I am trying to find some set of prime numbers $p$ such that the below identity holds true. I was hoping to seek advice as to finding some stronger or more useful relations that may help find such a set $p$.

Let there exist some prime number $P$ raised to some positive integer $N$. Let there further exist some prime numbers $p_0,..., p_i \in p$ such that $p_a \ne p_b$ for any $ a \ne b$, and positive integers $n_0,..., n_i \in n$ such that $n_a \leq n_b$ for any $a \gt b$, and such that they fulfill the below, equivalent identities. $$gcd(\prod_{j=0}^{i} \sum_{k=0}^{n_j} p_j^k,P^N)=P^N$$ $$\prod_{j=0}^{i} gcd(\sum_{k=0}^{n_i}p_j^k,P^N)=P^K,K \ge N$$

It may be useful to note the connection between this identity and multiperfect numbers, wherein this identity must hold for each factor $P^N$ in the prime factorization of some multiperfect number $x$. This becomes evident when considering the formula for the sum of the proper divisors of $x$ , $\sigma(x)=-x+\prod_{j=0}^{i} \sum_{k=0}^{n_j} p_j^k$, and recalling that $\frac{\sigma(x)}{x}$ must be an integer.

Any insight into the problem is appreciated.


Using the formula for summation of finite geometric sequences, we have: $$\prod\limits_{j=0}^i\sum\limits_{k=0}^{n_j}p_j^k = \prod\limits_{j=0}^i\frac{p_j^{n_j+1}-1}{p_j-1}$$

So you are asking for a set of distinct primes $p_0, \ldots, p_i$ and positive integers $n_0 \geq \ldots \geq n_i$ such that this quantity is divisible by $P^N$.

Here is one possible way to construct such a set (which works for primes $P \gt 2$).

First let $i = N-1$.

Now choose the $p_j$ to be distinct primes not equal to $P$ such that $p_j \not\equiv 1 \pmod P$.

This is possible because there are infinitely many primes in any of the congruence classes modulo $P$ with representatives that are coprime to $P$ (see Dirichlet's theorem on arithmetic progressions).

For each $j$, let $n_j = P-2$.

Then by Fermat's little theorem we have $p_j^{n_j+1} - 1 = p_j^{P-1} - 1 \equiv 0 \pmod P$, which means $p_j^{n_j+1}-1$ is divisible by $P$ for all $j$, and moreover $p_j - 1$ is not divisible by $P$ by the choice of $p_j$.

Therefore we obtain a product of $N$ terms, each of which is divisible by $P$, which means the product itself is divisible by $P^N$.

  • $\begingroup$ Another possibility would be to let $i = 0$, $p_0 \neq P$ prime such that $p_0 \not\equiv 1 \pmod P$ and $n_0 = \varphi(P^N) - 1 = P^{N-1}(P-1) - 1$. Then you can use Euler's theorem. $\endgroup$ – Tob Ernack Mar 5 '18 at 21:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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