Memristive Switching Behavior Research Articles

Carrier-doping-driven insulator-metal transition in disordered materials for memristive switching with high uniformity

Attaining highly uniform operations in a disordered system presents a persistent challenge. The utilization of ion migration in amorphous materials to trigger the resistive switching process of the material usually results in inferior uniformity of the memristive device. Here, we demonstrate that the resistive switching behavior can be activated through carrier doping in the disorder system, and highly ordered resistance modulation is achieved in Ag-doped albumen. By manipulating the doping level of the carrier, the localization of the free electron wavefunction can be tuned, leading to multi-level variations in resistance. This memristive switching behavior is in all electronic and displays excellent switching uniformity, holding great potential for applications in high-density memories and neuromorphic computing chips.

A physics-based predictive model for pulse design to realize high-performance memristive neural networks

Memristive neural networks have extensively been investigated for their capability in handling various artificial intelligence tasks. The training performance of memristive neural networks depends on the pulse scheme applied to the constituent memristors. However, the design of the pulse scheme in most previous studies was approached in an empirical manner or through a trial-and-error method. Here, we choose ferroelectric tunnel junction (FTJ) as a model memristor and demonstrate a physics-based predictive model for the pulse design to achieve high training performance. This predictive model comprises a physical model for FTJ that can adequately describe the polarization switching and memristive switching behaviors of the FTJ and an FTJ-based neural network that uses the long-term potentiation (LTP)/long-term depression (LTD) characteristics of the FTJ for the weight update. Simulation results based on the predictive model demonstrate that the LTP/LTD characteristics with a good trade-off between ON/OFF ratio, nonlinearity, and asymmetry can lead to high training accuracies for the FTJ-based neural network. Moreover, it is revealed that an amplitude-increasing pulse scheme may be the most favorable pulse scheme as it offers the widest ranges of pulse amplitudes and widths for achieving high accuracies. This study may provide useful guidance for the pulse design in the experimental development of high-performance memristive neural networks.

Field-Free Current-Induced Switching of L 1 0 - FePt Using Interlayer Exchange Coupling for Neuromorphic Computing

$L{1}_{0}$-phase $\mathrm{Fe}\mathrm{Pt}$ is well known for its unusually robust perpendicular magnetic anisotropy (PMA) properties arising from strong conduction-electron spin-orbit coupling (SOC) with the $\mathrm{Fe}$ orbital moment. The strong PMA enables stable magnetic storage and memory devices with ultrahigh capacity. Meanwhile, SOC is also the premise of the recently discovered spin-orbit-torque (SOT) effect, which opens avenues for possible electrical manipulation of magnetization for $L{1}_{0}$-$\mathrm{Fe}\mathrm{Pt}$. The bulk SOT of the $L{1}_{0}$-$\mathrm{Fe}\mathrm{Pt}$ single layer was discovered recently; this leads to the magnetization of $L{1}_{0}$-$\mathrm{Fe}\mathrm{Pt}$ reversibly switching on itself. However, deterministic SOT switching of bulk perpendicularly magnetized $\mathrm{Fe}\mathrm{Pt}$ magnets relies on an external magnetic field to break the symmetry. Here, the symmetry-breaking issue is resolved by interlayer exchange coupling, where the $\mathrm{Fe}\mathrm{Pt}$ layer is coupled with an in-plane magnetized $\mathrm{Ni}\mathrm{Fe}$ layer through a $\mathrm{Ti}\mathrm{N}$ spacer layer. Furthermore, our device also presents memristive or gradual switching behaviors, even without an external field, offering the potential for constructing spin synapses and spin neurons for neuromorphic computing. An artificial neural network with high accuracy (\ensuremath{\sim}91.17%) is realized based on the constructed synapses and neurons. Our work paves the way for field-free SOT switching of single bulk PMA magnets and their potential applications in neuromorphic computing.

Humidity‐Enabled Organic Artificial Synaptic Devices with Ultrahigh Moisture Resistivity

AbstractOrganic memristors with homogeneous resistive switching behaviors can emulate the functionalities of biological synapses, making them promising candidates for use in brain‐inspired neuromorphic computing. Among various environmental parameters, humidity affecting device operational stability is a critical issue that organic memristors face in practical applications. A high humidity resistant organic memristor is demonstrated here using a small molecule material F16CuPc. Unencapsulated devices show excellent moisture stability, maintaining reliable memristive switching behavior after 1 year of exposure to air at 60% high relative humidity and remaining operational when immersed in water for 96 h. A proton conduction mechanism is discovered responsible for the resistive switching effect. The continuously adjustable conductance in F16CuPc‐based memristive devices enables the mimicking of biological synapse functions. These results open up a path to a humidity tolerant neuromorphic computing system based on high humidity resistant artificial synapse devices.

Tailoring Neuromorphic Switching by CuNx -Mediated Orbital Currents

Current-induced spin-orbit torque (SOT) is regarded as a promising mechanism for driving neuromorphic behavior in spin-orbitronic devices. In principle, the strong SOT in a heavy-metal-based magnetic heterostructure is attributed to the spin-orbit coupling (SOC)-induced spin Hall effect and/or the spin Rashba-Edelstein effect. Recently, SOC-free mechanisms such as the orbital-angular-momentum-induced orbital Hall effect and/or the orbital Rashba-Edelstein effect have been proposed to generate sizable torques comparable to those from the conventional spin Hall mechanism. In this work, we show that the orbital current can be effectively generated by the nitrided light metal $\mathrm{Cu}$. The overall dampinglike SOT efficiency, which consists of both the spin and the orbital current contributions, can be tailored from ${\ensuremath{\xi}}_{\mathrm{DL}}\ensuremath{\approx}0.06\text{ to }0.4$ in a $\mathrm{Pt}/\mathrm{Co}/{\mathrm{Cu}\mathrm{N}}_{x}$ magnetic heterostructure by tuning the nitrogen doping concentration. Current-induced magnetization switching further verifies the efficacy of such orbital current with a critical switching current density as low as ${J}_{c}$ \ensuremath{\sim} 5 \ifmmode\times\else\texttimes\fi{} ${10}^{10}\phantom{\rule{0.25em}{0ex}}{\mathrm{A}/\mathrm{m}}^{2}$. Most importantly, orbital-current-mediated memristive switching behavior can be observed in such heterostructures, which reveals that gigantic SOT and efficient magnetization switching are the trade-offs for the applicable window of memristive switching. Our work provides insights into the role that orbital current might play in SOT neuromorphic devices for making energy-efficient spin-orbitronic devices.

Role of Defects and Power Dissipation on Ferroelectric Memristive Switching

AbstractAdvancement of information technology requires low power, high speed, and large capacity non‐volatile memory. Memristors have potential applications for not only information storage but also neuromorphic computation. Memristive devices are mostly focused on the use of binary oxides as the resistive switching materials. On the other hand, polarization assisted memristive devices based on ternary ferroelectric oxides are attracting more attention due to their unique switching properties. However, the underlying switching mechanisms and the current–voltage rotation direction are still not fully understood yet. By comparing stoichiometric BaTiO3, BiFeO3, and Bi1‐xFeO3‐δ ferroelectric memristors with different cation stoichiometry, it is found that off‐stoichiometry‐induced traps can play a critical role in controlling the ferroelectric memristive switching behavior. Ferroelectrics with slight off‐stoichiometry show greatly enhanced switching properties, and the switching on/off ratio is mainly determined by the trap energy levels and concentrations. The rotation direction of current–voltage hysteresis loop is affected by the defects, which can be controlled by synthesis and power dissipation. These findings provide insight in understanding the role of defects in ferroelectric memristors and offer guidance to design ferroelectric memristors with enhanced performance.

Memristive and biological synaptic behavior in transition metal dichalcogenide-WS2 nanostructures: A review

Recently, two dimensional (2D) layered materials are being identified as potential candidates for different device applications such as memristors and brain inspired computing. These materials exhibit unusual physical properties such as atomic thickness, mechanically robust, tunable electrical and optical properties. These fascinating properties of 2D materials are very difficult to achieve with traditional materials used for different electronic purposes. The 2D materials are promising for high performance synapses and artificial neurons because of their high energy efficiency, excellent scalability, and high integration density. The present article gives an overview on the properties of 2D transition metal dichalcogenides (WS2) based electronic devices for memristive switching and biological synaptic behavior. Moreover, its integration into large-scale devices and associated challenges are being summarized along with the future prospective.

Magnetization Switching by Spin–Orbit Torque in L10-FePt and Ta/FePt Films With Large Perpendicular Magnetic Anisotropy

The current induced perpendicular magnetization switching in L10-FePt single layer and Ta/FePt bilayer thin films was investigated. Here, L10-FePt (6 nm) single-layer film and Ta (1-5 nm)/FePt (6 nm) bilayer films were deposited on MgO substrate at 700°C, showing good perpendicular anisotropy with the out-of-plane coercivity of 7.5 kOe for FePt and around 5 kOe for Ta/FePt bilayers. It is found that current-induced partial magnetization-based switching was realized with the SOT switching ratio of about 3% in all films. For adding the Ta capping layer, the critical current density is reduced and the stability of the switching loop is relatively better as compared with the single FePt layer. Furthermore, harmonic Hall voltage results demonstrate that the SOT efficiency (βDL) of FePt single-layer is approximately 50 Oe/ (107A cm-2), significantly higher than that of Ta/FePt bilayers, which is several times higher than that of the traditional Ta/CoFeB/MgO structure reported before. More importantly, Ta/FePt bilayer structure exhibits very good memristive switching behavior, which is expected to be applied to the spin-synapse of neural networks in the future.

Electronic synapses mimicked in bilayer organic-inorganic heterojunction based memristor

Abstract Electronic synapses implementing in-memory computing system could overcome the developing limitation on the energy efficiency of traditional von Neumann architecture. Compared with the high sensitivity of biological synapses, lower responsivity of the memristive synapses was found via the electrical stimulations. Here, poly{2,2-(2,5-bis(2-octyldodecyl)-3,6-dioxo-2,3,5,6- tetrahydropyrrolo[3,4-c]pyrrole-1,4-diyl)-dithieno[3,2-b]thiophene-5,5-diyl-alt-thiophen-2,5-diyl} (PDPPBTT)/zinc oxide (ZnO) based heterojunction is found to exhibit stable memristive switching behavior, which originates from the confined formation/rupture of filament in the two-layer interface region as the ions migrate with different transport rates in two layers. The implementing synaptic functions in the sensitive memristive device can realize the short-term plasticity and long-term plasticity when stimulated by the applied electrical signals with different stimulating rate. Similar to the biological synapse, the memory loss, memory transition, and the critical role of stimulation rate on the transition process, can be achieved in the as-prepared memristor device. The systematic demonstrations on the synaptic emulation may facilitate building bio-inspired device-level neuromorphic systems.

Pulse-width and temperature dependence of memristive spin–orbit torque switching

It is crucial that magnetic memory devices formed from magnetic heterostructures possess sizable spin–orbit torque (SOT) efficiency and high thermal stability to realize both efficient SOT control and robust storage of such memory devices. However, most previous studies on various types of magnetic heterostructures have focused on only their SOT efficiencies, whereas the thermal stabilities therein have been largely ignored. In this work, we study the temperature-dependent SOT and stability properties of two types of W-based heterostructures, namely, W/CoFeB/MgO (standard) and CoFeB/W/CoFeB/MgO (field-free), from 25 °C (298 K) to 80 °C (353 K). Via temperature-dependent SOT characterization, the SOT efficacies for both systems are found to be invariant within the range of studied temperatures. Temperature-dependent current-induced SOT switching measurements further show that the critical switching current densities decrease with respect to the ambient temperature; thermal stability factors (Δ) are also found to degrade as temperature increases for both standard and field-free systems. The memristive SOT switching behaviors in both systems with various pulse-widths and temperatures are also examined. Our results suggest that, although the SOT efficacy is robust against thermal effects, the reduction of Δ at elevated temperatures could be detrimental to standard memory as well as neuromorphic (memristive) device applications.

In Situ Hydrogen Plasma Exposure for Varying the Stoichiometry of Atomic Layer Deposited Niobium Oxide Films for Use in Neuromorphic Computing Applications.

Niobium oxide (NbOx) materials of various compositions are of interest for neuromorphic systems that rely on memristive device behavior. In this study, we vary the composition of NbOx thin films deposited via atomic layer deposition (ALD) by incorporating one or more in situ hydrogen plasma exposure steps during the ALD supercycle. Films with compositions ranging from Nb2O5 to NbO2 were deposited, with film composition dependent on the duration of the plasma exposure step, the number of plasma exposure steps per ALD supercycle, and the hydrogen content of the plasma. The chemical and optical properties of the ALD NbOx films were probed using spectral ellipsometry, X-ray photoelectron spectroscopy, and optical transmission spectroscopy. Two-terminal electrical devices fabricated from ALD Nb2O5 and NbO2 thin films exhibited memristive switching behavior, with switching in the NbO2 devices achieved without a high-field electroforming step. The ability to controllably tune the composition of ALD-grown NbOx films opens new opportunities for realizing a variety of device structures relevant for neuromorphic computing and other emerging electronic and optoelectronic applications.

Self‐Assembled NiO Nanocrystal Arrays as Memristive Elements

AbstractMemristive switching, a nonlinear current–voltage (I–V) characteristic, has seen a tremendous surge in interest as an approach to achieve implementation of synaptic functions. The memristive switching behavior of self‐assembled NiO nanocrystals is investigated via scanning probe microscopy, based on first‐order reversal curve current–voltage spectroscopy. Synaptic switching is clearly observed as a direct consequence of filament growth (i.e., gradually increased conductance) in the nanocrystals. A spatial dependency of the conduction in the nanocrystals suggests that there is a localization of the switching filament. The current understanding of this localization ignores features related to local lateral variation current, which can generate an excessive local heat and temperature such as electrical dissipation. The observation of low electrical dissipation at the edge of the nanocrystals shows that less energy is wasted as heat such that the bias applied can be utilized more efficiently to assist the nucleation of the filament and thus reduces the power consumption. Electrical power dissipation is also found to scale with nanocrystal height and has spatial dependence within the nanocrystals. The combination of synaptic switching and high density of the nanocrystals demonstrate that it is feasible to exploit them to create a basic architecture for neuromorphic memory devices.

Bipolar Cu/HfO2/p++ Si Memristors by Sol-Gel Spin Coating Method and Their Application to Environmental Sensing

In this paper, the memristive switching behavior of Cu/ HfO2/p++ Si devices fabricated by an organic-polymer-assisted sol-gel spin-coating method, coupled with post-annealing and shadow-mask metal sputtering steps, is examined. HfO2 layers of about 190 nm and 80 nm, are established using cost-effective spin-coating method, at deposition speeds of 2000 and 4000 rotations per minute (RPM), respectively. For two types of devices, the memristive characteristics (Von, Ion, and Vreset) and device-to-device electrical repeatability are primarily discussed in correlation with the oxide layer uniformity and thickness. The devices presented in this work exhibit an electroforming free and bipolar memory-resistive switching behavior that is typical of an Electrochemical Metallization (ECM) I-V fingerprint. The sample devices deposited at 4000 RPM generally show less variation in electrical performance parameters compared to those prepared at halved spin-coating speed. Typically, the samples prepared at 4000 RPM (n = 8) display a mean switching voltage Von of 3.0 V (±0.3) and mean reset voltage Vreset of −1.1 V (±0.5) over 50 consecutive sweep cycles. These devices exhibit a large Roff/Ron window (up to 104), and sufficient electrical endurance and retention properties to be further examined for radiation sensing. As they exhibit less statistical uncertainty compared to the samples fabricated at 2000 RPM, the devices prepared at 4000 RPM are tested for the detection of soft gamma rays (emitted from low-activity Cs-137 and Am-241 radioactive sources), by assessing the variation in the on-state resistance value upon exposure. The analysis of the probability distributions of the logarithmic Ron values measured over repeated ON-OFF cycles, before, during and after exposing the devices to radiation, demonstrate a statistical difference. These results pave the way for the fabrication and development of cost-effective soft-gamma ray detectors.

A study on optical properties of Sb2Se3 thin films and resistive switching behavior in Ag/Sb2Se3/W heterojunctions

Abstract In this study, optical phase change property and bipolar memristive switching behavior were systematically characterized in Sb2Se3 thin films with different annealing temperatures and Ag/Sb2Se3/W heterojunction, respectively. The structural investigations reveal that the crystallinity of Sb2Se3 sample improves with increasing annealing temperature. Anomalous changes of both structure and optical properties are attributed to the increase of defect states at an excessively high annealing temperature. Additionally, the absorption edge shift towards longer wavelength after thermal treatment, indicating a decrease in the optical band gap. In the aspect of electrical characteristics, the Ag/Sb2Se3/W cell has a bipolar memristive switching behavior and shows reversible switching property. The study based on the optical properties and memristive switching behaviors of Sb2Se3 thin films offers an important understanding to the applications of Sb2Se3 and other chalcogenide semiconductor as promising materials for integrated optical system and chalcogenide-based memory devices.

Analog Memristive Characteristics and Conditioned Reflex Study Based on Au/ZnO/ITO Devices

As the fourth basic electronic component, the application fields of the memristive devices are diverse. The digital resistive switching with sudden resistance change is suitable for the applications of information storage, while the analog memristive devices with gradual resistance change are required in the neural system simulation. In this paper, a transparent device of ZnO films deposited by the magnetron sputtering on indium tin oxides (ITO) glass was firstly prepared and found to show typical analog memristive switching behaviors, including an I–V curve that exhibits a ‘pinched hysteresis loops’ fingerprint. The conductive mechanism of the device was discussed, and the LTspice model was built to emulate the pinched hysteresis loops of the I–V curve. Based on the LTspice model and the Pavlov training circuit, a conditioned reflex experiment has been successfully completed both in the computer simulation and the physical analog circuits. The prepared device also displayed synapses-like characteristics, in which resistance decreased and gradually stabilized with time under the excitation of a series of voltage pulse signals.

(Invited) 2D MoS2 Memristors with Variable Switching Characteristics

Memristors has been extensively studied as an important candidate device structure for realizing non-Von Neumann computers, neuromorphic and analog computing systems. The conventional memristors are made from transition metal oxides such as TiOx, TaOx, WOx, and HfOx.1 The conductance states of such memristors can be adjusted through modulating the distribution of oxygen vacancies in the oxide channels. Recently, memristor-like transport behaviors have been also observed in the electronic devices made from emerging 2D-layered transition metal dichalcogenides (TMDCs, such as MoS2 and WSe2).2 Such emerging 2D-material-based memristors could be used for constructing novel neural networks with a significantly improved level of connectivity and enabling bio-realistic emulation of brain neuron functions. Therefore, it is highly desirable to further perform nanofabrication and nanoelectronics research to fully explore the applicability and advantage of 2D-TMDC-based memristors for practical neuromorphic applications. Here, we present a study on the nanofabrication and characterization of memristive devices based on mechanically printed few-layer MoS2 structures. This work shows that mechanically printed MoS2 memristors with critical thicknesses exhibit a very low threshold electric field for initiating memristive-switching processes, which is essential for the successful construction and operation of energy-efficient memristor-based neural networks. More importantly, such MoS2 memristors exhibit both analogue (i.e., gradient) and discrete switching characteristics, which could be further exploited for making reconfigurable memristors for both analog and digital computing applications. This work leveraged the unique transport properties of 2D layered MoS2 for memristor-related applications and advanced the scientific/technical knowledge for controlling memristive switching behaviors.

MemSens: Memristor-Based Radiation Sensor

Resistive random-access memory (RRAM) technology has been gaining importance due to scalability, low power, non-volatility, and the ability to perform in-memory computing. The RRAM sensing applications have also emerged to enable single RRAM technology platforms which include sensing, data storage, and computing. This paper reports on sol-gel drop coated low-power <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu $ </tex-math></inline-formula> -thick Ag/TiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> /Cu memristor, named MemSens, developed for radiation sensing. MemSens exhibits a bipolar memristive switching behavior within a small voltage window, ranging up to +0.7 V for the turn-ON, and down to −0.2 V for the turn-OFF. Under these operating conditions, MemSens has 67% less switching voltage, 20% drop in ON switching current, 75% reduced active area and > 3x improved device endurance, compared to the best characteristics reported in the literature for <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu $ </tex-math></inline-formula> -thick memristors. The device is tested under direct exposure to ionizing Cs-137 662keV <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\gamma $ </tex-math></inline-formula> -rays, during which a significant increase in the electrical conductivity of the device is observed. MemSens circuit is proposed to allow a relatively real time and cost-effective radiation detection. This provides a first insight to the advancement of reliable memristors that could potentially be deployed in future low-power radiation sensing technologies for medical, personal protection, and other field applications.

Light-Gated Memristor with Integrated Logic and Memory Functions.

Memristive devices are able to store and process information, which offers several key advantages over the transistor-based architectures. However, most of the two-terminal memristive devices have fixed functions once made and cannot be reconfigured for other situations. Here, we propose and demonstrate a memristive device "memlogic" (memory logic) as a nonvolatile switch of logic operations integrated with memory function in a single light-gated memristor. Based on nonvolatile light-modulated memristive switching behavior, a single memlogic cell is able to achieve optical and electrical mixed basic Boolean logic of reconfigurable "AND", "OR", and "NOT" operations. Furthermore, the single memlogic cell is also capable of functioning as an optical adder and digital-to-analog converter. All the memlogic outputs are memristive for in situ data storage due to the nonvolatile resistive switching and persistent photoconductivity effects. Thus, as a memdevice, the memlogic has potential for not only simplifying the programmable logic circuits but also building memristive multifunctional optoelectronics.

Switching characteristics of microscale unipolar Pd/Hf/HfO2/Pd memristors

In this paper, we investigate the switching behavior of nano-thick microscale Pd/Hf/HfO2-10nm/Pd memristors. Several key electrical characteristics such as the forming voltage and the SET/RESET IV parameters are studied as a function of the device size, the hafnium-capping thickness and the environmental temperature. The fabricated stack exhibits a unipolar switching behavior that is characteristic of two interfacial oxide and buffer layers consisting of the same transition metal. The average forming voltage value decreases oppositely to the microscale device area while it increases with the Hf-capping layer thickness. The unipolar memristive switching behavior is found to be polarity-dependent due to diverse contributions of the top Hf-buffer layer and the bottom Pd electrode to the switching mechanism. Joule heating effects involving thermophoresis and diffusion are concluded to be predominantly governing the ON/OFF switching events in the positive polarity. Synergistic electric field effects are found to additionally control the unipolar switching behavior in the negative mode. The respective key roles of the top Hf-buffer layer and the bottom Pd electrode in the asymmetric unipolar memristive switching are elucidated at the molecular level according to well-established transport phenomena reported for oxygen vacancies. Given the importance of heat in controlling the conductance state of the proposed stack, the memristor's electrical sensitivity to changes in the ambient temperature is preliminarily investigated for potential temperature sensing applications.

Reliable Memristive Switching Memory Devices Enabled by Densely Packed Silver Nanocone Arrays as Electric-Field Concentrators.

Memristor devices based on electrochemical metallization operate through electrochemical formation/dissolution of nanoscale metallic filaments, and they are considered a promising future nonvolatile memory because of their outstanding characteristics over conventional charge-based memories. However, nanoscale conductive paths or filaments precipitated from the redox process of metallic elements are randomly formed inside oxides, resulting in unexpected and stochastic memristive switching parameters including the operating voltage and the resistance state. Here, we present the guided formation of conductive filaments in Ag nanocone/SiO2 nanomesh/Pt memristors fabricated by high-resolution nanotransfer printing. Consequently, the uniformity of the memristive switching behavior is significantly improved by the existence of electric-field concentrator arrays consisting of Ag nanocones embedded in SiO2 nanomesh structures. This selective and controlled filament growth was experimentally supported by analyzing simultaneously the surface morphology and current-mapping results using conductive atomic force microscopy. Moreover, stable multilevel switching operations with four discrete conduction states were achieved by the nanopatterned memristor device, demonstrating its potential in high-density nanoscale memory devices.