I've found this identity of a blog: $$\sum _{n=1}^{\infty } \frac{\left(\frac{4}{9}\right)^n \beta (2 n+1)}{n+1}=\frac{-9 \left(-\frac{2 C}{3}-1+\frac{1}{3} \pi \log \left(\sqrt{3}+2\right)\right)}{2 \pi }$$ Where $\beta$ denotes Dirichlet Beta function (i.e. A L-function modulo 4). Unfortunately, this result seems to be incorrect: numerical approximations failed disastrously. My question is: Is there a closed-form for this sum? After all, there are so many Zeta sums with fabulous closed-forms (see e.g. Srivastava's review), and this made me wonder if the given result can be modified to be correct. I tried reverse the order of summation but it lead to nothing valuable. Therefore I'd like you to help me. Thank you all!


Here is a sketch. Using, say, the link with Euler numbers, one gets $$\sum_{n=0}^{\infty}\beta(2n+1)x^{2n}=\frac{\pi}{4}\sec\frac{x\pi}{2},$$ which implies $$\sum_{n=0}^{\infty}\frac{\beta(2n+1)}{2n+2}x^{2n+2}=\frac{1}{\pi}f\left(\frac{x\pi}{2}\right),\quad f(y)=\int_0^y\frac{x\,dx}{\cos x}.$$ The given sum is then equal to $\dfrac{9}{2\pi}f\left(\dfrac{\pi}{3}\right)-\dfrac{\pi}{4}$, with the known value of $$f\left(\frac{\pi}{3}\right)=\frac{\pi}{3}\ln(2+\sqrt{3})-\frac{2}{3}G,$$ where $G$ is Catalan's constant. The latter formula can be derived from $$\int_0^y\ln\left|\tan\frac{x}{2}\right|\,dx=-2\sum_{n=0}^{\infty}\frac{\sin(2n+1)y}{(2n+1)^2}.$$

  • $\begingroup$ Is your $G$ the same as OP's $C$? $\endgroup$ – Gerry Myerson Oct 17 '19 at 11:26
  • $\begingroup$ A neat but clever answer. Thank you! $\endgroup$ – 雄凤山 Oct 17 '19 at 15:03
  • $\begingroup$ @GerryMyerson Yes it is. $\endgroup$ – 雄凤山 Oct 17 '19 at 15:04

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.