Non-volatile memory and synaptic learning properties of solution processable synthesized g-C₃N₄-WO₃ nanocomposite

Motivated by the growing demand for data storage and processing devices, researchers are synthesizing new functional materials. After a successful era of transition metal oxides (TMOs), carbon materials were utilized in electronic devices due to their superior chemical, thermal, and electrical properties. Given these properties, we fabricated composite films of graphitic carbon nitride (GCN) and tungsten trioxide (WO₃) and studied them for memory and synaptic learning applications. The hydrothermal reaction time of WO₃ was varied (4, 6, 8, and 10 h), and its effect on the GCNW nanocomposite was investigated. The Ag/GCNW/FTO-based devices exhibited bipolar resistive switching (RS) within ± 2 V, with the GCNW8 device demonstrating superior non-volatile memory performance, including an endurance of 7.5 × 10³ cycles, a retention time of 6 × 10³ s, and a memory window of ~ 36. Statistical analysis revealed tight distributions of switching voltages (V_SET and V_RESET), confirmed by Weibull distribution and low mean squared errors (MSEs < 1.2887 × 10^–2) in time series predictions. The device successfully mimicked synaptic functions, including potentiation-depression, excitatory postsynaptic current (EPSC), and paired-pulse facilitation (PPF) at the biological time scales (τ = 38.85 ms and 45.95 ms). Charge transport followed Ohmic and Child’s square law mechanisms. These results demonstrate the GCNW nanocomposite’s potential for high-performance memory and neuromorphic computing applications.