Evaluation of two self-adaptive random bit extractors for a quantum random number generator

We present a monolithic Quantum Random Number Generator (QRNG) implemented in 22-nm standard CMOS technology and based on the dark events generated by a Single-Photon Avalanche Diode (SPAD) in dark conditions. This work proposes a novel self-adjusting methodology based on the estimation of the median value to generate random bits. We outline a comparison of two approaches: one based on the inter-arrival time between events and the other based on the numbers of events counted in a fixed time interval. The research highlights the advantages and disadvantages of the two implementations in terms of speed and random numbers quality.