MoS₂-based Charge-trapping synaptic device with electrical and optical modulated conductance

Abstract

<jats:title>Abstract</jats:title> <jats:p>The synapse is one of the fundamental elements in human brain performing functions such as learning, memorizing, and visual processing. The implementation of synaptic devices to realize neuromorphic computing and sensing tasks is a key step to artificial intelligence, which, however, has been bottlenecked by the complex circuitry and device integration. We report a high-performance charge-trapping memory synaptic device based on two-dimensional (2D) MoS<jats:sub>2</jats:sub> and high-k Ta<jats:sub>2</jats:sub>O<jats:sub>5</jats:sub>–TiO<jats:sub>2</jats:sub> (TTO) composite to build efficient and reliable neuromorphic system, which can be modulated by both electrical and optical stimuli. Significant and essential synaptic behaviors including short-term plasticity, long-term potentiation, and long-term depression have been emulated. Such excellent synaptic behaviors originated from the good nonvolatile memory performance due to the high density of defect states in the engineered TTO composite. The 2D synaptic device also exhibits effective switching by incident light tuning, which further enables pattern recognition with accuracy rate reaching 100%. Such experimental demonstration paves a robust way toward a multitask neuromorphic system and opens up potential applications in future artificial intelligence and sensing technology.</jats:p>

Optoelectronic synapse using monolayer MoS2 field effect transistors

“…In traditional computing systems based on von Neumann architecture, memory and computational units are physically separated and connected by data bus. The bottleneck of von Neumann architecture arises from this data bus, which limits speed and efficiency and becomes inefficient while dealing with complex problems such as speech, image, and video data processing 2 . While neuromorphic systems can overcome this bottleneck by creating a network of artificial neurons and synapses, they do not respond directly to light stimulus.…”

“…A photosensitive layer, such as perovskites to absorb light and a conductive channel to transport the photogenerated carriers, with the interface enabling charge trapping has been used to generate synaptic properties 11 , 13 , 14 , 16 . Alternatively, efficient charge trapping by engineering a floating gate 2 or charge traps by functionalizing the channel/gate dielectric interface 17 have enabled the demonstration of optoelectronic memory and synapses using MoS 2 . Synaptic behavior has been demonstrated in pristine films of amorphous oxide semiconductors, such as indium-gallium-zinc-oxide (IGZO) by tapping their inherent persistent photoconductivity 18 .…”

Memristive Artificial Synapses for Neuromorphic Computing

“…These carriers can then be captured by various trap centers [75]. Mimicking of synaptic functions can be realized based on [78] the process of carriers capture and release [76,77]. The energy consumption based on this mechanism is expected to be lower than that based on ions migration and phase transition, in which high power supplies are usually needed.…”

Two-Dimensional Near-Atom-Thickness Materials for Emerging Neuromorphic Devices and Applications

“…For a negative V BG , electrons tunnel through the 4-nm-thick Al 2 O 3 barrier and are trapped in the TTO layer. This trapping mechanism enables the device to mimic optical synaptic properties ( Zhang et al., 2020 ). Similarly, John et al.…”