I am trying to compute the inverse Mellin transform of : $$\sum_{n=0}^{\infty}\frac{\Gamma(n+s)\zeta(n+s)\zeta(n+1+s)}{\zeta(s)n!}\left(-\omega\right)^{n}$$ w.r.t. the complex number $s$. $\omega$ being a real parameter.

My Attempt :

it can be easily verified that the function $\phi(s,\omega)$ given by : $$\phi(s,\omega)=\frac{(s-1)}{\Gamma(s)^{2}}\sum_{n=0}^{\infty}\frac{\Gamma(n+s)\zeta(n+s)\zeta(n+1+s)}{n!}\left(-\omega\right)^{n}$$

is entire in $s$. Thus, the Mellin inverse may be written as : $$\frac{1}{2\pi i }\int_{\sigma-i\infty}^{\sigma+i\infty}\frac{\Gamma(s)^{2}\phi(s,\omega)}{(s-1)\zeta(s)}x^{-s}ds$$ which can be computed using the residue theorem. The problem now is to find $\phi(s,\omega)$, and it's derivatives at negative integers, and the non-trivial zeros of the Riemann zeta function. hence my question.


$\phi(s,\omega)$ is complicated, it grows very fast on vertical lines, it won't help.

For $\Re(s) > 1$

$$\zeta(s)\zeta(s+1) = \sum_{n=1}^\infty a_n n^{-s}, a_n = \sum_{d | n} d^{-1}$$

$$F(s)=\Gamma(s)\zeta(s)\zeta(s+1) = \sum_{n=1}^\infty a_n \int_0^\infty x^{s-1} e^{-nx}dx=\int_0^\infty x^{s-1} \sum_{n=1}^\infty a_n e^{-nx}dx$$ For $|\omega| < 1$ $$\sum_{k=0}^\infty \frac{1}{k!} F(s+k) (-\omega)^k =\sum_{k=0}^\infty \frac{1}{k!} (-\omega)^k \int_0^\infty x^{s+k-1} \sum_{n=1}^\infty a_n e^{-nx}dx $$ $$= \int_0^\infty x^{s-1} \sum_{k=0}^\infty \frac{1}{k!} (-\omega x)^k \sum_{n=1}^\infty a_n e^{-nx}dx = \int_0^\infty x^{s-1} e^{-\omega x} \sum_{n=1}^\infty a_n e^{-nx}dx$$ Note the latter is analytic in $\omega, \Re(\omega) > -1$ even if the power series in $\omega$ diverges.

Can you finish from there ?

  • $\begingroup$ It might be the case that $\phi(s,\omega)$ grows very fast on vertical lines, but we have the $\Gamma(s)^{2}$ factor, which decays extremely fast in the half-plane $\Re(s)<0$ in any vertical strip. won't that cancel the growth of $\phi(s,\omega)$ ? $\endgroup$ – Mohammad Al Jamal Feb 18 at 8:11
  • $\begingroup$ @MohammadAlJamal Do you follow my answer ? Can you finish it or not ? If it was easier to use the residue theorem I would use it, but it is not because $\phi$ grows too fast on vertical lines and because there is a much simpler solution. $\endgroup$ – reuns Feb 19 at 5:03
  • $\begingroup$ i do follow your answer, and i can finish. the answer will be an infinite sum in terms of the Mobius function and the log of the Euler function - related to the Dedekind eta - but i am looking for an answer in terms of the non-trivial zeros of the zeta function. $\endgroup$ – Mohammad Al Jamal Feb 19 at 6:58
  • $\begingroup$ @MohammadAlJamal What do you get for $h_\omega$ the inverse Mellin transform of $H_\omega(s)=\sum_{n=0}^{\infty}\frac{\Gamma(n+s)\zeta(n+s)\zeta(n+1+s)}{\zeta(s)n!}\left(-\omega\right)^{n}$ then what do you want to know about $h$ and $H$. $\endgroup$ – reuns Feb 19 at 7:03
  • $\begingroup$ If $U$ is entire and $\frac{U(s) \Gamma(s)}{\zeta(s)}$ decreases fast enough as $Im(s) \to \infty$ and $\Re(s) \to \infty$ and the zeros of $\zeta(s)$ are all simple then what do you get for $\frac{1}{2\pi i }\int_{\sigma-i\infty}^{\sigma+i\infty} \frac{U(s)\Gamma(s)}{\zeta(s)} x^sds$ using the residue theorem. $\endgroup$ – reuns Feb 19 at 7:13

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.