Improvement of Resistance Change Memory Characteristics in Ferroelectric and Antiferroelectric (like) Parallel Structures

Abstract

Recently, considerable attention has been paid to the development of advanced technologies such as artificial intelligence (AI) and big data, and high-density, high-speed storage devices are being extensively studied to realize the technology. Ferroelectrics are promising non-volatile memory materials because of their ability to maintain polarization, even when an external electric field is removed. Recently, it has been reported that HfO2 thin films compatible with complementary metal–oxide–semiconductor (CMOS) processes exhibit ferroelectricity even at a thickness of less than 10 nm. Among the ferroelectric-based memories, ferroelectric tunnel junctions are attracting attention as ideal devices for improving integration and miniaturization due to the advantages of a simple metal–ferroelectric–metal two-terminal structure and low ultra-low power driving through tunneling. The FTJs are driven by adjusting the tunneling electrical resistance through partial polarization switching. Theoretically and experimentally, a large memory window in a broad coercive field and/or read voltage is required to induce sophisticated partial-polarization switching. Notably, antiferroelectrics (like) have different switching properties than ferroelectrics, which are generally applied to ferroelectric tunnel junctions. The memory features of ferroelectric tunnel junctions are expected to be improved through a broad coercive field when the switching characteristics of the ferroelectric and antiferroelectric (like) are utilized concurrently. In this study, the implementation of multiresistance states was improved by driving the ferroelectric and antiferroelectric (like) devices in parallel. Additionally, by modulating the area ratio of ferroelectric and antiferroelectric (like), the memory window size was increased, and controllability was enhanced by increasing the switchable voltage region. In conclusion, we suggest that ferroelectric and antiferroelectric (like) parallel structures may overcome the limitations of the multiresistance state implementation of existing ferroelectrics.

Spike Optimization to Improve Properties of Ferroelectric Tunnel Junction Synaptic Devices for Neuromorphic Computing System Applications

“…FTJ devices utilize the phenomenon of tunneling electrical resistance, which depends on the alignment of the internal polarization of the ferroelectric layer, to realize the memory states. In particular, FTJ devices based on polycrystalline HfO 2 thin films can determine multiple resistance states because the modulation of the asymmetric tunneling barrier by ferroelectric partial polarization is possible [12,13]. The tunneling current can be extracted in a non-destructive method by applying a read voltage sufficiently small enough not to modulate the tunneling barrier [14].…”