1 Proj in Algebraic Geometry

1.1 Motivation from projective geometry

Projective space is obtained by adding “points at infinity” to affine space. In algebraic geometry, this geometric idea is encoded by replacing coordinate rings by graded rings and then extracting a scheme-like object that reflects homogeneous (scaling-invariant) information. The Proj construction plays the role of turning such graded algebra into a projective scheme, so that homogeneous polynomials define global geometric loci in a coordinate-free way.

1.2 Relationship to Spec and affines

If a ring describes an affine scheme, then Spec converts algebraic data into geometric data via prime ideals. Proj is an analogous machine but suited to homogeneous coordinate rings. While Spec uses all prime ideals, Proj restricts attention to homogeneous prime ideals compatible with the grading, and it discards primes that contain the irrelevant ideal. This restriction is what allows Proj to behave like a projective, rather than affine, object.

1.3 Graded rings and geometric points

A graded ring \(S=\bigoplus_{n\ge 0} S_n\) organizes functions by “degree.” Geometric points of \(\mathrm{Proj}(S)\) correspond to homogeneous prime ideals that respect the grading’s structure. Intuitively, the construction identifies points by the vanishing of homogeneous elements, so that scaling of coordinates does not produce new information, matching the projective viewpoint.

2 Construction of Proj

2.1 Graded rings: basic setup

2.1.1 Homogeneous elements and ideals

A homogeneous element lies entirely in some graded piece \(S_n\). A homogeneous ideal is generated by homogeneous elements; equivalently, it is an ideal stable under taking homogeneous components. For Proj, homogeneous prime ideals are central: they are prime ideals that are also homogeneous, meaning they respect the grading.

2.1.2 The irrelevant ideal

Let \(S_+ := \bigoplus_{n>0} S_n\). This ideal collects all strictly positive-degree information. It is called the irrelevant ideal because, in the projective setting, it corresponds to data that should not affect the resulting geometric object (much like excluding the “origin” in homogeneous coordinates). Proj is formed by excluding homogeneous prime ideals that contain \(S_+\).

2.2 Homogeneous prime ideals used in Proj

The underlying set of \(\mathrm{Proj}(S)\) consists of homogeneous prime ideals \(\mathfrak p \subset S\) such that \(S_+\not\subset \mathfrak p\). Equivalently, some positive-degree elements survive modulo \(\mathfrak p\). This condition ensures that the point corresponds to a genuine projective location rather than a purely graded artifact.

2.3 Definition of \(\mathrm{Proj}(S)\)

With the set described above, \(\mathrm{Proj}(S)\) is defined by equipping these primes with a topology and a structure sheaf that recover geometric information. In this way, Proj becomes a scheme (typically a projective scheme when \(S\) is finitely generated and the grading is appropriate). The construction is designed so that standard opens correspond to localizing by homogeneous elements.

2.4 The topology on \(\mathrm{Proj}(S)\)

For a homogeneous element \(f\in S\), one defines the basic open set \[ D_+(f)=\{\mathfrak p\in \mathrm{Proj}(S)\mid f\notin \mathfrak p\}. \] Sets of this form form a basis for the topology. These opens reflect the fact that, on the projective chart where \(f\) does not vanish, homogeneous expressions can be normalized by dividing by powers of \(f\).

3 Affine Charts and Basic Opens

3.1 Standard opens \(D_+(f)\)

The open \(D_+(f)\) is the projective analogue of the affine open \(D(f)\subset \mathrm{Spec}(A)\). Here the grading is taken into account: only homogeneous \(f\) are used, ensuring compatibility with the projective scaling encoded in the grading.

3.2 Covering \(\mathrm{Proj}(S)\) by standard opens

If \(S\) is generated by degree-1 elements (or more generally if \(S_+\) is covered by the vanishing complements of finitely many homogeneous elements), then \(\mathrm{Proj}(S)\) is covered by finitely many standard opens \(D_+(f_i)\). Even without finite generation, the family of all \(D_+(f)\) for homogeneous \(f\) with \(S_+\not\subset \sqrt{(f)}\) provides a cover.

3.3 Structure sheaf on \(D_+(f)\)

On \(D_+(f)\), the structure sheaf is defined so that functions correspond to graded pieces localized by \(f\). A standard choice is to relate \(D_+(f)\) to an affine scheme built from the degree-zero part of a localization. This produces a sheaf of rings that agrees on overlaps, allowing gluing of local data into a global scheme structure.

3.4 Computing local rings via localization

For a homogeneous \(f\), the localization \(S_f\) inherits a grading, and one takes the degree-zero part to form an affine algebra controlling \(D_+(f)\). The local ring at a point represented by a homogeneous prime \(\mathfrak p\) is obtained by localizing this affine algebra further at the corresponding prime in the degree-zero piece. This parallels how local rings in \(\mathrm{Spec}\) arise from localization at primes.

4 Sheaves and Twisting Modules

4.1 The graded module \(\leftrightarrow\) quasi-coherent sheaf correspondence

Proj comes with a natural dictionary between graded modules over \(S\) and quasi-coherent sheaves on \(\mathrm{Proj}(S)\). Because the construction ignores the irrelevant ideal, not every graded module corresponds faithfully to a distinct sheaf; modules whose elements are annihilated “near the irrelevant locus” yield the same sheaf. This leads to a quotient-category viewpoint: one modds out by graded modules that are irrelevant in the projective sense.

4.2 The sheaves \(\widetilde{M}\) for graded modules \(M\)

Given a graded \(S\)-module \(M=\bigoplus_{n\in\mathbb Z} M_n\), one constructs a sheaf \(\widetilde{M}\) on \(\mathrm{Proj}(S)\) by describing it on each \(D_+(f)\) in terms of the localized module and then gluing. On standard opens, sections correspond to degree-zero elements of the localized graded module, reflecting the projective normalization by \(f\).

4.3 Twisting sheaves \(\mathcal{O}_{\mathrm{Proj}(S)}(n)\)

Twisting adjusts degrees. The invertible sheaf \(\mathcal{O}_{\mathrm{Proj}(S)}(n)\) is defined so that its sections correspond to shifting the grading by \(n\). Concretely, it is associated to the graded module \(S(n)\) where \((S(n))_m=S_{m+n}\). These sheaves form the basic toolkit for describing embeddings, divisors, and polynomial functions on projective space.

4.4 Cohomology and global sections (overview)

Cohomology of twisted sheaves measures how local descriptions of sections fail to globalize. In favorable cases (e.g., projective space over a field with standard gradings), cohomology groups can be computed explicitly and relate directly to homogeneous components of graded modules. More generally, global sections of \(\mathcal{O}(n)\) correspond to degree-\(n\) homogeneous elements, subject to the same irrelevant-ideal equivalences.

5 Morphisms Induced by Graded Maps

5.1 Maps of graded rings and induced morphisms of Proj

A homomorphism of graded rings can induce a morphism between Proj schemes when it respects grading and the projective relevance condition. Typically, a graded ring map \(S\to T\) (compatible with degree structures) yields a continuous map on homogeneous primes via extension and contraction, and the sheaf maps follow from localized graded module behavior.

5.2 Functoriality and compatibility with sheaves

Proj is functorial: compatible graded maps lead to compatible morphisms of schemes, and quasi-coherent sheaves transform accordingly. Under these morphisms, the correspondence \(M\mapsto \widetilde{M}\) behaves well with respect to pullback and pushforward in the appropriate quasi-coherent contexts, provided the morphism is defined in the standard scheme-theoretic way.

5.3 Closed immersions and homogeneous ideals

Closed subschemes in projective geometry correspond to homogeneous ideals. If \(I\subset S\) is a homogeneous ideal that defines an appropriate quotient, then \(\mathrm{Proj}(S/I)\) embeds as a closed subscheme of \(\mathrm{Proj}(S)\). The grading ensures that the ideal’s generators define equations in projective coordinates, and the irrelevant ideal condition ensures that the subscheme is not artificially concentrated at the excluded locus.

5.4 Rational maps vs morphisms (projective viewpoint)

Projective constructions naturally produce rational maps from graded data. A collection of homogeneous elements that generate an ideal away from a base locus can define a map where it is not everywhere defined; the Proj formalism tracks this by working on standard opens and examining where denominators vanish. When the base locus is empty (in the projective sense), the rational map extends to a genuine morphism.

6 Examples

6.1 \(\mathrm{Proj}(k[x_0,\dots,x_n])\) and projective space

Let \(S=k[x_0,\dots,x_n]\) with the standard grading by total degree. Then \(\mathrm{Proj}(S)\) is projective \(n\)-space \(\mathbb P^n_k\). In this case, standard opens correspond to sets where one coordinate does not vanish, and the structure sheaf recovers the usual geometric functions on projective space.

6.2 Hypersurfaces via homogeneous polynomials

A hypersurface in \(\mathbb P^n\) is defined by a single homogeneous polynomial \(f\in k[x_0,\dots,x_n]\). The corresponding closed subscheme is \(\mathrm{Proj}(S/(f))\). Different choices of \(f\) differing by scalar multiples define the same vanishing locus, reflecting the projective scaling inherent in homogeneous coordinates.

6.3 Veronese and Segre-type constructions (graded-algebra perspective)

Many projective transformations arise from manipulating gradings. The Veronese construction replaces \(S\) by a graded subring or regraded version to embed \(\mathrm{Proj}(S)\) using higher-degree homogeneous forms. The Segre-type constructions can be seen by forming graded tensor products and then applying Proj, yielding embeddings of products of projective varieties into larger projective spaces.

7 Key Properties

7.1 Projective schemes and properness (conceptual)

When \(S\) is a finitely generated graded ring over a base and the grading is suitably positive, \(\mathrm{Proj}(S)\) becomes a projective scheme over the base. Projective schemes have strong geometric finiteness properties, including properness over the base, which ensures controlled behavior of families of points and limits.

7.2 Irreducibility and connectedness criteria (graded viewpoint)

Irreducible components and connectedness can be studied via the graded ring’s structure. For instance, homogeneous prime ideals correspond to points, and the way \(S_+\) interacts with primes influences which components appear. Under standard finiteness hypotheses, properties such as irreducibility or connectedness can often be read from algebraic conditions on \(S\) and its graded modules.

The dimension of \(\mathrm{Proj}(S)\) is related to the dimension of \(S\) as an algebra, with a shift accounting for the grading. In many common settings, \(\dim \mathrm{Proj}(S)\) equals \(\dim S -1\), reflecting the fact that projective geometry removes one degree of freedom compared to affine cones.

7.4 Behavior under base change (high-level)

If the graded ring is constructed over a base ring, then changing the base (tensoring with another ring) can affect Proj in a controlled manner. Under suitable conditions (e.g., flatness assumptions or finite presentation), Proj commutes with base change in the sense that the new projective scheme can be obtained by applying Proj to the base-changed graded algebra, matching geometric intuition about “pulling back” projective data.

8.1 Relative Proj for graded \( \mathcal{O}_X \)-algebras

Relative Proj generalizes Proj to a setting where the graded ring varies over a base scheme \(X\). Instead of a single graded \(S\), one considers a graded \(\mathcal{O}_X\)-algebra that is quasi-coherent and satisfies finiteness conditions fiberwise. The resulting relative scheme parameterizes projective directions in each fiber, enabling construction of projective bundles and families of subschemes.

8.2 Proj vs Proj of a subring (graded truncations)

Regrading operations can change the embedding while keeping the underlying projective geometry closely related. Taking subrings or truncating degrees can yield Veronese-type relationships: while the projective scheme may remain isomorphic, the associated line bundle and the map to projective space can differ. These equivalences are frequently formalized by comparing how standard opens and twisting sheaves correspond.

8.3 Connections to Rees algebras (blow-up context)

Rees algebras encode filtrations of ideals and naturally lead to blow-ups. In this context, Proj appears by forming \(\mathrm{Proj}\) of a Rees algebra, producing the geometric space that separates directions at a center. This connects Proj to classical constructions of transforming schemes to resolve singularities or improve intersection behavior, with the grading capturing the order of vanishing along the chosen ideal.