Ultra-low power Hf0.5Zr0.5O2 based ferroelectric tunnel junction synapses for hardware neural network applications

Enhanced polarization and endurance properties of ZrO2-based ferroelectric capacitor using HfO2interfacial layer

Recently discovered ferroelectricity in fluorite-structure ZrO2thin film has attracted increasing and intense interest due to its lower crystallization temperature and higher content in nature in comparison to hafnium oxide. Here, the effect of HfO2interfacial layer on the ferroelectric properties of ZrO2thin films is investigated systematically by designing four types of interfacial structures. It is revealed that the ferroelectric orthorhombic phase, remanent polarization, and endurance can be improved in ZrO2thin film by inserting both a top- and bottom-HfO2interfacial layer. A maximal ferroelectric remanent polarization (2Pr) of 53.4μC cm-2and an optimal endurance performance of 3 × 107field cycles under frequency of 100 kHz are achieved in Pt/HfO2/ZrO2/HfO2/Pt capacitors, with ferroelectric stacks being crystallized at 450 °C via post-deposition annealing method. X-ray photoelectron spectroscopy analysis confirms that the HfO2bottom-layer plays a very important role in the formation of a higher ratio o-phase, thus enhancing the ferroelectricity. These results suggest that designing appropriate interfaces would help achieve excellent ferroelectric properties in ZrO2films.

A Zero-Voltage-Writing Artificial Nervous System Based on Biosensor Integrated on Ferroelectric Tunnel Junction

The artificial nervous system proves the great potential for the emulation of complex neural signal transduction. However, a more bionic system design for bio-signal transduction still lags behind that of physical signals, and relies on additional external sources. Here, this work presents a zero-voltage-writing artificial nervous system (ZANS) that integrates a bio-source-sensing device (BSSD) for ion-based sensing and power generation with a hafnium-zirconium oxide-ferroelectric tunnel junction (HZO-FTJ) for the continuously adjustable resistance state. The BSSD can use ion bio-source as both perception and energy source, and then output voltage signals varied with the change of ion concentrations to the HZO-FTJ, which completes the zero-voltage-writing neuromorphic bio-signal modulation. In view of in/ex vivo biocompatibility, this work shows the precise muscle control of a rabbit leg by integrating the ZANS with a flexible nerve stimulation electrode. The independence on external source enhances the application potential of ZANS in robotics and prosthetics.

CMOS-compatible Hf0.5Zr0.5O2-based ferroelectric memory crosspoints fabricated with damascene process

We report the fabrication of Hf0.5Zr0.5O2(HZO) based ferroelectric memory crosspoints using a complementary metal-oxide-semiconductor-compatible damascene process. In this work, we compared 12 and 56µm2crosspoint devices with the 0.02 mm2round devices commonly used as a benchmark. For all devices, a 9 nm thick ferroelectric thin film was deposited by plasma-enhanced atomic layer deposition on planarized bottom electrodes. The wake-up appeared to be longer for the crosspoint memories compared to 0.02 mm2benchmark, while all the devices reached a 2Prvalue of ∼50µC cm-2after 105cycles with 3 V/10µs squared pulses. The crosspoints stand out for their superior endurance, which was increased by an order of magnitude. Nucleation limited switching experiments were performed, revealing a switching time <170 ns for our 12 and 56µm2devices, while it remained in theµs range for the larger round devices. The downscaled devices demonstrate notable advantages with a rise in endurance and switching speed.

Resistive switching memories with enhanced durability enabled by mixed-dimensional perfluoroarene perovskite heterostructures

Hybrid halide perovskites are attractive candidates for resistive switching memories in neuromorphic computing applications due to their mixed ionic-electronic conductivity. Moreover, their exceptional optoelectronic characteristics make them effective as semiconductors in photovoltaics, opening perspectives for self-powered memory elements. These devices, however, remain unexploited, which is related to the variability in their switching characteristics, weak endurance, and retention, which limit their performance and practical use. To address this challenge, we applied low-dimensional perovskite capping layers onto 3D mixed halide perovskites using two perfluoroarene organic cations, namely (perfluorobenzyl)ammonium and (perfluoro-1,4-phenylene)dimethylammonium iodide, forming Ruddlesden-Popper and Dion-Jacobson 2D perovskite phases, respectively. The corresponding mixed-dimensional perovskite heterostructures were used to fabricate resistive switching memories based on perovskite solar cell architectures, showing that the devices based on perfluoroarene heterostructures exhibited enhanced performance and stability in inert and ambient air atmosphere. This opens perspectives for multidimensional perovskite materials in durable self-powered memory elements in the future.

A comprehensive review of advanced trends: from artificial synapses to neuromorphic systems with consideration of non-ideal effects

A neuromorphic system is composed of hardware-based artificial neurons and synaptic devices, designed to improve the efficiency of neural computations inspired by energy-efficient and parallel operations of the biological nervous system. A synaptic device-based array can compute vector-matrix multiplication (VMM) with given input voltage signals, as a non-volatile memory device stores the weight information of the neural network in the form of conductance or capacitance. However, unlike software-based neural networks, the neuromorphic system unavoidably exhibits non-ideal characteristics that can have an adverse impact on overall system performance. In this study, the characteristics required for synaptic devices and their importance are discussed, depending on the targeted application. We categorize synaptic devices into two types: conductance-based and capacitance-based, and thoroughly explore the operations and characteristics of each device. The array structure according to the device structure and the VMM operation mechanism of each structure are analyzed, including recent advances in array-level implementation of synaptic devices. Furthermore, we reviewed studies to minimize the effect of hardware non-idealities, which degrades the performance of hardware neural networks. These studies introduce techniques in hardware and signal engineering, as well as software-hardware co-optimization, to address these non-idealities through compensation approaches.

Physics, Structures, and Applications of Fluorite-Structured Ferroelectric Tunnel Junctions

The interest in ferroelectric tunnel junctions (FTJ) has been revitalized by the discovery of ferroelectricity in fluorite-structured oxides such as HfO2 and ZrO2 . In terms of thickness scaling, CMOS compatibility, and 3D integration, these fluorite-structured FTJs provide a number of benefits over conventional perovskite-based FTJs. Here, recent developments involving all FTJ devices with fluorite structures are examined. The transport mechanism of fluorite-structured FTJs is explored and contrasted with perovskite-based FTJs and other 2-terminal resistive switching devices starting with the operation principle and essential parameters of the tunneling electroresistance effect. The applications of FTJs, such as neuromorphic devices, logic-in-memory, and physically unclonable function, are then discussed, along with several structural approaches to fluorite-structure FTJs. Finally, the materials and device integration difficulties related to fluorite-structure FTJ devices are reviewed. The purpose of this review is to outline the theories, physics, fabrication processes, applications, and current difficulties associated with fluorite-structure FTJs while also describing potential future possibilities for optimization.

Filament-free memristors for computing

Memristors have attracted increasing attention due to their tremendous potential to accelerate data-centric computing systems. The dynamic reconfiguration of memristive devices in response to external electrical stimuli can provide highly desirable novel functionalities for computing applications when compared with conventional complementary-metal-oxide-semiconductor (CMOS)-based devices. Those most intensively studied and extensively reviewed memristors in the literature so far have been filamentary type memristors, which typically exhibit a relatively large variability from device to device and from switching cycle to cycle. On the other hand, filament-free switching memristors have shown a better uniformity and attractive dynamical properties, which can enable a variety of new computing paradigms but have rarely been reviewed. In this article, a wide range of filament-free switching memristors and their corresponding computing applications are reviewed. Various junction structures, switching properties, and switching principles of filament-free memristors are surveyed and discussed. Furthermore, we introduce recent advances in different computing schemes and their demonstrations based on non-filamentary memristors. This Review aims to present valuable insights and guidelines regarding the key computational primitives and implementations enabled by these filament-free switching memristors.

Emerging memristive artificial neuron and synapse devices for the neuromorphic electronics era

Growth of data eases the way to access the world but requires increasing amounts of energy to store and process. Neuromorphic electronics has emerged in the last decade, inspired by biological neurons and synapses, with in-memory computing ability, extenuating the 'von Neumann bottleneck' between the memory and processor and offering a promising solution to reduce the efforts both in data storage and processing, thanks to their multi-bit non-volatility, biology-emulated characteristics, and silicon compatibility. This work reviews the recent advances in emerging memristive devices for artificial neuron and synapse applications, including memory and data-processing ability: the physics and characteristics are discussed first, i.e., valence changing, electrochemical metallization, phase changing, interfaced-controlling, charge-trapping, ferroelectric tunnelling, and spin-transfer torquing. Next, we propose a universal benchmark for the artificial synapse and neuron devices on spiking energy consumption, standby power consumption, and spike timing. Based on the benchmark, we address the challenges, suggest the guidelines for intra-device and inter-device design, and provide an outlook for the neuromorphic applications of resistive switching-based artificial neuron and synapse devices.

Revival of Ferroelectric Memories Based on Emerging Fluorite-Structured Ferroelectrics

Over the last few decades, the research on ferroelectric memories has been limited due to their dimensional scalability and incompatibility with complementary metal-oxide-semiconductor (CMOS) technology. The discovery of ferroelectricity in fluorite-structured oxides revived interest in the research on ferroelectric memories, by inducing nanoscale nonvolatility in state-of-the-art gate insulators by minute doping and thermal treatment. The potential of this approach has been demonstrated by the fabrication of sub-30 nm electronic devices. Nonetheless, to realize practical applications, various technical limitations, such as insufficient reliability including endurance, retention, and imprint, as well as large device-to-device-variation, require urgent solutions. Furthermore, such limitations should be considered based on targeting devices as well as applications. Various types of ferroelectric memories including ferroelectric random-access-memory, ferroelectric field-effect-transistor, and ferroelectric tunnel junction should be considered for classical nonvolatile memories as well as emerging neuromorphic computing and processing-in-memory. Therefore, from the viewpoint of materials science, this review covers the recent research focusing on ferroelectric memories from the history of conventional approaches to future prospects.

From Ferroelectric Material Optimization to Neuromorphic Devices

Due to the voltage driven switching at low voltages combined with nonvolatility of the achieved polarization state, ferroelectric materials have a unique potential for low power nonvolatile electronic devices. The competitivity of such devices is hindered by compatibility issues of well-known ferroelectrics with established semiconductor technology. The discovery of ferroelectricity in hafnium oxide changed this situation. The natural application of nonvolatile devices is as a memory cell. Nonvolatile memory devices also built the basis for other applications like in-memory or neuromorphic computing. Three different basic ferroelectric devices can be constructed: ferroelectric capacitors, ferroelectric field effect transistors and ferroelectric tunneling junctions. In this article first the material science of the ferroelectricity in hafnium oxide will be summarized with a special focus on tailoring the switching characteristics towards different applications.The current status of nonvolatile ferroelectric memories then lays the ground for looking into applications like in-memory computing. Finally, a special focus will be given to showcase how the basic building blocks of spiking neural networks, the neuron and the synapse, can be realized and how they can be combined to realize neuromorphic computing systems. A summary, comparison with other technologies like resistive switching devices and an outlook completes the paper.

Compute in-Memory with Non-Volatile Elements for Neural Networks: A Review from a Co-Design Perspective

Deep learning has become ubiquitous, touching daily lives across the globe. Today, traditional computer architectures are stressed to their limits in efficiently executing the growing complexity of data and models. Compute-in-memory (CIM) can potentially play an important role in developing efficient hardware solutions that reduce data movement from compute-unit to memory, known as the von Neumann bottleneck. At its heart is a cross-bar architecture with nodal non-volatile-memory elements that performs an analog multiply-and-accumulate operation, enabling the matrix-vector-multiplications repeatedly used in all neural network workloads. The memory materials can significantly influence final system-level characteristics and chip performance, including speed, power, and classification accuracy. With an over-arching co-design viewpoint, this review assesses the use of cross-bar based CIM for neural networks, connecting the material properties and the associated design constraints and demands to application, architecture, and performance. Both digital and analog memory are considered, assessing the status for training and inference, and providing metrics for the collective set of properties non-volatile memory materials will need to demonstrate for a successful CIM technology.

Synaptic plasticity and non-volatile memory characteristics in TiN-nanocrystal-embedded 3D vertical memristor-based synapses for neuromorphic systems

Although vertical configurations for high-density storage require challenging process steps, such as etching high aspect ratios and atomic layer deposition (ALD), they are more affordable with a relatively simple lithography process and have been employed in many studies. Herein, the potential of memristors with CMOS-compatible 3D vertical stacked structures of Pt/Ti/HfOx/TiN-NCs/HfOx/TiN is examined for use in neuromorphic systems. The electrical characteristics (including I-V properties, retention, and endurance) were investigated for both planar single cells and vertical resistive random-access memory (VRRAM) cells at each layer, demonstrating their outstanding non-volatile memory capabilities. In addition, various synaptic functions (including potentiation and depression) under different pulse schemes, excitatory postsynaptic current (EPSC), and spike-timing-dependent plasticity (STDP) were investigated. In pattern recognition simulations, an improved recognition rate was achieved by the linearly changing conductance, which was enhanced by the incremental pulse scheme. The achieved results demonstrated the feasibility of employing VRRAM with TiN nanocrystals in neuromorphic systems that resemble the human brain.

Organic Memristor-Based Flexible Neural Networks with Bio-Realistic Synaptic Plasticity for Complex Combinatorial Optimization

Hardware neural networks with mechanical flexibility are promising next-generation computing systems for smart wearable electronics. Several studies have been conducted on flexible neural networks for practical applications; however, developing systems with complete synaptic plasticity for combinatorial optimization remains challenging. In this study, the metal-ion injection density is explored as a diffusive parameter of the conductive filament in organic memristors. Additionally, a flexible artificial synapse with bio-realistic synaptic plasticity is developed using organic memristors that have systematically engineered metal-ion injections, for the first time. In the proposed artificial synapse, short-term plasticity (STP), long-term plasticity, and homeostatic plasticity are independently achieved and are analogous to their biological counterparts. The time windows of the STP and homeostatic plasticity are controlled by the ion-injection density and electric-signal conditions, respectively. Moreover, stable capabilities for complex combinatorial optimization in the developed synapse arrays are demonstrated under spike-dependent operations. This effective concept for realizing flexible neuromorphic systems for complex combinatorial optimization is an essential building block for achieving a new paradigm of wearable smart electronics associated with artificial intelligent systems.

Improved Endurance of Ferroelectric Hf0.5Zr0.5O2 Using Laminated-Structure Interlayer

In this article, the endurance characteristic of the TiN/HZO/TiN capacitor was improved by the laminated structure of a ferroelectric Hf_0.5Zr_0.5O₂ thin film. Altering the HZO deposition ratio, the laminated-structure interlayer was formed in the middle of the HZO film. Although small remanent polarization reduction was observed in the capacitor with a laminated structure, the endurance characteristic was improved by two orders of magnitude (from 10⁶ to 10⁸ cycles). Moreover, the leakage current of the TiN/HZO/TiN capacitor with the laminated-structure interlayer was reduced by one order of magnitude. The reliability enhancement was proved by the Time-Dependent Dielectric Breakdown (TDDB) test, and the optimization results were attributed to the migration inhibition and nonuniform distribution of oxygen vacancies. Without additional materials and a complicated process, the laminated-structure method provides a feasible strategy for improving HZO device reliability.

Ligand-Triggered Self-Assembly of Flexible Carbon Dot Nanoribbons for Optoelectronic Memristor Devices and Neuromorphic Computing

Carbon dots (CDs) are widely utilized in sensing, energy storage, and catalysis due to their excellent optical, electrical and semiconducting properties. However, attempts to optimize their optoelectronic performance through high-order manipulation have met with little success to date. In this study, through efficient packing of individual CDs in two-dimensions, the synthesis of flexible CDs ribbons is demonstrated technically. Electron microscopies and molecular dynamics simulations, show the assembly of CDs into ribbons results from the tripartite balance of π-π attractions, hydrogen bonding, and halogen bonding forces provided by the superficial ligands. The obtained ribbons are flexible and show excellent stability against UV irradiation and heating. CDs ribbons offer outstanding performance as active layer material in transparent flexible memristors, with the developed devices providing excellent data storage, retention capabilities, and fast optoelectronic responses. A memristor device with a thickness of 8 µm shows good data retention capability even after 10⁴ cycles of bending. Furthermore, the device functions effectively as a neuromorphic computing system with integrated storage and computation capabilities, with the response speed of the device being less than 5.5 ns. These properties create an optoelectronic memristor with rapid Chinese character learning capability. This work lays the foundation for wearable artificial intelligence.

Synaptic Characteristic of Hafnia-Based Ferroelectric Tunnel Junction Device for Neuromorphic Computing Application

Owing to the 4th Industrial Revolution, the amount of unstructured data, such as voice and video data, is rapidly increasing. Brain-inspired neuromorphic computing is a new computing method that can efficiently and parallelly process rapidly increasing data. Among artificial neural networks that mimic the structure of the brain, the spiking neural network (SNN) is a network that imitates the information-processing method of biological neural networks. Recently, memristors have attracted attention as synaptic devices for neuromorphic computing systems. Among them, the ferroelectric doped-HfO₂-based ferroelectric tunnel junction (FTJ) is considered as a strong candidate for synaptic devices due to its advantages, such as complementary metal-oxide-semiconductor device/process compatibility, a simple two-terminal structure, and low power consumption. However, research on the spiking operations of FTJ devices for SNN applications is lacking. In this study, the implementation of long-term depression and potentiation as the spike timing-dependent plasticity (STDP) rule in the FTJ device was successful. Based on the measured data, a CrossSim simulator was used to simulate the classification of handwriting images. With a high accuracy of 95.79% for the Mixed National Institute of Standards and Technology (MNIST) dataset, the simulation results demonstrate that our device is capable of differentiating between handwritten images. This suggests that our FTJ device can be used as a synaptic device for implementing an SNN.

Progressive and Stable Synaptic Plasticity with Femtojoule Energy Consumption by the Interface Engineering of a Metal/Ferroelectric/Semiconductor

In the era of "big data," the cognitive system of the human brain is being mimicked through hardware implementation of highly accurate neuromorphic computing by progressive weight update in synaptic electronics. Low-energy synaptic operation requires both low reading current and short operation time to be applicable to large-scale neuromorphic computing systems. In this study, an energy-efficient synaptic device is implemented comprising a Ni/Pb(Zr_0.52 Ti_0.48 )O₃ (PZT)/0.5 wt.% Nb-doped SrTiO₃ (Nb:STO) heterojunction with a low reading current of 10 nA and short operation time of 20-100 ns. Ultralow femtojoule operation below 9 fJ at a synaptic event, which is comparable to the energy required for synaptic events in the human brain (10 fJ), is achieved by adjusting the Schottky barrier between the top electrode and ferroelectric film. Moreover, progressive domain switching in ferroelectric PZT successfully induces both low nonlinearity/asymmetry and good stability of the weight update. The synaptic device developed here can facilitate the development of large-scale neuromorphic arrays for artificial neural networks with low energy consumption and high accuracy.

Memristive technologies for data storage, computation, encryption, and radio-frequency communication

Memristive devices, which combine a resistor with memory functions such that voltage pulses can change their resistance (and hence their memory state) in a nonvolatile manner, are beginning to be implemented in integrated circuits for memory applications. However, memristive devices could have applications in many other technologies, such as non-von Neumann in-memory computing in crossbar arrays, random number generation for data security, and radio-frequency switches for mobile communications. Progress toward the integration of memristive devices in commercial solid-state electronic circuits and other potential applications will depend on performance and reliability challenges that still need to be addressed, as described here.

High-Speed Switching and Giant Electroresistance in an Epitaxial Hf0.5Zr0.5O2-Based Ferroelectric Tunnel Junction Memristor

HfO₂-based ferroelectric materials are good candidates for constructing next-generation nonvolatile memories and high-performance electronic synapses and have attracted extensive attention from both academia and industry. Here, a Hf_0.5Zr_0.5O₂-based ferroelectric tunnel junction (FTJ) memristor is successfully fabricated by epitaxially growing a Hf_0.5Zr_0.5O₂ film on a 0.7 wt % Nb-doped SrTiO₃ (001) substrate with a buffer layer of La_2/3Sr_1/3MnO₃ (∼1 u.c.). The FTJ shows a high switching speed of 20 ns, a giant electroresistance ratio of ∼834, and multiple states (eight states or three bits) with good retention >10⁴ s. As a solid synaptic device, tunable synapse functions have also been obtained, including long-term potentiation, long-term depression, and spike-timing-dependent plasticity. These results highlight the promising applications of Hf_0.5Zr_0.5O₂-based FTJ in ultrafast-speed and high-density nonvolatile memories and artificial synapses.

Constructing correlation between ferroelectricity and grain sizes in Hf0.5Zr0.5O2 ferroelectric thin films

The correlation between ferroelectricity and grain sizes in atomic layer deposition (ALD) derived Hf0.5Zr0.5O2 (HZO) thin films has been demonstrated through controlling process conditions. The influence of annealing temperature, annealing...

Selector-less Ferroelectric Tunnel Junctions by Stress Engineering and an Imprinting Effect for High-Density Cross-Point Synapse Arrays

In the quest for highly scalable and three-dimensional (3D) stackable memory components, ferroelectric tunnel junction (FTJ) crossbar architectures are promising technologies for nonvolatile logic and neuromorphic computing. Most FTJs, however, require additional nonlinear devices to suppress sneak-path current, limiting large-scale arrays in practical applications. Moreover, the giant tunneling electroresistance (TER) remains challenging due to their inherent weak polarization. Here, we present that the employment of a diffusion barrier layer as well as a bottom metal electrode having a significantly low thermal expansion coefficient has been identified as an important way to enhance the strain, stabilize the ferroelectricity, and manage the leakage current in ultrathin hafnia film, achieving a high TER of 100, negligible resistance changes even up to 10⁸ cycles, and a high switching speed of a few tens of nanoseconds. Also, we demonstrate that the usage of an imprinting effect in a ferroelectric capacitor induced by an ionized oxygen vacancy near the electrode results in highly asymmetric current-voltage characteristics with a rectifying ratio of 1000. Notably, the proposed FTJ exhibits a high density array size (>4k) with a securing read margin of 10%. These findings provide a guideline for the design of high-performance and selector-free FTJ devices for large-scale crossbar arrays in neuromorphic applications.

Ultralow Power Optical Synapses Based on MoS2 Layers by Indium-Induced Surface Charge Doping for Biomimetic Eyes

Biomimetic eyes, with their excellent imaging functions such as large fields of view and low aberrations, have shown great potentials in the fields of visual prostheses and robotics. However, high power consumption and difficulties in device integration severely restrict their rapid development. In this study, an artificial synaptic device consisting of a molybdenum disulfide (MoS₂ ) film coated with an electron injection enhanced indium (In) layer is proposed to increase the channel conductivity and reduce the power consumption. This artificial synaptic device achieves an ultralow power consumption of 68.9 aJ per spike, which is several hundred times lower than those of the optical artificial synapses reported in literature. Furthermore, the multilayer and polycrystalline MoS₂ film shows persistent photoconductivity performance, effectively resulting in short-term plasticity, long-term plasticity, and their transitions between each other. A 5 × 5 In/MoS₂ synaptic device array is constructed into a hemispherical electronic retina, demonstrating its impressive image sensing and learning functions. This research provides a new methodology for effective control of artificial synaptic devices, which have great opportunities used in bionic retinas, robots, and visual prostheses.

Atomic-scale fatigue mechanism of ferroelectric tunnel junctions

Ferroelectric tunnel junctions (FTJs) are promising candidates for next-generation memories due to fast read/write speeds and low-power consumptions. Here, we investigate resistance fatigue of FTJs, which is performed on Pt/BaTiO₃/Nb:SrTiO₃ devices. By direct observations of the 5–unit cell–thick BaTiO₃ barrier with high-angle annular dark-field imaging and electron energy loss spectroscopy, oxygen vacancies are found to aggregate at the Pt/BaTiO₃ interface during repetitive switching, leading to a ferroelectric dead layer preventing domain nucleation and growth. Severe oxygen deficiency also makes BaTiO₃ lattices energetically unfavorable and lastly induces a destruction of local perovskite structure of the barrier. Ferroelectric properties are thus degraded, which reduces barrier contrast between ON and OFF states and smears electroresistance characteristics of Pt/BaTiO₃/Nb:SrTiO₃ FTJs. These results reveal an atomic-scale fatigue mechanism of ultrathin ferroelectric barriers associated with the aggregation of charged defects, facilitating the design of reliable FTJs and ferroelectric nanoelectronic devices for practical applications.

An artificial synaptic transistor using an α-In2Se3 van der Waals ferroelectric channel for pattern recognition

Despite being widely investigated for their memristive behavior, ferroelectrics are barely studied as channel materials in field-effect transistor (FET) configurations. In this work, we use multilayer α-In₂Se₃ to realize a ferroelectric channel semiconductor FET, i.e., FeS-FET, whose gate-triggered and polarization-induced resistive switching is then exploited to mimic an artificial synapse. The FeS-FET exhibits key signatures of a synapse such as excitatory and inhibitory postsynaptic current, potentiation/depression, and paired pulsed facilitation. Multiple stable conductance states obtained by tuning the device are then used as synaptic weights to demonstrate pattern recognition by invoking a hidden layer perceptron model. Detailed artificial neural network (ANN) simulations are performed on binary scale MNIST data digits, invoking 784 input (28 × 28 pixels) and 10 output neurons which are used in the training of 42 000 MNIST data digits. By updating the synaptic weights with conductance weight values on 18 000 digits, we achieved a successful recognition rate of 93% on the testing data. Introduction of 0.10 variance of noise pixels results in an accuracy of more than 70% showing the strong fault-tolerant nature of the conductance states. These synaptic functionalities, learning rules, and device to system-level simulation results based on α-In₂Se₃ could facilitate the development of more complex neuromorphic hardware systems based on FeS-FETs.

Despite being widely investigated for their memristive behavior, ferroelectrics are barely studied as channel materials in field-effect transistor (FET) configurations.

Nonvolatile Reconfigurable 2D Schottky Barrier Transistors

Nonvolatile reconfigurable transistors can be used to implement highly flexible and compact logic circuits with low power consumption in maintaining the configuration. In this paper, we build nonvolatile reconfigurable transistors based on 2D CuInP₂S₆/MoTe₂ heterostructures. The ferroelectric polarization-induced electron and hole doping in the heterostructure are investigated. By introducing the ferroelectric doping into the source/drain contacts, we demonstrate reconfigurable Schottky barrier transistors, whose polarity (n-type or p-type) can be dynamically programmed, where the configuration is nonvolatile in nature. These transistors exhibit a tunable photoresponse, where the n-n doping state leads to negative photocurrent, whereas the p-p doping state gives rise to a positive photocurrent. The transistor with asymmetric (n-p or p-n) contacts exhibits a strong photovoltaic effect. These reconfigurable logic and optoelectronic transistors will enable a new type of device fabric for future computing systems and sensing networks.

Interlayer engineering for enhanced ferroelectric tunnel junction operations in HfOx-based metal-ferroelectric-insulator-semiconductor stack

Ferroelectric tunnel junction (FTJ) has been considered as a promising candidate for next-generation memory devices due to its non-destructive and low power operations. In this article, we demonstrate the interlayer (IL) engineering in the FTJs to boost device performances. Through the analysis on the material and electrical characteristics of the fabricated FTJs with engineered IL stacks, it is clearly found that the insertion of an Al₂O₃layer between the SiO₂insulator and the pure-HfO_xFE improves the read disturbance (2V_c = 2.2 V increased), the endurance characteristics (tenfold improvement), and the cell-to-cell TER variation simultaneously without the degradation of the ferroelectricity (less than 5%) and the polarization switching speeds through grain size modulation. Based on these investigations, the guidelines of IL engineering for low power ferroelectric devices were provided to obtain stable and fast memory operations.

Effect of ferroelectric and interface films on the tunneling electroresistance of the Al2O3/Hf0.5Zr0.5O2based ferroelectric tunnel junctions

Ferroelectric random-access memory (FRAM) based on conventional ferroelectric materials is a non-volatile memory with fast read/write operations, high endurance, and 10 years of data retention time. However, it suffers from destructive read-out operation and lack of CMOS compatibility. HfO₂-based ferroelectric tunnel junctions (FTJ) may compensate for the shortcomings of FRAM by its CMOS compatibility, fast operation speed, and non-destructive readout operation. In this study, we investigate the effect of ferroelectric and interface film thickness on the tunneling electroresistance or ON/OFF current ratio of the Hf_0.5Zr_0.5O₂/Al₂O₃based FTJ device. Integrating a thick ferroelectric layer (i.e. 12 nm Hf_0.5Zr_0.5O₂) with a thin interface layer (i.e. 1 nm Al₂O₃) resulted in an ON/OFF current ratio of 78. Furthermore, to elucidate the relationship between ON/OFF current ratio and interfacial properties, the Hf_0.5Zr_0.5O₂-Al₂O₃films and Ge-Al₂O₃interfaces are examined via time-of-flight secondary ion mass spectrometry depth profiling mode. A bilayer oxide heterostructure (Hf_0.5Zr_0.5O₂/Al₂O₃) is deposited by atomic layer deposition (ALD) on the Ge substrate. The ON/OFF current ratio is enhanced by an order of magnitude when the Hf_0.5Zr_0.5O₂film deposition mode is changed from exposure (H₂O) ALD to sequential plasma (sequential O₂-H₂) ALD. Moreover, the interfacial engineering approach based on thein situALD H₂-plasma surface pre-treatment of Ge increases the ON/OFF current ratio from 9 to 38 by reducing the interfacial trap density state at the Ge-Al₂O₃interface and producing Al₂O₃with fewer oxygen vacancies as compared to the wet etch (HF + H₂O rinse) treatment of the Ge substrate. This study provides evidence of strong coupling between Hf_0.5Zr_0.5O₂and Al₂O₃films in controlling the ON/OFF current ratio of the FTJ.

Defects in ferroelectric HfO2

The discovery of ferroelectricity in polycrystalline thin films of doped HfO2 has reignited the expectations of developing competitive ferroelectric non-volatile memory devices. To date, it is widely accepted that the performance of HfO2-based ferroelectric devices during their life cycle is critically dependent on the presence of point defects as well as structural phase polymorphism, which mainly originates from defects either. The purpose of this review article is to overview the impact of defects in ferroelectric HfO2 on its functional properties and the resulting performance of memory devices. Starting from the brief summary of defects in classical perovskite ferroelectrics, we then introduce the known types of point defects in dielectric HfO2 thin films. Further, we discuss main analytical techniques used to characterize the concentration and distribution of defects in doped ferroelectric HfO2 thin films as well as at their interfaces with electrodes. The main part of the review is devoted to the recent experimental studies reporting the impact of defects in ferroelectric HfO2 structures on the performance of different memory devices. We end up with the summary and perspectives of HfO2-based ferroelectric competitive non-volatile memory devices.

Evolution of Bio-Inspired Artificial Synapses: Materials, Structures, and Mechanisms

Artificial synapses (ASs) are electronic devices emulating important functions of biological synapses, which are essential building blocks of artificial neuromorphic networks for brain-inspired computing. A human brain consists of several quadrillion synapses for information storage and processing, and massively parallel computation. Neuromorphic systems require ASs to mimic biological synaptic functions, such as paired-pulse facilitation, short-term potentiation, long-term potentiation, spatiotemporally-correlated signal processing, and spike-timing-dependent plasticity, etc. Feature size and energy consumption of ASs need to be minimized for high-density energy-efficient integration. This work reviews recent progress on ASs. First, synaptic plasticity and functional emulation are introduced, and then synaptic electronic devices for neuromorphic computing systems are discussed. Recent advances in flexible artificial synapses for artificial sensory nerves are also briefly introduced. Finally, challenges and opportunities in the field are discussed.

Competing memristors for brain-inspired computing

The expeditious development of information technology has led to the rise of artificial intelligence (AI). However, conventional computing systems are prone to volatility, high power consumption, and even delay between the processor and memory, which is referred to as the von Neumann bottleneck, in implementing AI. To address these issues, memristor-based neuromorphic computing systems inspired by the human brain have been proposed. A memristor can store numerous values by changing its resistance and emulate artificial synapses in brain-inspired computing. Here, we introduce six types of memristors classified according to their operation mechanisms: ionic migration, phase change, spin, ferroelectricity, intercalation, and ionic gating. We review how memristor-based neuromorphic computing can learn, infer, and even create, using various artificial neural networks. Finally, the challenges and perspectives in the competing memristor technology for neuromorphic computing systems are discussed.

Excellent Reliability and High-Speed Antiferroelectric HfZrO2 Tunnel Junction by a High-Pressure Annealing Process and Built-In Bias Engineering

Hafnia-based ferroelectric tunnel junctions (FTJs) have great potential for use in logic in nonvolatile memory because of their complementary metal-oxide-semiconductor process compatibility, low power consumption, high scalability, and nondestructive readout. However, typically, ferroelectrics have a depolarization field, resulting in poor endurance owing to the early dielectric breakdown. Herein, an outstandingly reliable and high-speed antiferroelectric HfZrO tunnel junction (AFTJ) is probed to understand whether it is a promising candidate for next-generation nonvolatile memory applications. High-reliability AFTJ can be explained by less charge injection due to the low depolarized field. The formation of two stable nonvolatile states, even with antiferroelectric materials, is possible if asymmetric work function electrodes and fixed oxide charges are employed, generating a built-in bias and shifting the polarization-voltage curve. In addition, via high-pressure annealing, a critical voltage that determines the transition from the t-phase to the o-phase is effectively reduced (22%). The AFTJ shows a higher endurance property (>10⁹ cycles) and faster switching speed (<30 ns) than FTJ. Hence, it is proposed that with the help of internal bias modulation and high-pressure annealing, AFTJs can be employed in next-generation memory devices.

Emerging Materials for Neuromorphic Devices and Systems

Neuromorphic devices and systems have attracted attention as next-generation computing due to their high efficiency in processing complex data. So far, they have been demonstrated using both machine-learning software and complementary metal-oxide-semiconductor-based hardware. However, these approaches have drawbacks in power consumption and learning speed. An energy-efficient neuromorphic computing system requires hardware that can mimic the functions of a brain. Therefore, various materials have been introduced for the development of neuromorphic devices. Here, recent advances in neuromorphic devices are reviewed. First, the functions of biological synapses and neurons are discussed. Also, deep neural networks and spiking neural networks are described. Then, the operation mechanism and the neuromorphic functions of emerging devices are reviewed. Finally, the challenges and prospects for developing neuromorphic devices that use emerging materials are discussed.

Highly Controllable and Silicon-Compatible Ferroelectric Photovoltaic Synapses for Neuromorphic Computing

Ferroelectric synapses using polarization switching (a purely electronic switching process) to induce analog conductance change have attracted considerable interest. Here, we propose ferroelectric photovoltaic (FePV) synapses that use polarization-controlled photocurrent as the readout and thus have no limitations on the forms and thicknesses of the constituent ferroelectric and electrode materials. This not only makes FePV synapses easy to fabricate but also reduces the depolarization effect and hence enhances the polarization controllability. As a proof-of-concept implementation, a Pt/Pb(Zr_0.2Ti_0.8)O₃/LaNiO₃ FePV synapse is facilely grown on a silicon substrate, which demonstrates continuous photovoltaic response modulation with good controllability (small nonlinearity and write noise) enabled by gradual polarization switching. Using photovoltaic response as synaptic weight, this device exhibits versatile synaptic functions including long-term potentiation/depression and spike-timing-dependent plasticity. A simulated FePV synapse-based neural network achieves high accuracies (>93%) for image recognition. This study paves a new way toward highly controllable and silicon-compatible synapses for neuromorphic computing.

A 2D-SnSe film with ferroelectricity and its bio-realistic synapse application

Catering to the general trend of artificial intelligence development, simulating humans' learning and thinking behavior has become the research focus. Second-order memristors, which are more analogous to biological synapses, are the most promising devices currently used in neuromorphic/brain-like computing. However, few second-order memristors based on two-dimensional (2D) materials have been reported, and the inherent bionic physics needs to be explored. In this work, a second-order memristor based on 2D SnSe films was fabricated by the pulsed laser deposition technique. The continuously adjustable conductance of Au/SnSe/NSTO structures was achieved by gradually switching the polarization of a ferroelectric SnSe layer. The experimental results show that the bio-synaptic functions, including spike-timing-dependent plasticity, short-term plasticity and long-term plasticity, can be simulated using this two-terminal devices. Moreover, stimulus pulses with nanosecond pulse duration were applied to the device to emulate rapid learning and long-term memory in the human brain. The observed memristive behavior is mainly attributed to the modulation of the width of the depletion layer and barrier height is affected, at the SnSe/NSTO interface, by the reversal of ferroelectric polarization of SnSe materials. The device energy consumption is as low as 66 fJ, being expected to be applied to miniaturized, high-density, low-power neuromorphic computing.

Data retention and low voltage operation of Al2O3/Hf0.5Zr0.5O2 based ferroelectric tunnel junctions

Ferroelectric random-access memories based on conventional perovskite materials are non-volatile but suffer from lack of CMOS compatibility, scalability limitation, and a destructive reading scheme. On the other hand, ferroelectric tunnel junctions based on CMOS compatible hafnium oxide are a promising candidate for future non-volatile memory technology due to their simple structure, scalability, low power consumption, high operation speed, and non-destructive read-out operation. Herein, we report an efficient strategy based on the interface-engineering approach to improve upon the tunneling electroresistance effect and data retention by depositing bilayer oxide heterostructure (Al₂O₃/Hf_0.5Zr_0.5O₂) using atomic layer deposition (ALD) on Ge substrate which is treated in-situ ALD chamber with H₂-plasma before film deposition. Integrating a thin ferroelectric layer i.e. Hf_0.5Zr_0.5O₂ (8.4 nm) with a thin interface layer i.e. Al₂O₃ (1 nm) allowed us to reduce the operation (read and write) voltage to 1.4 V, and 4.3 V, respectively, while maintaining a good tunneling electroresistance or ON/OFF ratio above 10. Furthermore, an extrapolation to 1000 years at room temperature gives a residual ON/OFF ratio of 4.

In-Plane Ferroelectric Tin Monosulfide and Its Application in a Ferroelectric Analog Synaptic Device

Two-dimensional ferroelectrics is attractive for synaptic device applications because of its low power consumption and amenability to high-density device integration. Here, we demonstrate that tin monosulfide (SnS) films less than 6 nm thick show optimum performance as a semiconductor channel in an in-plane ferroelectric analogue synaptic device, whereas thicker films have a much poorer ferroelectric response due to screening effects by a higher concentration of charge carriers. The SnS ferroelectric device exhibits synaptic behaviors with highly stable room-temperature operation, high linearity in potentiation/depression, long retention, and low cycle-to-cycle/device-to-device variations. The simulated device based on ferroelectric SnS achieves ∼92.1% pattern recognition accuracy in an artificial neural network simulation. By switching the ferroelectric domains partially, multilevel conductance states and the conductance ratio can be obtained, achieving high pattern recognition accuracy.

The effects of thin film homogeneity on the performance of ferroelectric tunnel junctions

The compelling physical properties of the recently discovered ferroelectric phase in thin film Hf _x Zr_1-x O₂ have opened a window for applications such as non-volatile resistive switching memory devices with high retention known as ferroelectric tunnel junctions. In this article, we investigate the stability of these two-terminal, polarization induced resistance-switching devices with respect to the statistical reproducibility of constitutive electrical parameters based on surface thickness inhomogeneities. We provide a straightforward, quantitative model to estimate tunneling currents dependent on thickness variations, and the resulting tunneling electroresistance (TER) ratios and breakdown probability. An analytical expression for the probability distribution of tunneling currents for normally distributed thicknesses is given. Using material parameters of a TiN/HZO/Pt heterostructure, practical design requirements are deduced and an estimation with respect to the surface roughness is given for practical ferroelectric layer thicknesses and voltages below 4 nm and 1 V, respectively. In this regime, the simple model of a ballistic, direct tunneling mechanism can be used to adequately model the thickness and voltage dependence of the resistivity.

Oxygen vacancy control as a strategy to achieve highly reliable hafnia ferroelectrics using oxide electrode

Recently, hafnia ferroelectrics with two spontaneous polarization states have attracted marked attention for non-volatile, super-steep switching devices, and neuromorphic application due to their fast switching, scalability, and CMOS compatibility. However, field cycling-induced instabilities are a serious obstacle in the practical application of various low-power electronic devices that require a settled characteristic of polarization hysteresis. In this work, a large reduction in the field cycling-induced instabilities and significantly improved ferroelectric properties were observed in a Hf0.5Zr0.5O2 (HZO) thin film with a RuO2 oxide electrode. The oxide electrode can supply additional oxygen to the HZO film, consequently minimizing the oxygen vacancies at the interface which is the origin of low reliability. From the material and electrical analysis results, we verified that HZO with the RuO2 electrode has less non-ferroelectric dead layers and fewer oxygen vacancies at the interface, resulting in excellent switching properties and improved reliability. This result suggests a beneficial method to produce high-quality hafnia thin films free from interfacial defects and with stable field cycling electrical properties for actual applications.

Back-End, CMOS-Compatible Ferroelectric Field-Effect Transistor for Synaptic Weights

Neuromorphic computing architectures enable the dense colocation of memory and processing elements within a single circuit. This colocation removes the communication bottleneck of transferring data between separate memory and computing units as in standard von Neuman architectures for data-critical applications including machine learning. The essential building blocks of neuromorphic systems are nonvolatile synaptic elements such as memristors. Key memristor properties include a suitable nonvolatile resistance range, continuous linear resistance modulation, and symmetric switching. In this work, we demonstrate voltage-controlled, symmetric and analog potentiation and depression of a ferroelectric Hf_0.57Zr_0.43O₂ (HZO) field-effect transistor (FeFET) with good linearity. Our FeFET operates with low writing energy (fJ) and fast programming time (40 ns). Retention measurements have been performed over 4 bit depth with low noise (1%) in the tungsten oxide (WO_x) readout channel. By adjusting the channel thickness from 15 to 8 nm, the on/off ratio of the FeFET can be engineered from 1 to 200% with an on-resistance ideally >100 kΩ, depending on the channel geometry. The device concept is using earth-abundant materials and is compatible with a back end of line (BEOL) integration into complementary metal-oxide-semiconductor (CMOS) processes. It has therefore a great potential for the fabrication of high-density, large-scale integrated arrays of artificial analog synapses.

3D Stackable Synaptic Transistor for 3D Integrated Artificial Neural Networks

Although they have attracted enormous attention in recent years, software-based and two-dimensional hardware-based artificial neural networks (ANNs) may consume a great deal of power. Because there will be numerous data transmissions through a long interconnection for learning, power consumption in the interconnect will be an inevitable problem for low-power computing. Therefore, we suggest and report 3D stackable synaptic transistors for 3D ANNs, which would be the strongest candidate in future computing systems by minimizing power consumption in the interconnection. To overcome the problems of enormous power consumption, it might be necessary to introduce a 3D stackable ANN platform. With this structure, short vertical interconnection can be realized between the top and bottom devices, and the integration density can be significantly increased for integrating numerous neuromorphic devices. In this paper, we suggest and show the feasibility of monolithic 3D integration of synaptic devices using the channel layer transfer method through a wafer bonding technique. Using a low-temperature processible III-V and composite oxide (Al₂O₃/HfO₂/Al₂O₃)-based weight storage layer, we successfully demonstrated synaptic transistors showing good linearity (α_p/α_d = 1.8/0.5), a high transconductance ratio (6300), and very good stability. High learning accuracy of 97% was obtained in the training of 1 million MNIST images based on the device characteristics.

Memristor with a ferroelectric HfO2 layer: in which case it is a ferroelectric tunnel junction

New interest in the implementation of ferroelectric tunnel junctions has emerged following the discovery of ferroelectric properties in HfO2 films, which are fully compatible with silicon microelectronics technology. The coercive electric field to switch polarization direction in ferroelectric HfO2 is relatively high compared to classical perovskite materials, and thus it can cause the migration of non-ferroelectric charges in HfO2, namely charged oxygen vacancies. The charge redistribution would cause the change of the tunnel barrier shape and following change of the electroresistance effect. In the case of ambiguous ferroelectric properties of HfO2 ultrathin films, this oxygen-driven resistive switching effect can mimic the tunnel electroresistance effect. Here, we demonstrate two separate resistive switching regimes, depending on the applied voltage, in the same memristor device employing a ferroelectric Hf0.5Zr0.5O2 (4.5 nm) layer. The first regime originates from the polarization reversal, whereas the second one is attributed to the accumulation/depletion of the oxygen vacancies at the electrode interface. The modulation of the tunnel barrier causes the enhancement of R OFF/R ON ratio in ∼20 times compared to the tunnel electroresistance effect. The developed device was used to formulate the criteria for unambiguous discrimination between the ferroelectric-and non-ferroelectric resistive switching effects in HfO2-based ferroelectric tunnel junctions.

Ferroelectric Tunneling Junctions Based on Aluminum Oxide/ Zirconium-Doped Hafnium Oxide for Neuromorphic Computing

Ferroelectric tunneling junctions (FTJs) with tunable tunneling electroresistance (TER) are promising for many emerging applications, including non-volatile memories and neurosynaptic computing. One of the key challenges in FTJs is the balance between the polarization value and the tunneling current. In order to achieve a sizable on-current, the thickness of the ferroelectric layer needs to be scaled down below 5 nm. However, the polarization in these ultra-thin ferroelectric layers is very small, which leads to a low tunneling electroresistance (TER) ratio. In this paper, we propose and demonstrate a new type of FTJ based on metal/Al₂O₃/Zr-doped HfO₂/Si structure. The interfacial Al₂O₃ layer and silicon substrate enable sizable TERs even when the thickness of Zr-doped HfO₂ (HZO) is above 10 nm. We found that F-N tunneling dominates at read voltages and that the polarization switching in HZO can alter the effective tunneling barrier height and tune the tunneling resistance. The FTJ synapses based on Al₂O₃/HZO stacks show symmetric potentiation/depression characteristics and widely tunable conductance. We also show that spike-timing-dependent plasticity (STDP) can be harnessed from HZO based FTJs. These novel FTJs will have high potential in non-volatile memories and neural network applications.

Ferroelectric Tunneling Junctions Based on Aluminum Oxide/ Zirconium-Doped Hafnium Oxide for Neuromorphic Computing

Ferroelectric tunneling junctions (FTJs) with tunable tunneling electroresistance (TER) are promising for many emerging applications, including non-volatile memories and neurosynaptic computing. One of the key challenges in FTJs is the balance between the polarization value and the tunneling current. In order to achieve a sizable on-current, the thickness of the ferroelectric layer needs to be scaled down below 5 nm. However, the polarization in these ultra-thin ferroelectric layers is very small, which leads to a low tunneling electroresistance (TER) ratio. In this paper, we propose and demonstrate a new type of FTJ based on metal/Al₂O₃/Zr-doped HfO₂/Si structure. The interfacial Al₂O₃ layer and silicon substrate enable sizable TERs even when the thickness of Zr-doped HfO₂ (HZO) is above 10 nm. We found that F-N tunneling dominates at read voltages and that the polarization switching in HZO can alter the effective tunneling barrier height and tune the tunneling resistance. The FTJ synapses based on Al₂O₃/HZO stacks show symmetric potentiation/depression characteristics and widely tunable conductance. We also show that spike-timing-dependent plasticity (STDP) can be harnessed from HZO based FTJs. These novel FTJs will have high potential in non-volatile memories and neural network applications.

Characterization and Application of PVDF and Its Copolymer Films Prepared by Spin-Coating and Langmuir-Blodgett Method

Poly(vinylidene fluoride) (PVDF) and its copolymers are key polymers, displaying properties such as flexibility and electroactive responses, including piezoelectricity, pyroelectricity, and ferroelectricity. In the past several years, they have been applied in numerous applications, such as memory, transducers, actuators, and energy harvesting and have shown thriving prospects in the ongoing research and commercialization process. The crystalline polymorphs of PVDF can present nonpolar α, ε phase and polar β, γ, and δ phases with different processing methods. The copolymers, such as poly(vinylidene fluoride-trifluoroethylene) (P(VDF-TrFE)), can crystallize directly into a phase analogous to the β phase of PVDF. Since the β phase shows the highest dipole moment among polar phases, many reproducible and efficient methods producing β-phase PVDF and its copolymer have been proposed. In this review, PVDF and its copolymer films prepared by spin-coating and Langmuir-Blodgett (LB) method are introduced, and relevant characterization techniques are highlighted. Finally, the development of memory, artificial synapses, and medical applications based on PVDF and its copolymers is elaborated.

Stable Subloop Behavior in Ferroelectric Si-Doped HfO2

The recent demand for analogue devices for neuromorphic applications requires modulation of multiple nonvolatile states. Ferroelectricity with multiple polarization states enables neuromorphic applications with various architectures. However, deterministic control of ferroelectric polarization states with conventional ferroelectric materials has been met with accessibility issues. Here, we report unprecedented stable accessibility with robust stability of multiple polarization states in ferroelectric HfO₂. Through the combination of conventional voltage measurements, hysteresis temperature dependence analysis, piezoelectric force microscopy, first-principles calculations, and Monte Carlo simulations, we suggest that the unprecedented stability of intermediate states in ferroelectric HfO₂ is due to the small critical volume size for nucleation and the large activation energy for ferroelectric dipole flipping. This work demonstrates the potential of ferroelectric HfO₂ for analogue device applications enabling neuromorphic computing.

Ferroelectric Second-Order Memristor

While the conductance of a first-order memristor is defined entirely by the external stimuli, in the second-order memristor it is governed by the both the external stimuli and its instant internal state. As a result, the dynamics of such devices allows to naturally emulate the temporal behavior of biological synapses, which encodes the spike timing information in synaptic weights. Here, we demonstrate a new type of second-order memristor functionality in the ferroelectric HfO₂-based tunnel junction on silicon. The continuous change of conductance in the p⁺-Si/Hf_0.5Zr_0.5O₂/TiN tunnel junction is achieved via the gradual switching of polarization in ferroelectric domains of polycrystalline Hf_0.5Zr_0.5O₂ layer, whereas the combined dynamics of the built-in electric field and charge trapping/detrapping at the defect states at the bottom Si interface defines the temporal behavior of the memristor device, similar to synapses in biological systems. The implemented ferroelectric second-order memristor exhibits various synaptic functionalities, such as paired-pulse potentiation/depression and spike-rate-dependent plasticity, and can serve as a building block for the development of neuromorphic computing architectures.

Nanoscale resistive switching memory devices: a review

In this review the different concepts of nanoscale resistive switching memory devices are described and classified according to their I-V behaviour and the underlying physical switching mechanisms. By means of the most important representative devices, the current state of electrical performance characteristics is illuminated in-depth. Moreover, the ability of resistive switching devices to be integrated into state-of-the-art CMOS circuits under the additional consideration with a suitable selector device for memory array operation is assessed. From this analysis, and by factoring in the maturity of the different concepts, a ranking methodology for application of the nanoscale resistive switching memory devices in the memory landscape is derived. Finally, the suitability of the different device concepts for beyond pure memory applications, such as brain inspired and neuromorphic computational or logic in memory applications that strive to overcome the vanNeumann bottleneck, is discussed.

Flexible Memristors Based on Single-Crystalline Ferroelectric Tunnel Junctions

Ferroelectric tunnel junction (FTJ) based memristors exhibiting continuous electric field controllable resistance states have been considered promising candidates for future high-density memories and advanced neuromorphic computational architectures. However, the use of rigid single crystal substrate and high temperature growth of the epitaxial FTJ thin films constitutes the main obstacles to using this kind of heterostructure in flexible computing devices. Here, we report the integration of centimeter-scale single crystalline FTJs on flexible plastic substrates, by water-etching based epitaxial oxide membrane lift-off and the following transfer. The resulting highly flexible FTJ membranes retain the single-crystalline structure along with stable and switchable ferroelectric polarization as the grown-on single crystal substrate state. We show that the obtained flexible memristors, i.e., FTJs on plastic substrates, present high speed and low voltage mediated memristive behaviors with resistance changes over 500% and are stable against shape change. This work is an essential step toward the realization of epitaxial ultrathin ferroelectric oxide film-based electronics on large-area, flexible, and affordable substrates.

Atomic Layer Deposited Hf0.5Zr0.5O2-based Flexible Memristor with Short/Long-Term Synaptic Plasticity

Artificial synapses are the fundamental of building a neuron network for neuromorphic computing to overcome the bottleneck of the von Neumann system. Based on a low-temperature atomic layer deposition process, a flexible electrical synapse was proposed and showed bipolar resistive switching characteristics. With the formation and rupture of ions conductive filaments path, the conductance was modulated gradually. Under a series of pre-synaptic spikes, the device successfully emulated remarkable short-term plasticity, long-term plasticity, and forgetting behaviors. Therefore, memory and learning ability were integrated to the single flexible memristor, which are promising for the next-generation of artificial neuromorphic computing systems.