Spring 2016 - Problem 4

ring theory

Let AA be a commutative ring, SS a multiplicatively closed subset of AA, AA[S1]A \to A\br{S^{-1}} the localization.

  1. Which elements of AA map to zero in A[S1]A\br{S^{-1}}?
  2. Let p\mathfrak{p} be a prime ideal in AA. Show that the ideal generated by the image of p\mathfrak{p} in A[S1]A\br{S^{-1}} is prime if and only if the intersection of p\mathfrak{p} with SS is empty.
Solution.
  1. Suppose aa maps to 00 in A[S1]A\br{S^{-1}}; that is, a1=0\frac{a}{1} = 0. By definition, there exists sSs \in S so that as=0as = 0 in AA. Conversely, suppose aAa \in A is annihilated by SS, i.e., as=0as = 0 for some sSs \in S. Then in A[S1]A\br{S^{-1}}, we have

    a1=ass=0,\frac{a}{1} = \frac{as}{s} = 0,

    so the elements in AA which map to zero in the localization are precisely the elements annihilated by SS.

  2. "    \implies"

    Assume the ideal I=S1pI = S^{-1}\mathfrak{p} is prime. If there exists spSs \in \mathfrak{p} \cap S, then II is all of A[S1]A\br{S^{-1}}, but this means II is not prime. Thus, pS=\mathfrak{p} \cap S = \emptyset.

    "    \impliedby"

    Suppose pS=\mathfrak{p} \cap S = \emptyset, and assume asbtI\frac{a}{s} \cdot \frac{b}{t} \in I, where II is as above. Then abst=pr\frac{ab}{st} = \frac{p}{r} for some ppp \in \mathfrak{p}. Thus, rab=pstprab = pst \in \mathfrak{p}, and by assumption, rpr \notin \mathfrak{p}, so because p\mathfrak{p} is prime, we have abpab \in \mathfrak{p}. Thus, aa or bb is in p\mathfrak{p}, which implies as\frac{a}{s} or bt\frac{b}{t} is in II, so II is prime.