A regular affine ideal generator gives an exact frame of its ideal module
ProvedAlgebraicGeometry.Scheme.IdealSheafData.exists_frame_of_regular_affine_generatoralgebraic-geometrycartier-divisorsinvertible-sheavesphilippon-multiplicity
Let be a scheme, let be a quasi-coherent ideal, and let be an affine open. Suppose that is a non-zero-divisor and generates the ideal . Let be the registered ideal module, defined as the kernel of , with its inclusion . There is a section with which freely generates on every open subset of : for every , the map
is bijective. No nonemptiness, integrality, reducedness or Noetherian hypothesis is imposed on or .
This gives the local trivialization of an effective Cartier ideal with its actual defining equation as the image of the frame. In nodal modifications it identifies sections of a single global ideal module with the coefficient functions used in the local projective embeddings.
Preamble
import Mathlib import Definitions.Def_AlgebraicGeometry_IdealSheafModule import Definitions.Def_AlgebraicGeometry_ModulesSectionsTensor set_option autoImplicit false set_option backward.isDefEq.respectTransparency.types false open AlgebraicGeometry CategoryTheory universe u
Formal statement
theorem AlgebraicGeometry.Scheme.IdealSheafData.exists_frame_of_regular_affine_generator
{X : Scheme.{u}} (I : X.IdealSheafData) (W : X.affineOpens) (g : Γ(X, W))
(hg : g ∈ nonZeroDivisors Γ(X, W)) (hI : I.ideal W = Ideal.span {g}) :
∃ m : Γ(I.module, W), I.moduleι.app W m = g ∧ Scheme.Modules.IsFrameOn m W := by sorrySource
Derived local frame formulation of The Stacks Project, Lemma 31.14.2, tag 01WS, proof of (2) => (1), https://stacks.math.columbia.edu/tag/01WS , together with regularity on all opens as in Section 31.15, Lemma 31.15.7, https://stacks.math.columbia.edu/tag/0C4S . The formal statement uses the registered kernel realization of the ideal module and records the exact image of its frame. This is the local regular-generator argument used in the nodal Cartier-center step for the Philippon alteration frontier; it does not assert the full alteration theorem.