Resistive Switching Characteristic Improvement in a Single-Walled Carbon Nanotube Random Network Embedded Hydrogen...

In this study, we evaluated the improved memristive switching characteristics of hydrogen silsesquioxane (HSQ) nanocomposites embedded with a single-walled carbon nanotube (SWCNT) random network. A low-temperature solution process was implemented using a flexible memristor device on a polyethylene naphthalate (PEN) substrate. The difference in the resistive switching (RS) behavior due to the presence of the SWCNT random network was analyzed by the current transport mechanism. Such a random network not only improves the RS operation but also facilitates a stable multilevel RS performance. The multiple-resistance states exhibited highly reliable nonvolatile retention properties over 10⁴ s at room temperature (25 °C) and at a high temperature (85 °C), showing the possibility of an analog synaptic weight modulation. Consequently, the gradual weight potentiation/depression was realized through 3 × 10² synaptic stimulation pulses. These findings suggest that the embedded SWCNT random network can improve the synaptic weight modulation characteristics with high stability for an artificial synapse and hence can be used in future neuromorphic circuits.