Low‐Cost Fabricated MgSnO Electrolyte‐Gated Synaptic Transistor with Dual Modulation of Excitation and Inhibition

Abstract

challenges in terms of speed and power usage, while neural algorithms show the potential to bypass the Von Neumann bottleneck and exceed Moore's Limitation. [1] Compared to traditional computers, the biological brain excels in learning, generalization, abstraction, and applicability. [2] Scientists have created a range of synaptic devices inspired by the structure and function of the human brain, using physical processes, and functional materials. [3] Due to their cost-effectiveness, scalability, integrated density, and low power consumption, emerging two-terminal meristem devices are potential possibilities. [4] However, low on/off current ratios, nonlinear, and asymmetric conductance changes lead to low numbers of synaptic weight and inferior linearity of synaptic weight changes. [5] Furthermore, an additional selector component is needed to select the target cell. In comparison, threeterminal synaptic transistors can exhibit low read/write currents, parallel write and read operations, low power consumption, and quasi-linear weight change, which can improve the controllability of prominent weight thereby avoiding the crosstalk problems. [6] Moreover, the post-synaptic received signal cannot be easily disrupted by the pre-synaptic input signal in the three-terminal device, which is crucial to the design of multi-input devices. [7a] Three-terminal synaptic device can control synaptic weights more easily than two-terminal synaptic device because of its structural properties of independent terminals in both training and testing phases. The gate is used for training, and the source-drain and channel are used in the test phase. [5] This structural feature prevents synaptic weights from being destroyed during the testing phase. [7b] As a promising candidate for neuromorphic applications, three-terminal electrolyte-gated synaptic transistors (EGSTs) have drawn considerable attention owe to their low operation voltages, similarity to actual biological systems through the evolutionary structural design of synaptic transistor functions. [6a,b,8] The neural networks of the brain are built on the contributions of excitatory and inhibitory neurons. Presynaptic impulses can activate or inhibit post-synaptic neurons in signal processing, memory, and learning. [6e] Specifically, excitatory postsynaptic potentials cause postsynaptic neurons to produce action potentials. In contrast, inhibitory postsynaptic potentials counterbalance excitatory effects, making postsynaptic neurons less Artificial synapses have attracted extensive attention due to their capacity to circumvent the inherent restrictions of Von Neumann computing architecture. Although electrolyte-gated synaptic transistors (EGSTs) have been enabled to emulate the essential synaptic plasticity, the corresponding neurological applications have been restricted by the lack of inhibitory synaptic response and the subsequent low asymmetric non-linearity factor. Here, Mg-doped SnO 2 (Mg:SnO 2 ) EGSTs via a low-cost and facile solution method are fabricated. With ...

Synaptic and Gradual Conductance Switching Behaviors in CeO2/Nb–SrTiO3 Heterojunction Memristors for Electrocardiogram Signal Recognition

“…In contrast, the human brain could achieve low power consumption and high computing power at the same time. The neural network consists of a vast number of neurons and synapses which cooperate to complete the transmission and processing of information. , Therefore, researchers try to find artificial synaptic devices which could achieve such functions. , The memristor with two terminal structures is not only similar to the signal transmission model of natural synapse in terms of information transmission but also has high integrated density, low power consumption, and high scalability. − …”

“…5,6 Therefore, researchers try to find artificial synaptic devices which could achieve such functions. 7,8 The memristor with two terminal structures is not only similar to the signal transmission model of natural synapse in terms of information transmission but also has high integrated density, low power consumption, and high scalability. 9−11 The switching mechanism of memristors could be divided into several categories including the conductive filament mechanism, thermochemical transformation mechanism, phase transition mechanism, and Schottky barrier adjustment mechanism.…”

Synaptic Characteristics and Neuromorphic Computing Enabled by Oxygen Vacancy Migration Based on Porous In2O3 Electrolyte-Gated Transistors

“…In addition, as summarized in Table 1, the excitatory postsynaptic currents (EPSCs) of n-type semiconductor EGTs that were previously reported in the literature were all caused by the applied positive pulse, which can be explained by taking into account the EDL effect. 14,23,24 A detailed analysis of the underlying process is illustrated in Figure S2 and Supplementary Note 1. The migration of oxygen vacancies, which are one of the most common types of ionic defects, plays a crucial role in modulating the conductance properties of transition metal oxide.…”

“…On top of that, their evolutionary structural design permits the facile implementation of various synaptic functionalities similar to the operation of natural biological systems. , As summarized in Table S1, Supporting Information, the transfer curves ( I DS – V GS ) of n-type semiconductor EGTs that were previously reported in the literature are all anticlockwise (Figure S1a). ,− Figure S1b depicts the schematic illustration of the origin of the anticlockwise hysteresis. Particularly, during the forward sweeping process, the initial positive bias can attract anions toward the electrolyte/top gate interface and at the same time push cations toward the electrolyte/semiconductor interface.…”

Ferroelectrics‐Electret Synergetic Organic Artificial Synapses with Single‐Polarity Driven Dynamic Reconfigurable Modulation

“…[9,11] Recently, various neuromorphic circuits using single device to construct synapses have been presented. [12][13][14][15] However, most of them only mimic the static plasticity, which is far from achieving the complex behaviors of the human brain. [16,17] Complex behaviors are processed at the synaptic level by a variety of excitatory and inhibitory neurotransmitters being released at the synapses, which can result in excitatory and inhibitory postsynaptic potentials.…”