Encoding adaptive intelligence in molecular matter by design

For more than 50 years, scientists have sought to find alternatives to silicon to build molecular electronics. The vision was elegant; the reality proved far more complex. Within a device, molecules behave not as orderly textbook entities but as densely interacting systems where electrons flow, ions redistribute, interfaces evolve, and even subtle structural variations can induce strongly nonlinear responses. The promise was compelling, yet predictive control remained elusive. Meanwhile, neuromorphic computing – hardware inspired by the brain – has followed a parallel ambition: to discover a material that can store information, compute, and adapt within the same physical substrate and in real time. Yet today’s dominant platforms, largely based on oxide materials and filamentary switching mechanisms, continue to behave as engineered machines that emulate learning, rather than as matter that intrinsically embodies it.

A new study from the Indian Institute of Science (IISc) suggests that these two long-standing challenges may finally converge.

In a collaboration spanning chemistry, physics, and electrical engineering, a team led by Sreetosh Goswami, Assistant Professor at the Centre for Nano Science and Engineering (CeNSE) has created tiny molecular devices that can be tweaked to perform diverse functions. The same device can behave as a memory unit, a logic gate, a selector, an analog processor or an electronic synapse, depending on how it is stimulated. “It is rare to see adaptability at this level in electronic materials,” says Sreetosh Goswami. “Here, chemical design meets computation, not as an analogy, but as a working principle.”

This shapeshifting is powered by unique chemistry used to build and tweak these devices. The team synthesised 17 carefully designed ruthenium complexes and analysed how minute variations in molecular geometry and ionic surroundings sculpted electron behaviour. By carefully tweaking the ligands and ions arranged around the ruthenium molecules, the authors showed that the same device can exhibit many types of dynamic behaviour – switching from digital to analog, for instance – across a wide range of conductance values.

The molecular synthesis was carried out by Pradip Ghosh, Ramanujan Fellow and Santi Prasad Rath, former PhD student at CeNSE. Device fabrication was led by Pallavi Gaur, first author and PhD student at CeNSE. “What surprised me was how much versatility was hidden in the same system,” says Gaur. “With the right molecular chemistry and environment, a single device can store information, compute with it, or even learn and unlearn. That’s not something you expect from solid-state electronics.”

Understanding why this occurs required a long-missing element in molecular electronics: a rigorous theoretical foundation. The team developed a transport framework grounded in many-body physics and quantum chemistry, capable of predicting function from molecular structure. This enabled them to map how electrons traverse the molecular film, how individual molecules undergo oxidation and reduction, and how counterions rearrange within the molecular matrix, together governing the switching and relaxation dynamics and the stability of each molecular state.

Crucially, the unique adaptability of these complexes allows incorporating both memory and computation in the same material – this can lead to neuromorphic hardware in which learning can be encoded into the material itself. The team is already working on integrating such materials onto silicon chips, with the aim of developing future AI hardware that is both efficient and intrinsically intelligent.

“This work shows that chemistry can be an architect of computation, not just its supplier,” says Sreebrata Goswami, Visiting Scientist at CeNSE and co-author on the study who led the chemical design.