PHY2206 (Electromagnetic Fields)

Analytic Solutions to Laplaces Equation

Analytic Solutions to Laplaces Equation in 2-D

Cartesian Coordinates
When it works, the easiest way to reduce a partial differential equation to a set of ordinary ones is by
separating the variables
( x, y ) = X ( x )Y ( y ) so

2
d2 X
= Y ( y) 2
dx
x 2

2
d 2Y
= X( x) 2
dy
y2

and

2 2 1 d 2 X 1 d 2Y
+
=
+
=0
x 2 y 2 X dx 2 Y dy 2

and since x and y are independent, if this is to be true each term must be a constant and therefore
d2 X
= + 2 X and
dx 2

d 2Y
= 2Y .
dy 2

So the general solutions have the form

Y = cos y + sin y and

X = cosh x + sinh x

with only four of the five constants independent.

y

Example: Find the potential at an arbitrary point inside a

rectangular box of infinite extent in the z-direction with
conducting walls at potentials V1..V4.

V1

Solution: The problem is to choose the value of the

constants in the general solution above such that the
specified boundary conditions are met. Since the principle
of superposition applies to solutions of Laplaces equation
let 1 be the solution when V2 = V3 = V4 = 0 so

V2
V3
V4
0

x
a

1 (0, y ) = V1 = ( cos y + sin y ) 0

1 ( x,0 ) = 0 = ( cosh x + sinh x ) either ( = = 0) or ( = 0 )

Since = = 0 is a trivial solution set = 0 and then

ii

1 ( x,b ) = 0 = sin( b )( cosh x + sinh x ) = n / b for integer n

PHY2206 (Electromagnetic Fields)

Analytic Solutions to Laplaces Equation

ny
na + sinh na = coth na

cosh
b
b
b
b

1 ( a, y ) = 0 = sin

ny
b

1 (0, y ) = V1 = sin

which is not satisified by any single value of n, so use a Fourier series expansion

ny
2 a
ny
where n =
V1 ( y ) sin
dy

b
b
a 0

1 (0, y ) = n sin
n=1

and put it all together so that

ny
n x
n a
ny
cosh
coth
sinh
.
b
b
b
b

1 ( x, y ) = n sin
n=1

The procedure is repeated to find 2..4 and construct the final answer = 1 + 2 + 3 + 4 .
Although the general solution is simple in Cartesian coordinates, getting it to satisfy the boundary
conditions can be rather tedious.

Cylindrical Polar Coordinates

In cylindrical polar coordinates when there is no z-dependence 2 has the form
1 1 2
+
r
= 0.
r r r r 2 r 2
Separating variables = R(r ) ( ) so
1 R 1 2
2
r
r
=

=m
R r r 2
where the positive sign of the constant m2 anticipates the result
2
2
2 = m = cos( m ) + sin( m )

as must be periodic so m must be an integer.
r

and

R
r
= m 2 R so try R = Cr l
r r
l 2 Cr l = l 2 R = m 2 R l = m.

PHY2206 (Electromagnetic Fields)

Analytic Solutions to Laplaces Equation

Hence R = r m + r m is the general form for m 0 and R = 0 ln r + 0 when m = 0 and the most
general form of the solution is
i

(r, ) = 0 ln r + 0 +

[( mr m + mr m )( m cos(m ) + m sin(m ))]

m=1

including a redundant constant.

Example: A long conducting cylinder with its axis along the z-direction is placed into a field
E = E0 x . Find the resultant potential if the cylinder radius is a.
Solution: Far enough away from the cylinder the field will be unaffected so
as r E E0 x so E0 x = E0 cos( ) 0 = 0.
and symmetry about the x-axis implies that ( ) = ( ) therefore m = 0 (for m > 0). The increasing
powers of r would affect the potential at very large distances unless m = 0 (for m > 0). The required
form of the potential is therefore
= 0 E0 r cos +

r mm cos(m )

m=1

where the m have been absorbed into the m. On the surface of the conducting cylinder the potential
must be constant, say V0, so
(r = a) = V0

0 = ( 0 V0 ) + 1 E0 a cos( ) + mm cos( m ) .
a

m=2 a

Since {cos( m )} are linearly independent functions

0 = V0 ,

1 = E0 a 2 ,

(r, ) = V0 E0 r cos( ) +

m2 = 0

E0 a 2
cos( )
r

a2

= V 0 + E0 r cos( ).
r

Complex Variables
A function f(z) is analytic (also known as regular or holomorphic) if, at a point z0,
f ( z0 + z ) f ( z0 )
df
= lim
dz z0 z0
z
exists and has a single value. The quantity z can be in any direction in the complex plane. If

PHY2206 (Electromagnetic Fields)

Analytic Solutions to Laplaces Equation

f ( z ) = f ( x + iy) = u( x, y ) + iv( x, y)
then for z along the real axis z = x and
i

u( x + x, y ) u( x, y) + i[v( x + x, y ) v( x, y )] u
v
df
= lim
=
+i
dz x0
x
x
x
but when z is along the imaginary axis z = iy and
i

u v
df
= i
+ .
dz
y y
Equating the real and imaginary parts of these results yields the Cauchy-Riemann equations

u v
=
x y

and

u
v
=
y
x

which are both necessary and sufficient conditions for the function f to be analytic.
The Cauchy-Riemann equations can be differentiated

2u 2 v
=
x 2 x y

2u
2v
=
y2
y x

2u 2 v
=
y x y 2

2u
2v
= 2
x y
x

and adding the results together gives

2u 2u
+
= 0 and
x 2 y2
because

2v 2v
+
=0
x 2 y2

2g
2g
=
for a continuous function g(x,y).
x y y x

Therefore every analytic function provides two solutions to Laplaces equation in 2-dimensions, and
pairs of such solutions are known as conjugate harmonic functions. As the curves u = constant and
v = constant are perpendicular to each other, if one represents a contour of constant potential, then the
other is a flux line of the corresponding electric field. These interesting properties are the basis of
conformal mapping methods for solving 2-dimensional electrostatic problems. Details are given in
advanced text and reference books such as KJ Binns and PJ Lawrenson (1973) Analysis and
Computation of Electric and Magnetic Fields Problems 2nd edn, Pergamon Press.
i

Copyright CDH Williams

Exeter 1997, CW960317/2
4