Improvement in energy consumption and operational stability of electrolyte-gated synapse transistors using atomic-layer-deposited HfO2 thin films

Dong-Hee Kim,Sung-Min Yoon

Materials Science in Semiconductor Processing(2023)

引用 2|浏览0
暂无评分
摘要
Electrolyte-gated transistors (EGTs) are promising candidates for artificial synaptic devices, due to lower energy consumption by forming the electric double layer to modulate the channel conductance. In this work, we propose a three-terminal artificial synapse transistor employing HfO2 electrolyte-gate insulator (EGI). The inorganic HfO2 thin film with high dielectric constant was chosen as a good candidate of an EGI to improve the operational stability and energy consumption for the synapse transistor. Mobile hydrogen ions incorporated in the HfO2 EGI were revealed to be generated during the ALD process at a temperature as low as 100 °C. The synaptic plasticity of the fabricated synapse transistor could be gradually modulated by the movement of mobile ions in the EGI, which were examined to be dependent on the applied pulse conditions. The fabricated synapse transistor emulated biological synaptic functions, including excitatory/inhibitory post-synaptic current (EPSC/IPSC) and paired-pulse facilitation/depression (PPF/PPD). Furthermore, by controlling the pulse spike conditions, the conversion from short-term plasticity (STP) to long-term plasticity (LTP) was successfully verified. The LTP characteristics were demonstrated to be far enhanced with the EGI film thickness by means of magnifying the amounts of incorporated mobile ions. In addition, the long-term reliability and operational repeatability of the fabricated synapse transistor were well verified thanks to the choice of HfO2 EGI. The energy consumption required for single spike event was estimated to be a value as low as 0.26 pJ, thus demonstrating considerable potential as an artificial synapse device.
更多
查看译文
关键词
Neuromorphic,Synapse transistor,Solid electrolyte,Atomic-layer deposition,HfO2
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要