Let $(a,b)$ be a pair of coprime positive integers with $a$ being even. Are these conditions sufficient to prove that there exist infinitely many positive integers $n,$ such that $(a^n+1,b^n+1)=1$ ?

  • 2
    $\begingroup$ "Sufficient to prove that" is an interesting formulation :) If it is taken literally, then I am clueless about the answer since nobody has been able to prove this. In other words, this is a well known conjecture. However, using the work of Corvaja and Zannier based on Schmidt's Subspace theorem, one can at least prove the sub-exponential upper bound $\exp(\varepsilon n)$ on this gcd, for all $\varepsilon > 0$, and infinitely many $n$. (Indeed all large enough $n$ if additionally $a$ and $b$ are multiplicatively independent.) $\endgroup$ Jul 31, 2019 at 8:27
  • $\begingroup$ If you can prove there is one n, you should be able to prove there are infinitely many. Gerhard "Start With A Simpler Case" Paseman, 2019.07.31. $\endgroup$ Jul 31, 2019 at 8:33
  • $\begingroup$ Is it at least true that there are infinitely many values of $n$ with $(2^n+1,3^n+1)=1$? $\endgroup$
    – Seva
    Jul 31, 2019 at 8:51
  • $\begingroup$ Thank you very much for enlightening me. I was almost sure that it is a famous conjecture, but I couldn't find any papers about it. $\endgroup$ Jul 31, 2019 at 9:17
  • $\begingroup$ Is it known that for any set $P$ of primes of positive relative density ($|P\cap[1,x]\gg x/\log x$) there are infinitely many exponents $n$ such that $b^n+1$ is divisible only by the primes from $P$? Choosing $P$ to be the set of all primes $p\equiv 3\pmod 4$ such that $a$ is a quadratic residue modulo $p$, for every such $n$ we will have $(a^n+1,b^n+1)=1$ since $a^n+1$ is not divisible by primes $p\in P$ in view of $a^n\equiv -1\pmod p$. $\endgroup$
    – Seva
    Jul 31, 2019 at 12:41

1 Answer 1


I think that this sort of question was originally asked by Ailon and Rudnick, but they use $-1$ instead of $+1$ and asked if $\gcd(2^n-1,3^n-1)=1$ for infinitely many $n$. In this setting, for more general $a$ and $b$, the right question/conjecture would be $\gcd(a^n-1,b^n-1)=\gcd(a-1,b-1)$. They prove something stronger if you replace $\mathbb Z$ with $\mathbb C[t]$. Here's the Ailon-Rudnick paper:

N. Ailon and Z. Rudnick, ‘Torsion points on curves and common divisors of $a^k-1$ and $b^k-1$, Acta Arith., 113 (2004), no. 1, 31–38 (MSN).

There have been lots of articles on this and related problems. For example, a Google search on "Ailon Rudnick gcd" brings up recent articles such as:


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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