Excitability in a PhC Nanolaser with an Integrated Saturable Absorber
Maxime Delmulle, Bruno Garbin, Loredana Maria Massaro, Alexandre Bazin, Isabelle Sagnes, Konstantinos Pantzas, Sylvain Combrié, Alfredo De Rossi, Fabrice Raineri · 2023
Although machine learning (ML) algorithms are already applied in many fields (e.g. language recognition, temporal series prediction) using software computed on “Von Neumann” architectures, numerous technological advances (e.g. robotics, autonomous driving) require dedicated hardware. Integrated optics constitute a highly promising platform that could be harnessed for achieving portable ML chips with unprecedented power-efficiency and bandwidth [1]. Despite recent observations of ML key functionalities using integrated Silicon photonics, such as reconfigurable matrix multiplication [2] and nonlinear activation functions [3], excitability observation, the spiking mechanism of our biological neurons, remains incomplete in all-integrated systems; they are based on opto-thermal (slow) effects [4] that drastically limits its applicability in ML systems with high bandwidth requirements. Here, we use InP-based photonic crystal nanocavities heterogeneously integrated on top of a Silicon on Insulator (SOI) waveguide to demonstrate the first all-integrated (fast) excitable nanolaser.