I'm currently stuck on an explanation that works with a power series expansion of: $ \frac{a}{\eta}\: \frac{1}{e^{\frac{\eta}{a}}-1}$
As you can see, the function diverges around $\eta=0$, which is exactly where I need to investigate its behaviour.
This step is taken in the calculation of the Casimir force (quantum physics).
Sadly, I seem to be unable to find the correct way to expand this. The power series expansion is differentiated twice and multiplied by some prefactor, yielding something $ \propto \left( \frac{3a}{\eta^4} - \frac{1}{2\eta^2} - \frac{1}{720a^3} + \mathscr{O}\left[\eta\right] \right)$.
I hope this provides enough information. Any sort of answer including just an instruction would be greatly appreciated.