Oxide-based Resistive Random Access Memory Research Articles

Revealing the quantum nature of the voltage-induced conductance changes in oxygen engineered yttrium oxide-based RRAM devices

In this work, the quasi-analog to discrete transition occurring in the current–voltage characteristic of oxygen engineered yttrium oxide-based resistive random-access memory (RRAM) devices is investigated in detail. In particular, the focus of our research is not on the absolute conductance values of this characteristic but on the magnitude of its conductance changes occurring during the reset process of the device. It is found that the detected changes correspond to conductance values predominantly of the order of the quantum unit of conductance G0 = 2e2/h, where e is the electron charge and h the Planck constant. This feature is observed even at conductance levels far above G0, i.e. where electron transport is seemingly diffusive. It is also observed that such behavior is reproducible across devices comprising yttrium oxide layers with different oxygen concentrations and measured under different voltage sweep rates. While the oxygen deficiency affects the total number of quantized conductance states, the magnitude of the changes in conductance, close to 1 G0, is invariant to the oxygen content of the functional layer.

Enhancing simulation feasibility for multi-layer 2D MoS2 RRAM devices: reliability performance learnings from a passive network model.

While two-dimensional (2D) MoS2 has recently shown promise as a material for resistive random-access memory (RRAM) devices due to its demonstrated resistive switching (RS) characteristics, its practical application faces a significant challenge in industry regarding its limited yield and endurance. Our earlier work introduced an effective switching layer model to understand RS behavior in both mono- and multi-layered MoS2. However, functioning as a phenomenological percolation modeling tool, it lacks the capability to accurately simulate the intricate current-voltage (I-V) characteristics of the device, thereby hindering its practical applicability in 2D RRAM research. In contrast to the established conductive filament model for oxide-based RRAM, the RS mechanism in 2D RRAM remains elusive. This paper presents a novel simulator aimed at providing an intuitive, visual representation of the stochastic behaviors involved in the RS process of multi-layer 2D MoS2 RRAM devices. Building upon the previously proposed phenomenological simulator for 2D RRAM, users can now simulate both the I-V characteristics and the resistive switching behaviors of the RRAM devices. Through comparison with experimental data, it was observed that yield and endurance characteristics are linked to defect distributions in MoS2.

Physical model simulations of Hf oxide resistive random access memory device with a spike electrode structure

Resistive memory has become an attractive new memory type due to its outstanding performance. Oxide-based resistive random access memory is one type of widely used memory whose resistance can be transformed by applying current or voltage. Memristors are widely used in various kinds of memories and neural morphological calculations. Therefore, it is of vital importance to understand the physical change mechanism of an internal memristor under stimulation to improve electrical properties of the memristor. In our studies, a device model based on Hf oxide was proposed, then completely processes of the forming, reset and set were simulated. Meantime, the generation and recombination of oxygen vacancies were considered in all the processes, making the simulation more practical. In addition, a spike electrode structure was applied, a gathering electric field can be generated in the oxide layer so that the improved device has a faster forming voltage, lower forming current and lower instantaneous power consumption in the ON state. Finally, the effects of spike electrode length on the forming process were studied, the research results reveal that a longer probe electrode can engage a lower forming voltage and accelerate the formation of conductive filaments.

Sol-Gel-Processed Y2O3-Al2O3 Mixed Oxide-Based Resistive Random-Access-Memory Devices.

Herein, sol-gel-processed Y2O3-Al2O3 mixed oxide-based resistive random-access-memory (RRAM) devices with different proportions of the involved Y2O3 and Al2O3 precursors were fabricated on indium tin oxide/glass substrates. The corresponding structural, chemical, and electrical properties were investigated. The fabricated devices exhibited conventional bipolar RRAM characteristics without requiring a high-voltage forming process. With an increase in the percentage of Al2O3 precursor above 50 mol%, the crystallinity reduced, with the amorphous phase increasing owing to internal stress. Moreover, with increasing Al2O3 percentage, the lattice oxygen percentage increased and the oxygen vacancy percentage decreased. A 50% Y2O3-50% Al2O3 mixed oxide-based RRAM device exhibited the maximum high-resistance-state/low-resistance-state (HRS/LRS) ratio, as required for a large readout margin and array size. Additionally, this device demonstrated good endurance characteristics, maintaining stability for approximately 100 cycles with a high HRS/LRS ratio (>104). The HRS and LRS resistances were also retained up to 104 s without considerable degradation.

Spike-time dependent plasticity of tailored ZnO nanorod-based resistive memory for synaptic learning

Metal oxide resistive memory is a potential device that can substantially influence the current roadmap for nonvolatile memory and neuromorphic computing. However, common amorphous oxide-based resistive random-access memory suffers from high forming voltages that complicate circuit design and abrupt SET behavior incompatible with analog weight updates. To overcome such limitations, wurtzite ZnO nanorods were synthesized on a fluorine-doped tin oxide (FTO) substrate and a bipolar resistive memory with the Ag/w-ZnO/FTO stacking sequence was fabricated. The hexagonal NR morphology of w-ZnO with controlled vertical growth and nanochannel formation between the NRs were produced by in situ crystalline growth. This morphology enabled a forming-free switching and an analog switching effect that emulated neuromorphic functionalities such as potentiation–depression and complex spike-time dependent plasticity-based Hebbian learning rules. Importantly, the device exhibited nonabrupt switching behavior suitable for analog weight updates in neuromorphic computing in contrast to conventional resistive memory.

Improved synaptic performances with tungsten-doped indium-tin-oxide alloy electrode for tantalum oxide-based resistive random-access memory devices

Improved synaptic performances with tungsten-doped indium-tin-oxide alloy electrode for tantalum oxide-based resistive random-access memory devices

Thermal environment impact on HfOx RRAM operation: A nanoscale thermometry and modeling study

As the demand for computing applications capable of processing large datasets increases, there is a growing need for new in-memory computing technologies. Oxide-based resistive random-access memory (RRAM) devices are promising candidates for such applications because of their industry readiness, endurance, and switching ratio. These analog devices, however, suffer from poor linearity and asymmetry in their analog resistance change. Various reports have found that the temperature in RRAM devices increases locally by more than 1000 K during operation. Therefore, temperature control is of paramount importance for controlling their resistance. In this study, scanning thermal microscopy is used to map the temperature of Au/Ti/HfOx/Au devices at a steady power state and to measure temperature dynamics of the top electrode above the filament location during both resistive switching loops and voltage pulsing. These measurements are used to verify the thermal parameters of a multiphysics finite elements model. The model is then used to understand the impact of thermal conductivities and boundary conductances of constituent materials on resistance change during the first reset pulse in RRAM devices. It is found that the resistance change can be reduced significantly when the temperature in the titanium capping layer is reduced. We find that the greatest temperature reduction and, therefore, the lowest resistance change in the device are afforded by capping layers with increased thermal conductivities. This work links thermal properties to the resistance change in RRAM devices, providing critical insights into engineering devices with improved switching dynamics.

Understanding the Resistive Switching Mechanism of 2-D RRAM: Monte Carlo Modeling and a Proposed Application for Reliability Research

The 2-D materials have become promising candidates for resistive random access memory (RRAM) devices as more unique resistive switching (RS) characteristics have recently been revealed. However, endurance is a major challenge for industrialization. Unlike the well-developed and recognized conductive filament (CF) model for oxide-based RRAM, the RS mechanism for 2-D RRAM is not well understood. In this article, we first review the dissociation–diffusion–adsorption (DDA) model and the cluster model proposed in previous works on monolayer 2-D RRAM devices. The use of a Monte Carlo (MC) simulator for multilayer 2-D RRAM devices to expand the application of DDA and cluster models is then discussed. A simulator was designed to provide an intuitive physical view by visualizing the stochastic behaviors of the RS process in multilayer 2-D RRAM devices. By comparing the simulated results with experimental data, the endurance characteristic was found to be mainly determined by the formation and collapse of an effective switching layer. We also found that the thickness of the effective switching layer is independent of the total thickness of the multilayer 2-D material and the initial status of the device, which is consistent with the experimental observations. The model and results discussed in this work provide additional insights and guidance for improving the reliability of 2-D RRAM devices.

Modulating the resistive switching stability of HfO2-based RRAM through Gd doping engineering: DFT+U.

Oxide-based resistive random access memory (RRAM) is standing out in both non-volatile memory and the emerging field of neuromorphic computing, with the consequence of increasing performance demands. Rare-earth doping is often used as an effective means for performance modulation. In this work, the modulation mechanism of the resistive switching (RS) behaviors in trivalent rare-earth Gd-doped HfO2-based RRAM has been carefully investigated using first-principles calculations. The results of electronic structure analysis show that Gd doping would lead to a change in the local geometry of the m-HfO2 defect system and would enhance the Coulomb interaction between the atoms around Gd and oxygen vacancy (VO), which may be one of the reasons for the enhanced conductivity of the HfO2-based RRAM after Gd doping. Thermodynamic and kinetic study results indicate that there is a strong interaction between Gd and its surrounding VO defects, and this strong interaction would not only attract more oxygen vacancies (VOs) to be generated near the dopant Gd, but also increase the migration energy barrier of the +2 charged VOs around the Gd doping site, thus suppressing the random generation of VO filaments, which leads to a better uniformity of the switching parameters during the RS process and improves the performance stability of the devices. The results of this work will provide new insights into modulating the RS behaviors and improving the device performance of HfO2-based RRAM through doping engineering.

The effect of oxygen affinity electrode Ti on the performance of vanadium oxide-based valence change resistive random access memory

Two top electrode (TE) metals, Pt and Ti, are deposited on VOX/Pt to investigate the effect of oxygen affinity TE on resistive switching (RS). The oxygen affinity electrode Ti significantly improves the variability and durability of the devices compared to the inert TE Pt. The TiOY interfacial layer (IL) plays a vital role in RS as an oxygen reservoir between the Ti TE and the dielectric layer. During repeated RS, a rapid endurance degradation is observed for Pt/VOX/Pt. However, the resistance value in the high-resistance state (HRS) gradually decreases over the RS cycles for Ti/VOX/Pt. The oxygen ions in VOX are consumed and stored in TE Ti owing to the high oxygen affinity of Ti. This process causes an improvement in the uniformity and endurance for Ti/VOX/Pt.

A Unified Current-Voltage Model for Metal Oxide-Based Resistive Random-Access Memory.

Resistive random-access memory (RRAM) is essential for developing neuromorphic devices, and it is still a competitive candidate for future memory devices. In this paper, a unified model is proposed to describe the entire electrical characteristics of RRAM devices, which exhibit two different resistive switching phenomena. To enhance the performance of the model by reflecting the physical properties such as the length index of the undoped area during the switching operation, the Voltage ThrEshold Adaptive Memristor (VTEAM) model and the tungsten-based model are combined to represent two different resistive switching phenomena. The accuracy of the I-V relationship curve tails of the device is improved significantly by adjusting the ranges of unified internal state variables. Furthermore, the unified model describes a variety of electrical characteristics and yields continuous results by using the device's current-voltage relationship without dividing its fitting conditions. The unified model describes the optimized electrical characteristics that reflect the electrical behavior of the device.

Enhancement of Resistive and Synaptic Characteristics in Tantalum Oxide-Based RRAM by Nitrogen Doping.

Resistive random–access memory (RRAM) for neuromorphic systems has received significant attention because of its advantages, such as low power consumption, high–density structure, and high–speed switching. However, variability occurs because of the stochastic nature of conductive filaments (CFs), producing inaccurate results in neuromorphic systems. In this article, we fabricated nitrogen–doped tantalum oxide (TaOx:N)–based resistive switching (RS) memory. The TaOx:N–based device significantly enhanced the RS characteristics compared with a TaOx–based device in terms of resistance variability. It achieved lower device–to–device variability in both low-resistance state (LRS) and high–resistance state (HRS), 8.7% and 48.3% rather than undoped device of 35% and 60.7%. Furthermore, the N–doped device showed a centralized set distribution with a 9.4% variability, while the undoped device exhibited a wider distribution with a 17.2% variability. Concerning pulse endurance, nitrogen doping prevented durability from being degraded. Finally, for synaptic properties, the potentiation and depression of the TaOx:N–based device exhibited a more stable cycle–to–cycle variability of 4.9%, compared with only 13.7% for the TaOx–based device. The proposed nitrogen–doped device is more suitable for neuromorphic systems because, unlike the undoped device, uniformity of conductance can be obtained.

Nano-analytical investigation of the forming process in an HfO2-based resistive switching memory

Metal oxide-based resistive random access memory devices are highly attractive candidates for next-generation nonvolatile memories, but the resistive switching phenomena remain poorly understood. This article focuses on the microscopic understanding of the initial forming step, which is decisive for the switching process. The integrated resistive switching memory effect in Ti/HfO2/TiWN metal insulator metal structures is studied. After forming, transmission electron microscopy investigations pointed out the presence of a funnel-shaped region, in the ON state of the cell, where slightly oxidized Ti (TiOx) was present within HfO2 dielectric. Modeling of the measured ON state conductance of the cell with the semi-classical approximation is consistent with a conductive nanometric TiOx filament (or a sum of sub-nanometric TiOx filaments) present in the funnel-shaped region. The conductive area is likely formed by diffusion after the dielectric breakdown.

Low power multilevel resistive switching in titanium oxide-based RRAM devices by interface engineering

The stochastic nature of conductive filament formation and dissolution always leads to large fluctuations of key device parameters that hinder the practical applications of resistive random-access memories (RRAMs). Here, we report a simple bilayer oxide-based device structure of Al/TiOx/TiOy/FTO (x < y) employed to address this variability issue and improve the overall performance of the memory device. The bipolar resistive switching performance remarkably improved in these bilayer devices with lower forming voltage (~1 V), set/reset voltages of 0.4/-0.6 V, a programming current of 10 μA, an enlarged ON/OFF ratio (>103), longer retention (>103 s), and better uniformity as compared to the control Al/TiOy/FTO device. Moreover, the modulation of TiOx layer thickness enables tunability of switching voltages, currents, and the resistance window. In addition, reliable and reproducible multiple resistance states, more specifically different low resistance states, can be achieved in bilayer devices by controlling the programming current down to 10 μA, which is in contrast to the binary states of the control devices. These improved switching parameters with multilevel storage capability in bilayer devices are attributed to the incorporation of an oxygen vacancy-rich interfacial layer, namely TiOx, which serves as an oxygen vacancy reservoir and facilitates effective conductive filament formation inside the switching layer. Furthermore, the thermodynamics involving the dielectric constant and Gibb's free energy of oxide formation at the Al/TiOx and TiOx/TiOy interfaces play a pivotal role in the associated switching mechanism. The spatial variability of the operating voltage across these devices is found to be as low as 8%. These findings pave the way for low power, low cost, and high density data storage applications.

(Invited) Nanoscale Probing of Field-Driven Ion Migration in TaOx for Neuromorphic Memristor Applications

In semiconducting metal oxides, application of electric fields often causes migration of the constituent ions (oxygen ions, metal cations, and ionic vacancies) via ionic conduction and diffusion. The field-driven ion migration has been considered a key phenomenon in the recent development of oxide-based resistive random access memory (ReRAM) and memristive neuromorphic devices because their operations are generally made through the migration and resultant changes in the valence/resistance states of the oxide layers. Semiconducting amorphous TaOx is known as a practical material of switching layers in ReRAMs, and moreover, has recently been considered of critical importance in the development of neuromorphic hardware because of its variety of neuromorphic resistive switching functions. Many important functions, including multilevel (or analog) resistive switching, stochastic switching, and spike-rate/spike-timing dependent plasticity, have been demonstrated in junctions with a structure of metal/TaOx/metal, and the origin of the neuromorphic functionality has been ascribed to the high ionic conductivities (both for O2− and Ta5+) and degree of structural freedom of amorphous TaOx.Regarding the mechanisms of the neuromorphic switching, however, limited understanding has been achieved in the previous works. The mechanisms have so far been inferred from the switching behaviors in electrical measurements or numerical modeling because of the experimental difficulty in observing nanoscale ion migration during the neuromorphic resistive switching. To directly design of the neuromorphic functions and establish oxide-based neuromorphic computing, therefore, experimental probing of the ion migration processes is now importantly required for neuromorphic resistive switching of TaOx.In this study, we fabricated ultra-flat thin films of amorphous TaOx (3.0–10 nm thick) with a root-mean-square roughness of <0.2 nm by the pulsed laser deposition method, which will offer suitable conditions for ion migration analysis by probe microscopy. In the conductive tip atomic force microscopy investigations, we observed a stable resistive switching phenomenon at arbitrary positions on the TaOx surface by voltage applications from a probe. The positional independence of the switching characteristics ensures the experimental advantages of the ultra-flat films in the investigations. Additionally, this indicates that "easy spots" for resistive switching, of which the formation has been suggested in other metal oxides, are not present in the thin films of amorphous TaOx, and initial clustering of oxygen vacancies is not importantly involved in the switching. When multiple voltage stresses of >3.0 V were applied from a probe, we observed a multilevel switching phenomenon in the TaOx thin films. In the measurements, we observed that the multilevel switching appears with very small amount of structural changes of TaOx (<1.0 nm in the out-of-plane direction), and relatively large-scale structural deformations (2.0 nm in the out-of-plane direction and 150 nm in the in-plane direction) started to be observed only after breakdown-like behaviors in the resistance with formation of a 10 nm-sized conductive spot. This indicates that the intrinsic ion migration distance involved in the multilevel switching of TaOx is in a range of <1.0 nm in out-of-plane direction, which is smaller than previous predictions. From these results, the spatial extents of neuromorphic resistive switching of TaOx were revealed in this study, which will offer important insights into the process development for TaOx-based neuromorphic memristor.

RRAM Devices with Plasma Treated HfO2 with Ru as Top Electrode for In-Memory Computing Hardware

Recently, many non-volatile memory devices such as resistive random-access memory (RRAM) have shown promise as candidates for analog in-memory computing applications related to inference and training in artificial intelligence. A RRAM cell has an MIM (metal insulator metal) structure that exhibits reversible resistive switching on application of positive or negative voltage. HfO2-based dielectrics with controlled distribution of defects or oxygen vacancies can potentially enable low power switching and multi-resistance levels. However, detailed studies of the power consumption, repeatability and retention during multi-level operation have not been undertaken previously. Transition metal oxide-based RRAM, using HfO2, executes change in resistance (switching behavior) via electrochemical migration of oxygen vacancies. Lack of enough oxygen vacancies in stoichiometric HfO2 limits the low power operation. Therefore, HfOx (with x < 2) has been studied for many RRAM devices (1). A uniform distribution of defects in HfO2 may help in switching but may not help in reducing the power. The distribution of these defects throughout the bulk of HfOx is, therefore, not well-understood. It was recently reported that high oxygen vacancy concentration close to top electrode reduces switching power (2).This work investigates the role of extra oxygen vacancies, introduced by plasma for a 6 nm thick HfO2 during deposition to reduce the power consumption. Initially TiN, which is a commonly used metal in CMOS technology, was used as the top electrode for treated HfO2 using a Si/PVD Ti/PVD TiN/HfO2/PVD TiN structure. A range of compliance currents (CC) from 1 nA to 500 nA were used to evaluate the switching characteristics. The TiN devices switch at and above a CC of 50 nA and were tested for sustainability for hundreds of cycles. In addition to oxygen vacancy rich treated-HfO2 near the top electrode, TaN and Ru electrodes were explored to enhance the switching behavior and power consumption. The role of both TaN and Ru as bottom metal was also evaluated. With Ru as top metal the Si/PVD Ti/PVD TiN/HfO2/Ru/PVD TiN structure, switched at 1 nA CC and at higher CC. Whereas when Ru was used as a bottom electrode, a Si/PVD Ti/PVD TiN/Ru/HfO2/ PVD TiN structure, devices were unable to switch below a CC of 50 mA. For TaN as top metal electrode, devices switched at and above 1 mA CC whereas with TaN as bottom metal the initial switching was at CC of 2 mA. It was concluded that use of Ru as a top metal significantly reduced the switching energy of the treated HfO2 RRAM device but was ineffective when used as a bottom metal. Potential mechanism for this observation will be discussed. References Chen, W. Lu, B. Long, Y. Li, D. Gilmer, G. Bersuker, S. Bhunia, and R. Jha, Semiconductor Science and Technology 30, 075002 (2015).Misra, P. Zhao, D. H. Triyoso, V. Kaushik, K. Tapily, R. D. Clark, S. Consiglio,2, T. Hakamata, C. S. Wajda, and G. J. Leusink, ECS J. Solid State Sci. Technol. 9, 05300, (2020).

High performance and low power consumption resistive random access memory with Ag/Fe2O3/Pt structure

Due to magnetic field tunability and the abundance of iron in the Earth’s crust, iron oxide-based resistive random access memory (RRAM) is considered to be low cost and potential for multi-level storage. However, the relatively high operation voltage (>1 V) and small storage window (<100) limit its application. In this work, the devices with simple Ag/Fe2O3/Pt structure exhibit typical bipolar resistive switching with ultralow set voltage (V set) of 0.16 V, ultralow reset voltage (V reset) of −0.04 V, high OFF/ON resistance ratio of 103, excellent cycling endurance more than 104 and good retention time longer than 104 s. Each major parameter has about an order of magnitude improvement compared to the previous data. The devices demonstrate outstanding stable low power consumption quality. Based on the analysis of the experimental results, a percolation model of silver ion migration was established and confirmed that low operation voltage is attributed to the amorphous oxide layer with large porosity. During electrical testing, the compliance current (I c) and maximum reset voltage (V max) can also affect the device performance. This discovery suggests Fe2O3 memristor has significant potential for application and provides a new idea for the realization of high-performance low-power RRAM.

Clarifying the switching layer transformation through analysis of an abnormal I–V curves with increasing set compliance current in oxide-based resistive random access memory

This work investigates an abnormal rising current in high resistance state (HRS) during the reset process that occurs when increasing the sweeping set compliance current (SCC). When the SCC is increased, the HRS current read at a lower voltage (−0.1 V) remains similar, yet exhibits a gradual rise when read at a higher voltage (−0.8 V). In order to clarify this abnormal phenomenon, the current fittings were performed to obtain the corresponding switching layer (SL) properties after set operations with different SCCs. Furthermore, COMSOL electric field simulations with different SL properties were executed to verify the mechanism.

Physics-Based Stochastic Three-Dimensional Modeling for Metal–Oxide Resistive Random Access Memory

In this work, a fully physics-based stochastic 3-D modeling of metal–oxide-based resistive random access memory (RRAM) device has been conducted. In contrast to general 1-D or 2-D modeling of RRAM devices, 3-D modeling is able to provide a picture of the evolution of the 3-D random distributions of oxygen vacancies and temperature inside the device during the forming and SET/RESET processes. The modeling confirms that a tunneling gap between the top electrode (TE) which serves as an oxygen ion reservoir and the tip of the conductive filament may form at a high resistive state, which is assumed in the frequently used compact models of RRAM devices. The modeling shows that the sites with high temperature are concentrated in a cylindrical region at the center of the RRAM device, running from the inert bottom electrode to the TE. The modeling can reproduce both typical current-voltage characteristics of metal–oxide RRAM devices in forming and RESET and SET operations under a dc voltage sweep and the modulation of RRAM conductance by SET/RESET pulses.

Modulation of the Conductive Filament Growth During the Electroforming Process in Metal Oxide-Based RRAM

Metal oxide-based Resistive Random-Access Memories (RRAM) exhibit multiple intermediate resistive states, due to the formation/rupture of a conductive filament (CF) inside a switching layer. While there have been extensive studies on the resistive switching behavior caused by a pre-existing CF, the initial stage of the formation and growth of a CF, i.e., the electroforming process has been less understood. In this work, we developed a physical model to fundamentally understand the formation and growth behavior of a CF in a metal oxide layer constrained by the top/bottom electrodes. It is revealed that the CF formation is assisted by the supply of oxygen vacancies at the anode and the oxygen vacancies transport in the oxide layer. The effects of the external conditions (the oxygen exchange rate j 0 and the applied voltage) and the physical properties of the oxides (the electrical σ and thermal conductivity k th ) on the CF morphology evolution, current density evolution, and local temperature and electrical potential distribution have been systematically explored. It is found that larger oxygen vacancy exchange rate and relatively small voltage sweep rate result in a more uniform CF. Furthermore, by choosing oxides with lower Lorenz number (ratio of k th and σ ), CF with more homogeneous morphology can be realized. This work provides a fundamental understanding of the kinetic behaviors of the CF formation and growth during the electroforming process, and could potentially guide the oxide and electrode materials selection, synthesis and working conditions for switching to achieve improved RRAM stability and functionality.