1

Introduction

I was playing around with the zeta function when I thought of this series:

$$\sum_{n=2}^{\infty} \ln(\zeta(n))$$

According to WolframAlpha this sum converges and gives the following approximation:

$$\sum_{n=2}^{\infty} \ln(\zeta(n))\approx 0.83067$$


Questions

  1. Is this a "known" series from before? I find this series quite cool and I thought I would find it somewhere but I didn't.
  2. WolframAlpha says it confirmed the convergence by the root test, can anybody confirm this?
  3. What's the closed-form of what the series converges to (exact value)?


Extra: My attempts

I tried to solve the 3rd question, but I don't have much ideas. An attempt I made was to note that:

\begin{align} \sum_{n=2}^{\infty} \ln(\zeta(n)) &= \sum_{n=2}^{\infty} \ln(\prod_p \frac{1}{1-p^{-n}}) \\\\ & = \sum_{n=2}^{\infty} \sum_p \ln(\frac{1}{1-p^{-n}})) \\\\ & = -\sum_{n=2}^{\infty} \sum_p \ln(1-p^{-n})) \end{align}

...but it didn't get me anywhere. Wikipedia also lists some integrals that are equal to $\ln(\zeta(n))$, but I'm not that familiar with them. Here's an example:

$$\displaystyle \ln \zeta (s)=s\int _{0}^{\infty }{\frac {\pi (x)}{x(x^{s}-1)}}\,\mathrm {d} x$$ And also: $$\displaystyle \ln \zeta (s)=s\int _{0}^{\infty }J(x)x^{-s-1}\,\mathrm {d} x$$ where $J(x)$ is the Riemann prime-counting function. I doubt that these are going to help, but I'll just put them here.

I'm not experienced in evaluating infinite series so I don't have anymore ideas. It would be highly appreciated if some of you could shed a little bit of light on this question!

Thanks in advance!


EDIT 1:

As Enver noted, this question is quite similar because my series is equal to $\ln(v)$ in the other question.

EDIT 2: Using basically the same method as in this question, we obtain (using one of the equations above):

\begin{align} \sum_{n=2}^{\infty} \ln(\zeta(n)) &= \sum_{n=2}^{\infty} n\int _{0}^{\infty }J(x)x^{-n-1}\,\mathrm {d} \\\\ & = \int _{0}^{\infty}J(x)(\sum_{n=2}^{\infty}\frac{n}{x^{n+1}})\,\mathrm {d} x \\\\ &= \int_0^{\infty} J(x) \frac{2x-1}{(x-1)^2x^2}\,\mathrm {d} x\end{align}

This is a quite interesting result which I believe is correct, but I'm still not satisfied. We can do better!

  • 1
    See this: https://math.stackexchange.com/questions/3237676/what-can-be-said-about-prod-s-2-infty-zetas?rq=1 – eeen Apr 15 '20 at 12:39
  • @Enver I'm unsure how much that question helps this one, but yeah sure, I guess we have at least proved that it converges. – Casimir Rönnlöf Apr 15 '20 at 12:49
  • 1
    In the post I sent, your series is $\ln v$. I don't think there is a closed form for $\sum_{n=2}^\infty \ln(\zeta(n))$. – eeen Apr 15 '20 at 12:55
  • @Enver Oh woops, I'm sorry, I didn't see that. – Casimir Rönnlöf Apr 15 '20 at 12:57
  • @Enver I'm really unsatisfied now that this probably doesn't have a closed form :/ – Casimir Rönnlöf Apr 15 '20 at 13:00
  • I feel that man :( – eeen Apr 15 '20 at 13:01
  • You can get properly sized parentheses (and other paired delimiters) that adjust to the size of their content by preceding them with \left and \right. – joriki Apr 16 '20 at 15:15
  • 2
    It's probably a good idea to internalize the truth that very few convergent series have closed forms. We shouldn't expect a randomly chosen series to have a nice evaluation. – Greg Martin Apr 16 '20 at 15:34
  • This is essentially a duplicate of What can be said about $\prod_{s=2}^{\infty} \zeta(s) $? and should be closed as such. – joriki Apr 16 '20 at 15:36
  • @joriki Enver already posted the question here in the comments, thank you still. Thank you also for the Latex tip, I shall change the parantheses. – Casimir Rönnlöf Apr 16 '20 at 15:38
  • @GregMartin Yeah I know, I shall remember this in the future. It just felt like this was such a "basic" series that somebody would have figured out a closed form for it, so I was a little surprised. – Casimir Rönnlöf Apr 16 '20 at 15:41
  • 1
    @CasimirRönnlöf: Yes, I saw that, but he didn't vote to close as a duplicate. I did. In that case the system automatically generates a comment saying "Does this answer your question?". Since it had already been posted, that comment wouldn't have made much sense, so I edited it to explain that I think this should be closed as a duplicate. It makes things easier to find on the site if there's one place where a question is answered. Any new answers prompted by this question can be posted as answers of the other question. – joriki Apr 16 '20 at 15:41
  • @joriki Alright, thank you. I'm going to go trough the other question a little more carefully and then decide whether it answers my question (which it probably does). – Casimir Rönnlöf Apr 16 '20 at 15:48
  • @Joriki By the way do you have any idea if the stament in my 2nd Edit is correct? – Casimir Rönnlöf Apr 16 '20 at 15:49
  • @CasimirRönnlöf: It looks good to me. – joriki Apr 16 '20 at 16:00
  • @joriki thanks! – Casimir Rönnlöf Apr 16 '20 at 16:01
  • @joriki Here's what I think after reading a bit more carefully through it. The question is similar in the sense that my question is just $\ln(v)$, but (pardon me if this sounds stupid) I think that my question is more of a "sequence and series" question and asks more about the convergence of the series. I believe this series is still such a "simple" one, in the sense that others might consider it and have the same question (and therefore landing here), but that they may not find find the other question even though it's just my series raised to the $e$ power. – Casimir Rönnlöf Apr 16 '20 at 16:25
  • @joriki I don't have anything against closing the question and if you want it I can do it, but I think this question might lead some other people to the other question. – Casimir Rönnlöf Apr 16 '20 at 16:26
  • @CasimirRönnlöf: I don't think you can single-handedly close the question as a duplicate. As far as I'm aware, you can only delete it, which is not what I'm suggesting. If it's closed as a duplicate, it will still show up in the search results and will indeed lead people to the other question. It's just that all the answers will then be collected there and not spread over two questions. – joriki Apr 16 '20 at 16:52
  • @joriki Yeah okay, I misunderstood. I'll click the yes button on pop-up. Thanks for the help. – Casimir Rönnlöf Apr 16 '20 at 17:02
  • Ah, interesting, I always wondered what it means when it says that Community closed a question – so you got a popup asking you whether it's OK to close this as a duplicate? – joriki Apr 16 '20 at 17:42
  • @joriki The popup was above the title and basically just said something like "Some community members suggests this is a duplicate" and then it asked "Does this question (the link here) answer your question?". Then there were just the options Yes/No – Casimir Rönnlöf Apr 16 '20 at 17:47

0 Answers0