Temperature-dependent behavior of VO2-based artificial neurons DOI

Tiancheng Zhao,

Yuan Xu, Jia‐Cheng Liu

et al.

Applied Physics Letters, Journal Year: 2024, Volume and Issue: 125(21)

Published: Nov. 18, 2024

Temperature serves as a pivotal factor influencing information transmission and computational capacity in neurons, significantly affecting the function efficiency of neural networks. However, temperature dependence VO2-based artificial neuron, which is one highly promising has been hardly reported to date. Here, high-performance VO2 devices with NDR features are prepared by rapid annealing electroforming processes. We constructed neurons output properties similar those biological on basis Pearson–Anson oscillation circuit. The temperature-dependent behavior was fully investigated. Increasing leads decrease peak-to-peak value spikes neurons. spike period remains relatively stable near room temperature, but it decreases reaches above 26 °C. These ones suggesting natural advantage mimicking activity. findings contribute toward comprehending regulating based Mott memristor.

Language: Английский

Demonstration of Programmable Brain-Inspired Optoelectronic Neuron in Photonic Spiking Neural Network With Neural Heterogeneity DOI
Yun-Jhu Lee, Mehmet Berkay On, Luis El Srouji

et al.

Journal of Lightwave Technology, Journal Year: 2024, Volume and Issue: 42(13), P. 4542 - 4552

Published: Feb. 22, 2024

Photonic Spiking Neural Networks (PSNN) composed of the co-integrated CMOS and photonic elements can offer low loss, power, highly-parallel, high- throughput computing for brain-inspired neuromorphic systems. In addition, heterogeneity neuron dynamics also bring greater diversity expressivity to networks, potentially allowing implementation complex functions with fewer neurons. this paper, we design, fabricate, experimentally demonstrate an optoelectronic spiking that simultaneously achieve high programmability heterogeneous biological neural networks maintain high-speed computing. We our be programmed tune four essential parameters under 1GSpike/s input pattern signals. A single circuit tuned output three patterns, including chattering behaviors. The PSNN consisting a Mach-Zehnder interferometer (MZI) mesh synaptic network achieves 89.3% accuracy on Iris dataset. Our power consumption is 1.18 pJ/spike output, mainly limited by efficiency vertical-cavity-lasers, optical coupling efficiency, 45 nm platform used in experiment, predicted 36.84 fJ/spike 7 (e.g. ASAP7) integrated silicon photonics containing on-chip micron-scale lasers.

Language: Английский

Citations

5
Artificial tactile perception enabled by triboelectric effect DOI
Hao Lei,

Yihan Wei,

Jiayi Wang

et al.

Applied Physics Letters, Journal Year: 2025, Volume and Issue: 126(20)

Published: May 19, 2025

Artificial tactile receptors based on triboelectric nanogenerators (TENGs) hold great promise due to their high sensitivity and active pressure sensing. In this Perspective, we summarize the working mechanisms of sensors, highlighting applications in sliding perception. Additionally, tribotronic transistors TENGs synaptic have attracted attention neuromorphic computing capabilities for information. The are divided into potential model electron transfer model. physical by which forces change channel conductance states analyzed detail. Applications information perception modal identification presented show near-sensor computing. Finally, challenges faced large-scale recognition further discussed.

Language: Английский

Citations

0
Volatile and Nonvolatile Dual‐Mode Switching Operations in an Ag‐Ag2S Core‐Shell Nanoparticle Atomic Switch Network DOI Creative Commons
Oradee Srikimkaew, Saverio Ricci, M. Porzani

et al.

Advanced Electronic Materials, Journal Year: 2024, Volume and Issue: 10(10)

Published: July 10, 2024

Abstract This paper proposes a nanoparticle‐based atomic switch network memristive device, capable of both volatile and nonvolatile switching operations, which have not been previously reported for this material. The operational modes can be determined by altering the compliance current, demonstrating high stability over 100 cycles. Analysis conduction mechanism using I – V curves reveals characteristics consistent with space‐charge‐limited current during set process ohmic behavior in reset state. Furthermore, study analyzes these dual‐operational devices varying electrode spacings. results indicate that wider spacing necessitated higher volatile‐to‐nonvolatile transition, underscoring significance interconnection. These findings facilitate integration neuron synapse functions within single thereby advancing neuromorphic systems.

Language: Английский

Citations

2
Compact leak-integrate-fire neuron with auto-reset functionality based on a single spin–orbit torque magnetic tunnel junction device DOI Open Access
S. S. Wang,

Runjie Chen,

Chenyang Wang

et al.

Applied Physics Letters, Journal Year: 2024, Volume and Issue: 124(13)

Published: March 25, 2024

Leaky-integrate-fire (LIF) neurons are core components to construct a spiking neural network. The emulation of LIF has been implemented in spintronic devices, but typically suffers from challenges, such as relatively complex design and the requirement additional operations for resetting. In this Letter, we propose compact neuron device realized within single spin–orbit torque (SOT) magnetic tunnel junction device. Distinct standard memory input SOT current integrating process is applied manner that magnetization cannot cross hard plane. Consequently, can automatically reset its original state by combined effects anisotropy damping, which play vital role during leaky well. We verify proposal three types devices micromagnetic simulations, power consumption estimated 0.1 pJ/spike. auto-reset further captured our single-shot dynamic experiments. With state-of-the-art technology, work provides concise plausible scheme mimic neurons, practical interest neuromorphic computing.

Language: Английский

Citations

1
Mixed volatility in a single device: memristive non-volatile and threshold switching in SmNiO3/BaTiO3 devices DOI Creative Commons
Ruben Hamming‐Green,

M. van den Broek,

Laura Bégon‐Lours

et al.

Frontiers in Materials, Journal Year: 2024, Volume and Issue: 11

Published: May 9, 2024

Analog neuromorphic circuits use a range of volatile and non-volatile memristive effects to mimic the functionalities neurons synapses. Creating devices with combined is important for reducing footprint power consumption circuits. This work presents an epitaxial SmNiO 3 /BaTiO electrical device that displays switching either allow or block access threshold regime. behavior arises from coupling BaTiO ferroelectric polarization metal–insulator transition; in layer contact modifies resistance continuously controllable, manner. Additionally, state varies voltage at which Joule-heating-driven insulator-to-metal phase transition occurs nickelate, results negative differential curve produces sharp, switch. Reliable current oscillations stable frequencies, large amplitude, relatively low driving are demonstrated when placed Pearson–Anson-like circuit.

Language: Английский

Citations

1
A flexible thermal-coupled InGaZnO adaptive synapse DOI
Mingtao Xu, Haotian Long, Chuanyu Fu

et al.

Applied Physics Letters, Journal Year: 2024, Volume and Issue: 124(16)

Published: April 15, 2024

The development of neuromorphic sensory systems necessitates synaptic devices with adaptivity to a wide range stimuli. Furthermore, the introduction multimodal is highly favorable, which holds immense potential for improving processing capability system under complex environments. In this work, we report thermal-coupled adaptive synapse (TCAS) by integrating an IGZO-based transistor laser-induced graphene micro-heater. This enables active modulation nonlinear short-term plasticity gains through temperature and voltage co-mediated ion/electron coupling, facilitates high image denoising. images multilevel signals can be effectively denoised average reduction ∼84.0% in Euclidean distance comparison noisy images. outcome indicates effectiveness TCASs offers promising solution adaptability.

Language: Английский

Citations

0
A 1T1M Programmable Artificial Spiking Neuron via the Integration of FeFET and NbOₓ Mott Memristor DOI
Shujing Zhao, Chuanyu Han, Fengbin Tian

et al.

IEEE Electron Device Letters, Journal Year: 2024, Volume and Issue: 45(7), P. 1169 - 1172

Published: May 6, 2024

In this study, we present a one-transistor-one-memristor (1T1M) programmable artificial spiking neuron, achieved through the integration of Hf 0.5 Zr O xmlns:xlink="http://www.w3.org/1999/xlink">2 ferroelectric transistor (FeFET) and NbO xmlns:xlink="http://www.w3.org/1999/xlink">x Mott memristor. The FeFET's threshold voltage, configurable by gate write pulse ( Vpulse ), exhibits excellent retention properties, enabling storage data in multiple states. Simultaneously, memristor, characterized switching high stability, is driven FeFET, allowing for generation diverse spike rates corresponding to states FeFET. Consequently, neuron realized, with its precisely configured accurately transmit encoded neuromorphic spikes. This achievement lays groundwork development neural networks (SNNs).

Language: Английский

Citations

0
Demonstration of Neural Heterogeneity with Programmable Brain-Inspired Optoelectronic Spiking Neurons DOI
Yun-Jhu Lee, Mehmet Berkay On, Luis El Srouji

et al.

Optical Fiber Communication Conference (OFC) 2022, Journal Year: 2024, Volume and Issue: unknown, P. Tu3F.4 - Tu3F.4

Published: Jan. 1, 2024

Neural heterogeneity enables spiking neural networks to implement complex functions with fewer neurons. We designed, simulated, and demonstrated programmable optoelectronic neurons that can achieve multiple neuron characteristics based on external tuning voltages.

Language: Английский

Citations

0
Temperature-dependent behavior of VO2-based artificial neurons DOI

Tiancheng Zhao,

Yuan Xu, Jia‐Cheng Liu

et al.

Applied Physics Letters, Journal Year: 2024, Volume and Issue: 125(21)

Published: Nov. 18, 2024

Temperature serves as a pivotal factor influencing information transmission and computational capacity in neurons, significantly affecting the function efficiency of neural networks. However, temperature dependence VO2-based artificial neuron, which is one highly promising has been hardly reported to date. Here, high-performance VO2 devices with NDR features are prepared by rapid annealing electroforming processes. We constructed neurons output properties similar those biological on basis Pearson–Anson oscillation circuit. The temperature-dependent behavior was fully investigated. Increasing leads decrease peak-to-peak value spikes neurons. spike period remains relatively stable near room temperature, but it decreases reaches above 26 °C. These ones suggesting natural advantage mimicking activity. findings contribute toward comprehending regulating based Mott memristor.

Language: Английский

Citations

0