MATH 381

HOMEWORK 2 SOLUTIONS

Question 1 (p.86 #8). If g(x)[e2y − e2y ] is harmonic, g(0) = 0, g ′ (0) = 1, find g(x).
Solution. Let f (x, y) = g(x)[e2y − e2y ]. Then
∂2f
= g ′′ (x)[e2y − e2y ]
∂x2
∂2f
= 4g(x)[e2y − e2y ].
∂y 2
Since f (x, y) is harmonic, ∂2f
∂x2
+ ∂2f
∂y 2
= 0 and we require
g ′′ (x) + 4g(x) = 0.
Thus g(x) has the form A sin(2x) + B cos(2x) and by the initial conditions, A = 1�2 and B = 0. Therefore,

g(x) = sin(2x).
1
2
�

Question 2 (p.86 #12). Find the harmonic conjugate of tan−1 � xy � where −π < tan−1 � xy � ≤ π.

Solution. Write u(x, y) = tan−1 � xy �. Then by the Cauchy-Riemann equations,

y2 1
= 2 = =
∂u y ∂v
∂x x + y 2 y x2 + y 2 ∂y
(1)

y 2 −x
− =− 2 = =
∂u x ∂v
x + y 2 y 2 x2 + y 2 ∂x
(2) .
∂y
By (1),
v= log(x2 + y 2 ) + C(x),
1
2
and by (2)
= + C ′ (x) = 2
∂v x x
∂x x2 + y 2 x + y2
so C ′ (x) = 0 and C(x) is a constant, call it D. Therefore,

v(x, y) = log(x2 + y 2 ) + D.
1
2
�

Question 3. (p.86 #13) Show, if u(x, y) and v(x, y) are harmonic functions, that u + v must be a harmonic
function but that uv need not be a harmonic function. Is eu ev a harmonic function?
Solution. If u and v are harmonic, then u + v is harmonic since
∂ 2 (u + v) ∂ 2 (u + v) ∂2u ∂2v ∂2u ∂2v
2
+ 2
=� 2 + 2
�+� 2 + 2�
∂x ∂y ∂x ∂x ∂y ∂y
∂2u ∂2u ∂2v ∂2v
=� + � + � + � = 0.
∂x2 ∂y 2 ∂x2 ∂y 2
To show that uv is not necessarily harmonic, it suffices to show that there exists u, v harmonic such that
1 ∂ 2 (uv) ∂ 2 (uv)
� + �= + ≠ 0.
∂u ∂v ∂u ∂v
2 ∂x2 ∂y 2 ∂x ∂x ∂y ∂y
Any u = v harmonic where ∂u , ∂u ≠ 0 will suffice. For instance, taking u = v = x will work, since it’s harmonic (both
∂x ∂y
of its second-order partials vanish) but
+ = 12 ≠ 0.
∂u ∂v ∂u ∂v
∂x ∂x ∂y ∂y

Date: October 23, 2011.

1
2 MATH 381 HOMEWORK 2 SOLUTIONS

Now, in order for eu ev to be harmonic, we need

∂ 2 (eu ev ) ∂ 2 (eu ev ) ∂u ∂v 2 ∂u ∂v 2
2
+ 2
= eu+v �� + � +� + � � = 0.
∂x ∂y ∂x ∂x ∂y ∂y

Thus, the existence of any u, v harmonic such that � ∂u + � + � ∂u + � ≠ 0 will show that eu ev is not harmonic.
2
∂v 2 ∂v
∂x ∂x ∂y ∂y
Again, taking u = v = x gives us what we want as e 2x
is easily seen to be non-harmonic. �

Question 4 (p.106 #14). State the domain of analyticity of f (z) = eiz . Find the real and imaginary parts
u(x, y) and v(x, y) of the function, show that these satisfy the Cauchy-Riemann equations, and find f ′ (z) in
terms of z.
Solution. By definition,
f (z) = eiz = eix e−y = e−y [cos x + i sin x].
Therefore,
u(x, y) = e−y cos x
v(x, y) = e−y sin x.
These are continuous functions at all (x, y) ∈ R2 . Now,

= −e−y sin x =
∂u ∂v
∂x ∂y
= −e−y cos x = −
∂u ∂v
∂y ∂x
so u, v satisfy the C-R equations, and these derivatives are continuous for allx, y. Therefore, f (z) is entire. Further-
more,
f ′ (z) = −e−y sin x + i(e−y cos x) = i(e−y [cos x + i sin x]) = ieiz .
�

Question 5 (p.106 #16). State the domain of analyticity of f (z) = ee . Find the real and imaginary parts
z

u(x, y) and v(x, y) of the function, show that these satisfy the Cauchy-Riemann equations, and find f ′ (z) in
terms of z.
Solution. First, observe that f is an entire function of an entire function, so it is analytic everywhere. Now,
ee = ee (cos y+i sin y)
= ee � cos(ex sin y) + i sin(ex sin y)�,
z x x
cos y

so
u(x, y) = ee
x
cos y
cos(ex sin y)
v(x, y) = ee
x
cos y
sin(ex sin y)

= ee cos y (ex cos y) cos(ex sin y) − ee cos y (ex sin y) sin(ex sin y)
∂u x x

∂x
= ee cos y+x (cos y cos(ex sin y) − sin y sin(ex sin y))
x

= ee cos y (−ex sin y) sin(ex sin y) + ee cos y cos(ex sin y)(ex cos y)
∂v x x

∂y
= ee (cos y cos(ex sin y) − sin y sin(ex sin y))
x
cos y+x

= ee cos y (−ex sin y) cos(ex sin y) + ee cos y (ex cos y)(− sin(ex sin y))
∂u x x

∂y
= −ee (cos y sin(ex sin y) + sin y cos(ex sin y))
x
cos y+x

= ee cos y (ex cos y) sin(ex sin y) + ee cos y (ex sin y) cos(ex sin y)
∂v x x

∂x
= ee cos y+x (cos y sin(ex sin y) + sin y cos(ex sin y))
x

and f satisfies the C-R equations. Furthermore,

f ′ (z) = ee e �(cos y cos(ex sin y) − sin y sin(ex sin y)) + i(cos y sin(ex sin y) + sin y cos(ex sin y)�
x
cos y x

= ee �ex cos y� cos(ex sin y) + i sin(ex sin y)� + ex sin y�i cos(ex sin y) − sin(ex sin y)��
x
cos y

= ee � cos(ex sin y) + i sin(ex sin y)�ex (cos y + i sin y)

x
cos y

= ee ez .
z

�
 MATH 381 HOMEWORK 2 SOLUTIONS 3

Question 6 (p.106 #23).

(a) Prove the expression given in the text for the nth derivative of f (t) = t2t+1 = Re � t−i
1
�. (Note: t ∈ R).
(b) Find similar expressions for the n derivative of f (t) = t2 +1 = Im � t−i �.(Note: t ∈ R ).
th 1 1

Solution.
(a) By the Lemma, for n ≥ 1,

dn (−1)n n!
f (n) (t) = Re � � = � �
1
dtn t − i (t − i)n+1
Re

Now, observe that 1

t−i
= t+i
t2 +1
, so by the binomial theorem

(−1)n n! (−1)n n!(t + i)n+1 (−1)n n!(n + 1)! n+1 ik tn+1−k

= = �
(t − i)n+1 (t2 + 1)n+1 (t2 + 1)n+1 k=0 (n + 1 − k)!k!
.

But notice that we only get contributions to the real part of this expression when k is even; i.e. when ik ∈ R.
Summing over the even integers, k = 2m, we get for n odd that

(−1)n n!(n + 1)! 2

n+1
i2m tn+1−2m
f (n) (t) = �
(t + 1)
2 n+1
m=0 (n + 1 − 2m)!(2m)!

(−1)n!(n + 1)! 2 (−1)m tn+1−2m

n+1

= �
(t2 + 1)n+1 m=0 (n + 1 − 2m)!(2m)!

and for n even that

n!(n + 1)! 2
n
i2m tn+1−2m
f (n) (t) = �
(t + 1)
2 n+1
m=0 (n + 1 − 2m)!(2m)!

(b) In this case we want

dn (−1)n n!
f (n) (t) = Im � � = � �.
1
dtn t − i (t − i)n+1
Im

By the work above, we want the imaginary part of

(−1)n n!(n + 1)! n+1 ik tn+1−k

�
(t2 + 1)n+1 k=0 (n + 1 − k)!k!
.

In this case we get contributions when k is odd, so we take the the sum over k = 2m + 1 for m ≥ 0. Note that
i2m+1 = (−1)m i. It follows that when n is odd,

(−1)n!(n + 1)! 2 (−1)m tn−2m

n−1

f (n) (t) = �
(t + 1)
2 n+1
m=0 (n − 2m)!(2m + 1)!

and when n is even,

n!(n + 1)! n�2 (−1)m tn−2m
f (n) (t) = �
(t2 + 1)n+1 m=0 (n − 2m)!(2m + 1)!
.

Question 7 (p.106 #25). Let P (ψ) = ∑N −1 inψ

n=0 e .
(a) Show that

�P (ψ)� = � �.
sin(N ψ�2)
sin(ψ�2)
(b) Find limψ→0 �P (ψ)�.
(c) Plot �P (ψ)� for 0 ≤ ψ ≤ 2 and N = 3.
Solution.
(a) Note that
1 − eiN ψ
P (ψ) =
1 − eiψ
.
4 MATH 381 HOMEWORK 2 SOLUTIONS

Thus,
eiN ψ − 1
P (ψ) =
eiψ − 1
eiN ψ�2 eiN ψ�2 − e−iN ψ�2
= iψ�2 � iψ�2 �
e e − e−iψ�2
eiN ψ�2 cos(N ψ�2) + i sin(N ψ�2) − cos(−N ψ�2) − i sin(−N ψ�2)
= � �
eiψ�2 cos(ψ�2) + i sin(ψ�2) − cos(−ψ�2) − i sin(−ψ�2)
eiN ψ�2 2i sin(N ψ�2)
= .
eiψ�2 2i sin ψ�2
Thus,
eiN ψ�2 sin(N ψ�2)
�P (ψ)� = � �� �=� �.
sin(N ψ�2)
eiψ�2 sin ψ�2 sin ψ�2
(b) By l’Hopital’s rule we get
N �2 sin(N ψ�2)
= lim = N.
sin(N ψ�2)
lim
ψ→0 sin ψ�2 ψ→0 1�2 sin ψ�2
(c) If you have nothing else, just plug it in Wolfram Alpha.
�

Question 8 (p.112 #17). Show that sin z − cos z = 0 has solutions only for real values of z. What are the
solutions?
Solution. In other words, for z = x + iy we want
sin x cosh y + i cos x sinh y = cos x cosh y − i sin x sinh y.
Equating the real parts and imaginary parts we require
(3) sin x cosh y = cos x cosh y
(4) cos x sinh y = − sin x sinh y.
Suppose y ≠ 0 and hence sinh y ≠ 0 and cosh y ≠ 0. Then in order to have solutions, by (3), we need cos x = sin x and
by (4) we need cos x = − sin x. These equations are only satisfied for sin x = cos x = 0, but no solutions for x exists.
Therefore, if there are solutions to the original equation, we must have that y = 0.
Suppose y = 0. Then since cosh 0 = 1 and sinh 0 = 0 we simply need solutions to sin x = cos x. Thus we have
solutions if and only if
z = + kπ, k ∈ Z.
π
4
�

Question 9 (p.112 #21). Where does the function f (z) = √ 1

3 sin z−cos z
fail to be analytic?
√
Solution. Since sin z and cos z are both analytic, f (z) will fail to be analytic when 3 sin z −cos z = 0. In other words,
when we have solutions to
�
( sin x cosh y + i cos x sinh y) = cos x cosh y − i sin x sinh y.
Equating the real parts and imaginary parts we require
√
(5) 3 sin x cosh y = cos x cosh y
√
(6) 3 cos x sinh y = − sin x sinh y.
By the same argument as the previous question, there √ are no solutions when y ≠ 0. Suppose y = 0. Then since
cosh 0 = 1 and sinh 0 = 0 we simply need solutions to 3 sin x = cos x, that is to tan x = √13 . So f (z) is not analytic
when
z = + kπ, k ∈ Z.
π
6
�

Question 10 (p.112 #22). Let f (z) = sin � z1 �.

(a) Express this function in the form u(x, y) + iv(x, y). Where in the complex plane is this function
analytic?
(b) What is the derivative of f (z)? Where in the complex plane is f ′ (z) analytic?
Solution.
 MATH 381 HOMEWORK 2 SOLUTIONS 5

(a) Since sin z is entire, and 1

z
is analytic for z ≠ 0, it follows that f (z) is analytic for z ≠ 0.
x − iy
sin � � = sin � 2 �
1
z x + y2
−y −y
= sin � 2 � cosh � 2 � + i cos � 2 � sinh � 2 �
x x
x + y2 x + y2 x + y2 x + y2
= sin � 2 � cosh � 2 � − i cos � 2 � sinh � 2 �.
x y x y
x + y2 x + y2 x + y2 x + y2
(b) For z ≠ 0,
−1
sin � � = �cos � � 2 �
d 1 1
dz z z z
which is analytic for all z ≠ 0.
�
�
Question 11 (p.112 #25). Show that �cos z� = sinh2 y + cos2 x.
Solution.
�cos z� = �cos x cosh y − i sin x sinh y�
�
= cos2 x cosh2 y + sin2 x sinh2 y
�
= cos2 x(1 + sinh2 y) + sin2 x sinh2 y
�
= cos2 x + sinh2 y(cos x2 + sin2 x)
�
= cos2 x + sinh2 y
�

Question 12 (p.119 #16). Use logarithms to find solutions to ez = eiz .

Solution. We want solutions to ez(1−i) = 1, so taking logs on both sides we get for any k ∈ Z, z(1 − i) = 2πik, so
2πik (i + 1)2πik
z= = = (i − 1)kπ.
1−i 2
�

Question 13 (p.119 #18). Use logarithms to find solutions to ez = (ez − 1)2 .

Solution. In other words, we want solutions to e2z − 3ez + 1 = 0. By the quadratic formula, we get that
√ √
3± 4−9 3
e = = ±
z 5
.
2 2 2
√
Taking logs gives that

z = log � ± � + 2πik
3 5
2 2
for k ∈ Z. �

Question 14 (p.119 #21). Use logarithms to find solutions to ee = 1.

z

Solution. First, taking logs we get ez = 2πik for k ∈ Z. Now for k > 0, the argument of 2πik is π
+ 2πm where m ∈ Z,
and for k < 0, the argument of 2πik is + 2πm (again m ∈ Z). Thus, for k > 0,
2
3π
2

z = log(2πk) + i � + 2mπ�
π
2
and for k < 0,
−π
z = log(2πk) + i � + 2mπ� .
2
�

Question 15 (p.119 #23). Show that

Re �log(1 + eiθ )� = log �2 cos � ��

θ
2
where θ ∈ R and eiθ ≠ −1.
6 MATH 381 HOMEWORK 2 SOLUTIONS

Solution.
Re �log(1 + eiθ )� = log �1 + eiθ �

= log �(1 + cos θ)2 + sin2 θ�

1
2
= log(2 + 2 cos θ)
1
2
= log �2 cos2 � � + 2 sin2 � � + 2 cos2 � � − 2 sin2 � ��
1 θ θ θ θ
2 2 2 2 2
= log �2 cos � ��
θ
2
�

Question 16 (p.170 #9). Intergrate

−1 1
� dz
1 z
along �z� = 1, in the lower half plane.
Solution. Let z = eit , then we are integrating along the interval t ∈ [0, −π]. Now, dz = ieit dt so
−1 −π 1
dz = � iei tdt = −iπ.
1
� z ei t
1 0

Question 17 (p.170 #11). Show that x = 2 cos t, y = sin t, where t ranges from 0 to 2π, yields a parametric
representation of the ellipse x4 + y 2 = 1. Use this representation to evaluate ∫2 z̄dz along the portion of the
2 i

ellipse in the first quadrant.

Solution. Note that
(2 cos t)2
+ sin2 t = cos2 t + sin2 t = 1
4
and furthermore 2 cos 0 = 2 cos 2π = 2 and sin 0 = sin 2π = 0. To see that we get all of the ellipse, note that x = 2 cos t
has solutions t ∈ [0, 2π] for all x ∈ [−2, 2] and y = sin t has solutions t ∈ [0, 2π] for all y ∈ [−1, 1]. Furthermore, the
parametrization is 1:1 except for when x = 2, y = 0.
Setting z = x + iy = 2 cos t + i sin t, we get dz = (i cos t − 2 sin t)dt, and
π

z̄dz = � (2 cos t − i sin t)(i cos t − 2 sin t)dt

i
�
2

2 0
π

=� (2i − 3 sin t cos t)dt

2

= − + iπ.
3
2
�

Question 18 (p.170 #14). Consider I = ∫0 e dz taken along the line x = 2y. Without actually doing the
2+i z 2
√
integration, show that �I� ≤ 5e3 .

Solution. Let M be the maximal value attained by �ez � along the path of integration. Now, for x = 2y,
2

�ez � = �ex −y 2 +2ixy

� = e3y
2 2 2

which attains its maximum when y √ √

attains a maximum—that is, when z = 2+i. Therefore
√ M3 = e . By the pythagorean
3

theorem, the length of the path is 22 + 12 = 5, so by the ML inequality, �I� ≤ M L = 5e . �

Question 19 (p.170 #16). Consider I = ∫i ei log z̄ dz taken along the parabola y = 1 − x2 . Without doing the
1

integration, show that �I� ≤ 1.479eπ�2 .

Solution. Letting θ = arg z
�ei log z̄ � = �ei(log�z�−iθ) �

= �ei log�z� eθ � = eθ .

Along the given path, this attains a maximum when θ = π�2, so let M = eπ�2 .
 MATH 381 HOMEWORK 2 SOLUTIONS 7

Now, we need to find the length of the path of integration. So since dy = −2xdx,
�
dy 2
L=� 1+� � dx
1

dx
1√
0

=� 1 + 4x2 dx
0
< 1.479.
The ML inequality then gives the desired result. �

Question 20 (p.180 #2). Is the Cauchy-Goursat theorem directly applicable to ��z�=1 sin z
z+2i
dz?
Solution. Since sin z
z+2i
is analytic everywhere except for z = −2i which is not in the unit circle, the C-G theorem is
directly applicable. �

Question 21 (p.180 #6). Is the Cauchy-Goursat theorem directly applicable to ��z−i−1�=1 log zdz?
Solution. Since 0 is not in the unit circle about i + 1, log z is analytic in the desired region so the C-G theorem is
directly applicable. �

��z�=1�2 (z−1)4 +1 dz?

1
Question 22 (p.180 #7). Is the Cauchy-Goursat theorem directly applicable to

Solution. Observe that we have a singularity when z − 1 is a primitive 8th root of unity—that is, when (z − 1)4 = −1.
These roots of unity lie on the unit circle, so shifting over by 1, we need to determine if the roots closest to the origin,
at z = eiπ3�4 + 1 and z = eiπ5�4 + 1 have absolute value greater than 1�2. By geometry (right angle triangles), it can be
seen that at these points, �
�
�z� = ��1 − √ � + � √ � > 1�2
2 2
1 1
2 2
so the C-G theorem directly applies. �

Question 23 (p.180 #9). Is the Cauchy-Goursat theorem directly applicable to

1
�
�z�=b z + bz + 1
2
dz

where 0 < b < 1?

Solution. In this case the singularities are at the roots of the equation x2 + bx + 1, that is, when
√ √
−b ± b2 − 4 −b 4 − b2
z= = ±i .
2 2 2
Here, �
b2 4 − b2
�z� = + =1>b
4 4
therefore C-G applies directly. �
Question 24 (p.180 #13). Prove that

ecos θ � sin(sin θ + θ)�dθ = 0.

2π
�
0
Begin with� ez dz performed around �z� = 1. Use the parametric representation z = eiθ , 0 ≤ θ ≤ 2π. Separate
your equation into real and imaginary parts.
Solution. Let z = eiθ = cos θ + i sin θ, so dz = eiθ idθ. Since ez is analytic,
ez dz = � ecos θ+i sin θ eiθ idθ = 0.
2π
�
�z�=1 0

But then
ecos θ+i(sin θ+θ) dθ = � ecos θ (cos(θ + sin θ) + i sin(θ + sin θ))dθ = 0
2π 2π
�
0 0
so by equating the imaginary part with zero we get the desired result. �