Search

Abstract Electrolyte‐gated synaptic transistors (EGSTs) have attracted considerable attention as synaptic devices owing to their adjustable conductance, low power consumption, and multi‐state storage capabilities. To demonstrate high‐density EGST arrays, 2D materials are recommended owing to their excellent electrical properties and ultrathin profile. However, widespread implementation of 2D‐based EGSTs has challenges in achieving large‐area channel growth and finding compatible nanoscale solid electrolytes. This study demonstrates large‐scale process‐compatible, all‐solid‐state EGSTs utilizing molybdenum disulfide (MoS2) channels grown through chemical vapor deposition (CVD) and sub‐30 nm organic‐inorganic hybrid electrolyte polymers synthesized via initiated chemical vapor deposition (iCVD). The iCVD technique enables precise modulation of the hydroxyl group density in the hybrid matrix, allowing the modulation of proton conduction, resulting in adjustable synaptic performance. By leveraging the tunable iCVD‐based hybrid electrolyte, the fabricated EGSTs achieve remarkable attributes: a wide on/off ratio of 109, state retention exceeding 103, and linear conductance updates. Additionally, the device exhibits endurance surpassing 5 × 104 cycles, while maintaining a low energy consumption of 200 fJ/spike. To evaluate the practicality of these EGSTs, a subset of devices is employed in system‐level simulations of MNIST handwritten digit recognition, yielding a recognition rate of 93.2%.

Abstract Heterogeneous integration of dissimilar crystalline materials has recently attracted considerable attention due to its potential for high-performance multifunctional electronic and photonic devices. The conventional method for fabricating heterostructures is by heteroepitaxy, in which epitaxy is performed on crystallographically different materials. However, epitaxial limitations in monolithic growth of dissimilar materials prevent implementation of high quality heterostructures, such as complex-oxides on conventional semiconductor platforms (Si, III-V and III-N). In this work, we demonstrate gallium nitride (GaN) high-electron-mobility transistors with crystalline complex-oxide material enabled by heterogeneous integration through epitaxial lift-off and direct stacking. We successfully integrate high-κ complex-oxide SrTiO3 in freestanding membrane form with GaN heterostructure via a simple transfer process as the gate oxide. The fabricated device shows steep subthreshold swing close to the Boltzmann limit, along with negligible hysteresis and low dynamic on-resistance, indicating very low defect density between the SrTiO3 gate oxide and GaN heterostructure. Our results show that heterogeneous integration through direct material stacking is a promising route towards fabricating functional heterostructures not possible by conventional epitaxy.

Abstract Remote epitaxy, which was discovered and reported in 2017, has seen a surge of interest in recent years. Although the technology seemed to be difficult to reproduce by other labs at first, remote epitaxy has come a long way and many groups are able to consistently reproduce the results with a wide range of material systems including III-V, III-N, wide band-gap semiconductors, complex-oxides, and even elementary semiconductors such as Ge. As with any nascent technology, there are critical parameters which must be carefully studied and understood to allow wide-spread adoption of the new technology. For remote epitaxy, the critical parameters are the (1) quality of two-dimensional (2D) materials, (2) transfer or growth of 2D materials on the substrate, (3) epitaxial growth method and condition. In this review, we will give an in-depth overview of the different types of 2D materials used for remote epitaxy reported thus far, and the importance of the growth and transfer method used for the 2D materials. Then, we will introduce the various growth methods for remote epitaxy and highlight the important points in growth condition for each growth method that enables successful epitaxial growth on 2D-coated single-crystalline substrates. We hope this review will give a focused overview of the 2D-material and substrate interaction at the sample preparation stage for remote epitaxy and during growth, which have not been covered in any other review to date. Graphical Abstract

Internet‐of‐things (IoT) edge devices with a memristive neuromorphic system can more effectively enhance daily lives. However, cyberattacks remain critical concerns for smart IoT edge devices that process a vast body of information via networks. Herein, a highly secure neuromorphic system is reported, which can be implemented using a physically unclonable function (PUF) that exploits the high entropy achieved via the stochastic switching of a poly(1,3,5‐trivinyl‐1,3,5‐trimethyl cyclotrisiloxane) (pV3D3)‐based memristor. The excellent insulating property of pV3D3 enhances the stochasticity of the tunneling distance for randomly ruptured Cu filaments. The pV3D3 memristor‐based PUF (pV3D3‐PUF) achieves near‐ideal 50% averages for uniformity and uniqueness, excellent reliability under conditions of mechanical stress and water immersion, and reconfigurability‐bolstering security without additional hardware. Using stochastic in‐memory computing, the pV3D3‐PUF shows resilience to machine learning attacks. Furthermore, a cryptography protocol is demonstrated, which enables artificial intelligence service implementation without security issues for PUF‐integrated pV3D3 memristor‐based neuromorphic systems.

Epitaxial lift-off techniques, which aim to separate ultrathin single-crystalline epitaxial layers off of the substrate, are becoming increasingly important due to the need of lightweight and flexible devices for heterogeneously integrated ultracompact semiconductor platforms and bioelectronics. Remote epitaxy is a relatively newly discovered epitaxial lift-off technique that allows substrate-seeded epitaxial growth of ultrathin films through few layers of graphene. This universal epitaxial lift-off technique allows freestanding single-crystal membrane fabrication very quickly at low cost. However, the conventional method of remote epitaxy requires transfer of graphene grown on another substrate to the target single-crystalline substrate, which results in organic and metallic residues as well as macroscopic defects such as cracks and wrinkles, significantly reducing the yield of remote epitaxy. Here, we show that direct growth of thick graphene on the target single-crystalline substrate (SrTiO3 for this study) followed by atomic layer etching (ALE) of the graphene layers create a defect- and residue-free graphene surface for high yield remote epitaxy. We find that the ALE efficiently removes one atomic layer of graphene per cycle, while also clearing multi-dots (clumps of carbon atoms) that form during nucleation of the graphene layers. Our results show that direct-grown graphene on the desired substrate accompanied by ALE might potentially be an ideal pathway toward commercialization of remote epitaxy.

Deep learning-based emotion recognition using EEG has received increasing attention in recent years. The existing studies on emotion recognition show great variability in their employed methods including the choice of deep learning approaches and the type of input features. Although deep learning models for EEG-based emotion recognition can deliver superior accuracy, it comes at the cost of high computational complexity. Here, we propose a novel 3D convolutional neural network with a channel bottleneck module (CNN-BN) model for EEG-based emotion recognition, with the aim of accelerating the CNN computation without a significant loss in classification accuracy. To this end, we constructed a 3D spatiotemporal representation of EEG signals as the input of our proposed model. Our CNN-BN model extracts spatiotemporal EEG features, which effectively utilize the spatial and temporal information in EEG. We evaluated the performance of the CNN-BN model in the valence and arousal classification tasks. Our proposed CNN-BN model achieved an average accuracy of 99.1% and 99.5% for valence and arousal, respectively, on the DEAP dataset, while significantly reducing the number of parameters by 93.08% and FLOPs by 94.94%. The CNN-BN model with fewer parameters based on 3D EEG spatiotemporal representation outperforms the state-of-the-art models. Our proposed CNN-BN model with a better parameter efficiency has excellent potential for accelerating CNN-based emotion recognition without losing classification performance.