Canonical: https://knowlpedia.com/algebra-commutative/dedekind-domain/

# Dedekind domain

A Noetherian, integrally closed domain of Krull dimension one; equivalently, a nonfield domain with unique factorization of nonzero proper ideals into primes.

An [integral domain](https://knowlpedia.com/algebra-rings/integral-domain/) \(R\) is a **Dedekind domain** if:

1. \(R\) is a [Noetherian ring](https://knowlpedia.com/algebra-commutative/noetherian-ring/);
2. \(R\) is an [integrally closed domain](https://knowlpedia.com/algebra-commutative/integrally-closed-domain/);
3. \(R\) has [Krull dimension](https://knowlpedia.com/algebra-commutative/krull-dimension/) \(1\) (equivalently, \(R\) is not a field and every nonzero [prime ideal](https://knowlpedia.com/algebra-rings/prime-ideal/) is maximal).

We use the convention excluding fields. With this convention, every nonzero prime has [height](https://knowlpedia.com/algebra-commutative/height-of-prime/) one; some authors allow fields by requiring dimension at most one instead.

## Equivalent and characteristic properties

For a domain \(R\) that is not a field, each of the following characterizes a Dedekind domain:

- Every nonzero proper [ideal](https://knowlpedia.com/algebra-rings/ideal/) factors as a product of [prime ideals](https://knowlpedia.com/algebra-rings/prime-ideal/), and this factorization is unique up to ordering.
- The ring is Noetherian, and for every nonzero prime ideal \(\mathfrak p\), the localization \(R_\mathfrak p\) (see [localization at a prime](https://knowlpedia.com/algebra-commutative/localization-at-prime/)) is a [discrete valuation ring](https://knowlpedia.com/algebra-commutative/dvr/). The Noetherian hypothesis is required in this characterization.
- Every nonzero [fractional ideal](https://knowlpedia.com/algebra-commutative/fractional-ideal/) is [invertible](https://knowlpedia.com/algebra-commutative/invertible-fractional-ideal/).

## Examples

1. **The integers.**
   \(\mathbb{Z}\) is a Dedekind domain: it is Noetherian, integrally closed in \(\mathbb{Q}\), and has dimension \(1\). (More generally, any PID that is not a field is Dedekind; for instance, any [Euclidean domain](https://knowlpedia.com/algebra-rings/euclidean-domain/) that is not a field is a PID by [Euclidean ⇒ PID](https://knowlpedia.com/algebra-rings/euclidean-implies-pid/) and hence Dedekind.)

2. **A principal ideal domain of dimension one.**
   For a field \(k\), the [polynomial ring](https://knowlpedia.com/algebra-rings/polynomial-ring/) \(k[t]\) is a PID, hence Dedekind.

3. **Rings of integers in number fields.**
   If \(K/\mathbb{Q}\) is a [number field](https://knowlpedia.com/algebra-fields-galois/number-field/) and \(\mathcal O_K\) is its [ring of integers](https://knowlpedia.com/algebra-fields-galois/ring-of-integers/), then \(\mathcal O_K\) is a Dedekind domain (a key reason prime ideal factorization replaces unique factorization of elements).

## Non-example

- \(k[x,y]\) (two variables over a field \(k\)) is Noetherian and integrally closed, but it has Krull dimension \(2\), so it is not Dedekind.

## References

1. J. S. Milne, *Algebraic Number Theory*, [author’s text](https://www.jmilne.org/math/CourseNotes/ANTc.pdf), §3, Definition 3.3, Proposition 3.6, Theorem 3.7, and Theorem 3.20. The discussion after Proposition 3.6 explains the field convention and why the Noetherian hypothesis cannot be omitted.
