Nonconjugated Radical Polymers-Based High-Yield Organic Memory and Ethanol In-Sensor Computing

Abstract

Organic memristors are promising candidates for next-generation soft, biorealistic electronics due to their flexibility, biocompatibility, processability, and low power consumption. These features facilitate their seamless incorporation into mechanically compliant platforms, thereby expanding their potential for use in emerging applications such as flexible in-sensor computing architectures. However, conventional organic memristors based on conjugated polymers often suffer from low device yield and reliability due to their semicrystalline nature, which leads to film roughness and pinhole formation. In contrast, nonconjugated radical polymers offer amorphousness and intrinsic memristivity from stable redox activity, making them ideal for uniform memristive switching layers. Furthermore, the molecular tunability of nonconjugated radical polymers enables precise control over polymer properties, which significantly improves the reliability and fabrication yield of organic memristive devices. In this study, we demonstrate a high-yield organic memristor and a soft in-sensor computing system based on a radical polymer tailored through molecular design. This organic memristor arrays shows over 95% fabrication yield, excellent switching performance (on/off >106, retention >4 × 105 s, endurance >500 cycles), and mechanical durability over 1,000 bending cycles. Additionally, the device also shows chemical sensitivity toward ethanol, enabling multifunctional operation in soft in-sensor computing systems. This work demonstrates a polymer engineering strategy that enhances both the physical robustness and multifunctionality of organic memristive materials, advancing their potential in flexible and biointegrated electronics.