Chapter 7

Magnetostatic Fields

In electrostatics, the source of the field E is stationary charges described by ρv . In magnetostatics, the source

of the field B (in Tesla) is uniformly moving charges described by a steady current J. Note that, in general,

a steady current J gives rise to an electric field, in addition to the magnetic field. However, we will confine

our attention here to the magnetic field only. This subject is called magnetostatics. Here, we learn how

to compute the magnetic field field using (i) the so-called Biot-Savart’s law, (ii) Ampere’s law, and (iii) the

method of potentials. We then discuss magnetic dipoles.

7.1 Maxwell Equations of Magnetostatics

Maxwell equations of magnetostatics in differential form are


∇ × B = µ0 J


, (7.1)

∇•B=0



where µ0 = 4π × 10−7 H/m is the permeability of free space, and in integral form are
I 
B • dl = µ0 Ienc 


L , (7.2)
I

B dS = 0
•


S

where Ienc is the current enclosed by L. The equation ∇ × B = µ0 J implies that B does rotate (i.e., makes

closed loops) around its source, and the equation ∇ • B = 0 implies that B does not diverge from its source.

If the source occupies a region V ′ , and that region is suspended in an unbounded region, solution can be

1
ELEC 325 Hamad M. Alkhoori Fall 2024

found as
J(r′ ) × R
Z
µ0
B(r) = dV ′ , (7.3)
4π V′ R3

Notice that the input is a vector, and the output is another vector. Then, the Green function (not shown

clearly from the above equation), can’t be a scalar or a vector, by a dyadic.

For a surface current density K residing on a surface S ′ , the magnetic field is

K(r′ ) × R ′
Z
µ0
B(r) = dS , (7.4)
4π S′ R3

For a curve L′ that carries a current I, the magnetic field is found as

I(r′ ) × R ′
Z
µ0
B(r) = dl . (7.5)
4π L′ R3

Notice that we did not mention the magnetic field of a point charge Q moving with a uniform velocity v,

as it requires some background from special relativity (see Sec. 10.3 from Introduction to Electrodynamics

by D. J. Griffiths). As a summary, if v << c, where c is the speed of light, the magnetic field is found as

µ0 Qv × Ro
B(r) ≈ , (7.6)
4π Ro3

where Ro = r − ro . Notice that ro here is a constant vector.

Example 1: The line z ∈ [−L, L] carries a uniform current I = I0 âz . (a) Obtain B at any point in
dz ′ (z − z ′ )
Z
space. Hint: 2 ′ 2 3/2
= − p . (b) Obtain B for an infinite line. Hint:
[ρ + (z − z ) ] ρ2 ρ2 + (z − z ′ )2
Z∞
dz ′ 2
= 2.
[ρ + (z − z ′ )2 ]3/2
2 ρ
−∞
Ans: (a)

I(r′ ) × R ′ I×R ′
Z Z
µ0 µ0
B= 3
dl = dl
4π L′ R 4π L′ R3

dl′ = dz ′ âz ⇒ dl′ = dz ′

r = xâx + yây + zâz = ρâρ + zâz , r′ = z ′ âz

R = ρâρ + (z − z ′ )âz ⇒ R3 = [ρ2 + (z − z ′ )2 ]3/2

I × R = I0 âz × [ρâρ + (z − z ′ )âz ] = I0 ρâϕ

ZL " #
µ0 I0 ρ dz ′ µ0 I0 (z + L) (z − L)
⇒ B = âϕ = âϕ −p
[ρ2 + (z − z ′ )2 ]3/2
p
4π 4πρ ρ2 + (z + L)2 ρ2 + (z − L)2
−L

2
ELEC 325 Hamad M. Alkhoori Fall 2024

(b)
Z∞
µ0 I0 ρ dz ′ µ0 I0 ρ 2 µ0 I0
B = âϕ = âϕ = âϕ .
4π [ρ2 + (z − z ′ )2 ]3/2 4π ρ2 2πρ
−∞

Example 2: A loop of a radius a on the z = 0 plane carries a uniform current I = I0 â′ϕ . Obtain B at point

(0, 0, z).

Ans:
I(r′ ) × R ′ I×R ′
Z Z
µ0 µ0
B= 3
dl = dl
4π L′ R 4π L′ R3

dl′ = ρ′ dϕ′ â′ϕ ⇒ dl′ = ρ′ dϕ′

r = zâz , r′ = ρ′ â′ρ

R = −ρ′ â′ρ + zâz ⇒ R3 = (ρ′2 + z 2 )3/2

I × R = I0 â′ϕ × (−ρ′ â′ρ + zâz ) = I0 (zâ′ρ + ρ′ âz )

Z2π Z2π
µ0 I0 zâ′ρ + ρ′ âz ′ ′ µ0 I0 a zâ′ρ + aâz ′ µ0 I0 a2
⇒B= ρ dϕ = dϕ = âz
4π (ρ′2 + z 2 )3/2 ρ′ =a 4π (a2 + z 2 )3/2 2(a2 + z 2 )3/2
0 0

Example 3: The solenoid {ρ = a, z ∈ [−L, L]} has N turns, and carries a uniform current I = I0 â′ϕ . (a)

Obtain B at any point on the z axis where |z| > L. (b) Obtain B for an infinite solenoid.

Ans: (a) With the assumption that the turns are very close to each others, we can think of the problem
N
as a finite cylinder carrying a surface current K = I0 â′ϕ = nI0 â′ϕ on the surface ρ = a. Let us denote the
2L
surface z = L by S1 , the surface ρ = a by S2 , and the surface z = −L by S3 .

K(r′ ) × R ′
Z 
K×R ′ K×R ′
Z Z Z Z
µ0 µ0 µ0
B= 3
dS = 0+ dS + 0 = dS
4π S′ R 4π S1 S2 R3 S3 4π S2 R3

dS′ = ρ′ dϕ′ dz ′ â′ρ ⇒ dS ′ = ρ′ dϕ′ dz ′

r = zâz , r′ = ρ′ â′ρ + z ′ âz

R = −ρ′ â′ρ + (z − z ′ )âz ⇒ R3 = [ρ′2 + (z − z ′ )2 ]3/2

K × R = nI0 â′ϕ × [−ρ′ â′ρ + (z − z ′ )âz ] = nI0 [ρ′ âz + (z − z ′ )â′ρ ]

ZL Z2π ZL Z2π
µ0 nI0 ρ′ âz + (z − z ′ )â′ρ ′ ′ ′ µ0 nI0 a aâz + (z − z ′ )â′ρ
⇒B= ρ dϕ dz = dϕ′ dz ′
4π [ρ′2 + (z − z ′ )2 ]3/2 ρ′ =a 4π [a2 + (z − z ′ )2 ]3/2
−L 0 −L 0
ZL " #
µ0 nI0 a2 dz ′ µ0 nI (z + L) (z − L)
= âz = âz −p
[a2 + (z − z ′ )2 ]3/2
p
2 2 a2 + (z + L)2 a2 + (z − L)2
−L

3
ELEC 325 Hamad M. Alkhoori Fall 2024

(b) Since the expression we got is valid only along the z axis, and now the source is infinite along that

axis, the field we will find here is inside the solenoid along the z axis.
Z∞
µ0 nI0 a2 dz ′
B = âz = âz µ0 nI0
2 [a2 + (z − z ′ )2 ]3/2
−∞

Example 4: The infinite plane z = 0 carries a uniform surface current density K = K0 âx . Obtain B at
Z∞ Z∞ Z∞ Z∞
y dxdy dxdy 2π
point (0, 0, z > 0). Hint: 2 2 2 3/2
= 0 and 2 2 2 3/2
= .
(x + y + z ) (x + y + z ) z
−∞ −∞ −∞ −∞
Ans:
K(r′ ) × R ′ K×R ′
Z Z
µ0 µ0
B= dS = dS
4π S′ R3 4π S′ R3

dS′ = dx′ dy ′ âz ⇒ dS ′ = dx′ dy ′

r = zâz , r′ = x′ âx + y ′ ây

R = −x′ âx − y ′ ây + zâz ⇒ R3 = (x′2 + y ′2 + z 2 )3/2

K × R = K0 (−y ′ âz − zây )

Z∞ Z∞
µ0 K0 −y ′ âz − zây µ0 K0
⇒B= dx′ dy ′ = − ây
4π (x′2 + y ′2 + z 2 )3/2 2
−∞ −∞

In general, an infinite plane with a unit normal ân carrying a uniform current K will give a field

µ0
B= K × ân .
2

Further Readings:

ˆ Section 5.2 in “Introduction to Electrodynamics by David”.

ˆ Section 7.2 in “Elements of Electromagnetics by Matthew”. Note that the author uses H instead of

B
B. In free space, H = µ0 . I will show the role of H when we get to magnetostatic fields in mediums.

7.2 Ampere’s Law

Ampere found that for any path L that encloses a current Ienc , the magnetic field can be related to that

current by
I
B(r) • dl = µ0 Ienc . (7.7)
L

4
ELEC 325 Hamad M. Alkhoori Fall 2024

Note that Ienc is the total current enclosed by L. Equation (7.7) is the integral form of Ampere’s law.

It is true for any current distribution, but it is useful to obtain B only if symmetries present (just like the

integral form of Gauss’s law). In other words, if the current distribution and the shape of the object that

carries the current are such that B is constant along L, then it can be taken outside the integral sign, and

hence, it can be evaluated.

Although Ampere’s law has been obtained experimentally, it can be derived

ˆ by a completely theoretical approach (relativistic approach), or

ˆ from Biot-Savart law.

The integral form of Ampere’s law can be used as follows:

1. Determine the independent variables and the direction of B.

2. From the direction of B, choose your amperian path L accordingly. If the direction is âx , ây , or âz ,

then L will be a rectangular path. If the direction is âϕ , then L will be a circular path.

H
3. Evaluate L
B(r) • dl and Ienc .

H
4. Find B from L
B(r) • dl = µ0 Ienc .

Example 5: Find B for an infinite line that carries a uniform current I.

Ans: By intuition, B must be along âϕ , and thus, the amperian path here will be a circle of radius ρ.

Hence,
I
B • dl = 2πρB
L

Ienc = I

µ0 I
2πρB = µ0 I ⇒ B =
2πρ
µ0 I
⇒B= âϕ .
2πρ
Example 6: Three infinite cylinders of radii a, b, and c are coaxial as shown below. The cylinder of radius
I
a carries a uniform volume current density J = âz . The cylinder of radius c carries a uniform volume
πa2
I
current density J = − âz . Find B in all regions.
π(c2 − b2 )

5
ELEC 325 Hamad M. Alkhoori Fall 2024

Ans:
I
B • dl = 2πρB
L
Z2πZρ
ρ2
Z
I
0 < ρ < a : Ienc = J • dS = âz • âz ρdρdϕ = I
S πa2 a2
0 0
2
ρ µ0 Iρ µ0 Iρ
2πρB = µ0 I 2 ⇒ B = 2
⇒B= âϕ
a 2πa 2πa2
Z Z2πZa
I
a < ρ < b : Ienc = J • dS = âz • âz ρdρdϕ = I
S πa2
0 0
µ0 I µ0 I
2πρB = µ0 I ⇒ B = ⇒B= âϕ
2πρ 2πρ
Z2πZa Z2πZρ
ρ2 − b2
Z Z  
I I
b < ρ < c : Ienc = J dS +
• J dS =
• âz âz ρdρdϕ +
• − âz âz ρdρdϕ = I 1 − 2
•

S S πa2 π(c2 − b2 ) c − b2
0 0 0 b
ρ 2 − b2 ρ 2 − b2 ρ2 − b2
     
µ0 I µ0 I
2πρB = µ0 I 1 − 2 ⇒B= 1− 2 ⇒B= 1− 2 âϕ
c − b2 2πρ c − b2 2πρ c − b2
Z Z Z2πZa Z2πZc
I I
c < ρ < ∞ : Ienc = J • dS + J • dS = 2
âz • âz ρdρdϕ + − âz • âz ρdρdϕ = 0
S S πa π(c − b2 )
2
0 0 0 b
⇒B=0

µ0 Iρ
âϕ , 0<ρ<a


2πa2




µ0 I


âϕ , a<ρ<b



⇒B= 2πρ 
ρ 2 − b2

 µ0 I

 1− 2 âϕ , b < ρ < c
2πρ c − b2






c<ρ<∞

 0,

Example 7: A toroid of N turns carries a uniform current I. Find B in all regions.

6
ELEC 325 Hamad M. Alkhoori Fall 2024

Ans: I
B • dl = 2πρB
L

0 < ρ < a : Ienc = 0 ⇒ B = 0

µ0 N I
a < ρ < b : Ienc = N I ⇒ B = âϕ
2πρ

b < ρ < ∞ : Ienc = N I − N I = 0 ⇒ B = 0





 0, 0<ρ<a


µ0 N I

⇒B= âϕ , a < ρ < b


 2πρ


 0, b<ρ<∞


Example 8: For the infinite solenoid of radius a problem, we found the field inside as B = µ0 nI âz . Find B

outside.

Ans: Since I ∥ âϕ , then B ∥ âz . Hence, the amperian path will be a rectangle {ρ ∈ [b, c], z ∈ [0, L]}, where
H
c > b > a. In general, we could assume B = B(ρ)âz . Then, B • dl = [B(c) − B(b)]L. But since Ienc = 0,

we get [B(c) − B(b)]L = 0 ⇒ B(b) = B(c). This indicates that B will not depend on ρ, and hence, is a

constant. But very far away from the solenoid, B must be zero. So, since B(∞) = 0, then B(b) = B(c) = 0.

Therefore, 

 µ0 nI âz , 0 < ρ < a

⇒B=

 0,
 a<ρ<∞

Further Reading:

ˆ Section 7.4 in “Elements of Electromagnetics by Matthew”.

Further Practice:

7
ELEC 325 Hamad M. Alkhoori Fall 2024

ˆ Example 7.5 on page 304.

ˆ Practice exercise 7.5 on page 305.

ˆ End of chapter problems: 7.19, 7.22, and 7.28.

7.3 Magnetic Vector Potential

In electrostatics, ∇ × E = 0 allowed us to write E = −∇V , where V is a scalar potential. Then, from

∇ • E = −ρv /ε0 , we got ∇2 V = −ρv /ε0 . This equation is a scalar partial differential equation, whereby

obtaining its solution is, in principles, simpler than obtaining the solution of a vector partial differential

equation. The electric field can be obtained simply by taking the gradient of V .

In magnetostatics, ∇ • B = 0 allows us to write B as

B = ∇ × A, (7.8)

where A is called the magnetic vector potential. Substituting this back into the differential form of

Ampere’s law gives

∇ × ∇ × A = µ0 J, (7.9)

which is a complicated vector partial differential equation. If we let ∇ • A = 0, and we are permitted to do

this, Eq. (7.9) becomes

∇2 A = −µ0 J, (7.10)

which is a less complicated vector partial differential equation. Because we are in free space and if no

boundaries exist in the space surrounding the source J, the solution to Eq. (7.10) can be written as

J(r′ )
Z
µ0
A(r) = dV ′ , (7.11)
4π V′ R

where V ′ is a region containing J. For a surface current density K on a surface S ′ , A is found from

K(r′ ) ′
Z
µ0
A(r) = dS . (7.12)
4π S′ R

Finally, for a line current I along a path L′ , A is found from

I(r′ ) ′
Z
µ0
A(r) = dl . (7.13)
4π L′ R

8
ELEC 325 Hamad M. Alkhoori Fall 2024

The magnetic field can be obtained from B = ∇ × A.

7.4 Magnetic Dipole

For an arbitrary source J contained in a region V ′ , the magnetic vector potential is given by

J(r′ )
Z
µ0
A(r) = dV ′ . (7.14)
4π V′ R

With the assumption that r lies along the z axis, the term 1/R can be expanded as

∞  ′ n  ′
  ′

1 1X ′ r 1 ′ ′ r 1 ′r
= Pn (cos θ ) = P0 (cos θ ) + P1 (cos θ ) + . . . = 1 + cos θ + ... (7.15)
|r − r′ | r n=0 r r r r r

If |r| >> |r′ |, we may consider the first two terms and ignore the rest. Therefore,
Z Z
µ0 ′ ′ µ0
A≈ J(r ) dV + [r′ cos θ′ J(r′ )] dV ′
4πr V′ 4πr2 V ′
Z Z (7.16)
µ0 µ0
= J(r′ ) dV ′ + âr
• [r′ J(r′ )] dV ′ ,
4πr V ′ 4πr2 V′

where the quantity J(r′ )r′ is a 2nd-rank tensor. The first integral can be regarded as a monopole term

and the second integral as a dipole term.

For any source J, it happens that the first integral vanishes. Hence, magnetic monopole does not exist

theoretically. The second integral can be rearranged as

 Z 
µ0 1 ′ ′ ′
A(r) ≈ − âr × r × J(r ) dV . (7.17)
4πr2 2 V′

Upon defining m as
Z
1
m= r′ × J(r′ ) dV ′ , (7.18)
2 V′

Eq. (7.17) can be written as

µ0 m × âr
A(r) ≈ . (7.19)
4π r2

The quantity m is called the magnetic dipole moment. It it the magnetic analog to the electric dipole

moment p. Since field and source points do not enter into the expression of m, we just use unprimed

variables. Therefore,
Z
1
m= r × J(r) dV. (7.20)
2 V

9
ELEC 325 Hamad M. Alkhoori Fall 2024

For a surface current K, m is given by

Z
1
m= r × K(r) dS. (7.21)
2 S

Finally, for a line current I, m is given by

Z
1
m= r × I(r) dl. (7.22)
2 L

For line currents with closed loops, Eq. (7.22) can be written as

Z
m= I(r) dS, (7.23)
S

where S is the area bounded by L.

Example 9: For a small loop of radius a carrying a uniform current I situated at the z = 0 plane, find m,

A, and B.

Ans:
Z Z2πZa
m= I(r) dS = I âz ρdρdϕ = I(πa2 )âz = m0 âz
S
0 0

µ0 m × âr µ0 m0 µ0 m0 µ0 m0 sin θ
A= = (âz × âr ) = [(cos θâr − sin θâθ ) × âr ] = âϕ
4π r2 4πr2 4πr2 4πr2
µ0 m0
B=∇×A= (2 cos θâr + sin θâθ ).
4πr3
A two-dimensional plot of the magnetic field lines is shown below.

Example 10: Find m for the small rectangle {z = 0, x ∈ [0, a], y ∈ [0, b]} carrying a uniform current I.

Ans:
Z Zb Za
m= I(r) dS = I âz dxdy = Iabâz .
S
0 0

10