Ministry of Science and ICT

Oxygen vacancies and adsorbed oxygen species on metal oxide surfaces play important roles in various fields. However, existing methods for manipulating surface oxygen require severe settings and are ineffective for repetitive manipulation. We present a method to manipulate the amount of surface oxygen by modifying the oxygen adsorption energy by electrically controlling the electron concentration of the metal oxide. The surface oxygen control ability of the method is verified using first-principles calculations based on density functional theory (DFT), X-ray photoelectron spectroscopy (XPS), and electrical resistance analysis. The presented method is implemented by fabricating oxide thin film transistors with embedded microheaters. The method can reconfigure the oxygen vacancies on the In2O3, SnO2, and IGZO surfaces so that specific chemisorption dominates. The method can selectively increase oxidizing (e.g., NO and NO) and reducing gas (e.g., H2S, NH3, and CO) reactions by electrically controlling the metal oxide surface to be oxygen vacancy-rich or adsorbed oxygen species-rich. The proposed method is applied to gas sensors and overcomes their existing limitations. The method makes the sensor insensitive to one gas (e.g., H2S) in mixed-gas environments (e.g., NO2+H2S) and provides a linear response (R2 = 0.998) to the target gas (e.g., NO2) concentration within 3 s. We believe that the proposed method is applicable to applications utilizing metal oxide surfaces.

With the recently increasing prevalence of deep learning, both academia and industry exhibit substantial interest in neuromorphic computing, which mimics the functional and structural features of the human brain. To realize neuromorphic computing, an energy-efficient and reliable artificial synapse must be developed. In this study, the synaptic ferroelectric field-effect-transistor (FeFET) array is fabricated as a component of a neuromorphic convolutional neural network. Beyond the single transistor level, the long-term potentiation and depression of synaptic weights are achieved at the array level, and a successful program-inhibiting operation is demonstrated in the synaptic array, achieving a learning accuracy of 79.84% on the Canadian Institute for Advanced Research (CIFAR)-10 dataset. Furthermore, an efficient self-curing method is proposed to improve the endurance of the FeFET array by tenfold, utilizing the punch-through current inherent to the device. Low-frequency noise spectroscopy is employed to quantitatively evaluate the curing efficiency of the proposed self-curing method. The results of this study provide a method to fabricate and operate reliable synaptic FeFET arrays, thereby paving the way for further development of ferroelectric-based neuromorphic computing.

Abstract This article presents the design and realization of a compact ultra‐wideband (UWB) antenna with on‐demand WLAN band‐rejection. The antenna consists of a simple truncated rectangular patch with a U‐slot and a partial ground plane, which are both patterned on Taconic TLY‐5 substrate (ε r = 2.2). The lower corners of the patch are truncated with a semicircle to realize wideband characteristic, while the notch is obtained by etching a U‐slot on the radiating patch. The proposed antenna outperforms the existing UWB antennas owing to its compact size, radiation stability, and very wide impedance bandwidth. Simulated and measured results show that the novel antenna has a very wide operating bandwidth of 2.9‐23.5 GHz with a VSWR <2, and a notch band from 4.9 to 6.1 GHz to reject IEEE 802.11a and HIPERLAN/2 frequency band. The antenna offers promising performances including moderate gain ( G max = 6.1 dB), nearly omnidirectional stable radiation patterns, and a compact overall size of 13 × 22 × 0.8 mm 3 . Besides of the other advantages, this antenna design presents mechanical robustness, easy integration into circuit boards, and excellent low‐cost mass production suitability.

Abstract Artificial olfactory systems (AOSs) that mimic biological olfactory systems are of great interest. However, most existing AOSs suffer from high energy consumption levels and latency issues due to data conversion and transmission. In this work, an energy‐ and area‐efficient AOS based on near‐sensor computing is proposed. The AOS efficiently integrates an array of sensing units (merged field effect transistor (FET)‐type gas sensors and amplifier circuits) and an AND‐type nonvolatile memory (NVM) array. The signals of the sensing units are directly connected to the NVM array and are computed in memory, and the meaningful linear combinations of signals are output as bit line currents. The AOS is designed to detect food spoilage by employing thin zinc oxide films as gas‐sensing materials, and it exhibits low detection limits for H 2 S and NH 3 gases (0.01 ppm), which are high‐protein food spoilage markers. As a proof of concept, monitoring the entire spoilage process of chicken tenderloin is demonstrated. The system can continuously track freshness scores and food conditions throughout the spoilage process. The proposed AOS platform is applicable to various applications due to its ability to change the sensing temperature and programmable NVM cells.

Abstract Hardware neuromorphic systems are crucial for the energy‐efficient processing of massive amounts of data. Among various candidates, hafnium oxide ferroelectric tunnel junctions (FTJs) are highly promising for artificial synaptic devices. However, FTJs exhibit non‐ideal characteristics that introduce variations in synaptic weights, presenting a considerable challenge in achieving high‐performance neuromorphic systems. The primary objective of this study is to analyze the origin and impact of these variations in neuromorphic systems. The analysis reveals that the major bottleneck in achieving a high‐performance neuromorphic system is the dynamic variation, primarily caused by the intrinsic 1/ f noise of the device. As the device area is reduced and the read bias ( V Read ) is lowered, the intrinsic noise of the FTJs increases, presenting an inherent limitation for implementing area‐ and power‐efficient neuromorphic systems. To overcome this limitation, an adaptive read‐biasing (ARB) scheme is proposed that applies a different V Read to each layer of the neuromorphic system. By exploiting the different noise sensitivities of each layer, the ARB method demonstrates significant power savings of 61.3% and a scaling effect of 91.9% compared with conventional biasing methods. These findings contribute significantly to the development of more accurate, efficient, and scalable neuromorphic systems.

Smart healthcare systems integrated with advanced deep neural networks enable real-time health monitoring, early disease detection, and personalized treatment. In this work, a novel 3D AND-type flash memory array with a rounded double channel for computing-in-memory (CIM) architecture to overcome the limitations of conventional smart healthcare systems: the necessity of high area and energy efficiency while maintaining high classification accuracy is proposed. The fabricated array, characterized by low-power operations and high scalability with double independent channels per floor, exhibits enhanced cell density and energy efficiency while effectively emulating the features of biological synapses. The CIM architecture leveraging the fabricated array achieves high classification accuracy (93.5%) for electrocardiogram signals, ensuring timely detection of potentially life-threatening arrhythmias. Incorporated with a simplified spike-timing-dependent plasticity learning rule, the CIM architecture is suitable for robust, area- and energy-efficient in-memory arrhythmia detection systems. This work effectively addresses the challenges of conventional smart healthcare systems, paving the way for a more refined healthcare paradigm.

Neuromorphic computing (NC) architecture inspired by biological nervous systems has been actively studied to overcome the limitations of conventional von Neumann architectures. In this work, we propose a reconfigurable NC block using a flash-type synapse array, emerging positive feedback (PF) neuron devices, and CMOS peripheral circuits, and integrate them on the same substrate to experimentally demonstrate the operations of the proposed NC block. Conductance modulation in the flash memory enables the NC block to be easily calibrated for output signals. In addition, the proposed NC block uses a reduced number of devices for analog-to-digital conversions due to the super-steep switching characteristics of the PF neuron device, substantially reducing the area overhead of NC block. Our NC block shows high energy efficiency (37.9 TOPS/W) with high accuracy for CIFAR-10 image classification (91.80%), outperforming prior works. This work shows the high engineering potential of integrating synapses and neurons in terms of system efficiency and high performance.

A ferroelectric thin‐film transistor (FeTFT)‐based synaptic device with an indium–gallium–zinc oxide (IGZO) channel and a metal–ferroelectric–metal–insulator–semiconductor (MFMIS) structure is reported. The fabricated FeTFT exhibits a highly linear conductance response (| α | = 0.21) with a large dynamic range ( G max / G min ≈ 53.2), although identical program pulses are applied to the device. In addition, because the inner metal layer of the FeTFTs has an MFMIS structure, the electric field is uniformly applied to the entire IGZO channel, which reduces the cycle‐to‐cycle variation ( σ = 0.47%) in the conductance responses. In the system simulation with the measured synaptic characteristics, the high classification accuracy of ≈97.0% is achieved in the MNIST image set, verifying the feasibility of FeTFT‐based neuromorphic systems.

Abstract The pursuit of sub‐1‐nm field‐effect transistor (FET) channels within 3D semiconducting crystals faces challenges due to diminished gate electrostatics and increased charge carrier scattering. 2D semiconductors, exemplified by transition metal dichalcogenides, provide a promising alternative. However, the non‐idealities, such as excess low‐frequency noise (LFN) in 2D FETs, present substantial hurdles to their realization and commercialization. In this study, ideal LFN characteristics in monolayer MoS 2 FETs are attained by engineering the metal‐2D semiconductor contact and the subgap density of states (DOS). By probing non‐ideal contact resistance effects using CuS and Au electrodes, it is uncovered that excess contact noise in the high drain current ( I D ) region can be substantially reduced by forming a van der Waals junction with CuS electrodes. Furthermore, thermal annealing effectively mitigates sulfur vacancy‐induced subgap density of states (DOS), diminishing excess noise in the low I D region. Through meticulous optimization of metal‐2D semiconductor contacts and subgap DOS, alignment of 1/ f noise with the pure carrier number fluctuation model is achieved, ultimately achieving the sought‐after ideal LFN behavior in monolayer MoS 2 FETs. This study underscores the necessity of refining excess noise, heralding improved performance and reliability of 2D electronic devices.

The crossbar structure of the nonvolatile memory enables highly parallel and energy-efficient analog matrix-vector-multiply (MVM) operations. To exploit its efficiency, existing works design a mixed-signal deep neural network (DNN) accelerator, which offloads low-precision MVM operations to the memory array. However, they fail to accurately and efficiently support the low-precision networks due to their naive ADC designs. In addition, they cannot be applied to the latest technology nodes due to their premature RRAM-based memory array.In this work, we present 3D-FPIM, an energy-efficient and robust mixed-signal DNN acceleration system. 3D-FPIM is a full-stack 3D NAND flash-based architecture to accurately deploy low-precision networks. We design the hardware stack by carefully architecting a specialized analog-to-digital conversion method and utilizing the three-dimensional structure to achieve high accuracy, energy efficiency, and robustness. To accurately and efficiently deploy the networks, we provide a DNN retraining framework and a customized compiler. For evaluation, we implement an industry-validated circuit-level simulator. The result shows that 3D-FPIM achieves an average of 2.09x higher performance per area and 13.18x higher energy efficiency compared to the baseline 2D RRAM-based accelerator.

Ferroelectric (FE) materials are key to advancing electronic devices owing to their non-volatile properties, rapid state-switching abilities, and low-energy consumption. FE-based devices are used in logic circuits, memory-storage devices, sensors, and in-memory computing. However, the primary challenge in advancing the practical applications of FE-based memory is its reliability. To address this problem, a novel polarization pruning (PP) method is proposed. The PP is designed to eliminate weakly polarized domains by applying an opposite-sign pulse immediately after a program or erase operation. Significant improvements in the reliability of ferroelectric devices are achieved by reducing the depolarization caused by weakly polarized domains and mitigating the fluctuations in the ferroelectric dipole. These enhancements include a 25% improvement in retention, a 50% reduction in read noise, a 45% decrease in threshold voltage variation, and a 72% improvement in linearity. The proposed PP method significantly improves the memory storage efficiency and performance of neuromorphic systems.

Concerns about indoor and outdoor air quality, industrial gas leaks, and medical diagnostics are driving the demand for high-performance gas sensors. Owing to their structural variety and large surface area, reducible metal oxides hold great promise for constructing a gas-sensing system. While many earlier reports have successfully obtained a sufficient response to various types of target gases, the selective detection of target gases remains challenging. In this work, a novel method, low-frequency noise (LFN) spectroscopy is presented, to achieve selective detection using a single FET-type gas sensor. The LFN of the sensor is accurately modeled by considering the charge fluctuation in both the sensing material and the FET channel. Exposure to different target gases produces distinct corner frequencies of the power spectral density that can be used to achieve selective detection. In addition, a 3D vertical-NAND flash array is used with the fast Fourier transform method via in-memory-computing, significantly improving the area and power efficiency rate. The proposed system provides a novel and efficient method capable of selectively detecting a target gas using in-memory-computed LFN spectroscopy and thus paving the way for the further development in gas sensing systems.

Artificial neurons and synapses are crucial for efficiently implementing spiking neural networks (SNNs) in hardware. The distinct functional requirements of artificial neurons and synapses present significant challenges in the implementation of area- and energy-efficient SNNs. This study reports an all-ferroelectric SNN system through co-optimization of material properties and device configurations using wafer-scale atomic layer deposition. For the first time, a double-gate (DG) morphotropic phase boundary-based thin-film transistor (MPBTFT) is utilized for a leaky integrate-and-fire (LIF) neuron. The DG MPBTFT-based LIF neuron eliminates the need for capacitors and reset circuits, thereby enhancing area and energy efficiency. The DG configuration demonstrates various neuronal functions with high reliability. Co-optimizing materials and devices significantly enhance the performance and functional versatility of artificial neurons and synapses. Meticulous material engineering facilitates the seamless co-integration of DG MPBTFT-based neurons, ferroelectric thin-film transistor (TFT)-based synapses, and normal TFTs on a single wafer. All-ferroelectric SNN systems achieved a high classification accuracy of 94.9%, thereby highlighting the potential of DG MPBTFT-based LIF neurons for advanced neuromorphic computing.

We propose an accurate and effective method, low-frequency noise (LFN) spectroscopy, to examine the resistive switching mechanism in ferroelectric tunnel junctions (FTJs) based on pure hafnium oxide (HfOx). Contrary to previous studies that primarily focused on the ferroelectric (FE) resistive switching (RS) in HfOx-based FTJs, the results of this study demonstrate that non-FE RS affected by the redistribution of oxygen vacancies also plays a significant role in determining the performance of FTJs. LFN spectroscopy is conducted in different conditions by changing the operating temperature and inducing DC cycling stress. The results reveal that the RS mechanism changes from FE to non-FE RS with increased program bias in all conditions. This change is facilitated by the rise in temperature and the number of DC cycling stress.

Reinforcement learning (RL), exhibiting outstanding performance in various fields, requires large amounts of data for high performance. While exploration techniques address this requirement, conventional exploration methods have limitations: complexity of hardware implementation and significant hardware burden. Herein, in‐memory RL systems leveraging intrinsic 1/ f noise of synaptic ferroelectric field‐effect‐transistors (FeFETs) for efficient exploration are proposed. The electrical characteristics of fabricated FeFETs with low‐power operation capability verify their suitability for neuromorphic systems. The proposed system achieves comparable performance to the conventional exploration method without additional circuits. The intrinsic 1/ f noise of the FeFETs facilitates efficient exploration and offers significant advantages: efficiency in hardware implementation and simplicity in adjusting the 1/ f noise level for optimal performance. This approach effectively addresses the challenges of conventional exploration methods. The operation mechanism of the exploration method utilizing the 1/ f noise is systematically analyzed. The proposed in‐memory RL system demonstrates robustness and reliability to the device‐to‐device variation and the initial conductance distribution. This work provides further insights into the exploration methods of RL, paving the way for advanced in‐memory RL systems.

An AND-type flash synaptic array is cointegrated with CMOS circuits using a novel fabrication method. Electrical characteristics of the basic circuit blocks required for neural network operation are verified. By reducing the number of masks and fabrication steps required, the proposed fabrication method successfully integrates synaptic array and CMOS peripheral circuits, including integrate-and-fire (I&F) circuits and passive devices, on a single wafer. The proposed fabrication method provides a methodology for the efficient implementation of hardware-based neural networks as well as verification of excellent compatibility of the proposed synaptic array with CMOS technology.

We investigate the low-frequency noise characteristics of indium–gallium–zinc oxide ferroelectric thin-film transistors (FeTFTs) with a metal–ferroelectric–metal–insulator–semiconductor (MFMIS) structure. MFMIS FeTFTs are fabricated with different metal-to-FE area ratios (AM/AF's). It is revealed that the noise generation mechanism differs depending on the operation region [low and high drain current (ID) regions] and AM/AF. Excess noise in the low ID region is observed in the MFMIS FeTFTs with AM/AF's of 4 and 6 due to carrier mobility fluctuations. In the high ID region, the carrier number fluctuation generates the 1/f noise of the devices regardless of the AM/AF.

Recently, Nature-based Solutions (NbS) have increasingly been regarded as a new opportunity to maximize the synergies between nature, society, and the economy. In addition, especially for policymakers and practitioners engaged in climate technology transfer activities from developed to developing countries, this concept is promoted as a cost-effective, agile, and innovative way of tackling various climate challenges to achieve sustainable development goals (SDGs). Thus, in the present work, to enhance NbS as an innovative implement during the climate technology transfer, we first analyze previous NbS cases during the technical assistance activities for some SDGs accomplished by the United Nations Climate Technology Centre and Network (CTCN), such as coastal risk protection (to maximize ecosystems, Type 1), agroforestry (to restore ecosystems, Type 2) and green urban design (to create ecosystems, Type 3). Then, through in-depth interviews with NbS stakeholders, we identify dominant barriers to implementing each NbS Type in terms of innovation element: technology, market, and regulation. Finally, based on our staged innovation model considering the two-sided networks, we propose novel strategy for enhancing NbS by overcoming each barrier during the three stages of the climate technology transfer process: NbS technology assessment in the first eco-maximizing stage, blended finances for market creation in the second eco-restoring stage, and regulation incentivization in the third eco-creating stage.

Research on the efficient utilisation of electric arc furnace oxidising slag, a by-product of the steel industry, is necessary as it is similar to natural aggregates in terms of chemical composition. The slag is also being produced in increasing amounts every year. This study performed flexural experiments on simply support beams with aggregate type and examined their concrete compressive strength, tension reinforcement ratio and compressive reinforcement ratio as variables to assess the effects of compressive stress distribution of concrete with electric arc furnace oxidising slag aggregates on the flexural behaviour of reinforced concrete beams. Compared to specimens with natural aggregates, electric arc furnace oxidising slag aggregates showed a smaller deflection before the yielding of tension reinforcement and ductile behaviour after yielding due to differences in compressive stress distribution. In addition, the proposed equations for the equivalent rectangular stress block parameter were found to provide more accurate predictions than the criterion used by other countries.

Neuromorphic hardware using nonvolatile analog synaptic devices provides promising advantages of reducing energy and time consumption for performing large-scale vector-matrix multiplication (VMM) operations. However, the reported training methods for neuromorphic hardware have appreciably shown reduced accuracy due to the nonideal nature of analog devices, and use conductance tuning protocols that require substantial cost for training. Here, we propose a novel hybrid training method that efficiently trains the neuromorphic hardware using nonvolatile analog memory cells, and experimentally demonstrate the high performance of the method using the fabricated hardware. Our training method does not rely on the conductance tuning protocol to reflect weight updates to analog synaptic devices, which significantly reduces online training costs. When the proposed method is applied, the accuracy of the hardware-based neural network approaches to that of the software-based neural network after only one-epoch training, even if the fabricated synaptic array is trained for only the first synaptic layer. Also, the proposed hybrid training method can be efficiently applied to low-power neuromorphic hardware, including various types of synaptic devices whose weight update characteristics are extremely nonlinear. This successful demonstration of the proposed method in the fabricated hardware shows that neuromorphic hardware using nonvolatile analog memory cells becomes a more promising platform for future artificial intelligence.