CONSTRUCTION OF THE FINITE FIELDS Zp

S. R. DOTY

Elementary Number Theory

We begin with a bit of elementary number theory, which is concerned
solely with questions about the set of integers Z = {0, ±1, ±2, ±3, . . . }.
We are not concerned with real numbers or fractions here.
We need the concept of divisibility in Z, the division algorithm, and
the Euclidean algorithm.

A.1. Definition. Let a, b ∈ Z. Say that a divides b (written a | b) if

there exists some k ∈ Z such that b = ak. We also say that b is divisible
by a, or that b is a multiple of a, when this is so. If b is divisible by a,
we also say that a is a divisor or factor of b.
Note that any integer divides 0.

A.2. Lemma. (Division Algorithm) Let m and n be given positive

integers. There exist unique integers q, r such that
m = qn + r (0 ≤ r < n).
The integers q, r are called the quotient and remainder, respectively.
The procedure by which we obtain them is called long division, as you
undoubtedly recall.

A.3. Definition. A positive integer p > 1 is prime if it has no positive

integer divisors other than 1 and p.
Note that we explicitly exclude 1 from the set of prime numbers.
The number 1 is not a prime! The primes in Z are 2, 3, 5, 7, 11, 13,
17, 19, 23, 29, 31, 37, 41, 43, 47 and so on. There are infinitely many
primes in Z.

A.4. Definition. Let a, b ∈ Z. Any integer d such that d | a and d | b

is called a common divisor of a, b. The greatest common divisor (gcd)
of a, b is the largest of the common divisors. Write gcd(a, b) or more
simply just (a, b) for the gcd of a, b.
Note that the gcd of the pair 0, 0 is undefined, since there is no
largest common divisor, but the gcd is well-defined in every other case.
What is gcd(n, 0) when n 6= 0?
1
2 S. R. DOTY

People say that a, b are relatively prime if their gcd is 1. This means
that they have no prime factors in common, since if p | a and p | b then
(a, b) ≥ p is surely larger than 1.
One way to compute the gcd of two given positive integers is to make
a list of the divisors for each given number, and then the gcd is simply
the largest number common to both lists. Of course, this is virtually
impossible if the two numbers are large, but it works well enough for
small numbers. For instance, let us compute the gcd of the pair 51, 33
using this idea. The divisors of 49 are 1, 7, 49 and the divisors of 210
are 1, 2, 3, 5, 6, 7, 10, 14, 15, 21, 30, 42, 70, etc. The largest number
that is common to both lists is 7, so 7 = gcd(49, 210).
Another method is to factor each given number into its prime factor-
ization. Although this is virtually impossible for very large numbers,
it is pretty easy to do for small numbers. For instance, let us compute
the gcd of the pair 209, 1368 using this idea. You can check with your
calculator or by pen and paper that 209 = 11 · 19 and 1368 = 23 · 32 · 19.
Thus it follows immediately that gcd(209, 1368) = 19. Note that this
method, while better than the first method, still falls apart if we can-
not find the prime factors of the numbers. (Please tell me the prime
factors of 1273932017264838292093 if you doubt that this is a serious
obstacle.)
A third way to compute the gcd is to use the Euclidean algorithm.
It turns out that for large numbers this third way is far and away
the best, in terms of efficiency. For instance, to compute the gcd of
two random 100 digit numbers by a digital computer using the first or
second method would require so many divisions that even the fastest
computer would not compute an answer in a reasonable time. (It would
take many years, even for the fastest supercomputer.) However, even
a desktop PC could obtain an answer to the same problem using the
Euclidean algorithm in a second or so, because only a few hundred
divisions would be required! Now that you are convinced the Euclidean
algorithm is a great thing for mankind, let’s examine how it works.

A.5. Euclidean algorithm. Suppose you need to compute the gcd

of integers a, b. For simplicity, assume both a, b > 0 and that a is
the larger of the two. Repeatedly apply the division algorithm to get
equations:
a = bq + r (0 ≤ r < b)
b = rq1 + r1 (0 ≤ r1 < r)
r = r 1 q2 + r 2 (0 ≤ r2 < r1)
 CONSTRUCTION OF THE FINITE FIELDS Zp 3

r1 = r 2 q 3 + r 3 (0 ≤ r3 < r2)
..
.

The process terminates when you get a remainder of 0. The gcd of a, b

is the last nonzero remainder in the above procedure.
The Euclidean algorithm, probably one of the first algorithms ever
discovered, is the fastest known algorithm for computing the gcd of
any two given numbers. It has a surprising number of applications to
computer science, and is extremely useful in cryptography.
As an application of the Euclidean algorithm we have the following
important result.

A.6. Theorem. Let a, b ∈ Z, not both zero. There exist integers s, t

such that gcd(a, b) = sa + tb. (In other words, gcd(a, b) is expressible
as an integral linear combination of a, b.)

Proof. The Euclidean algorithm equations can be worked backwards,

with one substituted into another, until you reach this expression. To
be specific, start with the last equation in which the remainder is
nonzero. This can be rewritten to express that last nonzero remainder
(the gcd) as a linear combination of the two remainders which precede
it:
gcd(a, b) = rn = rn−2 − rn−1 qn .
This is the base step, and it expresses gcd(a, b) as a linear combination
of rn−1 and rn−2 .
Similarly, there is an equation which expresses rn−1 as a linear com-
bination of the two remainders which precede it:

rn−1 = rn−3 − rn−2 qn−1 .

and putting that linear combination into the gcd equation above and
grouping terms gives us

gcd(a, b) = rn−2 − (rn−3 − rn−2 qn−1 )qn

= (−qn )rn−3 + (1 + qn−1 qn )rn−2

which shows that gcd(a, b) is expressible as a linear combination of rn−2

and rn−3 .
One can repeat the process by induction until one arrives at a linear
combination of a and b. 
4 S. R. DOTY

A.7. Example. Here is an example which more clearly illustrates the

proof of the previous theorem. Let’s compute gcd(1521, 2004) using
the Euclidean algorithm. Divide the bigger number by the smaller to
get
(1) 2004 = 1 · 1521 + 483.
Here the quotient is 1 and the remainder is 483. Now we divide 1521
by 483 to obtain
(2) 1521 = 3 · 483 + 72.
Continuing, we divide 483 by 72 to obtain
(3) 483 = 6 · 72 + 51.
Dividing 72 by 51 we obtain
(4) 72 = 1 · 51 + 21.
Dividing 51 by 21 we obtain
(5) 51 = 2 · 21 + 9.
Finally, dividing 21 by 9 we obtain
(6) 21 = 2 · 9 + 3.
The process stops here because we will obtain a remainder of zero on
the next division. This proves that gcd(1521, 2004) = 3, which is the
last nonzero remainder obtained by the process.
Now we work the equations backwards to express gcd(1521, 2004) as
a linear combination of 1521 and 2004, as follows. By the last equation
(6) we have, solving for the remainder:
3 = 21 − 2 · 9.
Substitute 9 = 51 − 2 · 21, which comes from (5), to obtain
3 = 21 − 2(51 − 2 · 21) = (−2)51 + (5)21;
substitute 21 = 72 − 51, which comes from (4), to obtain
3 = (−2)51 + (5)(72 − 51) = (5)72 + (−7)51;
substitute 51 = 483 − 6 · 72, which comes from (3), to obtain
3 = (5)72 + (−7)(483 − 6 · 72) = (−7)483 + (47)72;
substitute 72 = 1521 − 3 · 483, which comes from (2), to obtain
3 = (−7)483 + (47)(1521 − 3 · 483) = (47)1521 + (−148)483;
substitute 483 = 2004 − 1521, which comes from (1), to obtain
3 = (47)1521 + (−148)(2004 − 1521) = (−148)2004 + (195)1521.
 CONSTRUCTION OF THE FINITE FIELDS Zp 5

This is the desired linear combination, since we have arrived at an

expression gcd(1521, 2004) = 1521s+2004t with s = 195 and t = −148.
Now we apply the theorem to prove a fundamental fact about divis-
ibility in the integers.

A.8. Lemma. (Euclid) If a | (bc) and gcd(a, b) = 1 then a | c.

Proof. By Theorem A.6, there exist integers s, t such that 1 = as + bt.
So c = cas + cbt. Since a | bc, both summands in the right-hand-side
of the preceding equality are divisible by a. Thus c is divisible by a, as
desired. 

A.9. Definition. Let a, b ∈ Z. A common multiple of a, b is any integer

m such that a | m, b | m. The least common multiple (lcm) of a, b is
the smallest positive common multiple.
For example, lcm(5, 12) = 60, lcm(16, 48) = 48, and lcm(18, 99) =
198. The relation between lcm and gcd is the following.

A.10. Theorem. Let a, b ∈ Z. Then lcm(a, b) = ab/ gcd(a, b).

Proof. Try to find your own proof of this theorem, as a useful test of
your understanding of divisibility. 

Modular Arithmetic
Next we come to the study of modular arithmetic and congruences.
This is an important application of our brief study of number theory.
Before giving the technical theory, it is perhaps worthwhile to make
some comments on the intuition behind modular arithmetic.
The basic idea in modular arithmetic is embodied in the ordinary
12 hour clock. We all know that if it is currently 11 o’clock, then two
hours from now will be 1 o’clock, because at 12 o’clock we start over
again at zero. We can symbolize this situation by writing the equation
11 + 2 = 1 (mod 12), or 11 + 2 ≡ 1 (mod 12). (Either notation is
common.)
To do modular arithmetic (that is, addition, subtraction, and mul-
tiplication) mod 12 we just compute as usual in Z taking into account
that all multiples of 12 are the same as zero. Thus we have 10 + 5 ≡ 3
(mod 12), 10 · 5 ≡ 2 (mod 12), and 5 − 10 ≡ −5 ≡ 7 (mod 12). Note
that the preferred answer for any modular arithmetic calculation is a
natural number between 0 and 11 (inclusive). These are the same as
the numbers on the face of the clock, except that 12 has been replaced
by zero.
6 S. R. DOTY

Modular arithmetic mod n (a positive integer) works exactly the

same way, except that we replace the number 12 by the number n.
Thus the preferred answer for a modular arithmetic calculation mod n
is an integer in the closed interval [0, n − 1]. Such integers are called
residues mod n.
A.11. Definition. (Gauss) Let n be a given fixed positive integer. For
a, b ∈ Z we say that a is congruent to b modulo n if n divides a − b.
Write a ≡ b (mod n) when this is so. Some folks write a = b (mod n)
instead.
A.12. Theorem. Congruence modulo n is an equivalence relation on
the set Z. That is:
(a) a ≡ a (mod n) (reflexivity)
(b) a ≡ b (mod n) ⇒ b ≡ a (mod n) (symmetry)
(c) a ≡ b (mod n), b ≡ c (mod n) ⇒ a ≡ c (mod n) (transitivity)
for all a, b, c ∈ Z.
Proof. The proof is routine, and left for you to check. 
A.13. Theorem. Suppose that a ≡ b (mod n) and c ≡ d (mod n).
Then:
(a) a + c ≡ b + d (mod n)
(b) ac ≡ bd (mod n).
Proof. Once again, the proof is a routine exercise. 
A.14. Theorem. If gcd(a, n) = 1 then the congruence ax ≡ b (mod n)
is solvable, and its solution is uniquely determined modulo n.
Proof. (Existence) Since gcd(a, n) = 1 we can find integers s, t such
that 1 = as + nt. Then b = bas + bnt. Thus b ≡ abs (mod n). This
proves that the integer x = bs is a solution.
(Uniqueness) Suppose that we have two solutions x, x0 to the given
congruence. Then ax ≡ b and ax0 ≡ b (mod n). Thus ax ≡ ax0
(mod n), so n | (ax − ax0), so n | a(x − x0). By Lemma A.8 we
conclude that n | (x − x0), and so x ≡ x0 (mod n). 
It is important to notice that the proof not only shows that the con-
gruence can be solved, but it actually tells us how to solve it, by means
of the Euclidean algorithm. For instance, let us solve the congruence
121x ≡ 45 (mod 390). By the Euclidean algorithm or otherwise, we
find that gcd(121, 390) = 1 and 1 = 9 · 390 − 29 · 121. Multiplying by
45 as in the above proof we get that 45 = 9 · 390 · 45 − 29 · 121 · 45,
 CONSTRUCTION OF THE FINITE FIELDS Zp 7

so in other words 121(−29 · 121) ≡ 45 (mod 390). So a solution to

121x ≡ 45 (mod 390) is given by x = −29 · 121) = −3509. Since
x ≡ 389 (mod 390) we can also use the simpler representative x = 389
to solve the congruence. (Any member of the equivalence class
E(−3509) = {m ∈ Z | m ≡ −3509 (mod 390)}
will solve the problem, and 389 is a member of this class as you can
easily check.)

A.15. Definition. Any integer x such that ax ≡ 1 (mod n) is called

a multiplicative inverse of a modulo n. When such an x exists we say
that a is invertible modulo n.
The preceding theorem says that whenever a is relatively prime to
n then a is invertible modulo n. In fact, also conversely: if a is not
relatively prime to n, then a is not invertible modulo n.

A.16. Definition. (The integers modulo n) Since congruence modulo

n is an equivalence relation ∼ on the set Z of integers, it determines a
partition of Z into disjoint equivalence classes. We write a = E(a) for
the equivalence class containing an integer a. Thus
a = {b ∈ Z | a ≡ b (mod n)} = {a + nk | k ∈ Z}.
Observe that an equivalence class simply collects together all the inte-
gers which leave the same remainder upon division by n.
Let Zn = Z/ ∼ be the set of equivalence classes mod n. We can
define binary operations, “addition” and “multiplication,” on the set
Zn as follows:
a + b = a + b; a b = ab.
where on the right-hand-side of each equality we add or multiply as
usual in Z. This definition formalizes the intuition of modular arith-
metic discussed previously. It may be verified that these binary oper-
ations are well-defined (i.e. independent of choice of representatives).
There are precisely n different equivalence classes in Zn . They cor-
respond bijectively with the possible remainders (i.e., the residues)
{0, 1, . . . , n − 1} obtainable by division of an integer by n. Sometimes
we use these numbers to denote the corresponding equivalence class.

EXAMPLE. Consider the integers mod 2. The equivalence classes

mod 2 are the even and odd integers: 0 = 0 = {2n | n ∈ Z} and
1 = 1 = {2n + 1 | n ∈ Z}. Mod 2 addition and multiplication can be
8 S. R. DOTY

described in the following addition and multiplication tables

+ 0 1 · 0 1
0 0 1 0 0 0
1 1 0 1 0 1
Mod 2 arithmetic is often called binary arithmetic. It is fundamental
in digital computers and mathematical logic.
EXAMPLE. For another example, consider the integers mod 4. The
equivalence classes mod 4 are the following:
0 = 0 = {4n | n ∈ Z}
1 = 1 = {4n + 1 | n ∈ Z}
2 = 2 = {4n + 2 | n ∈ Z}
3 = 3 = {4n + 3 | n ∈ Z}.
Mod 4 addition and multiplication are summarized by the following
addition and multiplication tables
+ 0 1 2 3 · 0 1 2 3
0 0 1 2 3 0 0 0 0 0
1 1 2 3 0 1 0 1 2 3
2 2 3 0 1 2 0 2 0 2
3 3 0 1 2 3 0 3 2 1

Commutative Rings and Number Fields

Next I want to introduce two useful formal algebraic structures. The
definitions are based on axioms abstracted from familiar properties of
ordinary integers and real numbers. We will obtain interesting exam-
ples of such structures from modular arithmetic. This was our main
objective all along, and the reason we studied number theory and mod-
ular arithmetic.

A.17. Definition. Let A be a given set. A binary operation on the set

A is a mapping f : A × A → A. Such a function has two arguments,
that is, we have to write f (x, y) to denote the function value. A binary
operation combines two given elements of A to give another element of
A.
Sometimes, if ∗ is a binary operation on a set A, we write x ∗ y
instead of ∗(x, y). For example, we always do that when ∗ is addition
or multiplication, because writing +(x, y) instead of the more familiar
x + y would look strange to us. Also, when ∗ is multiplication, we
often abbreviate x · y to xy, as usual. It is important to realize that
 CONSTRUCTION OF THE FINITE FIELDS Zp 9

although such conventions are adopted for convenience and comfort,

nevertheless we still have a binary operation, no matter how we choose
to write it.

A.18. Definition. A commutative ring is a set R with two binary op-

erations, multiplication · and addition +, such that for all a, b, c in the
set R we have:
(a) a + b = b + a; ab = ba (commutativity)
(b) a + (b + c) = (a + b) + c; a(bc) = (ab)c (associativity)
(c) a(b + c) = ab + ac (distributivity);
moreover, there exist distinguished elements 0, 1 in R such that
(d) a + 0 = a; a1 = a for all a (identity);
and finally, for every a ∈ R, there exists some element −a in R such
that
(e) a + (−a) = 0 (additive inverse).

The element 0 is called the additive identity, while 1 is known as the

multiplicative identity. The element −a is the additive inverse of a, or
the negative of a. We can subtract in a ring, because we can define
a − b = a + (−b). But division may not be possible in a ring. We can
view a commutative ring as a set of objects along with a way to add,
subtract, and multiply them, subject to the usual familiar properties
of numbers.
EXAMPLES: The set of integers Z is a commutative ring under or-
dinary addition and multiplication. So are the sets of rational numbers
Q, real numbers R, and complex numbers C.
The set Zn of equivalence classes mod n is a ring under modular
addition and modular multiplication, as you can easily check. This
ring is a finite ring since it contains just n different elements.
The set N of natural numbers under ordinary addition and multipli-
cation is not a ring, because axiom (e) fails.
The ring Z1 of integers mod 1 is quite strange: it lumps together all
of Z into just one equivalence class. We have just one element in this
ring, so 0 = 1 in this ring. This ring is sometimes called the zero ring
or the trivial ring. It is not of much interest.
Here are some formal consequences of the definition, which must hold
in any ring R, for all a, b, c in R:
(1) a0 = 0.
(2) a(−b) = −(ab) = (−a)b.
(3) (−a)(−b) = ab.
10 S. R. DOTY

(4) a(b − c) = ab − ac.

(5) (−1)a = −a.
The proof of these basic facts is an exercise. For instance, to prove (a)
one would begin with the equality 0 + 0 = 0 (which comes from axiom
(d) by taking a = 0 there), multiply both sides by a, and so forth.

A.19. Definition. An element a of a ring is said to be invertible if

there is some b in the ring such that ab = ba = 1. If such an element b
exists in the ring, we write a−1 = b. (Think of the inverse of a matrix.)
An invertible element is also called a unit in the ring.
EXAMPLES. In Z4 the invertible elements are 1 and 3. In Z the
invertible elements are 1 and −1. In the ring R of real numbers, every
nonzero element is invertible.

A.20. Definition. A field is a commutative ring F in which 0 6= 1,

and every nonzero element is invertible.
EXAMPLES: Q, R, and C are all fields. The ring of integers Z is
not a field.
Is the ring Zn ever a field? The answer, which is provided by the
following theorem, is quite satisfying.

A.21. Theorem. Zn is a field if and only if n is prime.

Proof. This follows immediately from Theorem A.14. Suppose that n is
prime. Let a 6= 0 in Zn . Then gcd(a, n) = 1 since a ∈ {1, 2, . . . , n − 1}
and thus p cannot divide a. Hence there exists some x ∈ Z such that
ax ≡ 1 (mod n), so x is the inverse of a in Zn . This shows that Zn is
a field if n is prime.
On the other hand, if n is not prime then we can factor n in the form
n = ab for some integers a, b < n. Then a will not be invertible in Zn
since ab = 0 in Zn. (If you assume a is invertible then you can easily
derive a contradiction from this.) 
NOTATION: In case n = p is prime, it is customary to write Fp = Zp
for this finite field of p elements.
Notice that the proof of the above theorem provides a procedure by
which we can compute the inverse of an invertible element of Zn, by
solving the corresponding congruence. Looking back at the proof of
A.14 we see that to solve the congruence we can use the Euclidean
algorithm. (For a small modulus, it is simpler to use trial and error
since there are only finitely many possibilities to check.)
 CONSTRUCTION OF THE FINITE FIELDS Zp 11

EXAMPLE. You can check that 7919 is prime. Let’s compute the
inverse of a = 210 in the finite field Z7919. The gcd of 210 and 7919
is 1, of course, and using the Euclidean algorithm we can find integers
s and t such that 1 = 210s + 7919t. In fact, s = 1169 and t = −31
as you can check. Thus x = 1169 is a solution to the congruence
210x ≡ 1 (mod 7919), so the inverse of 210 in Z7919 is the element
210−1 = 1169 = 1169. You can check this by computing the residue of
210 · 1169 (mod 7919), which should be 1.
In any field, we can define division by any nonzero element b, by
setting a/b = ab−1 . Intuitively, a field is a structure in which we have
all the ordinary operations of arithmetic: addition, subtraction, mul-
tiplication, and division, such that these operations satisfy the usual
properties of algebra. In addition to the familiar number fields Q, R,
C we have the novel finite fields Fp = Zp of p elements, for every prime
number p.

Exercises
(1) (a) Use the Euclidean algorithm to compute gcd(48157656, 541541).
(Use your calculator.)
(b) Find integers s, t such that
gcd(48157656, 541541) = 48157656s + 541541t.
(2) Show that if gcd(a, n) 6= 1 then a is not invertible modulo n.
(3) Compute the following:
(a) 2−1 (mod 7).
(b) 4−1 (mod 15).
(c) 23−1 (mod 275).
(d) 909−1 (mod 81799). (Use Euclidean algorithm.)
(4) Use your answers to the preceding exercise to solve the congruences:
(a) 2x ≡ 4 (mod 7).
(b) 4x ≡ 13 (mod 15).
(c) 23x ≡ 200 (mod 275).
(d) 909x ≡ 8910 (mod 81799).
Hint: Simply multiply by the inverse.
(5) Let R be a commutative ring. Prove, using only the axioms listed
in the definition of commutative ring, that for any elements a, b, c of
R we have:
(a) a0 = 0.
(b) a(−b) = −(ab) = (−a)b.
(c) (−a)(−b) = ab.
(d) a(b − c) = ab − ac.
(e) (−1)a = −a.
Note: b − c = b + (−c) (definition).
12 S. R. DOTY

(6) Show that the number of invertible elements in the ring Zn is ϕ(n),
for any n. Here ϕ(n) is the number of integers in the range 1, 2, . . . , n
which are relatively prime to n. The function ϕ is called Euler’s phi
function.
(7) Prove that congruence mod n is an equivalence relation on Z (The-
orem A.12).
(8) Prove Theorem A.13.
(9) Show that a congruence ax ≡ b (mod n) is solvable if and only if
gcd(a, n) divides b.
(10) Explain the contradiction at the end of the proof of A.21. In other
words, what is contradicted if you have elements a 6= 0, b 6= 0, and
c in a ring such that ac = 1 and ab = 0?
(11) Prove Theorem A.10.