English

On Ramanujan's formula for $\zeta(1/2)$ and $\zeta(2m+1)$

Number Theory 2021-06-10 v1

Abstract

Page 332 of Ramanujan's Lost Notebook contains a compelling identity for ζ(1/2)\zeta(1/2), which has been studied by many mathematicians over the years. On the same page, Ramanujan also recorded the series, \begin{align*} \frac{1^r}{\exp(1^s x) - 1} + \frac{2^r}{\exp(2^s x) - 1} + \frac{3^r}{\exp(3^s x) - 1} + \cdots, \end{align*} where ss is a positive integer and rsr-s is any even integer. Unfortunately, Ramanujan doesn't give any formula for it. This series was rediscovered by Kanemitsu, Tanigawa, and Yoshimoto, although they studied it only when rsr-s is a negative even integer. Recently, Dixit and the second author generalized the work of Kanemitsu et al. and obtained a transformation formula for the aforementioned series with rsr-s is any even integer. While extending the work of Kanemitsu et al., Dixit and the second author obtained a beautiful generalization of Ramanujan's formula for odd zeta values. In the current paper, we investigate transformation formulas for an infinite series, and interestingly, we derive Ramanujan's formula for ζ(1/2)\zeta(1/2), Wigert's formula for ζ(1/k)\zeta(1/k) as well as Ramanujan's formula for ζ(2m+1)\zeta(2m+1). Furthermore, we obtain a new identity for ζ(1/2)\zeta(-1/2) in the spirit of Ramanujan.

Keywords

Cite

@article{arxiv.2106.04797,
  title  = {On Ramanujan's formula for $\zeta(1/2)$ and $\zeta(2m+1)$},
  author = {Anushree Gupta and Bibekananda Maji},
  journal= {arXiv preprint arXiv:2106.04797},
  year   = {2021}
}

Comments

32 pages, Comments are welcome!