Induced map on spectra by a ring homomorphism

From Commalg
Revision as of 21:54, 2 February 2008 by Vipul (talk | contribs) (New page: Suppose <math>f:R \to S</math> is a homomorphism of commutative unital rings. Then, if <math>Spec(R)</math> and <math>Spec(S)</math> denote the spectra of the rings <math>...)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Suppose f:RS is a homomorphism of commutative unital rings. Then, if Spec(R) and Spec(S) denote the spectra of the rings R and S respectively, we have a functorially induced mapping:

f*:Spec(S)Spec(R)

sending a prime ideal of S to its contraction in R. The well-definedness of this map rests on the fact that the contraction of a prime ideal is prime.

Various ring-theoretic assumptions on the nature of f give set-theoretic or topological information about the map f*.

Surjective homomorphism yields injective map

If f:RS is surjective, the map f*:Spec(S)Spec(R) is injective. In fact, Spec(S) is identified with those elements of Spec(R) that contain the kernel of f.

Thus, the map is injective and its image is a closed subset. In fact, the map is a closed map with respect to the topologies on both sides.

This can be better understood in relation with the Galois correspondence between a ring and its spectrum.

Localization yields injective map

If S=U1R for some multiplicative subset U of R, and f is the natural map from R to S, then f* is again an injective map, but this time the image is not necessarily a closed set. f* identifies Spec(S) with those prime ideals of R that do not intersect U.

Finite morphism yields map with finite fibers