Artificial Synaptic Simulation Using Wurtzite AlScN-Based Ferroelectric Memristors

Ferroelectric memristors have significant potential as the key component in the neuromorphic computing technology that is critical in the post-Moore era. AlScN has emerged as a promising ferroelectric material with a wurtzite structure due to its large residual polarization, CMOS compatibility, great ability to be miniaturized, and high Curie temperature. Here, the Cu/Al0.73Sc0.27N/Pt with the coexistence of resistive switching (RS) and threshold switching behaviors was fabricated, showing a high Roff/Ron ratio (105%), great switching speed (<150 ns), and a long retention characteristic (>1500 s). By appropriately adjusting the compliance current (CC), the device exhibits gradual switching and multiple conductance states. In addition, biological synaptic functions including paired-pulse facilitation (PPF) and the short-term plasticity (STP) to long-term plasticity (LTP) transition were emulated successfully after verifying the synaptic plasticity of the device under the pulse stimuli. These findings may be applicable in designing next-generation ferroelectric memristors for efficient neuromorphic computing systems.