Synergistic enhancement of long-term plasticity in solid-state electrolyte-gated synaptic transistors realized by introducing an ion-capturing layer

As artificial intelligence (AI) technology rapidly advances, the need for fast processing of large amounts of data has highlighted the bottleneck of the existing von Neumann computing system. This realization has led to the emergence of alternative computing systems such as neuromorphic computing. In response, we introduce a novel solid-state sodium alginate (NaAlg) electrolyte-gated synaptic transistor (EGST) with a polyacrylic acid (PAA)/ indium-gallium-zinc-oxide (IGZO) channel layer designed to enhance long-term plasticity. A NaAlg electrolyte enables high-speed operation owing to its high ionic conductivity. PAA, as an ion-capturing layer, plays a role in the adsorption of Na ions from NaAlg. A key mechanism for this enhancement is the adsorption of Na ions from the electrolyte onto the oxygen functional groups in the PAA layer. The PAA/IGZO EGST shows synaptic behaviors such as a paired-pulse facilitation index of 124.5 % and 96 repetitions of 64 cycles of potentiationdepression with a dynamic range of 10.3. The handwritten pattern recognition accuracy is 88.42 %, approaching the ideal device accuracy of 93.73 %, and the image retention simulation results validate the device's synaptic performance. This study contributes to the advancement of AI technology through the development of synaptic transistors and the introduction of new materials and paradigms for neuromorphic computing.