On the concavity of Dirichlet's eta function and related functional inequalities (Q2257321)

Dirichlet's eta function on \((0, \infty)\) is defined by \[ \eta(s)= \sum_{j=1}^{\infty} \frac{(-1)^{j-1}}{j^s}. \] In [``The logarithmic concavity of \((1-2^{1-r})\zeta(r)\)'', J. Changsha Comm. Univ. 14, 1--15 (1998)], \textit{K. C. Wong} proved that \(\eta(s)\) is strictly logarithmically concave on \((0, \infty)\). The main result of the paper under review is that for every \(s \in (0, \infty)\), \[ \eta^{\prime \prime} (s) = \sum_{j=1}^{\infty} \frac{(-1)^{j-1} (\log j)^2}{ j^s} < 0, \] and thus it is established that \(\eta(s)\) is strictly concave on \((0, \infty)\). Several new inequalities satisfied by \(\eta(s)\) are also presented, for instance the following interesting double side inequality \[ \log(2) < \frac{\eta(x)^{\frac{1}{\sqrt{x}}} \eta(y)^{\frac{1}{\sqrt{y}}}}{\eta(xy)^{\frac{1}{\sqrt{xy}}}} <1 \] is established. In this inequality, it should be noted that both side bounds are sharp.