A note on flatness of some fiber type contractions
Eleonora Anna Romano

TL;DR
This paper investigates the flatness of certain fiber type contractions in complex smooth projective varieties, exploring their relation to conic bundle structures and considering cases with mild singularities.
Contribution
It provides new insights into the flatness properties of fiber type contractions and their connection to conic bundle structures, even in the presence of mild singularities.
Findings
Flatness of some fiber type contractions is characterized.
Relation between flatness and conic bundle structures established.
Results extend to varieties with mild singularities.
Abstract
We discuss the flatness property of some fiber type contractions of complex smooth projective varieties of arbitrary dimensions. We relate the flatness of some morphisms having one-dimensional fibers with their conic bundles structures, also in the general case in which some mild singularities of the varieties are admitted.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
A note on flatness of some fiber type contractions
E. A. Romano
Istitute of Mathematics, Faculty of Informatics, Mathematics and Mechanics, University of Warsaw, ul. Banacha 2, PL-02-097 Warszawa, Poland.
Abstract.
We discuss the flatness property of some fiber type contractions of complex smooth projective varieties of arbitrary dimensions. We relate the flatness of some morphisms having one-dimensional fibers with their conic bundles structures, also in the general case in which some mild singularities of the varieties are admitted.
Key words and phrases:
flatness, fiber type contractions, conic bundles.
1991 Mathematics Subject Classification:
14E08, 14E30, 14J35, 14J45
The project has been supported by Polish National Science Center grants 2013/08/A/ST1/00804 and 2016/23/G/ST1/04828. Thanks to Cinzia Casagrande for important discussions, and to the referees for their comments, expecially concerning Remark 2.2.
1. Introduction
Let be a complex normal projective variety of arbitrary dimension . A contraction of is a surjective morpism with connected fibers, where is a complex normal projective variety. We say that is -negative (or simply -negative) if is -Cartier and it has negative intersection with each curve contracted by .
In this note we deal with the case in which such a contraction is of fiber type, namely .
A conic bundle is a -negative fiber type contraction where is smooth and whose fibers are isomorphic to plane conics; i.e. every fiber is isomorphic as a scheme to a zero locus of a non-trivial section of . We refer the reader to [10, 11, 7] for a recent account on conic bundles, see also references therein. By [4, IV.15.4.2.]) a conic bundle is a flat morphism, i.e. for every the stalk is a flat -module.
In this note we show the flatness property of some -negative fiber type contractions. On the other hand, using the flatness of the morphisms in question, we prove that they have a conic bundle structure. The starting point is the following theorem which is due to Ando (see [1, Theorem 3.1 (ii)]) and it is a generalization in higher dimension of Mori’s result in dimension (see [8, Theorem 3.5, (3.5.1)]).
Theorem 1.1**.**
Let be a complex smooth projective variety and let be a -negative contraction where every fiber is one-dimensional. Then is also smooth and is a conic bundle.
The goal of this paper is to discuss an alternative proof of the above theorem by proving the flatness of the contraction. Indeed, this property is not analyzed in [1, Theorem 3.1 (ii)] but it represents a key point to deduce the smoothness of and the conic bundle structure of . See also our motivation explained in Remark 2.4. In particular, we are going to prove the following result, from which we discuss in Remark 2.4 how one can deduce Theorem 1.1.
Theorem 1.2**.**
Let be a complex smooth projective variety, and let be a -negative contraction where every fiber is one dimensional. Then is a flat morphism.
Finally, we focus on the more general case in which a complex normal projective variety with mild singularities admits a flat fiber type -negative contraction with one-dimensional fibers. Let us recall that given such a variety , it is Gorenstein if it is Cohen-Macaulay and is a Cartier divisor. Being Cohen-Macaulay is an algebraic condition on the local rings of ; we refer the reader to [5, II, 8] and [6, VII, 17] for its definition and properties. In the singular case, our main statement is as follows.
Proposition 1.3**.**
Let be a fiber type -negative contraction, where every fiber of is one-dimensional, is a Gorenstein projective variety with log-terminal singularities, and is smooth. Then is a locally free sheaf on of rank , can be embedded into , and through this immersion the fibers of are isomorphic to plane conics.
2. Flatness and conic bundles structures
The first part of this section is devoted to prove Theorem 1.2. To this end, the main idea is to consider a birational modification of to make flat the morphism. We are going to generalize to higher dimension the strategy used by Mori to prove [8, Lemma 3.25, Lemma 3.26]. We recall that has arbitrary dimension . The following construction represents the set-up for the proof of Theorem 1.2.
Construction 1** ([8], (3.24)).**
In the setting of Theorem 1.2, let us consider the maximal open set such that is a flat morphism.
Set . Denoting by the Hilbert scheme of , since is flat, we have an injective map . Up to restricting , this map gives an isomorphism to an open subset of . For simplicity of notation, we continue to denote this restriction by .
Let be the closure of in with the reduced subscheme structure.
Let us consider the universal family , and the two natural projections and , where is birational, being an isomorphism over , and is a flat morphism by construction. We have the following diagram:
[TABLE]
Moreover, by Lemma [1, Theorem 3.1 (i)] the general fiber of is isomorphic to . Notice that is an irreducible and reduced projective variety, because is flat, is irreducible and reduced, and the general fiber of is isomorphic to .
For every , set . We know that is a subscheme of , and as a -cycle is algebraically equivalent to the general fiber of , so that for every , is contracted to a point by . Hence, for every , , where is a fiber of . By our assumption , and , so that as -cycles, and .**
In order to prove Theorem 1.2 we need to study the fibers of . To make the exposition self contained we recall the following two lemmas. In the first one all the possibilities for such fibers are listed. The second lemma is a short version of [1, Lemma 1.5]. We refer the reader also to [9, Lemma 3.1.7, Lemma 3.1.8] for detailed proofs of the following results.
Lemma 2.1** ([8], Lemma (3.25)).**
Setting as in Construction 1. Take . Then does not have embedded points, and there are three possibilities for :
- (a)
; 2. (b)
, where are distinct components, both isomorphic to , and they intersect transversally at a point; 3. (c)
* as -cycles, with .*
In case , one has that , where denotes the ideal subsheaf of defining in , and similarly for .
Remark 2.2**.**
Assume that case of Lemma 2.1 holds, then . Indeed, is an -module of rank 1 without embedded points, that is, an -invertible sheaf. Since is flat, is independent of and because for a general geometric point . Hence as claimed by .
Lemma 2.3**.**
Setting as in Construction 1. Assume that as -cycles. There are two possibilities for the conormal sheaf of in :
- (a)
; 2. (b)
.
If holds, then . Otherwise, .
Proof of Theorem 1.2.
Let us consider the set-up as in Construction 1. We prove that is an isomorphism. Since is birational and is normal, it is enough to show that is finite. Assume by contradiction that there exists an irreducible curve such that with point of .
We see that gives a one-dimensional family of subschemes of . Let us consider such a family, which contains . Then .
Indeed, since , the -cycles for pass through , and they are contracted by . Thus there exists a fiber of such that , and for every . Denote by the general point of .
If is reduced, then which is a contradiction, because so that for the subschemes are distinct.
Then is not reduced for every , and with . By Lemma 2.1, we know that . We use the following exact sequence:
[TABLE]
where denotes the ideal subsheaf of defining in , and similarly for . Assume that we are in case of Lemma 2.3, so that one has , and by Remark 2.2 we know that . Then from (1) we get the following exact sequence
[TABLE]
By the isomorphisms , it follows that up to multiplication by scalars the map of (2) is unique.
Thus the subsheaf is uniquely determined, then also is uniquely determined. This contradicts the non constancy of the family . Repeating the same argument, we get a contradiction also when Lemma 2.3 holds.
Then is an isomorphism. Since is a flat morphism and is smooth, by [4, Proposition 17.3.3 (i)] it follows that is also smooth.
Applying Lemma 2.1, we deduce that and have the same fibers, so that using [3, Proposition 1.14], we find that , and . Then is smooth and is a flat morphism. ∎
Remark 2.4**.**
In [1, Theorem 3.1 (ii)] Ando proved Theorem 1.1. He showed that the fibers of have at most two irreducible components, and he analized the three possible cases for the fibers to prove that they are isomorphic to plane conics.
To this end, when the fiber as -cycle with , in [1, pag. 356, case 3] Ando claims that . We could not understand how to deduce that , without knowing that is flat. For this reason, to get Theorem 1.1, first we need to prove Theorem 1.2. Then the proof of Theorem 1.1 runs as done in [1, Theorem 3.1 (ii)].
Now we show Proposition 1.3. This is probably well-known to experts, but we include a proof for lack of references. To this end, we start with the following easy observation.
Remark 2.5**.**
Let be a flat morphism between projective varieties and let be a coherent locally free sheaf on . Then is flat over , namely for every , the stalk is a flat -module111Notice that we consider as an -module via the natural map ..
Proof.
By [5, III, Proposition 9.2 (e)] it follows that for every , is a flat -module. Since is flat, by definition is a flat - module, hence by the property of transitivity of flat sheaves (see [5, III, Proposition 9.2 (c)]), it follows that for every , is a flat -module, hence the statement. ∎
Proof of Proposition 1.3.
We notice that is an equidimensional morphism from a Cohen-Macaulay variety to a smooth variety, so that it is flat (see for instance [6, Corollary of Theorem 23.1]). Using the same proof of [1, Lemma 3.1 (i)], one can see that the general fiber of is isomorphic to . We show that the fibers are isomorphic to plane conics. Set . We prove that is a locally free sheaf on of rank 3. Set , that by Remark 2.5 is flat over . If we denote by the fiber over , since is a flat morphism, we know that is constant. Using the same argument of [8, (3.25.1)] it is easy to check that for every .
Then .
Hence by [5, Proposition 3.1.9] we deduce that is locally free sheaf on of rank 3, so that we have a -bundle . Since every fiber is one-dimensional, using [2, Corollary 1.11.1], it follows that is -very ample on so that gives an immersion over , and . Being , is embedded as a divisor of .
Finally, we prove that the restriction of to every fiber of is one-dimensional and belongs to . Denoting by a fiber of , we have that is the intersection between and a fiber of , hence . Now, using that is a divisor and that has degree 2 on every fiber of , we get , hence our claim. ∎
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] T. Ando . On extremal rays of the higher-dimensional varieties. Invent. Math. , 81:347–357, 1985.
- 2[2] M. Andreatta and J. A. Wiśniewski . A view on contractions of higher dimensional varieties. In Proceedings of Symposia in Pure Mathematics , volume 62, pages 153–184. American Mathematical Society, 1997.
- 3[3] O. Debarre . Higher-dimensional algebraic geometry . Universitext, Springer-Verlag, New York, 2001.
- 4[4] A. Grothendieck . Éléments de géometrié algébrique: IV. étude locale des schémas et des morphismes de schémas. Publications mathématiques de l’ IHÉS , 20:5–251, 1964.
- 5[5] R. Hartshorne . Algebraic Geometry . Springer-Verlag New York, 1997.
- 6[6] H. Matsumura and M. Reid . Commutative ring theory , volume 8. Cambridge University Press, 1989.
- 7[7] P. Montero and E. A. Romano . A characterization of some Fano 4-folds through conic fibrations. ar Xiv preprint ar Xiv:1803.09129 , 2018.
- 8[8] S. Mori . Threefolds whose canonical bundles are not numerically effective. In Algebraic Threefolds , pages 155–189. Springer, 1982.
