6.

complex analysis

6.6 Residues and poles

Lecture 17
18th November 2024 The concept of residue is one of the most powerful techniques in complex
analysis, which allows systematic evaluation of various definite integrals that
are otherwise impossible.

6.6.1 Residue

Isolated singularity Recall that we call a point z0 a singlar point of a function

Most of the singularities f , if f fails to be analytic at that point. If, in addition, there exists some
that we have encountered so
far are isolated singularities.
neighborhood of z0 in which f is analytic everywhere except at z0 , we call z0
But there exist functions an isolated singularity of f .
with singularities which are
not isolated. e.g., log(z). Let z0 be an isolated singularity. Then there exists a series expansion
1
X a 1 a 2
analsiicalsewhe
f (z) = an (z z0 ) n + + + ··· . (532)
z z0 (z z 0 )2
1
//f n=0

I
which is valid in some region r < |z z0 | < R for finite r and R. Such an
Zo hon -

analytic
expansion is called the Laurant series expansion. How do we calculate the
ex ) f(z ) =
π ,
Z0 = 0 coefficients an ’s? We can use the following useful identity
I
1
(z z0 )n k 1 dz = nk , (533)
22
ex
10gz fngolas 2⇡i
@

when is a contour that encircles z0 . (To prove this, let be a circle of radius
R centered at z0 . Evaluate the integral and send R ! 1.) This identity implies
that I
1 f (w)dw
tseio an = , n2Z. (534)
= Z
2⇡i (w z0 )n+1
z
.

xalwoys well - defined

L ircle around 2o = Z.
Eo :
f is an al ii a

Residue In particular, let us consider the coefficient of the (z z0 ) 1 term,

I
1
Annf (
'(
Eo ) fon
f)
a 1= f (z)dz . (535)
2⇡i
tay low ex pansion
.

Here is any closed contour around z0 as usual, which bounds the region
where f is analytic everywhere except at z0 . This number a 1 is called the
residue of f at the isolated singular point z0 . We use the following notation to
denote the residue at z0 :
I
2⇡i Resf (z) = f (z)dz . (536)
z=z0

= closed rure tha

.
L

Y bounds a negron D

98 D analyyis
where fC a7 is

throw gh out
.
 6.6 . residues and poles

What if the function f has other isolated singularties within the contour ?
One can show that the following theorem is true: If f (z) is analytic within
and on a closed contour , except at a finite number of isolated singularities
z1 , · · · zm interior to ,then
I m
X
f (z)dz = 2⇡i Resf (z) . (537)
z=zk
k=1

This is called the residue theorem. The proof can be done the Cauchy-Grousat
theorem:
I I I I
0 = f (z)dz = f (z)dz + f (z)dz + f (z)dz , (538)
C circles lines

where the contour C is chosen as in the figure below, so that f (z) is analytic
everywhere inside C.

Then it is clear that

I m I
X m
X
f (z)dz = f (z)dz = 2⇡i Resf (z) . (539)
z=zk
k=1 C k=1
k

Example Consider the integral

I
2z 3
dz , (540)
z(z 1)

where is the circle |z| = 2. This contour contains two isolated singularities at
z = 0 and z = 1. We can first expand around the point z = z0 , to find
2z 3 3 1 3
= = + 1 + z + ··· , (541)
z(z 1) z z 1 z

which is valid for |z| < 1. Therefore

Lbay =
3

99
6. complex analysis

Resf (z) = 3 . (542)

z=0

We can then expand the function around z = 1,

1
X
2z 3 3 1 1
= = +2 ( 1)n (z 1)n , (543)
z(z 1) z 1+1 z 1 z 1
k=0

which gives
Resf (z) = 1 . (544)
z=1

Therefore, we have
I
2z 3
dz = 2⇡i(3 1) = 4⇡i . (545)
z(z 1)

Let us come back to the Laurent series expansion of a function around an

isolated singularity z = z0 ,
1
X a 1 a 2
f (z) = an (z z0 ) n + + + ··· . (546)
z z0 (z z 0 )2
n=0

The terms with negative powers of (z z0 ) is called the principal part of f at

z0 . Depending on the bahaviour of the principal part, one can classify the
singularities:

• Pole of order m: We say that f (z) has a pole of order m if a n = 0 for all
n > m and a m is non-zero. In this case, the expansion takes the form
1
X a 1 a n
f (z) = an (z z0 ) n + + ··· + . (547)
z z0 (z z0 ) m
n=0
m tr on cosoro

• Simple pole: If m = 1, we call it a simple pole. If z0 is a simple pole, there

is a very useful formula to calculate the residue. It is clear that

Resf (z) = lim [(z z0 )f (z)] . (548)

z=z0 z!z0

• Essential singularity: If the Laurent series expansion of f (z) has infinite

number of terms in its principal part, we call it an essential singularity.
A prototype of functions with essential singularity is
1
X ✓ ◆
1 1
e1/z = . (549)
n! z n
n=0

100
 6.6 . residues and poles

To see how strange this function is, consider approaching z ! 0 by

setting
z = (ln a + 2n⇡i) 1 , (550)

for arbitary real number a and integer n, and taking the limit n ! 1.
Then we have
e1/z = ae2⇡ni = a , (551)

which implies that the function approaches arbitrarily close to any real
number.

The point at infinity In many cases, it is convenient to include the point at

infinity to the complex plane, denoted by z = 1. The complex plane, together
with this point,
C [ {1} , (552)

is sometimes called the extended complex plane. How should we think about
this point 1? Let us consider a sphere of radius 1 centered at the origin of
the complex plane, as in the figure below: To each point z on the complex
…

7 P- →
Map : 5

plane, there corresponds exactly one point P on the surface of the sphere.
This point P can be determined by the intersection of the line (indicated in
red) through the point z and the north pole N of the sphere with that surface.
For example, the origin of the complex plane z = 0 corresponds to the South
sphone
BTemann
pole of the sphere. In this way, one can construct a one-to-one map from the
Bloch spheve entire complex plane to the points on the sphere, except the north pole N. By
letting the point N of the sphere correspond to the point at infinity, we can
obtain a one-to-one correspondence between the points of the sphere and the
extended complex plane. This sphere is called the Riemann sphere, and the
correspondence is called the stereographic projection.
It is clear from this picture that the small region around the point N on
the sphere corresponds to the region exterior to the circle |z| = 1/✏. Therefore
it is reasonable to define the neighborhood of the point 1 to be the region
|z| > 1/✏. Using this concept, one can prove the following theorem:
✓ ◆
1
lim f (x) = w0 if and only if lim f = w0 . (553)
z!1 z!0 z

101
 6. complex analysis

One can also show that limz!z0 f (z) = 1 if and only if limz!z0 1/f (z) = 0.
Contour at infinity Consider the contour , which encircles the point at
infinity. How do we evaluate I
dz
? (554)
z

Note that it is natural to write this integral as the residue at infinity,

1
2⇡i Res . (555)
z=1 z

To evaluate this, we use the above theorem to change the variable

z = 1/w . (556)

It is important to note that we have dz = 1/w2 dw, which allows us to write

I I
dz dw
= , (557)
z w
0

where is the contour around w = 0. Therefore we can write

1 1
2⇡i Res = 2⇡i Res = 2⇡i . (558)
z=1 z w=0 w

6.6.2 Evaluation of integrals

The most important application of the residue integral is in the evaluation of

the definite integral of a real functions. The following lemma lies at the center
of this idea.
Λ
e"
Jordan’s lemma Let CR is a semicircle of radius R centered at the origin in
the upper half plane of C (See figure below). If f (z) is a function that decays
sufficiently fast (more precisely, if it uniformly goes to zero faster than 1/|z|)
as |z| ! 1, then Z
lim e i↵z f (z)dz = 0 , (559)
R!1
CR

for a non-negative real number ↵.

The proof is simple. The semi-circle Cr is parameterized by z = Re i✓ . Then

Z Z⇡
I R := e i↵z
f (z)dz = e i↵R cos ✓ ↵R sin ✓
i Rf (Re i✓ )d✓ (560)
CR 0

102
tvrys : E : IA ca ,expementially decay
1
 6.6 . residues and poles

α 20

α≤ O

Then
Z⇡
↵R sin ✓
|I R | < e R|f (Re i✓ )|d✓ . (561)
0

By assumption, there exists a positive number ✏(R), that remains finite as

R ! 1, such that
R|f (Re i✓ )| < ✏(R) . (562)

Therefore
Z⇡ ⇡/2
Z
↵R sin ✓ ↵R sin ✓
|I R | < ✏(R) e d✓ = 2✏(R) e d✓ . (563)
0 0

Since sin ✓ 2✓/⇡ for 0  ✓  ⇡/2, we can write

⇡/2
Z
(↵R/⇡)✓ ⇡✏(R) ↵R
|I R | < 2✏(R) e d✓ = (1 e ), (564)
↵R
0

which goes to zero as R gets sufficiently large. Note that the theorem is valid
for ↵ = 0 as well because the numerator of the RHS of the above equation
.

Jap fland a -
0

approaches ↵R as we take ↵ ! 0.

The power of this lemma will become clear once we work out the following
example:

Example Let us consider the real integral

Z1 .
*.
x2

.
I= dx . =

.
(565)
(x2 + 1)(x2 + 9)
0

To evaluate this, consider the following complex integral

103
6. complex analysis

I
z2

sordan
0 1
I = dz , = If (566)
2 (z 2 + 1)(z 2 + 9)
C
' s leama

where C is the contour given in the figure below. Once we show that the

integrand of this function decays sufficiently fast in the limit |z| ! 1, the
Jordan’s lemma implies that there’s no contribution from the semi-circle and
therefore I = I0 . Indeed, we have

R2 e2i✓ R3
R|f (Re i✓ )| = R = , (567)
(R2 e2i✓ + 1)(R2 w2i✓ + 9) |R2 e 2i✓ + 1| |R2 w2i✓ + 9|

which clearly vanishes as R ! 1. The integral (571) can be evaluated via the
residue integral formula. The function has four simple poles at x = ±i and
x = 3 ± i, and two of them are inside the contour. The residues are
1
Resf (z) = , (568)
z=i 16i

and
3
Res f (z) = , (569)
z=3i 16i

which gives ✓ ◆
1 3 ⇡
I = I0 = ⇡i + = . (570)
16i 16i 8

It is clear that the Jordan’s lemma holds if the contour CR is the semicircle
in the lower half plane of C, as long as the integrand decays sufficiently fast
on CR as R ! 1. One can try to evaluate the integral (565) by closing the
contour to the lower half-plane, i.e.,

104
 6.6 . residues and poles

I
1 00 z2
I=I = dz , (571)
2 (z 2 + 1)(z 2 + 9)
C0

where C0 is the contour given in the figure below.

It is important to remember that the orientation of the contour C0 is

clockwise. The residue theorem tells us that

I00 = 2⇡i Res + Res f (z) . (572)
z= i z= 3i

The residues at each poles are

1
Res f (z) = , (573)
z= i 16i

and
3
Res f (z) = , (574)
z= 3i 16i

which gives
⇡
I = I00 = , (575)
8

which is the same as the result (572). Why is this expected?

<
I dt z
to.
→
= ein .

+ corso

coo - s(
) do a
ndo
-

21 - t2 =
①
105
=
-

Besf ( z )