Brain-like optoelectronic artificial synapses with ultralow energy consumption based on MXene floating-gates for emotion recognition

Abstract

In the new generation of brain-like optoelectronics visual signal processing and artificial perception systems, floating-gate artificial synaptic devices based on two-dimensional materials represent a feasible route. However, the traditional optoelectronic...

Ferroelectricity-Defects Synergistic Artificial Synapses for High Recognition Accuracy Neuromorphic Computing

“…Figure 4i records the mapped images for the selected learning time, revealing a clear time-proportional relationship with the visualized pixel intensities. 54 For image "8", the visualized pixel intensity increases with the training period. These findings highlight the potential and possibilities of the devices for artificial synaptic applications, offering insights for future neurocomputing and brain-like simulations.…”

Technology and Integration Roadmap for Optoelectronic Memristor

“…The concept of neuromorphic computing, which takes its cues from the human brain's highly interconnected and energyefficient biological system, has received a great deal of attention as a potential new paradigm for not only traditional information processing and storage but also for more sophisticated forms of cognitive processing, interpretation, and learning. [205,214,256] The optoelectronic random-access memories (ORAMs), particularly those combining optical sensing and synaptic functions have shown promising results in optoelectronic neuromorphic computing and visual information processing. Artificial intelligence in the coming years will need to be capable of recognizing and interpreting olfactory and auditory data similar to the human brain, which has advantages for developing a highly energy-efficient processing system (shown in Figure 14a).…”

“…[ 115 ] The working mechanism they proposed is that VO ionizes during light stimulation and turns into VO + /VO 2+ , and the released electrons increase the channel conductance. [ 205 ] However, due to the existence of the energy barrier that VO + /VO 2+ transforms back to VO, it takes a long time for VO + /VO 2+ to be neutralized by electrons (that is, the conductance is reduced). In 2018, Wu et al.…”

“…[115] The working mechanism they proposed is that VO ionizes during light stimulation and turns into VO + /VO 2+ , and the released electrons increase the channel conductance. [205] However, due to the existence of the energy barrier that VO + /VO 2+ transforms back to VO, it takes a long time for VO + /VO 2+ to be neutralized by electrons (that is, the conductance is reduced). In 2018, Wu et al [206] found that the composition of the IGZO film affects the content of oxygen vacancies (as shown in Figure 10d,e), [206] which in turn affects the size of the photocurrent and the subsequent decay time, which further verified the important role of oxygen vacancies in the PPC effect of oxide semiconductors.…”

Inhibitory artificial synapses based on photoelectric co-modulation of graphene/WSe2 van der Waals heterojunctions

“…These challenges primarily stem from performance limitations associated with traditional materials, such as narrow wavelength coverage, indirect bandgaps, weak light-matter interactions, and surface/interface defects. In recent years, two-dimensional van der Waals (2D vdW) materials, such as graphene, transition metal dichalcogenides (TMDCs), MXenes, and black phosphorus (BP), have also been widely employed in optical synapses due to their unique structures and electrical properties [24][25][26][27][28][29][30][31]. The absence of dangling bonds on the surface and the atomically thin nature of 2D vdW materials facilitate the design of various subtle heterostructures without concerns of lattice mismatch or short-channel effects [32][33][34][35][36].…”