TetraNET advances neuromorphic computing by developing scalable and eco‑friendly ZnO tetrapod networks (TRL 3) that enable energy‑efficient in‑sensor preprocessing. As traditional computing faces limitations in energy use, scalability, and adaptability, neuromorphic hardware offers a promising low‑power alternative. However, current systems often depend on rigid, resource‑intensive fabrication methods, limiting flexibility and sustainability. TetraNET addresses this gap by introducing innovative materials and architectures designed for both performance and environmental responsibility.
Disordered neuromorphic networks have shown great potential due to their interconnected, adaptive behavior similar to biological neural systems. While nanoparticle and nanowire networks have demonstrated functional switching and distributed plasticity, each architecture has drawbacks—nanoparticles lack 3D connectivity, whereas nanowires require highly controlled structuring. ZnO tetrapod networks (ZnO‑TNs) provide an attractive intermediate solution, naturally forming 3D percolative networks with dense interaction points. Their inherent semiconductive properties and sensitivity to light and electric fields make them especially suited for hybrid sensing‑computing applications. Despite their promise, ZnO‑TN networks remain largely unexplored in neuromorphic devices.
TetraNET will demonstrate a proof‑of‑concept ZnO‑TN‑based in‑sensor preprocessor capable of real‑time feature extraction and adaptive signal processing, achieving low‑latency operation and ≤50 µJ energy per update (TRL 3). This technology will be evaluated on anomaly‑detection and edge‑computing tasks, establishing a roadmap toward higher TRLs. The project follows principles of green electronics, integrating sustainable fabrication methods and conducting a full life‑cycle assessment to minimize environmental impact. Through interdisciplinary collaboration, training, and secondments, TetraNET supports the development of a skilled research community and strengthens Europe’s position in sustainable neuromorphic computing.
