Chapter 2

Conformal mappings
2.1 Stereographic Projection
One known problem with numbers in the complex plane C = {(x, y) :
−∞ < x, y < +∞} do not have an ordering like the real numbers
on the real-axis R. Riemann’s (1826-1866) idea is to add an ideal
point, denoted by ∞, to C to obtain an extended complex plane Ĉ =
C ∪ {∞}. This construction can get around the problem of ordering.
The resulting Ĉ is compact which can be vasualised by the following
construction.
We show that there is an one-to-one correspondence between

S = {(x1 , x2 , x3 ) ∈ R3 : x21 + x22 + x23 = 1}

and Cc
= C ∪ {∞}.
Let N = (0, 0, 1) and z ∈ C. If we join the straight line between
N and z, the straight line intersects the sphere S at Z = (x1 , x2 , x3 )
say. The construction clearly exhibits an one-to-one correspondence
between S \ {N } and C. Note that Z → N as |z| → ∞. We may
associate N with ∞ and obtain the bijection between S and C. c
This
is known as the Stereographic projection.
Suppose P (x1 , x2 , x3 ) = Z ∈ S associates with z = (x, y) ∈ C. c

Then we may associate z the notation P with coordinate (x, y, 0).

48
CHAPTER 2. CONFORMAL MAPPINGS 49

N
b

S
b
Z

b z

Figure 2.1: Riemann sphere

Then we have, by considering similar triangles formed by the line seg-

ment N P and projecting onto the x−, y− and z−axes respectively,

|N P | x y 1
= = = , (2.1)
|N Z| x1 x2 1 − x3
so that
x1 + ix2
z = x + iy = .
1 − x3
Then
2 x21 + x22 1 + x3
|z| = = ,
(1 − x3 )2 1 − x3
hence
|z|2 − 1
x3 = .
|z|2 + 1
Then
z+z z−z
x1 = , x2 = .
1 + |z|2 i(1 + |z|2 )
This clearly shows a one-one correspondence between S\(0, 0, 1) and
C with the N = (0, 0, 1) corresponds to ∞. We also note that the
upper hemisphere where x3 > 0 corresponds to |z| > 1 and the lower
hemisphere of S corresponds to |z| < 1. An advantage with this Rie-
mann sphere model is that it puts all complex numbers including ‘∞’
in equal footing since any number can be rotated to N and vice-verse.
CHAPTER 2. CONFORMAL MAPPINGS 50

From a geometrical viewpoint, it is evident that every (infinite)

straight line in the z−plane is transformed into a circle on S that
passes through the North pole N , and conversely. Hence, every circle
(straight line included) on the z−plane corresponds to a circle/straight
line on S.
Theorem 2.1.1. A circle on the Riemann sphere is mapped under the
Stereographic projection into a circle (including a straight line) of the
C, and conversely.
Proof. Show that
1. a circle equation that lies on the Riemann sphere is an equation
of the form
ax1 + bx2 + cx3 = d
subject to 0 ≤ c < 1 and a2 + b2 + c2 = 1 (this is the intersection
of the plane and the unit sphere).
2. the above equation can be rewritten in the form
a(z + z̄) − ib(z − z̄) + c(|z|2 − 1) = d(|z|2 + 1)

3. the above equation can be further rewritten into the form

(d − c)(x2 + y 2 ) − 2ax − 2by + d + c = 0,

which is clearly a circle equation in the C and it becomes a straight

line. equation if and only if c = d.
That is, a circle on the Riemann sphere S corresponds to either
a circle or a straight line on C. In the case the circle on S passes
through the North pole N = (0, 0, 1), then the corresponding straight
line (also considered as an unbounded circle passes through) to ∞.

Exercise 2.1.1. Show that if z and w are two points in C so that

their images lie on two diametrically opposite points on the Riemann
sphere, then
wz + 1 = 0.
CHAPTER 2. CONFORMAL MAPPINGS 51

Theorem 2.1.2. The stereographic projection is isogonal (i.e., the

mapping preserves angles).

Proof. The statement of the theorem means that the tangents of two
curves in the C intersect at point z0 is equal to the angle made by two
tangents at the corresponding intersection point of two image curves
on the Riemann sphere. We shall make two assumptions:
1. that the Stereographic projection preserves tangents. We skip
the detail verification of this fact. But this is not difficult to see
since the Stereographic projection is a smooth map,
2. that without loss of generality that the two curves in C are (in-
finite) straight lines.
Suppose the two straight line equations are given by
a1 x + a2 y + a3 = 0 (x3 = 0);
(2.2)
b1 x + b2 y + b3 = 0. (x3 = 0)
It follows from (2.1) that the two plane equations become respectively,
a1 X1 + a2 X2 + a3 (X3 − 1) = 0;
b1 X1 + b2 X2 + b3 (X3 − 1) = 0.
In the limiting case when X3 = 1, we have the two tangent plane
equations
a1 X1 + a2 X2 = 0;
(2.3)
b1 X1 + b2 X2 = 0.
at N (0, 0. 1) parallel to the C. Clearly the angle between the two
curves in (2.2) is the same angle between the two lines in (2.3).
Note that any two intersecting circles in general positions on S can
be rotated so that the intersection point passes through the North pole
N . This consideration takes care of the preservation of the angle of
intersection of two curves in general position in C under the Stereo-
graphic projection.
CHAPTER 2. CONFORMAL MAPPINGS 52

Theorem 2.1.3. Let z1 , z2 be two points in C and Z1 , Z2 be their

images on the Riemann sphere S under the Stereographic projection.
We denote a(Z1 , Z2 ) to be arc length between Z1 and Z2 . Then
a(Z1 , Z2 ) 2
lim
z →z
= . (2.4)
2 1 |z − z |
1 2 1 + |z1 |2
That is, the ratio depends on position only. So the Stereographic pro-
jection is called a pure magnification.
We easily deduce from the above theorem that
Theorem 2.1.4. Let C = {z = z(s) : 0 ≤ s ≤ L} be a piecewise
smooth curve in C. Let Γ be the image curve of C on the Riemann
sphere under the Stereographic projection. Then the length `(Γ) of Γ
is given by
Z L 2|dz(s)|
`(Γ) = .
0 1 + |dz(s)|2

Let d(Z1 , Z2 ) denote the chordal distance between Z1 and Z2 on S.

We also write
χ(z1 , z2 ) := d(Z1 , Z2 ).
where z1 , z2 are the corresponding points in C.
Theorem 2.1.5. Let z1 , z2 ∈ C. Then
2|z1 − z2 |
χ(z1 , z2 ) = q q . (2.5)
1 + |z1 |2 1 + |z2 |2
Since
χ(z1 , z2 ) := d(Z1 , Z2 ) ≈ a(Z1 , Z2 )
as z1 → z2 . So the Theorem 2.1.3 follows from the equation (2.5) in
the limit z2 → z1 .
Proof. Let z1 = (x1 , y1 ) and z2 = (x2 , y2 ) and none equal to ∞. We
construct a plane passing through the following three points:

(0, 0, 1), (x1 , y1 , 0), (x2 , y2 , 0).

CHAPTER 2. CONFORMAL MAPPINGS 53

Figure 2.2: Riemann sphere slide

Then we have the above figure.

We deduce from the Riemann sphere S that
q q
d(N, z1 ) = 1 + |z1 |2 , d(N, z2 ) = 1 + |z2 |2 .

One can see from similar triangles consideration on the Riemman

sphere S that
x1 1 − x3 x2
= = .
x 1 y
Hence
x21 x22
1 + |z|2 = 1 + x2 + y 2 = 1 + +
(1 − x3 )2 (1 − x3 )2
2(1 − x3 ) 2
= = .
(1 − x3 )2 1 − x3
and
d(N, Z) 1 − x3 2
= = .
d(N, z) 1 1 + |z|2
CHAPTER 2. CONFORMAL MAPPINGS 54

holds. This gives raise to

2 2
d(N, Z1 ) = q , d(N, Z2 ) = q .
1 + |z1 |2 1 + |z2 |2

We conclude that

d(N, z1 )d(N, Z1 )= 2 =d(N, z2 )d(N, Z2 ).

Hence the triangle ∆N z1 z2 and ∆N Z1 Z2 are similar. Hence

d(Z1 , Z2 ) d(N, Z2 )
= .
d(z1 , z2 ) d(N, z1 )
It follows from the above consideration that
d(N, Z2 ) 2|z1 − z2 |
d(Z1 , Z2 ) = d(z1 , z2 ) · = q q .
d(N, z1 ) 1 + |z1 |2 1 + |z2 |2

as required.

We are ready to prove Theorem 2.1.3.

We observe the relation
d(Z1 , Z2 ) sin α
= ,
a(Z1 , Z2 ) α
holds, where α is the angle between the line segments N Z1 and N Z2
from the above figure. Hence
a(Z1 , Z2 ) d(Z1 , Z2 ) χ(z1 , z2 ) 2
= ≈ →
|z1 − z2 | |z1 − z2 | |z1 − z2 | 1 + |z1 |2
as z2 → z1 .
We also note that
2
χ(z1 , ∞) = zlim
→∞
χ(z1 , z2 ) = q ,
2
1 + |z1 |2
CHAPTER 2. CONFORMAL MAPPINGS 55

which follows from the Riemann sphere (geometric) or the Theorem

2.1.5 (algebraic) considerations. Thus we define the chordal distance
to be
2|z − z 0 |


, z, z 0 ∈ C



q q
 1 + |z|2 1 + |z 0 |2


χ(z, z 0 ) =  2
, z 0 = ∞.


 q
 1 + |z|2



Alternative derivation
of the chordal distance. Suppose (x1 , x2 , x3 ) ∈ S associates with z =
(x, y) ∈ C c
and (x01 , x02 , x03 ) ∈ S associates with z 0 ∈ C.
c

Then the distance or the length of the chord joining (x1 , x2 , x3 ) and
(x1 , x02 , x03 ) on S is given by
0
q
(x1 − x01 )2 + (x2 − x02 )2 + (x3 − x03 )2 .
On the other hand,
(x1 − x01 )2 + (x2 − x02 )2 + (x3 − x03 )2 = 2 − 2(x1 x01 + x2 x02 + x3 x03 ).

Exercise 2.1.2. Show that

x1 x01 + x2 x02 + x3 x03
(z + z̄)(z 0 + z̄ 0 ) − (z − z̄)(z 0 − z̄ 0 ) + (|z|2 − 1)(|z 0 | − 1)
=
(1 + |z|2 )(1 + |z 0 |2 )
(1 + |z|2 )(1 + |z 0 |2 ) − 2|z − z 0 |2
=
(1 + |z|2 )(1 + |z 0 |2 )
Exercise 2.1.3. Verify the formaula for chordal distance using the
above formala.
Exercise 2.1.4. Verify that χ(z1 , z2 ) = χ(z̄1 , z̄2 ) = χ(1/z1 , 1/z2 ).
Exercise 2.1.5. Describe a ε−neighbourhood of a pont z0 in the
chordal metric.
CHAPTER 2. CONFORMAL MAPPINGS 56

Metric space
The chordal distance χ(z1 , z2 ) defines a metric on Ĉ. This is because

1. χ(z1 , z2 ) ≥ 0 and with equality if and only if z1 = z2 ;

2. χ(z1 , z2 ) = χ(z2 , z1 );

3. χ(z1 , z3 ) ≤ χ(z1 , z2 ) + χ(z2 , z3 ),

where the third item follows from

Exercise 2.1.6. Let a, b, c ∈ C. Then

(a − b)(1 − c̄c) = (a − c)(1 + c̄b) + (c − b)(1 + c̄a).

Exercise 2.1.7. Show that the above metric space is complete.

2.2 Analyticity revisited

Local properties of one-one analytic functions
We recall that if f : E → P and there correspond only one point
in E for every point in P under this f , then we say the map f is
injective. This defines a function g on P , denoted by z = g(w), called
the inverse function or inverse mapping of f . In particular, we
see that g[f (z)] = z.

Let w = f (z) = u(x, y) + iv(x, y). Then one can view f as a

mapping R2 −→ R2 given by
   
x
  7−→ 
u(x, y)
.
y v(x, y)

What is a criterion that guarantee the existence of an inverse mapping

for the above mapping?
CHAPTER 2. CONFORMAL MAPPINGS 57

Standard material from calculus courses asserts that if

∂u ∂u
∂x ∂y 6= 0, at z0 = (x0 , y0 ),
∂v ∂v
∂x ∂y
then the Implicit function theorem asserts that an inverse function of
f exists there. That is, if the Jacobian is non-zero at z0 . But then the
Cauchy-Riemann equations give
∂u ∂u
∂x ∂y = ux −vx = u2 + v 2 = |f 0 (z )|2 .
∂v ∂v x x 0
vx ux
∂x ∂y

This leads to the following statement.

Theorem 2.2.1. Let f (z) be an analytic function on a domain D such

that f 0 (z0 ) 6= 0. Then there is an analytic function g(w) defined in a
neighbourhood N (w0 ) of w0 = f (z0 ) such that g(f (z)) = z throughout
this neighbourhood.

Proof. Since
∂u ∂u
∂x ∂y = |f 0 (z )|2 6= 0,
∂v ∂v 0

∂x ∂y
so the Implicit Function theorem asserts that is a neighbourhood N (w0 )
of w0 = f (z0 ) in which f has a local inverse at w0
   
u
  7−→ 
x(u, v)
.
v y(u, v)
Moreover, the analytic Implicit Function theorem asserts that the
stronger conclusion that since f is analytic at z0 so the g(w) is an-
alytic at w0 .
CHAPTER 2. CONFORMAL MAPPINGS 58

We prove that a strong form of converse of the above statement

also holds. Please note we could apply the Theorem 1.11.3 to prove
the theorem. But we prefer to apply the Rouché theorem instead.
Theorem 2.2.2. Let f (z) be an one-one analytic function on a domain
D. Then f 0 (z) 6= 0 on D.

Proof. We suppose on the contrary that f 0 (z0 ) = 0 for some z0 . We

first notice that f 0 (z) 6≡ 0. For otherwise, f (z) is identically a constant,
contradicting to the assumption that f (z) is one-one on D.
Since the zeros of f 0 (z) are isolated , so there is a ρ > 0 such that
f 0 (z) 6= 0 in {z : 0 < |z − z0 | < ρ}. Because of the assumption that f
is one-one, so
f (z) 6= f (z0 ) on |z − z0 | = ρ.
On the other hand, |f (z)| is continuous on the compact set |z − z0 | = ρ
so that we can find a δ > 0 such that
|f (z) − f (z0 )| ≥ δ > 0 on |z − z0 | = ρ.
Let w0 be an arbitrary point in {w : 0 < |w0 − w0 | < δ}. Then the
inequality
|f (z) − w0 | ≥ δ > |w0 − w0 |
holds, so that the Rouché theorem again implies that the function
f (z) − f (z0 ) = f (z) − w0 and the function
[f (z) − f (z0 )] + [f (z0 ) − w0 ] = f (z) − w0
have the same number of zeros inside {z : |z − z0 | < ρ}. But f 0 (z0 ) = 0
so f (z) − f (z0 ) has at least two zeros (counting multiplicity). Hence
f (z) − w0 also has at least two zeros (counting multiplicity) in {z :
|z − z0 | < ρ}. But f 0 (z) 6= 0 in {z : 0 < |z − z0 | < ρ}, so there are
at least two different zeros z1 and z2 in {z : |z − z0 | < ρ} so that
f (z1 ) = w0 and f (z2 ) = w0 , thus contradicting to the assumption that
f (z) is one-one.
Bibliography
[1] L. V. Ahlfors, Complex Analysis, 3rd Ed., McGraw-Hill, 1979.

[2] J.B. Conway, Functions of One Complex Variable, 2nd Ed.,

Springer-Verlag, 1978.

[3] F. T.-H. Fong, Complex Analysis, Lecture notes for MATH 4023,
HKUST, 2017.

[4] E. Hille, Analytic Function Theory, Vol. I, 2nd Ed., Chelsea Publ.
Comp., N.Y., 1982

[5] E. Hille, Analytic Function Theory, Vol. II, Chelsea Publ. Comp.,
N.Y., 1971.

[6] E. T. Whittaker & G. N. Watson, A Course of Modern Analysis,

4th Ed., Cambridge Univ. Press, 1927

59