Scientists uncover hidden mechanism that could improve future AI-inspired computing

A team of researchers in Japan has shed new light on a long-standing mystery in condensed matter physics, revealing how heat generated by electrical current can trigger a complex form of self-organisation inside an organic conductor. The discovery could have important implications for the development of future memory technologies and neuromorphic computing systems, which aim to mimic the structure and function of the human brain.

The research, led by Professor Tetsuaki Itou of the Tokyo University of Science (TUS), was recently published in Physical Review Applied and selected as an Editors’ Suggestion, a distinction reserved for papers judged to be particularly significant.

The study addresses a phenomenon known as volatile resistive switching, in which a material suddenly changes its electrical resistance when subjected to an electric current. Such behaviour has attracted growing interest because it could form the basis of next-generation electronic devices including resistive memories, advanced sensors and neuromorphic processors designed for artificial intelligence applications.

Understanding resistive switching

Many modern electronic devices depend on the ability to control electrical resistance. In conventional circuits, resistance typically follows Ohm’s Law, in which voltage and current maintain a predictable proportional relationship. Certain materials, however, exhibit far more unusual behaviour. In compounds that undergo a metal-insulator transition (MIT), the material can switch between conducting electricity efficiently and acting as an insulator. The trigger can be temperature, pressure, electrical current or other external factors.

Researchers have long known that some materials exhibit abrupt resistance changes close to these transition points. Yet the exact mechanism behind volatile resistive switching has remained unclear. According to the TUS researchers, one reason for this uncertainty is that much of the previous work has focused on thin-film inorganic materials. In these systems, heat generated by electrical currents rapidly dissipates into surrounding substrates, making it difficult to understand the precise relationship between temperature, electrical transport and phase transitions.

To address this challenge, the research team adopted a different strategy. Rather than investigating a conventional thin film, they focused on a bulk organic conductor called(d7-DMe-DCNQI)₂Cu, a deuterated derivative of N,N′-dicyanoquinonediimine. This material experiences a particularly sharp metal-insulator transition around 79 Kelvin, equivalent to approximately -194°C.   The researchers suspended a needle-like crystal of the material inside a Teflon tube and supplied current through gold wires attached to each end. This arrangement minimised heat loss, allowing the team to observe thermal effects that might otherwise remain hidden.

Crucially, the researchers employed proton nuclear magnetic resonance (¹H-NMR) spectroscopy to examine what was occurring inside the material during switching. NMR is more commonly associated with chemistry and medical imaging, but it can also provide highly detailed information about the microscopic state of materials.   The results revealed that the intermediate resistance state is not a simple transition from one phase to another. Instead, metallic and insulating regions coexist simultaneously within the material.

The surprising role of heat

One of the most significant discoveries involved the role of Joule heating, the heat produced when electrical current flows through a conductor. Traditionally, Joule heating has often been viewed as a secondary consequence of electrical activity. However, the new study suggests that it plays a much more active role. Using NMR measurements, the researchers observed a phenomenon they describe as temperature locking. Below the metal-insulator transition temperature, the material remained pinned close to the transition point despite changes in the surrounding temperature.  In effect, Joule heating raised the sample temperature above the ambient environment and maintained it near the critical transition temperature. This thermal balance created stable conditions for the coexistence of metallic and insulating phases, helping sustain the resistive-switched state.

The most unusual finding may be that the material displayed behaviour resembling an inverse Ohm’s Law. Rather than exhibiting a simple proportional relationship between voltage and current, the system generated an inverse relationship under certain conditions. According to the researchers, this emerged because Joule heating and heat dissipation continuously balanced one another while the material remained locked near the transition temperature.

The team found evidence that a metallic current filament formed inside the crystal. This conductive pathway effectively self-organised within the bulk material and adjusted its thickness as the applied current increased or decreased.  As the filament grew or shrank, the electrical properties of the material changed in a highly non-linear manner. This behaviour suggests that resistive switching is not simply the result of uniform heating but rather a dynamic interaction between heat flow, phase transitions and electrical conduction.

Why this matters for future computing

Although the research is rooted in fundamental physics, its implications extend into emerging computing technologies. One area of particular interest is resistive memory, sometimes referred to as memristive technology. Unlike traditional memory devices, resistive memories can store information through changes in resistance and may offer advantages in speed, energy consumption and durability.

The findings may also be relevant to neuromorphic computing, a field that seeks to build computer architectures inspired by biological neural networks. Conventional computing systems separate memory and processing functions. By contrast, the human brain integrates storage and processing in highly interconnected networks. Researchers hope that neuromorphic systems could deliver substantial improvements in energy efficiency while enabling more sophisticated forms of machine learning and artificial intelligence. Understanding how resistive switching arises, and how it can be controlled, is a crucial step towards making such technologies practical. Professor Itou notes that the newly discovered temperature-locking effect may offer a route to developing more durable and efficient switching devices.

Perhaps the most important message from the study is that resistive switching cannot be understood through simple models of electrical heating alone. The researchers conclude that the phenomenon represents a nonequilibrium steady state, in which heat flow, electrical transport and phase transitions become tightly coupled through thermal self-organisation. This insight not only advances understanding of organic conductors but could also influence future work on a broad class of materials used in electronic devices.