Excellent synaptic behavior of lithium-based nano-ionic transistor based on optimal WO<sub>2.7</sub> stoichiometry with high ion diffusivity

Lee, Jongwon; Nikam, Revannath Dnyandeo; Lim, Seokjae; Kwak, Myonghoon; Hwang, Hyunsang

doi:10.1088/1361-6528/ab793d

Cited by 34 publications

(41 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Consequently, a very thin active resistive switching layer is formed near interface during the potentiation, which results in rapid depression by the reverse flux due to its high concentration gradient. [ 31 ] Moreover, the low diffusivity can reduce the retention ability because the average conductivity can be changed as the diffusion proceeds, despite the total number of oxygen vacancies in the channel remains constant. This issue can be resolved by operating the device within a smaller dynamic range, leading to minimal ionic transport to barely discern the states, with a trade‐off vis‐à‐vis the noise budget of the peripheral circuit (Figure 5b).…”

Section: Resultsmentioning

confidence: 99%

“…As an alternative synaptic device, electrochemical random‐access memory (ECRAM) with a three‐terminal configuration has demonstrated superior programming characteristics, including linear and symmetric update and low programming energy. [ 24–35 ] The programming of ECRAM is based on ion (e.g., Li + , H + , and O 2− ) transport, with injection into or extraction from the channel through the electrolyte to tune the electronic conductance of the channel. Furthermore, metal‐oxide based ECRAM devices have shown complementary metal‐oxide‐semiconductor‐compatibility, enabling process integration with silicon‐based integrated circuits.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Elucidating Ionic Programming Dynamics of Metal‐Oxide Electrochemical Memory for Neuromorphic Computing

Jeong

Lee

Ryu

et al. 2021

Adv Elect Materials

View full text Add to dashboard Cite

Cross‐point arrays of synaptic devices have been investigated as a core platform for neuromorphic computing architectures. To achieve a significant speed boost in deep neural network computations compared to the von Neumann architecture, it is essential to develop synaptic devices with optimal performance for fully parallel vector‐matrix‐multiplication. Among various non‐volatile memory candidates, metal‐oxide based electrochemical random‐access memory (ECRAM) is considered as a promising analog synapse due to its superior switching characteristics and CMOS‐compatibility. However, the switching mechanisms of metal‐oxide ECRAM remain to be understood, impeding improvements of the synaptic characteristics and device performance. Here, ionic programming dynamics based on oxygen ion migration and associated redox‐reactions are investigated in metal‐oxide ECRAMs by considering ion transport via the electrolyte and diffusion in the channel. Additionally, the origins of update asymmetry and long‐term retention found in voltage‐pulse programming measurements are explained by non‐uniform distribution of the off‐stoichiometry in the channel layer. By exploiting the programming mechanism, ECRAMs can achieve desirable synaptic characteristics under voltage‐pulse mode. Finally, experimental and simulation studies consistently suggest that channel and electrolyte with high ionic transport properties can improve the performance of metal‐oxide ECRAMs, opening a path toward optimized synaptic device characteristics for the maximum computation performance.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Elucidating Ionic Programming Dynamics of Metal‐Oxide Electrochemical Memory for Neuromorphic Computing

Jeong

Lee

Ryu

et al. 2021

Adv Elect Materials

View full text Add to dashboard Cite

show abstract

“…Refs. [30], [80] and [82] use the same calculation method, while ref. [21] and this work use the same calculation method.…”

mentioning

confidence: 99%

Oxide‐Based Electrolyte‐Gated Transistors for Spatiotemporal Information Processing

et al. 2020

View full text Add to dashboard Cite

systems, which can be integrated with terminal sensors to form intelligent sensory systems. [3-7] Information processing in biological sensory nervous systems involves billions of neurons interconnected through trillions of synapses, constituting immense neural networks. [8] Compared with traditional von Neumannbased computing architecture, neural networks greatly reduce time-and energy consumption by taking the advantage of co-location of logic and memory, hyperconnectivity, robustness, and massively parallel processing. [9] As the third-generation artificial neural network, spiking neural networks (SNNs) are inspired by biological nervous systems. They employ spiking neurons as computational units that process information with timing of spikes. [10] Therefore, SNNs provide the potential for spatiotemporal information processing with high time-and energy-efficiency. The learning and recognition functions aiming at spatiotemporal patterns have been demonstrated in SNN formed by resistive switching synaptic devices. [11] However, this synaptic device suffers from several performance limitations such as the discrete weight update, write variation, and high energy programing related to the filamentary switching mechanism. [12-17] Recently, He et al. use capacitively coupled multi-terminal neurotransistors for spatiotemporal information processing, where the neurotransistors mimic the dendritic discriminability of different spatiotemporal input sequences. [18] To physically emulate neural networks in hardware, large-scale crossbar Spiking neural networks (SNNs) sharing large similarity with biological nervous systems are promising to process spatiotemporal information and can provide highly time-and energy-efficient computational paradigms for the Internet-of-Things and edge computing. Nonvolatile electrolyte-gated transistors (EGTs) provide prominent analog switching performance, the most critical feature of synaptic element, and have been recently demonstrated as a promising synaptic device. However, high performance, large-scale EGT arrays, and EGT application for spatiotemporal information processing in an SNN are yet to be demonstrated. Here, an oxide-based EGT employing amorphous Nb 2 O 5 and Li x SiO 2 is introduced as the channel and electrolyte gate materials, respectively, and integrated into a 32 × 32 EGT array. The engineered EGTs show a quasi-linear update, good endurance (10 6) and retention, a high switching speed of 100 ns, ultralow readout conductance (<100 nS), and ultralow areal switching energy density (20 fJ µm −2). The prominent analog switching performance is leveraged for hardware implementation of an SNN with the capability of spatiotemporal information processing, where spike sequences with different timings are able to be efficiently learned and recognized by the EGT array. Finally, this EGT-based spatiotemporal information processing is deployed to detect moving orientation in a tactile sensing system. These results provide an insight into oxide-based EGT devices for energy-efficient neuro...

show abstract

“…The G change ratio gradually increases from 0% to 50% as the pulse amplitude increases from 1 to 6 V, where the electrostatic modulation (V g < 3 V) and electrochemical intercalation (V g > 3 V) are mainly responsible for the short-term and long-term G change behaviors, respectively. In addition, EGT responses also depend on pulse width/frequency, device geometry, and material composition, etc., as reported in previous works [31,32,35,50,51] which provide a large number of hyperparameters that can be engineered to achieve the property of demand. Figure 1c shows the schematic of designed 1T1E synapse.…”

Section: Structure and Neural Characteristics Of One-transistor-one-electrolyte-gated Transistor Synapsementioning

confidence: 99%

One Transistor One Electrolyte‐Gated Transistor Based Spiking Neural Network for Power‐Efficient Neuromorphic Computing System

Xuan

et al. 2021

Adv Funct Materials

View full text Add to dashboard Cite

Neuromorphic computing powered by spiking neural networks (SNN) provides a powerful and efficient information processing paradigm. To harvest the advantage of SNNs, compact and low‐power synapses that can reliably practice local learning rules are required, posing significant challenges to the conventional silicon‐based platform in terms of area‐ and energy‐efficiency, as well as computing throughput. Here, electrolyte‐gated transistors (EGTs) paired with transistors are employed to implement power‐efficient neuromorphic computing systems. The one‐transistor‐one‐EGT (1T1E) synapse not only alleviates the self‐discharging of EGT but also provides a flexible and efficient way to practice the important spike‐timing‐dependent plasticity learning rule. Based on that, an SNN with a temporal coding scheme is implemented for associative memory that can learn and recover images of handwritten digits with high robustness. Thanks to the temporal coding scheme and low operation current of EGTs, the energy‐efficiency of 1T1E‐based SNN is ≈30× lower than that of the prevalent rate coding scheme, and the peak performance is estimated to be 2 pJ/SOP (picojoule per synaptic operation) at the training phase and 80 TOPs−1 W−1 (tera operations per second per watt) at inference phase, respectively. These results pave the way for power‐efficient neuromorphic computing systems with wide applications for edge computing.

show abstract

Excellent synaptic behavior of lithium-based nano-ionic transistor based on optimal WO_2.7 stoichiometry with high ion diffusivity

Cited by 34 publications

References 26 publications

Elucidating Ionic Programming Dynamics of Metal‐Oxide Electrochemical Memory for Neuromorphic Computing

Elucidating Ionic Programming Dynamics of Metal‐Oxide Electrochemical Memory for Neuromorphic Computing

Oxide‐Based Electrolyte‐Gated Transistors for Spatiotemporal Information Processing

One Transistor One Electrolyte‐Gated Transistor Based Spiking Neural Network for Power‐Efficient Neuromorphic Computing System

Contact Info

Product

Resources

About

Excellent synaptic behavior of lithium-based nano-ionic transistor based on optimal WO2.7 stoichiometry with high ion diffusivity

Cited by 34 publications

References 26 publications

Elucidating Ionic Programming Dynamics of Metal‐Oxide Electrochemical Memory for Neuromorphic Computing

Elucidating Ionic Programming Dynamics of Metal‐Oxide Electrochemical Memory for Neuromorphic Computing

Oxide‐Based Electrolyte‐Gated Transistors for Spatiotemporal Information Processing

One Transistor One Electrolyte‐Gated Transistor Based Spiking Neural Network for Power‐Efficient Neuromorphic Computing System

Contact Info

Product

Resources

About

Excellent synaptic behavior of lithium-based nano-ionic transistor based on optimal WO_2.7 stoichiometry with high ion diffusivity