Two Dirichlet's series related to the Divisor Summatory Function and to the Riemann's zeta-function, $\zeta(s)$

Your original observation is easily explained: $d(n)$ is odd if and only if $n$ is a square.


See my response here.