Publication | Closed Access
Tri-Gate Ferroelectric FET Characterization and Modelling for Online Training of Neural Networks at Room Temperature and 233K
27
Citations
0
References
2020
Year
Unknown Venue
EngineeringPhase Change MemorySymmetric MultilevelOnline TrainingFerroelectric ApplicationNanoelectronicsNeuromorphic EngineeringNeuromorphic DevicesNeurocomputersElectrical EngineeringGate LengthBias Temperature InstabilityNeural NetworksMicroelectronicsSpintronicsRoom TemperatureComputational NeuroscienceCondensed Matter PhysicsApplied PhysicsNeuronal NetworkSemiconductor MemoryBrain-like ComputingDevice Variability
This paper reports detailed analysis on switching dynamics and device variability over a wide range of temperatures for deeply scaled (40nm gate length) tri-gate ferroelectric FETs with 10nm HZO fabricated using gate first process on SOI wafers. Our experimental results manifest, 99% ferroelectric switching at room temperature and at 233K. A memory window over 5V and strong gate length dependence of memory window is observed. Highly linear and symmetric multilevel switching characteristics makes our ferroelectric FETs suitable for neuromorphic applications, as demonstrated with neural network online training simulations.