Documentation

TauCeti.AlgebraicGeometry.Normalization.IsIso

Recognizing the relative normalization #

Let f : X ⟶ Y be quasi-compact and quasi-separated, and factor it as f = f₁ ≫ f₂ with f₂ : T ⟶ Y integral. The universal property of the relative normalization gives a morphism f.normalizationDesc f₁ f₂ H : f.normalization ⟶ T over Y. This file characterizes when it is an isomorphism, that is, when T is the normalization of Y in X: exactly when, for every affine open U ⊆ Y, the restriction Γ(T, f₂ ⁻¹ U) → Γ(X, f ⁻¹ U) along f₁ is injective with image the integral closure of Γ(Y, U) in Γ(X, f ⁻¹ U).

The main application is to normal schemes. Let T be an integral scheme whose local rings are integrally closed, and g : T ⟶ Y an integral morphism. Then T is the normalization of Y in the function field of T, along the generic point Spec k(T) ⟶ T ⟶ Y. On a nonempty affine open V of T, the ring Γ(T, V) is an integrally closed domain with fraction field k(T) (AlgebraicGeometry.IsAffineOpen.isIntegrallyClosed), and it is integral over Γ(Y, U) when V = g ⁻¹ U, so it is the integral closure of Γ(Y, U) in k(T).

Main results #

References #

@[simp]

The preimage under f.normalizationDesc f₁ f₂ H of the preimage of U under f₂ is the preimage of U in the relative normalization.

Over an affine open U ⊆ Y, the sections of f.normalizationDesc f₁ f₂ H, read in the integral closure of Γ(Y, U) in Γ(X, f ⁻¹ U) through normalizationObjIso, are the restriction maps of f₁.

Over an affine open U ⊆ Y, the map of sections of f.normalizationDesc f₁ f₂ H is an isomorphism if and only if the restriction Γ(T, f₂ ⁻¹ U) → Γ(X, f ⁻¹ U) along f₁ is injective and its image contains every element integral over Γ(Y, U).

Recognizing the relative normalization. For f = f₁ ≫ f₂ with f₂ integral, the morphism f.normalizationDesc f₁ f₂ H : f.normalization ⟶ T is an isomorphism if and only if, for every affine open U ⊆ Y, the restriction Γ(T, f₂ ⁻¹ U) → Γ(X, f ⁻¹ U) along f₁ is injective and its image contains every element integral over Γ(Y, U). Its image is always contained in the integral closure, since f₂ is integral.

The relative normalization f.normalization ⟶ Y is an isomorphism if and only if, for every affine open U ⊆ Y, the map Γ(Y, U) → Γ(X, f ⁻¹ U) is injective and its image contains every element integral over Γ(Y, U): that is, Y is already the normalization of itself in X.

A normal scheme integral over Y is the normalization of Y in its function field. If T is an integral scheme whose local rings are integrally closed and g : T ⟶ Y is integral, then the morphism from the relative normalization of Y in Spec k(T), along Spec k(T) ⟶ T ⟶ Y, to T is an isomorphism.

The normalization of an integral scheme Y in its function field, along Spec k(Y) ⟶ Y, is an isomorphism over every affine open U whose ring of sections is integrally closed: the integral closure of Γ(Y, U) in its fraction field k(Y) is Γ(Y, U) itself.