COMPLEX ANALYSIS

SHAIVITE

15 January 2026

QUESTION PAPER 2019-20

Question 1(a)
Question: Find the radius of convergence of
∞
X n! n
z .
n=0
nn

Solution

Let
n!
an =.
nn
The radius of convergence R is given by the Cauchy–Hadamard formula
1 p
= lim sup n |an |.
R n→∞

Compute r
√ n n! (n!)1/n
n
an = n
= .
n n

Using Stirling’s approximation

√  n n
n! ∼ 2πn ,
e
we get
n
(n!)1/n ∼ .
e

1
Hence
√ (n!)1/n n/e 1
n
an = ∼ = .
n n e
Therefore
1 1
= ⇒ R = e.
R e

Practice Problems

Find the radius of convergence of the following power series:

∞
X n! n
1. n nn
z
n=1
2
∞
X (n!)2 n
2. z
n=1
n2n
∞
X n! n
3. n
z
n=1
(3n)
∞
X n!
4. n
(2z)n
n=1
n
∞
X n! n
5. n nn
z
n=1
5

Practice Solutions

Problem 1.
n! √ (n!)1/n n/e 1
an = , n
an = ∼ = .
2 nn
n 2n 2n 2e
Hence
R = 2e.

Problem 2.
(n!)2 √ (n!)2/n (n/e)2 1
an = , n
an = ∼ = 2.
n2n n 2 n 2 e
Thus
R = e2 .

Problem 3.
n! √ (n!)1/n n/e 1
an = , n
an = ∼ = .
(3n)n 3n 3n 3e
So
R = 3e.

Problem 4.
X n! X n!
n
(2z)n = 2n z n .
n nn
Thus
√ (n!)1/n 2 2
n
an = ∼ .
n e

2
Hence
e
R= .
2
Problem 5.
n! √ (n!)1/n n/e 1
an = , n
an = ∼ = .
5n nn 5n 5n 5e
Therefore
R = 5e.

Question 1(b)
Question: Let f (z) = U (x, y) + iv(x, y) be analytic in a region D, then show that
∂(U, v)
|f ′ (z)|2 = .
∂(x, y)

Solution

Since f (z) = U (x, y) + iv(x, y) is analytic, the Cauchy–Riemann equations hold:

Ux = Vy , Uy = −Vx .
The derivative of f is
f ′ (z) = Ux + iVx .
Hence
|f ′ (z)|2 = Ux2 + Vx2 .
Using the Cauchy–Riemann equations,
Vx = −Uy ,
so
|f ′ (z)|2 = Ux2 + Uy2 .

The Jacobian determinant of (U, V ) with respect to (x, y) is

∂(U, V ) U Uy
= x = Ux Vy − Uy Vx .
∂(x, y) Vx Vy
Using Vy = Ux and Vx = −Uy , we get
∂(U, V )
= Ux2 + Uy2 .
∂(x, y)
Therefore,
∂(U, V )
|f ′ (z)|2 = .
∂(x, y)

Practice Problems
∂(U, V )
1. For f (z) = z 2 , verify that |f ′ (z)|2 = .
∂(x, y)
2. For f (z) = ez , verify the same identity.
3. For f (z) = sin z, verify the identity.
4. Let f (z) = z 3 . Show that the Jacobian equals |f ′ (z)|2 .
5. Let f (z) = log z (principal branch). Verify the result away from the branch cut.

3
Practice Solutions

1. f (z) = z 2 = (x + iy)2 = x2 − y 2 + i(2xy), so

U = x2 − y 2 , V = 2xy.

Then
Ux = 2x, Uy = −2y, Vx = 2y, Vy = 2x.
Hence
∂(U, V )
= Ux Vy − Uy Vx = 4x2 + 4y 2 .
∂(x, y)
Also f ′ (z) = 2z, so |f ′ (z)|2 = 4(x2 + y 2 ). The result holds.
2. f (z) = ex+iy = ex (cos y + i sin y), so

U = ex cos y, V = ex sin y.

A direct computation gives

∂(U, V )
= e2x .
∂(x, y)
Since f ′ (z) = ez , we have |f ′ (z)|2 = |ez |2 = e2x .
3. f (z) = sin z = sin x cosh y + i cos x sinh y. Then

∂(U, V )
= cos2 x cosh2 y + sin2 x sinh2 y = | cos z|2 = |f ′ (z)|2 .
∂(x, y)

4. f (z) = z 3 , f ′ (z) = 3z 2 , so |f ′ (z)|2 = 9|z|4 . A Jacobian computation from U + iV = (x + iy)3 yields

the same value.
5. f (z) = log z, f ′ (z) = 1/z, hence |f ′ (z)|2 = 1/|z|2 . Computing the Jacobian from U = ln |z|, V = arg z
gives the same result.

Question 1(c)
Question: Evaluate I
2
ze1/z dz.
|z|=1

Solution

The integrand is
2
ze1/z .
Expand the exponential in a Laurent series about z = 0:
∞
2 X 1 1
e1/z = .
n=0
n! z 2n

Hence
∞
2 X 1 1−2n
ze1/z = z .
n=0
n!

4
 The residue at z = 0 is the coefficient of z −1 . We need

1 − 2n = −1 ⇒ n = 1.

For n = 1, the coefficient is

1
= 1.
1!
Therefore 2
Res ze1/z


= 1.
z=0

By Cauchy’s residue theorem, I

2
ze1/z dz = 2πi × 1 = 2πi.
|z|=1

Practice Problems

Evaluate the following integrals:

I
1. ze1/z dz
|z|=1
I
2
2. z 2 e1/z dz
|z|=1
I
2
3. e1/z dz
|z|=1

e1/z
I
4. dz
|z|=1 z
I
2
5. z 3 e1/z dz
|z|=1

Practice Solutions

1.
∞ ∞
X 1 1 X 1 1−n
ze1/z = z n
= z .
n=0
n! z n=0
n!
1
We need 1 − n = −1 ⇒ n = 2. Residue = 2! = 12 . Hence integral = πi.
2.
∞
2 X 1 2−2n
z 2 e1/z = z .
n=0
n!
We need 2 − 2n = −1, which has no integer solution. So residue = 0 and the integral is 0.
3.
∞
2 X 1 −2n
e1/z = z .
n=0
n!

No z −1 term occurs, so the integral is 0.

4.
∞
e1/z X 1 −n−1
= z .
z n=0
n!

5
The coefficient of z −1 corresponds to n = 0, which is 1. Hence the integral is 2πi.
5.
∞
2 X 1 3−2n
z 3 e1/z = z .
n=0
n!
1
We need 3 − 2n = −1 ⇒ n = 2. Residue = 2! = 12 . So the integral is πi.

Question 1(d)
Question: Find the singularities of
z+1
.
z sin z

Solution

The function
z+1
f (z) =
z sin z
is singular wherever the denominator vanishes, i.e. where

z sin z = 0.

This occurs when

z=0 or sin z = 0.
Since sin z = 0 at z = nπ, n ∈ Z, the singular points are

z = 0, ±π, ±2π, ±3π, . . .

Now classify them.

At z = 0, use the expansion
z3
sin z = z − + ···
6
so
z4
z sin z = z 2 − + ···
6
and hence
z+1
f (z) = 2
.
z 2 1 − z6 + · · ·
Since the numerator z + 1 is nonzero at z = 0, the function has a pole of order 2 at z = 0.
For z = nπ, n ̸= 0, we have

sin z = (z − nπ) cos(nπ) + · · · = (−1)n (z − nπ) + · · ·

so
z sin z = z(−1)n (z − nπ) + · · ·
which has a simple zero at z = nπ. Since the numerator z + 1 is nonzero at z = nπ, these are simple poles.
Thus,
z = 0 is a pole of order 2, z = nπ (n ̸= 0) are simple poles.

6
Practice Problems

Find and classify the singularities of:

z−1
1.
z sin z
z2
2.
z sin z
z+2
3.
z 2 sin z
1
4.
z sin z
z
5.
sin z

Practice Solutions

1. Singularities at z = 0, nπ. At z = 0 it is a pole of order 2. At z = nπ (n ̸= 0) simple poles.

2. z 2 /(z sin z) = z/ sin z has removable singularity at z = 0 and simple poles at z = nπ, n ̸= 0.
3. z = 0 is a pole of order 3. z = nπ, n ̸= 0 are simple poles.
4. z = 0 is a pole of order 2, z = nπ, n ̸= 0 simple poles.
5. z/ sin z has a removable singularity at z = 0 and simple poles at z = nπ, n ̸= 0.

Question 1(e)
Question: Find and sketch the region onto which the half–plane y > 0 is mapped by the transformation

w = (1 + i)z.

Solution

Let
z = x + iy, y > 0.
Then
w = (1 + i)(x + iy) = x + iy + ix − y = (x − y) + i(x + y).
Write
w = u + iv,
where
u = x − y, v = x + y.

We want to find the image of the region y > 0 in the uv–plane.

From
u = x − y, v = x + y,
solve for x, y:
u + v = 2x, v − u = 2y.

7
Thus
v−u
y= .
2
Since y > 0, we must have
v−u
>0 ⇒ v > u.
2
Hence the image of the half–plane y > 0 is the region

v > u,

which is the half–plane above the line v = u in the w–plane.

The boundary y = 0 maps to
v = u,
a straight line through the origin making an angle π/4 with the u–axis.

Practice Problems

Find the image of y > 0 under the transformations:

1. w = (1 − i)z
2. w = 2(1 + i)z

3. w = (2 + i)z
4. w = iz
5. w = (3 − 3i)z

Practice Solutions

1. w = (1 − i)(x + iy) = (x + y) + i(y − x). Hence u = x + y, v = y − x. Then y = (u + v)/2 > 0, so u + v > 0.

2. w = 2(1 + i)z scales the image of Question 1(e) by 2, so the image is still v > u.
3. w = (2 + i)(x + iy) = (2x − y) + i(2y + x). Solving gives y = (2v − u)/5 > 0, so 2v − u > 0.
4. w = iz = −y + ix, so u = −y, v = x. Since y > 0, we get u < 0, the left half–plane.
5. w = (3 − 3i)z = 3(1 − i)z, so the image is the same as in Problem 1: u + v > 0.

Question 1(f )
Question: If AB is an arc α ≤ θ ≤ β of the circle |z| = R and if

lim zf (z) = k (constant),

z→∞

then prove that Z

lim f (z) dz = i(β − α)k.
R→∞ AB

8
 Solution

On the circle |z| = R, write

z = Reiθ , α ≤ θ ≤ β.
Then
dz = iReiθ dθ.

Since
lim zf (z) = k,
z→∞

we have, for large |z|,

k
f (z) = + ε(z),
z
where ε(z) → 0 as |z| → ∞.
Thus on |z| = R,
k
f (z) = + ε(Reiθ ).
Reiθ
Now Z Z β  
k iθ
f (z) dz = + ε(Re ) iReiθ dθ.
AB α Reiθ
Hence Z Z β Z β
f (z) dz = ik dθ + iR ε(Reiθ )eiθ dθ.
AB α α

Since ε(z) → 0 as R → ∞, the second term tends to 0. Therefore,

Z
lim f (z) dz = ik(β − α).
R→∞ AB

Practice Problems
Z π
1. If limz→∞ zf (z) = 2, find lim f (Reiθ ) iReiθ dθ.
R→∞ 0

2. If limz→∞ zf (z) = 3i, find the limit over θ ∈ [0, 2π].

3. If limz→∞ zf (z) = 5, evaluate the limit on θ ∈ [−π/2, π/2].

4. If limz→∞ zf (z) = k, evaluate the limit on θ ∈ [0, π/3].
5. If limz→∞ zf (z) = −1, evaluate the limit on θ ∈ [π, 2π].

Practice Solutions

1. i(π − 0) · 2 = 2πi.
2. i(2π) · 3i = −6π.
3. i(π/2 − (−π/2)) · 5 = iπ · 5 = 5πi.
πik
4. i(π/3 − 0)k = 3 .

5. i(2π − π)(−1) = −πi.

9
Question 1(g)
Question: For what values of a, the equation

az n − ez = 0

has n roots inside |z| = 1?

Solution

Consider
az n − ez = 0.
Factor z:
z(az n−1 − e) = 0.
Hence one root is
z = 0,
which lies inside |z| < 1.
The remaining roots satisfy
e
az n−1 = e, z n−1 = .
a
Therefore
e
|z| n−1 = .
a
Hence 1
e n−1
|z| = .
a

All n − 1 roots of this equation will lie inside |z| < 1 if and only if
1
e n−1 e
<1 ⇐⇒ <1 ⇐⇒ |a| > e.
a a

Thus, for |a| > e, all n − 1 nonzero roots lie inside |z| < 1, together with z = 0, giving n roots in |z| < 1.
Hence the required condition is
|a| > e.

Practice Problems
1. For what values of a does az 5 − 2z = 0 have 5 roots in |z| < 1?

2. For what values of a does az 4 − 3z = 0 have all roots in |z| < 1?

3. For what values of a does az 6 − ez = 0 have 6 roots in |z| < 1?
4. For what values of a does az 3 − 5z = 0 have all roots in |z| < 1?

5. For what values of a does az n − cz = 0 have all n roots in |z| < 1?

10
Practice Solutions
2 2
1. z(az 4 − 2) = 0. Nonzero roots satisfy |z|4 = |a| . All inside |z| < 1 if |a| < 1 ⇒ |a| > 2.
3
2. z(az 3 − 3) = 0. Need |a| < 1 ⇒ |a| > 3.
e
3. z(az 5 − e) = 0. Need |a| < 1 ⇒ |a| > e.
5
4. z(az 2 − 5) = 0. Need |a| < 1 ⇒ |a| > 5.
|c|
5. z(az n−1 − c) = 0. All roots lie in |z| < 1 if < 1, i.e. |a| > |c|.
|a|

Question 2(a)
Question: Show that a power series of the form
∞
X
an (z − z0 )n
n=0

represents an analytic function S(z) at each point inside its circle of convergence.

Solution

Let
∞
X
S(z) = an (z − z0 )n
n=0

be a power series with radius of convergence R > 0. Then the series converges absolutely and uniformly on
every closed disk
|z − z0 | ≤ r < R.

To show that S(z) is analytic in |z − z0 | < R, we differentiate term by term.

Consider the series
∞
X
nan (z − z0 )n−1 .
n=1

By the ratio test, this series has the same radius of convergence R as the original series. Hence it converges
uniformly on |z − z0 | ≤ r < R.
Therefore
∞
X
S ′ (z) = nan (z − z0 )n−1
n=1

exists for all |z − z0 | < R.

Since S(z) has a complex derivative in this disk, S(z) is analytic at every point inside its circle of
convergence.

Practice Problems
zn
P∞
1. Show that n=0 n! is analytic in C.
P∞
2. Show that n=0 z n is analytic in |z| < 1.

11
 P∞ (z−2)n
3. Show that n=0 n+1 is analytic in its disk of convergence.
P∞
4. Show that n=0 nz n is analytic in |z| < 1.
P∞
5. Show that n=0 an (z − 1)n is analytic in |z − 1| < 1/|a|.

Practice Solutions

1. The radius of convergence is ∞. Termwise differentiation gives

∞ ∞
X z n ′ X z n−1
= ,
n=0
n! n=1
(n − 1)!

which converges everywhere, so the function is entire.

nz n−1 converges for |z| < 1, so the function
P
2. The radius of convergence is 1. The derivative series
is analytic there.
3. The radius of convergence is 1. The differentiated series converges in |z − 2| < 1, hence the function
is analytic there.
P 2 n−1
4. The radius of convergence is 1. The derivative series n z converges in |z| < 1, so the function
is analytic.
5. The radius of convergence is 1/|a|. The differentiated series converges in this disk, hence the function
is analytic there.

Question 2(b)
Question: Suppose that the power series
∞
X
f (z) = bn (z − z0 )n
n=0

can be differentiated term by term inside its circle of convergence |z − z0 | < R0 . Then show that
∞
X
f (n) (z) = k(k − 1) · · · (k − n + 1)bk (z − z0 )k−n , |z − z0 | < R0 .
k=n

Solution

Given
∞
X
f (z) = bk (z − z0 )k .
k=0

Differentiating term by term,

∞
X
′
f (z) = kbk (z − z0 )k−1 .
k=1

Differentiating again,
∞
X
f ′′ (z) = k(k − 1)bk (z − z0 )k−2 .
k=2

12
Proceeding inductively, after differentiating n times we obtain
∞
X
f (n) (z) = k(k − 1) · · · (k − n + 1)bk (z − z0 )k−n .
k=n

This is valid for all |z − z0 | < R0 .

Practice Problems
P∞
1. Find f ′′ (z) for f (z) = n=0 zn.
P∞ n
2. Find f (3) (z) for f (z) = n=0 zn! .
P∞
3. Find f ′ (z) for f (z) = n=0 (z − 1)n .
P∞
4. Find f ′′ (z) for f (z) = n=0 nz n .
P∞
5. Find f (n) (z) for f (z) = n=0 an (z − 2)n .

Practice Solutions

1.
∞
X 1 2
f (z) = zn = , f ′′ (z) = .
n=0
1−z (1 − z)3

2.
f (z) = ez , f (3) (z) = ez .

3.
∞
X ∞
X
f (z) = (z − 1)n , f ′ (z) = n(z − 1)n−1 .
n=0 n=1

4.
∞
X ∞
X
f (z) = nz n , f ′′ (z) = n(n − 1)z n−2 .
n=0 n=2

5.
∞
X
f (n) (z) = k(k − 1) · · · (k − n + 1)ak (z − 2)k−n .
k=n

Question 2(c)
Question: Show that the function f (z) defined by
 z
e − 1, z ̸= 0,
f (z) = z
1, z = 0,


is analytic.

13
Solution

Consider the Taylor series of ez about z = 0:

z2 z3
ez = 1 + z + + + ···
2! 3!
Then for z ̸= 0,
z2 3
ez − 1 z+ 2! + z3! + · · · z z2
= =1+ + + ···
z z 2! 3!
Hence
z z2 z3
f (z) = 1 + + + + ···
2! 3! 4!
for all z, including z = 0.
This is a power series with infinite radius of convergence, so it defines an analytic function in C. Therefore
f (z) is analytic everywhere.

Practice Problems
sin z
1. Show that f (z) = with f (0) = 1 is analytic.
z
ez − 1 − z
2. Show that f (z) = with f (0) = 12 is analytic.
z2
cos z − 1
3. Show that f (z) = with f (0) = − 12 is analytic.
z2
ez − 1
4. Show that f (z) = with f (0) = 1 is not analytic.
z2
sin z − z
5. Show that f (z) = with f (0) = − 16 is analytic.
z3

Practice Solutions
z3
1. sin z = z − 3! + · · · , hence
sin z z2
=1− + ···
z 3!
which is analytic.
z2
2. ez − 1 − z = 2! + · · · , dividing by z 2 gives a power series.
2
3. cos z − 1 = − z2! + · · · , dividing by z 2 gives a power series.
z2
4. ez − 1 = z + 2! + · · · , dividing by z 2 gives a pole at z = 0.
3
5. sin z − z = − z3! + · · · , dividing by z 3 gives a power series.

Question 3(a)
Question: Find all the bilinear transformations which transform the circle
|z| ≤ α
into the circle
|w| ≤ β
and verify it.

14
Solution

A bilinear (Möbius) transformation has the form

az + b
w= , ad − bc ̸= 0.
cz + d

Step 1. Map |z| ≤ α onto the unit disk |ζ| ≤ 1.

Define
z
ζ= .
α
Then
|z| ≤ α ⇐⇒ |ζ| ≤ 1.

Step 2. The general bilinear transformation mapping the unit disk onto itself is
ζ −a
ζ 7−→ eiθ , |a| < 1.
1 − aζ

Step 3. Map the unit disk onto |w| ≤ β by

w = βζ.

Combining all three mappings,

z
−a z − αa
w = βeiθ α z = βeiθ .
1−a α − az
α

Hence all bilinear transformations mapping |z| ≤ α onto |w| ≤ β are

z − αa
w = βeiθ , |a| < 1, θ ∈ R.
α − az

Verification

Let |z| ≤ α. Then

z
ζ= ⇒ |ζ| ≤ 1.
α
Since
ζ −a
1 − aζ
is an automorphism of the unit disk, it maps |ζ| ≤ 1 onto |η| ≤ 1. Hence

ζ −a
≤ 1.
1 − aζ

Multiplying by β,
ζ −a
|w| = β ≤ β.
1 − aζ
Therefore |z| ≤ α implies |w| ≤ β, and boundary maps to boundary. The transformation is verified.

15
Practice Problems
1. Find all bilinear maps sending |z| ≤ 2 onto |w| ≤ 5.
2. Find all bilinear maps sending |z| ≤ 1 onto |w| ≤ 3.

3. Find all bilinear maps sending |z| ≤ 4 onto |w| ≤ 1.

4. Find all bilinear maps sending |z| ≤ α onto |w| ≤ 1.
5. Find all bilinear maps sending |z| ≤ 1 onto |w| ≤ β.

Practice Solutions

1.
z − 2a
w = 5eiθ , |a| < 1.
2 − az

2.
z−a
w = 3eiθ , |a| < 1.
1 − az

3.
z − 4a
w = eiθ , |a| < 1.
4 − az

4.
z − αa
w = eiθ , |a| < 1.
α − az

5.
z−a
w = βeiθ , |a| < 1.
1 − az

Question 3(b)
Question: Find and sketch the images of
c1 < y < c2
under the transformation
w = z2,
where c1 and c2 are any real numbers.

Solution

Let
z = x + iy, w = u + iv = z 2 .
Then
w = (x + iy)2 = x2 − y 2 + 2ixy,
so
u = x2 − y 2 , v = 2xy.

The region c1 < y < c2 is the horizontal strip between the lines y = c1 and y = c2 in the z–plane.
First find the image of the boundary lines.

16
 For y = c (constant),
u = x 2 − c2 , v = 2cx.
Eliminate x:
v
x= .
2c
Substitute into u:  v 2 v2
u= − c2 = 2 − c2 .
2c 4c
Hence the image of the line y = c is the parabola
v2
u= − c2 .
4c2

Therefore, the lines y = c1 and y = c2 map into the parabolas

v2 v2
u= − c21 , u= − c22 .
4c21 4c22

Since c1 < y < c2 , the region maps into the region in the w–plane lying between these two parabolas.
Thus the image of the strip c1 < y < c2 under w = z 2 is the region bounded by the parabolas
v2 v2
u= − c21 and u= − c22 .
4c21 4c22

Practice Problems
1. Find the image of 0 < y < 1 under w = z 2 .
2. Find the image of −1 < y < 1 under w = z 2 .
3. Find the image of 1 < y < 2 under w = z 2 .
4. Find the image of y > 0 under w = z 2 .
5. Find the image of y < 0 under w = z 2 .

Practice Solutions

1. Bounded by
v2 v2
u= − 0, u= − 1,
0 4
v2
so the region lies between u = 4 − 1 and u = ∞.
2. Bounded by
v2 v2
u= −1 and u= − 1,
4 4
giving a symmetric region.
3. Bounded by
v2 v2
u= −1 and u= − 4.
4 16
4. Since y > 0, the image is the region to the right of the parabola
v2
u= .
4

17
 5. Since y < 0, the image is also the region to the right of the parabola

v2
u= .
4

Question 4(a)
Question: State and prove Casorati–Weierstrass theorem for essential singularity.

Solution

Statement (Casorati–Weierstrass Theorem). If z0 is an essential singularity of an analytic function

f (z), then in every neighborhood of z0 , the values of f (z) are dense in the complex plane C.
That is, for every ε > 0 and every complex number w, there exists z with

0 < |z − z0 | < ε

such that
|f (z) − w| < ε.

Proof.
Let z0 be an essential singularity of f (z). Suppose that the image of a punctured neighborhood of z0
does not fill the complex plane densely. Then there exists a complex number w0 and a neighborhood N (w0 )
such that
f (z) ̸= w0 for all 0 < |z − z0 | < r.
Define
1
g(z) = .
f (z) − w0
Then g(z) is analytic in 0 < |z − z0 | < r.
Since f (z) ̸= w0 there, g(z) is bounded near z0 . Hence by Riemann’s theorem on removable singularities,
g(z) has either a removable singularity or a pole at z0 . Therefore, f (z) has either a pole or removable
singularity at z0 , contradicting the assumption that z0 is an essential singularity.
Hence the values of f (z) are dense in C near z0 .

Practice Problems
1. Show that z = 0 is an essential singularity of e1/z .
2. Show that z = 0 is an essential singularity of sin(1/z).
3. Using Casorati–Weierstrass theorem, explain the behavior of e1/z near z = 0.

4. Determine the nature of z = 0 for f (z) = z1 e1/z .

1
5. Show that z = 0 is not an essential singularity of z2 .

18
Practice Solutions

1. e1/z has infinitely many negative powers in its Laurent expansion, so z = 0 is essential.
2. sin(1/z) has infinitely many negative powers, hence z = 0 is essential.
3. Near z = 0, e1/z takes values arbitrarily close to any complex number.
1 1/z
4. ze has infinitely many negative powers, so z = 0 is essential.
1
5. z2 has a pole of order 2, so it is not essential.

Question 4(b)
Question: Discuss the singularities of
ez−1
f (z) =
z(z − π)2
including infinity.

Solution

The function is
ez−1
f (z) = .
z(z − π)2

Finite singularities
The singular points occur when the denominator vanishes:

z = 0, z = π.

At z = 0: Since ez−1 is analytic and nonzero at z = 0, and the denominator has a simple zero at z = 0,

z=0

is a simple pole.
At z = π: The denominator has a factor (z − π)2 , while ez−1 is analytic and nonzero at z = π. Therefore

z=π

is a pole of order 2.
Singularity at infinity
To study z = ∞, consider
1
e w −1
 
1
F (w) = f =
2 .
w 1 1
−π
w w
Simplify:
1
e w −1 w3
F (w) = .
(1 − πw)2

As w → 0, the factor e1/w produces an essential singularity at w = 0. Hence z = ∞ is an essential

singularity of f (z).

z = 0 is a simple pole, z = π is a double pole, and z = ∞ is an essential singularity.

19
Practice Problems
ez
1. Find the singularities of .
z(z − 1)
1
2. Discuss the singularities of .
z 2 (z− 2)

e1/z
3. Discuss the singularities of .
z
sin z
4. Discuss the singularities of .
z3
1
5. Find the nature of infinity for .
z2 +1

Practice Solutions

1. Simple poles at z = 0 and z = 1, infinity is an essential singularity.

2. Pole of order 2 at z = 0, simple pole at z = 2, infinity is a removable singularity.
3. z = 0 is an essential singularity.
4. z = 0 is a pole of order 2.
5. Infinity is a pole of order 2.

Question 4(c)
Question: State and prove Jordan’s Lemma.

Solution

Statement (Jordan’s Lemma). Let f (z) be analytic in the upper half–plane and suppose that

M
|f (z)| ≤
|z|

for sufficiently large |z|. Then for a > 0,

Z
lim eiaz f (z) dz = 0,
R→∞ CR

where CR is the semicircle |z| = R, Im z ≥ 0.

Proof.
On CR , let z = Reiθ , 0 ≤ θ ≤ π. Then

|eiaz | = e−aR sin θ .

Hence Z Z π Z π
M
eiaz f (z) dz ≤ e−aR sin θ R dθ = M e−aR sin θ dθ.
CR 0 R 0

Since sin θ > 0 for 0 < θ < π, the integral tends to 0 as R → ∞. Hence Jordan’s lemma is proved.

20
Practice Problems
1. Verify Jordan’s lemma for f (z) = 1/z.
2. Verify Jordan’s lemma for f (z) = 1/(z 2 + 1).
R∞
3. Evaluate −∞ eiax dx using Jordan’s lemma.
R ∞ iax
4. Show that −∞ e x dx = πi for a > 0.
R ∞ iax
5. Evaluate −∞ xe2 +1 dx.

Practice Solutions

1. |1/z| ≤ 1/R, so the integral over CR tends to 0.

2. |1/(z 2 + 1)| ≤ 1/R2 , so Jordan’s lemma applies.
3. The integral is zero.
4. The integral equals πi.
5. The integral equals πe−a .

Question 5(a)
Question: State and prove the Argument Principle.

Solution

Statement (Argument Principle). Let f (z) be meromorphic inside and on a simple closed contour C,
and suppose f (z) has no zeros or poles on C. Let Z be the number of zeros and P the number of poles of
f (z) inside C, counted with multiplicities. Then
f ′ (z)
Z
1
dz = Z − P.
2πi C f (z)

Proof.
Let the zeros of f (z) inside C be a1 , . . . , am with multiplicities p1 , . . . , pm , and the poles be b1 , . . . , bn
with multiplicities q1 , . . . , qn .
Then
(z − a1 )p1 · · · (z − am )pm
f (z) = g(z),
(z − b1 )q1 · · · (z − bn )qn
where g(z) is analytic and nonzero inside C.
Taking logarithmic derivative,
m n
f ′ (z) X pk X qj g ′ (z)
= − + .
f (z) z − ak j=1 z − bj g(z)
k=1

Integrating around C,  
m n
f ′ (z)
Z X X
dz = 2πi  pk − qj  = 2πi(Z − P ).
C f (z) j=1
k=1

21
Practice Problems
1. Find the number of zeros of z 3 − 1 inside |z| = 2.
2. Find the number of zeros of z 4 + 1 inside |z| = 1.
3. Find the number of zeros of z 2 + z + 1 inside |z| = 1.
4. Find the number of zeros of ez − 1 inside |z| < 2.
z+1
5. Find the number of poles of z(z−1) inside |z| = 2.

Practice Solutions

1. All three zeros lie inside |z| = 2.

2. Two zeros lie inside |z| = 1.
3. All zeros lie on |z| = 1.
4. Only z = 0 is inside.
5. Poles at z = 0 and z = 1.

Question 5(b)
Question: Evaluate, by contour integration,
Z ∞
x sin(ax)
2 2
dx, c > 0, a ∈ R.
−∞ x + c

Solution

Consider the complex integral

∞
xeiax
Z
I= dx.
−∞ x2 + c 2
Then Z ∞
x sin(ax)
dx = Im(I).
−∞ x2 + c 2

Let
zeiaz
f (z) =.
z 2 + c2
For a > 0, close the contour in the upper half–plane. The poles of f (z) are at
z = ±ic.
Only z = ic lies inside the contour.
The residue at z = ic is
zeiaz ic e−ac 1
Res f (z) = lim (z − ic) = = e−ac .
z=ic z→ic (z − ic)(z + ic) 2ic 2

By the residue theorem,

∞
xeiax
Z
1
dx = 2πi · e−ac = πie−ac .
−∞ x2 + c 2 2

22
Taking imaginary parts, Z ∞
x sin(ax)
2 + c2
dx = πe−ac , a > 0.
−∞ x
For a < 0, the integral changes sign, hence
Z ∞
x sin(ax)
2 2
dx = πe−c|a| sgn(a).
−∞ x + c

Practice Problems
Z ∞
x sin 2x
1. dx
−∞ x2 + 1
Z ∞
x sin 3x
2. dx
−∞ x2 + 4
Z ∞
x sin x
3. dx
−∞ x2 + 9
Z ∞
x sin 5x
4. dx
−∞ x2 + 1
Z ∞
x sin ax
5. dx
−∞ x2 + 4

Practice Solutions

1.
πe−2 .

2.
πe−6 .

3.
πe−3 .

4.
πe−5 .

5.
πe−2|a| sgn(a).

Question 6(a)
Question: Consider the function f (z) defined on the disk
D = {z : |z − 1| < 1}
by the series expansion
1 1
f (z) = z 1/2 = 1 + (z − 1) − (z − 1)2 + · · · .
2 8
Show that f (z) has an analytic continuation around the circle
z(t) = eit , 0 < t < 2π.
Find the expression for ft (z).

23
Solution
√
The given series is the binomial expansion of z 1/2 about z = 1. It represents the principal branch of z in
the disk |z − 1| < 1, which does not include the negative real axis.
Let
z(t) = eit , 0 < t < 2π.
As t increases from 0 to 2π, z(t) makes one full turn around the origin.
The square root function can be written as
1
z 1/2 = e 2 log z ,
where log z = ln |z| + i arg z. Along the path z(t) = eit , we have
log z = it.
Hence
1
ft (z) = e 2 it = eit/2 .

After one full turn t = 2π, the value becomes

f2π (z) = eiπ = −1.
√
Thus the analytic continuation of z changes sign after one full circuit.
Hence
ft (z) = eit/2 .

Practice Problems
1. Find the analytic continuation of z 1/3 around z = eit .
√
2. Find the change in z after two turns around the origin.
3. Find ft (z) for z 1/4 .
4. Find the branch of log z along z = eit .
√
5. Find the monodromy of z.

Practice Solutions

1. ft (z) = eit/3 .
2. After two turns t = 4π, e2πi = 1, so it returns to the original value.
3. ft (z) = eit/4 .
4. log z = it.
√
5. After one turn, z changes sign.

Question 6(b)
Question: Using contour integration, evaluate
∞
x2
Z
dx.
0 1 + x6

24
Solution

Consider ∞
x2
Z
I= dx.
0 1 + x6
Write ∞
x2
Z
dx = 2I.
−∞ 1 + x6

Let
z2
f (z) = .
1 + z6
The poles occur when
1 + z6 = 0 ⇒ z 6 = −1 = eiπ .
Thus
zk = ei(π+2kπ)/6 , k = 0, 1, 2, 3, 4, 5.
The poles in the upper half–plane correspond to

z0 = eiπ/6 , z1 = eiπ/2 , z2 = ei5π/6 .

These are simple poles. The residue at each pole zk is

zk2 1
Res f (z) = 5 = 3.
z=zk 6zk 6zk

Hence  
X 1 1 1 1
Res f (z) = 3 + 3+ 3 .
upper half–plane 6 z0 z1 z2
Now
z03 = eiπ/2 = i, z13 = ei3π/2 = −i, z23 = ei5π/2 = i.
Thus
1 1 1 1 1 1
+ 3+ 3 = + + = −i + i − i = −i.
z03 z1 z2 i −i i
Hence X −i
Res = .
6
By the residue theorem,
∞
x2
 
−i
Z
π
dx = 2πi · = .
−∞ 1 + x6 6 3
Therefore
1 π π
I= · = .
2 3 6

Practice Problems
Z ∞
dx
1.
0 1 + x4
∞
x2
Z
2. dx
0 1 + x4
Z ∞
dx
3.
0 1 + x6

25
 ∞
x4
Z
4. dx
0 1 + x6
∞
x2
Z
5. dx
0 1 + x8

Practice Solutions
π
1. √ .
2
π
2. √ .
2 2
π
3. .
3
π
4. .
6
π
5. √ .
4 2

Question 7(a)
Question: State and prove Schwarz’s reflection principle for analytic continuation.

Solution

Statement (Schwarz Reflection Principle). Let f (z) be analytic in a domain D lying in the upper
half–plane and continuous on the real axis. If f (x) is real for all real x in D, then the function
(
f (z), Im z ≥ 0,
F (z) =
f (z), Im z < 0,

is analytic in the reflected domain.

Proof.
Define (
f (z), Im z ≥ 0,
F (z) =
f (z), Im z < 0.
Since f (x) is real on the real axis,
f (x) = f (x).
Hence F (z) is continuous across the real axis.
In the upper half–plane, F (z) = f (z) is analytic. In the lower half–plane, F (z) = f (z) is analytic.
Therefore F (z) is analytic everywhere in the reflected domain.

Practice Problems
1. Extend f (z) = z 2 from the upper half–plane.
2. Extend f (z) = ez from the upper half–plane.
3. Extend f (z) = sin z from the upper half–plane.

26
 4. Extend f (z) = z 3 from the upper half–plane.
5. Extend f (z) = log z along the real axis.

Practice Solutions

1. F (z) = z 2 .
2. F (z) = ez .
3. F (z) = sin z.
4. F (z) = z 3 .
5. F (z) = log |z| + i arg z.

Question 7(b)
Question: Find and sketch the branch points and branch lines for the function
log z 2 − 1 .

Solution

Consider
f (z) = log(z 2 − 1).
The logarithm is multivalued whenever its argument is zero or crosses the negative real axis.
First find where
z2 − 1 = 0 ⇒ z = ±1.
Hence
z = 1, z = −1
are branch points.
To define a single–valued branch, we introduce branch cuts so that z 2 − 1 does not encircle the origin.
The real values of z 2 − 1 are negative when
−1 < z < 1.
Hence the standard branch cut is taken along the interval
[−1, 1]
on the real axis.
Thus the branch points are at z = ±1 and the branch line is the segment joining them.

Practice Problems
1. Find branch points of log(z − 2).
2. Find branch points of log(z 2 ).
√
3. Find branch cuts for z 2 − 4.
4. Find branch points of log(z 3 − 1).
√
5. Find branch points of z 2 + 1.

27
Practice Solutions

1. z = 2.
2. z = 0.
3. Branch points at z = ±2.
4. Roots of z 3 = 1.
5. Branch points at z = ±i.

28