Thank you for the question and interest in the work. Charge injection etc could be an issue. This was part of the motivation for a differential implementation in order to at least partly reduce these effects. With regard to output signal size, these simulations were done with a very small input. At the sensitivity limit the output will be larger. For example with a gain of 55kOhms, an input of 20 uA leads to a pp differential swing of about 1 V. The modulation signal does need to be matched, but anything within a 45 degree range was close enough.

Zambia95 Can you please elaborate as to how that phenomenon works. I can't, for the life of me, figure out how the voltage spike drops back to the supply voltage after a moment. Also how does the spike at Vss go below 0 when Vss is 0.