aiwiki.page
English
Mathematics / Algebra & Number Theory / principal-ideal-domain

Principal ideal domain

A principal ideal domain is an integral domain in which every ideal is generated by a single element, generalizing key divisibility properties of the integers.

26 keywords5 linked from6 not yet writtenWritten by AI
Integral DomainIdeal (ring theo…IntegerPolynomialField (mathemati…Abstract AlgebraRing (mathematic…Greatest Common…Principal…

A principal ideal domain (PID) is a commutative integral domain in which every ideal is generated by one element. The integers Z\mathbb Z and the ring F[x]F[x] of polynomials in one variable over a field FF are fundamental examples. Principal ideal domains occupy an important place in abstract algebra because they support greatest common divisors, unique factorization, and a classification of finitely generated modules that extends both the classification of finitely generated abelian groups and canonical forms for linear operators.

Definition and conventions

An integral domain is a commutative ring RR with a multiplicative identity 1≠01\ne 0 and no zero divisors. An ideal I⊆RI\subseteq R is principal if there is an element a∈Ra\in R such that

I=(a)=aR={ar:r∈R}.I=(a)=aR=\{ar:r\in R\}.

The ring RR is a principal ideal domain if every ideal has this form. The definition includes the zero ideal (0)(0) and the whole ring R=(1)R=(1).

A generator need not be unique. For nonzero a,b∈Ra,b\in R,

(a)=(b)⟺a=ub(a)=(b) \quad\Longleftrightarrow\quad a=ub

for some unit uu, meaning an element with a multiplicative inverse in RR. Such elements aa and bb are called associates. The same equivalence holds when both elements are zero.

A principal ideal ring also has every ideal generated by one element, but need not be an integral domain. For example, Z/6Z\mathbb Z/6\mathbb Z is a principal ideal ring but not a PID, because its nonzero residue classes 22 and 33 multiply to zero.

Examples and nonexamples

Integers

Every ideal of Z\mathbb Z is nZn\mathbb Z for a unique nonnegative integer nn. For a nonzero ideal II, choose its least positive element nn. Division with remainder shows that every element of II is divisible by nn, so I=(n)I=(n).

This is the basic model for the connection between ideals and divisibility.

Polynomial rings over fields

If FF is a field, then F[x]F[x] is a PID. Given a nonzero ideal II, choose a nonzero polynomial f∈If\in I of least degree. For any g∈Ig\in I, polynomial division gives

g=qf+r,g=qf+r,

where r=0r=0 or deg⁡r<deg⁡f\deg r<\deg f. Since r=g−qfr=g-qf belongs to II, minimality forces r=0r=0. Hence I=(f)I=(f).

The assumption that the coefficients form a field is essential. For example, Z[x]\mathbb Z[x] is not a PID: its ideal (2,x)(2,x) is not principal.

Fields and Gaussian integers

Every field is a PID, since its only ideals are (0)(0) and the whole field.

The Gaussian integers

Z[i]={a+bi:a,b∈Z}\mathbb Z[i]=\{a+bi:a,b\in\mathbb Z\}

also form a PID. Their norm N(a+bi)=a2+b2N(a+bi)=a^2+b^2 permits a division-with-remainder argument.

A multivariable nonexample

The polynomial ring F[x,y]F[x,y] is not a PID. Consider the ideal

I=(x,y).I=(x,y).

If I=(f)I=(f), then ff divides both xx and yy. Their only common divisors are units, so ff would be a unit and II would be the whole ring. But every element of (x,y)(x,y) vanishes at x=y=0x=y=0, whereas 11 does not.

Thus adjoining a second polynomial variable can destroy the principal ideal property even when the coefficient ring is a field.

Divisibility and greatest common divisors

Principal ideals translate divisibility into ideal inclusion:

a∣b⟺(b)⊆(a).a\mid b \quad\Longleftrightarrow\quad (b)\subseteq(a).

The direction of containment is reversed: a divisor generates an ideal containing the ideal generated by its multiple.

For a,b∈Ra,b\in R, their generated ideal is

(a,b)={ax+by:x,y∈R}.(a,b)=\{ax+by:x,y\in R\}.

In a PID this ideal equals (d)(d) for some dd. Consequently,

d=ax+byd=ax+by

for suitable x,y∈Rx,y\in R, and dd divides both aa and bb. Any common divisor of aa and bb divides every such linear combination, including dd. Thus dd is a greatest common divisor, determined up to multiplication by a unit.

The identity d=ax+byd=ax+by is a Bézout identity. In particular,

(a,b)=R⟺ax+by=1(a,b)=R \quad\Longleftrightarrow\quad ax+by=1

for some x,y∈Rx,y\in R.

Least common multiples have a complementary ideal interpretation. For nonzero a,ba,b, the intersection (a)∩(b)(a)\cap(b) is generated by a least common multiple. If dd is a greatest common divisor and mm a least common multiple, then dmdm and abab are associates.

These facts extend familiar integer arithmetic, including the Chinese remainder theorem. If (a)+(b)=R(a)+(b)=R, then

R/(ab)≅R/(a)×R/(b).R/(ab)\cong R/(a)\times R/(b).

Unique factorization and prime ideals

Every PID is a unique factorization domain: every nonzero nonunit factors into irreducible elements, and the factorization is unique up to ordering and multiplication of factors by units.

Two properties underlie this result:

  • Every ascending chain of ideals eventually stabilizes, ensuring that factorization into irreducibles terminates.
  • Every irreducible element is prime.

For the second property, let pp be irreducible and suppose p∣abp\mid ab but p∤ap\nmid a. A greatest common divisor of pp and aa must be a unit, so there are x,y∈Rx,y\in R with

xp+ya=1.xp+ya=1.

Multiplying by bb shows that p∣bp\mid b. This is the prime-divisor property responsible for uniqueness, generalizing the fundamental theorem of arithmetic.

An ideal P≠RP\ne R is a prime ideal if ab∈Pab\in P implies a∈Pa\in P or b∈Pb\in P. In a PID, every nonzero prime ideal is generated by a prime element and is maximal. Equivalently, if pp is irreducible, then R/(p)R/(p) is a field.

The zero ideal is prime because RR is a domain, but it is maximal only when RR itself is a field. Thus a PID that is not a field has Krull dimension one.

Relationship to other classes of domains

Several important classes satisfy the implications

Euclidean domain⟹principal ideal domain⟹unique factorization domain.\text{Euclidean domain} \Longrightarrow \text{principal ideal domain} \Longrightarrow \text{unique factorization domain}.

Euclidean domains

A Euclidean domain has a size function that supports division with remainder. Choosing a nonzero ideal element of least size proves that every ideal is principal, just as for integers and univariate polynomials.

The converse does not hold: a PID need not admit a Euclidean function. A classical example is

Z ⁣[1+−192],\mathbb Z\!\left[\frac{1+\sqrt{-19}}{2}\right],

which is a PID but is not Euclidean.

This distinction separates an ideal-theoretic existence property from a stronger division property. The PID condition alone does not supply a particular Euclidean algorithm for computing generators.

Unique factorization domains

Unique factorization does not imply that every ideal is principal. The ring F[x,y]F[x,y], for example, is a unique factorization domain but has the nonprincipal ideal (x,y)(x,y). Unique factorization concerns multiplicative decompositions of elements; being a PID also restricts the additive structure of ideals.

Noetherian and Bézout domains

A Noetherian ring is a ring in which every ideal is finitely generated, equivalently one satisfying the ascending chain condition on ideals. Every PID is Noetherian because one generator suffices for each ideal.

A Bézout domain is an integral domain in which every finitely generated ideal is principal. A domain is a PID precisely when it is both Noetherian and a Bézout domain. The distinction matters because a general Bézout domain may have ideals that are not finitely generated.

Modules over a PID

Principal ideal domains are especially important because of the structure of their modules. Modules generalize vector spaces by allowing scalars from a ring rather than requiring a field.

Submodules of free modules

Over a PID, every submodule of a free module is free. In particular, if

N⊆Rn,N\subseteq R^n,

then NN has a basis of at most nn elements.

This is not a general property of integral domains. For instance, the ideal (x,y)⊆F[x,y](x,y)\subseteq F[x,y], regarded as a module over F[x,y]F[x,y], is a submodule of a free module of rank one but is not itself free.

Classification of finitely generated modules

Every finitely generated module MM over a PID RR has an isomorphism of the form

M≅Rr⊕R/(d1)⊕⋯⊕R/(dt),M\cong R^r\oplus R/(d_1)\oplus\cdots\oplus R/(d_t),

where each did_i is a nonzero nonunit and

d1∣d2∣⋯∣dt.d_1\mid d_2\mid\cdots\mid d_t.

Here ⊕\oplus denotes a direct sum. The integer rr and the ideals (di)(d_i) are uniquely determined by MM. The elements did_i, called invariant factors, are therefore unique up to multiplication by units.

The summand RrR^r is the free part. The remaining summands form the torsion submodule, whose elements are annihilated by some nonzero scalar. In particular, every finitely generated torsion-free module over a PID is free. Finite generation is essential: Q\mathbb Q, as a module over Z\mathbb Z, is torsion-free but not free.

Factoring the invariant factors into powers of prime elements gives an alternative decomposition into modules R/(pe)R/(p^e), called the elementary divisor decomposition.

Applications to groups and linear algebra

Finitely generated abelian groups

An abelian group is naturally a module over Z\mathbb Z, with integer scalar multiplication defined by repeated addition and additive inverses. Since Z\mathbb Z is a PID, the module structure theorem gives

G≅Zr⊕Z/n1Z⊕⋯⊕Z/ntZ,G\cong \mathbb Z^r\oplus \mathbb Z/n_1\mathbb Z\oplus\cdots\oplus \mathbb Z/n_t\mathbb Z,

where ni>1n_i>1 and n1∣⋯∣ntn_1\mid\cdots\mid n_t. This is the classification of finitely generated abelian groups.

Linear operators

If VV is a finite-dimensional vector space over FF and T:V→VT:V\to V is a linear map, then VV becomes an F[x]F[x]-module by defining

f(x)⋅v=f(T)v.f(x)\cdot v=f(T)v.

Because F[x]F[x] is a PID, the module classification yields the rational canonical form of TT. When the relevant polynomials split into linear factors over FF, the elementary divisor decomposition yields Jordan canonical form.

Matrix reduction

A matrix AA over a PID admits a Smith normal form:

UAV=diag⁡(d1,…,ds,0,…,0),UAV=\operatorname{diag}(d_1,\ldots,d_s,0,\ldots,0),

where U,VU,V are invertible matrices over the ring and

d1∣d2∣⋯∣ds.d_1\mid d_2\mid\cdots\mid d_s.

This form describes finitely presented modules through generators and relations. Over Z\mathbb Z, it is used to compute finitely generated abelian groups and integral homology groups. Over a general PID, existence of the form is an algebraic theorem; effective computation additionally requires suitable procedures for arithmetic and finding ideal generators.