⇥2 = cos
1
( .09) = 95⇥ .
2.90ay + 3.86az )
↵
25.66ay + 14.20az
· (.37ax + .56ay
60.49
153
.74az ) =
0.09
8.28. For values of B below the knee on the magnetization curve for silicon steel, approximate the curve
by a straight line with µ = 5 mH/m. The core shown in Fig. 8.17 has areas of 1.6 cm2 and lengths
of 10 cm in each outer leg, and an area of 2.5 cm2 and a length of 3 cm in the central leg. A coil
of 1200 turns carrying 12 mA is placed around the central leg. Find B in the:
a) center leg: We use mmf = ⇥R, where, in the central leg,
Rc =
Lin
3 ⇤ 10 2
=
µAin
(5 ⇤ 10 3 )(2.5 ⇤ 10
4)
= 2.4 ⇤ 104 H
4)
= 1.25 ⇤ 105 H
In each outer leg, the reluctance is
Ro =
Lout
10 ⇤ 10 2
=
µAout
(5 ⇤ 10 3 )(1.6 ⇤ 10
The magnetic circuit is formed by the center leg in series with the parallel combination of the
two outer legs. The total reluctance seen at the coil location is RT = Rc +(1/2)Ro = 8.65⇤104
H. We now have
mmf
14.4
⇥=
=
= 1.66 ⇤ 10 4 Wb
RT
8.65 ⇤ 104
The flux density in the center leg is now
B=
⇥
1.66 ⇤ 10
=
A
2.5 ⇤ 10
4
4
= 0.666 T
b) center leg, if a 0.3-mm air gap is present in the center leg: The air gap reluctance adds to the
total reluctance already calculated, where
Rair =
0.3 ⇤ 10 3
(4⌅ ⇤ 10 7 )(2.5 ⇤ 10
4)
= 9.55 ⇤ 105 H
Now the total reluctance is Rnet = RT + Rair = 8.56 ⇤ 104 + 9.55 ⇤ 105 = 1.04 ⇤ 106 . The
flux in the center leg is now
⇥=
14.4
= 1.38 ⇤ 10
1.04 ⇤ 106
and
B=
1.38 ⇤ 10
2.5 ⇤ 10
154
5
4
5
Wb
= 55.3 mT
8.29. In Problem 8.28, the linear approximation suggested in the statement of the problem leads to a flux
density of 0.666 T in the center leg. Using this value of B and the magnetization curve for silicon
.
steel, what current is required in the 1200-turn coil? With B = 0.666 T, we read Hin = 120 A · t/m
in Fig. 8.11. The flux in the center leg is ⇥ = 0.666(2.5 ⇤ 10 4 ) = 1.66 ⇤ 10 4 Wb. This divides
equally in the two outer legs, so that the flux density in each outer leg is
Bout =
1
2
1.66 ⇤ 10
1.6 ⇤ 10
4
4
= 0.52 Wb/m2
.
Using Fig. 8.11 with this result, we find Hout = 90 A · t/m We now use
⌘
H · dL = N I
to find
I=
1
(120)(3 ⇤ 10
(Hin Lin + Hout Lout ) =
N
155
2
) + (90)(10 ⇤ 10
1200
2
)
= 10.5 mA
8.30. A rectangular core has fixed permeability µr >> 1, a square cross-section of dimensions a ⇤ a, and
has centerline dimensions around its perimeter of b and d. Coils 1 and 2, having turn numbers N1
and N2 , are wound on the core. Consider a selected core cross-sectional plane as lying within the
xy plane, such that the surface is defined by 0 < x < a, 0 < y < a.
a) With current I1 in coil 1, use Ampere’s circuital law to find the magnetic flux density as
a function of position over the core cross-section: Along the midline of the core (at which
x = d/2), the path integral for H in Ampere’s law becomes
⌘
H · dL = (2b + 2d)H
At all other points in the core interior, but o⌅ the midline, the path integral becomes
⌘
H · dL = [2(d + a 2x) + 2(b + a 2x)] H = Iencl = N1 I1
The flux density magnitudes are therefore
B11 = B12 = µH =
µr µ0 N1 I1
2(d + b + 2a 4x)
in which we are assuming no y variation.
b) Integrate your result of part a to determine the total magnetic flux within the core: This will
be the integral of B over the core cross-section:
✓
✓
a
✓
a
µr µ0 N1 I1
dx dy =
4x)
s
0
0 2(d + b + 2a
⌦
↵
1
d + b + 2a
= µr µ0 N1 I1 a ln
Wb
8
d + b 2a
⇥m =
B · dS =
1
µr µ0 N1 I1 a ln [d + b + 2a
8
c) Find the self-inductance of coil 1:
L11
⌦
↵
N1 B11
1
d + b + 2a
2
=
= µr µ0 N1 a ln
H
I1
8
d + b 2a
d) find the mutual inductance between coils 1 and 2.
M12
⌦
↵
N2 B12
1
d + b + 2a
=M =
= µr µ0 N1 N2 a ln
H
I1
8
d + b 2a
156
⇧a
⇧
4x] ⇧
0
8.31. A toroid is constructed of a magnetic material having a cross-sectional area of 2.5 cm2 and an
e⌅ective length of 8 cm. There is also a short air gap 0.25 mm length and an e⌅ective area of 2.8
cm2 . An mmf of 200 A · t is applied to the magnetic circuit. Calculate the total flux in the toroid
if:
a) the magnetic material is assumed to have infinite permeability: In this case the core reluctance,
Rc = l/(µA), is zero, leaving only the gap reluctance. This is
Rg =
d
0.25 ⇤ 10 3
=
µ0 Ag
(4⌅ ⇤ 10 7 )(2.5 ⇤ 10
Now
⇥=
4)
= 7.1 ⇤ 105 H
mmf
200
=
= 2.8 ⇤ 10
Rg
7.1 ⇤ 105
4
Wb
b) the magnetic material is assumed to be linear with µr = 1000: Now the core reluctance is no
longer zero, but
8 ⇤ 10 2
Rc =
= 2.6 ⇤ 105 H
(1000)(4⌅ ⇤ 10 7 )(2.5 ⇤ 10 4 )
The flux is then
⇥=
mmf
200
=
= 2.1 ⇤ 10
Rc + Rg
9.7 ⇤ 105
4
Wb
c) the magnetic material is silicon steel: In this case we use the magnetization curve, Fig. 8.11,
and employ an iterative process to arrive at the final answer. We can begin with the value of
⇥ found in part a, assuming infinite permeability: ⇥(1) = 2.8 ⇤ 10 4 Wb. The flux density
(1)
in the core is then Bc = (2.8 ⇤ 10 4 )/(2.5 ⇤ 10 4 ) = 1.1 Wb/m2 . From Fig. 8.11, this
(1) .
corresponds to magnetic field strength Hc = 270 A/m. We check this by applying Ampere’s
circuital law to the magnetic circuit:
⌘
H · dL = Hc(1) Lc + Hg(1) d
(1)
where Hc Lc = (270)(8 ⇤ 10
199. But we require that
2
(1)
) = 22, and where Hg d = ⇥(1) Rg = (2.8 ⇤ 10
⌘
H · dL = 200 A · t
4
)(7.1 ⇤ 105 ) =
whereas the actual result in this first calculation is 199 + 22 = 221, which is too high. So, for
(2)
(2)
a second trial, we reduce B to Bc = 1 Wb/m2 . This yields Hc = 200 A/m from Fig. 8.11,
and thus ⇥(2) = 2.5 ⇤ 10 4 Wb. Now
⌘
H · dL = Hc(2) Lc + ⇥(2) Rg = 200(8 ⇤ 10 2 ) + (2.5 ⇤ 10 4 )(7.1 ⇤ 105 ) = 194
This is less than 200, meaning that the actual flux is slightly higher than 2.5 ⇤ 10
I will leave the answer at that, considering the lack of fine resolution in Fig. 8.11.
157
4
Wb.
8.32. a) Find an expression for the magnetic energy stored per unit length in a coaxial transmission line
consisting of conducting sleeves of negligible thickness, having radii a and b. A medium of relative
permeability µr fills the region between conductors. Assume current I flows in both conductors,
in opposite directions.
Within the coax, the magnetic field is H = I/(2⌅⇧) a⇤ . The energy density is then
wm =
1
µr µ0 I 2
B·H=
J/m3
2
8⌅ 2 ⇧2
The energy per unit length in z is therefore
Wm =
✓
wm dv =
v
✓
0
1
✓
2
0
✓
a
b
µr µ0 I 2
µr µ0 I 2
⇧
d⇧
d⌃
dz
=
ln
8⌅ 2 ⇧2
4⌅
b
a
J/m
b) Obtain the inductance, L, per unit length of line by equating the energy to (1/2)LI 2 .
L=
2Wm
µr µ0
=
ln
2
I
2⌅
b
a
H/m
8.33. A toroidal core has a square cross section, 2.5 cm < ⇧ < 3.5 cm, 0.5 cm < z < 0.5 cm. The
upper half of the toroid, 0 < z < 0.5 cm, is constructed of a linear material for which µr = 10,
while the lower half, 0.5 cm < z < 0, has µr = 20. An mmf of 150 A · t establishes a flux in the
a⇤ direction. For z > 0, find:
a) H⇤ (⇧): Ampere’s circuital law gives:
2⌅⇧H⇤ = N I = 150 ⌃ H⇤ =
b) B⇤ (⇧): We use B⇤ = µr µ0 H⇤ = (10)(4⌅ ⇤ 10
c) ⇥z>0 : This will be
⇥z>0 =
✓ ✓
B · dS =
= 5.0 ⇤ 10
7
✓
Wb
0
.005
✓
.035
.025
7
150
= 23.9/⇧ A/m
2⌅⇧
)(23.9/⇧) = 3.0 ⇤ 10
3.0 ⇤ 10
⇧
4
4
/⇧ Wb/m2 .
d⇧dz = (.005)(3.0 ⇤ 10
4
) ln
.035
.025
d) Repeat for z < 0: First, the magnetic field strength will be the same as in part a, since the
calculation is material-independent. Thus H⇤ = 23.9/⇧ A/m. Next, B⇤ is modified only by the
new permeability, which is twice the value used in part a: Thus B⇤ = 6.0 ⇤ 10 4 /⇧ Wb/m2 .
Finally, since B⇤ is twice that of part a, the flux will be increased by the same factor, since
the area of integration for z < 0 is the same. Thus ⇥z<0 = 1.0 ⇤ 10 6 Wb.
e) Find ⇥total : This will be the sum of the values found for z < 0 and z > 0, or ⇥total =
1.5 ⇤ 10 6 Wb.
158
8.34. Determine the energy stored per unit length in the internal magnetic field of an infinitely-long
straight wire of radius a, carrying uniform current I.
We begin with H = I⇧/(2⌅a2 ) a⇤ , and find the integral of the energy density over the unit
length in z:
✓
✓ 1✓ 2 ✓ a
1
µ0 ⇧2 I 2
µ0 I 2
We =
µ0 H 2 dv =
⇧
d⇧
d⌃
dz
=
J/m
2 4
16⌅
vol 2
0
0
0 8⌅ a
8.35. The cones ⇥ = 21⇥ and ⇥ = 159⇥ are conducting surfaces and carry total currents of 40 A, as shown
in Fig. 8.18. The currents return on a spherical conducting surface of 0.25 m radius.
a) Find H in the region 0 < r < 0.25, 21⇥ < ⇥ < 159⇥ , 0 < ⌃ < 2⌅: We can apply Ampere’s
circuital law and take advantage of symmetry. We expect to see H in the a⇤ direction and it
would be constant at a given distance from the z axis. We thus perform the line integral of
H over a circle, centered on the z axis, and parallel to the xy plane:
⌘
✓ 2
H · dL =
H⇤ a⇤ · r sin ⇥a⇤ d⌃ = Iencl. = 40 A
0
Assuming that H⇤ is constant over the integration path, we take it outside the integral and
solve:
40
20
H⇤ =
⌃ H=
a⇤ A/m
2⌅r sin ⇥
⌅r sin ⇥
b) How much energy is stored in this region? This will be
✓
✓ 2 ✓ 159 ✓ .25
✓
1
200µ0
100µ0 159 d⇥
2
2
WH =
µ0 H⇤ =
r sin ⇥ dr d⇥ d⌃ =
⌅
sin ⇥
⌅ 2 r2 sin2 ⇥
v 2
0
21
0
21
⌦
↵
100µ0
tan(159/2)
=
ln
= 1.35 ⇤ 10 4 J
⌅
tan(21/2)
8.36. The dimensions of the outer conductor of a coaxial cable are b and c, where c > b. Assuming µ = µ0 ,
find the magnetic energy stored per unit length in the region b < ⇧ < c for a uniformly-distributed
total current I flowing in opposite directions in the inner and outer conductors.
We first need to find the magnetic field inside the outer conductor volume. Ampere’s circuital
law is applied to a circular path of radius ⇧, where b < ⇧ < c. This encloses the entire center
conductor current (assumed in the positive z direction), plus that part of the z-directed
outer conductor current that lies inside ⇧. We obtain:
⌦ 2
↵
⌦ 2
↵
⇧
b2
c
⇧2
2⌅⇧H = I I 2
=I 2
c
b2
c
b2
So that
⌦ 2
I
c
H=
2⌅⇧ c2
⇧2
b2
↵
a⇤ A/m (b < ⇧ < c)
The energy within the outer conductor is now
⌦ 2
↵
✓
✓ 1✓ 2 ✓ c
1
µ0 I 2
c
2
2
Wm =
µ0 H 2 dv =
2c
+
⇧
⇧ d⇧ d⌃, dz
2 2
b2 )2 ⇧2
vol 2
0
0
b 8⌅ (c
⌦
↵
µ0 I 2
1
2 2
4 4
=
ln(c/b) (1 b /c ) + (1 b /c ) J
4⌅(1 b2 /c2 )2
4
159
8.37. Find the inductance of the cone-sphere configuration described in Problem 8.35 and Fig. 8.18.
The inductance is that o⌅ered at the origin between the vertices of the cone: From Problem 8.35,
the magnetic flux density is B⇤ = 20µ0 /(⌅r sin ⇥). We integrate this over the crossectional area
defined by 0 < r < 0.25 and 21⇥ < ⇥ < 159⇥ , to find the total flux:
⇥=
✓
159
21
✓
0.25
0
⌦
↵
20µ0
5µ0
tan(159/2)
5µ0
r dr d⇥ =
ln
=
(3.37) = 6.74 ⇤ 10
⌅r sin ⇥
⌅
tan(21/2)
⌅
6
Wb
Now L = ⇥/I = 6.74 ⇤ 10 6 /40 = 0.17 µH.
Second method: Use the energy computation of Problem 8.35, and write
2WH
2(1.35 ⇤ 10
=
2
I
(40)2
L=
4
)
= 0.17 µH
8.38. A toroidal core has a rectangular cross section defined by the surfaces ⇧ = 2 cm, ⇧ = 3 cm, z = 4
cm, and z = 4.5 cm. The core material has a relative permeability of 80. If the core is wound with
a coil containing 8000 turns of wire, find its inductance: First we apply Ampere’s circuital law to
a circular loop of radius ⇧ in the interior of the toroid, and in the a⇤ direction.
⌘
H · dL = 2⌅⇧H⇤ = N I
⌃ H⇤ =
NI
2⌅⇧
The flux in the toroid is then the integral over the cross section of B:
⇥=
✓ ✓
B · dL =
✓
.045
.04
✓
.03
.02
µr µ0 N I
µr µ0 N I
d⇧ dz = (.005)
ln
2⌅⇧
2⌅
The flux linkage is then given by N ⇥, and the inductance is
L=
N⇥
(.005)(80)(4⌅ ⇤ 10
=
I
2⌅
160
7
)(8000)2
ln(1.5) = 2.08 H
.03
.02
8.39. Conducting planes in air at z = 0 and z = d carry surface currents of ±K0 ax A/m.
a) Find the energy stored in the magnetic field per unit length (0 < x < 1) in a width w (0 <
y < w): First, assuming current flows in the +ax direction in the sheet at z = d, and in ax
in the sheet at z = 0, we find that both currents together yield H = K0 ay for 0 < z < d and
zero elsewhere. The stored energy within the specified volume will be:
WH =
✓
1
µ0 H 2 dv =
2
v
✓
d
0
✓
w
0
✓
1
0
1
1
µ0 K02 dx dy dz = wdµ0 K02 J/m
2
2
b) Calculate the inductance per unit length of this transmission line from WH = (1/2)LI 2 , where
I is the total current in a width w in either conductor: We have I = wK0 , and so
L=
2 wd
2 dw
µ0 d
µ0 K02 = 2 2
µ0 K02 =
H/m
2
I 2
w K0 2
w
c) Calculate the total flux passing through the rectangle 0 < x < 1, 0 < z < d, in the plane
y = 0, and from this result again find the inductance per unit length:
⇥=
✓
d
0
✓
0
1
µ0 Hay · ay dx dz =
Then
L=
✓
d
0
✓
1
µ0 K0 dx dy = µ0 dK0
0
⇥
µ0 dK0
µ0 d
=
=
H/m
I
wK0
w
8.40. A coaxial cable has conductor radii a and b, where a < b. Material of permeability µr = 1 exists in
the region a < ⇧ < c, while the region c < ⇧ < b is air-filled. Find an expression for the inductance
per unit length.
In both regions, the magnetic field will be H = I/(2⌅⇧) a⇤ A/m. So the flux per unit length
between conductors will be the sum of the fluxes in both regions. We integrate over a plane
surface of constant ⌃, unit length in z, and between radii a and b:
⇥m =
✓
s
B·dS =
✓
0
1
✓
a
c
µr µ0 I
d⇧ dz+
2⌅⇧
✓
1
0
✓
c
b
⌦
⌃c⌥
µ0 I
µ0 I
d⇧ dz =
µr ln
+ ln
2⌅⇧
2⌅
a
The inductance per unit length is then L = ⇥m /I (with one turn), or
L=
⌦
⌃c⌥
µ0
µr ln
+ ln
2⌅
a
161
b
c
↵
H/m
b
c
↵
Wb/m
8.41. A rectangular coil is composed of 150 turns of a filamentary conductor. Find the mutual inductance
in free space between this coil and an infinite straight filament on the z axis if the four corners of
the coil are located at
a) (0,1,0), (0,3,0), (0,3,1), and (0,1,1): In this case the coil lies in the yz plane. If we assume that
the filament current is in the +az direction, then the B field from the filament penetrates the
coil in the ax direction (normal to the loop plane). The flux through the loop will thus be
⇥=
✓
1
0
✓
3
1
µ0 I
µ0 I
ax · ( ax ) dy dz =
ln 3
2⌅y
2⌅
The mutual inductance is then
M=
N⇥
150µ0
=
ln 3 = 33 µH
I
2⌅
b) (1,1,0), (1,3,0), (1,3,1), and (1,1,1): Now the coil lies in the x = 1 plane, and the field from
the filament penetrates in a direction that is not normal to the plane of the coil. We write the
B field from the filament at the coil location as
B=
µ0 Ia⇤
2⌅
y2 + 1
The flux through the coil is now
⇥=
✓
1
0
=
✓
0
✓
3
2⌅
1
1
✓
1
µ0 Ia⇤
3
y2
+1
· ( ax ) dy dz =
✓
1
0
✓
3
1
µ0 I sin ⌃
2⌅
y2 + 1
dy dz
⇧3
µ0 Iy
µ0 I
⇧
2
dy
dz
=
ln(y
+
1)
⇧ = (1.6 ⇤ 10
2⌅(y 2 + 1)
2⌅
1
7
)I
The mutual inductance is then
M=
N⇥
= (150)(1.6 ⇤ 10
I
7
) = 24 µH
8.42. Find the mutual inductance between two filaments forming circular rings of radii a and a, where
a << a. The field should be determined by approximate methods. The rings are coplanar and
concentric.
We use the result of Problem 8.4, which asks for the magnetic field at the origin, arising from
a circular current loop of radius a. That solution is reproduced below: Using the Biot-Savart
law, we have IdL = Iad⌅ a⇤ , R = a, and aR = a⇥ . The field at the center of the circle is
then
✓ 2
✓ 2
Iad⌃ a⇤ ⇤ ( a⇥ )
Id⌃ az
I
Hcirc =
=
=
az A/m
2
4⌅a
4⌅a
2a
0
0
We now approximate that field as constant over a circular area of radius
flux linkage (for the single turn) as
µ0 I⌅( a)2
.
⇥m = ⌅( a)2 Bouter =
2a
162
⌃ M=
a, and write the
⇥m
µ0 ⌅( a)2
=
I
2a
8.43. a) Use energy relationships to show that the internal inductance of a nonmagnetic cylindrical wire
of radius a carrying a uniformly-distributed current I is µ0 /(8⌅) H/m. We first find the magnetic
field inside the conductor, then calculate the energy stored there. From Ampere’s circuital law:
⌅⇧2
I
⌅a2
2⌅⇧H⇤ =
Now
WH =
✓
v
1
µ0 H⇤2 dv =
2
✓
1
0
✓
⌃ H⇤ =
2
0
✓
a
0
I⇧
A/m
2⌅a2
µ0 I 2 ⇧2
µ0 I 2
⇧
d⇧
d⌃
dz
=
J/m
8⌅ 2 a4
16⌅
Now, with WH = (1/2)LI , we find Lint = µ0 /(8⌅) as expected.
2
b) Find the internal inductance if the portion of the conductor for which ⇧ < c < a is removed: The
hollowed-out conductor still carries current I, so Ampere’s circuital law now reads:
⌦ 2
↵
⌅(⇧2 c2 )
I
⇧
c2
2⌅⇧H⇤ =
⌃ H⇤ =
A/m
⌅(a2 c2 )
2⌅⇧ a2 c2
and the energy is now
✓ 1✓ 2 ✓
WH =
✓ a⌦
µ0 I 2 (⇧2 c2 )2
µ0 I 2
⇧ d⇧ d⌃ dz =
⇧3
2 ⇧2 (a2
2 )2
2
2 )2
8⌅
c
4⌅(a
c
0
0
c
c
⌦
⌃ a ⌥↵
µ0 I 2
1 4
=
(a
c4 ) c2 (a2 c2 ) + c4 ln
J/m
4⌅(a2 c2 )2 4
c
a
C4
2c ⇧ +
⇧
2
↵
d⇧
The internal inductance is then
Lint
⌦
2WH
µ0 a4
=
=
I2
8⌅
↵
4a2 c2 + 3c4 + 4c4 ln(a/c)
H/m
(a2 c2 )2
8.44. Show that the external inductance per unit length of a two-wire transmission line carrying equal
and opposite currents is approximately (µ/⌅) ln(d/a) H/m, where a is the radius of each wire and
d is the center-to-center wire spacing. On what basis is the approximation valid?
Suppose that one line is positioned along the z axis, with the other in the x-z plane at
x = d. With equal and opposite currents, I, and with the z axis wire current in the positive
z direction, the magnetic flux density in the x-z plane (arising from both currents) will be
µI
ay Wb/m2 (a < x < d
⌅x
B(x) =
a)
The flux per unit length in z between conductors is now:
⇥m =
✓
s
B · dS =
✓
0
1
✓
(d a)
a
µI
µI
ay · ay dx dz =
ln
⌅x
⌅
Now, the external inductance will be L = ⇥m /I, which becomes
. µ
L = ln
⌅
under the assumption that a << d.
163
d
a
H/m
d
a
a
CHAPTER 9
9.1. In Fig. 9.4, let B = 0.2 cos 120⌅t T, and assume that the conductor joining the two ends of the
resistor is perfect. It may be assumed that the magnetic field produced by I(t) is negligible.
Find:
a) Vab (t): Since B is constant over the loop area, the flux is ⇥ = ⌅(0.15)2 B = 1.41 ⇤
10 2 cos 120⌅t Wb. Now, emf = Vba (t) = d⇥/dt = (120⌅)(1.41 ⇤ 10 2 ) sin 120⌅t.
Then Vab (t) = Vba (t) = 5.33 sin 120⌅t V.
b) I(t) = Vba (t)/R = 5.33 sin(120⌅t)/250 = 21.3 sin(120⌅t) mA
9.2. In the example described by Fig. 9.1, replace the constant magnetic flux density by the timevarying quantity B = B0 sin t az . Assume that v is constant and that the displacement y of
the bar is zero at t = 0. Find the emf at any time, t.
The magnetic flux through the loop area is
⇥m =
s
B · dS =
vt
0
d
0
B0 sin t (az · az ) dx dy = B0 v t d sin t
Then the emf is
emf =
↵
E · dL =
d⇥m
=
dt
B0 d v [sin t +
t cos t] V
9.3. Given H = 300 az cos(3 ⇤ 108 t y) A/m in free space, find the emf developed in the general
a⌅ direction about the closed path having corners at
a) (0,0,0), (1,0,0), (1,1,0), and (0,1,0): The magnetic flux will be:
1
⇥=
0
Then
1
0
300µ0 cos(3 ⇤ 108 t
⇤
= 300µ0 sin(3 ⇤ 108 t
emf =
=
1)
y) dx dy = 300µ0 sin(3 ⇤ 108 t
⌅
sin(3 ⇤ 108 t) Wb
⇤
d⇥
= 300(3 ⇤ 108 )(4⌅ ⇤ 10 7 ) cos(3 ⇤ 108 t 1)
dt
⇤
⌅
1.13 ⇤ 105 cos(3 ⇤ 108 t 1) cos(3 ⇤ 108 t) V
⌅
cos(3 ⇤ 108 t)
b) corners at (0,0,0), (2⌅,0,0), (2⌅,2⌅,0), (0,2⌅,0): In this case, the flux is
⇥ = 2⌅ ⇤ 300µ0 sin(3 ⇤ 108 t
The emf is therefore 0.
164
y)|20 = 0
y)|10
9.4. A rectangular loop of wire containing a high-resistance voltmeter has corners initially at
(a/2, b/2, 0), ( a/2, b/2, 0), ( a/2, b/2, 0), and (a/2, b/2, 0). The loop begins to rotate
about the x axis at constant angular velocity , with the first-named corner moving in the
az direction at t = 0. Assume a uniform magnetic flux density B = B0 az . Determine the
induced emf in the rotating loop and specify the direction of the current.
The magnetic flux though the loop is found (as usual) through
⇥m =
s
B · dS, where S = n da
Because the loop is rotating, the direction of the normal, n, changing, and is in this case
given by
n = cos t az sin t ay
Therefore,
⇥m =
b/2
a/2
b/2
a/2
B0 az · (cos t az
sin t ay ) dx dy = abB0 cos t
The integral is taken over the entire loop area (regardless of its immediate orientation).
The important result is that the component of B that is normal to the loop area is varying
sinusoidally, and so it is fine to think of the B field itself rotating about the x axis in the
opposite direction while the loop is stationary. Now the emf is
emf =
↵
E · dL =
d⇥m
= ab B0 sin t V
dt
The direction of the current is the same as the direction of E in the emf expression.
It is easiest to picture this by considering the B field rotating and the loop fixed. By
convention, dL will be counter-clockwise when looking down on the loop from the upper
half-space (in the opposite direction of the normal vector to the plane). The current will
be counter-clockwise whenever the emf is positive, and will be clockwise whenever the
emf is negative.
9.5. The location of the sliding bar in Fig. 9.5 is given by x = 5t + 2t3 , and the separation of the
two rails is 20 cm. Let B = 0.8x2 az T. Find the voltmeter reading at:
a) t = 0.4 s: The flux through the loop will be
0.2
⇥=
0
x
0.8(x⇤ )2 dx⇤ dy =
0
0.16 3 0.16
x =
(5t + 2t3 )3 Wb
3
3
Then
emf =
d⇥
0.16
=
(3)(5t + 2t3 )2 (5 + 6t2 ) =
dt
3
(0.16)[5(.4) + 2(.4)3 ]2 [5 + 6(.4)2 ] =
b) x = 0.6 m: Have 0.6 = 5t + 2t3 , from which we find t = 0.1193. Thus
emf =
(0.16)[5(.1193) + 2(.1193)3 ]2 [5 + 6(.1193)2 ] =
165
.293 V
4.32 V
9.6. Let the wire loop of Problem 9.4 be stationary in its t = 0 position and find the induced emf
that results from a magnetic flux density given by B(y, t) = B0 cos( t
y) az , where and
are constants.
We begin by finding the net magnetic flux through the loop:
⇥m =
s
=
Now the emf is
emf =
↵
B · dS =
B0 a
b/2
a/2
b/2
a/2
[sin( t + b/2)
E · dL =
d⇥m
=
dt
B0 a
B0 cos( t
sin( t
y) az · az dx dy
b/2)]
[cos( t + b/2)
cos( t
b/2)]
Using the trig identity, cos(a ± b) = cos a cos b ⇧ sin a sin b, we may write the above result
as
emf = +2B0 a sin( t) sin( b/2) V
9.7. The rails in Fig. 9.7 each have a resistance of 2.2 ⇤/m. The bar moves to the right at a
constant speed of 9 m/s in a uniform magnetic field of 0.8 T. Find I(t), 0 < t < 1 s, if the bar
is at x = 2 m at t = 0 and
a) a 0.3 ⇤ resistor is present across the left end with the right end open-circuited: The flux
in the left-hand closed loop is
⇥l = B ⇤ area = (0.8)(0.2)(2 + 9t)
Then, emf l = d⇥l /dt = (0.16)(9) = 1.44 V. With the bar in motion, the loop
resistance is increasing with time, and is given by Rl (t) = 0.3 + 2[2.2(2 + 9t)]. The current
is now
emf l
1.44
Il (t) =
=
A
Rl (t)
9.1 + 39.6t
Note that the sign of the current indicates that it is flowing in the direction opposite that
shown in the figure.
b) Repeat part a, but with a resistor of 0.3 ⇤ across each end: In this case, there will be
a contribution to the current from the right loop, which is now closed. The flux in the
right loop, whose area decreases with time, is
⇥r = (0.8)(0.2)[(16
2)
9t]
and emf r = d⇥r /dt = (0.16)(9) = 1.44 V. The resistance of the right loop is Rr (t) =
0.3 + 2[2.2(14 9t)], and so the contribution to the current from the right loop will be
Ir (t) =
1.44
A
61.9 39.6t
The minus sign has been inserted because again the current must flow in the opposite
direction as that indicated in the figure, with the flux decreasing with time. The total
current is found by adding the part a result, or
⌦
1
1
IT (t) = 1.44
+
A
61.9 39.6t 9.1 + 39.6t
166
9.8. A perfectly-conducting filament is formed into a circular ring of radius a. At one point a
resistance R is inserted into the circuit, and at another a battery of voltage V0 is inserted.
Assume that the loop current itself produces negligible magnetic field.
a) Apply Faraday’s law, Eq. (4), evaluating each side of the equation carefully and independently to show the equality: With no B field present, and no time variation, the
right-hand side of Faraday’s law is zero, and so therefore
↵
E · dL = 0
This is just a statement of Kircho⌅’s voltage law around the loop, stating that the battery
voltage is equal and opposite to the resistor voltage.
b) Repeat part a, assuming the battery removed, the ring closed again, and a linearlyincreasing B field applied in a direction normal to the loop surface: The situation now
becomes the same as that shown in Fig. 9.4, except the loop radius is now a, and the
resistor value is not specified. Consider the loop as in the x-y plane with the positive
z axis directed out of the page. The a⌅ direction is thus counter-clockwise around the
loop. The B field (out of the page as shown) can be written as B(t) = B0 t az . With the
normal to the loop specified as az , the direction of dL is, by the right hand convention,
a⌅ . Since the wire is perfectly-conducting, the only voltage appears across the resistor,
and is given as VR . Faraday’s law becomes
↵
d⇥m
d
E · dL = VR =
=
B0 t az · az da = ⌅a2 B0
dt
dt s
This indicates that the resistor voltage, VR = ⌅a2 B0 , has polarity such that the positive
terminal is at point a in the figure, while the negative terminal is at point b. Current
flows in the clockwise direction, and is given in magnitude by I = ⌅a2 B0 /R.
9.9. A square filamentary loop of wire is 25 cm on a side and has a resistance of 125 ⇤ per meter
length. The loop lies in the z = 0 plane with its corners at (0, 0, 0), (0.25, 0, 0), (0.25, 0.25, 0),
and (0, 0.25, 0) at t = 0. The loop is moving with velocity vy = 50 m/s in the field Bz =
8 cos(1.5 ⇤ 108 t 0.5x) µT. Develop a function of time which expresses the ohmic power being
delivered to the loop: First, since the field does not vary with y, the loop motion in the y
direction does not produce any time-varying flux, and so this motion is immaterial. We can
evaluate the flux at the original loop position to obtain:
⇥(t) =
.25
8 ⇤ 10 6 cos(1.5 ⇤ 108 t 0.5x) dx dy
⇤
⌅
(4 ⇤ 10 6 ) sin(1.5 ⇤ 108 t 0.13) sin(1.5 ⇤ 108 t) Wb
0
=
.25
0
⇤
⌅
Now, emf = V (t) = d⇥/dt = 6.0 ⇤ 102 cos(1.5 ⇤ 108 t 0.13) cos(1.5 ⇤ 108 t) , The total
loop resistance is R = 125(0.25 + 0.25 + 0.25 + 0.25) = 125 ⇤. Then the ohmic power is
P (t) =
⇤
V 2 (t)
= 2.9 ⇤ 103 cos(1.5 ⇤ 108 t
R
167
0.13)
⌅2
cos(1.5 ⇤ 108 t) Watts
9.10 a) Show that the ratio of the amplitudes of the conduction current density and the displacement
current density is ⌃/ ⇥ for the applied field E = Em cos t. Assume µ = µ0 . First, D =
⇥E = ⇥Em cos t. Then the displacement current density is D/ t =
⇥Em sin t. Second,
Jc = ⌃E = ⌃Em cos t. Using these results we find |Jc |/|Jd | = ⌃/ ⇥.
b) What is the amplitude ratio if the applied field is E = Em e t/⇤ , where ⌥ is real? As before,
find D = ⇥E = ⇥Em e t/⇤ , and so Jd = D/ t = (⇥/⌥ )Em e t/⇤ . Also, Jc = ⌃Em e t/⇤ .
Finally, |Jc |/|Jd | = ⌃⌥ /⇥.
9.11. Let the internal dimension of a coaxial capacitor be a = 1.2 cm, b = 4 cm, and l = 40 cm.
The homogeneous material inside the capacitor has the parameters ⇥ = 10 11 F/m, µ = 10 5
H/m, and ⌃ = 10 5 S/m. If the electric field intensity is E = (106 /⇧) cos(105 t)a⇥ V/m, find:
a) J: Use
⌥
10
J = ⌃E =
cos(105 t)a⇥ A/m2
⇧
b) the total conduction current, Ic , through the capacitor: Have
Ic =
J · dS = 2⌅⇧lJ = 20⌅l cos(105 t) = 8⌅ cos(105 t) A
c) the total displacement current, Id , through the capacitor: First find
Jd =
Now
D
= (⇥E) =
t
t
(105 )(10 11 )(106 )
sin(105 t)a⇥ =
⇧
Id = 2⌅⇧lJd =
2⌅l sin(105 t) =
1
sin(105 t) A/m
⇧
0.8⌅ sin(105 t) A
d) the ratio of the amplitude of Id to that of Ic , the quality factor of the capacitor: This will
be
|Id |
0.8
=
= 0.1
|Ic |
8
168
9.12. Find the displacement current density associated with the magnetic field (assume zero conduction current):
H = A1 sin(4x) cos( t
z) ax + A2 cos(4x) sin( t
z) az
The displacement current density is given by
D
=
t
⇤ H = (4A2 + A1 ) sin(4x) cos( t
z) ay A/m2
9.13. Consider the region defined by |x|, |y|, and |z| < 1. Let ⇥r = 5, µr = 4, and ⌃ = 0. If
Jd = 20 cos(1.5 ⇤ 108 t bx)ay µA/m2 ;
a) find D and E: Since Jd = D/ t, we write
20 ⇤ 10 6
sin(1.5 ⇤ 108 bx)ay
1.5 ⇤ 108
= 1.33 ⇤ 10 13 sin(1.5 ⇤ 108 t bx)ay C/m2
D=
Jd dt + C =
where the integration constant is set to zero (assuming no dc fields are present). Then
E=
D
1.33 ⇤ 10 13
=
sin(1.5 ⇤ 108 t
⇥
(5 ⇤ 8.85 ⇤ 10 12 )
= 3.0 ⇤ 10
3
sin(1.5 ⇤ 108 t
bx)ay
bx)ay V/m
b) use the point form of Faraday’s law and an integration with respect to time to find B and
H: In this case,
⇤E=
Ey
az =
x
b(3.0 ⇤ 10
3
) cos(1.5 ⇤ 108 t
bx)az =
B
t
Solve for B by integrating over time:
B=
b(3.0 ⇤ 10 3 )
sin(1.5 ⇤ 108 t
1.5 ⇤ 108
Now
bx)az = (2.0)b ⇤ 10
11
sin(1.5 ⇤ 108 t
bx)az T
B
(2.0)b ⇤ 10 11
=
sin(1.5 ⇤ 108 t bx)az
µ
4 ⇤ 4⌅ ⇤ 10 7
= (4.0 ⇤ 10 6 )b sin(1.5 ⇤ 108 t bx)az A/m
H=
c) use
case
⇤ H = Jd + J to find Jd : Since ⌃ = 0, there is no conduction current, so in this
⇤H=
Hz
ay = 4.0 ⇤ 10
x
b cos(1.5 ⇤ 108 t
6 2
bx)ay A/m2 = Jd
d) What is the numerical value of b? We set the given expression for Jd equal to the result
of part c to obtain:
20 ⇤ 10
6
= 4.0 ⇤ 10
169
6 2
b
⌃ b=
5.0 m
1
9.14. A voltage source, V0 sin t, is connected between two concentric conducting spheres, r = a
and r = b, b > a, where the region between them is a material for which ⇥ = ⇥r ⇥0 , µ = µ0 , and
⌃ = 0. Find the total displacement current through the dielectric and compare it with the
source current as determined from the capacitance (Sec. 6.3) and circuit analysis methods:
First, solving Laplace’s equation, we find the voltage between spheres (see Eq. 39, Chapter
6):
(1/r) (1/b)
V (t) =
V0 sin t
(1/a) (1/b)
Then
E=
V =
V0 sin t
ar
r2 (1/a 1/b)
Now
Jd =
⌃ D=
⇥r ⇥0 V0 sin t
ar
r2 (1/a 1/b)
D
⇥r ⇥0 V0 cos t
= 2
ar
t
r (1/a 1/b)
The displacement current is then
Id = 4⌅r2 Jd =
4⌅⇥r ⇥0 V0 cos t
dV
=C
(1/a 1/b)
dt
where, from Eq. 6, Chapter 6,
C=
4⌅⇥r ⇥0
(1/a 1/b)
9.15. Let µ = 3⇤10 5 H/m, ⇥ = 1.2⇤10 10 F/m, and ⌃ = 0 everywhere. If H = 2 cos(1010 t
x)az
A/m, use Maxwell’s equations to obtain expressions for B, D, E, and : First, B = µH =
6 ⇤ 10 5 cos(1010 t
x)az T. Next we use
H
ay = 2 sin(1010 t
x
⇤H=
x)ay =
D
t
from which
D=
2 sin(1010 t
x) dt + C =
2
cos(1010 t
1010
x)ay C/m2
where the integration constant is set to zero, since no dc fields are presumed to exist. Next,
E=
D
=
⇥
2
(1.2 ⇤ 10
10 )(1010 )
Now
⇤E=
cos(1010 t
Ey
az = 1.67
x
2
x)ay =
1.67 cos(1010 t
sin(1010 t
x)az =
x)ay V/m
B
t
So
B=
1.67
2
sin(1010 t
x)az dt = (1.67 ⇤ 10
10
)
2
cos(1010 t
x)az
We require this result to be consistent with the expression for B originally found. So
(1.67 ⇤ 10
10
)
2
= 6 ⇤ 10
170
5
⌃
= ±600 rad/m
9.16. Derive the continuity equation from Maxwell’s equations: First, take the divergence of both
sides of Ampere’s circuital law:
·
⇣
where we have used
✏
0
⇤H=
⌘
·J+
t
·D=
·J+
⇧v
=0
t
· D = ⇧v , another Maxwell equation.
9.17. The electric field intensity in the region 0 < x < 5, 0 < y < ⌅/12, 0 < z < 0.06 m in free
space is given by E = C sin(12y) sin(az) cos(2 ⇤ 1010 t) ax V/m. Beginning with the ⇤ E
relationship, use Maxwell’s equations to find a numerical value for a, if it is known that a is
greater than zero: In this case we find
⇤E=
Ex
ay
z
Ez
az
y
= C [a sin(12y) cos(az)ay
12 cos(12y) sin(az)az ] cos(2 ⇤ 1010 t) =
B
t
Then
H=
=
1
µ0
⇤ E dt + C1
C
[a sin(12y) cos(az)ay
µ0 (2 ⇤ 1010
12 cos(12y) sin(az)az ] sin(2 ⇤ 1010 t) A/m
where the integration constant, C1 = 0, since there are no initial conditions. Using this result,
we now find
⌦
Hz
Hy
C(144 + a2 )
D
⇤H=
ax =
sin(12y) sin(az) sin(2 ⇤ 1010 t) ax =
y
z
µ0 (2 ⇤ 1010 )
t
Now
E=
D
=
⇥0
1
⇥0
⇤ H dt + C2 =
C(144 + a2 )
sin(12y) sin(az) cos(2 ⇤ 1010 t) ax
µ0 ⇥0 (2 ⇤ 1010 )2
where C2 = 0. This field must be the same as the original field as stated, and so we require
that
C(144 + a2 )
=1
µ0 ⇥0 (2 ⇤ 1010 )2
Using µ0 ⇥0 = (3 ⇤ 108 )
2
, we find
(2 ⇤ 1010 )2
a=
(3 ⇤ 108 )2
171
⌦1/2
144
= 66 m
1
9.18. The parallel plate transmission line shown in Fig. 9.7 has dimensions b = 4 cm and d = 8 mm,
while the medium between plates is characterized by µr = 1, ⇥r = 20, and ⌃ = 0. Neglect fields
outside the dielectric. Given the field H = 5 cos(109 t
z)ay A/m, use Maxwell’s equations
to help find:
a) , if > 0: Take
Hy
E
⇤H=
ax = 5 sin(109 t
z)ax = 20⇥0
z
t
So
5
E=
sin(109 t
z)ax dt =
cos(109 t
z)ax
20⇥0
(4 ⇤ 109 )⇥0
Then
2
Ex
H
⇤E=
ay =
sin(109 t
z)ay = µ0
9
z
(4 ⇤ 10 )⇥0
t
So that
H=
2
(4 ⇤ 109 )µ0 ⇥0
sin(109 t
z)ax dt =
2
(4 ⇤ 1018 )µ0 ⇥0
cos(109 t
z)
= 5 cos(109 t
z)ay
where the last equality is required to maintain consistency. Therefore
2
(4 ⇤ 1018 )µ0 ⇥0
=5 ⌃
= 14.9 m
1
b) the displacement current density at z = 0: Since ⌃ = 0, we have
⇤ H = Jd =
=
5 sin(109 t
z) =
74.5 sin(109 t
14.9z)ax
74.5 sin(109 t)ax A/m at z = 0
c) the total displacement current crossing the surface x = 0.5d, 0 < y < b, and 0 < z < 0.1
m in the ax direction. We evaluate the flux integral of Jd over the given cross section:
0.1
⇤
⌅
Id = 74.5b
sin(109 t 14.9z) ax · ax dz = 0.20 cos(109 t 1.49) cos(109 t) A
0
9.19. In the first section of this chapter, Faraday’s law was used to show that the field E =
1
kt
kt
2 kB0 ⇧e a⌅ results from the changing magnetic field B = B0 e az .
a) Show that these fields do not satisfy Maxwell’s other curl equation: Note that B as stated
is constant with position, and so will have zero curl. The electric field, however, varies
with time, and so ⇤H = ⌥D
⌥t would have a zero left-hand side and a non-zero right-hand
side. The equation is thus not valid with these fields.
b) If we let B0 = 1 T and k = 106 s 1 , we are establishing a fairly large magnetic flux
density in 1 µs. Use the ⇤ H equation to show that the rate at which Bz should (but
does not) change with ⇧ is only about 5 ⇤ 10 6 T/m in free space at t = 0: Assuming
that B varies with ⇧, we write
Hz
1 dB0 kt
E
1
⇤H=
a⌅ =
e = ⇥0
=
⇥0 k2 B0 ⇧ekt
⇧
µ0 d⇧
t
2
Thus
dB0
1
1012 (1)⇧
= µ0 ⇥0 k2 ⇧B0 =
= 5.6 ⇤ 10 6 ⇧
d⇧
2
2(3 ⇤ 108 )2
which is near the stated value if ⇧ is on the order of 1m.
172
9.20. Given Maxwell’s equations in point form, assume that all fields vary as est and write the
equations without explicitly involving time: Write all fields in the general form A(r, t) =
A0 (r)est , where r is a position vector in any coordinate system. Maxwell’s equations become:
⇤ E0 (r) est =
t
⇤ H0 (r) est = J0 (r)est +
t
⇥
B0 (r) est =
sB0 (r) est
⇥
D0 (r) est = J0 (r)est + sD0 (r) est
· D0 (r) est = ⇧0 (r) est
· B0 (r) est = 0
In all cases, the est terms divide out, leaving:
⇤ E0 (r) =
sB0 (r)
⇤ H0 (r) = J0 (r) + sD0 (r)
· D0 (r) = ⇧0 (r)
· B0 (r) = 0
9.21. a) Show that under static field conditions, Eq. (55) reduces to Ampere’s circuital law. First
use the definition of the vector Laplacian:
2
A=
⇤
⇤A+
(
· A) =
µJ
which is Eq. (55) with the time derivative set to zero. We also note that
state (from Eq. (54)). Now, since B = ⇤ A, (55) becomes
⇤B=
· A = 0 in steady
⇤H=J
µJ ⌃
b) Show that Eq. (51) becomes Faraday’s law when taking the curl: Doing this gives
⇤E=
⇤
V
The curl of the gradient is identially zero, and
⇤E=
173
t
⇤A
⇤ A = B. We are left with
B/ t
9.22. In a sourceless medium, in which J = 0 and ⇧v = 0, assume a rectangular coordinate system in
which E and H are functions only of z and t. The medium has permittivity ⇥ and permeability
µ.
a) If E = Ex ax and H = Hy ay , begin with Maxwell’s equations and determine the second
order partial di⌅erential equation that Ex must satisfy.
First use
B
Ex
Hy
⇤E=
⌃
ay = µ
ay
t
z
t
in which case, the curl has dictated the direction that H must lie in. Similarly, use the
other Maxwell curl equation to find
⇤H=
D
t
⌃
Hy
Ex
ax = ⇥
ax
z
t
Now, di⌅erentiate the first equation with respect to z, and the second equation with
respect to t:
2
2
2
2
Ex
Hy
Hy
Ex
=
µ
and
=
⇥
z2
t z
z t
t2
Combining these two, we find
2
2
Ex
Ex
=
µ⇥
z2
t2
b) Show that Ex = E0 cos( t
Substituting, we find
2
Ex
=
z2
2
z) is a solution of that equation for a particular value of :
E0 cos( t
z) and µ⇥
2
Ex
=
t2
2
µ⇥E0 cos( t
These two will be equal provided the constant multipliers of cos( t
c) Find
=
z)
z) are equal.
as a function of given parameters. Equating the two constants in part b, we find
µ⇥.
174
9.23. In region 1, z < 0, ⇥1 = 2 ⇤ 10 11 F/m, µ1 = 2 ⇤ 10 6 H/m, and ⌃1 = 4 ⇤ 10 3 S/m; in
region 2, z > 0, ⇥2 = ⇥1 /2, µ2 = 2µ1 , and ⌃2 = ⌃1 /4. It is known that E1 = (30ax + 20ay +
10az ) cos(109 t) V/m at P1 (0, 0, 0 ).
a) Find EN 1 , Et1 , DN 1 , and Dt1 : These will be
EN 1 = 10 cos(109 t)az V/m Et1 = (30ax + 20ay ) cos(109 t) V/m
DN 1 = ⇥1 EN 1 = (2 ⇤ 10
Dt1 = ⇥1 Et1 = (2 ⇤ 10
11
11
)(10) cos(109 t)az C/m2 = 200 cos(109 t)az pC/m2
)(30ax + 20ay ) cos(109 t) = (600ax + 400ay ) cos(109 t) pC/m2
b) Find JN 1 and Jt1 at P1 :
JN 1 = ⌃1 EN 1 = (4 ⇤ 10
Jt1 = ⌃1 Et1 = (4 ⇤ 10
3
3
)(10 cos(109 t))az = 40 cos(109 t)az mA/m2
)(30ax + 20ay ) cos(109 t) = (120ax + 80ay ) cos(109 t) mA/m2
c) Find Et2 , Dt2 , and Jt2 at P1 : By continuity of tangential E,
Et2 = Et1 = (30ax + 20ay ) cos(109 t) V/m
Then
Dt2 = ⇥2 Et2 = (10
Jt2 = ⌃2 Et2 = (10
11
)(30ax + 20ay ) cos(109 t) = (300ax + 200ay ) cos(109 t) pC/m2
3
)(30ax + 20ay ) cos(109 t) = (30ax + 20ay ) cos(109 t) mA/m2
d) (Harder) Use the continuity equation to help show that JN 1 JN 2 = DN 2 / t
DN 1 / t
and then determine EN 2 , DN 2 , and JN 2 : We assume the existence of a surface charge
layer at the boundary having density ⇧s C/m2 . If we draw a cylindrical “pillbox” whose
top and bottom surfaces (each of area a) are on either side of the interface, we may use
the continuity condition to write
(JN 2
where ⇧s = DN 2
JN 1 ) a =
⇧s
a
t
DN 1 . Therefore,
JN 1
JN 2 =
t
(DN 2
DN 1 )
In terms of the normal electric field components, this becomes
⌃1 EN 1
⌃2 EN 2 =
t
(⇥2 EN 2
⇥1 EN 1 )
Now let EN 2 = A cos(109 t) + B sin(109 t), while from before, EN 1 = 10 cos(109 t).
175
9.23d (continued)
These, along with the permittivities and conductivities, are substituted to obtain
(4 ⇤ 10 3 )(10) cos(109 t) 10 3 [A cos(109 t) + B sin(109 t)]
⇤
⌅
=
10 11 [A cos(109 t) + B sin(109 t)] (2 ⇤ 10 11 )(10) cos(109 t)
t
= (10 2 A sin(109 t) + 10 2 B cos(109 t) + (2 ⇤ 10 1 ) sin(109 t)
We now equate coe⌥cients of the sin and cos terms to obtain two equations:
4 ⇤ 10
10
3
2
10
3
A = 10
B=
10
2
A + 2 ⇤ 10
2
B
1
These are solved together to find A = 20.2 and B = 2.0. Thus
Then
and
⇤
⌅
EN 2 = 20.2 cos(109 t) + 2.0 sin(109 t) az = 20.3 cos(109 t + 5.6⇥ )az V/m
DN 2 = ⇥2 EN 2 = 203 cos(109 t + 5.6⇥ )az pC/m2
JN 2 = ⌃2 EN 2 = 20.3 cos(109 t + 5.6⇥ )az mA/m2
176
9.24. A vector potential is given as A = A0 cos( t
possible are zero, find H, E, and V ;
kz) ay . a) Assuming as many components as
With A y-directed only, and varying spatially only with z, we find
H=
1
µ
1 Ay
ax =
µ z
⇤A=
kA0
sin( t
µ
kz) ax A/m
Now, in a lossless medium we will have zero conductivity, so that the point form of
Ampere’s circuital law involves only the displacement current term:
⇤H=
D
E
=⇥
t
t
Using the magnetic field as found above, we find
⇤H=
Hx
k2 A0
ay =
cos( t
z
µ
Now,
E=
A
⌃
t
V
or
V = A0
Integrating over y we find
kz) ay =
⌦
k2
sin( t
2 µ⇥
y 1
kz) ay V/m
⌦
A
+E
t
V =
⌦
k2
sin( t
2 µ⇥
1
V = A0
E
k2 A0
⌃ E=
sin( t
t
µ⇥
kz) ay = ⇥
V
ay
y
kz) + C
where C, the integration constant, can be taken as zero. In part b, it will be shown that
k=
µ⇥, which means that V = 0.
b) Specify k in terms of A0 , , and the constants of the lossless medium, ⇥ and µ. Use the
other Maxwell curl equation:
⇤E=
so that
H
=
t
1
µ
⇤E=
B
=
t
µ
1 Ey
ax =
µ z
H
t
k3 A0
cos( t
µ2 ⇥
kz) ax
Integrate over t (and set the integration constant to zero) and require the result to be
consistant with part a:
H=
k3 A0
sin( t
2 µ2 ⇥
kz) ax =
kA0
sin( t
µ
⇣
✏
kz) ax
⌘
from part a
We identify
k=
177
µ⇥
9.25. In a region where µr = ⇥r = 1 and ⌃ = 0, the retarded potentials are given by V = x(z
V and A = x[(z/c) t]az Wb/m, where c = 1/ µ0 ⇥0 .
a) Show that
·A=
ct)
µ⇥( V / t):
First,
Az
x
= = x µ0 ⇥0
z
c
·A=
Second,
V
=
t
x
µ0 ⇥0
cx =
so we observe that ·A = µ0 ⇥0 ( V / t) in free space, implying that the given statement
would hold true in general media.
b) Find B, H, E, and D:
Use
B=
Then
H=
Now,
E=
A
=
t
V
Then
⇧
Ax
ay = t
x
⇤A=
B
1 ⇧
=
t
µ0
µ0
(z
ct)ax
z⌃
ay T
c
z⌃
ay A/m
c
xaz + xaz = (ct
D = ⇥0 E = ⇥0 (ct
z)ax V/m
z)ax C/m2
c) Show that these results satisfy Maxwell’s equations if J and ⇧v are zero:
i.
· D = · ⇥0 (ct z)ax = 0
ii.
iii.
·B=
· (t
z/c)ay = 0
Hy
1
ax =
ax =
z
µ0 c
⇤H=
which we require to equal D/ t:
D
= ⇥0 cax =
t
iv.
Ex
ay =
z
⇤E=
which we require to equal
⇥0
ax
µ0
B/ t:
B
= ay
t
So all four Maxwell equations are satisfied.
178
ay
⇥0
ax
µ0
9.26. Write Maxwell’s equations in point form in terms of E and H as they apply to a sourceless
medium, where J and ⇧v are both zero. Replace ⇥ by µ, µ by ⇥, E by H, and H by E,
and show that the equations are unchanged. This is a more general expression of the duality
principle in circuit theory.
Maxwell’s equations in sourceless media can be written as:
⇤E=
H
t
µ
⇤H=⇥
E
t
· ⇥E = 0
· µH = 0
(1)
(2)
(3)
(4)
In making the above substitutions, we find that (1) converts to (2), (2) converts to (1),
and (3) and (4) convert to each other.
179
CHAPTER 10
10.1. In Fig. 10.4, let B = 0.2 cos 120πt T, and assume that the conductor joining the two ends
of the resistor is perfect. It may be assumed that the magnetic field produced by I(t) is
negligible. Find:
a) Vab (t): Since B is constant over the loop area, the flux is Φ = π(0.15)2 B = 1.41 ×
10−2 cos 120πt Wb. Now, emf = Vba (t) = −dΦ/dt = (120π)(1.41 × 10−2 ) sin 120πt.
Then Vab (t) = −Vba (t) = −5.33 sin 120πt V.
b) I(t) = Vba (t)/R = 5.33 sin(120πt)/250 = 21.3 sin(120πt) mA
10.2. In Fig. 10.1, replace the voltmeter with a resistance, R.
a) Find the current I that flows as a result of the motion of the sliding bar: The current is
found through
1
I=
R
!
E · dL = −
1 dΦm
R dt
Taking the normal to the path integral as az , the path direction will be counter-clockwise
when viewed from above (in the −az direction). The minus sign in the equation indicates
that the current will therefore flow clockwise, since the magnetic flux is increasing with
time. The flux of B is Φm = Bdvt, and so
|I| =
1 dΦm
Bdv
=
(clockwise)
R dt
R
b) The bar current results in a force exerted on the bar as it moves. Determine this force:
F=
"
IdL × B =
"
0
d
Idxax × Baz =
"
0
d
Bdv
B 2 d2 v
ax × Baz = −
ay N
R
R
c) Determine the mechanical power required to maintain a constant velocity v and show
that this power is equal to the power absorbed by R. The mechanical power is
Pm = F v =
The electrical power is
Pe = I 2 R =
(Bdv)2
W
R
(Bdv)2
= Pm
R
10.3. Given H = 300 az cos(3 × 108t − y) A/m in free space, find the emf developed in the general
aφ direction about the closed path having corners at
a) (0,0,0), (1,0,0), (1,1,0), and (0,1,0): The magnetic flux will be:
" 1" 1
Φ=
300µ0 cos(3 × 108t − y) dx dy = 300µ0 sin(3 × 108t − y)|10
0
0
$
#
= 300µ0 sin(3 × 108t − 1) − sin(3 × 108t) Wb
Then
#
$
dΦ
= −300(3 × 108)(4π × 10−7) cos(3 × 108t − 1) − cos(3 × 108t)
dt
#
$
= −1.13 × 105 cos(3 × 108t − 1) − cos(3 × 108t) V
emf = −
b) corners at (0,0,0), (2π,0,0), (2π,2π,0), (0,2π,0): In this case, the flux is
Φ = 2π × 300µ0 sin(3 × 108t − y)|2π
0 =0
The emf is therefore 0.
10.4. Conductor surfaces are located at ρ = 1cm and ρ = 2cm in free space. The volume 1 cm <
ρ < 2 cm contains the fields Hφ = (2/ρ) cos(6 × 108πt − 2πz) A/m and Eρ = (240π/ρ) cos(6 ×
108πt − 2πz) V/m.
a) Show that these two fields satisfy Eq. (6), Sec. 10.1: Have
∇×E=
∂Eρ
2π(240π)
480π 2
aφ =
sin(6 × 108πt − 2πz) aφ =
sin(6 × 108πt − 2πz)aφ
∂z
ρ
ρ
Then
2µ0 (6 × 108)π
∂B
=
sin(6 × 108πt − 2πz) aφ
−
∂t
ρ
(8π × 10−7)(6 × 108)π
480π 2
=
sin(6 × 108πt − 2πz) =
sin(6 × 108πt − 2πz) aφ
ρ
ρ
b) Evaluate both integrals in Eq. (4) for the planar surface defined by φ = 0, 1cm < ρ < 2cm,
0 < z < 0.1m, and its perimeter, and show that the same results are obtained: we take
the normal to the surface as positive aφ , so the the loop surrounding the surface (by the
right hand rule) is in the negative aρ direction at z = 0, and is in the positive aρ direction
at z = 0.1. Taking the left hand side first, we find
!
" .01
240π
cos(6 × 108πt) aρ · aρ dρ
E · dL =
ρ
.02
" .02
240π
cos(6 × 108πt − 2π(0.1)) aρ · aρ dρ
+
ρ
.01
% &
% &
1
2
8
8
+ 240π cos(6 × 10 πt − 0.2π) ln
= 240π cos(6 × 10 πt) ln
2
1
$
#
8
8
= 240(ln 2) cos(6 × 10 πt − 0.2π) − cos(6 × 10 πt)
10.4b (continued). Now for the right hand side. First,
"
"
B · dS =
"
0.1
0
"
.02
.01
0.1
8π × 10−7
cos(6 × 108πt − 2πz) aφ · aφ dρ dz
ρ
(8π × 10−7) ln 2 cos(6 × 108πt − 2πz) dz
0
#
$
= −4 × 10−7 ln 2 sin(6 × 108πt − 0.2π) − sin(6 × 108πt)
=
Then
−
"
d
dt
$
#
B · dS = 240π(ln 2) cos(6 × 108πt − 0.2π) − cos(6 × 108πt) (check)
10.5. The location of the sliding bar in Fig. 10.5 is given by x = 5t + 2t3, and the separation of the
two rails is 20 cm. Let B = 0.8x2 az T. Find the voltmeter reading at:
a) t = 0.4 s: The flux through the loop will be
Φ=
"
0.2
0
"
x
0.8(x′ )2 dx′ dy =
0
0.16 3 0.16
x =
(5t + 2t3)3 Wb
3
3
Then
emf = −
dΦ
0.16
=
(3)(5t + 2t3)2 (5 + 6t2 ) = −(0.16)[5(.4) + 2(.4)3]2 [5 + 6(.4)2 ] = −4.32 V
dt
3
b) x = 0.6 m: Have 0.6 = 5t + 2t3, from which we find t = 0.1193. Thus
emf = −(0.16)[5(.1193) + 2(.1193)3]2 [5 + 6(.1193)2 ] = −.293 V
10.6. A perfectly conducting filament containing a small 500-Ω resistor is formed into a square, as
illustrated in Fig. 10.6. Find I(t) if
a) B = 0.3 cos(120πt − 30◦ ) az T: First the flux through the loop is evaluated, where the
unit normal to the loop is az . We find
Φ=
"
loop
B · dS = (0.3)(0.5)2 cos(120πt − 30◦ ) Wb
Then the current will be
I(t) =
emf
1 dΦ
(120π)(0.3)(0.25)
=−
=
sin(120πt − 30◦ ) = 57 sin(120πt − 30◦ ) mA
R
R dt
500
b) B = 0.4 cos[π(ct − y)] az µT where c = 3 × 108 m/s: Since the field varies with y, the flux
is now
Φ=
"
loop
B · dS = (0.5)(0.4)
"
0
.5
cos(πy − πct) dy =
0.2
[sin(πct − π/2) − sin(πct)] µWb
π
The current is then
emf
1 dΦ
−0.2c
=−
=
[cos(πct − π/2) − cos(πct)] µA
R
R dt
500
−0.2(3 × 108)
[sin(πct) − cos(πct)] µA = 120 [cos(πct) − sin(πct)] mA
=
500
I(t) =
10.7. The rails in Fig. 10.7 each have a resistance of 2.2 Ω/m. The bar moves to the right at a
constant speed of 9 m/s in a uniform magnetic field of 0.8 T. Find I(t), 0 < t < 1 s, if the bar
is at x = 2 m at t = 0 and
a) a 0.3 Ω resistor is present across the left end with the right end open-circuited: The flux
in the left-hand closed loop is
Φl = B × area = (0.8)(0.2)(2 + 9t)
Then, emf l = −dΦl /dt = −(0.16)(9) = −1.44 V. With the bar in motion, the loop
resistance is increasing with time, and is given by Rl (t) = 0.3 + 2[2.2(2 + 9t)]. The current
is now
Il (t) =
emf l
−1.44
=
A
Rl (t)
9.1 + 39.6t
Note that the sign of the current indicates that it is flowing in the direction opposite that
shown in the figure.
b) Repeat part a, but with a resistor of 0.3 Ω across each end: In this case, there will be
a contribution to the current from the right loop, which is now closed. The flux in the
right loop, whose area decreases with time, is
Φr = (0.8)(0.2)[(16 − 2) − 9t]
and emf r = −dΦr /dt = (0.16)(9) = 1.44 V. The resistance of the right loop is Rr (t) =
0.3 + 2[2.2(14 − 9t)], and so the contribution to the current from the right loop will be
Ir (t) =
−1.44
A
61.9 − 39.6t
10.7b (continued). The minus sign has been inserted because again the current must flow in the
opposite direction as that indicated in the figure, with the flux decreasing with time. The
total current is found by adding the part a result, or
'
1
1
IT (t) = −1.44
+
61.9 − 39.6t 9.1 + 39.6t
(
A
10.8. Fig. 10.1 is modified to show that the rail separation is larger when y is larger. Specifically, let
the separation d = 0.2 + 0.02y. Given a uniform velocity vy = 8 m/s and a uniform magnetic
flux density Bz = 1.1 T, find V12 as a function of time if the bar is located at y = 0 at t = 0:
The flux through the loop as a function of y can be written as
Φ=
"
B · dS =
"
0
y
"
.2+.02y ′
1.1 dx dy ′ =
0
"
y
1.1(.2 + .02y ′ ) dy ′ = 0.22y(1 + .05y)
0
Now, with y = vt = 8t, the above becomes Φ = 1.76t(1 + .40t). Finally,
V12 = −
dΦ
= −1.76(1 + .80t) V
dt
10.9. A square filamentary loop of wire is 25 cm on a side and has a resistance of 125 Ω per meter
length. The loop lies in the z = 0 plane with its corners at (0, 0, 0), (0.25, 0, 0), (0.25, 0.25, 0),
and (0, 0.25, 0) at t = 0. The loop is moving with velocity vy = 50 m/s in the field Bz =
8 cos(1.5 × 108t − 0.5x) µT. Develop a function of time which expresses the ohmic power being
delivered to the loop: First, since the field does not vary with y, the loop motion in the y
direction does not produce any time-varying flux, and so this motion is immaterial. We can
evaluate the flux at the original loop position to obtain:
"
.25" .25
8 × 10−6cos(1.5 × 108t − 0.5x) dx dy
0
0
#
$
= −(4 × 10−6) sin(1.5 × 108t − 0.13x) − sin(1.5 × 108t) Wb
Φ(t) =
$
#
Now, emf = V (t) = −dΦ/dt = 6.0×102 cos(1.5 × 108t − 0.13x) − cos(1.5 × 108t) , The total
loop resistance is R = 125(0.25 + 0.25 + 0.25 + 0.25) = 125 Ω. Then the ohmic power is
P (t) =
#
$
V 2 (t)
= 2.9 × 103 cos(1.5 × 108t − 0.13x) − cos(1.5 × 108t) Watts
R
10.10a. Show that the ratio of the amplitudes of the conduction current density and the displacement
current density is σ/ωϵ for the applied field E = Em cos ωt. Assume µ = µ0 . First, D =
ϵE = ϵEm cos ωt. Then the displacement current density is ∂D/∂t = −ωϵEm sin ωt. Second,
Jc = σE = σEm cos ωt. Using these results we find |Jc |/|Jd | = σ/ωϵ.
b. What is the amplitude ratio if the applied field is E = Em e−t/τ , where τ is real? As before,
find D = ϵE = ϵEm e−t/τ , and so Jd = ∂D/∂t = −(ϵ/τ )Em e−t/τ . Also, Jc = σEm e−t/τ .
Finally, |Jc |/|Jd | = στ /ϵ.
10.11. Let the internal dimension of a coaxial capacitor be a = 1.2 cm, b = 4 cm, and l = 40 cm.
The homogeneous material inside the capacitor has the parameters ϵ = 10−11 F/m, µ = 10−5
H/m, and σ = 10−5 S/m. If the electric field intensity is E = (106/ρ) cos(105t)aρ V/m, find:
a) J: Use
J = σE =
%
10
ρ
&
cos(105t)aρ A/m2
b) the total conduction current, Ic , through the capacitor: Have
Ic =
" "
J · dS = 2πρlJ = 20πl cos(105t) = 8π cos(105t) A
c) the total displacement current, Id , through the capacitor: First find
Jd =
∂D
1
∂
(105)(10−11 )(106)
= (ϵE) = −
sin(105t)aρ = − sin(105t) A/m
∂t
∂t
ρ
ρ
Now
Id = 2πρlJd = −2πl sin(105t) = −0.8π sin(105t) A
d) the ratio of the amplitude of Id to that of Ic , the quality factor of the capacitor: This will
be
|Id |
0.8
=
= 0.1
|Ic |
8
10.12. Show that the displacement current flowing between the two conducting cylinders in a lossless
coaxial capacitor is exactly the same as the conduction current flowing in the external circuit
if the applied voltage between conductors is V0 cos ωt volts.
From Chapter 7, we know that for a given applied voltage between the cylinders, the
electric field is
E=
ϵV0 cos ωt
V0 cos ωt
aρ V/m ⇒ D =
aρ C/m2
ρ ln(b/a)
ρ ln(b/a)
Then the displacement current density is
∂D
−ωϵV0 sin ωt
=
aρ
∂t
ρ ln(b/a)
Over a length ℓ, the displacement current will be
" "
∂D
∂D
2πℓωϵV0 sin ωt
dV
Id =
· dS = 2πρℓ
=
=C
= Ic
∂t
∂t
ln(b/a)
dt
where we recall that the capacitance is given by C = 2πϵℓ/ ln(b/a).
10.13. Consider the region defined by |x|, |y|, and |z| < 1. Let ϵr = 5, µr = 4, and σ = 0. If
Jd = 20 cos(1.5 × 108t − bx)ay µA/m2 ;
a) find D and E: Since Jd = ∂D/∂t, we write
"
20 × 10−6
sin(1.5 × 108 − bx)ay
1.5 × 108
= 1.33 × 10−13sin(1.5 × 108t − bx)ay C/m2
D=
Jd dt + C =
where the integration constant is set to zero (assuming no dc fields are present). Then
D
1.33 × 10−13
=
sin(1.5 × 108t − bx)ay
ϵ
(5 × 8.85 × 10−12 )
= 3.0 × 10−3sin(1.5 × 108t − bx)ay V/m
E=
b) use the point form of Faraday’s law and an integration with respect to time to find B and
H: In this case,
∇×E=
∂B
∂Ey
az = −b(3.0 × 10−3) cos(1.5 × 108t − bx)az = −
∂x
∂t
Solve for B by integrating over time:
B=
b(3.0 × 10−3)
sin(1.5 × 108t − bx)az = (2.0)b × 10−11 sin(1.5 × 108t − bx)az T
1.5 × 108
10.13b (continued). Now
(2.0)b × 10−11
B
=
sin(1.5 × 108t − bx)az
µ
4 × 4π × 10−7
= (4.0 × 10−6)b sin(1.5 × 108t − bx)az A/m
H=
c) use ∇ × H = Jd + J to find Jd : Since σ = 0, there is no conduction current, so in this
case
∇×H=−
∂Hz
ay = 4.0 × 10−6b2 cos(1.5 × 108t − bx)ay A/m2 = Jd
∂x
d) What is the numerical value of b? We set the given expression for Jd equal to the result
of part c to obtain:
20 × 10−6 = 4.0 × 10−6b2 ⇒ b =
√
5.0 m−1
10.14. A voltage source, V0 sin ωt, is connected between two concentric conducting spheres, r = a
and r = b, b > a, where the region between them is a material for which ϵ = ϵr ϵ0 , µ = µ0 , and
σ = 0. Find the total displacement current through the dielectric and compare it with the
source current as determined from the capacitance (Sec. 5.10) and circuit analysis methods:
First, solving Laplace’s equation, we find the voltage between spheres (see Eq. 20, Chapter
7):
V (t) =
(1/r) − (1/b)
V0 sin ωt
(1/a) − (1/b)
Then
E = −∇V =
V0 sin ωt
ar
2
r (1/a − 1/b)
⇒ D=
ϵr ϵ0 V0 sin ωt
ar
r2 (1/a − 1/b)
Now
Jd =
∂D
ϵr ϵ0 ωV0 cos ωt
= 2
ar
∂t
r (1/a − 1/b)
The displacement current is then
Id = 4πr2 Jd =
4πϵr ϵ0 ωV0 cos ωt
dV
=C
(1/a − 1/b)
dt
where, from Eq. 47, Chapter 5,
C=
The results are consistent.
4πϵr ϵ0
(1/a − 1/b)
10.15. Let µ = 3×10−5 H/m, ϵ = 1.2×10−10 F/m, and σ = 0 everywhere. If H = 2 cos(1010 t−βx)az
A/m, use Maxwell’s equations to obtain expressions for B, D, E, and β: First, B = µH =
6 × 10−5cos(1010 t − βx)az T. Next we use
∇×H=−
∂D
∂H
ay = 2β sin(1010 t − βx)ay =
∂x
∂t
from which
D=
"
2β sin(1010 t − βx) dt + C = −
2β
cos(1010 t − βx)ay C/m2
1010
where the integration constant is set to zero, since no dc fields are presumed to exist. Next,
E=
2β
D
=−
cos(1010 t − βx)ay = −1.67β cos(1010 t − βx)ay V/m
ϵ
(1.2 × 10−10 )(1010 )
Now
∇×E=
So
B=−
"
∂Ey
∂B
az = 1.67β 2 sin(1010 t − βx)az = −
∂x
∂t
1.67β 2 sin(1010 t − βx)az dt = (1.67 × 10−10 )β 2 cos(1010 t − βx)az
We require this result to be consistent with the expression for B originally found. So
(1.67 × 10−10 )β 2 = 6 × 10−5 ⇒ β = ±600 rad/m
10.16. Derive the continuity equation from Maxwell’s equations: First, take the divergence of both
sides of Ampere’s circuital law:
∂
∂ρv
=0
∇·∇×H=∇·J+ ∇·D=∇·J+
) *+ ,
∂t
∂t
0
where we have used ∇ · D = ρv , another Maxwell equation.
10.17. The electric field intensity in the region 0 < x < 5, 0 < y < π/12, 0 < z < 0.06 m in free
space is given by E = C sin(12y) sin(az) cos(2 × 1010 t) ax V/m. Beginning with the ∇ × E
relationship, use Maxwell’s equations to find a numerical value for a, if it is known that a is
greater than zero: In this case we find
∇×E=
∂Ex
∂Ez
ay −
az
∂z
∂y
= C [a sin(12y) cos(az)ay − 12 cos(12y) sin(az)az ] cos(2 × 1010 t) = −
∂B
∂t
10.17 (continued). Then
"
1
∇ × E dt + C1
H=−
µ0
C
=−
[a sin(12y) cos(az)ay − 12 cos(12y) sin(az)az ] sin(2 × 1010 t) A/m
µ0 (2 × 1010
where the integration constant, C1 = 0, since there are no initial conditions. Using this result,
we now find
'
∂Hy
∂Hz
−
∇×H=
∂y
∂z
(
ax = −
C(144 + a2 )
∂D
sin(12y) sin(az) sin(2 × 1010 t) ax =
10
µ0 (2 × 10 )
∂t
Now
E=
D
=
ϵ0
"
1
C(144 + a2 )
∇ × H dt + C2 =
sin(12y) sin(az) cos(2 × 1010 t) ax
ϵ0
µ0 ϵ0 (2 × 1010 )2
where C2 = 0. This field must be the same as the original field as stated, and so we require
that
Using µ0 ϵ0 = (3 × 108)−2 , we find
'
C(144 + a2 )
=1
µ0 ϵ0 (2 × 1010 )2
(2 × 1010 )2
− 144
a=
(3 × 108)2
(1/2
= 66 m−1
10.18. The parallel plate transmission line shown in Fig. 10.8 has dimensions b = 4 cm and d = 8
mm, while the medium between plates is characterized by µr = 1, ϵr = 20, and σ = 0. Neglect
fields outside the dielectric. Given the field H = 5 cos(109t − βz)ay A/m, use Maxwell’s
equations to help find:
a) β, if β > 0: Take
∇×H=−
So
E=
Then
"
−5β
β
sin(109t − βz)ax dt =
cos(109t − βz)ax
20ϵ0
(4 × 109)ϵ0
∇×E=
So that
"
∂Hy
∂E
ax = −5β sin(109t − βz)ax = 20ϵ0
∂z
∂t
∂Ex
β2
∂H
ay =
sin(109t − βz)ay = −µ0
9
∂z
(4 × 10 )ϵ0
∂t
−β 2
β2
9
sin(10
t
−
βz)a
dt
=
cos(109t − βz)
x
(4 × 109)µ0 ϵ0
(4 × 1018)µ0 ϵ0
= 5 cos(109t − βz)ay
H=
10.18a (continued) where the last equality is required to maintain consistency. Therefore
β2
= 5 ⇒ β = 14.9 m−1
(4 × 1018)µ0 ϵ0
b) the displacement current density at z = 0: Since σ = 0, we have
∇ × H = Jd = −5β sin(109t − βz) = −74.5 sin(109t − 14.9z)ax
= −74.5 sin(109t)ax A/m at z = 0
c) the total displacement current crossing the surface x = 0.5d, 0 < y < b, and 0 < z < 0.1
m in the ax direction. We evaluate the flux integral of Jd over the given cross section:
Id = −74.5b
"
0
0.1
#
$
sin(109t − 14.9z) ax · ax dz = 0.20 cos(109t − 1.49) − cos(109t) A
10.19. In the first section of this chapter, Faraday’s law was used to show that the field E =
− 12 kB0 ρekt aφ results from the changing magnetic field B = B0 ekt az .
a) Show that these fields do not satisfy Maxwell’s other curl equation: Note that B as stated
is constant with position, and so will have zero curl. The electric field, however, varies
with time, and so ∇×H =
∂D
∂t
would have a zero left-hand side and a non-zero right-hand
side. The equation is thus not valid with these fields.
b) If we let B0 = 1 T and k = 106 s−1 , we are establishing a fairly large magnetic flux
density in 1 µs. Use the ∇ × H equation to show that the rate at which Bz should (but
does not) change with ρ is only about 5 × 10−6 T/m in free space at t = 0: Assuming
that B varies with ρ, we write
∇×H=−
Thus
∂E
∂Hz
1 dB0 kt
1
aφ = −
e = ϵ0
= − ϵ0 k 2 B0 ρekt
∂ρ
µ0 dρ
∂t
2
dB0
1012 (1)ρ
1
= 5.6 × 10−6ρ
= µ0 ϵ0 k 2 ρB0 =
dρ
2
2(3 × 108)2
which is near the stated value if ρ is on the order of 1m.
10.20. Point C(−0.1, −0.2, 0.3) lies on the surface of a perfect conductor. The electric field intensity
at C is (500ax − 300ay + 600az ) cos 107t V/m, and the medium surrounding the conductor is
characterized by µr = 5, ϵr = 10, and σ = 0.
a) Find a unit vector normal to the conductor surface at C, if the origin lies within the
conductor: At t = 0, the field must be directed out of the surface, and will be normal to
it, since we have a perfect conductor. Therefore
n=
+E(t = 0)
5ax − 3ay + 6az
= √
= 0.60ax − 0.36ay + 0.72az
|E(t = 0)|
25 + 9 + 36
b) Find the surface charge density at C: Use
ρs = D · n|surf ace = 10ϵ0 [500ax − 300ay + 600az ] cos(107t) · [.60ax − .36ay + .72az ]
= 10ϵ0 [300 + 108 + 432] cos(107t) = 7.4 × 10−8cos(107t) C/m2
= 74 cos(107t) nC/m2
10.21. a) Show that under static field conditions, Eq. (55) reduces to Ampere’s circuital law. First
use the definition of the vector Laplacian:
∇2 A = −∇ × ∇ × A + ∇(∇ · A) = −µJ
which is Eq. (55) with the time derivative set to zero. We also note that ∇ · A = 0 in steady
state (from Eq. (54)). Now, since B = ∇ × A, (55) becomes
−∇ × B = −µJ ⇒ ∇ × H = J
b) Show that Eq. (51) becomes Faraday’s law when taking the curl: Doing this gives
∇ × E = −∇ × ∇V −
∂
∇×A
∂t
The curl of the gradient is identially zero, and ∇ × A = B. We are left with
∇ × E = −∂B/∂t
10.22. In a sourceless medium, in which J = 0 and ρv = 0, assume a rectangular coordinate system in
which E and H are functions only of z and t. The medium has permittivity ϵ and permeability
µ. (a) If E = Ex ax and H = Hy ay , begin with Maxwell’s equations and determine the second
order partial differential equation that Ex must satisfy.
First use
∇×E=−
∂B
∂t
⇒
∂Hy
∂Ex
ay = −µ
ay
∂z
∂t
in which case, the curl has dictated the direction that H must lie in. Similarly, use the
other Maxwell curl equation to find
∇×H=
∂D
∂t
⇒ −
∂Ex
∂Hy
ax = ϵ
ax
∂z
∂t
Now, differentiate the first equation with respect to z, and the second equation with
respect to t:
∂ 2 Ex
∂ 2 Hy
∂ 2 Hy
∂ 2 Ex
=
−µ
and
=
−ϵ
∂z 2
∂t∂z
∂z∂t
∂t2
Combining these two, we find
∂ 2 Ex
∂ 2 Ex
=
µϵ
∂z 2
∂t2
b) Show that Ex = E0 cos(ωt − βz) is a solution of that equation for a particular value of β:
Substituting, we find
∂ 2 Ex
∂ 2 Ex
2
=
−β
E
cos(ωt
−
βz)
and
µϵ
= −ω 2 µϵE0 cos(ωt − βz)
0
∂z 2
∂t2
These two will be equal provided the constant multipliers of cos(ωt − βz) are equal.
c) Find β as a function of given parameters. Equating the two constants in part b, we find
√
β = ω µϵ.
10.23. In region 1, z < 0, ϵ1 = 2 × 10−11 F/m, µ1 = 2 × 10−6 H/m, and σ1 = 4 × 10−3 S/m; in
region 2, z > 0, ϵ2 = ϵ1 /2, µ2 = 2µ1 , and σ2 = σ1 /4. It is known that E1 = (30ax + 20ay +
10az ) cos(109t) V/m at P1 (0, 0, 0− ).
a) Find EN 1 , Et1 , DN 1 , and Dt1 : These will be
EN 1 = 10 cos(109t)az V/m
Et1 = (30ax + 20ay ) cos(109t) V/m
DN 1 = ϵ1 EN 1 = (2 × 10−11 )(10) cos(109t)az C/m2 = 200 cos(109t)az pC/m2
10.23a (continued).
Dt1 = ϵ1 Et1 = (2 × 10−11 )(30ax + 20ay ) cos(109t) = (600ax + 400ay ) cos(109t) pC/m2
b) Find JN 1 and Jt1 at P1 :
JN 1 = σ1 EN 1 = (4 × 10−3)(10 cos(109t))az = 40 cos(109t)az mA/m2
Jt1 = σ1 Et1 = (4 × 10−3)(30ax + 20ay ) cos(109t) = (120ax + 80ay ) cos(109t) mA/m2
c) Find Et2 , Dt2 , and Jt2 at P1 : By continuity of tangential E,
Et2 = Et1 = (30ax + 20ay ) cos(109t) V/m
Then
Dt2 = ϵ2 Et2 = (10−11 )(30ax + 20ay ) cos(109t) = (300ax + 200ay ) cos(109t) pC/m2
Jt2 = σ2 Et2 = (10−3)(30ax + 20ay ) cos(109t) = (30ax + 20ay ) cos(109t) mA/m2
d) (Harder) Use the continuity equation to help show that JN 1 − JN 2 = ∂DN 2 /∂t − ∂DN 1 /∂t
and then determine EN 2 , DN 2 , and JN 2 : We assume the existence of a surface charge layer
at the boundary having density ρs C/m2 . If we draw a cylindrical “pillbox” whose top and
bottom surfaces (each of area ∆a) are on either side of the interface, we may use the continuity
condition to write
(JN 2 − JN 1 )∆a = −
∂ρs
∆a
∂t
where ρs = DN 2 − DN 1 . Therefore,
JN 1 − JN 2 =
∂
(DN 2 − DN 1 )
∂t
In terms of the normal electric field components, this becomes
σ 1 E N 1 − σ 2 EN 2 =
∂
(ϵ2 EN 2 − ϵ1 EN 1 )
∂t
Now let EN 2 = A cos(109t) + B sin(109t), while from before, EN 1 = 10 cos(109t).
10.23d (continued)
These, along with the permittivities and conductivities, are substituted to obtain
(4 × 10−3)(10) cos(109t) − 10−3[A cos(109t) + B sin(109t)]
$
∂ # −11
=
10 [A cos(109t) + B sin(109t)] − (2 × 10−11 )(10) cos(109t)
∂t
= −(10−2 A sin(109t) + 10−2 B cos(109t) + (2 × 10−1 ) sin(109t)
We now equate coefficients of the sin and cos terms to obtain two equations:
4 × 10−2 − 10−3A = 10−2 B
−10−3B = −10−2 A + 2 × 10−1
These are solved together to find A = 20.2 and B = 2.0. Thus
Then
#
$
EN 2 = 20.2 cos(109t) + 2.0 sin(109t) az = 20.3 cos(109t + 5.6◦ )az V/m
DN 2 = ϵ2 EN 2 = 203 cos(109t + 5.6◦ )az pC/m2
and
JN 2 = σ2 EN 2 = 20.3 cos(109t + 5.6◦ )az mA/m2
10.24. In a medium in which ρv = 0, but in which the permittivity is a function of position, determine
the conditions on the permittivity variation such that
a) ∇ · E = 0: We first note that ∇ · D = 0 if ρv = 0, where D = ϵE. Now
∇ · D = ∇ · (ϵE) = E · ∇ϵ + ϵ∇ · E = 0
or
∇·E + E·
We see that ∇ · E = 0 if ∇ϵ = 0.
∇ϵ
=0
ϵ
.
b) ∇ · E = 0: From the development in part a, ∇ · E will be approximately zero if ∇ϵ/ϵ is
negligible.
10.25. In a region where µr = ϵr = 1 and σ = 0, the retarded potentials are given by V = x(z − ct)
√
V and A = x[(z/c) − t]az Wb/m, where c = 1/ µ0 ϵ0 .
a) Show that ∇ · A = −µϵ(∂V /∂t):
First,
∇·A=
∂Az
x
√
= = x µ0 ϵ0
∂z
c
Second,
∂V
x
= −cx = − √
∂t
µ0 ϵ0
so we observe that ∇·A = −µ0 ϵ0 (∂V /∂t) in free space, implying that the given statement
would hold true in general media.
b) Find B, H, E, and D:
Use
B=∇×A=−
Then
∂Ax
z.
ay = t −
ay T
∂x
c
1 z.
B
=
ay A/m
t−
H=
µ0
µ0
c
Now,
E = −∇V −
∂A
= −(z − ct)ax − xaz + xaz = (ct − z)ax V/m
∂t
Then
D = ϵ0 E = ϵ0 (ct − z)ax C/m2
c) Show that these results satisfy Maxwell’s equations if J and ρv are zero:
i. ∇ · D = ∇ · ϵ0 (ct − z)ax = 0
ii. ∇ · B = ∇ · (t − z/c)ay = 0
iii.
∂Hy
1
∇×H=−
ax =
ax =
∂z
µ0 c
which we require to equal ∂D/∂t:
∂D
= ϵ0 cax =
∂t
/
ϵ0
ax
µ0
/
ϵ0
ax
µ0
10.25c (continued).
iv.
∇×E=
∂Ex
ay = −ay
∂z
which we require to equal −∂B/∂t:
∂B
= ay
∂t
So all four Maxwell equations are satisfied.
10.26. Let the current I = 80t A be present in the az direction on the z axis in free space within the
interval −0.1 < z < 0.1 m.
a) Find Az at P (0, 2, 0): The integral for the retarded vector potential will in this case assume
the form
A=
where R =
√
80µ0
Az =
4π
"
.1
−.1
µ0 80(t − R/c)
az dz
4πR
z 2 + 4 and c = 3 × 108 m/s. We obtain
'"
.1
t
"
.1
(
1
0
.1
.1
1
8 × 10−6 1
1
1
z
dz = 8 × 10−6t ln(z + z 2 + 4)1 −
1
c
3 × 108 −.1
−.1
√
dz −
z2 + 4
−.1
3
2
√
.1 + 4.01
√
− 0.53 × 10−14 = 8.0 × 10−7t − 0.53 × 10−14
= 8 × 10−6ln
−.1 + 4.01
−.1
#
$
So finally, A = 8.0 × 10−7t − 5.3 × 10−15 az Wb/m.
b) Sketch Az versus t over the time interval −0.1 < t < 0.1 µs: The sketch is linearly increasing
with time, beginning with Az = −8.53 × 10−14 Wb/m at t = −0.1 µs, crossing the time axis
and going positive at t = 6.6 ns, and reaching a maximum value of 7.46 × 10−14 Wb/m at
t = 0.1 µs.
CHAPTER 11
11.1. Show that Exs = Aejk0 z+ is a solution to the vector Helmholtz equation, Sec. 11.1, Eq. (30),
✏
for k0 =
µ0 ⌅0 and any and A: We take
d2
Aejk0 z+ = (jk0 )2 Aejk0 z+ =
dz 2
k02 Exs
11.2. A 100-MHz uniform plane wave propagates in a lossless medium for which ⌅r = 5 and µr = 1.
Find:
✏
✏
a) vp : vp = c/ ⌅r = 3 ⇤ 108 / 5 = 1.34 ⇤ 108 m/s.
b) ⇥: ⇥ = /vp = (2 ⇤ 108 )/(1.34 ⇤ 108 ) = 4.69 m
1
.
c) ⌥: ⌥ = 2 /⇥ = 1.34 m.
d) Es : Assume real amplitude E0 , forward z travel, and x polarization, and write
Es = E0 exp( j⇥z)ax = E0 exp( j4.69z) ax V/m.
✏
✏
e) Hs : First, the intrinsic impedance of the medium is ⇧ = ⇧0 / ⌅r = 377/ 5 = 169 ⇥.
Then Hs = (E0 /⇧) exp( j⇥z) ay = (E0 /169) exp( j4.69z) ay A/m.
f) < S >= (1/2)Re {Es ⇤ H⇥s } = (E02 /337) az W/m2
12.3. An H field in free space is given as H(x, t) = 10 cos(108 t ⇥x)ay A/m. Find
a) ⇥: Since we have a uniform plane wave, ⇥ = /c, where we identify = 108 sec
⇥ = 108 /(3 ⇤ 108 ) = 0.33 rad/m.
1
. Thus
b) ⌥: We know ⌥ = 2 /⇥ = 18.9 m.
c) E(x, t) at P (0.1, 0.2, 0.3) at t = 1 ns: Use E(x, t) = ⇧0 H(x, t) = (377)(10) cos(108 t
⇥x) = 3.77 ⇤ 103 cos(108 t ⇥x). The vector direction of E will be az , since we require
that S = E ⇤ H, where S is x-directed. At the given point, the relevant coordinate is
x = 0.1. Using this, along with t = 10 9 sec, we finally obtain
E(x, t) =
3.77 ⇤ 103 cos[(108 )(10
=
9
)
(0.33)(0.1)]az =
3.76 ⇤ 10 az V/m
3
3.77 ⇤ 103 cos(6.7 ⇤ 10
2
)az
11.4. Small antennas have low e⌃ciencies (as will be seen in Chapter 14) and the e⌃ciency increases
with size up to the point at which a critical dimension of the antenna is an appreciable fraction
of a wavelength, say ⌥/8.
a) An antenna is that is 12cm long is operated in air at 1 MHz. What fraction of a wavelength
long is it? The free space wavelength will be
⌥air =
c
3.0 ⇤ 108 m/s
1.2
=
= 300 m, so that the fraction =
= 4.0 ⇤ 10
6
1
f
10 s
300
3
b) The same antenna is embedded in a ferrite material for which ⌅r = 20 and µr = 2, 000.
What fraction of a wavelength is it now?
⌥air
300
1.2
⌥f errite = ✏
=⌧
= 1.5m ⌃ fraction =
= 0.8
µr ⌅r
1.5
(20)(2000)
221
11.5. A 150-MHz uniform plane wave in free space is described by Hs = (4 + j10)(2ax + jay )e
A/m.
j⇥z
a) Find numerical values for , ⌥, and ⇥: First, = 2 ⇤150⇤106 = 3 ⇤ 108 sec 1 . Second,
for a uniform plane wave in free space, ⌥ = 2 c/ = c/f = (3 ⇤ 108 )/(1.5 ⇤ 108 ) = 2 m.
Third, ⇥ = 2 /⌥ = rad/m.
b) Find H(z, t) at t = 1.5 ns, z = 20 cm: Use
H(z, t) = Re{Hs ej t } = Re{(4 + j10)(2ax + jay )(cos( t ⇥z) + j sin( t ⇥z)}
= [8 cos( t ⇥z) 20 sin( t ⇥z)] ax [10 cos( t ⇥z) + 4 sin( t ⇥z)] ay
8
. Now at the given position and
✏ time, t ⇥z = (3 ⇤ 10 )(1.5 ⇤ 10
And cos( /4) = sin( /4) = 1/ 2. So finally,
H(z = 20cm, t = 1.5ns) =
1
✏ (12ax + 14ay ) =
2
8.5ax
c) What is |E|max ? Have |E|max = ⇧0 |H|max , where
⌧
|H|max = Hs · H⇥s = [4(4 + j10)(4 j10) + (j)( j)(4 + j10)(4
9
)
(0.20) = /4.
9.9ay A/m
1/2
j10)]
= 24.1 A/m
Then |E|max = 377(24.1) = 9.08 kV/m.
11.6. A uniform plane wave has electric field Es = (Ey0 ay Ez0 az ) e x e j⇥x V/m. The intrinsic
impedance of the medium is given as ⇧ = |⇧| ej , where is a constant phase.
a) Describe the wave polarization and state the direction of propagation: The wave is linearly
polarized in the y-z plane, and propagates in the forward x direction (from the e j⇥x
factor).
b) Find Hs : Each component of Es , when crossed into its companion component of H⇥s ,
must give a vector in the positive-x direction of travel. Using this rule, we find
↵
↵
Ey
Ez
Ey0
Ez0
Hs =
az +
ay =
az +
ay e x e j e j⇥x A/m
⇧
⇧
|⇧|
|⇧|
⇧
⌃
c) Find E(x, t) and H(x, t): E(x, t) = Re Es ej t = [Ey0 ay
⇧
⌃
H(x, t) = Re Hs ej t = [Ey0 az + Ez0 ay ] e
Ez0 az ] e
x
cos( t
x
cos( t
⇥x
⇥x)
)
where all amplitudes are assumed real.
d) Find < S > in W/m2 :
< S >=
⇥
1
1
2
2
Re {Es ⇤ H⇥s } =
Ey0
+ Ez0
e
2
2
2 x
cos ax W/m2
e) Find the time-average power in watts that is intercepted by an antenna of rectangular
cross-section, having width w and height h, suspended parallel to the yz plane, and at a
distance d from the wave source. This will be
⌫ ⌫
⇥ 2 d
1
2
2
P =
< S > · dS = | < S > |x=d ⇤ area = (wh) Ey0
+ Ez0
e
cos W
2
plate
222
11.7. The phasor magnetic field intensity for a 400-MHz uniform plane wave propagating in a
certain lossless material is (2ay j5az )e j25x A/m. Knowing that the maximum amplitude
of E is 1500 V/m, find ⇥, ⇧, ⌥, vp , ⌅r , µr , and H(x, y, z, t): First, from the phasor expression,
we identify ⇥✏= 25 m 1 ✏
from the argument
of the exponential function.
✏
✏ Next, we evaluate
⇥
2
2
H0 = |H| = H · H = 2 + 5 = 29. Then ⇧ = E0 /H0 = 1500/ 29 = 278.5 ⇥. Then
⌥ = 2 /⇥ = 2 /25 = .25 m = 25 cm. Next,
vp =
⇥
=
2 ⇤ 400 ⇤ 106
= 1.01 ⇤ 108 m/s
25
Now we note that
⇧ = 278.5 = 377
And
µr
⌅r
⌃
c
vp = 1.01 ⇤ 108 = ✏
µr ⌅r
µr
= 0.546
⌅r
⌃ µr ⌅r = 8.79
We solve the above two equations simultaneously to find ⌅r = 4.01 and µr = 2.19. Finally,
⇧
H(x, y, z, t) = Re (2ay
j5az )e
j25x j t
e
= 2 cos(2 ⇤ 400 ⇤ 106 t
= 2 cos(8 ⇤ 108 t
⌃
25x)ay + 5 sin(2 ⇤ 400 ⇤ 106 t
25x)ay + 5 sin(8 ⇤ 108 t
25x)az
25x)az A/m
11.8. An electric field in free space is given in spherical coordinates as Es (r) = E0 (r)e
jkr
a⇧ V/m.
a) find Hs (r) assuming uniform plane wave behavior: Knowing that the cross product of Es
with the complex conjugate of the phasor Hs field must give a vector in the direction of
propagation, we obtain,
E0 (r) jkr
Hs (r) =
e
a A/m
⇧0
b) Find < S >: This will be
< S >=
1
E 2 (r)
Re {Es ⇤ H⇥s } = 0
ar W/m2
2
2⇧0
c) Express the average outward power in watts through a closed spherical shell of radius r,
centered at the origin: The power will be (in this case) just the product of the power
density magnitude in part b with the sphere area, or
P = 4 r2
E02 (r)
W
2⇧0
where E0 (r) is assumed real.
d) Establish the required functional form of E0 (r) that will enable the power flow in part c
to be independent of radius: Evidently this condition is met when E0 (r) ⌥ 1/r
223
11.9. A certain lossless material has µr = 4 and ⌅r = 9. A 10-MHz uniform plane wave is propagating
in the ay direction with Ex0 = 400 V/m and Ey0 = Ez0 = 0 at P (0.6, 0.6, 0.6) at t = 60 ns.
a) Find ⇥, ⌥, vp , and ⇧: For a uniform plane wave,
✏
µ⌅ =
⇥=
2 ⇤ 107 ⌧
✏
µr ⌅r =
(4)(9) = 0.4 rad/m
c
3 ⇤ 108
Then ⌥ = (2 )/⇥ = (2 )/(0.4 ) = 5 m. Next,
vp =
⇥
=
2 ⇤ 107
= 5 ⇤ 107 m/s
4 ⇤ 10 1
Finally,
µ
= ⇧0
⌅
⇧=
4
= 251 ⇥
9
µr
= 377
⌅r
b) Find E(t) (at P ): We are given the amplitude at t = 60 ns and at y = 0.6 m. Let the
maximum amplitude be Emax , so that in general, Ex = Emax cos( t ⇥y). At the given
position and time,
Ex = 400 = Emax cos[(2 ⇤ 107 )(60 ⇤ 10
= 0.99Emax
So Emax = (400)/( 0.99) =
9
)
(4 ⇤ 10
1
403 V/m. Thus at P, E(t) =
)(0.6)] = Emax cos(0.96 )
403 cos(2 ⇤ 107 t) V/m.
c) Find H(t): First, we note that if E at a given instant points in the negative x direction,
while the wave propagates in the forward y direction, then H at that same position and
time must point in the positive z direction. Since we have a lossless homogeneous medium,
⇧ is real, and we are allowed to write H(t) = E(t)/⇧, where ⇧ is treated as negative and
real. Thus
H(t) = Hz (t) =
Ex (t)
=
⇧
403
cos(2 ⇤ 10
251
7
t) = 1.61 cos(2 ⇤ 10
7
t) A/m
11.10. In a medium characterized by intrinsic impedance ⇧ = |⇧|ej , a linearly-polarized plane wave
propagates, with magnetic field given as Hs = (H0y ay + H0z az ) e x e j⇥x . Find:
a) Es : Requiring orthogonal components of Es for each component of Hs , we find
Es = |⇧| [H0z ay
b) E(x, t) = Re {Es ej t } = |⇧| [H0z ay
c) H(x, t) = Re {Hs e
j t
d) < S >=
H0y az ] e
H0y az ] e
} = [H0y ay + H0z az ] e
x
x
x
j⇥x j
e
cos( t
cos( t
⌅
1
1 ⇤ 2
2
Re{Es ⇤ H⇥s } = |⇧| H0y
+ H0z
e
2
2
224
e
⇥x + ).
⇥x).
2 x
cos ax W/m2
11.11. A 2-GHz uniform plane wave has an amplitude of Ey0 = 1.4 kV/m at (0, 0, 0, t = 0) and is
propagating in the az direction in a medium where ⌅⌅⌅ = 1.6 ⇤ 10 11 F/m, ⌅⌅ = 3.0 ⇤ 10 11
F/m, and µ = 2.5 µH/m. Find:
a) Ey at P (0, 0, 1.8cm) at 0.2 ns: To begin, we have the ratio, ⌅⌅⌅ /⌅⌅ = 1.6/3.0 = 0.533. So
=
µ⌅⌅
2
✏
⌅⌅⌅
⌘ 1+
⌅⌅
⇣1/2
⌦2
1✓
(2.5 ⇤ 10
= (2 ⇤ 2 ⇤ 109 )
Then
⇥=
Thus in general,
µ⌅⌅
2
6 )(3.0
2
⇤ 10
✏
⌅⌅⌅
⌘ 1+
Ey (z, t) = 1.4e
28.1z
⌅⌅
⇠⌧
1 + (.533)2
11 )
⌦2
⇣1/2
+ 1✓
⇡1/2
1
= 28.1 Np/m
= 112 rad/m
cos(4 ⇤ 109 t
112z) kV/m
Evaluating this at t = 0.2 ns and z = 1.8 cm, find
Ey (1.8 cm, 0.2 ns) = 0.74 kV/m
b) Hx at P at 0.2 ns: We use the phasor relation, Hxs =
⇧=
So now
µ
⌧
⌅⌅ 1
Hxs =
Then
1
=
j(⌅⌅⌅ /⌅⌅ )
Eys
=
⇧
2.5 ⇤ 10
3.0 ⇤ 10
6
11
(1.4 ⇤ 103 )e 28.1z e
271ej14
Hx (z, t) =
5.16e
28.1z
⌧
1
j112z
Eys /⇧ where
1
= 263 + j65.7 = 271⇧ 14⇤ ⇥
j(.533)
=
5.16e
cos(4 ⇤ 10
9
t
28.1z
112z
This, when evaluated at t = 0.2 ns and z = 1.8 cm, yields
Hx (1.8 cm, 0.2 ns) =
225
3.0 A/m
e
j112z
14⇤ )
e
j14
A/m
11.12. Describe how the attenuation coe⌃cient of a liquid medium, assumed to be a good conductor,
could be determined through measurement of wavelength in the liquid at a known frequency.
What restrictions apply? Could this method be used to find the conductivity as well? In a
good conductor, we may use the approximation:
µ↵
2
where ⇥ =
2
⌥
.
.
=⇥=
Therefore, in the good conductor approximation,
could also find
. 4
↵= 2
⌥ fµ
.
= 2 /⌥. From the above formula, we
which would work provided that again, we are certain that we have a good conductor, and
that the permeability is known.
11.13. Let jk = 0.2 + j1.5 m 1 and ⇧ = 450 + j60 ⇥ for a uniform plane wave propagating in the az
direction. If = 300 Mrad/s, find µ, ⌅⌅ , and ⌅⌅⌅ : We begin with
µ
⌧
⌅⌅ 1
⇧=
and
⇧⇧ ⇥ =
and
j(⌅⌅⌅ /⌅⌅ ) = 0.2 + j1.5
µ
1
⌧
= (450 + j60)(450
⌅
⌅
1 + (⌅⌅⌅ /⌅⌅ )2
(jk)(jk)⇥ =
2
µ⌅⌅
Taking the ratio of (2) to (1),
(jk)(jk)⇥
=
⇧⇧ ⇥
Then with
⌧ ⌧
µ⌅⌅ 1
jk = j
Then
1
= 450 + j60
j(⌅⌅⌅ /⌅⌅ )
= 3 ⇤ 108 ,
(⌅⌅ )2 =
2
j60) = 2.06 ⇤ 105
⌧
1 + (⌅⌅⌅ /⌅⌅ )2 = (0.2 + j1.5)(0.2
⇥
(⌅⌅ )2 1 + (⌅⌅⌅ /⌅⌅ )2 =
j1.5) = 2.29
2.29
= 1.11 ⇤ 10
2.06 ⇤ 105
1.11 ⇤ 10 5
1.23 ⇤ 10 22
=
(3 ⇤ 108 )2 (1 + (⌅⌅⌅ /⌅⌅ )2 )
(1 + (⌅⌅⌅ /⌅⌅ )2 )
(1)
(2)
5
(3)
Now, we use Eqs. (35) and (36). Squaring these and taking their ratio gives
⌧
1 + (⌅⌅⌅ /⌅⌅ )2
(0.2)2
=⌧
=
2
⇥
(1.5)2
1 + (⌅⌅⌅ /⌅⌅ )2
2
We solve this to find ⌅⌅⌅ /⌅⌅ = 0.271. Substituting this result into (3) gives ⌅⌅ = 1.07 ⇤ 10 11
F/m. Since ⌅⌅⌅ /⌅⌅ = 0.271, we then find ⌅⌅⌅ = 2.90 ⇤ 10 12 F/m. Finally, using these results in
either (1) or (2) we find µ = 2.28 ⇤ 10 6 H/m. Summary: µ = 2.28 ⇤ 10 6 H/m,
⌅⌅ = 1.07 ⇤ 10 11 F/m, and ⌅⌅⌅ = 2.90 ⇤ 10 12 F/m.
226
11.14. A certain nonmagnetic material has the material constants ⌅⌅r = 2 and ⌅⌅⌅ /⌅⌅ = 4 ⇤ 10 4 at
= 1.5 Grad/s. Find the distance a uniform plane wave can propagate through the material
before:
a) it is attenuated by 1 Np: First, ⌅⌅⌅ = (4 ⇤ 104 )(2)(8.854 ⇤ 10 12 ) = 7.1 ⇤ 10 15 F/m.
Then, since ⌅⌅⌅ /⌅⌅ << 1, we use the approximate form for , given by Eq. (51) (written
in terms of ⌅⌅⌅ ):
.
=
⌅⌅⌅
2
µ
(1.5 ⇤ 109 )(7.1 ⇤ 10
=
⌅⌅
2
15
) 377
✏ = 1.42 ⇤ 10
2
3
Np/m
b) the power level is reduced by one-half: The governing relation is e
z1/2 = ln 2/2 = ln 2/2(1.42 ⇤ 10 3) = 244 m.
2 z1/2
The required distance is now z1 = (1.42 ⇤ 10
3
)
1
= 706 m
= 1/2, or
⇤
c) the phase shifts
is defined
✏as one wavelength, where ⌥ = 2 /⇥
⌧ 360 : This distance
= (2 c)/(
⌅⌅r ) = [2 (3 ⇤ 108 )]/[(1.5 ⇤ 109 ) 2] = 0.89 m.
11.15. A 10 GHz radar signal may be represented as a uniform plane wave in a su⌃ciently small
region. Calculate the wavelength in centimeters and the attenuation in nepers per meter if
the wave is propagating in a non-magnetic material for which
a) ⌅⌅r = 1 and ⌅⌅⌅r = 0: In a non-magnetic material, we would have:
✏
⇣1/2
⌦2
µ0 ⌅0 ⌅⌅r ⌘
⌅⌅⌅r
=
1+
1✓
2
⌅⌅r
and
⇥=
✏
µ0 ⌅0 ⌅⌅r ⌘
1+
2
⌅⌅⌅r
⌅⌅r
⌦2
⇣1/2
+ 1✓
With the given values of ⌅⌅r and ⌅⌅⌅r , it is clear that ⇥ =
✏
µ0 ⌅0 =
/c, and so
⌥ = 2 /⇥ = 2 c/ = 3 ⇤ 1010 /1010 = 3 cm. It is also clear that
= 0.
. ⌧ ⌅
b) ⌅⌅r = 1.04 and ⌅⌅⌅r = 9.00⇤10 4 : In this case ⌅⌅⌅r /⌅⌅r << 1, and so ⇥ =
⌅r /c = 2.13 cm
Thus ⌥ = 2 /⇥ = 2.95 cm. Then
✏
⌅⌅⌅r µ0 ⌅0
⌅⌅⌅r
2 ⇤ 1010 (9.00 ⇤ 10 4 )
. ⌅⌅⌅ µ
⌧
⌧
✏
=
=
=
=
2
⌅⌅
2
2c ⌅⌅r
2 ⇤ 3 ⇤ 108
⌅⌅r
1.04
= 9.24 ⇤ 10
2
Np/m
c) ⌅⌅r = 2.5 and ⌅⌅⌅r = 7.2: Using the above formulas, we obtain
✏
⇣1/2
✏
⌦2
2 ⇤ 1010 2.5 ⌘
7.2
✏
⇥=
1+
+ 1✓
= 4.71 cm
2.5
(3 ⇤ 1010 ) 2
1
and so ⌥ = 2 /⇥ = 1.33 cm. Then
✏
✏
2 ⇤ 10
2.5
✏ ⌘ 1+
=
8
(3 ⇤ 10 ) 2
10
227
7.2
2.5
⌦2
⇣1/2
1✓
= 335 Np/m
1
.
11.16. Consider the power dissipation term, E · Jdv in Poynting’s theorem (Eq. (70)). This gives
the power lost to heat within a volume into which electromagnetic waves enter. The term
pd = E · J is thus the power dissipation per unit volume in W/m3 . Following the same
reasoning that resulted in Eq. (77), the time-average power dissipation per volume will be
< pd >= (1/2)Re {Es · J⇥s }.
a) Show that in a conducting medium, through which a uniform plane wave of amplitude E0
propagates in the forward z direction, < pd >= (↵/2)|E0 |2 e 2 z : Begin with the phasor
expression for the electric field, assuming complex amplitude E0 , and x-polarization:
Es = E0 e
Then
z
e
Js = ↵Es = ↵E0 e
j⇥z
z
e
ax V/m2
j⇥z
ax A/m2
So that
⇧
< pd >= (1/2)Re E0 e
z
e
j⇥z
ax · ↵E0⇥ e
z +j⇥z
e
⌃
ax = (↵/2)|E0 |2 e
2 z
b) Confirm this result for the special case of a good conductor by using the left hand side of
Eq. (70), and consider a very small volume. In a good conductor, the intrinsic impedance
is, from Eq. (85), ⇧c = (1 + j)/(↵⇤), where the skin depth, ⇤ = 1/ . The magnetic field
phasor is then
Es
↵
Hs =
ay =
E0 e z e j⇥z ay A/m
⇧c
(1 + j)
The time-average Poynting vector is then
< S >=
1
Re {Es ⇤ H⇥s } =
|E0 |2 e
2
4↵
2 z
az W/m2
Now, consider a rectangular volume of side lengths, x, y, and z, all of which are
very small. As the wave passes through this volume in the forward z direction, the power
dissipated will be the di⇤erence between the power at entry (at z = 0), and the power
.
that exits the volume (at z = z). With small z, we may approximate e 2 z = 1 2 z,
and the dissipated power in the volume becomes
Pd = Pin
Pout =
⇠
4↵
|E0 |2
⇡
x y
⇠
4↵
|E0 |2 (1
2
⇡
z) x y =
2↵
|E0 |2 ( x y z)
This is just the result of part a, evaluated at z = 0 and multiplied by the volume. The
relation becomes exact as z ⇧ 0, in which case < pd >⇧ (↵/2)|E0 |2 .
It is also possible to show the relation by using Eq. (69) (which involves taking the
divergence of < S >), or by removing the restriction of a small volume and evaluating
the integrals in Eq. (70) without approximations. Either method is straightforward.
228
11.17. Let ⇧ = 250 + j30 ⇥ and jk = 0.2 + j2 m 1 for a uniform plane wave propagating in the az
direction in a dielectric having some finite conductivity. If |Es | = 400 V/m at z = 0, find:
a) < S > at z = 0 and z = 60 cm: Assume x-polarization for the electric field. Then
1
1
400
Re {Es ⇤ H⇥s } = Re 400e z e j⇥z ax ⇤ ⇥ e z ej⇥z ay
2
2
⇧
1
1
1
= (400)2 e 2 z Re
az = 8.0 ⇤ 104 e 2(0.2)z Re
⇥
2
⇧
250 j30
<S>=
= 315 e
2(0.2)z
az
az W/m2
Evaluating at z = 0, obtain < S > (z = 0) = 315 az W/m2 ,
and at z = 60 cm, Pz,av (z = 0.6) = 315e 2(0.2)(0.6) az = 248 az W/m2 .
b) the average ohmic power dissipation in watts per cubic meter at z = 60 cm: At this point
a flaw becomes evident in the problem statement, since solving this part in two di⇤erent
ways gives results that are not the same. I will demonstrate: In the first method, we
use Poynting’s theorem in point form Eq.(69), which we modify for the case of timeindependent fields to read: ⇣· < S >=< J · E >, where the right hand side is the
average power dissipation per volume. Note that the additional right-hand-side terms in
Poynting’s theorem that describe changes in energy stored in the fields will both be zero
in steady state. We apply our equation to the result of part a:
< J · E >=
⇣· < S >=
d
315 e
dz
2(0.2)z
= (0.4)(315)e
2(0.2)z
= 126e
0.4z
W/m3
At z = 60 cm, this becomes < J · E >= 99.1 W/m3 . In the second method,
for
✏ ⌅ ⌧we solve
2
⌅⌅
the conductivity and evaluate < J · E >= ↵ < E >. We use jk = j
µ⌅ 1 j(⌅ /⌅⌅ )
and
µ
1
⌧
⇧=
⌅
⌅
1 j(⌅⌅⌅ /⌅⌅ )
We take the ratio,
↵
jk
⌅
=j ⌅ 1
⇧
j
⌅⌅⌅
⌅⌅
Identifying ↵ = ⌅⌅⌅ , we find
↵ = Re
jk
⇧
= Re
⌦
= j ⌅⌅ + ⌅⌅⌅
0.2 + j2
250 + j30
= 1.74 ⇤ 10
3
S/m
Now we find the dissipated power per volume:
↵ < E >= 1.74 ⇤ 10
2
3
1
2
⌦
400e
⇥
0.2z 2
At z = 60 cm, this evaluates as 109 W/m3 . One can show that consistency between the
two methods requires that
1
↵
Re
=
⇥
⇧
2
This relation does not hold using the numbers as given in the problem statement and the
value of ↵ found above. Note that in Problem 11.13, where all values are worked out, the
relation does hold and consistent results are obtained using both methods.
229
11.18. Given, a 100MHz uniform plane wave in a medium known to be a good dielectric. The phasor
electric field is Es = 4e 0.5z e j20z ax V/m. Not stated in the problem is the permeability,
which we take to be µ0 . Determine:
a) ⌅⌅ : As a first step, it is useful to see just how much of a good dielectric we have. We use
the good dielectric approximations, Eqs. (60a) and (60b), with ↵ = ⌅⌅⌅ . Using these, we
take the ratio, ⇥/ , to find
⇥
20
=
=
0.5
⌅
✏ ⌅⇤
µ⌅ 1 + (1/8)(⌅⌅⌅ /⌅⌅ )2
⌧
=2
( ⌅⌅⌅ /2) µ/⌅⌅
⌅⌅
⌅⌅⌅
⌦
+
1
4
⌅⌅⌅
⌅⌅
⌦
This becomes the quadratic equation:
⌅⌅⌅
⌅⌅
⌦2
⌅⌅⌅
⌅⌅
160
⌦
+8=0
The solution to the quadratic is (⌅⌅⌅ /⌅⌅ ) ✏
= 0.05, which
⌧ means that we can neglect the
.
second term in Eq. (60b), so that ⇥ =
µ⌅⌅ = ( /c) ⌅⌅r . With the given frequency of
⌧
100 MHz, and with µ = µ0 , we find ⌅⌅r = 20(3/2 ) = 9.55, so that ⌅⌅r = 91.3, and finally
⌅⌅ = ⌅⌅r ⌅0 = 8.1 ⇤ 10 10 F/m.
b) ⌅⌅⌅ : Using Eq. (60a), the set up is
= 0.5 =
⌅⌅⌅
2
µ
⌅⌅
⌃ ⌅⌅⌅ =
⌅⌅
10 8 ✏
=
91.3 = 4.0 ⇤ 10
µ
2 (377)
2(0.5)
2 ⇤ 108
11
F/m
c) ⇧: Using Eq. (62b), we find
.
⇧=
↵
µ
1
1+j
⌅
⌅
2
⌅⌅⌅
⌅⌅
⌦
377
=✏
(1 + j.025) = (39.5 + j0.99) ohms
91.3
d) Hs : This will be a y-directed field, and will be
Hs =
Es
4
ay =
e
⇧
(39.5 + j0.99)
0.5z
e
j20z
ay = 0.101e
0.5z
e
j20z
e
j0.025
ay A/m
e) < S >: Using the given field and the result of part d, obtain
< S >=
1
(0.101)(4)
Re{Es ⇤ H⇥s } =
e
2
2
2(0.5)z
cos(0.025) az = 0.202e
z
az W/m2
f) the power in watts that is incident on a rectangular surface measuring 20m x 30m at
z = 10m: At 10m, the power density is < S >= 0.202e 10 = 9.2 ⇤ 10 6 W/m2 . The
incident power on the given area is then P = 9.2 ⇤ 10 6 ⇤ (20)(30) = 5.5 mW.
230
11.19. Perfectly-conducting cylinders with radii of 8 mm and 20 mm are coaxial. The region between
the cylinders is filled with a perfect dielectric for which ⌅ = 10 9 /4 F/m and µr = 1. If E in
this region is (500/⌦) cos( t 4z)a⌥ V/m, find:
a)
, with the help of Maxwell’s equations in cylindrical coordinates: We use the two curl
equations, beginning with ⇣ ⇤ E = ◆B/◆t, where in this case,
◆E⌥
2000
a =
sin( t
◆z
⌦
⇣⇤E=
So
B =
Then
⌫
H =
2000
sin( t
⌦
4z)dt =
4z)a =
◆B
a
◆t
2000
cos( t
⌦
B
2000
=
cos( t
µ0
(4 ⇤ 10 7 ) ⌦
4z) T
4z) A/m
We next use ⇣ ⇤ H = ◆D/◆t, where in this case
⇣⇤H=
◆H
1 ◆(⌦H )
a⌥ +
az
◆z
⌦ ◆⌦
where the second term on the right hand side becomes zero when substituting our H .
So
◆H
8000
◆D⌥
⇣⇤H=
a⌥ =
sin( t 4z)a⌥ =
a⌥
◆z
(4 ⇤ 10 7 ) ⌦
◆t
And
D⌥ =
⌫
8000
sin( t
(4 ⇤ 10 7 ) ⌦
4z)dt =
8000
(4 ⇤ 10 7 )
2⌦
cos( t
4z) C/m2
Finally, using the given ⌅,
E⌥ =
D⌥
8000
=
cos( t
⌅
(10 16 ) 2 ⌦
4z) V/m
This must be the same as the given field, so we require
8000
500
=
16
2
(10 ) ⌦
⌦
= 4 ⇤ 108 rad/s
⌃
b) H(⌦, z, t): From part a, we have
H(⌦, z, t) =
2000
cos( t
(4 ⇤ 10 7 ) ⌦
4z)a =
4.0
cos(4 ⇤ 108 t
⌦
4z)a A/m
c) S(⌦, , z): This will be
S(⌦, , z) = E ⇤ H =
=
2.0 ⇤ 10
⌦2
500
cos(4 ⇤ 108 t
⌦
3
cos2 (4 ⇤ 108 t
231
4z)a⌥ ⇤
4.0
cos(4 ⇤ 108 t
⌦
4z)az W/m2
4z)a
11.19d) the average power passing through every cross-section 8 < ⌦ < 20 mm, 0 < < 2 . Using
the result of part c, we find < S >= (1.0 ⇤ 103 )/⌦2 az W/m2 . The power through the given
cross-section is now
⌦
⌫ 2⌃ ⌫ .020
1.0 ⇤ 103
20
3
P=
⌦ d⌦ d = 2 ⇤ 10 ln
= 5.7 kW
⌦2
8
0
.008
11.20. Voltage breakdown in air at standard temperature and pressure occurs at an electric field
strength of approximately 3 ⇤ 106 V/m. This becomes an issue in some high-power optical
experiments, in which tight focusing of light may be necessary. Estimate the lightwave power
in watts that can be focused into a cylindrical beam of 10µm radius before breakdown occurs.
Assume uniform plane wave behavior (although this assumption will produce an answer that
is higher than the actual number by as much as a factor of 2, depending on the actual beam
shape).
The power density in the beam in free space can be found as a special case of Eq. (76)
(with ⇧ = ⇧0 , ⌃⌅ = = 0):
|<S>|=
E02
(3 ⇤ 106 )2
=
= 1.2 ⇤ 1010 W/m2
2⇧0
2(377)
To avoid breakdown, the power in a 10-µm radius cylinder is then bounded by
P < (1.2 ⇤ 1010 )( ⇤ (10
232
) ) = 3.75 W
5 2
11.21. The cylindrical shell, 1 cm < ⌦ < 1.2 cm, is composed of a conducting material for which
↵ = 106 S/m. The external and internal regions are non-conducting. Let H = 2000 A/m at
⌦ = 1.2 cm.
a) Find H everywhere: Use Ampere’s circuital law, which states:
H · dL = 2 ⌦(2000) = 2 (1.2 ⇤ 10
2
)(2000) = 48 A = Iencl
Then in this case
48
1.00) ⇤ 10
I
az =
Area
(1.44
J=
4
az = 1.09 ⇤ 106 az A/m2
With this result we again use Ampere’s circuital law to find H everywhere within the
shell as a function of ⌦ (in meters):
H 1 (⌦) =
1
2 ⌦
⌫
2⌃
0
⌫
⌥
.01
1.09 ⇤ 106 ⌦ d⌦ d =
54.5
(104 ⌦2
⌦
1) A/m (.01 < ⌦ < .012)
Outside the shell, we would have
H 2 (⌦) =
48
= 24/⌦ A/m (⌦ > .012)
2 ⌦
Inside the shell (⌦ < .01 m), H = 0 since there is no enclosed current.
b) Find E everywhere: We use
E=
J
1.09 ⇤ 106
=
az = 1.09 az V/m
↵
106
which is valid, presumeably, outside as well as inside the shell.
c) Find S everywhere: Use
P = E ⇤ H = 1.09 az ⇤
=
59.4
(104 ⌦2
⌦
54.5
(104 ⌦2
⌦
1) a
1) a⌥ W/m2 (.01 < ⌦ < .012 m)
Outside the shell,
S = 1.09 az ⇤
24
a =
⌦
233
26
a⌥ W/m2 (⌦ > .012 m)
⌦
11.22. The inner and outer dimensions of a copper coaxial transmission line are 2 and 7 mm, respectively. Both conductors have thicknesses much greater than ⇤. The dielectric is lossless and
the operating frequency is 400 MHz. Calculate the resistance per meter length of the:
a) inner conductor: First
⇤=✏
1
1
=⌧
f µ↵
(4 ⇤ 108 )(4 ⇤ 10
7 )(5.8
⇤ 107 )
= 3.3 ⇤ 10
6
m = 3.3µm
Now, using (90) with a unit length, we find
Rin =
1
=
2 a↵⇤
2 (2 ⇤ 10
3 )(5.8
1
⇤ 107 )(3.3 ⇤ 10
6)
= 0.42 ohms/m
b) outer conductor: Again, (90) applies but with a di⇤erent conductor radius. Thus
Rout =
a
2
Rin = (0.42) = 0.12 ohms/m
b
7
c) transmission line: Since the two resistances found above are in series, the line resistance
is their sum, or R = Rin + Rout = 0.54 ohms/m.
11.23. A hollow tubular conductor is constructed from a type of brass having a conductivity of
1.2 ⇤ 107 S/m. The inner and outer radii are 9 mm and 10 mm respectively. Calculate the
resistance per meter length at a frequency of
a) dc: In this case the current density is uniform over the entire tube cross-section. We
write:
L
1
R(dc) =
=
= 1.4 ⇤ 10 3 ⇥/m
7
↵A
(1.2 ⇤ 10 ) (.012 .0092 )
b) 20 MHz: Now the skin e⇤ect will limit the e⇤ective cross-section. At 20 MHz, the skin
depth is
⇤(20MHz) = [ f µ0 ↵]
1/2
= [ (20 ⇤ 106 )(4 ⇤ 10
7
)(1.2 ⇤ 107 )]
1/2
= 3.25 ⇤ 10
5
m
This is much less than the outer radius of the tube. Therefore we can approximate the
resistance using the formula:
R(20MHz) =
L
1
1
=
=
7
↵A
2 b⇤
(1.2 ⇤ 10 )(2 (.01))(3.25 ⇤ 10
5)
= 4.1 ⇤ 10
2
⇥/m
c) 2 GHz: Using the same formula as in part b, we find the skin depth at 2 GHz to be ⇤ =
3.25 ⇤ 10 6 m. The resistance (using the other formula) is R(2GHz) = 4.1 ⇤ 10 1 ⇥/m.
234
11.24 a) Most microwave ovens operate at 2.45 GHz. Assume that ↵ = 1.2 ⇤ 106 S/m and µr = 500
for the stainless steel interior, and find the depth of penetration:
⇤=✏
1
1
=⌧
9
f µ↵
(2.45 ⇤ 10 )(4 ⇤ 10
7 )(1.2
⇤
106 )
= 9.28 ⇤ 10
6
m = 9.28µm
b) Let Es = 50⇧ 0⇤ V/m at the surface of the conductor, and plot a curve of the amplitude
of Es vs. the angle of Es as the field propagates
into the stainless steel: Since the
.
. ✏
conductivity is high, we use (82) to write = ⇥ =
f µ↵ = 1/⇤. So, assuming that the
direction into the conductor is z, the depth-dependent field is written as
Es (z) = 50e
z
e
j⇥z
= 50e
z/⇤
e
jz/⇤
= 50 exp( z/9.28) exp( j z/9.28)
!
"
!
"
amplitude
angle
where z is in microns. Therefore, the plot of amplitude versus angle is simply a plot of
e x versus x, where x = z/9.28; the starting amplitude is 50 and the 1/e amplitude (at
z = 9.28 µm) is 18.4.
11.25. A good conductor is planar in form and carries a uniform plane wave that has a wavelength
of 0.3 mm and a velocity of 3 ⇤ 105 m/s. Assuming the conductor is non-magnetic, determine
the frequency and the conductivity: First, we use
f=
v
3 ⇤ 105
=
= 109 Hz = 1 GHz
⌥
3 ⇤ 10 4
Next, for a good conductor,
⇤=
⌥
1
=✏
2
f µ↵
⌃ ↵=
4
=
⌥2 f µ
(9 ⇤ 10
235
4
8 )(109 )(4
⇤ 10
7)
= 1.1 ⇤ 105 S/m
11.26. The dimensions of a certain coaxial transmission line are a = 0.8mm and b = 4mm. The outer
conductor thickness is 0.6mm, and all conductors have ↵ = 1.6 ⇤ 107 S/m.
a) Find R, the resistance per unit length, at an operating frequency of 2.4 GHz: First
⇤=✏
1
1
=⌧
8
f µ↵
(2.4 ⇤ 10 )(4 ⇤ 10
7 )(1.6
⇤ 107 )
= 2.57 ⇤ 10
6
m = 2.57µm
Then, using (90) with a unit length, we find
Rin =
1
=
2 a↵⇤
2 (0.8 ⇤ 10
3 )(1.6
1
⇤ 107 )(2.57 ⇤ 10
6)
= 4.84 ohms/m
The outer conductor resistance is then found from the inner through
Rout =
a
0.8
Rin =
(4.84) = 0.97 ohms/m
b
4
The net resistance per length is then the sum, R = Rin + Rout = 5.81 ohms/m.
b) Use information from Secs. 6.3 and 8.10 to find C and L, the capacitance and inductance
per unit length, respectively. The coax is air-filled. From those sections, we find (in free
space)
2 ⌅0
2 (8.854 ⇤ 10 12 )
C=
=
= 3.46 ⇤ 10 11 F/m
ln(b/a)
ln(4/.8)
µ0
4 ⇤ 10 7
ln(b/a) =
ln(4/.8) = 3.22 ⇤ 10 7 H/m
2
2
⌧
c) Find and ⇥ if + j⇥ = j C(R + j L): Taking real and imaginary parts of the given
expression, we find
L=
⇢⌧
= Re
j C(R + j L) =
✏
✏
LC ⌘
✏
1+
2
R
L
⌦2
⇢⌧
⇥ = Im
j C(R + j L) =
✏
✏
LC ⌘
✏
1+
2
R
L
⌦2
and
These can be found by writing out
= Re
⇣1/2
1✓
⇣1/2
+ 1✓
⇢⌧
⌧
j C(R + j L) = (1/2) j C(R + j L) + c.c.
where c.c denotes the complex conjugate. The result is squared, terms collected, and the
square root taken. Now, using the values of R, C, and L found in parts a and b, we find
= 3.0 ⇤ 10 2 Np/m and ⇥ = 50.3 rad/m.
236
11.27. The planar surface at z = 0 is a brass-Teflon interface. Use data available in Appendix C to
evaluate the following ratios for a uniform plane wave having = 4 ⇤ 1010 rad/s:
a) Tef / brass : From the appendix we find ⌅⌅⌅ /⌅⌅ = .0003 for Teflon, making the material a
good dielectric. Also, for Teflon, ⌅⌅r = 2.1. For brass, we find ↵ = 1.5 ⇤ 107 S/m, making
brass a good conductor at the stated frequency. For a good dielectric (Teflon) we use the
approximations:
. ↵
=
2
µ
=
⌅⌅
⌅⌅⌅
⌅⌅
⌦
1
2
↵
1
. ⌧ ⌅
⇥=
µ⌅ 1 +
8
For brass (good conductor) we have
.
. ⌧
=⇥=
f µ↵brass =
1
2
⌦
⌦
⌦
⌧
⌧
1 ⌅⌅⌅
⌅
µ⌅ =
⌅⌅r
⌅
2 ⌅
c
⌦
⌧
⌅⌅⌅
. ✏
=
µ⌅
=
⌅⌅r
⌅⌅
c
(4 ⇤ 1010 )(4 ⇤ 10
7 )(1.5
⇤ 107 ) = 6.14 ⇤ 105 m
1
Now
Tef
brass
⌧
✏
1/2 (⌅⌅⌅ /⌅⌅ ) ( /c) ⌅⌅r
(1/2)(.0003)(4 ⇤ 1010 /3 ⇤ 108 ) 2.1
✏
=
=
= 4.7 ⇤ 10
6.14 ⇤ 105
f µ↵brass
8
b)
c)
✏
⌥Tef
(2 /⇥Tef )
⇥brass
c f µ↵brass
(3 ⇤ 108 )(6.14 ⇤ 105 )
⌧ ⌅
✏
=
=
=
=
= 3.2 ⇤ 103
10
⌥brass
(2 /⇥brass )
⇥Tef
⌅r Tef
(4 ⇤ 10 ) 2.1
vTef
( /⇥Tef )
⇥brass
=
=
= 3.2 ⇤ 103 as before
vbrass
( /⇥brass )
⇥Tef
11.28. A uniform plane wave in free space has electric field given by Es = 10e
V/m.
j⇥x
az + 15e
j⇥x
ay
a) Describe the wave polarization: Since the two components have a fixed phase di⇤erence
(in this case zero) with respect to time and position, the wave has linear polarization,
with the field vector in the yz plane at angle = tan 1 (10/15) = 33.7⇤ to the y axis.
b) Find Hs : With propagation in forward x, we would have
Hs =
10
e
377
j⇥x
ay +
15
e
377
j⇥x
az A/m =
26.5e
j⇥x
ay + 39.8e
j⇥x
az mA/m
c) determine the average power density in the wave in W/m2 : Use
Pavg
↵
1
1 (10)2
(15)2
⇥
= Re {Es ⇤ Hs } =
ax +
ax = 0.43ax W/m2 or Pavg = 0.43 W/m2
2
2 377
377
237
11.29. Consider a left-circularly polarized wave in free space that propagates in the forward z direction. The electric field is given by the appropriate form of Eq. (100).
a) Determine the magnetic field phasor, Hs :
We begin, using (100), with Es = E0 (ax + jay )e j⇥z . We find the two components of
Hs separately, using the two components of Es . Specifically, the x component of Es is
associated with a y component of Hs , and the y component of Es is associated with a
negative x component of Hs . The result is
Hs =
E0
(ay
⇧0
jax ) e
j⇥z
b) Determine an expression for the average power density in the wave in W/m2 by direct
application of Eq. (77): We have
⌦
1
1
E0
Pz,avg = Re(Es ⇤ H⇥s ) = Re E0 (ax + jay )e j⇥z ⇤
(ay jax )e+j⇥z
2
2
⇧0
2
E
= 0 az W/m2 (assuming E0 is real)
⇧0
11.30. In an anisotropic medium, permittivity varies with electric field direction, and is a property
seen in most crystals. Consider a uniform plane wave propagating in the z direction in such a
medium, and which enters the material with equal field components along the x and y axes.
The field phasor will take the form:
Es (z) = E0 (ax + ay ej
⇥z
)e
j⇥z
where ⇥ = ⇥x ⇥y is the di⇤erence in phase constants for waves that are linearly-polarized
in the x and y directions. Find distances into the material (in terms of ⇥) at which the field
is:
a) Linearly-polarized: We want the x and y components to be in phase, so therefore
⇥zlin = m
⌃ zlin =
m
, (m = 1, 2, 3...)
⇥
b) Circularly-polarized: In this case, we want the two field components to be in quadrature
phase, such that the total field is of the form, Es = E0 (ax ± jay )e j⇥z . Therefore,
⇥zcirc =
(2n + 1)
2
⌃ zcirc =
(2n + 1)
2 ⇥
, (n = 0, 1, 2, 3...)
c) Assume intrinsic impedance ⇧ that is approximately constant with field orientation and
find Hs and < S >: Magnetic field is found by looking at the individual components:
Hs (z) =
< S >=
E0
ay
⇧
ax ej
⇥z
⇥
e
j⇥z
and
1
E2
Re {Es ⇤ H⇥s } = 0 az W/m2
2
⇧
where it is assumed that E0 is real. ⇧ is real because the medium is evidently lossless.
238
11.31. A linearly-polarized uniform plane wave, propagating in the forward z direction, is input to a
lossless anisotropic material, in which the dielectric constant encountered by waves polarized
along y (⌅ry ) di⇤ers from that seen by waves polarized along x (⌅rx ). Suppose ⌅rx = 2.15,
⌅ry = 2.10, and the wave electric field at input is polarized at 45⇤ to the positive x and y axes.
Assume free space wavelength ⌥.
a) Determine the shortest length of the material such that the wave as it emerges from the
output end is circularly polarized: With the input field at 45⇤ , the x and y components are
of equal magnitude, and circular polarization will result if the phase di⇤erence between
the components is /2. Our requirement over length L is thus ⇥x L ⇥y L = /2, or
L=
2(⇥x
⇥y )
=
✏
2 ( ⌅rx
c
✏
⌅ry )
With the given values, we find,
L=
(58.3) c
⌥
= 58.3 = 14.6 ⌥
2
4
b) Will the output wave be right- or left-circularly-polarized? With the dielectric constant
greater for x-polarized waves, the x component will lag the y component in time at the output. The field can thus be written as E = E0 (ay jax ), which is left circular polarization.
11.32. Suppose that the length of the medium of Problem 11.31 is made to be twice that as determined
in the problem. Describe the polarization of the output wave in this case: With the length
doubled, a phase shift of radians develops between the two components. At the input, we
can write the field as Es (0) = E0 (ax + ay ). After propagating through length L, we would
have,
Es (L) = E0 [e j⇥x L ax + e j⇥y L ay ] = E0 e j⇥x L [ax + e j(⇥y ⇥x )L ay ]
where (⇥y ⇥x )L =
(since ⇥x > ⇥y ), and so Es (L) = E0 e j⇥x L [ax ay ]. With the reversal
of the y component, the wave polarization is rotated by 90⇤ , but is still linear polarization.
11.33. Given a wave for which Es = 15e j⇥z ax + 18e
characterized by complex intrinsic impedance, ⇧.
e ay V/m, propagating in a medium
j⇥z j
a) Find Hs : With the wave propagating in the forward z direction, we find:
Hs =
⌅
1⇤
18ej ax + 15ay e
⇧
j⇥z
A/m
b) Determine the average power density in W/m2 : We find
Pz,avg =
1
1
Re {Es ⇤ H⇥s } = Re
2
2
239
(15)2
(18)2
+
⇧⇥
⇧⇥
= 275 Re
1
⇧⇥
W/m2
11.34. Given the general elliptically-polarized wave as per Eq. (93):
Es = [Ex0 ax + Ey0 ej ay ]e
j⇥z
a) Show, using methods similar to those of Example 11.7, that a linearly polarized wave
results when superimposing the given field and a phase-shifted field of the form:
Es = [Ex0 ax + Ey0 e
j
ay ]e
j⇥z j⇤
e
where ⇤ is a constant: Adding the two fields gives
⇤
⇥
⇥ ⌅
Es,tot = Ex0 1 + ej⇤ ax + Ey0 ej + e j ej⇤ ay e
✏
⌥
⌥
◆
j⇤/2
j⇤/2
j⇤/2
j⇤/2
=◆
E
e
e
+
e
a
+
E
e
e
x0
x
y0
⌘
!
"
!
j⇥z
j⇤/2 j
2 cos(⇤/2)
This simplifies to Es,tot = 2 [Ex0 cos(⇤/2)ax + Ey0 cos(
linearly polarized.
e
+e
2 cos(
j
⇤/2)
⇣

ej⇤/2 ay 
✓e
"
⇤/2)ay ] ej⇤/2 e
j⇥z
j⇥z
, which is
b) Find ⇤ in terms of such that the resultant wave is polarized along x: By inspecting the
part a result, we achieve a zero y component when 2
⇤ = (or odd multiples of ).
240
CHAPTER 12
+
+
12.1. A uniform plane wave in air, Ex1
= Ex10
cos(1010 t ⇥z) V/m, is normally-incident on a
copper surface at z = 0. What percentage of the incident power density is transmitted into
the copper? We need to find the reflection coe⌥cient. The intrinsic impedance of copper (a
good conductor) is
⌅c =
j µ
= (1 + j)
µ
2
⌧
= (1 + j)
1010 (4 ⇥ 107 )
= (1 + j)(.0104)
2(5.8 ⇥ 107 )
Note that the accuracy here is questionable, since we know the conductivity to only two
significant figures. We nevertheless proceed: Using ⌅0 = 376.7288 ohms, we write
=
⌅c ⌅0
.0104 376.7288 + j.0104
=
=
⌅c + ⌅0
.0104 + 376.7288 + j.0104
Now | |2 = .9999, and so the transmitted power fraction is 1
is transmitted.
.9999 + j.0001
| |2 = .0001, or about 0.01%
12.2. The plane z = 0 defines the boundary between two dielectrics. For z < 0, ⇤r1 = 9, ⇤⌅⌅r1 = 0,
+
and µ1 = µ0 . For z > 0, ⇤⌅r2 = 3, ⇤⌅⌅r2 = 0, and µ2 = µ0 . Let Ex1
= 10 cos( t 15z) V/m and
find
⇡ ⌅
⇡ ⌅
⇡
a) : We have ⇥ =
µ0 ⇤1 =
⇤r1 /c = 15. So = 15c/ ⇤⌅r1 = 15 ⇥ (3 ⇥ 108 )/ 9 =
1.5 ⇥ 109 s 1 .
⇡
⇡
b) < S+
µ0 /⇤⌅1 = ⌅0 / ⇤⌅r1 = 377/ 9 = 126 ohms. Next
1 >: First we need ⌅1 =
we apply Eq. (76), Chapter 11, to evaluate the Poynting vector (with no loss and
consequently with no phase di⌅erence between electric and magnetic fields). We find
2
2
2
< S+
1 >= (1/2)|E1 | /⌅1 az = (1/2)(10) /126 az = 0.40 az W/m .
c) < S1 >: First, we need to evaluate the reflection coe⌥cient:
⇡
⇡
⇡ ⌅
⇡ ⌅
⌅0 / ⇤⌅r2 ⌅0 / ⇤⌅r1
⇤r1
⇤
⌅2 ⌅1
⇡ ⌅
⇡ ⌅ =⇡ ⌅
⇡ r2
=
=
=
⌅2 + ⌅1
⌅0 / ⇤r2 + ⌅0 / ⇤r1
⇤r1 + ⇤⌅r2
Then < S1 >=
| |2 < S+
1 >=
(0.27)2 (0.40) az =
9
9+
3
= 0.27
3
0.03 az W/m2 .
d) < S+
2 >: This will be the remaining power, propagating in the forward z direction, or
2
< S+
2 >= 0.37 az W/m .
241
12.3. A uniform plane wave in region 1 is normally-incident on the planar boundary separating
regions 1 and 2. If ⇤⌅⌅1 = ⇤⌅⌅2 = 0, while ⇤⌅r1 = µ3r1 and ⇤⌅r2 = µ3r2 , find the ratio ⇤⌅r2 /⇤⌅r1 if 20% of
the energy in the incident wave is reflected at the boundary. There are two possible answers.
First, since | |2 = .20, and since both permittivities and permeabilities are real, = ±0.447.
we then set up
⇡
⇡
⌅0 (µr2 /⇤⌅r2 ) ⌅0 (µr1 /⇤⌅r1 )
⌅2 ⌅1
⇡
= ±0.447 =
= ⇡
⌅2 + ⌅1
⌅0 (µr2 /⇤⌅r2 ) + ⌅0 (µr1 /⇤⌅r1 )
⇡
⇡
(µr2 /µ3r2 )
(µ /µ3 )
µ
µr2
⇡ r1 r1 = r1
=⇡
3
3
µ
+
µr2
r1
(µr2 /µr2 ) + (µr1 /µr1 )
Therefore
µr2
1 ⌅ 0.447
=
= (0.382, 2.62) ⌃
µr1
1 ± 0.447
⇤⌅r2
=
⇤⌅r1
µr2
µr1
3
= (0.056, 17.9)
12.4. A 10-MHz uniform plane wave having an initial average power density of 5W/m2 is normallyincident from free space onto the surface of a lossy material in which ⇤⌅⌅2 /⇤⌅2 = 0.05, ⇤⌅r2 = 5,
and µ2 = µ0 . Calculate the distance into the lossy medium at which the transmitted wave
power density is down by 10dB from the initial 5W/m2 :
First, since ⇤⌅⌅2 /⇤⌅2 = 0.05 << 1, we recognize region 2 as a good dielectric. Its intrinsic
impedance is therefore approximated well by Eq. (62b), Chapter 11:
⌦
↵
µ0
1 ⇤⌅⌅2
377
⌅2 =
1+j ⌅ =
[1 + j0.025]
⌅
⇤2
2 ⇤2
5
The reflection coe⌥cient encountered by the incident wave from region 1 is therefore
=
⌅2 ⌅1
(377/ 5)[1 + j.025] 377
(1
=
=
⌅2 + ⌅1
(377/ 5)[1 + j.025] + 377
(1 +
5) + j.025
=
5) + j.025
0.383 + j0.011
The fraction of the incident power that is reflected is then | |2 = 0.147, and thus the
fraction of the power that is transmitted into region 2 is 1 | |2 = 0.853. Still using the
good dielectric approximation, the attenuation coe⌥cient in region 2 is found from Eq.
(60a), Chapter 11:
.
=
⇤⌅⌅2
2
µ0
= (2 ⇥ 107 )(0.05 ⇥ 5 ⇥ 8.854 ⇥ 10
⇤⌅2
12
)
377
= 2.34 ⇥ 10
2 5
2
Np/m
Now, the power that propagates into region 2 is expressed in terms of the incident power
through
< S2 > (z) = 5(1
| |2 )e
2 z
= 5(.853)e
2(2.34⇥10
2
)z
= 0.5 W/m2
in which the last equality indicates a factor of ten reduction from the incident power, as
occurs for a 10 dB loss. Solve for z to obtain
z=
ln(8.53)
2(2.34 ⇥ 10
242
2)
= 45.8 m
12.5. The region z < 0 is characterized by ⇤⌅r = µr = 1 and ⇤⌅⌅r = 0. The total E field here is given
as the sum of the two uniform plane waves, Es = 150e j10z ax + (50⇧ 20⇤ )ej10z ax V/m.
a) What is the operating frequency? In free space, ⇥ = k0 = 10 = /c = /3 ⇥ 108 . Thus,
= 3 ⇥ 109 s 1 , or f = /2 = 4.7 ⇥ 108 Hz.
b) Specify the intrinsic impedance of the region z > 0 that would provide the appropriate
reflected wave: Use
⇥
⇥
Er
50ej20
1
⌅ ⌅0
=
=
= ej20 = 0.31 + j0.11 =
Einc
150
3
⌅ + ⌅0
Now
⌅ = ⌅0
1+
1
= 377
1 + 0.31 + j0.11
1 0.31 j0.31
= 691 + j177 ⇤
c) At what value of z ( 10 cm < z < 0) is the total electric field intensity a maximum
amplitude? We found the phase of the reflection coe⌥cient to be ↵ = 20⇤ = .349rad, and
we use
↵
.349
zmax =
=
= 0.017 m = 1.7 cm
2⇥
20
12.6. In the beam-steering prism of Example 12.8, suppose the anti-reflective coatings are removed,
leaving bare glass-to-air interfaces. Calcluate the ratio of the prism output power to the input
power, assuming a single transit.
In making the transit, the light encounters two interfaces at normal incidence, at which
loss will occur. The reflection coe⌥cient at the front surface (air to glass) is
f
=
⌅g ⌅0
⌅0 /ng ⌅0
1 ng
=
=
⌅g + ⌅0
⌅0 /ng + ⌅0
1 + ng
Taking the glass index, ng , as 1.45, we find f = 0.18. The interface on exit from the
prism is glass to air, and so the reflection coe⌥cient there will be equal and opposite to
f ; i.e., b =
f.
Now, the wave power that makes it through (assuming total reflection at the 45⇤ interface)
will be
Pout = Pin (1 | f |2 )(1 | b |2 ) = Pin (1 |0.18|2 )2 = 0.93Pin
So we have 93% net transmission.
243
12.7. The semi-infinite regions z < 0 and z > 1 m are free space. For 0 < z < 1 m, ⇤⌅r = 4, µr = 1,
and ⇤⌅⌅r = 0. A uniform plane wave with = 4 ⇥ 108 rad/s is travelling in the az direction
toward the interface at z = 0.
a) Find the standing wave ratio in each of the three regions: First we find the phase constant
in the middle region,
⇡
⇤⌅r
2(4 ⇥ 108 )
=
= 2.67 rad/m
c
3 ⇥ 108
⇥2 =
Then, with the middle layer thickness ⇡
of 1 m, ⇥2 d = 2.67 rad. Also, the intrinsic
impedance of the middle layer is ⌅2 = ⌅0 / ⇤⌅r = ⌅0 /2. We now find the input impedance:
⌅in = ⌅2
⌦
↵
⌦
↵
⌅0 cos(⇥2 d) + j⌅2 sin(⇥2 d)
377 2 cos(2.67) + j sin(2.67)
=
= 231 + j141
⌅2 cos(⇥2 d) + j⌅0 sin(⇥2 d)
2
cos(2.67) + j2 sin(2.67)
Now, at the first interface,
12
=
⌅in ⌅0
231 + j141 377
=
=
⌅in + ⌅0
231 + j141 + 377
.176 + j.273 = .325⇧ 123⇤
The standing wave ratio measured in region 1 is thus
s1 =
1+|
1 |
12 |
12 |
=
1 + 0.325
= 1.96
1 0.325
In region 2 the standing wave ratio is found by considering the reflection coe⌥cient for
waves incident from region 2 on the second interface:
23
=
⌅0 ⌅0 /2
1 1/2
1
=
=
⌅0 + ⌅0 /2
1 + 1/2
3
Then
s2 =
1 + 1/3
=2
1 1/3
Finally, s3 = 1, since no reflected waves exist in region 3.
b) Find the location of the maximum |E| for z < 0 that is nearest to z = 0. We note that
the phase of 12 is ↵ = 123⇤ = 2.15 rad. Thus
zmax =
↵
2.15
=
=
2⇥
2(4/3)
.81 m
12.8. A wave starts at point a, propagates 1m through a lossy dielectric rated at dB = 0.1dB/cm,
reflects at normal incidence at a boundary at which = 0.3 + j0.4, and then returns to point
a. Calculate the ratio of the final power to the incident power after this round trip: Final
power, Pf , and incident power, Pi , are related through
Pf = Pi 10
0.1
dB L
| |2 10
0.1
dB L
⌃
Pf
= |0.3 + j0.4|2 10
Pi
244
0.2(0.1)100
= 2.5 ⇥ 10
3
12.9. Region 1, z < 0, and region 2, z > 0, are both perfect dielectrics (µ = µ0 , ⇤⌅⌅ = 0). A
uniform plane wave traveling in the az direction has a radian frequency of 3 ⇥ 1010 rad/s. Its
wavelengths in the two regions are ⌃1 = 5 cm and ⌃2 = 3 cm. What percentage of the energy
incident on the boundary is
a) reflected; We first note that
2 c
⌃1
⇤⌅r1 =
2
and ⇤⌅r2 =
2 c
⌃2
2
Therefore ⇤⌅r1 /⇤⌅r2 = (⌃2 /⌃1 )2 . Then with µ = µ0 in both regions, we find
⇡
⇡
⇡
⌅0 1/⇤⌅r2 ⌅0 1/⇤⌅r1
⇤⌅ /⇤⌅
1
⌅2 ⌅1
(⌃2 /⌃1 ) 1
⇡
⇡ r1 r2
=
= ⇡ ⌅
=
=
⌅
⌅
⌅
⌅2 + ⌅1
(⌃2 /⌃1 ) + 1
⌅0 1/⇤r2 + ⌅0 1/⇤r1
⇤r1 /⇤r2 + 1
⌃2 ⌃1
3 5
1
=
=
=
⌃2 + ⌃1
3+5
4
The fraction of the incident energy that is reflected is then | |2 = 1/16 = 6.25 ⇥ 10
2
.
b) transmitted? We use part a and find the transmitted fraction to be
1 | |2 = 15/16 = 0.938.
c) What is the standing wave ratio in region 1? Use
s=
1+| |
1 + 1/4
5
=
= = 1.67
1 | |
1 1/4
3
12.10. In Fig. 12.1, let region 2 be free space, while µr1 = 1, ⇤⌅⌅r1 = 0, and ⇤⌅r1 is unknown. Find ⇤⌅r1 if
a) the amplitude of E1 is one-half that of E+
1 : Since region 2 is free space, the reflection
coe⌥cient is
⇡
⇡ ⌅
⌅0 ⌅0 / ⇤⌅r1
⇤
1
|E1 |
⌅0 ⌅1
1
⇡ ⌅ = ⇡ r1
= + =
=
=
⌃ ⇤⌅r1 = 9
⌅
⌅
+
⌅
2
|E1 |
⌅0 + ⌅0 / ⇤r1
⇤r1 + 1
0
1
.
b) < S1 > is one-half of < S+
1 >: This time
⇧⇡
⇧2
⇧ ⇤⌅
⇧
1
1
⇧
⇧
| |2 = ⇧ ⇡ r1
⇧ =
⌅
⇧ ⇤r1 + 1 ⇧
2
c) |E1 |min is one-half |E1 |max : Use
|E1 |max
1+| |
=s=
=2 ⌃ | |=
|E1 |min
1 | |
245
⌃ ⇤⌅r1 = 34
⇡ ⌅
⇤
1
1
= = ⇡ r1
⌅
3
⇤r1 + 1
⌃ ⇤⌅r1 = 4
12.11. A 150 MHz uniform plane wave in normally-incident from air onto a material whose intrinsic
impedance is unknown. Measurements yield a standing wave ratio of 3 and the appearance of
an electric field minimum at 0.3 wavelengths in front of the interface. Determine the impedance
of the unknown material: First, the field minimum is used to find the phase of the reflection
coe⌥cient, where
1
zmin =
(↵ + ) = 0.3⌃ ⌃ ↵ = 0.2
2⇥
where ⇥ = 2 /⌃ has been used. Next,
| |=
s 1
3 1
1
=
=
s+1
3+1
2
So we now have
= 0.5ej0.2⇤ =
We solve for ⌅u to find
⌅u ⌅0
⌅u + ⌅0
⌅u = ⌅0 (1.70 + j1.33) = 641 + j501 ⇤
12.12. A 50MHz uniform plane wave is normally incident from air onto the surface of a calm ocean.
For seawater, = 4 S/m, and ⇤⌅r = 78.
a) Determine the fractions of the incident power that are reflected and transmitted: First
we find the loss tangent:
⇤⌅
=
4
2 (50 ⇥
106 )(78)(8.854
⇥ 10
12 )
= 18.4
This value is su⌥ciently greater than 1 to enable seawater to be considered a good conductor at 50MHz. ⇡
Then, using the approximation (Eq. 85, Chapter 11), the intrinsic
impedance is ⌅s =
f µ/ (1 + j), and the reflection coe⌥cient becomes
⇡
where
is
⇡
f µ/ =
=⇡
⇡
f µ/ (1 + j)
⌅0
f µ/ (1 + j) + ⌅0
(50 ⇥ 106 )(4 ⇥ 10 7 )/4 = 7.0. The fraction of the power reflected
⇡
[
f µ/
⌅0 ]2 + f µ/
Pr
[7.0 377]2 + 49.0
2
=| | = ⇡
=
= 0.93
Pi
[7.0 + 377]2 + 49.0
[
f µ/ + ⌅0 ]2 + f µ/
The transmitted fraction is then
Pt
=1
Pi
| |2 = 1
0.93 = 0.07
b) Qualitatively, how will these answers change (if at all) as the frequency is increased?
Within the limits of our good conductor approximation (loss tangent greater than about
ten), the reflected power fraction, using the formula derived in part a, is found to decrease
with increasing frequency. The transmitted power fraction thus increases.
246
12.13. A right-circularly-polarized plane wave is normally incident from air onto a semi-infinite slab
of plexiglas (⇤⌅r = 3.45, ⇤⌅⌅r = 0). Calculate the fractions of the incident power that are reflected
and transmitted. Also, describe the polarizations of the reflected and transmitted waves. First,
the impedance of the plexiglas will be ⌅ = ⌅0 / 3.45 = 203 ⇤. Then
=
203 377
=
203 + 377
0.30
The reflected power fraction is thus | |2 = 0.09. The total electric field in the plane of
the interface must rotate in the same direction as the incident field, in order to continually satisfy the boundary condition of tangential electric field continuity across the interface.
Therefore, the reflected wave will have to be left circularly polarized in order to make this
happen. The transmitted power fraction is now 1 | |2 = 0.91. The transmitted field will be
right circularly polarized (as the incident field) for the same reasons.
12.14. A left-circularly-polarized plane wave is normally-incident onto the surface of a perfect conductor.
a) Construct the superposition of the incident and reflected waves in phasor form: Assume
positive z travel for the incident electric field. Then, with reflection coe⌥cient, = 1,
the incident and reflected fields will add to give the total field:
Etot = Ei + Er = E0 (ax + jay )e j⇥z E0 (ax + jay )e+j⇥z
⇣
⌘
⇥
⇥

= E0 ✓ e j⇥z ej⇥z ax + j e j⇥z ej⇥z ay ◆ = 2E0 sin(⇥z) [ay
2j sin(⇥z)
jax ]
2j sin(⇥z)
b) Determine the real instantaneous form of the result of part a:
⇤
⌅
E(z, t) = Re Etot ej⌅t = 2E0 sin(⇥z) [cos( t)ay + sin( t)ax ]
c) Describe the wave that is formed: This is a standing wave exhibiting circular polarization
in time. At each location along the z axis, the field vector rotates clockwise in the xy
plane, and has amplitude (constant with time) given by 2E0 sin(⇥z).
247
12.15. Sulfur hexafluoride (SF6 ) is a high-density gas that has refractive index, ns = 1.8 at a specified
pressure, temperature, and wavelength. Consider the retro-reflecting prism shown in Fig.
12.16, that is immersed in SF6 . Light enters through a quarter-wave antireflective coating
and then totally reflects from the back surfaces of the glass. In principle, the beam should
experience zero loss at the design wavelength (Pout = Pin ).
a) Determine the minimum required value of the glass refractive index, ng , so that the
interior beam will totally reflect: We set the critical angle of total reflection equal to 45⇤ ,
which gives
ns
1
sin ⇧c =
= sin(45⇤ ) =
⌃ ng = ns 2 = 2.55
ng
2
b) Knowing ng , find the required refractive index of the quarter-wave film, nf : For a quarterwave section, we know that the film intrinsic impedance will be
⇡
⌅f = ⌅s ⌅g ⌃ nf = ns ng = (1.80)(2.55) = 2.14
c) With the SF6 gas evacuated from the chamber, and with the glass and film values as
previously found, find the ratio, Pout /Pin . Assume very slight misalignment, so that the
long beam path through the prism is not re-traced by reflected waves. The beam loses
power at the two normal-incidence boundaries, whereas the back reflections at 45⇤ will
still be lossless, as that angle is now greater than ⇧c with the reduced surrounding index.
At the first normal incidence boundary (from air to film to glass), the input intrinsic
impedance is
⌅f2
⌅02 /n2f
ng
⌅in1 =
=
= ⌅0
⌅g
⌅0 /ng
n2f
At the second normal incidence boundary at the prism exit (glass to film to air), the input
intrinsic impedance is
⌅f2
⌅02 /n2f
1
⌅in2 =
=
= ⌅0
⌅0
⌅0
n2f
The reflection coe⌥cients at the two boundaries will be
1
=
ng n2f
⌅in1 ⌅0
2.55 (2.14)2
=
=
=
2
⌅in1 + ⌅0
ng + nf
2.55 + (2.14)2
2
=
ng n2f
⌅in2 ⌅g
=
=
⌅in2 + ⌅g
ng + n2f
2
1|
⇥
0.285
1
The power ratio will be:
Pout
= 1
Pin
|
1
|
248
2
2|
⇥
= 1
(0.285)2
⇥2
= 0.845
12.16. In Fig. 12.5, let regions 2 and 3 both be of quarter-wave thickness. Region 4 is glass, having
refractive index, n4 = 1.45; region 1 is air.
a) Find ⌅in,b : Since region 3 is a quarter-wave layer, ⇥3 lb = /2, and (47) reduces to
⌅in,b =
⌅32
⌅4
b) Find ⌅in,a : Again, with region 2 of quarter-wave thickness, ⇥2 la = /2 and (48) becomes
⌅in,a =
⌅22
⌅ 2 ⌅4
= 22
⌅in,b
⌅3
c) Specify a relation between the four intrinsic impedances that will enable total transmission
of waves incident from the left into region 4: At the front surface, we need to have a zero
reflection coe⌥cient, so the input impedance there must match that of free space:
⌅in,a = ⌅0 ⌃ ⌅22 ⌅4 = ⌅32 ⌅0
d) Specify refractive index values for regions 2 and 3 that will accomplish the condition of
part c: We can rewrite the part c result as
⌅02
n22
⌅0
⌅4
=
⌅02
n23
⌅0 ⌃ n4 =
n23
n22
So any combination of indices that satisfy this result will work. One combination, for
example, would be n2 = 1.10 and n3 = 1.33. It is better to have the indices ascending
(or descending) monotonically in value from layer to layer because the high transmission
feature is then less sensitive to changes in wavelength (as an exercise for fun, show this).
e) Find the fraction of incident power transmitted if the two layers were of half-wave thickness instead of quarter-wave: For any half-wave layer, we know that the input impedance
is equal to that of the load. Therefore, ⌅in,b = ⌅in,a = ⌅4 . The reflection coe⌥cient at
the front surface is therefore
in
=
⌅in,a ⌅0
⌅4 ⌅0
1 n4
1 1.45
=
=
=
=
⌅in,a + ⌅0
⌅4 + ⌅0
1 + n4
1 + 1.45
The transmitted power fraction is then
Pt
=1
Pin
|
2
in |
249
=1
(0.184)2 = 0.97
0.184
12.17. A uniform plane wave in free space is normally-incident onto a dense dielectric plate of thickness ⌃/4, having refractive index n. Find the required value of n such that exactly half the
incident power is reflected (and half transmitted). Remember that n > 1.
In this problem, ⌅1 = ⌅3 = ⌅0 , and the quarter-wave section input impedance is therefore
⌅22
⌅ 2 /n2
⌅0
= 0
= 2
⌅3
⌅0
n
⌅in =
The reflection coe⌥cient at the front surface is then
in
⌅in ⌅0
1 n2
=
⌅in + ⌅0
1 + n2
=
For half-power reflection, we require that | in |2 = 0.5, or in = ±1/ 2. Since n must be
greater than 1, we choose the minus sign option and write:
1 n2
=
1 + n2
1
⌃ n=
2
2+1
2 1
✏1/2
= 2.41
12.18. A uniform plane wave is normally-incident onto a slab of glass (n = 1.45) whose back surface
is in contact with a perfect conductor. Determine the reflective phase shift at the front surface
of the glass if the glass thickness is: (a) ⌃/2; (b) ⌃/4; (c) ⌃/8.
With region 3 being a perfect conductor, ⌅3 = 0, and Eq. (36) gives the input impedance
to the structure as ⌅in = j⌅2 tan ⇥⌫. The reflection coe⌥cient is then
=
⌅in ⌅0
j⌅2 tan ⇥⌫ ⌅0
⌅ 2 tan2 ⇥⌫ ⌅02 + j2⌅0 ⌅2 tan ⇥⌫
=
= 2
=
⌅in + ⌅0
j⌅2 tan ⇥⌫ + ⌅0
⌅22 tan2 ⇥⌫ + ⌅02
r
+j
i
where the last equality occurs by multiplying the numerator and denominator of the
middle term by the complex conjugate of its denominator. The reflective phase is now
↵ = tan
1
i
= tan
r
1
⌦
↵
2⌅2 ⌅0 tan ⇥⌫
= tan
⌅22 tan2 ⇥⌫ ⌅02
1
⌦
(2.90) tan ⇥⌫
tan ⇥⌫ 2.10
↵
where ⌅2 = ⌅0 /1.45 has been used. We can now evaluate the phase shift for the three
given cases. First, when ⌫ = ⌃/2, ⇥⌫ = , and thus ↵(⌃/2) = 0. Next, when ⌫ = ⌃/4,
⇥⌫ = /2, and
↵(⌃/4) ⇧ tan 1 [2.90] = 71⇤
as ⌫ ⇧ ⌃/4. Finally, when ⌫ = ⌃/8, ⇥⌫ = /4, and
↵(⌃/8) = tan
1
⌦
↵
2.90
=
1 2.10
250
69.2⇤ (or 291⇤ )
12.19. You are given four slabs of lossless dielectric, all with the same intrinsic impedance, ⌅, known
to be di⌅erent from that of free space. The thickness of each slab is ⌃/4, where ⌃ is the
wavelength as measured in the slab material. The slabs are to be positioned parallel to one
another, and the combination lies in the path of a uniform plane wave, normally-incident. The
slabs are to be arranged such that the air spaces between them are either zero, one-quarter
wavelength, or one-half wavelength in thickness. Specify an arrangement of slabs and air
spaces such that
a) the wave is totally transmitted through the stack: In this case, we look for a combination
of half-wave sections. Let the inter-slab distances be d1 , d2 , and d3 (from left to right).
Two possibilities are i.) d1 = d2 = d3 = 0, thus creating a single section of thickness ⌃,
or ii.) d1 = d3 = 0, d2 = ⌃/2, thus yielding two half-wave sections separated by a halfwavelength.
b) the stack presents the highest reflectivity to the incident wave: The best choice here
is to make d1 = d2 = d3 = ⌃/4. Thus every thickness is one-quarter wavelength. The
impedances transform as follows: First, the input impedance at the front surface of the last
slab (slab 4) is ⌅in,1 = ⌅ 2 /⌅0 . We transform this back to the back surface of slab 3, moving
through a distance of ⌃/4 in free space: ⌅in,2 = ⌅02 /⌅in,1 = ⌅03 /⌅ 2 . We next transform this
impedance to the front surface of slab 3, producing ⌅in,3 = ⌅ 2 /⌅in,2 = ⌅ 4 /⌅03 . We continue
in this manner until reaching the front surface of slab 1, where we find ⌅in,7 = ⌅ 8 /⌅07 .
Assuming ⌅ < ⌅0 , the ratio ⌅ n /⌅0n 1 becomes smaller as n increases (as the number of
slabs increases). The reflection coe⌥cient for waves incident on the front slab thus gets
close to unity, and approaches 1 as the number of slabs approaches infinity.
251
12.20. The 50MHz plane wave of Problem 12.12 is incident onto the ocean surface at an angle to the
normal of 60⇤ . Determine the fractions of the incident power that are reflected and transmitted
for
a) s polarization: To review Problem 12, we first we find the loss tangent:
⇤⌅
=
4
2 (50 ⇥
106 )(78)(8.854
⇥ 10
12 )
= 18.4
This value is su⌥ciently greater than 1 to enable seawater to be considered a good conductor at 50MHz. Then, using the
⇡ approximation (Eq. 85, Chapter 11), and with µ = µ0 ,
the intrinsic impedance is ⌅s =
f µ/ (1 + j) = 7.0(1 + j). Next we need the angle of
refraction, which means that we need to know the refractive index of seawater at 50MHz.
For a uniform plane wave in a good conductor, the phase constant is
⇥=
nsea
c
. ⇡
=
fµ
.
⌃ nsea = c
µ
= 26.8
4 f
Then, using Snell’s law, the angle of refraction is found:
sin ⇧2 =
nsea
sin ⇧1 = 26.8 sin(60⇤ ) ⌃ ⇧2 = 1.9⇤
n1
.
This angle is small enough so that cos ⇧2 = 1. Therefore, for s polarization,
s
7.0(1 + j) 377/ cos 60⇤
. ⌅s2 ⌅s1
=
=
=
⌅s2 + ⌅s1
7.0(1 + j) + 377/ cos 60⇤
The fraction of the power reflected is now |
0.04.
2
s|
0.98 + j0.018 = 0.98⇧ 179⇤
= 0.96. The fraction transmitted is then
b) p polarization: Again, with the refracted angle close to zero, the relection coe⌥cient for
p polarization is
p
7.0(1 + j) 377 cos 60⇤
. ⌅p2 ⌅p1
=
=
=
⌅p2 + ⌅p1
7.0(1 + j) + 377 cos 60⇤
The fraction of the power reflected is now |
0.14.
252
2
p|
0.93 + j0.069 = 0.93⇧ 176⇤
= 0.86. The fraction transmitted is then
12.21. A right-circularly polarized plane wave in air is incident at Brewster’s angle onto a semi-infinite
slab of plexiglas (⇤⌅r = 3.45, ⇤⌅⌅r = 0, µ = µ0 ).
a) Determine the fractions of the incident power that are reflected and transmitted: In
plexiglas, Brewster’s angle is ⇧B = ⇧1 = tan 1 (⇤⌅r2 /⇤⌅r1 ) = tan 1 ( 3.45) = 61.7⇤ . Then
the angle of refraction is ⇧2 = 90⇤ ⇧B (see Example 12.9), or ⇧2 = 28.3⇤ . With incidence
at Brewster’s angle, all p-polarized power will be transmitted — only s-polarized power
will be reflected. This is found through
s
=
⌅2s ⌅1s
.614⌅0 2.11⌅0
=
=
⌅2s + ⌅1s
.614⌅0 + 2.11⌅0
0.549
where ⌅1s = ⌅1 sec ⇧1 = ⌅0 sec(61.7⇤ ) = 2.11⌅0 ,
and ⌅2s = ⌅2 sec ⇧2 = (⌅0 / 3.45) sec(28.3⇤ ) = 0.614⌅0 . Now, the reflected power fraction
is | |2 = ( .549)2 = .302. Since the wave is circularly-polarized, the s-polarized component represents one-half the total incident wave power, and so the fraction of the total
power that is reflected is .302/2 = 0.15, or 15%. The fraction of the incident power that
is transmitted is then the remainder, or 85%.
b) Describe the polarizations of the reflected and transmitted waves: Since all the p-polarized
component is transmitted, the reflected wave will be entirely s-polarized (linear). The
transmitted wave, while having all the incident p-polarized power, will have a reduced
s-component, and so this wave will be right-elliptically polarized.
12.22. A dielectric waveguide is shown in Fig. 12.16 with refractive indices as labeled. Incident light
enters the guide at angle ↵ from the front surface normal as shown. Once inside, the light
totally reflects at the upper n1 n2 interface, where n1 > n2 . All subsequent reflections from
the upper an lower boundaries will be total as well, and so the light is confined to the guide.
Express, in terms of n1 and n2 , the maximum value of ↵ such that total confinement will
occur, with n0 = 1. The quantity sin ↵ is known as the numerical aperture of the guide.
From the illustration we see that ↵1 maximizes when ⇧1 is at its minimum value. This minimum
will be the critical angle for the n1 n2 interface, where sin ⇧c = sin ⇧1 = n2 /n1 . Let the
refracted angle to the right of the vertical interface (not shown) be ↵2 , where n0 sin ↵1 =
n1 sin ↵2 . Then we see that ↵2 + ⇧1 = 90⇤ , and so sin ⇧1 = cos ↵2 . Now, the numerical aperture
becomes
⇢
⇢
⇡
n1
2
2
sin ↵1max =
sin ↵2 = n1 cos ⇧1 = n1 1 sin ⇧1 = n1 1 (n2 /n1 ) = n21 n22
n0
⌃⇡
⌥
Finally, ↵1max = sin 1
n21 n22 is the numerical aperture angle.
12.23. Suppose that ↵1 in Fig. 12.16 is Brewster’s angle, and that ⇧1 is the critical angle. Find n0 in
terms of n1 and n2 : With the incoming ray at Brewster’s angle, the refracted angle of this ray
(measured from the inside normal to the front surface) will be 90⇤ ↵1 . Therefore, ↵1 = ⇧1 ,
and thus sin ↵1 = sin ⇧1 . Thus
⇢
n1
n2
sin ↵1 = ⇡ 2
=
sin
⇧
=
⌃
n
=
(n
/n
)
n21 n22
1
0
1
2
2
n
1
n0 + n1
Alternatively, we
⇡ could have used the result of Problem 12.22, in which it was found that
sin ↵1 = (1/n0 ) n21 n22 , which we then set equal to sin ⇧1 = n2 /n1 to get the same result.
253
12.24. A Brewster prism is designed to pass p-polarized light without any reflective loss. The prism
of Fig. 12.17 is made of glass (n = 1.45), and is in air. Considering the light path shown,
determine the vertex angle, : With entrance and exit rays at Brewster’s angle (to eliminate
reflective loss), the interior ray must be horizontal, or parallel to the bottom surface of the
prism. From the geometry, the angle between the interior ray and the normal to the prism
surfaces that it intersects is /2. Since this angle is also Brewster’s angle, we may write:
= 2 sin
1
1
1 + n2
= 2 sin
1
1
⇡
1 + (1.45)2
= 1.21 rad = 69.2⇤
12.25. In the Brewster prism of Fig. 12.17, determine for s-polarized light the fraction of the incident
power that is transmitted through the prism, and from this specify the dB insertion loss,
defined as 10 log10 of that number:
We use
and
s
= (⌅s2
⌅s1 )/(⌅s2 + ⌅s1 ), where
⌅s2 =
⌅2
⌅2
⌅0 ⇡
=
= 2 1 + n2
cos(⇧B2 )
n
n/ 1 + n2
⌅s1 =
⇡
⌅1
⌅1
=
= ⌅0 1 + n2
cos(⇧B1 )
1/ 1 + n2
Thus, at the first interface, = (1 n2 )/(1 + n2 ). At the second interface, will be equal
but of opposite sign to the above value. The power transmission coe⌥cient through each
interface is 1 | |2 , so that for both interfaces, we have, with n = 1.45:
Ptr
= 1
Pinc
⇥
2 2
| |
= 1
n2 1
n2 + 1
✏
2 2
= 0.76
The insertion loss is now
⌫i (dB) = 10 log10 (0.76) =
254
1.19 dB
12.26. Show how a single block of glass can be used to turn a p-polarized beam of iight through 180⇤ ,
with the light su⌅ering, in principle, zero reflective loss. The light is incident from air, and
the returning beam (also in air) may be displaced sideways from the incident beam. Specify
all pertinent angles and use n = 1.45 for glass. More than one design is possible here.
The prism below is designed such that light enters at Brewster’s angle, and once inside, is
turned around using total reflection. Using the result of Example 12.9, we find that with
glass, ⇧B = 55.4⇤ , which, by the geometry, is also the incident angle for total reflection at the
back of the prism. For this to work, the Brewster angle must be greater than or equal to the
critical angle. This is in fact the case, since ⇧c = sin 1 (n2 /n1 ) = sin 1 (1/1.45) = 43.6⇤ .
12.27. Using Eq. (79) in Chapter 11 as a starting point, determine the ratio of the group and phase
velocities of an electromagnetic wave in a good conductor. Assume conductivity does not vary
with frequency: In a good conductor:
⇥=
Thus
d
=
d⇥
d⇥
d
⇡
fµ =
1
=2
µ
2
2
= vg
µ
⇧
and
Therefore vg /vp = 2.
255
d⇥
1⌫ µ ⇠
=
d
2
2
vp =
⇥
=⇡
1/2
µ
2
µ /2
=
2
µ
12.28. Over a small wavelength range, the refractive index of a certain material varies approximately
.
linearly with wavelength as n(⌃) = na + nb (⌃ ⌃a ), where na , nb , and ⌃a are constants, and
where ⌃ is the free space wavelength.
a) Show that d/d = (2 c/ 2 )d/d⌃: With ⌃ as the free space wavelength, we use ⌃ =
2 c/ , from which d⌃/d = 2 c/ 2 . Then d/d = (d⌃/d ) d/d⌃ = (2 c/ 2 ) d/d⌃.
b) Using ⇥(⌃) = 2 n/⌃, determine the wavelength-dependent (or independent) group delay
over a unit distance: This will be
⌦
↵
⌦
↵
1
d⇥
d 2 n(⌃)
2 c d 2
tg =
=
=
=
[na + nb (⌃ ⌃a )]
2 d⌃
vg
d
d
⌃
⌃
⌦
↵
2 c
2
2
=
[na + nb (⌃ ⌃a )] +
nb
2
⌃2
⌃
⌦
↵
⌃2
2 na
2 nb ⌃a
1
=
+
= (na nb ⌃a ) s/m
2
2
2 c
⌃
⌃
c
c) Determine ⇥2 from your result of part b: ⇥2 = d2 ⇥/d 2 |⌅0 . Since the part b result is
independent of wavelength (and of frequency), it follows that ⇥2 = 0.
d) Discuss the implications of these results, if any, on pulse broadening: A wavelengthindependent group delay (leading to zero ⇥2 ) means that there will simply be no pulse
broadening at all. All frequency components arrive simultaneously. This sort of thing
happens in most transparent materials – that is, there will be a certain wavelength, known
as the zero dispersion wavelength, around which the variation of n with ⌃ is locally linear.
Transmitting pulses at this wavelength will result in no pulse broadening (to first order).
12.29. A T = 5 ps transform-limited pulse propagates in a dispersive channel for which ⇥2 =
10 ps2 /km. Over what distance
⇡ will the pulse spread to twice its initial width? After propagation, the width is T ⌅ = T 2 + (⇥⌦ )2 = 2T . Thus ⇥⌦ = 3T , where ⇥⌦ = ⇥2 z/T .
Therefore
⇥2 z
3T 2
3(5 ps)2
= 3T or z =
=
= 4.3 km
T
⇥2
10 ps2 /km
12.30. A T = 20 ps transform-limited pulse propagates through 10 km of a dispersive channel for
which ⇥2 = 12 ps2 /km. The pulse then propagates through a second 10 km channel for which
⇥2 = 12 ps2 /km. Describe the pulse at the output of the second channel and give a physical
explanation for what happened.
Our theory of pulse spreading will allow for changes in ⇥2 down the length of the channel. In
fact, we may write in general:
1 L
⇥⌦ =
⇥2 (z) dz
T 0
Having ⇥2 change sign at the midpoint, yields a zero ⇥⌦ , and so the pulse emerges from the
output unchanged! Physically, the pulse acquires a positive linear chirp (frequency increases
with time over the pulse envelope) during the first half of the channel. When ⇥2 switches sign,
the pulse begins to acquire a negative chirp in the second half, which, over an equal distance,
will completely eliminate the chirp acquired during the first half. The pulse, if originally
transform-limited at input, will emerge, again transform-limited, at its original width. More
generally, complete dispersion compensation is achieved using a two-segment channel when
⇥2 L = ⇥2⌅ L⌅ , assuming dispersion terms of higher order than ⇥2 do not exist.
256
CHAPTER 13
13.1. The conductors of a coaxial transmission line are copper ( c = 5.8 ⇤ 10 7 S/m) and the dielectric
is polyethylene (⌅⌅r = 2.26, / ⌅⌅ = 0.0002). If the inner radius of the outer conductor is 4 mm,
find the radius of the inner conductor so that (assuming a lossless line):
a) Z0 = 50 ⇤: Use
1
Z0 =
2⌦
✓
µ
ln
⌅⌅
⇣
⇤ ⌅
⇤ ⌅
2⌦ ⌅⌅r (50)
b
b
= 50 ⌃ ln
=
= 1.25
a
a
377
Thus b/a = e1.25 = 3.50, or a = 4/3.50 = 1.142 mm
b) C = 100 pF/m: Begin with
2⌦⌅⌅
C=
= 10
ln(b/a)
10
⇤ ⌅
b
⌃ ln
= 2⌦(2.26)(8.854 ⇤ 10
a
2
) = 1.257
So b/a = e1.257 = 3.51, or a = 4/3.51 = 1.138 mm.
c) L = 0.2 µH/m: Use
µ0
L=
ln
2⌦
⇤ ⌅
b
= 0.2 ⇤ 10
a
6
⇤ ⌅
b
2⌦(0.2 ⇤ 10 6 )
⌃ ln
=
=1
a
4⌦ ⇤ 10 7
Thus b/a = e1 = 2.718, or a = b/2.718 = 1.472 mm.
257
13.2. Find R, L, C, and G for a coaxial cable with a = 0.25 mm, b = 2.50 mm, c = 3.30 mm, ⌅r = 2.0,
µr = 1, c = 1.0 ⇤ 107 S/m, = 1.0 ⇤ 10 5 S/m, and f = 300 MHz.
First, we note that the metal is a good conductor, as confirmed by the loss tangent:
c
⌅⌅
=
1.0 ⇤ 107
2⌦(3.00 ⇤ 108 )(2)(8.854 ⇤ 10
12 )
= 3.0 ⇤ 108 >> 1
So the skin depth into the metal is
⇤=
✓
2
µ
=
c
◆
2⌦(3.00 ⇤
108 )(4⌦
2
⇤ 10
7 )(1.0
⇤ 107 )
= 9.2 µm
The current can be said to exist in layers of thickness ⇤ just beneath the inner conductor
radius, a, and just inside the outer conductor at its inner radius, b. The outer conductor far
radius, c, is of no consequence. Both conductors behave essentially as thin films of thickness
⇤. We are thus in the high frequency regime, and the equations in Sec. 13.1.1 apply. The
calculations for the primary constants are as follows:
1
R=
2⌦⇤
c
⇤
1 1
+
a b
⌅
=
1
2⌦(9.2 ⇤ 10
6 )(1.0
⇤ 107 )
⇤
1
0.25 ⇤ 10
3
1
+
2.5 ⇤ 10
using Eq. (12).
L = Lext
µ
=
ln
2⌦
using Eq. (11).
C=
⇤ ⌅
b
4⌦ ⇤ 10
=
a
2⌦
7
ln
2⌦⌅⌅
2⌦(2)(8.854 ⇤ 10
=
ln(b/a)
ln(2.5/0.25)
using Eq. (9).
G=
⇤
2.5
0.25
12
)
⌅
= 0.46 µH/m
= 48 pF/m
2⌦
2⌦(1.0 ⇤ 10 5 )
=
= 27 µS/m
ln(b/a)
ln(2.5/0.25)
using Eq. (10).
258
3
⌅
= 7.6 ⇤/m
13.3. Two aluminum-clad steel conductors are used to construct a two-wire transmission line. Let
7
6
Al = 3.8 ⇤ 10 S/m, St = 5 ⇤ 10 S/m, and µSt = 100 µH/m. The radius of the steel wire is 0.5
in., and the aluminum coating is 0.05 in. thick. The dielectric is air, and the center-to-center wire
separation is 4 in. Find C, L, G, and R for the line at 10 MHz: The first question is whether we
are in the high frequency or low frequency regime. Calculation of the skin depth, ⇤, will tell us.
We have, for aluminum,
1
⇤=⌦
⌦f µ0
Al
1
=⇣
7
⌦(10 )(4⌦ ⇤ 10
7 )(3.8
⇤
107 )
= 2.58 ⇤ 10
5
m
so we are clearly in the high frequency regime, where uniform current distributions cannot be
assumed. Furthermore, the skin depth is considerably less than the aluminum layer thickness, so
the bulk of the current resides in the aluminum, and we may neglect the steel. Assuming solid
aluminum wires of radius a = 0.5 + 0.05 = 0.55 in = 0.014 m, the resistance of the two-wire line
is now
1
1
R=
=
= 0.023 ⇤/m
⌦a⇤ Al
⌦(.014)(2.58 ⇤ 10 5 )(3.8 ⇤ 107 )
Next, since the dielectric is air, no leakage will occur from wire to wire, and so G = 0 S/m. Now
the capacitance will be
C=
⌦⌅0
⌦ ⇤ 8.85 ⇤ 10 12
=
= 1.42 ⇤ 10
1
cosh (d/2a)
cosh 1 (4/(2 ⇤ 0.55))
11
F/m = 14.2 pF/m
Finally, the inductance per unit length will be
L=
µ0
4⌦ ⇤ 10
cosh(d/2a) =
⌦
⌦
7
cosh (4/(2 ⇤ 0.55)) = 7.86 ⇤ 10
259
7
H/m = 0.786 µH/m
13.4. Find R, L, C, and G for a two-wire transmission line in polyethylene at f = 800 MHz. Assume
copper conductors of radius 0.50 mm and separation 0.80 cm. Use ⌅r = 2.26 and /( ⌅⌅ ) =
4.0 ⇤ 10 4 .
From the loss tangent, we find the conductivity of polyethylene:
= (4.0 ⇤ 10
4
)(2⌦ ⇤ 8.00 ⇤ 108 )(2.26)(8.854 ⇤ 10
12
) = 4.0 ⇤ 10
5
S/m
As polyethylene is a good dielectric, its penetration depth is (using Eq. (60a), Chapter 11):
1
⇤p =
=
2
p
◆
⌦
⌅⌅
2⌅⌅r
2 2.26
=
=
= 199 m
µ
⇧0
(4.0 ⇤ 10 5 )(377)
Therefore, we can assume field distributions over a any cross-sectional plane to be the same
as those of the lossless line. On the other hand, within the copper conductors (for which
7
c = 5.8 ⇤ 10 ) the skin depth will be (from Eq. (82), Chapter 11):
⇤c =
1
⌦f µ
=
c
⌦(8.00 ⇤
108 )(4⌦
1
⇤ 10
7 )(5.8
⇤ 107 )
= 2.3 µm
As this value is much less than the overall conductor dimensions, the line operates in the high
frequency regime. The primary constants are found using Eqs. (20) - (23). We have
C=
⌦⌅⌅
⌦(2.26)(8.854 ⇤ 10
=
1
cosh (d/2a)
cosh 1 (8)
L = Lext =
µ
cosh
⌦
G=
R=
1
⌦a⇤c
=
c
1
(d/2a) =
⌦
cosh
1
(d/2a)
⌦(5.0 ⇤ 10
=
4⌦ ⇤ 10
⌦
cosh
)
1
= 22.7 pF/m
(8) = 1.11 µH/m
⌦(4.0 ⇤ 10 5 )
= 46 µS/m
cosh 1 (8)
4 )(2.3
260
7
12
1
⇤ 10
6 )(5.8
⇤ 107 )
= 4.8 ohms/m
13.5. Each conductor of a two-wire transmission line has a radius of 0.5mm; their center-to-center
distance is 0.8cm. Let f = 150M Hz and assume and c are zero. Find the dielectric constant
of the insulating medium if
a) Z0 = 300 ⇤: Use
1
300 =
⌦
✓
µ0
cosh
⌅⌅r ⌅0
1
⇤
d
2a
⌅
⇣
120⌦
⌃
⌅⌅r =
cosh
300⌦
1
⇤
8
2(.5)
⌅
= 1.107 ⌃ ⌅⌅r = 1.23
b) C = 20 pF/m: Use
20 ⇤ 10
12
=
⌦⌅⌅
cosh 1 (d/2a)
⌃ ⌅⌅r =
20 ⇤ 10
⌦⌅0
12
cosh
1
(8) = 1.99
c) vp = 2.6 ⇤ 108 m/s:
1
1
c
⇣
vp = ⌦
=⇣
=
µ0 ⌅0 ⌅⌅r
⌅⌅r
LC
⌃
⌅⌅r
=
⇤
3.0 ⇤ 108
2.6 ⇤ 108
⌅2
= 1.33
13.6. The transmission line in Fig. 6.8 is filled with polyethylene. If it were filled with air, the capacitance
would be 57.6 pF/m. Assuming that the line is lossless, find C, L, and Z0 .
The line cross-section is of little consequence here because we have its capacitance and we
know that it is lossless. The capacitance with polyethylene added will be just the air-filled
line capacitance multiplied by the dielectric constant of polyethylene, which is ⌅⌅r = 2.26, or
Cp = ⌅⌅r Cair = 2.26(57.6) = 130 pF/m
The inductance is found from the capacitance and the wave velocity, which in the air-filled
line is just the speed of light in vacuum:
1
1
1
vp = c = 3⇤108 m/s = ⌦
⌃ L= 2
=
8
2
c Cair
(3.00 ⇤ 10 ) (57.6 ⇤ 10
LCair
Finally, the characteristic impedance in the dielectric-filled line will be
Z0 =
◆
L
=
Cp
✓
1.93 ⇤ 10
1.30 ⇤ 10
7
10
= 38.5 ohms
where the inductance is una⌅ected by the incorporation of the dielectric.
261
12 )
= 0.193 µH/m
13.7. Pertinent dimensions for the transmission line shown in Fig. 13.2 are b = 3 mm, and d = 0.2 mm.
The conductors and the dielectric are non-magnetic.
a) If the characteristic impedance of the line is 15 ⇤, find ⌅⌅r : We use
✓
Z0 =
⇤ ⌅
⇤
⌅2
d
377
.04
⌅
= 15 ⌃ ⌅r =
= 2.8
b
15
9
µ
⌅⌅
b) Assume copper conductors and operation at 2 ⇤ 108 rad/s. If RC = GL, determine the loss
tangent of the dielectric: For copper, c = 5.8 ⇤ 107 S/m, and the skin depth is
⇤=
✓
2
µ0
=
c
Then
◆
2
(2 ⇤ 108 )(4⌦ ⇤ 10
7 )(5.8
2
2
=
7
(5.8 ⇤ 10 )(1.2 ⇤ 10
c ⇤b
R=
Now
C=
⌅⌅ b
(2.8)(8.85 ⇤ 10
=
d
0.2
and
L=
12
)(3)
⇤ 107 )
5 )(.003)
= 1.2 ⇤ 10
= 0.98 ⇤/m
= 3.7 ⇤ 10
µ0 d
(4⌦ ⇤ 10 7 )(0.2)
=
= 8.4 ⇤ 10
b
3
8
10
F/m
H/m
Then, with RC = GL,
G=
Thus
d
= (4.4 ⇤ 10
3
RC
(.98)(3.7 ⇤ 10 10 )
=
= 4.4 ⇤ 10
L
(8.4 ⇤ 10 8 )
)(0.2/3) = 2.9 ⇤ 10
l.t. =
d
⌅⌅
=
(2 ⇤
4
S/m =
db
d
S/m. The loss tangent is
2.9 ⇤ 10
4
108 )(2.8)(8.85
262
3
⇤ 10
12 )
= 5.85 ⇤ 10
2
5
m
13.8. A transmission line constructed from perfect conductors and an air dielectric is to have a maximum
dimension of 8mm for its cross-section. The line is to be used at high frequencies. Specify its
dimensions if it is:
a) a two-wire line with Z0 = 300 ⇤: With the maximum dimension of 8mm, we have, using (24):
1
Z0 =
⌦
✓
µ
cosh
⌅⌅
1
⇤
8
2a
2a
⌅
= 300 ⌃
Solve for a to find a = 0.56 mm. Then d = 8
8
2a
= cosh
2a
⇤
300⌦
120⌦
⌅
= 6.13
2a = 6.88 mm.
b) a planar line with Z0 = 15 ⇤: In this case our maximum dimension dictates that
So, using (8), we write
Z0 =
✓
µ
⌅⌅
⌦
⇣
64 b2
= 15 ⌃
64
b
b2 =
⌦
d2 + b2 = 8.
15
b
377
Solving, we find b = 7.99 mm and d = 0.32 mm.
c) a 72 ⇤ coax having a zero-thickness outer conductor: With a zero-thickness outer conductor,
we note that the outer radius is b = 8/2 = 4mm. Using (13), we write
1
Z0 =
2⌦
✓
µ
ln
⌅⌅
⇤ ⌅
⇤ ⌅
b
b
2⌦(72)
= 72 ⌃ ln
=
= 1.20 ⌃ a = be
a
a
120⌦
1.20
= 4e
1.20
= 1.2
Summarizing, a = 1.2 mm and b = 4 mm.
13.9. A microstrip line is to be constructed using a lossless dielectric for which ⌅⌅r = 7.0. If the line is to
have a 50-⇤ characteristic impedance, determine:
a) ⌅r,ef f : We use Eq. (34) (under the assumption that w/d > 1.3) to find:
⌅r,ef f = 7.0 [0.96 + 7.0(0.109
0.004 ⇤ 7.0) log10 (10 + 50)
1]
1
= 5.0
b) w/d: Still under the assumption that w/d > 1.3, we solve Eq. (33) for d/w to find
⌥
⇤
d
⌅⌅r 1
= (0.1)
⌅r,ef f
w
2
= (0.1) 3.0 (5.0
4.0)
1
⌅⌅r + 1
2
✏1/0.555
263
⌅
1 1/0.555
0.10
0.10 = 0.624 ⌃
w
= 1.60
d
13.10. Two microstrip lines are fabricated end-to-end on a 2-mm thick wafer of lithium niobate (⌅⌅r = 4.8).
Line 1 is of 4mm width; line 2 (unfortunately) has been fabricated with a 5mm width. Determine
the power loss in dB for waves transmitted through the junction.
We first note that w1 /d1 = 2.0 and w2 /d2 = 2.5, so that Eqs. (32) and (33) apply. As the
first step, solve for the e⌅ective dielectric constants for the two lines, using (33). For line 1:
⌅r1,ef f
For line 2:
⌅r2,ef f
⇧
⇤ ⌅⌃
5.8 3.8
1
=
+
1 + 10
2
2
2
⇧
⇤
⌅⌃
5.8 3.8
1
=
+
1 + 10
2
2
2.5
0.555
= 3.60
0.555
= 3.68
We next use Eq. (32) to find the characteristic impedances for the air-filled cases. For line 1:
air
Z01
and for line 2:
⌦
⇤ ⌅ ◆ ⇤ ⌅2
1
1
= 60 ln ↵4
+ 16
+ 2 = 89.6 ohms
2
2
air
Z02
= 60 ln ↵4
⇤
1
2.5
⌅
+
◆
16
⇤
1
2.5
⌅2
⌦
+ 2 = 79.1 ohms
The actual line impedances are given by Eq. (31). Using our results, we find
Z air
89.6
Z air
79.1
Z01 = ⌦ 01
=⌦
= 47.2 ⇤ and Z02 = ⌦ 02
=⌦
= 41.2 ⇤
⌅r1,ef f
⌅r2,ef f
3.60
3.68
The reflection coe⌥cient at the junction is now
=
Z02 Z01
47.2 41.2
=
= 0.068
Z02 + Z01
47.2 + 41.2
The transmission loss in dB is then
PL (dB) =
10 log10 (1
| |2 ) =
264
10 log10 (0.995) = 0.02 dB
13.11. A parallel-plate waveguide is known to have a cuto⌅ wavelength for the m = 1 TE and TM modes
of c1 = 0.4 cm. The guide is operated at wavelength = 1 mm. How many modes propagate?
The cuto⌅ wavelength for mode m is cm = 2nd/m, where n is the refractive index of the guide
interior. For the first mode, we are given
c1
=
2nd
0.4
0.2
= 0.4 cm ⌃ d =
=
cm
1
2n
n
Now, for mode m to propagate, we require
⌅
2nd
0.4
=
m
m
⌃ m⌅
0.4
=
0.4
=4
0.1
So, accounting for 2 modes (TE and TM) for each value of m, and the single TEM mode, we will
have a total of 9 modes.
13.12. A parallel-plate guide is to be constructed for operation in the TEM mode only over the frequency
range 0 < f < 3 GHz. The dielectric between plates is to be teflon (⌅⌅r = 2.1). Determine the
maximum allowable plate separation, d: We require that f < fc1 , which, using (41), becomes
f<
c
2nd
⌃ dmax =
c
3 ⇤ 108
= ⌦
= 3.45 cm
2nfmax
2 2.1 (3 ⇤ 109 )
13.13. A lossless parallel-plate waveguide is known to propagate the m = 2 TE and TM modes at
frequencies as low as 10GHz. If the plate separation is 1 cm, determine the dielectric constant of
the medium between plates: Use
fc2 =
c
3 ⇤ 1010
=
= 1010 ⌃ n = 3 or ⌅r = 9
nd
n(1)
13.14. A d = 1 cm parallel-plate guide is made with glass (n = 1.45) between plates. If the operating
frequency is 32 GHz, which modes will propagate? For a propagating mode, we require f > fcm
Using (41) and the given values, we write
f>
mc
2nd
⌃ m<
2f nd
2(32 ⇤ 109 )(1.45)(.01)
=
= 3.09
c
3 ⇤ 108
The maximum allowed m in this case is thus 3, and the propagating modes will be TM1 , TE1 ,
TM2 , TE2 , TM3 , and TE3 .
265
13.15. For the guide of Problem 13.14, and at the 32 GHz frequency, determine the di⌅erence between the
group delays of the highest order mode (TE or TM) and the TEM mode. Assume a propagation
distance of 10 cm: From Problem 13.14, we found mmax = 3. The group velocity of a TE or TM
mode for m = 3 is
◆
⇤
⌅2
c
fc3
3(3 ⇤ 1010 )
vg3 =
1
where fc3 =
= 3.1 ⇤ 1010 = 31 GHz
n
f
2(1.45)(1)
Thus
vg3 =
3 ⇤ 10
1.45
10
◆
⇤
1
31
32
⌅2
= 5.13 ⇤ 109 cm/s
For the TEM mode (assuming no material dispersion) vg,T EM = c/n = 3 ⇤ 1010 /1.45 = 2.07 ⇤ 1010
cm/s. The group delay di⌅erence is now
⇥tg = z
⇤
1
vg3
1
vg,T EM
⌅
= 10
⇤
1
5.13 ⇤ 109
1
2.07 ⇤ 1010
⌅
= 1.5 ns
13.16. The cuto⌅ frequency of the m = 1 TE and TM modes in an air-filled parallel-plate guide is
known to be fc1 = 7.5 GHz. The guide is used at wavelength
= 1.5 cm. Find the group
velocity of the m = 2 TE and TM modes. First we know that fc2 = 2fc1 = 15 GHz. Then
f = c/ = 3 ⇤ 108 /.015 = 20 GHz. Now, using (57),
vg2 =
c
n
◆
⇤
1
fc2
f
⌅2
=
c
(1)
◆
1
⇤
15
20
⌅2
= 2 ⇤ 108 m/s
13.17. A parallel-plate guide is partially filled with two lossless dielectrics (Fig. 13.25) where ⌅⌅r1 = 4.0,
⌅⌅r2 = 2.1, and d = 1 cm. At a certain frequency, it is found that the TM1 mode propagates through
the guide without su⌅ering any reflective loss at the dielectric interface.
a) Find this frequency: The ray angle is such that the wave is incident on the interface at
Brewster’s angle. In this case
⌃B = tan
The ray angle is thus ⌃ = 90
1
✓
2.1
= 35.9⇤
4.0
35.9 = 54.1⇤ . The cuto⌅ frequency for the m = 1 mode is
fc1 =
c
3 ⇤ 1010
⇣ ⌅ =
= 7.5 GHz
2(1)(2)
2d ⌅r1
The frequency is thus f = fc1 / cos ⌃ = 7.5/ cos(54.1⇤ ) = 12.8 GHz.
b) Is the guide operating at a single TM mode at the frequency found in part a? The cuto⌅
frequency for the next higher mode, TM2 is fc2 = 2fc1 = 15 GHz. The 12.8 GHz operating
frequency is below this, so TM2 will not propagate. So the answer is yes.
266
13.18. In the guide of Figure 13.25, it is found that m = 1 modes propagating from left to right totally
reflect at the interface, so that no power is transmitted into the region of dielectric constant ⌅⌅r2 .
a) Determine the range of frequencies over which this will occur: For total reflection, the ray
angle measured from the normal to the interface must be greater than or equal to the critical
angle, ⌃c , where sin ⌃c = (⌅⌅r2 /⌅⌅r1 )1/2 . The maximum mode ray angle is then ⌃1 max = 90⇤ ⌃c .
Now, using (39), we write
90
⇤
⌃c = cos
1
⇤
⌦
kmax d
⌅
= cos
1
⇤
⌦
2⌦fmax d 4
Now
cos(90
⌦c
⌃c ) = sin ⌃c =
◆
⌅
= cos
1
⇤
c
4dfmax
⌅
⌅⌅r2
c
=
⌅⌅r1
4dfmax
⌦
⌦
Therefore fmax = c/(2 2.1d) = (3 ⇤ 108 )/(2 2.1(.01)) = 10.35 GHz. The frequency range is
thus f < 10.35 GHz.
b) Does your part a answer in any way
⌦ relate to the cuto⌅ frequency for m = 1 modes in any
region? We note that fmax = c/(2 2.1d) = fc1 in guide 2. To summarize, as frequency is
lowered, the ray angle in guide 1 decreases, which leads to the incident angle at the interface
increasing to eventually reach and surpass the critical angle. At the critical angle, the refracted
angle in guide 2 is 90⇤ , which corresponds to a zero degree ray angle in that guide. This defines
the cuto⌅ condition in guide 2. So it would make sense that fmax = fc1 (guide 2).
13.19. A rectangular waveguide has dimensions a = 6 cm and b = 4 cm.
a) Over what range of frequencies will the guide operate single mode? The cuto⌅ frequency for
mode mp is, using Eq. (101):
fc,mn
c
=
2n
✓
m ⇥2
p ⇥2
+
a
b
where n is the refractive index of the guide interior. We require that the frequency lie between
the cuto⌅ frequencies of the T E10 and T E01 modes. These will be:
fc10 =
c
3 ⇤ 108
2.5 ⇤ 109
=
=
2na
2n(.06)
n
fc01 =
c
3 ⇤ 108
3.75 ⇤ 109
=
=
2nb
2n(.04)
n
Thus, the range of frequencies for single mode operation is 2.5/n < f < 3.75/n GHz
b) Over what frequency range will the guide support both T E10 and T E01 modes and no others?
We note first that f must be greater than fc01 to support both modes, but must be less than
the cuto⌅ frequency for the next higher order mode. This will be fc11 , given by
fc11
c
=
2n
◆⇤
1
.06
⌅2
+
⇤
1
.04
⌅2
=
The allowed frequency range is then
3.75
4.5
GHz < f <
GHz
n
n
267
30c
4.5 ⇤ 109
=
2n
n
13.20. Two rectangular waveguides are joined end-to-end. The guides have identical dimensions, where
a = 2b. One guide is air-filled; the other is filled with a lossless dielectric characterized by ⌅⌅r .
a) Determine the maximum allowable value of ⌅⌅r such that single mode operation can be simultaneously ensured in both guides at some frequency: Since a = 2b, the cuto⌅ frequency for any
mode in either guide is written using (101):
fcmp =
✓
mc ⇥2
pc ⇥2
+
4nb
2nb
⇣
where n = 1 in guide 1 and n =
⌅⌅r in guide 2. We see that, with a = 2b, the next
modes (having the next higher cuto⌅ frequency) above TE10 with be TE20 and TE01 . We
also see that in general, fcmp (guide 2) < fcmp (guide 1). To assure single mode operation
in both guides, the operating frequency must be above cuto⌅ for TE10 in both guides, and
below cuto⌅ for the next mode in both guides. The allowed frequency
range is therefore
⇣
fc10 (guide 1) < f < fc20 (guide 2). This leads to c/(2a) < f < c/(a ⌅⌅r ). For this range to be
viable, it is required that ⌅⌅r < 4.
b) Write an expression for the frequency range over which single mode operation will occur in
both guides; your answer should be in terms of ⌅⌅r , guide dimensions as needed, and other
known constants: This was already found in part a:
c
c
<f < ⇣ ⌅
2a
⌅r a
where ⌅⌅r < 4.
13.21. An air-filled rectangular waveguide is to be constructed for single-mode operation at 15 GHz.
Specify the guide dimensions, a and b, such that the design frequency is 10% higher than the
cuto⌅ frequency for the TE10 mode, while being 10% lower than the cuto⌅ frequency for the next
higher-order mode: For an air-filled guide, we have
fc,mp =
✓
mc ⇥2
pc ⇥2
+
2a
2b
For TE10 we have fc10 = c/2a, while for the next mode (TE01 ), fc01 = c/2b. Our requirements
state that f = 1.1fc10 = 0.9fc01 . So fc10 = 15/1.1 = 13.6 GHz and fc01 = 15/0.9 = 16.7 GHz.
The guide dimensions will be
a=
c
3 ⇤ 1010
c
3 ⇤ 1010
=
=
1.1
cm
and
b
=
=
= 0.90 cm
2fc10
2(13.6 ⇤ 109 )
2fc01
2(16.7 ⇤ 109 )
268
13.22. Using the relation Pav = (1/2)Re{Es ⇤ H⇥s }, and Eqs. (106) through (108), show that the average
power density in the TE10 mode in a rectangular waveguide is given by
Pav =
⇥10 2 2
E sin (⌥10 x) az
2 µ 0
W/m2
Inspecting (106) through (108), we see that (108) includes a factor of j, and so would lead to an
imaginary part of the total power when the cross product with Ey is taken. Therefore, the real
power in this case is found through the cross product of (106) with the complex conjugate of (107),
or
1
⇥10 2 2
Pav = Re {Eys ⇤ H⇥xs } =
E sin (⌥10 x) az
W/m2
2
2 µ 0
13.23. Integrate the result of Problem 13.22 over the guide cross-section 0 < x < a, 0 < y < b, to show
that the power in Watts transmitted down the guide is given as
P =
⇥10 ab 2
ab 2
E0 =
E sin ⌃10 W
4 µ
4⇧ 0
⇣
where ⇧ = µ/⌅, and ⌃10 is the wave angle associated with the TE10 mode. Interpret. First, the
integration:
b
a
⇥10 2 2
⇥10 ab 2
P =
E0 sin (⌥10 x) az · az dx dy =
E
4 µ 0
0
0 2 µ
⌦
Next, from (54), we have ⇥10 =
µ⌅ sin ⌃10 , which, on substitution, leads to
ab 2
P =
E sin ⌃10 W with ⇧ =
4⇧ 0
✓
µ
⌅
The sin ⌃10 dependence demonstrates the principle of group velocity as energy velocity (or power).
This was considered in the discussion leading to Eq. (57).
13.24. Show that the group dispersion parameter, d2 ⇥/d
gular waveguide is given by
d2 ⇥
=
d 2
n
c
c
2
, for given mode in a parallel-plate or rectan-
⇥2 ⇧
1
c
⇥2 ⌃
3/2
where c is the radian cuto⌅ frequency for the mode in question (note that the first derivative
form was already found, resulting in Eq. (57)). First, taking the reciprocal of (57), we find
⇧
d⇥
n
=
1
d
c
c
⇥2 ⌃
Taking the derivative of this equation with respect to
d2 ⇥
n
=
2
d
c
⇤
1
2
⌅⇧
1
c
⇥2 ⌃
3/2
⇤
2
269
2
c
3
⌅
1/2
leads to
=
n
c
c
⇥2 ⇧
1
c
⇥2 ⌃
3/2
13.25. Consider a transform-limited pulse of center frequency f = 10 GHz and of full-width 2T = 1.0 ns.
The pulse propagates in a lossless single mode rectangular guide which is air-filled and in which the
10 GHz operating frequency is 1.1 times the cuto⌅ frequency of the T E10 mode. Using the result of
Problem 13.24, determine the length of the guide over which the
⇣ pulse broadens to twice its initial
⌅
width: The broadened pulse will have width given by T = T 2 + (⇥ )2 , where ⇥ = ⇥2 L/T
for a transform limited pulse (assumed gaussian). ⇥2 is the Problem 13.24 result evaluated at the
operating frequency, or
d2 ⇥
⇥2 =
|
d 2
=10 GHz
= 6.1 ⇤ 10
19
=
1
10
(2⌦ ⇤ 10 )(3 ⇤ 108 )
s2 /m = 0.61 ns2 /m
⇤
1
1.1
⌅2 ⌥
1
⇤
1
1.1
⌅2
3/2
Now ⇥ = 0.61L/0.5 = 1.2L ns. For the pulse width to double, we have T ⌅ = 1 ns, and
⇣
(.05)2 + (1.2L)2 = 1 ⌃ L = 0.72 m = 72 cm
What simple step can be taken to reduce the amount of pulse broadening in this guide, while
maintaining the same initial pulse width? It can be seen that ⇥2 can be reduced by increasing the
operating frequency relative to the cuto⌅ frequency; i.e., operate as far above cuto⌅ as possible,
without allowing the next higher-order modes to propagate.
13.26. A symmetric dielectric slab waveguide has a slab thickness d = 10 µm, with n1 = 1.48 and
n2 = 1.45. If the operating wavelength ⇣
is = 1.3 µm, what modes will propagate? We use the
condition expressed through (141): k0 d n21 n22 ⇧ (m 1)⌦. Since k0 = 2⌦/ , the condition
becomes
⌘
2d
2(10) ⇣
n21 n22 ⇧ (m 1) ⌃
(1.48)2 (1.45)2 = 4.56 ⇧ m 1
1.3
Therefore, mmax = 5, and we have TE and TM modes for which m = 1, 2, 3, 4, 5 propagating (ten
total).
13.27. A symmetric slab waveguide is known to support only a single pair of TE and TM modes at
wavelength
= 1.55 µm. If the slab thickness is 5 µm, what is the maximum value of n1 if
n2 = 3.30? Using (142) we have
2⌦d
⌘
n21
n22
< ⌦ ⌃ n1 <
◆⇤
2d
⌅2
+
n22
=
◆⇤
1.55
2(5)
⌅2
+ (3.30)2 = 3.304
13.28. In a symmetric slab waveguide, n1 = 1.50, n2 = 1.45, and d = 10 µm.
a) What is the phase velocity of the m = 1 TE or TM mode at cuto⌅? At cuto⌅, the mode
propagates in the slab at the critical angle, which means that the phase velocity will be equal
to that of a plane wave in the upper or lower media of index n2 . Phase velocity will therefore
be vp (cuto⌅) = c/n2 = 3 ⇤ 108 /1.45 = 2.07 ⇤ 108 m/s.
b) What is the phase velocity of the m = 2 TE or TM modes at cuto⌅? The reasoning of part a
applies to all modes, so the answer is the same, or 2.07 ⇤ 108 m/s.
270
13.29. An asymmetric slab waveguide is shown in Fig. 13.26. In this case, the regions above and below
the slab have unequal refractive indices, where n1 > n3 > n2 .
a) Write, in terms of the appropriate indices, an expression for the minimum possible wave angle,
⌃1 , that a guided mode may have: The wave angle must be equal to or greater than the critical
angle of total reflection at both interfaces. The minimum wave angle is thus determined by
the greater of the two critical angles. Since n3 > n2 , we find ⌃min = ⌃c,13 = sin 1 (n3 /n1 ).
b) Write an expression for the maximum phase velocity a guided mode may have in this structure,
using given or known parameters: We have vp,max = /⇥min , where ⇥min = n1 k0 sin ⌃1,min =
n1 k0 n3 /n1 = n3 k0 . Thus vp,max = /(n3 k0 ) = c/n3 .
13.30. A step index optical fiber is known to be single mode at wavelengths > 1.2 µm. Another fiber
is to be fabricated from the same materials, but is to be single mode at wavelengths > 0.63 µm.
By what percentage must the core radius of the new fiber di⌅er from the old one, and should it be
larger or smaller? We use the cuto⌅ condition, given by (159):
⌘
2⌦a
> c=
n21 n22
2.405
With reduced, the core radius, a, must also be reduced by the same fraction. Therefore, the
percentage reduction required in the core radius will be
%=
1.2 .63
⇤ 100 = 47.5%
1.2
13.31. Is the mode field radius greater than or less than the fiber core radius in single-mode step-index
fiber?
The answer to this can be found by inspecting Eq. (164). Clearly the mode field radius
decreases with increasing V , so we can look at the extreme case of V = 2.405, which is the
upper limit to single-mode operation. The equation evaluates as
↵0
1.619
2.879
= 0.65 +
+
= 1.10
3/2
a
(2.405)6
(2.405)
Therefore, ↵0 is always greater than a within the single-mode regime, V < 2.405.
13.32. The mode field radius of a step-index fiber is measured as 4.5 µm at free space wavelength = 1.30
µm. If the cuto⌅ wavelength is specified as c = 1.20 µm, find the expected mode field radius at
= 1.55 µm.
In this problem it is helpful to use the relation V = 2.405( c / ), and rewrite Eq. (164) to
read:
⇤ ⌅3/2
⇤ ⌅6
↵0
= 0.65 + 0.434
+ 0.015
a
c
c
At = 1.30 µm, / c = 1.08, and at 1.55 µm, / c = 1.29. Using these values, along with
our new equation, we write
⇧
⌃
0.65 + 0.434(1.29)3/2 + 0.015(1.29)6
↵0 (1.55) = 4.5
= 5.3 µm
0.65 + 0.434(1.08)3/2 + 0.015(1.08)6
271
CHAPTER 14
14.1. A short dipole carrying current I0 cos ↵t in the az direction is located at the origin in free space.
a) If k = 1 rad/m, r = 2 m, ⌅ = 45⇤ , = 0, and t = 0, give a unit vector in rectangular components that shows the instantaneous direction of E: In spherical coordinates, the components
of E are given by (13a) and (13b):
Ers =
E
=
s
I0 d
⇤ cos ⌅ e
2
I0 d
⇤ sin ⌅ e
4
1
1
+
r2
jkr3
(13a)
jk
1
1
+ 2+
r
r
jkr3
(13b)
jkr
jkr
Since we want a unit vector at t = 0, we need only the relative amplitudes of the two components, but we need the absolute phases. Since ⌅ = 45⇤ , sin ⌅ = cos ⌅ = 1/ 2. Also, with
k = 1 = 2 /⇧, it follows that ⇧ = 2 m. The above two equations can be simplified by these
substitutions, while dropping all amplitude terms that are common to both. Obtain
1
1
+ 3
2
r
jr
Ar =
A =
1
2
j
e
jr
1
1
1
+ 2+ 3
r r
jr
e
jr
Now with r = 2 m, we obtain
Ar =
A =
j
1
4
1 1
+
4 8
j
1
8
e
j
1
16
j2
=
e
1
(1.12)e
4
j2
=
j26.6
e
j2
1
(0.90)ej56.3 e
4
j2
The total vector is now A = Ar ar + A a . We can normalize the vector by first finding the
magnitude:
1⌘
|A| = A · A⇥ =
(1.12)2 + (0.90)2 = 0.359
4
Dividing the field vector by this magnitude and converting 2 rad to 114.6⇤ , we write the
normalized vector as
AN s = 0.780e j141.2 ar + 0.627e 58.3 a
In real instantaneous form, this becomes
⇥
AN (t) = Re AN s ej⌃t = 0.780 cos(↵t
141.2⇤ )ar + 0.627 cos(↵t
58.3⇤ )a
We evaluate this at t = 0 to find
AN (0) = 0.780 cos(141.2⇤ )ar + 0.627 cos(58.3⇤ )a =
272
0.608ar + 0.330a
14.1 a) (continued)
⌘
Dividing by the magnitude, (0.608)2 + (0.330)2 = 0.692, we obtain the unit vector at t = 0:
aN (0) = 0.879ar + 0.477a . We next convert this to rectangular components:
aN x = aN (0) · ax =
0.879 sin ⌅ cos + 0.477 cos ⌅ cos
aN y = aN (0) · ay =
aN z = aN (0) · az =
=
1
( 0.879 + 0.477) =
2
0.879 sin ⌅ sin + 0.477 cos ⌅ sin
0.879 cos ⌅
The final result is then
0.477 sin ⌅ =
aN (0) =
0.284ax
= 0 since
1
( 0.879
2
0.477) =
0.284
=0
0.959
0.959az
b) What fraction of the total average power is radiated in the belt, 80⇤ < ⌅ < 100⇤ ? We use the
far-zone phasor fields, (22) and (23), with k = 2 /⇧, and first find the average power density:
Pavg =
1
I 2 d2 ⇤
⇥
Re[E s H⇧s
] = 0 2 2 sin2 ⌅ W/m2
2
8⇧ r
We integrate this over the given belt, an at radius r:
2⌅
Pbelt =
0
100
80
I02 d2 ⇤
I02 d2 ⇤
2
2
sin
⌅
r
sin
⌅
d⌅
d
=
8⇧2 r2
4⇧2
100
sin3 ⌅ d⌅
80
Evaluating the integral, we find
Pbelt
I 2 d2 ⇤
= 0 2
4⇧
⌦
↵
⇥ 100
1
I 2 d2 ⇤
2
cos ⌅ sin ⌅ + 2
= (0.344) 0 2
3
4⇧
80
The total power is found by performing the same integral over ⌅, where 0 < ⌅ < 180⇤ . Doing
this, it is found that
I 2 d2 ⇤
Ptot = (1.333) 0 2
4⇧
The fraction of the total power in the belt is then f = 0.344/1.333 = 0.258.
273
14.2. For the Hertzian dipole, prepare a curve, r vs. ⌅ in polar coordinates, showing the locus in the
= 0 plane where:
a) The radiation field |E s | is one-half of its value at r = 104 m, ⌅ = /2: Assuming the far field
approximation, we use (22) with k = 2 /⇧ to set up the equation:
|E s | =
I0 d⇤
1
I0 d⇤
sin ⌅ = ⇤
2⇧r
2 2 ⇤ 104 ⇧
⌥ r = 2 ⇤ 104 sin ⌅
b) The average radiated power density, Pr,av , is one-half of its value at r = 104 m, ⌅ = /2. To
find the average power, we use (22) and (23) in
Pr,av =
1
1 I02 d2 ⇤
1 1 I02 d2 ⇤
2
⇥
Re{E s H⇧s
}=
sin
⌅
=
⇤
2
2 4⇧2 r2
2 2 4⇧2 (108 )
The polar plots for field (r = 2 ⇤ 104 sin ⌅) and power (r =
Both are circles.
274
⌥ r=
2 ⇤ 104 sin ⌅
2 ⇤ 104 sin ⌅) are shown below.
14.3. Two short antennas at the origin in free space carry identical currents of 5 cos ↵t A, one in the az
direction, one in the ay direction. Let ⇧ = 2 m and d = 0.1 m. Find Es at the distant point:
a) (x = 0, y = 1000, z = 0): This point lies along the axial direction of the ay antenna, so its
contribution to the field will be zero. This leaves the az antenna, and since ⌅ = 90⇤ , only the
E s component will be present (as (136) and (137) show). Since we are in the far zone, (138)
applies. We use ⌅ = 90⇤ , d = 0.1, ⇧ = 2 , ⇤ = ⇤0 = 120 , and r = 1000 to write:
Es = E s a = j
= j(1.5 ⇤ 10
I0 d⇤
sin ⌅e
2⇧r
2
)e
j1000
j2⌅r/⇥
a =
a =j
5(0.1)(120 )
e
4 (1000)
j(1.5 ⇤ 10
2
)e
j1000
j1000
a
az V/m
b) (0, 0, 1000): Along the z axis, only the ay antenna will contribute to the field. Since the
distance is the same, we can apply the part a result, modified such the the field direction is
in ay : Es = j(1.5 ⇤ 10 2 )e j1000 ay V/m
c) (1000, 0, 0): Here, both antennas will contribute. Applying the results of parts a and b, we
find Es = j(1.5 ⇤ 10 2 )(ay + az ).
d) Find E at (1000, 0, 0) at t = 0: This is found through
⇥
E(t) = Re Es ej⌃t = (1.5 ⇤ 10
Evaluating at t = 0, we find
E(0) = (1.5 ⇤ 10 2 )[ sin(1000)](ay + az ) =
2
) sin(↵t
(1.24 ⇤ 10
2
1000)(ay + az )
)(ay + az ) V/m.
e) Find |E| at (1000, 0, 0) at t = 0: Taking the magnitude of the part d result, we find |E| =
1.75 ⇤ 10 2 V/m.
275
14.4. Write the Hertzian dipole electric field, whose components are given in Eqs. (15) and (16), in the
near-zone in free space, where kr << 1. In this case, only a single term in each of the two equations
survives, and the phases, r and , simplify to a single value. Construct the resulting electric field
vector and compare your result to the static dipole result (Eq. (35) (not (36)) in Chapter 4). What
relation must exist between the static dipole charge, Q, and the current amplitude, I0 , so that the
two results are identical?
First, we evaluate the phase terms under the approximation, kr << 1: Using (17b) and (18)
we have
.
.
1
(0)
=
and
= tan 1 (
) =
r = tan
2
2
2
These are used in (15) and (16). In those equations, the (kr)
the (kr) 4 term is kept in (16). The results are
2
term is retained in (15) and
.
Ers =
I0 d
⇤ cos ⌅ exp( j /2) =
2 r2 (kr)
j
I0 ⇤d
cos ⌅
2 kr3
.
=
I0 kd
⇤ sin ⌅ exp( j /2) =
4 r(kr)2
j
I0 ⇤d
sin ⌅
4 kr3
E
s
where, in free space
⌘
µ0 /⇥0
⇤
1
=
=
k
↵ µ0 ⇥0
↵⇥0
The field vector can now be constructed as
Es =
( jI0 /↵)d
[2 cos ⌅ ar + sin ⌅ a ]
4 ⇥0 r3
This expression is equivalent to the static dipole field (Eq. (35) in Chapter 4), provided that
I0 = j↵Q, which is just a statement that the current is the time derivative of the time-harmonic
charge.
276
14.5. Consider the term in Eq. (14) (or in Eq. (10)) that gives the 1/r2 dependence in the Hertzian
dipole magnetic field. Assuming this term dominates and that kr << 1, show that the resulting
magnetic field is the same as that found by applying the Biot-Savart law (Eq. (2), Chapter 7) to
a current element of di⇤erential length d, oriented along the z axis, and centered at the origin.
We begin by writing Eq. (14) under the condition kr << 1, for which the phase term in (14)
becomes, using Eq. (17a):
exp[ j(kr
⇧ )]
= exp[ j(kr
tan
1
.
(kr))] = exp[ j(kr
kr)] = 0
With kr << 1, the 1/(kr)2 term in (14) dominates, and the field becomes:
I0 d
. I0 kd 1
H⇧ =
sin ⌅ exp(j0) =
sin ⌅
4 r (kr)
4 r2
Now, using the Biot-Savart law for a short element of assumed di⇤erential length, d, we have
I0 dL ⇤ aR
4 R2
H=
where, with the element at the origin and oriented along z, we have R = r, aR = ar , and
dL = daz . With these subsitutions:
H=
I0 d (az ⇤ ar )
I0 d
=
sin ⌅ a⇧
2
4 r
4 r2
(done)
14.6. Evaluate the time-average Poynting vector, < S >= (1/2)Re {Es ⇤ H⇥s } for the Hertzian dipole,
assuming the general case that involves the field components as given by Eqs. (10), (13a), and
(13b). Compare your result to the far-zone case, Eq. (26).
With radial and theta components of E, and with only a phi component of H, the Poynting
vector becomes
⇧
⌃
1
⇥
⇥
< S >= Re E s H⇧s
ar
Ers H⇧s
a
2
Substituting (10), (13a) and (13b), this becomes:
1
< S > = Re
2
2
⌦
k2
k
⇤ sin ⌅ 2 j 3
r
r
⌦
(I0 d)2
k
1
1
sin ⌅ cos ⌅ j 3
+ 4
2
4
8
r
r
r
I0 d
4
2
1
k
1
+j 3 + 4
r4
r
r
↵
k
j 5 a
r
↵
k
j 5 ar
r
In taking the real part, the imaginary terms that do not cancel are removed, other real terms
cancel, and only the first term in the radial component remains. The final result is
< S >=
1
2
I0 d k
4 r
which we see to be identical to Eq. (26).
277
2
⇤ sin ar W/m2
14.7. A short current element has d = 0.03⇧. Calculate the radiation resistance for each of the following
current distributions:
a) Uniform: In this case, (30) applies directly and we find
Rrad = 80
2
2
d
⇧
= 80
2
(.03)2 = 0.711 ⇥
b) Linear: I(z) = I0 (0.5d |z|)/0.5d: Here, the average current is 0.5I0 , and so the average
power drops by a factor of 0.25. The radiation resistance therefore is down to one-fourth the
value found in part a, or Rrad = (0.25)(0.711) = 0.178 ⇥.
c) Step: I0 for 0 < |z| < 0.25d and 0.5I0 for 0.25d < |z| < 0.5d: In this case the average current
on the wire is 0.75I0 . The radiated power (and radiation resistance) are down to a factor of
(0.75)2 times their values for a uniform current, and so Rrad = (0.75)2 (0.711) = 0.400 ⇥.
14.8. Evaluate the time-average Poynting vector, < S >= (1/2)Re {Es ⇤ H⇥s } for the magnetic dipole
antenna in the far-zone, in which all terms of order 1/r2 and 1/r4 are neglected in Eqs. (48), (49),
and (50). Compare your result to the far-zone power density of the Hertzian dipole, Eq. (26). In
this comparison, and assuming equal current amplitudes, what relation between loop radius, a,
and dipole length, d, would result in equal radiated powers from the two devices?
First, neglecting the indicated terms in (48)-(50), only the terms in 1/r survive. We find:
.
E⇧s =
j
↵µ0 a2 I0 k
sin ⌅ exp[ j(kr
4 r
. ↵µ0 a2 I0 k 1
=j
sin ⌅ exp[ j(kr
4 r
⇤
.
where, in the approximation, ⇧ = = tan 1 (kr). So now
H
s
< S >= (1/2)Re {E⇧s ⇤ H⇥s } =
⇧ )]
)]
1
Re {E⇧s H ⇥s } ar
2
Using the above field expressions, we find
| < S > | = Sr(md) =
1
2
2
I0 k
4 r
[↵µ0 a2 ]2
1
sin2 ⌅
⇤
which we now compare to the power density from the Hertzian dipole, Eq. (26):
Sr(Hd) =
1
2
I0 k
4 r
2
d2 ⇤ sin2 ⌅
For equal power densities from the two structures, we thus require (assuming free space):
[↵µ0 a2 ]2
1
↵µ0
= d2 ⇤ ⌥ d = a2
= a2 (↵ µ0 ⇥0 ) = a2
⇤
⇤
2
⇧
or
d=
(2 a)2
2⇧
(the square of the loop circumference over twice the wavelength)
278
14.9. A dipole antenna in free space has a linear current distribution. If the length is 0.02⇧, what value
of I0 is required to:
a) provide a radiation-field amplitude of 100 mV/m at a distance of one mile, at ⌅ = 90⇤ : With
a linear current distribution, the peak current, I0 , occurs at the center of the dipole; current
decreases linearly to zero at the two ends. The average current is thus I0 /2, and we use Eq.
(138) to write:
|E | =
I0 d⇤0
I0 (0.02)(120 )
sin(90⇤ ) =
= 0.1 ⌥ I0 = 85.4 A
4⇧r
(4)(5280)(12)(0.0254)
b) radiate a total power of 1 watt? We use
Pavg =
1
4
1 2
I Rrad
2 0
where the radiation resistance is given by Eq. (140), and where the factor of 1/4 arises from
the average current of I0 /2: We obtain Pavg = 10 2 I02 (0.02)2 = 1 ⌥ I0 = 5.03 A.
14.10. Show that the chord length in the E-plane plot of Fig. 14.4 is equal to b sin ⌅, where b is the circle
diameter.
Referring to the construction below, half the chord length, c/2, is given by
c
b
= cos(90
2
2
⌅) =
b
sin ⌅ ⌥ c = b sin ⌅ (done)
2
279
14.11. A monopole antenna in free space, extending vertically over a perfectly conducting plane, has a
linear current distribution. If the length of the antenna is 0.01⇧, what value of I0 is required to
a) provide a radiation field amplitude of 100 mV/m at a distance of 1 mi, at ⌅ = 90⇤ : The image
antenna below the plane provides a radiation pattern that is identical to a dipole antenna of
length 0.02⇧. The radiation field is thus given by (22) in free space, where k = 2 /⇧, ⌅ = 90⇤ ,
and with an additional factor of 1/2 included to account for the linear current distribution:
|E | =
1 I0 d⇤0
2 2⇧r
⌥ I0 =
4r|E |
4(5289)(12 ⇤ .0254)(100 ⇤ 10
=
(d/⇧)⇤0
(.02)(377)
3
)
= 85.4 A
b) radiate a total power of 1W: For the monopole over the conducting plane, power is radiated
only over the upper half-space. This reduces the radiation resistance of the equivalent dipole
antenna by a factor of one-half. Additionally, the linear current distribution reduces the
radiation resistance of a dipole having uniform current by a factor of one-fourth. Therefore,
Rrad is one-eighth the value obtained from (30), or Rrad = 10 2 (d/⇧)2 . The current magnitude
is now
⌦
↵1/2 ⌦
↵1/2
2Pav
2(1)
2
I0 =
=
=
= 7.1 A
Rrad
10 2 (d/⇧)2
10 (.02)
14.12. Find the zeros in ⌅ for the E-plane pattern of a dipole antenna for which (using Fig. 14.8 as a
guide):
a)
= ⇧: We look for zeros in the pattern function, Eq. (59), for which k = (2 /⇧)⇧ = 2 .
F (⌅) =
⌦
↵ ⌦
↵
cos(k cos ⌅) cos(k )
cos(2 cos ⌅) cos(2 )
=
sin ⌅
sin ⌅
Zeros will occur whenever
cos(2 cos ⌅) = 1 ⌥ ⌅ = (0, 90⇤ , 180⇤ , 270⇤ )
You are encouraged to show (using small argument approximations for cosine and sine) that
F (⌅) does approach zero when ⌅ approaches zero or 180⇤ (despite the sin ⌅ term in the denominator).
b) 2 = 1.3⇧: In this case k = (2 /⇧)(1.3⇧/2) = 1.3 . The pattern function becomes
F (⌅) =
⌦
cos(1.3 cos ⌅) cos(1.3 )
sin ⌅
↵
Zeros of this function will occur at ⌅ = (0, 57.42⇤ , 122.58⇤ , 180⇤ ).
280
14.13. The radiation field of a certain short vertical current element is E
it is located at the origin in free space.
a) Find E
find
s
at P (r = 100, ⌅ = 90⇤ ,
E
s
=
s
= (20/r) sin ⌅ e
j10⌅r
V/m if
= 30⇤ ): Substituting these values into the given formula,
20
sin(90⇤ )e
100
j10⌅(100)
= 0.2e
j1000⌅
V/m
b) Find E s at P if the vertical element is located at A(0.1, 90⇤ , 90⇤ ): This places the element
on the y axis at y = 0.1. As a result of moving the antenna from the origin to y = 0.1, the
change in distance to point P is negligible when considering the change in field amplitude, but
is not when considering the change in phase. Consider lines drawn from the origin to P and
from y = 0.1 to P . These lines can be considered essentially parallel, and so the di⇤erence in
.
their lengths is l = 0.1 sin(30⇤ ), with the line from y = 0.1 being shorter by this amount. The
construction and arguments are similar to those used in the discussion of the electric dipole
in Sec. 4.7. The electric field is now the result of part a, modified by including a shorter
distance, r, in the phase term only. We show this as an additional phase factor:
E
s
= 0.2e
j1000⌅ j10⌅(0.1 sin 30
e
= 0.2e
j1000⌅ j0.5⌅
e
V/m
c) Find E s at P if identical elements are located at A(0.1, 90⇤ , 90⇤ ) and B(0.1, 90⇤ , 270⇤ ): The
original element of part b is still in place, but a new one has been added at y = 0.1. Again,
constructing a line between B and P , we find, using the same arguments as in part b, that
the length of this line is approximately 0.1 sin(30⇤ ) longer than the distance from the origin
to P . The part b result is thus modified to include the contribution from the second element,
whose field will add to that of the first:
E
s
= 0.2e
j1000⌅
ej0.5⌅ + e
j0.5⌅
⇥
= 0.2e
j1000⌅
2 cos(0.5 ) = 0
The two fields are out of phase at P under the approximations we have used.
14.14. For a dipole antenna of overall length 2 = ⇧, evaluate the maximum directivity in dB, and the
half-power beamwidth.
Dmax is found using Eq. (64) with k = , and involves a numerical integration. This I did
using a Mathematica code to find Dmax = 2.41, and in decibels this is 10 log10 (2.41) = 3.82 dB,
with the maximum occurring in the direction ⌅ = 90⇤ .
With k = , the pattern function, Eq. (59), reaches a maximum value of 2 at ⌅ = 90⇤ . The
half-power beamwidth is found numerically by setting the pattern function equal to 2 and
thus solving
⌦
↵
cos( cos ⌅) cos( )
F (⌅) =
= 2
sin ⌅
The result is ⌅ = 66.1⇤ , which means that the deviation angle from the direction along the
maximum is ⌅1/2 /2 = 90 66.1 = 23.9⇤ . The half-power beamwidth is therefore ⌅1/2 =
2 ⇤ 23.9 = 47.8⇤ .
281
14.15. For a dipole antenna of overall length 2 = 1.3⇧, determine the locations in ⌅ and the peak
intensity of the sidelobes, expressed as a fraction of the main lobe intensity. Express your result
as the sidelobe level in decibels, given by Ss [dB] = 10 log10 (Sr,main /Sr,sidelobe ). Again, use Fig.
14.8 as a guide.
Here, k = (2 /⇧)(1.3⇧/2) = 1.3 . The angular intensity distribution is given by the square
of the pattern function, Eq. (59), with k = 1.3 substituted:
|<S>|
⌦
cos(1.3 cos ⌅) cos(1.3 )
[F (⌅)] =
sin ⌅
2
↵2
To find the maxima, the easiest way is to create a polar plot of the function, as shown in
Fig. 14.8, and then numerically determine the relative intensities of the peaks by searching
for local maxima at locations indicated in the plot. The main lobe occurs at ⌅ = 90⇤ , and has
value [F (⌅)]2 = 2.52. The secondary maxima occur at ⌅ = 33.8⇤ and 146.2⇤ , and both peaks
have values [F (⌅)]2 = 0.47. The sidelobe level in decibels is thus
Ss [dB] = 10 log10
282
2.52
0.47
= 7.3 dB
14.16. For a dipole antenna of overall length, 2 = 1.5⇧:
a) Evaluate the locations in ⌅ at which the zeros and maxima in the E-plane pattern occur: In
this case, k = (2 /⇧)(1.5⇧/2) = 1.5 . The angular intensity distribution is given by the
square of the pattern function, Eq. (59), with k = 1.3 substituted:
|<S>|
⌦
cos(1.5 cos ⌅) cos(1.5 )
[F (⌅)] =
sin ⌅
2
↵2
⌦
cos(1.5 cos ⌅)
=
sin ⌅
↵2
Zeros occur whenever cos(1.5 cos ⌅) = 0, which in turn occurs whenever 1.5 cos ⌅ is an odd
multiple of /2 (positive or negative). All zeros over the range 0 ⇧ ⌅ ⇧ are found with just
the first two odd multiples, or
1
cos ⌅z = ± , ±1 ⌥ ⌅z = (0, 70.5⇤ , 109.5⇤ , 180⇤ )
3
The function does indeed zero at ⌅ = 0 and 180⇤ , despite the presence of the sin ⌅ term in
the denominator, as can be demonstrated using small argument approximations in the trig
functions (show this!). The maxima are found numerically, as was done in Problem 14.15. The
results are ⌅max,m = (42.6⇤ , 137.4⇤ ) (main lobes) and ⌅max,s = 90⇤ (sidelobe). The pattern in
plotted below.
b) Determine the sidelobe level, as per the definition in Problem 14.15: At the main lobe angle
(⌅ = 42.6⇤ ), [F (⌅)]2 evaluates as 1.96. The sidelobe at ⌅ = 90⇤ gives a value of 1.00. The
sidelobe level in decibles is thus
Ss [dB] = 10 log10
1.96
1.00
= 2.9 dB
c) Determine the maximum directivity: This is found evaluating Eq. (64) numerically (which I
did using a Mathematica code), and the result is Dmax = 2.23, or in decibels this becomes
10 log10 (2.23) = 3.5 dB.
283
14.17. Consider a lossless half-wave dipole in free space, with radiation resistance, Rrad = 73 ohms, and
maximum directivity Dmax = 1.64. If the antenna carries a 1-A current amplitude,
a) How much total power (in watts) is radiated? We use the definition of radiation resistance,
Eq. (29), and write the radiated power:
Pr =
1 2
1
I0 Rrad = (1)2 (73) = 36.5 W
2
2
b) How much power is intercepted by a 1 m2 aperture situated at distance r = 1 km away. The
aperture is on the equatorial plane and squarely faces the antenna. Assume uniform power
density over the aperture: The maximum directivity is given by Dmax = 4 Kmax /Pr , where
the maximum radiation intensity, Kmax , is related to the power density in W/m2 at distance
r from the antenna as Sr = Kmax /r2 . The power intercepted by the aperture is then the
power density times the aperture area, or
Prec = Sr ⇤ Area =
Dmax Pr
1.64(36.5) 2
⇤ Area =
(1) = 4.77 µW
2
4 r
4 (103 )2
14.18. Repeat Problem 14.17, but with a full-wave antenna (2 = ⇧). Numerical integrals may be necessary. In this case, k = , and the radiation resistance is found by applying Eq. (65):
⌅
Rrad = 60
0
⌦
cos( cos ⌅) cos( )
sin ⌅
↵2
sin ⌅ d⌅ = 199 ohms
where the answer was found by numerical integration. With the same 1-A current amplitude as
before, the total radiated power is now
Pr =
1 2
1
I0 Rrad = (1)2 (199) = 99.5 W
2
2
The maximum directivity is the result of Problem 14.14 (involving a numerical integration in Eq.
(64) with k = ), and the result was found to be Dmax = 2.41. The received power is now found
as in Problem 14.17:
Prec =
Dmax Pr
2.41(99.5) 2
⇤ Area =
(1) = 19.1 µW
4 r2
4 (103 )2
284
14.19. Design a two-element dipole array that will radiate equal intensities in the = 0, /2, , and 3 /2
directions in the H-plane. Specify the smallest relative current phasing, ⌥, and the smallest element
spacing, d.
The array function is given by Eq. (81), with n = 2:
⌥
⌥
1 ⌥⌥ sin(⌦) ⌥⌥
|A2 (⌦)| = ⌥
2 sin(⌦/2) ⌥
This has periodic maxima occurring at ⌦ = 0, ±2m , and in the H-plane, ⌦ = ⌥ + kd cos .
Now, for broadside operation, we have maxima at
= ± /2, at which ⌦ = ⌥, so we set
⌥ = 0 to get the array function principal maximum at ⌦ = 0. Having done this, we now have
⌦ = kd cos , which we need to have equal to ±2 when = 0, , in order to get endfire
operation. This will happen when kd = 2 d/⇧ = 2 , so we set d = ⇧.
14.20. A two-element dipole array is configured to provide zero radiation in the broadside ( = ±90⇤ )
and endfire ( = 0, 180⇤ ) directions, but with maxima occurring at angles in between. Consider
such a set-up with ⌦ = at = 0 and ⌦ = 3 at = , with both values determined in the
H-plane.
a) Verify that these values give zero broadside and endfire radiation: For two elements, the array
function is given by Eq. (81), with n = 2:
⌥
⌥
1 ⌥⌥ sin(⌦) ⌥⌥
|A2 (⌦)| = ⌥
2 sin(⌦/2) ⌥
This is clearly zero at ⌦ =
and at ⌦ =
3 .
b) Determine the required relative current phase, ⌥: We know that ⌦ = ⌥ + kd cos
H-plane. Setting up the two given conditions, we have:
in the
= ⌥ + kd cos(0) = ⌥ + kd
and
3 = ⌥ + kd cos( ) = ⌥
kd
Adding these two equations gives ⌥ =
, which means that the contributions from the two
elements will in fact completely cancel in the broadside direction, regardless of the element
spacing.
c) Determine the required element spacing, d: We can now use the first condition, for example,
substituting values that we know:
⌦( = 0) =
=
+ kd cos(0) =
+ kd
Therefore, kd = 2 /⇧ = 2 , or d = ⇧.
d) Determine the values of at which maxima in the radiation pattern occur: With the values
as found, the array function now becomes
⌥
⌥
1 ⌥⌥ sin[ (2 cos
1)] ⌥⌥
|A2 ( )| = ⌥
2 sin[ ⌅2 (2 cos
1)] ⌥
This function maximizes at
= 60⇤ , 120⇤ , 240⇤ , and 300⇤ , as found numerically.
285
14.21. In the two-element endfire array of Example 14.4, consider the e⇤ect of varying the operating
frequency, f , away from the original design frequency, f0 , while maintaining the original current
phasing, ⌥ =
/2. Determine the values of at which the maxima occur when the frequency is
changed to
a) f = 1.5f0 : The original phase and length from the example are ⌥ =
/2 and d = ⇧0 /4,
where ⇧0 = c/f0 . Thus kd = (2 /⇧0 )(⇧0 /4) = /2. The H-plane array function in that case
was found to be
⌦ ↵
⌦
↵
⌥
✏
⇣
⌦
⌥
kd
⌥
A( /2, )⌥ = cos
= cos
+
cos
= cos
+ cos
2
2
2
4
4
⇥0
Now, if we change the frequency to 1.5f0 , while leaving ⌥ fixed (and the antenna physical
length is unchanged), we obtain ⇧ = ⇧0 /1.5, and kd = (2 /⇧0 )(1.5)(⇧0 /4) = 3 /4. The new
array function is then:
⌥
⌥
A( /2, )⌥
⇥0 /1.5
= cos
⌦
⌥
kd
+
cos
2
2
↵
= cos
⌦
This function amplitude maximizes when 3 cos
4
+
↵
3
cos
8
2 = 0, or
= cos
= cos
1
✏
8
(3 cos
⇣
2)
(2/3) = ±48.2⇤ .
b) f = 2f0 : Now, ⇧ = ⇧0 /2 and kd = (2 /⇧0 )(2)(⇧0 /4) = . The array function becomes
⌥
⌥
A( /2, )⌥
⇥0 /2
⌦
⌥
kd
= cos
+
cos
2
2
↵
= cos
This function amplitude maximizes when 2 cos
286
✏
4
+
2
cos
1 = 0, or
⇣
= cos
= cos
1
✏
4
(2 cos
⇣
1)
(1/2) = ±60.0⇤ .
14.22. Revisit Problem 14.21, but with the current phase allowed to vary with frequency (this will automatically occur if the phase di⇤erence is established by a simple time delay between the feed
currents). Now, the current phase di⇤erence will be ⌥ ⌅ = ⌥ f /f0 , where f0 is the original (design)
frequency. Under this condition, radiation will maximize in the = 0 direction regardless of frequency (show this). Backward radiation (along = ) will develop however as the frequency is
tuned away from f0 . Derive an expression for the front-to-back ratio, defined as the ratio of the
radiation intensities at = 0 and = , expressed in decibels. Express this result as a function
of the frequency ratio f /f0 .
With the current phase and wavenumber both changing with frequency, we would have
⌦⌅ =
With ⌥ =
f
f0
⌥+
f
f0
kd cos
/2 and d = ⇧0 /4, The H-plane array function at the original frequency, f0 , is
⌥
⌥
A( /2, )⌥
f0
⌦ ↵
⌦
↵
✏
⌦
⌥
kd
= cos
= cos
+
cos
= cos (cos
2
2
2
4
⇣
1)
The e⇤ect on this of changing the frequency is then
⌦ ⌅↵
⌦
⌥
⌦
f
⌥
A( /2, )⌥ = cos
= cos
2
f0
f
⌥
kd
+
cos
2
2
↵
= cos
⌦
f
f0
4
(cos
↵
1)
The condition for the zero argument (which maximizes A) is (cos
1) = 0 This term is
una⇤ected by changing the frequency, and so the array function will always maximize in the
= 0 direction.
In the forward direction ( = 0), the value of |A|2 (proportional to the radiation intensity) is
unity. In the backward direction ( = 180⇤ ), we find
|A( = )|2 = cos2
f
f0 2
= cos2
f
2f0
The front-to-back ratio is then
Rf b = 10 log10
⌦
↵
⌦
↵
|A( = 0)|2
1
= 10 log10
|A( = )|2
cos2 ( f /2f0 )
Evaluate the front-to-back ratio for a) f = 1.5f0 , b) f = 2f0 , c) f = 0.75f0 . Substituting
these values, we find:
⌦
↵
1
f = 1.5f0 : Rf b = 10 log10
= 3 dB
cos2 (1.5 /2)
⌦
↵
1
f = 2f0 : Rf b = 10 log10
= 0 dB
cos2 ( )
⌦
↵
1
f = 0.75f0 : Rf b = 10 log10
= 8.3 dB
cos2 ( /2.67)
287
14.23. A turnstile antenna consists of two crossed dipole antennas, positioned in this case in the xy plane.
The dipoles are identical, lie along the x and y axes, and are both fed at the origin. Assume
that equal currents are supplied to each antenna, and that a zero phase reference is applied to
the x-directed antenna. determine the relative phase, ⌥, of the y-directed antenna so that the net
radiated electric field as measured on the positive z axis is a) left circularly-polarized; b) linearly
polarized along the 45⇤ axis between x and y.
When looking at the field along the z axis, the expression for E can be constructed using Eq.
(57), evaluated at ⌅ = /2 (so that we consider the direction normal to the antenna), and in
which r is replaced by z. An additional “array” term includes the two polarization directions
(unit vectors) which are out of phase by ⌥:
E(z, ⌅ = /2) = J
I0 ⇤
[1
2 z
cos(k )] e
jkz
⇤
⌅
ax + ay ej⇤
a) To achieve left circular polarization for propagation in the forward z direction, the vector
array function must be
⇤
⌅
ax + ay ej⇤ = [ax + jay ] ⌥ ⌥ = /2
b) To achieve 45⇤ polarization, the vector array function must be
⇤
⌅
ax + ay ej⇤ = [ax ± ay ] ⌥ ⌥ = 0,
14.24. Consider a linear endfire array, designed for maximum radiation intensity at = 0, using ⌥ and d
values as suggested in Example 14.5. Determine an expression for the front-to-back ratio (defined
in Problem 14.22) as a function of the number of elements, n, if n is an odd number.
From Example 14.5, we use ⌥ =
/2 and d = ⇧/4, so that kd = /2. Then in the H-plane
⌦ = ⌥ + kd cos
=
2
(cos
1)
The array function is then:
A(⌦) =
⌦
↵
⌦
1 sin(n⌦/2)
1 sin[(n /4)(cos
=
n sin(⌦/2)
n sin[( /4)(cos
1)]
1)]
↵
In the forward direction ( = 0), A = 1. In the backward direction ( = ), A = ±1/n, where
n is an odd integer. The front-to-back ratio is therefore:
Rf b = 10 log10
|A( = 0)|2
|A( = )|2
= 10 log10 (n2 ) dB
in which performance is clearly better as n increases.
288
14.25. A six-element linear dipole array has element spacing d = ⇧/2.
a) Select the appropriate current phasing, ⌥, to achieve maximum radiation along
With d = ⇧/2, we have kd = and in the H-plane,
⌦ = ⌥ + kd cos
= ±60⇤ :
= ⌥ + cos
We set the principal maximum to occur at ⌦ = 0, and with = 60⇤ , the condition for this
becomes
⌦ = 0 = ⌥ + cos(60⇤ ) = ⌥ +
⌥ ⌥=
2
2
b) With the phase set as in part a, evaluate the intensities (relative to the maximum) in the
broadside and endfire directions. With the above result, the array function (Eq. (81)) in the
H-plane becomes
⌦
↵
⌦
↵
1
sin[3(⌥ + kd cos )]
1 sin[(3 /2)(2 cos
1)]
A( /2, ) =
=
6 sin[(1/2)(⌥ + kd cos )]
6 sin[( /4)(2 cos
1)]
In the broadside direction ( = 90⇤ ) the array function becomes
⌦
↵
1 sin[ 3 /2]
2
A( /2, /2) =
=
6 sin[ /4]
6
The intensity in this direction (relative to that along
= 60⇤ ) is the square of this or 1/18
In the endfire direction ( = 0⇤ ) the array function becomes
⌦
↵
1 sin[3 /2]
2
A( /2, 0) =
=
6 sin[ /4]
6
The intensity in this direction (relative to that along = 60⇤ ) is the square of this or 1/18,
as before. The same result is found for = 180⇤ and for = 90⇤ . The intensity pattern is
shown below:
289
14.26. In a linear endfire array of n elements, a choice of current phasing that improves the directivity is
given by the Hansen-Woodyard condition:
⌥=±
2 d
+
⇧
n
where the plus or minus sign choices give maximum radiation along = 180⇤ and 0⇤ respectively.
Applying this phasing may not necessarily lead to unidirectional endfire operation (zero backward
radiation), but will do so with the proper choice of element spacing, d.
a) Determine this required spacing as a function of n and ⇧: We first construct the expression
for ⌦ in the H-plane, using the given current phase expression, and in which the minus sign is
chosen:
2 d
2 d
⌦ = ⌥ + kd cos =
+
cos
⇧
n
⇧
Note that ⌦( = 0) =
/n. The array function, Eq. (81) is:
⌦
↵
⌦
↵
1 sin(n⌦/2)
1 sin[(n d/⇧)(cos
1)
/2]
A(⌥, ) =
=
n sin(⌦/2)
n sin[( d/⇧)(cos
1) ( /2n)]
This function is equal to 1 when
function becomes
= 0, regardless of the choice of d. When
⌦
↵
1 sin[( 2n d/⇧) ( /2)]
A(⌥, ) =
n sin[( 2 d/⇧) ( /2n)]
which we require to be zero. This will happen when ⌦( = ) =
for ⌦, we set up the condition:
⌦( = ) =
4 d
⇧
n
=
⌥ d=
= , the array
. Using the above formula
1
n
n
⇧
4
b) Show that the spacing as found in part a approaches ⇧/4 for a large number of elements: It
is readily seen from the part a result that when n ⌃ , d ⌃ ⇧/4.
c) Show that an even number of elements is required: This can be shown by substituting d as
found in part a into the array function at = . The result is
⌦
↵
1 sin( n /2)
A(⌥, ) =
n sin( /2)
We require this result to be zero, which will only happen if n is an even integer. This can be
more quickly seen from the general expression:
⌦
↵
1 sin(n⌦/2)
A(⌥, ) =
n sin(⌦/2)
wherein if ⌦ =
, we will get a zero result only if n is even.
290
14.27. Consider an n-element broadside linear array. Increasing the number of elements has the e⇤ect
of narrowing the main beam. Demonstrate this by evaluating the separation in between the
zeros on either side of the principal maximum at = 90⇤ . Show that for large n this separation is
.
.
approximated by
= 2⇧/L, where L = nd is the overall length of the array.
For broadside operation, ⌥ = 0, and, with kd = 2 d/⇧, we have
⌦ = ⌥ + kd cos
=
2 d
cos
⇧
With this condition, the array function, Eq. (81) is:
⌦
↵
⌦
↵
1 sin(n⌦/2)
1 sin[(n d/⇧) cos ]
A(⌥, ) =
=
n sin(⌦/2)
n sin[( d/⇧) cos ]
Zeros in this function will occur whenever
n d
cos
⇧
= ±m
where m is an integer. Now, on either side of the principal maximum at ⌦ = 0, zeros will
occur at + and
(giving a 2 separation), which we can associate with the two angles +
and
respectively, which lie on either side of 90⇤ . We can therefore write:
n d
cos
⇧
+
cos
Using a trig identity, this becomes:
⌦
2n d
1
sin (
⇧
2
+
+
◆
⇥
=2
↵
⌦
1
) sin (
2
+
↵
) =2
1
Because we are considering the broadside direction, we have (
first sine term is just unity. We find
⌦
1
sin (
2
+
+
+
)/2 = /2 and so the
↵
⇧
) =
nd
Now as n gets large, both sides of the equation become << 1, so that we may approximate
the sine term by just its argument:
1
(
2
Defining
=
+
+
. ⇧
)=
nd
, the above expression becomes
. 2⇧ . 2⇧
=
=
(done)
nd
L
291
14.28. A large ground-based transmitter radiates 10kW, and communicates with a mobile receiving station
that dissipates 1mW on the matched load of its antenna. The receiver (not having moved) now
transmits back to the ground station. If the mobile unit radiates 100W, what power is received
(at a matched load) by the ground station?
We can use Eq. (93) and the fact that Z21 = Z12 to write:
PL2
|Z21 |2
|Z12 |2
PL1
=
=
=
Pr1
R11 R22
R11 R22
Pr2
We have
PL2
10 3
=
= 10
Pr1
104
So
PL1 = Pr2 ⇤ 10
7
= 100 ⇤ 10
7
7
=
PL1
Pr2
= 10
5
= 10 µW
14.29. Signals are transmitted at a 1m carrier wavelength between two identical half-wave dipole antennas
spaced by 1km. The antennas are oriented such that they are exactly parallel to each other.
a) If the transmitting antenna radiates 100 watts, how much power is dissipated by a matched
load at the receiving antenna? To find the dissipated power, use Eq. (106):
PL2 =
Pr1 ⇧2
D1 (⌅1
(4 r)2
1 ) D2 (⌅2 2 )
Since the antennas face each other squarely, the directivities are both the maximum values for
the two, where for the half-wave dipole, we know that Dmax = 1.64 So
PL2 =
100(1)2
(1.64)2 = 1.7 µW
(4 ⇤ 103 )2
b) Suppose the receiving antenna is tilted by 45⇤ while the two antennas remain in the same plane.
What is the received power in this case? With the receiving antenna tilted, its directivity is
reduced accordingly. Using Eq. (63), with the help of the results of Example 14.2, we have
⌥
⌥
⌥ cos[( /2) cos(45⇤ )] ⌥2
⌥
⌥ = 0.643
D2 (⌅2 = 45 ) = Dmax ⇤ |F (45 )| = 1.64 ⇤ ⌥
⌥
sin(45⇤ )
⇤
So now,
⇤
PL2 (45⇤ ) =
2
100(1)2
(1.64)(0.643) = 672 nW
(4 ⇤ 103 )2
292
14.30. A half-wave dipole antenna is known to have a maximum e⇤ective area, given as Amax .
a) Write the maximum directivity of this antenna in terms of Amax and wavelength ⇧: From Eq.
(104), it follows that
4
Dmax = 2 Amax
⇧
b) Express the current amplitude, I0 , needed to radiate total power, Pr , in terms of Pr , Amax ,
and ⇧: Combining Eqs. (64) and (65), we can write
2Pr
120[F (⌅)]2max
=
I02
Dmax
where, for a half-wave dipole, [F (⌅)]2max = 1. Then, using the part a result, we find
I0 =
✓
Pr Amax
15⇧2
c) At what values of ⌅ and will the antenna e⇤ective area be equal to Amax ? These will be the
same angles over which the half-wave dipole directivity maximizes, or at which F (⌅) for the
antenna reaches its maximum. We know from Example 14.2 that this happens at ⌅ = /2,
and at all values of .
293