The Stacks project

Lemma 91.15.1. Let $R' \to R$ be a surjection of rings whose kernel is an ideal $I$ of square zero. For every $K \in D^-(R)$ there is a canonical map

\[ \omega (K) : K \longrightarrow K \otimes _ R^\mathbf {L} I[2] \]

in $D(R)$ with the following properties

  1. $\omega (K) = 0$ if and only if there exists $K' \in D(R')$ with $K' \otimes _{R'}^\mathbf {L} R = K$,

  2. given $K \to L$ in $D^-(R)$ the diagram

    \[ \xymatrix{ K \ar[d] \ar[rr]_-{\omega (K)} & & K \otimes ^\mathbf {L}_ R I[2] \ar[d] \\ L \ar[rr]^-{\omega (L)} & & L \otimes ^\mathbf {L}_ R I[2] } \]

    commutes, and

  3. formation of $\omega (K)$ is compatible with ring maps $R' \to S'$ (see proof for a precise statement).

Proof. Choose a bounded above complex $K^\bullet $ of free $R$-modules representing $K$. Then we can choose free $R'$-modules $(K')^ n$ lifting $K^ n$. We can choose $R'$-module maps $(d')^ n_ K : (K')^ n \to (K')^{n + 1}$ lifting the differentials $d^ n_ K : K^ n \to K^{n + 1}$ of $K^\bullet $. Although the compositions

\[ (d')^{n + 1}_ K \circ (d')^ n_ K : (K')^ n \to (K')^{n + 2} \]

may not be zero, they do factor as

\[ (K')^ n \to K^ n \xrightarrow {\omega ^ n_ K} K^{n + 2} \otimes _ R I = I(K')^{n + 2} \to (K')^{n + 2} \]

because $d^{n + 1} \circ d^ n = 0$. A calculation shows that $\omega ^ n_ K$ defines a map of complexes. This map of complexes defines $\omega (K)$.

Let us prove this construction is compatible with a map of complexes $\alpha ^\bullet : K^\bullet \to L^\bullet $ of bounded above free $R$-modules and given choices of lifts $(K')^ n, (L')^ n, (d')^ n_ K, (d')^ n_ L$. Namely, choose $(\alpha ')^ n : (K')^ n \to (L')^ n$ lifting the components $\alpha ^ n : K^ n \to L^ n$. As before we get a factorization

\[ (K')^ n \to K^ n \xrightarrow {h^ n} L^{n + 1} \otimes _ R I = I(L')^{n + 1} \to (L')^{n + 2} \]

of $(d')^ n_ L \circ (\alpha ')^ n - (\alpha ')^{n + 1} \circ (d')_ K^ n$. Then it is an pleasant calculation to show that

\[ \omega ^ n_ L \circ \alpha ^ n = (d_ L^{n + 1} \otimes \text{id}_ I) \circ h^ n + h^{n + 1} \circ d_ K^ n + (\alpha ^{n + 2} \otimes \text{id}_ I) \circ \omega ^ n_ K \]

This proves the commutativity of the diagram in (2) of the lemma in this particular case. Using this for two different choices of bounded above free complexes representing $K$, we find that $\omega (K)$ is well defined! And of course (2) holds in general as well.

If $K$ lifts to $K'$ in $D^-(R')$, then we can represent $K'$ by a bounded above complex of free $R'$-modules and we see immediately that $\omega (K) = 0$. Conversely, going back to our choices $K^\bullet $, $(K')^ n$, $(d')^ n_ K$, if $\omega (K) = 0$, then we can find $g^ n : K^ n \to K^{n + 1} \otimes _ R I$ with

\[ \omega ^ n = (d_ K^{n + 1} \otimes \text{id}_ I) \circ g^ n + g^{n + 1} \circ d_ K^ n \]

This means that with differentials $(d')^ n_ K - g^ n : (K')^ n \to (K')^{n + 1}$ we obtain a complex of free $R'$-modules lifting $K^\bullet $. This proves (1).

Finally, part (3) means the following: Let $R' \to S'$ be a map of rings. Set $S = S' \otimes _{R'} R$ and denote $J = IS' \subset S'$ the square zero kernel of $S' \to S$. Then given $K \in D^-(R)$ the statement is that we get a commutative diagram

\[ \xymatrix{ K \otimes _ R^\mathbf {L} S \ar[d] \ar[rr]_-{\omega (K) \otimes \text{id}} & & (K \otimes ^\mathbf {L}_ R I[2]) \otimes _ R^\mathbf {L} S \ar[d] \\ K \otimes _ R^\mathbf {L} S \ar[rr]^-{\omega (K \otimes _ R^\mathbf {L} S)} & & (K \otimes _ R^\mathbf {L} S) \otimes ^\mathbf {L}_ S J[2] } \]

Here the right vertical arrow comes from

\[ (K \otimes ^\mathbf {L}_ R I[2]) \otimes _ R^\mathbf {L} S = (K \otimes _ R^\mathbf {L} S) \otimes _ S^\mathbf {L} (I \otimes _ R^\mathbf {L} S)[2] \longrightarrow (K \otimes _ R^\mathbf {L} S) \otimes _ S^\mathbf {L} J[2] \]

Choose $K^\bullet $, $(K')^ n$, and $(d')^ n_ K$ as above. Then we can use $K^\bullet \otimes _ R S$, $(K')^ n \otimes _{R'} S'$, and $(d')^ n_ K \otimes \text{id}_{S'}$ for the construction of $\omega (K \otimes _ R^\mathbf {L} S)$. With these choices commutativity is immediately verified on the level of maps of complexes. $\square$


Comments (2)

Comment #8725 by Shiji Lyu on

In the second last paragraph, it looks like we should have , not . Am I right?


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