The manufacture and disposal of modern electronic devices impose a significant environmental burden, while processing large-scale sensory data requires substantial energy consumption. Achieving truly sustainable electronics therefore demands a multifaceted approach: (i) low-waste manufacturing; (ii) material selection that prioritizes green synthesis and end-of-life recyclability; and (iii) adoption of energy-efficient edge-computing architectures enabling event-driven, parallel information processing. The advances in fully additive manufacturing, sustainable materials, and neuromorphic device engineering each address one key bottleneck towards fully sustainable electronics. However, the convergence of these separate domains, tackling the problem from cradle to grave, remains a critical technological gap. In this context, we present an array of fully printed, flexible, and sustainably manufactured synaptic Organic Electrochemical Transistors (OECTs), serving as the fundamental elements for parallel, asynchronous (thus, energy-efficient) processing of a spiking neural network (SNN) for sensory data interpretation. The presented OECTs emulate synaptic behavior with high bio-plausibility. Presynaptic current pulses at the gate induce analog conductance tuning with high linearity and repeatability among OECTs. This consistency is typically hard to achieve in printed devices, for which network simulations reported in literature are based on single device characteristics, while it is crucial to achieve an SNN with high accuracy [1].

The synaptic OECTs fabrication is governed by sustainability criteria. Manufacturing is based on low-waste inkjet printing, and the material is governed by non-toxicity and green ink formulation. The material stack features a flexible bio-based ethyl cellulose substrate, electrodes printed with water-based gold inks, and a printed channel of green-synthesized Organic Mixed Ionic–Electronic Conductor (OMIEC) poly[2-(3,3′-bis(2-(2-(2-methoxyethoxy)ethoxy)ethoxy)-[2,2′-bithiophen]-5-yl)thieno[3,2-b]thiophene] (p(g2T-TT)).

The so-fabricated synaptic OECTs exhibit analog modulation of the channel conductance in a large window (I_ON/I_OFF > 10³), demonstrating potentiation and depression analogous to biological synapses. The magnitude of the conductance change ΔG scales with pulse amplitude, duration, number, and frequency, thus implementing synaptic plasticity functions needed real-time adaptive learning. All devices show a common window of more than 40 linear conductance states (with deviation from linearity <5%). This, together with the symmetric conductance update, where potentiation and depression occur at symmetric ΔG steps of opposite sign depending on the presynaptic pulse polarity, enables easy and consistent programming, needed for an agile and accurate SNN.

These results establish the presented green, flexible synaptic OECT as a promising building block for smart bio-inspired artificial sensory systems—including electronic skins, human–machine interfaces, and soft robotics—where sustainable materials and on-device learning converge to enable intelligent, eco-friendly electronics.