Grothendieck’s proof of Hirzebruch-Riemann-Roch theorem
Giacomo Graziani · 2025
Read Grothendieck’s proof of Hirzebruch-Riemann-Roch theorem 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-2511-06880.txt.
Introduction
Giacomo Graziani presents a self-contained guide to one of algebraic geometry's cornerstone results, the Hirzebruch-Riemann-Roch theorem, following Grothendieck's proof. The theorem connects algebraic data from sheaf cohomology with geometric data from intersection theory. Graziani starts from the classic versions for curves and surfaces before building the modern machinery needed for the general statement.
Two tools carry the proof: the Grothendieck group, which sorts sheaves so the Euler characteristic becomes natural, and the Chow ring, which organizes cycles for intersection theory. The Chern character and Todd class bridge these two worlds, and combining them yields the full Hirzebruch-Riemann-Roch formula. Graziani closes by showing the general formula reproduces the familiar classic theorems for curves and surfaces, making the survey usable as an introduction.
Catalog record
| CREATOR | Giacomo Graziani |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 80 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-06880
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Grothendieck’s proof of Hirzebruch-Riemann-Roch theorem Giacomo Graziani Abstract The Riemann-Roch Theorem is one of the cornerstones of algebraic geometry, connecting algebraic data (sheaf cohomology) with geometric ones (intersection theory). This survey paper provides a self-contained introduction and a complete proof of the Hirzebruch-Riemann-Roch (HRR) Theorem for smooth projective varieties over an algebraically closed field. Starting from the classical formulations for curves and surfaces, we introduce the two modern tools necessary for the generalization: the Grothendieck group K 0 ( X ) K_{0}(X) as the natural setting for the Euler characteristic, and the Chow ring A ∙ ( X ) A_{\bullet}(X) as the setting for cycles and intersection theory. We then construct the fundamental bridge between these two worlds—the Chern character ( ch \mathrm{ch} ) and the Todd class ( td \mathrm{td} ) —culminating in a full proof of the HRR formula: χ ( X , ℰ ) = ∫ X ch ( ℰ ) ⋅ td ( X ) \chi(X,\mathcal{E})=\int_{X}\mathrm{ch}(\mathcal{E})\cdot\mathrm{td}(X) We conclude by showing how this general formula recovers the classical theorems for curves and surfaces. 1 Introduction Given an algebraic variety X , we can adopt (at least) two perspectives in its study: the first one stems from the algebraic world, namely Morita’s theory, homological and commutative algebra which allow us to identify X with its category of coherent \mathcal{O}_{X} -modules. On the other hand of course we have the geometry of X which is usually phrased in terms of property of particular embeddings (e.g. in \mathbb{P}^{n} ) or, and this is intrinsic in our variety, in terms of its subvarieties and their mutual relations, that is, in its intersection theory. The real interest in algebraic geometry comes from the interplay between these points of view and the Hirzebruch-Riemann-Roch theorem is one bright example of such a relation. The quest for this kind of relations has been ultimately started by the work of Bernhard Riemann and Gustav Roch to address the problem of computing the dimension of the space of meromorphic functions on a compact Riemann surface with prescribed zeros and poles. On the one side one finds that, given a divisor D 1 1 1 At this point, a divisor may be thought of as a compact way of describing the zero (and polar) locus we’re interested in. Namely here D=\sum_{\mbox{finite}}n_{i}P_{i} is a formal sum where P_{i} are points and n_{i}\in\mathbb{Z} . In this case we simply define \deg\left(D\right)=\sum n_{i} . , the relation is l\left(D\right)-l\left(K-D\right)=\deg\left(D\right)+1-g, where l\left(D\right) is the dimension we were looking for, K is a canonically defined divisor and g is the genus (a topological invariant). Phrased in more modern terms, one can associate to every coherent \mathcal{O}_{X} -module \mathcal{F} a sequence of finite dimensional vector spaces H^{i}\left(X,\mathcal{F}\right) and, under this perspective, the integer \ell\left(D\right) is just the dimension of H^{0}\left(X,\mathcal{O}_{X}\left(D\right)\right) . Working harder one can show that \ell\left(K-D\right)=\dim_{\Bbbk}H^{1}\left(X,\mathcal{O}_{X}\left(D\right)\right), this is a particular case of the Serre duality 2 2 2 For a proof of Riemann-Roch theorem in these terms, you can look at [ 7 , Theorem IV.1.3, pag. 295] . At this point one is lead to guess how a generalization to higher dimensions looks like: the obvious generalization of the left hand side is given by the (finite) sum \chi\left(X,\mathcal{O}_{X}\left(D\right)\right)=\sum_{i\geq 0}\left(-1\right)^{i}\dim_{\Bbbk}H^{i}\left(X,\mathcal{O}_{X}\left(D\right)\right), this has the nice property of being additive on bounded exact sequences. The tough part is to understand the right hand side. The two-dimensional case, that is for smooth surfaces, was worked out essentially by Guido Castelnuovo and can be written in cohomological terms as 3 3 3 For a proof, see [ 7 , Theorem V.1.6, pag. 362] . \chi\left(S,\mathcal{O}_{S}\left(D\right)\right)=\frac{1}{2}D\cdot\left(D-K\right)+\chi\left(S,\mathcal{O}_{S}\right) where the \cdot denotes the intersection product between divisors. This can be seen as an evidence that the right hand side has to do with the intersection theory of our variety, and this will be addressed in Section 3 . The main ideas are essentially two: in the first part we introduce the Grothendieck group of an exact category, this will be a universal group for “functions” which are additive on short exact sequences, therefore it is the right source both for \chi\left(X,\bullet\right) and for the so-called Chern character \mathrm{ch}\left(\bullet\right) , on the other hand one can associate to X a ring A_{\bullet}\left(X\right) made up of classes of subvarieties modulo a suitable equivalence relation and whose product is given essentially by intersecting such varieties. One can wonder whether this can be further generalized, and the answer is yes, but at the cost of another strong idea, which is ultimately one of the cornerstones of Grothendieck’s approach to geometry: the focus of the theorem should be a morphism between two schemes rather than a scheme itself. This leads to the general Grothendieck-Riemann-Roch theorem and it is essentially the content of SGA6. What we actually prove here is the so-called Hirzebruch-Riemann-Roch theorem, although this name usually refers to the complex case, where also transcendental methods are applied in the course of the proof. Notation . Let us fix some notation and terminology: \Bbbk will denote an algebraically closed fields of arbitrary characteristic 4 4 4 Even if much of the work done here can be adapted with little effort to arbitrary fields and all geometric objects will be defined over \Bbbk . When we say that X is a variety, we mean that X is a reduced and irreducible. 2 The K -theory Summary For the reader interested primarily in the proof of the Hirzebruch-Riemann-Roch Theorem, the technical construction of K-theory can be treated as a toolbox whose main properties we now outline. To every smooth projective variety X , we associate an abelian group K_{0}(X) generated by isomorphism classes of locally free sheaves (vector bundles). Since X is smooth, any coherent sheaf admits a finite resolution by vector bundles; this fundamental result allows us to identify K_{0}(X) with the group G_{0}(X) generated by all coherent sheaves (Theorem 2.8 ). This group carries a commutative ring structure induced by the tensor product: [\mathcal{F}]\cdot[\mathcal{G}]=[\mathcal{F}\otimes\mathcal{G}] (Remark 2.10 ). A crucial feature of the theory is its functoriality: any projective morphism f:X\to Y induces a pushforward homomorphism f_{!}:K_{0}(X)\to K_{0}(Y) , defined via the alternating sum of higher direct images (Definition 2.14 ). This operation recovers classical invariants in specific cases: if f is the map to a point, f_{!}[\mathcal{F}] is the Euler characteristic \chi(X,\mathcal{F}) (Example 2.16 ); if i:Z\hookrightarrow X is a closed embedding, i_{!} simplifies to the classical direct image i_{*} (Proposition 2.17 ). Finally, explicitly computing these groups is possible: for the projective space \mathbb{P}^{n} , the ring K_{0}(\mathbb{P}^{n}) is K_{0}(\mathbb{P}^{n})\simeq\frac{\mathbb{Z}[x]}{(1-x)^{n+1}} where x corresponds to [\mathcal{O}_{\mathbb{P}^{n}}(1)] (Theorem 2.19 ). 2.1 The Grothendieck group Here we introduce the basics about the Grothendieck group of rings and schemes. Definition 2.1 . A category with exact sequences is a full additive subcategory \mathcal{P} of an abelian category \mathcal{C} which is closed under extensions, that is, given an exact sequence 0\to P^{\prime}\to P\to P^{\prime\prime}\to 0 in \mathcal{C} with P^{\prime},P^{\prime\prime} in \mathcal{P} , then also P is an object of \mathcal{P} . An exact sequence in \mathcal{P} is an exact sequence in \mathcal{C} involving only objects of \mathcal{P} . 5 5 5 One should also add some smallness condition \mathcal{P} : in [