The Size of a Finite Field
A finite field has prime characteristic , contains , and is a vector space of some dimension over it — so it has exactly elements. Conversely, for every prime power there is a field with elements, unique up to isomorphism, written or .
This is a complete classification: one field for each prime power, none for any other size. To build , take any irreducible polynomial of degree over and form as in the previous topics. Different irreducible polynomials give the same field up to a relabeling of its elements.
is irreducible over (no root), so , with . The powers of are : all seven nonzero elements.
is not a prime power, so no field has elements. Indeed, such a field would have characteristic or and would be a vector space over or , so its size would be a power of or of .
The field with elements is NOT : in , , so is not even an integral domain. is , which has characteristic () — a completely different ring.
Which of are sizes of finite fields?
Show the answerHide the answer
The prime powers: — five of them. ( have two different prime factors.)
The Multiplicative Group Is Cyclic
The nonzero elements of form a CYCLIC group of order under multiplication. A generator is called a primitive element; there are exactly of them. In , primitive elements are the primitive roots mod .
This brings the group theory of the first half straight into fields: , so everything known about cyclic groups applies — the number of elements of each order, the subgroups, and the powers. Every nonzero element is a power of a primitive element, which makes multiplication in a finite field as easy as adding exponents mod .
The powers of mod are : all six nonzero residues. So generates , and there are primitive roots in all ( and ).
has order , a prime, so every element except is primitive: . has order , with primitive elements. In the primitive roots are — of them.
The ADDITIVE group of is usually not cyclic: in , every element satisfies , so the additive group is . It is the MULTIPLICATIVE group of nonzero elements that is always cyclic.
Is a primitive root mod ?
Show the answerHide the answer
Its powers mod : — all ten nonzero residues appear, so yes.
Subfields of a Finite Field
The subfields of are exactly the fields with , one for each divisor . So has subfields, where is the number of divisors of .
This follows from the tower law: if then , so . Conversely the elements with form a subfield of size whenever . The subfield lattice of is therefore the divisor lattice of .
, and the divisors of are . So has exactly subfields: .
and , but . So does not contain , even though both have characteristic and .
Containment depends on the EXPONENT dividing, not the size: (), but . And fields of different characteristics never contain one another.
How many subfields does have?
Show the answerHide the answer
One for each divisor of : — subfields.
Counting Irreducible Polynomials
Over , is the product of ALL monic irreducible polynomials whose degree divides , each appearing once. Comparing degrees: if is the number of monic irreducible polynomials of degree , then .
The roots of are exactly the elements of , each once. Every such element has a minimal polynomial over , irreducible of degree dividing , and grouping the roots by minimal polynomial gives the factorization. The degree identity lets you count irreducible polynomials recursively.
The divisors of are and , so : the two irreducibles of degree and the two of degree . Degree check: .
For : . With and : , so . For : , so .
Do not divide by directly: the elements of smaller subfields have minimal polynomials of smaller degree and must be removed first. For , : , but the correct count is .
How many monic irreducible quadratics are there over ?
Show the answerHide the answer
with , so (they are , , ).
Why It Works
A finite field has prime characteristic (characteristic would make all different). The multiples of form a copy of , and is a vector space over it of some finite dimension . A basis of vectors gives possible coefficient choices, one for each element.
Let , of order . For each , the elements of order are roots of , which has at most roots; if one element of order exists, its powers already give roots, so the elements of order are among , and there are exactly of them. So the number of elements of order is or . Since , every count must be — in particular there are elements of order .
If sits inside , then is a vector space over of some dimension , so and .