MA 205 Complex Analysis: Lecture 3

U. K. Anandavardhanan
IIT Bombay

August 05, 2019

[Power Series]

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Announcements

Your first tutorial is tomorrow at 5:30 PM. There are six batches

LT 001, LT 002, LT 003, LT 004, LT 005, LT 006.

There’s a file named “Tutorial Batches” on Moodle where you can

find out your batch. Ctrl F for your name or roll number. The
tutorial sheet for this week is also there on Moodle. Have a look
before you go to the tutorial class.

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Introduction

In the last lecture, you saw holomorphic functions in some detail.

If f = u + ıv is holomorphic in Ω, then (i) both u and v satisfy
Cauchy-Riemann equations, and (ii) f (x, y ) = (u(x, y ), v (x, y )) is
real differentiable. You also saw that though neither (i) nor (ii) is
sufficient to guarantee holomorphicity, both (i) and (ii) together do
guarantee holomorphicity of f . You also studied harmonic functions
which are closely related to holomorphic functions. The key name
here was the Laplacian. The notion of a harmonic conjugate of a
given harmonic function was defined, and it was stated that the
existence of a harmonic conjugate for every harmonic function is
guaranteed if and only if the domain is “simply connected”.
Remember that we asserted this but did not prove it.

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Cauchy, Riemann, Laplace [Source: Wikipedia]

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Cauchy, Riemann, Laplace [Source: Wikipedia]

“More concepts and theorems have been named for Cauchy than for any other mathematician (in elasticity alone

there are sixteen concepts and theorems named for Cauchy).”

“When Riemann was 33, he developed the famous Riemann hypothesis. This was an article that was only 8 pages

long. Mathematicians ever since have struggled to try to prove what Riemann wrote.”

“Geometrician of the first rank, Laplace was not long in showing himself a worse than average administrator; from

his first actions in office we recognized our mistake. Laplace did not consider any question from the right angle: he

sought subtleties everywhere, conceived only problems, and finally carried the spirit of ”infinitesimals” into the

administration.” (Napoleon, 1799)

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Polynomials
Today, we’ll discuss the so called analytic functions. To warm up,
let’s first look at the simplest of all functions. What’s the most
trivial function? f (z) = a0 , a constant. Then, may be various
powers of z; easily differentiable, integrable etc. Thus the simplest
class of functions is polynomials: f (z) = a0 + a1 z + . . . + an z n ,
ai ∈ C. These are clearly holomorphic everywhere in C. The same
polynomial f (z) can be expanded along any point z0 . What do I
mean by this? f (z) can be written as

b0 + b1 (z − z0 ) + . . . + bn (z − z0 )n .

The complex numbers bi can be easily calculated. Of course, you

could expand powers of (z − z0 ) using binomial theorem and
compare coefficients to get bi in terms of ai . Perhaps a smarter
(i)
way would be to notice that bi = f i!(z0 ) , which then can be
Xn
calculated from f (z) = ai z i .
i=0
U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3
Power Series

A polynomial, by definition, is a finite polynomial; i.e., it comes

with a finite degree. As the next simplest class of functions, why
can’t we consider infinite polynomials? It’ll look like:

f (z) = a0 + a1 z + a2 z 2 + . . . ,

or more generally,
∞
X
ai (z − z0 )i .
i=0

Of course no one calls this a polynomial; it’s called a power series.

You’ll agree that now we need to be a bit careful. There can be
convergence questions. For example, f (z) = 1 + z + z 2 + . . . makes
sense for all z such that |z| < 1, but not when |z| > 1. (Why?)
We say that this power series has radius of convergence = 1.

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Power Series

It’s a beautiful fact that the radius of convergence exists for any
power series; i.e., there exists R such that
∞
X
ai (z − z0 )i
i=0

converges when |z − z0 | < R, and diverges when |z − z0 | > R. In

other words, the radius of convergence is the largest R such that
the given power series converges inside a disc of radius R. We’ll
soon give a formula for R in terms of the coefficients of the given
power series.

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Power Series
Before we prove the existence of radius of convergence, let’s recall
a few definitions and observations.
∞
X
We write a = ai , ai ∈ C, if the sequence of partial sums {sn },
i=1
where sn = a1 + . . . + an , converges, and lim si = a. The series
i→∞
∞
X ∞
X
ai is said be absolutely convergent if |ai | is convergent.
i=1 i=1
Exercise:
1. Absolute convergence =⇒ convergence.
∞
X
2. Comparison Test: If bi is absolutely convergent, and if
i=1
∞
X
|ai | ≤ |bi | for all large enough i, then ai is absolutely
i=1
convergent.
U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3
Power Series

Recall upper limit: for a sequence of real numbers x1 , x2 , . . ., let yn

be the supremum of the set {xn , xn+1 , . . .}. Then the sequence
y1 , y2 , . . . either diverges to ±∞ or has a finite limit. This is called
the upper limit of the sequence {xi }. It can be ∞. If limit exists,
then the upper limit coincides with the usual limit.

Examples:
1. the sequence 1, 2, 3, . . . has upper limit ∞.
2. the sequence 1, 12 , 13 , . . . has upper limit 0.
3. the sequence 1, −1, 1, −1, . . . has upper limit 1.

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Power Series

Theorem (Cauchy’s Root Test)

∞
X p
For a series ai , let C = lim sup i |ai |. Then the series converges
i=1 i→∞
absolutely if C < 1 and it diverges if C > 1.
p
Proof: If C < 1, then we can choose a k such that i |ai | < k < 1
after a stage (by the definition of the upper limit). Thus, after a
X∞
i
stage, |ai | < k < 1. Now k i converges absolutely, and therefore
i=1
∞
X
ai is absolutely convergent (by comparison test). If C > 1, then
i=1 p
for infinitely many i, i |ai | > 1. Hence |ai | is bigger than 1 for
X∞
infinitely many i. Thus, lim ai 6= 0. So ai diverges. (Why?)
i→∞
i=1

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Power Series
Theorem (Ratio Test)
∞  
X  ai+1 
For a series ai , let L = lim sup 
 . Then, if L < 1, the series
i→∞ ai 
i=1
converges
  absolutely. The series diverges if there exists N such
that  aai+1  > 1 for i ≥ N.
 
i

Proof: Let L < 1. Let r be such that L < r < 1. Then after a
stage, say for i ≥ N, |ai+1 | < r |ai |. So |ai+k | < r k |ai |. Now
∞
X N
X ∞
X N
X ∞
X
|ai | = |ai | + |ai | = |ai | + |aN+i |
i=1 0 N+1 0 1
N ∞ N
X X
i
X r
< |ai | + |aN | r = |ai | + |aN | < ∞.
1−r
0 1 0
In the other case, |ai+1 | > |ai | for all large enough i, so
lim ai 6= 0. Therefore the series diverges.
n→∞ U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3
Power Series

Theorem (Existence of Radius of Convergence)

∞
X
For the power series ai (z − z0 )i , let R = 1p
. Then
i=1
lim sup i |ai |
i→∞
the power series converges absolutely if |z − z0 | < R and diverges
if |z − z0 | > R.

Proof: Apply the root test.

 
 ai+1 
If lim 
  exists, then by applying the ratio test instead of the
i→∞ ai   
 ai 
root test, it follows that R = lim  .
i→∞ ai+1 

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Power Series

Remark: If a series converges by the ratio test, then it converges

by the root test as well. But not conversely. Thus the root test is
better than the ratio test. But the ratio test is much easier to use
whenever it succeeds.

In fact:
 
p
i
 ai+1 
lim sup |ai | ≤ lim sup 
 
i→∞ i→∞ ai 

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Power Series

Examples:
∞ i  
X z  ai 
 = lim i = ∞; i.e., the
1. . Apply ratio test. lim 

i! i→∞ ai+1  i→∞
i=1
series converges everywhere.
3 5
2. z − z3 + z5 − . . . Radius of convergence is 1. Both the tests
apply here.

2
2
3. 12 + 13 z + 12 z 2 + 13 z 3 + . . .. Check that the ratio test
fails. Apply root test to show that the radius of convergence is √12 .

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Power Series

Power series can be added, subtracted, and multiplied in the

obvious way. It can also be differentiated and integrated term by
term, in its domain of convergence. Indeed,
f (z) = a0 + a1 (z − z0 ) + a2 (z − z0 )2 + . . . , then,
X  [(z − z0 + h)n − (z − z0 )n ]

f (z + h) − f (z)
lim = an lim
h→0 h n
h→0 h
X
= nan (z − z0 )n−1 .
n

Similarly, for integration. Apply root test to check that a given

power series, the differentiated series and the integrated series, all
have the same radius of convergence.

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Analytic Functions

A function f : Ω → C is said to be analytic if it is locally given by

a convergent power series; i.e., every z0 ∈ Ω has a neighbourhood
contained in Ω such that there exists a power series centered at z0
which converges to f (z) for all z in that neighbourhood. Clearly,
analytic functions are infinitely differentiable; you only have to
differentiate the power series term by term. Also, if
∞
(i)(z )
X
f (z) = ai (z − z0 )i , then ai = f i! 0 . Thus, an analytic function
i=1
is given by its Taylor series. We’ll later prove:

holomorphic =⇒ analytic.

This would prove our statement from Lecture 1 that once

differentiable is always differentiable!

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3

Analytic Functions

Just as in the complex case, power series and analytic functions

can be defined in the real case too. But unlike in the complex
case, differentiable does not mean real analytic. In fact, even
infinitely differentiable does not mean real analytic. For example,
f : R → R given by
( 1
e − x 2 if x > 0
f (x) =
0 if x ≤ 0,

is infinitely differentiable but not real analytic. In this example,

f (i) (0) = 0 for all i, and thus the Taylor series of f is the zero
function.

If you are interested in some math general knowledge you may

google for “Fabius function”.

U. K. Anandavardhanan IIT Bombay MA 205 Complex Analysis: Lecture 3