Let R be a commutative ring, I an ideal of R, and a∈R such that a is not a zero divisor. Then ReesR(I)≅ReesR(aI).
Proof
First, recall that (aI)n=anIn. This holds in any commutative ring.
We will construct the isomorphism φ:ReesR(I)→ReesR(aI) explicitly.
First, let’s define a family of maps φn:In→anIn as φn:x↦anx. Clearly, this map is an R-module homomorphism. We also define ψn:anIn→In as ψn:anx→x. We can do this, since multiplication by a is injective and all elements of anIn are multiples of an.
These maps are inverses of each other and an R-module homomorphism with an inverse (set-) function is an isomorphism. Now, let φ be the direct sum of all φn. Similarly, let ψ be the direct sum of all ψn.
Is it possible that φ fails to be the inverse of ψ? For every n, φn∘ψn=idanIn and similarly ψn∘φn=idIn. So φ∘ψ and ψ∘φ are both direct sums of identity arrows and thus must themselves be identities.
Now we need to check that φ is an R-algebra homomorphism. Recall (5.17) that (1,0,...) is the multiplicative identity of any Rees algebra. Since all components of φ preserve 0, and since its zeroth component is the identity arrow on R, φ preserves the multiplicative identity.
To verify that ψ preserves multiplication, let u,v∈ReesR(aI) and observe that