Segre Class of Schemes with Regularly Embedded Components

Guanxi Li · 2025

Read Segre Class of Schemes with Regularly Embedded Components by Guanxi Li (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-2511-06799.txt.

Introduction

Guanxi Li works in intersection theory, a branch of algebraic geometry that counts and classifies how geometric objects meet. The paper generalizes Fulton's Residual Intersection Theorem for the Segre class, an invariant that measures how a scheme sits inside its surroundings. The setting is schemes whose parts are regularly embedded, a well-behaved class of pieces.

Li expresses the Segre classes of such schemes through the Chern classes of the normal bundles of the parts and their intersections. Formulas are given for two cases: when the parts meet transversely. When the ideal sheaf of the scheme, after blowing up along one part, factors into the exceptional divisor part and the residual scheme part. The results give geometers concrete tools for computing these invariants.

Catalog record

CREATORGuanxi Li
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE43 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-06799

Browse the full library · About the Tower of Babel

Full text

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

Segre Class of Schemes with Regularly Embedded Components Guanxi Li Abstract . We generalize Fulton’s Residual Intersection Theorem for the Segre class and express the Segre classes of schemes with regularly embedded components in terms of the Chern classes of the normal bundles to the components and their intersections. More specifically, we provide formulas for the following situations: when the components of the scheme intersect transversely and when the ideal sheaf of the scheme, after the blowup along a component, is the product of the ideal sheaves of the exceptional divisor and the residual scheme. 1. Introduction Let X be a closed subscheme of a scheme Z with an ideal sheaf \mathscr{F} . Let C_{X}Z:=Spec(\sum_{n=0}^{\infty}\mathscr{F}^{n}/\mathscr{F}^{n+1}) denote the normal cone to X in Z . Consider its projective completion P(C_{X}Z\oplus 1) . Let \mathscr{O}(1) be the dual to tautological line bundle, and let q:P(C_{X}Z\oplus 1)\rightarrow X be the projection. The Segre class of X is then defined as: s(X,Z):=q_{*}(\sum_{i\geq 0}c_{1}(\mathscr{O}(1))^{i}\cap[P(C_{X}Z\oplus 1)]) [ 4 , 4.2] . The Segre class of a vector bundle E on X is defined as follows: s(E)=c(E)^{-1}, where c(\cdot) denotes the Chern class, and we write s(E)=s_{1}(E)+s_{2}(E)+s_{3}(E)\cdots , where s_{i}(E):A_{*}(X)\rightarrow A_{*-i}(X) . In the case when X is regularly embedded in Z , N_{X}Y:=C_{X}Y denotes the normal bundle to X . The Segre class is then expressed as: s(X,Z)=c(N_{X}Z)^{-1}\cap[X]=s(N_{X}Z)\cap[X]. A problem of interest is that of residual intersection. Suppose X\hookrightarrow Z is regularly embedded and V\subseteq Z is a purely k -dimensional subscheme. Additionally, let W=V\cap X. Then the intersection product is X\cdot V=\{c(N_{X}Y)\cap s(W,V)\}_{k-d}, where d=codim_{Z}(X) ( [ 4 ] , 6.1). Here, the subscript of the bracket is interpreted as taking the dimension (k-d) -term of the class. Furthermore, if U\subseteq W is a closed subscheme, the problem of residual intersection involves expressing X\cdot V as a class on U with a residual class \mathbb{R} : X\cdot V=\{c(N_{X}Y)\cap s(U,V)\}_{k-d}+\mathbb{R}. At the level of Segre classes, this problem is equivalent to computing s(W,Z) in terms of the sum of s(U,Z) and some residual terms. In the case when U is a Cartier divisor, we have Fulton’s residual intersection formula: Proposition 1.1 ( [ 4 ] Proposition 9.2) . Let D\subseteq W\subseteq Z be closed embeddings, with Z a k -dimension variety, and D a Cartier divisor. Let R be a subscheme in W such that W=D\cup R set theoretically and \mathscr{W}=\mathscr{D}\cdot\mathscr{R} , where \mathscr{W},\mathscr{D},\mathscr{R} are ideal sheaves of W,D,R . Then, s(W,Z)_{m}=s(D,Z)_{m}+\sum_{j=0}^{k-m}\binom{k-m}{j}(-D)^{j}\cdot s(R,Z)_{m+j} in A_{m}(W) . The following definition and corollary are from [ 3 ] and [ 1 ] : Definition 1.2 . Let s(X,Z)=\sum_{i}\alpha^{i} , where \alpha^{i} is the term of codimension i in X of the Segre class. Then, the \mathscr{L} -tensored Segre class of X in Z is s(X,Z)^{\mathscr{L}}=\sum_{i}\frac{\alpha^{i}}{c(\mathscr{L})^{i+1}} The expression for the Segre class as a whole is the following: Corollary 1.3 . With the above notation, we have: s(W,Z)=s(D,Z)+s(R,Z)^{\mathscr{O}(D)} In this paper, we aim to generalize Fulton’s Residual Intersection formula beyond the case of Cartier divisors. We study the case when W has irreducible components W_{i} , and we aim to express the Segre class s(W,Z) in terms of s(W_{i},Z) . We introduce the following notation: Definition 1.4 . Let Z be a scheme. Let E,F be vector bundles on Z . We define the residual product of the Segre class of vector bundles as follows: s(E)\odot s(F):=\sum_{k=0}^{\infty}\sum_{\begin{subarray}{c}i+j=k\\ i,j\geq 1\end{subarray}}\binom{k}{i}s_{i-rank(E)}(E)s_{j-rank(F)}(F), where s(E)\odot s(F):A_{*}(X)\rightarrow A_{*}(X) . Then, note that s(E)\odot s(F)=\sum_{k=0}^{\infty}\{s(E)\odot s(F)\}^{k}, where \{s(E)\odot s(F)\}^{k} is defined to be the operator from A_{*}(X) to A_{*-k}(X) . Then, \{s(E)\odot s(F)\}^{k}=\sum_{\begin{subarray}{c}i+j=k\\ i,j\geq 1\end{subarray}}\binom{k}{i}s_{i-rank(E)}(E)s_{j-rank(F)}(F). Definition 1.5 . Let Y_{1},\dots,Y_{n} be regularly embedded subschemes of Z . For all subsets S\subseteq\{1,2,\dots,n\} , let Y_{S}=\bigcap_{i}Y_{i} . We say that they intersect transversely, if for all subsets S\subseteq\{1,2,\dots,n\} , N_{Y_{S}/Z}=\bigoplus_{i}N_{Y_{i}/Z}|_{Y_{S}} . Note that the definition of trnaverse intersection that we will be using in this paper requires neither the subvarieties Y_{i} nor the scheme Z to be nonsingular. Theorem 1.6 . If X,Y are regularly embedded in the scheme Z with ideal sheaves \mathscr{I},\mathscr{J} . Furthermore if X,Y intersect transversely in Z , and let W be the scheme supported on X\cup Y defined by the ideal sheaf \mathscr{I}\cdot\mathscr{J} , then, s(W,Z)=s(X,Z)+s(Y,Z)-s(N_{X}Z)\odot s(N_{Y}Z)\cap[X\cap Y] in A_{*}(W) . The obvious push forward of inclusion is omitted. Transversality is a very strong condition and Segre classes are usually calculated through blowups. Therefore, in practice, one would like to obtain an analogous statement by replacing the transversality condition with some given properties of the scheme after blowups. Theorem 1.7 . Let X,Y be regularly embedded in the scheme Z , with X\cap Y regularly embedded in both X and Y . Let W be supported on X\cup Y , and consider the blowup \pi:Bl_{Y}Z\rightarrow Z. If \pi^{-1}(W) is supported on E\cup\tilde{X} , such that the proper transform of X , \tilde{X} , is residual to the exceptional divisor E . Then, s(W,Z)=s(X,Z)+s(Y,Z)-\frac{c(N_{X\cap Y}X)c(N_{X\cap Y}Y)}{c(N_{X\cap Y}Z)}\cdot(s(N_{X\cap Y}Y)\odot s(N_{X\cap Y}Y))\cap[X\cap Y] in A_{*}(W) . The obvious pushforwards of inclusion is omitted. 2. Proof of Theorem 1.6 We will first prove a lemma regarding the properties of our notion of transverse intersection. Lemma 2.1 . Let X,Y be regularly embedded subvarieties of the scheme Z , intersects transversely. X\cap Y is regularly embedded in X and Y . Let \pi:Bl_{Y}Z\rightarrow Z denote the blowup of Z at Y , \tilde{Z}=Bl_{Y}Z and \tilde{X} be the proper transform of X . Then: (a) N_{\tilde{X}}\tilde{Z}=\pi^{*}N_{X}Z . (b) \pi^{-1}(X)=\tilde{X} Proof. To show (a), since we have the normal sheaves \mathscr{N}_{X}Y=(\mathscr{I}_{X}/\mathscr{I}^{2}_{X})^{\vee} and \mathscr{N}_{\tilde{X}}\tilde{Z}=(\mathscr{I}_{\tilde{X}}/\mathscr{I}^{2}_{\tilde{X}})^{\vee} Since \mathscr{N}_{X}Y is locally free, then \pi^{*}(\mathscr{N}_{X}Y)\simeq(\pi^{*}(\mathscr{I}_{X}/\mathscr{I}^{2}_{X}))^{\vee} . Therefore, it is enough to show that \pi^{*}\mathscr{I}_{X}\simeq\mathscr{I}_{\tilde{X}}. Note that this would imply (b) as well. We will check the above isomorphism at the level of the local ring at the closed point. Note that, if p\notin X\cap Y then the blowup does not affect X , and \pi is the identity. Therefore \pi^{*}\mathscr{I}_{X,p}\simeq\mathscr{I}_{\tilde{X},\pi^{-1}(p)} . Hence, it is sufficient to check the closed point p in X\cap Y , and we wish to show that for all points q\in\pi^{-1}(p) , (\pi^{*}\mathscr{I}_{X})_{q}\simeq\mathscr{I}_{\tilde{X},q} . Let A be \mathscr{O}_{Z,p} , where p is a closed point in X\cap Y , Since X,Y are regularly embedded, the stalks of their ideals sheaves over the closed points can be generated by regular sequences. Based on the proof of previous lemma, we can assume that I_{X}:=\mathscr{I}_{X,p} is generated by the regular sequence (a_{1},a_{2},\cdots,a_{n}) and I_{Y}:=\mathscr{I}_{Y,p} is generated by the regular sequence (b_{1},b_{2},\cdots b_{m}) By [ 4 ] B.6.10, over any closed point q such that \pi(q)=p , the structure sheaf of the blowup evaluated at q can be written as the following, \mathscr{O}_{\tilde{Z},q}=A[x_{1},x_{2},\cdots,x_{m-1}]/(b_{1}-b_{m}x_{1},b_{2}-b_{m}x_{2},\cdots,b_{m-1}-b_{m}x_{m-1}) Now consider \tilde{X}=Bl_{X\cap Y}X . Since X\cap Y\hookrightarrow X is regular embedding, we have I_{X\cap Y}\otimes\mathscr{O}_{X,p}:=(\bar{b}_{1},\bar{b}_{2},\cdots,\bar{b}_{m}) also forms a regular sequence. \mathscr{O}_{\tilde{X},q}=\mathscr{O}_{X,p}[x_{1},x_{2},\cdots,x_{m-1}]/(\bar{b}_{1}-\bar{b}_{m}x_{1},\cdots,\bar{b}_{m-1}-\bar{b}_{m}x_{m-1}) =A[x_{1},x_{2},\cdots,x_{m-1}]/(a_{1},\cdots,a_{n},b_{1}-b_{m}x_{1},\cdots,b_{m-1}-b_{m}x_{m-1}) Then the inclusion map \tilde{X}\hookrightarrow\tilde{Z} induces surjection \mathscr{O}_{\tilde{Z},q}\rightarrow\mathscr{O}_{\tilde{X},q} , given by x_{i}\mapsto x_{i} . The kernel is exactly (a_{1},a_{2},\cdots,a_{n}) . This is exactly (\pi^{*}\mathscr{I}_{X})_{q} . Thus, we have the equality \mathscr{I}_{\tilde{X},q}\simeq(\pi^{*}\mathscr{I}_{X})_{q} . ∎ With the above lemma, we can prove 1.6 : Proof. It is equivalent to proving the following: s(W,Z)_{d}=s(X,Z)_{d}+s(Y,Z)_{d}-\sum_{i=1}^{dim(Z)-d-1}\binom{dim(Z)-d}{i}s_{dim(Y)-d-i}(N_{Y}Z)s_{dim(X)+i-dim(Z)}(N_{X}Z). By Proposition 2.2 we have the following commutative diagram: {\lx@inpgf@ignorespaces Bl_{X\cup Y}Z} {\lx@inpgf@ignorespaces Bl_{X}Z} {\lx@inpgf@ignorespaces Z} \scriptstyle{\lx@inpgf@ignorespaces\pi_{Y}} \scriptstyle{\lx@inpgf@ignorespaces\pi} \scriptstyle{\lx@inpgf@ignorespaces\pi_{X}} Let E_{X} be the exceptional divisor of

Read the complete text