CHAPTER 9

Conformal Mapping and Bilinear Transformation

BY

Dr. Pulak Sahoo

Assistant Professor
Department of Mathematics
University of Kalyani
West Bengal, India
E-mail : sahoopulak1@gmail.com

1
Module-2: Bilinear Transformation -
Basic Properties

Another important class of elementary mappings was studied by Augustus Ferdinand

Möbius. These mappings are conveniently expressed as the quotient of two linear expres-
sions and is defined as follows.

Definition 1. If a, b, c, d are complex constants then the transformation

az + b
w = f (z) = (1)
cz + d
 
 
 a b 
where   = ad − bc 6= 0 is called a Bilinear Transformation or Möbius Transforma-

 c d 
tion or linear fractional transformation. The expression ad − bc is called the determinant
of the transformation.

Note 1. The transformation (1) can also be written as

Azw + Bz + Cw + D = 0, AD − BC 6= 0.

Since this is linear in both the variables z and w, (1) deserves to be termed bilinear
transformation.

Remark 1. When c = 0, (1) represents simply a linear transformation. When c 6= 0,

then the transformation (1) can be written as
a
az + b c
+ d) − ad
(cz c
+b
w = =
cz + d cz + d
a ad − bc 1
= − · .
c c cz + d
a
If ad − bc = 0, then w = c
= constant. Thus the condition ad − bc 6= 0 means that the
function w = f (z) is non-constant.

2
Theorem 1. The inverse of a bilinear transformation is also a bilinear transformation.

Proof. Let

az + b
w = , ad − bc 6= 0
cz + d

be a bilinear transformation. Solving for z we obtain from above

−dw + b
z = , (2)
cw − a

where the determinant of the transformation is ad−bc which is not zero. Thus the inverse
of a bilinear transformation is also a bilinear transformation.

Remark 2. From the bilinear transformation (1) and its inverse (2) it follows that to
every z other than z = −d/c (w has a simple pole at z = −d/c) there corresponds
only one value of w and to every value of w other than w = a/c (z has a simple pole at
w = a/c) corresponds just one value of z. We suppose that the point at infinity in the
w-plane corresponds to the point z = −d/c, and that the point at infinity in the z-plane
is mapped into the point w = a/c. Thus if c 6= 0, z = ∞ corresponds to w = a/c and
z = −d/c corresponds to w = ∞. When c = 0, the point z = ∞ corresponds to w = ∞.
Therefore, for c 6= 0, we have

az+b


 cz+d
, if z 6= −d/c, z 6= ∞

w = f (z) = ∞, if z = −d/c


a
if z = ∞.


c

It now follows that the bilinear transformation (1) set up a one-one correspondence be-
tween the points of the extended z-plane and the points of the extended w-plane.

Theorem 2. A bilinear transformation is a conformal mapping for all finite z except

z = −d/c.

az+b
Proof. Let w = f (z) = cz+d
, ad − bc 6= 0 be a bilinear transformation. Then

a(cz + d) − c(az + b) ad − bc
f 0 (z) = = 6= 0 f or z 6= −d/c,
(cz + d)2 (cz + d)2

and so w = f (z) is a conformal mapping for all finite z except z = −d/c.

Theorem 3. The composition of two bilinear transformation is again a bilinear trans-

formation.

3
Proof. Let

a1 z + b1
ζ = , a 1 d 1 − b1 c1 =
6 0
c1 z + d1
a2 ζ + b2
and w = , a 2 d 2 − b2 c2 =6 0
c2 ζ + d2

be two bilinear transformations. Substituting we obtain

a2 ac11z+d
z+b1
1
+ b2 (a1 a2 + b2 c1 )z + (a2 b1 + b2 d1 )
w = =
c2 ac11z+d
z+b1
1
+ d2 (a1 c2 + c1 d2 )z + (b1 c2 + d1 d2 )
az + b
= ,
cz + d

where a = a1 a2 + b2 c1 , b = a2 b1 + b2 d1 , c = a1 c2 + c1 d2 , d = b1 c2 + d1 d2 . Again
    
    
 a b   a1 b1   a2 b2 
ad − bc =   =
 


 6= 0.

 c d   c1 d1   c2 d2 

Thus, the composition of two bilinear transformation is again a bilinear transformation.

Theorem 4. The identity mapping w = z is a bilinear transformation.

Proof. We have

1·z+0
w =z = ,
0·z+1

which is obviously a bilinear transformation.

Theorem 5. The associative law for composition of bilinear transformation holds.

Proof. Let

a1 z + b1
T1 : ζ = , a1 d1 − b1 c1 6= 0,
c1 z + d1
a2 ζ + b2
T2 : λ = , a2 d2 − b2 c2 6= 0
c2 ζ + d 2
a3 λ + b3
and T3 : w = , a3 d3 − b3 c3 6= 0
c3 λ + d 3

be three bilinear transformations. Then

a2 ac11z+d
z+b1
1
+ b2
T2 T1 : λ =
c2 ac11z+d
z+b1
1
+ d2
(a1 a2 + b2 c1 )z + (a2 b1 + b2 d1 )
= .
(a1 c2 + c1 d2 )z + (b1 c2 + d1 d2 )

4
Therefore,

a3 (a 1 a2 +b2 c1 )z+(a2 b1 +b2 d1 )

(a1 c2 +c1 d2 )z+(b1 c2 +d1 d2 )
+ b3
T3 (T2 T1 ) : w =
c3 (a1 a2 +b2 c1 )z+(a2 b1 +b2 d1 )
(a1 c2 +c1 d2 )z+(b1 c2 +d1 d2 )
+ d3
(a1 a2 a3 + a3 b2 c1 + a1 b3 c2 + b3 c1 d2 )z + (a2 a3 b1 + a3 b2 d1 + b1 b3 c2 + b3 d1 d2 )
=
(a1 a2 c3 + b2 c1 c3 + a1 c2 d3 + c1 d2 d3 )z + (a2 b1 c3 + b2 c3 d1 + b1 c2 d3 + d1 d2 d3 )
az + b
= , say.
cz + d

Again

a3 ac22ζ+d
ζ+b2
2
+ b3
T3 T2 : w =
c3 ac22ζ+d
ζ+b2
2
+ d3
(a2 a3 + b3 c2 )ζ + (a3 b2 + b3 d2 )
= .
(a2 c3 + c2 d3 )ζ + (b2 c3 + d2 d3 )

So

(a2 a3 + b3 c2 ) ac11z+d
z+b1
1
+ (a3 b2 + b3 d2 )
(T3 T2 )T1 : w =
(a2 c3 + c2 d3 ) ac11z+d
z+b1
1
+ (b2 c3 + d2 d3 )
(a1 a2 a3 + a3 b2 c1 + a1 b3 c2 + b3 c1 d2 )z + (a2 a3 b1 + a3 b2 d1 + b1 b3 c2 + b3 d1 d2 )
=
(a1 a2 c3 + b2 c1 c3 + a1 c2 d3 + c1 d2 d3 )z + (a2 b1 c3 + b2 c3 d1 + b1 c2 d3 + d1 d2 d3 )
az + b
= .
cz + d

This shows that T3 (T2 T1 ) = (T3 T2 )T1 and so the associative property holds for the
composition of bilinear transformation.

Example 1. Show by an example that bilinear transformations are not commutative

under composition.

Solution. We consider the bilinear transformations

z z−1
T1 (z) = ; T2 (z) = .
z+1 z−2

Therefore,
 
z−1 z−1
T1 T2 (z) = T1 (T2 (z)) = T1 = ,
z−2 2z − 3
 
z 1
T2 T1 (z) = T2 (T1 (z)) = T2 = .
z+1 z+2

Hence, T1 T2 (z) 6= T2 T1 (z). This shows that bilinear transformations do not satisfy the
commutative property.

5
Note 2. The set of all bilinear transformations form a non-commutative group with
respect to the composition of maps.

Theorem 6. Every bilinear transformation maps circles and lines into circles and lines
(a line is a circle of infinite radius).

az+b
Proof. Let w = f (z) = cz+d
, ad − bc 6= 0 be a bilinear transformation. If c = 0, then

a b a b
f (z) = z + = Az + B, A = and B = .
d d d d

Clearly, Az + B, being linear, maps circles and lines into circles and lines.
If c 6= 0, then
a
+ d) − ad
c
(cz c
+b
f (z) =
cz + d
a bc − ad 1
= + 2
· .
c c z + d/c

Assigning

1 bc − ad
z1 = z + d/c, z2 = , z3 = z2
z1 c2
a
we obtain f (z) = c
+ z3 . It is clear that the above transformations are of the form

1
w1 = z + α, w2 = , w3 = βz.
z

This establishes the fact that every bilinear transformation is the resultant of bilinear
transformations with simple geometric imports. Thus, a bilinear transformation maps
circles and lines into circles. This proves the theorem.

2z+3
Example 2. Show that the transformation w = z−4
maps the circle x2 + y 2 − 4x = 0
onto the line 4u + 3 = 0.

Solution. Given transformation is clearly a bilinear transformation. The inverse trans-

formation is given by

4w + 3
z = .
w−2

The equation of the circle can be written as

x2 + y 2 − 4x = 0

6
 i.e. | z |2 −4Re z = 0

i.e. zz − 2(z + z) = 0.

Putting the value of z and noting that w = u + iv, we obtain from above
 
4w + 3 4w + 3 4w + 3 4w + 3
· −2 + =0
w−2 w−2 w−2 w−2

i.e. (4w + 3)(4w + 3) − 2{(4w + 3)(w − 2) + (4w + 3)(w − 2)} = 0

i.e. 2(w + w) + 3 = 0

i.e. 4u + 3 = 0,

which is the required line.

Example 3. Show that the line x = 3y is mapped onto the circle under the bilinear
iz+2
transformation w = 4z+i
. Find the centre and radius of the image circle.

Solution. Given transformation is clearly a bilinear transformation. The inverse trans-

formation is given by

−iw + 2
z = .
4w − i

Putting z = x + iy and w = u + iv we obtain

(v + 2) − iu
x + iy =
4u + i(4v − 1)
[(v + 2) − iu][4u − i(4v − 1)]
=
16u2 + (4v − 1)2
9u − i(4u2 + 4v 2 + 7v − 2)
= .
16u2 + (4v − 1)2

Comparing real and imaginary parts we obtain

9u 4u2 + 4v 2 + 7v − 2
x = , y =− .
16u + (4v − 1)2
2 16u2 + (4v − 1)2

Putting these values in the equation x = 3y we get

9u 4u2 + 4v 2 + 7v − 2
= −3
16u2 + (4v − 1)2 16u2 + (4v − 1)2
3 7 1
i.e. u2 + v 2 + u + v − = 0,
4 4 2
q
3 5
which represents a circle with centre at (−3/8, −7/8) and radius 4 2
.

7
Example 4. Find the image of the annulus {z : 1 <| z |< 2} under the bilinear
z
transformation w = 1−z
.

z
Solution. Here the transformation is w = 1−z
. Solving for z we obtain

w
z = .
1+w

Now

| z | > 1 ⇔| w |2 >| 1 + w |2 , i.e. 0 > 1 + 2Re w

and | z | < 2 ⇔| w |2 < 4 | 1 + w |2 , i.e. 0 < 3[| w + 4/3 |2 −4/9].

Therefore, it is easily seen that

(i) | z |> 1 is mapped into Re w < −1/2 and
(ii) | z | < 2 is mapped onto | w + 4/3 | > 2/3.
Thus the required image is {w : Re w < −1/2} ∩ {w : | w + 4/3 | > 2/3}.