Abstract In this paper, we demonstrated the feasibility of Aerosol Deposition (AD) method which can be adapted as a future fabrication process for flexible electronic devices. On basis method’s noticeable advantages such room-temperature processing, suitability mass production, wide material selectivity, and direct on substrate, fabricated evaluated conductive bridge random access memory (CBRAM) to confirm method. The CBRAM was by AD-method, novel film formation mechanism observed analyzed....

An ionic electrochemical-reaction-based two-terminal synaptic device (Ionic Synapse) is developed for improved weight updating in hardware-based neuromorphic systems. The synapse consists of an anode and a cathode that have the typical structure rechargeable battery. It can migrate metal ions from (cathode) to (anode) under electrical bias, yielding total conductance changes. Analysis fabricated reveals changes are strongly dependent on amounts migrated ions, which be sensitively controlled...

Abstract To implement a neuromorphic computing system capable of efficiently processing vast amounts unstructured data, significant number synapse and neuron devices are needed, resulting in increased area demands. Therefore, we developed nanoscale vertically structured device that supports high-density integration. realize this device, the interface effects between resistive switching layer electrode were investigated utilized. Electrical physical analyses conducted to comprehend...

In this study, a highly reliable Li based two-terminal electrochemical synapse device was developed to investigate the synaptic characteristics. Based on identified fabrication conditions, solid-state electrolyte interposed between anode and cathode prevent self-injection phenomenon. Consequently, improved reliabilities, such as endurance retention, were achieved. These results demonstrate significant potential of in hardware implementation neuromorphic computing systems.

In this letter, we report an all-solid-state Pr0.7 Ca0.3MnO3 (PCMO)-based three-terminal synapse device for neuromorphic computing. structure, PCMO channel conductance can be precisely tuned to analog levels by modulating the O ion concentration in using identical voltage pulses. It allows realization of excellent synaptic characteristics such as linear and symmetric changes stable states. These properties are important achieve nearly ideal performance system avoid complex architecture. The...

An oxygen-based ionic synaptic transistor (O-IST) is a promising element for neuromorphic computing. In this study, we demonstrated that the density of electrolyte plays key role in achieving excellent characteristics an O-IST. Pr0.7Ca0.3MnO3-based O-IST, precisely controlled HfOx and found low-density could improve ion mobility. Owing to improved mobility migration, characteristics, such as wide dynamic range, linear weight update, low operating voltage operations, stable cyclic operation,...

Abstract In this study, we investigated the effect of an Al 2 O 3 barrier layer in all-solid-state inorganic Li-based nano-ionic synaptic transistor (LST) with Li PO 4 electrolyte/WO x channel structure. Near-ideal behavior ultralow conductance range (∼50 nS) was obtained by controlling abrupt ion migration through introduction a sputter-deposited thin (∼3 nm) interfacial layer. A trade-off relationship between weight update linearity and on/off ratio varying thickness also observed. To...

The origins of the nonlinear and asymmetric synaptic characteristics TiO x -based synapse devices were investigated. Based on origins, a microstructural electrode was utilized to improve characteristics. Under an identical pulse bias, device exhibited saturated conductance changes, which led formation interfacial layer between layer, can limit consecutive oxygen migration chemical reactions, considered as main origin saturation behavior. To achieve structural engineering utilized. resultant...

Due to the rapid progress of artificial intelligence technology based on neural networks, amount required computation has been increasing dramatically. To keep up with ever-increasing demand, novel analog neuromorphic computing architectures have intensively studied, where cross-point arrays resistive memory devices are utilized for high-speed and power-efficient computation. Among various synaptic device candidates, a metal oxide-based electrochemical random-access (MO-ECRAM) attractive due...

Abstract Lately, there has been a rapid increase in the use of software-based deep learning neural networks (S-DNN) for analysis unstructured data consumption. For implementation S-DNN, synapse-device-based hardware DNN (H-DNN) proposed as an alternative to typical Von-Neumann structural computing systems. In H-DNN, various numerical values such synaptic weight, activation function, and etc., have be realized through electrical device or circuit. Among them, weight that should both positive...

Hardware neural networks (HNNs) which use synapse device (SD) arrays show promise as an approach to energy efficient parallel computation of massive vector-matrix multiplication. To maximize the inference accuracy application-specific HNNs, we propose a highly reliable 2-terminal SD with fixed resistance based on WOx films. First, investigate requirements array-based HNN through MATLAB and SPICE simulations taking into account parasitic effects in array. On top that, fabricate utilize...

To overcome the performance degradation in hardware neural networks (NNs) with non-ideal synapse devices, we proposed a novel neuromorphic architecture both TiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</sub> -based interfacial RRAM and CBRAM-based filamentary for highly accurate NN training long-term inference reliability. We used threshold-triggered scheme, which RRAMs were programmed complementary fashion. This took advantage of long...

To achieve both excellent analog switching for training and retention inference simultaneously, we investigated elevated-temperature (ET) of Pr

In this letter, we propose a compact frequency programmable oscillator with an NbO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -based insulator-metal transition (IMT) device and three-terminal Li-based electro-chemical RAM (Li-ECRAM) for coupled oscillators-based temporal pattern recognition system. Owing to the non-volatility, multilevel characteristics, linear conductance modulation of Li-ECRAM, our proposed exhibits large number...

Abstract In this research, we propose a method that can significantly improve the linearity of current–voltage characteristics ( L–IV ) synapse devices. Considering analog input data are dependent on , devices having non-linear result in drastic conductance variations during inference operations. It means is one key parameters device. To triode region metal oxide semiconductor field effect transistor (MOSFET) was utilized with Li-ion-based memristor as gate voltage divider, which results...

In hardware neural networks (HNNs), different operating temperatures cause variation in conductance of resistive arrays, and they can significantly distort the information synaptic weights, leading to a considerable loss pattern recognition accuracy. this study, WO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</sub> -based device is characterized with varying ambient temperatures, 1k-bit synapse arrays are evaluated. A systematic analysis...

We demonstrated <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{Pr}_{0.7}\text{Ca}_{0.3}\text{MnO}_{3-\mathrm{x}}$</tex> (PCMO) based Electro-Chemical Random-Access Memory (ECRAM) to achieve excellent on-chip training efficiency and inference accuracy. Accelerated oxygen ion migration through electrolyte at elevated temperature (ET) provides linearity symmetry of weight update, multi-level conductance states as well retention (10 <sup...

To improve a pattern recognition accuracy of synaptic device-based neuromorphic system, we tried to obtain symmetric conductance changes between increase process (potentiation) and decrease (depression). By utilizing compensational voltage division, achieved more gradual during the depression, which led potentiation depression. On basis changes, obviously improved was obtained on multilayer perceptron structural simulation.

We propose a titanium oxide (TiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</sub> )-based temperature sensor that can be operated in two ways, namely, by showing either abrupt or gradual metal-insulator transition (MIT) characteristics. For an as-prepared reliable TiOx-based MIT device, both and characteristics were observed under varying bias conditions. These results strongly demonstrate different types of (critical differential sensor)...

To realize a step-shaped ternary transistor capable of high-speed operation, NbO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> threshold switch (TS) device was integrated on the drain side conventional MOSFET. The fast switching speed(<63ps), excellent reliability, and moderate OFF -current level TS enable sub-10ns clock operations CMOS inverter. For comparison, various types devices were evaluated for applications. Ag-based is...

A novel 3D vertical structure that can store both information and energy is proposed in this letter. For the fabrication of structural devices, an isolation layer (IL) essential because it prevents interference between each cell device. Given occupies a large volume structure, IL be more effectively utilized. In particular, for efficiency, energy-storage device adopted into IL; thus, stored same using hybrid structure. The demonstration showed strong potential various fields stackable...

Neuromorphic computing has gained a considerable research interest due to its potential in realizing highly efficient parallel computations. However, the existing neuromorphic architectures face various drawbacks. In this study, we present an integrate-and-fire (I&F) neuron using Li-based electrochemical random access memory (Li-ECRAM) achieve exceptional area efficiency and low-power computing. The proposed Li-ECRAM employs significantly reduced number of transistors when compared other...