Neuromorphic computing based on integrated photonic circuits on-chip is a burgeoning area aiming to achieve high-speed, energy-efficient, and low-latency data processing to alleviate artificial intelligence-related applications, such as autonomous driving and speech recognition. Outstanding properties of phase change materials, such as reversible and fast switching of the states and high index contrast together with a wide spectral region integrated in nanophotonic devices provide a unique on-chip hybrid system to obtain fast modulators, switches and to realize neuromorphic systems on-chip.
We designed on-chip reconfigurable broadband nanocrystalline graphene-assisted waveguide switch and memory unit covering the whole telecommunication C-band based on absorption modulation of integrated Ge2Sb2Te5 (GST) PCM cells. GST state switching is triggered via patterned nanocrystalline graphene external microheaters placed on top of silicon nitride waveguide and beneath PCM cell.
|