Contradictory Majorants: Analyzing Integral Bounds and the Riemann Hypothesis

Download Full Article

This article is available as a downloadable PDF with complete code listings and syntax highlighting.

Introduction

The Riemann Hypothesis (RH) remains the most significant open problem in analytic number theory, asserting that all non-trivial zeros of the Riemann zeta function ζ(s) lie on the critical line Re(s) = 1/2. This conjecture is inextricably linked to the behavior of the Mertens function, M(x), which is the summatory function of the Möbius function μ(n). Historically, RH is known to be equivalent to the growth bound M(x) = O(x^1/2+ε) for any ε > 0.

In the research paper arXiv:hal-01845847v2, a novel refutation is proposed by examining the integral representation of the reciprocal of the zeta function. The analysis focuses on the tail of the Mellin transform of M(x) and introduces a sophisticated framework of majorants to test the consistency of the critical line. By comparing the lower bounds derived from the discrete nature of square-free integers with the upper bounds required by the Riemann Hypothesis, the paper identifies a scaling conflict that suggests the supremum of the real parts of the zeros, β*, must be equal to 1.

This article provides a technical deep dive into these integral structures, evaluating the majorant classes G and L, the role of phase alignment in Dirichlet tails, and the resulting asymptotic contradictions that challenge the current consensus on the distribution of zeta zeros.

Mathematical Background

The foundation of this analysis rests on the relationship between the zeta function and the Mertens function. For Re(s) > 1, the reciprocal of the zeta function is expressed as a Dirichlet series involving the Möbius function: 1/ζ(s) = ∑ μ(n)n^-s. Through Abel summation, this is transformed into the integral: 1/ζ(s) = s ∫₁^∞ M(u)u^-(s+1) du.

The source paper defines β* as the supremum of the real parts of the non-trivial zeros ρ. If the Riemann Hypothesis holds, β* = 1/2. The paper focuses on the quantity J(x, s) = ∫_x^∞ M(u)u^-(s+1) du, which represents the tail of this integral transform starting from a point x. Two critical function classes are introduced:

The G-class: A set of functions G(x) that bound the magnitude of J(x, s) uniformly for all imaginary parts t.
The L-class: A set of functions B(x) that bound J(x, s) weighted by the factor sqrt(σ² + t²).

The relationship between these classes is governed by the infimum function α(x), which represents the tightest possible uniform majorant. The paper establishes that σα(x) = A(x), where A(x) is the supremum over t of the weighted integral magnitude.

Main Technical Analysis

The Lower Bound and Phase Alignment

The most provocative aspect of arXiv:hal-01845847v2 is the derivation of a lower bound for |J(x, s)|. This is achieved by identifying a set of values T_ε(x) in the imaginary domain where the phases of the Dirichlet series nearly align. For a given ε > 0, there exist values of t such that the magnitude of the sum ∑ μ(n)n^-s for n > x is at least (1-ε) times the sum of the absolute values ∑ |μ(n)|n^-σ.

Since |μ(n)| is the indicator function for square-free numbers, and the density of square-free numbers is known to be 6/π², the lower bound for the integral magnitude can be approximated as: |J(x, s)| ≥ [(1-ε)/sqrt(σ²+t²)] * [ (6 x^1-σ)/(π²(σ-1)) + O(x^1/2-σ) ] - (m x^β-σ)/sqrt(σ²+t²), where m is a constant bounding the Mertens function.

The Upper Bound and Asymptotic Contradiction

Conversely, if we assume RH or any bound where β* < 1, we can establish an upper bound based on the growth of the Mertens function M(u) = O(u^β). This yields: |J(x, s)| ≤ (σ m x^β-σ) / ((σ - β) sqrt(σ² + t²)).

The contradiction emerges when comparing the ratio of these two bounds as x tends to infinity. If β < 1, the lower bound (which decays at a rate related to x^1-σ) eventually exceeds the upper bound (which decays at a rate related to x^β-σ). Specifically, the ratio contains a factor of x^1-β, which grows without bound. This leads the author to conclude that the initial assumption β < 1 is false, thereby implying β* = 1 and refuting the Riemann Hypothesis.

Novel Research Pathways

1. Quantitative Large-Values Theory for Dirichlet Tails

A primary avenue for future research is the rigorous quantification of the set T_ε(x). While the paper assumes this set is never empty, a diophantine analysis of the frequencies log(n) is required to determine how large t must be to achieve near-alignment. If the required t grows exponentially with x, the factor 1/sqrt(σ²+t²) might suppress the lower bound faster than the x-terms can compensate, potentially resolving the contradiction in favor of RH.

2. Higher-Order Moment Estimates

The integral bound methodology can be extended from first-order integrals of M(u) to higher-order moments. By analyzing the integral of M(u)^k, researchers could establish a hierarchy of constraints on the zero-density of the zeta function. This would involve defining generalized majorant classes G_k and investigating if the contradiction persists across all moments, which would provide a more robust refutation or a clearer identification of the analytic flaw.

3. Functional-Analytic Study of the Infimum Majorant

The function α(x), defined as the infimum of the majorant class G, can be studied as a maximal operator in a Hardy space. By treating the mapping from t to the integral tail as a boundary value of an analytic function, one could use the theory of extremal functions to prove whether α(x) must satisfy specific growth conditions (such as Omega-bounds) that are independent of the phase-alignment argument.

Computational Implementation

The following Wolfram Language implementation demonstrates the numerical tension between the empirical integral tails of the Mertens function and the theoretical upper bounds derived from the Riemann Hypothesis (assuming β = 1/2).

(* Section: Mertens Integral and Bound Discrepancy *)
(* Purpose: This code computes the tail of the Mertens function integral and compares it against upper bounds derived from the Riemann Hypothesis (beta = 1/2) *)

ClearAll[mertensList, integralTail, sigma, t, x, beta, mVal, results];

(* Parameters for the analysis *)
sigma = 1.1;
t = 14.13472514173469; (* Imaginary part of the first Riemann zero *)
beta = 0.5; (* RH assumption *)
limit = 5000;

(* 1. Compute the Mertens function values efficiently *)
mertensList = Accumulate[Table[MoebiusMu[n], {n, 1, limit}]];

(* 2. Define the integral tail approximation J(x, s) *)
(* J(x, s) = integral from x to infinity of M(u)u^-(s+1) du *)
integralTail[x_Integer, s_Complex] := Module[{sum},
  sum = NSum[mertensList[[Floor[u]]] * u^(-(s + 1)), {u, x, limit, 1}, 
    Method -> "SequenceLimit", WorkingPrecision -> 20];
  Abs[sum]
];

(* 3. Define the theoretical upper bound based on M(u) = O(u^beta) *)
mVal = Max[Table[Abs[mertensList[[n]]]/n^beta, {n, 1, limit}]];
upperBound[x_, sig_, b_] := (sig * mVal * x^(b - sig)) / ((sig - b) * Sqrt[sig^2 + t^2]);

(* 4. Generate comparison data for x in [50, 500] *)
results = Table[
  {x, integralTail[x, sigma + I*t], upperBound[x, sigma, beta]}, 
  {x, 50, 500, 50}
];

(* 5. Visualize the results *)
Print["Mertens Integral Analysis at t = ", t];
Print[TableForm[results, TableHeadings -> {None, {"x", "|J(x,s)|", "Upper Bound"}}]];

ListLinePlot[{results[[All, {1, 2}]], results[[All, {1, 3}]]},
  PlotStyle -> {Blue, {Red, Dashed}},
  PlotLegends -> {"|Integral Tail|", "RH Upper Bound"},
  AxesLabel -> {"x", "Value"},
  PlotLabel -> "Comparison of Mertens Integral and RH Bounds",
  ImageSize -> Large]

Conclusions

The analysis of arXiv:hal-01845847v2 reveals a significant structural tension between the discrete distribution of square-free integers and the analytic constraints of the Riemann zeta function. The majorant framework provides a rigorous way to isolate the influence of the Mertens function's fluctuations on the location of zeros. While the refutation of the Riemann Hypothesis remains a radical claim, the integral contradictions identified here suggest that the cancellation in the Möbius sequence must be extremely precise to maintain the critical line. The most promising next step is the quantification of the t-dependence in the phase alignment sets to determine if the identified contradiction is an asymptotic reality or a diophantine artifact.

References

arXiv:hal-01845847v2: Réfutation de l'hypothèse de Riemann.
G. H. Hardy and E. M. Wright, An Introduction to the Theory of Numbers, 5th edition, Clarendon Press.
E. C. Titchmarsh, The Theory of the Riemann Zeta-function, Oxford University Press.