Distribution Of Oxygen Vacancies Research Articles

Unraveling the Correlation between the Interface Structures and Tunable Magnetic Properties of La1-xSrxCoO3-δ/La1-xSrxMnO3-δ Bilayers Using Deep Learning Models.

Perovskite oxides are gaining significant attention for use in next-generation magnetic and ferroelectric devices due to their exceptional charge transport properties and the opportunity to tune the charge, spin, lattice, and orbital degrees of freedom. Interfaces between perovskite oxides, exemplified by La1-xSrxCoO3-δ/La1-xSrxMnO3-δ (LSCO/LSMO) bilayers, exhibit unconventional magnetic exchange switching behavior, offering a pathway for innovative designs in perovskite oxide-based devices. However, the precise atomic-level stoichiometric compositions and chemophysical properties of these interfaces remain elusive, hindering the establishment of surrogate design principles. We leverage first-principles simulations, evolutionary algorithms, and neural network searches with on-the-fly uncertainty quantification to design deep learning model ensembles to investigate over 50,000 LSCO/LSMO bilayer structures as a function of oxygen deficiency (δ) and strontium concentration (x). Structural analysis of the low-energy interface structures reveals that preferential segregation of oxygen vacancies toward the interfacial La0.7Sr0.3CoO3-δ layers causes distortion of the CoOx polyhedra and the emergence of magnetically active Co2+ ions. At the same time, an increase in the Sr concentration and a decrease in oxygen vacancies in the La0.7Sr0.3MnO3-δ layers tend to retain MnO6 octahedra and promote the formation of Mn4+ ions. Electronic structure analysis reveals that the nonuniform distributions of Sr ions and oxygen vacancies on both sides of the interface can alter the local magnetization at the interface, showing a transition from ferromagnetic (FM) to local antiferromagnetic (AFM) or ferrimagnetic regions. Therefore, the exotic properties of La1-xSrxCoO3-δ/La1-xSrxMnO3-δ are strongly coupled to the presence of hard/soft magnetic layers, as well as the FM to AFM transition at the interface, and can be tuned by changing the Sr concentration and oxygen partial pressure during growth. These insights provide valuable guidance for the precise design of perovskite oxide multilayers, enabling tailoring of their functional properties to meet specific requirements for various device applications.

Machine learning-based prediction of polaron-vacancy patterns on the TiO2(110) surface

The multifaceted physics of oxides is shaped by their composition and the presence of defects, which are often accompanied by the formation of polarons. The simultaneous presence of polarons and defects, and their complex interactions, pose challenges for first-principles simulations and experimental techniques. In this study, we leverage machine learning and a first-principles database to analyze the distribution of surface oxygen vacancies (VO) and induced small polarons on rutile TiO2(110), effectively disentangling the interactions between polarons and defects. By combining neural-network supervised learning and simulated annealing, we elucidate the inhomogeneous VO distribution observed in scanning probe microscopy (SPM). Our approach allows us to understand and predict defective surface patterns at enhanced length scales, identifying the specific role of individual types of defects. Specifically, surface-polaron-stabilizing VO-configurations are identified, which could have consequences for surface reactivity.

Synapse Neurotransmitter Channel-Inspired AlOx Memristor with "V" Type Oxygen Vacancy Distribution.

Memristor possesses great potential and advantages in neuromorphic computing, while consistency and power consumption issues have been hindering its commercialization. Low cost and accuracy are the advantages of human brain, so memristors can be used to construct brain-like synaptic devices to solve these problems. In this work, a five-layer AlOx device with a V-shaped oxygen distribution is used to simulate biological synapses. The device simulates synapse structurally. Further, under electrical stimulation, O2- moves to the Ti electrode and oxygen vacancy (Vo) moves to the Pt electrode, thus forming a conductive filament (CF), which simulates the Ca2+ flow and releases neurotransmitters to the postsynaptic membrane, thus realizing the transmission of information. By controlling applied voltage, the regulation of Ca2+ gated pathway is realized to control the Ca2+ internal flow and achieve different degrees of information transmission. Long-term Potentiation (LTP)/Long-term Depression (LTD), Spike Timing Dependent Plasticity (STDP), these basic synaptic performances can be simulated. The AlOx device realizes low power consumption (56.7 pJ/392 fJ), high switching speed (25ns/60ns), and by adjusting the window value, the nonlinearity is improved (0.133/0.084), a high recognition accuracy (98.18%) is obtained in neuromorphic simulation. It shows a great prospect in multi-value storage and neuromorphic computing.

Lattice strain engineered reactive oxygen species generation of NaNbO3 ferroelectric

The generation of reactive oxygen species (ROS) using ambient mechanical energy with piezoelectric materials is emerging as a promising method for producing renewable green oxidants. However, the low piezoelectricity and limited electron-hole pair separation efficiency significantly hinder its productivity. Here, non-stoichiometric Na1-xNbO3-δ perovskites with lattice strain and ferroelectric enhancement were successfully designed to boost the piezo-catalytic activity. Significantly, hydrogen peroxide (H2O2), superoxide radicals (•O2-), and hydroxyl radicals (•OH) were powerfully generated by the Na0.95NbO2.975 piezo-catalyst through redox reaction with rates of 362.4, 69.3, and 65.0 µmol∙g-1∙h-1, respectively, and the yield of H2O2 was approximately three times higher than that of the stoichiometric NaNbO3. The elevated piezoelectricity and polarization electric field effects induced by lattice strain could promote the generation of electron-hole pairs and drive them to separate facilely. The distribution of sodium vacancies and oxygen vacancies in the crystal lattice enhanced the diffusion of carriers for fast migration to the opposite surface of the catalyst. Furthermore, the presence of Na defects enhanced the adsorption of O2 on the Na1-xNbO3-δ surface and contributed to the catalytic generation of H2O2. This work provides a novel and effective methodology for the design of high-performance piezo-catalysts with lattice-strained engineering and broadens the strategy for improved H2O2 evolution efficiency.

Multi-level, forming and filament free, bulk switching trilayer RRAM for neuromorphic computing at the edge

CMOS-RRAM integration holds great promise for low energy and high throughput neuromorphic computing. However, most RRAM technologies relying on filamentary switching suffer from variations and noise, leading to computational accuracy loss, increased energy consumption, and overhead by expensive program and verify schemes. We developed a filament-free, bulk switching RRAM technology to address these challenges. We systematically engineered a trilayer metal-oxide stack and investigated the switching characteristics of RRAM with varying thicknesses and oxygen vacancy distributions to achieve reliable bulk switching without any filament formation. We demonstrated bulk switching at megaohm regime with high current nonlinearity, up to 100 levels without compliance current. We developed a neuromorphic compute-in-memory platform and showcased edge computing by implementing a spiking neural network for an autonomous navigation/racing task. Our work addresses challenges posed by existing RRAM technologies and paves the way for neuromorphic computing at the edge under strict size, weight, and power constraints.

Unveiling the mechanism of alkali etched CuMn2O4-δ spinel catalysis enhance for acetone catalytic Oxidation: The roles of Metal-Oxygen bond and oxygen defect site

Spinel is a powerful non-precious metal oxide catalyst that can be used to catalyze the combustion of volatile organic pollutants, potentially replacing precious metal catalysts. To enhance the catalytic activity and stability of CuMn2O4, we constructed gradient distribution of oxygen vacancies (Vo) on the CuMn2O4-δ surface utilizing alkali treatment method. Alkali treatment dissolves some of the Cu atoms of the CuMn2O4 catalyst, leading to the transformation of part of the Mn-O-Cu structure into the Mn-O structure, which was results in the lengthening of the Mn-O bond and the production of cation defects. The Mn-O bond breaks, releasing a large amount of reactive oxygen and forming oxygen vacancies. Highly reactive lattice oxygen and cation defects is more likely to participate in oxidation reactions. In the catalytic combustion of acetone, the alkali-treated catalyst showed high activity and stability, and the oxidation rate of acetone at 170 °C was increased by 2.86 times. Particularly, the alkali treatment catalysts can convert acetone that has been adsorbed into acetic acid and formic acid using catalyst lattice oxygen in anoxic environment. In the presence of an oxygen stream, the acetic and formic acids underwent direct oxidation, resulting in the formation of water and carbon dioxide. This work promotes the development of oxygen vacancy engineering on the surface of transition metal oxides and provides a new approach to improve the activity and stability of CuMn2O4 in real environments.

Enhancing the piezoelectric properties and thermal stability of PMN-PMS-PSZT high-power piezoelectric ceramics through defect engineering

The development of high-power piezoelectric ceramics with superior piezoelectric properties and broad temperature usage ranges is vital for the thriving progress of electromechanical applications. However, achieving high piezoelectricity, high mechanical quality factor, and temperature stability often presents a trade-off. In this study, we present an advanced ceramic material with excellent high-power piezoelectric properties and superior temperature stability. The piezoelectric coefficient d33 reaches 348 pC N−1, the electromechanical coupling factor kp is 54%, the mechanical quality factor Qm is 1501, the dielectric loss tanδ is 0.35%, and the d33 variation is less than 10% within 25–210 °C in 0.05Pb(Mg1/3Nb2/3)O3-0.05Pb(Mn1/3Sb2/3)O3-0.9Pb0.95Sr0.05(Zr0.48Ti0.52)O3 ceramic. The excellent piezoelectricity is attributed to the enhancement of intrinsic ionic displacement and reversible domains wall motion because of symmetry-conforming short-range distribution of oxygen vacancy, while the high Qm is a result of increased domain size and oxygen vacancy. The outstanding temperature stability is related to the stable non-180°domains structure due to the oxygen vacancy pinning effect. These properties hold great potential for advanced applications, particularly for piezoelectric high-power applications requiring constant d33 over a broad temperature range.

Study on the Sodium-Doped Titania Interface-Type Memristor.

Memristors integrated into a crossbar-array architecture (CAA) are promising candidates for analog in-memory computing accelerators. However, the relatively low reliability of the memristor device and sneak current issues in CAA remain the main obstacles. Alkali ion-based interface-type memristors are promising solutions for implementing highly reliable memristor devices and neuromorphic hardware. This interface-type device benefits from self-rectifying and forming-free resistive switching (RS), and exhibits relatively low variation from device to device and cycle to cycle. In a previous report, we introduced an in situ grown Na/TiO2 memristor using atomic layer deposition (ALD) and proposed a RS mechanism from experimentally measured Schottky barrier modulation. Self-rectifying RS characteristics were observed by the asymmetric distribution of Na dopants and oxygen vacancies as the Ti metal used as the adhesion layer for the bottom electrode diffuses over the Pt electrode at 250 °C during the ALD process and is doped into the TiO2 layer. Here, we theoretically verify the modulation of the Schottky barrier at the TiO2/Pt electrode interface by Na ions. This study fabricated a Pt/Na/TiO2/Pt memristor device and confirmed its self-rectifying RS characteristics and stable retention characteristics for 24 h at 85 °C. Additionally, this device exhibited relative standard deviations of 27 and 7% in the high and low resistance states, respectively, in terms of cycle-to-cycle variation. To verify the RS mechanism, we conducted density functional theory simulations to analyze the impact of Na cations at interstitial sites on the Schottky barrier. Our findings can contribute to both fundamental understanding and the design of high-performance memristor devices for neuromorphic computing.

Novel Phase YBa2Cu3O5 derived from the High Tc Superconductor YBa2Cu3O7-x

The orthorhombic unit cell of high Tc oxide superconductor YBa2Cu3O7–x is indicated by two-dimensional CuO2 planes spanned in a-b directions. Partial removal of oxygen, down to x > 1, creates an insulating state and randomly distributed oxygen vacancies. An attempt has been made to increase the concentration of vacancies up to some critical value when their ordering could take place, while balance with copper cations will still maintain the original tetragonal structure. Removal of two oxygen atoms per unit cell results in the YBa2Cu3O5 compound and formation of linear Cu1+– O chains. The novel phase is an insulator, and doping in a low (50 –200 mbar) oxygen atmosphere at T < 800 K converts it to a conducting state.

Bi-directional threshold voltage shift of amorphous InGaZnO thin film transistors under alternating bias stress

Amorphous InGaZnO (a-IGZO) has attracted a lot of attention as a high-mobility channel material for thin film transistors (TFTs). However, the instability mechanism involving threshold voltage and subthreshold swing (SS) in a-IGZO TFTs still requires further investigation. In this study, we investigated the electrical instability of amorphous InGaZnO TFTs subjected to alternating positive and negative bias stresses. Based on the respective mechanisms under positive and negative bias stresses, including ionization and spatial movement of oxygen vacancies, bi-directional threshold voltage shifts were observed under alternating bias stress. The SS values vary with the bias stress polarity, reflecting the presence and distribution of oxygen vacancies. Our findings reveal a complementary mechanism based on oxygen vacancies, elucidating the behavior under complex bias stress schemes and extending our understanding of instability mechanisms beyond monotonous bias stress.

Low-frequency noise behaviors of quasi-two-dimensional electron systems based on complex oxide heterostructures

We report the low-frequency noise behaviors in quasi-two-dimensional (quasi-2D) electron systems based on complex oxide heterostructures. First, the surface 2D electron gas (2DEG) on SrTiO3 (STO) exhibits the 1/fα-type current power spectral density (PSD) with α∼1.39. The non-unity exponent α indicates the discrepancy between the depth distributions of electrons and oxygen vacancies in the STO substrate. Second, the amorphous LaAlO3/KTaO3 (LAO/KTO) interface, another quasi-2D electron system, shows Lorentzian components of PSD at a high-frequency region around 1 kHz, implying that the amorphous overlayer can provide additional shallow charge-trapping sites to the quasi-2D electrons in the crystalline KTO substrate. Lastly, ultrathin SrRuO3 (SRO) film grown on STO substrate exhibits the Lorentzian components of PSD at a low-frequency region around 200 Hz. The slight suppression of the fast charge trapping is attributed to the intrinsic band bending at the interface between film and substrate. These results will provide a guideline for understanding the defect-induced charge trapping and the relevant electron dynamics in the quasi-2D electron systems as well as the oxide-based electronic materials in general.

Reliability effects of lateral filament confinement by nano-scaling the oxide in memristive devices.

Write-variability and resistance instability are major reliability concerns impeding implementation of oxide-based memristive devices in neuromorphic systems. The root cause of the reliability issues is the stochastic nature of conductive filament formation and dissolution, whose impact is particularly critical in the high resistive state (HRS). Optimizing the filament stability requires mitigating diffusive processes within the oxide, but these are unaffected by conventional electrode scaling. Here we propose a device design that laterally confines the switching oxide volume and thus the filament to 10 nm, which yields reliability improvements in our measurements and simulations. We demonstrate a 50% decrease in HRS write-variability for an oxide nano-fin device in our full factorial analysis of modulated current-voltage sweeps. Furthermore, we use ionic noise measurements to quantify the HRS filament stability against diffusive processes. The laterally confined filaments exhibit a change in the signal-to-noise ratio distribution with a shift to higher values. Our complementing kinetic Monte Carlo simulation of oxygen vacancy (re-)distribution for confined filaments shows improved noise behavior and elucidates the underlying physical mechanisms. While lateral oxide volume scaling down to filament sizes is challenging, our efforts motivate further examination and awareness of filament confinement effects in regards to reliability.

Entropic Stabilization in Lithium Rich Transition Metal Oxide Cathodes

It is well known that lithium rich transition metal oxide cathodes (LRM) experience voltage fade and capacity loss over the course of cycling, and it is generally known that these are a result of the LRM materials’ general instability 1–7. However, the mechanisms behind these losses are still unknown, due in large part to the structural ambiguity and controversy associated with the LRM material structure. In this perspective we consider the role entropy plays in LRM materials’ structural stability and electrochemical properties. With this mind set we argue that due to their metastability LRM cathodes’ structure and properties are not synthesis agnostic, and therefore establishment of consistent synthetic procedures is of paramount importance. We also propose that LRM materials could be categorized as entropically stabilized oxides through exploring their configurational entropies, phase behaviors, and thermal responses. The entropic stabilization of LRM materials is also found to be consistent with the ordering behaviors of LRM cathode transitional metal layers, with this ordering presenting both compositional, thermal, and electrochemical dependencies. The impact of these behaviors on the redox and electrochemical structure and activation of LRM materials is also explored. Finally, we seek to put these findings into the context of the wider battery field. Fell, C. R. et al. Correlation Between Oxygen Vacancy, Microstrain, and Cation Distribution in Lithium-Excess Layered Oxides During the First Electrochemical Cycle. Chem. Mater. 1621–1629 (2013) doi:10.1021/cm4000119.Lu, Z., Beaulieu, L. Y., Donaberger, R. A., Thomas, C. L. & Dahn, J. R. Synthesis, Structure, and Electrochemical Behavior of Li [Ni x Li(1/2-2x/3) Mn(2/3-x/3)]O2. J. Electrochem. Soc. 149, A778–A791 (2002).Wu, Y. & Manthiram, A. High Capacity, Surface-Modified Layered Li†Li„1−x.../3Mn„2−x.../3Nix/3Cox/3‡O2 Cathodes with Low Irreversible Capacity Loss. Electrochem. Solid-State Lett. 9, A221–A224 (2006).Dupré, N., Cuisinier, M., Legall, E., War, D. & Guyomard, D. Contribution of the oxygen extracted from overlithiated layered oxides at high potential to the formation of the interphase. J. Power Sources 299, 231–240 (2015).Jarvis, K. A. et al. Formation and effect of orientation domains in layered oxide cathodes of lithium-ion batteries. Acta Mater. 108, 264–270 (2016).Jiang, M., Key, B., Meng, Y. S. & Grey, C. P. Electrochemical and Structural Study of the Layered, ‘Li-Excess’; Lithium-Ion Battery Electrode Material Li[Li 1/9 Ni 1/3 Mn 5/9 ]O 2. Chem. Mater. 21, 2733–2745 (2009).Hy, S. et al. Performance and design considerations for lithium excess layered oxide positive electrode materials for lithium ion batteries. Energy Environ. Sci. 9, 1931–1954 (2016).

Mechanical-electrical-chemical coupling study on the stabilization of a hafnia-based ferroelectric phase

The metastable polar orthorhombic phase is believed to be the origin of the ferroelectricity of hafnia-based films. The adjustment of stain, oxygen vacancies and dopant during film deposition and the wake-up electric cycling are common strategies to induce the ferroelectricity in hafnia. However, none of them could independently render the ferroelectric phase to be the most stable phase from the theoretical calculation results. The exact external conditions to stabilize orthorhombic phase still remain elusive. In this paper, we investigate the effects of the type, distribution, concentration, and charge state characteristics of oxygen vacancies and the uniaxial strain on the crystal’ energy, dielectric constant and spontaneous polarization (Ps); In addition, the impact of the applied electric field parallel to the Ps on the crystal’ energy is explored by first-principles calculations. It is challenging to independently stabilize the ferroelectric phase of hafnia-based films by a single component owing to the rather strict conditions. Surprisingly, the ferroelectricity can be easily obtained when simultaneously considering the effects of oxygen vacancies, uniaxial strain, and applied electric fields, suggesting the extremely important mechanical-electrical-chemical coupling effects. This work provides an explanation for the typical wake-up phenomenon in hafnia and a guidance for film applications.

Ambipolar surface conduction in oxygen sub-stoichiometric molybdenum oxide films

The surface electric conduction in amorphous and crystallized molybdenum oxide films was studied as a function of electronic structure by current–voltage and simultaneous spectroscopic ellipsometry measurements on structures of the kind Al/Molybdenum oxide (MoOx)/Al, at temperatures up to 400 °C and in ambient air. At room temperature, both amorphous and crystalline MoOx samples were found to be sub-stoichiometric in oxygen. The random distribution of oxygen vacancies and the imperfect atomic ordering induced the creation of an intermediate band (IB) located near the valence band and of individual electronic gap states. At temperatures below 300 °C, the conduction was found to exhibit ambipolar character in which electrons and holes participated, the former moving in the conduction band and the latter in the IB and though gap states. Above 300 °C, due to samples gradual oxidation and improvement of atomic ordering (samples crystallization), the density of states in the IB and the gap gradually decreased. The above in their turn resulted in the gradual suppression of the ambipolar character of the conduction, which at 400 °C was completely suppressed and became similar to that of ordinary n-type semiconductor. The above phenomena were found to be reversible, so as the semiconducting MoOx samples were returning to room temperature the ambipolarity of the conduction was gradually re-appearing giving rise to an unusual phenomenon of “metallic” temperature variation of electrical resistance when electrons were injected.

Electrocatalytic Properties of Oxygen-Deficient Perovskites Ca3Fe3-xMnxO8 (x = 1-2) for the Hydrogen Evolution Reaction.

We have demonstrated a systematic trend in the electrocatalytic activity for the hydrogen evolution reaction (HER) and its correlations with transition-metal type, structural order, and electrical conductivity. The materials studied in this work, Ca3FeMn2O8 (CaFe1/3Mn2/3O3-1/3), Ca3Fe1.5Mn1.5O8, and Ca3Fe2MnO8, belong to the family of oxygen-deficient perovskites and show a gradual increase in the ordering of oxygen vacancies. Ca3FeMn2O8 (CaFe1/3Mn2/3O3-1/3) contains randomly distributed oxygen vacancies, which begin to order in Ca3Fe1.5Mn1.5O8, and are fully ordered in Ca3Fe2MnO8. The gradual increase in the structural order is associated with a systematic enhancement of the electrocatalytic activity for HER in acidic conditions, Ca3FeMn2O8 < Ca3Fe1.5Mn1.5O8 < Ca3Fe2MnO8. While the improvement of the HER activity is also associated with an increase in the Fe content, we have shown that the type of structural order plays a more important role. We demonstrated this effect by control experiments on an analogous material where all Mn was substituted by Fe, leading to a different type of structural order and showing an inferior HER activity compared to the above three materials. Furthermore, electrical conductivity studies in a wide range of temperatures, 25-800 °C, indicate that the trend in the electrical conductivity is the same as that of the HER activity. These findings reveal several important structure-property relationships and highlight the importance of synergistic effects in enhancing the electrocatalytic properties.

Achieving High Performance of ZnSnO Thin-Film Transistor via Homojunction Strategy.

The zinc-tin-oxide (ZTO) thin-film transistor (TFT) is one of the most promising candidates for advanced display applications, though its popularity is limited by its performances. In this work, a heterojunction channel strategy was adopted to regulate the electron transport behaviors and the TFT performances by manipulating the concentration and the distribution of oxygen vacancies, and a reasonable physical model was proposed based on experimental and simulation results. It is difficult to mediate the contradiction between mobility and threshold voltage for the single channel. Via a heterojunction channel strategy, desirable TFT performances, with mobility of 12.5 cm2/Vs, threshold voltage of 1.2 V and Ion/Ioff of 3 × 109, are achieved when the oxygen-vacancy-enriched layer gets close to the gate insulator (GI). The enhanced performances can be mainly attributed to the formation of two-dimensional electron gas (2DEG), the insensitive potential barrier and the reasonable distribution of oxygen vacancy. On the contrary, when the oxygen-vacancy-enriched layer stays away from GI, all the main performances degenerate due to the vulnerable potential well. The findings may facilitate the development and application of heterojunction channels for improving the performances of electronic devices.

Superionic conduction of self-assembled heterostructural LSCrF-CeO2 electrolyte for solid oxide fuel cell at 375–550 °C

Insufficient ionic conductivity of electrolyte materials and high manufacturing costs create bottlenecks in the commercialization of low-temperature solid oxide fuel cells (LT-SOFCs). Developing the semiconductor-ionic electrolyte represents a new research strategy for high-performance LT-SOFCs because of their attractive ion conduction and associated low polarization loss. In this study, a self-assembly approach is demonstrated by incorporating ABO3 perovskite La0.7Sr0.3Cr0.5Fe0.5O3-δ (LSCrF) into the fluorite structure CeO2, for developing 1LSCrF-xCeO2 composite as semiconductor-ionic heterostructure for LT-SOFCs with superior electrochemical efficiency at the temperature below 550 °C and even down to 370 °C over unmodified and mechanically mixed samples. Self-assembly enables abundant nanoscale hetero-interfaces accompanied by widely distributed oxygen vacancies, which leads to the superior ionic conductivity of 0.018–0.112 S cm−1, lower activation energy of 0.54 eV and better peak power density of 136–735 mW cm−2 over control samples at 370–550 °C, and stable operation for 50 h of resultant LT-SOFCs under a galvanostatic mode. This work highlights the important role of self-assembly in designing heterostructural semiconductor based electrolytes for high-performance LT-SOFCs.

Improved Switching Uniformity of SCLC-RRAM Using the Vertically Formed Nanoscale 2DEG Electrode and Control of Oxygen Vacancy Distribution

Improved Switching Uniformity of SCLC-RRAM Using the Vertically Formed Nanoscale 2DEG Electrode and Control of Oxygen Vacancy Distribution

Definition of a Localized Conducting Path via Suppressed Charge Injection in Oxide Memristors for Stable Practical Hardware Neural Networks.

Oxide-based memristors have been demonstrated as suitable options for memory components in neuromorphic systems. In such devices, the resistive switching characteristics are caused by the formation of conductive filaments (CFs) comprising oxygen vacancies. Thus, the electrical performance is primarily governed by the CF structure. Despite various approaches for regulating the oxygen vacancy distributions in oxide memristors, controlling the CF structure without modifying the device configuration related to material compatibility is still a challenge. This study demonstrates an effective strategy for localizing CF distributions in memristors by suppressing charge injection during the formation of conducting paths. As the injected charge quantity is reduced in the electroforming process of the oxide memristor, the CF distributions become narrower, leading to more reproducible and stable resistive switching characteristics in the device. Based on these findings, a reliable hardware neural network comprising oxide memristors is constructed to recognize complex images. The developed memristor has been employed as a synaptic memory component in systems without degradation for a long time. This promising concept of oxide memristors acting as stable synaptic components holds great potential for developing practical neuromorphic systems and their expansion into artificial intelligent systems.