An introduction to p-adic L-functions — Lean blueprint

9.1. Notation for the cyclotomic tower🔗

For n \in \N set F_n := \Q(\mu_{p^n}) and K_n := \Qp(\mu_{p^n}), with maximal totally real subfields F_n^+, K_n^+ (the fixed fields of complex conjugation c). Write \sU_n := \cO_{K_n}^{\times}, \qquad \sU_n^+ := \cO_{K_n^+}^{\times} for the local units. The extensions K_n/\Qp are Galois and totally ramified at p, of degree (p-1)p^{n-1}. Passing to the limit, K_\infty := \Qp(\mu_{p^\infty}) = \bigcup_{n\ge 1} K_n.

Definition9.1.1
uses 0used by 1XL∃∀N

Sending a primitive p^nth root of unity to a primitive p^nth root of unity gives an isomorphism \chi_n : \Gal(F_n/\Q) \xrightarrow{\sim} (\Z/p^n\Z)^{\times}, characterised by \sigma(\xi) = \xi^{\chi_n(\sigma)} for any \xi \in \mu_{p^n}. Taking the inverse limit yields the cyclotomic character \chi := \varprojlim_n \chi_n : \GCal := \Gal(F_\infty/\Q) \xrightarrow{\sim} \Zpx, an isomorphism of profinite groups, where F_\infty := \Q(\mu_{p^\infty}). It induces \GCal^+ := \Gal(F_\infty^+/\Q) \cong \Zpx/\{\pm 1\}.

Definition9.1.2
uses 1used by 1L∃∀N

We fix once and for all a compatible system of roots of unity (\xi_{p^n})_{n \in \N}: each \xi_{p^n} is a primitive p^nth root of unity with \xi_{p^{n+1}}^p = \xi_{p^n}. Put \pi_n := \xi_{p^n} - 1, a uniformiser of K_n. The infinite-level local units are \sU_\infty := \varprojlim_n \sU_n, the inverse limit taken with respect to the norm maps N_{n,n-1} : K_n \to K_{n-1}. Together with its norm-1 submodule \sU_{\infty,1} := \varprojlim_n \sU_{n,1} (where \sU_{n,1} = \{u \in \sU_n : u \equiv 1 \pmod{\mathfrak p_n}\}) it is a compact \Zp-module carrying a continuous action of \GCal, hence of the Iwasawa algebra \Lam(\GCal). This Definition 9.1.1 \GCal-action is the main reason for passing to infinite level.

In the formalisation the tower lives inside ℂ_[p] (matching the B(0,1)-framing below): the compatible system is zetaSys (built by recursion from algebraic closedness), K n is the intermediate field ℚ_p(ξ_{p^n}) with [K n : ℚ_p] = φ(p^n) proved by an Eisenstein argument, O n its norm-unit-ball, levelNorm the relative norm N_{n+1,n} (with the collapse N(ξ^b_{p^{n+1}} − 1) = ξ^b_{p^n} − 1), and 𝒰_∞ is the structure NormCompatUnits of norm-compatible unit families. The 𝒢-action and 𝒰_{∞,1} are deferred to the §11 pass (plan.md); the cyclotomic-character node above stays prose for the same reason.

Lean code for Definition9.1.26 definitions
  • defdefined in PadicLFunctions/Coleman/Tower.lean
    complete
    def PadicLFunctions.Coleman.zetaSys (p : ) [hp : Fact (Nat.Prime p)] :
        ℂ_[p]
    def PadicLFunctions.Coleman.zetaSys (p : )
      [hp : Fact (Nat.Prime p)] :   ℂ_[p]
    The fixed compatible system `n ↦ ξ_{p^n}` (RJW TeX 2507). 
  • defdefined in PadicLFunctions/Coleman/Tower.lean
    complete
    def PadicLFunctions.Coleman.K (p : ) [hp : Fact (Nat.Prime p)] (n : ) :
      IntermediateField ℚ_[p] ℂ_[p]
    def PadicLFunctions.Coleman.K (p : )
      [hp : Fact (Nat.Prime p)] (n : ) :
      IntermediateField ℚ_[p] ℂ_[p]
    R9: the local cyclotomic field `K_n = ℚ_p(μ_{p^n})`, realised inside
    `ℂ_p` as `ℚ_p(ξ_{p^n})` (RJW TeX 2473). 
  • defdefined in PadicLFunctions/Coleman/Tower.lean
    complete
    def PadicLFunctions.Coleman.pi (p : ) [hp : Fact (Nat.Prime p)] (n : ) :
      ℂ_[p]
    def PadicLFunctions.Coleman.pi (p : )
      [hp : Fact (Nat.Prime p)] (n : ) :
      ℂ_[p]
    R9: the uniformiser `π_n = ξ_{p^n} − 1` of `K_n` (RJW TeX 2507). 
  • defdefined in PadicLFunctions/Coleman/Tower.lean
    complete
    def PadicLFunctions.Coleman.O (p : ) [hp : Fact (Nat.Prime p)] (n : ) :
      Subring ℂ_[p]
    def PadicLFunctions.Coleman.O (p : )
      [hp : Fact (Nat.Prime p)] (n : ) :
      Subring ℂ_[p]
    R9: the integer ring `O_n = O_{K_n}` — the norm-unit-ball of `K_n`
    (equivalently the integral closure of `ℤ_p`; the identification is the
    Eisenstein-monogenicity ticket). RJW TeX 2474. 
  • defdefined in PadicLFunctions/Coleman/Tower.lean
    complete
    def PadicLFunctions.Coleman.levelNorm (p : ) [hp : Fact (Nat.Prime p)]
      (n : ) : ℂ_[p]  ℂ_[p]
    def PadicLFunctions.Coleman.levelNorm (p : )
      [hp : Fact (Nat.Prime p)] (n : ) :
      ℂ_[p]  ℂ_[p]
    The relative field norm `N_{K_{n+1}/K_n} : K_{n+1} → K_n`, viewed as a map
    `ℂ_p → ℂ_p` (junk-extended by `0` off `K_{n+1}`). For `x ∈ K_{n+1}` it is the
    `Algebra.norm` of the corresponding element of
    `IntermediateField.extendScalars (K_le_succ p n)` (whose carrier is `K_{n+1}`
    seen as a `K_n`-algebra), coerced back into `ℂ_p` via `K_n ↪ ℂ_p`.
    RJW TeX 2503. 
  • structure(4 fields)defined in PadicLFunctions/Coleman/Tower.lean
    complete
    structure PadicLFunctions.Coleman.NormCompatUnits (p : )
      [hp : Fact (Nat.Prime p)] : Type
    structure PadicLFunctions.Coleman.NormCompatUnits
      (p : ) [hp : Fact (Nat.Prime p)] : Type
    `𝒰_∞`, the norm-inverse-limit of the local unit groups (RJW TeX 2503): a
    compatible system of units, each in its integer ring together with its inverse,
    matched by the level norms `N_{n+1,n}`. The `compat` field is only imposed for
    `n ≥ 1` (the level norm `N_{n+1,n}` carries the `n ≥ 1` degree-`p` step). 

    Fields

    elems :   ℂ_[p]ˣ
    The unit at level `n`, `u_n ∈ K_n^×`. 
    mem :  (n : ), (self.elems n)  PadicLFunctions.Coleman.O p n
    Each `u_n` lies in the integer ring `𝒪_n`. 
    inv_mem :  (n : ), (↑(self.elems n))⁻¹  PadicLFunctions.Coleman.O p n
    Each inverse `u_n⁻¹` lies in `𝒪_n` (so `u_n ∈ 𝒪_n^×`). 
    compat :  (n : ), 1  n  PadicLFunctions.Coleman.levelNorm p n (self.elems (n + 1)) = (self.elems n)
    Norm compatibility `N_{n+1,n}(u_{n+1}) = u_n` for `n ≥ 1`. 

The norm-1 units \sU_{n,1} are genuine \Zp-modules: for u \in \sU_{n,1} and a \in \Zp, the series u^a = \sum_{k\ge 0}\binom{a}{k}(u-1)^k converges. By contrast the full local units \sU_n are only \Z-modules.