An international collaboration led by researchers at The University of Manchester’s National Graphene Institute has today announced a dramatic new step in neuromorphic computing. They have developed a new class of programmable nanofluidic memristors that effectively mimic the memory functions of the human brain. This breakthrough technology has the potential to power the next generation of computing systems that think and reason more like human cognition.
Professor Radha Boya heads the research, delving into the ways in which nanofluidic devices can switch between different memory loop patterns. This magic of transformation happens by retuning various experimental knobs. This research demonstrates just some of the amazing potential of these devices. It gives us a clean structure through which we can begin to parse and ultimately design future nanofluidic memory architectures.
Mechanisms Behind the Technology
These electrolytes are confined within dense arrays of thin nanochannels formed from two-dimensional materials, such as molybdenum disulfide (MoS₂) and hexagonal boron nitride (hBN). In contrast to conventional solid-state memristors based on electron transport, these new devices utilize liquid electrolytes to carry out their operation.
Such behavior can be actively manipulated, meaning researchers have control over how the nanofluidic devices behave. They accomplish this by adjusting parameters such as electrolyte composition, pH levels, voltage frequency, and channel geometry. The study reveals that the same device can exhibit four distinct memory loop styles: two “crossing” types and two “non-crossing” types.
Professor Radha Boya noted, “This is the first time all four memristor types have been observed in a single device.” This latest find showcases the incredible adaptability of these nanofluidic systems and their ability to reproduce elaborate brain-like behavior.
Implications for Neuromorphic Computing
This research has very real implications that go well beyond academic use. The ability of these nanofluidic memristors to “forget” information over time or retain it for days, depending on applied voltage and electrolyte conditions, opens up exciting possibilities for low-power, adaptive computing systems.
Dr. Abdulghani Ismail, the lead author of the study, emphasized the significance of these findings: “It opens up exciting possibilities for low-power, adaptive computing systems that operate more like the human brain.” This adaptive capability makes these devices prime candidates for the next wave of artificial intelligence and machine learning breakthroughs.
A Unified Framework for Future Research
This platonic research has created a model that duplicates easily and effectively all four memristive loop types. This accomplishment lays an exciting groundwork for subsequent research into advanced nanofluidic memory architectures. Incorporating a unifying framework like this one could help researchers design more efficient and versatile memory devices that better emulate the biological processes.
This study was recently published in the journal Nature Communications and is freely available online under DOI 10.1038/s41467-025-61649-6.
