Jackson 5.

3 Homework Problem Solution

PROBLEM:
A right-circular solenoid of finite length L and radius a has N turns per unit length and carries a current
I. Show that the magnetic induction on the cylinder axis in the limit NL → ∞ is

0 N I
Bz= cos  1cos  2 
2

where the angles are defined in the figure.

θ1 θ2
z

SOLUTION:
Let us first find the on-axis magnetic induction produced by one circular loop of current. Place the
observation point such that the line from the observation point to the loop makes an angle θi with the
solenoid's axis and is a distance ri away. In this set-up, then the observation point can be considered at
the origin of a spherical coordinate system.

ri
θi
z

If the current is flowing into the paper at the bottom and out of the paper at the top, the current density
in spherical coordinates becomes:

 r −r i 
J= I 
 −i  
r

Plug this into the Biot-Savart Law:

 0 J x '×x−x ' 3
B x=
4
∫ ∣ ∣3 d x '
x−x '

B x=
0
∫ ∫∫
2  ∞
[
 r '−r i 
I
r'
 × x−x '
 '− i ' ] r '2 sin  ' d r ' d ' d '
4  0 0 0  r 2r '2−2 r r ' cos  cos  'sin  sin  ' cos − '3 /2
Spherical unit vectors are not constant or fixed, but change as we integrate. The safest thing to do is
express the vector directions in Cartesian coordinates, using relations such as ϕ̂ '=−sin ϕ ' ̂i+cos ϕ ' ̂j
and

x−x '=r sin  cos −r ' sin  ' cos  ' i r sin  sin −r ' sin  ' sin ' jr cos −r ' cos  ' k

0 2   ∞ 2
B x= ∫ ∫
4 0 0 0
∫ [
r ' sin  ' d r ' d  ' d ' I
r '−r i
r' ]
  '− i −sin  ' icos ' j ×

r sin cos −r ' sin ' cos ' ir sin  sin −r ' sin  ' sin ' jr cos −r ' cos ' k 
r 2r '2−2 r r 'cos  cos  'sin  sin  ' cos − '3/ 2

Now evaluate the Dirac delta's

I 0 2
B x= ∫ r sin i d ' [ −sin ' icos  ' j ]×
4 0 i
r sin  cos −r i sin  i cos ' ir sin  sin −r i sin i sin ' jr cos −r i cos  i  k 
2 2 3/ 2
r r i −2 r r i cos cos i sin  sin  i cos−'

Evaluate the cross product

I 0 2 r cos −r i cos i cos  ' isin  ' j r i sin i −r sin  cos−' k 
B x=
4 0
∫ r i sin i d '
r 2r 2i −2 r r i cos  cos i sin  sin  i cos − '3 /2

This expression is general for any point in space in spherical coordinates. All that is left to do is
perform the integral over the azimuthal angle and we would have the final solution to the total
magnetic field (magnetic induction) B as any point in space created by a loop of current. The integral is
messy however, and we don't need to do it for this particular problem. For an observation point at the
origin r =0 , this reduces to:

[ ]
2 2 2
I  0 sin i   
B x= −cos  i   i ∫ cos ' d ' j ∫ sin  ' d 'sin  i k ∫ d '
4  ri 0 0 0

We can now perform the integrals and find

 0 I sin 2 i 
B x= k
2 ri

If we recognize that the ring has some fixed radius no matter where we place the ring, we can simplify
this using r i sin  i=a

 0 I a 2 1
B x=k
2 r 3i

or
  0 I a 2 1
B x=k
2 a z i23 /2
2

Now we can use the principle of superposition. If we have many of these wire loops very close together
so that they fill the length Δz = zi+1 - zi, but are in a small bundle still at this same location, and are
spaced in a linear density N, then there are total NΔz loops in this bundle. We simply multiply the
equation above by to take into account that they are all at the same location but that there is no loops
instead of one.

 0 I a 2 1
B x=k N  z i1−z i
2 a z i23 /2
2

Using the principle of superposition again, we can add together all the magnetic fields due to many
such bundles along the length of the solenoid to get the total field.
NL 2
 Ia 1
B x=∑ k 0 N  z i1− z i
i=1 2 a z i23 /2
2

This is as far as we can get for a finite number of loops. However, if the number of loops is large, we
can approximate the number of loops as approaching infinity, NL → ∞. We can now shrink each bundle
of width (zi+1 – zi) as small as we want. By definition, the sum becomes an integral and Δz → dz
z2 2
 Ia 1
B x=∫ k 0 N dz
z 1
2 a z 23/ 2
2

z2
 N I a2
B x=k 0
2
∫ 2 1 2 3 / 2 dz
z a  z  1

2 z2
 N Ia
B x=k 0
2 [ z
a a  z
2 2 2 ]
z1

B x=k
0 N I a 2
2 [ z2
−
z1
a 2  a 2 z 22 a 2  a 2z 12 ]
 NI
B x=k 0
2
[ cos 2 −cos *1 ]

The star on the angle 1 is to remind us that because we have used a fixed coordinate system, both
angles are measured relative to the positive z-axis. But the problem asks us to define angle 1 as
measured from the negative z-axis. We must make a transformation to get the angle defined in the way
they want:  *1  − 1
  NI
B x=k 0 [ cos 2cos 1 ]
2

We can check our answer by taking limiting cases. If the solenoid's length becomes infinite, both angles
become zero and we get our usual answer for a long solenoid B x=k  0 N I