Presentation Information
[8p-E217-11]Pad Size-Dependent Capacitance and Reservoir Computing Performance in Ag/Ag2S Nanoparticle Devices
〇(M2)Kanyarak Klaichit1, Yusuke Nakaoka1, Oradee Srikimkaew2, Alif Syafiq Kamarol Zaman1, Muzhen Xu3, Yuki Usami1,3, Hirofumi Tanaka1,3 (1.LSSE, Kyushu Inst. Tech. (Kyutech), 2.FSRC, School of Science, Walailak Univ., 3.Neumorph Center, Kyutech)
Keywords:
Materials computation,CMOS integration,Low-power consumption
Brain-inspired computing has gained attention for energy-efficient in-material reservoir computing. However, the influence of electrode geometry on capacitance and temporal information processing remains unclear. Here, Ag/Ag2S nanoparticles were deposited on Pt/Ti electrode arrays with pad side lengths of 100–1000 µm. Capacitance was evaluated by electrochemical impedance spectroscopy using the Hsu–Mansfeld model. AgNP-device capacitance increased with pad length, indicating greater charge storage for larger electrodes. Reservoir performance was assessed using MC, IPC, and NARMA2 tasks. These metrics showed no monotonic dependence on pad length or capacitance, suggesting that performance arises from a balance among memory, high-dimensional mapping, and nonlinear dynamics rather than capacitance alone. More devices should be tested to confirm reproducibility and identify the pad length that best balances computing performance and power efficiency.