6
$\begingroup$

I am interested in obtaining an upper bound for $\prod_{p|N} (1 + 1/\sqrt{p})$ when $N$ is squarefree. It's not too hard to show that $$ \prod_{p\mid N} (1 + 1/\sqrt{p}) \ll C^{\omega(N)} \ll N^{\varepsilon} $$ for any $\varepsilon > 0$. I was wondering is it possibly to prove an upper bound $\ll \log N$ or some power of $\log N$? Any comments would be appreciated. Thank you.

$\endgroup$
12
$\begingroup$

The $N$ for which this product is largest will be of the form $N = \prod_{p<y} p$, so that $\log N \sim y$. For this $N$, $$ \log \prod_{p\mid N} \bigg( 1+\frac1{\sqrt p} \bigg) = \sum_{p\mid N} \log \bigg( 1+\frac1{\sqrt p} \bigg) \sim \sum_{p\mid N} \frac1{\sqrt p} = \sum_{p<y} \frac1{\sqrt p} \sim \frac{\sqrt y}{\log y} \sim \frac{\sqrt{\log N}}{\log\log N}, $$ and so $$ \prod_{p\mid N} \bigg( 1+\frac1{\sqrt p} \bigg) = \exp\bigg( (1+o(1)) \frac{\sqrt{\log N}}{\log\log N} \bigg) $$ for these $N$ (and the right-hand side will be an upper bound for all $N$). In particular, an upper bound of $(\log N)^A$ is too ambitious.

Reality check: replacing $1+1/\sqrt p$ by $1+1$ and carrying this computation through will recover the more familiar bound $$ 2^{\omega(N)} \le \exp\bigg( (\log2+o(1)) \frac{\log N}{\log\log N} \bigg). $$

$\endgroup$

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.