Complex Mappings

Deepesh K P

NIT Calicut

2024
Introduction

▶ We need two separate planes to plot the domain and codomain

of complex functions.
▶ What happens to points and geometrical shapes under special
analytic functions?
▶ We are basically interested in what happens to standard simple
shapes like straight lines, rays, circles etc. under some analytic
mappings.
▶ Image of z = x + i y as w = u + i v in cartesian form
▶ Image of z = reiθ will be represented as w = Reiϕ in polar form
The mapping f (z) = z 2

▶ Easy to deal in polar form

▶ w = Rei ϕ = f (rei θ ) = (rei θ )2 = r 2 ei 2θ
▶ R = r 2 and ϕ = 2θ
▶ A point at r distance from origin and at θ angle will be mapped to
a point at r 2 distance from origin and at 2θ angle from positive X
axis.
▶ Plot!
−1
▶ Point 4 + 3i = 5etan (3/4) will be mapped to a point 25 distance
away from origin at angle 2 tan−1 (3/4)
The mapping f (z) = z 2

What happens to the circle r = a? (0 ≤ θ ≤ 2π)

▶ w = Rei ϕ = f (aei θ ) = (aei θ )2 = a2 ei 2θ

▶ R = a2 , a constant. So its a circle.
▶ ϕ = 2θ. So 0 ≤ ϕ ≤ 4π
▶ Hence when the domain points rotate once around the circle,
image circle get rotated twice.
▶ Circle of radius 3 will be mapped to a circle of radius 9.
▶ Image of unit circle gets no visible change (actually twice traced)
The mapping f (z) = z 2

What happens to the ray θ = a? (0 ≤ r < ∞)

▶ w = Rei ϕ = f (rei a ) = (rei a )2 = r 2 ei 2a

▶ ϕ = 2a, a constant. So its on a ray, but at a bigger angle!
▶ R = r 2 . So 0 ≤ R < ∞
▶ Hence when the domain points move along a ray, image points
move along a different ray at double the angle and moves at
square speed.
▶ Ray at angle π3 will be mapped to a ray at angle 2 π3 .
▶ Its portion of size 3 will be elongated to be of size 9!
The mapping f (z) = z 2

▶ What happens to a sector?

▶ It becomes another sector!
The mapping f (z) = z 2

What happens to the line x = a (∞ < y < ∞)

▶ w = u + iv = f (x + iy) = (x + iy )2 = x 2 − y 2 + i 2xy
▶ u = x 2 − y 2 and v = 2xy
▶ x = a ⇒ u = a2 − y 2 and v = 2ay
▶ To plot the figure in UV plane, we need the relation between u
and v .
▶ Eliminate y from the data.
2
▶ u = a2 − 4av 2
2 ⇒ v = 4a2 (a2 − u)
▶ Its a parabola: axis is parallel to U axis; open to negative side
(u ≤ a2 )
The mapping f (z) = z 2

What happens to the line y = a (∞ < x < ∞)

▶ w = u + iv = f (x + iy) = (x + iy )2 = x 2 − y 2 + i 2xy
▶ u = x 2 − y 2 and v = 2xy
▶ y = a ⇒ u = x 2 − a2 and v = 2xa
▶ To plot the figure in UV plane, we need the relation between u
and v .
▶ Eliminate x from the data.
2
v 2 2
▶ u = 4a 2 − a ⇒ v = 4a2 (u + a2 )
▶ Its a parabola: axis is parallel to U axis; open to positive side
(u ≥ −a2 )
The mapping f (z) = z 2

▶ What happens to area between two lines x = a and x = b?

▶ It becomes area between two parabolas!
The mapping f (z) = z + c, where c = α + i β

▶ z = x + i y ⇒ w = (x + α) + i (y + β)
▶ x → x + α and y → y + β
▶ Every point will be moved to α distance to the right and β
distance vertically
▶ lines are shifted to lines
▶ Circles are shifted to circles
▶ It is called shifting map
The mapping f (z) = cz where c = α + i β = r0 ei θ0

Consider a point z = rei θ

▶ w = cz = rr0 ei(θ+θ0 )
▶ R = rr0 and ϕ = θ + θ0
▶ A point moves to rr0 distance from origin and gets rotated by
angle θ + θ0
▶ so both magnification/compression and rotation happens.
▶ If |c| = 1 only rotation happens.
▶ If c is a real number, only magnification happens
The mapping f (z) = cz where c = α + i β = r0 ei θ0

Consider a cirlce r = a
▶ w = cz = ar0 ei(θ+θ0 )
▶ R = ar0 . So image is a circle of radius ar0 .
Consider a ray θ = a
▶ w = cz = rr0 ei(a+θ0 )
▶ ϕ = a + θ0 . So image is a ray at angle a + θ0 .
Consider a line x = a
▶ w = (α + i β)(a + i y) = (αa − βy ) + i (αy + βa)
▶ So u = (αa − βy ) and v = αy + βa
▶ Solving, we get v = mu + k , a line.
▶ Similarly, image of y = a is also a line.
Conformal Mapping

Definition
A mapping is said to be conformal at a point z0 if it preserves the
angle between any two oriented curves intersecting at z0 .
▶ Angle between tangents to the curve are considered.
▶ Angle measured anticlockwise
▶ If domain curves C1 and C2 are at angle θ0 from C1 to C2 , angle
measured from f (C1 ) and f (C2 ) should be θ0 itself.
▶ Conformal mappings does not change the angle (so some shape
preservation is there!)
▶ f (z) = z 2 is not conformal at 0. See the image of two rays!!
Conformal Mapping Identification

Theorem
An analytic function f (z) will be conformal at all points where
f ′ (z) ̸= 0..
▶ Helps identifying conformal/non conformal points of analytic
functions.
▶ f (z) = z n ?
▶ f (z) = ez ?
▶ f (z) = z + c

How to prove the Theorem ?(1)

The Mapping f (z) = z n , for positive integer n

Images of x = a and y = a are complicated, depending on n. But we

can think about the images of circle r = a and ray θ = a.
▶ r = a ⇒ w = r n ei nθ = an ei nθ ⇒ R = an , ϕ = nθ
▶ So image of circle of radius a is a circle of radius an , but rotates n
times faster!
▶ Image of ray θ = a is

w = r n ei nθ = r n ei na ⇒ R = r n , ϕ = na

So image is again a ray, but at angle na.

The mapping f (z) = ez

▶ Easy to deal with cartesion in the domain and polar form in

codomain
▶ w = Rei ϕ = f (x + iy) = ex+iy = ex ei y
▶ R = ex and ϕ = y
▶ The point (x, y ) is mapped to a point which is at ex distance from
origin and at y angle from positive X axis.
▶ For example, the point 3.14i on Y axis will be mapped to a point
at unit distance from origin and will be placed at angle π
approximately (on -X axis)!
▶ Plot!
The mapping f (z) = ez

Image of the line x = a and y = a

▶ x = a ⇒ w = Rei ϕ = ea ei y ⇒ R = ea , ϕ = y.
▶ So image is a circle of radius ea .
▶ One circle will be completed when y varies from 0 to 2π in the
domain. So for the full line x = a, circle is traced infinitely many
times.
▶ Y axis will be mapped to unit circle!
Image of the line y = a:
▶ y = a ⇒ w = Rei ϕ = ex ei a ⇒ R = ex , ϕ = a.
▶ So image is a ray ϕ = a.
What will be the image of a rectangle bounded by
x = a, x = b, y = c, y = d?
The mapping f (z) = ez

Image of the circle r = a

iθ
▶ w = Reiϕ = ere = ercosθ eir sin θ ⇒ R = er cosθ , ϕ = r sin θ.
▶ r = a ⇒ R = ea cos θ , ϕ = a sin θ
Image of the ray θ = a:
ia
▶ w = Reiϕ = ere = ercos a eir sin a ⇒ R = er cos a , ϕ = r sin a. In
particular,
▶ θ = −π ⇒ R = e−r &ϕ = 0, which is a part of the initial ray (0,1].
▶ θ = π ⇒ R = e−r &ϕ = 0, same as above.
▶ θ = 0 ⇒ R = er &ϕ = 0, which is a part of the initial ray [1,∞).
▶ We know that ez has the inverse Ln z, which is valid for all points
except 0. Hence it maps the complex plane to C \ {0}.
▶ The image of −π < θ ≤ π is C \ {0}.
Ploting
Ploting
Linear Fractional Transformation(Mobius)

Definition
A complex mapping of the form

az + b
w = f (z) = , with ad − bc ̸= 0, a, b, c, d ∈ C
cz + d
is called a linear fractional transformation
▶ f ′ (z) = ad−bc
(cz+d)2
̸= 0. So conformal(condition is for this!).
▶ There are four main types of linear fractional transformations:
▶ w = z + b (Translation)
▶ w = az (Rotation)
▶ w = az + b (Linear maps)
▶ w = z1 (inversion in the unit circle)
▶ A linear fractional transformation maps circles and straightlines in
the z-plane to circles and straightlines in the w−plane.
fixed points of a linear fractional transformation

Definition
Points which are mapped to themselves are called fixed points. That
is if f (z0 ) = z0 , then z0 is said to be a fixed point of the linear
fractional transformation.
▶ Identity function f (z) = z has all points as fixed points.
▶ All other linear fractional transformations will have atmost two
fixed points:

az + b
z= ⇒ cz 2 − (a − d)z − b = 0
cz + d
is a quadratic equation for c ̸= 0, will have atmost two roots.
▶ Find the fixed points of z−1
z+1 .
▶ Find the LFT whose fixed points are 0 and 1.
Inverse of a Bilinear transformation

▶ w = ad+b
cz+d can be written as czw + dw − az − b = 0
▶ the function f (z, w) = czw + dw − az − b is linear in both
variables (bilinear)

Result
A bilinear transformation is invertible if considered as
f : C − {− dc } → C − { ac }. The inverse of f is

dw − b
f −1 (w) =
−cw + a
which is again a bilinear transformation.
▶ If we introduce one more point called ∞ in the domain as well as
the codomain, we can treat f : C ∪ {∞} → C ∪ {∞} as an
invertible map by mapping − dc → ∞ and ∞ → ac .
▶ C ∪ {∞} is called extended complex plane.
Bilinear transformation

Result
Every bilinear transformation is a combination of translation, rotation
and magnification, and inversion maps.

" #
az + b a(z + ba ) a(z + dc + ba − dc ) a ( ba − dc )
w= = = = 1+
cz + d c(z + dc ) c(z + dc ) c (z + dc )

a bc − ad 1
w= + d
c c2 z+ c

1 1 1
z →z +α→ →β →γ+β
z +α z +α z +α
Hence lines/circles are mapped to lines/circles under this map
Importance of cross-ratio in bilinear transformation

Definition
If z1 , z2 , z3 , z4 are four points in C ∪ {∞}, their cross-ratio is denoted
as (z1 , z2 , z3 , z4 ) and is defined as

(z4 − z2 )(z3 − z1 )
(z1 , z2 , z3 , z4 ) =
(z4 − z1 )(z3 − z2 )

Result
Every bilinear transformation preserves the cross-ratio.
Proof:
az+b
Consider a bilinear transformation w = cz+d
▶ Suppose z1 , z2 , z3 , z4 get mapped to w1 , w2 , w3 , w4
az4 +b az2 +b (ad−bc)(z4 −z2 )
▶ w4 − w2 = cz4 +d − cz2 +d = (cz4 +d)(cz2 +d)
(ad−bc)(z3 −z1 )
▶ w3 − w1 = (cz3 +d)(cz1 +d) , w4 − w1 , w3 − w2 etc.
▶ Easy to Verify that (z1 , z2 , z3 , z4 ) = (w1 , w2 , w3 , w4 )
Fixing a bilinear transformation

The cross-ratio preservation can be used to fix a bilinear

transformation!

Result
There is a unique bilinear transformation mapping z1 , z2 , z3 to
w1 , w2 , w3 .
▶ Solve (z, z1 , z2 , z3 ) = (w, w1 , w2 , w3 )
▶ Easy to see that this gives a bilinear transformation
▶ Observe z = z1 ⇒ w = w1 ...
▶ Why unique ?(2)
 1
The mapping f (z) = z

▶ Easy to deal with polar form

▶ w = Rei ϕ = f (reiθ ) = 1r e−iθ
▶ R= 1
r and ϕ = −θ
▶ The point (r , θ) is mapped to a point which is at 1r distance from
origin and at −θ angle from positive X axis.
▶ For example, the point 2i on Y axis will be mapped to a point at
1 π
2 distance from origin and will be placed at angle − 2
▶ Distance from origin gets inverted and then the point is reflected
about the X axis.
▶ Plot!
Ploting
 1
Image of circles and straight lines under f (z) = z

Equation of straightlines and circles in z plane is

A(x 2 + y 2 ) + Bx + Cy + D = 0

If A = 0, then it is a straightline. Otherwise it is a circle.

▶ Azz + B z+z z−z
2 + C 2i + D = 0
▶ Convert z into w using z = 1
w
1
A + w1 1
− w1
▶ ww +B w
2 +C w
2i +D =0
▶ Multiply with ww ⇒

w +w w −w
A+B +C + Dww = 0
2 2i
▶ But u = Re(w) = w+w w−w
2 and v = Im(w) = 2i . So

A + Bu − Cv + D(u 2 + v 2 ) = 0,

which is a circle or a straightline.

 1
Image straight lines x = a and y = a under f (z) = z

Image of x = a:
▶ Consider A(x 2 + y 2 ) + Bx + Cy + D = 0
▶ Take A = 0, B = 1, C = 0, D = −a
▶ Image is A + Bu − Cv + D(u 2 + v 2 ) = 0
▶ That is u − a(u 2 + v 2 ) = 0 which is a circle, if a ̸= 0.
▶ a(u 2 + v 2 ) − u = 0 ⇒ au 2 − u + av 2 = 0
1 2 1 2
▶ Write it in standard form: (u − 2a ) + v 2 = ( 2a )
1
▶ Circle with center at ( 2a 1
, 0) and radius 2a .
▶ So it passes through the origin.
Do the same for y = a.
Ploting
 1
Image circle x 2 + y 2 = a2 under f (z) = z

Image of x 2 + y 2 = a2 :
▶ Consider A(x 2 + y 2 ) + Bx + Cy + D = 0
▶ Take A = 1, B = 0, C = 0, D = −a2
▶ Image is 1 + 0u − 0v − a2 (u 2 + v 2 ) = 0
▶ That is u 2 + v 2 = a12 which is a circle with center at origin and
radius a1 .
 1
Image circle x 2 + y 2 = a2 under f (z) = z polar method

Image of x 2 + y 2 = a2 . That is r = a:
▶ R = 1r = a1 , ϕ = −θ
▶ It is a circle with center at origin and radius a1 but traced
clockwise.
▶ But the cartesian method gives all circles (with any center and
any radius)
Ploting
The mapping f (z) = sin z

▶ f (z) = sin(x + iy) = sin x cos(iy) + cos x sin(iy) =

sin x cosh y + icos x sinhy
▶ So u = sin x cosh y and v = cos x sinhy
▶ sin z is periodic, with period 2π
▶ So we need to consider only one period of sine. That is take the
domain as z = x + iy with − 12 π ≤ x ≤ 21 π.
▶ The map is not conformal at the end points − π2 , π2 (f ′ (z) = 0)
The mapping f (z) = sin z

Image of straightline y = a.
▶ u = sin x cosh a and v = cos x sinh a
▶ Use the identity cos2 t + sin2 t = 1 for all t, of trigonometric
functions
▶ Eliminating x using the above:

u2 v2
+ = 1,
cosh2 a sinh2 a
which is an ellipse whose major axis is of length cosh a and
minor axis of length sinh a. (Assuming a to be such that
sinh a ̸= 0)
2 2
▶ Note: xa2 + yb2 = 1 is an ellipse, whose major and minor axis are
of lengths a and b, respectively.
The mapping f (z) = sin z

Image of straightline x = a.
▶ u = sin a coshy and v = cos a sinhy
▶ Use the identity cosh2 t − sinh2 t = 1 for all t, of hyperbolic
functions
▶ Eliminating y using the above:

u2 v2
− = 1,
sin a cos2 a
2

which is a hyperbola whose asymptotes are y = ± cos a

sin a x
(Assuming sin a, cos a ̸= 0)
2 2
▶ Note: xa2 − yb2 = 1 is a hyperbola, whose asymptotes at infinity
are y = ba x and y = − ba x
* Dealing with circles is difficult in this map!
Example

Image of straightline y = 1.
▶ u = sin x cosh 1 and v = cos x sinh 1
▶ Use the identity cos2 t + sin2 t = 1 for all t, of trigonometric
functions
▶ Eliminating x using the above:

u2 v2
+ = 1,
cosh 1 sinh2 1
2

which is an ellipse whose major axis is of length cosh 1 ≈ 1.54

and minor axis of length sinh 1 ≈ 1.17.
Example

Image of straightline x = 1.
▶ u = sin 1 coshy and v = cos 1 sinhy
▶ Eliminating y using the above:

u2 v2
− = 1,
sin 1 cos2 1
2

which is a hyperbola whose asymptotes are y = ± cos 1

sin 1 x and it
cuts the u axis at (±sina, 0).
Proof(1) of Conformal mapping
Consider an analytic function f : C → C and two parametric curves
C1 , C2 in the domain plane intersecting at z0 . Let the parametric
equations of the curves be z1 (t) and z2 (t) where t ∈ [a, b].
We need to see the relation between their angles and their images
angle at an intersecting point z0 = z1 (t0 ) = z2 (t0 ).
▶ w1 (t) = f (z1 (t)) ⇒ dw ′ dz1
dt |t0 = f (z0 ) dt |t0
1

Arg(w1′ (t0 )) = Arg(f ′ (z0 )) + Arg(z1′ (t0 ))

Same thing happens for the second curve also

Arg(w2′ (t0 )) = Arg(f ′ (z0 )) + Arg(z2′ (t0 ))

▶ So the different in angles (angle between the tangents) are equal

for curves and their images:

Arg(w2′ (t0 )) − Arg(w1′ (t0 )) = Arg(z2′ (t0 )) − Arg(z1′ (t0 ))

▶ Note that Arg(f ′ (z0 )) is not defined when f ′ (z0 ) = 0

Proof(2) of uniqueness of bilinear transformation

Suppose there are two bilinear maps f (z) = ac11 z+d

z+b1
1
, g(z) = a2 z+b2
c2 z+d2
taking z1 , z2 , z3 to w1 , w2 , w3 .
▶ We need to show that f (z) = g(z) for all z
▶ We know that f (z1 ) = g(z1 ), . . .
▶ We have seen that g −1 is also a bilinear transformation taking
w1 , w2 , w3 to z1 , z2 , z3
▶ g −1 f is also a bilinear transformation
▶ g −1 (f (z1 )) = z1 , g −1 (f (z2 )) = z2 , g −1 (f (z3 )) = z3
▶ Then g −1 f has three fixed points z1 , z2 , z3
▶ So g −1 f (z) = z for all z, identity. That is f = g.
Example

Find the bilinear map which maps 1, i, −1 to i, 1, −i

▶
THANK YOU