Module 4 The Fundamental Theorem of Arithmetic

The Fundamental Theorem

of Arithmetic
4.1 The Theorem of Ages
From the earliest grades, teachers of mathematics try to enlighten their students to the basic
principles of arithmetic. Every natural number, we are told, can be factored as a product of
primes. The primes themselves, we are told, are distinguished by the property that they cannot
be factored as a product of strictly smaller natural numbers. The Fundamental Theorem of
Arithmetic asserts something very special about the factorization of natural numbers into prime
factors. It asserts that the factorization is essentially unique, which is to say that if we ignore the
order of the prime factors, then there is one and only one way to accomplish the factorization.
The Fundamental Theorem of Arithmetic is widely taken for granted. Evidence suggests that
it was known to all cultural groups, at all stages of civilization, as far back as recorded history
permits us to probe. It enjoys the distinction of being the oldest and most cosmopolitan of all
fundamental principles in mathematics. It has been called the Theorem of Ages.

4.2 Elements of Arithmetic

To begin, we must establish exactly what is meant by the phrase “ b is a factor of a ” or, equiv-
alently, “ b is a divisor of a ”. By definition, we say that b ∈ Z is a divisor of a ∈ Z iff there
exists c ∈ Z such that a = bc . If b is a divisor (or factor) of a , then we denote this relation
symbolically as b | a . It is customary to read the symbolic string “ b | a ” as “ b divides a ”.
Equivalently, it is also legitimate to say that “ a is divisible by b ”.
Let a , b and c be integers. The following basic properties of the divisibility relation are easily
verified:
(a) If c | b and b | a , then c | a .
(b) If c | a and c | b , then c | ax + by for all x, y ∈ Z .
(c) If x ∈ Z and x | 1 , then x = ±1 .
Occasionally, we may want to talk about the set of all positive divisors of an integer a ∈ Z .
For this purpose, we introduce the notation div(a) = {b ∈ N | b is a positive divisor of a} . For
instance, the set of all positive divisors of a = 6 is div(6) = {1, 2, 3, 6} .
Scattered among the set of natural numbers N = {1, 2, 3, . . .} , there are special numbers called
primes, which are distinguished by the property that they cannot be factored as a product of
strictly smaller natural numbers. By definition, a natural number p ≥ 2 is prime iff whenever
p = ab , where a, b ∈ N , then either a = 1 and b = p or a = p and b = 1 . In terms of the
above symbolism, a natural number p is prime iff div(p) = {1, p} . It may seem fairly obvious
that every natural n ≥ 2 must possesses at least one prime factor, but without witnessing a
rigorous proof, no self-respecting student of mathematics should take this for granted.

31
Module 4 The Fundamental Theorem of Arithmetic

Theorem 4.1: Every natural number n ≥ 2 has at least one prime divisor.
Proof: If n is prime, then there is nothing to prove. Otherwise, if n is not prime, then it
must possess at least one non-trivial divisor d such that 1 < d < n . Let d0 be the smallest
non-trivial divisor of n . If d0 is prime, then the proof is complete. Otherwise, If d0 is not
prime, then d0 must possess a divisor d1 such that 1 < d1 < d0 . But, if so, then d1 must also
be a non-trivial divisor of n . But since d1 < d0 , this contradicts our assumption that d0 is the
smallest non-trivial divisor of n . Thus d0 must be prime, and the proof is complete.

Does the set of primes {2, 3, 5, 7, 11, 13, . . .} go on forever, or do we eventually encounter a
largest prime pmax beyond which no more primes can be found? The first proof in recorded
history of the fact that the set of primes is infinite is due to Euclid of Alexandria (circa 365-275
B.C.). It is one of the most famous proofs in the history of mathematics.
Theorem 4.2: (Euclid) The set of primes is infinite.
Proof: Suppose that pmax is the largest prime in the world. We will show that this supposition
leads to a contradiction. Consider the natural number q obtained by multiplying together all
the primes in the world, from smallest to the largest and adding 1 to the product. That is
q = 2 × 3 × 5 × 7 × 11 × · · · × pmax + 1.
Now plainly q > pmax . Since pmax is the largest prime, q itself cannot be prime. Thus, by
Theorem 4.1, q possesses at least one prime factor, say p . Since q − 1 is the product of all
primes in the world, it follows that p | q − 1 . Now p | q and p | q − 1 implies that p | 1 , which
is a contradiction. Therefore there can be no largest prime.

The primes {2, 3, 5, 7, 11, 13, . . .} are scattered among the natural numbers in a very irregular
way. Any attempt to generate them by means of a simple formula or to try to fit them to a
recognizable pattern is bound to fail. However, an efficient algorithm does exist for generating
the list of all primes up to a given natural number n . The algorithm is named after Eratosthenes
of Cyrene (276-194 B.C.), as he is the first person in recorded history to have explicitly described
it in his writings. The sieve of Eratosthenes (described below) is a consequence of the following
very simple theorem.
√
Theorem 4.3: Let n ≥ 2 . If n is not prime, then n possesses a prime divisor p ≤ n.
Proof:
√ Since n is not prime, it must possess a prime divisor, say p , such√that 1 < p < n . If
p ≤ n , then there is nothing further
√ to show. Otherwise, suppose p > n . Let n1 = n/p .
Then plainly n1 | n and
√ 1 < n 1 < n . Let p1 be any prime divisor of n1 . Then p1 is also a
divisor of n and p1 < n .

To generate a list of all primes up to given natural number n , proceed as follows.√First, you may
assume that you have already managed to generate a list of all primes up to n . To extend
the list all the way up to n , proceed down the list 1, 2, 3, 4, . . . , n , crossing out all multiples of
2 (except 2 itself), then crossing out all multiples of 3 (except 3 itself), then crossing out all
multiples of 5 (except 5 itself), and so on.√Continue in this fashion until you have crossed out
all the multiples of the largest prime p ≤ n . At this point, by Theorem 4.3, you have crossed
out all the non-prime (composite) numbers in the list 1, 2, 3, 4, . . . , n , leaving only the primes.

32
Module 4 The Fundamental Theorem of Arithmetic

Every respectable discussion of arithmetic must, at some point, turn to the idea of the “greatest
common divisor” of two integers. Given a, b ∈ Z , the greatest common divisor of a and b ,
denoted by gcd(a, b) , is defined as the largest member of the set div(a) ∩ div(b) . That is

gcd(a, b) = max (div(a) ∩ div(b)) .

If gcd(a, b) = 1 , then we say that a and b are relatively prime.

Ex: By definition, gcd(40, 90) = max (div(40) ∩ div(90)) . By direct calculation, we have
div(40) = {1, 2, 4, 5, 8, 10, 20, 40}
div(90) = {1, 2, 3, 5, 6, 9, 10, 15, 18, 30, 45, 90}
div(40) ∩ div(90) = {1, 2, 5, 10}
Therefore gcd(40, 90) = max{1, 2, 5, 10} = 10 .

Ex: By definition, gcd(40, 91) = max (div(40) ∩ div(91)) . By direct calculation, we have
div(40) = {1, 2, 4, 5, 8, 10, 20, 40}
div(91) = {1, 7, 13, 91}
div(40) ∩ div(90) = {1}
Therefore gcd(40, 91) = max{1} = 1 . Thus, 40 and 91 are relatively prime.

The above examples serve to illustrate how to calculate gcd(a, b) directly from the definition.
This method is extremely inefficient, as it forces us to find complete factorizations for each of
the integers a and b , a process which, from a computational standpoint, happens to be very
costly. A far more efficient technique is that known as the Euclidean algorithm. The Euclidean
algorithm is aptly named after Euclid of Alexandria, the first individual in recorded history
to have documented it, and quite possibly the first to have discovered it. Given two integers
a > b > 0 , to calculate gcd(a, b) by the Euclidean algorithm, we proceed as follows:
Step 1: a = q1 b + r1 ( 0 ≤ r1 < b )
Step 2: b = q2 r1 + r2 ( 0 ≤ r2 < r1 )
Step 3: r1 = q3 r2 + r3 ( 0 ≤ r3 < r2 )
Step 4: r2 = q4 r3 + r4 ( 0 ≤ r4 < r3 )
..
.
Step n : rn−2 = qn rn−1 + rn ( 0 ≤ rn < rn−1 )
Step n + 1 : rn−1 = qn+1 rn + rn+1 ( rn+1 = 0 ).

In each of the above steps, say step k , what we are doing is dividing the remainder (rk−2 )
obtained two steps earlier by the remainder (rk−1 ) obtained one step earlier, to obtain a new
remainder (rk ) . The successive quotients are denoted q1 , q2 , q3 , and so on. Since successive
remainders are non-negative and decreasing, that is b > r1 > r2 > r3 > · · · > rk ≥ 0 , the process
must eventually terminate at a remainder of 0. The last non-zero remainder rn is equal to
gcd(a, b) . To see why, suppose d is a divisor of rn . By tracing the Euclidean algorithm in
reverse, we deduce that d must be a divisor of both a and b , hence of gcd(a, b) . Conversely,
if d is any divisor of gcd(a, b) , then, by tracing the Euclidean algorithm in forward order, we
deduce that d is a divisor of rn . Thus div(gcd(a, b)) = div(rn ) , which implies gcd(a, b) = rn .

33
Module 4 The Fundamental Theorem of Arithmetic

Ex: To calculate gcd(91, 40) by the Euclidean algorithm, we proceed as follows:

91 = 2 × 40 + 11
40 = 3 × 11 + 7
11 = 1 × 7 + 4
7=1×4+3
4=1×3+1
3=3×1+0
Therefore gcd(91, 40) = 1 .

From the Euclidean algorithm, we can infer the following fundamental result, whose importance
in elementary number theory is second only to the Fundamental Theorem of Arithmetic:
Theorem 4.4: Let a, b ∈ Z and suppose gcd(a, b) = d . Then there exist integers x and y
such that ax + by = d .
Proof: Without loss of generality, we may assume that a and b are positive, for, if not, the
signs of a and b can be adjusted by absorbing them into the coefficients x and y . By the
Euclidean algorithm, we have d = rn , where rn is the last non-zero remainder in the algorithm.
Now, using the the n -th step in the algorithm we can express d as a linear combination of rn−1
and rn−2 . Similarly, using the (n − 1) -th step of the algorithm, we can express express rn−1
as a linear combination of rn−2 and rn−3 , so that d becomes a linear combination of rn−2 and
rn−3 . Similarly, using the (n − 2) -th step of the algorithm, we can express rn−2 as a linear
combination of rn−3 and rn−4 , so that d becomes a linear combination of rn−3 and rn−4 .
Continuing in the same manner, we see that, at every step of the algorithm, d can be expressed
as a linear combination of rk and rk−1 . Ultimately, back at step 1, we have d = ax + by .

Corollary 4.5: Two integers a and b are relatively prime ( gcd(a, b) = 1 ) if and only if there
exist integers x and y such that ax + by = 1 .
Proof: If gcd(a, b) = 1 , then, by Theoren 4.4, gcd(a, b) = ax + by = 1 . Conversely, suppose
ax + by = 1 . If gcd(a, b) = d , then d | a and d | b , so d | ax + by = 1 , which implies d = 1 .

Ex: Since gcd(40, 91) = 1 , there exist integer coefficients x and y such that 40x + 91y = 1 .
To find them, we proceed, as in the proof of Theorem 4.4, by unraveling the Euclidean
algorithm in reverse. To make it easier to distinguish the remainders in each step, we adorn
them by dressing them in square brackets:
[4] = 1 × [3] + [1] ⇒ [1] = −[3] + [4]
[7] = 1 × [4] + [3] ⇒ [3] = −[4] + [7] ⇒ [1] = 2[4] − [7]
[11] = 1 × [7] + [4] ⇒ [4] = −[7] + [11] ⇒ [1] = −3[7] + 2[11]
[40] = 3 × [11] + [7] ⇒ [7] = −3[11] + [40] ⇒ [1] = 11[11] − 3[40]
[91] = 2 × [40] + [11] ⇒ [11] = −2[40] + [91] ⇒ [1] = −25[40] + 11[91]
Therefore 40(−25) + 91(11) = 1 . That is x = −25 and y = 11 .

34
Module 4 The Fundamental Theorem of Arithmetic

Corollary 4.6: If gcd(a, b) = 1 then gcd(a2 , b2 ) = 1 .

Proof: By Theorem 4.4, we can find x, y ∈ Z such that ax + by = 1 . Raising both to the 3-
rd power, we get x3 a3 + 3x2 a2 yb + 3xay 2 b2 + y 3 b3 = 1 , which is equivalent to a2 x1 + b2 y1 = 1 ,
where x1 = x3 a + 3x2 yb and y1 = 3xay 2 + y 3 b . Thus, by Corollary 4.5, gcd(a2 , b2 ) = 1 .

To the mathematicians of antiquity, especially Pythagoras and his followers (circa 550 B.C.),
it came as a great shock to discover that some numbers are not rational (cannot be expressed
as the ratio of two whole numbers). Numbers that are not rational are called irrational. Using
√
Corollary 4.6, we can show that p is irrational for every prime p .
√
Theorem 4.7: If p is prime, then p is irrational.
√
Proof: Suppose p is rational. We will show that this supposition leads to a contradiction.
√ √
If p is rational, then it can be expressed as a fraction in lowest terms, say p = a/b where
gcd(a, b) = 1 . By Corollary 4.6, we have gcd(a2 , b2 ) = 1 and we are assured that integers x1
√
and y1 exist such that a2 x1 + b2 y1 = 1 . The relation p = a/b implies a2 = b2 p , so that
b2 px1 + b2 y1 = 1 . Therefore b2 | 1 , which implies b2 = 1 . Thus p = a2 , which contradicts the
assumption that p is prime.

Using Theorem 4.4, we can derive an alternate characterization of what it means for a natural
number n to be prime. This alternate characterization of primality plays an essential role in
the proof of the Fundamental Theorem of Arithmetic.
Theorem 4.8: A natural number n is prime if and only if n possesses Property P: whenever
n | ab , where a, b ∈ N , then n | a or n | b .
Proof: If n is not prime (i.e. composite), then n = ab for some natural numbers a and b ,
where 1 < a < n and 1 < b < n . So n 6 | a and n 6 | b . In particular, this means that if n is
not prime, then n does not possess Property P. Conversely, suppose n is prime and n | ab .
Then ab = nk for some k ∈ Z . If n 6 | a , then gcd(n, a) = 1 . Thus, by Theorem 4.4, there exist
integers x and y such that nx + ay = 1 . So nbx + aby = b , which implies that nbx + nky = b ,
which implies that n | b . Thus, if n is prime, then n possesses Property P.

Corollary 4.9: If p is prime and p | a1 a2 a3 · · · am , then p | aj for some 1 ≤ j ≤ m .

Proof: We proceed by induction on m . By Theorem 4.8, the claim is valid for m = 2 . Suppose
the claim is valid for m = k . If p | a1 a2 a3 · · · ak ak+1 , then p | ak+1 or p | a1 a2 a3 · · · ak . That
is, p | ak+1 or p | aj for some 1 ≤ j ≤ k .

Corollary 4.10: If p is prime and p | am , where m ≥ 1 , then p | a .

Proof: Apply Corollary 4.9 with a = a1 = a2 = · · · = am .

Corollary 4.11: If p and q are primes and p | q m , where m ≥ 1 , then p = q .

Proof: By Corollary 4.10, p | q . However, since div(q) = {1, q} , it follows that p = q .

35
Module 4 The Fundamental Theorem of Arithmetic

4.3 The Fundamental Theorem of Arithmetic

Suppose we are given natural number n and suppose we are able to factor it as a product of
primes, say n = p1 p2 p3 · · · . In general, there are many ways to write the product, depending
on how we choose to order the factors and depending on how we choose to arrange the factors
in terms of prime powers. For instance, 360 = 2 · 2 · 2 · 3 · 3 · 5 = 3 · 5 · 2 · 3 · 2 · 2 = 32 · 51 · 23 .
The preferred form of factorization is in terms of prime powers, where the distinct primes in the
factorization are arranged in increasing order from left to right. That is
n = pr11 pr22 · · · prkk where p1 < p2 < p3 < · · ·
A factorization like the above is called a product of prime powers in standard arrangement. For
instance, if n = 360 is expressed as a product of prime powers in standard arrangement, we get
360 = 23 · 32 · 51 .

Theorem 4.12: (The Fundamental Theorem of Arithmetic) Every natural number n ≥ 2 can
be expressed in exactly one way as a product of prime powers in standard arrangement.
Proof: First we must show that n can be factored as a product of primes. By Theorem 4.1,
there exists a prime p1 such that n = p1 n1 , where n > n1 ≥ 1 . Similarly, by Theorem 4.1, there
exists a prime p2 such that n1 = p2 n2 , where n > n1 > n2 ≥ 1 . Continuing in this manner, we
see that n can be expressed as n = p1 p2 p3 · · · pk nk where n > n1 > n2 > · · · > nk ≥ 1 . Since
the sequence of numbers nj are decreasing and positive, they must eventually terminate, at
which point the factorization of n as a product of primes is complete. By regrouping the factors
as prime powers in standard arrangement, we obtain n = pr11 pr22 · · · prkk . Now, we still have to
show that this representation is unique. Suppose an alternate representation of n as a product
of prime powers in standard arrangement is possible, say n = q1s1 q2s2 · · · qm
sm
. We will show that
these two representations are, in fact, one and the same. By Corollaries 4.9-4.11, every prime
s
pi in the first factorization divides some prime power qj j in the second factorization, so that
r
pi = qj . Similarly, every prime qi in the second factorization divides some prime power pj j in
the first factorization, so that qi = pj . Thus, the primes that appear in the first factorization
are exactly the same as the primes that appear in the second factorization. As a result, we have
n = pr11 pr22 · · · prkk = ps11 ps22 · · · pskk .
At this point, we must still show that corresponding exponents are equal. Suppose, by way
of contradiction, that ri < si for some i , 1 ≤ i ≤ k . If we divide the first factorization by
pri i , then the prime pi no longer occurs in the factorization of n/pri i . On the other hand,
if we divide the second factorization by pri i , then, since ri < si , the prime pi still occurs in
the factorization of n/pri i with non-zero exponent si − ri . This implies that pi divides one of
r
the prime power factors pj j in the first factorization where pi 6= pj , which, by Corollary 4.11,
implies that pi = pj , a contradiction. Therefore, all corresponding exponents must be equal,
and the proof of the theorem is complete.

Ex: Using the Fundamental Theorem of Arithmetic, we show that log72 (175) is irrational. By
way of contradiction, suppose that log72 (175) is rational, say log72 (175) = m/n where m
and n are positive whole numbers. Then 175 = 72m/n , which implies 175n = 72m , so that
n m
(52 · 7) = (23 · 32 ) . Hence 52n · 7n = 23m · 32m , which violates the unique factorization
principle of the Fundamental Theorem of Arithmetic.

36
Module 4 The Fundamental Theorem of Arithmetic

4.4 Exercises
1. Decide if the claim is true or false.
(a) 9 | 4203 .
(b) 12 | 4 .
(c) 7 | 0 .
(d) 3 | −18 .
(e) −6 ∈ div(18) .
(f) −3 ∈ div(−6) .
(g) div(5) ∩ div(7) = ∅ .
√
98
(h) √ is a rational number.
3 8
(i) The smallest positive prime number is 1.
(j) A prime number cannot be even.
2. Use the Euclidean algorithm to find gcd(a, b) .
(a) a = 495 and b = 147
(b) a = 729 and b = 512
3. Use the Euclidean algorithm in reverse to find integers x and y such that ax + by = gcd(a, b) .
(a) a = 495 and b = 147
(b) a = 729 and b = 512
4. Express n as a product of prime powers in standard arrangement.
(a) n = 337500
(b) n = 11307375
(c) n = 510510
(d) n = 216 − 1
(e) n = 15!
5. Prove or disprove the claim that every integer of the form n2 − n + 41 is prime for all n ≥ 0 .
6. Show that log3850 (4760) is irrational.
7. Show that if gcd(a, b) = 1 then gcd(a3 , b3 ) = 1 .
√
8. Prove that 3 p is irrational for every prime p .
9. Show that gcd(3x + 8, 5x + 13) = 1 for every integer x .
x y
10. Find two non-zero integers x and y such that + = 1.
81 25

37