Affine Schemes

The triple (Spec R, O, R) — assembling topology, sheaf, and ring into a geometric object

Affine Schemes

An affine scheme is the triple (Spec(R), 𝒪Spec(R), R) — a topological space with a structure sheaf, built from a single commutative ring R. The remarkable theorem is that this construction gives an equivalence of categories between affine schemes and commutative rings (with arrows reversed). Every ring is a geometric space, and every morphism of rings is a geometric map. This is the heart of Grothendieck's revolution: algebra and geometry become two languages for the same mathematics.

A striking consequence is that affine schemes can carry nilpotentelements — things like ε with ε² = 0. The "fat point" Spec(k[ε]/(ε²)) has only one topological point but remembers a tangent direction. Classical varieties, which are always reduced (no nilpotents), cannot see this infinitesimal structure. Scheme theory can, and that extra information is essential for intersection theory, deformation theory, and moduli problems.

Interactive: Fat Point Visualizer

Compare a reduced point Spec(k) with the fat point Spec(k[e]/(e^2)) on a curve — see how nilpotent structure remembers a tangent direction invisible to classical geometry.
Why fat points matter: A morphism from Spec(k[ε]/(ε²)) to a variety Y is exactly a point of Y together with a tangent vector at that point. The dual numbers k[ε]/(ε²) encode first-order infinitesimal data — the "derivative" of a map — in a purely algebraic way. Classical varieties can't see this; schemes can.

The Ring–Scheme Dictionary

The functor Spec establishes a contravariant equivalence: ring homomorphisms R → S correspond to scheme morphisms Spec(S) → Spec(R). Surjections R ↠ R/I correspond to closed immersions V(I) ↪ Spec(R). Localizations R → Rf correspond to open immersions D(f) ↪ Spec(R). The entire geometric structure — subschemes, intersections, fibers — is encoded in ring-theoretic operations.

Global sections of the structure sheaf recover the ring: 𝒪X(Spec(R)) = R. This means the ring R is both the input to the construction and the output of taking global functions. No information is lost in the passage from algebra to geometry and back.

Nilpotents, Fat Points, and Infinitesimal Geometry

Consider the ring k[ε]/(ε²). It has one prime ideal (ε), so Spec(k[ε]/(ε²)) is a single point — but it is not the same as Spec(k). The nilpotent element ε remembers a tangent vector. A map from this "fat point" into a scheme X is the same data as a point of X plus a tangent vector at that point. This is the scheme-theoretic definition of the tangent space.

A scheme is reduced if its structure sheaf has no nilpotent sections — equivalently, if every stalk R𝔭 has no nilpotents. The reduced scheme Xred associated to X has the same topological space but "kills" all nilpotents. Classical varieties are always reduced, so scheme theory strictly generalizes the classical setting by allowing nilpotent (infinitesimal) structure.

Key Takeaways

  • Every ring IS a geometric space: the functor Spec gives a contravariant equivalence between commutative rings and affine schemes.
  • Fat point = point + tangent direction: Spec(k[ε]/(ε²)) has one point but carries infinitesimal data invisible to classical geometry.
  • Nilpotents encode infinitesimal geometry: they appear naturally in intersections, fibers, and deformation theory, recording multiplicity and tangency.
  • Reduced means no nilpotents: a reduced scheme is the scheme-theoretic analog of a classical variety; Xred strips away all infinitesimal fuzz.