Confinement of zeta zeros in a finite-height band
ProvedZetaZeroConfinement.zero_in_bandarda-rh65k-transfernumber-theoryriemann-zeta
Let , , and , where is the defined effective zero-free-region constant. If and with , then
This supplies the real-part confinement used by the finite-height box certificate.
Preamble
import Definitions.Def_ArdaRH65K_Turing121_DlvpZetaRateEffective
import Definitions.Def_ArdaRH65K_Turing121_StripRepr
import Definitions.Def_ArdaRH65K_Turing121_StripReprR1
import Mathlib
import Theorems.Thm_ZeroFreeBridge_riemannZeta_ne_zero_region
import Theorems.Thm_ZetaZeroConfinement_zeta_zero_iff_completed_zero_of_im_ne
/- CONFINEMENT (kernel): every nontrivial zeta zero up to height T lies in the band [a, 1-a].
This is the "confinement" half of the "all zeros up to height T are on the line" project:
box-localization (all zeros IN a box on the line, PR #312) + confinement (all zeros ARE in the
box, this file). We prove the confinement:
`zero_in_band` : a nontrivial zero `ρ` (`ζ ρ = 0`, `0 < ρ.im ≤ T`) with `a ≤ dlvpRateC/log T`
satisfies `a ≤ ρ.re ≤ 1 - a`.
Both edges come from the SELF-CONTAINED effective de la Vallée Poussin zero-free region
`ZeroFreeBridge.riemannZeta_ne_zero_region` (PR #318 -- no multiplicity input, β≥3/4 discharged
internally): `ζ(β+iγ) ≠ 0` for `|γ| ≥ 55/16` and `β > 1 - dlvpRateC/log|γ|`.
- RIGHT edge `ρ.re ≤ 1-a`: contrapositive of the region at `ρ` gives `ρ.re ≤ 1 - dlvpRateC/log|γ|`,
and `dlvpRateC/log|γ| ≥ dlvpRateC/log T ≥ a` (log monotone, `|γ| = ρ.im ≤ T`; `dlvpRateC > 0`).
- LEFT edge `a ≤ ρ.re`: reflect through the functional equation. `ζ ρ = 0` ⟹ `Λ ρ = 0`
(unconditional strip bridge: `ρ.im ≠ 0` ⟹ `Gammaℝ ρ ≠ 0`) ⟹ `Λ (1-ρ) = 0`
(`completedRiemannZeta_one_sub`) ⟹ `ζ (1-ρ) = 0` (strip bridge again, `(1-ρ).im = -ρ.im ≠ 0`).
The reflected zero has real part `1-ρ.re`, height `-ρ.im`, `|-ρ.im| = |ρ.im| ≥ 55/16`. The
region contrapositive on it gives `1-ρ.re ≤ 1 - dlvpRateC/log|γ| ≤ 1 - a`, i.e. `ρ.re ≥ a`.
RESIDUAL (documented, honest -- a genuine Mathlib gap, NOT a fake). Mathlib has no "no zeta
zeros with `|Im| < 14`" fact (the first nontrivial zero is at height ~14.13 > 55/16 ~ 3.44), so
the region's height-floor hypothesis `55/16 ≤ |ρ.im|` is carried as an ADDED hypothesis `hγ` on
`zero_in_band`. It is a small classical fact, discharged at the concrete T=100 instantiation
from the on-line sweep (all 29 zeros up to 100 have height ≥ 14 > 55/16). No multiplicity input,
no `β ≥ 3/4` casework, no `0 < ρ.re` strip hypothesis (the FE strip bridge here is unconditional
for `ρ.im ≠ 0`).
conjecture1_proved = False (a kernel-verified reduction of confinement to the effective dVP
region + the functional equation, NOT a proof of RH).
-/
open Complex
section
namespace ZetaZeroConfinement
end ZetaZeroConfinement
open ZetaZeroConfinement
Formal statement
theorem ZetaZeroConfinement.zero_in_band (a T : ℝ) (haC : a ≤ ZeroFreeBridge.dlvpRateC / Real.log T)
(ha0 : 0 < a) (hT : 100 ≤ T)
{ρ : ℂ} (hzero : riemannZeta ρ = 0) (him0 : 0 < ρ.im) (himT : ρ.im ≤ T)
(hγ : 55 / 16 ≤ |ρ.im|) :
a ≤ ρ.re ∧ ρ.re ≤ 1 - a := by sorry
Source