Section 3

Prime Numbers and Prime

Factorisation

Definition 3.1 A prime number is an integer p > 1 whose only positive

divisors are 1 and p.

Example 3.2 The first few prime numbers are:

2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, . . . .

See the Prime Pages (http://www.utm.edu/research/primes/) for longer

lists of primes and much other interesting information.
The primes are useful since they form the ‘building blocks’ from which
all other integers are made.

Theorem 3.3 (Fundamental Theorem of Arithmetic) Any integer n

with n > 1 can be written uniquely in the form

n = pk11 pk22 . . . pkr r

where the pi are prime numbers with p1 < p2 < · · · < pr and the ki are
positive integers.

Example 3.4 180 = 22 · 32 · 5.

Proof of Theorem 3.3: We shall need some auxiliary results before we

can establish the uniqueness part, but we can establish the existence part
straight away.

We proceed by induction on n. If n is a prime (including n = 2, the base

case in the induction), then already n is a product of prime powers (namely
a single prime) so there is nothing to prove. Assume then that n > 2 and

16
that n factorises, say n = st where 1 < s, t < n. By induction, both s and t
can be written as a product of prime powers. Hence, upon multiplying these
expressions together, we see that n = st is also a product of prime powers.
This establishes existence of a prime decomposition. ¤

For the uniqueness part we prove:

Lemma 3.5 Let p be a prime number.

(i) If p | ab, then either p | a or p | b.

(ii) If p | a1 a2 . . . as , then p | ai for some i.

(iii) If p | q1 q2 . . . qt where each qi is a prime number, then p = qj for

some j.

Proof: (i) Assume p | ab. If p | a then the result holds. So assume that
p does not divide a. Then gcd(p, a) = 1 (since the only divisors of p are
1 and p and the latter does not divide a). Now part (ii) of Theorem 2.6 tells
us that
1 = up + va
for some u, v ∈ Z. Hence
b = ubp + vab.
Now p | ab (by assumption), so we deduce p | (ubp + vab); that is, p | b, as
required.
(ii) Proceed by induction on s. If s = 1, then p | a1 and there is nothing
to prove. Assume then that s > 1. We have p | bas where b = a1 a2 . . . as−1 .
Hence, by (i), either p | b or p | as . If the first holds, that is, p | b, then
by induction p divides one of a1 , a2 , . . . , as−1 . Hence we deduce p | ai for
some i, completing the inductive step.
(iii) Apply (ii). We deduce p | qj for some j. But as qj is prime, its only
divisors are 1 and qj . Hence p = qj . ¤

We now return to the uniqueness part of Theorem 3.3. Assume that we

have expressed n in the required form in two different ways:

n = pk11 pk22 . . . pkr r = q1l1 q2l2 . . . qtlt . (3.1)

Note that pi | n = q1l1 q2l2 . . . qtlt . Hence by Lemma 3.5 we have pi = qj for
some j. By the same argument, each ql is equal to some pm . We conclude
that
r = t, p1 = q1 , p2 = q2 , . . . , pr = qr .
Then Equation (3.1) becomes

pk11 pk22 . . . pkr r = pl11 pl22 . . . plrr .

17
Assume ki 6= li for some i. We may assume without loss of generality that
ki > li . Then dividing through by plii gives
k i−1 ki −li k l l
pk11 . . . pi−1 pi i+1
pi+1 . . . pkr r = pl11 . . . pi−1
i−1 i+1
pi+1 . . . plrr .
l l
Hence pi divides pl11 . . . pi−1
i−1 i+1
pi+1 . . . plrr . This implies that pi = pj (with
i 6= j) by Lemma 3.5(iii). This is a contradiction. Thus ki = li for all i and
we have established the required uniqueness. ¤

We have established that a positive integer can be factorised uniquely as

a product of prime numbers. This decomposition can also be used to find
all the divisors. Suppose
n = pk11 pk22 . . . pkr r
where the pi are prime numbers with p1 < p2 < · · · < pr and ki ∈ N for all i.
Let m | n with m positive, so write
n = mu
for some integer u. We can decompose m as a product of prime powers
m = q1l1 q2l2 . . . qsls .
Then each qi | m, so qi | n and Lemma 3.5 gives
qi = pj for some j.
Thus m is a product of powers of some of the primes occurring in the prime
factorisation of n. The same argument can be applied to u: it too is a
product of powers of some of the pi . Let
m = ps11 ps22 . . . psrr , u = pt11 pt22 . . . ptrr
where si > 0 and ti > 0 for all i. Then
n = mu = ps11 +t1 ps22 +t2 . . . prsr +tr .
The uniqueness of the expression implies
si + ti = ki for all i
and hence
0 6 si = ki − ti 6 ki for all i.

Theorem 3.6 Let n = pk11 pk22 . . . pkr r be a positive integer expressed as a

product of prime powers (so each pi is a prime number and each ki is a
positive integer). The positive divisors of n are precisely the numbers of the
form
ps11 ps22 . . . psrr
where 0 6 si 6 ki for i = 1, 2, . . . , r. ¤

18
Example 3.7 The divisors of 180 are:

1 = 20 · 30 · 50 9 = 20 · 32 · 50 15 = 20 · 31 · 51
2 = 21 · 30 · 50 18 = 21 · 32 · 50 30 = 21 · 31 · 51
4 = 22 · 30 · 50 36 = 22 · 32 · 50 60 = 22 · 31 · 51
3 = 20 · 31 · 50 5 = 20 · 30 · 51 45 = 20 · 32 · 51
6 = 21 · 31 · 50 10 = 21 · 30 · 51 90 = 21 · 32 · 51
12 = 22 · 31 · 50 20 = 22 · 30 · 51 180 = 22 · 32 · 51

We now turn to consider the properties of the prime numbers. We have

already observed that they are the basic building blocks from which all other
positive integers are constructed via multiplication. Furthermore, they have
been recognised as important and have been studied since ancient times.
Many deep theorems have been proved about them, often involving startling
and surprising methods. On the other hand, there are many questions about
them which are easy to state but which have still not been answered. For
the rest of this lecture I shall begin to examine some of these theorems and
questions. We begin with an important fact which was known by Euclid.

Theorem 3.8 There are infinitely many prime numbers.

Proof: Assume for a contradiction that there are only finitely many primes,
say p1 , p2 , . . . , pn . Consider

N = p1 p2 . . . pn + 1.

This number N must be divisible by a prime, say pi divides N . Then

pi | (N − p1 p2 . . . pn ) = 1,

which is a contradiction. ¤

The argument can be modified in a number of ways to show that there

are infinitely many primes of certain forms. First note that as 2 is the only
even prime, it follows that all primes greater than 2 either have the form
4k + 1 or 4k + 3. We can prove:

Theorem 3.9 There are infinitely many prime numbers of the form 4k + 3.

Proof: Suppose that p1 , p2 , . . . , pn are all the primes of the form 4k + 3.

Consider
N = 4p1 p2 . . . pn − 1 = 4(p1 p2 . . . pn − 1) + 3.
Then N is a product of prime numbers, all of which must be odd. Note that

(4k + 1)(4l + 1) = 4kl + 4k + 4l + 1 = 4(kl + k + l) + 1.

19
Hence if N were a product of primes all of which had the form 4k + 1, then
N would also have this form. Therefore one of the prime divisors of N must
have the form 4k + 3, so is pi for some i. Then

pi | (4p1 p2 . . . pn − N ) = 1,

a contradiction. This completes the proof. ¤

Similarly it is possible to show that every prime greater than 3 has the
form 6k + 1 or 6k + 5 and modifying the above argument will show that
there are infinitely many primes of the form 6k + 5. In fact, there are also
infinitely many primes of the form 4k + 1 and infinitely many primes of the
form 6k + 1, but this is much harder to prove. These are all special cases of
the following theorem proved by Dirichlet:

Theorem 3.10 (Dirichlet 1837) If a and b are coprime positive integers

then there are infinitely many prime numbers of the form ak + b (k = 0, 1,
2, . . . ).
The proof of this theorem is well beyond what we can hope to prove
in this course, but it does at least give a flavour of the development of
mathematics.
Consider the sequence of numbers of the form 4k + 3:

3, 7, 11, 15, 19, 23, 27, 31, 35, . . . .

We are not claiming that they are all prime, but just that infinitely many
of them are: primes will contine to occur in this list no matter how far we
go along it.
In fact there is no known simple formula which yields only prime numbers
whatever we substitute into it. For some time mathematicians believed that

f (n) = n2 + n + 41

was such a formula having checked that f (n) is prime for n = 1, 2, 3, . . . , 39.
(Here f (39) = 1601.) However this is, of course, not good enough and we
find:

f (40) = 402 + 40 + 41 = 40(40 + 1) + 41 = 40 · 41 + 41 = 412 ,

which is not prime. In fact we can prove that there is no polynomial formula
which generates primes.

Theorem 3.11 There is no polynomial f (n) with integer coefficients which

is not constant and which takes only prime values for all non-negative inte-
gers n.

20
Proof: Assume that

f (n) = ak nk + ak−1 nk−1 + · · · + a1 n + a0

is such a polynomial. Then f (0) = a0 is prime, and so is f (ta0 ) for all

choices of t = 1, 2, . . . . But

f (ta0 ) = ak tk ak0 + ak−1 tk−1 ak−1

0 + · · · + a1 ta0 + a0

and therefore a0 | f (ta0 ) for all t. Since f (ta0 ) is prime this forces

f (ta0 ) = a0 for t = 1, 2, . . . .

Hence the polynomial f (n) takes the value a0 infinitely many times and
therefore f (n) must be the constant polynomial taking value a0 . ¤

The other thing that we might ask about is the distribution of primes.
Although we have observed it is difficult to generate primes, one can actually
say quite a lot about their distribution.

Theorem 3.12 Let pn denote the nth prime number. Then

n−1
pn 6 22 .

This result is pointing to the fact that primes do occur reasonably often.

Proof: We proceed by induction on n. To start with we have

1−1
p1 = 2 = 22 .

The inductive step relies on the method of proof of Theorem 3.8. We know
that there is a prime number q which divides

p1 p2 . . . pn + 1

and this cannot be one of the first n prime numbers. Hence

pn+1 6 q 6 p1 p2 . . . pn + 1
n−1
6 2 · 22 · . . . · 2 2 +1
1+2+···+2n−1
=2 +1
2n −1
=2 +1
2n −1 n −1 n
=2 + 22 = 22 .

This completes the proof. ¤

21
 A much stronger (and far harder to prove) result is that

n log n
lim = 1.
n→∞ pn
This is one of the equivalent formulations of the famous Prime Number
Theorem (proved in 1896 by Hadamard and de la Vallée Poussin). It can
be interpreted as saying that for large values of n, the nth prime is roughly
nearby the number n log n. (This being the natural logarithm.)
It is certainly worth mentioning a few open questions that mathemati-
cians have still yet to solve:

Goldbach’s Conjecture: Is it true that every even number greater than 2

can be written as the sum of two prime numbers?

Twin Primes Conjecture: Is it true that there are infinitely many prime
numbers p such that p + 2 is also prime?

We can interpret the Prime Number Theorem as saying that there are
many prime numbers and that they occur rather regularly. The above con-
jectures also assert similar things. However, we finish our discussion of
primes by showing that there are large gaps where no primes occur. Namely
we prove:

Theorem 3.13 For every positive integer n, there is a sequence of n con-

secutive composite numbers.
(‘Composite’ means a number that is not prime; i.e., composed as a
product of more than one prime number.)

Proof: Consider the following numbers:

(n + 1)! + 2 divisible by 2
(n + 1)! + 3 divisible by 3
..
.
(n + 1)! + (n + 1) divisible by n + 1.

These n consecutive numbers are all composite. ¤

22