THE UNIVERSITY OF HONG KONG

DEPARTMENT OF MATHEMATICS
Year 2023-2024

MATH3403

Functions of a Complex Variable

Last update: January 15, 2024

By Kwok Kin WONG
 THE UNIVERSITY OF HONG KONG
DEPARTMENT OF MATHEMATICS
MATH3403: Functions of a Complex Variable

I. Complex Numbers

The field of complex numbers, polar representation, identities and inequalities, sequences and
series, absolute convergence.

II. Functions on Plane Domains

Continuous functions, complex-differentiability, analytic functions, Cauchy-Riemann equa-

tions, real-differentiability, harmonic functions, harmonic conjugates.

III. Power Series

Radius of convergence, the Cauchy-Hadamard Formula, analytic functions defined by conver-

gent power series. Exponential function, natural logarithm.

IV. Path and Contour Integration

Path integration, primitives, Cauchy’s Theorem, the Cauchy Integral Formula, mean-value
inequality, Maximum Principle, Cauchy’s estimates, Liouville’s Theorem.

V. Function Theory

Taylor’s expansion, zeros of analytic functions and factorization, argument principle. mero-
morphic functions: poles, principal parts and residues, Laurent series expansions on annuli,
singularities of analytic functions on punctured disks.

VI. Residue Calculus

The Residue Theorem, calculation of contour integrals, application to the calculation of some
improper integrals.

References:

1. L.V. Ahlfors, Complex Analysis (McGraw-Hill, 3rd edition).

2. K. Kodaira, Introduction to Complex Analysis (Cambridge).

3. J. Bak and D. J. Newman, Complex Analysis, Undergraduate Texts in Mathematics (Springer-

Verlag, 2nd edition).

4. R.B. Ash and W.P. Novinger: Complex Variables (Dover, 2nd edition).
http://www.math.uiuc.edu/∼r-ash/CV.html
Contents

1 Complex Numbers 1
1.1 Basic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Limit and Continuity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

i
 1

§1
Complex Numbers

1.1 Basic Properties

In the set R of real numbers, there is no x ∈ R satisfying the equation x2 + 1 = 0. An imagined
number √
i := −1
is taken to be a solution to the equation x2 + 1 = 0. Thus
i2 = −1.
We agree that the equation x2 + 1 = 0 has two solutons and the other solution is −i.
The set of complex numbers is defined to be
C := {x + iy : x, y ∈ R}.
We call x the real part of z, written x = Re(z); and y the imaginary part of z, written y = Im(z).

1.1.1 Arithmetic operations

We equip C with the following operations: for z = x + iy, w = u + iv ∈ C,
Addition: z + w := (x + iy) + (u + iv) := (x + u) + i(y + v);
Multiplication: z · w := (x + iy)(u + iv) := (xu − yv) + i(xv + yu).

For a complex number z = x + iy, the number −z is defined by the identity z + (−z) = 0. Thus
−z = −x − iy which is called the additive inverse of z. The subtraction is now defined by
z − w := z + (−w) = (x + iy) + (−u − iv) = (x − u) + i(y − v).
1 1 1
For a complex number z, we write for a complex number satisfying z · = 1. If exists it
z z z
will be called the multiplicative inverse of z, or the reciprocal of z. Write z = x + iy. Observe
the important identity
(x + iy)(x − iy) = (x2 + y 2 ) + i(−xy + xy) = x2 + y 2
which is a nonnegative real number. Thus, for z = x + iy ̸= 0, we have
x2 + y 2
 
x − iy
(x + iy) = = 1.
x2 + y 2 x2 + y 2
1.1. Basic Properties 2

1
In other words, for z ̸= 0, the reciprocal exists and it is given by the formula
z
1 x − iy
= 2 .
z x + y2
When z = 0, 0 · w = 0 for any complex number w, so there is no multiplicative inverse. Now
division on complex numbers is given by
 
w 1 x − iy (u + iv)(x − iy) (ux + vy) + i(vx − uy)
:= w · = (u + iv) 2 2
= 2 2
= .
z z x +y x +y x2 + y 2

1.1.2 Complex conjugation

For z = x + iy; x, y ∈ R, define
z := x − iy,
called the (complex) conjugate of z. The operation ‘ · ’ is called (complex) conjugation.
Observe that
(a) z + w = z + w;
(b) zw = z w.
These properties of complex conjugation, together with the obvious facts 0 = 0, 1 = 1, imply that
(c) z − w = z − w;
w w
(d) = , whenever z ̸= 0.
z z
Proposition 1.1. Let a0 , · · · , an be real numbers. Suppose z ∈ C satisfies the equation an z n +
an−1 z n−1 + · · · + a1 z + a0 = 0. Then z ∈ C satisfies the equation an z n + an−1 z n−1 + · · · a1 z + a0 = 0
Proof. Taking conjugates we have
0 = 0 = an z n + an−1 z n−1 + · · · + a1 z + a0
= an z n + an−1 z n−1 + · · · + a1 z + a0
Now, by an induction argument z k = z k for any positive integer k. Moreover ak = ak since ak is
real for k = 0, 1, · · · , n. Hence
0 = an z n + an−1 z n−1 + · · · + a1 z + a0 .

Remark. For the case of quadratic polynomials P (z) = az 2 + bz + c; a, b, c ∈ R, we have

√
−b ± ∆
z= , where ∆ = b2 − 4ac (discriminant).
2a
There are 3 cases:
(a) ∆ > 0, and we have 2 distinct real roots
(b) ∆ = 0 and we have a single double root, which is real.
(c) ∆ < 0, ∆ = −|∆| and we have two distinct complex roots
p
−b ± i |∆|
z= ,
2a
which are complex conjugates of each other.
1.1. Basic Properties 3

1.1.3 Polar representation

A point z = x + iy ∈ C can be identified with the ordered pair (x, y) ∈ R2 . Therefore C is usually
called the complex plane. The visualization of C as R2 is called Argand’s diagram.
We can write
z = x + iy = r cos θ + (r sin θ)i =: r cis θ .
The pair (r, θ) is called the polar coordinates. By convention we will take r ≥ 0. The polar
p y
radius r ≥ 0 is given by r = x2 + y 2 , and the polar angle θ satisfies tan θ = . Note that
x
r2
r2 = x2 + y 2 = (x + iy)(x − iy) = z z; i.e., z =
z
whenever z ̸= 0. The polar radius
|z| := r
is also called the modulus, or norm, or absolute value of z.
For z = r cis θ, w = ρ cis φ with r, ρ ≥ 0 and θ, φ ∈ R, since

cis θ · cis φ = (cos θ + i sin θ)(cos φ + i sin φ)

= (cos θ cos φ − sin θ sin φ) + i(cos θ sin φ + sin θ cos φ)
= cos(θ + φ) + i sin(θ + φ) = cis (θ + φ),

we have
z · w = rρ cis θ · cis φ.
As a consequence, we have

Theorem 1.1 (de Moivre’s Law). Let z = r cis θ in polar representation. Then z n = rn cis (nθ)
for any positive integer n. If z ̸= 0 we also have z n = rn cis (nθ) for any integer n.

Proof. We prove by induction. Let n be a positive integer. For n = 1 the theorem is obvious.
Suppose z k = rk cis (kθ) for 1 ≤ k ≤ n − 1. Then

z n = z (n−1)+1 = z n−1 · z = rn−1 cis ((n − 1)θ) · r cis θ

= r(n−1)+1 cis ((n − 1)θ + θ) = rn cis (nθ) .

For the last statement suppose z ̸= 0, i.e., r ̸= 0. z 0 = 1 by convention. For n = −m < 0, we have
 m
1
z n = z −m = .
z
1 1 1
Now for z = r cis θ ̸= 0, we have = · .
z r cis θ
1
Since cis θ · cis (−θ) = cis (θ − θ) = cis 0 = 1, we have cis (−θ) = . Hence,
cis θ
1 1 1 cis (−θ)
= · = ;
z r cis θ r
 m  m
1 1
n
z = = cis (−mθ) = r−m cis (−mθ) = rn cis (nθ),
z r
as desired.
1.1. Basic Properties 4

1.1.4 Roots of unity

Let n be a positive integer. An n-th root of unity is a complex number z satisfying z n = 1.

Problem. Find all n-th roots of unity, i.e., solve z n = 1.

Solution. Write z = r cis θ. By de Moivre’s Law,

1 = z n = rn cis (nθ)

Hence, rn = 1, cis (nθ) = 1. Now, cis (nθ) = cos(nθ) + i sin(nθ), so cos(nθ) = 1, sin(nθ) = 0, i.e.,
nθ = 2kπ for some integer k. Thus
2kπ
θ= ; k = · · · , −2, −1, 0, 1, 2, · · ·
n
2π 4π (2n − 2)π
θ ≡ 0, , ,··· , mod 2π .
n n n
Here we write α ≡ β mod 2π if and only if α − β = 2ℓ π for some ℓ ∈ Z. Note that cis α = cis β
whenever α ≡ β mod 2π.      
2π 4π (2n − 2)π
Thus, the n distinct numbers cis 0 = 1, cis , cis , · · · , cis on the unit
n n n
circle divides the unit circle into n circular arcs of equal length, and they give all the n-the roots
of unity. ■

1.1.5 Identities and inequalities on complex numbers

Recall z = x + iy := Re(z) + i Im(z). We have the following preliminary facts.

(i) | Re(z)| ≤ |z|; | Im(z)| ≤ |z| (which follows from Pythagoras’ Theorem, i.e., |z|2 = x2 + y 2 );

(ii) |z|2 = |x − iy|2 = x2 + (−y)2 = x2 + y 2 = |z|2 , i.e., |z| = |z|.

We collect in what follows basic identities and inequalities (a)-(d) on complex numbers.

(a) zz = |z|2

(b) |zw| = |z||w|

(c) Triangular inequality: |z + w| ≤ |z| + |w|.

(c′ ) Variant form of the triangular inequality: |z + w| ≥ |z| − |w| .

(d) Parallelogram law: |z + w|2 + |z − w|2 = 2|z|2 + 2|w|2 .

Thus, the sum of the squares of lengths of the two diagonals of a parallelogram is equal to
the sum of squares of lengths of the four edges.

Proof. (a) z z̄ = r cis(θ)r cis(−θ) = r2 = |z|2 .

(b) We have zw = (r cis θ)(ρ cis φ) = rρ cis θ cis φ = rρ cis (θ + φ). Thus,

|zw| = rρ = |z||w| .
1.2. Limit and Continuity 5

(c)
|z + w|2 = (z + w)(z + w) = zz + zw + wz + ww
= |z|2 + 2 Re(zw) + |w|2 ;
(|z| + |w|)2 = |z|2 + 2|z||w| + |w|2 .
We have

|z + w|2 ≤ |z|2 + 2|zw| + |w|2 = |z|2 + 2|z||w| + |w|2 = (|z| + |w|)2 .

Hence
|z + w| ≤ |z| + |w| .

(c′ ) Write z = (z + w) + (−w). Then, by (c)

|z| ≤ |z + w| + | − w| = |z + w| + |w|

Thus
|z + w| ≥ |z| − |w|.
Interchanging z and w we have |z + w| ≥ |w| − |z|. Hence,

|z + w| ≥ max (|z| − |w|, |w| − |z|) = |z| − |w| .

(d)
|z + w|2 = |z|2 + 2 Re (zw) + |w|2 ;
|z − w|2 = |z|2 − 2 Re (zw) + |w|2 .
Thus,
|z + w|2 + |z − w|2 = 2|z|2 + 2|w|2 .

Remark. The variant form of the triangular inequality can be used to give an upper bound on
reciprocals. Thus, if |z| > |w|, we have
1 1
≤
|z + w| |z| − |w|

since |z + w| ≥ |z| − |w| > 0.

1.2 Limit and Continuity

1.2.1 Sequence of complex numbers
A sequence of complex numbers is a mapping ζ : N → C from the set of natural numbers to the set
of complex numbers. It is normally written as (zn )n=∞
n=1 , or just (zn ), where zn := ζ(n).

Remark. In place of starting from n = 1, one may start with n = 0, or any integer.
1.2. Limit and Continuity 6

A sequence (zn ) is said to converge to L ∈ C, written lim zn = L or zn → L as n → ∞, if and

n→∞
only if the following holds true:
Given ε > 0, there exists N = N (ε) ∈ N such that |zn − L| < ε whenever n ≥ N .
If (zn ) is not convergent, then we say that (zn ) is divergent.
Proposition 1.2.
(
limn→∞ Re zn = Re L
lim zn = L ⇐⇒ .
n→∞ limn→∞ Im zn = Im L

Proof. Write zn = xn + iyn , L = a + ib. Then

(=⇒) zn − L = (xn + iyn ) − (a + ib) = (xn − a) + i(yn − b). Thus,

|xn − a| = | Re(zn − L)| ≤ |zn − L|;

|yn − b| = | Im(zn − L)| ≤ |zn − L| .

Now, zn → L is equivalent to limn→∞ |zn − L| = 0. Hence,

0 ≤ |xn − a| ≤ |zn − L| → 0 as n → ∞;
0 ≤ |yn − b| ≤ |zn − L| → 0 as n → ∞.

Hence, zn → L implies xn → a and yn → b, as desired.

ε ε
(⇐=) For any > 0, there exists some N = N (ε) ∈ N such that for all n ≥ N , |xn − a| < and
2 2
ε
|yn − b| < . Then,
2
|zn − L| = |(xn − a) + i(yn − b)| ≤ |xn − a| + |yn − b| < ε

by the Triangular Inequality. Hence, xn → a and yn → b =⇒ zn → L = a + ib.

Basic rules
(a) zn → L, zn′ → L′ =⇒ zn ± zn′ → L ± L′

(b) zn → L and λ ∈ C =⇒ λzn → λL

(c) zn → L, zn′ → L′ =⇒ zn zn′ → LL′

′
zn L′
(d) zn → L ̸= 0, zn′ → L′ =⇒ zn →
L
Rules (a) and (b) are straightforward. We give here the proofs for (c) and (d).

Proof. (c) Write wn = zn zn′ . Then

wn − LL′ = zn zn′ − LL′ = (zn zn′ − Lzn′ ) + (Lzn′ − LL′ ) = (zn − L)zn′ + L(zn′ − L′ ).

ε ε
Given ε > 0, choose N ∈ N such that |zn − L| < ′
, |zn′ − L′ | < and
|L| + |L | + 1 |L| + |L′ | + 1
|zn′ − L′ | < 1 whenever n ≥ N .
1.2. Limit and Continuity 7

From |zn′ − L′ | < 1 it follows that |zn′ | < |L′ | + 1. Hence,

|wn − LL′ | ≤ |zn − L||zn′ | + |L||zn′ − L′ |

ε ε
< ′
(|L′ | + 1) + (|L|) = ε.
|L| + |L | + 1 |L| + |L′ | + 1

Thus, lim zn zn′ = lim wn = LL′ , as desired.

n→∞ n→∞

(d) Assume (c). Then

zn′ 1 1 L′
= zn′ · → L′ · =
zn zn L L
1 1
if we can show that lim = .
n→∞ zn L
1 1
Thus, to prove (d) it suffices to prove lim = .
n→∞ zn L
1 1 (L − zn )
In this case we have − = .
zn L zn · L
|L| |L|
Choose N large enough so that |zn − L| < whenever n > N . Hence |zn | ≥ .
2 2
1 1 |zn − L| 2
Then, − < |L|
= |z − L|.
2 n
zn L |L|( 2 ) |L|
ε|L|2
Since zn → L, given ε > 0 and choosing N large enough we have |zn − L| < whenever
2
n ≥ N.
ε|L|2
 
1 1 2
Then, for n ≥ N we have − < = ε.
zn L |L|2 2
1 1
Thus, lim = , as desired.
n→∞ zn L

1.2.2 Cauchy sequences

A sequence of complex numbers (zn ) is said to be a Cauchy sequence if and only if the following
statement holds: For all ε > 0, there exists N = N (ε) such that

|zn − zm | < ε, whenever n, m ≥ N.

Recall that a sequence (un ) of real numbers is convergent if and only if it is a Cauchy sequence.

Lemma 1.1. Write zn = xn + iyn ; xn = Re(zn ), yn = Im(zn ). Then (zn ) is a Cauchy sequence if
and only if both (xn ) and (yn ) are Cauchy sequences.

Proof. The proof follows readily from

|xn − xm |, |yn − ym | ≤ |zn − zm | ≤ |xn − xm | + |yn − ym |.

Theorem 1.2. A sequence of complex numbers is convergent if and only if it is a Cauchy sequence.
1.2. Limit and Continuity 8

Proof. (=⇒) Obvious.

(⇐=) Suppose (zn ) is a Cauchy sequence. Then, for any ε > 0, there exists N ∈ N such that
|zn − zm | < ε whenever, n, m ≥ N . By

|xn − xm | ≤ |zn − zm |, |yn − ym | ≤ |zn − zm |;

we know that (xn ) and (yn ) are Cauchy sequences.

Hence, each of (xn ) and (yn ) are convergent, i.e., xn → a, for some a ∈ R; yn → b for some b ∈
R. Hence
zn = xn + iyn → a + ib.

A sequence (zn ) in C is said to be bounded if and only if there exists a non-negative real
number M such that |zn | ≤ M for every n ∈ N.

Remark. If (zn ) is convergent, then (zn ) is bounded.

A subsequence of (zn ) is a sequence (vn ) defined by vn = zσ(n) for some injective strictly
increasing function σ : N → N. The Bolzano-Weierstrass theorem over R says that any bounded
sequence (un ) of real numbers has a convergent subsequence (uσ(n) ). Similarly,

Theorem 1.3 (Bolzano-Weierstrass over C). Any bounded sequence (zn ) of complex numbers has
a convergent subsequence (zσ(n) ).

Proof. Write zn = xn + iyn ; xn , yn ∈ R. Since |xn |, |yn | ≤ |zn |, (xn ) and (yn ) are bounded real
sequences. By Bolzano-Weierstrass Theorem for bounded real sequences, there exists a convergent
subsequence (xµ(n) ) of (xn ).
Consider now the bounded sequence (yµ(n) ) of real numbers. There exists a convergent subse-
quence (yµ(τ (n)) ) of (yµ(n) ), τ : N → N being some strictly increasing mapping.
Write σ(n) := µ(τ (n)). We have lim yσ(n) := b for some b ∈ R. On the other hand lim xµ(n) :=
n→∞ n→∞
a ∈ R, and hence
lim xσ(n) = lim xµ(τ (n)) = lim xµ(n) = a.
n→∞ n→∞ n→∞

Therefore,
lim zσ(n) = lim xσ(n) + i lim yσ(n) = a + ib,
n→∞ n→∞ n→∞

proving the theorem.

1.2.3 Continuous functions

Definition 1.1. Let f be a complex-valued function defined on Ω ⊂ C. The function f is said to
have a limit L at a ∈ Ω, written as
lim f (z) = L,
z→a

if and only if the following holds true:

Given any ε > 0, there exists δ > 0 such that |f (z) − L| < ε whenever |z − a| < δ.
1.2. Limit and Continuity 9

The function f is said to be continuous at a ∈ Ω if and only if

lim f (z) = f (a).

z→a

If f is continuous at all points of Ω, f is said to be continuous.

Let f, g be continuous complex-valued functions defined on Ω ⊂ C. Then f + g, f · g are

continuous and f /g is continuous at points a where g(a) ̸= 0. Moreover, Re f, Im f, |f | are all
continuous.

Domains in C
An open disk in C is a subset defined by

D(a; r) := {z ∈ C : |z − a| < r}

for some a ∈ C and 0 < r < ∞. A subset G ⊂ C is said to be open if it can be written as the
union of a collection of open disks. In particular, an open disk is open. A subset G ⊂ C is open if
and only if for any given a ∈ G, there exists ra > 0 such that D(a; r) ⊂ G.
An open subset Ω ⊂ C is said to be connected if and only if it cannot be decomposed into the
union of two disjoint non-empty open subsets. A connected open subset is said to be a domain.