Due to a rapid increase in the amount of data, there is a huge demand for the development of new memory technologies as well as emerging computing systems for high-density memory storage and efficient computing. As the conventional transistor-based storage devices and computing systems are approaching their scaling and technical limits, extensive research on emerging technologies is becoming more and more important. Among other emerging technologies, CBRAM offers excellent opportunities for future memory and neuromorphic computing applications. The principles of the CBRAM are explored in depth in this review, including the materials and issues associated with various materials, as well as the basic switching mechanisms. Furthermore, the opportunities that CBRAMs provide for memory and brain-inspired neuromorphic computing applications, as well as the challenges that CBRAMs confront in those applications, are thoroughly discussed. The emulation of biological synapses and neurons using CBRAM devices fabricated with various switching materials and device engineering and material innovation approaches are examined in depth.

The interface region between Ga-face n-type GaN and Al₂O₃ dielectric (achieved via atomic-layer deposition or ALD) is investigated by X-ray photoelectron spectroscopy (XPS). An increase in the Ga-O to Ga-N bond intensity ratio following Al₂O₃ deposition implies that the growth of an interfacial gallium sub-oxide (GaO_x) layer occurred during the ALD process. This finding may be ascribed to GaN oxidation, which may still happen following the reduction of a thin native GaO_x by trimethylaluminum (TMA) in the initial TMA-only cycles. The valence band offset between GaN and Al₂O₃, obtained using both core-level and valence band spectra, is found to vary with the thickness of the deposited Al₂O₃. This observation may be explained by an upward energy band bending at the GaN surface (due to the spontaneous polarization induced negative bound charge on the Ga-face GaN) and the intrinsic limitation of the XPS method for band offset determination.

A major intrinsic limitation of the reaction-diffusion (R-D) model for negative-bias temperature instability (NBTI) is revealed through dynamic stress experiments. We found no evidence of self-limiting recovery, one of the key features of the transport-based R-D model, after repeating the stress and relaxation cycles alternately for many times. The amount of recovery per cycle of the parameter of interest (e.g., threshold voltage shift, change in the charge-pumping (CP) current, etc.) is shown to remain constant, independent of the number of stress/recovery cycles. Under repeated cycling of the test device between stress and recovery, it is also found that the amount of parametric shift induced by the stress cycle becomes nearly identical to that recovered during the relaxation cycle, i.e., the parametric evolution under a fixed set of stress and recovery intervals is cyclic in nature. In conjunction with the thermal activation result, this cyclic behavior of the dynamic NBTI is ascribed to an ensemble of switching hole traps having broad spectra of characteristic trapping and detrapping time constants. The same group of traps responds under a fixed set of experimental conditions, giving rise to the cyclic behavior. The interface state generation was also investigated using a CP current measurement and is found to be permanent within the range of timing examined. It is also shown that the variation in the power-law exponent of the as-measured change in the CP current with temperature could be consistently explained by considering the different thermal activation of the hole trapping and interface state components. In view of these new evidences, previous claims of consistency between the generation/recovery of the interface states and the R-D model or its dispersive counterpart must be reviewed.

Growth of data eases the way to access the world but requires increasing amounts of energy to store and process. Neuromorphic electronics has emerged in the last decade, inspired by biological neurons and synapses, with in-memory computing ability, extenuating the ‘von Neumann bottleneck’ between the memory and processor and offering a promising solution to reduce the efforts both in data storage and processing, thanks to their multi-bit non-volatility, biology-emulated characteristics, and silicon compatibility. This work reviews the recent advances in emerging memristive devices for artificial neuron and synapse applications, including memory and data-processing ability: the physics and characteristics are discussed first, i.e., valence changing, electrochemical metallization, phase changing, interfaced-controlling, charge-trapping, ferroelectric tunnelling, and spin-transfer torquing. Next, we propose a universal benchmark for the artificial synapse and neuron devices on spiking energy consumption, standby power consumption, and spike timing. Based on the benchmark, we address the challenges, suggest the guidelines for intra-device and inter-device design, and provide an outlook for the neuromorphic applications of resistive switching-based artificial neuron and synapse devices.

The state‐of‐the‐art conception of a bionic/robotic eye is a somewhat bulky multipart system comprising a video camera connected to a processing unit that in turn communicates data through a wireless transmitter to either an in vivo retinal implant or a computer system. An artificial cogni‐retina is a millimeter‐scale, intelligent apparatus designed as a replacement for these systems, while executing simple image processing tasks. As a bionic limb, it can connect directly to the optic nerve and perform rudimentary cognitive functions such as perceiving, learning, remembering, and classifying elementary visual data. This theoretical system presents a quantum leap in terms of size, power consumption, and speed to both prosthetic human eyes and robotic vision in artificial intelligence‐based platforms such as autonomous vehicles. Recently, an increasing number of publications have used interesting materials in artificial synaptic devices that drive this idea closer toward becoming a real‐world application. Such devices may form a basis for hardware‐based deep learning artificial neural networks that can potentially execute image processing tasks within a single clock cycle compared to software algorithms running on conventional von Neumann machines that require millions of cycles to perform image sensor interfacing, memory fetch operations, and data path propagation.