Holomorphic Functions and Power Series

Having studied the basic properties of the complex field, we now study functions of a complex
variable. We define what it means for a complex-valued function defined on a domain in the com-
plex plane to be complex-differentiable. We then study the relationship between this new notion of
differentiability and the already familiar notion of real-differentiability. Functions that are complex-
differentiable at every point in a domain are called holomorphic functions. For various reasons, which
we shall study in great detail, a function being holomorphic is far more restrictive than the function
being real-differentiable.
The focal point of this chapter will be power series and functions that have local power series
expansions, i.e., analytic functions. The reason for this is two-fold:
1. The most important functions that arise in practice are analytic.
2. All holomorphic functions are analytic!
The second point above will be one of the central theorems we will prove in this course. We end
this chapter with a brief exposition of the exponential and trigonometric functions.
Notation: Throughout this chapter, U will be a domain in C with a ∈ U and f : U → C. will be a
function.

1 Holomorphic functions
1.1 Definitions and examples
Definition 1. Let U ⊂ C be open and f : U → C be a map and a ∈ U . We say f is complex-
differentiable (C-differentiable at a) if
f (z) − f (a)
lim
0,z→a z −a
exists and in which case we denote the limit by f 0 (a) and call it the complex derivative of f at a.
We say that f is holomorphic on U if f is C-differentiable at every point of U . A holomorphic
function on C is called an entire function.
Remarks.
1. It is easy to see that if f is C-differentiable at a then f is continuous at a.
2. If f and д are C- differentiable then so are f ± д, f д and c f , c ∈ C. The proofs of these facts
follows mutatis mutandis from the proof in the real case.
3. If f is C-differentiable at a and f 0 (a) , 0, then f1 is C- differentiable at a and its derivative at a
is f 0−1
(a) 2
.

Examples and non-examples

We will provide several examples and non-examples to facilitate the digestion of the definitions of
C-differentiability and holomorphicity.

1
 1 Holomorphic functions

1. The function z n is entire with derivative nz n−1 . Combined with fact that sums and products of
C-differentiable functions are C-differentiable we see that any polynomial in z is entire.
2. Quotients of C-differentiable functions are C-differentiable wherever they are defined.
3. The function f (z) = z is not C-differentiable at any point. To see this, observe that the definition
gives
f (z 0 + h − f (z 0 )) h
=
h h
which has different limits as h → 0 along the real and imaginary axes.
4. Consider the function f (z) = |z| 2 = z · z. In real coordinates, we can consider at as the function
(x, y) 7→ x 2 + y 2 . This is polynomial function in x, y and hence C ∞ -smooth. Again computing
as before, we see that the difference quotient if of the form a + ah/h + h. As h 7→ 0, the above
difference quotient has a limit iff a = 0. Hence |z| 2 is C-differentiable only at the point 0.

The last two examples show that f being C-differentiable at a point a is a severe restriction. The
limit of the difference quotient must exist and agree on all possible approaches to the point a. We
could approach a along any line, or even complicated curves like spirals and the limiting value must
exist and agree on any such approach. This is the crucial difference between real-differentiability of a
function defined on an interval and C-differentiability. We will now explore this connection in some
detail.

1.2 Relationship between the complex derivative and the real derivative
Let f be C-differentiable at a. Let R(h) (remainder) be the function defined as follows:

 f (a+h)−f

h
(a)
− f 0 (a) if h , 0
R(h) =   0 if h = 0 .

Observe that R is well-defined for h sufficiently close to 0 and that R is continuous. The definition of
C-differentiability can now be restated as follows: f is C-differentiable at a iff we can find a complex
number α and a continuous function R : D → C, where D ⊂ C is a suitably small neighborhood of 0
such that R(0) = 0 and
f (a + h) = f (a) + αh + R(h)h.
Recall that f is R-differentiable at a if we can find a linear transformation d fa : R2 → R2 such that:

k f (a + h) − f (a) − d fa hk
lim → 0.
h→0 khk
If f is C-differentiable at a then clearly f is R-differentiable at a with the role of d fa played the
C-linear (and hence R-linear map) h 7→ f 0 (a)h. Furthermore, if f is R-differentiable at a and the
derivative d fa is C-linear then from our discussion in the previous chapter, it follows that d fa is of
the form h 7→ αh for some complex number α ∈ C. Hence, f is C-differentiable at a.

1.3 The Cauchy–Riemann equations

Let U , f , a be as before and write z as x + iy, x, y ∈ R and f = u + iv where u and v are real-
valued. Assume that a, the partial derivatives u x , uy , v x , vy exist. We define the Wirtinger derivatives

2
 2 Power series

as follows:
∂f 1 ∂f ∂f
!
(a) := −i (a)
∂z 2 ∂x ∂y
∂f 1 ∂f ∂f
!
(a) := +i (a).
∂z 2 ∂x ∂y

Remark 2. Contrast the definition above with the second characterization of an R-linear map given
in the last chapter.

Theorem 3. With notations as above, the following conditions are all equivalent.

1. f is C-differentiable at a.
∂f ∂f
2. f is real-differentiable at a and ∂y (a) = i ∂x (a).
∂f
3. f is real-differentiable at a and ∂z (a) = 0.
4. f is real-differentiable at a and u x = vy and uy = −v x .

Proof. Let us see how the map d fa acts on the two elements 1, i ∈ C. Now 1, i are nothing but the
∂f ∂f
vectors (1, 0) and (0, 1), respectively, in R2 . This d fa (1) = ∂x (a) and d fa (i) = ∂y (a). Now, d fa is
C-linear iff d fa (i) = id fa (1) and this proves the equivalence. 

Remark 4. The above proof is very slick but comes with the risk of deluding you into a false sense of
understanding. Please mull over the proof keeping in mind the discussion on linear mappings in the
last chapter.

2 Power series
Our development of the theory is so far severely deficient in examples. We remedy this situation with
a dose of power series. We first recall some deeper facts about convergent infinite series.

2.1 Infinite series

We will assume that the reader is sufficiently adept at identifying whether a (complex) series is (abso-
lutely) convergent or not using the standard convergence tests such as the comparison test, root test,
ratio test, etc.
We will now sketch the proof of one-part of a famous theorem of Riemann on the behavior of
absolutely convergent series.

Theorem 5. Let zn be an absolutely convergent series. Then given any bijection σ : N → N, the new
P
series z σ (n) is convergent and has the same sum as zn .
P P

Proof. Let sn denote the n-th partial sum of the series zn and sn0 that of the series z σ (n) . Let ε > 0.
P P
First choose N 0 so large so that n=N0 |zn | < ε and then choose N 1 > N 0 so that σ ({1, 2, . . . , N 1 }) ⊃
P∈
{1, 2, . . . , N 0 }. Then it is clear that |sn − sn0 | < ε proving that the series z σ (n) converges to the same
P
sum as zn .
P


Definition 6 (Cauchy product). Let an and bn be two series. We define the Cauchy product of
P P
the two series to be the new series c k , where c k = n+m=k am bn .
P P

3
 2 Power series

Theorem 7. If the series an and bn are both absolutely convergent then so is their Cauchy product.
P P
Moreover the Cauchy product converges to the product of the sums of the two series.

Proof. Let |an | = α and |bn | = β. Then observe that

P P

N
X
|c k | ≤ (|a 0 | + |a 1 | + · · · + |a N |)(|b0 | + |b1 | + . . . |b N |) ≤ α β .
k =0

This shows that the Cauchy product is absolutely convergent. Now,

X N
X ∞
X ∞
X
ck − (a 0 + a 1 + · · · + an )(b0 + b1 + · · · + bn ) ≤ α |bk | + β |ak |.
k n=0 k=N +1 K =N +1

The last term above can be made as small as desired. 

2.2 Power series and radius of convergence

Definition 8. A formal power series about z 0 ∈ C with coefficients c n ∈ C is a series of the form

c n (z − z 0 ) n .
X

n ∈N

Remark 9. We may substitute different complex numbers in the place of z to convert the above formal
series to an honest one. However, apart from z = z 0 , it is not at all obvious that such a substitution
leads to a convergent series.
Given a formal power series n ∈N c n (z − z 0 ) n , let us consider the set of convergence
P

c n (z − z 0 ) n converges}.
X
C := {z ∈ C :
n ∈N

Our objective now is to study the geometry of the set C . Define

R := sup{R ∈ R+ : ∃M R ∀n ∈ N |an |R n < M R }.

Theorem 10. Given a formal power series n ∈N c n (z − z 0 ) n and R as defined as above, the series con-
P
verges absolutely for z ∈ D(z 0 , R) and diverges whenever |z| > R. Moreover, if f : D (z 0 , R) is defined by
the sum of the series then sequence fk := kn=0 c n (z − z 0 ) n converges uniformly on compact subsets of
P
D (z 0 , R) to f .

Proof. If |z| > R then the series cannot converge as the n-th term does not go to 0. On the other hand,
if |z| < r < R we have
!n X |z − z 0 | ! n
X
n
X
n |z − z 0 |
|c n ||z − z 0 | = r |c n | ≤ Mr .
r r

By comparison test, the above series converges. By Weierstrass M-test, the assertion regarding uni-
form convergence on compacts follows.


Examples.

4
 2 Power series

1. The geometric series 1 + z + z 2 + . . . converges to the function 1−z

1
whenever |z| < 1. The
behaviour on the unit circle is very simple in this case. We have divergence at all points since
the n-th term does not go to 0.

As one expects, power series can be differentiated term-by-term.

Theorem 11. Let c n (z − z 0 ) n be a convergent power series with radius of convergence R > 0 and let
P
f : D(z 0 , R) → C be the sum. Then f is holomorphic on D (a, R) and its derivative is given by

nc n (z − z 0 ) n−1 .
X

Proof. First of all, the derived power series has radius of convergence R. To see this, observe that if
n|c n |r n−1 < Mr then |c n |r n < r Mr . showing that R 0 ≤ R (R 0 is the radius of convergence of the derived
power series). On the other hand, if 0 < r < ρ < R then
!n !n
r r
n|c n |r n−1 = n/r |c n |ρ n ≤ n/r M ρ ( )
ρ ρ
 n
But, n ρr is a bounded sequence and therefore we can find an upper bound for n|c n |r n−1 showing
that R ≤ R 0.
Now, we must show that for each ε > 0 and z ∈ D(z 0 , R), we have

f (z) − f (w ) X
− c n (z − z 0 ) n−1 < ε
z −w n
when w is in a suitably small neighborhood of z. To see this, we employ a standard technique in
analysis: break up the series into the first few terms and the tail which we know can be made very
small. To this end, write f (z) = S N (z) + R(z) where S N the partial sum of the power series of f . We
will choose N suitably to force the above required inequality. The difference quotient becomes
N
* S N (z) − S N (w ) − R N (z) − R N (w )
!
nc n (z − z 0 ) n−1 + − nc n (z − z 0 ) n−1 +
X X
.
, z − w n=0 - n >N z −w

Call the three terms I, II and III, respectively. The second term can be made as small as desired (being
the tail of a convergent series) by choosing N to be suitably large. On the other hand, the first term
can be made as small as desired by choosing w in a suitably small neighborhood of z. The only term
to worry about is III. We have

c n (z − z 0 ) n−1 + (z − z 0 ) n−2 (w − z 0 ) + · · · + (w − w 0 ) n−1

X  

n >N

where we expanded out (z − z 0 ) n−1 − (w − z 0 ) n−1 and canceled the z − w in the denominator. We may
choose z and w so that |z − z 0 | < r < R and |w − z 0 | < R whence the above sum becomes

nc n ρ n−1
X
≤
n >N

which is the tail of a convergent series and can be made as small as desired by choosing N suitably
large (here it is irrelevant what z and w are as long as both are in D(z 0 , R)). In conclusion, we first
choose N so large that the second and third term become lesser than ε/3 and then we choose a small
enough neighborhood of z so that I also becomes smaller than ε/3. 

5
 2 Power series

Corollary 12. If f is the limit of a convergent power series c n (z − z 0 ) r in D(z 0 , R) then f is infinitely
P
differentiable and we have
f n (z 0 )
cn = .
n!
Proof. Straightforward. 

Many of the nice properties of holomorphic functions stem from the following deep theorem which
is the central theorem of the course. We first need a definition.

Definition 13. Let f : U → C be a function. We say that f is complex-analytic if for each z 0 ∈ U ,

we can find a power series
c n (z − z 0 ) n
X

such that on some disk D(z 0 , r ) ⊂ U , r > 0, the series converges to f |D (z0,r ) .

Theorem 14. Any holomorphic function is automatically complex-analytic!

We will spend the better part of two chapters developing the theory of contour integration to prove
this result. After that, the rest of the course will be spent on obtaining an endless stream of wonderful
results! To whet you appetite, here is one particularly famous one:

Theorem 15 (Open mapping theorem). Let f : U → C be holomorphic. Then either f is constant or f

is an open map.