Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Not sure if this belongs to MO or not. Are there any lower bound on radical of $2^n \pm 1$ (Recall that radical of integer $rad(k)$ is a product of primes which divide integer $k$)?

As an example If abc-conjecture is true in the form $max(|a|,|b|,|c|) \leq rad(abc) ^2 $ then $$rad(2^n \pm 1) \geq 2^{n/2 - 1}$$ I wonder if this estimate is proven (or perhaps conjectured) by anyone? Are there any nontrivial results here?

share|improve this question
    
The inequalities in the comment seem to be intended in the other direction. –  quid Jan 15 '13 at 18:15
    
$2^n + 1 = (2^n + 1) \rightarrow 2^n+1 \leq rad (1\times 2\times (2^n+1))^2 \rightarrow 2rad(2^n+1) \geq \sqrt{2^n+1} $ Thank you, quid, sorry for that –  Ostap Chervak Jan 15 '13 at 18:23
    
Yyou could use the Cunningham project for numerical evidence to help decide what the conjecture should be. Gerhard "Ask Me About System Design" Paseman, 2013.01.15 –  Gerhard Paseman Jan 15 '13 at 18:34
5  
There are certainly nontrivial results, e.g., from work of Stewart and Yu towards the ABC conjecture it seems one gets $\mathrm{rad}(2^n\pm 1) \geq n^{3-o(1)}$ as $n\to\infty$. –  Anonymous Jan 16 '13 at 17:33
1  
According to 'On the largest prime factor of the Mersenne Numbers', written by Ford, Luca, and Shparlinski, Schinzel proves in 'On primitive prime factors of $a^n - b^n$' that the largest prime factor of $2^n-1$ is at least $2n-1$ for all $n \ge 13$. –  Woett Jan 17 '13 at 0:32

1 Answer 1

The paper arXiv:1409.2974v1 proves a lemma regarding $2^L-1$.

Let $n$ be a positive integer, and define $L = \text{lcm}[1, 2, \dots n]$ and $t = \lfloor \frac{\log(n)}{\log(2)} \rfloor$. Then $L*\text{rad}(2^L-1) \le 2^t(2^L-1)$.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.