The Stacks project

Lemma 10.99.16. Let $A$ be a ring, let $M$ be an $A$-module, and let $f \in A$. If

  1. $f$ is a nonzerodivisor on $A$ and $M$,

  2. $M_ f$ is a flat $A_ f$-module, and

  3. $M/fM$ is a flat $A/fA$-module,

Then $M$ is a flat $A$-module. Same with “flat” replaced by “faithfully flat”.

Proof. Since $0 \to A \to A \to A/fA \to 0$ and $0 \to M \to M \to M/fM \to 0$ are exact, we find that $\text{Tor}_ i^ A(M, A/fA) = 0$ for $i = 1$ and $i = 2$. By Lemma 10.99.8 we conclude that $\text{Tor}_1^ A(M, N) = 0$ for all $A$-modules $N$ annihilated by $f$ (this uses the flatness of $M/fM$ over $A/fA$). Given an $A$-module $N$ annihilated by $f$ we may choose a short exact sequence $0 \to N' \to F \to N \to 0$ of $A$-modules where $F$ is a direct sum of copies of $A/fA$. From the exact sequence

\[ \text{Tor}_2^ A(M, F) \to \text{Tor}_2^ A(M, N) \to \text{Tor}_1^ A(M, N') \]

we conclude that $\text{Tor}_2^ A(M, N) = 0$ for all $A$-modules $N$ annihilated by $f$. Next, let $K$ be an arbitrary $A$-module. We may break the map $f : K \to K$ into two short exact sequences

\[ 0 \to K[f] \to K \to K' \to 0 \quad \text{and}\quad 0 \to K' \to K \to K/fK \to 0 \]

where $K' = K/K[f] \cong fK$. Applying the exact sequences of Tor we obtain exact sequences

\[ \text{Tor}_1^ A(K[f], M) \to \text{Tor}_1^ A(K, M) \to \text{Tor}_1^ A(K', M) \]

and

\[ \text{Tor}_2^ A(K/fK, M) \to \text{Tor}_1^ A(K', M) \to \text{Tor}_1^ A(K, M) \]

Using the vanishing of $\text{Tor}_1^ A(K[f], M)$ and $\text{Tor}_2^ A(K/fK, M)$ we conclude that $f : K \to K$ induces an injective map on $\text{Tor}^ A_1(M, K)$. In other words, we see that $\text{Tor}^ A_1(M, K)$ is zero if and only if $\text{Tor}^ A_1(M, K) \otimes _ A A_ f$ is zero. By Lemma 10.76.1 we have

\[ \text{Tor}^ A_1(M, K) \otimes _ A A_ f = \text{Tor}^{A_ f}_1(M_ f, K_ f) = 0 \]

The vanishing by the flatness of $M_ f$ over $A_ f$ (Lemma 10.75.8). We conclude that $\text{Tor}_1^ A(M, K) = 0$ for all $A$-modules $K$. Hence $M$ is flat over $A$ by Lemma 10.75.8.

We omit the argument for the case of faithfully flat modules. $\square$


Comments (0)


Post a comment

Your email address will not be published. Required fields are marked.

In your comment you can use Markdown and LaTeX style mathematics (enclose it like $\pi$). A preview option is available if you wish to see how it works out (just click on the eye in the toolbar).

Unfortunately JavaScript is disabled in your browser, so the comment preview function will not work.

All contributions are licensed under the GNU Free Documentation License.




In order to prevent bots from posting comments, we would like you to prove that you are human. You can do this by filling in the name of the current tag in the following input field. As a reminder, this is tag 0H7N. Beware of the difference between the letter 'O' and the digit '0'.