0
$\begingroup$

Hello,

I would like to know if there is some kind of heuristics according to which the sequence $(S_n)_{n\in\mathbb{N}}$ would be a divisibility sequence, where $S_n$ is the $n$-th superperfect number (i.e a number $m$ such that $\sigma(\sigma(m))=2m$).
Thanks in advance.

$\endgroup$
2
  • $\begingroup$ Read "divisibility" instead of "disibility". Sorry for the typo. $\endgroup$ Dec 15, 2012 at 20:49
  • $\begingroup$ Don't apologize. Fix. Gerhard "Ask Me About System Design" Paseman, 2012.12.15 $\endgroup$ Dec 15, 2012 at 20:55

1 Answer 1

2
$\begingroup$

The sequence of superperfect numbers is a dvisibility sequence if and only if no odd superperfect numbers exist.

To see this just note that the even superperfect numbers are known to be all powers of two, precisely they are $2^{k-1}$ such that $2^k -1$ is a (Mersenne) prime.

Then, if all are even it is clearly a divisibility sequence, indeed each one divides all subsequent ones (not just those whose index is a multiple, as is the definition of a divisibilty sequence).

By contrast, if an odd should exists its index would need to be a prime $p$, as otherwise it would be divisible by some preceding superpefect number--an even number; yet then the $2p$-th superperfect number can be neither odd, being divisible by the second superperfect number, nor a power of two, being divisible by the odd $p$-th superperfect number.

It seems to be believed that there are no odd superperfect numbers (and thus the sequence is a divisibility sequence), and partial results in this direction (somewhat resembling those on odd perfect numbers). Various literature on this is available (freely); for example "On finiteness of odd superperfect numbers" by Tomohiro Yamada (see http://arxiv.org/abs/0803.0437 ), showing in particular that for any fixed number of distinct prime divisors there can only be finitely many.

In brief, the heuristics for this being a divisibility sequence are literally the same as those for there not being an odd superperfect number, and there can be no additional/other heuristics.

$\endgroup$
1
  • $\begingroup$ I know that even superperfect numbers are powers of 2, but I was expecting some other reasons explaining why it should be a divisibility sequence. I accept your answer though, as you gave me the link of Yamada's paper. $\endgroup$ Dec 15, 2012 at 22:44

Your Answer

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

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