THE UNIVERSITY OF HONG KONG

DEPARTMENT OF MATHEMATICS

MATH1013 University Mathematics II

Tutorial 3 Solutions

1. (a) By the binomial expansion (or direct expansion), we have

(cos θ + i sin θ)5

= cos5 θ + 5 cos4 θ(i sin θ) + 10 cos3 θ(i sin θ)2 + 10 cos2 θ(i sin θ)3 + 5 cos θ(i sin θ)4
+ (i sin θ)5
= (cos5 θ − 10 cos3 θ sin2 θ + 5 cos θ sin4 θ) + i(5 cos4 θ sin θ − 10 cos2 θ sin3 θ + sin5 θ).

(b) By De Moivre’s theorem, we have (cos θ + i sin θ)5 = cos 5θ + i sin 5θ. We compare the real
parts of this and the expression found in part (a). This gives

cos 5θ = cos5 θ − 10 cos3 θ sin2 θ + 5 cos θ sin4 θ

= cos5 θ − 10 cos3 θ(1 − cos2 θ) + 5 cos θ(1 − cos2 θ)2
= 16 cos5 θ − 20 cos3 θ + 5 cos θ.

Remark. Similarly, by comparing the imaginary parts, we have

sin 5θ = 5 cos4 θ sin θ − 10 cos2 θ sin3 θ + sin5 θ

= 5(1 − sin2 θ)2 sin θ − 10(1 − sin2 θ) sin3 θ + sin5 θ
= 16 sin5 θ − 20 sin3 θ + 5 sin θ.

Using the same idea, one can show that cos nθ can be expressed as a polynomial of cos θ,
called the Chebyshev polynomial of the first kind. The Chebyshev polynomial of the
second kind is related to the expansion of sin nθ.

2. (a) Let z = r cis ϕ. Then z cis θ = r cis (ϕ + θ) by proposition 2.4. Therefore, z cis θ has the
same distance as z to 0. Also, the angle measured from the segment joining 0 and z to
the segment joining 0 and z cis θ is (ϕ + θ) − ϕ = θ. This shows z cis θ is the image under
the rotation.
Im

Im z′ − w

z cis θ z θ
Re
ϕ+θ ′ 0 z−w
z
ϕ
Re θ
0
w z

1
 (b) Let z ′ be the resultant image. We perform a translation by the complex number −w
(which means adding −w to every number). The centre of rotation becomes w − w = 0,
while z becomes z − w. We rotate z − w about 0 by θ radians to obtain (z − w) cis θ
by part (a). This is equal to z ′ − w according to the translation. Therefore, we have
z ′ = (z − w) cis θ + w.

(c) Note that the length of the segment joining a and b is the same as the length of the segment
π
joining b and c. Also, the interior angle at b is . We first consider the configuration as
3
π
shown. In this case, a is obtained from c by a rotation of radians (anticlockwise) about
3
centre b. By part (b), we have
√ √ √
π 1 3 1 − 3i 1 + 3i
a = (c − b) cis + b = (c − b) ( + i) + b = b+ c.
3 2 2 2 2
√ √
−1 + 3i 2 −1 − 3i
Let ω = . Then ω = , so that the above relation becomes a+ωb+ω 2 c = 0.
2 2
Lastly, we check that ω 3 = 1, which means ω is a cube root of unity by definition.

a c

π
3

b
Next, there is another configuration where b and c are interchanged.
√ In this case, we
√
1 + 3i 1 − 3i
simply swap b and c in the above steps, and obtain a = b+ c. This time
√ √ 2 2
−1 − 3i −1 + 3i
we define ω = . Then we check that ω 2 = and ω 3 = 1. Again, ω is a
2 2
cube root of unity, and we have a + ωb + ω 2 c = 0.

3. (a) Let z = x + yi. Then the condition means ∣x∣ + ∣y∣ ⩽ 5. For the first quadrant, we have
x + y ⩽ 5. This is the region lying below the straight line x + y = 5. Therefore, it is
the triangle with vertices 0, 5, 5i. The parts in the other quadrants can be obtained by
reflecting this triangle about the imaginary axis and the real axis (which corresponds to
flipping the signs of x and y). Therefore, the resultant figure is a square with vertices
±5, ±5i.

Im
5i

Re
−5 5

−5i

2
 z−1
(b) Recall arg = arg (z − 1) − arg (i − 1) (up to a difference of 2π). Analogous to example
i−1
2.7, arg (z − 1) measures the angle from the positive real axis to the line joining z and 1,
while arg (i − 1) measures the angle from the positive real axis to the line joining i and 1.
Thus, arg (z − 1) − arg (i − 1) measures the angle from the line joining i and 1 to the line
π
joining z and 1. In other words, z can be any point such that this angle is . This gives
3
a ray with endpoint 1 as shown. The point 1 may be included or excluded.

Im

i
π
3
Re
1
z

(c) Firstly, ∣z − 1 − i∣ = 1 represents the circle Γ with centre 1 + i and radius 1. Secondly,
multiplying z by a complex number r cis θ means to multiply its modulus by r and rotate
√ π √
it about 0 by θ radians. As 1 + i = 2 cis , we need to enlarge Γ by a ratio of 2 and
4
π
rotate it about 0 by radians. Clearly, the result is still a circle. The new centre is
4 √ √
(1 + i)(1 + i) = 2i, and the new radius is (1)( 2) = 2.

Im

Re

4. (a) Let z1 and z2 be two distinct points on the straight line. Then the line is the locus of
z − z1
point z such that z − z1 is a real multiple of z2 − z1 . This means ∈ R, and hence
z2 − z1
z − z1 z − z1
( )= . Now,
z2 − z1 z2 − z1

z − z1 z − z1
( )=
z2 − z1 z2 − z1
z − z1 z − z1
⇔ =
z2 − z1 z2 − z1
⇔ (z2 − z1 )(z − z1 ) = (z2 − z1 )(z − z1 )
⇔ (z1 − z2 )z + (z2 − z1 )z + (z1 z2 − z2 z1 ) = 0.

This gives an equation of the desired form because z1 and z2 are fixed constants.

3
 (b) No. For example, consider the equation z + 2z + 1 = 0. Let z = x + yi. The equation
becomes x + yi + 2(x − yi) + 1 = 0, i.e. 3x + 1 − yi = 0. By comparing the real parts and
1
the imaginary parts, we must have 3x + 1 = 0 and −y = 0. This gives x = − and y = 0.
3
Therefore, the equation represents a single point instead of a straight line.

Additional problems:

5. (a) Note that

ei(2−i) = e1+2i = e ⋅ e2i = e cis 2.
Therefore, we have

ei(2−i) cis (π − 2) = e cis (2 + π − 2) = e cis π = −e.

√ √√ √
2 −1
(b) Let 3 − i = r cis θ. Then we have r =3 + (−1)2 = 2 and tan θ = √ . Since 3 − i
3
π
lies in the fourth quadrant of the Argand diagram, we have θ = tan−1 √ = − . By De
−1
3 6
Moivre’s theorem, we obtain
√ π 100 100π 4π √
( 3 − i)100 = (2 cis (− )) = 2100 cis (− ) = 2100 cis = −299 (1 + 3i).
6 6 3

√ √
(c) Let 1 + i = r cis θ. Then we have r = (1)2 + (1)2 =
2 and tan θ = 1. Since 1 + i lies in the
π
first quadrant of the Argand diagram, we have θ = tan−1 1 = . By De Moivre’s theorem,
4
we obtain
1000
√ π
(1 + i)1000 = ( 2 cis ) = 2500 cis 250π = 2500 .
4
By the hint (with r = 1 + i), we deduce

(1 + i)1000 − 1
1 + (1 + i) + (1 + i)2 + ⋯ + (1 + i)999 =
(1 + i) − 1
2500 − 1
=
i
= −(2500 − 1)i.

Remark. The identity in the hint can be proved by verifying

(1 + r + r2 + ⋯ + rn−1 )(r − 1) = rn − 1

via expansion.

(d) Using the hint of part (c), since ω ≠ 1, we have

ωn − 1
ω n−1 + ω n−2 + ⋯ + 1 = .
ω−1
ωn − 1
By definition, we have ω n = 1. This implies = 0 as desired.
ω−1

4
 1 2 1
6. (a) Let z = x+yi. Then the condition means x2 +y 2 ⩽ x. This is the same as (x − ) +y 2 ⩽ .
2 4
1 1
Therefore, the set is the closed disc with centre and radius .
2 2
Im

Re

Remark. In mathematics, a disc means a circular region, while a circle only means the
boundary of a disc (without the interior part).

(b) The equation means the distance between z and 1 is equal to the distance between z and
3 − 2i. Therefore, it represents the perpendicular bisector of 1 and 3 − 2i.

Remark. One can find more details of the perpendicular bisector as follows. Firstly,
1 + (3 − 2i)
it passes through the midpoint = 2 − i of 1 and 3 − 2i. Secondly, since it is
2
perpendicular to the line in the direction (3 − 2i) − 1 = 2 − 2i, it is parallel to the line in the
direction 1 + i (this requires some knowledge in coordinate geometry or vector geometry,
i.e. (a, −b) and (b, a) point in two perpendicular directions). Thus, the perpendicular
bisector is the line passing through 2 − i and 2 − i + (1 + i) = 3, for example.

Im

1
Re
2−i
3 − 2i

π π
(c) The equation means arg i + arg z = (up to a difference of 2π). Since arg i = , it means
2 2
arg z = 0. Therefore, it represents the positive real axis.

(d) Let z = x + yi. Then

z−i x + (y − 1)i [x + (y − 1)i][(x − 1) − yi]

0 = Re ( ) = Re ( ) = Re ( )
z−1 (x − 1) + yi [(x − 1) + yi][(x − 1) − yi]
[x(x − 1) + y(y − 1)] + (1 − x − y)i x(x − 1) + y(y − 1)
= Re ( )= .
(x − 1)2 + y 2 (x − 1)2 + y 2

5
 1 2 1 2 1
This is the same as x(x − 1) + y(y − 1) = 0, i.e. (x − ) + (y − ) = . This gives a circle
2 2 2
1 1 1
with centre ( , ) and radius √ .
2 2 2

1 √ 1
7. Let − √ + i = r cis θ. Then we have tan θ = − 3. Since − √ + i lies in the second quadrant
3 3
−1
√ 2π
of the Argand diagram, we have θ = tan (− 3) + π = . In view of the figure, we see that
3
π 1 2π
< arg z < arg (− √ + i) = .
2 3 3
4π
ˆ By De Moivre’s theorem, we have arg (z 2 ) = 2 arg z. Therefore, π < arg (z 2 ) < , which
3
means z 2 lies in the third quadrant.
3π
ˆ Similarly, we have arg (z 3 ) = 3 arg z, and hence < arg (z 3 ) < 2π. This means z 3 lies in
2
the fourth quadrant.

Im
z
i

Re
− √13 a

2kπ
8. Let ωk = cis be a 5th root of unity where k = 0, 1, 2, 3, 4. As in example 2.12, we have
5
z 5 + 32(z + 1)5 = 0
⇔ z 5 = −32(z + 1)5
⇔ z 5 = [−2(z + 1)]5
⇔ z = −2(z + 1)ωk for some k
⇔ z = −2ωk z − 2ωk
⇔ z + 2ωk z = −2ωk
⇔ (1 + 2ωk )z = −2ωk
−2ωk
⇔ z= .
1 + 2ωk
−2 cis 2kπ
5
The solutions are given by z = .
1 + 2 cis 2kπ
5

9. (a) By Euler’s formula, we have

eiθ + e−iθ = (cos θ + i sin θ) + (cos (−θ) + i sin (−θ))
= (cos θ + i sin θ) + (cos θ − i sin θ)
= 2 cos θ.

6
 (b) By part (a), we have

eiθ + e−iθ
5
5
cos θ = ( )
2
1
= (e5iθ + 5e3iθ + 10eiθ + 10e−iθ + 5e−3iθ + e−5iθ ) (by the binomial expansion)
32
1
= [(ei(5θ) + e−i(5θ) ) + 5(ei(3θ) + e−i(3θ) ) + 10(eiθ + e−iθ )]
32
1
= (2 cos 5θ + 10 cos 3θ + 20 cos θ)
32
1 5 5
= cos 5θ + cos 3θ + cos θ.
16 16 8
1 5 5
Therefore, a = , b= and c = .
16 16 8

10. (a) Let cis α, cis β and cis γ be the polar forms of a, b and c respectively. Since C is one of
the midpoints of the arc AB, we have γ − α = β − γ + 2nπ for some integer n. This implies
2γ = α + β + 2nπ, and hence

c2 = (cis γ)2 = cis 2γ = cis (α + β) = cis α cis β = ab.

Similarly, we have d2 = ab by symmetry (as the above proof does not specify whether the
major arc or the minor arc is used).

Im
C
B
A

β γ
α Re
O

D
π 2π
(b) It is easy to deduce a = cis and b = cis . By part (a), c and d are the roots of the
4 3
equation
π 2π 11π
z 2 = ab = cis ( + ) = cis .
4 3 12
By proposition 2.5, we know that
11π 35π
c, d = cis , cis .
24 24
Note that a lies in the first quadrant and b lies in the second quadrant. Therefore,
π 2π 11π 35π
< arg c < . This implies c = cis and d = cis .
4 3 24 24