riemann

The Lean code claims a proof of the Riemann Hypothesis. This site has not checked the claim yet. See what is checked.

The Riemann Hypothesis from the right half-plane

drafted

Statement

Suppose that ζ(s)≠0\zeta(s) \neq 0 for each s∈ℂs \in \mathbb{C} with Re⁡(s)>12\operatorname{Re}(s) > \tfrac12 and s≠1s \neq 1. Then the Riemann Hypothesis is true: each zero ss of ζ\zeta that is not a trivial zero and not 11 has Re⁡(s)=12\operatorname{Re}(s) = \tfrac12.

Lean

theorem Formalism.VW.rh_of_right (hz : ∀ s : ℂ, 1 / 2 < s.re → s ≠ 1 → riemannZeta s ≠ 0) : RiemannHypothesis

The declaration is in Formalism/GammaFactorNonvanishing/RhOfRight.lean, line 5. Its page has the full list of what it uses and what uses it, and a link to the source.

Proof

Let ss be a zero of ζ\zeta, not a trivial zero, with s≠1s \neq 1. Compare Re⁡(s)\operatorname{Re}(s) with 12\tfrac12. There are three cases.

Case Re⁡(s)>12\operatorname{Re}(s) > \tfrac12. The hypothesis gives ζ(s)≠0\zeta(s) \neq 0. This is a contradiction, so this case does not occur.

Case Re⁡(s)=12\operatorname{Re}(s) = \tfrac12. This is the claim.

Case Re⁡(s)<12\operatorname{Re}(s) < \tfrac12. We show that this case does not occur. Put u=1−su = 1 - s. Then Re⁡(u)=1−Re⁡(s)>12\operatorname{Re}(u) = 1 - \operatorname{Re}(s) > \tfrac12.

First, s≠0s \neq 0, because ζ(0)=−12\zeta(0) = -\tfrac12 (Mathlib, riemannZeta_zero). So u≠1u \neq 1. Also uu is not 0,−1,−2,…0, -1, -2, \dots, because Re⁡(u)>12\operatorname{Re}(u) > \tfrac12.

The functional equation of Mathlib (riemannZeta_one_sub) says: if uu is not 0,−1,−2,…0, -1, -2, \dots and u≠1u \neq 1, then

ζ(1−u)=2 (2π)−u Γ(u) cos⁡ ⁣(πu2)ζ(u).\zeta(1 - u) = 2\,(2\pi)^{-u}\,\Gamma(u)\,\cos\!\left(\frac{\pi u}{2}\right)\zeta(u).

Here 1−u=s1 - u = s, so the left side is ζ(s)=0\zeta(s) = 0. We show that each factor on the right is not zero.

  • 2≠02 \neq 0, and (2π)−u≠0(2\pi)^{-u} \neq 0 because 2π≠02\pi \neq 0.
  • Γ(u)≠0\Gamma(u) \neq 0, because Re⁡(u)>0\operatorname{Re}(u) > 0 (Mathlib, Complex.Gamma_ne_zero_of_re_pos).
  • ζ(u)≠0\zeta(u) \neq 0, by the hypothesis, because Re⁡(u)>12\operatorname{Re}(u) > \tfrac12 and u≠1u \neq 1.
  • cos⁡(πu/2)≠0\cos(\pi u / 2) \neq 0. Suppose that it is zero. Then πu/2=(2k+1)π/2\pi u / 2 = (2k + 1)\pi / 2 for an integer kk, so u=2k+1u = 2k + 1 and s=−2ks = -2k. If k<0k < 0, then Re⁡(s)=−2k≥2\operatorname{Re}(s) = -2k \geq 2, which is not less than 12\tfrac12. If k=0k = 0, then s=0s = 0, which we excluded. If k>0k > 0, then s=−2k=−2((k−1)+1)s = -2k = -2((k - 1) + 1) is a trivial zero, which we excluded.

So the right side is not zero, but the left side is zero. This is a contradiction.

Notes

This step is classical and uses only Mathlib. The Lean proof follows the argument above line by line. The case split on kk and the conversions between integers, natural numbers and complex numbers take most of its lines.

Mathlib writes the functional equation with cos⁡\cos and Γ(u)\Gamma(u), not with the completed zeta function Λ(s)=π−s/2Γ(s/2)ζ(s)\Lambda(s) = \pi^{-s/2}\Gamma(s/2)\zeta(s). The two forms are equivalent.

Uses

The proof cites no other step.

Used by

Status

Drafted. The text follows the Lean code, but no person has checked it against the code with the rules of spec 05.7 yet. If you check it, tell us.

A model wrote the first draft (writer@claude-opus-5-5). A person checks it before its status goes up.

The source of this page. Markdown. Something is wrong? Report it.