Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You and Your Team.

Learn More →

On the Kodaira vanishing theorem for log del Pezzo surfaces in positive characteristic

On the Kodaira vanishing theorem for log del Pezzo surfaces in positive characteristic We investigate the vanishing of H (X , O (−D)) for a big and nef Q-Cartier Z-divisor D on a log del Pezzo pair (X,) over a perfect field of positive characteristic p. Keywords Kodaira vanishing · Log del pezzo surfaces · Positive characteristic Mathematics Subject Classification 14E30 · 14F17 · 14J45 · 13A35 Introduction It has long been known that the (Kodaira and Kawamata–Viehweg) vanishing theorems, so fundamental to birational geometry in characteristic zero, in general fail for surfaces in positive characteristic [20]. Several people have investigated different classes of surfaces over perfect fields of positive characteristic for which the (Kodaira and Kawamata–Viehweg) vanishing theorems may or may not hold. The question is subtle; for example, every smooth rational surface over a perfect field satisfies Kodaira vanishing, however over any alge- braically closed field of positive characteristic, there are smooth rational surfaces that violate Kawamata–Viehweg vanishing [4]. On the other hand a smooth surface with ample anti- canonical bundle satisfies Kawamata–Viehweg vanishing theorem over any algebraically closed field [4, Proposition A.1]. This however is no longer true if (klt) singularities are allowed and counterexamples (even to Kodaira vanishing) have been constructed over alge- braically closed fields of characteristic 2 [5]and 3[2]. The situation simplifies in large characteristics. In [6] the authors prove the existence of an integer p such that over an algebraically closed field of characteristic p > p every 0 0 log del Pezzo surface satisfies Kawamata–Viehweg vanishing theorem. Finding an effective bound for this p is a central open question in positive characteristic birational geometry. For example over an algebraically closed field of characteristic p > max{5, p } athree The author was supported by SNF Grant #200021/169639. B Emelie Arvidsson emelie.arvidsson@epfl.ch EPFL SB MATH CAG, MA C3 615 (Batiment MA), Station 8, 1015 Lausanne, Switzerland 123 E. Arvidsson dimensional klt singularity is rational [12, Theorem 1.1], and threefolds satisfy a refined version of the basepoint free theorem [3, Theorem 1.1]. The construction of the integer p in [6] is implicit. During the course of the proof of [6, Theorem 1.1] the authors consider log del Pezzo surfaces belonging to a bounded family over Spec(Z). The construction of the integer p does in particular depend on this family in an implicit way. It is therefore natural to ask if it is possible to prove vanishing theorems in large explicit characteristics depending on some explicit numerical invariant of a bounded family of log del Pezzo surfaces. In this direction, we show that Kodaira vanishing (for big and nef divisors) holds on a del Pezzo surface of bounded index I in characteristic p > p (I ) where p (I ) is an explicit polynomial in the index I . We also determine an explicit bound for p 0 0 (for Kodaira vanishing) in terms of > 0for an -klt log del Pezzo surface. Our main results in this direction are the following. Theorem A Let X be a projective klt surface over a perfect field k of characteristic p > 0 with ample anti-canonical divisor −K of Cartier index I . Let D be a big and nef Q-Cartier Z-divisor on X. (1) If I = 1, i.e., if −K is an ample Cartier divisor then: • if p ≥ 9221,then H (X , O (−D)) = 0 • if p ≥ 5,then H (X , O (D)) = 0. 2 3 2 (2) If I ≥ 2 and p ≥ (13 − 45I ) (2I + 4I + 2I ) + 1, then: 1 1 H (X , O (D)) = H (X , O (−D)) = 0. X X −1/2 Theorem B Let 0 < < 3 be a real number. Let (X,) be a projective -klt surface with −(K + ) ample over a perfect field k of characteristic (128) 128 25 p ≥ 2 + 1. Then Kodaira vanishing holds on X, i.e., for all ample Q-Cartier Z-divisors D on X we have that H (X , O (D + K )) = 0, for all i > 0. X X In this note, we use a technique due to Ekedahl [7]. When trying to prove Kodaira vanishing in explicit characteristics using this technique, the difficulties arise from Weil divisors which are far from being Cartier. For example, on a log del Pezzo surface over an algebraically closed field of characteristic p ≥ 4c + 1 we can prove that H (X , O (K + A)) = 0 X X for every ample Q-Cartier Weil divisor A of Cartier index ≤ c (see Remark 3.5). Another strategy is therefore to control the Cartier index of a Weil divisor D in terms of the Cartier index I of (K + ). We prove, in this way, that a big and nef Z-divisors D of big volume relative to the index I in characteristic p ≥ 5 satisfies H (X , O (−D)) = 0. Theorem C Let (X,) be a projective klt surface over a perfect field k of characteristic p ≥ 5 with −(K + ) an ample Q-Cartier Q-divisor of Cartier index I . Let D be a big and nef Q-Cartier Z-divisor. Then the following holds true: 2 1 (1) If I = 1 and D ≥ 9,then H (X , O (−D)) = 0. 2 3 2 1 (2) If I ≥ 2 and D ≥ 2I + 4I + 2I, then H (X , O (−D)) = 0. During the preparation of this note we were informed that J. Lacini has classified all log del Pezzo surfaces of Picard rank one over an algebraically closed field of characteristic 123 On the Kodaira vanishing theorem for log... p ≥ 5[17, Theorem 1.1]. From this classification it follows that all log del Pezzo surfaces of Picard rank one over an algebraically closed field of characteristic p > 5 admit a log resolution that lifts to W (k) [17, Theorem 7.2]. It therefore follows from his work that Kodaira vanishing holds for log del Pezzo surfaces of Picard rank one over an algebraically closed field of characteristic p > 5(see[6, Lemma 6.1]). Combining his result with the techniques exploited in this paper, one can prove the Kodaira vanishing theorem on a log del Pezzo surface over an algebraically closed field of characteristic p > 5. At the end of this note we illustrate this argument. Building on the liftability of rank one log del Pezzo surfaces in characteristic p > 5 by Lacini we can prove the following theorem. Theorem D Let (X,) be a projective log del Pezzo surface over a perfect field of charac- teristic p > 5. Then Kodaira vanishing holds on X, i.e., for all ample Z-divisors D on X we have H (X , O (D + K )) = 0, for all i > 0. X X After this note was made public, the author has, in joint work with Fabio Bernasconi and Justin Lacini, extended the above result. In [1] we show that log del Pezzo surfaces over a perfect field of characteristic p > 5 satisfy the Kawamata Viehweg vanishing theorem. In [1] we also construct a del Pezzo surface over an algebraically closed field of characteristic p = 5 on which the Kodaira vanishing theorem fails. This implies that the characteristic p > 5inTheorem D is optimal. 1 Preliminaries By a variety we mean a finite type integral separated scheme over a field. We will work exclusively over a perfect field k. A Q-divisor D is said to be Q-Cartier if there exists an integer m such that mD is Cartier. A Q-divisor is ample/nef/big if it is Q-Cartier and an integer multiple is ample/nef/big as a line bundle. If X is a normal variety, then the reflexive sheaves on X are determined (up to isomorphism) by their restriction to a big open subset U (an open subset U of X is big if codim (X −U ) ≥ 2). If i : U → X is the inclusion of a big open subset and M is a reflexive sheaf on U then i M is a reflexive sheaf on X, moreover for any reflexive sheaf G on X we have an isomorphism G = i G [10, Proposition 1.6]. Let U denote the smooth locus of X.Since X ∗ | is normal U is big. Therefore, for any Weil divisor D on X the restriction of D to the smooth locus defines a reflexive sheaf i O (U ∩ D) of rank one on X. Conversely, any reflexive ∗ U sheaf of rank one F on X defines a Weil divisor D by setting D to be the closure of a Weil divisor E on U satisfying F = O (E ). This defines a one to one correspondence between | U the reflexive sheaves of rank one on X up to isomorphism and the set of Weil divisors on X up to rational equivalence [10]. We denote by O (D) the reflexive sheaf corresponding to a divisor D on X. We use the notation K for the class corresponding to the closure of a canonical divisor K on U in X, hence O (K ) i O (K ). If X is normal and projective U X X ∗ U U then O (K ) can be seen to be a dualizing sheaf for X [15, Proposition 5.75]. We use the X X notation ω for a dualizing sheaf on X. In general if D and D are Weil divisors on a normal variety X,then O (D ) ⊗ 1 2 X 1 O (D ) is not reflexive. However, the dual M of a coherent sheaf M is always reflex- X 2 ive [10, Corollary 1.2] and we may define a product structure on the set of reflexive ∨∨ sheaves by defining O (D ) ⊗ O (D ) := (O (D ) ⊗ O (D )) .Wehavethe fol- X 1 X 2 X 1 X 2 lowing identities for Weil divisors D , D and a Cartier divisor C: O (D + D ) = 1 2 X 1 2 123 E. Arvidsson ∨∨ ∨ ∼ ∼ (O (D ) ⊗ O (D )) , O (−D ) = O (D ) , O (D + C ) = O (D ) ⊗ O (C ) X 1 X 2 X 1 X 1 X 1 X 1 X and Hom (O (D ), O (D )) O (D − D ). This product structure makes the set of O X 1 X 2 X 2 1 reflexive sheaves on X (up to isomorphism) into a group such that the correspondence between Weil divisors and reflexive sheaves on X, as described above, becomes an isomorphism of groups. By a log pair (X,) we shall mean a normal variety X together with an effective Q-divisor on X such that K +  is Q-Cartier. By a klt (respectively lc) pair we shall mean a log pair with klt (respectively lc) singularities in the sense of [15, Section 2]. By [23, Corollary 4.11] if X is a surface and (X,) is a klt pair then X is Q-factorial. A normal surface X is said to be a del Pezzo surface if it has klt singularities and −K is Q-Cartier and ample. A log pair (X , B) is said to be a log del Pezzo surface if it has klt singularities and −(K + B) is ample. 1.1 Serre vanishing and Serre duality for Z-divisors One of the most fundamental vanishing theorems in algebraic geometry is that of Serre. It naturally implies a Weil divisor version. Theorem 1.1 Let X be a normal projective variety. If D is an ample Q-Cartier Z-divisor and F is a coherent sheaf on X then H (X , F ⊗ O (lD)) = 0 for all l big enough and all i > 0. Proof Let m be an integer such that mD is Cartier. By Serre vanishing there exists, for all j ⊗n 1 ≤ i ≤ m,an n such that H (X , F ⊗ O (iD) ⊗ O (mD) ) = 0, for all n ≥ n i X X i ⊗n and all j > 0. Since O (mD) is Cartier O (iD) ⊗ O (mD) = O ((i + mn)D).Set X X X X N = max n .For all l ≥ (m + 1)N , we can write l = i +mk,for 1 ≤ i ≤ m − 1, where 1≤i ≤m i i i ⊗k k ≥ N . Therefore, we have H (X , F ⊗O (lD)) = H (X , F ⊗O (iD)⊗O (mD) ) = 0, X X X for all l ≥ (m + 1)N . Corollary 1.2 Let D be a big and nef integral divisor on a normal projective surface X, then H (X , O(−mD)) = 0 for all m >> 0. Proof Let r be the Cartier index of D.Let n > 0 be such that H (X , O(lD)) is non zero for all l ≥ n, this is possible since D is big. By Fujita vanishing [9, Theorem 10] there exists an integer m such that H (X , O(−(kr + l)D)) = 0for all l ≥ n and all k ≥ m, this proves the Corollary. We will repeatedly need to use a form of Serre Duality valid for reflexive sheaves on a normal surface. Theorem 1.3 (Serre Duality for CM-sheaves) Let X be a projective scheme of pure dimension n over a field k. Let F be a CM sheaf on X such that Supp(F ) is of pure dimension n. Then i n−i H (X , F ) is dual to H (X , Hom (F,ω )). O X Proof E.g., [15, Theorem 5.71] In particular, by [10, Proposition 1.3] for any Z-divisor D on a normal projective surface i 2−i X the vector space H (X , O (D)) is dual to H (X , O (K − D)) for i = 0, 1. X X X 123 On the Kodaira vanishing theorem for log... Remark 1.4 If D is a big and nef Z-divisor on a normal projective surface X then 0 2 H (X , O (−D)) = 0 and hence by duality H (X , O (D + K )) = 0. Therefore, Kodaira X X X 1 1 vanishing is equivalent to the vanishing of H (X , O (−D)) = H (X , O (K + D)). X X X 1.2 Frobenius techniques Let X be a scheme over a positive characteristic base S,let F denote the absolute Frobenius on X. The absolute Frobenius F is not a morphism of S-schemes unless F on S is the X S identity. In particular we have the following commuting diagram: X /S X ×SX SS . where F denotes the relative Frobenius. When S = Spec(k) for a perfect field k then X /S the absolute Frobenius on S is an isomorphism and X ×SX is an isomorphism over S if we consider X as a scheme over S by post-composing the structure morphism to S −1 with F on S. We often make this identification, under which the relative and the absolute Frobenius coincide, and denote by F : X → X the corresponding morphism. Let X be a normal variety over an algebraically closed field of positive characteristic. If D is an integral ∗ ∨∨ divisor on X then (F O (D)) = O (pD), since they agree on the regular locus which is X X a big open subset of X. There is a natural map (unit of adjunction) O (D) → F F O (D). X ∗ X This induces a morphism: i i ∗ i ∗ ∨∨ H (X , O (D)) → H (X , F O (D)) → H (X , F O (D) ). X X X Therefore, the Frobenius on X induces morphisms for all m: i m i m+1 H (X , O (−p D)) → H (X , O (−p D)). X X If D is an ample divisor whose associated Weil sheaf is CM, then these groups eventually vanish for m >> 0 by Theorem 1.3 and Corollary 1.1. Therefore, Kodaira vanishing theorem i m for such divisors D hold true on X if and only if the morphisms H (X , O (−p D)) → i m+1 H (X , O (−p D)) are injective for all m and all i < dim(X ). 1.3 Non-vanishing and the associated ˛ -torsor O (D ) X 0 Let X be a projective normal surface and D be a Q-Cartier Z-divisor on X.Let X ⊂ X denote the smooth locus of X and D := D ∩ X be the restriction of D to the smooth 0 0 locus. Let F denote the Frobenius on X . We will assume that H (X , O (−D)) = 0and 0 0 X 1 1 that H (X , O (−pD)) = 0. Let α ∈ H (X , O (−D)) denote a non-trivial element of X X i i the kernel of the map H (X , O (−D)) → H (X , O (−pD)) described in Sect. 1.2.Since X X Z := X − X is the complement of a big open subset of X and O (nD) is reflexive (and so 0 X 0 0 in particular, H (X , O (nD)) H (X , O (nD ))), we see from the local cohomology X 0 X 0 0 long exact sequence that we have an inclusion: 1 1 H (X , O (−nD)) → H (X , O (−nD )) X 0 X 0 123 E. Arvidsson for all n. By our assumptions, α therefore defines a non-trivial element: α ∈ H (X , O (−D )), 0 0 X 0 which belongs to the kernel of the morphism: 1 1 H (X , O (−D )) → H (X , O (−pD )) 0 X 0 0 X 0 0 0 induced by F . This kernel has a geometric description on X which is described in the next 0 0 section. 1.3.1 ˛ -torsors An invertible sheaf L on a scheme X naturally defines a sheaf of groups under addition −i representable by the affine group scheme L := Spec ( L ). Let X be defined over a field X i k of characteristic p > 0. Then there exists a homomorphism of sheaves of additive groups L → L , defined by raising a local section to its p-power. This is a purely characteristic p p p p > 0 phenomenon, since for any two local sections s and t of L we have (s + t ) = s + t in characteristic p. Relative to the Zariski site of X the above morphism of sheaves is not in general a surjection. However, as a morphism of X -sheaves it is, i.e., the corresponding morphism fppf of sheaves of groups on the flat site of X is surjective [18, II.2.18]. The kernel of this morphism is a sheaf of groups on X which we denote by α . By construction α is representable fppf L L p−1 −i by an affine group scheme over X, denoted α . In fact,wehave α = Spec ( L ). L L i =0 The short exact sequence of sheaves of groups (relative to the flat- topology on X) 0 α LL 0 induces a long exact sequence on cohomology [18,4,Prop4.5]: 0 p 1 1 1 p ··· → H (X , L ) → H (X,α ) → H (X , L) → H (X , L ) → ... fppf 1 1 p Therefore, a non-trivial element α in the kernel of H (X , L) → H (X , L ) defines a 1 1 ˇ ˇ non-trivial element α ∈ H (X,α ). The group H (X,α ) has a geometric meaning, L L fppf fppf i.e., a non-trivial element α ∈ H (X,α ) corresponds to a non-trivial α -torsors Y → X, L L fppf see for example, [18, Proposition 4.6]. The torsor Y corresponding to α ∈ H (X,α ) fppf constructed as above were studied by Ekedahl. Proposition 1.5 ([7, pp. 106–107], [19, Theorem 2.11]) Let X be a normal variety over an algebraically closed field k of characteristic p > 0.Let L ∈ Pic(X ). A non-trivial element of 1 1 p the kernel under the Frobenius action: H (X , L) → H (X , L ), gives rise to a non-trivial −1 α -torsor β : Y → X. Locally over U = Spec(A), the morphism β U → U is given by A → A[x ]/(x − a) for some element a ∈ A which is not a p-power. Therefore, β is purely inseparable of degree p. If X is a G and S variety, then Y is also a G and S variety which 1 2 1 2 ∗ p−1 satisfies: ω = β (ω ⊗ L ). Y X Here S denotes Serre’s condition and G denotes Gorenstein in codimension one. 2 1 Remark 1.6 Y is in general not normal. 123 On the Kodaira vanishing theorem for log... 1.4 Geometric construction for Weil-divisors Let X be a projective normal surface over an algebraically closed field of charcteristic p > 0and let D a Q-Cartier ample Weil divisor on X such that H (X , O (−D)) = 0and H (X , O (−pD)) = 0. Let X denote the smooth locus of X and let D = D ∩ X . X 0 0 0 Let α be a non-trivial element of H (X,α ) coming from a non-trivial element 0 O (D ) fppf X 0 α ∈ H (X , O (D)) as described in point Sect. 1.3.Let π : Y → X be the corresponding X 0 0 0 non-trivial α -torsor. Then π O is a locally free O -algebra. Let i : X → X O (D ) 0∗ Y X 0 X 0 0 0 denote the inclusion of the regular locus. The multiplication on X extends (because i π O 0 ∗ 0∗ Y is reflexive) to make i π O into a sheaf of O -algebras. We may therefore define Y := ∗ 0∗ Y X Spec (i π O ).Let π : Y → X denote the natural morphism. ∗ 0∗ Y X 0 Lemma 1.7 In the situation above π : Y → X is a finite degree p morphism and Y is a projective S and G surface that satisfies 2 1 K = π (K − (p − 1)D). Y X Proof Since π O = i π O is S we have by [15, Proposition 5.4] that Y is S . Since X ∗ Y ∗ 0∗ Y 2 2 and X agree in codimension one, Y is G . Since everything is S and G we may replace 0 1 2 1 ∗ ∗ X with X by [11, Theorem 1.12], therefore π (K ) and π (D) are well-defined and the 0 X formula K = π (K − (p − 1)D) follows from Proposition 1.5. Y X Lemma 1.8 With the notation as above, let η : Y → Y be the normalization of Y and let π : Y → X denotes the induced morphism. There exists an effective Z-divisor  ≥ 0 on Y such that K +  = π (K − (p − 1)D). Proof Since everything is S and G we may replace Y with a big open set and assume that Y 2 1 ∼ ∼ is Gorenstein and Y is regular. Affine locally we may assume that ω = O and ω = O . Y Y Y Y In this situation [19, Lemma 2.14] proves that the natural map ω → η ω has image equal to the conductor ideal, where η ω is identified with O . 1.5 Bend and break By studying different properties of the variety Y appearing in Lemma 1.8 one hopes to arrive at a contradiction to the assumed non-vanishing, H (X , O (−D)) = 0. This strategy was successfully employed by Ekedahl [7] and more recently in [19]. The main tool is the use of bend and break, together with the expression for the canonical divisor given by Lemma 1.8. Together, this can be used in order to derive inequalities comparing intersection numbers on X with the characteristic p of the base field. We will repeatedly use the following: Theorem 1.9 [16, Chapter II, Theorem 5.8] Let X be a projective variety over an alge- braically closed field, C a smooth, projective and irreducible curve, f : C → X a morphism and M any nef R-divisor. Assume that X is smooth along f (C ) and K · C > 0. Then for every x ∈ f (C ) there is a rational curve ⊂ X containing x such that M · C M · ≤ 2dim(X ) . −K · C 123 E. Arvidsson 2 Kodaira vanishing for divisors of big volume We will use the following terminology: We will say that a curve C on a surface X is big and basepoint free if the linear system |C | defines a birational morphism. If C is big and basepoint free, then for any finite number of points in X there exists C ∈|C | such that C avoids all those points. Lemma 2.1 Let X be a projective normal surface. Suppose that there exists an effective Q- divisor E such that −(K + E ) is a Q-Cartier ample divisor of Cartier index I . If D is a 2 2 2 big Q-Cartier semiample Z-divisor such that D ≥ I (K + E ) then there exists an ample Cartier divisor A and a big basepoint free curve C on X such that D · C ≥ A · C. Proof Suppose, in order to arrive at a contradiction, that for every big basepoint free curve C and every ample Cartier divisor A of X we have D · C < A · C.For m divisible enough, −m(K + E ) is very ample. Therefore, for A =−I (K + E ) and C ∈| − m(K + E )| X X X we find −(K + E ) · D < I (K + E ) . Similarly, for n divisible enough, nD is basepoint X X free and we may consider a curve C ∈|nD| to find that D < −I (K + E ) · D. Putting this 2 2 2 together, we find that D < I (K + E ) . Theorem 2.2 Let X be a projective normal surface over an algebraically closed field k of characteristic p ≥ 5. Suppose that there exists an effective Q-divisor E such that −(K + E ) is a Q-Cartier ample divisor of Cartier index I . If D is a Q-Cartier nef and big semiample 2 2 2 1 Z-divisor such that D ≥ I (K + E ) then H (X , O (−D)) = 0. X X Proof Suppose that the theorem is not true. Let D be a nef and big semiample Z-divisor 2 2 2 1 on X such D ≥ I (K + E ) and assume that H (X , O (−D)) = 0. By Corollary X X 1 j 1.2 there exists some j ≥ 1suchthat H (X , O (−p D)) = 0. If we replace D with j −1 j −1 p D the inequality will be satisfied for p D. Suppose we can prove the theorem for j −1 1 j −1 p D then H (X , O (−p D)) = 0 and, hence, by descending induction on j,wehave proven that H (X , O (−D)) = 0. Therefore, there is no loss of generality to assume that H (X , O (−pD)) = 0. Let π : Y → X beadegree p-cover corresponding to a non-zero element y ∈ H (X , O (−D)) as in Sect. 1.3. By replacing Y with its normalisation, we may assume that Y is normal and that ∗ ∗ K = π K + (1 − p)π D −  (2.1) Y X where  is effective (Lemma 1.8). By Lemma 2.1 there exists a big basepoint free curve C on X and an ample Cartier divisor A such that D · C ≥ A · C.Let π : Y → X be as above. By replacing C with a large enough multiple we see that there exists a curve C on Y such that: • C is contained in the smooth locus of Y • C is not contained in a component of • C = π C for some curve C on X not contained in a component of E andsuchthat ∗ Y D · C ≥ A · C . By Eq. (2.1)wehave −K · C =−K · C + (p − 1)D · C +  · C ≥−K · C,where Y Y X Y X the last inequality comes from the assumption that D is nef and that C is not contained in a component of . By assumption −(K + E ) is ample, hence −K · C > E · C ≥ 0. X X Therefore −K · C > 0. By Theorem 1.9 for C and π A on Y , there exists a rational Y Y Y 123 On the Kodaira vanishing theorem for log... curve passing through a point x ∈ C such that: x Y π A · C π A · ≤ 4 . −K · C Y Y By Equation (2.1), we find: ∗ ∗ ∗ ∗ π A · (−π K + (p − 1)π D + ) · C ≤ 4π A · C . x X Y Y However, since A is an ample Cartier divisor and π is finite π A · ≥ 1. Moreover, ∗ ∗ ∗ −π K · C =−K · C > 0and C ·  ≥ 0. By assumption, π A · C ≤ π D · C , X Y X Y Y Y therefore: ∗ ∗ ∗ ∗ (p − 1)π A · C < π A · (−π K + (p − 1)π D + ) · C . Y x X Y Putting these inequalities together we find: ∗ ∗ (p − 1)π A · C < 4π A · C Y Y and, therefore, p < 5. Theorem 2.2 implies, together with results of [13], our main theorem for divisors of big volume. Proof of Theorem C By base change to an algebraic closure of k we can assume that k is algebraically closed. By the Basepoint free theorem [22, Theorem 1.3], D is semiample. Note that a klt surface of index I,necessarily,is -klt. By [13, Theorem 1.3], we therefore 2 2 have that (K + ) ≤ max{9, 2I + 4 + }. The result therefore follows from Theorem 2.2. Remark 2.3 Recently Kawakami improved both Theorems A and C in the case when I = 1. In particular, let X be a del Pezzo surface with canonical singularities over an algebraically closed field of positive characteristic p and let D be a big and nef Q-divisor. In [14]itis 1 1 shown that H (X , O (−D)) = 0if p = 2[14, Theorem 1.1] and that H (X , O (D)) = 0 X X in all characteristic [14, Proposition 4.10]. See [14] for the precise statements which are slightly more general. Corollary 2.4 Let X be a klt del Pezzo surface over an algebraically closed field of charac- teristic p ≥ 5, such that −K is ample of Cartier index I . Let D be a big and nef Z-divisor of Cartier index c. Then H (X , O (D)) = 0 if either of the following holds: 2 2 2 • K ≥ (c − 1) D and D is ample 2 2 2 • D ≥ (I − 1) K . Proof The assumptions imply that D − K is ample and H (X , O (K − D)) is dual to X X X H (X , O (D)). Assume that the claimed vanishing does not hold. Since the characteristic of the base field is greater than or equal to five, we may argue as in the proof of Proposition 2.2 to assume that for every big basepoint free curve C and every ample Cartier divisor A we have (D − K ) · C < A · C . The first inequality comes from setting A = cD and C =−mK and X X C = lD respectively, for large and divisible enough l and m. The second inequality comes from setting A =−IK and C =−mK and C = lD respectively, for large and divisible X X enough l and m. 123 E. Arvidsson 3 Kodaira vanishing in large characteristic In this section we prove Kodaira vanishing in large characteristic for log del Pezzo surfaces of bounded index. That bounded families of log del Pezzo surfaces over Spec(Z) satisfy Kodaira vanishing in large characteristic (depending only on the family) has been proven by other methods in [6]. In Proposition 3.1 below we show that the existence of a very ample divisor A of bounded volume r on a normal projective surface X over a field of characteristic p ≥ 4r + 1 implies Kodaira Vanishing whenever −K · A > 0. In particular, we do not make any further assumption on the singularities of X. Proposition 3.1 Let r > 0 be an integer. Let X be a projective normal surface over an algebraically closed field k of characteristic p ≥ 4r + 1. If there exists a very ample divisor A on X such that A ≤ r and −K · A > 0 then Kodaira vanishing (even for big and nef Q-Cartier Z-divisors) holds on X. Proof Suppose that D is a big and nef Z-divisor and suppose that H (X , O (−D)) = 0. j 1 By replacing D with p D for some j ≥ 1 we may assume that H (X , O (−pD)) = 0 (Corollary 1.2). By Ekedahl’s construction we may assume that there exists a normal variety Y and a purely inseparable degree p morphism π : Y → X such that K = π K − (p − 1)D − E (3.1) Y X for E ≥ 0. Let A be a very ample general curve on X as above. There exists an integer l > 0 such that lπ A is very ample on Y . Therefore there exists a curve A on Y such that: • A is contained in the smooth locus of Y • A is not contained in a component of E • A = π A for some A ∈|nA| where n > 0is an integer. ∗ Y By Eq. (3.1)wehave −K · A = n(−K · A + (p − 1)D · A) + E · A ≥−nK · A. Y Y X Y X By assumption −K · A > 0. Therefore −K · A > 0. By Theorem 1.9 for A and π (A) X Y Y Y on Y , there exists for every point x ∈ A a rational curve passing through x such that Y x π (A) · A π (A) · ≤ 4 . −K · A Y Y However, since A is an ample Cartier divisor and π is finite, π A · ≥ 1. By Eq. (3.1)and the assumption A ≤ r, we therefore have: ∗ ∗ (−π K + (p − 1)π D + E )A ≤ 4rn. X Y Since −K · A > 0and E · A ≥ 0, we find that: X Y (p − 1)π D · A < 4rn. Since A ∈|nA| we have π D · A ≥ n. The result follows. Remark 3.2 If (X,) is a log pair such that −(K + ) is ample then −K · A > · A ≥ 0 X X for any ample divisor A. Proof of Theorem A By base change to an algebraic closure of k we can assume that k is algebraically closed. By [24] the divisor (13 − 45I )IK is very ample. By [13, Theorem 1.3] there is an inequality K ≤ max{9, 2I + 4 + }. The result therefore follows from Proposition 3.1 and Theorem C. 123 On the Kodaira vanishing theorem for log... For the boundary version of the theorem above we need to use a common bound for the Cartier indices of K and K + . X X Theorem 3.3 Let (X,) be a log del Pezzo surface over an algebraically closed field k of characteristic p > 0.Let m ≥ 2 be such that m(K + ) and m K are both Cartier. Let D X X be a big and nef Z-divisor on X. If 8 7 5 4 3 p ≥ 173056m − 86528m + 16200m + 32400m + 16200m + 1 then H (X , O (−D)) = 0. Proof Let H = 13mK − 45m (K + ).Accordingto[24], H is a very ample Cartier X X 2 5 divisor. By [24, Lemma 6.2], we have inequalities (K +)·K ≥ 0and K ≤ 128m (2m − X X 1).By[13, Theorem 1.3], we have an inequality (K + ) ≤ 2m + 4 + . Therefore: 2 2 7 2 5 4 3 H ≤ (13) 128m (2m − 1) + (45) (2m + 4m + 2m ). We conclude by Proposition 3.1. Remark 3.4 A log del Pezzo surface (X,) where (K +) has Cartier index I is necessarily -klt for  = . As in the proof below, the Cartier index of K can therefore be bounded in terms of the Cartier index of (K + ). Proof of Theorem B By base change to an algebraic closure of k we can assume that k is algebraically closed. By assumption X is -klt, therefore the Q-factorial index q at any point x ∈ X satisfies q ≤ 2 by [24, Prop 6.1]. Moreover, by the same proposition, the Picard rank of a minimal resolution (128) 128 2 25 of X is bounded by . Hence, for every Z-divisor D, there exists some c ≤ 2 , depending on D, such that cD is Cartier. 1 k−1 Suppose that H (X , O (−D)) = 0. By replacing D with p D for some k > 1, we may assume that H (X , O (−pD)) = 0. We may, therefore, assume that there exists a normal variety Y and a degree p inseparable morphism Y → X such that K = π (K + Y X (1 − p)D) − E,for some E ≥ 0. Let C be a general complete intersection curve on Y , not contained in the support of E, such that π C = C is not contained in the support of . ∗ Y Then −K · C > 0. By Theorem 1.9 applied to C and Y Y Y ∗ ∗ π (−(K + ) + (p − 1)D) ≡−(K + E ) − π (), X Y there exists a rational curve through a point on C such that: (−(K + E ) − π ()) · C Y Y π (−(K + ) + (p − 1)D) · ≤ 4 . −K · C Y Y ∗ ∗ Since −π (K + ) · > 0, E · C ≥ 0and π () · C ≥ 0, we find that X Y Y (p − 1)D · < 4. Since cD is Cartier we have an inequality D · ≥ . Therefore: (128) 128 25 p < 4c + 1 ≤ 2 + 1. 123 E. Arvidsson Remark 3.5 From the proof of Theorem B we see that on a log del Pezzo surface (X,) over an algebraically closed field k of characteristic p ≥ 4c + 1wehave H (X , O (−D)) = 0, for every ample Z-divisor D of Cartier index ≤ c. 4 Kodaira vanishing in char p> 5 using the liftability of log del Pezzo surfaces of Picard rank one In this section we deduce the Kodaira vanishing theorem for log del Pezzo surfaces from the liftability in characteristic p > 5 of del Pezzo surfaces of Picard rank one, together with the Kodaira vanishing theorem for surfaces Y which admit a Mori fiber space structure f : Y → B onto a curve. In fact, for a surface admitting a Mori fiber space structure onto a curve the difficulties arising from non-Cartier divisors in our previous arguments are not present. We learned the main ideas in Proposition 4.1 from Zsolt Patakfalvi. Proposition 4.1 Let X be a normal projective surface over an algebraically closed field k of characteristic p > 0 which admits a proper birational morphism g : X → Yto a normal projective surface Y admitting a K -negative fibration to a curve B. That is, there exists a proper morphism f : Y → B such that f O = O and such that the ∗ Y B general fiber is isomorphic to P .Let D bea Q-Cartier big and nef Z-divisor on X. Then H (X , O (−D)) = 0. Proof According to Corollary 1.2 by replacing D with p D, we may assume that 1 1 H (X , O (−pD)) = 0. Suppose that H (X , O (−D)) = 0. We apply Ekedahl’s con- X X struction 1.3 to get a degree p finite morphism π : Z → X,where Z is normal, and ∗ ∗ K + E = π K + (1 − p)π D Z X for an effective divisor E ≥ 0. We illustrate the situation with a diagram: XY Let F be a fiber of f not containing any point which is the image of a curve contracted by g.Then F can be identified with a fiber of f ◦ g : X → B. By abuse of notation we denote a general fiber of f ◦ g by F. A general fiber of the composition f ◦ g avoids the exceptional locus of g since the curves contracted by g map to a finite number of points on Y . We compute: ∗ ∗ ∗ K · π F = π ((K − (p − 1)D) · F ) − E · π F . Z X The morphism π is finite of order p, therefore, we have: ∗ ∗ K · π F = p((K − (p − 1)D) · F − E · π F . Z X Since two general fibers F do not intersect we have F = 0 and therefore K · F =−2. Since D is big and F is general, we have D · F > 0. 123 On the Kodaira vanishing theorem for log... Moreover, since E is effective, we see (upon varying F) that the intersection number E · π F ≥ 0. We conclude that: K · π F ≤ p(K − (p − 1)D) · F Z X = p(deg(K ) − (p − 1)D · F ) (4.1) =−2 p − p(p − 1)D · F . Denote by G = π F. We claim that there is an inequality −2 p ≤ K · G.Let F = G Z red then we have −2 ≤ K · F . It is therefore sufficient to show that if G is non-reduced then it is of multiplicity p, i.e., G = pF . This is clear since π is of degree p. This together with the inequality (4.1) implies that −2 p ≤ K · G ≤−2 p − p(p − 1)D · F , but −p(p − 1)D · F < 0 which leads to a contradiction. The above proposition treats the case when running a K -MMP on a log del Pezzo surface X ends with a Mori fiber space f : Y → B. We will use Lacini’s results on liftability (Theorem 4.3) in the case when Y is a del Pezzo surface of Picard rank one. We first recall the definition of liftability to W (k). Definition 4.2 [8, Definition 8.11] Let k be a perfect field of positive characteristic and let X be a scheme smooth over Spec(k).Let D = D be a reduced simple normal crossing divisor on X. A lifting of D = D ⊂ X to W (k) consists of a scheme X and i 2 i =1 subschemes D all defined and flat over W (k) such that: X = X × Spec(k) and D = D × Spec(k), Spec(W (k)) i Spec(W (k)) 2 i 2 for 1 ≤ i ≤ n. For a pair (X , D) consisting of a smooth surface and a reduced simple normal crossing divisor as above, [6] defines the notion of liftability to characteristic zero over a smooth base. We refer the reader to [6, Definition 2.15] for the definition. If the pair (X , D) lifts to characteristic zero over a smooth base then the pair (X , D) lifts to W (k) [6, Remark 2.16]. We restate below the liftability theorem for log del Pezzo surfaces of Picard rank one proven by Justin Lacini. Theorem 4.3 [17, Theorem 7.2] Let S be a log del Pezzo surface of Picard rank one over an algebraically closed field of characteristic p > 5. Then there exists a log resolution μ : V → S such that (V , Exc(μ)) lifts to characteristic zero over a smooth base. We believe the following lemma to be well-known to experts, however since we did not find a reference we include it here. We are indebted to Maciej Zdanowicz for helping us with its proof. Lemma 4.4 Let X be a smooth surface and D = D be a reduced simple normal i =1 crossing divisor. Suppose that (X , D) lifts to W (k).Let x ∈ X be a closed point. If π : Y → −1 X is the blow up of X at x then (Y , Exc(π ) + π D) lifts to W (k). Proof Let (X , D ) be a lift of (X , D) to W (k).Let x ∈ X be a closed point. By formal smoothness there exists a lift x ∈ X of x. We claim the following: Claim If J ⊂{1, 2,..., n} is an index set such that x ∈ D but x ∈ / D for k ∈ / J , i k i ∈J then, there exists a lift x ∈ X , such that x ∈ D ,but x ∈ / D for k ∈ / J . i ∈J i k 123 E. Arvidsson Proof If x ∈ X − Supp(D) we can choose a lift x ∈ X − Supp(D ) and hence any such lift x satisfies the claim. Suppose that x ∈ Supp(D).By[8, Lemma 8.13 d)], we may assume that there is an open subset U containing x and a diagram with all squares being Cartesian: UU 2 2 A A k W (k) Spec(k) Spec(W (k)) 2 2 where A = Spec(k[t , t ]), A = Spec(W (k)[t , t ]), U and U are open subsets of 1 2 2 1 2 k W (k) X and X respectively and the morphisms e and e are étale. By [8, Lemma 8.14 e)], we may assume that we have chosen local parameters: φ = e (t ) ∈ O and φ = e (t ) ∈ O for i = 1, 2, i i U i U such that x = V((φ ,φ )) and such that (upon relabeling the components of D) there exists 1 2 a subset I of {1, 2}, such that D ∩ U = V( φ ) and D ∩ U = V( φ ).Let i ∈I i ∈I x = V((φ ,φ )),then x satisfies the claim. 1 2 Let π : Y → X be the blow up of x ∈ X ,where x is chosen as in Claim 4.We −1 −1 want to prove that (Y ,π D + Exc(π )) is a lift of (Y , Exc(π ) + π D) to W (k).If ∗ ∗ −1 x ∈ / Supp(D), for any choice of x ∈ / Supp(D ) it is clear that (Y ,π D + Exc(π )) −1 is a lift of (Y , Exc(π ) + π D) to W (k). We therefore assume that x = V((φ ,φ )), 2 1 2 x = V((φ ,φ )) and that U , U , D and D are as before. By the commutativity of blowing 1 2 up with flat base change [21, Lemma 085S] there exists, for any ring S and an étale morphism T → A , a natural Cartesian diagram: ∗ ∗ Bl (T ) Bl (A ) V ((e (t ),e (t ))) V ((t ,t )) 1 2 1 2 T A 2 ∗ where A = Spec(S[t , t ]). The strict transform of V( e (t )) is the base change to T 1 2 i i ∈I of the strict transform of V( t ) for any subset I ⊂{1, 2}. Using the notation introduced i ∈I in the beginning or this proof we therefore have: Bl  (U ) = Bl (A ) × U x V(t ,t ) 1 2 W (k) A W (k) and Bl (U ) = Bl (A ) × U . x V(t ,t ) 1 2 k A Since the base change of an étale morphism is étale, Bl  (U ) → Bl A is étale. x V (t ,t ) 1 2 W (k) Consequently, we are reduced to the following situation: 123 On the Kodaira vanishing theorem for log... 2 2 Bl A Bl A V(t ,t ) V (t ,t ) 1 2 1 2 k W (k) 2 2 A A k W (k) Spec(k) Spec(W (k)) The top square, and therefore the whole diagram, is clearly Cartesian. In this situation it is evident that Bl A is smooth over W (k) and that the strict transform of each of the V(t ,t ) 2 1 2 W (k) coordinate axes, as well as the exceptional divisor of the blow up, are flat and hence smooth over W (k) [8, Lemma 8.13]. Since the base change of an étale morphism is étale this shows −1 that Y as well as the support of π D ,for 1 ≤ i ≤ n,and Exc(π ) are all flat over W (k). ∗ i −1 −1 Consequently (Y ,π D + Exc(π )) is a lift of (Y , Exc(π ) + π D) to W (k). ∗ ∗ Theorem 4.5 (Theorem D) Let X be a log terminal projective surface with K not pseudo- effective, over a perfect field k of characteristic p > 5. Then Kodaira vanishing holds on X, i.e., for all ample Z-divisors D on X we have H (X , O (D + K )) = 0, for all i > 0. X X Proof By base change to an algebraic closure of k we can assume that k is algebraically closed. Let D be an ample Q-Cartier Z-divisor. It is sufficient to prove that 0 = H (X , O (−D)) = H (X , O (K + D)).Werun a K -MMP g : X → Y,where Y is either a klt del Pezzo X X X surface of Picard rank one or Y admits a Mori fiber space structure f : Y → B onto a curve B. (1) If f : Y → B is a K -MFS onto a curve, then the result follows from Proposition 4.1. (2) If there exists a birational morphism X → Y where Y is a log del Pezzo surface of Picard rank one then by Theorem 4.3 there exists a log resolution μ : (V , Exc(μ)) → Y such that (V , Exc(μ)) lifts to characteristic zero over a smooth base where Exc(μ) denotes the reduced divisor supported at the exceptional locus of μ. There is an induced rational map μ : V X.Let π : V → V be a resolution of the indeterminacy locus of μ . The morphism π : V → V can be factorised as a composition of blowups at points. By −1 Lemma 4.4, (V ,π Exc(μ) + Exc(π )) lifts to W (k). The diagram below illustrates the situation: −1 X (V ,π Exc(μ) + Exc(π )) μ ˜ Y (V , Exc(μ)) −1 The support of Exc(μ) ˜ is contained in the support of π Exc(μ) + Exc(π ). Therefore μ ˜ : (V , Exc(μ)) ˜ → X is a log-resolution that lifts to W (k).By[6, Lemma 6.1] the result follows. Acknowledgements I would like to thank my PhD-advisor Zsolt Patakfalvi for the help he has given me during this work. I would like to thank Jakub Witaszek for answering questions about [6]. I also thank Fabio Bernasconi and Maciej Zdanowicz for very useful discussions and for reading earlier drafts of this paper. Funding Open Access funding provided by EPFL Lausanne. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, 123 E. Arvidsson and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. References 1. Arvidsson, E., Bernasconi, F., Lacini, J.: On the Kawamata–Viehweg vanishing for log del Pezzo surfaces in positive characteristic (2020). arXiv:2006.03571 2. Bernasconi, F.: Kawamata–Viehweg vanishing fails for log del Pezzo surfaces in characteristic 3 (2017). arXiv:1709.09238 3. Bernasconi, F.: On the base point free theorem for klt threefolds in large characteristic (2019). arXiv:1907.10396 4. Cascini, P., Tanaka, H.: Smooth rational surfaces violating Kawamata–Viehweg vanishing. Eur. J. Math. 4(1), 162–176 (2018) 5. Cascini, P., Tanaka, H.: Purely log terminal threefolds with non-normal centres in characteristic two. Am. J. Math. 141(4), 941–979 (2019) 6. Cascini, P., Tanaka, H., Witaszek, J.: On log del Pezzo surfaces in large characteristic. Compos. Math. 153(4), 820–850 (2017) 7. Ekedahl, T.: Canonical models of surfaces of general type in positive characteristic. Inst. Hautes Études Sci. Publ. Math. 67, 97–144 (1988) 8. Esnault, H., Viehweg, E.: Lectures on Vanishing Theorems. DMV Seminar, vol. 20. Birkhäuser Verlag, Basel (1992) 9. Fujita, T.: Vanishing theorems for semipositive line bundles. In: Algebraic Geometry (Tokyo/Kyoto, 1982), Lecture Notes in Math., vol. 1016, pp. 519–528. Springer, Berlin (1983) 10. Hartshorne, R.: Stable reflexive sheaves. Math. Ann. 254(2), 121–176 (1980) 11. Hartshorne, R.: Generalized divisors on Gorenstein schemes. In: Proceedings of Conference on Algebraic Geometry and Ring Theory in honor of Michael Artin, Part III (Antwerp, 1992), vol. 8, pp. 287–339 (1994) 12. Hacon, C., Witaszek, J.: On the rationality of Kawamata log terminal singularities in positive characteristic. Algebr. Geom. 6(5), 516–529 (2019) 13. Jiang, C.: Bounding the volumes of singular weak log del Pezzo surfaces. Int. J. Math. 24(13), 1350110 (2013) 14. Kawakami, T.: On Kawamata–Viehweg type vanishing for three dimensional Mori fiber spaces in positive characteristic (2020). arXiv:2004.07125 15. Kollár, J., Mori, S.: Birational Geometry of Algebraic Varieties, Cambridge Tracts in Mathematics, vol. 134. Cambridge University Press, Cambridge (1998) (With the collaboration of C. H. Clemens and A. Corti, Translated from the 1998 Japanese original) 16. Kollár, J.: Rational Curves on Algebraic Varieties, Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics, vol. 32 [Results in Mathematics and Related Areas. 3rd Series. A Series of Modern Surveys in Mathematics]. Springer, Berlin (1996) 17. Lacini, J.: On rank one log del Pezzo surfaces in characteristic different from two and three (2020). arXiv:2005.14544 18. Milne, J.S.: Étale Cohomology. Princeton Mathematical Series, vol. 33. Princeton University Press, Princeton (1980) 19. Patakfalvi, Z., Waldron, J.: Singularities of General Fibers and the LMMP. arXiv:1708.04268 20. Raynaud, M.: Contre-exemple au “vanishing theorem” en caractéristique p>0. In: C. P. Ramanujam—a tribute, Tata Inst. Fund. Res. Studies in Math., vol. 8, pp. 273–278. Springer, Berlin (1978) 21. The Stacks project authors.: The stacks project (2020). https://stacks.math.columbia.edu 22. Tanaka, H.: Minimal models and abundance for positive characteristic log surfaces. Nagoya Math. J. 216, 1–70 (2014) 23. Tanaka, H.: Minimal model program for excellent surfaces. Ann. Inst. Fourier (Grenoble) 68(1), 345–376 (2018) 24. Witaszek, J.: Effective bounds on singular surfaces in positive characteristic. Mich. Math. J. 66(2), 367– 388 (2017) Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Mathematische Zeitschrift Springer Journals

On the Kodaira vanishing theorem for log del Pezzo surfaces in positive characteristic

Mathematische Zeitschrift , Volume OnlineFirst – May 6, 2021

Loading next page...
 
/lp/springer-journals/on-the-kodaira-vanishing-theorem-for-log-del-pezzo-surfaces-in-AJxaws1IG1
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
ISSN
0025-5874
eISSN
1432-1823
DOI
10.1007/s00209-021-02742-6
Publisher site
See Article on Publisher Site

Abstract

We investigate the vanishing of H (X , O (−D)) for a big and nef Q-Cartier Z-divisor D on a log del Pezzo pair (X,) over a perfect field of positive characteristic p. Keywords Kodaira vanishing · Log del pezzo surfaces · Positive characteristic Mathematics Subject Classification 14E30 · 14F17 · 14J45 · 13A35 Introduction It has long been known that the (Kodaira and Kawamata–Viehweg) vanishing theorems, so fundamental to birational geometry in characteristic zero, in general fail for surfaces in positive characteristic [20]. Several people have investigated different classes of surfaces over perfect fields of positive characteristic for which the (Kodaira and Kawamata–Viehweg) vanishing theorems may or may not hold. The question is subtle; for example, every smooth rational surface over a perfect field satisfies Kodaira vanishing, however over any alge- braically closed field of positive characteristic, there are smooth rational surfaces that violate Kawamata–Viehweg vanishing [4]. On the other hand a smooth surface with ample anti- canonical bundle satisfies Kawamata–Viehweg vanishing theorem over any algebraically closed field [4, Proposition A.1]. This however is no longer true if (klt) singularities are allowed and counterexamples (even to Kodaira vanishing) have been constructed over alge- braically closed fields of characteristic 2 [5]and 3[2]. The situation simplifies in large characteristics. In [6] the authors prove the existence of an integer p such that over an algebraically closed field of characteristic p > p every 0 0 log del Pezzo surface satisfies Kawamata–Viehweg vanishing theorem. Finding an effective bound for this p is a central open question in positive characteristic birational geometry. For example over an algebraically closed field of characteristic p > max{5, p } athree The author was supported by SNF Grant #200021/169639. B Emelie Arvidsson emelie.arvidsson@epfl.ch EPFL SB MATH CAG, MA C3 615 (Batiment MA), Station 8, 1015 Lausanne, Switzerland 123 E. Arvidsson dimensional klt singularity is rational [12, Theorem 1.1], and threefolds satisfy a refined version of the basepoint free theorem [3, Theorem 1.1]. The construction of the integer p in [6] is implicit. During the course of the proof of [6, Theorem 1.1] the authors consider log del Pezzo surfaces belonging to a bounded family over Spec(Z). The construction of the integer p does in particular depend on this family in an implicit way. It is therefore natural to ask if it is possible to prove vanishing theorems in large explicit characteristics depending on some explicit numerical invariant of a bounded family of log del Pezzo surfaces. In this direction, we show that Kodaira vanishing (for big and nef divisors) holds on a del Pezzo surface of bounded index I in characteristic p > p (I ) where p (I ) is an explicit polynomial in the index I . We also determine an explicit bound for p 0 0 (for Kodaira vanishing) in terms of > 0for an -klt log del Pezzo surface. Our main results in this direction are the following. Theorem A Let X be a projective klt surface over a perfect field k of characteristic p > 0 with ample anti-canonical divisor −K of Cartier index I . Let D be a big and nef Q-Cartier Z-divisor on X. (1) If I = 1, i.e., if −K is an ample Cartier divisor then: • if p ≥ 9221,then H (X , O (−D)) = 0 • if p ≥ 5,then H (X , O (D)) = 0. 2 3 2 (2) If I ≥ 2 and p ≥ (13 − 45I ) (2I + 4I + 2I ) + 1, then: 1 1 H (X , O (D)) = H (X , O (−D)) = 0. X X −1/2 Theorem B Let 0 < < 3 be a real number. Let (X,) be a projective -klt surface with −(K + ) ample over a perfect field k of characteristic (128) 128 25 p ≥ 2 + 1. Then Kodaira vanishing holds on X, i.e., for all ample Q-Cartier Z-divisors D on X we have that H (X , O (D + K )) = 0, for all i > 0. X X In this note, we use a technique due to Ekedahl [7]. When trying to prove Kodaira vanishing in explicit characteristics using this technique, the difficulties arise from Weil divisors which are far from being Cartier. For example, on a log del Pezzo surface over an algebraically closed field of characteristic p ≥ 4c + 1 we can prove that H (X , O (K + A)) = 0 X X for every ample Q-Cartier Weil divisor A of Cartier index ≤ c (see Remark 3.5). Another strategy is therefore to control the Cartier index of a Weil divisor D in terms of the Cartier index I of (K + ). We prove, in this way, that a big and nef Z-divisors D of big volume relative to the index I in characteristic p ≥ 5 satisfies H (X , O (−D)) = 0. Theorem C Let (X,) be a projective klt surface over a perfect field k of characteristic p ≥ 5 with −(K + ) an ample Q-Cartier Q-divisor of Cartier index I . Let D be a big and nef Q-Cartier Z-divisor. Then the following holds true: 2 1 (1) If I = 1 and D ≥ 9,then H (X , O (−D)) = 0. 2 3 2 1 (2) If I ≥ 2 and D ≥ 2I + 4I + 2I, then H (X , O (−D)) = 0. During the preparation of this note we were informed that J. Lacini has classified all log del Pezzo surfaces of Picard rank one over an algebraically closed field of characteristic 123 On the Kodaira vanishing theorem for log... p ≥ 5[17, Theorem 1.1]. From this classification it follows that all log del Pezzo surfaces of Picard rank one over an algebraically closed field of characteristic p > 5 admit a log resolution that lifts to W (k) [17, Theorem 7.2]. It therefore follows from his work that Kodaira vanishing holds for log del Pezzo surfaces of Picard rank one over an algebraically closed field of characteristic p > 5(see[6, Lemma 6.1]). Combining his result with the techniques exploited in this paper, one can prove the Kodaira vanishing theorem on a log del Pezzo surface over an algebraically closed field of characteristic p > 5. At the end of this note we illustrate this argument. Building on the liftability of rank one log del Pezzo surfaces in characteristic p > 5 by Lacini we can prove the following theorem. Theorem D Let (X,) be a projective log del Pezzo surface over a perfect field of charac- teristic p > 5. Then Kodaira vanishing holds on X, i.e., for all ample Z-divisors D on X we have H (X , O (D + K )) = 0, for all i > 0. X X After this note was made public, the author has, in joint work with Fabio Bernasconi and Justin Lacini, extended the above result. In [1] we show that log del Pezzo surfaces over a perfect field of characteristic p > 5 satisfy the Kawamata Viehweg vanishing theorem. In [1] we also construct a del Pezzo surface over an algebraically closed field of characteristic p = 5 on which the Kodaira vanishing theorem fails. This implies that the characteristic p > 5inTheorem D is optimal. 1 Preliminaries By a variety we mean a finite type integral separated scheme over a field. We will work exclusively over a perfect field k. A Q-divisor D is said to be Q-Cartier if there exists an integer m such that mD is Cartier. A Q-divisor is ample/nef/big if it is Q-Cartier and an integer multiple is ample/nef/big as a line bundle. If X is a normal variety, then the reflexive sheaves on X are determined (up to isomorphism) by their restriction to a big open subset U (an open subset U of X is big if codim (X −U ) ≥ 2). If i : U → X is the inclusion of a big open subset and M is a reflexive sheaf on U then i M is a reflexive sheaf on X, moreover for any reflexive sheaf G on X we have an isomorphism G = i G [10, Proposition 1.6]. Let U denote the smooth locus of X.Since X ∗ | is normal U is big. Therefore, for any Weil divisor D on X the restriction of D to the smooth locus defines a reflexive sheaf i O (U ∩ D) of rank one on X. Conversely, any reflexive ∗ U sheaf of rank one F on X defines a Weil divisor D by setting D to be the closure of a Weil divisor E on U satisfying F = O (E ). This defines a one to one correspondence between | U the reflexive sheaves of rank one on X up to isomorphism and the set of Weil divisors on X up to rational equivalence [10]. We denote by O (D) the reflexive sheaf corresponding to a divisor D on X. We use the notation K for the class corresponding to the closure of a canonical divisor K on U in X, hence O (K ) i O (K ). If X is normal and projective U X X ∗ U U then O (K ) can be seen to be a dualizing sheaf for X [15, Proposition 5.75]. We use the X X notation ω for a dualizing sheaf on X. In general if D and D are Weil divisors on a normal variety X,then O (D ) ⊗ 1 2 X 1 O (D ) is not reflexive. However, the dual M of a coherent sheaf M is always reflex- X 2 ive [10, Corollary 1.2] and we may define a product structure on the set of reflexive ∨∨ sheaves by defining O (D ) ⊗ O (D ) := (O (D ) ⊗ O (D )) .Wehavethe fol- X 1 X 2 X 1 X 2 lowing identities for Weil divisors D , D and a Cartier divisor C: O (D + D ) = 1 2 X 1 2 123 E. Arvidsson ∨∨ ∨ ∼ ∼ (O (D ) ⊗ O (D )) , O (−D ) = O (D ) , O (D + C ) = O (D ) ⊗ O (C ) X 1 X 2 X 1 X 1 X 1 X 1 X and Hom (O (D ), O (D )) O (D − D ). This product structure makes the set of O X 1 X 2 X 2 1 reflexive sheaves on X (up to isomorphism) into a group such that the correspondence between Weil divisors and reflexive sheaves on X, as described above, becomes an isomorphism of groups. By a log pair (X,) we shall mean a normal variety X together with an effective Q-divisor on X such that K +  is Q-Cartier. By a klt (respectively lc) pair we shall mean a log pair with klt (respectively lc) singularities in the sense of [15, Section 2]. By [23, Corollary 4.11] if X is a surface and (X,) is a klt pair then X is Q-factorial. A normal surface X is said to be a del Pezzo surface if it has klt singularities and −K is Q-Cartier and ample. A log pair (X , B) is said to be a log del Pezzo surface if it has klt singularities and −(K + B) is ample. 1.1 Serre vanishing and Serre duality for Z-divisors One of the most fundamental vanishing theorems in algebraic geometry is that of Serre. It naturally implies a Weil divisor version. Theorem 1.1 Let X be a normal projective variety. If D is an ample Q-Cartier Z-divisor and F is a coherent sheaf on X then H (X , F ⊗ O (lD)) = 0 for all l big enough and all i > 0. Proof Let m be an integer such that mD is Cartier. By Serre vanishing there exists, for all j ⊗n 1 ≤ i ≤ m,an n such that H (X , F ⊗ O (iD) ⊗ O (mD) ) = 0, for all n ≥ n i X X i ⊗n and all j > 0. Since O (mD) is Cartier O (iD) ⊗ O (mD) = O ((i + mn)D).Set X X X X N = max n .For all l ≥ (m + 1)N , we can write l = i +mk,for 1 ≤ i ≤ m − 1, where 1≤i ≤m i i i ⊗k k ≥ N . Therefore, we have H (X , F ⊗O (lD)) = H (X , F ⊗O (iD)⊗O (mD) ) = 0, X X X for all l ≥ (m + 1)N . Corollary 1.2 Let D be a big and nef integral divisor on a normal projective surface X, then H (X , O(−mD)) = 0 for all m >> 0. Proof Let r be the Cartier index of D.Let n > 0 be such that H (X , O(lD)) is non zero for all l ≥ n, this is possible since D is big. By Fujita vanishing [9, Theorem 10] there exists an integer m such that H (X , O(−(kr + l)D)) = 0for all l ≥ n and all k ≥ m, this proves the Corollary. We will repeatedly need to use a form of Serre Duality valid for reflexive sheaves on a normal surface. Theorem 1.3 (Serre Duality for CM-sheaves) Let X be a projective scheme of pure dimension n over a field k. Let F be a CM sheaf on X such that Supp(F ) is of pure dimension n. Then i n−i H (X , F ) is dual to H (X , Hom (F,ω )). O X Proof E.g., [15, Theorem 5.71] In particular, by [10, Proposition 1.3] for any Z-divisor D on a normal projective surface i 2−i X the vector space H (X , O (D)) is dual to H (X , O (K − D)) for i = 0, 1. X X X 123 On the Kodaira vanishing theorem for log... Remark 1.4 If D is a big and nef Z-divisor on a normal projective surface X then 0 2 H (X , O (−D)) = 0 and hence by duality H (X , O (D + K )) = 0. Therefore, Kodaira X X X 1 1 vanishing is equivalent to the vanishing of H (X , O (−D)) = H (X , O (K + D)). X X X 1.2 Frobenius techniques Let X be a scheme over a positive characteristic base S,let F denote the absolute Frobenius on X. The absolute Frobenius F is not a morphism of S-schemes unless F on S is the X S identity. In particular we have the following commuting diagram: X /S X ×SX SS . where F denotes the relative Frobenius. When S = Spec(k) for a perfect field k then X /S the absolute Frobenius on S is an isomorphism and X ×SX is an isomorphism over S if we consider X as a scheme over S by post-composing the structure morphism to S −1 with F on S. We often make this identification, under which the relative and the absolute Frobenius coincide, and denote by F : X → X the corresponding morphism. Let X be a normal variety over an algebraically closed field of positive characteristic. If D is an integral ∗ ∨∨ divisor on X then (F O (D)) = O (pD), since they agree on the regular locus which is X X a big open subset of X. There is a natural map (unit of adjunction) O (D) → F F O (D). X ∗ X This induces a morphism: i i ∗ i ∗ ∨∨ H (X , O (D)) → H (X , F O (D)) → H (X , F O (D) ). X X X Therefore, the Frobenius on X induces morphisms for all m: i m i m+1 H (X , O (−p D)) → H (X , O (−p D)). X X If D is an ample divisor whose associated Weil sheaf is CM, then these groups eventually vanish for m >> 0 by Theorem 1.3 and Corollary 1.1. Therefore, Kodaira vanishing theorem i m for such divisors D hold true on X if and only if the morphisms H (X , O (−p D)) → i m+1 H (X , O (−p D)) are injective for all m and all i < dim(X ). 1.3 Non-vanishing and the associated ˛ -torsor O (D ) X 0 Let X be a projective normal surface and D be a Q-Cartier Z-divisor on X.Let X ⊂ X denote the smooth locus of X and D := D ∩ X be the restriction of D to the smooth 0 0 locus. Let F denote the Frobenius on X . We will assume that H (X , O (−D)) = 0and 0 0 X 1 1 that H (X , O (−pD)) = 0. Let α ∈ H (X , O (−D)) denote a non-trivial element of X X i i the kernel of the map H (X , O (−D)) → H (X , O (−pD)) described in Sect. 1.2.Since X X Z := X − X is the complement of a big open subset of X and O (nD) is reflexive (and so 0 X 0 0 in particular, H (X , O (nD)) H (X , O (nD ))), we see from the local cohomology X 0 X 0 0 long exact sequence that we have an inclusion: 1 1 H (X , O (−nD)) → H (X , O (−nD )) X 0 X 0 123 E. Arvidsson for all n. By our assumptions, α therefore defines a non-trivial element: α ∈ H (X , O (−D )), 0 0 X 0 which belongs to the kernel of the morphism: 1 1 H (X , O (−D )) → H (X , O (−pD )) 0 X 0 0 X 0 0 0 induced by F . This kernel has a geometric description on X which is described in the next 0 0 section. 1.3.1 ˛ -torsors An invertible sheaf L on a scheme X naturally defines a sheaf of groups under addition −i representable by the affine group scheme L := Spec ( L ). Let X be defined over a field X i k of characteristic p > 0. Then there exists a homomorphism of sheaves of additive groups L → L , defined by raising a local section to its p-power. This is a purely characteristic p p p p > 0 phenomenon, since for any two local sections s and t of L we have (s + t ) = s + t in characteristic p. Relative to the Zariski site of X the above morphism of sheaves is not in general a surjection. However, as a morphism of X -sheaves it is, i.e., the corresponding morphism fppf of sheaves of groups on the flat site of X is surjective [18, II.2.18]. The kernel of this morphism is a sheaf of groups on X which we denote by α . By construction α is representable fppf L L p−1 −i by an affine group scheme over X, denoted α . In fact,wehave α = Spec ( L ). L L i =0 The short exact sequence of sheaves of groups (relative to the flat- topology on X) 0 α LL 0 induces a long exact sequence on cohomology [18,4,Prop4.5]: 0 p 1 1 1 p ··· → H (X , L ) → H (X,α ) → H (X , L) → H (X , L ) → ... fppf 1 1 p Therefore, a non-trivial element α in the kernel of H (X , L) → H (X , L ) defines a 1 1 ˇ ˇ non-trivial element α ∈ H (X,α ). The group H (X,α ) has a geometric meaning, L L fppf fppf i.e., a non-trivial element α ∈ H (X,α ) corresponds to a non-trivial α -torsors Y → X, L L fppf see for example, [18, Proposition 4.6]. The torsor Y corresponding to α ∈ H (X,α ) fppf constructed as above were studied by Ekedahl. Proposition 1.5 ([7, pp. 106–107], [19, Theorem 2.11]) Let X be a normal variety over an algebraically closed field k of characteristic p > 0.Let L ∈ Pic(X ). A non-trivial element of 1 1 p the kernel under the Frobenius action: H (X , L) → H (X , L ), gives rise to a non-trivial −1 α -torsor β : Y → X. Locally over U = Spec(A), the morphism β U → U is given by A → A[x ]/(x − a) for some element a ∈ A which is not a p-power. Therefore, β is purely inseparable of degree p. If X is a G and S variety, then Y is also a G and S variety which 1 2 1 2 ∗ p−1 satisfies: ω = β (ω ⊗ L ). Y X Here S denotes Serre’s condition and G denotes Gorenstein in codimension one. 2 1 Remark 1.6 Y is in general not normal. 123 On the Kodaira vanishing theorem for log... 1.4 Geometric construction for Weil-divisors Let X be a projective normal surface over an algebraically closed field of charcteristic p > 0and let D a Q-Cartier ample Weil divisor on X such that H (X , O (−D)) = 0and H (X , O (−pD)) = 0. Let X denote the smooth locus of X and let D = D ∩ X . X 0 0 0 Let α be a non-trivial element of H (X,α ) coming from a non-trivial element 0 O (D ) fppf X 0 α ∈ H (X , O (D)) as described in point Sect. 1.3.Let π : Y → X be the corresponding X 0 0 0 non-trivial α -torsor. Then π O is a locally free O -algebra. Let i : X → X O (D ) 0∗ Y X 0 X 0 0 0 denote the inclusion of the regular locus. The multiplication on X extends (because i π O 0 ∗ 0∗ Y is reflexive) to make i π O into a sheaf of O -algebras. We may therefore define Y := ∗ 0∗ Y X Spec (i π O ).Let π : Y → X denote the natural morphism. ∗ 0∗ Y X 0 Lemma 1.7 In the situation above π : Y → X is a finite degree p morphism and Y is a projective S and G surface that satisfies 2 1 K = π (K − (p − 1)D). Y X Proof Since π O = i π O is S we have by [15, Proposition 5.4] that Y is S . Since X ∗ Y ∗ 0∗ Y 2 2 and X agree in codimension one, Y is G . Since everything is S and G we may replace 0 1 2 1 ∗ ∗ X with X by [11, Theorem 1.12], therefore π (K ) and π (D) are well-defined and the 0 X formula K = π (K − (p − 1)D) follows from Proposition 1.5. Y X Lemma 1.8 With the notation as above, let η : Y → Y be the normalization of Y and let π : Y → X denotes the induced morphism. There exists an effective Z-divisor  ≥ 0 on Y such that K +  = π (K − (p − 1)D). Proof Since everything is S and G we may replace Y with a big open set and assume that Y 2 1 ∼ ∼ is Gorenstein and Y is regular. Affine locally we may assume that ω = O and ω = O . Y Y Y Y In this situation [19, Lemma 2.14] proves that the natural map ω → η ω has image equal to the conductor ideal, where η ω is identified with O . 1.5 Bend and break By studying different properties of the variety Y appearing in Lemma 1.8 one hopes to arrive at a contradiction to the assumed non-vanishing, H (X , O (−D)) = 0. This strategy was successfully employed by Ekedahl [7] and more recently in [19]. The main tool is the use of bend and break, together with the expression for the canonical divisor given by Lemma 1.8. Together, this can be used in order to derive inequalities comparing intersection numbers on X with the characteristic p of the base field. We will repeatedly use the following: Theorem 1.9 [16, Chapter II, Theorem 5.8] Let X be a projective variety over an alge- braically closed field, C a smooth, projective and irreducible curve, f : C → X a morphism and M any nef R-divisor. Assume that X is smooth along f (C ) and K · C > 0. Then for every x ∈ f (C ) there is a rational curve ⊂ X containing x such that M · C M · ≤ 2dim(X ) . −K · C 123 E. Arvidsson 2 Kodaira vanishing for divisors of big volume We will use the following terminology: We will say that a curve C on a surface X is big and basepoint free if the linear system |C | defines a birational morphism. If C is big and basepoint free, then for any finite number of points in X there exists C ∈|C | such that C avoids all those points. Lemma 2.1 Let X be a projective normal surface. Suppose that there exists an effective Q- divisor E such that −(K + E ) is a Q-Cartier ample divisor of Cartier index I . If D is a 2 2 2 big Q-Cartier semiample Z-divisor such that D ≥ I (K + E ) then there exists an ample Cartier divisor A and a big basepoint free curve C on X such that D · C ≥ A · C. Proof Suppose, in order to arrive at a contradiction, that for every big basepoint free curve C and every ample Cartier divisor A of X we have D · C < A · C.For m divisible enough, −m(K + E ) is very ample. Therefore, for A =−I (K + E ) and C ∈| − m(K + E )| X X X we find −(K + E ) · D < I (K + E ) . Similarly, for n divisible enough, nD is basepoint X X free and we may consider a curve C ∈|nD| to find that D < −I (K + E ) · D. Putting this 2 2 2 together, we find that D < I (K + E ) . Theorem 2.2 Let X be a projective normal surface over an algebraically closed field k of characteristic p ≥ 5. Suppose that there exists an effective Q-divisor E such that −(K + E ) is a Q-Cartier ample divisor of Cartier index I . If D is a Q-Cartier nef and big semiample 2 2 2 1 Z-divisor such that D ≥ I (K + E ) then H (X , O (−D)) = 0. X X Proof Suppose that the theorem is not true. Let D be a nef and big semiample Z-divisor 2 2 2 1 on X such D ≥ I (K + E ) and assume that H (X , O (−D)) = 0. By Corollary X X 1 j 1.2 there exists some j ≥ 1suchthat H (X , O (−p D)) = 0. If we replace D with j −1 j −1 p D the inequality will be satisfied for p D. Suppose we can prove the theorem for j −1 1 j −1 p D then H (X , O (−p D)) = 0 and, hence, by descending induction on j,wehave proven that H (X , O (−D)) = 0. Therefore, there is no loss of generality to assume that H (X , O (−pD)) = 0. Let π : Y → X beadegree p-cover corresponding to a non-zero element y ∈ H (X , O (−D)) as in Sect. 1.3. By replacing Y with its normalisation, we may assume that Y is normal and that ∗ ∗ K = π K + (1 − p)π D −  (2.1) Y X where  is effective (Lemma 1.8). By Lemma 2.1 there exists a big basepoint free curve C on X and an ample Cartier divisor A such that D · C ≥ A · C.Let π : Y → X be as above. By replacing C with a large enough multiple we see that there exists a curve C on Y such that: • C is contained in the smooth locus of Y • C is not contained in a component of • C = π C for some curve C on X not contained in a component of E andsuchthat ∗ Y D · C ≥ A · C . By Eq. (2.1)wehave −K · C =−K · C + (p − 1)D · C +  · C ≥−K · C,where Y Y X Y X the last inequality comes from the assumption that D is nef and that C is not contained in a component of . By assumption −(K + E ) is ample, hence −K · C > E · C ≥ 0. X X Therefore −K · C > 0. By Theorem 1.9 for C and π A on Y , there exists a rational Y Y Y 123 On the Kodaira vanishing theorem for log... curve passing through a point x ∈ C such that: x Y π A · C π A · ≤ 4 . −K · C Y Y By Equation (2.1), we find: ∗ ∗ ∗ ∗ π A · (−π K + (p − 1)π D + ) · C ≤ 4π A · C . x X Y Y However, since A is an ample Cartier divisor and π is finite π A · ≥ 1. Moreover, ∗ ∗ ∗ −π K · C =−K · C > 0and C ·  ≥ 0. By assumption, π A · C ≤ π D · C , X Y X Y Y Y therefore: ∗ ∗ ∗ ∗ (p − 1)π A · C < π A · (−π K + (p − 1)π D + ) · C . Y x X Y Putting these inequalities together we find: ∗ ∗ (p − 1)π A · C < 4π A · C Y Y and, therefore, p < 5. Theorem 2.2 implies, together with results of [13], our main theorem for divisors of big volume. Proof of Theorem C By base change to an algebraic closure of k we can assume that k is algebraically closed. By the Basepoint free theorem [22, Theorem 1.3], D is semiample. Note that a klt surface of index I,necessarily,is -klt. By [13, Theorem 1.3], we therefore 2 2 have that (K + ) ≤ max{9, 2I + 4 + }. The result therefore follows from Theorem 2.2. Remark 2.3 Recently Kawakami improved both Theorems A and C in the case when I = 1. In particular, let X be a del Pezzo surface with canonical singularities over an algebraically closed field of positive characteristic p and let D be a big and nef Q-divisor. In [14]itis 1 1 shown that H (X , O (−D)) = 0if p = 2[14, Theorem 1.1] and that H (X , O (D)) = 0 X X in all characteristic [14, Proposition 4.10]. See [14] for the precise statements which are slightly more general. Corollary 2.4 Let X be a klt del Pezzo surface over an algebraically closed field of charac- teristic p ≥ 5, such that −K is ample of Cartier index I . Let D be a big and nef Z-divisor of Cartier index c. Then H (X , O (D)) = 0 if either of the following holds: 2 2 2 • K ≥ (c − 1) D and D is ample 2 2 2 • D ≥ (I − 1) K . Proof The assumptions imply that D − K is ample and H (X , O (K − D)) is dual to X X X H (X , O (D)). Assume that the claimed vanishing does not hold. Since the characteristic of the base field is greater than or equal to five, we may argue as in the proof of Proposition 2.2 to assume that for every big basepoint free curve C and every ample Cartier divisor A we have (D − K ) · C < A · C . The first inequality comes from setting A = cD and C =−mK and X X C = lD respectively, for large and divisible enough l and m. The second inequality comes from setting A =−IK and C =−mK and C = lD respectively, for large and divisible X X enough l and m. 123 E. Arvidsson 3 Kodaira vanishing in large characteristic In this section we prove Kodaira vanishing in large characteristic for log del Pezzo surfaces of bounded index. That bounded families of log del Pezzo surfaces over Spec(Z) satisfy Kodaira vanishing in large characteristic (depending only on the family) has been proven by other methods in [6]. In Proposition 3.1 below we show that the existence of a very ample divisor A of bounded volume r on a normal projective surface X over a field of characteristic p ≥ 4r + 1 implies Kodaira Vanishing whenever −K · A > 0. In particular, we do not make any further assumption on the singularities of X. Proposition 3.1 Let r > 0 be an integer. Let X be a projective normal surface over an algebraically closed field k of characteristic p ≥ 4r + 1. If there exists a very ample divisor A on X such that A ≤ r and −K · A > 0 then Kodaira vanishing (even for big and nef Q-Cartier Z-divisors) holds on X. Proof Suppose that D is a big and nef Z-divisor and suppose that H (X , O (−D)) = 0. j 1 By replacing D with p D for some j ≥ 1 we may assume that H (X , O (−pD)) = 0 (Corollary 1.2). By Ekedahl’s construction we may assume that there exists a normal variety Y and a purely inseparable degree p morphism π : Y → X such that K = π K − (p − 1)D − E (3.1) Y X for E ≥ 0. Let A be a very ample general curve on X as above. There exists an integer l > 0 such that lπ A is very ample on Y . Therefore there exists a curve A on Y such that: • A is contained in the smooth locus of Y • A is not contained in a component of E • A = π A for some A ∈|nA| where n > 0is an integer. ∗ Y By Eq. (3.1)wehave −K · A = n(−K · A + (p − 1)D · A) + E · A ≥−nK · A. Y Y X Y X By assumption −K · A > 0. Therefore −K · A > 0. By Theorem 1.9 for A and π (A) X Y Y Y on Y , there exists for every point x ∈ A a rational curve passing through x such that Y x π (A) · A π (A) · ≤ 4 . −K · A Y Y However, since A is an ample Cartier divisor and π is finite, π A · ≥ 1. By Eq. (3.1)and the assumption A ≤ r, we therefore have: ∗ ∗ (−π K + (p − 1)π D + E )A ≤ 4rn. X Y Since −K · A > 0and E · A ≥ 0, we find that: X Y (p − 1)π D · A < 4rn. Since A ∈|nA| we have π D · A ≥ n. The result follows. Remark 3.2 If (X,) is a log pair such that −(K + ) is ample then −K · A > · A ≥ 0 X X for any ample divisor A. Proof of Theorem A By base change to an algebraic closure of k we can assume that k is algebraically closed. By [24] the divisor (13 − 45I )IK is very ample. By [13, Theorem 1.3] there is an inequality K ≤ max{9, 2I + 4 + }. The result therefore follows from Proposition 3.1 and Theorem C. 123 On the Kodaira vanishing theorem for log... For the boundary version of the theorem above we need to use a common bound for the Cartier indices of K and K + . X X Theorem 3.3 Let (X,) be a log del Pezzo surface over an algebraically closed field k of characteristic p > 0.Let m ≥ 2 be such that m(K + ) and m K are both Cartier. Let D X X be a big and nef Z-divisor on X. If 8 7 5 4 3 p ≥ 173056m − 86528m + 16200m + 32400m + 16200m + 1 then H (X , O (−D)) = 0. Proof Let H = 13mK − 45m (K + ).Accordingto[24], H is a very ample Cartier X X 2 5 divisor. By [24, Lemma 6.2], we have inequalities (K +)·K ≥ 0and K ≤ 128m (2m − X X 1).By[13, Theorem 1.3], we have an inequality (K + ) ≤ 2m + 4 + . Therefore: 2 2 7 2 5 4 3 H ≤ (13) 128m (2m − 1) + (45) (2m + 4m + 2m ). We conclude by Proposition 3.1. Remark 3.4 A log del Pezzo surface (X,) where (K +) has Cartier index I is necessarily -klt for  = . As in the proof below, the Cartier index of K can therefore be bounded in terms of the Cartier index of (K + ). Proof of Theorem B By base change to an algebraic closure of k we can assume that k is algebraically closed. By assumption X is -klt, therefore the Q-factorial index q at any point x ∈ X satisfies q ≤ 2 by [24, Prop 6.1]. Moreover, by the same proposition, the Picard rank of a minimal resolution (128) 128 2 25 of X is bounded by . Hence, for every Z-divisor D, there exists some c ≤ 2 , depending on D, such that cD is Cartier. 1 k−1 Suppose that H (X , O (−D)) = 0. By replacing D with p D for some k > 1, we may assume that H (X , O (−pD)) = 0. We may, therefore, assume that there exists a normal variety Y and a degree p inseparable morphism Y → X such that K = π (K + Y X (1 − p)D) − E,for some E ≥ 0. Let C be a general complete intersection curve on Y , not contained in the support of E, such that π C = C is not contained in the support of . ∗ Y Then −K · C > 0. By Theorem 1.9 applied to C and Y Y Y ∗ ∗ π (−(K + ) + (p − 1)D) ≡−(K + E ) − π (), X Y there exists a rational curve through a point on C such that: (−(K + E ) − π ()) · C Y Y π (−(K + ) + (p − 1)D) · ≤ 4 . −K · C Y Y ∗ ∗ Since −π (K + ) · > 0, E · C ≥ 0and π () · C ≥ 0, we find that X Y Y (p − 1)D · < 4. Since cD is Cartier we have an inequality D · ≥ . Therefore: (128) 128 25 p < 4c + 1 ≤ 2 + 1. 123 E. Arvidsson Remark 3.5 From the proof of Theorem B we see that on a log del Pezzo surface (X,) over an algebraically closed field k of characteristic p ≥ 4c + 1wehave H (X , O (−D)) = 0, for every ample Z-divisor D of Cartier index ≤ c. 4 Kodaira vanishing in char p> 5 using the liftability of log del Pezzo surfaces of Picard rank one In this section we deduce the Kodaira vanishing theorem for log del Pezzo surfaces from the liftability in characteristic p > 5 of del Pezzo surfaces of Picard rank one, together with the Kodaira vanishing theorem for surfaces Y which admit a Mori fiber space structure f : Y → B onto a curve. In fact, for a surface admitting a Mori fiber space structure onto a curve the difficulties arising from non-Cartier divisors in our previous arguments are not present. We learned the main ideas in Proposition 4.1 from Zsolt Patakfalvi. Proposition 4.1 Let X be a normal projective surface over an algebraically closed field k of characteristic p > 0 which admits a proper birational morphism g : X → Yto a normal projective surface Y admitting a K -negative fibration to a curve B. That is, there exists a proper morphism f : Y → B such that f O = O and such that the ∗ Y B general fiber is isomorphic to P .Let D bea Q-Cartier big and nef Z-divisor on X. Then H (X , O (−D)) = 0. Proof According to Corollary 1.2 by replacing D with p D, we may assume that 1 1 H (X , O (−pD)) = 0. Suppose that H (X , O (−D)) = 0. We apply Ekedahl’s con- X X struction 1.3 to get a degree p finite morphism π : Z → X,where Z is normal, and ∗ ∗ K + E = π K + (1 − p)π D Z X for an effective divisor E ≥ 0. We illustrate the situation with a diagram: XY Let F be a fiber of f not containing any point which is the image of a curve contracted by g.Then F can be identified with a fiber of f ◦ g : X → B. By abuse of notation we denote a general fiber of f ◦ g by F. A general fiber of the composition f ◦ g avoids the exceptional locus of g since the curves contracted by g map to a finite number of points on Y . We compute: ∗ ∗ ∗ K · π F = π ((K − (p − 1)D) · F ) − E · π F . Z X The morphism π is finite of order p, therefore, we have: ∗ ∗ K · π F = p((K − (p − 1)D) · F − E · π F . Z X Since two general fibers F do not intersect we have F = 0 and therefore K · F =−2. Since D is big and F is general, we have D · F > 0. 123 On the Kodaira vanishing theorem for log... Moreover, since E is effective, we see (upon varying F) that the intersection number E · π F ≥ 0. We conclude that: K · π F ≤ p(K − (p − 1)D) · F Z X = p(deg(K ) − (p − 1)D · F ) (4.1) =−2 p − p(p − 1)D · F . Denote by G = π F. We claim that there is an inequality −2 p ≤ K · G.Let F = G Z red then we have −2 ≤ K · F . It is therefore sufficient to show that if G is non-reduced then it is of multiplicity p, i.e., G = pF . This is clear since π is of degree p. This together with the inequality (4.1) implies that −2 p ≤ K · G ≤−2 p − p(p − 1)D · F , but −p(p − 1)D · F < 0 which leads to a contradiction. The above proposition treats the case when running a K -MMP on a log del Pezzo surface X ends with a Mori fiber space f : Y → B. We will use Lacini’s results on liftability (Theorem 4.3) in the case when Y is a del Pezzo surface of Picard rank one. We first recall the definition of liftability to W (k). Definition 4.2 [8, Definition 8.11] Let k be a perfect field of positive characteristic and let X be a scheme smooth over Spec(k).Let D = D be a reduced simple normal crossing divisor on X. A lifting of D = D ⊂ X to W (k) consists of a scheme X and i 2 i =1 subschemes D all defined and flat over W (k) such that: X = X × Spec(k) and D = D × Spec(k), Spec(W (k)) i Spec(W (k)) 2 i 2 for 1 ≤ i ≤ n. For a pair (X , D) consisting of a smooth surface and a reduced simple normal crossing divisor as above, [6] defines the notion of liftability to characteristic zero over a smooth base. We refer the reader to [6, Definition 2.15] for the definition. If the pair (X , D) lifts to characteristic zero over a smooth base then the pair (X , D) lifts to W (k) [6, Remark 2.16]. We restate below the liftability theorem for log del Pezzo surfaces of Picard rank one proven by Justin Lacini. Theorem 4.3 [17, Theorem 7.2] Let S be a log del Pezzo surface of Picard rank one over an algebraically closed field of characteristic p > 5. Then there exists a log resolution μ : V → S such that (V , Exc(μ)) lifts to characteristic zero over a smooth base. We believe the following lemma to be well-known to experts, however since we did not find a reference we include it here. We are indebted to Maciej Zdanowicz for helping us with its proof. Lemma 4.4 Let X be a smooth surface and D = D be a reduced simple normal i =1 crossing divisor. Suppose that (X , D) lifts to W (k).Let x ∈ X be a closed point. If π : Y → −1 X is the blow up of X at x then (Y , Exc(π ) + π D) lifts to W (k). Proof Let (X , D ) be a lift of (X , D) to W (k).Let x ∈ X be a closed point. By formal smoothness there exists a lift x ∈ X of x. We claim the following: Claim If J ⊂{1, 2,..., n} is an index set such that x ∈ D but x ∈ / D for k ∈ / J , i k i ∈J then, there exists a lift x ∈ X , such that x ∈ D ,but x ∈ / D for k ∈ / J . i ∈J i k 123 E. Arvidsson Proof If x ∈ X − Supp(D) we can choose a lift x ∈ X − Supp(D ) and hence any such lift x satisfies the claim. Suppose that x ∈ Supp(D).By[8, Lemma 8.13 d)], we may assume that there is an open subset U containing x and a diagram with all squares being Cartesian: UU 2 2 A A k W (k) Spec(k) Spec(W (k)) 2 2 where A = Spec(k[t , t ]), A = Spec(W (k)[t , t ]), U and U are open subsets of 1 2 2 1 2 k W (k) X and X respectively and the morphisms e and e are étale. By [8, Lemma 8.14 e)], we may assume that we have chosen local parameters: φ = e (t ) ∈ O and φ = e (t ) ∈ O for i = 1, 2, i i U i U such that x = V((φ ,φ )) and such that (upon relabeling the components of D) there exists 1 2 a subset I of {1, 2}, such that D ∩ U = V( φ ) and D ∩ U = V( φ ).Let i ∈I i ∈I x = V((φ ,φ )),then x satisfies the claim. 1 2 Let π : Y → X be the blow up of x ∈ X ,where x is chosen as in Claim 4.We −1 −1 want to prove that (Y ,π D + Exc(π )) is a lift of (Y , Exc(π ) + π D) to W (k).If ∗ ∗ −1 x ∈ / Supp(D), for any choice of x ∈ / Supp(D ) it is clear that (Y ,π D + Exc(π )) −1 is a lift of (Y , Exc(π ) + π D) to W (k). We therefore assume that x = V((φ ,φ )), 2 1 2 x = V((φ ,φ )) and that U , U , D and D are as before. By the commutativity of blowing 1 2 up with flat base change [21, Lemma 085S] there exists, for any ring S and an étale morphism T → A , a natural Cartesian diagram: ∗ ∗ Bl (T ) Bl (A ) V ((e (t ),e (t ))) V ((t ,t )) 1 2 1 2 T A 2 ∗ where A = Spec(S[t , t ]). The strict transform of V( e (t )) is the base change to T 1 2 i i ∈I of the strict transform of V( t ) for any subset I ⊂{1, 2}. Using the notation introduced i ∈I in the beginning or this proof we therefore have: Bl  (U ) = Bl (A ) × U x V(t ,t ) 1 2 W (k) A W (k) and Bl (U ) = Bl (A ) × U . x V(t ,t ) 1 2 k A Since the base change of an étale morphism is étale, Bl  (U ) → Bl A is étale. x V (t ,t ) 1 2 W (k) Consequently, we are reduced to the following situation: 123 On the Kodaira vanishing theorem for log... 2 2 Bl A Bl A V(t ,t ) V (t ,t ) 1 2 1 2 k W (k) 2 2 A A k W (k) Spec(k) Spec(W (k)) The top square, and therefore the whole diagram, is clearly Cartesian. In this situation it is evident that Bl A is smooth over W (k) and that the strict transform of each of the V(t ,t ) 2 1 2 W (k) coordinate axes, as well as the exceptional divisor of the blow up, are flat and hence smooth over W (k) [8, Lemma 8.13]. Since the base change of an étale morphism is étale this shows −1 that Y as well as the support of π D ,for 1 ≤ i ≤ n,and Exc(π ) are all flat over W (k). ∗ i −1 −1 Consequently (Y ,π D + Exc(π )) is a lift of (Y , Exc(π ) + π D) to W (k). ∗ ∗ Theorem 4.5 (Theorem D) Let X be a log terminal projective surface with K not pseudo- effective, over a perfect field k of characteristic p > 5. Then Kodaira vanishing holds on X, i.e., for all ample Z-divisors D on X we have H (X , O (D + K )) = 0, for all i > 0. X X Proof By base change to an algebraic closure of k we can assume that k is algebraically closed. Let D be an ample Q-Cartier Z-divisor. It is sufficient to prove that 0 = H (X , O (−D)) = H (X , O (K + D)).Werun a K -MMP g : X → Y,where Y is either a klt del Pezzo X X X surface of Picard rank one or Y admits a Mori fiber space structure f : Y → B onto a curve B. (1) If f : Y → B is a K -MFS onto a curve, then the result follows from Proposition 4.1. (2) If there exists a birational morphism X → Y where Y is a log del Pezzo surface of Picard rank one then by Theorem 4.3 there exists a log resolution μ : (V , Exc(μ)) → Y such that (V , Exc(μ)) lifts to characteristic zero over a smooth base where Exc(μ) denotes the reduced divisor supported at the exceptional locus of μ. There is an induced rational map μ : V X.Let π : V → V be a resolution of the indeterminacy locus of μ . The morphism π : V → V can be factorised as a composition of blowups at points. By −1 Lemma 4.4, (V ,π Exc(μ) + Exc(π )) lifts to W (k). The diagram below illustrates the situation: −1 X (V ,π Exc(μ) + Exc(π )) μ ˜ Y (V , Exc(μ)) −1 The support of Exc(μ) ˜ is contained in the support of π Exc(μ) + Exc(π ). Therefore μ ˜ : (V , Exc(μ)) ˜ → X is a log-resolution that lifts to W (k).By[6, Lemma 6.1] the result follows. Acknowledgements I would like to thank my PhD-advisor Zsolt Patakfalvi for the help he has given me during this work. I would like to thank Jakub Witaszek for answering questions about [6]. I also thank Fabio Bernasconi and Maciej Zdanowicz for very useful discussions and for reading earlier drafts of this paper. Funding Open Access funding provided by EPFL Lausanne. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, 123 E. Arvidsson and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. References 1. Arvidsson, E., Bernasconi, F., Lacini, J.: On the Kawamata–Viehweg vanishing for log del Pezzo surfaces in positive characteristic (2020). arXiv:2006.03571 2. Bernasconi, F.: Kawamata–Viehweg vanishing fails for log del Pezzo surfaces in characteristic 3 (2017). arXiv:1709.09238 3. Bernasconi, F.: On the base point free theorem for klt threefolds in large characteristic (2019). arXiv:1907.10396 4. Cascini, P., Tanaka, H.: Smooth rational surfaces violating Kawamata–Viehweg vanishing. Eur. J. Math. 4(1), 162–176 (2018) 5. Cascini, P., Tanaka, H.: Purely log terminal threefolds with non-normal centres in characteristic two. Am. J. Math. 141(4), 941–979 (2019) 6. Cascini, P., Tanaka, H., Witaszek, J.: On log del Pezzo surfaces in large characteristic. Compos. Math. 153(4), 820–850 (2017) 7. Ekedahl, T.: Canonical models of surfaces of general type in positive characteristic. Inst. Hautes Études Sci. Publ. Math. 67, 97–144 (1988) 8. Esnault, H., Viehweg, E.: Lectures on Vanishing Theorems. DMV Seminar, vol. 20. Birkhäuser Verlag, Basel (1992) 9. Fujita, T.: Vanishing theorems for semipositive line bundles. In: Algebraic Geometry (Tokyo/Kyoto, 1982), Lecture Notes in Math., vol. 1016, pp. 519–528. Springer, Berlin (1983) 10. Hartshorne, R.: Stable reflexive sheaves. Math. Ann. 254(2), 121–176 (1980) 11. Hartshorne, R.: Generalized divisors on Gorenstein schemes. In: Proceedings of Conference on Algebraic Geometry and Ring Theory in honor of Michael Artin, Part III (Antwerp, 1992), vol. 8, pp. 287–339 (1994) 12. Hacon, C., Witaszek, J.: On the rationality of Kawamata log terminal singularities in positive characteristic. Algebr. Geom. 6(5), 516–529 (2019) 13. Jiang, C.: Bounding the volumes of singular weak log del Pezzo surfaces. Int. J. Math. 24(13), 1350110 (2013) 14. Kawakami, T.: On Kawamata–Viehweg type vanishing for three dimensional Mori fiber spaces in positive characteristic (2020). arXiv:2004.07125 15. Kollár, J., Mori, S.: Birational Geometry of Algebraic Varieties, Cambridge Tracts in Mathematics, vol. 134. Cambridge University Press, Cambridge (1998) (With the collaboration of C. H. Clemens and A. Corti, Translated from the 1998 Japanese original) 16. Kollár, J.: Rational Curves on Algebraic Varieties, Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics, vol. 32 [Results in Mathematics and Related Areas. 3rd Series. A Series of Modern Surveys in Mathematics]. Springer, Berlin (1996) 17. Lacini, J.: On rank one log del Pezzo surfaces in characteristic different from two and three (2020). arXiv:2005.14544 18. Milne, J.S.: Étale Cohomology. Princeton Mathematical Series, vol. 33. Princeton University Press, Princeton (1980) 19. Patakfalvi, Z., Waldron, J.: Singularities of General Fibers and the LMMP. arXiv:1708.04268 20. Raynaud, M.: Contre-exemple au “vanishing theorem” en caractéristique p>0. In: C. P. Ramanujam—a tribute, Tata Inst. Fund. Res. Studies in Math., vol. 8, pp. 273–278. Springer, Berlin (1978) 21. The Stacks project authors.: The stacks project (2020). https://stacks.math.columbia.edu 22. Tanaka, H.: Minimal models and abundance for positive characteristic log surfaces. Nagoya Math. J. 216, 1–70 (2014) 23. Tanaka, H.: Minimal model program for excellent surfaces. Ann. Inst. Fourier (Grenoble) 68(1), 345–376 (2018) 24. Witaszek, J.: Effective bounds on singular surfaces in positive characteristic. Mich. Math. J. 66(2), 367– 388 (2017) Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Journal

Mathematische ZeitschriftSpringer Journals

Published: May 6, 2021

There are no references for this article.