Let A be a noetherian ring, F a closed subset of Spec(A), U the complementary open set, and u:M→N a homomorphism of finitely generated A-modules.
a) Suppose that up:Mp→Np is injective for every p∈U and that one has profF(M)⩾1. Then u is injective.
b) Suppose that up:Mp→Np is bijective for every p∈U and that one has profF(M)⩾2 and profF(N)⩾1. Then u is bijective.
a) The hypotheses imply Supp(Keru)⊂F, hence HomA(Keru,M)=0 (No. 5, Remark 1); therefore one has Keru=0.
b) One already knows that u is injective, and one has Supp(Cokeru)⊂F. By loc. cit., one has HomA(Cokeru,N)=0 and ExtA1(Cokeru,M)=0. From the exact sequence of extension modules
HomA(Cokeru,N)→HomA(Cokeru,Cokeru)→ExtA1(Cokeru,M)
one deduces HomA(Cokeru,Cokeru)=0, which implies Cokeru=0.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.
Nothing.
Nothing.