Light‐Stimulated Artificial Synapses Based on 2D Organic Field‐Effect Transistors

Abstract

the working principles of the brain, which may offer a new alternative paradigm for next-generation computation systems toward artificial intelligence. [3-8] In the nervous system, the transmission of information between neurons is usually performed in the chemical form of releasing neurotransmitters or in the electrical form of spikes, which is achieved through synapses. The ability of synapse to strengthen or weaken its connection strength between two neurons over time provides the physiological basis for synaptic computing and learning. [9,10] Therefore, the study of electronic devices with synaptic function is of great significance for constructing brainlike computing systems. [11-15] Thus far, lots of devices have been reported for simulating synaptic functions, including three/multi-terminal transistors [16-19] and two terminal resistant switching memories. [20-23] Compared with the two terminal devices, the transistorbased three/multi-terminal devices can write and read information synchronously. [24] Furthermore, external stimulus (e.g., light, pressure) can be easily converted into electrical signal in the transistor through proper material selection and device structure design. [25-28] So, it is possible to construct complex neuron networks with fewer transistor-based neuron elements. Previous studies on transistor-based artificial synapses mainly used traditional semiconductors as the functional layer to achieve synaptic functions. However, with the integrating development in neuromorphic chips, short channel effects will inevitably occur which strongly hinder the performance of devices, [29,30] such as the lowering in drain-induced-barrier and the increase in tunneling currents. [31,32] 2D materials with atomically layered scaling structure have only a limited vertical dimension and flat surfaces free from defects, [33,34] which have the potential to be immune to short channel effects. [35] It has been proved that it is feasible to fabricate sub-10 nm channel length devices with 2D semiconductors. [36] Thus, the emerging 2D semiconductor materials have attracted intensive research attention due to their unique size advantages, which may provide a feasible way for extending Moore's Law. [33] In addition, the 2D structure helps the active layer to gain good flexibility and optical transparency that the bulk semiconductors are sometimes hard to get. Ultrathin channels facilitate fast heat dissipation and quick respond to external stimuli, which is critical for the practical use of photosensitive electronic devices. [35,37] So far, reports on 2D active layer based 2D organic semiconductors (OSCs) with atomically layered scaling structure have been attracting intensive attention in recent years. Benefiting from their unique size advantages, 2D materials have the potential to be immune to short-channel effects. High-performance photoresponsive transistors based on 2D OSC films with excellent light-stimulated synaptic properties are reported. They exhibit a high I photo /I dark (up to 1.7 × 10 5), a competit...

Recent Progress in Synaptic Devices Based on 2D Materials

“…With an accompanying increasing research in 2D materials, a number of research groups have reported works utilizing graphene and other 2D materials in volatile and nonvolatile memory . Arnold et al reported a synaptic device through engineering of hysteresis in MoS 2 transistors to mimic the key functions in a chemical synapse .…”

“…Together with the inorganic 2D materials, 2D organic layers have attracted the attention of researchers because the solution‐based process and flexibility in 2D organic layer materials enable low fabrication cost for large‐area production. In particular, 2D organic molecules modulate the response to external signals, providing a potential platform for realizing sensory synapses in multidimensions …”

Interface Modification in Three‐Terminal Organic Memory and Synaptic Device

“…By comparing the devices with and without interfacial layer, it could be found that the interfacial charge trapping effect was the key to mimic the biological synapse. [ 30 ] Synaptic weight was an important parameter in artificial synapses, the change of which represented the plasticity of synapses after external operation (electrical pulse or light pulse) was applied. The change of synaptic weight can be expressed as the following formulation where I was the synaptic weight, I 0 was the initial synaptic weight, and I 1 was the synaptic weight after external operation was applied.…”

“…[4][5][6] Recently, in order to enrich the variety of devices, some new materials including nanoparticles, [7] polymer, [8] ferroelectric, [9] organic small molecules, [10] 2D materials, [11] and hybrid composites [12] were developed and applied in three-terminal organic memory devices, contributing to the development of three new kinds of OFET memory devices, electrodes/semiconductor layer; therefore, the interface plays an important role in these devices. [30] In this review, we focus on the interface engineering in three-terminal organic memory devices (Figure 1). First, the factors affecting the electrical properties will be examined from the perspective of interface, including four main factors morphology, surface diploes, adhesion, and surface energy.…”

“…[ 28,29 ] Similar to the mechanism of three‐terminal organic memory devices, the properties of artificial synaptic devices based on three‐terminal structure are also related to the interface between dielectric layer/semiconductor layer and source‐drain electrodes/semiconductor layer; therefore, the interface plays an important role in these devices. [ 30 ] In this review, we focus on the interface engineering in three‐terminal organic memory devices ( Figure ). First, the factors affecting the electrical properties will be examined from the perspective of interface, including four main factors morphology, surface diploes, adhesion, and surface energy.…”