Talk:Idempotents in group rings (Ex)
Recall that by the representation theory of finite groups we have
since all irreducibles are -dimensional as the group is abelian. Hence the idempotents of are precisely -tupels with entries either or , since fulfills if and only if each fulfills which implies the claim.
Let us turn to the other two rings and let us focus on the caseTex syntax errorfirst. It is a standard fact that this ring is isomorphic to
Tex syntax error
where the last equality holds since we are in characteristic .
Now it is a fact from commutative algebra that the ringTex syntax erroris local. This can be shown as follows. Suppose
Tex syntax erroris a maximal ideal. We let
Tex syntax errorbe its preimage under the canonical projection. It is a prime ideal that contains the ideal and hence must contain the ideal . But since is maximal in
Tex syntax errorwe must have equality, i.e., and hence is the image of the ideal generated by which determines uniquely. Thus
Tex syntax errorhas a unique maximal ideal. So we have shown that
Tex syntax erroris a local ring. In particular every element fulfills that either or is invertible (in a local ring the complement of the maximal ideal consists of the units). Thus if is an idempotent it must be or . We can put this in more geometric terms as follows. The ring
Tex syntax erroris a local, artinian ring (this ring is not only finitely generated as an
Tex syntax error-algebra but also as an
Tex syntax error-module and is thus artinian) and hence its spectrum consists of precisely one point (every prime ideal is maximal). But a non-trivial idempotent for any ring decomposes the ring into the product and hence the spectrum of is a disjoint union of two non-empty subsets. Of course, this cannot be the case if the spectrum consists only of one point.
We now claim that also in the ring for general prime numbers there are only the trivial idempotents. This follows from the previous considerations as follows. Suppose we have an idempotent . Then we have just shown that the reduction mod of is either or . Without loss of generality we may assume that it is zero. But then it follows that for some . But since is an idempotent it fulfills for all . Calculating shows that is divisible by for all which implies that .