Ruled and rational surfaces and their models

Giacomo Graziani · 2025

Read Ruled and rational surfaces and their models by Giacomo Graziani (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-06896.txt.

Introduction

Giacomo Graziani surveys the theory of ruled and rational surfaces, two classic classes in algebraic geometry. Rational surfaces are those birationally equivalent to the projective plane, and ruled surfaces are swept out by classes of lines, loosely like a cylinder or cone. A powerful idea in classifying such objects is the model: choosing, within an isomorphism class, a representative with nice features.

The survey defines and studies suitable models for these surfaces in the smooth complex case. With good models in hand, features of the surfaces themselves become easier to read off and compare. Graziani presents the constructions and the theory around them as a coherent account, giving readers a clear entry into how modern algebraic geometry tames these classic objects through their best representatives.

Catalog record

CREATORGiacomo Graziani
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE190 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-06896

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Ruled and rational surfaces and their models Giacomo Graziani Abstract One of the most powerful ideas in the study and classification of algebraic varieties is the notion of a model —that is, to single out an object, in the appropriate isomorphism class, with nice properties. This survey aims to define and study suitable models of rational and, more generally, ruled surfaces in the smooth complex case, and to use them to study such surfaces. A key consequence of Castelnuovo’s contractibility Theorem is that every surface \mathcal{S} admits a relatively minimal model : a birational morphism \mathcal{S}\to\mathcal{S}^{\prime} where \mathcal{S}^{\prime} has no \left(-1\right) -curves. We will see that the relatively minimal models of surfaces ruled over an irrational curve are exactly the \mathbb{P}^{1} -bundles, i.e., the projectivization of rank 2 vector bundles over that curve. It is also possible to introduce a numerical invariant e for such bundles, and this invariant will be shown to completely classify the indecomposable relatively minimal models over an elliptic curve. Nevertheless, the notion of a relatively minimal model turns out to be not completely satisfactory since, in general, such a model is not unique. To solve this problem, we define minimal models as surfaces with a nef canonical system; it is straightforward to see that such a model, when it exists, is unique. Dealing with minimal models requires the introduction of a new biregular invariant of a smooth surface, namely the notion of type . The existence of this invariant provides an analogue of Kodaira’s classification for smooth complex surfaces. From these results, we can deduce many important statements, such as the existence of a minimal model for every non-ruled surface and an extended description of relatively minimal models in the general ruled case. This leads to a broad characterization of ruled and rational surfaces, which contains as particular cases both Castelnuovo’s criterion for rationality and an affirmative answer to Lüroth’s problem for complex surfaces . In the final part of this survey, we use the theory we have developed to study the embeddings of ruled surfaces into projective spaces. Namely, we will show that for every pair of integers k,n with k>n\geq 0 we can embed \mathbb{F}_{n} as a rational scroll of degree 2k-n in \mathbb{P}^{2k-n+1} . We also show that for every n\geq e+3\geq 2 there exists an elliptic scroll in \mathbb{P}^{2n-e-1} of degree 2n-e and invariant e . These results, in analogy with the case of curves, suggest that the possible degrees of an embedding are related to the rationality of the surface. After showing that the degree of an irreducible non-degenerate surface in \mathbb{P}^{n} is bounded below by n-1 , we give a complete classification of smooth surfaces that can be embedded in \mathbb{P}^{n} with degree n-1 or n . As it turns out, all such surfaces are rational. The material presented in this survey follows the classical treatments of [ 1 ] and [ 4 ] , while we refer to [ 5 ] for background material. 1 Technical Tools 1.1 Ampleness and very ampleness Lemma 1.1 . Let X be an affine variety and f\in\mathbb{C}\left[X\right] . Let U_{f} be the open set defined by f\neq 0 and \mathcal{F} a quasi-coherent sheaf on X . Then 1. if s\in\Gamma\left(X,\mathcal{F}\right) is a global section such that s_{|U_{f}}=0 , then there exists an integer n>0 with f^{n}s=0 ; 2. if t\in\Gamma\left(U_{f},\mathcal{F}\right) , there exists an integer n>0 such that f^{n}t extends to a global section of \mathcal{F} . Proof. Let us note that, since \mathcal{F} is quasi-coherent on X, we can find open sets \left(U_{g_{i}}\right) with g_{i}\in\mathbb{C}\left[X\right] such that they form a \emph{finite} open cover of X and \mathcal{F}_{|U_{g_{i}}}=\tilde{M_{i}} , where M_{i} is a \mathbb{C}\left[X\right]_{g_{i}} -module. Moreover U_{g_{i}}\cap U_{g_{j}}=U_{g_{i}g_{j}} . Let s be a global section with s_{|U_{f}}=0 , we denote with s_{i} the restriction s_{|U_{g_{i}}} . For every i , since \mathcal{F}_{|U_{f}\cap U_{g_{i}}}=\tilde{\left(M_{i}\right)_{f}} , the image of s_{i} in \left(M_{i}\right)_{f} is zero, thus there exists n_{i}\in\mathbb{N} such that f^{n_{i}}s_{i}=0 . Since the open cover is finite, we can find n such that f^{n}s_{i}=0 for every i , and this proves 1) . To prove 2) , we note that, for every i , there exists t_{i}\in M_{i}=\Gamma\left(U_{g_{i}},\mathcal{F}\right) such that the image of t_{i} on U_{f}\cap U_{g_{i}}=U_{fg_{i}} is f^{n_{i}}t for some n_{i} . As in the preceding point, we can take n working for every i . On the intersection U_{g_{i}g_{j}} we have two sections t_{i} and t_{j} which agree on U_{fg_{i}g_{j}} , from point 1) there exists m such that f^{m}\left(t_{i}-t_{j}\right)=0 on U_{g_{i}g_{j}} and we can take m big enough to work for every i,j . Finally, local sections f^{m}t_{i}\in\Gamma\left(U_{g_{i}},\mathcal{F}\right) glue to give a global section whose restriction to U_{f} is f^{n+m}t . ∎ Lemma 1.2 . Let X be an algebraic variety, \mathcal{L}\in\mathrm{Pic}\left(X\right) and f\in\Gamma\left(X,\mathcal{L}\right) . We set X_{f}=\left\{x\in X\,\,|\,\,f_{x}\notin\mathfrak{m}_{x}\cdot\mathcal{L}_{x}\right\}. For every quasi-coherent sheaf \mathcal{F} on X : 1. if X is quasi-compact and s\in\Gamma\left(X,\mathcal{F}\right) is a section such that s_{|X_{f}}=0 , then there exists n>0 with f^{n}s=0 as a global section of \mathcal{F}\otimes\mathcal{L}^{n} ; 2. moreover, if \mathcal{L} is locally free over an open cover \left(U_{i}\right) such that the intersections U_{i}\cap U_{j} are quasi-compact, for every t\in\Gamma\left(X_{f},\mathcal{F}\right) , there exists an integer n>0 such that f^{n}t\in\Gamma\left(X_{f},\mathcal{F}\otimes\mathcal{L}^{n}\right) extend to a global section. Proof. It is a straightforward consequence of Lemma 1.1 . We simply remark that the hypothesis U_{i}\cap U_{j} quasi-compact is used to apply point 1) . ∎ Proposition 1.3 ( [ 4 , Exercise 5.5] ) . If f:X\to Y is a finite morphism between projective varieties and \mathcal{F} is a \emph{coherent} sheaf on X , then f_{\ast}F is a coherent sheaf on Y . In particular, if i:X\hookrightarrow Y is a closed embedding, then i_{\ast}F is coherent for every coherent \mathcal{O}_{X} -module \mathcal{F} . Proof. Let \left(U_{i}\right) be an affine cover of Y , then \left(f^{-1}\left(U_{i}\right)\right) is an affine cover of X and thus, for every i , \mathcal{F}_{|f^{-1}\left(U_{i}\right)}\simeq\tilde{M_{i}} for a finitely generated \mathbb{C}\left[f^{-1}\left(U_{i}\right)\right] -module M_{i} . Then, in view of [ 5 , Lemma 5.6.1] , we have f_{\ast}F_{|U_{i}}\simeq\left(M_{i}\otimes_{\mathbb{C}\left[U_{i}\right]}\mathbb{C}\left[f^{-1}\left(U_{i}\right)\right]\right)\tilde{}. Thus f_{\ast}F is coherent since \mathbb{C}\left[f^{-1}\left(U_{i}\right)\right] is a finitely generated \mathbb{C}\left[U_{i}\right] -module for every i . Moreover it is clear that a closed embedding is a finite morphism. ∎ We have all the tools to prove Theorem 1.4 (Serre) . Let X\subseteq\mathbb{P}^{N} be a projective variety and let \mathcal{F} be a coherent \mathcal{O}_{X} -module. Then, for n\gg 0 , the sheaf \mathcal{F}\left(n\right) is generated by global sections. Proof. Let j be the inclusion X\hookrightarrow\mathbb{P}^{N} , we have that j^{\ast}\left(\mathcal{O}\left(1\right)\right)=\mathcal{O}_{X}\left(1\right). Moreover, in view of Proposition 1.3 , j_{\ast}F is coherent and in view of projection formula we deduce \displaystyle j_{\ast}\left(\mathcal{F}\left(1\right)\right) \displaystyle= \displaystyle j_{\ast}\left(\mathcal{F}\otimes\mathcal{O}_{X}\left(1\right)\right)=j_{\ast}\left(\mathcal{F}\otimes j^{\ast}\left(\mathcal{O}\left(1\right)\right)\right)= \displaystyle= \displaystyle j_{\ast}\mathcal{F}\otimes\mathcal{O}\left(1\right)=\left(j_{\ast}\mathcal{F}\right)\left(1\right) from which we get that j_{\ast}\left(\mathcal{F}\left(n\right)\right)=\left(j_{\ast}\mathcal{F}\right)\left(n\right) for every n . Clearly \mathcal{F}\left(n\right) is generated by global sections if and only if \left(j_{\ast}\mathcal{F}\right)\left(n\right) is generated by global sections, thus we reduce to the case when X=\mathbb{P}^{N} . Let \left(U_{i}\right) be the standard affine open cover of \mathbb{P}^{N} , then \mathcal{F}_{|U_{i}}=\tilde{M_{i}} where M_{i} is a finitely generated A_{i} -module, where A_{i}=\mathbb{C}\left[\frac{x_{0}}{x_{i}},\dots,\frac{x_{N}}{x_{i}}\right] . For every i , let s_{ij} , for j=1,\dots,k_{i} be generators for M_{i} over A_{i} . In view of Lemma 1.2 there exists n such that x_{i}^{n}s_{ij} extends to a global section

Read the complete text