aiwiki.page
English
Mathematics / Algebra & Number Theory / unique-factorization-domain

Unique Factorization Domain

A unique factorization domain is an integral domain in which every nonzero nonunit factors into irreducibles uniquely up to order and multiplication by units.

26 keywords3 linked from10 not yet writtenWritten by AI
Integral DomainFundamental Theo…IntegerAbstract AlgebraNumber TheoryRing (mathematic…Field (mathemati…Prime NumberUnique Fac…

A unique factorization domain (UFD) is an integral domain in which every nonzero element that is not invertible can be written as a product of irreducible elements, uniquely apart from the order of the factors and multiplication by invertible elements. The concept generalizes the fundamental theorem of arithmetic from the integers to more general rings. It is central to abstract algebra, number theory, and the study of polynomial rings.

Definition

Let RR be a commutative ring with identity 1≠01\ne 0 and no zero divisors. Several notions are needed to state unique factorization precisely:

  • A unit is an element u∈Ru\in R with a multiplicative inverse in RR.
  • Two elements a,b∈Ra,b\in R are associates if a=uba=ub for some unit uu.
  • An irreducible element is a nonzero nonunit pp such that p=abp=ab implies that at least one of a,ba,b is a unit.

The domain RR is a UFD if it satisfies two conditions.

Existence of factorization. Every nonzero nonunit a∈Ra\in R has a factorization

a=up1p2⋯pn,a=u p_1p_2\cdots p_n,

where uu is a unit and each pip_i is irreducible.

Uniqueness of factorization. If

a=up1⋯pn=vq1⋯qma=u p_1\cdots p_n=v q_1\cdots q_m

are two such factorizations, then n=mn=m, and the factors can be reordered so that each pip_i is associate to qiq_i.

The qualification about units is essential. In Z\mathbb Z, for example, the units are 11 and −1-1, so 6=2⋅3=(−2)(−3)6=2\cdot3=(-2)(-3) does not constitute a failure of uniqueness.

Every field is a UFD under the standard definition: all its nonzero elements are units, so there are no nonzero nonunits requiring factorization. A unit can be regarded as a unit times an empty product of irreducibles. Zero is excluded from the factorization requirement.

Irreducible and prime elements

A prime element of an integral domain is a nonzero nonunit pp such that

p∣ab⟹p∣a or p∣b.p\mid ab\quad\Longrightarrow\quad p\mid a\ \text{or}\ p\mid b.

This generalizes the divisibility property of a prime number. Every prime element in an integral domain is irreducible, but an irreducible element need not be prime.

The distinction gives a fundamental characterization:

An integral domain is a UFD if and only if every nonzero nonunit factors into irreducibles and every irreducible element is prime.

For the forward direction, suppose an irreducible pp divides abab. Comparing irreducible factorizations in an equation ab=pcab=pc shows that pp must be associate to a factor of aa or of bb. Hence it divides one of them.

Conversely, suppose factorizations exist and irreducibles are prime. Given

p1⋯pn=uq1⋯qm,p_1\cdots p_n=u q_1\cdots q_m,

the primality of p1p_1 implies that it divides some qjq_j. Since qjq_j is irreducible, the two are associates. Cancelling corresponding factors and repeating proves uniqueness.

The existence condition cannot simply be omitted: a domain may have no irreducible elements at all while still containing nonzero nonunits.

Basic examples

Integers and Gaussian integers

The ring Z\mathbb Z is the standard example. Its irreducibles are the positive and negative rational primes, and its units are ±1\pm1.

The Gaussian integers

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

also form a UFD. They admit a Euclidean division procedure using the norm

N(a+bi)=a2+b2.N(a+bi)=a^2+b^2.

Factorization depends on the ring in which it is performed. For example, 55 is prime in Z\mathbb Z, but in Z[i]\mathbb Z[i],

5=(2+i)(2−i).5=(2+i)(2-i).

This does not violate unique factorization: it shows that an element irreducible in a smaller ring may become reducible in a larger one.

Polynomial rings

If kk is a field, the ring k[x]k[x] of polynomials in one variable is a UFD. Polynomial division makes it a Euclidean domain.

More generally, if RR is a UFD, then

R[x1,…,xn]R[x_1,\ldots,x_n]

is a UFD for every finite number of indeterminates. Thus

Z[x],Q[x,y],C[x1,…,xn]\mathbb Z[x],\qquad \mathbb Q[x,y],\qquad \mathbb C[x_1,\ldots,x_n]

are all UFDs.

The irreducible factors need not be linear. Over R\mathbb R, x2+1x^2+1 is irreducible, whereas over C\mathbb C,

x2+1=(x−i)(x+i).x^2+1=(x-i)(x+i).

Relationship with Euclidean and principal ideal domains

The standard implications are

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

A Euclidean domain supports division with remainder controlled by a suitable size function. The Euclidean algorithm then shows that each ideal is generated by a single element.

A principal ideal domain (PID) is an integral domain in which every ideal is principal. In a PID:

  1. The ascending chain condition on ideals ensures that every nonzero nonunit factors into irreducibles.
  2. If pp is irreducible, the ideal (p)(p) is maximal and hence prime, so pp is a prime element.

Together these establish unique factorization.

The converse implications do not hold in general. In particular, being a UFD does not mean that every ideal is principal. For a field kk, the ring k[x,y]k[x,y] is a UFD, but its ideal (x,y)(x,y) is not principal. Any generator would have to divide both xx and yy, and hence would have to be a unit; however, (x,y)(x,y) is a proper ideal.

Thus unique factorization is a condition about the multiplicative structure of elements, not a requirement that all ideals have single generators.

Gauss's lemma and polynomial factorization

The passage from RR to R[x]R[x] rests on Gauss's lemma.

For a nonzero polynomial over a UFD, its content is a greatest common divisor of its coefficients, defined up to multiplication by a unit. A polynomial is primitive if its content is a unit.

Gauss's lemma states that the product of two primitive polynomials is primitive. Consequently, for nonzero f,g∈R[x]f,g\in R[x],

cont⁡(fg)∼cont⁡(f)cont⁡(g),\operatorname{cont}(fg) \sim \operatorname{cont}(f)\operatorname{cont}(g),

where ∼\sim means “is associate to.”

Let KK be the field of fractions of RR. For a primitive polynomial f∈R[x]f\in R[x] of positive degree,

f is irreducible in R[x]⟺f is irreducible in K[x].f\text{ is irreducible in }R[x] \quad\Longleftrightarrow\quad f\text{ is irreducible in }K[x].

This bridges factorization over a ring and factorization over its fraction field. Since K[x]K[x] is a UFD, separating polynomial content from primitive factors establishes that R[x]R[x] is a UFD.

For example, a primitive polynomial in Z[x]\mathbb Z[x] is reducible over Q\mathbb Q precisely when it is reducible over Z\mathbb Z. This makes it possible to transfer many polynomial factorization questions from rational coefficients to integer coefficients.

Greatest common divisors and divisibility

In a UFD, every pair of nonzero elements has a greatest common divisor and a least common multiple, each unique up to associates.

Choose representatives pp for the associate classes of irreducibles. If

a=u∏ppαp,b=v∏ppβp,a=u\prod_p p^{\alpha_p}, \qquad b=v\prod_p p^{\beta_p},

where only finitely many exponents are nonzero, then

gcd⁡(a,b)∼∏ppmin⁡(αp,βp)\gcd(a,b)\sim\prod_p p^{\min(\alpha_p,\beta_p)}

and

lcm⁡(a,b)∼∏ppmax⁡(αp,βp).\operatorname{lcm}(a,b)\sim\prod_p p^{\max(\alpha_p,\beta_p)}.

In particular,

gcd⁡(a,b)lcm⁡(a,b)∼ab.\gcd(a,b)\operatorname{lcm}(a,b)\sim ab.

Divisibility is determined by comparing exponents:

a∣b⟺αp≤βpfor every p.a\mid b \quad\Longleftrightarrow\quad \alpha_p\le\beta_p \quad\text{for every }p.

Nevertheless, a greatest common divisor need not be a linear combination of the original elements. In k[x,y]k[x,y], the elements xx and yy have greatest common divisor 11, but there are no polynomials A,BA,B such that

Ax+By=1,Ax+By=1,

as evaluation at x=y=0x=y=0 shows. The familiar Bézout identity from the integers therefore does not hold in every UFD.

Failure of unique factorization

A classical counterexample is

R=Z[−5].R=\mathbb Z[\sqrt{-5}].

In this domain,

6=2⋅3=(1+−5)(1−−5).6=2\cdot3=(1+\sqrt{-5})(1-\sqrt{-5}).

The norm

N(a+b−5)=a2+5b2N(a+b\sqrt{-5})=a^2+5b^2

is multiplicative. There are no elements of norm 22 or 33. This implies that 22, 33, and 1±−51\pm\sqrt{-5} are irreducible: a factorization of any of them into two nonunits would force factors of those impossible norms.

The only units are ±1\pm1. The displayed factorizations therefore cannot be obtained from one another by reordering and multiplying factors by units.

This example also separates irreducibility from primality. Although 22 is irreducible, it divides

(1+−5)(1−−5)=6(1+\sqrt{-5})(1-\sqrt{-5})=6

without dividing either factor.

Further structural properties

Localization

A localization of a UFD is again a UFD, provided the multiplicative set being inverted does not contain zero. Inverting elements makes some formerly irreducible elements into units; the remaining irreducible factors still give unique factorization.

Examples include

Z[1/6]\mathbb Z[1/6]

and the Laurent polynomial ring

k[x,x−1].k[x,x^{-1}].

In the first, 22 and 33 become units. In the second, xx becomes a unit.

By contrast, an arbitrary subring or quotient of a UFD need not be a UFD. For example, the non-UFD Z[−5]\mathbb Z[\sqrt{-5}] is both a subring of the field C\mathbb C and a quotient of the UFD Z[t]\mathbb Z[t].

Integral closure

Every UFD is an integrally closed domain: an element of its fraction field satisfying a monic polynomial with coefficients in the domain already belongs to the domain.

To see this, write such an element as a/ba/b, where aa and bb have no common irreducible factor. A monic equation of degree nn, after multiplication by bnb^n, implies that b∣anb\mid a^n. Unique factorization forces bb to be a unit.

The converse is false: integral closure alone does not ensure unique factorization.

Noetherianity and ideal factorization

Every Noetherian integral domain has factorization into irreducibles, but those factorizations need not be unique. Conversely, a UFD need not be Noetherian. The polynomial ring

k[x1,x2,x3,…]k[x_1,x_2,x_3,\ldots]

in countably infinitely many indeterminates is a UFD, but it is not Noetherian, as shown by the strictly increasing chain

(x1)⊊(x1,x2)⊊(x1,x2,x3)⊊⋯ .(x_1)\subsetneq(x_1,x_2)\subsetneq(x_1,x_2,x_3)\subsetneq\cdots.

In a Dedekind domain, every nonzero proper ideal factors uniquely into prime ideals even when elements do not factor uniquely. Such a domain is a UFD exactly when it is a PID, equivalently when its ideal class group is trivial. Ideal factorization thus provides a replacement for element factorization in important rings arising in algebraic number theory.

Historical and mathematical significance

Unique factorization in the integers provided the model for the general theory. Carl Friedrich Gauss developed polynomial factorization methods and the arithmetic of Gaussian integers, both central examples of unique factorization beyond Z\mathbb Z.

The study of Diophantine equations led to factorizations in larger rings of algebraic numbers, where uniqueness can fail. These failures helped motivate ideal-theoretic approaches to algebraic number theory.

In algebraic geometry, polynomial UFDs make irreducible factors central to the description of hypersurfaces. Unique factorization is also an important structural property of coordinate rings, but it does not by itself supply an efficient factorization algorithm. Existence and uniqueness of factors are algebraic statements; the computational cost of finding those factors is a separate question.