Primordial Black Holes (PBHs) have recently attracted much attention as they may explain some of the LIGO/Virgo/KAGRA observations and significantly contribute to the dark matter in our universe. The next generation of Gravitational Wave (GW) detectors will have the unique opportunity to set stringent bounds on this putative population of objects. Focusing on the Einstein Telescope (ET), in this paper we analyse in detail the impact of systematics and different detector designs on our future capability of observing key quantities that would allow us to discover and/or constrain a population of PBH mergers. We also perform a population analysis, with a mass and redshift distribution compatible with the current observational bounds. Our results indicate that ET alone can reach an exquisite level of accuracy on the key observables considered, as well as detect up to tens of thousands of PBH binaries per year, but for some key signatures (in particular high-redshift sources) the cryogenic instrument optimised for low frequencies turns out to be crucial, both for the number of observations and the error on the parameters reconstruction. As far as the detector geometry is concerned, we find that a network consisting of two separated L-shaped interferometers of 15 (20) km arm length, oriented at 45° with respect to each other performs better than a single triangular-shaped instrument of 10 (15) km arm length, for all the metrics considered.