OctoInformBeta

A counterexample to the small Cohen–Macaulay module conjecture

For a Noetherian local ring (R,𝔪)(R,\mathfrak m), a small Cohen-Macaulay module is a nonzero finitely generated RR-module MM with depth⁡RM=dim⁡R\operatorname{depth}_R M=\dim R. The domain form of the small Cohen-Macaulay module conjecture asks whether every complete Noetherian local domain admits such a module. For a smooth integral projective complex surface XX and an ample globally generated divisor HH, the section ring is T(X,H)=⨁j≥0H0(X,𝒪X(jH))T(X,H)=\bigoplus_{j\geq0}H^0(X,\mathcal O_X(jH)). The vertex is the maximal ideal T(X,H)>0T(X,H)_{>0} of positive-degree elements of T(X,H)T(X,H). The completed local ring at the vertex is R(X,H)=T(X,H)T(X,H)>0̂R(X,H)=\widehat{T(X,H)_{T(X,H)_{>0}}}. Write KXK_X for the canonical divisor of XX. Write c2(X)c_2(X) for the second Chern number of XX.

Hochster formulated the finite-module existence problem. As recalled by Hochster, every complete local domain of dimension at most two admits a small Cohen-Macaulay module. The theorem of Hartshorne, Peskine-Szpiro, and Hochster gives nonzero finitely generated graded modules of depth three over three-dimensional finitely generated nonnegatively graded domains whose degree-zero part is a perfect field of positive characteristic.

The self-intersection number of KXK_X is denoted by KX2K_X^2. The self-intersection number of HH is denoted by H2H^2. The preprint proves that KX2−2c2(X)≤6H2K_X^2-2c_2(X)\leq6H^2 whenever R(X,H)R(X,H) admits a small Cohen-Macaulay module. The preprint constructs a pair (X,H)(X,H) for which R(X,H)R(X,H) is a three-dimensional complete Noetherian normal local domain with residue field ℂ\mathbb C and no nonzero finitely generated module of depth three. For this pair, the preprint also proves that the uncompleted local ring at the vertex has no small Cohen-Macaulay module, a consequence not contained in the Lean library.

The Lean library SurfaceCones proves the ring properties in the counterexample statement. The Lean library does not prove that the completed ring has no small Cohen-Macaulay module. A nonzero finite module of depth three over the constructed ring is a hypothesis of one of the Lean library’s results. Lean has checked every result of the Lean library, so the Lean library is the reference for the statements on these pages. Every result is stated as in the Lean library and linked to its statement there. Where the preprint differs, its statement is given beside the Lean library’s statement. The preprint is a guide only for the order and names of the results.

The Lean library: SurfaceCones

SurfaceCones is a Lean library: 16 items, in seven groups. Each has a page written by OctoInform, listed below.

1. Introduction 2 items
  1. Section 1 (setting) A completed cone and the counterexample proposition.
  2. Theorem 1.1 (ring properties only) A complete normal local domain exists.
2. From a completed cone to its exceptional surface 3 items
  1. Section 2.1 (setting) Graded series and completion of the cone.
  2. Section 2.2 (setting) Defining the projective projection and completed cone morphism.
  3. Section 2.3 (setting) Finite depth-three modules admit exceptional models.
3. Balancing slopes by modifications 1 item
  1. Lemma 3.1 (elementary-transform portion only) A Cartier transform retains its puncture framing.
4. A Chern-character inequality from a divisor 4 items
  1. Section 4 (setting) A Cartier divisor defines coherent layers.
  2. Lemma 4.1 (cleared free lattices) Clearing a framing gives free lattices.
  3. Lemma 4.1 (finite pushforward rank) The divisor pushforward has the parameter rank.
  4. Lemma 4.1 (scalar layer inequalities) Bounded layer ranks satisfy quadratic inequalities.
5. The numerical obstruction 1 item
  1. Section 5 (proof of Theorem 1.2) Variance decomposes in finite weighted mixtures.
6. An explicit surface and its polarization 4 items
  1. Section 6 (setting) A surface from seventh-root sections.
  2. Proposition 6.1 (geometric portions only) A smooth surface from seventh roots.
  3. Lemma 6.2 (Kummer field degree) Five line radicals have maximal extension degree.
  4. Lemma 6.2 (smooth coefficient charts) Good coefficient charts are smooth.
7. The counterexample 1 item
  1. Section 7 (proof of Theorem 1.1) A strict comparison at exponent seven.

No item matches.