6 Laurent’s Theorem

Theorem 6.1 Let A = {z : R < |z − a| < S} and suppose f is holomorphic

on A. Then ∞
cn (z − a)n
X
f (z) =
n=−∞

for z ∈ A where
1 Z f (w)
cn = dw
2πi γ(a;r) (w − a)n+1
for R < r < S. The cn are unique.

Proof 6.1 WLOG a = 0. Fix z ∈ A and choose P and Q so that R < P <
˜ then
|z| < Q < S. Choose γ̃ and γ̃;

1 Z f (w)
f (z) = dw
2πi γ̃ w − z
(by Cauchy integral formula)

1 Z f (w)
0= dw
2πi γ̃˜ w − z
(by Cauchy’s theorem) Hence

1 Z f (w) 1 Z f (w)
f (z) = dw − dw
2πi γ(0;Q) w − z 2πi γ(0;P ) w − z
∞ ∞
1 Z X zn 1 Z X −wm
= f (w)dw − f (w)dw
2πi γ(0;Q) n=0 wn+1 2πi γ(0;P ) m=0 z m+1
using the binomial expansion. Use the Uniform Convergence Theorem to
interchange summation and integration. This gives
∞
!
1 Z f (w)
dw z n
X
f (z) = n+1
n=0 2πi γ(0;Q) w
∞
!
1 Z
f (w)wm dw z −m−1 .
X
+
m=0 2πi γ(0;P )
Use the deformation theorem to replace γ(0; Q) and γ(0; P ) by γ(a; r) as in
the statement.

1
Example 6.1
1
f (z) =
z(1 − z)
is holomorphic on A1 and A2 , where
A1 = {z : 0 < |z| < 1}
and
A2 = {z : |z| > 1}.
On A1 ,
∞
f (z) = z −1 + (1 − z)−1 = zn
X

n=−1
On A2 , we have
−2
−1 −1 −1 −1
−z n .
X
f (z) = z − z (1 − z ) =
n=−∞

Example 6.2
1
f (z) =
z(1 − z)2
is holomorphic on 0 < |z − 1| < 1. On this region it is equal to
1 1 1 
2

= 1 − (z − 1) + (z − 1) − . . .
(z − 1)2 1 + (z − 1) (z − 1)2
So ∞
(−1)n (z − 1)n .
X
f (z) =
n=−2

Example 6.3
∞
cn z n
X
csc(z) =
n=−∞

on 0 < |z| < π. Since

z3
sin(z) = z − + ...
3!
also
(z 2 )
!
1
csc(z) = 1− + O(z 4 ))−1
z 3!
z2
!
1
= 1+ + ...
z 3!

2
Example 6.4
 
cot(z) = 1 − z 2 /z! + . . . (1/z + z/3! + . . .)

1 
= 1 + z 2 (−1/2 + 1/6) + O(z 4 ) .
z

6.1 Singularities
Definition 6.2 A point a is a regular point of f if f is holomorphic at a. It
is a singularity of f if a is a limit point of regular points which is not itself
regular.

Definition 6.3 f has an isolated singularity at a if f is holomorphic in a

punctured disc D(a; r) \ {0}; if a is a singular point that does not satisfy this
condition, it is called a non-isolated singularity.

If f has an isolated singularity at a, f is holomorphic in the annulus {z :

0 < |z − a| < r} and has a unique Laurent expansion f (z) = ∞ n=−∞ cn (z −
P
n
a) . The singularity a is: removable singularity if cn = 0∀n < 0; pole of order
m if c−m 6= 0, cn = 0∀n < −m; isolated essential singularity if there does not
exist m such that cn = 0 ∀n < −m.
In D0 (a, r), f (z) = −1 n P∞ n
n=−∞ cn (z − a) + n=0 cn (z − a)
P

Definition 6.4 The principal part of the Laurent expansion is

−1
cn (z − a)n .
X

n=−∞

6.2 Zeros
Suppose f is holomorphic in D(a; r) and f (a) = 0. Assume f is not identi-
cally zero in D(a; r) (in other words f is not zero everywhere in D(a; r)).
Then by Taylor’s theorem,
∞
cm (z − a)m
X
f (z) =
n=m

for some m ≥ −1, cm 6= 0.

The order of zero of f at a is m if and only if f (a) = f 0 (a) = . . . =
(m−1)
f (a) = 0 but f (m) (a) 6= 0.

3
Theorem 6.5 Suppose f is holomorphic in D(a; r). Then f has a zero of
order m at a if and only if limz→a (z − a)−m f (z) = C for some constant
C 6= 0.

Theorem 6.6 (Theorem 2) Suppose f is holomorphic on D0 (a; r). Then f

has a pole of order m at a if and only if

lim (z − a)m f (z) = D

z→a

for a nonzero constant D.

Example 6.5 z sin(z) has a zero of order 2 at z = 0 and has zeros of order
1 at z = nπ, n 6= 0.

Proof of Theorem 2 =⇒ Suppose a is a pole of order m. For z ∈ D(a; r),

z 6= a,
∞
cn (z − a)n , c−m 6= 0
X
f (z) =
n=−m

In D0 (a; r),
∞
(z − a)m f (z) = cn−m (z − a)n .
X

n=0

The series on the right hand side defines a function continuous at z = a.

Hence
lim (z − a)n f (z) = c−m 6= 0.
z→a

By Laurent’s theorem,
∞
cn (z − a)n
X
f (z) =
n=−∞

where
1 Z f (w)
cn = dw
2πi γ(a;s) (w − a)n+1
(for 0 < s < r). We need cn = 0 (n < −m) and c−m 6= 0. Since limz→a (z −
a)m f (z) = D 6= 0, there is δ > 0 such that

|(w − a)m f (w) − D| < 

4
where 0 < |w − a| < δ. Take 0 < s < min(δ, r). Then if |w − a| = s, then
|(w − a)m f (w)| ≤ |D| + Hence(w − a)−n−1 f (w)| ≤ (|D| + ) s−m−n−1 . So
using the estimation theorem,
|cn | ≤ (|D| + )s−n−m .
If n < −m then s−n−m can be made arbitrarily small, but cn is independent
of s so cn = 0. Hence
∞
cn (z − a)n .
X
f (z) =
n=−m

As in the proof of (⇒),

c−m = lim (z − a)m f (z) = D 6= 0.
z→a

6.3 Behaviour near an isolated singularity

Case 1: Removable singularity. If f (z) = ∞ n 0
n=0 cn (z − a) in D (a; 4). then
P

f (z) → c0 as z → a. Redefining f (a) to c0 we find f is holomorphic in

D(a; r).
sin(z)
Example 6.6 f (z) = z

Case 2: Pole If f has a pole at a, then |f (z)| → ∞ as z → a.

Case 3: Essential singularity:
Casorati-Weierstrass theorem (not proved): If f has an isolated essential
singularity at a, for any w ∈ C there exists < an > such that an → a and
f (an ) → w. In fact according to Picard’s theorem, in any D0 (a, r), f assumes
every complex value except possibly one. For example, e1/z has an essential
singularity at 0; the value not assumed is 0.
Definition 6.7 The extended complex plane Ĉ is C ∪ {∞} (add an extra
point at ∞).
Define this by identifying
Ĉ = U ∪ V / ∼
where U = C, V = C. On C \ {0}, identify U with V via u ∈ U \ {0}
∼ v ∈ V where v = 1/u. So as u → ∞, v → 0 and as u → 0, v → ∞. We
can also write Ĉ as {[z, w]}/ ∼ where (z, w) ∼ (λz, λw) for any λ ∈ C \ {0}.
Thus if z 6= 0, (z, w) ∼ (1, w/z) and if w 6= 0, then (z, w) ∼ (z/w, 1). These
are in correspondence with the sets U and V .
Uniqueness

5
Theorem 6.8 Suppose f is holomorphic on A with
∞
bn (z − a)n .
X
f (z) =
n=0

Suppose also
∞
cn (z − a)n .
X
f (z) =
n=0

Then bn = cn .

Proof 6.2 Assume a = 0. Choose r with R < r < S. Then

Z
2πicn = f (w)w−n−1 dw
γ(0;r)

Z ∞
bk wk−n−1 dw
X
=
γ(0;r) k=−∞

Z ∞ Z ∞
bk wk−n−1 dw = b−m w−m−n−1 dw
X X
=
γ(0;r) k=0 γ(0;r) m=−1

Using the theorem on uniform convergence to interchange the sum with

the integral,
∞ Z
wk−n−1 dw = 2πibn .
X
2πicn = bk
k=−∞ γ(0,r)

6.4 Meromorphic functions

Definition 6.9 A C-valued function which is holomorphic in an open set
G ⊂ Ĉ except possibly for poles is called meromorphic in G.

Theorem 6.10 If f is holomorphic on Ĉ, then f is constant.

(Proof: Use Liouville’s theorem)

Theorem 6.11 If f is meromorphic on Ĉ then f is a rational function

p(z)/q(z) for some polynomials p and q.

Example 6.7 1. (z − a)−2 has a double pole at z = 1

6
 2. (1 − cos(z))/z 2 is holomorphic except at z = 0. At z = 0, 1 − cos(z) =
z 2 /2 + . . . so the singularity is removable.
1 1 1
3. sin(z)
= z−z 3 /3!+...
= z(1−z 2 /3!+...)
so there is a simple pole at 0.

4.
cos(z)/ sin(z) = (1 − z 2 /2 + . . .)(1/z)(1 + z 2 /3! + O(z 4 ))
cot(z) has a simple pole at z = 0 Similarly since cot(z − kπ) = cot(z),
cot(z) has a simple pole at kπ.
P∞ n z −(2n+1)
5. sin(1/z) = n=0 (−1) (2n+1)! has an isolated essential singularity at 0.
1
6. is singular when sin(1/z) = 0, in other words when 1/z = kπ or
sin(1/z)
z = 1/(kπ) for some integer k 6= 0.

Remark 6.1 If a is a limit point of the singularities of a function defined on

a subset of C, then f cannot be holomorphic on any punctured disk with centre
at a, and cannot have a Laurent expansion about a. So a is not an isolated
singularity or a regular point; it is a non-isolated essential singularity.

Example 6.8
1
f (z) = .
z3 cos(1/z)
This has poles at 1/z = (2n+1)π/2, where cos(1/z) has zeros (in other words
2
z = (2n+1)π ). This expression tends to 0 as n → ∞ so 0 is a limit point of
the poles, or a non-isolated essential singularity. It follows that f does not
have a Laurent expansion about 0.

If f is meromorphic in an open subset G of C̃, then the set of poles of f

has no limit point in G, and f can have at most finitely many poles in any
closed subset of G.

6.5 Behaviour of functions at ∞

Zeros and poles of a function f at ∞ are studied by studying the function
fˆ(w) = f (1/w). f has a pole of order m at ∞ if and only if fˆ has a pole of
order m at 0.
f has a zero of order m at ∞ if and only if fˆ has a zero of order m at 0.
Examples:

7
 1.
f (z) = z 3
fˆ(w) = f (1/2) = w−3
At w = 0, fˆ has a pole of order 3.

2. f (z) = 1
z2
sin z1 . fˆ(w) = w2 sin(w). fˆ(w) has a zero of order 3 at w = 0.

3. f (z) = z sin(1/z)
fˆ(w) = sin(w)/w has a removable singularity at w = 0 if and only if
z = ∞, in other words w = 0.

Example 6.9
(z − 1)2 cos(πz)
f (z) =
(2z − 1)(z 2 + 1)5 sin3 (πz)
The denominator is zero at z = 1/2, z = ±i and z = k ∈ Z.
z = 1/2 :
(z 2 + 1)5 sin3 (πz) 6= 0
and (z − 1)2 6= 0 but cos(πz) = 0. So z = 1/2 is a removable singularity.
z = ±i: 2z − 1 6= 0, sin3 (πz) 6= 0, z − 1 6= 0, cos(πz) 6= 0. So z 2 + 1 =
(z + i)(z − i) and z = ±i are poles of f (z) of order 5.
z = k, k 6= 1: cos(πz) 6= 0, z − 1 6= 0
sin(πz) = (πz − πk)(1 + higherorder)
so 2z − 1 6= 0, z 2 + 1 6= 0.
So z = k is a pole of order 3, and z = 1 is a pole of order 1.

6.6 Meromorphic functions

Definition 6.12 A C-valued function which is holomorphic in an open set
G ⊂ C except possibly for poles is called meromorphic in G.

Theorem 6.13 If f is holomorphic on C̃, then f is constant.

(Proof: Use Liouville’s theorem)

Theorem 6.14 If f is meromorphic on Ĉ then f is a rational function
p(z)/q(z) for some polynomials p and q.

Example 6.10 1. (z − a)−2 has a double pole at z = 1

8
 2. (1 − cos(z))/z 2 is holomorphic except at z = 0. At z = 0, 1 − cos(z) =
z 2 /2 + . . . so the singularity is removable.
1 1 1
3. sin(z)
= z−z 3 /3!+...
= z(1−z 2 /3!+...)
so there is a simple pole at 0.

4.
cos(z)/ sin(z) = (1 − z 2 /2 + . . .)(1/z)(1 + z 2 /3! + O(z 4 ))
cot(z) has a simple pole at z = 0 Similarly since cot(z − kπ) = cot(z),
cot(z) has a simple pole at kπ.
P∞ n z −(2n+1)
5. sin(1/z) = n=0 (−1) (2n+1)! has an isolated essential singularity at 0.
1
6. is singular when sin(1/z) = 0, in other words when 1/z = kπ or
sin(1/z)
z = 1/(kπ) for some integer k 6= 0.

Remark 6.2 If a is a limit point of the singularities of a function defined on

a subset of C, then f cannot be holomorphic on any punctured disk with centre
at a, and cannot have a Laurent expansion about a. So a is not an isolated
singularity or a regular point; it is a non-isolated essential singularity.

Example 6.11
1
f (z) = .
z3 cos(1/z)
This has poles at 1/z = (2n+1)π/2, where cos(1/z) has zeros (in other words
2
z = (2n+1)π ). This expression tends to 0 as n → ∞ so 0 is a limit point of
the poles, or a non-isolated essential singularity. It follows that f does not
have a Laurent expansion about 0.

If f is meromorphic in an open subset G of C̃, then the set of poles of f

has no limit point in G, and f can have at most finitely many poles in any
closed subset of G.
Behaviour of functions at ∞
Zeros and poles of a function f at ∞ are studied by studying the function
fˆ(w) = f (1/w). f has a pole of order m at ∞ if and only if fˆ has a pole of
order m at 0.
f has a zero of order m at ∞ if and only if fˆ has a zero of order m at 0.
Examples:

9
 1.
f (z) = z 3
fˆ(w) = f (1/2) = w−3
At w = 0, fˆ has a pole of order 3.
2. f (z) = 1
z2
sin z1 . fˆ(w) = w2 sin(w). fˆ(w) has a zero of order 3 at w = 0.

3. f (z) = z sin(1/z) fˆ(w) = sin(w)/w has a removable singularity at

w = 0 if and only if z = ∞.
Example 6.12
(z − a)2 cos(πz)
f (z) =
(2z − 1)(z 2 + 1)5 sin3 (πz)
The denominator is zero at z = 1/2, z = ±i and z = k ∈ Z.
z = 1/2 :
(z 2 + 1)5 sin3 (πz) 6= 0
and (z − 1)2 6= 0 but cos(πz) = 0. So z = 1/2 is a removable singularity.
z = ±i: 2z − 1 6= 0, sin3 (πz) 6= 0, z − 1 6= 0, cos(πz) 6= 0. So z 2 + 1 =
(z + i)(z − i) and z = ±i are poles of F (z) of order 5.
z = k, k 6= 1: cos(πz) 6= 0, z −1 6= 0 sin(πz) = (πz −πk)(1+higherorder)
so 2z − 1 6= 0, z 2 + 1 6= 0. So z = k is a pole of order 3, and z = 1 is a pole
of order 1.
Example 6.13 f (z) = z sin(z) has zeros at z = nπ. So 1/f has poles at
z = kπ. At z = 0, sin(z) = z − z 3 /3! + . . . = z(1 − z 2 /3! + O(z 4 )) so
1/ sin(z) = 1/z(1−z 2 /3!+O(z 4 ))−1 = (1/z)(1+(z 2 /3!+. . .)+(z 2 /3!+. . .)2 +. . .
= (1/z)(1 + z 2 /3! + O(z 4 ))
So 1/(z sin(z)) has a double pole at z = 0. At z = kπ, k 6= 0,
sin(z) = (−1)k sin(z − kπ)
and
z = kπ + (z − kπ)
So ∞
1 1 1 X z − kπ n
= = (−1)n ( ) .
z kπ (1 + (z − kπ)/kπ) kπ n=0 kπ
So 1/(z sin(z)) has a simple pole at z = kπ when k 6= 0.

10
Example 6.14
f (z) = cot(z) = cos(z)/ sin(z)
Since sin(z) = 0 when z = kπ and cos(kπ) 6= 0, and sin(z) has a simple zero
at z = kπ,
sin(z) = (z − kπ)(1 − (z − kπ)2 /3! + . . .)
we find cot(z) has simple poles at these values.

Non-isolated singularities are always essential. (Non-isolated singularity

means there is no punctured disk {z : 0 ≤ |z − a| < r} where f is holomor-
phic.) If a is an isolated singularity (f is holomorphic in D0 (a; r) for some
r), there is always a Laurent expansion
∞
cn (z − a)n
X
f (z) =
n=−∞

(i) Isolated essential singularity: Infinitely many nonzero cj for j < 0

(ii) Pole: f (z) = c−m (z − a)−m + c−m+1 (z − a)−m+1 + . . . of order m.
(iii) Removable singularity: All cn = 0 if n < 0.
1 1 1
= 2 = 3
z2 sin(z) 3 5
z (z − z /3! + z /5! + . . .) z (1 − A)

where A = z 2 /3! − z 4 /5! + . . .

1
= 3
(1 + A + A2 + . . .)
z
Expand 1 + A + . . . to order z 2 in z.

11