11
$\begingroup$

This is a second attempt (see Primes $p$ such that $432 p +1$ is prime)

Is the set of squarefree numbers $n$ such that $n(432 n+1)$ is also squarefree known to be infinite?

Fact: the number of such numbers $n$ such that $n\leq 10^6$ is precisely $553095$. Do we expect that $$ \lim_{x\to \infty}\frac{\# \{n \leq x\ : \ n \ \textrm{is squarefree integer, and } 432n+1 \ \textrm{is squarefree}\}}{x} = 1/2?$$

$\endgroup$
13
  • 7
    $\begingroup$ I don't have time to work this out, but the density is the product of local densities, which is surely not $1/2$ but some more complicated number (probably transcendental). That is, for each prime $p$ let $c_p$ denote the density of $n$'s such that $n(432n+1)$ is not divisible by $p^2$, and then the density in the post will be $\prod_p c_p$. $\endgroup$
    – GH from MO
    Aug 5, 2014 at 13:51
  • 3
    $\begingroup$ What is so special about 432? $\endgroup$ Aug 5, 2014 at 13:52
  • $\begingroup$ See my response here how to do this for $n(n+1)$: mathoverflow.net/questions/172009/… $\endgroup$
    – GH from MO
    Aug 5, 2014 at 13:54
  • 1
    $\begingroup$ I think you should put more about the motivation in the question, then. $\endgroup$ Aug 5, 2014 at 13:55
  • 1
    $\begingroup$ Actually, thinking a bit about it, the problem is not so simple because the error terms can add up to a constant times $x$. At any rate, Estermann proved in 1931 that for any nonzero $k$ there are infinitely many $n$'s such that $n^2+k$ is square-free. I am sure the approach there can be easily modified to yield what you need. I found Estermann's paper in a more recent paper by Heath-Brown (arxiv.org/pdf/1010.6217.pdf). Estermann's paper is also online: digizeitschriften.de/dms/img/… $\endgroup$
    – GH from MO
    Aug 5, 2014 at 14:45

1 Answer 1

22
$\begingroup$

Here is a quick proof that the density in question exists and equals $$ c:=\frac{2}{3}\prod_{p\geq 5}\left(1-\frac{2}{p^2}\right)\approx 0.553087\ . $$

Let $f(d)$ denote the number of solutions of the congruence $n(432n+1)\equiv 0\pmod{d}$. Note that, for $p$ prime, $f(p^2)=1$ when $p<5$ and $f(p^2)=2$ when $p\geq 5$. Let $P\geq 2$ be fixed. By a simple inclusion-exclusion sieve combined with the Chinese remainder theorem, we see that the number of $n\leq x$ such that $n(432n+1)$ is not divisible by the square of any prime $p\leq P$, equals $$ x\prod_{p\leq P}\left(1-\frac{f(p^2)}{p^2}\right)+o_P(x)=x\prod_p\left(1-\frac{f(p^2)}{p^2}\right)+o_P(x)+O(x/P).$$ Observe that the infinite product on the right equals $c$. On the other hand, the number of $n\leq x$ such that $n(432n+1)$ is divisible by the square of some prime $p>P$, is at most $$ \sum_{p>P}O(x/p^2)=O(x/P).$$ Altogether we see that the number of $n\leq x$ such that $n(432n+1)$ is square-free, equals $cx+o_P(x)+O(x/P)$. In particular, both the lower and the upper density of these numbers equals $c+O(1/P)$. These quantities are independent of $P$, hence upon letting $P\to\infty$, we see that they both equal $c$. So the density exists and also equals $c$.

For related comments see the introduction of this 1953 paper by Erdős.

$\endgroup$

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.