Let R be a commutative integral domain. R is a unique factorization domain (or UFD) if
1. Every element of R which is not zero or a unit can be written as a product of irreducible elements of R.
2. This expression is unique in the sense that if we write c=a1...ar=b1...bs in R, for some nonzero non-unit c, with each ai and bj irreducible, then we have r=s and after renumbering the ai if necessary we have that each ai is an associate of bj.

Of course, the fundamental theorem of arithmetic asserts that the ring of integers Z is a UFD.

Proposition Any principal ideal domain is a UFD.

After the Proposition and the principal ideal domain writeup we know that besides the ring of integers we have several examples of UFDs. The ring of Gaussian integers Z[i] is a UFD and so is the polynomial ring k[x], for any field k.

Proof of the Proposition Let R be a PID. I first claim that every nonzero nonunit a in R can be factorized as product of irreducibles. If a is irreducible there is nothing to prove. If not then we can write a=a1b1 with a1,b1 non-units. If a1,b1 are both irreducible we are done. Suppose that a1 is not. Then we factorize a1=a2b2, as a product. We repeat this until we have expressed a as a product of irreducibles. Thus we will have a=a1b1=a2b1b2...

We have to show that this process terminates. Write a0=a. Then associated to our factorization attempt we have the increasing sequence of ideals a0R < a1R < a1R < .... Our factorisation attempt terminates if this chain terminates.

Let I be the union of these ideals. Because this sequence is increasing it is not hard to see that I is also an ideal. Since R is a PID this means that I=bR. Now b has to be in one of the ideals aiR (by the definition of I). Thus b=air, say. But ai must lie in I. It follows from this that b and ai are associates. Clearly then ajR=aiR for any j>=i, as required to show the existence of factorizations.

We now have to show that factorizations are unique. But we know (since we have a PID) that the concepts of primeness and irreducibility coincide. Thus it is easy to adapt the familiar proof of the fundmental theorem of arithmetic to this case.

Log in or register to write something here or to contact authors.