Documentation

LeanPool.MassFormula.EisensteinMonogenic

Auxiliary file: Eisenstein monogenicity—integers L is generated by x #

For a root x of an Eisenstein polynomial over 𝒪[K] in SeparableClosure K, the integral closure of 𝒪[K] in K⟮x⟯ is generated by x itself ([Serre 1979, Chap. I, §6, Prop. 17][Serre1979]). The proof is the classical two-step denominator chase, powered by two Mathlib lemmas: Algebra.discr_mul_isIntegral_mem_adjoin clears an integral element into the adjoin after multiplication by the discriminant of the power basis—nonzero by separability, which is free inside SeparableClosure K—and mem_adjoin_of_smul_prime_pow_smul_of_minpoly_isEisensteinAt strips prime powers, the discriminant being a unit times a power of the uniformizer in the discrete valuation ring 𝒪[K]. This file sits below First.lean and Discriminant.lean in the import graph: the total-ramifiedness of K⟮x⟯ (isTotallyRamified_adjoin, below) and the discriminant bound n - 1 ≤ d L (sub_one_le_d_adjoin) are computations inside 𝒪[K][x] made possible by this identification.

References #

Instances Mathlib does not reach #

Two links of the tower 𝒪[K] → L → SeparableClosure K that instance search no longer finds on its own: IntermediateField.isScalarTower wants the base to act on K, and FaithfulSMul over SeparableClosure K is only reachable through IsFractionRing 𝒪[K] K, which is not a global instance. Both are needed by minpoly and isIntegral_algebraMap_iff throughout this file.

The constant coefficient of the Eisenstein minimal polynomial is an associate of the uniformizer.

A root of an Eisenstein polynomial is monogenic: the integral closure of 𝒪[K] in K⟮x⟯ is generated by x itself.

Total ramifiedness of the adjoined root field #

Three abstract commutative-algebra lemmas—kept over opaque rings so that unification stays structural—followed by the concrete theorem. The route avoids constructing any valuation on L = K⟮x⟯: from the Eisenstein relation, the constant coefficient satisfies a₀ * u = -ξ ^ n for a unit u = 1 + ξ * t, so the extended maximal ideal is exactly the n-th power of (ξ); the ideal (ξ) is maximal as the kernel of the constant-coefficient-then-residue map through the monogenic presentation of integers L as 𝒪[K][X] modulo (g) from integers_eq_adjoin; and Ideal.ramificationIdx'_spec turns the two power bounds into e = n, which is the degree.

If evaluation at z is a surjection R[X] → A with kernel (g), where the maximal ideal of the local ring R is generated by the constant coefficient of g, whose image lies in (z), then (z) is maximal: it is the kernel of the induced map onto the residue field of R. (Not private: RootLifting.lean reuses it to make the integral closure local.)

Total ramifiedness of the adjoined root field #

The unit relation behind total ramifiedness, exposed for the discriminant bound of sub_one_le_d_adjoin: the constant coefficient of the minimal polynomial becomes a unit multiple of -ξ ^ n inside the integral closure.