53.4 Duality
In this section we work out the consequences of the very general material on dualizing complexes and duality for proper 1-dimensional schemes over fields. If you are interested in the analogous discussion for higher dimension proper schemes over fields, see Duality for Schemes, Section 48.27.
Lemma 53.4.1. Let X be a proper scheme of dimension \leq 1 over a field k. There exists a dualizing complex \omega _ X^\bullet with the following properties
H^ i(\omega _ X^\bullet ) is nonzero only for i = -1, 0,
\omega _ X = H^{-1}(\omega _ X^\bullet ) is a coherent Cohen-Macaulay module whose support is the irreducible components of dimension 1,
for x \in X closed, the module H^0(\omega _{X, x}^\bullet ) is nonzero if and only if either
\dim (\mathcal{O}_{X, x}) = 0 or
\dim (\mathcal{O}_{X, x}) = 1 and \mathcal{O}_{X, x} is not Cohen-Macaulay,
for K \in D_\mathit{QCoh}(\mathcal{O}_ X) there are functorial isomorphisms1
\mathop{\mathrm{Ext}}\nolimits ^ i_ X(K, \omega _ X^\bullet ) = \mathop{\mathrm{Hom}}\nolimits _ k(H^{-i}(X, K), k)
compatible with shifts and distinguished triangles,
there are functorial isomorphisms \mathop{\mathrm{Hom}}\nolimits (\mathcal{F}, \omega _ X) = \mathop{\mathrm{Hom}}\nolimits _ k(H^1(X, \mathcal{F}), k) for \mathcal{F} quasi-coherent on X,
if X \to \mathop{\mathrm{Spec}}(k) is smooth of relative dimension 1, then \omega _ X \cong \Omega _{X/k}.
Proof.
Denote f : X \to \mathop{\mathrm{Spec}}(k) the structure morphism. We start with the relative dualizing complex
\omega _ X^\bullet = \omega _{X/k}^\bullet = a(\mathcal{O}_{\mathop{\mathrm{Spec}}(k)})
as described in Duality for Schemes, Remark 48.12.5. Then property (4) holds by construction as a is the right adjoint for f_* : D_\mathit{QCoh}(\mathcal{O}_ X) \to D(\mathcal{O}_{\mathop{\mathrm{Spec}}(k)}). Since f is proper we have f^!(\mathcal{O}_{\mathop{\mathrm{Spec}}(k)}) = a(\mathcal{O}_{\mathop{\mathrm{Spec}}(k)}) by definition, see Duality for Schemes, Section 48.16. Hence \omega _ X^\bullet and \omega _ X are as in Duality for Schemes, Example 48.22.1 and as in Duality for Schemes, Example 48.22.2. Parts (1) and (2) follow from Duality for Schemes, Lemma 48.22.4. For a closed point x \in X we see that \omega _{X, x}^\bullet is a normalized dualizing complex over \mathcal{O}_{X, x}, see Duality for Schemes, Lemma 48.21.1. Assertion (3) then follows from Dualizing Complexes, Lemma 47.20.2. Assertion (5) follows from Duality for Schemes, Lemma 48.22.5 for coherent \mathcal{F} and in general by unwinding (4) for K = \mathcal{F}[0] and i = -1. Assertion (6) follows from Duality for Schemes, Lemma 48.15.7.
\square
Lemma 53.4.2. Let X be a proper scheme over a field k which is Cohen-Macaulay and equidimensional of dimension 1. The module \omega _ X of Lemma 53.4.1 has the following properties
\omega _ X is a dualizing module on X (Duality for Schemes, Section 48.22),
\omega _ X is a coherent Cohen-Macaulay module whose support is X,
there are functorial isomorphisms \mathop{\mathrm{Ext}}\nolimits ^ i_ X(K, \omega _ X[1]) = \mathop{\mathrm{Hom}}\nolimits _ k(H^{-i}(X, K), k) compatible with shifts for K \in D_\mathit{QCoh}(X),
there are functorial isomorphisms \mathop{\mathrm{Ext}}\nolimits ^{1 + i}(\mathcal{F}, \omega _ X) = \mathop{\mathrm{Hom}}\nolimits _ k(H^{-i}(X, \mathcal{F}), k) for \mathcal{F} quasi-coherent on X.
Proof.
Recall from the proof of Lemma 53.4.1 that \omega _ X is as in Duality for Schemes, Example 48.22.1 and hence is a dualizing module. The other statements follow from Lemma 53.4.1 and the fact that \omega _ X^\bullet = \omega _ X[1] as X is Cohen-Macualay (Duality for Schemes, Lemma 48.23.1).
\square
Here is a sanity check for the dualizing complex.
Lemma 53.4.4. Let X be a proper scheme of dimension \leq 1 over a field k. Let \omega _ X^\bullet and \omega _ X be as in Lemma 53.4.1.
If X \to \mathop{\mathrm{Spec}}(k) factors as X \to \mathop{\mathrm{Spec}}(k') \to \mathop{\mathrm{Spec}}(k) for some field k', then \omega _ X^\bullet and \omega _ X satisfy properties (4), (5), (6) with k replaced with k'.
If K/k is a field extension, then the pullback of \omega _ X^\bullet and \omega _ X to the base change X_ K are as in Lemma 53.4.1 for the morphism X_ K \to \mathop{\mathrm{Spec}}(K).
Proof.
Denote f : X \to \mathop{\mathrm{Spec}}(k) the structure morphism. Assertion (1) really means that \omega _ X^\bullet and \omega _ X are as in Lemma 53.4.1 for the morphism f' : X \to \mathop{\mathrm{Spec}}(k'). In the proof of Lemma 53.4.1 we took \omega _ X^\bullet = a(\mathcal{O}_{\mathop{\mathrm{Spec}}(k)}) where a be is the right adjoint of Duality for Schemes, Lemma 48.3.1 for f. Thus we have to show a(\mathcal{O}_{\mathop{\mathrm{Spec}}(k)}) \cong a'(\mathcal{O}_{\mathop{\mathrm{Spec}}(k)}) where a' be is the right adjoint of Duality for Schemes, Lemma 48.3.1 for f'. Since k' \subset H^0(X, \mathcal{O}_ X) we see that k'/k is a finite extension (Cohomology of Schemes, Lemma 30.19.2). By uniqueness of adjoints we have a = a' \circ b where b is the right adjoint of Duality for Schemes, Lemma 48.3.1 for g : \mathop{\mathrm{Spec}}(k') \to \mathop{\mathrm{Spec}}(k). Another way to say this: we have f^! = (f')^! \circ g^!. Thus it suffices to show that \mathop{\mathrm{Hom}}\nolimits _ k(k', k) \cong k' as k'-modules, see Duality for Schemes, Example 48.3.2. This holds because these are k'-vector spaces of the same dimension (namely dimension 1).
Proof of (2). This holds because we have base change for a by Duality for Schemes, Lemma 48.6.2. See discussion in Duality for Schemes, Remark 48.12.5.
\square
Lemma 53.4.5. Let X be a proper scheme of dimension \leq 1 over a field k. Let i : Y \to X be a closed immersion. Let \omega _ X^\bullet , \omega _ X, \omega _ Y^\bullet , \omega _ Y be as in Lemma 53.4.1. Then
\omega _ Y^\bullet = R\mathop{\mathcal{H}\! \mathit{om}}\nolimits (\mathcal{O}_ Y, \omega _ X^\bullet ),
\omega _ Y = \mathop{\mathcal{H}\! \mathit{om}}\nolimits (\mathcal{O}_ Y, \omega _ X) and i_*\omega _ Y = \mathop{\mathcal{H}\! \mathit{om}}\nolimits _{\mathcal{O}_ X}(i_*\mathcal{O}_ Y, \omega _ X).
Proof.
Denote g : Y \to \mathop{\mathrm{Spec}}(k) and f : X \to \mathop{\mathrm{Spec}}(k) the structure morphisms. Then g = f \circ i. Denote a, b, c the right adjoint of Duality for Schemes, Lemma 48.3.1 for f, g, i. Then b = c \circ a by uniqueness of right adjoints and because Rg_* = Rf_* \circ Ri_*. In the proof of Lemma 53.4.1 we set \omega _ X^\bullet = a(\mathcal{O}_{\mathop{\mathrm{Spec}}(k)}) and \omega _ Y^\bullet = b(\mathcal{O}_{\mathop{\mathrm{Spec}}(k)}). Hence \omega _ Y^\bullet = c(\omega _ X^\bullet ) which implies (1) by Duality for Schemes, Lemma 48.9.7. Since \omega _ X = H^{-1}(\omega _ X^\bullet ) and \omega _ Y = H^{-1}(\omega _ Y^\bullet ) we conclude that \omega _ Y = \mathop{\mathcal{H}\! \mathit{om}}\nolimits (\mathcal{O}_ Y, \omega _ X). This implies i_*\omega _ Y = \mathop{\mathcal{H}\! \mathit{om}}\nolimits _{\mathcal{O}_ X}(i_*\mathcal{O}_ Y, \omega _ X) by Duality for Schemes, Lemma 48.9.3.
\square
Lemma 53.4.6. Let X be a proper scheme over a field k which is Gorenstein, reduced, and equidimensional of dimension 1. Let i : Y \to X be a reduced closed subscheme equidimensional of dimension 1. Let j : Z \to X be the scheme theoretic closure of X \setminus Y. Then
Y and Z are Cohen-Macaulay,
if \mathcal{I} \subset \mathcal{O}_ X, resp. \mathcal{J} \subset \mathcal{O}_ X is the ideal sheaf of Y, resp. Z in X, then
\mathcal{I} = i_*\mathcal{I}' \quad \text{and}\quad \mathcal{J} = j_*\mathcal{J}'
where \mathcal{I}' \subset \mathcal{O}_ Z, resp. \mathcal{J}' \subset \mathcal{O}_ Y is the ideal sheaf of Y \cap Z in Z, resp. Y,
\omega _ Y = \mathcal{J}'(i^*\omega _ X) and i_*(\omega _ Y) = \mathcal{J}\omega _ X,
\omega _ Z = \mathcal{I}'(i^*\omega _ X) and i_*(\omega _ Z) = \mathcal{I}\omega _ X,
we have the following short exact sequences
\begin{align*} 0 \to \omega _ X \to i_*i^*\omega _ X \oplus j_*j^*\omega _ X \to \mathcal{O}_{Y \cap Z} \to 0 \\ 0 \to i_*\omega _ Y \to \omega _ X \to j_*j^*\omega _ X \to 0 \\ 0 \to j_*\omega _ Z \to \omega _ X \to i_*i^*\omega _ X \to 0 \\ 0 \to i_*\omega _ Y \oplus j_*\omega _ Z \to \omega _ X \to \mathcal{O}_{Y \cap Z} \to 0 \\ 0 \to \omega _ Y \to i^*\omega _ X \to \mathcal{O}_{Y \cap Z} \to 0 \\ 0 \to \omega _ Z \to j^*\omega _ X \to \mathcal{O}_{Y \cap Z} \to 0 \end{align*}
Here \omega _ X, \omega _ Y, \omega _ Z are as in Lemma 53.4.1.
Proof.
A reduced 1-dimensional Noetherian scheme is Cohen-Macaulay, so (1) is true. Since X is reduced, we see that X = Y \cup Z scheme theoretically. With notation as in Morphisms, Lemma 29.4.6 and by the statement of that lemma we have a short exact sequence
0 \to \mathcal{O}_ X \to \mathcal{O}_ Y \oplus \mathcal{O}_ Z \to \mathcal{O}_{Y \cap Z} \to 0
Since \mathcal{J} = \mathop{\mathrm{Ker}}(\mathcal{O}_ X \to \mathcal{O}_ Z), \mathcal{J}' = \mathop{\mathrm{Ker}}(\mathcal{O}_ Y \to \mathcal{O}_{Y \cap Z}), \mathcal{I} = \mathop{\mathrm{Ker}}(\mathcal{O}_ X \to \mathcal{O}_ Y), and \mathcal{I}' = \mathop{\mathrm{Ker}}(\mathcal{O}_ Z \to \mathcal{O}_{Y \cap Z}) a diagram chase implies (2). Observe that \mathcal{I} + \mathcal{J} is the ideal sheaf of Y \cap Z and that \mathcal{I} \cap \mathcal{J} = 0. Hence we have the following exact sequences
\begin{align*} 0 \to \mathcal{O}_ X \to \mathcal{O}_ Y \oplus \mathcal{O}_ Z \to \mathcal{O}_{Y \cap Z} \to 0 \\ 0 \to \mathcal{J} \to \mathcal{O}_ X \to \mathcal{O}_ Z \to 0 \\ 0 \to \mathcal{I} \to \mathcal{O}_ X \to \mathcal{O}_ Y \to 0 \\ 0 \to \mathcal{J} \oplus \mathcal{I} \to \mathcal{O}_ X \to \mathcal{O}_{Y \cap Z} \to 0 \\ 0 \to \mathcal{J}' \to \mathcal{O}_ Y \to \mathcal{O}_{Y \cap Z} \to 0 \\ 0 \to \mathcal{I}' \to \mathcal{O}_ Z \to \mathcal{O}_{Y \cap Z} \to 0 \end{align*}
Since X is Gorenstein \omega _ X is an invertible \mathcal{O}_ X-module (Duality for Schemes, Lemma 48.24.4). Since Y \cap Z has dimension 0 we have \omega _ X|_{Y \cap Z} \cong \mathcal{O}_{Y \cap Z}. Thus if we prove (3) and (4), then we obtain the short exact sequences of the lemma by tensoring the above short exact sequence with the invertible module \omega _ X. By symmetry it suffices to prove (3) and by (2) it suffices to prove i_*(\omega _ Y) = \mathcal{J}\omega _ X.
We have i_*\omega _ Y = \mathop{\mathcal{H}\! \mathit{om}}\nolimits _{\mathcal{O}_ X}(i_*\mathcal{O}_ Y, \omega _ X) by Lemma 53.4.5. Again using that \omega _ X is invertible we finally conclude that it suffices to show \mathop{\mathcal{H}\! \mathit{om}}\nolimits _{\mathcal{O}_ X}(\mathcal{O}_ X/\mathcal{I}, \mathcal{O}_ X) maps isomorphically to \mathcal{J} by evaluation at 1. In other words, that \mathcal{J} is the annihilator of \mathcal{I}. This follows from the above.
\square
Comments (0)