Search

Cation-based resistive switching (RS) devices, dominated by conductive filaments (CF) formation/dissolution, are widely considered for the ultrahigh density nonvolatile memory application. However, the current-retention dilemma that the CF stability deteriorates greatly with decreasing compliance current makes it hard to decrease operating current for memory application and increase driving current for selector application. By centralizing/decentralizing the CF distribution, this current-retention dilemma of cation-based RS devices is broken for the first time. Utilizing the graphene impermeability, the cation injecting path to the RS layer can be well modulated by structure-defective graphene, leading to control of the CF quantity and size. By graphene defect engineering, a low operating current (≈1 µA) memory and a high driving current (≈1 mA) selector are successfully realized in the same material system. Based on systematically materials analysis, the diameter of CF, modulated by graphene defect size, is the major factor for CF stability. Breakthrough in addressing the current-retention dilemma will instruct the future implementation of high-density 3D integration of RS memory immune to crosstalk issues.

Resistive Random Access Memory (RRAM) can be regarded as a promising candidate on the manipulation of both electrical and magnetic properties. There is a widespread concern about the electrical manipulation of magnetic properties in RRAM devices. In our work, Co/HfO 2 /Pt RRAM device with magnetic conductive filament (CF) is designed and fabricated. Then the anisotropic magnetoresistance (AMR) effect on the magnetic CF is studied in order to find out the underlying nature of its transport. We measured the AMR by means of rotating the direction of the magnetic field and sweeping its magnitude, respectively. The results indicate that the direction of CF's easy-axis is perpendicular to the top electrode's plane. The maximum of magnetoresistance is found to appear when the angle between the direction of magnetic field and that of electric current in the CF is about 30°, and it varies slightly at different temperature.

Tail bits in the RESET process influence the distribution of resistive parameters, and will dramatically decrease the uniformity in high electric stimuli region. In this study, such phenomena can be explained as non-uniform distribution of defects in the insulating layer and high defects density in the bottleneck region of conductive filament (CF). Then departing from defects' distribution function, an improved statistical analysis is developed, exhibiting better feasibility than traditional Weibull method. Based on the results of statistical analysis, the predictable distribution of tail bits can be realized. Thus, targeted approaches at defect control are expected to implement on the improvement of RESET behavior in the future studies.

This paper presents a 0.13μm 64Mb HfOx ReRAM for embedded storage in IoT device. The configurable ramped voltage write and low read-disturb sensing techniques are proposed to address the reliability challenges in ReRAM. Experimental results show that, the ReRAM chip achieves more than 107 cycles' endurance and 10 years' retention at 75°C. In addition, its full function and superior random write performance are demonstrated on a developed evaluation board.

Conductive bridge random access memory (CBRAM) has been extensively studied as a next-generation non-volatile memory. The conductive filament (CF) shows rich physical effects such as conductance quantization and magnetic effect. But so far, the study of filaments is not very sufficient. In this work, Co/HfO2/Pt CBRAM device with magnetic CF was designed and fabricated. By electrical manipulation with a partial-RESET method, we controlled the size of ferromagnetic metal filament. The resistance-temperature characteristics of the ON-state after various partial-RESET behaviors have been studied. Using two kinds of magnetic measurement methods, we measured the anisotropic magnetoresistance (AMR) of the CF at different temperatures to reflect the magnetic structure characteristics. By rotating the direction of the magnetic field and by sweeping the magnitude, we obtained the spatial direction as well as the easy-axis of the CF. The results indicate that the easy-axis of the CF is along the direction perpendicular to the top electrode plane. The maximum magnetoresistance was found to appear when the angle between the direction of magnetic field and that of the electric current in the CF is about 30°, and this angle varies slightly with temperature, indicating that the current is tilted.

In this study, we present a bilayer resistive switching memory device with Pt/Ta2O5/HfO2-x /Hf structure, which shows sub-1 μA ultralow operating current, median switching voltage, adequate ON/OFF ratio, and simultaneously containing excellent self-rectifying characteristics. The control sample with single HfO2-x structure shows bidirectional memory switching properties with symmetrical I–V curve in low resistance state. After introducing a 28-nm-thick Ta2O5 layer on HfO2-x layer, self-rectifying phenomena appeared, with a maximum self-rectifying ratio (RR) of ~4 × 103 observed at ±0.5 V. Apart from being a series resistance for the cell, the Ta2O5 rectifying layer also served as an oxygen reservoir which remains intact during the whole switching cycle.

We investigated the endurance characteristics of a Cu-doped HfO2 selector device in one transistor-one selector (1T1S) structure, which is fully compatible with standard BEOL process. The device exhibits high endurance of 1010 under 10 μΑ compliance current. However, reduced endurance (105) was observed as increasing the compliance up to 100 μΑ. Under the condition of high operation, intrinsic defect in the HfO2 layer was possibly generated and resulted in endurance degradation.

Neuron and synapse are the basic constitutional units of biological neural network which has high efficiency in information processing. On account of which, the bio-inspired neuromorphic computing has attracted considerable attention. In this letter, we demonstrate an integration and fire artificial neuron (neuristor) based on threshold switching memristor. The neuristor captures 4 crucial features for action-potential-based computing: all-or-nothing spiking of an action potential, threshold-driven spiking, a refractory period, and the strength-modulated frequency response. A system-level simulation of the neuristor as post-synaptic neuron show that the validity to be used for a pattern recognition application. The experimental and simulation result exhibit that the neurisor can implement the basic functions of spiking neurons and has a great potential for neuromorphic computing.

In resistive random access memories, modeling conductive filament growing dynamics is important to understand the switching mechanism and variability. In this paper, a universal Monte Carlo simulator is developed based on a cell switching model and a tunneling-based transport model. Driven by external electric field, the growing process of the nanoscale filament occurring in the gap region is actually dominated by cells’ conductive/insulating switching, modeled through a phenomenological physics-based probability function. The electric transport through the partially formed CF is considered as current tunneling in the framework of the Quantum Point Contact model, and the potential barrier is modulated during cells’ evolution. To demonstrate the validity and universality of our simulator, various operation schemes are simulated, with the simulated I − V characteristics well explaining experimental observations. Furthermore, the statistical analyses of simulation results in terms of Weibull distribution and conductance evolution also nicely track previous experimental results. Representing a simulation scale that links atomic-scale simulations to compact modeling, our simulator has the advantage of being much faster comparing with other atomic-scale models. Meanwhile, our simulator shows good universality since it can be applied to various operation signals, and also to different electrodes and dielectric layers dominated by different switching mechanisms.

A physical model is developed for electroforming process in valence change resistive random access memory. In the developed model, electric field- and temperature-dependent vacancy generation, drift/diffusion in the forming process are considered. Based on the proposed model, the intrinsic nature of pulse amplitude dependence of forming time is attributed to combined effects of vacancy generation and migration. The evolution of microscopic vacancy concentration during forming operation is calculated and the effect of vacancy migration on forming process is quantitatively evaluated. The simulated pulse amplitude dependence of forming time agrees well with the experimental data.

Volatile threshold switching (TS) and non-volatile memory switching (MS) are two typical resistive switching (RS) phenomena in oxides, which could form the basis for memory, analog circuits, and neuromorphic applications. Interestingly, TS and MS can be coexistent and converted in a single device under the suitable external excitation. However, the origin of the transition from TS to MS is still unclear due to the lack of direct experimental evidence. Here, conversion between TS and MS induced by conductive filament (CF) morphology in Ag/SiO2/Pt device is directly observed using scanning electron microscopy and high-resolution transmission electron microscopy. The MS mechanism is related to the formation and dissolution of CF consisting of continuous Ag nanocrystals. The TS originates from discontinuous CF with isolated Ag nanocrystals. The results of current–voltage fitting and Kelvin probe force microscopy further indicate that the TS mechanism is related to the modulation of the tunneling barrier between Ag nanocrystals in CF. This work provides clearly experimental evidence to deepen understanding of the mechanism for RS in oxide-electrolyte-based resistive switching memory, contributing to better control of the two RS behaviors to establish high-performance emerging devices.

Reversible chemical and structural changes induced by ionic motion and reaction in response to electrical stimuli leads to resistive switching effects in metal-insulator-metal structures. Filamentary switching based on the formation and rupture of nanoscale conductive filament has been applied in non-volatile memory and volatile selector devices with low power consumption and fast switching speeds. Before the mass production of resistive switching devices, great efforts are still required to enable stable and reliable switching performances. The conductive filament, a bridge of microscopic metal-insulator-metal structure and macroscopic resistance states, plays an irreplaceable part in resistive switching behavior, as unreliable performance often originates from unstable filament behavior. In this Review, departing from the filamentary switching mechanism and the existing issues, recent advances of the switching performance improvement through the conductive filament modulation are discussed, in the sequence of material modulation, device structure design and switching operation scheme optimization. In particular, two-dimensional (2D) nanomaterials with excellent properties including and beyond graphene, are discussed with emphasis on performance improvement by their active roles as the switching layer, insertion layer, thin electrode, patterned electrode, and edge electrode, etc.

We report a novel TaOx-based selector with trapezoidal band structure, formed by rapid thermal annealing in O 2 plasma. Salient features were successfully achieved, such as high current density (1MA/cm2), high selectivity (5×104), low off-state current (∼10 pA), robust endurance (>1010), self-compliance and excellent uniformity. The device is composed of fully CMOS-compatible materials and has no thermal budget compatibility concerns. Furthermore, the selector was fabricated in 1kb crossbar array and the integrated 1S1R device shows high nonlinearity in low resistance state (LRS), which is quite effective to solve the sneaking current issue. The demonstrated high performance selector device here shows high potential on manufacturing large scale crossbar array.

Multilevel per cell (MLC), achieved by controlling the compliance current during SET operation, is a common approach to realize high-density storage in resistive random access memory (RRAM). In this letter, we investigated the failure mechanism of the MLC storage in one transistor and one resistor structure. By commonly modulating the amplitudes of gate bias to achieve the MLC, we found some unexpected failed SET operations, which caused the shrinkage of the MLC margin. In situ monitoring of the dynamic voltage drops on both transistor and memory cell revealed that there was an abnormal rise of source potential of the transistor, resulting in the increase of threshold voltage of the access transistor. If the applied gate bias was below the increased threshold voltage, the transistor would not program the RRAM cell successfully. Finally, possible improvement approaches to solve this problem are suggested.

Data retention is one crucial reliability aspect of resistive random access memory (RRAM). The retention failure mechanism of the low-resistance state (LRS) for conductive bridge RAM is generally originated from the lateral diffusion of metal ions/atoms from the filament to its surrounding medium. In this letter, we proposed an effective method to improve the LRS retention by controlling the formation of the single filament. For a certain LRS, the effective surface area for metal ions/atoms diffusion in single filament is less than that of multi-filament. Thus, better LRS retention characteristics can be achieved by reducing the metal species diffusion. The validity of this method is verified by the superior retention characteristics of the LRS programmed by current mode, in comparison with voltage programming mode. The former tends to generate a single filament, while the later grows multi-filament. This letter provides a possible way to enhance the retention characteristics of RRAM.

In this summary, the resistive random access memory (RRAM) with the structure of Pt/HfO 2 /Ti is investigated for applications in radiation circumstance. The heavy ion 86Kr26+ of HIRFL (The Heavy Ion Research Facility in Lanzhou) is used as the radiation source. The energy of 86Kr26+ is 25 MeV/u, the LET is 37.6 MeV/(cm2/mg), and the fluence of 5e11 is achieved after 2 hours radiation. Basic performance of the Pt/HfO 2 /Ti is compared before and after radiation. An obvious decrease is shown in the original resistance value after radiation, and the forming process is no longer needed. The HRS, LRS, transition voltage and endurance are still stable after radiation.

In this paper, the characteristics of the set time (t set ) correlated with the initial off-state resistance (R off ) were studied using a statistical method based on a Ti/ZrO 2 /Pt RRAM device. The data were collected by the width-adjusting pulse operation (WAPO) method. The Weibull distribution is used to analyze t set variation. Both the Weibull slope (β t ) and scale factor (t set6 3%) of t Set distributions increase logarithmically with R off . An analytical cell-based model was developed to explain the experimental statistics. Our result provides an inspiration on the switching uniformity and optimization of the tradeoff between the set speed-disturb dilemma.

Negative-SET behavior is observed in various cation-based memories, which degrades the device reliability. Transmission electron microscopy results demonstrate the behavior is caused by the overgrowth of the conductive filament (CF) into the Pt electrode. The CF overgrowth phenomenon is suppressed and the negative-SET behavior is eliminated by inserting an impermeable graphene layer. The graphene-based devices show high reliability and satisfying performance.

In this paper, we present an statistical characterization of set/reset parameters for the Ti/HfO 2 /Pt Resistive Random Access Memory (RRAM) cells. Devices are tested under a variety of current compliance (CC) for a certain cycles. Weibull distribution has been employed to explore the variability of switching parameters which is caused by the stochastic nature of the filament. Cell-based set and reset statistics models can be used to reveal the experimental statistical properties.

An ideal selector device should have ultra-low leakage current, high reliability and fast recovery speed. In this work, we present a Cu BEOL compatible threshold switching (TS) selector being capable of maintaining ultra-low leakage for hours and recovering low leakage after write operation rapidly. High reliability and uniformity are achieved in both read operation and write operation.

In low cost vertical resistive switching memory (VRRAM), the inter-layer leakage becomes a serious problem, primarily resulting from the ultimate scaling in the vertical dimension. In this work, for the first time, we present a novel approach of fabricating 3D VRRAM using self-aligned self-selective RRAM to effectively address such challenge. By successfully suppressing the inter-layer leakage, the scaling limit of VRRAM could be extended beyond 5 nm. Other benefits, such as high nonlinearity (>103), low power consumption (sub-μA), robust endurance and excellent disturbance immunity, were also demonstrated.

The resistive switching (RS) process of resistive random access memory (RRAM) is dynamically correlated with the evolution process of conductive path or conductive filament (CF) during its breakdown (rupture) and recovery (reformation). In this study, a statistical evaluation method is developed to analyze the filament structure evolution process in the reset operation of Cu/HfO2/Pt RRAM device. This method is based on a specific functional relationship between the Weibull slopes of reset parameters’ distributions and the CF resistance (R on). The CF of the Cu/HfO2/Pt device is demonstrated to be ruptured abruptly, and the CF structure of the device has completely degraded in the reset point. Since no intermediate states are generated in the abrupt reset process, it is quite favorable for the reliable and stable one-bit operation in RRAM device. Finally, on the basis of the cell-based analytical thermal dissolution model, a Monte Carlo (MC) simulation is implemented to further verify the experimental results. This work provides inspiration for RRAM reliability and performance design to put RRAM into practical application.

Developing high performance self-selective cell (SSC) is one of the most critical issues of the integration of 3D vertical RRAM (V-RRAM). In this work, a four-layer V-RRAM array, with high performance HfO2/mixed ionic and electronic conductor (MIEC) bilayer SSC, was demonstrated for the first time. Several salient features were achieved, including ultra-low half-select leakage ( 103), low operation current (nA level), self-compliance, high endurance (>107), and robust read/write disturbance immunity.

Resistive switching memory is a very promising technology for emerging nonvolatile memory applications. Generally, the switching behavior is triggered by a sweep or pulse voltage. In this paper, a charged capacitor is proposed to be used as the external electrical source to program a resistive switching element. From theoretical analyses based on a set behavior model, the capacitor approach can program the device efficiently. Compared with sweep or pulse mode, capacitor driving method can greatly reduce over-programming after the set event, and thus improve the resistance uniformity. The experimental results performed on Cu/ZrO2:Cu/Pt device support this conclusion quite well. The proposed methodology has great value for achieving a reliable resistive switching, which is important for high density or embedded application such as memory, FPGA, DSP, and even neuromorphic systems.

Resistive switching mechanism of zirconium oxide-based resistive random access memory (RRAM) devices composed of Cu-doped ZrO2 film sandwiched between an oxidizable electrode and an inert electrode was investigated. The Ag/ZrO2:Cu/Pt RRAM devices with crosspoint structure fabricated by e-beam evaporation and e-beam lithography show reproducible bipolar resistive switching. The linear I–V relationship of low resistance state (LRS) and the dependence of LRS resistance (RON) and reset current (Ireset) on the set current compliance (Icomp) indicate that the observed resistive switching characteristics of the Ag/ZrO2:Cu/Pt device should be ascribed to the formation and annihilation of localized conductive filaments (CFs). The physical origin of CF was further analyzed by transmission electron microscopy (TEM) and energy dispersive X-ray spectroscopy (EDS). CFs were directly observed by cross-sectional TEM. According to EDS and elemental mapping analysis, the main chemical composition of CF is determined by Ag atoms, coming from the Ag top electrode. On the basis of these experiments, we propose that the set and reset process of the device stem from the electrochemical reactions in the zirconium oxide under different external electrical stimuli.

The CuxO films grown by plasma oxidation are composed of an insulating CuO layer and a conductive gradient CuxO layer. We found that the surface CuO layer influenced the switching behaviors greatly. Giant improvement of reliable endurance was achieved after annealing the device in the N2 atmosphere, resulting from the transition of CuO to Cu2O. The possible mechanism for this improvement is attributed to the alleviation of over-programming during forming process. The result shows that for resistance switching Cu2O is much more preferred than CuO. After further reducing the thickness of Cu2O layer, the forming voltage can be totally eliminated.

Resistive memory (ReRAM) based on a solid-electrolyte insulator is a promising nanoscale device and has great potentials in nonvolatile memory, analog circuits, and neuromorphic applications. The underlying resistive switching (RS) mechanism of ReRAM is suggested to be the formation and rupture of nanoscale conductive filament (CF) inside the solid-electrolyte layer. However, the random nature of the nucleation and growth of the CF makes their formation difficult to control, which is a major obstacle for ReRAM performance improvement. Here, we report a novel approach to resolve this challenge by adopting a metal nanocrystal (NC) covered bottom electrode (BE) to replace the conventional ReRAM BE. As a demonstration vehicle, a Ag/ZrO(2)/Cu NC/Pt structure is prepared and the Cu NC covered Pt BE can control CF nucleation and growth to provide superior uniformity of RS properties. The controllable growth of nanoscale CF bridges between Cu NC and Ag top electrode has been vividly observed by transmission electron microscopy (TEM). On the basis of energy-dispersive X-ray spectroscopy (EDS) and elemental mapping analyses, we further confirm that the chemical contents of the CF are mainly Ag atoms. These testing/metrology results are consistent with the simulation results of electric-field distribution, showing that the electric field will enhance and concentrate on the NC sites and control location and orientation of Ag CFs.

In this letter, MOS capacitors with bilayer dielectrics consisted of large bandgap Al2O3 and high-k HfO2 in different stacking order on n-type doped (100) β-Ga2O3 are investigated through C − V and J − V measurement. The C − V measurement results reveal that incoming HfO2 makes both bilayer structures attain an increasing dielectric constant, which means a better gate control ability in transistors comparing with single Al2O3. Additionally, the interface state density extracted by high-low frequency capacitance method suggests that Al2O3/(100)β-Ga2O3 with no treatment shows a comparative Dit value (8.0 × 1012 cm-2eV-1 to 2.2 × 1011 cm-2eV-1) with HfO2/(100)β-Ga2O3 (8.4 × 1012 cm-2eV-1 to 1.0 × 1011 cm-2eV-1) in energy range of 0.2 to 0.9 eV. Furthermore, HfO2/Al2O3/Ga2O3 showing a bigger forward breakdown voltage of 11.0 V than 7.8 V of Al2O3/HfO2/Ga2O3 demonstrates that inserted larger bandgap Al2O3 insulator between Ga2O3 semiconductor and high-k HfO2 dielectric can prevent gate leakage current more effectively. Accordingly, the HfO2/Al2O3/Ga2O3 can enhance gate control ability with an acceptable gate breakdown voltage and become an alternative choice in the design of the gate structure for Ga2O3 MOSFETs.

Pt/SiO2:metal nanoparticles/Pt sandwich structure is fabricated with the method of metal ion (Ag) implantation. The device exhibits multilevel storage with appropriate R off/R on ratio, good endurance and retention properties. Based on transmission electron microscopy and energy dispersive spectrometer analysis, we confirm that Pt nanoparticles are spurted into SiO2 film from Pt bottom electrode by Ag implantation; during electroforming, the local electric field can be enhanced by these Pt nanoparticles, meanwhile the Ag nanoparticles constantly migrate toward the Pt nanoparticles. The implantation induced nanoparticles act as trap sites in the resistive switching layer and play critical roles in the multilevel storage, which is evidenced by the negative differential resistance effect in the current-voltage (I-V) measurements.

β-Ga2O3 is an ultra-wide bandgap semiconductor with applications in power electronic devices. Revealing the transport characteristics of β-Ga2O3 devices at various temperatures is important for improving device performance and reliability. In this study, we fabricated a Pt/β-Ga2O3 Schottky barrier diode with good performance characteristics, such as a low ON-resistance, high forward current, and a large rectification ratio. Its temperature-dependent current–voltage and capacitance–voltage characteristics were measured at various temperatures. The characteristic diode parameters were derived using thermionic emission theory. The ideality factor n was found to decrease from 2.57 to 1.16 while the zero-bias barrier height Φb0 increased from 0.47 V to 1.00 V when the temperature was increased from 125 K to 350 K. This was explained by the Gaussian distribution of barrier height inhomogeneity. The mean barrier height Φ ¯ b0 = 1.27 V and zero-bias standard deviation σ0 = 0.13 V were obtained. A modified Richardson plot gave a Richardson constant A* of 36.02 A·cm−2·K−2, which is close to the theoretical value of 41.11 A·cm−2·K−2. The differences between the barrier heights determined using the capacitance–voltage and current–voltage curves were also in line with the Gaussian distribution of barrier height inhomogeneity.

RRAM exhibit highly stable characteristics under TID radiation. But different radiation effects are shown in the RRAM based different oxide. The radiation effects in RRAM based HfO2 and SiO2 are test and analyzed.

Memristive devices, with a fusion of memory and logic functions, provide good opportunities for configuring new concepts computing. However, progress towards paradigm evolution has been delayed due to the limited understanding of the underlying operating mechanism. The stochastic nature and fast growth of localized conductive filament bring difficulties to capture the detailed information on its growth kinetics. In this work, refined programming scheme with real-time current regulation was proposed to study the detailed information on the filament growth. By such, discrete tunneling and quantized conduction were observed. The filament was found to grow with a unit length, matching with the hopping conduction of Cu ions between interstitial sites of HfO2 lattice. The physical nature of the formed filament was characterized by high resolution transmission electron microscopy. Copper rich conical filament with decreasing concentration from center to edge was identified. Based on these results, a clear picture of filament growth from atomic view could be drawn to account for the resistance modulation of oxide electrolyte based electrochemical memristive elements.

High density 3-dimensional (3D) crossbar resistive random access memory (RRAM) is one of the major focus of the new age technologies. To compete with the ultra-high density NAND and NOR memories, understanding of reliability mechanisms and scaling potential of 3D RRAM crossbar array is needed. Thermal crosstalk is one of the most critical effects that should be considered in 3D crossbar array application. The Joule heat generated inside the RRAM device will determine the switching behavior itself and for dense memory arrays, the temperature surrounding may lead to a consequent resistance degradation of neighboring devices. In this work, thermal crosstalk effect and scaling potential under thermal effect in 3D RRAM crossbar array are systematically investigated. It is revealed that the reset process is dominated by transient thermal effect in 3D RRAM array. More importantly, thermal crosstalk phenomena could deteriorate device retention performance and even lead to data storage state failure from LRS (low resistance state) to HRS (high resistance state) of the disturbed RRAM cell. In addition, the resistance state degradation will be more serious with continuously scaling down the feature size. Possible methods for alleviating thermal crosstalk effect while further advancing the scaling potential are also provided and verified by numerical simulation.

In this work, we investigate the temperature, gate voltage, and electrode metal dependences of the contact resistance, in amorphous InGaZnO thin film transistors (a-IGZO TFTs). The measurement of the contact resistance is conducted by transfer length method using Lakeshore cryogenic probe station providing vacuum and low temperature environment. It is found that the contact resistance would decrease with the increase of the gate voltage and temperature, and vary with the contact electrode of different work function. The intrinsic temperature dependent field effect mobility of the a-IGZO TFTs can also be obtained by getting rid of the influence the contact resistance. Using transmission line model, the experimental data can be well described. The results provide a more intrinsic physics in a-IGZO TFTs.

The variation of switching parameters in RRAM is impacted by the resistance states through a statistical analysis. We proposed some improved program/erase (P/E) methods to precisely control the fluctuation of the resistance states. In DC P/E operation, combing current-sweep-induced gradual SET and voltage-sweep-induced gradual RESET, uniform resistance states can be achieved. In pulse P/E operation, width/height-adjusting method was proposed to control the resistance value after switching. After using this new method, the variance of the R ON and R OFF has decreased by 40% and 81%, respectively, and the resistive switching endurance has increased from 103 cycles to more than 106 cycles.

We proposed a new statistical evaluation method to justify the microstructure evolution process of conductive filament (CF) in the reset operation of resistive switching memory (RRAM) according to the dependence of the Weibull slope of reset parameters on the CF size. We demonstrated that in Cu/HfO 2 /Pt device the CF can be controlled to be ruptured abruptly, which is more advantageous to the reliable operation of RRAM since the intermediate states are avoided. A Monte Carlo (MC) simulator based on the thermal dissolution model is conducted to further prove our experimental results.

Conductive bridge random access memory is considered to be a promising candidate or complementary component to the traditional charge based memory. In-depth understanding of the failure behavior is essential for performance optimization. In this work, the failure mechanisms of endurance and retention of HfO 2 -based cell were studied in a 1 kb array. The degradation of high resistance state was found to exist in the majority cases of endurance failure (stuck at low resistance state). Based on high resolution transmission electron microscopy analysis, the copper accumulation in the filament was found to be the dominant reason behind the endurance failure. The mechanism of retention failure was found related with the lateral diffusion of copper species from the filament to its surrounding (or to the tunneling gap from the filament tip to counter electrode). The retention of high resistance state was degraded with the switching cycle, whereas that of low resistance state was improved. This work provides a basic understanding on the failure mechanism of electrochemical resistive switching memory.