Abstract Vertical field effect transistors (VFETs) using graphene and transition metal dichalcogenides (TMDs) heterostructures are promising for downsizing the channel length to a monolayer TMD thickness of 0.65 nm. However, graphene/monolayer TMD/metal VFETs struggle with a low on/off ratio due to gate field screening by the graphene layer and a high off-state tunneling current caused by the large contact area. Here, we propose a 0.65 nm channel length VFET with a very high on/off current ratio made by cross-stacking top and bottom carbon nanotubes (CNTs) with a monolayer TMD in between. The ultra-narrow junction area in the CNT/monolayer TMD/CNT VFET can significantly reduce the off-state tunneling current. Additionally, the gate field is transmitted from the sidewall of the bottom CNT to the monolayer MoS2 vertical channel between the two CNTs without field screening, thus achieving very strong gate modulation. Unlike the BH change (< 92 meV) of the graphene/MoS2/metal junction, which is fully dependent on the Fermi level (EF) shift of graphene, the CNT/MoS2/CNT junction exhibits a larger BH change (370 meV) than the typical EF shift (20 meV with Vg = -30 ~ 20 V) of semi-metallic CNTs. As a result, our CNT/monolayer MoS2/CNT VFETs exhibit about 105 times higher on/off ratio (= 106), 105 times lower off current (= 10− 13 A), and 100 times lower SS (= 0.4 V.dec− 1) compared to graphene/monolayer TMD/metal VFETs. In the comparison between multilayer MoS2 and monolayer MoS2 VFETs, rigid multilayer MoS2 forms a large air gap at the multilayer MoS2/CNT/substrate heterostructure, which reduces electric field transmission. In contrast, monolayer MoS2 bends significantly along the sidewall of the CNT, resulting in minimal air gap formation and enhancing the electric field effect in the channel. As a result, CNT/monolayer MoS2/CNT VFET shows 10 times higher on-current saturation and on/off ratio compared to the CNT/multilayer MoS2/CNT VFET.
2D transition metal dichalcogenides (TMDs) exhibit intriguing properties for applications in optoelectronics and electronics, among which memtransistors received extensive attention as multifunctional devices. For practical applications of 2D TMDs, large-area fabrication of the materials via reliable processes, which is in trade-off with their quality, has been a long-standing issue. Here, a simple and effective way is proposed to fabricate large-area and high-quality molybdenum disulfide thin films using MoS2 colloidal ink through a spray coating, followed by a postsulfurization process. High-quality MoS2 thin films exhibit excellent optical and electrical properties that can be utilized in field-effect transistors (FETs) and memtransistor arrays. The MoS2 FETs show an average on/off ratio of 5 × 106 and a high electron mobility of 10.34 cm2 V-1 s-1 , which can be understood by the healing of sulfur vacancies, recrystallization, and the removal of the carbon contamination of the MoS2 . These MoS2 -based memtransistors present stable operations with a high switching ratio tuned by back gate and light illumination, which is promising for multiple-levels memory and complex neuromorphic computing. This study demonstrates a new strategy to fabricate 2D TMDs with large-area and high quality for integrated optoelectronic and memory device applications.
This review covers recent advancements and future directions in 2DM-based devices for in-sensor computing, focusing on unique physical mechanisms for sensory responses, biomimetic synaptic features, and potential applications.
Abstract Time‐series analysis and forecasting play a vital role in the fields of economics and engineering. Neuromorphic computing, particularly recurrent neural networks (RNNs), has emerged as an effective approach to address these tasks. Reservoir computing (RC), a type of RNN, offers a powerful and efficient solution for handling nonlinear information in high‐dimensional spaces and addressing temporal tasks. CuInP 2 S 6 (CIPS), a van der Waals material with ion conductivity, shows promise for sequential task processing. Here, a synapse device based on CIPS is demonstrated that exhibits temporal dynamics under electrical stimulation. By controlling Cu + ion migration, this study successfully emulates synaptic performance, including potentiation and depression characteristics, and RC. Migration of Cu + ions is confirmed using piezoresponse and Kelvin probe force microscopy. The device achieves low normalized root mean square errors (NRMSE) of 0.04762 and 0.01402 for the Hénon map and Mackey‐Glass series tasks, respectively. For real‐life time‐series prediction based on the Jena temperature database, an overall NRMSE of 0.03339 is achieved. These results highlight the potential of CIPS ion conductivity for real‐time signal processing in machine learning, expanding applications in neuromorphic computing.
Abstract In‐memory computing, particularly neuromorphic computing, has emerged as a promising solution to overcome the energy and time‐consuming challenges associated with the von Neumann architecture. The ferroelectric field‐effect transistor (FeFET) technology, with its fast and energy‐efficient switching and nonvolatile memory, is a potential candidate for enabling both computing and memory within a single transistor. In this study, the capabilities of an integrated ferroelectric HfO 2 and 2D MoS 2 channel FeFET in achieving high‐performance 4‐bit per cell memory with low variation and power consumption synapses, while retaining the ability to implement diverse learning rules, are demonstrated. Notably, this device accurately recognizes MNIST handwritten digits with over 94% accuracy using online training mode. These results highlight the potential of FeFET‐based in‐memory computing for future neuromorphic computing applications.
Spiking neural network (SNN), where the information is evaluated recurrently through spikes, has manifested significant promises to minimize the energy expenditure in data-intensive machine learning and artificial intelligence. Among these applications, the artificial neural encoders are essential to convert the external stimuli to a spiking format that can be subsequently fed to the neural network. Here, a molybdenum disulfide (MoS2 ) hafnium oxide-based ferroelectric encoder is demonstrated for temporal-efficient information processing in SNN. The fast domain switching attribute associated with the polycrystalline nature of hafnium oxide-based ferroelectric material is exploited for spike encoding, rendering it suitable for realizing biomimetic encoders. Accordingly, a high-performance ferroelectric encoder is achieved, featuring a superior switching efficiency, negligible charge trapping effect, and robust ferroelectric response, which successfully enable a broad dynamic range. Furthermore, an SNN is simulated to verify the precision of the encoded information, in which an average inference accuracy of 95.14% can be achieved, using the Modified National Insitute of Standards and Technology (MNIST) dataset for digit classification. Moreover, this ferroelectric encoder manifests prominent resilience against noise injection with an overall prediction accuracy of 94.73% under various Gaussian noise levels, showing practical promises to reduce the computational load for the neural network.
Reservoir computing (RC) architecture which mimics the human brain is a fundamentally preferred method to process dynamical systems that evolve with time. However, the difficulty in generating rich reservoir states using two‐terminal devices remains challenging, which hinders its hardware implementation. Herein, the 1D array of ferroelectric field‐effect transistor (Fe‐FET) based on α ‐In 2 Se 3 channel, which shows volatile memory effect for realizing various RC systems, is demonstrated. The fading effect in α ‐In 2 Se 3 is sufficiently investigated by polarization dynamic model. The proposed Fe‐FET is capable of experimentally classifying images using MNIST dataset with a high accuracy of 91%. Furthermore, time‐series real‐life chaotic system, for example, Earth's weather, can be accurately forecasted using our Ferro‐RC based on the Jena climate dataset recorded in a 1 year period. Remarkable determination coefficient ( R 2 ) of 0.9983 and normalized root mean square error (NRMSE) of 8.3 × 10 −3 are achieved using a minimized readout network. The demonstration of integrated memory and computation opens a route for realizing a compact RC hardware system.
'%3e%3cg id='b'%3e%3cpath id='a' stroke='%23000' stroke-width='2' d='M-6-25H6m-12 3H6m-12 3H6'/%3e%3cuse xlink:href='%23a' y='44'/%3e%3c/g%3e%3cpath stroke='%23fff' d='M0 17v10'/%3e%3ccircle r='12' fill='%23cd2e3a'/%3e%3cpath fill='%230047a0' d='M0-12A6 6 0 0 0 0 0a6 6 0 0 1 0 12 12 12 0 0 1 0-24z'/%3e%3c/g%3e%3cg transform='rotate(-123.69)'%3e%3cuse xlink:href='%23b'/%3e%3cpath stroke='%23fff' d='M0-23.5v3M0 17v3.5m0 3v3'/%3e%3c/g%3e%3c/svg%3e)
'/%3e%3cuse xlink:href='%23a' transform='translate(288.889 -214.211)'/%3e%3cuse xlink:href='%23a' transform='translate(108 342.749)'/%3e%3cuse xlink:href='%23a' transform='translate(469.189 342.749)'/%3e%3c/svg%3e)