1
$\begingroup$

Let $\sigma(x)=\sigma_1(x)$ be the classical sum of divisors of the positive integer $x$.

My question in the present post is closely related to this one in MO:

QUESTION

Does there exist a nontrivial prime power $q^k$ such that $\sigma(n^2)/n = q^k$ for some $n$?

I tried checking for examples of numbers $2 \leq n \leq {10}^6$ satisfying the divisibility constraint $$n \mid \sigma(n^2)$$ using a Pari-GP script, via Sage Cell Server:

for(n=2, 1000000, if((Mod(sigma(n^2),n) == 0),print(n,factor(n))))

Here is the output:

39[3, 1; 13, 1]
793[13, 1; 61, 1]
2379[3, 1; 13, 1; 61, 1]
7137[3, 2; 13, 1; 61, 1]
13167[3, 2; 7, 1; 11, 1; 19, 1]
76921[13, 1; 61, 1; 97, 1]
78507[3, 2; 11, 1; 13, 1; 61, 1]
230763[3, 1; 13, 1; 61, 1; 97, 1]
238887[3, 2; 11, 1; 19, 1; 127, 1]
549549[3, 2; 7, 1; 11, 1; 13, 1; 61, 1]
692289[3, 2; 13, 1; 61, 1; 97, 1]
863577[3, 2; 11, 2; 13, 1; 61, 1]

The Pari-GP interpreter of Sage Cell Server crashes as soon as a search limit of ${10}^7$ is specified.

OEIS sequence A232354 lists $187$ additional examples.

I skimmed through the list of the first $199$ examples (not including $1$) in OEIS and noted that all of them are odd. Additionally, here are the corresponding integer values for $\sigma(n^2)/n$ for $2 \leq n \leq {10}^6$: $$\frac{\sigma({39}^2)}{39} = 61,$$ $$\frac{\sigma({793}^2)}{793} = 873 = {3^2} \times {97},$$ $$\frac{\sigma({2379}^2)}{2379} = 3783 = 3 \times {13} \times {97},$$ $$\frac{\sigma({7137}^2)}{7137} = 11737 = {11}^2 \times {97},$$ $$\frac{\sigma({13167}^2)}{13167} = 26543 = {11} \times {19} \times {127},$$ $$\frac{\sigma({76921}^2)}{76921} = 85563 = {3^3} \times {3169},$$ $$\frac{\sigma({78507}^2)}{78507} = 141911 = 7 \times {11} \times {19} \times {97},$$ $$\frac{\sigma({230763}^2)}{230763} = 370773 = {3^2} \times {13} \times {3169},$$ $$\frac{\sigma({238887}^2)}{238887} = 417263 = 7 \times {11} \times {5419},$$ $$\frac{\sigma({549549}^2)}{549549} = 1155561 = 3 \times {11} \times {{19}^2} \times {97},$$ $$\frac{\sigma({692289}^2)}{692289} = 1150347 = 3 \times {{11}^2} \times {3169},$$ $$\frac{\sigma({863577}^2)}{863577} = 1562185 = 5 \times {97} \times {3221}.$$

(I did the computations one at a time using WolframAlpha.)

Note that $61$ is prime. However, it is not a nontrivial prime power.

I was hoping somebody with more computing power (and better programming skills) could automate the computation and thereby yield my desired result, if there is one at all. (I would be content if this computation were done over this list in OEIS.)

Thank you!

$\endgroup$
13
  • $\begingroup$ This inquiry is also related to this answer to a closely related question, which currently has a pending bounty. $\endgroup$ Feb 6, 2022 at 13:09
  • $\begingroup$ How large is the search range in OEIS ? $\endgroup$
    – Peter
    Feb 6, 2022 at 13:39
  • $\begingroup$ Hold on, let me check, @Peter. $\endgroup$ Feb 6, 2022 at 13:47
  • 2
    $\begingroup$ That is a huge search limit ! If no solution is within this search limit, I do not expect a solution at all. Let us think about useful restrictions for $n$ ... $\endgroup$
    – Peter
    Feb 6, 2022 at 13:51
  • 1
    $\begingroup$ Then it looks more like programming question (for stack overflow?)... Anyway I have loaded the list into Maple and checked that only $n=39$ there leads to single $p^e$ component. $\endgroup$
    – Sil
    Feb 9, 2022 at 12:42

1 Answer 1

1
$\begingroup$

I would be content if this computation were done over this list in OEIS

Quick & dirty Maple code below checks that only $n=39$ in the list above corresponds to a prime power $\frac{\sigma({39}^2)}{39} = 61$.

vals:=[seq(v[2],v in readdata("b232354.txt",[integer,integer]))]:
for n in vals do:
  m:=NumberTheory:-sigma(n^2)/n:
  ifs:=ifactors(m)[2]:
  comps:=nops(ifs):
  if comps=1 then:
    print(n,ifs[1][1],ifs[1][2]):
  fi:
od:

Output:

39, 61, 1

Though I don't think this helps to conclude anything (that's why this was originally posted just as a comment), because even if the list covers all $n\leq 677004885711$, it is still a list with just $200$ elements. $\sigma(n^2)/n$ being and integer is strict by itself, so adding another condition (it being a prime power) can simply lead to larger example (I am sure someone here could provide a heuristic for this). I don't think you can move on without understanding or at least restricting the $n$ for which $\sigma(n^2)/n$ is an integer.

$\endgroup$
4
  • $\begingroup$ Thank you for your time and attention, and your answer. I actually predict that $61$ does divide an odd perfect number, which led me to this inquiry. (I think the proof can be done via a factor chain implementation.) That $61$ does divide an odd perfect number $q^k n^2$ with special prime $q$ (yes, using the same notation as in my original post here) will then result to a contradiction, @Sil. $\endgroup$ Feb 11, 2022 at 23:53
  • $\begingroup$ My apologies for being stubborn, @Sil, but could you check (via Maple) what is the proportion of numbers $n$ satisfying $n \mid \sigma(n^2)$ and ${61} \mid n$, provided $n$ is in this list? I conjecture that this percentage is at least $90 \%$. (I do not have Maple installed on my machine, and I do not want to install a pirated copy.) Thanks! $\endgroup$ Feb 12, 2022 at 1:30
  • 1
    $\begingroup$ @ARNIEBEBITA-DRIS All of them satisfy $n \mid \sigma(n^2)$ by definition, and $67$ of them satisfy $61 \mid n$. $\endgroup$
    – Sil
    Feb 12, 2022 at 8:52
  • 1
    $\begingroup$ Thanks for checking, @Sil! $\endgroup$ Feb 12, 2022 at 15:03

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .