In a groundbreaking development towards advancing neuromorphic computing, engineers at Ecole Polytechnique Fédérale de Lausanne (EPFL) have achieved a significant milestone: executing logic operations using artificial nanofluidic synapses, heralding a potential paradigm shift in computational memory storage. Published in Nature Electronics, their research showcases the potential of nanofluidic-based neuromorphic computing, a brain-inspired approach that leverages ions instead of electrons for data processing.

Memory, a fundamental aspect of computing and cognitive function, is pivotal in facilitating the storage and retrieval of information. Unlike traditional computing, which relies on separate memory and processing units, the EPFL team’s approach aims to emulate the brain’s efficient computation directly on stored data, thereby mitigating the inefficiencies associated with the von Neumann bottleneck.

Lead researcher Aleksandra Radenovic, from EPFL’s Laboratory of Nanoscale Biology (LBEN), embarked on an ambitious endeavor to create a functional nanofluidic memristive device capable of mimicking the brain’s energy-efficient information processing. The device, developed by LBEN postdoctoral researcher Théo Emmerich and team, represents a significant advancement in scalability and performance compared to previous attempts.

At the heart of their innovation lies the utilization of ions, rather than electrons, to enable memory storage and computation. Memristors, capable of switching between two conductance states, have been adapted to exploit the diverse properties of ions. By immersing the device in an electrolyte water solution containing potassium ions, the researchers demonstrated the ability to modulate memory by varying ion compositions.