75.11 The coherator
Let S be a scheme. Let X be an algebraic space over S. The coherator is a functor
Q_ X : \textit{Mod}(\mathcal{O}_ X) \longrightarrow \mathit{QCoh}(\mathcal{O}_ X)
which is right adjoint to the inclusion functor \mathit{QCoh}(\mathcal{O}_ X) \to \textit{Mod}(\mathcal{O}_ X). It exists for any algebraic space X and moreover the adjunction mapping Q_ X(\mathcal{F}) \to \mathcal{F} is an isomorphism for every quasi-coherent module \mathcal{F}, see Properties of Spaces, Proposition 66.32.2. Since Q_ X is left exact (as a right adjoint) we can consider its right derived extension
RQ_ X : D(\mathcal{O}_ X) \longrightarrow D(\mathit{QCoh}(\mathcal{O}_ X)).
Since Q_ X is right adjoint to the inclusion functor \mathit{QCoh}(\mathcal{O}_ X) \to \textit{Mod}(\mathcal{O}_ X) we see that RQ_ X is right adjoint to the canonical functor D(\mathit{QCoh}(\mathcal{O}_ X)) \to D(\mathcal{O}_ X) by Derived Categories, Lemma 13.30.3.
In this section we will study the functor RQ_ X. In Section 75.19 we will study the (closely related) right adjoint to the inclusion functor D_\mathit{QCoh}(\mathcal{O}_ X) \to D(\mathcal{O}_ X) (when it exists).
Lemma 75.11.1. Let S be a scheme. Let f : X \to Y be an affine morphism of algebraic spaces over S. Then f_* defines a derived functor f_* : D(\mathit{QCoh}(\mathcal{O}_ X)) \to D(\mathit{QCoh}(\mathcal{O}_ Y)). This functor has the property that
\xymatrix{ D(\mathit{QCoh}(\mathcal{O}_ X)) \ar[d]_{f_*} \ar[r] & D_\mathit{QCoh}(\mathcal{O}_ X) \ar[d]^{Rf_*} \\ D(\mathit{QCoh}(\mathcal{O}_ Y)) \ar[r] & D_\mathit{QCoh}(\mathcal{O}_ Y) }
commutes.
Proof.
The functor f_* : \mathit{QCoh}(\mathcal{O}_ X) \to \mathit{QCoh}(\mathcal{O}_ Y) is exact, see Cohomology of Spaces, Lemma 69.8.2. Hence f_* defines a derived functor f_* : D(\mathit{QCoh}(\mathcal{O}_ X)) \to D(\mathit{QCoh}(\mathcal{O}_ Y)) by simply applying f_* to any representative complex, see Derived Categories, Lemma 13.16.9. For any complex of \mathcal{O}_ X-modules \mathcal{F}^\bullet there is a canonical map f_*\mathcal{F}^\bullet \to Rf_*\mathcal{F}^\bullet . To finish the proof we show this is a quasi-isomorphism when \mathcal{F}^\bullet is a complex with each \mathcal{F}^ n quasi-coherent. The statement is étale local on Y hence we may assume Y affine. As an affine morphism is representable we reduce to the case of schemes by the compatibility of Remark 75.6.3. The case of schemes is Derived Categories of Schemes, Lemma 36.7.1.
\square
Lemma 75.11.2. Let S be a scheme. Let f : X \to Y be a morphism of algebraic spaces over S. Assume f is quasi-compact, quasi-separated, and flat. Then, denoting
\Phi : D(\mathit{QCoh}(\mathcal{O}_ X)) \to D(\mathit{QCoh}(\mathcal{O}_ Y))
the right derived functor of f_* : \mathit{QCoh}(\mathcal{O}_ X) \to \mathit{QCoh}(\mathcal{O}_ Y) we have RQ_ Y \circ Rf_* = \Phi \circ RQ_ X.
Proof.
We will prove this by showing that RQ_ Y \circ Rf_* and \Phi \circ RQ_ X are right adjoint to the same functor D(\mathit{QCoh}(\mathcal{O}_ Y)) \to D(\mathcal{O}_ X).
Since f is quasi-compact and quasi-separated, we see that f_* preserves quasi-coherence, see Morphisms of Spaces, Lemma 67.11.2. Recall that \mathit{QCoh}(\mathcal{O}_ X) is a Grothendieck abelian category (Properties of Spaces, Proposition 66.32.2). Hence any K in D(\mathit{QCoh}(\mathcal{O}_ X)) can be represented by a K-injective complex \mathcal{I}^\bullet of \mathit{QCoh}(\mathcal{O}_ X), see Injectives, Theorem 19.12.6. Then we can define \Phi (K) = f_*\mathcal{I}^\bullet .
Since f is flat, the functor f^* is exact. Hence f^* defines f^* : D(\mathcal{O}_ Y) \to D(\mathcal{O}_ X) and also f^* : D(\mathit{QCoh}(\mathcal{O}_ Y)) \to D(\mathit{QCoh}(\mathcal{O}_ X)). The functor f^* = Lf^* : D(\mathcal{O}_ Y) \to D(\mathcal{O}_ X) is left adjoint to Rf_* : D(\mathcal{O}_ X) \to D(\mathcal{O}_ Y), see Cohomology on Sites, Lemma 21.19.1. Similarly, the functor f^* : D(\mathit{QCoh}(\mathcal{O}_ Y)) \to D(\mathit{QCoh}(\mathcal{O}_ X)) is left adjoint to \Phi : D(\mathit{QCoh}(\mathcal{O}_ X)) \to D(\mathit{QCoh}(\mathcal{O}_ Y)) by Derived Categories, Lemma 13.30.3.
Let A be an object of D(\mathit{QCoh}(\mathcal{O}_ Y)) and E an object of D(\mathcal{O}_ X). Then
\begin{align*} \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ Y))}(A, RQ_ Y(Rf_*E)) & = \mathop{\mathrm{Hom}}\nolimits _{D(\mathcal{O}_ Y)}(A, Rf_*E) \\ & = \mathop{\mathrm{Hom}}\nolimits _{D(\mathcal{O}_ X)}(f^*A, E) \\ & = \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(f^*A, RQ_ X(E)) \\ & = \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ Y))}(A, \Phi (RQ_ X(E))) \end{align*}
This implies what we want.
\square
Lemma 75.11.3. Let S be a scheme. Let X be an affine algebraic space over S. Set A = \Gamma (X, \mathcal{O}_ X). Then
Q_ X : \textit{Mod}(\mathcal{O}_ X) \to \mathit{QCoh}(\mathcal{O}_ X) is the functor which sends \mathcal{F} to the quasi-coherent \mathcal{O}_ X-module associated to the A-module \Gamma (X, \mathcal{F}),
RQ_ X : D(\mathcal{O}_ X) \to D(\mathit{QCoh}(\mathcal{O}_ X)) is the functor which sends E to the complex of quasi-coherent \mathcal{O}_ X-modules associated to the object R\Gamma (X, E) of D(A),
restricted to D_\mathit{QCoh}(\mathcal{O}_ X) the functor RQ_ X defines a quasi-inverse to (75.5.1.1).
Proof.
Let X_0 = \mathop{\mathrm{Spec}}(A) be the affine scheme representing X. Recall that there is a morphism of ringed sites \epsilon : X_{\acute{e}tale}\to X_{0, Zar} which induces equivalences
\xymatrix{ \mathit{QCoh}(\mathcal{O}_ X) \ar@<1ex>[r]^{{\epsilon _*}} & \mathit{QCoh}(\mathcal{O}_{X_0}) \ar@<1ex>[l]^{{\epsilon ^*}} }
see Lemma 75.4.2. Hence we see that Q_ X = \epsilon ^* \circ Q_{X_0} \circ \epsilon _* by uniqueness of adjoint functors. Hence (1) follows from the description of Q_{X_0} in Derived Categories of Schemes, Lemma 36.7.3 and the fact that \Gamma (X_0, \epsilon _*\mathcal{F}) = \Gamma (X, \mathcal{F}). Part (2) follows from (1) and the fact that the functor from A-modules to quasi-coherent \mathcal{O}_ X-modules is exact. The third assertion now follows from the result for schemes (Derived Categories of Schemes, Lemma 36.7.3) and Lemma 75.4.2.
\square
Next, we prove a criterion for when the functor D(\mathit{QCoh}(\mathcal{O}_ X)) \to D_\mathit{QCoh}(\mathcal{O}_ X) is an equivalence.
Lemma 75.11.4. Let S be a scheme. Let X be a quasi-compact and quasi-separated algebraic space over S. Suppose that for every étale morphism j : V \to W with W \subset X quasi-compact open and V affine the right derived functor
\Phi : D(\mathit{QCoh}(\mathcal{O}_ U)) \to D(\mathit{QCoh}(\mathcal{O}_ W))
of the left exact functor j_* : \mathit{QCoh}(\mathcal{O}_ V) \to \mathit{QCoh}(\mathcal{O}_ W) fits into a commutative diagram
\xymatrix{ D(\mathit{QCoh}(\mathcal{O}_ V)) \ar[d]_\Phi \ar[r]_{i_ V} & D_\mathit{QCoh}(\mathcal{O}_ V) \ar[d]^{Rj_*} \\ D(\mathit{QCoh}(\mathcal{O}_ W)) \ar[r]^{i_ W} & D_\mathit{QCoh}(\mathcal{O}_ W) }
Then the functor (75.5.1.1)
D(\mathit{QCoh}(\mathcal{O}_ X)) \longrightarrow D_\mathit{QCoh}(\mathcal{O}_ X)
is an equivalence with quasi-inverse given by RQ_ X.
Proof.
We first use the induction principle to prove i_ X is fully faithful. More precisely, we will use Lemma 75.9.6. Let (U \subset W, V \to W) be an elementary distinguished square with V affine and U, W quasi-compact open in X. Assume that i_ U is fully faithful. We have to show that i_ W is fully faithful. We may replace X by W, i.e., we may assume W = X (we do this just to simplify the notation – observe that the condition in the statement of the lemma is preserved under this operation).
Suppose that A, B are objects of D(\mathit{QCoh}(\mathcal{O}_ X)). We want to show that
\mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, B) \longrightarrow \mathop{\mathrm{Hom}}\nolimits _{D(\mathcal{O}_ X)}(i_ X(A), i_ X(B))
is bijective. Let T = |X| \setminus |U|.
Assume first i_ X(B) is supported on T. In this case the map
i_ X(B) \to Rj_{V, *}(i_ X(B)|_ V) = Rj_{V, *}(i_ V(B|_ V))
is a quasi-isomorphism (Lemma 75.10.7). By assumption we have an isomorphism i_ X(\Phi (B|_ V)) \to Rj_{V, *}(i_ V(B|_ V)) in D(\mathcal{O}_ X). Moreover, \Phi and {-}|_ V are adjoint functors on the derived categories of quasi-coherent modules (by Derived Categories, Lemma 13.30.3). The adjunction map B \to \Phi (B|_ V) becomes an isomorphism after applying i_ X, whence is an isomorphism in D(\mathit{QCoh}(\mathcal{O}_ X)). Hence
\begin{align*} \mathop{\mathrm{Mor}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, B) & = \mathop{\mathrm{Mor}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, \Phi (B|_ V)) \\ & = \mathop{\mathrm{Mor}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ V))}(A|_ V, B|_ V) \\ & = \mathop{\mathrm{Mor}}\nolimits _{D(\mathcal{O}_ V)}(i_ V(A|_ V), i_ V(B|_ V)) \\ & = \mathop{\mathrm{Mor}}\nolimits _{D(\mathcal{O}_ X)}(i_ X(A), Rj_{V, *}(i_ V(B|_ V))) \\ & = \mathop{\mathrm{Mor}}\nolimits _{D(\mathcal{O}_ X)}(i_ X(A), i_ X(B)) \end{align*}
as desired. Here we have used that i_ V is fully faithful (Lemma 75.11.3).
In general, choose any complex \mathcal{B}^\bullet of quasi-coherent \mathcal{O}_ X-modules representing B. Next, choose any quasi-isomorphism s : \mathcal{B}^\bullet |_ U \to \mathcal{C}^\bullet of complexes of quasi-coherent modules on U. As j_ U : U \to X is quasi-compact and quasi-separated the functor j_{U, *} transforms quasi-coherent modules into quasi-coherent modules (Morphisms of Spaces, Lemma 67.11.2). Thus there is a canonical map \mathcal{B}^\bullet \to j_{U, *}(\mathcal{B}^\bullet |_ U) \to j_{U, *}\mathcal{C}^\bullet of complexes of quasi-coherent modules on X. Set B'' = j_{U, *}\mathcal{C}^\bullet in D(\mathit{QCoh}(\mathcal{O}_ X)) and choose a distinguished triangle
B \to B'' \to B' \to B[1]
in D(\mathit{QCoh}(\mathcal{O}_ X)). Since the first arrow of the triangle restricts to an isomorphism over U we see that B' is supported on T. Hence in the diagram
\xymatrix{ \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, B'[-1]) \ar[r] \ar[d] & \mathop{\mathrm{Hom}}\nolimits _{D(\mathcal{O}_ X)}(i_ X(A), i_ X(B')[-1]) \ar[d] \\ \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, B) \ar[r] \ar[d] & \mathop{\mathrm{Hom}}\nolimits _{D(\mathcal{O}_ X)}(i_ X(A), i_ X(B)) \ar[d] \\ \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, B'') \ar[r] \ar[d] & \mathop{\mathrm{Hom}}\nolimits _{D(\mathcal{O}_ X)}(i_ X(A), i_ X(B'')) \ar[d] \\ \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, B') \ar[r] & \mathop{\mathrm{Hom}}\nolimits _{D(\mathcal{O}_ X)}(i_ X(A), i_ X(B')) }
we have exact columns and the top and bottom horizontal arrows are bijective. Finally, choose a complex \mathcal{A}^\bullet of quasi-coherent modules representing A.
Let \alpha : i_ X(A) \to i_ X(B) be a morphism between in D(\mathcal{O}_ X). The restriction \alpha |_ U comes from a morphism in D(\mathit{QCoh}(\mathcal{O}_ U)) as i_ U is fully faithful. Hence there exists a choice of s : \mathcal{B}^\bullet |_ U \to \mathcal{C}^\bullet as above such that \alpha |_ U is represented by an actual map of complexes \mathcal{A}^\bullet |_ U \to \mathcal{C}^\bullet . This corresponds to a map of complexes \mathcal{A} \to j_{U, *}\mathcal{C}^\bullet . In other words, the image of \alpha in \mathop{\mathrm{Hom}}\nolimits _{D(\mathcal{O}_ X)}(i_ X(A), i_ X(B'')) comes from an element of \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, B''). A diagram chase then shows that \alpha comes from a morphism A \to B in D(\mathit{QCoh}(\mathcal{O}_ X)). Finally, suppose that a : A \to B is a morphism of D(\mathit{QCoh}(\mathcal{O}_ X)) which becomes zero in D(\mathcal{O}_ X). After choosing \mathcal{B}^\bullet suitably, we may assume a is represented by a morphism of complexes a^\bullet : \mathcal{A}^\bullet \to \mathcal{B}^\bullet . Since i_ U is fully faithul the restriction a^\bullet |_ U is zero in D(\mathit{QCoh}(\mathcal{O}_ U)). Thus we can choose s such that s \circ a^\bullet |_ U : \mathcal{A}^\bullet |_ U \to \mathcal{C}^\bullet is homotopic to zero. Applying the functor j_{U, *} we conclude that \mathcal{A}^\bullet \to j_{U, *}\mathcal{C}^\bullet is homotopic to zero. Thus a maps to zero in \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, B''). Thus we may assume that a is the image of an element of b \in \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, B'[-1]). The image of b in \mathop{\mathrm{Hom}}\nolimits _{D(\mathcal{O}_ X)}(i_ X(A), i_ X(B')[-1]) comes from a \gamma \in \mathop{\mathrm{Hom}}\nolimits _{D(\mathcal{O}_ X)}(A, B''[-1]) (as a maps to zero in the group on the right). Since we've seen above the horizontal arrows are surjective, we see that \gamma comes from a c in \mathop{\mathrm{Hom}}\nolimits _{D(\mathit{QCoh}(\mathcal{O}_ X))}(A, B''[-1]) which implies a = 0 as desired.
At this point we know that i_ X is fully faithful for our original X. Since RQ_ X is its right adjoint, we see that RQ_ X \circ i_ X = \text{id} (Categories, Lemma 4.24.4). To finish the proof we show that for any E in D_\mathit{QCoh}(\mathcal{O}_ X) the map i_ X(RQ_ X(E)) \to E is an isomorphism. Choose a distinguished triangle
i_ X(RQ_ X(E)) \to E \to E' \to i_ X(RQ_ X(E))[1]
in D_\mathit{QCoh}(\mathcal{O}_ X). A formal argument using the above shows that i_ X(RQ_ X(E')) = 0. Thus it suffices to prove that for E \in D_\mathit{QCoh}(\mathcal{O}_ X) the condition i_ X(RQ_ X(E)) = 0 implies that E = 0. Consider an étale morphism j : V \to X with V affine. By Lemmas 75.11.3 and 75.11.2 and our assumption we have
Rj_*(E|_ V) = Rj_*(i_ V(RQ_ V(E|_ V))) = i_ X(\Phi (RQ_ V(E|_ V))) = i_ X(RQ_ X(Rj_*(E|_ V)))
Choose a distinguished triangle
E \to Rj_*(E|_ V) \to E' \to E[1]
Apply RQ_ X to get a distinguished triangle
0 \to RQ_ X(Rj_*(E|_ V)) \to RQ_ X(E') \to 0[1]
in other words the map in the middle is an isomorphism. Combined with the string of equalities above we find that our first distinguished triangle becomes a distinguished triangle
E \to i_ X(RQ_ X(E')) \to E' \to E[1]
where the middle morphism is the adjunction map. However, the composition E \to E' is zero, hence E \to i_ X(RQ_ X(E')) is zero by adjunction! Since this morphism is isomorphic to the morphism E \to Rj_*(E|_ V) adjoint to \text{id} : E|_ V \to E|_ V we conclude that E|_ V is zero. Since this holds for all affine V étale over X we conclude E is zero as desired.
\square
Proposition 75.11.5. Let S be a scheme. Let X be a quasi-compact algebraic space over S with affine diagonal over \mathbf{Z} (as in Properties of Spaces, Definition 66.3.1). Then the functor (75.5.1.1)
D(\mathit{QCoh}(\mathcal{O}_ X)) \longrightarrow D_\mathit{QCoh}(\mathcal{O}_ X)
is an equivalence with quasi-inverse given by RQ_ X.
Proof.
Let V \to W be an étale morphism with V affine and W a quasi-compact open subspace of X. Then the morphism V \to W is affine as W has affine diagonal over \mathbf{Z} and V is affine (Morphisms of Spaces, Lemma 67.20.11). Lemma 75.11.1 then guarantees that the assumption of Lemma 75.11.4 holds. Hence we conclude.
\square
Lemma 75.11.6. Let S be a scheme and let f : X \to Y be a morphism of algebraic spaces over S. Assume X and Y are quasi-compact and have affine diagonal over \mathbf{Z} (as in Properties of Spaces, Definition 66.3.1). Then, denoting
\Phi : D(\mathit{QCoh}(\mathcal{O}_ X)) \to D(\mathit{QCoh}(\mathcal{O}_ Y))
the right derived functor of f_* : \mathit{QCoh}(\mathcal{O}_ X) \to \mathit{QCoh}(\mathcal{O}_ Y) the diagram
\xymatrix{ D(\mathit{QCoh}(\mathcal{O}_ X)) \ar[d]_\Phi \ar[r] & D_\mathit{QCoh}(\mathcal{O}_ X) \ar[d]^{Rf_*} \\ D(\mathit{QCoh}(\mathcal{O}_ Y)) \ar[r] & D_\mathit{QCoh}(\mathcal{O}_ Y) }
is commutative.
Proof.
Observe that the horizontal arrows in the diagram are equivalences of categories by Proposition 75.11.5. Hence we can identify these categories (and similarly for other quasi-compact algebraic spaces with affine diagonal) and then the statement of the lemma is that the canonical map \Phi (K) \to Rf_*(K) is an isomorphism for all K in D(\mathit{QCoh}(\mathcal{O}_ X)). Note that if K_1 \to K_2 \to K_3 \to K_1[1] is a distinguished triangle in D(\mathit{QCoh}(\mathcal{O}_ X)) and the statement is true for two-out-of-three, then it is true for the third.
Let \mathcal{B} \subset \mathop{\mathrm{Ob}}\nolimits (X_{spaces, {\acute{e}tale}}) be the set of objects which are quasi-compact and have affine diagonal. For U \in \mathcal{B} and any morphism g : U \to Z where Z is a quasi-compact algebraic space over S with affine diagonal, denote
\Phi _ g : D(\mathit{QCoh}(\mathcal{O}_ U)) \to D(\mathit{QCoh}(\mathcal{O}_ Z))
the derived extension of g_*. Let P(U) = “for any K in D(\mathit{QCoh}(\mathcal{O}_ U)) and any g : U \to Z as above the map \Phi _ g(K) \to Rg_*K is an isomorphism”. By Remark 75.9.5 conditions (1), (2), and (3)(a) of Lemma 75.9.4 hold and we are left with proving (3)(b) and (4).
Checking condition (3)(b). Let U be an affine scheme étale over X. Let g : U \to Z be as above. Since the diagonal of Z is affine the morphism g : U \to Z is affine (Morphisms of Spaces, Lemma 67.20.11). Hence P(U) holds by Lemma 75.11.1.
Checking condition (4). Let (U \subset W, V \to W) be an elementary distinguished square in X_{spaces, {\acute{e}tale}} with U, W, V in \mathcal{B} and V affine. Assume that P holds for U, V, and U \times _ W V. We have to show that P holds for W. Let g : W \to Z be a morphism to a quasi-compact algebraic space with affine diagonal. Let K be an object of D(\mathit{QCoh}(\mathcal{O}_ W)). Consider the distinguished triangle
K \to Rj_{U, *}K|_ U \oplus Rj_{V, *}K|_ V \to Rj_{U \times _ W V, *}K|_{U \times _ W V} \to K[1]
in D(\mathcal{O}_ W). By the two-out-of-three property mentioned above, it suffices to show that \Phi _ g(Rj_{U, *}K|_ U) \to Rg_*(Rj_{U, *}K|_ U) is an isomorphism and similarly for V and U \times _ W V. This is discussed in the next paragraph.
Let j : U \to W be a morphism X_{spaces, {\acute{e}tale}} with U, W in \mathcal{B} and P holds for U. Let g : W \to Z be a morphism to a quasi-compact algebraic space with affine diagonal. To finish the proof we have to show that \Phi _ g(Rj_*K) \to Rg_*(Rj_*K) is an isomorphism for any K in D(\mathit{QCoh}(\mathcal{O}_ U)). Let \mathcal{I}^\bullet be a K-injective complex in \mathit{QCoh}(\mathcal{O}_ U) representing K. From P(U) applied to j we see that j_*\mathcal{I}^\bullet represents Rj_*K. Since j_* : \mathit{QCoh}(\mathcal{O}_ U) \to \mathit{QCoh}(\mathcal{O}_ X) has an exact left adjoint j^* : \mathit{QCoh}(\mathcal{O}_ X) \to \mathit{QCoh}(\mathcal{O}_ U) we see that j_*\mathcal{I}^\bullet is a K-injective complex in \mathit{QCoh}(\mathcal{O}_ W), see Derived Categories, Lemma 13.31.9. Hence \Phi _ g(Rj_*K) is represented by g_*j_*\mathcal{I}^\bullet = (g \circ j)_*\mathcal{I}^\bullet . By P(U) applied to g \circ j we see that this represents R_{g \circ j, *}(K) = Rg_*(Rj_*K). This finishes the proof.
\square
Comments (0)