The quotient of the ring of integers by an ideal always seems to have the same structure

952 Views Asked by At

Can the following be proven or does there exist a counter example:

The quotient of the ring of integers of a numnber field by an ideal is the direct sum of rings that are either isomorphic to $\mathbb{Z}_m$ or to $\mathbb{F}_q$, where $m$ is an integer and $q$ is the power of a prime.

I don't know if the following outline of a proof is correct. Let $Q$ be the quotient under consideration, and $q_1, \ldots, q_k$ the set of idempotents that are nozero and not the sum of other idempotents (they correspond to the elements of the form $(0,\ldots,1,\ldots,0)$ of Q. We have that $Q=Q.q_1+\ldots+Q.q_k$. Now for a certain $i$ consider the map $Q \rightarrow Q_i=q_1.Q$ that maps $q_1, \ldots, q_k$ to zero except $q_i \mapsto q_i$. The composition of the original quotient map with this map has an ideal as kernel. Either this ideal is maximal (and thus prime) and $Q_i$ is a finite field, either this ideal is not maximal and so $Q_i$ is not a field and the prime ideals of $Q_i$ determine the the prime factorization of the integer $m$.

Marc Bogaerts

1

There are 1 best solutions below

4
On

A simple counterexample is $\mathbf Z[i]/(2)$. It has size 4, characteristic 2, is not a field, and is not a direct sum of two or more nonzero rings since it is a local ring (i.e., it has one maximal ideal) while the direct sum of two or more nonzero rings is not a local ring.

More generally, suppose for a number field $K$ there is a prime number $p$ such that the ideal $(p)$ in the integers of $K$ is divisible by a prime ideal power $\mathfrak p^r$ where $r > 1$. Then the quotient ring $\mathcal O_K/{\mathfrak p^r}$ is a counterexample: it is not a field since $r>1$ and it has characteristic $p$ since $\mathfrak p^r \mid (p)$, so if it were isomorphic to some $\mathbf Z/(m)$ then that $m$ would have to be $p$, but that would make it a field. It is not isomorphic to a direct sum of the two types of rings you put in your question since it is a local ring, and local rings never decompose as a direct sum of two or more nonzero rings. Every number field other than $\mathbf Q$ has a ramified prime $p$, and (by definition) such a prime is divisible by the square of some prime ideal, so in fact the ring of integers of every number field besides the rationals is a counterexample!