CMOS Compatible Low Power Consumption Ferroelectric Synapse for Neuromorphic Computing
Zhenhai Li, Jialin Meng, Jiajie Yu, Yongkai Liu, Tianyu Wang, Pei Liu, Shiyou Chen, Hao Zhu, Qingqing Sun, David Wei Zhang, Lin Chen
Abstract
With the development of bioelectronics, brain-inspired artificial synapses become more and more important. To simulate artificial synapse, a HfAlO ferroelectric tunnel junction (FTJ) was fabricated, which can simulate short-term synaptic plasticity for neuromorphic computing. The devices realize the synaptic function with low power consumption of about 7.15 aJ per synaptic event. Moreover, to explore the effect of oxygen defects on ferroelectric properties of HfAlO-based device, the first-principle analysis was further carried out. These results pave the way of hafnium-based ferroelectric synaptic devices.
Topics & Concepts
Neuromorphic engineeringBioelectronicsSynapseFerroelectricityMaterials sciencePower consumptionCMOSComputer scienceSynaptic plasticityMemristorOptoelectronicsNeuroscienceElectrical engineeringPower (physics)NanotechnologyArtificial neural networkEngineeringArtificial intelligencePhysicsChemistryPsychologyBiosensorBiochemistryReceptorDielectricQuantum mechanicsAdvanced Memory and Neural ComputingFerroelectric and Negative Capacitance DevicesSemiconductor materials and devices