Photocatalytic Reduction of Graphene Oxide–TiO₂ Nanocomposites for Improving Resistive‐Switching Memory Behaviors

Abstract

Graphene oxide (GO)-based resistive-switching (RS) memories offer the promise of low-temperature solution-processability and high mechanical flexibility, making them ideally suited for future flexible electronic devices. The RS of GO can be recognized as electric-field-induced connection/disconnection of nanoscale reduced graphene oxide (RGO) conducting filaments (CFs). Instead of operating an electrical FORMING process, which generally results in high randomness of RGO CFs due to current overshoot, a TiO -assisted photocatalytic reduction method is used to generate RGO-domains locally through controlling the UV irradiation time and TiO concentration. The elimination of the FORMING process successfully suppresses the RGO overgrowth and improved RS memory characteristics are achieved in graphene oxide-TiO (Go-TiO ) nanocomposites, including reduced SET voltage, improved switching variability, and increased switching speed. Furthermore, the room-temperature process of this method is compatible with flexible plastic substrates and the memory cells exhibit excellent flexibility. Experimental results evidence that the combined advantages of reducing the oxygen-migration barrier and enhancing the local-electric-field with RGO-manipulation are responsible for the improved RS behaviors. These results offer valuable insight into the role of RGO-domains in GO memory devices, and also, this mild photoreduction method can be extended to the development of carbon-based flexible electronics.

Toward a generalized Bienenstock-Cooper-Munro rule for spatiotemporal learning via triplet-STDP in memristive devices

“…B rain-inspired neuromorphic computing systems are attracting strong interest because of their massive parallelism, high energy efficiency, good error tolerance, and good ability to implement cognitive functions [1][2][3][4][5][6] . Hardware implementations of neuromorphic computing can take advantage of novel nanodevices that emulate the biological synapses with inherent learning functions [7][8][9][10][11][12][13] .…”

Controlled Nonvolatile Transition in Polyoxometalates‐Graphene Oxide Hybrid Memristive Devices

“…As is well known that the work function (WF) of graphene are tunable by adjusting the amount of oxygen‐containing group, and the value of WF can be varied even larger than 2.5 eV depending on the oxygen functional groups attached to the graphene domain . In order to investigate the interface characteristic, ultraviolet photoelectron spectroscopy (UPS) measurement of GO and RGO species was carried out to obtain the exact WF values, with the result displayed in Figure S7 in the Supporting Information.…”

Overview of Resistive Random Access Memory (RRAM): Materials, Filament Mechanisms, Performance Optimization, and Prospects

“…Graphene oxide devices that can achieve the bidirectional progressive conductance were also used to reveal the impacts of different pulses on conductance control and the potential relationship between pulse modulation and energy applied; their performance parameters (switching time, pulse scheme, and voltage) are compared with others as shown in Table . On the basis of the programming current statistics of RRAM functional materials in recent years, it is found that the programming current of the 2D‐perovskite device can reach a minimum of 10 pico‐Amperes, which is at least one order of magnitude lower than general RRAM devices, thus achieving the goal of low power consumption . Due to the extremely low programming current, the resistive switching memory demonstrated 400 fJ/spike synaptic operation, which is very close to the energy consumption of biological synapses.…”