Processing math: 100%

The Stacks project

57.5 Characterizing coherent modules

This section is in some sense a continuation of the discussion in Derived Categories of Schemes, Section 36.34 and More on Morphisms, Section 37.69.

Before we can state the result we need some notation. Let k be a field. Let n \geq 0 be an integer. Let S = k[X_0, \ldots , X_ n]. For an integer e denote S_ e \subset S the homogeneous polynomials of degree e. Consider the (noncommutative) k-algebra

R = \left( \begin{matrix} S_0 & S_1 & S_2 & \ldots & \ldots \\ 0 & S_0 & S_1 & \ldots & \ldots \\ 0 & 0 & S_0 & \ldots & \ldots \\ \ldots & \ldots & \ldots & \ldots & \ldots \\ 0 & \ldots & \ldots & \ldots & S_0 \end{matrix} \right)

(with n + 1 rows and columns) with obvious multiplication and addition.

Lemma 57.5.1. With k, n, and R as above, for an object K of D(R) the following are equivalent

  1. \sum _{i \in \mathbf{Z}} \dim _ k H^ i(K) < \infty , and

  2. K is a compact object.

Proof. If K is a compact object, then K can be represented by a complex M^\bullet which is finite projective as a graded R-module, see Differential Graded Algebra, Lemma 22.36.6. Since \dim _ k R < \infty we conclude \sum \dim _ k M^ i < \infty and a fortiori \sum \dim _ k H^ i(M^\bullet ) < \infty . (One can also easily deduce this implication from the easier Differential Graded Algebra, Proposition 22.36.4.)

Assume K satisfies (1). Consider the distinguished triangle of trunctions \tau _{\leq m}K \to K \to \tau _{\geq m + 1}K, see Derived Categories, Remark 13.12.4. It is clear that both \tau _{\leq m}K and \tau _{\geq m + 1} K satisfy (1). If we can show both are compact, then so is K, see Derived Categories, Lemma 13.37.2. Hence, arguing on the number of nonzero cohomology modules of K we may assume H^ i(K) is nonzero only for one i. Shifting, we may assume K is given by the complex consisting of a single finite dimensional R-module M sitting in degree 0.

Since \dim _ k(M) < \infty we see that M is Artinian as an R-module. Thus it suffices to show that every simple R-module represents a compact object of D(R). Observe that

I = \left( \begin{matrix} 0 & S_1 & S_2 & \ldots & \ldots \\ 0 & 0 & S_1 & \ldots & \ldots \\ 0 & 0 & 0 & \ldots & \ldots \\ \ldots & \ldots & \ldots & \ldots & \ldots \\ 0 & \ldots & \ldots & \ldots & 0 \end{matrix} \right)

is a nilpotent two sided ideal of R and that R/I is a commutative k-algebra isomorphic to a product of n + 1 copies of k (placed along the diagonal in the matrix, i.e., R/I can be lifted to a k-subalgebra of R). It follows that R has exactly n + 1 isomorphism classes of simple modules M_0, \ldots , M_ n (sitting along the diagonal). Consider the right R-module P_ i of row vectors

P_ i = \left( \begin{matrix} 0 & \ldots & 0 & S_0 & \ldots & S_{i - 1} & S_ i \end{matrix} \right)

with obvious multiplication P_ i \times R \to P_ i. Then we see that R \cong P_0 \oplus \ldots \oplus P_ n as a right R-module. Since clearly R is a compact object of D(R), we conclude each P_ i is a compact object of D(R). (We of course also conclude each P_ i is projective as an R-module, but this isn't what we have to show in this proof.) Clearly, P_0 = M_0 is the first of our simple R-modules. For P_1 we have a short exact sequence

0 \to P_0^{\oplus n + 1} \to P_1 \to M_1 \to 0

which proves that M_1 fits into a distinguished triangle whose other members are compact objects and hence M_1 is a compact object of D(R). More generally, there exists a short exact sequence

0 \to C_ i \to P_ i \to M_ i \to 0

where C_ i is a finite dimensional R-module whose simple constituents are isomorphic to M_ j for j < i. By induction, we first conclude that C_ i determines a compact object of D(R) whereupon we conclude that M_ i does too as desired. \square

Lemma 57.5.2. Let k be a field. Let n \geq 0. Let K \in D_\mathit{QCoh}(\mathcal{O}_{\mathbf{P}^ n_ k}). The following are equivalent

  1. K is in D^ b_{\textit{Coh}}(\mathcal{O}_{\mathbf{P}^ n_ k}),

  2. \sum _{i \in \mathbf{Z}} \dim _ k H^ i(\mathbf{P}^ n_ k, E \otimes ^\mathbf {L} K) < \infty for each perfect object E of D(\mathcal{O}_{\mathbf{P}^ n_ k}),

  3. \sum _{i \in \mathbf{Z}} \dim _ k \mathop{\mathrm{Ext}}\nolimits ^ i_{\mathbf{P}^ n_ k}(E, K) < \infty for each perfect object E of D(\mathcal{O}_{\mathbf{P}^ n_ k}),

  4. \sum _{i \in \mathbf{Z}} \dim _ k H^ i(\mathbf{P}^ n_ k, K \otimes ^\mathbf {L} \mathcal{O}_{\mathbf{P}^ n_ k}(d)) < \infty for d = 0, 1, \ldots , n.

Proof. Parts (2) and (3) are equivalent by Cohomology, Lemma 20.50.5. If (1) is true, then for E perfect the derived tensor product E \otimes ^\mathbf {L} K is in D^ b_{\textit{Coh}}(\mathcal{O}_{\mathbf{P}^ n_ k}) and we see that (2) holds by Derived Categories of Schemes, Lemma 36.11.3. It is clear that (2) implies (4) as \mathcal{O}_{\mathbf{P}^ n_ k}(d) can be viewed as a perfect object of the derived category of \mathbf{P}^ n_ k. Thus it suffices to prove that (4) implies (1).

Assume (4). Let R be as in Lemma 57.5.1. Let P = \bigoplus _{d = 0, \ldots , n} \mathcal{O}_{\mathbf{P}^ n_ k}(-d). Recall that R = \text{End}_{\mathbf{P}^ n_ k}(P) whereas all other self-Exts of P are zero and that P determines an equivalence - \otimes ^\mathbf {L} P : D(R) \to D_\mathit{QCoh}(\mathcal{O}_{\mathbf{P}^ n_ k}) by Derived Categories of Schemes, Lemma 36.20.1. Say K corresponds to L in D(R). Then

\begin{align*} H^ i(L) & = \mathop{\mathrm{Ext}}\nolimits ^ i_{D(R)}(R, L) \\ & = \mathop{\mathrm{Ext}}\nolimits ^ i_{\mathbf{P}^ n_ k}(P, K) \\ & = H^ i(\mathbf{P}^ n_ k, K \otimes P^\vee ) \\ & = \bigoplus \nolimits _{d = 0, \ldots , n} H^ i(\mathbf{P}^ n_ k, K \otimes \mathcal{O}(d)) \end{align*}

by Differential Graded Algebra, Lemma 22.35.4 (and the fact that - \otimes ^\mathbf {L} P is an equivalence) and Cohomology, Lemma 20.50.5. Thus our assumption (4) implies that L satisfies condition (2) of Lemma 57.5.1 and hence is a compact object of D(R). Therefore K is a compact object of D_\mathit{QCoh}(\mathcal{O}_{\mathbf{P}^ n_ k}). Thus K is perfect by Derived Categories of Schemes, Proposition 36.17.1. Since D_{perf}(\mathcal{O}_{\mathbf{P}^ n_ k}) = D^ b_{\textit{Coh}}(\mathcal{O}_{\mathbf{P}^ n_ k}) by Derived Categories of Schemes, Lemma 36.11.8 we conclude (1) holds. \square

Lemma 57.5.3. Let X be a scheme proper over a field k. Let K \in D^ b_{\textit{Coh}}(\mathcal{O}_ X) and let E in D(\mathcal{O}_ X) be perfect. Then \sum _{i \in \mathbf{Z}} \dim _ k \mathop{\mathrm{Ext}}\nolimits ^ i_ X(E, K) < \infty .

Proof. This follows for example by combining Derived Categories of Schemes, Lemmas 36.11.7 and 36.13.5. Alternative proof: combine Derived Categories of Schemes, Lemmas 36.11.6 and 36.11.3. \square

Lemma 57.5.4.reference Let X be a proper scheme over a field k. Let K \in \mathop{\mathrm{Ob}}\nolimits (D_\mathit{QCoh}(\mathcal{O}_ X)). The following are equivalent

  1. K \in D^ b_{\textit{Coh}}(\mathcal{O}_ X), and

  2. \sum _{i \in \mathbf{Z}} \dim _ k \mathop{\mathrm{Ext}}\nolimits ^ i_ X(E, K) < \infty for all perfect E in D(\mathcal{O}_ X).

Proof. The implication (1) \Rightarrow (2) follows from Lemma 57.5.3. The implication (2) \Rightarrow (1) follows from More on Morphisms, Lemma 37.69.6 (see Derived Categories of Schemes, Example 36.35.2 for the meaning of a relatively perfect object over a field); the easier proof in the projective case is in the next paragraph.

Assume (2) and X projective over k. Choose a closed immersion i : X \to \mathbf{P}^ n_ k. It suffices to show that Ri_*K is in D^ b_{\textit{Coh}}(\mathbf{P}^ n_ k) since a quasi-coherent module \mathcal{F} on X is coherent, resp. zero if and only if i_*\mathcal{F} is coherent, resp. zero. For a perfect object E of D(\mathcal{O}_{\mathbf{P}^ n_ k}), Li^*E is a perfect object of D(\mathcal{O}_ X) and

\mathop{\mathrm{Ext}}\nolimits ^ q_{\mathbf{P}^ n_ k}(E, Ri_*K) = \mathop{\mathrm{Ext}}\nolimits ^ q_ X(Li^*E, K)

Hence by our assumption we see that \sum _{q \in \mathbf{Z}} \dim _ k \mathop{\mathrm{Ext}}\nolimits ^ q_{\mathbf{P}^ n_ k}(E, Ri_*K) < \infty . We conclude by Lemma 57.5.2. \square


Comments (0)


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.