Let $p$ be a prime number, I think when $p^2+p+1=q^a$, where $q$ is a prime number, then $a=1$. But I can't prove it. Is it true?

2$\begingroup$ Please use TeX on this site. $\endgroup$ – GH from MO May 18 '15 at 23:29
This question is answered (affirmatively and somewhat more generally) in the following paper: Chat Yin Ho, Projective planes with a regular collineation group and a question about powers of a prime, J. Algebra 154 (1993), no. 1, 141–151. The proof there uses the ring of Eisenstein integers. (The issue is available online via Open Access).

$\begingroup$ Thanks a lot Mr Geoff Robinson, I saw the paper and it was exactly what I wanted. Very nice $\endgroup$ – darya May 18 '15 at 16:22


$\begingroup$ I am not sure. It may be that Ho's proof can be adapted to that case. If it does not, it may point to a counterexample. $\endgroup$ – Geoff Robinson May 18 '15 at 16:57

$\begingroup$ @ Geoff Robinson thank you very much for your help and advice. $\endgroup$ – darya May 18 '15 at 17:28

10
The equation $$ \frac{x^k1}{x1}=y^m$$ is known as the NagellLjunggren equation. It is conjectured that for $x\geq 2$, $y\geq 2$, $k\geq 3$, $m\geq 2$, the only solutions are $$ \frac{3^51}{31}=11^2,\qquad \frac{7^41}{71}=20^2,\qquad \frac{18^31}{181}=7^3.$$ For $3\mid k$, the equation was solved by Ljunggren (Norsk. Mat. Tidsskr. 25 (1943), 1720). For more details see also here.
It follows that $p^2+p+1=q^a$ for any integers $p,q,a\geq 2$ implies $p=18$, $q=7$, $a=3$.

1

$\begingroup$ @knsam: Thank you, especially for the scan of Ljunggren's paper. In fact I emphasized the case $m=2$ in my earlier related post (mathoverflow.net/questions/177952/…), and the linked paper by Bugeaud and Mihailescu mentions other solved cases (by NagellLjunggren) as well. BTW I would love to read a complete account of these results in English. Under my other MO response I linked here, it was remarked that Ribenboim's book about Catalan's conjecture discusses some cases in detail (starting with page 110), but I don't have access to this book. $\endgroup$ – GH from MO May 19 '15 at 0:03

1$\begingroup$ Yes, actually Ho notes at the end of his paper that his Theorem A is covered by Ljunggren's work. $\endgroup$ – Geoff Robinson May 19 '15 at 0:38

1$\begingroup$ @knsam: Having a full copy of the book would be wonderful. Having the relevant pages on NagellLjunggren only (pp.110) would be great, too! Thanks for your efforts. $\endgroup$ – GH from MO May 19 '15 at 0:55

1$\begingroup$ @darya: Instead of asking a new question in a comment, you probably should pose this as another question. Also, you should tell why you are interested in this specific equation. $\endgroup$ – Peter Mueller May 19 '15 at 11:53