2
$\begingroup$

I am interested in the rational zeta series as title. WolframAlpha gives the result: $$ \sum_{n=2}^{\infty}\frac{\zeta(n)-1}{n-1}=0.78853056591150896106... $$ This is the Lüroth analog of Khintchine’s constant, which is defined as: $$ \sum_{n=1}^{\infty}\frac{\ln (n)}{n(n+1)}=-\sum_{n=2}^{\infty}(-1)^{n}\zeta^{'}(n) $$ Why this rational zeta series is not in terms of $\gamma$, $\ln2$, $\ln \pi$ or normal constant that usually seen, but comes out an unusual constant? How to evaluate it to a closed form?

Also how to evaluate its companion zeta series to a closed form? $$ \sum_{n=2}^{\infty}\frac{(-1)^{n}\left ( \zeta(n)-1 \right )}{n-1}=0.56459970638442432059... $$ Or there are no closed forms for both zeta series?

$\endgroup$
6
  • $\begingroup$ Why would you sysoect a closed form exists? $\endgroup$ – Mark Viola Jul 12 '20 at 1:41
  • $\begingroup$ I am not sure if a closed form exists. It's a relatively simple form zeta series, and should have been studied in depth. That's why I asked this question to see if some experts know its closed form, or know that there is no closed form. $\endgroup$ – Nanhui Lee Jul 12 '20 at 1:51
  • $\begingroup$ Series much simpler than this have no (known) closed form, Nanhui, e.g., $\sum^{\infty}n^{-3}$. $\endgroup$ – Gerry Myerson Jul 12 '20 at 1:58
  • 1
    $\begingroup$ The first series is discussed at oeis.org/A085361 where there is a link to 1,000 decimal places & some other information but no closed form is given (which strongly suggests none is known). $\endgroup$ – Gerry Myerson Jul 12 '20 at 2:01
  • 1
    $\begingroup$ There is a simple rule (works most of the time) to determine whether such series have a "closed form": for infinite sum $$\sum_n \frac{\log^k(n+b)}{(n+a)^s}$$ there is no simple answer when $a\neq b$. In your case, you have $\sum_n \log n / (1+n)$, this precludes nice closed-form. You need to resort to some peculiar function like poly-Hurwitz zeta for this. $\endgroup$ – pisco Aug 4 '20 at 5:56
1
$\begingroup$

Although there are no closed forms for both rational zeta series, we can evaluate them with definite integrals.

Set $f(x)$ as \begin{align*} f(x)&=\sum_{n=2}^{\infty}\frac{\zeta(n)-1}{n-1}x^{n-1}=\sum_{n=2}^{\infty}\sum_{k=2}^{\infty}\frac{1}{k^n}\int_{0}^{x}t^{n-2}\,dt=\int_{0}^{x}\sum_{n=2}^{\infty}\sum_{k=2}^{\infty}\frac{t^{n-2}}{k^n}\,dt \\ &=\int_{0}^{x}\sum_{k=2}^{\infty}\frac{1}{k^2}\sum_{n=2}^{\infty}\left( \frac{t}{k} \right)^{n-2}dt=\int_{0}^{x}\sum_{k=2}^{\infty}\frac{1}{k^2}\frac{1}{\left(1-\frac{t}{k}\right)}\,dt \\ &=\int_{0}^{x}\left(\frac{1}{t-1}+\sum_{k=1}^{\infty}\frac{1}{k}\frac{1}{(k-t)} \right)dt=\int_{0}^{x}\frac{1}{t-1}\,dt-\int_{0}^{x}\frac{\gamma+\psi(1-t)}{t}\,dt \\ &=\ln(1-x)+\left [ \frac{\ln\Gamma(1-t))}{t} \right ]_{t=0}^{t=x}+\int_{0}^{x}\left ( \frac{\ln\Gamma(1-t)}{t^2}-\frac{\gamma}t{} \right )dt \\ &=\ln(1-x)+\frac{\ln\Gamma(1-x)}{x}-\gamma+\int_{0}^{x}\left ( \frac{\ln\Gamma(1-t)}{t^2}-\frac{\gamma}t{} \right )dt \\ \\ f(1)&=\sum_{n=2}^{\infty}\frac{\zeta(n)-1}{n-1} \\ &=\lim_{x\rightarrow 1^{-}}\left ( \ln(1-x))+\frac{\ln\Gamma(1-x))}{x} \right )-\gamma+\int_{0}^{1}\left ( \frac{\ln\Gamma(1-t)}{t^2}-\frac{\gamma}t{} \right )dt \\ &=-\gamma+\int_{0}^{1}\left ( \frac{\ln\Gamma(1-t)}{t^2}-\frac{\gamma}t{} \right )dt \approx0.7885306 \\ \end{align*} In the same way, set $f(x)$ as \begin{align*} f(x)&=\sum_{n=2}^{\infty}\frac{(-1)^{n}(\zeta(n)-1)}{n-1}x^{n-1}=\sum_{n=2}^{\infty}\sum_{k=2}^{\infty}\frac{(-1)^{n}}{k^n}\int_{0}^{x}t^{n-2}\,dt=\int_{0}^{x}\sum_{n=2}^{\infty}\sum_{k=2}^{\infty}\frac{(-t)^{n-2}}{k^n}\,dt \\ &=\int_{0}^{x}\sum_{k=2}^{\infty}\frac{1}{k^2}\sum_{n=2}^{\infty}\left( -\frac{t}{k} \right)^{n-2}dt=\int_{0}^{x}\sum_{k=2}^{\infty}\frac{1}{k^2}\frac{1}{\left(1+\frac{t}{k}\right)}\,dt \\ &=\int_{0}^{x}\left(-\frac{1}{t+1}+\sum_{k=1}^{\infty}\frac{1}{k}\frac{1}{(k+t)} \right)dt=-\int_{0}^{x}\frac{1}{t+1}\,dt+\int_{0}^{x}\frac{\gamma+\psi(1+t)}{t}\,dt \\ &=-\ln(1+x)+\left [ \frac{\ln\Gamma(1+t))}{t} \right ]_{t=0}^{t=x}+\int_{0}^{x}\left ( \frac{\ln\Gamma(1+t)}{t^2}+\frac{\gamma}t{} \right )dt \\ &=-\ln(1+x)+\frac{\ln\Gamma(1+x)}{x}+\gamma+\int_{0}^{x}\left ( \frac{\ln\Gamma(1+t)}{t^2}+\frac{\gamma}t{} \right )dt \\ \\ f(1)&=\sum_{n=2}^{\infty}\frac{(-1)^{n}(\zeta(n)-1)}{n-1} \\ &= -\ln(2)+\gamma+\int_{0}^{1}\left ( \frac{\ln\Gamma(1+t)}{t^2}+\frac{\gamma}t{} \right )dt \approx0.5645997 \\ \end{align*}

$\endgroup$

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.