On Congruences for Iterates of the Sum--Power Divisor Function and Conditional Implications for the Riemann Hypothesis
Abstract
Inspired by Cohen and te Riele~Cohen1996, who computationally verified that for every n ≤ 400 there exists k such that σk(n) 0 n (where σk denotes the k-fold iteration of the sum-of-divisors function), this paper resolves their reverse question negatively: no integer n > 1 satisfies σk(n) 0 n for all k ≥ 1. The proof eliminates prior gaps via Lenstra's density-zero bounds σk(m) m / m combined with Robin's RH-equivalent criterion σ(n) < eγ n n + 0.6483 n / n (n ≥ 5041), showing universal metaperfect divisibility implies RH-violating σ growth or low-lying zeta zeros near s=1. Among multiperfect n with prime L = lcm(1+ep : p n), only n=6 satisfies the congruence for all odd k, with Shannon entropy H(σk(6) 6) 2 reflecting periodic order. We analyze bifurcation phenomena in the dynamics σk(n) n, where high-entropy chaotic residues for other n mirror GUE statistics of zeta zeros ( T / 2π near s=1/2, >41\% verified on critical line), contrasting the ordered n=6 case. Zero rates near s=1 (simple pole) and s=1/2 bound iterated σ distributions, linking to RH via divisor sums and dynamical bifurcations; we conjecture n=6 uniquely achieves odd-k divisibility with small period dividing L.
Turn this paper into a lesson
ArcXiv compiles a structured reading guide from this paper's metadata: plain-English importance, contributions, prerequisite concepts, which sections to read first, flashcards, and a quiz. Grounded in the abstract, never invented.