Synaptic characteristics of ferroelectric capacitors for neuromorphic systems

The neuromorphic system is a hardware-level biosimulation system that implements neuron and synaptic elements. It has the same advantages as the human brain, such as low power consumption and small size. We proposed the synapse consisting of a memcapacitor for low power consumption. In this study, we evaluated the characteristics of ferroelectric capacitor type synapses by simulation. Moreover, we created a capacitor-type synapses with ferroelectric thin films. Electrical characteristics were measured as a device evaluation, and they showed usefulness as a synapse. It is expected that if character correction learning is performed using the capacitor-type synapses, the modified pattern will match the memorized pattern, indicating successful training.