Memristive Devices Research Articles

Adaptive spiking, itinerancy, and quantum effects in artificial neuron circuit hardware with niobium–hafnium oxide-niobium memristor devices inserted

To improve artificial intelligence/autonomous systems and help with treating neurological conditions, there is a requirement for the discovery and design of artificial neuron hardware that mimics the advanced functionality and operation of the neural networks available in biological organisms. We examine experimental artificial neuron circuits that we designed and built in hardware with memristor devices using 4.2 nm of hafnium oxide and niobium metal inserted in the positive and negative feedback of an oscillator. At room temperature, these artificial neurons have adaptive a spiking behavior and hybrid non-chaotic/chaotic modes. When networked, they output with strong itinerancy, and we demonstrate a four-neuron learning network and modulation of signals. The superconducting state at 8.1 K results in Josephson tunneling with signs that the hafnium oxide ionic states are influenced by quantum control effects in accordance with quantum master equation calculations of the expectation values and correlation functions with a calibrated time-dependent Hamiltonian. These results are of importance to continue advancing neuromorphic hardware technologies that integrate memristors and other memory devices for many biological-inspired applications and beyond that can function with adaptive-itinerant spiking and quantum effects in their principles of operation.

Effects of Group IVA Elements on the Electrical Response of a Ge2Se3-Based Optically Gated Transistor.

The optically gated transistor (OGT) has been previously demonstrated as a viable selector device for memristor devices, and may enable optical addressing within cross-point arrays. The OGT current-voltage response is similar to a MOSFET device, with light activating the gate instead of voltage. The OGT also provides a naturally built-in compliance current for a series resistive memory element, determined by the incident light intensity on the gate, thus keeping the integrated periphery circuitry size and complexity to a minimum for a memory array. The OGT gate comprises an amorphous Ge2Se3 material that can readily be doped with other elements to alter the transistor's electrical properties. In this work, we explore the operation of the OGT when the Ge2Se3 gate material is doped with the Group IVA elements C, Si, Sn, and Pb. The dopant atoms provide changes to the optical and electrical properties that allow key electrical properties such as the dark current, photocurrent, switching speed, and threshold voltage to be tuned.

Manganite memristive devices: recent progress and emerging opportunities

Manganite-based memristive devices have emerged as promising candidates for next-generation non-volatile memory and neuromorphic computing applications, owing to their unique resistive switching behavior and tunable electronic properties. This review explores recent innovations in manganite-based memristive devices, with a focus on materials engineering, device architectures, and fabrication techniques. We delve into the underlying mechanisms governing resistive switching in manganite thin films, elucidating the intricate interplay of oxygen vacancies, charge carriers, and structural modifications. This review underscores breakthroughs in harnessing manganite memristors for a range of applications, from high-density memory storage to neuromorphic computing platforms that mimic synaptic and neuronal functionalities. Additionally, we discuss the role of characterization techniques and the need for a unified benchmark for these devices. We provide insights into the challenges and opportunities associated with the co-integration of manganite-based memristive devices with more mature technologies, offering a roadmap for future research directions.

Research on Residual Gas Adsorption on Surface of Hexagonal Boron Nitride-Based Memristor.

As a promising nonvolatile memory device with two ends, the memristor has received extensive attention for its industrial manufacture. Density functional theory was used to analyze the adsorption properties of residual gas on hexagonal boron nitride (h-BN)-based memristor model surfaces with Stone-Wales-5577 grain boundary defects [h-BN(SW)]. First, by calculating the adsorption energy, geometric parameters, and charge transfer, we identified the most stable adsorption sites for hydrogen atoms (H-TB1) and H2 molecules (H2-TN2). We observed a tendency toward chemisorption for hydrogen atoms and physical adsorption for H2 molecules at these sites. Furthermore, two coadsorption configurations were formed by introducing H2 molecules and hydrogen atoms into single adsorption configurations: namely H-TB1_H2-TN1TN2 and H2-TN2_H-TB1TN1TN3. In the case of hydrogen-based configuration, there is weak dissociation of the H2 molecule, which does not facilitate hydrogen atom adsorption. However, adjacent hydrogen atoms tend to form stable dimers, while excess hydrogen atoms have a tendency to weakly chemisorb in the case of H2-based configuration. The pristine h-BN surface is more favorable for hydrogen atom migration compared to the h-BN(SW) surface due to its higher adsorption energy. On the h-BN(SW) surface, hydrogen atoms tend to migrate inward from the center of adjacent heptagonal boron nitride rings while coadsorption has a minimal impact on their vertical migration as well as that of H2 molecules. This work provides theoretical insights into the H/H2 trace gas interaction during h-BN wafer-level fabrication for memristor devices.

Influence of stop and gate voltage on resistive switching of 1T1R HfO2-based memristors, a modeling and variability analysis

Memristive devices, particularly resistive random access memory (RRAM) cells based on hafnium oxide (HfO₂) dielectrics, exhibit promising characteristics for a wide range of applications. In spite of their potential, issues related to intrinsic variability and the need for precise simulation tools and modeling methods remain a medium-term hurdle. This study addresses these challenges by investigating the resistive switching (RS) behavior of different 1T1R HfO₂-based memristive devices under various experimental conditions. Through a comprehensive experimental analysis, we extract RS parameters using different numerical techniques to understand the cycle-to-cycle (C2C) and device-to-device (D2D) variability. Additionally, we employ advanced simulation methodologies, including circuit breaker-based 3D simulation, to shed light on our experimental findings and provide a theoretical framework to disentangle the switching phenomena. Our results offer valuable insights into the RS mechanisms and variability, contributing to the improvement of robust parameter extraction methods, which are essential for the industrial application of memristive devices.

Impact of Oxidation Layer in the Resistive Switching Behavior of Nitride-Based Memristor Devices

Impact of Oxidation Layer in the Resistive Switching Behavior of Nitride-Based Memristor Devices

Importance of growth method and substrate-induced crystalline quality in Al/Gd0.2Ca0.8MnO3/Au memristor devices

We report on the impact of the growth method and substrate-induced crystalline quality on the performance of planar Al/Gd0.2Ca0.8MnO3/Au memristor devices. Structural, magnetic, and resistive properties were thoroughly examined for fundamental characterizations, with a particular emphasis on their correlation with the memristive properties of fabricated devices. Our findings suggest that memristor structures grown on single crystal SrTiO3 substrates using pulsed laser deposition (PLD) consistently exhibit superior crystalline quality compared to those fabricated using chemical solution deposition and on silicon-based substrates. Despite variations in growth method and substrate, all memristor structures display typical resistive switching (RS) behaviour, distinguishing between high-resistance and low-resistance states. However, endurance and retention measurements demonstrate that memristor structures produced via PLD on single crystalline SrTiO3 demonstrate the most favourable RS properties. To elucidate the mechanisms underlying the differences in RS behavior across substrates and deposition methods, we extensively discuss these issues in the context of structural distortion and conduction mechanisms.

Building an Analog Circuit Synapse for Deep Learning Neuromorphic Processing

In this article, we propose a circuit to imitate the behavior of a Reward-Modulated spike-timing-dependent plasticity synapse. When two neurons in adjacent layers produce spikes, each spike modifies the thickness in the shared synapse. As a result, the synapse’s ability to conduct impulses is controlled, leading to an unsupervised learning rule. By introducing a reward signal, reinforcement learning is enabled by redirecting the growth and shrinkage of synapses based on signal feedback from the environment. The proposed synapse manages the convolution of the emitted spike signals to promote either the strengthening or weakening of the synapse, represented as the resistance value of a memristor device. As memristors have a conductance range that may differ from the available current input range of typical CMOS neuron designs, the synapse circuit can be adjusted to regulate the spike’s amplitude current to comply with the neuron. The circuit described in this work allows for the implementation of fully interconnected layers of neuron analog circuits. This is achieved by having each synapse reconform the spike signal, thus removing the burden of providing enough power from the neurons to each memristor. The synapse circuit was tested using a CMOS analog neuron described in the literature. Additionally, the article provides insight into how to properly describe the hysteresis behavior of the memristor in Verilog-A code. The testing and learning capabilities of the synapse circuit are demonstrated in simulation using the Skywater-130 nm process. The article’s main goal is to provide the basic building blocks for deep neural networks relying on spiking neurons and memristors as the basic processing elements to handle spike generation, propagation, and synaptic plasticity.

Effect of dimensional confinement on the memristive properties of the perovskite-inspired novel CuxAgBiI4+x halide compound

Low dimensional halide perovskite materials have shown significant performance enhancement in memristor devices. In search of highly stable and robust perovskites, several perovskite-inspired structures are proposed. This present study demonstrates the fabrication of novel lead-free all-inorganic CuxAgBiI4+x (x = 1 or 2) as the active material sandwiched between gold and indium doped tin oxide electrodes. Initial studies revealed that the memristor devices consisting of two-dimensional Cu2AgBiI6 have higher current on/off ratio as compared to the three-dimensional CuAgBiI5. This is ascribed to the elevated Schottky barrier at the Au/Cu2AgBiI6 interface as suggested by ultraviolet photoelectron spectroscopy, which resulted low current in the high resistance state (HRS) of the device. Kelvin probe force microscopy shows that the charge accumulation property is similar in both the materials although the CuAgBiI5 possesses less vacancies. However, the high switching window of the Cu2AgBiI6-based device succeeded in retaining the stable on/off ratio in the cycling endurance. Artificial synaptic characterizations have been carried out in the Cu2AgBiI6 based device due to low HRS current. The device could mimic paired-pulsed facilitation, transition from short-term to long-term potentiation and thus establishes the fact that low dimensional perovskite-type materials possess great potential for memristor application.

Unraveling the Roles of Switching and Relaxation Times in Volatile Electrochemical Memristors to Mimic Neuromorphic Dynamical Features

AbstractComputing through ensembles of interacting dynamical elements is the next frontier of the diverse field of neuromorphic computing. Spiking neural networks are one of the possible examples. Computation through dynamics and through time requires the development of novel technologies for devices with rich dynamics. Among the various candidates, the most promising ones are volatile electrochemical memristive systems that switch from high to low resistance state by voltage application and self‐recover the high resistance state after a tunable relaxation time. Such devices can perform a wide variety of computational primitives. However, a clear comprehensive picture of their possible dynamics and their physical interpretation is still missing. In the present manuscript, prototypical electrochemical silver/silicon oxide/platinum (Ag/SiOx/Pt) memristive devices are characterized to identify dynamical aspects, like integrative effects and stochastic switching. Integrative effects are evidenced both in high and low resistance states, associated to wake‐up phase and cumulative switching. All the dynamical aspects are related to characteristic switching times and relaxation times, and with reference to the electrochemical and physical processes involved in device operation. The various dynamical aspects are linked to short‐term memory effects and basic temporal processing functions based on paired‐pulse effects that are relevant for neuromorphic applications.

Reservoir computing with a random memristor crossbar array.

Physical implementations of reservoir computing (RC) based on the emerging memristors have become promising candidates of unconventional computing paradigms. Traditionally, sequential approaches by time-multiplexing volatile memristors have been prevalent because of their low hardware overhead. However, they suffer from the problem of speed degradation and fall short of capturing the spatial relationship between the time-domain inputs. Here, we explore a new avenue for RC using memristor crossbar arrays (MCAs) with device-to-device variations, which serve as physical random weight matrices of the reservoir layers, enabling faster computation thanks to the parallelism of matrix-vector multiplication as an intensive operation in RC. To achieve this new RC architecture, ultralow-current, self-selective memristors are fabricated and integrated without the need of transistors, showing greater potential of high scalability and three-dimensional integrability compared to the previous realizations. The information processing ability of our RC system is demonstrated in tasks of recognizing digit images and waveforms. This work indicates that the "nonidealities" of the emerging memristor devices and circuits are a useful source of inspiration for new computing paradigms.

Volatile and Nonvolatile Dual‐Mode Switching Operations in an Ag‐Ag2S Core‐Shell Nanoparticle Atomic Switch Network

AbstractThis paper proposes a nanoparticle‐based atomic switch network memristive device, capable of both volatile and nonvolatile switching operations, which have not been previously reported for this material. The operational modes can be determined by altering the compliance current, demonstrating high stability over 100 cycles. Analysis of the conduction mechanism using I–V curves reveals switching characteristics consistent with space‐charge‐limited current conduction during the set process and ohmic behavior in the reset state. Furthermore, this study analyzes these dual‐operational modes in devices with varying electrode spacings. The results indicate that a wider spacing necessitated a higher compliance current for the volatile‐to‐nonvolatile transition, underscoring the significance of interconnection. These findings facilitate the integration of neuron and synapse functions within a single atomic switch network device, thereby advancing neuromorphic systems.

Implementation of 8-bit reservoir computing through volatile ZrOx-based memristor as a physical reservoir

In this study, we employed a sputtering process to construct a memristive device within the ITO/ZrOx/TaN structure for implementing neuromorphic computing. Initially, we scanned the basic electrical properties of the ITO/ZrOx/TaN device using a DC voltage sweep on the top ITO electrode. A highly uniform gradual resistive switching phenomenon was observed over 100 cycles. The current decay in the low-resistance state was effectively controlled by the volatile memory properties. Gradual conductance changes for potentiation and depression were achieved by applying electrical pulses, enabling the establishment of multi-level conductance states. In addition, the emulation of various synaptic functions was achieved by following the learning rules of SRDP, EPSC, STDP, ADSP, Pavlovian associative learning, and PPF. Finally, 8-bit reservoir computing was demonstrated in cost-effective pattern generation and recognition, highlighting the ITO/ZrOx/TaN device’s advantageous memory storage properties for synaptic characteristics.

Gas phase alloyed crystalline S–Se dielectrics with high ionic mobility

The advancement of future electronic devices necessitates dielectric materials with enhanced compositional complexity and improved capabilities. We here demonstrate a gas-phase alloying approach that yields ultrathin and crystalline dielectrics with attractive properties for the integration into electronics. A surface-selective deposition process was shown to produce sulfur (S) and selenium (Se) alloys with large-scale uniformity. Through combination of experimental diffraction analysis and materials modeling, we establish the crystallinity of the alloy with a modified lattice structure compared to the host materials. The resulting lattice arrangement endows the alloy dielectric with high ionic mobility as validated by electrochemical impedance spectroscopy. Leveraging this innovative feature, we fabricate memristive devices exhibiting promising performance characteristics. Our findings demonstrate the feasibility of utilizing gas-phase alloying to engineer dielectrics with superior properties and functionality for future device integration.

Unraveling Conductive Filament Formation in High Performance Halide Perovskite Memristor

AbstractHalide perovskites (HPs) are promising materials for memristor devices because of their unique characteristics. In this study, nonvolatile resistive switching memory devices based on thick MAPbI3 perovskite (800 nm) films with structure FTO/MAPbI3/polymethyl methacrylate (PMMA)/Ag are presented. Reproducible and reliable bipolar switching characteristics are demonstrated with an ultra‐low operating voltage (−0.1 V), high ON/OFF ratio (106), endurance (>2 × 103 times) and a record retention time (>105 s). The I–V curve of the first cycle exhibits self‐formed conductive filaments. These are attributed to the presence of metallic Pb resulting from an excess of PbI2 in the perovskite film. The subsequent activation process involves the formation of conductive filaments, consisting of either iodide vacancies or migrated charged metals. Numerical simulations are then carried out to understand the nature of these conductive filaments and the role of the internal electric field in the migration of iodide ions, iodide vacancies, and Ag cations. Finally, an exhaustive model is proposed that explains the set and reset processes governing the first voltage cycle and the steady state, at different voltage ranges. In summary, this work offers a novel and thorough perspective of the complete resistive switching (RS) behavior in a MAPbI3/buffer/Ag memristor, supported by numerical simulations.

Effect of neural firing pattern on NbOx/Al2O3 memristor-based reservoir computing system

The implementation of reservoir computing using resistive random-access memory as a physical reservoir has attracted attention due to its low training cost and high energy efficiency during parallel data processing. In this work, a NbOx/Al2O3-based memristor device was fabricated through a sputter and atomic layer deposition process to realize reservoir computing. The proposed device exhibits favorable resistive switching properties (>103 cycle endurance) and demonstrates short-term memory characteristics with current decay. Utilizing the controllability of the resistance state and its variability during cycle repetition, electrical pulses are applied to investigate the synapse-emulating properties of the device. The results showcase the functions of potentiation and depression, the coexistence of short-term and long-term plasticity, excitatory post-synaptic current, and spike-rate dependent plasticity. Building upon the functionalities of an artificial synapse, pulse spikes are categorized into three distinct neural firing patterns (normal, adapt, and boost) to implement 4-bit reservoir computing, enabling a significant distinction between “0” and “1.”

Facile synthesis of iodine variants-incorporated lithium perrhenate crystal and film: Resistive switching device and synaptic behavior

This study investigates the integration of iodine as a dopant into lithium perrhenate (LiReO4) using rhenium oxide (Re2O7) as a precursor, aiming to enhance the electrical conductivity of the compound, with implications for various material science applications. The iodine variants-incorporated LiReO4 (Iv@LiReO4·H2O) exhibits a notable improvement in the average electrical conductivity (∼ 2 S/cm) and light absorption across a broad spectrum, extending to the near-infrared range. Additionally, the doping process induces resistive switching and synaptic behaviors due to iodine diffusion. The presence of iodine and the partial reduction of Re oxide is confirmed by X-ray photoelectron spectroscopy, while UV-Vis absorption spectroscopy indicates the existence of incorporated iodine variants species. Density functional theory calculations further elucidate the mechanism by which iodine doping triggers a significant enhancement in electrical conductance, primarily by introducing a mixed-valence oxidation state in rhenium, leading to the narrowing of the band gap from 3.68 eV to 0.33 eV. This research suggests potential uses for Iv@LiReO4·H2O in applications to memristive neuromorphic devices.

Memristive feature and mechanism induced by laser-doping in defect-free 2D semiconductor materials

Memristors as non-volatile memory devices have gained numerous attentions owing to their advantages in storage, in-memory computing, synaptic applications, etc. In recent years, two-dimensional (2D) materials with moderate defects have been discovered to exist memristive feature. However, it is very difficult to obtain moderate defect degree in 2D materials, and studied on modulation means and mechanism becomes urgent and essential. In this work, we realized memristive feature with a bipolar switching and a configurable on/off ratio in a two-terminal MoS2 device (on/off ratio ~100), for the first time, from absent to present using laser-modulation to few-layer defect-free MoS2 (about 10 layers), and its retention time in both high resistance state and low resistance state can reach 2 × 104 s. The mechanism of the laser-induced memristive feature has been cleared by dynamic Monte Carlo simulations and first-principles calculations. Furthermore, we verified the universality of the laser-modulation by investigating other 2D materials of TMDs. Our work will open a route to modulate and optimize the performance of 2D semiconductor memristive devices.

Digital-analog hybrid control of an inverted pendulum robot based on memristor

Abstract Single inertial measurement unit sensor signals are inaccurate, and traditional digital processing systems with separate storage and computation architectures cannot directly handle analog sensor signals, leading to significant delays and high power consumption. In that case, the control capability for self-balancing robots has been limited. In this work, a three-terminal MoS2-based memristor device was fabricated and successfully integrated into a memristor-based hybrid control inverted pendulum robot system. This system achieved continuous multi-sensor analog data fusion computations and analog PD control computations by combining the least squares method and a gradient descent algorithm based on a reward mechanism in digital circuits to control the memristors dynamically. Additionally, simulation testing and real, inverted pendulum robot control confirmed the efficacy of its data fusion and the accuracy of its PD control. The comparative experiments demonstrated that the hybrid control system based on memristors was approximately 86.41 mJ lower than the digital system. This system holds significant importance for the processing of large-scale sensor data and robot control.

Analog Implementation of a Spiking Neuron with Memristive Synapses for Deep Learning Processing

Analog neuromorphic prototyping is essential for designing and testing spiking neuron models that use memristive devices as synapses. These prototypes can have various circuit configurations, implying different response behaviors that custom silicon designs lack. The prototype’s behavior results can be optimized for a specific foundry node, which can be used to produce a customized on-chip parallel deep neural network. Spiking neurons mimic how the biological neurons in the brain communicate through electrical potentials. Doing so enables more powerful and efficient functionality than traditional artificial neural networks that run on von Neumann computers or graphic processing unit-based platforms. Therefore, on-chip parallel deep neural network technology can accelerate deep learning processing, aiming to exploit the brain’s unique features of asynchronous and event-driven processing by leveraging the neuromorphic hardware’s inherent parallelism and analog computation capabilities. This paper presents the design and implementation of a leaky integrate-and-fire (LIF) neuron prototype implemented with commercially available components on a PCB board. The simulations conducted in LTSpice agree well with the electrical test measurements. The results demonstrate that this design can be used to interconnect many boards to build layers of physical spiking neurons, with spike-timing-dependent plasticity as the primary learning algorithm, contributing to the realization of experiments in the early stage of adopting analog neuromorphic computing.