Unveiling the dual role of chemically synthesized copper doped zinc oxide for resistive switching applications

Abstract

In this study, efforts were devoted to unveiling the dual role of single crystalline Cu (5%) doped ZnO (Cu:ZnO) synthesized by a simple and low-cost chemical process and to investigate its efficacy on resistive switching (RS) applications. It was found that when Cu:ZnO was annealed at a lower temperature of 450 °C and integrated onto ITO/glass for RS applications, only oxygen mediated vacancies were responsible for its resistive switching. However, ferroelectric properties have been observed when the same Cu:ZnO was annealed at a higher temperature of 800 °C and integrated onto Nb doped SrTiO 3 . X-ray diffraction, high resolution transmission electron microscope, x-ray photoelectron spectroscopy, UV-VIS-near infrared spectrometer, and piezoelectric force microscopy (PFM) were employed to study the crystallinity, interfaces, chemical compositions, bandgap, and domains in Cu:ZnO thin films, respectively. The bandgap of Cu:ZnO was found to be 3.20 eV. PFM study exhibits the domain inversion with 180°polarization inversion by applying an external bias, evidencing its effectiveness for memory applications. When the electrical characteristics were concerned, the RS device based on this ferroelectric Cu:ZnO offers better performance, such as lower SET/RESET voltages (∼1.40 V), higher retention (up to 10 6 s) without distortion, and higher ON/OFF ratio (2.20 × 10 3 ), as compared to the former lower temperature annealed Cu:ZnO devices. A band-diagram was proposed, and transport studies were developed to understand the operational mechanism of these devices. This study explains both the limits and scopes of Cu:ZnO RS devices and formulates an idea which may accelerate the design of future generation devices.

Implementation of fast ICA using memristor crossbar arrays for blind image source separations

“…Here, w i j is the weight of one memristor. The weights of the algorithm are stored in the memristor itself due to having its nonvolatile property [28].…”

“…Here, is the weight of one memristor. The weights of the algorithm are stored in the memristor itself due to having its non‐volatile property [28]. In order to realise matrix multiplication of the input data matrix with the weight matrix of the used neural network algorithm, a Cu:ZnO (memristor using VTEAM model, which mimic exactly the same output performances to the fabricated devices) based crossbar array has been developed (shown in Fig.…”

“…Recently, memristors have been found to be useful in emulating important features of biological synapses owing to their nanoscale dimensions, their ability to store multiple bits of information per element and the low energy required to write distinct states [30]. A Pt/Cu:ZnO/Nb:STO memristor was fabricated and the detailed fabrication procedures are provided in our earlier publication [28].…”

“…Recently, memristors have been found to be useful in emulating important features of biological synapses owing to their nanoscale dimensions, their ability to store multiple bits of information per element and the low energy required to write distinct states [30]. A Pt/Cu:ZnO/Nb:STO memristor was fabricated and the detailed fabrication procedures are provided in our earlier publication [28]. Since the resistance of a memristor varies with the amount of current passing through the memristor, it can store the weights of the neural network and update them using simple voltage pulses [31].…”

“…For the memristor, a new alternative non‐filamentary Cu (5%) doped ZnO (Cu:ZnO) based Pt/Cu:ZnO/Nb:STO memristor was utilised. The detailed working mechanism of Pt/Cu:ZnO/Nb:STO memristors and their suitable electrical performances such low Set/Reset voltages, high retention, fast switching and non‐filamentary conduction can be found in our earlier work [28]. The main contribution of this paper is to introduce an innovative memristor‐based advanced neural network proposed by Hyvarinen [19], which is also known as Fast ICA, and this non‐conventional attempt has been initiated for the first time to the best of our knowledge, though it possesses several advantages over conventional ICA as already discussed in the previous paragraph.…”