Japanese Journal of Applied Physics

Anomalously large heat generation phenomena that cannot be explained by any known chemical processes were discovered: Ni-based nano-structured multilayer metal composites were preloaded with hydrogen gas and heated rapidly to diffuse hydrogen and trigger the heat generation reaction. Maximum energy released per total hydrogen absorption was over 10 keV H^–1 and no gamma rays or neutrons, which are harmful to the human body, were observed. It is possible to intentionally induce the heat burst phenomenon, which can increase the amount of heat generated without any new energy input. This can be applied to reaction control as well as to improving the accuracy of heat generation evaluation. A common feature, that regions of very high oxygen concentrations are observed in places, was observed in the heat-producing samples. At this time, however, the discussion of this oxygen concentration as nuclear in origin must exclude the possibility of many chemical processes.

High-NA extreme ultraviolet (EUV) lithography is currently in development. Fabrication of exposure tools and optics with a numerical aperture (NA) equal to 0.55 has started at ASML and Carl Zeiss. Lenses with such high NA will have very small depths-of-focus, which will require improved focus systems and significant improvements in wafer flatness during processing. Lenses are anamorphic to address mask 3D issues, which results in wafer field sizes of 26 mm × 16.5 mm, half that of lower NA EUV tools and optical scanners. Production of large die will require stitching. Computational infrastructure is being created to support high-NA lithography, including simulators that use Tatian polynomials to characterize the aberrations of lenses with central obscurations. High resolution resists that meet the line-edge roughness and defect requirements for high-volume manufacturing also need to be developed. High power light sources will also be needed to limit photon shot noise.

Power semiconductor devices are key components in power conversion systems. Silicon carbide (SiC) has received increasing attention as a wide-bandgap semiconductor suitable for high-voltage and low-loss power devices. Through recent progress in the crystal growth and process technology of SiC, the production of medium-voltage (600–1700 V) SiC Schottky barrier diodes (SBDs) and power metal–oxide–semiconductor field-effect transistors (MOSFETs) has started. However, basic understanding of the material properties, defect electronics, and the reliability of SiC devices is still poor. In this review paper, the features and present status of SiC power devices are briefly described. Then, several important aspects of the material science and device physics of SiC, such as impurity doping, extended and point defects, and the impact of such defects on device performance and reliability, are reviewed. Fundamental issues regarding SiC SBDs and power MOSFETs are also discussed.

Understanding the fundamental relationships between physics and its information-processing capability has been an active research topic for many years. Physical reservoir computing is a recently introduced framework that allows one to exploit the complex dynamics of physical systems as information-processing devices. This framework is particularly suited for edge computing devices, in which information processing is incorporated at the edge (e.g. into sensors) in a decentralized manner to reduce the adaptation delay caused by data transmission overhead. This paper aims to illustrate the potentials of the framework using examples from soft robotics and to provide a concise overview focusing on the basic motivations for introducing it, which stem from a number of fields, including machine learning, nonlinear dynamical systems, biological science, materials science, and physics.

Breakdown voltage (BV) is arguably one of the most critical parameters for power devices. While avalanche breakdown is prevailing in silicon and silicon carbide devices, it is lacking in many wide bandgap (WBG) and ultra-wide bandgap (UWBG) devices, such as the gallium nitride high electron mobility transistor and existing UWBG devices, due to the deployment of junction-less device structures or the inherent material challenges of forming p-n junctions. This paper starts with a survey of avalanche and non-avalanche breakdown mechanisms in WBG and UWBG devices, followed by the distinction between the static and dynamic BV. Various BV characterization methods, including the static and pulse I–V sweep, unclamped and clamped inductive switching, as well as continuous overvoltage switching, are comparatively introduced. The device physics behind the time- and frequency-dependent BV as well as the enabling device structures for avalanche breakdown are also discussed. The paper concludes by identifying research gaps for understanding the breakdown of WBG and UWBG power devices.

Benefitted from progress on the large-diameter Ga₂O₃ wafers and Ga₂O₃ processing techniques, the Ga₂O₃ power device technology has witnessed fast advances toward power electronics applications. Recently, reports on large-area (ampere-class) Ga₂O₃ power devices have emerged globally, and the scope of these works have gone well beyond the bare-die device demonstration into the device packaging, circuit testing, and ruggedness evaluation. These results have placed Ga₂O₃ in a unique position as the only ultra-wide bandgap semiconductor reaching these indispensable milestones for power device development. This paper presents a timely review on the state-of-the-art of the ampere-class Ga₂O₃ power devices (current up to >100 A and voltage up to >2000 V), including their static electrical performance, switching characteristics, packaging and thermal management, and the overcurrent/overvoltage ruggedness and reliability. Exciting research opportunities and critical technological gaps are also discussed.

The development of a high-power EUV light source is very important in EUV lithography to overcome the stochastic effects for higher throughput and higher numerical aperture (NA) in the future. We have designed and studied a high-power EUV free-electron laser (FEL) based on energy-recovery linac (ERL) for future lithography. We show that the EUV-FEL light source has many advantages, such as extremely high EUV power without tin debris, upgradability to a Beyond EUV (BEUV) FEL, polarization controllability for high-NA lithography, low electricity consumption, and low construction and running costs per scanner, as compared to the laser-produced plasma source used for the present EUV lithography exposure tool. Furthermore, the demonstration of proof of concept (PoC) of the EUV-FEL is in progress using the IR-FEL in the Compact ERL (cERL) at the High Energy Accelerator Research Organization. In this paper, we present the EUV-FEL light source for future lithography and progress in the PoC of the EUV-FEL.

InGaN-based LEDs are efficient light sources in the blue–green light range and have been successfully commercialized in the last decades. Extending their spectral range to the red region causes a significant reduction in LED efficiency. This challenge hinders the integration of red, green, and blue LEDs based on III-nitride materials, especially for full-color micro-LED displays. We review our recent progress on InGaN-based red LEDs with different chip sizes from hundreds to tens of micrometers, including the epitaxial structures, device fabrication, and optical performance (peak wavelength, full-width at half-maximum, light output power, efficiency, temperature stability, and color coordinates).

Energy harvesting technology is attracting attention as "enabling technology" that expands the use and opportunities of IoT utilization, enriches lives and enhances social resilience. This technology harvests energy that dissipates around us, in the form of electromagnetic waves, heat, vibration, etc. and converts it into easy-to-use electric energy. This paper describes the features of these technologies, recent topics and major challenges, and boldly predicts the future prospects of the development.

Two metal-induced lateral crystallization (MILC) methods are proposed as candidate techniques to enhance cell current in future ultra-high-density NAND-type 3D flash memory devices. The channel crystallinity differs depending on the MILC method. In a single MILC, the channel is composed of single-crystal Si, whereas in a regional MILC, the channel comprises multiple crystal grains that are larger than those of the conventional polycrystalline Si. Using transmission electron microscopy, the inhibiting factor of MILC was modeled to reveal that the two MILC approaches result in different cell current distributions that are related to their degree of crystallinity. A comparison of these two cell current distributions in a 3D flash memory with over 900 word-line stacks showed that the single MILC delivers a higher median cell current with outliers on the lower side. In contrast, the regional MILC delivers a lower median cell current without outliers on the lower side.

This study focuses on enhancing the bendability of flexible interconnects with out-of-plane corrugation for flexible hybrid electronics. We propose two typical configurations of 3D corrugated interconnects: serpentine and trapezoidal. Three methods are introduced to fabricate these corrugated interconnects. The advantages and drawbacks of each fabrication strategy are discussed, and the impact of the 3D corrugation geometry and material on bendability is elucidated. In addition, the material properties of two types of negative photosensitive materials, SU-8 and F-PD (flexible-photoimageable dielectric), are compared. Results show that the resistance increase of 3D corrugated interconnects after a 5 mm radius bending test is drastically lower (by approximately 1900%–2000%) than that of conventional 2D planar interconnects.

A 100 nm wide superconducting niobium (Nb) interconnect was fabricated by a 300 mm wafer process for Cryo-CMOS and superconducting digital logic applications. A low pressure and long throw sputtering was adopted for the Nb deposition, resulting in good superconductivity of the 50 nm thick Nb film with a critical temperature (T_c) of 8.3 K. The interconnects had a titanium nitride (TiN)/Nb stack structure, and a double-layer hard mask was used for the dry etching process. The exposed area of Nb film was minimized to decrease the effects of plasma damage during fabrication and atmosphere. The developed 100 nm wide and 50 nm thick Nb interconnect showed good superconductivity with a T_c of 7.8 K and a critical current of 3.2 mA at 4.2 K. These results are promising for Cryo-CMOS and superconducting digital logic applications in the 4 K stage.

Photomixing is one of the promising THz-wave generation methods. A conventional photomixing system with two lasers results in increasing the volume and power consumption of the system. For downsizing the system and reducing its power consumption, we have previously proposed a method to generate a THz-wave by a single wavelength-tunable laser. This technique can generate a quasi-continuous THz wave as well as a pulsed THz wave, based on which we are studying toward pulse frequency-amplitude modulation as a multi-level modulation method. In this paper, we have successfully generated a THz wave with frequency modulation between 286 GHz and 322 GHz by a single wavelength-tunable laser.

In semiconductor processes, precise control of the wafer-in-process is a key parameter closely related to production yield, and the development of electrostatic chuck (ESC) continues towards higher chucking voltage with higher backside cooling gas. This study aims to determine the target temperature and uniformity of the wafer surface by varying the contact ratio of the ceramic-embossing facing the wafer-in-process. A computational fluid dynamics model with a thin wall boundary condition is considered to interpret the flow of the rarefied gas between the wafer and ceramic surface of the ESC. Through 3D simulations conducted with ANSYS Fluent, we observed temperature changes as the backside gas pressure varied from 1 to 9 Torr. The ESC with the highest contact ratio performed exceptionally well with an average temperature of 295 K and a coefficient of variation of 0.04%.

We are on the eve of the next big step in lithography technology with the introduction of high numerical aperture EUV. The change from NA 0.33–0.55 in EUV lithography is an increase of 67%, which is the largest jump in the last decades, and puts tight requirements on focus and edge placement. Moreover, the lithography system has changed from fully isomorphic, i.e. same demagnification in all directions, to an anamorphic system, i.e. the demagnification in scan direction has doubled with respect to the slit direction. At imec we are fostering the ecosystem surrounding the lithography tool. In this paper we focus on the imaging and mask innovations supporting the EUV ecosystem, which are categorized into four areas: novel absorber masks, stitching, mask variability, and innovative imaging solutions. The current drivers of IC manufacturers implementing (high NA) EUV lithography (EUVL) are reduction of the EUV exposure dose and decrease in wafer stochastics. We discuss how these four areas have the potential to deliver in EUVL an increase in productivity, an improvement in the process window and a reduction in stochasticity at wafer level.

A quartz crystal microbalance (QCM) sensor can detect various physical and chemical properties, including biomolecules, gases, external forces, and so on, through changes in its resonance frequency. Because of the extremely high temperature stability of the resonance frequency, no thermostatic device is required, making the entire system compact. The sensitivity is governed by the thinness of the quartz resonator, and a wireless-electrodeless approach has achieved much thinner resonators. This review introduces recent advances in wireless-electrodeless QCM sensors for studying real-time biomolecules and target-gas detection.

High-speed atomic force microscopy (HS-AFM) is a technique that enables real-time imaging of nanoscale phenomena in solution. It was originally developed to visualize biomolecules, whose dynamics in solution significantly affect the manifestation of their functions, and has contributed to the understanding of molecular mechanisms based on the observation of single-molecule dynamics of proteins. In recent years, its application has broadened to include not only biomolecules, but also the structural dynamics of supramolecular assemblies that associate and dissociate in solution, as well as the evaluation of synthetic molecules such as polymer gels that swell in solution. In this paper, we review some of our recent studies on the application of HS-AFM to supramolecular polymers and hydrogel particles.

The biological membrane is a dynamic supramolecular architecture that plays vital roles in the cell. However, understanding the physicochemical properties and functions of the membrane supramolecular system is difficult. We have developed an integrated model system of the biological membrane comprising patterned polymeric and natural lipid bilayers. The polymeric bilayer acts as a framework to support embedded natural membranes. The embedded natural membranes retain important characteristics of the biological membrane such as fluidity, and reproduces the physical states and functions of the biological membrane. Membrane proteins can be reconstituted into the model membrane for analyzing their functions in a controlled lipid membrane environment. Three-dimensional structures can be constructed by attaching micro-/nano-fabricated structures to the polymeric bilayer framework. The integrated model membrane realizes a versatile platform to study membrane functions, and should open new opportunities in fundamental biological sciences as well as biomedical/analytical applications.

Since the last decade, research on plasma catalysis has attracted keen attention as an emerging type of low-carbon technology. An advantage of plasma is to facilitate non-equilibrium reaction fields on a large scale, which is inaccessible by conventional thermal approaches. Stable molecules such as CO₂ and CH₄ are activated by electrical energy, paving the way for low-temperature chemistry that departs from energy-intensive heat-dependent systems. Moreover, the power-to-chemical concept could gain momentum with plasma technologies that are driven by renewable energy. Currently, research is accelerating with application initiatives, but at the same time the importance of scientific understanding of plasma catalytic reactions is being recognized more than ever. This review article offers an overview of various plasma technologies in the "plasma alone" and "plasma–catalyst combination" context. Plasma–catalyst combination technology, known as "plasma catalysis", is discussed further to dry methane reforming (CH₄ + CO₂ = 2CO + 2H₂) and the reverse water gas shift reaction (CO₂ + H₂ = CO + H₂O) for a mechanistic insight.

In this work, a series of technology computer-aided design (TCAD) device simulations have been carried out to investigate the effects of gate underlap and overlap structures on device performances of vertical-channel MOSFET. The device characterizations were conducted in both aspects of DC and high-frequency operations for higher completeness of this work since both are not usually optimized at the same time under the same structural and processing conditions. Under the underlap condition, slight degradation in the on-state current Ion drivability was observed. A noticeable off-state current Ioff increases were witnessed under the underlap conduction. It is explicitly demonstrated that excessive gate underlap results in non-ideal effects including degradation of subthreshold swing (S), worsening of drain-induced barrier lowering (DIBL), lowering of maximum transconductance (gm,Max). Although fT and fmax were sustained to be high under overlap and gate-drain alignment conditions,

This paper reports on a fabrication process suitable for ultra-low resonant frequency inertial MEMS sensors. The low resonant frequency is achieved by electrically tunable springs and a heavy mass formed by through-silicon deep reactive-ion etching (DRIE) applied to a Silicon-on-Glass (SOG). A thermal issue of through-silicon DRIE (TSD) stemming from the low-resonant-frequency structure is circumvented by two methods: introducing cooling time between the DRIE steps, and adopting a metal hard mask. A blade dicing method suited for this process is also presented. To monitor the verticality of TSD, a non-destructive taper detection method that utilizes a capacitance-voltage (CV) curve is proposed and verified.

In this study, we used femtosecond laser-assisted etching (FLAE) to drill through glass vias (TGVs) in 0.3 mm thick non-alkali glass substrates. &#xD;In FLAE, the focus of the femtosecond laser pulses is scanned to modify the material along a preprogrammed pattern, and the modified region is preferentially removed by chemical etching. &#xD;We found that the scanning strategy affected the etching rate along the laser-modified lines. &#xD;Among four types of scanning strategies tested, the strategy <du>—that is, scanning in a downward direction followed by an upward direction—obtained the highest etching rate. &#xD;In this case, the etching rate along the laser-modified line was approximately 10 times larger than that of the unmodified region.

A ScAlN thin film is one of the key materials of MEMS and high-frequency filters used in new-generation communication devices. Piezoelectricity can be improved by increasing Sc concentration. However, abnormal grains often appear at high Sc concentrations, degrading crystallinity and piezoelectricity. Herein, we demonstrated that underlayer roughness considerably affects the emergence of abnormal grains in a Sc0.4Al0.6N thin film formed via reactive DC sputtering. Dry etching with Ar plasma can effectively reduce the surface roughness of amorphous SiN and polycrystalline Si. Sc0.4Al0.6N thin films deposited on amorphous SiN and polycrystalline Si with sufficient flat surfaces exhibited a low density of abnormal grains, high crystallinity and piezoelectricity, and low loss tangent. Moreover, such high-quality thin films were obtained on a borophosphosilicate glass flattened using a reflow process without Ar etching. Therefore, underlayer roughness played an important role. The findings can help enable the large-scale production of highly doped ScAlN thin films.

In this work the integration of ferroelectric (FE) devices for advanced in-memory computing applications is demonstrated based on the FeMFET memory cell concept. In contrast to FeFET having the FE layer directly embedded in the gate-stack, the FeMFET consists of a separated ferroelectric capacitor which can be integrated in the chip-interconnect layers.&#xD;Optimization of the FE material stack under such lower thermal budget constraints will be discussed as well as the significant performance improvement and reduction of variability by application of superlattice FE-stacks and further optimization knobs. The low memory state variability is important for accurate multiply-accumulate (MAC) operation. Such improvements are demonstrated on a memory array test chip including functional verification of MAC operation along a FeMFET-based array column with good accuracy over high dynamic current range. &#xD;

We investigated the effectiveness of poling processes for Pb(Mg_1/3Nb_2/3)O₃-PbTiO₃ single crystals (SCs) produced using a continuous feeding Bridgman method, which is known to produce a high property uniformity. The four studied poling processes are: (I) standard direct current poling (STD-DCP); (II) low-voltage field-cooling DCP (LV-FCP); (III) high-voltage field-cooling DCP (HV-FCP); and (IV) mid-temperature alternating current poling (MT-ACP). The highest free dielectric constant (ε₃₃^T/ε₀) and piezoelectric constant (d₃₃) were obtained by MT-ACP (ε₃₃^T/ε₀ = 11 000, d₃₃ = 3000 pC/N), followed by LV-FCP (ε₃₃^T/ε₀ = 7500, d₃₃ = 2400 pC/N), HV-FCP (ε₃₃^T/ε₀ = 6250, d₃₃ = 1850 pC/N), and STD-DCP (ε₃₃^T/ε₀ = 6200, d₃₃ = 1800 pC/N). The LV-FCP SC showed a 21% and 33% increase in ε₃₃^T/ε₀ and d₃₃ compared to that of the STD-DCP SC; however, this was not as much as the 77% and 67% improvement of the MT-ACP SC. These results provide guidance for SC transducers.

This study applies Bayesian super-resolution to X-ray photoelectron spectroscopy (XPS), achieving up to a 20-fold reduction in measurement time while preserving data quality. Traditional XPS, crucial for surface analysis, typically requires extensive measurement durations. Our methodology significantly accelerates the process, as demonstrated with glass and Polytetrafluoroethylene samples, where we reduced measurement times by up to 1/20th without compromising spectral accuracy. This approach decreases noise levels and retains spectral integrity, offering a highly efficient solution for XPS. This innovation is particularly valuable in material science, enabling rapid, reliable surface analysis.

The effective quality factor of the cantilever plays a fundamental role in dynamic mode atomic force microscopy. Here we present a technique to modify the quality factor of an atomic force microscopy cantilever within a Fabry-Perot optical interferometer. The experimental setup uses two separate laser sources to detect and excite the oscillation of the cantilever. While the intensity modulation of the excitation laser drives the oscillation of the cantilever, the average intensity can be used to modify the quality factor via optomechanical force without changing the fiber-cantilever cavity length. The technique enables users to optimize the quality factor for different types of measurements without influencing the deflection measurement sensitivity. An unexpected frequency shift was observed and modelled as temperature dependence of the cantilever's Young's modulus, which was validated using finite element simulation. The model was used to compensate for the thermal frequency shift. The simulation provided relations between optical power, temperature, and frequency shift.

In this study, we present a low thermal budget elevated-laser-liquid-phase-epitaxy technique designed for the precise fabrication of single-crystal islands (SCIs) intended for use in middle-end-of-line (MEOL) FinFETs. Each of these SCIs features a (100) orientation tended from Si seeding structure and is successfully integrated as channel materials in the MEOL circuit of a monolithic 3D IC (3DIC). This technique effectively mitigates the typical performance disparities associated with poly-Si channel materials in upper tiers, addressing a significant challenge in advanced electronic device fabrication and potentially enhancing the performance and reliability of MEOL FinFETs in monolithic 3DIC.

In photonic quantum applications, optical routers are required to handle single photons with low loss, high speed, and preservation of their quantum states. Single-photon routing with maintained polarization states is particularly important for utilizing them as qubits. Here, we demonstrate a polarization-maintaining electro-optic router compatible with single photons. Our custom electro-optic modulator is embedded in a configuration of a Mach–Zehnder interferometer, where each optical component achieves polarization-maintaining operation. We observe the performance of the router with 2%–4% loss, 20 dB switching extinction ratio, 2.9 ns rise time, and >99% polarization process fidelity to an ideal identity operation.

A 100 nm wide superconducting niobium (Nb) interconnect was fabricated by a 300 mm wafer process for Cryo-CMOS and superconducting digital logic applications. A low pressure and long throw sputtering was adopted for the Nb deposition, resulting in good superconductivity of the 50 nm thick Nb film with a critical temperature (T_c) of 8.3 K. The interconnects had a titanium nitride (TiN)/Nb stack structure, and a double-layer hard mask was used for the dry etching process. The exposed area of Nb film was minimized to decrease the effects of plasma damage during fabrication and atmosphere. The developed 100 nm wide and 50 nm thick Nb interconnect showed good superconductivity with a T_c of 7.8 K and a critical current of 3.2 mA at 4.2 K. These results are promising for Cryo-CMOS and superconducting digital logic applications in the 4 K stage.

Ge-on-Insulator (GOI) is considered to be a necessary structure for novel Ge-based devices. This paper proposes an alternative approach for fabricating GOI based on the Ge-on-Nothing (GeON) template. In this approach, a regular macropore array is formed by lithography and dry etching. These pores close and merge upon annealing, forming a suspended monocrystalline Ge membrane on one buried void. GOI is fabricated by direct bonding of GeON on Si carrier substrates, using an oxide bonding interface, and subsequent detachment. The fabricated GOI shows uniform physical properties as demonstrated using micro-photoluminescence measurements. Its electrical characteristics and cross-sectional structure are superior to those of Smart-Cut^TM GOI. To demonstrate its application potential, back-gate GOI capacitors and MOSFETs are fabricated. Their characteristics nicely agree with the theoretically calculated one and show typical MOSFET operations, respectively, which indicates promising Ge crystallinity. This method, therefore, shows the potential to provide high-quality GOI for advanced Ge application devices.

We developed a compact thermo-optic Mach–Zehnder interferometer switch with a direct heating heater using multimode interference and achieved a sufficiently low thermal crosstalk performance. Large-scale switch systems, such as optical neural networks, require thermo-optical switches with low power consumption, fast switching speed, compact size, and low thermal crosstalk. This switch is equipped with a heater that directly heats the Si core waveguide, which is a structure that connects non-doped Si wires between phase shifters and a heatsink. As a result, a significant miniaturization with a phase shifter length of approximately 7 μm, low π-phase shift power consumption of less than 20 mW, and fast switching in sub-microseconds were achieved. The improved phase shifter showed a very small figure of merit of 8.89 mWμs. Simultaneously, transmission spectrum measurements of nearby ring resonators show that the thermal crosstalk is significantly reduced even at a distance of only 30 μm. This device can contribute to the overall circuit performance and footprint reduction in large-scale optical integrated circuits and optical neural network configurations.

Two metal-induced lateral crystallization (MILC) methods are proposed as candidate techniques to enhance cell current in future ultra-high-density NAND-type 3D flash memory devices. The channel crystallinity differs depending on the MILC method. In a single MILC, the channel is composed of single-crystal Si, whereas in a regional MILC, the channel comprises multiple crystal grains that are larger than those of the conventional polycrystalline Si. Using transmission electron microscopy, the inhibiting factor of MILC was modeled to reveal that the two MILC approaches result in different cell current distributions that are related to their degree of crystallinity. A comparison of these two cell current distributions in a 3D flash memory with over 900 word-line stacks showed that the single MILC delivers a higher median cell current with outliers on the lower side. In contrast, the regional MILC delivers a lower median cell current without outliers on the lower side.

This paper comprehensively analyses dual integration of approximate random weight generator (ARWG) and computation-in-memory for event-based neuromorphic computing. ARWG can generate approximate random weights and perform multiply-accumulate (MAC) operation for reservoir computing (RC) and random weight spiking neural network (SNN). Because of using device variation to generate random weights, ARWG does not require any random number generators (RNGs). Because RC and random weight SNN allow approximate randomness, ARWG only needs to generate approximate random weights, which does not require error-correcting code to correct weights to make the randomness accurate. Moreover, ARWG has a read port for MAC operation. In this paper, the randomness of random weights generated by the proposed ARWG is evaluated by Hamming distance and Hamming weight. As a result, this paper reveals that the randomness required for ARWG is much lower than that for physically unclonable functions and RNGs, and thus the proposed ARWG achieves high recognition accuracy.

By modulating a ζ potential of graphene FET (G-EFT), the sensitivity of G-FET could be enhanced than that without modulation. Therefore, 1 × 10⁷ FFU ml⁻¹ SARS-CoV-2 was detected using G-FET modified with the ζ potential modulator which is the cation polymer with the positive charge. This method is based on the relationship between the surface charge and the sensitivity, in which the highest sensitivity is obtained when the ζ potential is 0 and/or the surface charge is almost 0. In this study, the microfluidic channel was installed on G-FET to get the precise result because it could wash away the free-floating virus and the physical adsorbed virus. 32 G-FETs including the reference FETs were integrated on the silicon substrate and the precise results were obtained by subtracting the noise terms.

This paper proposes a design methodology for a compact edge vision transformer (ViT) Computation-in-Memory (CiM). ViT has attracted much attention for its high inference accuracy. However, to achieve high inference accuracy, the conventional ViT requires fine-tuning many parameters with pre-trained models on large datasets and a large number of matrix multiplications in inference. Thus, to map ViT to non-volatile memory (NVM)-based CiM compactly for edge applications (IoT/Mobile devices) in inference, this paper analyses fine-tuning in training, clipping, and quantization in inference. The proposed compact edge ViT CiM can be optimized by three design methods according to use cases considering the required fine-tuning time, ease of setting memory bit precision, and memory error tolerance of ViT CiM. As a result, in CIFAR-10, the most compact type successfully reduces the total memory size of ViT by 85.8% compared with the conventional ViT. Furthermore, the high accuracy type and high error-tolerant type improve inference accuracy by 4.4% and memory-error tolerance by more than four times compared with convolutional neural networks, respectively.

Plasmonic color is a structural color generated via preferential light absorption and scattering in dielectric nanostructures. In this study, a large plasmonic color image was successfully fabricated by an electron beam lithography (EBL) system. A software program, referred to as P-color in this study, was developed to facilitate the conversion of a desired color bitmap image to a GDS file composed of multiple nano-patterns to realize plasmonic color. The relationship between the color, width, and pitch of the pattern structures was investigated under different area-dose conditions during EBL as basic data for plasmonic color image design. After establishing conversion techniques for both the large-capacity GDS and EBL files, a plasmonic color image sample with a size of 60 mm × 40 mm area (which is difficult to fabricate using a conventional point-type EBL system) was successfully fabricated.

Sc-based contacts to Si:P have shown great potential for NMOS devices. However, the promising properties of this material system are not yet fully understood. This work provides new insights into the crystallinity and composition of annealed TiN/Sc/Si:P stacks. After silicidation, two distinct phases are evidenced, with orthorhombic ScSi lying atop a thin Sc_1−x−ySi_xP_y interfacial layer that shares a commensurate interface with the underlaid Si:P, hypothetically resulting in a low interface defectivity. The formed ScSi phase is observed to be thermally stable between ∼450 °C and 700 °C, which is suitable for most device applications. The impact of additional thermal budgets within this temperature range is investigated, revealing potential origins for thermally induced degradation of the contact properties.

There is a growing demand for physical reservoirs that operate with low power consumption and low computational cost. We have conducted research on the basic properties of Ag₂S reservoirs, which are a type of physical reservoir. However, little research has been conducted on their applications. In this study, as a first step toward the practical application of Ag₂S reservoirs, we implemented two types of rock-paper-scissors judgment systems using Ag₂S reservoirs. In these experiments, we were able to demonstrate fast learning in the reservoir by comparing the results with methods using a single-layer perceptron and a convolutional neural network. In addition, we could obtain a maximum accuracy rate of about 98%.