Ferroelectric Memristor Based on Hf_0.5Zr_0.5O₂ Thin Film Combining Memristive and Neuromorphic Functionalities

Abstract

As a new type of semiconductor memory device, memristors receive wide attention because of their ability to realize multilevel memory and synapse-like learning. However, memristors made of traditional materials often demonstrate dispersion and unstable switching. Ferroelectric memristors can solve this problem and realize the regulation of synaptic plasticity. Herein, memristors based on Hf 0.5 Zr 0.5 O 2 are fabricated on Si substrates. The I-V curve shows that the turn-on voltage of the device is lower for Hf 0.5 Zr 0.5 O 2 film than for other compositions in Hf x Zr 1Àx O y films, and the ratio of R OFF /R ON can reach 10 4 , implying excellent device performance. Furthermore, the train of pulse with different parameters can influence the modulation of device conductance. Also, the synaptic plasticity behavior of the device is studied. These results lay the foundation for the development of ferroelectric thin films with neuron-like artificial synapses.

High-Performance Artificial Synapse Developed by HZO on (110) NSTO

“…The impact of artificial synapses on conductance can be studied by applying pulse signals to the device as well. 29 By adjusting the pulse amplitude, frequency, and width while keeping other conditions constant, we investigate how the conductance of the device is modulated by the pulse signal, as depicted in Figure 3d−f. Figure 3d illustrates the relationship between pulse amplitude and conductance.…”

Nanoscale Probing of Electrical Memory Effects in van der Waals Layered PdSe2

“…Of particular interest are neuromorphic computing approaches that are underpinned by performing data storage and processing in the same physical space, in contrast to traditional computing architectures where data have to be constantly transferred between processors and memory, requiring additional time and energy. , It is therefore important to evaluate materials in terms of their suitability to be used in neuromorphic chips. One of the main desired characteristics is the possibility to encode memory states into the material structure, as well as the stability and accessibility of these states across a broad range of time scales specific to a desired circuit function. − In this regard, a wide range of functional mechanisms for changes in conductivity and capacitance have been studied, including electrochemical random access memory in which the resistivity can be gradually changed by varying the ion concentration, gradual formation of conductive filaments, ,− electrochemical metallization of cells, ferroelectric modulation due to polarization switching, , structural phase changes, metastable polarization states, and charge shielding . Emanating from their unique structural and electronic properties in addition to the ability to create ultrathin functional devices with low power consumption, two-dimensional materials are highly suitable candidates for neuromorphic hardware elements. ,, …”

Surface Functionalization of TiO2 Nanotubes Modified with a Thin Film of BiFeO3

“…Improving materials for resistive switching devices and field-effect transistors based on ferroelectrics is closely related to optimizing the surfaces of these materials [1][2][3][4][5]. There is growing interest in combining these elements as field-effect transistors based on TiO 2 memristors with a programmable set/reset for neuromorphic computing [6].…”