Sorry I'm really out of any sort of number theory... is this a known product? \begin{align} P & = \prod_{n=2}^\infty \left(1-\dfrac1{n^p}\right) \end{align} As a special case I would be interested in $p=3$. Thank you very much

  • 1
    $\begingroup$ It can be written in terms of the gamma function. $\endgroup$ – Sangchul Lee Feb 18 '16 at 7:02
  • $\begingroup$ that would work, can you tell me how? $\endgroup$ – Dac0 Feb 18 '16 at 7:03
  • 1
    $\begingroup$ The basic idea is to factor $(1-n^{-p})$ into linear terms and utilize the Weierstrass product for $\Gamma(z)$ to simply the resulting expression. For a technical reason it is easier to work with partial products first. $\endgroup$ – Sangchul Lee Feb 18 '16 at 7:08
  • $\begingroup$ Of course the valuable fact is an understandable proof, which I hope see from an user to learn it. But here you have a tool for the future (type product (1-1/k^3), k=2 to infinity) in INFINITE PRODUCTS here wolframalpha.com/input/… $\endgroup$ – user243301 Feb 18 '16 at 7:25
  • $\begingroup$ Just give the infinite product to maple and you will get a closed form! $\endgroup$ – Mhenni Benghorbal Feb 18 '16 at 7:26

From the Weierstrass product for $\Gamma(z)$, we can easily find that

$$ \frac{e^{(\gamma-1)z}}{\Gamma(2-z)} = \prod_{n=2}^{\infty} \left( 1 - \frac{z}{n} \right) e^{z/n}. $$


$$ \prod_{n=2}^{\infty} \left(1 - \frac{1}{n^p}\right) = \prod_{n=2}^{\infty} \prod_{\omega : \omega^p = 1} \left( 1 - \frac{\omega}{n} \right) e^{\omega/n} $$

and interchanging the order of summation, we have

$$ \prod_{n=2}^{\infty} \left(1 - \frac{1}{n^p}\right) = \prod_{\omega : \omega^p = 1} \frac{1}{\Gamma(2-\omega)}. $$

For some special choice of $p$ (such as $p = 2$ or $p = 3$), this can be simplified further.

  • $\begingroup$ Great Answer!!! $\endgroup$ – Dac0 Feb 18 '16 at 7:44

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.