MTH6128 Number Theory

Notes 2 Spring 2019

2 Algebraic numbers
2.1 Algebraic numbers and algebraic integers
Pythagoras and his school discovered that the square root of 2 is not a rational
number. (You saw the proof of this in Introduction to Algebra.) However, it
is an easy number to describe geometrically: it is the ratio of the diagonal of a
square to its side.
The number π has a more complicated geometric description: it is the ratio
of the circumference of a circle to its diameter, but there is no simple method
to construct a straight line which is equal to the circumference of a given circle.
(We know now, for example, that such a line cannot be constructed with the
traditional geometric instruments of “ruler and compass”.)
We make a distinction between algebraic numbers (which are roots of poly-
nomials with rational coefficients) and transcendental numbers (which are not):

Definition Let α be a complex number. Then:

• α is an algebraic number if there is a non-zero polynomial f (x) with rational

coefficients such that f (α) = 0;

• α is a transcendental number if α is not an algebraic number. Moreover,

• α is an algebraic integer if there is a non-zero monic polynomial f (x) with

integer coefficients such that f (α) = 0.

(A polynomial is monic if the coefficient of the highest power of x is equal to 1,

i.e., it is of the form f (x) = xn +an−1 xn−1 +. . .+a1 x+a0 , with a0 , . . . , an−1 ∈ Z.)
Note that there are two extra conditions in the definition of an algebraic
integer. The polynomial should be monic, and it should have integer coefficients.

1
Just asking one of these conditions would just define algebraic numbers over
again. For, given any non-zero polynomial f (x),

• we can make it monic by dividing by the coefficient of the highest power

of x; and

• we can make the coefficients integers by multiplying by the least common

multiple of the denominators.

Example
√
• Let α = 2. Then α2 − 2 = 0, so α is a root of the polynomial x2 − 2.
This is monic and has integer coefficients. Thus α is an algebraic integer.
√ √
• Let α = 21 2 = 1/ 2. Then α2 = 12 , so α is a root of the monic rational
polynomial x2 − 12 , or (equivalently) of the non-monic integer polynomial
2x2 −1. Thus, α is an algebraic number; but we haven’t decided yet whether
it is an algebraic integer. We will see soon that it is not.

• π is transcendental (that is, it is not an algebraic number). This was proved

by Lindemann about 150 years ago.

Any integer is an algebraic integer; the integer n satisfies the polynomial

x − n = 0. Similarly, any rational number is an algebraic number. In the other
direction, we have:

Proposition 2.1 A rational number is an algebraic integer if and only if it is

an integer.

For this reason, we sometimes call the ordinary integers “rational integers”.

Proof We have seen that integers are algebraic integers; we have to prove that
a rational number which is an algebraic integer is an integer.
Let q = a/b be a rational number (a ∈ Z, b ∈ Z, b ≥ 1) in its lowest terms
(so that gcd(a, b) = 1). Suppose that q satisfies a monic polynomial with integer
coefficients, say

f (q) = q n + cn−1 q n−1 + · · · + c1 q + c0 = 0.

Putting q = a/b, and multiplying this equation by bn , we obtain

an + cn−1 an−1 b + · · · + c1 abn−1 + c0 bn = 0.

2
Now every term in this equation except the first is divisible by b, so b divides
an . Now assume b > 1, and choose a prime factor p of b. Since b divides an , p
divides an as well. Applying Lemma 1.2, we find that p | a. But also p | b, so
p | gcd(a, b) = 1. This is a contradiction. Therefore b = 1, so that q is an integer.


There is a result, which I will not prove, which makes things like this much
easier. This is known as Gauss’s Lemma. It can be stated in many different
ways. But the following will do for our purposes.

Definition Let α be an algebraic number. The minimal polynomial of α is

the non-zero, monic polynomial f (x) of smallest possible degree with rational
coefficients such that f (α) = 0.
(Any algebraic number satisfies a monic polynomial with rational coefficients,
and we can certainly choose one of smallest degree. Why is it unique? Suppose
that f1 (x) and f2 (x) were two different monic polynomials of the same (smallest)
degree satisfied by α, and let g(x) = f1 (x) − f2 (x). Then f1 (α) = f2 (α) = 0,
so g(α) = 0, but g has smaller degree than f1 and f2 ; we can make it monic by
dividing by the leading coefficient. But this contradicts the minimality of the
degree of f1 and f2 .)

Theorem 2.2 (A version of Gauss’ Lemma) The algebraic number α is an

algebraic integer if and only if its minimal polynomial has integer coefficients.

Now let q be a rational number. It satisfies the polynomial x − q = 0,

and clearly this is monic and has smallest possible degree, so it is the minimal
polynomial of q. So q is an algebraic integer if and only if the coefficients 1 and
−q of this polynomial are both integers, i.e. if√and only if q is an integer.
Let us now reconsider the number α = 12 2. We saw that it satisfies the
monic polynomial x2 − 12 . This polynomial is irreducible, and its coefficients are
not all integers. So α is not an algebraic integer. This confirms what we claimed
earlier.

2.2 Quadratic numbers and quadratic integers

We now define a quadratic number or a quadratic integer, to be an algebraic num-
ber or algebraic integer which is not rational but satisfies a quadratic equation.
More formally:

3
Definition An algebraic number is a quadratic number if its minimal polyno-
mial is of degree 2.
An algebraic number is a quadratic integer if its minimal polynomial is of
degree 2 and has integer coefficients.

Remark A quadratic number is irrational. Therefore, we also call quadratic

numbers quadratic irrationals.
An algebraic number is a quadratic integer if and only if it is a quadratic
number and an algebraic integer.

Real quadratic numbers will be very important to us in this course, so let us

2
see what they look
√ like. The solutions to the quadratic equation x + bx + c = 0
are x = (−b ± b2 − 4c)/2. We can write this in simpler form as follows.

Definition An integer d is said to be squarefree if for every prime p, p | d ⇒

p2 6 | d.
Said otherwise, d is squarefree if no prime divides it more than once, so a
squarefree number is - up to sign - a product of distinct primes.
Every integer is of the form c2 d for some c, d ∈ Z, d squarefree.
√
√
Now for any rational number q, we can write q = v d where v ∈ Q and d
is a squarefree integer. For example,
r r
2√
r
280 23 · 5 · 7 23 · 3 · 5 · 7 · 11
= = = 2310.
297 33 · 11 34 · 112 99
Here 2310 = 2 · 3 · 5 · 7 · 11 is squarefree.
The general procedure is as in this example: multiply top and bottom by all
the primes occurring to odd powers in the denominator; take all even powers of
primes out of the square root; and what is left is of the right form.
So we conclude:

Proposition
√ 2.3 A complex number α is a quadratic number if and only if α =
u + v d where u, v ∈ Q, v 6= 0 and d is a squarefree integer not equal to 1.

Proof ”⇐”: f (x) = (x − u)2 − v 2 d satisfies f (α) = 0. q

2
2
”⇒”: Let f (x) = x + px + q satisfy f (α) = 0. Then α = − p2 ± p4 − q.
2 p2
Write p4 − q = ab for some a, b ∈ Z, b > 0, so that 4
− q = ab = b12 ab = b12 c2 d for
some c, d ∈ Z, c > 0, d 6= 1, d squarefree. Set u = − p2 , v = ± cb . 

4
 √
Remark √ The expression
√ u + v d for α is unique, in the sense that if we have
u1 + v1 d1 = u2 + v2 d2 for some u1 , u2 , v1 , v2 ∈ Q and squarefree integers d1 , d2
not equal to 1, then u1 = u2 , v1 = v2 and d1 = d2 .
√
2.3 The ring of integers of Q( d)
√ √
Recall that Q( d) = {u + v d : u, v ∈ Q} is a field. For d = 1 this is just the
rational numbers, which contains Z, the ring of integers.
√ In this section we will
introduce the analog of the integers for the field Q( d).
√
Definition Suppose d is -free. The ring of integers of Q( d) denoted OQ(√d)
is defined as
√
OQ(√d) = {α ∈ Q( d) : α is an algebraic integer}.

It turns out that OQ(√d) is a ring, however this is not clear from the definition.
√
Also, note that Z ⊂ OQ(√d) since Z ⊂ Q( d) and every integer is an algebraic
integer.
In the following proposition we will characterize the quadratic integers.

Proposition
√ 2.4 A quadratic number α is a quadratic integer if and only if
α = u + v d, where d is a squarefree integer not equal to 1 and either

• u, v ∈ Z, v 6= 0, or

• u − 21 , v − 1
2
∈ Z and d ≡ 1 (mod 4).
√
Example The golden ratio φ = (1 + 5)/2 is a quadratic integer, with u =
v = 12 and d = 5. It is a root of the polynomial x2 − x − 1.
√
Example The complex number ω = (−1 + −3)/2 is a quadratic integer.
Indeed, it is an algebraic integer since it is a cube root of unity; that is, it is a
root of x3 − 1. Now x3 − 1 = (x − 1)(x2 + x + 1), so ω has minimal polynomial
x2 + x + 1.
√ √
Example 1/ 2 = 12 2 is an algebraic number but not an algebraic integer,
√ √
since it has the form u + v d with u = 0, v = 12 and d = 2. Similarly, (3 + 7)/2
is an algebraic number but not an algebraic integer; we have u = 3/2, v = 1/2,
and d = 7 6≡ 1 (mod 4).

5
 The proof of Proposition 2.4 is not difficult, but for sake of brevity has been
moved to an appendix. I won’t expect you to know√the proof.
A naive guess would be that OQ(√d) = {m + n d : m, n ∈ Z}, but if d = 5
√
in the example above we saw that φ = (1 + 5)/2 ∈ OQ(√5) so the guess is not
quite correct.
Using the proposition we can now describe OQ(√d) .

Proposition 2.5 For d square-free we have

√ √
OQ(√d) = {m + n d : m, n ∈ Z} =: Z[ d] if d ≡ 2, 3 (mod 4)

and
  √   " √ !#
 1+ d  1+ d
OQ(√d) = m+n : m, n ∈ Z =: Z if d ≡ 1 (mod 4)
 2  2

Example For d = −1 we get OQ(√−1) = Z[i], which is referred to as the

Gaussian integers.
√ h √ i
1+ 5
Example We saw the golden ratio φ = (1 + 5)/2 ∈ OQ(√5) =Z 2
.

Remark Using Proposition 2.4 it is a straightforward exercise to prove that

OQ(√d) is a ring.
√
Proof If d ≡ 2, 3 (mod 4) then by Proposition 2.4 α = u + v d ∈ OQ(√d) iff
√
u, v ∈ Z iff α ∈ Z[ d].
The case d ≡ 1 (modh4) is slightly harder.
√ i
1+ d
Step 1. Show that Z 2
⊂ OQ(√d) (Homework).
h √ i
Step 2. Show that OQ(√d) ⊂ Z 1+2 d . If α ∈ OQ(√d) and α ∈ Z then
h √ i √
clearly α ∈ Z 1+2 d . By Proposition 2.4 if α = u + v d ∈ OQ(√d) and is not
√
in Z then α = u − 21 + (v − 21 ) d where u, v ∈ Z. Let a = 2u − 1 and b = 2v − 1
so a, b are odd and
√ ! " √ !#
a b√ b−a 1+ d 1+ d
α= + d= +b ∈Z .
2 2 2 2 2

6
2.4 Units in OQ(√d) and Pell’s equation
By analogy with the integers number theorists are also interested in the arith-
metic structure of OQ(√d) . One of the most fundamental questions one can ask
is: ”What are the units in OQ(√d) ?”.
Recall that in a ring R an element α ∈ R is a unit provided that α · β = 1R
for some β ∈ R.

Example The units in Z are ±1.

Example In Z[i], ±1, ±i are units. In fact, these are all the units.
√ √
Definition Given α = u + v d ∈ Q( d) the conjugate of α, denoted by α is
√
α = u − v d.
√ √ √
√For example the conjugate of 1 + 2 is 1 − 2, and the conjugate of 5 is
− 5. Also, α is rational if and only if α = α (why?). Here are some properties
of conjugates.
√
Lemma 2.6 Suppose α, β ∈ Q( d) where d is squarefree then
a) α = β if and only if α = β
√ √
b) αα ∈ Z if α = u + v d ∈ Z[ d]
c) αβ = αβ
√ √
Proof Write
√ α = u +
√ v d and β = r + s d. To prove a) notice that if α = β
then u + v d = r + s d so
√
(u − r) = (v − s) d
since the left hand side is a rational number we must have v − s = 0 (otherwise
the right
√ hand side√ would be irrational). So u = r and v = s. This implies
u − v d = r − s d as desired.
Now consider
√ √
αα = (u + v d)(u − v d) = u2 − dv 2 ∈ Z
whenever u, v ∈ Z. This proves b).
Finally,
√ √ √ √ √
αβ = (u + v d)(r + s d) = ur +vsd−(us+vr) d = (u−v d)(r −s d) = αβ,
which proves c). 

7
 √ 2.7 Suppose that d ≡ 2, 3 (mod 4) is squarefree. An element
Proposition
α = u + v d ∈ OQ(√d) is a unit if and only if

u2 − dv 2 = ±1.

Remark The equation

x2 − dy 2 = ±1
is known as Pell’s equation. Finding the solutions of this equation will be a major
topic of this course. The above proposition shows that finding solutions to Pell’s
√ d ≡ 2, 3 (mod 4) corresponds to finding units
equation for squarefree d, with
in the ring of integers of Q( d).

Example In Z[i], α = u + vi is a unit iff u2 + v 2 = ±1. The only way this can
happen is if u = 0, v = ±1 or u = ±1 and v = 0. Hence the units in Z[i] are
precisely ±1, ±i.
√
Remark
√ It turns out that Z[ 3] has infinitely many units! We’ll see that
(2 + 3)n is a unit for each n ∈ N .
√ √
Proof √(⇒) Suppose that α = u + v d ∈ Z[ d] is a unit. Then there exists
β ∈ Z[ d] such that αβ = 1 so by the lemma, part a), αβ = 1. Combining these
equations gives
1 = αβαβ = (αα)(ββ)
by part c) of the lemma. By part b) we know ββ ∈ Z and also αα = u2 − dv 2 so
we get
1 = (u2 − dv 2 ) · ” an integer”.
Hence u2 − dv 2 is a unit in Z so u2 − dv 2 = ±1.
(⇐) Suppose that u2 − dv 2 = ±1. Then

(αα)2 = (u2 − dv 2 )2 = 1
√
so α(α · α2 ) = 1 so that α is a unit in Z[ d]. 

2.5 Appendix: The proof of Proposition 2.4

Proof Proof of Proposition 2.4√Let α be a quadratic number. By the Proposi-
tion 2.3, we can write α = u + v d, where u, v ∈ Q and d is a squarefree integer
not equal to 1.

8
 Let f (x) = x2√
− ax + b be the minimal polynomial√of α. So one root of
f (x) is α =√u + v d, the other
√ one has to be = u − v d. Therefore, f (x) =
(x − (u + v d))(x − (u − v d)). We conclude that

• a = 2u, and

• b = u2 − dv 2 .

So α is an algebraic integer if and only if 2u, u2 − dv 2 ∈ Z. In particular, 2u ∈ Z

means that either u ∈ Z, or u − 21 ∈ Z, depending on whether 2u is an even or
an odd integer.

Case 1: u ∈ Z. Then dv 2 ∈ Z. Since d is squarefree, this implies that v ∈ Z.

Suppose not. Write v = ab for a, b ∈ Z, b > 1, gcd(a, b) = 1. By assumption,
2
dv 2 = da
b2
∈ Z. So b2 | da2 . Let p be a prime with p | b. Such a prime exists if
b > 1. Then p2 divides b2 . Thus p2 | da2 . But we also have p 6 | a as gcd(a, b) = 1.
It follows that p2 | d. This is a contradiction since d is squarefree.

1 1
Case 2: u − 2
∈ Z. Say u = s + 2
where s ∈ Z. We have

1
s2 + s + − dv 2 ∈ Z,
4
so dv 2 − 41 ∈ Z, and in particular 4v 2 d − 1 ∈ 4Z.
The last condition shows that v ∈ / Z; for if it is, then 4v 2 d − 1 would be odd.
Again, write v = ab for a, b ∈ Z, b > 0, gcd(a, b) = 1. By assumption,
2
4dv 2 = 4da
b2
∈ Z. So b2 | 4da2 . Let p be a prime, n ≥ 1 an integer with pn | b.
Then p2n divides b2 . Thus p2n | 4da2 . But we also have p 6 | a as gcd(a, b) = 1. It
follows that p2n | 4d. As d is squarefree, this implies p2n−1 | 4. Hence p = 2 and
n = 1. So the denominator of v must be 2, and v = t + 12 for some t ∈ Z. Then

u2 − dv 2 = s2 + s + 41 − d(t2 + t + 41 ),

so (1 − d)/4 ∈ Z, whence d ≡ 1 (mod 4).