Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated


For a nonzero integer xx and a prime pp, ord⁡px\operatorname{ord}_p x is the exponent of pp in xx, and ω(m)\omega(m) is the number of distinct prime factors of mm.

Statement

Theorem 1.2 (printed p. 295). Let aa and bb be integers with a>b>0a>b>0. There is a number C1C_1, effectively computable in terms of ω(ab)\omega(ab), with the following property: for every prime pp that does not divide abab and exceeds C1C_1, and every integer n≥2n\ge2,

ord⁡p(an−bn)<pexp⁡ ⁣(−log⁡p52log⁡log⁡p)log⁡a+ord⁡pn.(1.9)\operatorname{ord}_p(a^n-b^n)< p\exp\!\left(-\frac{\log p}{52\log\log p}\right)\log a+\operatorname{ord}_p n. \tag{1.9}

Immediately after the theorem the paper records the case n=p−1n=p-1: if a>b>0a>b>0 are integers and pp is an odd prime not dividing abab with p>C1p>C_1, then

ord⁡p(ap−1−bp−1)<pexp⁡ ⁣(−log⁡p52log⁡log⁡p)log⁡a.\operatorname{ord}_p(a^{p-1}-b^{p-1})< p\exp\!\left(-\frac{\log p}{52\log\log p}\right)\log a .

The printed hypotheses of this consequence also include "nn is an integer with n⩾2n\geqslant 2" (p. 295), although nn does not occur in its inequality.

The paper says the theorem follows from a special case of Lemma 4.3, the pp-adic estimate that yields a crucial step in the proof of Theorem 1.1. It then cites Yamada's estimate ord⁡p(ap−1−1)<C2 p(log⁡p)−2log⁡a\operatorname{ord}_p(a^{p-1}-1)<C_2\,p(\log p)^{-2}\log a, with C2C_2 effectively computable in terms of ω(a)\omega(a), display (1.10) on p. 296.

Source and proof pointer

Cameron L. Stewart, On divisors of Lucas and Lehmer numbers, Acta Mathematica 211 (2013), 291--314, as identified on the source card. The theorem and (1.9) are on printed p. 295 (physical p. 5), with the case n=p−1n=p-1 below them. The proof is Section 6, printed p. 311 (physical p. 21).

In the arXiv:1008.1274v1 manuscript the result is Theorem 2 with display (9), physical p. 4. There the remark that follows also states the intermediate inequality ord⁡p(an−bn)≤ord⁡p(ap−1−bp−1)+ord⁡pn\operatorname{ord}_p(a^n-b^n)\le\operatorname{ord}_p(a^{p-1}-b^{p-1})+\operatorname{ord}_p n for odd p∤abp\nmid ab and n≥2n\ge2, before the case n=p−1n=p-1. The proof is Section 6, physical pp. 16--17.

In outline, the proof reduces to coprime a,ba,b, shows that for an odd prime pp the pp-adic order of an−bna^n-b^n is at most that of ap−1−bp−1a^{p-1}-b^{p-1} plus ord⁡pn\operatorname{ord}_p n (the paper's (6.5)), and then applies Lemma 4.3 with exponent p−1p-1. The proof is not transcribed here.

Read depth. Claims checked: the statement, its hypotheses, constants, label and page were read clause by clause on the printed page. The proof was not checked line by line.

Bears on

  • Problem 977: the theorem is not the result that settles the problem. It follows from a special case of Lemma 4.3, which yields a crucial step in the paper's proof of Theorem 1.1, whose specialization (1.8) with a=2a=2, b=1b=1 gives P(2n−1)/n→∞P(2^n-1)/n\to\infty.