The Stacks project

Lemma 91.16.4. Let $\mathcal{C}$ be a site. Let $\mathcal{O} \to \mathcal{O}_0$ be a surjection of sheaves of rings whose kernel is an ideal sheaf $\mathcal{I}$ of square zero. For every object $K_0$ in $D^-(\mathcal{O}_0)$ the following are equivalent

  1. the class $\omega (K_0) \in \mathop{\mathrm{Ext}}\nolimits ^2_{\mathcal{O}_0}(K_0, K_0 \otimes _{\mathcal{O}_0} \mathcal{I})$ constructed in Lemma 91.16.2 is zero,

  2. there exists $K \in D^-(\mathcal{O})$ with $K \otimes _\mathcal {O}^\mathbf {L} \mathcal{O}_0 = K_0$ in $D(\mathcal{O}_0)$.

Proof. Let $K$ be as in (2). Then we can represent $K$ by a bounded above complex $\mathcal{F}^\bullet $ of flat $\mathcal{O}$-modules. Then $\mathcal{F}_0^\bullet = \mathcal{F}^\bullet \otimes _{\mathcal{O}} \mathcal{O}_0$ represents $K_0$ in $D(\mathcal{O}_0)$. Since $d_{\mathcal{F}^\bullet } \circ d_{\mathcal{F}^\bullet } = 0$ as $\mathcal{F}^\bullet $ is a complex, we see from the very construction of $\omega (K_0)$ that it is zero.

Assume (1). Let $\mathcal{F}^ n$, $d : \mathcal{F}^ n \to \mathcal{F}^{n + 1}$ be as in the construction of $\omega (K_0)$. The nullity of $\omega (K_0)$ implies that the map

\[ \omega = d \circ d : \mathcal{F}_0^\bullet \longrightarrow (\mathcal{F}_0^\bullet \otimes _{\mathcal{O}_0} \mathcal{I})[2] \]

is zero in $D(\mathcal{O}_0)$. By definition of the derived category as the localization of the homotopy category of complexes of $\mathcal{O}_0$-modules, there exists a quasi-isomorphism $\alpha : \mathcal{G}_0^\bullet \to \mathcal{F}_0^\bullet $ such that there exist $\mathcal{O}_0$-modules maps $h^ n : \mathcal{G}_0^ n \to \mathcal{F}_0^{n + 1} \otimes _\mathcal {O} \mathcal{I}$ with

\[ \omega \circ \alpha = d_{\mathcal{F}_0^\bullet \otimes \mathcal{I}} \circ h + h \circ d_{\mathcal{G}_0^\bullet } \]

We set

\[ \mathcal{H}^ n = \mathcal{F}^ n \times _{\mathcal{F}^ n_0} \mathcal{G}_0^ n \]

and we define

\[ d' : \mathcal{H}^ n \longrightarrow \mathcal{H}^{n + 1},\quad (f^ n, g_0^ n) \longmapsto (d(f^ n) - h^ n(g_0^ n), d(g_0^ n)) \]

with obvious notation using that $\mathcal{F}_0^{n + 1} \otimes _{\mathcal{O}_0} \mathcal{I} = \mathcal{F}^{n + 1} \otimes _\mathcal {O} \mathcal{I} = \mathcal{I}\mathcal{F}^{n + 1} \subset \mathcal{F}^{n + 1}$. Then one checks $d' \circ d' = 0$ by our choice of $h^ n$ and definition of $\omega $. Hence $\mathcal{H}^\bullet $ defines an object in $D(\mathcal{O})$. On the other hand, there is a short exact sequence of complexes of $\mathcal{O}$-modules

\[ 0 \to \mathcal{F}_0^\bullet \otimes _{\mathcal{O}_0} \mathcal{I} \to \mathcal{H}^\bullet \to \mathcal{G}_0^\bullet \to 0 \]

We still have to show that $\mathcal{H}^\bullet \otimes _\mathcal {O}^\mathbf {L} \mathcal{O}_0$ is isomorphic to $K_0$. Choose a quasi-isomorphism $\mathcal{E}^\bullet \to \mathcal{H}^\bullet $ where $\mathcal{E}^\bullet $ is a bounded above complex of flat $\mathcal{O}$-modules. We obtain a commutative diagram

\[ \xymatrix{ 0 \ar[r] & \mathcal{E}^\bullet \otimes _\mathcal {O} \mathcal{I} \ar[d]^\beta \ar[r] & \mathcal{E}^\bullet \ar[d]^\gamma \ar[r] & \mathcal{E}_0^\bullet \ar[d]^\delta \ar[r] & 0 \\ 0 \ar[r] & \mathcal{F}_0^\bullet \otimes _{\mathcal{O}_0} \mathcal{I} \ar[r] & \mathcal{H}^\bullet \ar[r] & \mathcal{G}_0^\bullet \ar[r] & 0 } \]

We claim that $\delta $ is a quasi-isomorphism. Since $H^ i(\delta )$ is an isomorphism for $i \gg 0$, we can use descending induction on $n$ such that $H^ i(\delta )$ is an isomorphism for $i \geq n$. Observe that $\mathcal{E}^\bullet \otimes _\mathcal {O} \mathcal{I}$ represents $\mathcal{E}_0^\bullet \otimes _{\mathcal{O}_0}^\mathbf {L} \mathcal{I}$, that $\mathcal{F}_0^\bullet \otimes _{\mathcal{O}_0} \mathcal{I}$ represents $\mathcal{G}_0^\bullet \otimes _{\mathcal{O}_0}^\mathbf {L} \mathcal{I}$, and that $\beta = \delta \otimes _{\mathcal{O}_0}^\mathbf {L} \text{id}_\mathcal {I}$ as maps in $D(\mathcal{O}_0)$. This is true because $\beta = (\alpha \otimes \text{id}_\mathcal {I}) \circ (\delta \otimes \text{id}_\mathcal {I})$. Suppose that $H^ i(\delta )$ is an isomorphism in degrees $\geq n$. Then the same is true for $\beta $ by what we just said and Lemma 91.16.3. Then we can look at the diagram

\[ \xymatrix{ H^{n - 1}(\mathcal{E}^\bullet \otimes _\mathcal {O} \mathcal{I}) \ar[r] \ar[d]^{H^{n - 1}(\beta )} & H^{n - 1}(\mathcal{E}^\bullet ) \ar[r] \ar[d] & H^{n - 1}(\mathcal{E}_0^\bullet ) \ar[r] \ar[d]^{H^{n - 1}(\delta )} & H^ n(\mathcal{E}^\bullet \otimes _\mathcal {O} \mathcal{I}) \ar[r] \ar[d]^{H^ n(\beta )} & H^ n(\mathcal{E}^\bullet ) \ar[d] \\ H^{n - 1}(\mathcal{F}_0^\bullet \otimes _\mathcal {O} \mathcal{I}) \ar[r] & H^{n - 1}(\mathcal{H}^\bullet ) \ar[r] & H^{n - 1}(\mathcal{G}_0^\bullet ) \ar[r] & H^ n(\mathcal{F}_0^\bullet \otimes _\mathcal {O} \mathcal{I}) \ar[r] & H^ n(\mathcal{H}^\bullet ) } \]

Using Homology, Lemma 12.5.19 we see that $H^{n - 1}(\delta )$ is surjective. This in turn implies that $H^{n - 1}(\beta )$ is surjective by Lemma 91.16.3. Using Homology, Lemma 12.5.19 again we see that $H^{n - 1}(\delta )$ is an isomorphism. The claim holds by induction, so $\delta $ is a quasi-isomorphism which is what we wanted to show. $\square$


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 0DIW. Beware of the difference between the letter 'O' and the digit '0'.