Electrospun Nanofiber-Based Synaptic Transistor with Tunable Plasticity for Neuromorphic Computing

Biological synapses are the operational connection of the neurons for signal transmission in neuromorphic networks and hardware implementation combined with electrospun 1D nanofibers have realized its functionality for complicated computing tasks in basic three-terminal field-effect transistors with gate-controlled channel conductance. However, it still lacks the fundamental understanding that how the technological parameters influence the signal intensity of the information processing in the neural systems for the nanofiber-based synaptic transistors. Here, by tuning the electrospinning parameters and introducing the channel surface doping, an electrospun ZnO nanofiber-based transistor with tunable plasticity is presented to emulate the changing synaptic functions. The underlying mechanism of influence of carrier concentration and mobility on the device's electrical and synaptic performance is revealed as well. Short-term plasticity behaviors including paired-pulse facilitation, spike duration-dependent plasticity, and dynamic filtering are tuned in this fiber-based device. Furthermore, Perovskite-doped devices with ultralow energy consumption down to ≈0.2554 fJ and their handwritten recognition application show the great potential of synaptic transistors based on a 1D nanostructure active layer for building next-generation neuromorphic networks.     

A Reliable All-2D Materials Artificial Synapse for High Energy-Efficient Neuromorphic Computing

High-performance artificial synaptic devices are indispensable for developing neuromorphic computing systems with high energy efficiency. However, the reliability and variability issues of existing devices such as nonlinear and asymmetric weight update are the major hurdles in their practical applications for energy-efficient neuromorphic computing. Here, a two-terminal floating-gate memory (2TFGM) based artificial synapse built from all-2D van der Waals materials is reported. The 2TFGM synaptic device exhibits excellent linear and symmetric weight update characteristics with high reliability and tunability. In particular, the high linearity and symmetric synaptic weight realized by simple programming with identical pulses can eliminate the additional latency and power consumption caused by the peripheral circuit design and achieve an ultralow energy consumption for the synapses in the neural network implementation. A large number of states up to ≈3000, high switching speed of 40 ns and low energy consumption of 18 fJ for a single pulse have been demonstrated experimentally. A high classification accuracy up to 97.7% (close to the software baseline of 98%) has been achieved in the Modified National Institute of Standards and Technology (MNIST) simulations based on the experimental data. These results demonstrate the potential of all-2D 2TFGM for high-speed and low-power neuromorphic computing.     

Retina-Inspired Artificial Synapses with Ultraviolet to Near-Infrared Broadband Responses for Energy-Efficient Neuromorphic Visual Systems

Neuromorphic visual system with image perception, memory, and preprocessing functions is expected to simulate basic features of the human retina. Organic optoelectronic synaptic transistors emulating biological synapses may be promising candidates for constructing neural morphological visual system. However, the sensing wavelength range of organic optoelectronic synaptic transistors usually limits their potential in artificial multispectral visual perception. Here, retina-inspired optoelectronic synaptic transistors that present broadband responses covering ultraviolet, visible, and near-infrared regions are demonstrated, which leverage the wide-range photoresponsive charge trapping layer and the heterostructure formed between PbS quantum dots and organic semiconductor. Simplified neuromorphic visual arrays are developed to simulate comprehensive image perception, memory, and preprocessing functions. Benefitting from the flexibility of the charge trapping and organic semiconductor layers, a flexible neuromorphic visual array can be fabricated, having an ultralow power consumption of 0.55 fJ per event under a low operating voltage of −0.01 V. More significantly, an accelerating image preprocessing effect can be observed in a wide wavelength range even beyond the perception range of the human visual system, due to the gate-adjustable synaptic plasticity. These devices are highly promising for implementing neuromorphic visual systems with broadband perception, increasing image processing efficiency, and promoting the development of artificial vision.     

Bio-Inspired Photoelectric Artificial Synapse based on Two-Dimensional Ti3C2Tx MXenes Floating Gate

The highly parallel artificial neural systems based on transistor-like devices have recently attracted widespread attention due to their high-efficiency computing potential and the ability to mimic biological neurobehavior. For the past decades, plenty of breakthroughs related to synaptic transistors have been investigated and reported. In this work, a kind of photoelectronic transistor that successfully mimics the behaviors of biological synapses has been proposed and systematically analyzed. For the individual device, MXenes and the self-assembled titanium dioxide on the nanosheet surface serve as floating gate and tunneling layers, respectively. As the unit electronics of the neural network, the typical synaptic behaviors and the reliable memory stability of the synaptic transistors have been demonstrated through the voltage test. Furthermore, for the first time, the UV-responsive synaptic properties of the MXenes floating gated transistor and its applications, including conditional reflex and supervised learning, have been measured and realized. These photoelectric synapse characteristics illustrate the great potential of the device in bio-imitation vision applications. Finally, through the simulation based on an artificial neural network algorithm, the device successfully realizes the recognition application of handwritten digital images. Thus, this article provides a highly feasible solution for applying artificial synaptic devices to hardware neuromorphic networks.     

A Flexible Mott Synaptic Transistor for Nociceptor Simulation and Neuromorphic Computing

Designing transparent flexible electronics with multi-biological neuronal functions and superior flexibility is a key step to establish wearable artificial intelligence equipment. Here, a flexible ionic gel-gated VO₂ Mott transistor is developed to simulate the functions of the biological synapse. Short-term and long-term plasticity of the synapse are realized by the volatile electrostatic carrier accumulation and nonvolatile proton-doping modulation, respectively. With the achievement of multi-essential synaptic functions, an important sensory neuron, nociceptor, is perfectly simulated in our synaptic transistors with all key characteristics of threshold, relaxation, and sensitization. More importantly, this synaptic transistor exhibits high tolerance to the bending deformation, and the cycle-to-cycle variations of multi-conductance states in potentiation and depression properties are maintained within 4%. This superior stability further indicates that our flexible device is suitable for neuromorphic computing. Simulation results demonstrate that high recognition accuracy of handwritten digits (>95%) can be achieved in a convolution neural network built from these synaptic transistors. The transparent and flexible Mott transistor based on electrically-controlled VO₂ metal-insulator transition is believed to open up alternative approaches to developing highly stable synapses for future flexible neuromorphic systems.     

Organic Field Effect Transistor-Based Photonic Synapses: Materials,Devices, and Applications

Photonic artificial synapses-based neuromorphic computing systems have been regarded as promising candidates for replacing von Neumann-based computing systems due to the high bandwidth, ultrafast signal transmission, low energy consumption, and wireless communication. Although significant progress has been made in developing varied device structures for synaptic emulation, organic field-effect transistors (OFETs) hold the compelling advantages of facile preparation, liable integration, and versatile structures. As a powerful and effective platform for photonic synapses, OFETs can fulfill not only the simulation of simple synaptic functions, but also complex photoelectric dual modulation and simulation of the visual system. Herein, an overview of OFET-based photonic synapses, including functional materials, device configurations, and innovative applications is provided. Meanwhile, rules for selecting materials, mechanism of photoelectric conversion, and fabrication techniques of devices are also highlighted. Finally, challenges and opportunities are all discussed, providing solid guidance for multilevel memory, multi-functional tandem artificial neural system, and artificial intelligence.     

Printed Neuromorphic Devices Based on Printed Carbon Nanotube Thin‐Film Transistors

Hardware implementation of artificial synapse/neuron by electronic/ionic hybrid devices is of great interest for brain‐inspired neuromorphic systems. At the same time, printed electronics have received considerable interest in recent years. Here, printed dual‐gate carbon‐nanotube thin‐film transistors with very high saturation field‐effect mobility (≈269 cm² V^?1 s^–1) are proposed for artificial synapse application. Some important synaptic behaviors including paired‐pulse facilitation (PPF), and signal filtering characteristics are successfully emulated in such printed artificial synapses. The PPF index can be modulated by spike width and spike interval of presynaptic impulse voltages. The results present a printable approach to fabricate artificial synaptic devices for neuromorphic systems.     

Nanowires for UV–vis–IR Optoelectronic Synaptic Devices

Simulating biological synaptic functionalities through artificial synaptic devices opens up an innovative way to overcome the von Neumann bottleneck at the device level. Artificial optoelectronic synapses provide a non-contact method to operate the devices and overcome the shortcomings of electrical synaptic devices. With the advantages of high photoelectric conversion efficiency, adjustable light absorption coefficient, and broad spectral range, nanowires (NWs)-based optoelectronic synapses have attracted wide attention. Herein, to better promote the applications of nanowires-based optoelectronic synapses for future neuromorphic systems, the functionalities of optoelectronic synaptic devices and the current progress of NWs optoelectronic synaptic devices in UV–vis–IR spectral range are introduced. Furthermore, a bridge between NWs-based optoelectronic synaptic device and the neuromorphic system is established. Challenges for the forthcoming development of NWs optoelectronic synapses are also discussed. This review may offer a vision into the design and neuromorphic applications of NWs-based optoelectronic synaptic devices.     

Electrolyte-gated transistors for neuromorphic applications

Von Neumann computers are currently failing to follow Moore’s law and are limited by the von Neumann bottleneck.To enhance computing performance,neuromorphic computing systems that can simulate the function of the human brain are being developed.Artificial synapses are essential electronic devices for neuromorphic architectures,which have the ability to perform signal processing and storage between neighboring artificial neurons.In recent years,electrolyte-gated transistors(EGTs)have been seen as promising devices in imitating synaptic dynamic plasticity and neuromorphic applications.Among the various electronic devices,EGT-based artificial synapses offer the benefits of good stability,ultra-high linearity and repeated cyclic symmetry,and can be constructed from a variety of materials.They also spatially separate“read”and“write”operations.In this article,we provide a review of the recent progress and major trends in the field of electrolyte-gated transistors for neuromorphic applications.We introduce the operation mechanisms of electric-double-layer and the structure of EGT-based artificial synapses.Then,we review different types of channels and electrolyte materials for EGT-based artificial synapses.Finally,we review the potential applications in biological functions.     

Tuning the synaptic behaviors of biocompatible synaptic transistor through ion-doping

Biodegradable and environmentally friendly artificial synapse devices are essential for the future development of neuromorphic computing. The emergence of synaptic transistors based on biocompatible polymer materials provides an ideal approach to achieve green electronics. However, modulating the synaptic properties in a wide range in a fixed biocompatible synaptic transistor is still challengeable, while it is vitally important for improving the adaptability of the synaptic device to achieve neuro-prosthetics in the future. Here, we reported the regulation of the synaptic behavior of biocompatible synaptic transistor through ion-doping, which allows to adjusting the response of the synaptic device according to a specific function. The ions doped into the insulating layer strengthen the formation of electric double layers (EDLs), which enables a remarkable regulation effect on post-synaptic current. Moreover, basic synaptic properties, including excitatory/inhibitory post-synaptic current (EPCS/IPSC), paired-pulse facilitation/depression (PPF/PPD), short-term/long-term memory (STM/LTM) are successfully demonstrated. In addition, high-pass and low-pass filtering functions are also implemented in a single synaptic device, indicating that the synapse attenuation can be effectively transformed according to the needs of the function. More importantly, this is the first work to demonstrate that the accuracy of pattern recognition of synaptic transistors, an important indicator of neuromorphic calculations, can be significantly improved via ion doping (as high as 75.96% relative to undoped devices of 41.68%). Our research provides a feasible strategy for precisely controlling synaptic behaviors, which has a profound impact on improving the adaptability of artificial synaptic devices in the field of neuromorphic computing.     

Artificial Synapse Based on Oxygen Vacancy Migration in Ferroelectric-Like C-Axis-Aligned Crystalline InGaSnO Semiconductor Thin-Film Transistors for Highly Integrated Neuromorphic Electronics

Artificial synapses are a key component of neuromorphic computing systems. To achieve high-performance neuromorphic computing ability, a huge number of artificial synapses should be integrated because the human brain has a huge number of synapses (≈10¹⁵). In this study, a coplanar synaptic, thin-film transistor (TFT) made of c-axis-aligned crystalline indium gallium tin oxide (CAAC–IGTO) is developed. The electrical characteristics of the biological synapses such as inhibitory postsynaptic current (IPSC), paired-pulse depression (PPD), short-term plasticity (STP), and long-term plasticity at V_DS = 0.1 V, are demonstrated. The measured synaptic behavior can be explained by the migration of positively charged oxygen vacancies (V_o⁺/V_o⁺⁺) in the CAAC–IGTO layer. The mechanism of implementing synaptic behavior is completely new, compared to previous reports using electrolytes or ferroelectric gate insulators. The advantage of this device is to use conventional gate insulators such as SiO₂ for synaptic behavior. Previous studies use chitosan, Ta₂O₃, SiO₂ nanoparticles , Gd₂O₃, and HfZrO_x for gate insulators, which cannot be used for high integration of synaptic devices. The metal–oxide TFTs, widely used in the display industry, can be applied to the synaptic transistors. Therefore, CAAC–IGTO synaptic TFT can be a good candidate for application as an artificial synapse for highly integrated neuromorphic chips.     

Ultrasensitive Freestanding and Mechanically Durable Artificial Synapse with Attojoule Power Based on Na-Salt Doped Polymer for Biocompatible Neuromorphic Interface

The human brain, with high energy-efficient and parallel processing ability, inspires to mitigate power issues perplexing von Neumann architecture. As one of the essential components constructing the human brain, the emulation of biological synapses exploiting electronic devices consuming power at a biological level lays the foundation for the implementation of energy-efficient neuromorphic computing. Besides, signal matching between biologically-related stimuli and the driving voltage of artificial synapses helps to realize intelligent neuromorphic interfaces and sustainable energy. Here, ultra-sensitive artificial synapse stimulated at 1 mV with energy consumption of 132 attojoule/synaptic event is demonstrated. Biological signal matching and low power application are realized simultaneously based on sodium acetate (NaAc) doped polyvinyl alcohol (PVA) electrolyte. The biphasic current, which comprises the electrical- and ion-mediation current component, contributes to enrich synaptic functions compared to monophasic synaptic behavior. Moreover, freestanding NaAc-doped PVA membrane functions as both dielectric layer and mechanical support and facilitates to achieve flexible, transferable artificial synapse, which maintains functional stability at an ultralow voltage and power even after bending tests. Thus, encompassing superior sensitivity, low energy, and multiple functionalities with flexible, self-supported, biocompatible property, takes a step to construct energetically-efficient, complex neuromorphic systems for wearable, implantable medicines as well as smart bio-electronic interfaces.     

Tailoring Neuroplasticity in a Ferroelectric-Gated Multi-Terminal Synaptic Transistor by Bi-Directional Modulation for Improved Pattern Edge Recognition

The dynamic modulation of the plasticity of artificial neuromorphic devices facilitates a wide range of neuromorphic functions. However, integrating diverse plasticity modulation techniques into a single device presents a challenge due to limitations in the device structure design. Here, a multiterminal artificial synaptic device capable of bi-directional modulation on its plasticity is proposed. Significantly, the conversion of inhibitory and excitatory synaptic plasticity can be achieved not only by modifying the polarity of the presynaptic voltage spike but also by exchanging its input terminal between top and bottom gate while maintaining the same presynaptic stimuli. This unique bi-directional modulation of synaptic plasticity has been attributed to two distinct physical mechanisms: nonvolatile ferroelectric polarization and interface charge trap-induced memory characteristics. Additionally, the effective dynamic modulation of the synaptic behaviors is quantified under different back-gate bias and verified in the constructed neural network perceptron. Further, a visual simulation demonstrates the enhanced clarity and precision of edge recognition through the back-gate modulation in the artificial synapses. This study provides a strategy to fulfill diversified modulation on synaptic plasticity in ferroelectric-gated transistors, thereby prompting efficient and controllable neuromorphic visual systems.     

Adaptive Crystallite Kinetics in Homogenous Bilayer Oxide Memristor for Emulating Diverse Synaptic Plasticity

A critical routine for memristors applied to neuromorphic computing is to approximate synaptic dynamic behaviors as closely as possible. A type of homogenous bilayer memristor with a structure of W/HfO_y/HfO_x/Pt is designed and constructed in this paper. The memristor replicates the structure and oxygen vacancy (V_O) distribution of a complete synapse and its Ca²⁺ distribution, respectively, after the forming process. The detailed characterizations of its atomic structure and phase transformation in and near the conductive channel demonstrate that the crystallite kinetics are adaptively coupled with the V_O migration prompted by directional external bias. The extrusion (injection) of the V_Os and the subsequent crystallite coalescence (separation), phase transformation, and alignment (misalignment) resemble closely the Ca²⁺ flux and neurotransmitter dynamics in chemical synapses. Such adaptation and similarity allow the memristor to emulate diverse synaptic plasticity. This study supplies a kinetic process of conductive channel theory for bilayer memristors. In addition, our memristor has very low energy consumption (5–7.5 fJ per switching for a 0.5 µm diameter device, compatible with a synaptic event) and is therefore suitable for large‐scale integration used in neuromorphic networks.     

Recent Developments of Optoelectronic Synaptic Devices Based on Metal Halide Perovskites

Massive data processing with high computing efficiency and low operating power is required owing to the rapid development of artificial intelligence and information technology. However, the von Neumann structure computing system with the separated memory and processor can cause large energy consumption and a low running speed during massive data processing. Therefore, the brain-inspired neuromorphic computing system is developed, that can provide hardware support for emulating biological synaptic functions and realizing highly intensive data processing with low power consumption. As a neuromorphic device, the optoelectronic synaptic device (OSD) is regarded as an ideal device to replace the von Neumann-based computer because of its ultrafast signal transmission, large bandwidth, low energy consumption, and wireless communication. Owing to their unique optoelectronic property, metal halide perovskites (MHPs) have received growing attention as effective photosensitive materials in OSDs. Therefore, the review introduces the recent progress on OSDs based on MHPs (MHPs-OSDs) including the structures and properties of MHPs, and the architectures and performance characteristics of MHPs-OSDs. Furthermore, applications of MHPs-OSDs are presented. Finally, the outlook and opportunity of MHPs-OSDs are discussed.     

Advanced Optoelectronic Devices for Neuromorphic Analog Based on Low-Dimensional Semiconductors

Neuromorphic systems can parallelize the perception and computation of information, making it possible to break through the von Neumann bottleneck. Neuromorphic engineering has been developed over a long period of time based on Hebbian learning rules. The optoelectronic neuromorphic analog device combines the advantages of electricity and optics, and can simulate the biological visual system, which has a very strong development potential. Low-dimensional materials play a very important role in the field of optoelectronic neuromorphic devices due to their flexible bandgap tuning mechanism and strong light-matter coupling efficiency. This review introduces the basic synaptic plasticity of neuromorphic devices. According to the different number of terminals, two-terminal neuromorphic memristors, three-terminal neuromorphic transistors and artificial visual system are introduced from the aspects of the action mechanism and device structure. Finally, the development prospect of optoelectronic neuromorphic analog devices based on low-dimensional materials is prospected.     

Mechanoplastic Tribotronic Floating‐Gate Neuromorphic Transistor

The emulation of synaptic plasticity to achieve sophisticated cognitive functions and adaptive behaviors is critical to the evolution of neuromorphic computation and artificial intelligence. More feasible plastic strategies (e.g., mechanoplasticity) are urgent to achieve comparable, versatile, and active cognitive complexity in neuromorphic systems. Here, a versatile mechanoplastic artificial synapse based on tribotronic floating‐gate MoS₂ synaptic transistors is proposed. Mechanical displacement can induce triboelectric potential coupling to the floating‐gate synaptic transistor, trigger a postsynaptic current signal, and modulate the synaptic weights, which realizes the synaptic mechanoplasticity in an active and interactive way. Typical synaptic plasticity behaviors including potentiation/inhibition and paired pulse facilitation/depression are successfully imitated. Assistant with the charge trapping by floating gate, the artificial synapse can realize mechanical displacement derived short‐term and long‐term plasticity simultaneously. A facile artificial neural network is also constructed to demonstrate an adding synaptic weight and neuromorphic logic switching (AND, OR) by mechanoplasticity without building complex complementary metal oxide semiconductor circuits. The proposed mechanoplastic artificial synapse offers a favorable candidate for the construction of mechanical behavior derived neuromorphic devices to overcome the von Neumann bottleneck and perform advanced synaptic behaviors.     

Redox MXene Artificial Synapse with Bidirectional Plasticity and Hypersensitive Responsibility

Artificial synapses are key elements for the nervous system which is an emulation of sensory and motor neuron signal transmission. Here, the design and fabrication of redox-behavior the metal carbide nanosheets, termed MXene artificial synapse, which uses a highly-conductive MXene electrode, are reported. Benefiting from the special working mechanism of ion migration with adsorption and insertion, the device achieves world-record power consumption (460 fW) of two-terminal synaptic devices, and so far, the bidirectionally functioned synaptic device could effectively respond to ultra-small stimuli at an amplitude of ±80 mV, even exceeding that of a biological synapse. Potential applications have also been demonstrated, such as dendritic integration and memory enhancement. The special strategy and superior electrical characteristics of the bidirectionally functioned electronic device pave the way to high-power-efficiency brain-inspired electronics and artificial peripheral systems.     

Recent Advances in Ambipolar Transistors for Functional Applications

Ambipolar transistors represent a class of transistors where positive (holes) and negative (electrons) charge carriers both can transport concurrently within the semiconducting channel. The basic switching states of ambipolar transistors are comprised of common off‐state and separated on‐state mainly impelled by holes or electrons. During the past years, diverse materials are synthesized and utilized for implementing ambipolar charge transport and their further emerging applications comprising ambipolar memory, synaptic, logic, and light‐emitting transistors on account of their special bidirectional carrier‐transporting characteristic. Within this review, recent developments of ambipolar transistor field involving fundamental principles, interface modifications, selected semiconducting material systems, device structures, ambipolar characteristics, and promising applications are highlighted. The existed challenges and prospective for researching ambipolar transistors in electronics and optoelectronics are also discussed. It is expected that the review and outlook are well timed and instrumental for the rapid progress of academic sector of ambipolar transistors in lighting, display, memory, as well as neuromorphic computing for artificial intelligence.     

Piezotronic Graphene Artificial Sensory Synapse

The human somatosensory system, consisting of receptors, transmitters, and synapses, functions as the medium for external mechanical stimuli perception and sensing signal delivery/processing. Developing sophisticated artificial sensory synapses with a high performance, uncomplicated fabrication process, and low power consumption is still a great challenge. Here, a piezotronic graphene artificial sensory synapse developed by integrating piezoelectric nanogenerator (PENG) with an ion gel–gated transistor is demonstrated. The piezopotential originating from PENG can efficiently power the synaptic device due to the formation of electrical double layers at the interface of the ion gel/electrode and ion gel/graphene. Meanwhile, the piezopotential coupling is capable of linking the spatiotemporal strain information (strain amplitude and duration) with the postsynaptic current. The synaptic weights can be readily modulated by the strain pulses. Typical properties of a synapse including excitation/inhibition, synaptic plasticity, and paired pulse facilitation are successfully demonstrated. The dynamic modulation of a sensory synapse is also achieved based on dual perceptual presynaptic PENGs coupling to a single postsynaptic transistor. This work provides a new insight into developing piezotronic synaptic devices in neuromorphic computing, which is of great significance in future self‐powered electronic skin with artificial intelligence, a neuromorphic interface for neurorobotics, human–robot interaction, an intelligent piezotronic transistor, etc.