Conductive Bridge Random Access Memory Research Articles

Investigation of deposition technique and thickness effect of HfO2 film in bilayer InWZnO-based conductive bridge random access memory

In recent research on memristor of brain-mimicking integrated circuit (IC) architecture, the electrical performance of electrochemical metallization memory can be significantly improved by applying a bilayer switching structure. In this work, a novel amorphous semiconductor InWZnO (IWZO) is used as the main switching layer, and a bilayer structure with HfO2 is fabricated by different deposition technique. By surface and chemical composition analysis, the effect of deposition process and thickness of HfO2 have been fully explored. The bilayer memory inserted with 3 nm HfO2 formed by atomic layer deposition process exhibits excellent electrical properties, such as high endurance cycle (more than 2 × 103), better retention time (up to 2 × 104 s, 85 °C), relatively uniform resistance state distribution and low compliance current operation. Consequently, the low power electrochemical metallization memory with optimized HfO2 layer has great potential for next generation memory in pixel and three-dimensional brain-mimicking IC technology.

A 10-nm-thick silicon oxide based high switching speed conductive bridging random access memory with ultra-low operation voltage and ultra-low LRS resistance

In this paper, a silicon oxide based conductive bridging random access memory (CBRAM) with an ultra-low operation voltage, a high switching speed, and an ultra-low resistance at low resistance state (LRS) is reported. The CBRAM has a sandwich structure with platinum and copper as electrode layers and an ultra-thin 10-nm-thick silicon oxide film as an insulating switching layer. The CBRAMs are fabricated with CMOS compatible materials and processes. DC I–V sweep characterizations show an ultra-low SET/RESET voltage of 0.35 V/−0.05 V, and the RESET voltage is the lowest among all ultra-low voltage CBRAMs. The CBRAM is capable of withstanding endurance tests with over 106 pulses of +0.4 V/−0.1 V with 1 μs pulse width, with the resistance at LRS maintaining at an ultra-low value of only 20 Ω, which is the lowest among all CBRAMs to date, and it is reduced by at least 2.95 times compared with prior studies. Meanwhile, the switching ratio between high resistance state and LRS is more than 1.49 × 104. Moreover, the switching time characterization of the CBRAM demonstrates an ultra-short SET/RESET time of 7/9 ns. The CBRAM has potential applications in high-speed, ultra-low voltage, and ultra-low power electronics.

CBRAM technology: transition from a memory cell to a programmable and non-volatile impedance for new radiofrequency applications

Electrical resistance control programming of conductive bridging random access memory (CBRAM) radio frequency (RF) switches could benefit the development of electronically controlled non-volatile RF attenuators and other reconfigurable devices. The object of this study is to adapt a conventional CBRAM based memory cell to be used as an RF switch, and to demonstrate the feasibility of programming non-volatile RF CBRAM switches to achieve specific target resistances within a range of continuous values. The memory-RF technologic transition implies a drastic increase of the geometry in order to handle a much higher power, a decrease of the transition capacitance in order to operate at much higher frequencies, and a decrease of the LRS to a few ohms, which is critical for RF applications. These studies are initially performed on an in-house made RF CBRAM cell array at DC frequency, and then extended successfully to a co-planar waveguide (CPW) based shunt mode RF switch with an integrated CBRAM cell. Reliability of the proposed technique is validated through detailed analysis of factors like repeatability of the process, time stability of programmed states, and statistics of time taken to converge to a desired resistance value for an arbitrary RF CBRAM switch.

Density functional simulations of a conductive bridging random access memory cell: Ag filament formation in amorphousGeS2

Density functional/molecular dynamics simulations have been performed to shed light on the drift of Ag atoms in an amorphous $\mathrm{Ge}{\mathrm{S}}_{2}$ solid-state electrolyte between Ag and Pt electrodes in the presence of a finite electric field. The system models a conductive bridging random access memory device, where the electric field induces the formation of conductive filaments across the chalcogenide. Simulations of a 1019-atom structure under an external electrostatic potential of 0.20 eV/\AA{} at 480 and 680 K show significant atomic diffusion within 500 ps. Ag migration and the formation of percolating filaments occur in both cases. Three simulations for a smaller model (472 atoms) confirm the formation of percolating Ag strings. Significantly reduced mobility of Ag cations at 380 K means that Ag migration to the Pt electrode did not occur within 1 ns. The electronic structure analysis of selected snapshots shows that dissolved Ag atoms become markedly cationic, which changes when Ag clusters form at the Pt electrode. The electrolyte does not conduct, despite percolating single-atom Ag wire segments. Sulfur becomes anionic during the migration as a result of Ag-S bonding, and the effect is most pronounced near the active electrode. The formation of conductive filaments requires a percolating network of Ag clusters to grow from the Pt interface, and the weakest link of this network is at the Ag electrode.

Formation of a ternary oxide barrier layer and its role in switching characteristic of ZnO-based conductive bridge random access memory devices

The insertion of a metal layer between an active electrode and a switching layer leads to the formation of a ternary oxide at the interface. The properties of this self-formed oxide are found to be dependent on the Gibbs free energy of oxide formation of the metal (ΔGf°). We investigated the role of various ternary oxides in the switching behavior of conductive bridge random access memory (CBRAM) devices. The ternary oxide acts as a barrier layer that can limit the mobility of metal cations in the cell, promoting stable switching. However, too low (higher negative value) ΔGf° leads to severe trade-offs; the devices require high operation current and voltages to exhibit switching behavior and low memory window (on/off) ratio. We propose that choosing a metal layer having appropriate ΔGf° is crucial in achieving reliable CBRAM devices.

Power efficient transistors with low subthreshold swing using abrupt switching devices

With the rapid development of transparent integrated circuits, transistors with extremely low subthreshold swing (SS) is becoming a necessary requirement. Here, we fabricated three transparent device structures that show abrupt electrical switching and make their series connection to the source terminal of the conventional field effect transistors (FET) to lower the SS value. Firstly, we demonstrate an environment friendly, disposable, and transparent conductive bridge random access memory (CBRAM) device composed of a cellulose nanocrystals active layer. Our CBRAM consists of a silver (Ag) electrochemically active top electrode and a cellulose nanocrystals-based switching layer on the FTO coated glass substrate. Devices with CBRAM can enable FET with an ultra-steep slope that is SS < 0.24 mV/dec and has a significantly high on/off ratio (~105) by switching the Ag metallic filament between on and off. Niobium oxide (NbO2) based threshold switching devices and zinc oxide (ZnO) based flexible Schottky diodes that show electrical breakdown were also stacked with FET, which gave SS values < 0.74 mV/dec and < 5.20 mV/dec, respectively. Comparatively, a nano-watt transistor called filament transistor (FET + CBRAM stack) can significantly improve the SS slope value with the lowest leakage current (~nA) and a record low turn on-voltage (~0.2 V) with a set power of only ~197 nW compared to the other series stack, which thereby attracts the attention of low power operations.

Reliable multilevel memristive neuromorphic devices based on amorphous matrix via quasi-1D filament confinement and buffer layer.

Conductive-bridging random access memory (CBRAM) has garnered attention as a building block of non–von Neumann architectures because of scalability and parallel processing on the crossbar array. To integrate CBRAM into the back-end-of-line (BEOL) process, amorphous switching materials have been investigated for practical usage. However, both the inherent randomness of filaments and disorders of amorphous material lead to poor reliability. In this study, a highly reliable nanoporous–defective bottom layer (NP–DBL) structure based on amorphous TiO2 is demonstrated (Ag/a-TiO2/a-TiOx/p-Si). The stoichiometries of DBL and the pore size can be manipulated to achieve the analog conductance updates and multilevel conductance by 300 states with 1.3% variation, and 10 levels, respectively. Compared with nonporous TiO2 CBRAM, endurance, retention, and uniformity can be improved by 106 pulses, 28 days at 85°C, and 6.7 times, respectively. These results suggest even amorphous-based systems, elaborately tuned structural variables, can help design more reliable CBRAMs.

Conductive Bridging-Based Memristive RF Switches on a Silicon Substrate

Three RF switches and a phase shifter are proposed using the conductive bridging random access memory (CBRAM) technique. The fabrication process is developed with the Nafion membrane on a high-resistivity silicon substrate. The equivalent circuit model is established for each memristive RF switch. The proposed switches and the phase shifter have been validated by the good agreement achieved among the simulated, measured, and modeled results. The experimental results show that the series planar switch has the ON-state insertion loss (IL) below 1.69 dB and the OFF-state isolation over 25 dB up to 10 GHz, with the actuation voltages of 6.1~7.8/-1.5~-2 V. The parallel metal-insulator-metal (MIM) switch has the small ON-state IL below 0.45 dB and the OFF-state isolation over 18.4 dB when biased with 0.8~3.7/-0.5~-0.9 V. The series MIM switch has the ON-state IL below 0.61 dB and the OFF-state isolation over 21.9 dB with 0.7~2.3/-0.5~-0.9 V. The series MIM switch shows a good comprehensive performance of low actuation voltage, good uniformity, low IL, moderate isolation, and independent control. The phase shifter is realized with four series MIM switches, which provides a phase shift of 45° ± 5° from 3.8 to 4.3 GHz with the in-band IL of 2.25 dB. Compared with the conventional RF switchable devices, the proposed devices have the advantages of nonvolatility and low power consumption. Without depending on the electron beam lithography process and high-temperature phase change, the proposed technique is promising for reconfigurable RF circuits with mass production and feasibility to various substrates.

Rational Design on Controllable Cation Injection with Improved Conductive-Bridge Random Access Memory by Glancing Angle Deposition Technology toward Neuromorphic Application.

A conductive-bridge random access memory (CBRAM) has been considered a promising candidate for the next-generation nonvolatile memory technology because of its excellent performance, for which the resistive switching behavior depends on the formation/dissolution of conducting filaments in an electrolyte layer originated by the cation injection from the active electrode with electrochemical reactions. Typically, the controllability of cations into the electrolyte layer is a main issue, leading to stable switching reliability. In this work, an architecture combining spike-shaped Ag electrodes created by Al2O3 nanopillar arrays as a physical diffusion barrier by glancing angle deposition technology was proposed to localize Ag cation injection for the formation of controllable filaments inside TiOx as the switching layer. Interestingly, the dimension of the Ag plugs defined by the topography of Al2O3 nanopillar arrays can control Ag cation injection to influence the dimensionality of conductive filaments. Compared to the typical planar-Ag/TiOx/Pt device, the spiked-Ag/Al2O3 nanopillar arrays/TiOx/Pt device shows improvement of endurance and voltage disturbance. With enhanced multilevel characteristics, the spiked active-metal-based CBRAM device can be expected to serve as an analogue synapse for neuromorphic applications.

Enhanced Linearity in CBRAM Synapse by Post Oxide Deposition Annealing for Neuromorphic Computing Applications

Artificial synapse with good linearity is a critical issue in conductive bridging random access memory (CBRAM) synaptic device to accomplish an efficient learning approach in the artificial intelligence system. In this work, we investigate a novel approach to enhance the linearity of CBRAM synapse. The linearity of a memristive synapse can be improved by the high-temperature vacuum annealing process. The annealed device not only improves reliability such as endurance characteristics but also improves the synaptic characteristics including multilevel characteristics with varying RESET stop voltages from −0.60 to −1.40 V. The nonlinearities of potentiation and depression are 1.36 and −2.18 with 500 conductance pulses, respectively, and the device exhibits 720 training epochs with a total number of 720 000 pulse numbers. The post oxide annealed CBRAM device with analog switching behavior and excellent reliability is potential to be an artificial synapse for neuromorphic computing. In addition, the experimental potentiation and depression data are employed to train HNN for image processing of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$30\times30$ </tex-math></inline-formula> pixels comprising 900 synapses. It is found that the HNN can be successfully trained to recognize the input image with a training accuracy of ~98% in 18 iterations.

Impact of O2 plasma treatment on novel amorphous oxide InWZnO on conductive bridge random access memory

The impact of O2 plasma treatment on novel amorphous oxide InWZnO (IWZO) as conductive bridge random access memory (CBRAM) was investigated. A high-quality film on the surface of IWZO can be obtained by using remote O2 plasma treatment. The uniformity of O2 plasma sample is better than control sample, and also the set and reset voltage are more uniform and smaller to suitable for memory operation. Moreover, the O2 plasma sample shows excellent memory performance, such as high switching endurance cycles (up to 3 × 103), long retention time for 104 s at 85 °C. These results show that the surface modification with O2 plasma on IWZO CBRAM device is a critical technique for next generation memory applications.

Oxygen Concentration Effect on Conductive Bridge Random Access Memory of InWZnO Thin Film.

In this study, the influence of oxygen concentration in InWZnO (IWZO), which was used as the switching layer of conductive bridge random access memory, (CBRAM) is investigated. With different oxygen flow during the sputtering process, the IWZO film can be fabricated with different oxygen concentrations and different oxygen vacancy distribution. In addition, the electrical characteristics of CBRAM device with different oxygen concentration are compared and further analyzed with an atomic force microscope and X-ray photoelectron spectrum. Furthermore, a stacking structure with different bilayer switching is also systematically discussed. Compared with an interchange stacking layer and other single layer memory, the CBRAM with specific stacking sequence of bilayer oxygen-poor/-rich IWZO (IWZOx/IWZOy, x < y) exhibits more stable distribution of a resistance state and also better endurance (more than 3 × 104 cycles). Meanwhile, the memory window of IWZOx/IWZOy can even be maintained over 104 s at 85 °C. Those improvements can be attributed to the oxygen vacancy distribution in switching layers, which may create a suitable environment for the conductive filament formation or rupture. Therefore, it is believed that the specific stacking bilayer IWZO CBRAM might further pave the way for emerging memory applications.

Multi‐Level Control of Conductive Filament Evolution and Enhanced Resistance Controllability of the Cu‐Cone Structure Embedded Conductive Bridge Random Access Memory

AbstractMemristor is considered as one of the key components of neuromorphic hardware aiming to overcome the limitations of classical von Neumann computers, but the continuous control of the device conduction remains a challenge. Random and stochastic dynamics of weak conducting filaments (CF) hinder the precise control of the resistance states, particularly in the conductive bridge random access memory (CBRAM). The problem of random and stochastic resistive switching of the Cu‐based CBRAM is largely mitigated by inserting the Cu‐cone as a cation source into the memory cell (Cu‐cone device). The electric field concentrated on the Cu‐cone induces stronger CF with a limited number near the tip region, making the device feature distinctive from the planar Cu CBRAM. The Cu‐cone device exhibits enhanced resistance controllability attributed to the well‐controlled CF, improved switching reliability improvement, and multiple intermediate resistance states. Such features are used to explore its usefulness as an artificial synapse, where fluent potentiation and depression can be achieved using voltage pulses. Further research is still required to emulate the artificial synapses fully, but the feasibility is presented in this work by exploiting the similarities between the Cu cations in the CBRAM device and the Ca2+ dynamics in a biological synapse.

Stochastic based compact model to predict highly variable electrical characteristics of organic CBRAM devices

A compact model is proposed using a stochastic approach to capture the resistive switching behavior of conductive-bridge random access memory (CBRAM) device featuring a solid polymer electrolyte consisting of Polyethylene Oxide (PEO). This model considers the statistical distribution of five electrical parameters used to describe the resistive switching observed in experimental data. A switching probability is defined to control the change of resistive state. This approach gives the model the stochastic behavior of current–voltage characteristics observed in this kind of devices. A good agreement between the simulation and the experimental curves is observed despite the unusual variability cycle-to-cycle for this type of ReRAM.

Impact of annealing environment on performance of InWZnO conductive bridge random access memory

In this paper, the characteristics of conductive bridge random access memory (CBRAM) with novel material tungsten doped indium zinc oxide (W doped InZnO, IWZO) acting as switching material in different annealing atmosphere have been investigated. The endurance cycles of annealed IWZO CBRAM devices are obviously increased to 104 cycles, while even longer retention time can be achieved at 85 °C, compared with the as-deposited IWZO CBRAM device. Besides, after annealing in nitrogen atmosphere, the uniformity of the resistance state has been greatly improved. By detailed X-ray photoelectron spectroscopy analysis, the better electrical properties can be attributed to the generated oxygen vacancies during nitrogen annealing. As a result, that notable improvement in IWZO CBRAM may show its potential for future integrated display circuit applications, just by utilizing simple annealing process.

Radiation Effects in Advanced and Emerging Nonvolatile Memories

Despite hitting major roadblocks in 2-D scaling, NAND flash continues to scale in the vertical direction and dominate the commercial nonvolatile memory market. However, several emerging nonvolatile technologies are under development by major commercial foundries or are already in small volume production, motivated by storage-class memory and embedded application drivers. These include spin-transfer torque magnetic random access memory (STT-MRAM), resistive random access memory (ReRAM), phase change random access memory (PCRAM), and conductive bridge random access memory (CBRAM). Emerging memories have improved resilience to radiation effects compared to flash, which is based on storing charge, and hence may offer an expanded selection from which radiation-tolerant system designers can choose from in the future. This review discusses the material and device physics, fabrication, operational principles, and commercial status of scaled 2-D flash, 3-D flash, and emerging memory technologies. Radiation effects relevant to each of these memories are described, including the physics of and errors caused by total ionizing dose, displacement damage, and single-event effects, with an eye toward the future role of emerging technologies in radiation environments.

Effect of Ag source layer thickness on the switching mechanism of TiN/Ag/SiN x /TiN conductive bridging random access memory observed at sub-µA current

Experiments are conducted to compare the resistive switching characteristics for several samples with different amounts of Ag deposition in TiN/Ag/SiN x /TiN conductive bridging random access memory (CBRAM). The compliance current in TiN/Ag/SiN x /TiN CBRAM determines the volatile/non-volatile memory operation as the current level controls the strength of the filament made of Ag. The transient measurement showed that the effective thickness of Ag source layer in the TiN/Ag/SiN x /TiN controls the supply of the Ag atoms into the insulating layer, affecting the strength of the conductive bridge. The mechanism for the switching characteristics and the volatility trend with the amount of Ag deposition is closely investigated using transmission electron microscopy and scanning electron microscopy images. The device shows the conductance potentiation by a voltage pulse train under 1 µA current level, and the higher potentiation rate is observed in the CBRAM with thick Ag source layer.

Impact of Active Electrode on the Synaptic Properties of SiO2-Based Forming-Free Conductive Bridge Memory

The influence of active top electrode (TE) material is thoroughly investigated in order to shed light into the manifestation of various switching modes and the concomitant repercussion on a variety of synaptic properties in conductive-bridge random access memory (CBRAM) SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -based devices. While both threshold and bipolar switching effects were captured when Ag was selected for TE, only the latter switching mode was confirmed for the case of Cu. The SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -based memristors exhibit not only conductance tuning properties at low bias conditions, but can emulate the synaptic potentiation and depression procedures under the same pulse polarity (Ag TE). Moreover, by modulating the time interval difference of the pulses train, we can obtain short-term (STP) to long-term potentiation (LTP) transition properties along Spike-Timing Dependent Plasticity (STDP) performance. Our results indicate thus, that by leveraging the rich internal dynamic nature of the resistive switching effect, several on-demand neuromorphic properties can be attained.

The Origin of CBRAM With High Linearity, On/Off Ratio, and State Number for Neuromorphic Computing

Conductive bridge random access memory (CBRAM) has been concentrated recently for its ultrasmall size, low power consumption, synaptic characteristics, and application in neuromorphic computing. However, the accuracy of the CBRAM array-based neural network is not high enough due to the low linearity, limited ON/OFF ratio, and the number of states. The original illustration and the optimization methods are still paucity. In this work, the origin of the characteristics of CBRAM has been revealed from the filament distribution of the devices, which inspires us to design an inserted graphene structure of CBRAM and preset seeds leading to high linearity (0.995), ON/OFF ratio (26.4), and the number of states (63). The Monte Carlo simulation results reveal that the CBRAM with more seeds can promote a larger number of potential advantage path (PAP) conducing better characteristics. Moreover, the PAP can be modulated by the number of preset seeds. Finally, a handwritten recognition neural network has been realized by using a 1T-1R array, and high recognition accuracy (92%) has been obtained, which shows that devices with higher PAP can eventually promote higher recognition accuracy.

Self-assembling crystalline peptide microrod for neuromorphic function implementation

Self-assembling crystalline peptide microrod for neuromorphic function implementation