Polymer-Based Linear and Symmetric Artificial Synaptic Memristors for Accurate and Reliable Neuromorphic Computing Applications.

The rapid expansion of artificial intelligence has intensified the demand for hardware systems capable of emulating brain-like information processing with high accuracy, energy efficiency, and reliability. Neuromorphic computing based on memristive artificial synapses has emerged as a promising approach to overcome the limitations of conventional von Neumann architectures. Although inorganic and oxide-based synaptic memristors have been widely explored for neuromorphic systems, they often suffer from poor linearity, asymmetric potentiation/depression behavior, limited conductance states, and device variability, which restrict learning accuracy and scalability. In contrast, polymer-based memristors have gained significant attention owing to their intrinsic advantages, including mechanical flexibility, molecular tunability, controllable electronic/ionic transport, low-temperature processability, and compatibility with large-area fabrication. This review critically examines recent advances in polymer-based memristive materials and devices for achieving linear and symmetric artificial synaptic behavior. Polymer synapses are classified into pure polymer, polymer composite, and polymer-hybrid systems through a mechanism to function framework. Rather than providing a general compilation of organic memristor studies, this review analyzes how polymer chemistry, ion-migration control, trap state distribution, redox activity, electrode selection, active layer thickness, and interface engineering govern conductance update linearity, symmetry, and uniformity. Fundamental switching mechanisms, material classifications, device architectures, key synaptic characteristics, and system-level neuromorphic performance, including pattern-recognition applications, are critically discussed. By explicitly linking material and device design to conductance update fidelity, learning accuracy, training convergence, and pattern-recognition reliability, this review provides practical design guidelines and future perspectives for next-generation polymer-based neuromorphic hardware with improved linearity, symmetry, reliability, and scalability.