Professional Documents
Culture Documents
Comp PDF
Comp PDF
Michaelmas 2013
The shortest path between two truths in the real domain passes through the complex domain.
Jacques Hadamard (1865-1963)
ax2 + bx + c = 0
are
b b2 4ac
x= . (1)
2a
When b2 4ac > 0, these two roots are real and distinct; graphically they are where the curve y =
ax2 + bx + c cuts the x-axis. When b2 4ac = 0, then we have one real root and the curve just touches
the x-axis here. But what happens when b2 4ac < 0? In this case there are no real solutions to the
equation, as no real number squares to give the negative b2 4ac. From the graphical point of view, the
curve y = ax2 + bx + c lies either entirely above or entirely below the x-axis.
If we imagine 1 to exist, and that it behaves much like other numbers, then the two roots of the
quadratic ax2 + bx + c = 0 can be written in the form
x = A B 1 (2)
where A = b/2a and B = 4ac b2 /2a are real numbers.
Notation 1 We shall from now on write i for 1. This is standard notation amongst mathematicians,
though many books, particularly those written for engineers and physicists, use j instead.
z = a + bi
where a and b are real numbers. The real number a is known as the real part of z and b as the imaginary
part. We write a = Re z and b = Im z. Two complex numbers are equal precisely when their real and
imaginary parts are equal; that is, a + bi = c + di if and only if a = c and b = d. This is called
comparing real and imaginary parts.
1
Note that we can regard real numbers as complex: a real number is simply a complex number with
zero imaginary part.
Definition 4 Let z = a + bi where a, b R. The (complex) conjugate of z is the number a bi, and this
is denoted z (or in some books z ).
Proof Let us prove one of these statements; the remaining ones are left as exercises.
(5): let z = a + bi and w = c + di. Then
zw = (ac bd) + (bc + ad) i
= (ac bd) (bc + ad) i
= (a bi) (c di) = z w.
Note from equation (2) that when the real quadratic equation ax2 + bx + c = 0 has complex roots
which are not real, then these roots are conjugates of each other. More generally, we have
2
Corollary 6 The complex (non-real) roots of a real polynomial come in complex conjugate pairs. Equiv-
alently, if C satisfies the polynomial equation ak z k + ak1 z k1 + + a0 = 0, where each ai is real,
then is also a root of this equation.
Proof Note from the algebraic properties of the conjugate function, proved in the previous proposition,
that
k k1 k k1
ak () + ak1 () + + a1 + a0 = ak () + ak1 () + + a1 + a0
k k1
= ak () + ak1 () + + a1 + a0 [since each ai is real]
k k1
= ak () + ak1 () + + a0
= 0 [since is a root]
= 0.
We needed a special symbol i for 1, but we proceed to show that no further symbols are needed
to find the square root of i. Suppose that z = a + bi, where a and b are real. Then z 2 = i if and only if
2
i = (a + bi) = a2 b2 + 2abi.
a2 b2 = 0 and 2ab = 1.
4
So b = 1/2a from the second
equation, and
substituting for b into the first equation gives a = 1/4, which
has real solutions a = 1/ 2 or a = 1/ 2. So the two complex numbers z which satisfy z 2 = i (i.e. the
two square roots of i) are
1+i 1 i
and .
2 2
Similarly any non-zero complex number has exactly two complex square roots.
In the same spirit, the quadratic formula (1) is also valid for any complex coefficients a, b, c with a 6= 0,
provided that appropriate sense is made of the square roots of the complex number b2 4ac.
Example 7 We can use the quadratic formula (1) to find the two solutions of
z 2 (3 + i) z + (2 + i) = 0.
3
Knowing i = (1 + i) / 2, from the previous problem, we have
b b2 4ac
x =
2a
(3 + i) 2i
=
2
(3 + i) 2 i
=
2
(3 + i) (1 + i)
=
2
4 + 2i 2
= or
2 2
= 2 + i or 1.
0110
0000000000
1111111111
1010
a + bi
i0000000000
1111111111
0000000000
1111111111 r
0000000000
1111111111
1010
real axis
0000000000
1111111111
0000000000
1111111111
1
imaginary axis
A complex number z in the complex plane can be represented by Cartesian coordinates (that is, its
4
real and imaginary parts), but equally useful is the representation of z by polar coordinates. If we let r
be the distance of z from the origin, and if, for z 6= 0, we define to be the angle that the line connecting
the origin to z makes with the positive real axis (measured in the clockwise direction), then we can write
The relationship between the Cartesian and polar coordinates of z is simple we see that
z = cos + i sin
have modulus p
|z| = (cos )2 + (sin )2 = 1.
Note also that arg z is defined only up to the addition of integer multiples of 2. For example, the
argument of 1 + i could be taken as /4 or 9/4 or 7/4, etc. For simplicity, in these notes we shall
give all arguments in the range 0 < 2, so that /4 would be the preferred choice here.
We now come to some useful algebraic properties of the modulus and argument functions.
Moreover, up to the addition of integer multiples of 2, the following equations also hold for z, w 6= 0.
Proof A selection of the above statements is proved here; the remaining ones are left as exercises.
5
(8): let z = a + bi and w = c + di. Then zw = (ac bd) + (bc + ad) i, so that
q
2 2
|zw| = (ac bd) + (bc + ad)
p
= a 2 c 2 + b2 d 2 + b2 c 2 + a 2 d 2
p
= (a2 + b2 ) (c2 + d2 )
p p
= a2 + b2 c2 + d2 = |z| |w| .
(12): let
z = r (cos + i sin )
and
w = R (cos + i sin ) .
Then
We can read off that |zw| = rR = |z| |w| which gives us a second proof of (8), and also that
An important geometric property of the modulus is
|z + w| |z| + |w| .
Note that the shortest distance between 0 and z + w is the modulus of z + w. This is shorter in length
than the path which goes from 0 to z to z + w. The total length of this second path is |z| + |w|.
For an algebraic proof, note that for any complex number z + z = 2 Re z and Re z |z| . So for
z, w C,
z w + zw
= Re (z w) |z w| = |z| |w| = |z| |w| .
2
Then
2
|z + w| = (z + w) (z + w)
= (z + w) (z + w)
= z z + z w + zw + ww
2 2 2
|z| + 2 |z| |w| + |w| = (|z| + |w|) ,
6
000000000000000000000001010
11111111111111111111111 z+w
00000000000000000000000
11111111111111111111111 |w|
00000000000000000000000
11111111111111111111111
z
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
|z + w|
00000000000000000000000
11111111111111111111111 |z|
00000000000000000000000
11111111111111111111111
|z|
00000000000000000000000
11111111111111111111111
00000000000000000000000
11111111111111111111111
01
00000000000000000000000
11111111111111111111111
|w| w
00000000000000000000000
11111111111111111111111
Figure 2: Triangle inequality on the Argand plane
for any c, d R. Our new notation is consistent with earlier uses of the exponential function, as the
following proposition shows:
Proposition 11 If z = rei and w = sei then zw = rsei(+) , and in particular z n = rn ein for any
positive integer n.
Proof The first identity is simply a re-statement of (8) and (12) of Proposition 9. The second follows
by repeated application of the first.
Finally, here is a useful result summarising multiple angle formulas:
Corollary 12 (De Moivres Theorem) For a real number and integer n we have that
n
cos n + i sin n = (cos + i sin ) .
7
Proof Using the last formula in Proposition 11,
n
(cos + i sin ) = (ei )n = ein = cos n + i sin n.
1100000000
0011111111
z+w
00000000
11111111
00000000
11111111
translation by z
00000000
11111111
111010
0000000000
11111111 w
reflection
1010
1110
00 w
z = x + iy = rei .
w 7 z w
w 7 ei w
8
and
w 7 rw.
The latter, with r positive real, is very easy to understand: it is just dilation by magnitude r. Turning to
the former, recall that ei is a complex number with modulus one, and, using the formula in Proposition 11,
w = sei 7 ei w = sei(+) .
So the modulus of ei w is the same as that of w, whereas the argument is increased by . Thus this
transformation is a rotation of the complex plane by angle in the anti-clockwise direction around the
origin.
0110ei w
0000
1111
0000
1111 00
11
rotation
0000
1111 1010 11
00
000000000000000000
111111111111111111
0000
1111
000000000000000000
111111111111111111
0000
1111
rw
000000000000000000
111111111111111111
0000
1111
w dilation
000000000000000000
111111111111111111
a0 + a1 x + a2 x 2 + + an x n = 0
The proof of this theorem is far beyond the scope of these notes. Note that for any given polynomial
with complex coefficients the theorem only guarantees the existence of the a root somewhere in C, unlike
the quadratic formula, which gives us a formula for the two roots. The Fundamental Theorem gives no
hint as to where in C a root is to be found.
a0 + a1 x + a2 x2 + + an xn = 0,
9
with complex coefficients ai C and an 6= 0, there are n (not necessarily distinct) complex numbers
1 , . . . , n such that
a0 + a1 x + a2 x2 + + an xn = an (x 1 ) (x 2 ) (x n ) .
In particular, this shows that any complex polynomial of degree n has, counting repetitions, exactly n
roots in C.
This statement is not hard to derive from the Fundamental Theorem, although the Fundamental
Theorem itself is very hard to prove.
6 Roots of Unity
Problem 15 Let n be a natural number. Find all those complex numbers z such that z n = 1.
We know from the Fundamental Theorem of Algebra that there are (counting repetitions) n solutions:
these are known as the nth roots of unity. Let us first solve z n = 1 directly for n = 2, 3, 4.
When n = 2 we have
0 = z 2 1 = (z 1) (z + 1)
and so the square roots of 1 are 1.
When n = 3 we can factorise as follows:
0 = z 3 1 = (z 1) z 2 + z + 1 .
Plotting these roots on the Argand plane we can see a pattern developing.
Returning to the general case, suppose that z = r (cos + i sin ) satisfies z n = 1. Then by Proposi-
tion 9, z n has modulus rn and has argument n. As 1 has modulus 1 and argument 0, we can compare
their moduli to find rn = 1 giving r = 1 (remember r is a non-negative real number). Comparing argu-
ments, we see n = 0 up to the addition of integer multiples of 2. Thus n = 2k for some integer k,
giving = 2k/n. So the roots of z n = 1 are
2k 2k
z = cos + i sin where k is an integer.
n n
10
01
01 10
1+ 3i i
2
01 01
10 10
1 1
0110
1 3i
2
01 i
At first glance there seem to be an infinite number of roots, but since cos and sin have period 2 these
roots repeat with period n. So the nth roots of unity are the n complex numbers
2k 2k
z = cos + i sin where k = 0, 1, 2, . . . , n 1,
n n
0110
0110
01
0110 0110
11
00 0110
00
11 0110
Figure 6: Roots of unity on the Argand plane for large n
11
If we write 2 + 2i in its polar form, we have
3 3
2 + 2i = 8 cos + i sin .
4 4
Historical notes
It is only comparatively recently that mathematicians have become comfortable with the roots of the quadratic
equation ax2 + bx + c = 0 when b2 4ac < 0. During the Renaissance the quadratic would have been considered
unsolvable, or its roots would have been called imaginary.
The term imaginary was first used by the French Mathematician Rene Descartes (1596-1650). Whilst he is
known more as a philosopher, Descartes made many important contributions to mathematics and helped found
coordinate geometry hence the naming of Cartesian coordinates.
But what meaning can imaginary roots have? This philosophical point pre-occupied mathematicians until
the start of the 19th century; afterwards these imaginary numbers started proving so useful (especially in the
work of Cauchy and Gauss) that these philosophical concerns were essentially forgotten.
The i notation was first introduced by the Swiss mathematician Leonhard Euler (1707-1783). Much of our
modern notation is due to him including e and . Euler was a giant in 18th century mathematics and the most
prolific mathematician ever. His most important contributions were in analysis (eg. on infinite series, calculus of
variations). The study of topology arguably dates back to his solution of the Konigsberg Bridge Problem.
The term complex number is due to the German mathematician Carl Gauss (1777-1855). Gauss is considered
by many the greatest mathematician ever. He made major discoveries in almost every area of mathematics from
number theory and non-Euclidean geometry, to astronomy and magnetism. His name precedes a multitude of
theorems and definitions throughout mathematics. In 1799, he proved the Fundamental Theorem of Algebra, one
of the first major results concerning complex numbers, which conclusively demonstrated their usefulness.
The Argand plane is named after the Swiss mathematician Jean-Robert Argand (1768-1822).
Abraham De Moivre (1667-1754) was a French protestant who moved to England. He is best remembered
for this formula, but his major contributions were in probability and appeared in his The Doctrine Of Chances
(1718).
12