An algebraically closed field is a field such that every nonconstant polynomial with coefficients in has at least one root in . Equivalently, every such polynomial factors completely into linear factors over . The field of complex numbers is the principal familiar example, but algebraically closed fields also exist in positive characteristic. They provide a fundamental setting for the study of polynomial equations, algebraic geometry, and parts of model theory.
Definition and equivalent characterizations
Write for the ring of polynomials in one variable with coefficients in . Algebraic closedness means that, for every of positive degree, there is an such that .
Once one root is found, polynomial division gives
Applying the same argument repeatedly yields a complete factorization. Thus, if has degree , then
where and . The roots need not be distinct: an algebraically closed field guarantees roots counted with multiplicity, not distinct roots.
The following statements about a field are equivalent:
- is algebraically closed.
- Every nonconstant polynomial in splits into linear factors.
- Every irreducible polynomial in has degree one.
- has no proper algebraic field extension.
- has no proper finite field extension.
An extension is algebraic if every element of satisfies a nonzero polynomial with coefficients in . The equivalences follow from the correspondence between irreducible polynomials and simple algebraic extensions. An irreducible polynomial of degree greater than one produces a proper finite extension ; conversely, an element in a proper algebraic extension has an irreducible minimal polynomial of degree greater than one.
The word “closed” here is algebraic, not topological. No topology or notion of convergence is required.
Examples and nonexamples
Complex numbers
The fundamental theorem of algebra states that is algebraically closed. In particular, every complex polynomial of degree has exactly complex roots when multiplicities are included.
The real numbers are not algebraically closed: has no real root. Adjoining an element satisfying gives
Although this single quadratic extension suffices for , an arbitrary field need not have an algebraic closure obtained by adjoining just one element.
Algebraic numbers
The field of all complex algebraic numbers is algebraically closed. Indeed, if a complex number is a root of a polynomial whose coefficients are algebraic over , the transitivity of algebraicity implies that the root is itself algebraic over .
Thus algebraic closedness does not require a field to contain every complex number. The field is a countable set, whereas is uncountable. Numbers transcendental over , such as , do not belong to .
The rational numbers themselves are not algebraically closed; for example, has no rational root.
Finite fields and their closures
No finite field is algebraically closed. If has elements, every satisfies , so
has no root in .
Nevertheless, every finite field has an algebraic closure. Within a fixed algebraic closure of , where is prime, there is a unique subfield with elements for each positive integer , and
Here the union is taken inside that fixed closure; the finite subfields form a directed family under inclusion. Consequently, algebraically closed fields can be infinite and countable even in positive characteristic.
Algebraic closures
An algebraic closure of a field is an extension satisfying two conditions:
- is algebraically closed.
- Every element of is algebraic over .
The second condition distinguishes an algebraic closure from an arbitrary algebraically closed extension. For example, is an algebraically closed extension of , but it is not an algebraic closure of , because it contains transcendental elements.
Every field has an algebraic closure in the usual set-theoretic foundations with the axiom of choice. One construction repeatedly adjoins roots of polynomials, arranging that every polynomial over the resulting union eventually splits. Existence arguments commonly use a maximality principle or a well-ordering construction.
Any two algebraic closures of are isomorphic by an isomorphism fixing every element of . This isomorphism need not be unique. For example, complex conjugation restricts to a nonidentity automorphism of fixing .
More generally, if is an algebraically closed extension of , the elements of algebraic over form an algebraic closure of .
Characteristic and repeated roots
The characteristic of an algebraically closed field is either zero or a prime number , as for every field.
In characteristic , the Frobenius map
is an automorphism. It is injective in every field of characteristic , and algebraic closedness makes it surjective because has a root for every .
Consequently, every algebraically closed field is a perfect field. In particular, irreducible polynomials over it are separable—here trivially so, since they are linear.
This does not mean that every polynomial over an algebraically closed field has distinct roots. In characteristic ,
whenever . This polynomial has only one distinct root, with multiplicity . Algebraic closedness concerns the existence and location of roots, while separability concerns their multiplicities.
Polynomial equations in several variables
Algebraic closedness is defined using one-variable polynomials, but its most important consequences extend to systems of polynomial equations.
Let be algebraically closed. The weak form of Hilbert’s Nullstellensatz states that every maximal ideal of
has the form
for some point . Equivalently, every proper ideal in this polynomial ring has a common zero in .
The strong form states
where is the common zero set of , is the ideal of polynomials vanishing on that set, and
is the radical of . This establishes a precise correspondence between algebraic sets and radical ideals.
These statements do not assert that every system of polynomial equations has a solution: inconsistent systems generate the whole polynomial ring. They assert that, over an algebraically closed field, a proper ideal cannot fail to have a common zero merely because additional algebraic roots are missing from the field.
Linear algebra
Algebraic closedness simplifies linear algebra because every characteristic polynomial splits.
If is a nonzero finite-dimensional vector space over an algebraically closed field , every linear operator has an eigenvalue in . Repeatedly passing to invariant subspaces or quotients shows that has an upper-triangular matrix in a suitable basis. It also admits a Jordan normal form.
An algebraically closed scalar field does not guarantee diagonalizability. For example,
is not diagonalizable over any field: its eigenspace for the sole eigenvalue has dimension one.
Classification and model theory
A transcendence basis of an extension is a maximal algebraically independent subset ; the extension is then algebraic. The cardinality of is the transcendence degree.
Up to field isomorphism, an algebraically closed field is determined by:
- its characteristic;
- its transcendence degree over its prime subfield.
The prime subfield is in characteristic zero and in characteristic . Given a transcendence basis , an algebraically closed field is an algebraic closure of the rational function field generated by over its prime subfield.
In first-order logic, algebraically closed fields are described by the field axioms together with an axiom for each positive degree asserting that every polynomial of that degree has a root. For each fixed characteristic, this theory is complete and admits quantifier elimination in the language of rings. Completeness means that any two algebraically closed fields of that characteristic satisfy the same first-order sentences in this language, even when their transcendence degrees differ.
Limits of the concept
Algebraic closedness is an existence property, not a general formula for computing roots. It does not imply that polynomial roots can be expressed by radicals; the obstructions studied in Galois theory remain relevant.
It is also independent of completeness in the sense of analysis. For example, is algebraically closed but is not complete under the usual absolute value inherited from . Conversely, is complete under its usual absolute value but is not algebraically closed.
Finally, algebraic closedness governs polynomial equations, not arbitrary equations involving exponential, analytic, or other additional functions. Such questions require structure beyond the field operations.