Table of contents

Wide-bandgap semiconductors are expected to be applied to solid-state lighting and power devices, supporting a future energy-saving society. While GaN-based white LEDs have rapidly become widespread in the lighting industry, SiC- and GaN-based power devices have not yet achieved their popular use, like GaN-based white LEDs for lighting, despite having reached the practical phase. What are the issues to be addressed for such power devices? In addition, other wide-bandgap semiconductors such as diamond and oxides are attracting focusing interest due to their promising functions especially for power-device applications. There, however, should be many unknown phenomena and problems in their defect, surface, and interface properties, which must be addressed to fully exploit their functions. In this review, issues of wide-bandgap semiconductors to be addressed in their basic properties are examined toward their "full bloom".

Many important processes take place at solid/liquid interfaces. To understand these processes, in situ real-time evaluation of the geometric, electronic, and molecular structures at solid/liquid interfaces at the atomic and molecular levels is essential. Owing to the presence of the liquid, however, techniques such as electron microscopy and low-energy electron diffraction, which are powerful tools for surface structural analysis in vacuum, cannot be used for solid/liquid interfaces. In this review, various techniques applicable to solid/liquid interfaces, including scanning probe microscopy, synchrotron-radiation-based X-ray techniques, and nonlinear spectroscopy, are briefly described. The characterization of the electrodeposition process of Pd layers on Au single-crystal electrode surfaces is presented as an example to demonstrate the importance of using multiple techniques in an integrated manner to understand the processes at solid/liquid interfaces.

In this paper, surface-activation-based nanobonding technology and its applications are described. This bonding technology allows for the integration of electronic, photonic, fluidic and mechanical components into small form-factor systems for emerging sensing and imaging applications in health and environmental sciences. Here, we describe four different nanobonding techniques that have been used for the integration of various substrates — silicon, gallium arsenide, glass, and gold. We use these substrates to create electronic (silicon), photonic (silicon and gallium arsenide), microelectromechanical (glass and silicon), and fluidic (silicon and glass) components for biosensing and bioimaging systems being developed. Our nanobonding technologies provide void-free, strong, and nanometer scale bonding at room temperature or at low temperatures (<200 °C), and do not require chemicals, adhesives, or high external pressure. The interfaces of the nanobonded materials in ultra-high vacuum and in air correspond to covalent bonds, and hydrogen or hydroxyl bonds, respectively.

This paper focuses on the application of low temperature bonding to the fabrication of three-dimensional (3D) massively parallel signal processors for high performance infrared imagers. We review two generations of the 3D heterogeneous integration process. The first generation process, compatible with pixel sizes in the 20 to 30 µm range, relies on low temperature epoxy bonding that is followed by the formation of copper-filled through-silicon vias (TSVs). The second generation process, scalable to pixel sizes of 10 µm and smaller, employs solid–liquid diffusion bonding of copper–tin to copper at 250 °C; the bonding follows TSV fabrication. To demonstrate the second generation process, we fabricated 3D test vehicles in the form of 640 × 512 arrays of vertical interconnects composed of TSVs and metal–metal bonds on a 10 µm pitch. We characterized electrical conductivity of the interconnects, the isolation resistance between the interconnects, and the operability and yield of the arrays. The successful demonstration of the interconnect technology paves the way to a functional demonstration of 3D signal processors in infrared imagers with 10 µm pixels.

Three-dimensional (3D) integration requires vertical stacking of dies while forming permanent electrical and mechanical connections between the input/output pins of the devices. How to enable stacking thermal sensitive devices at low temperature gains interest. This paper presents a systematic study of Cu/Sn bonding at 150–200 °C, during which intermetallic compounds were formed by solid state inter-diffusion. It was found that below the lower-limit pressure of 20 MPa it was hard to make good contact between the rough joint surfaces and hence electrical connection was lost. However, beyond the upper-limit of 150 MPa Sn squeezed out leading to electrical shorting between adjacent bumps. Oxides removal was another key factor for good bonding. Finally, this Cu/Sn solid state diffusion bonding together with Cu through-silicon-via (TSV) was used for making die to die vertical interconnection. The measured resistance of single Cu/Sn solder joint and Cu TSV was in the range of 12–25 mΩ.

Room-temperature microjoining in air ambient has been achieved by using a cone-shaped Au microbump with ultrasonic application. In situ observation of ultrasonic bonding is performed using a high-speed camera to investigate the dynamics of the bonding process. "Softening" of the bump under ultrasonic application is observed. It is suggested that bonding is achieved within 50 ms.

Bump-shaped vertically aligned carbon nanotubes were fabricated as bump interconnect structures on flexible substrates for flexible multilayer substrates. These structures were bonded and transferred by the surface activated bonding method. In this paper, the fabrication process and mechanical properties of these structures are reported.

Twelve-channel vertical-cavity surface-emitting laser (12-ch VCSEL) chips are heterogeneously self-assembled on Si and glass wafers using water surface tension as a driving force. The VCSEL chips have a high length-to-width aspect ratio, that is, 3 mm long and 0.35 mm wide. The VCSEL chips are precisely self-assembled with alignment accuracies within 2 µm even when they are manually placed on liquid droplets provided on the host substrate. After the self-assembly of the VCSEL chips and the subsequent thermal compression, the chips successfully emit 850 nm light and exhibit no degradation of their current–voltage (I–V) characteristics.

Thermal management of high-power semiconductor lasers is of great importance since the output power and beam quality are affected by the temperature rise of the gain region. Thermal simulations of a vertical-external-cavity surface-emitting laser by a finite-element method showed that the solder layer between the semiconductor thin film consisting of the gain region and a heat sink has a strong influence on the thermal resistance and direct bonding is preferred to achieve effective heat dissipation. To realize thin-film semiconductor lasers directly bonded on a high-thermal-conductivity substrate, surface-activated bonding using an argon fast atom beam was applied to the bonding of gallium arsenide (GaAs) and silicon carbide (SiC) wafers. The GaAs/SiC structure was demonstrated in the wafer scale (2 in. in diameter) at room temperature. The cross-sectional transmission electron microscopy observations showed that void-free bonding interfaces were achieved.

Current-injected light emission was confirmed for metal organic vapor phase epitaxy (MOVPE) grown (Ga)InAs/InP quantum dots (QDs) on directly bonded InP/Si substrate. The InP/Si substrate was prepared by directly bonding of InP thin film and a Si substrate using a wet-etching and annealing process. A p–i–n LED structure including Stranski–Krastanov (Ga)InAs/InP QDs was grown by MOVPE on an InP/Si substrate. No debonding between Si substrate and InP layer was observed, even after MOVPE growth and operation of the device under continuous wave conditions at RT. The photoluminescence, current/voltage, and electroluminescence characteristics of the device grown on the InP/Si substrate were compared with reference grown on an InP substrate.

A method to integrate III–V compound semiconductor and SOI-CMOS on a common Si substrate is demonstrated. The SOI-CMOS layer is temporarily bonded on a Si handle wafer. Another III–V/Si substrate is then bonded to the SOI-CMOS containing handle wafer. Finally, the handle wafer is released to realize the SOI-CMOS on III–V/Si hybrid structure on a common substrate. Through this method, high temperature III–V materials growth can be completed without the presence of the temperature sensitive CMOS layer, hence damage to the CMOS layer is avoided.

We fabricated p⁺-, p-, and p⁻-Si/n-4H-SiC junctions by surface activated bonding (SAB). We investigated their electrical properties by measuring their current–voltage (I–V) characteristics at raised ambient temperatures, capacitance–voltage (C–V) characteristics at various frequencies, and capacitance–frequency (C–f) characteristics at room temperature. The activation energy of their reverse-bias current and the flat-band voltage in their C–V characteristics, which were estimated to be 0.97–1.01 eV and 0.83–0.84 V, respectively, were insensitive to the concentrations of acceptors in Si substrates. The relaxation times estimated from the C–f characteristics were 0.8 and 1.5 µs for the p-Si/n-4H-SiC and p⁻-Si/n-4H-SiC junctions, respectively. The results are explained by a scheme wherein Fermi level pinning occurs at the Si/4H-SiC interfaces fabricated by SAB.

The electrical properties of p-GaAs/n⁺-Si, p⁺-Si/n-GaAs, p⁺-GaAs/n⁺-Si, p⁺-Si/n⁺-GaAs, n⁺-Si/n⁺-GaAs, and p⁺-Si/p⁺-GaAs junctions fabricated by surface-activated bonding (SAB) were investigated. An amorphous layer with a thickness of 3 nm was found across the bonding interface without annealing. The current–voltage (I–V) characteristics of p⁺-GaAs/n⁺-Si, p⁺-Si/n⁺-GaAs, n⁺-Si/n⁺-GaAs, and p⁺-Si/p⁺-GaAs junctions showed excellent linearity. The interface resistance of n⁺-Si/n⁺-GaAs junctions was found to be 0.112 Ω·cm², which is the smallest value observed in all the samples. The resistance decreased with increasing annealing temperature and decreased to 0.074 Ω·cm² after the junction annealing at 400 °C. These results demonstrate that n⁺-Si/n⁺-GaAs junctions are suitable for the connection of subcells in the fabrication of tandem solar cells.

Effects of annealing on surface-activated bonding (SAB)-based Si/Si junctions were investigated by transmission electron microscopy (TEM) observations and current–voltage (I–V) measurements. We observed an amorphous-like layer at the bonding interface, which was recrystallized by annealing. We extracted the potential barrier heights at Si/Si interfaces annealed at different temperatures from the results of I–V measurements at various ambient temperatures. For p-Si/p-Si junctions, the barrier height increased as the annealing temperature increased from 200 to 400 °C and decreased from 400 to 1000 °C. For n-Si/n-Si junctions, the barrier height increased as the annealing temperature increased from 200 to 600 °C and decreased from 600 to 1000 °C. By using the charge neutral level (CNL) model, we estimated the energy of CNL, E_CNL, and the density of interface states, D_it, at each annealing temperature. D_it decreased as the annealing temperature increased from 400 to 1000 °C. E_CNL showed values larger than the reported ones.

This paper reports the mechanical and electrical characteristics of Ge/Ge interfaces prepared by room-temperature surface-activated bonding (SAB). Bonded Ge/Ge wafer pairs with high bonding strength equivalent to that of the bulk material were achieved without any heat treatment. It was found that the bonding of Ge wafers was not sensitive to the background vacuum pressure in a wafer-bonding chamber compared with the bonding of Si wafers. The current–voltage characteristics and microstructures of bonded interfaces formed by SAB and low-temperature plasma activation bonding (PAB) were compared. It was demonstrated that junctions with very low resistivity can be obtained by SAB at room temperature.

4H-SiC wafer bonding has been achieved by the modified surface activated bonding (SAB) method without any chemical-clean treatment and high temperature annealing. Strong bonding between the SiC wafers with tensile strength greater than 32 MPa was demonstrated at room temperature under 5 kN force for 300 s. Almost the entire wafer has been bonded very well except a small peripheral region and few voids. The interface structure was analyzed to verify the bonding mechanism. It was found an amorphous layer existed as an intermediate layer at the interface. After annealing at 1273 K in vacuum for 1 h, the bonding tensile strength was still higher than 32 MPa. The interface changes after annealing were also studied. The results show that the thickness of the amorphous layer was reduced to half after annealing.

MEMS-compatible fabrication of nanoporous gold and the application for low temperature bonding are demonstrated. A cyanide-free electroplating solution is prepared for the Au–Sn alloy deposition. To investigate the influence of electroplating on Au–Sn alloy and nanoporous gold, different plating parameters and various sizes of patterns are designed and discussed. The optimized electroplating condition realizes 40 to 720 µm line width patterns fabricated on the same substrate. Low temperature substrate bonding at 200 °C is achieved with nanoporous gold and gold film, which has shear bond strength more than 60 MPa. The fracture inspection of the bonded area after shear tests verifies the bonding success. This study gives a study for fabricating on-chip nanostructure, and the results indicate the high feasibility of nanoporous gold for low temperature substrate bonding.

In this study, the effect of the metal salt generation bonding technique on the strength of a direct-bonded copper–copper interface was investigated. Copper surfaces were modified by boiling in several types of organic acids, and direct bonding was performed at a bonding temperature of 423–673 K under a load of 588 N (for a bonding time of 0.9 ks). As a result of the surface modification, bonded joints were obtained at bonding temperatures of 150 K (after treatment with formic acid) and 100 K (after citric acid treatment) lower than that required for the unmodified surfaces. In addition, the duration of the modification effects was investigated by exposing the modified surface to an air atmosphere furnace kept at 323 K. The bonding strength of the citric acid-modified surface remained unchanged even after 168 h, whereas that of the surface modified with formic acid decreased within 6 h.

This study developed a low-temperature low-vacuum direct bonding process for dissimilar metals via surface modification with formic acid vapor. Robust Cu/Ag and Cu/Zn bonding with a shear strength higher than 25 MPa can be achieved by thermal compression at 275 and 300 °C, respectively. CuZn₅ and Cu₅Zn₈ formed at the interface of Cu/Zn joints, while no distinct interdiffusion layers appeared at the Cu/Ag interface. At elevated temperatures, the shear strength of Cu/Zn joints decreased significantly and turned to be weaker than Cu/Ag at 250 °C due to the softening of Zn. All the joints performed well subjected to thermal cycling up to 1000 times. However, compared with Cu/Ag joints with stable mechanical performance suffering aging at 250 °C, the shear strength of Cu/Zn degraded drastically up to 200 h, and after that it remained almost constant, which can be ascribed to the competitive growth between CuZn₅ and Cu₅Zn₈, resulting in collapse and oxidation of CuZn₅.

Low-temperature hydrophilic SiO₂–SiO₂ wafer bonding has been performed in vacuum by a new combined surface-activated bonding (SAB) technique. In this technique, wafers are irradiated by ion beam bombardment and simultaneously deposited with silicon by in situ silicon sputter deposition, and then terminated with Si–OH groups by water vapor exposure prior to bonding in vacuum. A surface energy of more than 1 J/m² was achieved by 200 °C postbonding annealing. A void-free oxide intermediate layer with a thickness of about 15 nm was observed at the bonding interface by transmission electron microscopy (TEM). The increased bonding energy can be attributed to the greater number of Si–OH formed through hydroxylation of the silicon deposited on the SiO₂ surfaces.

Thermocouples on a trench sidewall fronting a flow are fabricated by three-dimensional (3D) photolithography. The conventional thermocouples on the wafer top surface are also fabricated. The performances of these devices are compared. Without the flow inside the microchannel, the thermocouple on the trench sidewall shows the same output voltage as that on the wafer top surface positioned 40 µm from the channel. As a static response, when the microchannel is heated and room-temperature air flows inside the channel, the thermocouple on the sidewall shows a lower voltage. As a dynamic response, when hot air flows inside the channel and replaces the room-temperature air, the thermocouple on the sidewall shows a faster response, increasing its output voltage, and the local temperature of the flow can be measured more precisely.

Non-evaporable getter (NEG) thin films based on alloys of transition metals have been studied by various authors for vacuum control in wafer-level packages of micro electro mechanical systems (MEMS). These materials have typically a relatively high activation temperature (300–450 °C) which is incompatible with some temperature sensitive MEMS devices. In this work we investigate the potential of Au/Ti system with a thin or ultrathin non oxidizable Au layer as a low activation temperature getter material. In this bilayer system, gettering activation is produced by thermal outdiffusion of titanium atoms through the gold film. The outdiffusion kinetics of titanium was modelled and characterized by scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS), and Rutherford backscattering spectrometry (RBS) at various temperatures. Results confirm that Au/Ti bilayer is a promising getter material for wafer-level packaging with an activation temperature below 300 °C for 1 h annealing time.

In order to enhance the thermal stability of corundum-structured gallium oxide (α-Ga₂O₃), which is attractive for use in wide-band-gap heterostructure devices and amenable to band gap and function engineering but suffers from phase transformation in high-temperature growth (>500 °C) and treatments (>550 °C), we attempted aluminum (Al) doping. The thermal stability of the films was enhanced by increasing the Al doping concentration, and under the best doping conditions where the Al concentration was negligible compared with the basic chemical composition of Ga₂O₃, the growth and successive thermal treatment temperatures were increased to as high as 650 and 750 °C, respectively, without the marked appearance of the β-gallia phase. Under the doping conditions above, the inclusion of Al was not negligible at the growth temperature of 800 °C and the film composition was expressed as an alloy of α-(Al_0.2Ga_0.8)₂O₃, but this film remained as the α-phase at annealing temperatures up to 900 °C. Enhanced thermal stability widens the device process windows, contributing to the formation of various high-performance devices.

Atmospheric-pressure-plasma nitriding of titanium alloy Ti–6Al–4V has been achieved by using a pulsed-arc plasma jet with a N₂/H₂ gas mixture, where the plasma jet plume is sprayed onto the titanium surface under atmospheric pressure. We successfully formed a titanium nitride layer on the sample surface. Moreover, the diffusion layer was also formed, the hardness of which was increased from that of as-received titanium. The nitride layer growth was found to be diffusion-controlled, as in other conventional nitriding methods.

Local excitation and emission dynamics of an isolated "Type-I₁" basal-plane stacking-fault (BSF) in very low dislocation density GaN were studied using spatio-time-resolved cathodoluminescence. The low temperature lifetime of the BSF emission was quantified to be 640 ps. The carrier diffusion length was estimated by observing the temporal delay of the BSF peak relative to the free-exciton signal as a function of distance from the BSF. The results indicate that the near-band-edge emission leads to subsequent optical excitation of the BSF that increases the apparent diffusion length. Limiting the observation volume can improve the spatial resolution.

Epitaxial graphene growth on SiC by Si selective etching using tetrafluorosilane (SiF₄) is introduced, where SiF₄ in Ar ambient selectively etches Si from the SiC surface at temperatures 1400 °C or above, leaving the C as graphene. Raman spectra of SiC treated in Ar for 60 min at 300 Torr did not show a graphene G-peak. However, with the addition of SiF₄, a clear G-peak was observed for the surface treated for only 1 min, demonstrating faster Si removal using SiF₄. Si selective etching of SiC is explained by the Gibbs free energy, where Si removal is more favorable compared to C removal by SiF₄.

To clarify the instability of the ferrimagnetism which is the fundamental magnetism of ferrite, numerical-diagonalization study is carried out for the two-dimensional S = 1/2 Heisenberg antiferromagnet with frustration. We find that the ferrimagnetic ground state has the spontaneous magnetization in small frustration; due to a frustrating interaction above a specific strength, the spontaneous magnetization discontinuously vanishes so that the ferrimagnetic state appears only under some magnetic fields. We also find that, when the interaction is increased further, the ferrimagnetism disappears even under magnetic field.

We grow a boron (B)-doped BaSi₂ (0.7 µm)/undoped n-BaSi₂ (1.7 µm) layered structure on a p-Si(111) substrate by molecular beam epitaxy, and observe the cross-sectional potential profile across the junction by Kelvin probe force microscopy (KFM). The potential increases when the KFM tip is moved from the B-doped BaSi₂ to the n-BaSi₂, and decreases in the p-Si. Inflection points are clearly observed in the potential profile at the B-doped BaSi₂/n-BaSi₂ and n-BaSi₂/p-Si interfaces. Secondary ion mass spectrometry reveals that B atoms scarcely diffuse to the n-BaSi₂ layer. These results show the formation of a pn junction at the B-doped BaSi₂/n-BaSi₂.

The effect of spatial modulation of the uniaxial anisotropy (K) and exchange stiffness (A) parameters on the domain wall thickness was theoretically studied. We derived the Euler–Lagrange equation and the Landau–Lifshitz–Gilbert equation considering the modulation of K and A, and showed that the modulation of A gives rise to an additional term consisting of the first derivatives of A and the magnetization unit vector. Owing to this term, the modulation of A is more effective to modify the domain wall thickness than K. The condition for domain wall pinning by controlling its thickness through the modulation of K and A was also obtained.

Temperature-dependent photoluminescence (PL) and photoreflectance (PR) spectroscopy and room-temperature Raman spectroscopy and X-ray diffraction have been utilized to investigate the optical properties, electron concentration, crystalline quality, and electronic band structures, especially valence-band splittings, of InN films grown by plasma-assisted molecular beam epitaxy (PAMBE) and metal–organic chemical vapor deposition (MOCVD). The smaller thermal activation energies imply the PAMBE-grown InN film exhibits low-density localized states from band tail states. PR signals of the InN film are detectable when the temperature is below about 100 K due to the cooling down of free electrons to trap states. For the MOCVD-grown InN film, no PR signal is observed even at 15 K due to the higher free electron concentration. To analyze the energetic positions of the features in the PR spectra without ambiguity, the moduli of individual PR resonances are considered. Based on the PR results and appropriate Hamiltonian, the values of the crystal-field splitting and the spin–orbit splitting in InN are experimentally determined as 26.8 and 14.5 meV, respectively. Theoretical and experimental reports are compared and discussed to verify this result.

We analyze breakdown characteristics and current collapse in AlGaN/GaN HEMTs, with the passivation layer's relative permittivity ε_r as a parameter. It is shown that the off-state breakdown voltage is considerably enhanced by introducing a high-k passivation layer because the electric field at the drain edge of the gate is weakened and the buffer leakage current is reduced. The breakdown voltage in the high-ε_r region increases when the gate voltage is changed from −8 to −10 V, because the buffer leakage current is reduced. It is also shown that drain lag and current collapse in AlGaN/GaN HEMTs could be reduced by introducing a high-k thick passivation layer, because the electric field at the drain edge of the gate is reduced, leading to less electron injection into the buffer layer and weaker buffer trapping effects.

Epitaxial Sr(Sn_1−xTi_x)O₃ (SSTO, x = 0–1) thin films were grown on MgO substrates by a pulsed laser deposition technique. The effects of composition on the structural and optical properties of SSTO films were investigated. X-ray diffraction studies show that the lattice parameter decreases from 4.041 to 3.919 Å gradually with increasing Ti content from 0 to 1 in SSTO films. Optical spectra analysis reveals that the band gap energy E_g decreases continuously from 4.44 to 3.78 eV over the entire doping range, which is explained by the decreasing degree of octahedral tilting distortion and thus the increasing tolerance factor caused by the increasing small-Ti-ion doping concentration.

Recent studies on single- and multilayer molybdenum disulfide (MoS₂) devices have revealed their promising characteristics as semiconductor devices. Understanding the transport properties at metal/MoS₂ interfaces may be crucial for their implementation. In this study, we measured the electrical characteristics of field effect transistors (FETs) with a MoS₂ channel from room temperature to 30 mK. A high on/off ratio (up to 10⁷ at 1 K) was observed at all temperatures. Below 1 K, we observed for the first time an anomalously large hysteresis in the transfer characteristics of the MoS₂ FET. We hypothesize that this hysteresis results from the slow injection of electrons via quantum tunneling through the Schottky barrier at the contacts. The size of the hysteresis increased with increase in the scan rate of the gate voltage, which is consistent with the possibility of slow injection of electrons.

We studied a novel method of increasing the efficiency of solar cells using BaSi₂ as a semiconductor. BaSi₂ could be deposited by RF magnetron sputtering using a polycrystalline BaSi₂ target, followed by annealing at 500 °C for 30 min in N₂ ambient. Furthermore, Schottky-type solar cells using BaSi₂ were fabricated. The crucial point is that Al–Nd was used to form the Schottky junction between the BaSi₂ film and the Al–Nd electrode. Additionally, Si₃N₄ (3–5 nm) was used as an oxidation prevention layer. Under irradiation at 90 K, resulting in a short-circuit current density (J_sc) of 3.19 mA/cm², an open-circuit voltage (V_oc) of 0.76 V, and a fill factor (FF) of 0.28 were obtained.

To explore the thermoelectric transport nature of photo-excited carriers, the electrical conductivity and the Seebeck coefficient are measured under ultraviolet illumination in the wide-gap semiconductor ZnS near room temperature. The conductivity increases linearly as against the photon flux density with little dependence on temperature, indicating the conduction under illumination is mostly governed by the photo-doped carriers. We have found that, in high contrast to the temperature-insensitive photoconductivity, the temperature dependence of the Seebeck coefficient is dramatically varied by illumination, which is unexplained from a simple photo-doping effect for one majority carrier. Such a distinct difference in the transport quantities is rather understood within a two-carrier model, in which only the Seebeck coefficient is strongly affected by photo-excited minority carriers. The present result is also compared with earlier reports of the photo-Hall experiments to discuss the underlying photo-transport mechanism.

This paper reports on sputter-deposited SiGeSb thin films and their application for four-terminal chalcogenide switch devices. The microstructures and electrical properties of the SiGeSb films were highly dependent on antimony concentration and annealing temperature. Microstructural changes such as surface roughening and formation of antimony grains were observed only for the Sb-rich SiGeSb films after annealing at 400 °C and higher. The sheet resistance of the SiGeSb films containing a small amount of antimony changed sporadically with annealing temperature because of a trade-off between activation and surface depletion of antimony. The resistance of the SiGeSb heating electrodes was varied by changing sputtering power for the antimony target and by changing the annealing temperature. Four-terminal chalcogenide switch devices were fabricated with SiGeSb heating electrodes of varying resistance. It was found that the switching voltage of the fabricated switch device was proportional to the resistance of the SiGeSb heating electrode. This indicates that the SiGeSb films with tunable sheet resistance are of great importance in fabricating chalcogenide switch devices and the optimization of the resistance of the SiGeSb film is essential to ensure proper switch operation.

A Cu/SiO₂/Pt structure was fabricated to investigate its resistive switching characteristics. The application of DC voltages with different polarities allowed for the reversible manipulation of the structure's resistance. This resistive switching phenomenon is the result of the formation and rupture of Cu conducting filaments near the Cu/SiO₂ interface. However, significant switching dispersion occurred during successive switching cycles, which resulted in operational difficulties and switching failure. In this study, a voltage prestress was applied to the structure in an attempt to minimize the switching dispersion. A statistical technique was used to analyze the status of formation/rupture sites, and a schematic model is proposed to explain the reason for the dispersion improvement. It is suggested that the voltage prestress builds nonconnected filaments and reduces the number of sites of filament formation/rupture. This reduction in the number of sites leads to reduced switching dispersion.

The electromagnetic interaction of atoms with an evanescent light field is investigated with photon energy density and photon energy flux density using the Fresnel–Maxwell equations. The theory provides expressions of absorption signal intensities in attenuated total reflection spectroscopy for the p- and s-polarizations of the incident light, and relates them to the Goos–Hänchen shift. For an electric dipole interaction, the energy stored in the electric field of an evanescent light governs the absorption process, and hence the photon energy density and the photon energy flux density should be estimated from the energy stored in the electric field. The violation of equality between the portions of the energy stored respectively in the electric and magnetic fields yields the difference of the absorption signal intensity and its dependence on the incident angle between the s- and p-polarizations.

The gain form of a 1.5 µm multimode laser diode was derived by phenomenologically adding an intrinsic gain saturation term to the linearized gain form described in our previous report. Although the proposed gain form was simple, it almost perfectly matched the complex laser diode gain in a direct transition model implicitly incorporating both band filling and intrinsic gain saturation effects. Using laser diode rate equations including the proposed gain form, the characteristic power spectrum of a gain-switched pulse from a 1.5 µm multimode laser diode was successfully simulated.

Difference of conduction band minimum (E_C) between transparent conductive oxide (TCO) and absorber, named ΔE_C-TA, in thin-film solar cell is investigated for high cell performance using device simulator. According to the simulation, the optimized ΔE_C-TA value is different, depending on the carrier density in buffer layer, N_D-B. With ΔE_C-TA above 0.6 eV for both N_D-Bs of 1.0 × 10¹³ and 1.0 × 10¹⁸ cm⁻³, the spike is formed at the TCO/buffer interface, thus decreasing cell performances, especially short-circuit current density owing to impeding photo-generated carriers to TCO. On the other hand, with ΔE_C-TAs below −0.2 and −0.4 eV for N_D-Bs of 1.0 × 10¹³ and 1.0 × 10¹⁸ cm⁻³, the solar cells demonstrate double diode characteristics, thereby decreasing cell efficiency. Eventually, the optimized ΔE_C-TA values for high cell performance are proposed to be in the ranges from −0.2 to 0.6 eV and from −0.4 to 0.6 eV for N_D-Bs of 1.0 × 10¹³ and 1.0 × 10¹⁸ cm⁻³, respectively.

We realized a long-scanning-range and high-resolution interferometry in a time-domain full-field microscopic scheme by adopting a simple configuration. A reference mirror was synchronously scanned with an objective lens, which was installed in a common path, to prevent lateral resolution degradation due to defocus at the mirror. High axial resolution was obtained using a broadband supercontinuum (SC) generated by a 1.55 µm pump. The SC was generated by propagating a femtosecond pulse at 1.55 µm through a highly nonlinear dispersion shifted fiber with a small dispersion slope. We designed and constructed an interferometer carefully to utilize the entire bandwidth. The broad bandwidth of the interferometer achieved an axial resolution of 2.50 µm in air. The synchronous scanning maintained a lateral resolution longer than 1 mm. The system successfully yielded a cross-sectional image of two layers of scotch tape along the 400-µm-depth and 90-nm-step surface profiles.

We have developed a simple, high-sensitivity optical-fiber temperature sensor based on multimode interference (MMI). The fabricated MMI structure comprises three segmented fibers: a single-mode fiber (SMF); a large-core multimode fiber (MMF), whose outer surface is coated with a temperature-sensitive material; and another SMF. Fluoroacrylate and silicone rubber are tested as temperature-sensitive cladding materials. The silicone rubber coating exhibits a large shift in interference wavelength with temperature, producing a very fine temperature resolution as low as 0.01 °C.

By simultaneously employing GaAs and Cr⁴⁺:YAG saturable absorbers, a diode-pumped doubly passively Q-switched and mode-locked (DP-QML) Nd:Gd₃Ga₅O₁₂ (Nd:GGG)/KTiOPO₄ (KTP) green laser has been presented. At the maximum incident pump power of 7.69 W, the obtained average output power, the pulse duration of the Q-switched envelope, the mode-locked pulse width and the pulse repetition rate are 62 mW, 16.8 ns, 352 ps, and 15.5 kHz, respectively, corresponding to a pulse energy of 4.2 µJ and a peak power of 0.25 kW. In comparison with singly passively Q-switched and mode-locked (SP-QML) green laser with GaAs or Cr⁴⁺:YAG saturable absorbers, the DP-QML green laser can generate shorter pulse width, deeper modulation depth and higher peak power. The coupled equations for diode-pumped doubly passively QML Nd:GGG/KTP green laser are given and the numerical simulations are in good agreement with the experimental results.

Magnetoresistance in antiferromagnetically coupled GaMnAs/GaAs:Be multilayers exhibits a unique step feature caused by sequential flips of magnetization in individual GaMnAs layers. Hysteresis loops corresponding to such magnetization flips in specific layers were measured by adjusting the range of the field scan to where the flip occurs. Using the hysteresis shifts of such partial loops, we were able to obtain the strength of the interlayer exchange coupling exerted on a given GaMnAs layer by the rest of the multilayers. This also allowed us to quantitatively establish the magnitude of the coupling of each GaMnAs layer with its adjacent neighbors.

A partially ordered Fe₁₆N₂ thin film, which exhibits a higher saturation magnetization than a bcc-Fe thin film, was grown on a Au(001) texture on a GaAs(001) substrate for studies of crystalline structure, electronic structure, and magnetic properties. Fe 2p_3/2 and 2p_1/2 X-ray photoelectron spectroscopies (XPS) reveal the electronic hybridization between the Fe atoms and the adjacent N atoms, whereas a multipeak analysis suggests the charge-transfer-induced electronic rearrangement of electronic configuration in Fe(8h) and Fe(4e) geometrical sites. These results are consistent with the previous model and help explain the saturation magnetization enhancement in the α-FeN system.

Longitudinal and in-plane electron g-factors, and a nuclear spin polarization (NSP) have been evaluated precisely in a CdTe/Cd_0.85Mg_0.15Te single quantum well by using the time-resolved Kerr rotation and double lock-in detection techniques. Resident electron spin polarization (RESP) was formed via the negative trion formation and recombination, and RESP gave rise to NSP in an oblique magnetic field configuration. We observed the effective nuclear field of a few mT which was weak compared with that in III–V semiconductor nanostructures as expected, but the nuclear field can be converted to the maximal NSP of 12% in Faraday geometry.

We have developed waveguide-type low-noise superconducting hot-electron bolometer (HEB) mixers for astronomical observations in the 1.3–1.5 THz region by using a relatively thick NbTiN superconducting film (10.8 nm). We have achieved a receiver noise temperature of 490 K (DSB: double side band) at 1.475 THz. This noise temperature corresponds to seven times the quantum noise. According to gain bandwidth measurements, the contribution of diffusion cooling is found to be responsible for such a good noise performance.

In this paper, we analyzed the amplitude of variable junction leakage currents caused by the interaction between two interface states in MOS transistors. For the first time, an analytical equation for the ratio between the junction leakage current before and after electron capture into the slow state was derived with consideration of both the change in the capture cross-section and the electric field. Also, the correct equation for the electric field after electron trapping was derived and used. The distance between the two interface states was extracted from the equation and measurement data. The extracted distance was interpreted under the framework of the inter-atomic distance in the silicon lattice structure at the Si/SiO₂ interface.

The bistability characteristics of GaN/AlN resonant tunneling diodes (RTDs) grown on a sapphire substrate by metalorganic vapor phase epitaxy (MOVPE) were investigated to better understand their physical origin and explore their use in nonvolatile memories. The bistability current–voltage (I–V) characteristics of GaN/AlN RTDs, which were due to intersubband transitions and electron accumulation in the quantum well, were clearly observed over a wide temperature range between 50 and 300 K. However, the I–V characteristics sometimes degraded at temperatures above 250 K. Complex staircase structures were observed in the voltage region showing a negative differential resistance in the I–V curve, and the forward current increased or decreased rapidly as the forward-bias voltage increased. Repeated measurements of the I–V characteristics over the wide temperature range between 50 and 300 K revealed that the bistability characteristics of GaN/AlN RTDs degraded owing to the leakage of electrons accumulating in the quantum well through a deep level in the AlN barrier associated with crystal defects such as dislocations and impurities. Therefore, reduction in crystal defect and impurity densities in the AlN barrier, and a careful design that considers deep levels are important for realizing realize ultrafast nonvolatile memories based on the bistability characteristics of GaN/AlN RTDs.

Scaling the gate dielectric is a key to improving the steep switching characteristics of tunnel field-effect transistors (TFETs). The effect of the gate leakage current caused by this scaling has not fully been investigated until now. In this work, gate leakage current paths in a p-channel TFET are experimentally investigated by designing three types of measurement setup for separating the current paths. Our measurements reveal that the so-called gate-edge leakage current that directly flows from the gate into the source adversely affects the subthreshold characteristics of the source current. We also report that the areal leakage current tunneling from the gate to the channel causes the OFF current to increase. A device simulation is performed to separate the contribution of electron gate tunneling from that of hole gate tunneling and to understand the mechanism underlying these tunnelings, which reveals that electron tunneling occurring in the nanometer-scale source–gate overlap determines the gate leakage. Structural engineering around the source–channel overlap plays an important role in achieving high-performance TFETs.

We propose a silicon-nanowire-based photodetector with a radial metal junction that is predicted to exhibit a significantly reduced response time. The width of the radial depletion layer across the nanowire can be controlled by adjusting the doping concentration of the silicon nanowire. We calculated depletion region widths for silicon nanowires of various diameters and doping concentrations, and then calculated the photogenerated carrier transit time, the RC time constant, and the diffusion time in the nanowire structure. We found that by using the radial junction configuration we could significantly improve the response time to 81 ps. We also found that the diffusion time for the photogenerated carriers depends strongly on the nanowire length and doping concentration.

The effect of Si-doping on the phase separation of wurtzite (WZ) and zinc-blende (ZB) phases in catalyst-free Si-doped GaAs nanowires (NWs) grown on a Si(111) substrate was investigated using transmission electron microscope (TEM), high-resolution X-ray diffraction (HR-XRD), low-temperature photoreflectance (PR), and photoluminescence (PL) techniques. The appearance of WZ structure with an increase in the amount of Si dopant was observed through TEM, and the results showed that the thicknesses of ZB and WZ structures were random. Furthermore, all NW samples exhibited HR-XRD diffraction peaks at the (0002) and (111) planes, which correspond to the WZ and ZB structures, respectively. Their peak intensity ratio [WZ/(WZ + ZB)] increased with the amount of Si doping. The PR modulus and PL spectra at 4 K for the sample with the middle amount of Si doping in three samples exhibited peaks at 1.43, 1.49, and 1.51 eV. The peaks at 1.51 and 1.49 eV were presumed to result from band-to-band and conduction-band-to-Si-acceptor transitions, respectively. In accordance with the prediction by a theoretical band alignment calculation of the conduction- and valence-bands discontinuities, the transition energy of 1.43 eV was due to the interband transition at the WZ-ZB interface. We also found that the 1.43 eV PR and PL peaks became dominant when the amount of Si doping increased. This indicate that this interband transition became significant when the amount of WZ phase increased, which resulted from the increased Si doping. The appearance of type-II band structures induced by Si doping was also confirmed.

We demonstrate the possible candidate dispersion agents that can uniformly disperse carbon nanotubes (CNTs) into organic solvent, from among commercially available polymers. We find that CNTs were well dispersed into dimethylacetamide with the use of polystyrene, poly(vinyl chloride), and poly(vinyl pyrrolidone) as dispersion agents. Theoretical calculations revealed that the dispersibility of these polymers arises from the moderate strength and preferential directionality of the interactions between the CNTs and the polymers.

Ion concentration polarization (ICP) is a distinctive electrochemical phenomenon that occurs near an ion-exchange membrane with an applied DC electric field, generating a significant concentration gradient in back and forth on the membrane. To date, however, there have been only a few attempts to introduce unconventional materials for ion transport in micro–nanofluidic systems. Here, we describe the development of a novel ICP system using an entangled single-wall carbon nanotube (SWNT) film as an ion-selective membrane instead of a Nafion membrane, for investigating the detailed relationship between electrical properties, i.e., ionic conductance through nanojunctions, and nonlinear electrokinetic behavior.

In this study the organic resistive switching devices having sandwich structure of indium tin oxide (ITO)-coated glass/poly(4-vinylphenol) (PVP)–graphene composite/silver (Ag) were fabricated and characterized. The active layers were fabricated using blended, semiblend and layer-by-layer approaches, sandwiched between two electrodes. The film thicknesses of the active layers were measured to be about 200 nm. The surface morphology was characterized by field-emission scanning electron microscopy. Electrical current–voltage (I–V) analyses confirmed the memristive behavior of the sandwich devices. The effect of active layer fabrication approach was analyzed by comparing the resistive switching characteristics. The devices showed characteristic OFF to ON (high resistance to low resistance) transition at low voltages, when operated between ±2 V, characterized at 100 µA and 5 mA compliance currents. The memristive behavior of PVP–graphene active layer fabricated by blended approach showed more stability and robustness compared to non-blended approaches. The devices fabricated by blended approach exhibited a room temperature V–I hysteresis and R_OFF/R_ON ≈ 5.

Although layered polysilane (LPS) has a graphitelike structure and a higher electric capacity than Si powder, LPS shows poor cycling performance. Therefore, to improve its cyclability, the synthesis of composite materials of LPS and carbon was attempted. Carbon-coated LPS (C-LPS) was prepared by pulsed laser deposition (PLD), and a silicon–carbon (Si–C) composite consisting of silicon plates coated onto carbon particles was obtained by reacting LPS with sucrose. The cycling performance characteristics of the C-LPS and Si–C composite electrodes were better than that of the LPS electrode. On galvanostatic discharge–charge curves, although the Si–C composite showed no plateau at 0.3 V on the first discharge, the LPS and C-LPS electrodes showed a 0.3 V plateau that was due to the dissociation of the Si–H bonds of LPS, which led to a high initial irreversible capacity.

Epitaxial BaTiO₃ (BTO) thin films of less than 100 nm thickness were grown on Pt(001)/MgO(001) substrates at growth temperatures of 500–700 °C with a low deposition rate of about 25 nm/h by metal–organic chemical vapor deposition (MOCVD). The BaTiO₃ thin films were epitaxialy grown with (001) orientation. These films show quadrangular grains and a dense cross-sectional structure. The relative permittivities of these films grown at 500, 600, and 700 °C with thicknesses of 62, 65, and 82 nm were 338, 455, and 566 at 1 kHz, respectively. These relative permittivities were higher than those of BTO films prepared by other methods and BTO ceramics with thicknesses less than 100 nm.

The interdiffusion of magnesium and iron in gallium nitride (GaN), i.e., magnesium–iron interdiffusion, was investigated using magnesium-doped GaN layers on iron-doped GaN substrates. The investigation confirms that the magnesium–iron interdiffusion strongly depends on the concentrations of magnesium and iron, that is, it occurs when the iron and magnesium concentrations are high (magnesium: 2 × 10²⁰ cm⁻³; iron: 2 × 10¹⁹ cm⁻³). It also confirms that diffused iron in the magnesium-doped GaN layer acts as a nonradiative recombination center in GaN.

An ion-filtered inductively coupled plasma (IF-ICP) is proposed to reduce ion bombardment and provide high metastable species density for chemical vapor deposition. Argon plasma, which has simple reaction mechanism, is simulated to show the effects of ion filter. Compared to typical ICP, the maximum density of ions of IF-ICP is lower while that of metastable species is higher. The filter can absorb ions effectively and relatively small amount of metastable species, with the absorption coefficient proportional to its surface area. A proper gap between filter and substrate can achieve more metastable species and less ions on the substrate. The pressure and RF power need to be optimized based on the tradeoff between deposition rate and ion damage. The density of ions on the substrate can be reduced by two orders of magnitude while that of metastable species are maintained in the order of 10¹⁷ m⁻³ under the optimized conditions.

The CF₄ and C₄F₈ gas etching profiles of oxide films were compared by multiscale simulation that comprises gas reaction, sheath, and surface reaction models. The densities of CF₃, CF₂, and CF radicals in CF₄/Ar or C₄F₈/Ar gas were measured and compared with those obtained by simulation using the gas reaction model. From this comparison, the electron temperatures were determined to be 2.8 and 4.5 eV for CF₄ and C₄F₈ gases, respectively. In the sheath model, the behavior of ions in a sheath was simulated for sheath lengths calculated from these electron temperatures. In the surface reaction model, we simulated the formation of a polymer and active layers by CF₂ radicals, determined the depth of etching resulting from ion bombardment, and obtained the etching profiles of the oxide films. The profile of a contact hole with a depth of 820 nm and an aperture diameter of 160 nm was simulated. The results showed that the photoresist height was approximately 70 nm greater and the bowing diameter was approximately 10 nm smaller in the case of using C₄F₈ gas than in the case of using CF₄ gas. This is because the CF₂ density in the C₄F₈ gas is approximately 30 times higher than that in the CF₄ gas and the polymer layer more strongly protects the underlying film. When the etching profiles were simulated with a fixed density of positive ions but with various CF₂ density, the bottom diameter was constant but the bowing diameter changed for CF₂ densities between 10¹³ and 10¹⁴ cm⁻³.

We report on the effects of UV light intensity on the photo assisted electrochemical wet etching of SiC(0001) underneath an epitaxially grown graphene for the fabrication of suspended structures. The maximum etching rate of SiC(0001) was 2.5 µm/h under UV light irradiation in 1 wt % KOH at a constant current of 0.5 mA/cm². The successful formation of suspended structures depended on the etching rate of SiC. In the Raman spectra of the suspended structures, we did not observe a significant increase in the intensity of the D peak, which originates from defects in graphene sheets. This is most likely explained by the high quality of the single-crystalline graphene epitaxially grown on SiC.

We show the systematical investigation results of the effects of the implanted ion dose of P or As under various solid-phase epitaxy (SPE) conditions on the local stress in channel regions in metal–oxide–semiconductor field-effect transistor (MOSFET) structures, and on sheet resistance and strain in carbon-doped source/drain (Si:C-S/D) layers. P or As substitution is in conflict with C substitution in Si:C layers during SPE. Furthermore, the amount of P incorporated instead of C into the Si lattice site is larger than that of As incorporated instead of C. Therefore, low-resistivity Si:C layers with low stress in the case of using P and high-resistivity Si:C layers with high stress in the case of using As are formed by single-step C₇H_x implantation with rapid thermal annealing and nonmelt laser annealing, respectively. As a countermeasure, we demonstrate that cascade C₇H_x implantation to control the C profiles in Si:C layers is effective for achieving high-strain channels and low-resistivity Si:C-S/D layers. Control of C profiles is a key technology for state-of-the-art complementary MOS devices with Si:C-S/D.

The impact of low-k dielectric benzocyclobutane (BCB) encapsulation on the electrical performance and structural stability of AlGaN/GaN HEMTs on Si were investigated. After BCB encapsulation, devices exhibited no degradation in their drain current density, extrinsic transconductance and small signal microwave performances. The curing temperature (280 °C) of BCB layer had no influence on the device electrical performances. Compared to devices without BCB encapsulation, the BCB encapsulated devices achieved ∼2 orders of magnitude lower gate- and drain-leakage current. An order of magnitude lower surface leakage current was also observed by BCB encapsulation between the two adjacent mesas. Due to the reduction of leakage currents, ∼2-times increase of OFF-state breakdown voltage was observed. In addition, the 9-µm-thick BCB encapsulation layer also helps to have structurally stable air bridges. This work demonstrates the low-k dielectric BCB as a viable solution for the complete encapsulation of GaN HEMTs and ICs.

We have developed a wafer-scale layer-transfer technique for transferring GaAs and Ge onto Si wafers of up to 300 mm in diameter. Lattice-matched GaAs or Ge layers were epitaxially grown on GaAs wafers using an AlAs release layer, which can subsequently be transferred onto a Si handle wafer via direct wafer bonding and patterned epitaxial lift-off (ELO). The crystal properties of the transferred GaAs layers were characterized by X-ray diffraction (XRD), photoluminescence, and the quality of the transferred Ge layers was characterized using Raman spectroscopy. We find that, after bonding and the wet ELO processes, the quality of the transferred GaAs and Ge layers remained the same compared to that of the as-grown epitaxial layers. Furthermore, we realized Ge-on-insulator and GaAs-on-insulator wafers by wafer-scale pattern ELO technique.

Extreme ultraviolet (EUV) lithography is the most promising candidate technique for the high-volume production of semiconductor devices with half-pitches of sub-10 nm. An anion-bound polymer, in which the anion part of onium salts is polymerized, has attracted much attention from the viewpoint of the control of acid diffusion. In this study, we modeled the acid generation processes in the anion-bound chemically amplified resists upon exposure to EUV radiation and developed a Monte Carlo simulation code. Using the developed simulation code, the dependence of the quantum efficiency of acid generation on the concentration of acid generator units was calculated. The calculated quantum efficiencies well agreed with the experimental values with a fitting error of less than 10%. The thermalization distance was considered to be approximately 3 nm. The blur of proton distribution intrinsic to the reaction mechanisms of anion-bound chemically amplified resists was roughly estimated to be 4.5–6.5 nm.

Line edge roughness (LER) rapidly increases in the sub-10-nm-half-pitch region of resist processes used for the fabrication of semiconductor devices. Sub-10-nm fabrication with high throughput is a challenging task. In this study, the stochastic effects (LER and stochastic defect generation) of chemically amplified resist processes in the sub-10-nm-half-pitch node were investigated, assuming the use of extreme ultraviolet (EUV) lithography. The latent images were calculated by a Monte Carlo method on the basis of the sensitization and reaction mechanisms of chemically amplified EUV resists. 7-nm-half-pitch fabrication by chemically amplified resist processes is considered to be feasible. However, significant improvement in the efficiencies of the conversion processes from optical images to resist images is required.

The mechanical properties of compliant single cells are significantly associated with various cell functions. It is thus crucially important in the identification and sorting of cells to characterize not only the complex moduli that exhibit power-law rheology but also the cell-to-cell variation at the single cell level. Atomic force microscopy (AFM) can be used to measure cellular mechanical properties and to quantify cell-to-cell variation. However, less is known about how precisely and routinely the cell-to-cell variation is obtained when the complex moduli vary substantially in different cell samples. Here, we investigate the storage modulus G' for single cells measured at controlled positions by AFM. We find that the spatial dependence of the frequency-dependent component of the cell-to-cell variation is preserved even if the spatial heterogeneity of G' is changed with the cell sample. The invariance of the frequency-dependent cell-to-cell variation indicates the robustness of AFM for the mechanical diagnosis of single cells.

Magnetic particle imaging is a novel method of imaging the spatial distribution of magnetic nanoparticles. When considering the practical application of magnetic particle imaging, it is important to correct the inhomogeneous sensitivity of the receiving coil together with the feedthrough interference. In this study, we developed a simple and practical method for these corrections in which projection data are multiplied by correction factors obtained by fitting projection data acquired in a blank scan to a sixth-degree polynomial. Phantom experiments suggest that our method can be simply and easily implemented to realize the above corrections.

FeS₂ is potentially well-suited for the absorber layer of a thin-film solar cell. Since it usually has p-type conductivity, a pn heterojunction cell can be fabricated by combining it with an n-type material. In this work, the band alignment in the heterostructure based on FeS₂ is investigated on the basis of the first-principles calculation. CdS, the most popular buffer-layer material for thin-film solar cells, is selected as the partner in the heterostructure. The results indicate that there is a large conduction band offset (0.65 eV) at the interface, which will hinder the flow of photogenerated electrons from FeS₂ to CdS. Thus an n-type material with the conduction band minimum positioned lower than that of CdS will be preferable as the partner in the heterostructure.

Al₂O₃ passivation by thermal oxidation of aluminium on AlGaN/GaN high-electron-mobility transistors (HEMTs) without Al₂O₃ etching is proposed. The deposition of a 5-nm-thick Al film was carried out, followed by a lift-off process to remove Al from the ohmic and contact pad area. Subsequently, the Al film was annealed under O₂ ambient. When the gate bias was −7 V, the gate leakage currents of a conventional nonpassivated HEMT, a surface-passivated Schottky-gate HEMT, and a surface-passivated MOS-HEMT were determined as 140, 96, and 4.1 µA/mm, respectively. The current collapse phenomenon in the Al₂O₃-surface-passivated devices was evidently suppressed compared with that in the nonpassivated HEMT.

We have investigated the temperature dependence of current-induced magnetic domain wall (DW) motion in a Co/Ni nanowire, where the spin Hall torque is responsible for the DW motion. The threshold current density (J_th) for DW motion is found to be dependent on the temperature T and the depinning magnetic field H_dep, which is different from the symmetric Co/Ni system where J_th is insensitive to T and H_dep. This result indicates that an extrinsic pinning governs the DW motion when it is driven by the spin Hall torque. Our work therefore suggests that reducing the extrinsic pinning is a key for achieving a low-power-consumption device.