MATH 300 REVIEW

1. The Arithmetic of Complex Numbers

1.1. The Algebra of Complex Numbers.
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)

(10)
(11)
(12)
(13)

(14)
(15)
(16)

1, 2, 3, many
1, 2, 3, . . .
Operations on positive integers; addition and multiplication: a + b and ab
The Greeks thought of numbers in terms of lengths or areas; the number line; interpreting numbers in terms of Euclidean geometry.
Solving the linear equation x + b = a : solution x = b a.
The number 0 and negative numbers.
Solving the linear equation bx = a : solution x = a/b.
The rational numbers Q.
Arithmetic properties of Q : commutativity of addition and multiplication; existence
of 0 and 1; inverses for addition and multiplication; associativity of addition and multiplication; the distributive law.
Are there more numbers?

The diagonal of the unit square; irrationality of 2; the number field Q( 2).
The real numbers R; interpret in terms of the number line.
Are there more numbers? Consider
quadratic equations ax2 + bx + c = 0, where a 6= 0.

b b2 4ac
The solution is x =
. If b2 4ac < 0 this makes no sense, unless of
2a
course we introduce some new numbers. In particular will stand for a particular
square root of 1.
Adjoin to R to get the complex number field C; a+b, where a and b are real numbers.
Arithmetic properties of the complex number field.
Algebraic numbers, transcendental numbers.

1.2. Point Representation of Complex Numbers.

(1) The Argand diagram.
(2) Real and imaginary axes; real and imaginary parts of a complex number z = x + y are
Re(z) = x, Im(z) = y.
p
(3) The modulus of z = x + y is |z| := x2 + y 2 ; the conjugate of a complex number,
z = x + y = x y; algebraic properties of conjugation; z = z z is real.
(4) |z z0 | = r represents a circle.
1.3. Vectors and Polar Forms.
(1) The complex number z = x + y can be identified with the vector (x, y).
(2) The triangle inequality |z1 + z2 | |z1 | + |z2 |.
1

(3)
(4)
(5)
(6)

Exercise |z2 | |z1 | |z2 z1 |.

The arguments of a complex number; the principal argument ( < .)
z = x + y = r(cos + sin ), where r = |z|.
cis = cos + sin ;

z1 z2 = r1 cis1 r2 cis2 = r1 r2 cis(1 + 2 ).

1+
1
= cis5/12.
(7) Examples 1 + 3 = 2cis/3;
3
2
1.4. The Complex Exponential.
1 2 1 3
x + x + converges for all x.
2!
3!
How do we define ez = ex+y for a complex number z? Want the usual laws of exponents
to hold for the complex exponential function, for example ez1 ez2 = ez1 +z2 . The correct
definition is
1
1
exp(z) = ex = 1 + z + z 2 + z 3 +
2!
3!
This converges for all z.
ex+y = ex (cos y + sin y); |ez | = ex .
Polar form of a complex number ez = re .
z1 z2 = r1 r2 e(1 +2 ) .
De Moivres formula (cos + sin )n = cos n + sin n.

(1) exp(x) = ex = 1 + x +
(2)

(3)
(4)
(5)
(6)

1.5. Powers and Roots.

(1) De Moivres formula allows us to find simple formulas for the nth roots of a complex
number z = re . There are n roots, they are
k = r1/n e(+2k)/n , k = 0, 1, . . . , n 1.
(2) The nth roots of unity on the unit circle.
(3) Examples: the 3th , 4th and 6th roots of unity; the 8th roots of unity.
We end this section by explaining how to solve a cubic equation a0 z 3 + a1 z 2 + a2 z + a3 = 0.
By simple substitutions we can assume the cubic has the form z 3 + az + b = 0. Then we make
the substitution
z = u + v, where we choose u, v to satisfy 3uv + a = 0.
Thus
z 3 + az + b = (u + v)3 + a(u + v) + b
= u3 + 3u2 v + 3uv 2 + v 3 + a(u + v) + b
= u3 + v 3 + (3uv + a)(u + v)) + b
a3
+b
= u3
27u3

Therefore
a3
a3
6
3
z 3 + az + b = 0 u3
+
b
=
0

u
+
bu

= 0
27u3
27
!1/3
p
p
2 + 4a3 /27
2 + 4a3 /27
b

b
b
b

u3 =
u =
2
2
Theorem 1. The solutions of the cubic z 3 + az + b = 0 are
!1/3
!1/3
p
p
b b2 + 4a3 /27
b + b2 + 4a3 /27
+
z1 =
2
2
!1/3
!1/3
p
p
2 + 4a3 /27
b + b2 + 4a3 /27
b

b
z2 =
+ 2
2
2
!1/3
!1/3
p
p
2 + 4a3 /27
2 + 4a3 /27
b
+
b
b

b
z3 = 2
+
2
2
where = e2/3 and the cube roots are chosen so that
!1/3
!1/3
p
p
b + b2 + 4a3 /27
b b2 + 4a3 /27
a

=
2
2
3
Example: If we apply this theorem to the cubic z 3 15z 4 = 0 we find that the solutions
are
z1 = (2 + 11)1/3 + (2 11)1/3
z2 = (2 + 11)1/3 + 2 (2 11)1/3
z3 = 2 (2 + 11)1/3 + (2 11)1/3 ,
where the cube roots are chosen so that (2 + 11)1/3 (2 11)1/3 = 5.
We end this section with an important theorem.
Theorem 2. The Fundamental Theorem of Algebra The solutions of any polynomial
equation a0 z n + a1 z n1 + + an1 z + an = 0 all lie in C.
1.6. Planar Sets.
(1)
(2)
(3)
(4)
(5)
(6)
(7)

Functions of a real variable are defined on intervals.

|z z0 | > r is an open disc.
Interior points of a set S in C; open subsets
of C;


Examples |z| > R; |z| R; {z = x + y  y 6= 0}
An open set is connected if any 2 points can be joined by a polygonal path
A domain is an open connected set.
The boundary or frontier of a subset; closed subsets.

1.7. The Riemann Sphere and Stereographic projection.

(1) The unit sphere in R3 is given by x21 + x22 + x23 = 1; a sketch of the unit sphere and its
equatorial plane; the line through the north pole (0, 0, 1) and a point z = x + y in the
complex plane C.
(2) A one-to-one correspondence between points Z = (x1 , x2 , x3 ) on the unit sphere, other
than the north pole, and points z = x + y in the complex plane C
(3)
2y
x2 + y 2 1
2x
,
x
=
,
x
=
x1 = 2
2
3
x + y2 + 1
x2 + y 2 + 1
x2 + y 2 + 1
(4) The inverse transformation is
x1
x2
x=
, y=
1 x3
1 x3
(5) Circles on the Riemann sphere correspond to circles and lines in the complex plane C.
:= C{}.
(6) The north pole (0, 0, 1) corresponds to a point at in the complex plane; C
(7) There is only one point at infinity in C.
(8) Some topology of the extended complex plane.
2. Analytic Functions
2.1. Analytic Functions. In this section we study complex valued functions of a complex
variable, that is w = f (z), where z C and w C.
If z = x + y we write w = f (z) = u(x, y) + v(x, y).
In particular we want to differentiate and integrate such functions.
Example 1. Describe the image of the unit disc under the map f (z) = z n .
Example 2. The mapping f (z) = 1/z corresponds to rotation by about the x1 axis on the
Riemann sphere.
2.2. Limits and Continuity.
Definition 1. A sequence of complex numbers {zn }n1 converges to z0 if
given any  > 0 N such that |zn z0 | <  if n N
Definition 2. Suppose f (z) is a function defined in some open neighbourhood of z0 , except
possibly at z0 itself. We say that the limit of f (z) as z approaches z0 is w0 , and write
limzz0 f (z) = w0 (equivalently; f (z) w0 as z z0 ) if, given any  > 0 > 0 such
that
|z z0 | < = |f (z) f (z0 )| < .
Definition 3. Suppose f (z) is defined in a neighbourhood of z0 . Then f (z) is continuous at
z0 if limzz0 = f (z0 ).

Some obvious arithmetic properties of limits and continuity. So for example all polynomials
are continuous on C.
2.3. Analyticity. The basic notion here is the concept of the derivative of a complex valued
function w = f (z) :
Definition 4.

df
f (z0 + z) f (z0 )
= f 0 (z0 ) = lim
z0
dz
z

In this definition we are allowed to let z approach zero in any way possible.
Example 3. f (z) = z is nowhere differentiable. On the other hand the power functions
z z n , for n an integer, are differentiable everywhere.
Rules of differentiation. All polynomials are differentiable everywhere. Differentiable functions
are continuous.
Definition 5. A function f (z) is analytic on an open set if f 0 (z) z .
2.4. The Cauchy-Riemann Equations.
Theorem 3. A necessary condition for a function w = f (z) = u(x, y) + v(x, y) to be analytic
at a point z = x + y is that the Cauchy-Riemann equations
u
v u
v
=
,
=
(Cauchy-Riemann equations)
x
y y
x
are satisfied at that point.
To derive the Cauchy-Riemann equations let z 0 through real values (z = x) or
complex values (z = y). This gives
f 0 (z) =

u
v
v v
+
=

x
x
y x

Theorem 4. Suppose f (z) = u(x, y) + v(x, y) is defined in an open set containing z0 . If all
first partial derivatives of u and v exist in , are continuous at z0 and the Cauchy-Riemann
equations hold at z0 then f (z) is differentiable at z0 .

2.5. Harmonic Functions. Solutions of Laplaces equation 2 :=

2 2
+
= 0 are called
x2 y 2

harmonic functions.
Example: The 2-dimensional heat equation on a region is
 2

u
2u
u
2
=c
+
, (x, y)
t
x2 y 2
Steady state heat flow is governed by Laplaces equation. We often impose boundary conditions
of the form
u(x, y) = some given function on the boundary of
Definition 6. A real valued function (x, y) is called harmonic in a domain if all of its
second order partials are continuous on and Lalaces equation holds in .
Theorem 5. If f (z) = u(x, y) + v(x, y) is analytic in then u(x, y) and v(x, y) are harmonic
in .
Proof. Assuming analyticity implies continuity of all second partials this follows from the
Cauchy-Riemann equations.

Examples:
(1) z 3 and ez .
(2) Show that u = x3 3xy 2 + y is harmonic and find a harmonic conjugate v(x, y).
(3) Find a function (x, y) that is harmonic in the region = {z  x > 0, 2 x2 y 2 4}
and which satisfies

if x > 0 and x2 y 2 = 2
u(x, y) =
if x > 0 and x2 y 2 = 4

(4) Find a function (x, y) that is harmonic in the region = {z  y > 0, 2 xy 4} and
which satisfies

if y > 0 and xy = 2
u(x, y) =
if y > 0 and xy = 4
Theorem 6. If f (z) = y(x, y) + v(x, y) is analytic then the family curves u(x, y) = const.
is mutually perpendicular to the family of curves v(x, y) = const at each point of intersection
(x, y) if f 0 (z) 6= 0 at the point of intersection.
Level curves of a harmonic function and one of its conjugates intersect at right angles at all
points where the gradients are 6= 0.
Examples:
(1) f (z) = z 2 = x2 y 2 + 2xy.

2.6. Steady State Temperature as a Harmonic Function. Omit this section.

2.7. Iterated Maps: Julia and Mandelbrot Sets. Omit this section.
3. Elementary Functions
3.1. Polynomials and Rational Functions. The general polynomial function is
P (z) = a0 + a1 z + a2 z 2 + + an z n , where the ai C and an 6= 0
The degree of P (z) is deg(P ) = n.
Definition 7. A rational function R(z) is a function of the form
R(z) =

P (z)
a0 + a1 z + a2 z 2 + + am z m
, where am 6= 0, bn 6= 0
=
Q(z)
b0 + b1 z + b2 z 2 + + bn z n

We assume that P (z) and Q(z) have no common factor. By the fundamental theorem of
algebra we have
P (z)
am (z z1 )d1 (z z2 )d2 (z zr )dr
R(z) =
=
Q(z)
bn (z 1 )e1 (z 2 )e2 (z s )es
s
r
X
X
di = m and
ei = n. There is no intersection amongst the sets {z1 , . . . , zr } and
where
i=1

i=1

{1 , . . . , s }.
Definition 8. We say that i is a pole of R(z) of order ei .
1
= .
z (z )e
Lemma 1. limzi R(z) = .
Example: lim

Example: The equation ez = w, where w is any non-zero complex number, has infinitely many
solutions in any neighbourhood of . Let w = re and z = x + y. Then
ex+y = ex ey = re ex = r and ey = e x = ln r and y = + 2k.
Exercise: Show that given any complex number w = re 6= 0 a sequence {zn } such that
all zn 6= 0, zn 0 as n , and e1/z = w.
Partial Fraction Decomposition:
A1,e1 2
A1,e1 1
A1,0
+ +
+
+
R(z) =
2
z 1
(z 1 )
(z 1 )e1
A2,e2 1
A2,e2 2
A2,0
+
+ +
+
2
z 2
(z 2 )
(z 2 )e2
As,es 1
As,es 2
As,0
+
+
+ +
2
z s
(z s )
(z s )es

The coefficients are computed by the formula


1 dj
ek 
Ak,j =
, 1 k s, 0 j ek 1
R(z)(z

)

k
j! dz j
z=k
z2 1
1
Example: The partial fraction decomposition of the functions
.
2 and n
2
z 1
(z + 1)
3.2. The Exponential, Trigonometric and Hyperbolic Functions.
ez = ex+y := ex ey = ex cos y + (ex sin y)
1
1
1
= 1 + z + z2 + z3 + + zn +
2!
3!
n!
z1
z2
Lemma 2. e = e z1 = z1 + 2k for some integer k.
Thus the function ez is periodic with period 2. A fundamental region for ez is any one of the
following sets:

Sn = {x + y  < x < , (2n 1) < y (2n + 1)}
Definition 9.
ez + ez
ez ez
, sin z :=
2
2
z
z
z
e +e
e ez
cosh z :=
, sinh z :=
2
2
cos z :=

From these definitions all the usual formulas for single variable calculus follow (formulas for
derivatives, various identities, etc.)
3.3. The Logarithmic Function. Single valued functions vs multi-valued functions. Taking
the inverse of a single valued function often leads to a multi-valued function.
Example:
z = ew log(z) = w.
The function ew is a single valued function of w, but log z is multi-valued. If w = u + v then
log(z) = u + (v + 2k). The periodicity of the exponential function becomes multi-valuedness
for the inverse function.
Definition 10. For any complex number z 6= 0 we define the multi-valued function log(z) by
log(z) = ln |z| + argz, where arg(z) is any argument for z.
Example: ln(1) = (2n + 1) for any integer n. The usual laws of logarithms hold for the
complex valued logarithm so long as we remember the fact that log z is multi-valued. For
example log(z1 z2 ) = log z1 = log z2 means that some value of log(z1 z2 ) equals some value of
log(z1 ) + log(z2 ), and vice-versa.

Example: log(1) = log(1)(1) = log(1) + log(1). The values of log 1 are 2k, k =
0, 1, 2, . . . , whereas the values of log(1) + log(1) are (2m + 1) + (2n + 1).
Definition 11. The principal branch of the logarithm is defined on the cut domain

= C {x + y  x 0, y = 0} by the formula Log(z) = Log|z| + Arg(z).
Recall that the possible arguments for the principal argument satisfy < .
d
1
Logz = .
dz
z
Corollary 1. Log|z| and Arg(z) are harmonic in .
Theorem 7. Log(z) is analytic in , and

3.4. Complex Powers and Inverse Trigonometric Functions. We define the various
inverse trig functions in the standard way. For example
arcsin z = w z = sin w
arccos z = w z = cos w
arctan z = w z = tan w
Periodicity of the trig functions sinw, cos w, tan w means that the corresponding inverse trig
fuynctions are multi-valued.

Example: The multi-valued function f (z) = z a has a single valued branch on any cut
domain that prevents going once around the point z = a.

Example: The function f (z) = z 2 1 has a single valued branch of the cut domain

= {z = x + y  if 1 x 1 then y 6= 0}.
Exercise: Prove that the function in the last example is an odd function.
Theorem 8. The inverse trig functions are given by



arcsin z = log z + 1 z 2



arccos z = log z + z 2 1
arctan z =

1
1 + z
log
2
1 z

Definition 12. The power function is defined by z = e log z .

Thus the power function will in general have multiple values.
Exercise: Prove the following statements:
(1) If z = re then z = r e(+2k) , where k is any integer.
(2) z has a single value if is an integer.

10

(3) z has finitely many values if is a rational number.

(4) z has finitely infinitely many values in all other cases.
Example 4. The values of .
Example 5. The solutions of cos z = 2.
ez + ez
= 2 e2z 4ez + 1 = 0
2

4 16 4
z
e =
= (2 5)
2
z = log((2 5))

The values
of log((2 + 5)) are ln(2 + 5) + (/2 + 2k) and the values of log((2 5))
are ln( 5 2) + (3/2 + 2k) , where k is any integer. Therefore

cos z = 2 z = ln(2 + 5) /2 2k, z = ln( 5 2) 3/2 2k

cos z

4. Complex Integration
4.1. Contours.
(1) A smooth arc is given by a continuously differentiable function z(t), a t b, such
that: z 0 (t) 6= 0 for a t b and z(t) is one-to-one for a < t < b. A smooth closed
curve is the range of a smooth arc satisfying: z(a) = z(b) and the right hand derivative
of z(t) at t = a equals the left hand derivative of z(t) at t = b.
(2) Any smooth arc or smooth closed curve has 2 natural orientations.
(3) A contour is a sequence of oriented smooth arcs 1 , 2 , . . . , k such that the end point
of i is the beginning point of i+1 , i = 1, . . . , k 1.
(4) Parametrizations of contours.
(5) Jordan Curve Theorem: Any simple closed contour in C separates C into 2 domains,
the interior of the contour and the exterior.
4.2. Contour Integrals. Let f (z)Z be a continous function in some domain D and let C be
a contour in D. Then the integral

f (z)dz is defined to be a limit of Riemann sums in the

C

obvious way. This integal exists even if f (z) is only piecewise continuous. An immediate
consequence of this definition is the so-called M L inequality.
Theorem 9. Suppose f (z) is continuous
on
Z
 a contour C of lenth L and a constant M such


that |f (z)| M z C. Then  f (z)dz  M L.
C

Theorem 10. (The Fundamental Theorem of Calculus)

11

Suppose C is a smooth curve parametrized by z(t), a t b, and suppose F (z) is an antiderivative of f (z), that is F 0 (z) = f (z). Then
Z
Z t=b
f (z)dz =
f (z(t))z 0 (t)dt = F (z(b)) F (z(a))
C

t=a

Z
In particular

f (z)dz is independent of the parametrization.

C

4.3. Independence of Path. The fundamental theorem of calculus can be expressed in the
form
Theorem 11. Suppose f (z) is continuous in a domain D and F (z) is an antiderivative. Then
Z
f (z)dz = F (Q) F (P ), where C is any contour in D with beginning point P and ending
C

point Q.
This is saying that integration is independent of path.
Corollary
2. Suppose f (z) is continuous in a domain D and has an antiderivative. Then
Z
f (z)dz = 0 for all closed contours C in D.
C

The following corollary is easy to prove.

CorollaryZ 3. Suppose f (z) is continuous on a domain D. Then integration is independent of
path
f (z)dz = 0 for all closed contours C in D.
C

Theorem 12. Suppose f (z) is continuous on a domain D. Then the following statements are
equivalent:
(1) Z
a function F (z) defined on D and satisfying F 0 (z) = f (z) z D.
f (z)dz = 0 closed contours C in D.
(2)
C
Z
Z
f (z)dz =
f (z)dz for all piecewise
(3) Integration is independent of path, that is
C1

C2

smooth curves C1 , C2 in D having the same beginning and ending points.

Proof. The above theorem and corollaries establish the implications (a) = (b) =Z(c), so we
need only prove that (c) = (a). Define a function F (z), for z D, by F (z) =

f ()d,
C

where C is any smooth path from a fixed reference point z0 D to z. Then F 0 (z) = f (z). 

12

4.4. Cauchys Integral Theorem. Deforming one contour C0 in a domain D into another
C1 . Lots of examples of deformations. Simply connected domains.
Theorem 13. (The
Cauchy Integral Theorem) If f (z) is analytic in a simply connected
Z
domain D then
f (z)dz = 0 for all closed contours C in D.
C

Proof. Let z = x + y and f (z) = u(x, y) + v(x, y). Then dz = dx + dy and

Z
Z
Z
Z
f (z)dz = (u + v)(dx + dy) = (udx vdy) + (vdx + udy)
C

u
v u
v
=
,
= . Let
x
y y
x
be the interior of C. Then applying Greens Theorem from Vector Calculus (Math 317) we
have

Since f (z) is analytic u, v satisfy the Cauchy-Riemann equations:

(u + v)(dx + dy) = (udx vdy) + (vdx + udy)

C
C


Z Z 
Z Z 
u v
v u

dA +

dA = 0
=

x y
x y

f (z)dz =
C


Corollary 4. In a simply connected domain an analytic function has an antiderivative.
Theorem 14. Suppose f (z) is an analytic function on the domain
Z D. If C0 and
Z C1 are closed
contours in D that can be deformed into each other in D then
f (z)dz =
f (z)dz.
C0

C1

4.5. Cauchys Integral Formula and Consequences. An important corollary of Cauchys

Integral Theorem is the Cauchy Integral Formula:
Theorem 15. Suppose f (z) is analytic in some simply connected domain D, C is a positively
oriented simple closed contour in D, and z0 is a point interior to C. Then
Z
f (z)
1
dz.
f (z0 ) =
2 C z z0
This theorem is rather amazing since it is saying that the values of an analytic function inside
a simple closed contour are determined from the values of f (z) for z C in a very simple way.
Z

Z
f (z)
f (z)
dz =
dz, where Cr is the positively oriented circle
Proof. First note that
C z z0
Cr z z0
|z z0 | = r and r is chosen sufficiently small so that the disc {|z z0 | r} is interior to C.

13

This follows from the fact that in a simply connected domain we can always deform one simple
closed curve into any other one. Then
Z
Z
Z
Z
f (z)
f (z0 )
f (z) f (z0 )
f (z)
dz =
dz =
dz +
z z0
Cr z z0
Cr z z0
Cr
C z z0
Z
f (z) f (z0 )
= 2f (z0 ) +
dz
z z0
Cr
Z
Z
f (z)
f (z) f (z0 )
dz must be independent of r since it equals
dz 2f (z0 ),
Thus
z z0
C z z0
Cr
which is obviously independent of r. But then
Z
Z
f (z) f (z0 )
f (z) f (z0 )
dz = lim+
dz = 0 (by the M L inequality)
r0
z z0
z z0
Cr
Cr
f (z) f (z0 )
for r small. Moreover
z z0
the length of the contour Cr is 2r, which tends to 0 as r 0+ .


Note: since f (z) is analytic in D We can find a bound for

Theorem 16. Let Zg() be a function which is continuous on some contour

Z C. Then the function
g()
g()
defined by G(z) =
d.
d is analytic z 6 C. Moreover G0 (z) =
2
C z
C ( z)
d
This theorem is saying that
dz
inside the integral.

Z
C

g()
d =
z

Z
C

d
dz

g()
z


d, that is we can differentiate

Corollary 5. If f (z) is analytic in some domain D then f (z) has derivatives of all orders.
This follows from the last theorem and the Cauchy Integral Formula. In fact the derivatives
are given by:
Theorem 17. Suppose f (z) is analytic inside and on the positively oriented simple closed
contour C and z is any point interior to C. In fact the nth derivative is given by
Z
f ()
n!
(n)
d
f (z) =
2 C ( z)n+1
4.6. Bounds for Analytic Functions. A simple corollary of the last theorem is
Corollary 6. Suppose f (z) is analytic inside and on the cirlce CR = {|z z0 | < R} and
|f (z)| M z CR . Then
n!M
|f (n) (z0 )| n , n = 1, 2, . . .
R

14

Proof. Parametrize CR by z = z0 + Re where goes from = 0 to = 2. Then



Z =2

Z

 n!



n!
f
(z
+
Re
)
f
(z)
0
(n)

=


|f (z0 )| = 
dz
Re
d

2 CR (z z0 )n+1  2  =0 R(n+1) e(n+1)
Z =2

n! 
f (z0 + Re ) 
n!
n!M
=

d
M

2
=
(M L inequality)
 2Rn
2Rn  =0
en
Rn

Corollary 7. (Liouvilles Theorem) A bounded entire function is a constant.
Corollary 8. (Fundamental Theorem of Algebra) Any polynomial of degree 1 has a
root.
Theorem 18. (Maximum Modulus Principle) Suppose f (z) is analytic inside and on the
simple closed contour C. Then the maximum value of |f (z)| occurs on C.
Theorem 19. Suppose f (z) is a non-constant function analytic inside and on a simple closed
contour C. Then maxC |f ()| > |f (z)| z interior to C.
5. Series Representations for Analytic Functions
5.1. Sequences and Series.
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)

Sequences and their limits.

Convergence of series
Absolute convergence of series.
The geometric series.
The comparison Test.
The Ratio Test.
Pointwise convergence of a sequence of functions {fn (z)}n=
n=1 on a domain D.
n=
Uniform convergence of a sequence of functions {fn (z)}n=1 on a domain D.

Definition 13.
Suppose fn (z) is a sequence of functions on a domain . We say that the sequence converges
pointwise to a function f (z) if given any  > 0 and any z , a positive integer N such that
n N = |fn (z) f (z)| < 
Definition 14. Suppose fn (z) is a sequence of functions on a domain . We say that the
sequence converges uniformly to a function f (z) if given any  > 0, a positive integer N such
that
n N = |fn (z) f (z)| <  for all z
This is definitely stronger than pointwise convergence. In the case of pointwise convergence the
N we choose depends on both  and the point z , but in the case of uniform convergence
N depends only on . The choice of N is uniform for all z .

15

5.2. Taylor Series. Let f (z) be a function which is infinitely differentiable at some point z0 .
N
X
f (j) (z0 )
Then the Taylor nth Taylor polynomial of f (z) at z0 is Pn (z) =
(z z0 )j . Pn (z) is
j!
j=0
the unique polynomial of degree n such that f (j) (z0 ) = P (j) (z0 ) for j n.
The Taylor series of f (z) at z = z0 is

X
f (n) (z0 )

n!

n=0

(z z0 ) = lim Pn (z)
n

Theorem 20. If f (z) is analytic in the disc D = {|z z0 | < R} then the Taylor series
converges to f (z) z D. Moreover this convergence is uniform on the domain {|z z0 | R0 }
for any R0 < R.
5.3. Power Series.
Definition 15. A power series is a series of the form

aj (z z0 ) := lim

j=0

We set f (z) =

N
X

aj (z z0 )j .

j=0

aj (z z0 )j if this series converges for z.

j=0

Questions:
(1) For what values of z does the series converge?
(2) Is f (z) analytic in its domain of convergence?
(3) What is the relationship between Taylor series and power series?
Lemma 3. If a power series

aj (z z0 )j converges for some z1 with |z1 z0 | = r then it

j=0

converges for all z with |z z0 | < r.

An easy corollary of this lemma is:
Corollary 9. Given any power series

aj (z z0 )j a number R such that

j=0

(1)

aj (z z0 )j converges for all z with |z z0 | < R.

j=0

(2)

X
j=0

aj (z z0 )j diverges for all z with |z z0 | > R.

16

No conclusion is possible if |zz0 | = R. Moreover the convergence is uniform on any subdomain

{z  |z z0 | R0 }, where R0 is any number satisfying R0 < R.
R is called the radius of convergence of the power series. It can be computed using the ratio
test.
an+1
= L. Then the radius of convergence of the power series
Theorem 21. Suppose lim
an

X
z j is R = 1/L.
j=0

Example 6. The geometric series

z converges for |z| < 1, and in fact

j=0

zj =

j=0

1
1z

1
makes sense for all z 6= 1.
|z| < 1. It diverges for |z| 1. Notice that
1z

X
zj
Example 7. The radius of convergence of the power series
is R = 1. In this case the
2
j
j=0

X
1
power series converges for all |z| 1. To see this use make a comparison with the series
.
2
j
j=1

32n
z 4n . To compute the radius of
2
(n
+
1)
n=0
convergence we consider ratios of consecutive non-zero terms. That is we compute
 2n+2 4n+4

3
z
/(n + 2)2 
1

= 9|z 4 | < 1 |z| < .
lim  2n 4n

2
n
3 z /(n + 1)
3

Thus the radius of convergence is r = 1/ 3.

Example 8. Consider the power series f (z) =

Theorem 22. Suppose fn (z) is a sequence of continuous functions on some domain , converging uniformly to a function f (z) on . Then f (z) is continuous on .
Proof. Let z0 be any point in . To prove continuity of f (z) at z = z0 we must show that given
any  > 0, there exists > 0 such that |f (z) f (z0 )| <  if |z z0 | < .
Since the sequence fn (z) converges uniformly to f (z) on we can find a positive integer N
such that |fn (z) f (z)| < /3 for all n N and all z in . Now the function fN (z) is
continuous at z = z0 and therefore > 0 such that |fN (z) fN (z0 )| < /3 if |z z0 | < .
Then
|f (z) f (z0 )| = |f (z) fN (z) + fN (z) fN (z0 ) + fN (z0 ) f (z0 )|
|f (z) fN (z)| + |fN (z) fN (z0 )| + |fN (z0 ) f (z0 )|



<
+ + =
3 3 3

17


Corollary 10. Suppose fn (z) is a sequence of continuous functions
converging uniformly
to a
Z
Z
function f (z) on a domain . If C is a contour in then
fn (z)dz converges to
f (z)dz.
C

Proof. Let  be any positive number. Then N such that


|fn (z) f (z)| < n N and z . where L is the length of C.
L

Z




Now we apply the M L inequality:  (fn (z) f (z))dz  L = .
L

Z
fn (z)dz =

Remark: This is saying that lim

lim fn (z)dz.

C n

Corollary 11. Suppose fn (z) is a sequence of analytic functions converging uniformly to a

function f (z) on a domain . Then f (z) is analytic.
Z
Proof. By the Cauchy integral theorem
fn (z)dz = 0 for any closed contour C in . Then
Z
ZC
the last corollary gives
f (z)dz = lim
fn (z)dz = 0. By Moreras theorem (see page 210)
C

it follows that f (z) is analytic.

From this it follows that a power series

aj (z z0 )j with radius of convergence R converges

j=0

to a function f (z) which is analytic for |z z0 | < R. Moreover we must have an =

n = 0, 1, 2, . . .

f (n) (z0 )
for
n!

5.4. Mathematical Theory of Convergence. Omit this section.

5.5. Laurent Series.

Theorem 23. Suppose f (z) is analytic in the annulus = {z  r < |z z0 | < R}. Then
Z

X
1
f ()
j
f (z) =
aj (z z0 ) , where aj =
d, < j < ,
2 C ( z0 )j+1
j=
and C is any positively oriented simple closed contour going once around the annulus. Moreover
the series converges uniformly in any proper annulus.

18

Such a series is called a Laurent series. It splits into a singular part (the part with negative
powers of z z0 ) and an analytic part (the part with non-negative powers of z z0 ):
f (z) =

1
X
j=

aj (z z0 ) +

aj (z z0 )j .

j=0

Definition 16. A function f (z) has an isolated singularity at z = z0 if some R > 0 such
that f (z) is analytic in the annulus {z  0 < |z z0 | < R}. The function may or may not be
analytic at z = z0 .
By the above theorem a function f (z) with an isolated singularity at z0 has a Laurent series
representation at z0 since we simply choose r = 0.
5.6. Zeros and Singularities.
Definition 17. Let f (z) be analytic in a neighbourhood of z0 . Then f (z) has a zero of order
m at z0 if its Taylor series at z0 has the form
f (z) = am (z z0 )m + am+1 (z z0 )m+1 + am+2 (z z0 )m+2 + , where am 6= 0.
This is equivalent to saying that f (n) (z0 ) = 0 for n = 0, 1, . . . , m 1, but f (m) (z0 ) 6= 0. It is
also equivalent to saying that in a neighbourhood of z0 we have f (z) = (z z0 )m g(z), where
g(z) is analytic in a neighbourhood of z0 and g(z0 ) 6= 0.
Definition 18. Suppose f (z) has an isolated singularity at z0 and f (z) =

aj (z z0 )j is

j=

its Laurent series. Then

(1) f (z) has a removable singularity at z0 if aj = 0 for all j < 0. In this case there are no
negative powers in the Laurent series and f (z) is analytic at z0 .
(2) f (z) has a pole at z0 if a negative integer N such that aj = 0 for all j < N and
aN 6= 0. In this case there are some negative powers in the Laurent series, but only
finitely many, say
f (z) = an (z z0 )n + an+1 (z z0 )n+1 + + a1 (z z0 )1 + a0 +
where an 6= 0. We say that the order of the pole at z0 is n.
(3) f (z) has an essential singularity at z0 if aj 6= 0 for infinitely many negative j. Thus
there are infinitely many negative powers in the Laurent series.
Example 9.
P (z)
, where P (z), Q(z) are polynomials with no common factors, has
Q(z)
poles at the zeros of the denominator Q(z). If z0 is a zero of order n of Q(z) then z0 is a pole
of order n of f (z). On the other hand the function f (z) = e1/z has an essential singularity at
z = 0.

The function f (z) =

19

Theorem 24. If f (z) has a pole at z0 then lim |f (z)| = +, or equivalently lim f (z) =
zz0

zz0

in the extended complex plane.

Contrast this with the following theorem, which we will not prove.
Theorem 25. (Picard) Suppose f (z) has an essential singularity at z = z0 . Let  be any
positive number. Then for all w in the complex plane C, except possibly for one value of w,
the equation f (z) = w has a solution satisfying 0 < |z z0 | < .

In other words, in every neighbourhood of z0 , no matter how small, the set {f (z)  z }
is either the entire complex plane C or the complex plane C minus a single point! This is an
amazing result.
Theorem 26. If f (z) has a pole of order n at z0 then an analytic function g(z) such that
g(z)
and g(z0 ) 6= 0
f (z) =
(z z0 )n
Poles arent so bad.