This is a short introduction to complex numbers written primarily
for students aged from about 14 or 15 to 18 or 19. To understand
the first few sections, it would be helpful to be familiar with
polynomial equations (for example, solving $x^2 - 3x+2 = 0$),
basic geometry (angles and lengths) and basic trigonometry (sine
and cosine functions). For the later sections it would be helpful
to be familiar with more advanced trigonometry (double angle
formulae, for example), the exponential and logarithm functions
($e^x$ and $\log(x)$), vectors and matrices, and power series.
The difficult bits are marked. (For students in the UK, the early
sections should be accessible to anyone doing GCSE or above, but
some of the later sections are A-level standard.) There are
exercises throughout for you to test your understanding, with
answers at the back. Any corrections or questions about or beyond
what I've written here can be emailed to
dog@fcbob.demon.co.uk
I've tried to make the exercises less like the standard
computational ones you get at school, but this means that some of
them are quite hard.
Throughout this article, I've used both radians and degrees to
measure angles. You might not have seen radians used before,
they're just another way of measuring angles. Instead of
measuring angles from $0^{\circ}$ to $360^{\circ}$ you measure
them from 0 to $2\pi$. If you haven't seen this before, don't
worry. Where possible I've included degrees and radians (although
some bits only make sense in radians, you can skip these
bits).
You can use the Maths
Thesaurus to look up any unfamiliar words or concepts.
Introduction
If you've done any quadratic equations, you'll know that there
is a nice formula for the solution of the quadratic equation $a
x^2 + b x + c = 0$, given by: $$x = \frac { -b\pm \sqrt{b^2 - 4a
c}} {2a}$$
However, you'll also know that if $b^2 - 4ac$< 0 then there is no solution to the quadratic
equation. Mathematicians like to have answers, so this annoyed
them.
Imagine that the equation $x^2 + 5 = 0$ had a solution $\alpha$.
If there was a solution (ignore the fact that there isn't a
solution for the moment), we could work out all sorts of things
without actually knowing what the solution is. For instance, what
would $8\alpha^3 + 4\alpha^2 + 40\alpha + 20$ be? Well, $\alpha^2
+ 5 = 0$, so $\alpha^2 = -5$. Multiplying this by $\alpha$, we
get $\alpha^3 = -5\alpha$ . Putting this into the equation, we
get:
\begin{eqnarray} 8\alpha^3 + 4\alpha^2 + 40\alpha + 20
&=& 8\times(-5\alpha) + 4\times(- 5)+40a +20 \qquad
&(\textrm{because} \quad \alpha^2 = -5 \quad \textrm{and}
\quad \alpha^3 = -5\alpha)& \\ &=& -40\alpha - 20 +
40\alpha + 20 \qquad &\textrm{(multiplying)}& \\
&=& 0 \qquad &\textrm{(simplifying)}&
\end{eqnarray}
What happens if we substitute $x = 2\alpha$ into the polynomial
$x^3 + x^{2} + 20x + 20$? Well, we get:
\begin{eqnarray} x^3 + x^2 + 20x + 20 &=& (2\alpha)^3
+(2\alpha)^2 + 20\times2\alpha + 20
&\textrm{(substituting)}& \\ &=& 8\alpha^3 +
4\alpha^2 + 40\alpha + 20 &\textrm{(expanding)}&\\
&=& 0 &\textrm{(using the equation above)}&
\end{eqnarray}
So, what other equations can we solve armed with our new number
$\alpha$? Strangely enough, it turns out that we can actually
solve any polynomial equation with this new number, but proving
this is very difficult. If you do maths at university, this is
one of the things you'll learn.
The Basics of Complex
Numbers
If $i$ was a solution to the equation $x^2 + 1 = 0$ (again,
ignore the fact that there is no solution) , then $i^2 + 1 = 0$,
so $i^2 = -1$. If $\beta = i\sqrt{5}$, then $$\beta^2 =
(i\sqrt{5})^2 = i^2 \times 5 = - 5 = \alpha^2$$
As you know, if $\alpha^2 = \beta^2$ then $\alpha = \pm \beta$.
In other words, the number $\alpha$ above is just a multiple of
$i$, $\alpha = \pm i \sqrt{5}$. So, a solution to the equation
$x^3 + x^2 + 20x + 20 = 0$ is $x = 2\alpha = \pm 2 i \sqrt{5}$.
What we do now is pretend that the number $i$ really does exist
(and why not?), but that it is very unlike any number we've seen
before. This new number $i$ is called an imaginary number,
because you have to imagine that it exists, and then you can do
things with it. A complex number is a number $a+i b$, where $a$
and $b$ are the numbers you're familiar with (they're called real
numbers). We can add two complex numbers to get a new complex
number, $(a+i b)+(c+i d) = (a+c)+i(b+d)$. We can multiply them,
$(a+i b)(c+i d) = a c+i b c+i a d+i^2 b d$ = $(a c-b d)+i(b c+a
d)$.
Exercise 1 Work out $i^3
, i^4 , i^5 , i^6 , i^7$ and $i^8$. Do you see a pattern? What do
you think $i^{443}$ is?
Exercise 2 Factorise the
polynomial $x^2 - a^2$ (you don't need complex numbers for this)
Exercise 3 Factorise the
polynomial $x^2 + a^2$ (you do need complex numbers for this)
Exercise 4 (easier) Work
out $(1+i)^2$ and $(1+i)^3$.
Amazingly, we can also divide them, although it's not so obvious
how to do that. What is $1/i$ for example? Well, we can multiply
the top and bottom by the same number and get the same answer,
because $a/i a$ = $1/i$ (the $a$'s cancel). Now set $a = i$ to
get $1/i = i /i^2 = i /-1 = - i$. So $1 /i = - i$. Another way of
seeing this is that $i^2 = - 1$, so if we divide both sides by
$i$ we get $i = - 1 /i$ . What about $1 /1+i$ though? We can't
multiply the top and bottom by $i$ now, because the bottom would
still be a complex number which is no good. However, if we
multiply the top and bottom by $1 - i$, we get: $$\frac{1}{1+i} =
\frac{1- i}{(1+i)(1- i)} = \frac{1- i}{1+i- i+1} = \frac{1- i}{2}
= \frac{1}{2} - i\frac{1}{2}$$
Excellent, but what about in general? What about $1/a+i b$? To
answer this, we introduce something called the complex conjugate
. We'll write the
complex
conjugate of $a+i b$ as $\textrm{Conj}(a+i b)$. We define
it to mean $\textrm{Conj}(a+i b) = a - i b$. If $z = a+i b$, then
$z\textrm{Conj}(z) = (a+i b)(\textrm{Conj}(a+i b)) = (a+i b)(a- i
b) = a^2 + b^2$. But $a$ and $b$ are real numbers, so
$z\textrm{Conj}(z)$ is a real number. Now we can answer the
question of what is $[1/z]$ where $z$ is a complex number. $1 /z
= \textrm{Conj}(z) /z\textrm{Conj}(z)$, where the top bit is a
complex number and the bottom bit is a real number. Expanding it
out we get $$ \frac {1}{a + i b} = \frac {a -i b }{a^2 + b^2} $$
Exercise 5 Work out $$
\frac {a^2 + b^2 }{a -i b} $$ If a complex number $z = a+i b$, we
say that the real part of $z$, written $\textrm{Re}(z)$, is $a$,
and the imaginary part of $z$, written $\textrm{Im}(z)$, is $b$.
In other words, $\textrm{Re}(a+i b) = a$ and $\textrm{Im}(a+i b)
= b$, if $a$ and $b$ are real numbers. Two complex numbers $z$
and $w$ are equal if $\textrm{Re}(z) = \textrm{Re}(w)$ and
$\textrm{Im}(z) = \textrm{Im}(w)$, or in other words $a+i b = c+i
d$ if $a = c$ and $b = d$.
Exercise 6 Express
$\textrm{Re}(z^2)$ and $\textrm{Im}(z^2)$ in terms of
$\textrm{Re}(z)$ and $\textrm{Im}(z)$.
Going back to our old friend the quadratic equation, if $a$, $b$
and $c$ are real numbers, the solution to $a x^2 + b x + c = 0$
is: $$ x = \frac { -b\pm \sqrt{b^2 - 4a c}} {2a} $$ But now this
formula works if $b^2 - 4ac$
< 0 .
Write $\Delta$ for $b^2 - 4a c$ (so $\Delta$
< 0 ). $-\Delta
> 0
$ so we can work out $\sqrt{-\Delta}$ in the normal way. Also,
$i^2 = -1$ so $i = \pm\sqrt{-1}$. Now $\sqrt{\Delta} =
\sqrt{(-1)(-\Delta)} = \sqrt{-1}\sqrt{-\Delta} = \pm{i}\sqrt{(4a
c - b^2)}$. So, if $b^2 - 4a c$
< 0
, then $$ x = \frac { -b\pm i \sqrt{4a c - b^2}} {2a} $$
Amazingly, the formula also works if $a$, $b$ and $c$ are complex
numbers! However, we'll have to wait a bit to see what that
means.
Exercise 7 (harder) If $w
= 1 + i/\sqrt{2}$, work out $w^8$ [hint, work out $w^2$ first,
then square that, then square that $((w^2)^2)^2 = w^8$]
Exercise 8 Find a
solution to the equation $X^2 + 1 = 0$, using exercise 1.
Exercise 9 Find a
solution to the equation $X^4 + 1 = 0$, using exercise 1. If you
want a challenge, try finding all 4 solutions.
Exercise 10 (harder) Find
a solution to the equation $X^2 + 2X + 2 = 0$.
Exercise 11 (very hard, for
those who have done some trigonometry and used radians instead of
degrees to measure angles only) Expand and simplify
$(\cos(2\pi /2 )+ i\sin(2\pi /2 ))^2$. Expand and simplify
$(\cos(2\pi /3 )+ i\sin(2\pi /3 ))^3$. Expand and simplify
$(\cos(2\pi /4 )+ i\sin(2\pi /4 ))^4$. Can you guess what happens
if you expand and simplify $(\cos(2\pi /n )+ i\sin(2\pi /n ))^n$?
For those of you who have met mathematical induction , can you
prove this?
The Argand Plane
We haven't yet got on to the most amazing thing about complex
numbers, the geometric interpretation. As you know, a complex
number $z$ can be written $a+i b$ where $a$ and $b$ are real
numbers. If you've ever done vectors, this will look very
familiar, a 2D vector can be written $a\mathbf{i} +b\mathbf{j}$
where $\mathbf{i}$ and $\mathbf{j}$ are the unit vectors. So, a
complex number $z = a+i b$ corresponds to a point in a 2D plane,
given by $a\mathbf{i} +b\mathbf{j}$. If you haven't done vectors
using the notation above, $a+ib$ just corresponds to the point in
2D with $x$-coordinate $a$ and $y$-coordinate $b$.

What about the sum of two complex numbers, $z+w$. It turns out
that adding two complex numbers is the same as adding two
vectors. If you don't know about how to add two vectors, look at
the following picture:

So, adding vectors corresponds to adding complex numbers. The
2D plane of complex numbers is called the Argand plane or Argand
diagram. Well, that's nice, but not that great. The amazing thing
is what happens when you multiply two vectors. Before we get on
to that, we need a couple of new ideas.
The modulus $|z|$ of a
complex number $z = a+i b$ is given by $|z|$ = $\sqrt{(a^2 +
b^2)}$. By Pythagoras' theorem, this is just the length of the
vector corresponding to $z$. There's another formula for $|z|$,
using the complex conjugate we met earlier, namely $|z| =
\sqrt{z\textrm{Conj}(z)}$ (just expand the right hand side out
and see).
The argument $\arg(z)$ of a complex number $z$ is the angle
between the vector corresponding to $z$ and the positive
$x$-axis. If you've done some trigonometry, $\arg(z) = tan^{-1}
(\textrm{Im}(z) /\textrm{Re}(z) )$ (unless $\textrm{Re}(z) = 0$).
The following picture hopefully explains the modulus and argument
of a complex number:

Exercise 12 Let $w =
1+i/\sqrt{2}.$ What are$|w|$ and $\arg(w)$?
Exercise 13 What is
$|z|-|\textrm{Conj}(z)|$? Explain geometrically what
$\textrm{Conj}(z)$ is in terms of $z$. Using this, what is
$\arg(\textrm{Conj}(z))+\arg(z)$?
Now we can discuss the amazing thing about the geometry of
complex numbers. If $z$ and $w$ are two complex numbers, then $|z
w| = |z| |w|$ and $\arg(z w) = \arg(z)+\arg(w)$. In other words,
if you multiply two complex numbers, you multiply their lengths
and add their angles. We won't be able to prove the second
equation until after the next section, but we can prove the first
one. If $z = a+i b$ and $w = x+i y$, then $z w = (a x- b y)+i(a
y+b x)$. So
\begin{eqnarray} |z w|^2 &=& (a x-b y)^2 +(a y+b x)^2
\\ &=& a^2x^2 -2a b x y+b^2y^2+a^2y^2 +2a b x y+b^2x^2
\\ &=& (a^2 + b^2)(x^2 + y^2) \\ &=& |z|^2
|w|^2 \end{eqnarray}
So $|z w| = |z| |w|$.
Exercise 14 What is
$|z/w|$ in terms of $|z|$ and $|w|$? What is $\arg(z/w)$ in terms
of $\arg(z)$ and $\arg(w)$? [Hint: $z = (z/w)(w)$]
Exercise 15 What is
$|z^n|$ in terms of $|z|$? What is $\arg(z^n)$ in terms of
$\arg(z)$? If you know about mathematical induction, prove your
result.
Exercise 16 (for those who have
done some trigonometry) If $z = r(\cos{\theta}
+i\sin{\theta})$, what are $|z|$ and $\arg(z)$? Can you write $z$
in terms of $|z|$ and $\arg(z)$ for any complex number $z$? Can
you use this to prove $\arg(z w) = \arg(z)+\arg(w)$?
Exercise 17 (very hard, for
those who have done some trigonometry) What is
$(\cos{\theta} +i\sin{\theta})^n$ in terms of $\cos{n\theta}$ and
$\sin{n\theta}$?
Exercise 18 (very hard, for
those who have completed the previous question only) Find
a solution (not $z$ = 1) to the equation $z^n - 1 = 0$ in terms
of $\cos(2\pi /n ) = \cos(360^{\circ}/n )$ and $\sin(2\pi /n ) =
\sin(360^{\circ}/n )$. Find all the $n$ solutions in terms of
this solution.
Polar Coordinates and De
Moivre's Formula
There is another way of writing complex numbers apart from $a+i
b$. Since a complex number is like a point in the complex plane,
we can work out the distance of this point from the origin, $r$,
and the angle that the line from the origin to the point makes
with the $x$-axis, $\theta$. Once we've worked these out we can
write a complex number as $(r,\theta)$, this is called the
Polar Coordinate
notation . These numbers $r$ and $\theta$ are just the
modulus and argument of $x+i y$ that we met above.
So a complex number $z = a+i b$ can be written as $z =
(|z|,\arg(z))$. Also, given a complex number $(r,\theta)$ we can
change it into $x+i y$ notation (this is what Q1 is about) as
$(r, \theta) = r(\cos\theta +i \sin\theta)$.
OK, so we can switch between two different notations for writing
complex numbers, but what use is it? Well, it's very useful for a
reason that will become apparent if you read the next section
(which is a bit harder), but is also useful for a couple of other
reasons.
I've already mentioned that $|z w| = |z| |w|$ and $\arg(z w) =
\arg(z)+\arg(w)$. This makes it very easy to multiply and divide
complex numbers written in polar coordinates, since
$(r_1,\theta_1)(r_2, \theta_2) = (r_1 r_2 , \theta_1 +
\theta_2)$.
Suppose $z = (r, \theta)$, what is $z^2$ in polar coordinates?
Well, using the formula above, $z^2 = (r^2 ,2\theta)$. How about
$z^3$, $z^4$, etc.? If we repeatedly use the above formula, we
get that $z^3 = (r^3 ,3\theta)$, $z^4 = (r^4 ,4\theta)$. You can
probably guess how the pattern continues, $z^n = (r^n ,n\theta)$.
If $|z| = r = 1$ then $z = \cos\theta +i\sin\theta$, and so the
formula in this case gives us: $$(\cos\theta +i\sin\theta)^n =
\cos(n\theta) +i\sin(n\theta)$$
This is known as
De Moivre's
Formula and is the answer to exercise 17.
Exercise 19 Write $1+i$
in polar coordinates.
Exercise 20 Using the
previous question, calculate $(1+i)^8$ in a couple of lines of
working.
Exercise 21 (a bit
tricky) Write $\cos(3\theta)$ in terms of $\cos(\theta)$.
[Hint: use De Moivre's formula and the fact that $\cos^2\theta +
\sin^2\theta = 1$]
Exponentials and Trigonometry
(Advanced)
This section is much more difficult, you need to be able to
understand measuring angles in radians instead of degrees. If you
don't know about radians, skip to the next section.
Some functions have what is called a power series . If you've met
these before, this section shouldn't be too difficult. If not,
all you need to know is that for every number $\theta$ (an angle
measured in radians) $$\cos\theta = 1 - \frac{\theta^2}{2!} +
\frac{\theta^4}{4!} - \frac{\theta^6}{6!} + \dots$$ and
$$\sin\theta = \theta - \frac{\theta^3}{3!} + \frac{\theta^5}{5!}
- \frac{\theta^7}{7!} + \dots$$
In fact, this is the reason that we use radians instead of
degrees. You also need to know that the "exponential function''
$e^x$ (which is the number $e \approx 2.718281828\dots$ to the
power of $x$) has a power series given by: $$e^x = 1 + x +
\frac{x^2}{2!} + \frac{x^3}{3!} + \frac{x^4}{4!} + \dots$$ What
happens if we use this power series to calculate $e^{i\theta}$?
So far, if you've used power series at all, you've probably only
used them for real numbers, but they work just as well for
complex numbers. Expanding $e^{i\theta}$ we get:
\begin{eqnarray} e^{i\theta} &=& 1 + (i\theta) +
\frac{(i\theta)^2}{2!} + \frac{(i\theta)^3}{3!} +
\frac{(i\theta)^4}{4!} + \dots \\ &=& 1 + i\theta -
\frac{\theta^2}{2!} - i\frac{\theta^3}{3!} +
\frac{\theta^4}{4!} + \dots \\ &=& (1 -
\frac{\theta^2}{2!} + \frac{\theta^4}{4!} + \dots) + i(\theta -
\frac{\theta^3}{3!} + \frac{\theta^5}{5!} + \dots) \\
&=& \cos(\theta) + i\sin(\theta) \end{eqnarray}
The last line follows from looking at the power series for
$\cos(\theta)$ and $\sin(\theta)$ above. So, we've proved that
(for $\theta$ measured in radians), $e^{i\theta} = \cos\theta +
i\sin\theta$ ! This is sometimes referred to as Euler's formula . You can use this
to prove one of the most famous formulae in mathematics, which
is $e^{i\pi} = - 1$, because $e^{i\pi} = \cos\pi +i\sin\pi = -1
+ 0i = -1$.
So, what does this have to do with the polar coordinate notation?
If we write $z = (r,\theta)$ then we can also write $z =
{r}{e^{i\theta}}$ using the proof above.
Exercise 22 Prove that
$|z w| = |z| |w|$ and $\arg(z w) = \arg(z)+\arg(w)$ using the
${r}{e^{i\theta}}$ formula.
Exercise 23 Find a
solution to the equation $z^n = -1$ using the ${r}{e^{i\theta}}$
formula and $e^{i\pi} = - 1$.
Exercise 24 (a bit
tricky) The hyperbolic cosine is defined to be $\cosh(z) =
[e^z + e^{-z}] /2$. Find a power series expansion for $\cosh(z)$
and prove that $\cosh(i z) = \cos(z)$.
Cool stuff
Now we know the basics of how complex numbers work, what can we
do with them?
Suppose we have a complex number $z = (r, \theta)$ in polar
coordinates, and another complex number $w = (1, \phi)$ with
modulus 1. The product of these two complex numbers is $z w = (r,
\theta + \phi)$. In other words, $z w$ is $z$ rotated by an angle
of $\phi$. We can use these to work out the matrix which rotates
a vector by an angle $\phi$. Suppose $z = x+i y$ and $w = (1,
\phi) = \cos\phi + i\sin\phi$, then $z w = (x+i y)(\cos\phi +
i\sin\phi) = (x\cos\phi - y\sin\phi) + i(x\sin\phi + y\cos\phi)$.
In other words, the $x$-coordinate (which is equivalent to the
real part of a complex number) of a vector rotated by an angle
$\phi$ is $x\prime = x\cos\phi - y\sin\phi$ and the
$y$-coordinate (the imaginary part) is $y\prime = x\sin\phi -
y\cos\phi$. We can represent this as a product of matrices:
\left( \begin{array}{c} x\prime \\ y\prime\end{array}\right) =
\left(\begin{array}{cc} \cos\phi & -\sin\phi \\ \sin\phi
& \cos\phi \end{array}\right) \left(\begin{array}{c} x \\ y
\end{array}\right)
In other words, the rotation matrix $M$ which rotates a
vector \textbf{v} to a vector $M\mathbf{v}$ by an angle $\phi$
is:
\mathbf{M} = \left(\begin{array}{cc} \cos\phi & -\sin\phi
\\ \sin\phi & \cos\phi \end{array}\right)
I mentioned at the beginning that you can solve any
polynomial equation using complex numbers, this is known as the
Fundamental Theorem of
Algebra and a famous mathematician known as
Gauss gave about 9 different proofs
of this hundreds of years ago. More precisely, the fundamental
theorem of algebra states that if $p(z) = z^n + a_{n-1}z^{n-1} +
a_{n-2}z^{n-2} + \dots + a_{2}z^{2} + a_1z + a_0$ (where each of
the $a_i$ is a complex number) then there is a complex number $w$
such that $p(w) = 0$. This is quite amazing really, since in
creating complex numbers what we did was "invent'' a new number,
a "solution'' to the equation $z^2 +1 = 0$, and we get the
solutions to all polynomial equations for free!
Exercise 25 [hard] Use
the fundamental theorem of algebra (stated above) and the
Remainder Theorem (stated below) to prove, by induction, that a
polynomial $p(z) = z^n + a_{n-1}z^{n-1} + a_{n-2}z^{n-2} + \dots
+ a_1z + a_0$ can be written $p(z) = (z - w_1)(z - w_2)\dots(z -
w_n)$ for some complex numbers $w_i$. The degree of a polynomial
is the highest power of $z$ in the polynomial, for example the
degree of $z^3 + z^2 + z$ is 3 and the degree of $z^{18} + 2$ is
18. The Remainder Theorem says (or almost does) that if $p(w)$ =
0 then $p(z) = (z-w)q(z)$ for some polynomial $q(z)$ whose degree
is one less than the degree of $p(z)$.
Cool stuff (Advanced)
This section is quite difficult, and is for those who understood
the section on exponentials and trigonometry.
If you know about the function $e^z$ and its power series, then
you've probably also heard of the log function, which is its
inverse. If not, the function $\log(z)$ is the unique (for real
numbers) function such that $\log(e^z) = z$ and $e^{\log(z)} =
z$.
Using this function, we can show that $x^y = e^{y\log(x)}$ for
any positive real numbers $x$ and $y$. This is because (by the
definition of log) $x = e^{\log(x)}$ , so $x^y = (e^{\log(x)})^y
= e^{y\log(x)}$.
We can extend the idea of the log function to complex numbers,
but unfortunately it is not unique (and there are other problems
with it that I won't go into). For example, we know that
$e^{2{\pi}{i}} = e^0 = 1$, so for complex numbers log(0) can be
$2{\pi}{i}$ or 0! Since $x^y = e^{y\log(x)}$, this also means
that $x^y$ is not unique when $x$ and $y$ are complex numbers.
However, there is a solution to this problem. We know that any
complex number can be written $z = {r}{e^{i\theta}}$ , suppose
that $\log(z) = x+i y$, then $e^{\log(z)} = z = {r}{e^{i\theta}}
= e^{x+i y} = e^{x} e^{i y}$. So the modulus and argument of
$e^{x} e^{i y}$ must be $r$ and $\theta$ respectively. In other
words, $\log(z) = \log|z| +i\arg(z)$ where $\log|z|$ is unique
because $|z|$ is a positive real number. I glossed over this
problem earlier, but the argument of a complex number is not
unique, the argument of $i$ could be $90^{\circ}$ or it could be
$450^{\circ}$ , or in fact $(90+360n)^{\circ}$ for any integer
$n$.
We say that the
principal
argument of $z$, written $\textrm{Arg}(z)$, is the unique
angle $- \pi < \textrm{Arg}(z) \geq \pi$. Similarly, we can
define the
principal
branch of the logarithm $\textrm{Log}(z) = \log|z| +
i\textrm{Arg}(z)$. Using this, we can define the
principal branch of complex
exponentiation to be $x^y = e^{y\textrm{Log}(x)}$.
As an example, let's use this to work out the
principal value of $i^i$ . We know
that $e^{i\pi /2} = i$, and so $\textrm{Log}(i) = i\pi /2$ . So
$i^i = e^{i\textrm{Log}(i)} = e^{i \times i\pi /2} = e^{-\pi
/2}$, and this is a real number. In other words, we take one
complex number to the power of another complex number and we get
a real number!
Exercise 26 [very difficult
indeed] Find all complex number solutions to the equation
$x^y = y^x$ . [Hint: write $y = t x$ and see where you can go
from there.]
Answers
The Basics of Complex
Numbers
Answer 1
$i^3 = - i, i^4 = 1, i^5 = i, i^6 = - 1, i^7 = - i, i^8 = 1,$
etc. In general $i^{4n} = 1, i^{4n+1} = i, i^{4n+2} = - 1,
i^{4n+3} = - i$, so $i^{443} = - i$.
Answer 2 $\quad (x-
a)(x+a)$.
Answer 3 $ \quad (x- i
a)(x+i a)$.
Answer 4 $ \quad (1+i)^2
= 2i, (1+i)^3 = 2i- 2$.
Answer 5
$$ \frac {a^2 + b^2 }{a -i b}\ = \frac {(a^2 + b^2)(a + i b)}{a^2
+ b^2}\ = a + i b $$
Answer 6 $\quad z =
\textrm{Re}(z)+i\textrm{Im}(z)$, so $z^2 = \textrm{Re}(z)^2 -
\textrm{Im}(z)^2 +2i(\textrm{Re}(z)\textrm{Im}(z))$. So
$\textrm{Re}(z^2) = \textrm{Re}(z)^2 - \textrm{Im}(z)^2$ and
$\textrm{Im}(z^2) = 2\textrm{Re}(z)\textrm{Im}(z).$
Answer 7 $ \quad w^8 =
1$. This is called an
eighth
root of unity .
Answer 8 $ \quad X = \pm
i$.
Answer 9 $ \quad X = \pm
w = \pm 1+i /\sqrt{2}$. The other solutions are $- w$ , $i w$ and
$- i w$.
Answer 10 The solution
is, using the quadratic formula, $$ x = \frac { -2\pm i
\sqrt{4-8}} {2}\ = 1\pm \sqrt{-1}\ = -1 \pm i $$
Answer 11 Expanding each
of the equations and simplifying gives 1 in each case (including
the general case). Look at the section on De Moivre's formula for
the proof.
The Argand Plane
Answer 12 $ \quad |w| =
1$ and $\arg(w) = 45^{\circ} = \pi/4$.
Answer 13 $ \quad|z|^2 =
x^2 + y^2 = x^2 +(- y)^2 = |\textrm{Conj}(z)|^2$, so $|z| =
|\textrm{Conj}(z)|$ and $|z| - |\textrm{Conj}(z)| = 0$. The
complex conjugate is a reflection in the $x$-axis, so
$\arg(\textrm{Conj}(z)) = - \arg(z)$ and hence
$\arg(z)+\arg(\textrm{Conj}(z)) = 0$.
Answer 14 $ \quad |z^n| =
|z|^n$ , and we can prove this using induction. If $n = 1$ then
it is obvious. If it is true for $n$, in other words we know that
$|z^n| = |z|^n$ then $|z^{n+1}| = |z^{n} z| = |z^n| |z| = |z|^n
|z| = |z|^{n+1}$, and so by induction it is true for all $n$.
Similarly, $\arg(z^n) = n\arg(z)$.
Answer 15 Using the hint
that $z = (z /w )w$ and the fact that $|u v| = |u| |v|$ we get
that $|z| = |(z /w )| |w|$ , and so $|(z /w )| = |z| /|w|$ .
Similarly, taking the argument of both sides, we get $\arg(z) =
\arg(z /w )+\arg(w)$ and so $\arg(z /w ) = \arg(z)- \arg(w)$.
Answer 16 $\quad |z| = r$
and $\arg(z) = \theta$. So $z = |z|
(\cos(\arg(z))+i\sin(\arg(z)))$. We can now prove that $|z w| =
|z| |w|$ and $\arg(z w) = \arg(z)+\arg(w)$ by expanding out the
brackets using this expansion of $z$ and $w$ and the
trigonometric formula for $\cos(A+B) = \cos(A)\cos(B)-
\sin(A)\sin(B)$ and $\sin(A+B) = \cos(A)\sin(B)+\sin(A)\sin(B)$.
So, if $|z| = r_1, |w| = r_2$, $\arg(z) = \theta_1$ and $\arg(w)
= \theta_2$ then $z = r_1 (\cos\theta_1 +i\sin\theta_1)$ and $w =
= r_2 (\cos\theta_2 +i\sin\theta_2)$. Expanding out
\begin{eqnarray} z w &=& r_{1} r_{2}\cos\theta_{1}
+i\sin\theta_{1})(\cos\theta_{2} +i\sin\theta_{2})\\
&=& r_{1} r_{2} ((\cos\theta_{1} \cos\theta_{2} -
\sin\theta_{1} \sin\theta_{2} )+i(\cos\theta_{1} \sin\theta_{2}
+ \cos\theta_{2} \sin\theta_{1} )) \\ &=& r_1 r_2
(\cos(\theta_{1} + \theta_{2} )+i\sin(\theta_{1} + \theta_{2}
)) \end{eqnarray}
So $|z w| = r_1 r_2$ and $\arg(z w) = \theta_1 + \theta_2$.
Answer 17 There are
various ways of answering this, you could calculate it directly
using the formulas for $\cos(A+B)$ and $\sin(A+B)$. See the
section on De Moivre's formula for an easier way.
Answer 18 The solution is
$z = \cos(2\pi /n )+i\sin(2\pi /n) =
cos(360^{\circ}/n)+i\sin(360^{\circ}/n)$ (which you can check
works using the previous question). The other $n$ solutions are
$z^2$, $z^3$,\dots, $z^n = 1$.
Polar Coordinates
Answer 19 $\quad 1+i =
(\sqrt{2},45^{\circ}) = (\sqrt{2}, \pi /4 )$.
Answer 20 $\quad (1+i)^8
= (\sqrt{2},45^{\circ})^8 = (\sqrt{2^8} , 8 \times 45^{\circ}) =
(16,360^{\circ}) = 16$.
Answer 21 Using De
Moivre's formula, we have that $\cos(3\theta) + i\sin(3\theta) =
(\cos\theta +i\sin\theta)^3$. We can expand out the right hand
side to get $\cos^3\theta +i\cos^2\theta \sin\theta - \cos\theta
\sin^2\theta - i\sin^3\theta$ which has real part $\cos^3\theta -
\cos\theta \sin^2\theta$. We also know that $\sin^2\theta = 1 -
\cos^2\theta$ , so the real part of $(\cos\theta +i\sin\theta)^3$
is $\cos^3\theta - \cos\theta (1 - \cos^2\theta) = 2\cos^3\theta
- \cos\theta$. So $\cos(3\theta) = 2\cos^3\theta - \cos\theta$.
You can use the same method to express $\cos(n\theta)$ or
$\sin(n\theta)$ as a polynomial in $\cos\theta$ or $\sin\theta$.
Exponentials and Trigonometry
(Advanced)
Answer 22 If $z = r_1
e^{i\theta_1}$ and $w = r_2 e^{i\theta_2}$ then $z w = r_1 r_2
e^{i\theta_1} e^{i\theta_2} = r_1 r_2 e^{i\theta_1 + \theta_2}$
and so $|z w| = |z| |w|$ and $\arg(z w) = \arg(z)+\arg(w)$.
Answer 23 We know
$e^{i\pi} = - 1$ so $z^n = - 1$ means that $z^n = e^{i\pi}$. If
$z = r e^{i\theta}$ then $z^n = r^n e^{{i}{n}\theta}$. So we need
to solve $r^n e^{{i}{n}\theta} = e^{i\pi}$. Clearly $r = 1$ and
$q = \pi /n$ work, so $z = e^{i\pi/n}$ is a solution.
Answer 24 By adding and
expanding the series for $e^z$ and $e^{-z}$ we get that
$$\cosh{z} = 1 + \frac{z^2}{2!} + \frac{z^4}{4!} + \dots$$ and so
$$\cosh{i z} = 1 - \frac{z^2}{2!} + \frac{z^4}{4!} + \dots$$
which is just the power series for $\cos(z)$. You can use the
same method to prove that if the hyperbolic cosine $\sinh(z) =
[e^z - e^{-z}] /2$ then $\sinh(i z) = i\sin(z)$. Also, $\cos(i z)
= \cosh(z)$ and $\sin(i z) = i\sinh(z)$.
Cool Stuff
Answer 25 We're trying to
prove it by induction, so we have to first prove that it is true
for $n = 1$. If $n = 1$ then $p(z) = z + a_0$ for some complex
number $a_0$. In this case, $p(z)$ can be written $p(z) = (x- (-
a_0 ))$ and so the theorem is true. Next we have to prove that if
it is true for $n$ then it is true for $n+1$.
Suppose that $p(z)$ is a polynomial of degree $n+1$. By the
fundamental theorem of algebra, there is a solution $w_{n+1}$. By
the Remainder Theorem, $p(z) = (z- w_{n+1})q(z)$ where the degree
of $q(z)$ is $n$. By induction, $q(z) = (z - w_1) \dots (z -
w_n)$ and hence $p(z) = (z - w_1) \dots (z - w_n)(z - w_{n+1})$
and so the theorem is true by induction.
As a further exercise, you might like to try and prove the
Remainder Theorem.
Cool Stuff (Advanced)
Answer 26 As I said, we
start by writing $y = t x$, because if $x$ isn't zero (if $x = 0$
then $y = 0$ is the only solution, but $0^0$ is undefined so it
isn't a solution) and $y$ is a solution then $y = y /x \times x$,
or $t = y /x$. You might wonder why we introduce this spurious
$t$, but just wait a minute and you'll see how useful it is.
Substituting $t x$ for $y$ in the equation $x^y = y^x$ we get
$x^{t x} = (t x)^x$. Taking logs of both sides we get $t x\log(x)
= x(\log(t)+\log(x))$. We know that $x \neq 0$ so we can divide
both sides by $x$ to get $t\log(x) = \log(t)+\log(x)$.
Rearranging, $\log(x)(t - 1) = \log(t)$, so $\log(x) = \log(t) /
t- 1$. Taking exponentials, $x = e^{\log(t)/(t- 1)} = t^{1/(t-
1)}$. Also, $y = t x = t^{1/(t- 1)+1} = t^{t/(t - 1)}$. So we've
found the solutions in terms of a parameter $t$.