1.

Holomorphic functions
1.1. Complex-valued functions.
At the start of the study of calculus, we usually consider real-valued functions f
of a real variable x. Now we want to replace real -valued functions f by complex -
valued functions F , and we want to replace the real variable x by a complex variable
z = x + iy. Thus we consider functions F defined on the complex plane (or a
subregion of the complex plane) which takes on complex values. Examples of such
functions are:
F1 (x + iy) = (x + iy)2 , or equivalently F1 (z) = z 2
F2 (x + iy) = ex cos(y) + iex sin(y), or equivalently F2 (z) = ez
F3 (x + iy) = x − iy, or equivalently F3 (z) = z̄.
A complex-valued function has a real part and an imaginary part. Thus every
complex-valued function F can be written
F (x + iy) = u(x + iy) + iv(x + iy)
where u and v are real -valued functions of a complex variable. For example
F1 (x + iy) = (x + iy)2 = (x2 − y 2 ) + i (2xy),
so in this case
u(x + iy) = x2 − y 2
and
v(x + iy) = 2xy.
It follows that a complex-valued function F of a complex variable is really the same
as a pair of real-valued functions (u, v) of a complex variable. Also, since a complex
number z is determined by giving its real part x and its imaginary part y, we can
think of a real-valued functions u and v of a complex variable as the same as a pair
of real-valued functions of two real variables (x, y).

A complex-valued function F of a complex variable z is really the same as a

pair (u, v) of real-valued functions of a pair (x, y) of real variables. We can write


F (x + iy) = u(x + iy) + iv(x + iy), or equivalently, F (x, y) = u(x, y), v(x, y) .

1.2. Derivatives.
In the differential calculus we define the derivative of a real-valued function f
of a real variable x. We think of f as defined on an interval of the real axis with
values on the real axis, and we write:
f (x + h) − f (x)
f 0 (x) = lim .
h→0 h
We would like to do the same thing for complex -valued functions of a complex
variable. The key point is that if F is a complex-valued function of a complex
variable, and if h is a non-zero complex number, the expression
F (z + h) − F (z)
h
 2

is the quotient of two complex numbers, and hence is itself a complex number.

It is important to realize that this situation is very special to the case of complex-
valued functions of a complex variable. If we consider functions F which map R3
to R3 , and if x̄ and h̄ are vectors in R3 , then F (x̄ + h̄) − F (x̄) is a vector in R3 ,
and the quotient
F (x̄ + h̄) − F (x̄)
h̄
makes no sense because division of vectors is not defined!

Now just as with real-valued functions of a real variable, we say that the function
F is (complex) differentiable at the point z if the limit as h → 0 of this expression
exists, and we write
F (z + h) − F (z)
F 0 (z) = lim .
h→0 h

Definition 1.1. A complex-valued function F of a complex variable is holomorphic

in a region Ω ⊂ C if F is complex differentiable at each point of Ω.
Examples:

(i) Consider F1 (z) = z 2 . We have

F1 (z + h) − F1 (z) (z + h)2 − z 2 2zh + h2
= = = 2z + h
h h h
and so
F1 (z + h) − F1 (z)
lim = lim [2z + h] = 2z.
h→0 h h→0
It follows that F1 (z) = z 2 is a holomorphic function, and its derivative is
F10 (z) = 2z.

(ii) Consider F3 (z) = z̄. We have

F3 (z + h) − F3 (z) (z + h) − z z̄ + h̄ − z̄ h̄
= = = .
h h h h
Now if h is real, h̄ = h, and so
F3 (z + h) − F3 (z) h
lim = lim = 1.
h→0 h h→0 h
h real h real

On the other hand, if h is imaginary, h̄ = −h, and so

F3 (z + h) − F3 (z) −h
lim = lim = −1.
h→0 h h→0 h
h imaginary h imaginary

It follows that
F3 (z + h) − F3 (z)
lim does not exist,
h→0 h
and so F3 (z) = z̄ is not a holomorphic function.

Exercise: Show that F2 (z) = ez and F4 (z) = z −1 are holomorphic functions.

 3

1.3. The Cauchy-Riemann equations.

We now examine the meaning of complex differentiablity in terms of the real
and imaginary parts of a complex valued function. Thus suppose that F (z) =
u(z) + iv(z) where u and v are real valued functions, and suppose that F is complex
differentiable. Then if h is a complex number,
   
F (z + h) − F (z) u(z + h) − u(z) + i v(z + h) − v(z)
= .
h h
If we let z = x + iy = (x, y), and we take h = (h, 0) to be a (small) real number, it
follow that
u(x + h, y) − u(x, y) v(x + h, y) − v(x, y)
F 0 (z) = lim + i lim
h→0 h h→0 h
∂u ∂v
= (x, y) + i (x, y).
∂x ∂x
On the other hand, if we take h = ik = (0, k) to be a small imaginary number, it
follows that
u(x, y + k) − u(x, y) v(x, y + k) − v(x, y)
F 0 (z) = lim + i lim
k→0 ik h→0 ik
1 ∂u ∂v
= (x, y) + (x, y).
i ∂y ∂y
It follows that we must have the following equalities: ∂u ∂v
∂x (x, y) = ∂y (x, y) and
∂v ∂u
∂x (x, y) = − ∂y (x, y). These are called the Cauchy-Riemann equations. We have
established
Lemma 1.2. If F (x + iy) = u(x + iy) + iv(x + iy) is a holomorphic function of a
complex variable z = x + iy, then the real and imaginary parts u and v must satisfy
∂u ∂v
(x, y) = + (x, y),
∂x ∂y
∂v ∂u
(x, y) = − (x, y).
∂x ∂y

1.4. Line integrals.

We now consider line integrals of complex-valued functions. We need to introduce
some notation. We let
F (z) = u(x, y) + iv(x, y),
dz = dx + idy,
so that formally,



F (z) dz = u(x, y) + iv(x, y) dx + idy
   
= u(x, y) dx − v(x, y) dy + i u(x, y) dy + v(x, y) dx .
Z
Then if C is a curve in the complex plane, we define the line integral F (z) dz to
C
be:
Z Z Z
F (z) dz = u(x, y) dx − v(x, y) dy + i u(x, y) dy + v(x, y) dx.
C C C
 4

Theorem 1.3. Suppose Ω ⊂ C = R2 is a region to which we can apply Green’s

theorem. Suppose that F is holomorphic and continuously differentiable on Ω and
its boundary. Then
I
F (z) dz = 0.
∂Ω

Proof. From the definition, we have

I I I
F (z) dz = u(x, y) dx − v(x, y) dy + i u(x, y) dy + v(x, y) dx.
∂Ω ∂Ω ∂Ω

Using Green’s theorem, we show that

I
(a) u(x, y) dx − v(x, y) dy = 0,
∂Ω
I
(b) u(x, y) dy + v(x, y) dx = 0.
∂Ω

For (a), we have

I ZZ h
∂u ∂v i
u(x, y) dx − v(x, y) dy = − (x, y) − (x, y) dx dy
∂Ω ∂y ∂x
Z ZΩ
= 0 dx dy = 0
Ω
∂v
from the second Cauchy-Riemann equation ∂x (x, y) = − ∂u
∂y (x, y). For (b), we have
I ZZ h
∂u ∂v i
u(x, y) dy + v(x, y) dx = (x, y) − (x, y) dx dy
∂Ω ∂x ∂y
Z ZΩ
= 0 dx dy = 0
Ω
∂u ∂v
from the first Cauchy-Riemann equation ∂x (x, y) = + ∂y (x, y). This completes the
proof. 

Theorem 1.4 (Cauchy’s integral formula). Suppose Ω ⊂ C = R2 is a region to

which we can apply Green’s theorem. Suppose that F is holomorphic and continu-
ously differentiable on Ω and its boundary. If z ∈ Ω, then
I
1 F (w)
F (z) = dw.
2πi ∂Ω w − z
I
F (w)
Proof. We first consider the line integral dw where C is the circle cen-
C w −z
tered at z = x + iy of radius . We can parameterize C by
w(t) = z + eit = x +  cos(t), y +  sin(t) , 0 ≤ t ≤ 2π.

Then
w − z = eit ,
dw = ieit dt,
 5

and so
I Z 2π
F (w)
dw = i F (z + eit ) dt
C w−z 0
Z 2π Z 2π
=i F (z) dt + i [F (z + eit ) − F (z)] dt
0 0
Z 2π
= 2πiF (z) + i [F (z + eit ) − F (z)] dt.
0

Since F is continuous at z, we can make |F (z + eit ) − F (z)| as small as we like

(for all t) by making  sufficiently small. It follows that
I
F (w)
lim dw = 2πiF (z).
→0 C w − z


But now it follows from Theorem 1.3 that for any  > 0,
I I
F (w) F (w)
dw = dw.
∂Ω w − z C w − z

This completes the proof. 

1.5. Evaluation of definite integrals.

The Cauchy integral formula can be used to evaluate some definite integrals. We
consider several examples.
Z 2π
dθ
Example 1: Evaluate with 0 ≤ a < 1.
0 1 + a cos(θ)
1
We can write cos(θ) = eiθ + e−iθ , and so

2
1 1
=
1 + a2 eiθ + e−iθ
 
1 + a cos(θ)
2eiθ
=  .
2eiθ + a e2iθ + 1
If we put z = eiθ , then dz = ieiθ dθ, and so
Z 2π
2 2π ieiθ dθ
Z
dθ
=
0 1 + a cos(θ) i 0 ae2iθ + 2eiθ + a
I
2 dz
=
i az 2 + 2z + a
I
1 dz
= (4π) 2
2πi az + 2z + a
where the line integral is take around the boundary
√ of the unit circle. √ But the
polynomial az 2 + 2z + a has two roots: r1 = −1 − 1 − a2 and r2 = −1 + 1 − a2 .
Moreover, r1 < −1 < r2 < 0. Thus one root r2 is inside the circle, and the other is
outside the circle. We can write
az 2 + 2z + a = a(z − r1 )(z − r2 ).
 6

Thus
Z 2π I
dθ 1 dz
= (4π)
0 1 + a cos(θ) 2πi (z − r1 )(z − r2 )
1
= (4π)
r2 − r1
2π
=√ .
1 − a2
Z +∞
dx
Example 2: Evaluate 2 + 4)(x2 + 9)
−∞ (x
The function
1 1
f (z) = =
(z 2 2
+ 4)(z + 9) (z + 2i)(z − 2i)(z + 3i)(z − 3i)
is holomorphic everywhere except at the four points ±2i and ±3i. We consider
I
dz
(z 2 + 4)(z 2 + 9)
C(R1 ,R2 ,N )

where C(R1 , R2 , N ) is the square with vertices at {R1 , R1 + iN, −R2 + iN, −R2 },
where N > 3 Then the points 2i and 3i are inside this square, and the points −2i
and −3i are outside.
We can replace this line integral with the sum of two line integrals, one going
around a small circle C2 centered at 2i and the other around a small circle C3
centered at 3i. Using the Cauchy integral formula, we have
I
dz  −1 2πi π
2 + 4)(z 2 + 9)
= 2πi (2i + 2i)(2i + 3i)(2i − 3i) = 3
= ;
C (z −20i 10
I 2
dz  −1 2πi π
2 + 4)(z 2 + 9)
= 2πi (3i + 2i)(3i − 2i)(3i + 3i) = 3
=− .
C3 (z 30i 15
Thus I
dz π π π
= − = .
C(R1 ,R2 ,N ) (z 2 2
+ 4)(z + 9) 10 15 30
dz
H
But C(R1 ,R2 ,N ) (z2 +4)(z 2 +9) is the sum of four integrals I1 + I2 + I3 + I4 . The

first is parameterized by z = x with −R2 ≤ x ≤ R1 , and gives

Z R1
dx
I1 = 2 2
.
−R2 (x + 4)(x + 9)
The second is parameterized by z = R1 + it with 0 ≤ t ≤ N , and gives
Z N
i dt
I2 = .
0 (R1 + it) + 4)(R1 + it)2 + 9)
2

The third is paramterized by z = t + iN with R1 ≥ t ≥ −R2 , and gives

Z −R2
dt
I3 = 2 + 4)(t + iN )2 + 9)
.
R1 (t + iN )
The fourth is parameterized by z = −R2 + it with N ≥ t ≥ 0, and gives
Z 0
i dt
I4 = 2 + 4)(−R + it)2 + 9)
.
N (−R 1 + it) 2
 7

Now we need to make estimates of the absolute values of I2 , I3 , and I4 . We can

show
N
|I2 | ≤
(R12 − 4)(R12 − 9)
Z +∞
1 dt π
|I3 | ≤ 2 2
≤
N −∞ t + 9 2N 2
N
|I4 | ≤ .
(R22 − 4)(R22 − 9)
We first let R1 and R2 go to positive infinity, which makes |I2 | and |I4 | go to zero.
We then let N go to positive infinity, which makes |I3 | go to zero. In this process,
I1 goes to the desired integral, and so we get
Z +∞
dx π
2 2
= .
−∞ (x + 4)(x + 9) 30
Z +∞
sin(x)
Example 3: Show that dx = π.
−∞ x

1.6. Holomorphic functions and power series.

We want to prove:
Theorem 1.5.n Suppose F (z) ois holomorphic in an open set containing the closed
disk D(R) = z ∈ C  |z| ≤ R . Then for |z| < R we can represent F (z) by a

convergent power series
∞
X
F (z) = an z n .
n=0
Moreover, the coefficients are given by
I
1 1 F (w)
an = F (n) (0) = dw.
n! 2πi ∂D(R) wn+1
Proof. For any |z| < R, we can use the Cauchy integral formula to write
I
1 F (w)
F (z) = dw
2πi ∂D(R) w − z
z −1
I
1 F (w) 
= 1− t dw
2πi ∂D(R) w w
I ∞
1 F (w) X  z n
= dw
2πi ∂D(R) w n=0 w
∞
" I #
X 1 F (w)
= dw z n .
n=0
2πi ∂D(R) wn+1