Table of contents

See pdf for details

The development of the high electron mobility transistor (HEMT) provides a good illustration of the way a new device emerges and evolves toward commercialization. This article will focus on these events that the author feels might be of interest to young researchers. Recent progress and future trends in HEMT technology are also described.

Photocatalysis has recently become a common word and various products using photocatalytic functions have been commercialized. Among many candidates for photocatalysts, TiO₂ is almost the only material suitable for industrial use at present and also probably in the future. This is because TiO₂ has the most efficient photoactivity, the highest stability and the lowest cost. More significantly, it has been used as a white pigment from ancient times, and thus, its safety to humans and the environment is guaranteed by history. There are two types of photochemical reaction proceeding on a TiO₂ surface when irradiated with ultraviolet light. One includes the photo-induced redox reactions of adsorbed substances, and the other is the photo-induced hydrophilic conversion of TiO₂ itself. The former type has been known since the early part of the 20th century, but the latter was found only at the end of the century. The combination of these two functions has opened up various novel applications of TiO₂, particularly in the field of building materials. Here, we review the progress of the scientific research on TiO₂ photocatalysis as well as its industrial applications, and describe future prospects of this field mainly based on the present authors' work.

We implanted As and P ions in a 110-nm-thick HfO₂ layer, subjected the substrates to various thermal processes, and evaluated their diffusion coefficients by comparing experimental and numerical data. We found that the diffusion coefficients of As and P in HfO₂ are almost the same as that of B and are much higher, by two orders, than that of B in SiO₂. The impurity penetration through the HfO₂ gate insulator is much more severe than that of SiO₂ even though a thicker HfO₂ layer is available.

We present two-dimensional simulation of quantum tunneling across a potential barrier with surface roughness using quantum lattice–gas automata. The impact of the nonuniformity of the barrier thickness on the transmission coefficient is discussed by comparing the results of one- and two-dimensional tunneling simulations. The dependence of the transmission coefficient on the parallel momentum of the incident electron is also investigated, and it is demonstrated that the scattering by the surface roughness on the incident side of the interface causes the violation of the parallel momentum conservation. We discuss the effect of the obtained results on the gate current modeling for the scaled metal–oxide–semiconductor field-effect transistors.

Heavily doped areas and lattice strains in device structures can provide effective traps for metals and compete with intentionally introduced gettering sites for impurities. In this study, the effect of device structures on Fe gettering by p⁺ substrates was investigated by reverse current measurements and deep level transient spectroscopy (DLTS) of n⁺–p junction diodes. A significant improvement in the gettering efficiency was achieved by optimizing the cooling rate in annealing. DLTS detected the gap states, which act as minority carrier traps, in the wafers without optimization of annealing conditions. The gap states correlated to the perimeter length of n⁺–p junctions, which were defined by local oxidation of silicon (LOCOS). The leakage current and the density of gap states were greatly reduced by optimizating the cooling rate. We suggest that the Fe atoms gettered by lattice strain due to LOCOS generate the gap states, and that these gap states are the origin of excess leakage current.

An ultrathin SiO₂ overlayer on a Si(001) surface formed by a 5 eV O-atom beam at room temperature was analyzed by synchrotron radiation photoelectron spectroscopy (SR-PES). SR-PES spectra clearly indicated that the SiO₂ layer formed by a hyperthermal O-atom beam at room temperature contains a small amount of suboxides compared with that formed by high-temperature oxidation in O₂ atmosphere. Quantitative analysis suggests that the thickness of the structural transformation layer was less than a monolayer and the amount of suboxides was independent of the film thickness. The translational energy dependence of SR-PES spectra suggests that the reaction probability with a Si-atom increases with the translational energy of the O atoms in the range between 1 to 5 eV. The role of inverse diffusion of interstitial Si atoms in the kinetics of hyperthermal O-atom-beam oxidation is suggested.

Comprehensive investigations of the various static and microwave performances of InAlAs/InGaAs/InP high-electron-mobility transistor (HEMT) with a linearly graded In_xGa_1-xAs channel (LGC-HEMT) have been conducted. LGC-HEEMT was compared with the same HEMT having a conventional lattice-matched In_0.53Ga_0.47As channel (LM-HEMT). Improved carrier transport characteristics and confinement capability achieved by employing a linearly graded channel have contributed to a high extrinsic transconductance (g_m) of 319 mS/mm, a high unity-gain cutoff frequency ( f_t) of 37 GHz, and a maximum oscillation frequency ( f_max) of 51 GHz at 300 K for a gate length of 0.65 µm. The improved gate-voltage swing, turn-on and output power characteristics of LGC-HEMT have also been discussed.

The effect of indium on photoluminescence properties of InGaPN layers was investigated and compared with that of GaPN layers. Two phenomena involving photoluminescence properties were observed in the InGaPN layers: (i) an S-shape of photoluminescence (PL) peak energy as a function of temperature, caused by spatial fluctuation of bandgap energy related to In and N content; and (ii) red shifts of the PL peak energy at 18 K in the InGaPN layers after rapid thermal annealing (RTA), caused by the increase of N- and In-rich region with increasing RTA temperature. It was also found that integrated PL intensity in the InGaPN layers was higher than that in the GaPN layers, and that PL quenching became more insensitive to the change in temperature resulting from the decrease in nonradiative centers with increasing RTA temperature.

In order to extract electrical properties of multicrystalline silicon (Mx-Si) solar cells precisely, current–voltage (I–V) measurement in the dark condition was carried out. A new two-diode model, which includes a diffusion-current dominant area (DCA) and a recombination-current dominant area (RCA), with nonequivalent series resistances, was proposed as a new equivalent circuit. Electrical properties as fitting parameters were successfully extracted using a successive approximation method. This characterization derived J₀₁ and J₀₂, which are the diode saturation current densities for n=1 and 2 diodes, respectively. To ensure this diode model, temperature characteristics were measured, and the validity of this model was shown through calculation of the energy band gap from J₀₁. J₀₁ and J₀₂ indicate a power factor and a loss factor of solar cells, respectively, which are newly proposed to classify solar cell performance.

We investigate the impact of varying the grain boundary (GB) position on the output (I_d–V_d) characteristics of submicron single GB polysilicon thin film transistors (TFTs), by two-dimensional (2D), drift-diffusion based, device simulation. We employ a localized GB trapping model with a distribution of both donor-like and acceptor-like trap states over the forbidden energy gap of the GB region. We show that for devices with channel lengths in the deep submicron regime, significant variations in output conductance (g_d) occur as the GB position is varied. Specifically, we find that output conductance increases as the GB approaches the drain edge. Furthermore, the sensitivity of output conductance to the GB position increases as channel length decreases. The findings have important implications for any future analogue three-dimensional (3D) IC design that uses polysilicon as a device material.

The etching of α(4H,6H)–SiC{0001} substrates by a photoelectrochemical method using HF (0.015–4.5 wt %) and HF+HNO₃ (0.006–1.2 wt %) electrolytes was studied. The dependences of etching rate on polytype, polarity, and pH were measured. In the case of the HF electrolyte, an etching rate of 15–27 nm/min was achieved over a pH range from 0.5 to 4.5 under a photocurrent density of 1 mA/cm². By optimizing etching conditions, the surface roughness of the Si face could be improved to 0.9 nm (4H) and 0.4 nm (6H) compared with the initial surface roughnesses of 4H (8.9 nm) and 6H–SiC (6.5 nm). In the case of the HF+HNO₃ electrolyte, a thin oxide film 2–3 µm thick was formed after 60 min. The oxidized layer was two orders of magnitude thicker than that obtained using the thermal method. The pH of the electrolyte decreased after the electrochemical process, indicating an increase in the concentration of H⁺ ions. Therefore, holes and H₂O have a strong influence on the rate of oxidation reactions in electrochemical methods. Electrochemical etching proceeds by the competitive processes of formation and removal of oxide films.

Excess carrier lifetime in bulk 2-in. SiC wafers was measured by microwave photoconductivity decay (µ-PCD). The mapping technique was used to obtain the lifetime distribution in the entire wafer. We observed the birefringence image and X-ray topograph of the wafers in order to determine the structural defect distribution, and the net donor concentration distribution was also observed by capacitance–voltage measurements. By comparison of lifetime maps with the structural defect distribution, it was found that relatively long lifetime regions correspond to regions with high-density structural defects. The net donor concentration did not show a clear influence on the carrier lifetimes. We confirmed that surface recombination has a negligible effect on the carrier lifetimes, and therefore the lifetimes obtained from mapping measurements are mainly dominated by carrier recombination behavior in the bulk of the wafers.

We have investigated the transformation processes of the first (n=1) and second (n=2) minibands to the Wannier–Stark (WS) localization states in a GaAs (6.8 nm)/AlAs (0.9 nm) superlattice embedded in a p–i–n structure by electroreflectance spectroscopy. The high sensitivity of electroreflectance enabled us to observe the interband optical transitions associated with the n=2 minibands in addition to those associated with the n=1 minibands. The systematic results of the electroreflectance spectra as a function of electric field strength demonstrate that the electric field strength for the formation of the WS-localization state of the n=2 miniband is higher than that of the n=1 miniband, which reflects a difference in the energy widths of the n=1 and n=2 minibands. The experimental results are quantitatively discussed on the basis of the electric-field-strength dependence of the energies and envelope functions of the n=1 and n=2 minibands calculated using a transfer-matrix method. We find a clear correlation between the miniband widths and the critical electric field strengths for the formation of the WS-localization states.

An optimum design of an antireflective (AR) coating for a spherical Si solar cell with a reflector cup has been analyzed with a three-dimensional (3D) ray-tracing simulation. The 3D ray tracing is required to estimate the short-circuit current density (J_sc) because of an incident light angle distribution on a solar cell surface. In the simulation, the solar cell consists of a spherical Si solar cell with a diameter of 1 mm covered by an AR coating without surface texturing, and a hemispherical reflector cup with an aperture of 2 mm. The calculations of AR coating thickness (t) and refractive index (n) dependences on J_sc revealed that an optimum J_sc of 39.7 mA/cm² can be obtained at t=78 nm and n=2.0. Also, a higher design tolerance in an optical film thickness of the AR coating is demonstrated, which is especially advantageous for spherical Si solar cells which inherently have some technical difficulties in a uniform AR coating formation.

Spherical Si solar cells were fabricated based on polycrystalline Si spheres with a diameter of 1 mm produced by a dropping method. To decrease the cooling rates of Si spheres by decreasing the convection heat transfer to ambient, the Si spheres were dropped in a free-fall tower at a pressure of 0.2 atm. The conversion efficiency of low-pressure spherical Si solar cells was higher than that of normal-pressure spherical Si solar cells. Both Si spheres were polycrystals that consisting of crystal grains of about 200 µm. The distribution between electric active defects and the crystal quality were characterized by electron beam induced current (EBIC) measurements and transmission electron microscopy (TEM). By EBIC measurements, the low-pressure spherical Si solar cells were clearly observed to have the lower recombination sinks than the normal-pressure spherical Si solar cells. By TEM, the dislocation density in the low-pressure spherical Si solar cells was observed to be more reduced than that in the normal-pressure spherical Si solar cells. The dislocation density in the low-pressure Si spheres decreased because the reduction in the stress generated in the crystal grain.

A 2.4 and 5.2 GHz dual-band monopole antenna fabricated on a flexible parylene membrane for Bluetooth and wireless local area network (WLAN) modules was designed and fabricated. The measured results show a good antenna performance and agree very well with the design simulations in both return loss and radiation characteristics. Parylene is a flexible polymer film; it can be deposited on various substrates and peeled off after completing device processes as stand-alone thin-and-flexible-substrate-containing devices. Owing to its excellent physical and chemical properties such as its low dielectric constant, light weight, low conductivity, and inertness to chemicals, parylene is also suitable for fabrication of many other RF devices.

The electrical behavior of a decananometer double-gate (DG) metal–oxide–semiconductor field-effect transistor (MOSFET) with high-permittivity (high-κ) and stacked-gate dielectrics has been investigated using two-dimensional (2D) quantum numerical simulation. We show that in spite of a quasi-ideal control of the channel by the gates in the double-gate structure, the device performances can be significantly degraded when using high-κ dielectrics due to two important electrostatic limitations of high-κ materials: i) the fringing field-induced barrier lowering effect (FIBL) and ii) the presence of discrete fixed charges in the gate stack. The FIBL compromises the performance of short-channel devices when simultaneously increasing the dielectric constant and its physical thickness, whereas the charges trapped in the high-κ layer induce 2D potential fluctuations in the structure and degrades the subthreshold behaviour of the drain current.

Improvements in the electrical and structural properties of tetraethylorthosilicate (TEOS) SiO₂ films fabricated by plasma-enhanced chemical vapor deposition (PECVD) method were investigated using high-pressure H₂O vapor heat treatment. The density of interface trap states was reduced from 3.3×10¹² (initial) to 5.1×10¹⁰ cm^-2 eV^-1 by 1.3×10⁶ Pa H₂O vapor heat treatment at 260°C for 9 h. The density of fixed charges was also reduced from 6.1×10¹¹ to 1.3×10¹¹ cm^-2. The full width at half-maximum (FWHM) of the optical absorption band corresponding to vibration of Si–O–Si bonding was reduced from 82.9 to 78.1 cm^-1. Narrowing in FWHM of the Si 2p core level peak measured by X-ray photoelectron spectroscopy (XPS) was also observed. The reduction in the FWHM probably results from improvement of the Si–O bonding network.

A metal–oxide–semiconductor (MOS) capacitor was fabricated using Pt and a 6H–SiC substrate, and the interface state was evaluated in oxygen and hydrogen ambients under high-temperature conditions by the AC conductance technique. The relationship among interface state density (D_it), time constant (τ_it) and energy level (E_c-E_t) was obtained. The atmosphere was repeatedly changed between hydrogen and oxygen. Some levels of a narrow region near the conduction band, levels near 0.4 eV, and levels of a wide region near the band center were observed. D_it in the deeper levels increases in O₂ atmosphere and an increase in τ_it accompanies it. In H₂ atmosphere, D_it in the deeper wide energy region near the band center decreases. This change is almost reversible. In the range of 300–500°C, the D_it near the band center in O₂ atmosphere increases with temperature.

One-dimensional energy band simulation has been performed in order to understand chemical-vapor-deposition (CVD) diamond surface conductivity. It was found that the presence of shallow-level acceptors in the subsurface region and defect states at the surface causes a steep rise in the valence band toward the Fermi level, which causes accumulation of holes in the valence band in the subsurface and near-surface regions. An artificial negative charge accumulation (NCA) layer is introduced in the simulation to examine the effect of possible negatively charged adsorbates on surface conductivity. By adjusting the thickness of NCA layers, we have reproduced quantitatively both the surface conductivity change and Fermi-level change found in previous experiments [Kono et al.: Diamond Relat. Mater. 14 (2005) 459; Riedel et al.: ibid.13 (2004) 746].

Chemical mechanical polishing (CMP) is carried out using slurry particles in contact with a wafer and a pad. The size and distribution of particles between the wafer and the pad play a crucial role in achieving desired CMP performance. Polishing rates and friction forces were measured as a function of particle size and solids loading, and surface finishes of silica wafers polished with colloidal silica particles were analyzed to validate the polishing mechanism. On the basis of polishing rate, friction force and surface finish, polishing occurring at the pad-particles-wafer interface was analyzed and an interfacial contact model was proposed. Understanding the polishing mechanism using colloidal particles makes it possible to achieve desired CMP performance.

Low-resistive ultrashallow n⁺/p junctions were formed with Sb as a dopant by heat-assisted laser annealing. A process window platted against laser energy density and substrate temperature was obtained in which the sheet resistance was about 500 Ω/sq. while the junction depth was maintained at about 20 nm, which was equal to that of the as implanted condition. Cross-sectional transmission electron microscopy (XTEM) images showed that substrate heating before laser irradiation enhanced recrystallization of the amorphized layer. This causes non melt laser annealing using the complete recrystallization of the amorphized layer. From current–voltage characteristics of n⁺/p diodes in the reverse direction, leakage current reduction accompanying the heat-assisted annealing was confirmed.

After chemical mechanical planarization (CMP) processing of a Cu/low-k structure device, defects are often observed and some of them induce problems in manufacturing very large scale integrated circuit (VLSI) devices. As an example of defects, watermarks and protrusions on the Cu are detected. We found that the number of watermarks or protrusions is strongly affected by the cleaning conditions. The energy dispersive X-ray analysis (EDX) showed that these protrusions were composed of Cu and O. Moreover, atomic force microscopy (AFM) observations revealed that these protrusions grew during the storage time after the postcleaning. Electrochemical measurements also indicated that the protrusions were oxidized copper formed in the cleaning solutions due to the difference in corrosion current densities for various conditions of the Cu surface. Therefore, optimization of the post-CMP cleaning processing is a key issue for the reduction of defects such as protrusions.

Porous low-k materials are required for the construction of 45-nm-node LSI devices. However, the extremely low Young's modulus values of these materials result in the stress corrosion cracking (SCC) of the Cu interconnects during chemical mechanical planarization (CMP). We performed finite element method analyses of the stress at each step during the CMP. The results showed that the horizontal tensile stress was especially concentrated at the edges of the isolated fine wiring, and that higher tensile stresses appeared at the step of the barrier CMP. Moreover, the maximum values of the tensile stress increased with a decrease in Young's modulus in the low-k films. The cause of the horizontal tensile stress was the downward CMP pressure, which indented the low-k films. These results suggest that CMP with a lower downward pressure and an LSI structure with a Cu dummy pattern were effective for avoiding SCC.

In extreme ultraviolet (EUV) lithography, off-axis incident light on a reflective mask causes asymmetry of contrast in the edges of patterns with the critical dimension (CD) and a shift in the center position of the printed image of a line-and-space pattern, even if the line-and-space pattern has perfect symmetry and is correctly placed on a mask. In this study, we analyzed the asymmetry of the pattern edge contrast of line-and-space patterns using light intensity distribution functions which is obtained by taking inverse Fourier transform of the diffracted rays passing through the pupil. The results showed that, when only 0th- and ±1st-order diffracted rays pass through the pupil, the printed image tends to be symmetrical, regardless of any asymmetry in the amplitude and phase of the diffracted rays; but when 0th-, ±1st-, and ±nth-order (n≧2) diffracted rays pass through the pupil, the printed image tends to be asymmetrical.

The level of agglomeration in the cerium precursor and its effect on the physicochemical properties of the synthesized ceria particles and how these properties influence shallow trench isolation chemical mechanical polishing (STI CMP) performance were investigated. Two different types of ceria particles were synthesized from cerium precursors of different degrees of agglomeration. The crystallinity and particle size distribution of the synthesized ceria particles were markedly different between these two types of particles. The ceria particles synthesized from agglomerated cerium precursors had a smaller crystallite size than the other particles due to the incomplete decarbonation reaction, which resulted in large agglomerations of particles. The different physical characteristics of the ceria particles resulted in remarkable discrepancies between the STI CMP performances of the ceria slurries, such as the oxide removal rate, the selectivity and the uniformity.

A novel design for a low-power DC–DC converter using high-efficiency charge pump circuits for flat panel displays (FPDs) is proposed. The designed converter uses a variable voltage level shifter to decrease the conduction power loss of the charge pump circuit. The output voltage and the output current of the proposed DC–DC converter consisting of five charge pump circuits operating at 3.3 V are 17 V and 10 mA, respectively. The power efficiency and the ripple voltage of the designed charge pump circuits are 84.9% and 65 mV, respectively, which is a significant enhancement over those of Dickson charge pump circuits. These results indicate that our newly designed, low-power DC–DC converter using high-efficiency charge pump circuits holds promise for potential applications in mobile devices fabricated using FPDs operating at low voltages.

Amorphous GaN (a-GaN) films were deposited at a low temperature below 500°C by compound-source molecular beam epitaxy (CS-MBE). The relationship between excess Ga and its oxidation in the deposited GaN films is reported. X-ray photoelectron spectroscopy (XPS) revealed that the excess Ga in deposited films was oxidized in the air and converted to gallium oxide. By increasing the substrate temperature, the total amount of gallium oxide in the deposited films decreased due to the reduction of the excess Ga. Cathodoluminescence (CL) intensity from the UV to the blue spectral regions increased with as the amount of gallium oxide in the deposited films decreased.

The DC and RF characteristics of short-gate high-electron-mobility transistors (HEMTs) formed on four AlGaN/GaN wafers grown on sapphire substrates with different crystal quality have been investigated. Atomic force microscopy observation revealed many pits and trenches on the AlGaN surface, and the morphology of each sample was distinct. There were also differences in electron mobility and sheet carrier concentration. However, the gate length dependences of the measured transconductance and cut-off frequency were virtually the same. For a more detailed investigation, we subtracted the source resistance and AlGaN thickness contributions from measured DC and performed a delay time analysis for the RF characteristics. The results indicate that the intrinsic performance of HEMTs was independent of the surface morphology and that effective electron velocity ranged from 1.4 to 1.8×10⁷ cm/s.

Front and back illumination of a metal–semiconductor–metal structure on a 2-µm-thick GaN layer showed obvious differences in the spectral responsivity in the wavelength range of 300–500 nm. Pt/Au interdigitated electrodes on an unintentionally doped n-GaN were confirmed to be of extremely low leakage Schottky type, and simulations of the electrostatic potential distribution have revealed that the depletion regions do not prevail throughout the thick GaN layer even at a bias of 10 V. The difference observed in the wavelength region shorter than the fundamental absorption edge is due to incomplete depletion of the GaN layer off the Schottky contacts in conjunction with short optical penetration depths, while the back-incidence responsivity in the longer wavelength region reflects extrinsic optical absorptions characteristic to the epitaxial crystal.

The buffer thickness dependence of strain field distribution was investigated in SiGe heterostructures by micro-Raman spectroscopy. Crosshatch-like strain fluctuations were clearly observed in strained-Si and SiGe buffer layers, and the fluctuation wavelength was found to increase almost linearly with increasing buffer thickness. It was also found that SiGe homoepitaxial growth on planarized SiGe buffer layers gave rise to crosshatch roughness on the surface with almost the same morphology as the strain distribution, indicating that the strain fluctuation caused the roughness formation associated with growth kinetics. The strain fluctuation still remained on the buffer layer thicker than 7 µm, which should be taken into consideration for device applications.

The effect of the plasma nitridation of the SiO₂(1.2 µm)/Ti-foil substrate on the quality of the GaN thin films grown by the reactive evaporation method was investigated. Highly c-axis oriented GaN(0002) thin films were obtained. In contrast to the case without nitridation, the quality of GaN(0002) crystals with substrate nitridation was improved. This can be attributed to the flattened surface of the intermediate SiO₂ layer and the passivation of Si and O dangling bonds during the substrate nitridation. Photoluminescence measurements show that the emission at the near-band edge (about 3.15 eV) is detected when nitridation occurs whereas only deep level emission is observed when no nitridation occurs.

ZnO films were fabricated on glass substrates by metal organic decomposition (MOD). X-ray diffraction, reflection high-energy electron diffraction, and energy-dispersive X-ray spectroscopy revealed that the obtained ZnO films are stoichiometric polycrystalline with a wurtzite structure. Strong near-band-edge emission in the UV region without any deep-level emissions was observed from the ZnO films at room temperature. The results show that MOD is a promising growth method for obtaining high-quality ZnO films, which paves the way for the fabrication of electronic and optoelectronic devices using ZnO at low cost.

In this study, we have systematically investigated the thermal stability of strained silicon–germanium (SiGe) and bulk Si p-metal–oxide–semiconductor field-effect transistors (MOSFETs) in dynamic-threshold (DT) mode and standard mode operation, respectively, from 77 to 400 K. Possessing advantages of good carrier confinement and higher carrier mobility in the strained-SiGe material, SiGe DT-MOSFETs exhibit not only enhanced drive current but also better thermal stability for threshold voltage, substrate sensitivity, and low-frequency noise over the entire operating temperature range than their counterpart Si devices. That is very important for radio-frequency micropower analogue and digital integrated circuits applications.

The surface structure and magnetic state of ultrathin Co films with Volmer–Weber growth (V–W growth) on α-Al₂O₃(0001) have been investigated as a function of Co thickness. Due to V–W growth, Co forms particles with a diameter of approximately 5 nm. Originating from the particle structure, the magnetic state of Co is superparamagnetism below a thickness of 1.1 nm, and evolves to ferromagnetism as the Co thickness increases. The transition occurs for a wide thickness range, i.e., 1 to 3 nm. Due to the wide-range transition, it enables to observe the coexistence of superparamagnetism and ferromagnetism at a thickness of around 2 nm. The wide range transition of the magnetic state is explained by the slow coalescence of Co particles. The structural transition would be dominated by the large interface energy between Co and α-Al₂O₃(0001).

A jumping search method is applied to the design of a normal-incidence antireflection coating for the 400–750 nm spectrum regions. A search algorithm that includes jump turning and optimal elimination is successively used to optimize the visible antireflection performance and to simplify the design structure. It is shown that the average visible reflectivity of the jumping search antireflection coating designs of a two-material 40-layer system is reduced to approximately 0.064–0.079%, and the final designs are reduced to 14–16-layered structures.

A blue light 420–485 nm output with a linewidth of 0.3 nm and a pump-to-signal energy conversion efficiency of up to 29% at an injected seeding energy of 400 µJ has been achieved. The optimum crystal length from 4 to 2 mm at a pump intensity of 500 MW/cm² is studied using computer simulation. The experimental designs of the optical parametric generation of type-II noncritical-phase-matched LiB₃O₅ and the amplification of type-I critical-phase-matched β-BaB₂O₄ are presented in detail. The experimental setup is pumped using the 355 nm 30 ps-duration output pulse of an active-passive model-locked Nd:YAG laser, is reported.

We generated frequency-tunable terahertz (THz) waves with stable carrier-envelope phases (CEPs) by exciting a photoconductive antenna with intensity-modulated laser pulses. The modulation of a laser pulse by using a birefringent crystal and a grating pair produced two chirped laser pulses with a fixed time separation, or in other words, an intensity-modulated laser pulse. The stability of the CEP of the THz waves was governed by the stability of the time separation of the two laser pulses generated by the birefringent crystal. Because the crystal thickness was fixed, the CEP of the THz wave was stable and not affected by the mechanical vibration of the optical components. We also demonstrated the CEP control of the THz wave.

We prepared an electrically poled film of poly(β-phenylpropyl L-aspartate) and performed second-harmonic generation (SHG) measurements to detect its thermo-reversible helix–helix transition between right-handed α_R- and left-handed ω_L-helices. In this poled film, the helical molecules are highly oriented and SHG intensity profiles were measured with various angular combinations of polarizer and analyzer. The in-plane anisotropies of SHG were interpreted by the polar C₆ symmetry through the temperature region from 25 to 160°C, and then clearly reflected the thermo reversible transformation between the α_R-helix and ω_L-helix at 155°C on heating and at 142°C on cooling. The β_lmn ratios of molecules evaluated from the χ components were found to be characteristic for each helical conformation; β_XZX,α:β_ZXX,α:β_ZZZ,α=8.7:2.0:1.0 for the α_R-helix (at 145°C) and β_XZX,ω:β_ZXX,ω:β_ZZZ,ω=5.3:3.3:0.7 for the ω_L-helix (at 160°C).

The codoping method is applied to fabricate efficient blue organic light-emitting diodes (OLEDs). With the same structure of indium–tin oxide (ITO)/N,N'-bis(1-naphthyl)-N,N'-diphenyl-1,1'-biphenyl-4,4'diamine (NPB)(80 nm)/light-emitting layer (30 nm)/tris-(8-hydroxy-quinoline)aluminum (Alq₃) (20 nm)/LiF (1 nm)/Al (120 nm), a set of three devices was manufactured for comparison. For Devices 1, 2, and 3, the light-emitting layers are 9,10-di(2-naphthyl)anthracene (ADN):4,4'-(1,4-phenylenedi-2,1-ethene diyl)bis[N,N-bis(4-methylphenyl)-benzenamine] (DPAVB) (1 wt %), ADN:2,5,8,11-tetra-(t-butyl)-perylene (TBPE) (1 wt %), and ADN:DPAVB (0.3 wt %):TBPE (0.7 wt %), respectively. It is found that the codoped Device 3 has the highest maximum luminance, Electroluminescence (EL) quantum efficiency and color saturation. Further study on the effect of the codopants was through a relative photoluminescence (PL) quantum efficiency measurement. The result shows that the relative PL efficiencies of Devices 1, 2, and 3 are 15.6, 19.3, and 24%, respectively, as determined using an integrating sphere system excited at 375 nm. The codoping method improves the EL efficiency intrinsically. Codopants of the heterogeneous light-emitting molecules may decrease the possibility of self-quenching from the interaction of the homogenous molecules at the same total doping concentration. Furthermore, the decrease in the interaction of homogenous molecules suppresses the light emission from the aggregations thus narrowing the emission spectrum, and results in saturated blue light emission.

We examined three commercially available Fabry–Perot laser diodes with different oscillation wavelengths (660, 800, and 1550 nm) to suppress side fringes appearing under gain- or loss-modulation operation in low-coherence interferometric fringe measurements. For the shorter-wavelength laser diodes (660 and 800 nm), the amplitude of the maximum side fringe was effectively suppressed to 4% of the main fringe amplitude by utilizing the linewidth broadening effect of gain-switched laser diodes due to the generation of frequency chirping, and also by superimposing a multimode spectrum from another laser diode of the same kind. For the 1550 nm laser diode, however, the above approaches were less effective because noticeable gain saturation in the laser diode disturbed the transition of the spectral shape from a characteristic multimode spectrum to a continuum.

In this paper, we introduce a new four-ary run-length-limited (d,k)=(2,12) code for multilevel optical recording channels. This code has a simple one-codeword look-ahead encoder and a sliding block decoder. The code rate is 4/5 (bits/symbol), and the density ratio is 2.4 bits/T_min, which is larger than that of binary codes used for typical optical recording systems. The encoding and decoding rules of the proposed code are simple and easy for implementation. With the bit error rate (BER) performance evaluation, this code shows feasible to be applied in high-density multilevel optical storage systems.

We report numerical simulation of conversion efficiency and beam quality factor (M²) in second harmonic generation taking into account diffraction and pump depletion. Compared with the results in Boyd and Kleinman's study [J. Appl. Phys. 39 (1968) 3597], which treated the no-depletion case, with the presence of pump depletion, conversion efficiency is maximized under substantially weaker focusing. Depletion also leads to the degradation of the beam quality factor of both the unconverted fundamental and the generated second harmonic radiation.

Soliton switching in a three-core Kerr-coupler is studied numerically by taking into account intermodal dispersion (IMD), third-order dispersion and higher-order nonlinear effects, like the Raman and self-steepening effects. It is shown that the weak coupling regime is the most suitable for the purpose. Also, it is shown that IMD has negligible influence on switching, while, out of the higher-order nonlinear effects, Raman intrapulse scattering has the dominant effect.

The effects of rf power and substrate temperature on the crystal quality of ZnO thin films deposited on sapphire (001) substrates by rf magnetron sputtering are investigated by X-ray diffraction (XRD), scanning electron microscopy (SEM), atomic force microscopy (AFM) and photoluminescence (PL) spectroscopy analyses. A ZnO thin film with the highest crystal quality is obtained by rf magnetron sputtering at a substrate temperature of 600°C and an rf power of 80 W. The crystal quality of the film is markedly improved by annealing. Excess rf power deteriorates the crystallinity and surface roughness of the film. PL spectroscopy analysis results confirm that rf magnetron sputtering yields ZnO films with a low density of crystallographic defects. Therefore, high-quality ZnO films can be obtained by optimizing the substrate temperature and rf power when using the rf magnetron sputtering technique and by annealing.

Y-branch-type polymeric optical waveguides with a large core size of 1000 µm, branching angles from 2 to 10° and a branching top-part radius of 200 µm were fabricated by hot embossing. An output power ratio of 1:1 from each output port and an excessive loss of ∼1.0 dB depending on branching angle were realized. A compact size Y-branch-type waveguide device with a wide branching angle was proposed.

The structure and optical band-gap energies of Ba_0.5Sr_0.5TiO₃ (BST0.5) thin films prepared on SiO₂/Si and fused quartz substrates by RF magnetron plasma sputtering were studied in terms of deposition temperature and film thickness. Highly (100)-oriented BST0.5 thin films were successfully sputtered on a Si substrate with an approximately 1.0-µm-thick SiO₂ layer at a deposition temperature of above 450°C. The optical transmittance of BST0.5 thin films weakly depended on the magnitude of X-ray diffraction (XRD) peak intensity. This is very helpful for monolithic integration of BST0.5 films for electrooptical functions directly onto a SiO₂/Si substrate. The band-gap energies showed a strong dependence on the deposition temperature and film thickness. It was mainly related to the quantum size effect and the influence of the crystallinity of thin films, such as grain boundaries, grain size, oriented growth, and the existence of an amorphous phase. The band-gap energy values, which were much larger than those of single crystals, decreased with the increase in the deposition temperature and the thickness of BST0.5 thin films. The band-gap energy of 311-nm-thick amorphous BST0.5 thin film was about 4.45 eV and that of (100)-oriented BST0.5 thin film with a thickness of 447 nm was about 3.89 eV. It is believed that the dependence of the band-gap energies of the thin films on the crystallinity for various values of deposition temperature and film thickness means that there could be applications in integrated optical devices.

Highly-pure amorphous silica slides were implanted by 200 keV Ag ions with doses ranged from 1×10¹⁶ to 2×10¹⁷ ions/cm². Optical absorption spectra show that Ag nanoclusters with various sizes have been formed. Enhancement of surface enhanced Raman scattering signal by a factor up to about 10³ was obtained by changing the Ag particle size. The silica was damaged by the implanted Ag ions, and the large compression stress on the silica leads to the shift of Raman peaks. New bands at 1368 and 1586 cm^-1, which are attributed to the vibration of Ag–O bond and O₂ molecules in silica, are observed in the samples with doses higher than 1×10¹⁷ ions/cm².

A new group of ABO₃-type lead-free piezoelectric ceramics, [Bi_0.5(Na_1-xAg_x)_0.5]_1-yBa_yTiO₃, was developed, and the corresponding invention patent was submitted. The ceramics were synthesized by the conventional ceramic sintering technique using electronic grade raw materials, and the preparation techniques are very stable and convenient. The crystalline phase, microstructure and electric properties of the ceramics were also investigated. All the ceramics have high densities of about 5.70–5.84 g/cm³, which are more than 95% of the theoretical values. This system provides high piezoelectric performances: d₃₃=168 pC/N, k_p=0.31 when x=0.06, y=0.06. Moreover, the samples doped with a moderate amount of Mn could increase the mechanical quality factor Q_m and reduce the dielectric loss tgδ simultaneously. The temperature dependence of piezoelectric properties measured show that at up to 180°C, d₃₃ can still remain 126 pC/N for [Bi_0.5(Na_0.96Ag_0.04)_0.5]_0.90Ba_0.10TiO₃ ceramics, which has a d₃₃ of 137 pC/N at room temperature.

In this study, different proportions of silver–palladium alloy used as an inner electrode are adopted to fabricate (Zn,Mg)TiO₃-based multilayer ceramic capacitors. Effects of sintering temperature, and measuring frequency on the dielectric properties of the samples with different proportions of the Pd–Ag inner electrode are investigated. The continuity of the inner electrode and the proportion of Pd–Ag of the inner electrode of samples sintered at different temperatures play important roles in determining the dielectric properties.

Bismuth ferrite (BiFeO₃) thin films were fabricated by depositing sol–gel solutions on Pt/Ti/SiO₂/Si(100) structures. X-ray diffraction (XRD) showed that a polycrystalline phase and also a small fraction of a secondary phase, Bi₂Fe₄O₉, were present in the film. The nonperovskite secondary phase decreased with increasing thickness, which showed the influence of volume effects on the film. Improved leakage current density and enhanced polarization in BiFeO₃ films were observed. A 400-nm-thick film showed a leakage current on the order of 10^-8 A/cm² at room temperature. The remanent polarization was approximately 90 µC/cm² at 80 K and the piezoelectric coefficient d₃₃ was approximately 50 pm/V.

The size dependence of the dielectric properties of BaTiO₃ nano-grains was investigated using films prepared by RF-plasma chemical vapor deposition. BaTiO₃ nanoparticles were directly deposited on Pt/Al₂O₃/SiO₂/Si substrates; then they were annealed at temperatures between 600 and 900°C in order to control the grain size. The dielectric constants of BaTiO₃ nano-grains smaller than 30 nm were measured to be 200–300, and reached approximately 1000 for grain sizes of 50–60 nm. According to capacitance–voltage (C–V) curves and hysteresis loops, the phase transformation from paraelectric to ferroelectric occurred at a grain size of approximately 40 nm. The dielectric constants of the films were considerably stable against the temperatures ranging from 5 to 160°C, even though the dielectric constants were approximately 1000. The breakdown electric fields of the films increased with decreasing grain sizes and a value of 649 kV/cm was achieved at a grain size of 27 nm.

The interaction process between elastic and dielectric energy in a piezoelectric transducer, which has been treated on the basis of lumped parameters in the framework of electrical equivalent circuit, is treated on a distributed-parameter basis. The concept of coupling between elastic and dielectric energy modes is introduced, and the superposition of the energy mode after the coupling is considered, which allows us to determine the frequency characteristics of the transducer with an electromechanical coupling. Mathematically, the coupling effect is represented by the exponential of a matrix, and the superposition of the energy mode is calculated with the infinite geometric series of a matrix, Neumann series, which provides physically reasonable results.

BaTiO₃ thin films (<200 nm) were directly deposited on Si wafers using an oxygen-ion-beam-assisted deposition technique. Si surfaces were oxidized during the deposition. The thicknesses of the formed SiO₂ layers were proportional to the ion beam current and were almost independent of the ion beam energy and the N₂ flux. The BaTiO₃ films with thin SiO₂ layers (20 nm) had larger remanent polarizations (∼2 µC/cm²) than the BaTiO₃ films with thick SiO₂ layers (25 nm). Many prepared films had a coercive field of about 50 kV/cm.

We have fabricated a plastic liquid crystal display using the pixel-isolated liquid crystal (PILC) mode and characterized the mechanical stability of the sample against mechanical pressure and bending. Since the liquid crystal molecules are fully isolated in the pixels by the phase-separated polymer walls in the PILC mode, the electrooptic properties of the sample show good mechanical stability. In addition, a polymer layer formed on the upper substrate enhances the adhesion between two plastic substrates.

The temperature dependence of the azimuthal anchoring strength of the nematic liquid crystal (LC) 4'-n-pentyl-4-cyanobiphenyl (5CB) on parallel grooved glass substrates has been studied. The U-shaped grooves are prepared by reactive ion etching. Two parallel grooved substrates with chiral doped 5CB sandwiched in between are used to form an LC cell. The azimuthal anchoring strength in the nematic temperature range is determined by measuring the twist angle in the LC cells using an optical method at two laser wavelengths. The anchoring strength decreases steadily with increasing temperature. The change in the anchoring strength is attributed to the change in the elastic constant K₂₂ unless the temperature is close to the clearing point.

We have studied the crystal structure of the misfit-layered crystal Bi_1.91Sr₂Rh_1.77O_x by electron diffraction and high-resolution electron microscopy. This compound consists of two interpenetrating subsystems of a RhO₂ sheet and a distorted four-layered rock-salt-type (Bi,Sr)O block. Both subsystems have common a-axes, c-axes, and β-angles with a=5.28 Å, c=29.77 Å, and β=93.7°. On the other hand, the crystal structure is incommensurated parallel to the b-axis, b₁=3.07 Å for the RhO₂ sheet and b₂=4.88 Å for the (Bi,Sr)O block. The axis ratio, b₁/b₂∼0.63, characterizes the structural analogue as [Bi_1.79Sr_1.98O_y]_0.63[RhO₂]. This compound has two modulation vectors, q₁=a^*+0.63b₁^* and q₂=0.17b₁^*+c^*, and the superspace group is assigned as the Cc(1β0,0µ1) type from electron diffraction patterns. High-resolution images taken with the incident electron beam parallel to the a- and c-axes clearly exhibit modulated displacive as well as compositional atomic arrangements. A tentative structure model has been proposed, and the calculated image reproduces observed characteristic features reasonably well.

MnSi layers in Ge-doped MnSi_∼1.7 increased with increasing Ge content up to x=0.00133, began to break at x=0.00265 and finally disappeared at x=0.00530. An experimental equation for the growth of MnSi was proposed for the interval between the MnSi layers and amount of doped Ge content. The crystallinity of Ge-doped MnSi_∼1.7 increased initially with increasing doped Ge content and saturated at high Ge content. Thermoelectric transport properties along the c-axis of Ge-doped MnSi_∼1.7 were measured as a function of Ge content at room temperature. Electrical conductivity and thermoelectric power of Ge-doped MnSi_∼1.7 were compared to those of Al-doped MnSi_∼1.7 in our previous work. A maximum in the electrical conductivity and a minimum in the thermoelectric power of Ge-doped MnSi_∼1.7 were observed at x=0.00133, reflecting a change in hole density which was influenced by the volume ratio of MnSi. Hole mobility depended on the existence of MnSi layers and/or of interfaces between MnSi_∼1.7 and MnSi and on the crystallinity of MnSi_∼1.7. The thermal conductivity of Ge-doped MnSi_∼1.7 had a maximum at x=0.00053. The increase in thermal conductivity at low Ge doping can be explained by the increase in the amount of MnSi segregated in doped MnSi_∼1.7, while the decrease at high Ge content was caused by the increase in phonon scattering of Ge. A maximum figure of merit of Ge-doped MnSi_∼1.7 was obtained at x=0.00974, reflecting a maximum power factor.

Europium-doped yttrium oxysulfide (Y₂O₂S:Eu) phosphor was successfully synthesized at room temperature from yttrium oxide, europium oxide, and sulfur. The method employs high-energy ball milling to enable a substitution reaction between oxygen and sulfur, unlike conventional methods, such as heating in a sulfurizing atmosphere. It was found that the material is fluorescent through X-ray irradiation, and the luminescence spectra exhibit four peaks in the wavelength region from 500 to 800 nm.

The purpose of this work was to develop a design methodology for the fabrication of waveguides in LiNbO₃ operating in the single-mode regime at several wavelengths, with specific characteristics required to optimize integrated devices. To achieve this, it is necessary to obtain the relations between the optical characteristics of the waveguides and their respective fabrication conditions, and to introduce models of the waveguide formation process. The relations between fabrication conditions and optical characteristics of planar waveguides realized by titanium in-diffusion are documented extensively in the literature. However, reports on the characterization of waveguide fabrication processes, performed in a systematic way, could not be found, resulting in the need to combine information from several sources. Discrepancies among results from different researches are evident, resulting from different experimental methodologies and calibrations of equipment. Therefore, aiming at extracting a consistent data set, optical characterization techniques for the refractive index profile were employed to study series of samples.

The stereochemical control of surface reactions is one of the ultimate goals of surface scientists. An oriented-molecular-beam technique based on the Stark effect of a molecule in an inhomogeneous hexapole electrostatic field is a potential tool for achieving such a goal. This technique allows us to select a specific rotational quantum state and also an orientation of a reagent molecule. We have designed, built and tuned up a new UHV-compatible oriented-molecular-beam machine for the elucidation of the reaction dynamics on surfaces and for surface manufacturing application. In the dissociative adsorption of CH₃Cl on a Si{100} surface, we found a dynamical steric effect on the initial sticking probability (S₀) using the new machine. S₀ in Cl-end collision is larger than that in CH₃-end collision at an incident energy of 120 meV. To our knowledge, this is the first measurement of the steric effect in the chemisorption of a molecule on a Si surface.

The surface of implantable biomaterials directly contacts the host tissue and is critical in determining biocompatibility. To improve implant integration, interfacial reactions must be controlled to minimize nonspecific adsorption of proteins, and tissue-healing phenomena can be controlled. The purpose of this study was to develop a new method of functionalizing titanium surfaces by plasma treatment. The covalent immobilization of bioactive organic molecules and the bioactivities in vitro were assessed by transmission electron microscopy (TEM), atomic force spectroscopy (AFM), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), and 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay as indices of cellular cytotoxicity. Argon plasma removed all of the adsorbed contaminants and impurities. Plasma-cleaned titanium surfaces showed better bioactive performances than untreated titanium surfaces. The analytical results reveal that plasma-cleaned titanium surfaces provide a clean and reproducible starting condition for further plasma treatments to create well-controlled surface layers. Allylamine was ionized by plasma treatment, and acted as a medium to link albumin. Cells demonstrated a good spread, and a wide attachment was attained on the Albu-Ti plate. Cell attachment and growth were shown to be influenced by the surface properties. The plasma treatment process plays an important role in facilitating tissue healing. This process not only provides a clean titanium surface, but also leads to surface amination on plasma-treated titanium surfaces. Surface cleaning by ion bombardment and surface modification by plasma polymerization are believed to remove contamination on titanium surfaces and thus promote tissue healing.

Ba_0.65Sr_0.35TiO₃ thin films have been prepared by RF magnetron sputtering. The crystallization and microstructure of the films were characterized by X-ray diffraction (XRD), scan electronic microstructure (SEM) and atom force microstructure (AFM). As-deposited thin films were found to be amorphous. The more intense characteristic diffraction peaks and improved crystallization can be observed in (Ba,Sr)TiO₃ (BST) thin films deposited at high temperatures and annealed at higher than 650°C. Optical constants were determined from transmittance spectra by using the envelope method. The refractive index increased from 1.778 to 1.961 as the substrate temperature increased from 560 to 650°C. Both the refractive index and extinction coefficient increased with annealing temperature. The refractive index and extinction coefficient increased when the oxygen-to-argon ratio increased from 1:4 to 1:1. The dispersion of relation of the extinction coefficient vs wavelength was also investigated. The optical band gap of BST thin films was found to be about 3.56 eV, which decreased apparently with increasing annealing temperature.

Lead zirconate (PbZrO₃) thin films are grown on Pt(111)/Ti/SiO₂/Si(100) substrates by sol–gel using precursor solutions with stoichiometric and 20 mol % excess Pb. Films with no preferred orientation and [111] pseudocubic texture (denoted as [111]_pc) are obtained by changing the drying temperature at the pyrolysis stage. Randomly oriented films were found to have a two-phase microstructure consisting of large rosette-type perovskite regions with a nanocrystalline phase located in between. The ratio of the rosettes increase to cover the entire surface of the films with the increase of Pb content. The films with [111]_pc orientation have a uniform microstructure with micron size grains. The electrical properties of the films were influenced markedly by the microstructure and orientation of the films. The [111]_pc oriented films exhibit a square-like double hysteresis loop with maximum polarization (P_max) reaching 61×10^-6 C/cm² under 550 kV/cm, whereas stoichiometric films with no preferred orientation have a P_max of 36×10^-6 C/cm² with slimmer hysteresis curves.

Superlattice thin films of the perovskite-type oxide proton conductor SrZr_0.95Y_0.05O₃/SrTiO₃ was fabricated by pulsed laser deposition. Their structural and proton transport properties were reported. X-ray diffraction analysis and selected area electron diffraction revealed that the thin films were epitaxially grown on MgO(001) substrate. High-density edge dislocations and a columnar structure were observed in the films by high-resolution electron microscopy. The in-plane electrical conductivity of the thin films was determined by impedance spectroscopy. The contribution of proton transport to the total conductivity of the films was confirmed by H₂O/D₂O exchange measurement. The conductivity of superlattice films was increased by introducing heterointerfaces. The high activation energy (E_a=1.0 eV) was explained by the grain-boundary effect of the columnar structure in the films.

Subsurface dopant atoms are investigated by scanning tunneling microscopy (STM) and barrier height (BH) imaging. The obtained BH image shows a local reduction in BH -(1–2) eV at positively charged donor sites while it shows a local increase in BH (1–5) eV at negatively charged surface Ga vacancies. Our results indicate that the charge-sensitive imaging can be accomplished by utilizing the STM–BH technique. A simple one-dimensional simulation of tunneling current through a STM junction on the GaAs(110) surface indicates that the results of BH imaging well reflect the charge state of the subsurface dopant or charged surface defects under appreciable electron accumulation conditions.

The magnetic field distribution around sub-µm-wide current paths was investigated by magnetic force microscopy (MFM) as a candidate for the current mapping in fine structures. In particular, an undesirable electrostatic force working between an MFM tip and the current path was dynamically eliminated utilizing an extra ac bias to observe the magnetic field correctly. We observed magnetic force signals around current paths consisting of branching or closely aligned metal wires, and the results were compared with results of the numerical simulation of the magnetic field. We found that spatial resolution of magnetic force detection by our method was better than 0.2 µm. The calculation results also indicate that the oscillatory motion of the MFM tip in tapping operation influences spatial resolution.

Projected ranges of electrons in some elemental solids and gases of atomic number between 4 and 18 (Be, C, N, Na, Al, Si and Ar) have been obtained by a method based on a discontinuous slowing down model in the slowing down of an electron from an certain energy E to a minimum energy E_min. The incident energies of electrons considered are from 0.06 to 3 MeV. The obtained projected ranges of electrons have been compared to the values in previous work.

Plasma processing is widely used in the mass production of industrial devices. A RF sputtering can be used for insulator and dielectric materials. When capacitive coupled plasma (CCP) is used for RF sputtering, the state of discharge has to be researched. The relationship between the ratio of areas of the electrodes and the ratio of DC bias voltages in magnetron sputtering was investigated, because it determines the acceleration voltage for ions, and may play an important role in sputtering. The imperfection of plasma control leads to problems in mass production. The relationship between the ratio of areas of the electrodes and the ratio of DC bias voltages in magnetron sputtering was investigated in this study. Moreover, the simulation results of some models that are different in chamber size or gas pressure were obtained. These results were compared with the experimental results and the difference was discussed. The results of simulations regarding the relationship between a bias voltage of a target (V_dc) and gas pressure with the same chamber, and between V_dc and chamber size correspond to the experimental results qualitatively. However, the changes are much less than in the experiments, especially with respect to chamber size. Considering distributions of neutral gases and radicals, the accuracy for V_dc may be improved.

Next-generation lithography will require an extreme-ultraviolet (EUV) light source that ensures high radiation intensities at a wavelength of around 13.5 nm. The characteristics of pinch dynamics and emission in this spectral range were studied experimentally for xenon capillary Z-pinch plasma driven by different dI/dt discharge current pulses. The pinch dynamics of the capillary Z-pinch plasma were examined by employing a high-speed camera, and the spectral emission from plasma was inspected using an EUV photodiode, a mini calorimeter and spectrometers. Our results confirm that high-dI/dt discharge current has better performance in comparison with the low one in terms of plasma dynamics, EUV power output and debris generation.

The discharge characteristics of high-efficiency AC-plasma display panels (PDPs) using a separate discharge mode were investigated using both numerical simulation and experiments. New types of segmented sustain electrodes have been proposed in order to increase the luminous efficiency with low power consumption and without major changes in the manufacturing process. We have obtained the higher efficiency with a wider separation between the left-side and right-side discharges in a cell. This effect can be explained in that a wider gap between the tiles is beneficial because the discharge remains confined to the walls, thus efficiently exploiting the phosphor on the walls. The discharge images of an intensified charge-coupled device (ICCD) camera in a tiled sustain electrode structure support these results.

A phenomenological model has been developed to simulate the feature profile evolution for nanometer-scale control of the profile and critical dimension during plasma etching. Attention was focused on the feature profile evolution of infinitely long trenches etched in Si with chlorine chemistries. The model takes into account the transport of ions and neutrals in microstructures, multilayer surface reactions through ion-enhanced etching, and the resulting feature profile evolution, where the transport is analyzed by a two-dimensional particle simulation based on successively injected single-particle trajectories with three velocity components. To incorporate an atomistic picture into the model, the substrates are taken to consist of a large number of small cells or lattices in the entire computational domain of interest, and the evolving interfaces are modeled by using the cell removal method; the Si atoms are allocated in the respective two-dimensional square lattices of atomic scale. Moreover, the Monte Carlo calculation is employed for the trajectory of incident Cl⁺ ions that penetrate into substrates. The present model has a prominent feature to phenomenologically simulate the multilayer surface reaction, the surface roughness, and also the feature profile evolution during etching. The etching of planar Si substrates was simulated for a test of validity of the present model, showing the structure of surface reaction layers, the distribution of Cl atoms therein, and the surface roughness that depend on incident neutral-to-ion flux ratio and ion energy. The etch yield as a function of neutral-to-ion flux ratio for different ion energies gave a similar tendency to the known experimental data, indicating that the present model properly reflects synergistic effects between neutral reactants and energetic ions in the ion-enhanced etching. The feature profile evolution during etching was then simulated for sub-100 nm line-and-space patterns of Si, exhibiting the reactive ion etching (RIE) lag that occurs depending on neutral-to-ion flux ratio and ion energy. The degree of RIE lag was found to be more significant at higher flux ratios and higher energies, being associated with the difference in surface chlorination at the feature bottom; in effect, for narrow pattern features of the order of sub-100 nm, the bottom surfaces tend to starve for neutral reactants owing to severe effects of the geometrical shadowing.

The characteristics of the phenomena caused by laser irradiation to an electrostatic probe in plasmas are studied to avoid the disturbance of the laser photodetachment signals for negative ion density measurement. In helium–hydrogen and hydrogen–methane plasmas, a probe surface ablation phenomenon was observed as an anomalous excess electron current in response to the laser irradiation to the electrostatic probe, while the phenomenon was not observed in pure hydrogen plasmas. Contaminations of the probe surface appear to be the mechanism causing the ablation phenomena. In order to clean the probe surface, a filament-type heated probe, which is the same type of conventional emissive probe, is applied to the laser photodetachment technique. When the surface is cleaned by heating the probe, the ablation phenomenon disappears, and the negative ion density can be evaluated at a sufficiently high laser pulse energy to saturate the photodetachment rate of negative ions. The method developed in this paper is useful for the measurement of negative ion density in plasmas where the probe surface is easily contaminated.

Electron energy probability functions (EEPFs) are investigated in inductive C₂F₆/Ar and C₂F₆/O₂ discharges. The structure of the EEPFs changes from bi-Maxwellian to Druyvesteyn-like distributions through a Maxwellian one with increasing Ar content, whereas the EEPFs form the bi-Maxwellian structure at any oxygen content except for pure oxygen. Dependences of both the fluorine atom density measured by actinometry and the relative variation of CF₂ density determined from an optical emission at 251.9 nm on the dilution gas content are also investigated. The densities of F and CF₂ are independent of the Ar content, while they depend strongly on the oxygen content.

Spectral imaging for electroluminescence (EL) characterization of a light-emitting diode based on blends of poly[2,7-(9,9-di-n-octylfluorene)] (PFO) and poly[2,7-(9,9-di-n-octylfluorene)-alt-(1,4-phenylene-((4-sec-butylphenyl)amino)-1,4-phenylene)] (TFB) was performed using a two-dimensional microspectroscopy imaging system. We found that EL spectral images changed markedly with increasing applied voltage. Such a variation is presumed to have originated from the transfer of emission sites in the polymer blend layer.

A diazonaphthoquinone/novolak resist on a 42-alloy substrate was irradiated by the second harmonic wave (532 nm) of a pulsed Nd³⁺YAG laser. The resist was removed, despite the existence of hexamethyldisilane (HMDS). There was no apparent damage to the substrate. In contrast, the resist on a Si wafer could not perfectly be removed. In some cases, there was damage to the substrate. The peeling strength of the resist with HMDS was about three times than that without HMDS. The width of the resist removed by laser irradiation without HMDS was 1.02–1.03 times larger than that with HMDS. The use of this process will benefit the environment, since expensive and toxic chemicals are not used.

Organic double-layer p–n junction photovoltaic devices using copper phthalocyanine (CuPc) as a p-type layer and 3,4,9,10-perylenetetracarboxylic-bis-benzimidazole (PTCBI) as an n-type layer were prepared. Molecular orientation was introduced in the PTCBI layer by the rubbing method, whereas no orientation was introduced in the CuPc layer. The power conversion efficiency of the device under white polarized light illumination was 0.1%. Photocurrent anisotropy was observed at the spectral region of 400 to 750 nm. The photocurrent ratio when the light was parallel and perpendicularly polarized against the molecular orientation axis was as high as 1.57 at 540 nm. The observed photocurrent anisotropy, however, was much smaller than that expected from the anisotropy of optical absorption of the aligned PTCBI layer, that is 4.76, because of filtering effects.

The technique of X-ray scattering topography (XST) is useful for imaging imperfect single crystals that cannot be properly observed by ordinary X-ray topography and has been successfully applied to the mapping of crystal-lattice orientation. Using scanning-type XST with a finely collimated X-ray beam and a pinhole slit and applying the technique of Laue orientation imaging using a charged-coupled device (CCD) camera, gnomonic projection, and Hough transform data processing, we have taken pictures of the scattering topographs of a C₆₀-fullerene single crystal with a CCD camera that enable the relative angular difference of the zone axis to be observed. C₆₀-fullerene fcc single crystal samples were grown from vapour employing a double temperature gradient. The results indicate that crystal growth began at the central region and proceeded to the periphery. These observations provide valuable information about crystal growth.

An on-site method of correcting the image distortion and nonuniform response of a charge-coupled device (CCD)-based X-ray detector was developed using the response of the imaging plate as a reference. The CCD-based X-ray detector consists of a beryllium-windowed X-ray image intensifier (Be-XRII) and a CCD as the image sensor. An image distortion of 29% was improved to less than 1% after the correction. In the correction of nonuniform response due to image distortion, subpixel approximation was performed for the redistribution of pixel values. The optimal number of subpixels was also discussed. In an experiment with polystyrene (PS) latex, it was verified that the correction of both image distortion and nonuniform response worked properly. The correction for the "contrast reduction" problem was also demonstrated for an isotropic X-ray scattering pattern from the PS latex.

An asymmetric electrostatic-quadrupole lens (AEQL) system for high definition field emission displays (HD-FEDs) was proposed. It was applied to the double-gated structure where the emitters are a thick layer of carbon nanotube paste such as a flat surface emitter. The AEQL structure was designed with two opposing planar electrodes of noncircular apertures which generate the quadrupole electric field. Utilizing a design of field emitter arrays (FEAs) with AEQL, an optimized beam shape with horizontal reduction and vertical elongation was obtained. According to three-dimensional (3D) simulation results, this AEQL structure exhibited excellent focusing effects that satisfied the aspects of pixel size and shape in HD-FEDs.

In this paper, the effect of a pulsed low-energy electron beam on bacteria has been explored. The experiment has been carried out using an electron beam with a pulse duration of 5 µs and an acceleration voltage of 80 kV. It is shown that homogeneous irradiation with an electron beam decontaminates surfaces. Completely Sterilization is achieved at a low concentration of bacteria. With increasing concentration of bacteria or for a wet target, the survivability of a microorganism increases. A spectroscopic experiment and also scanning electron microscopy (SEM) images show that the low-energy electron beam does not break the outer structure of a bacterial cell and that bacteria inactivation occurs through internal chemical or genetic changes.

Both the transverse and longitudinal beam dynamics in a photoinjector were investigated experimentally for high-brightness electron beam generation. The transverse emittance growth and the energy spread, both due to the rf and the space-charge effects in the rf gun, were investigated with the laser injection phase. A 31.7 MeV electron beam at 1.0 nC with a normalized rms transverse emittance of 3.2±0.2 mm·mrad, an rms bunch length of 1.8±0.2 ps, and an rms relative energy spread of 0.04±0.01% was obtained in the photoinjector with a 5 ps (full width at half maximum; FWHM) laser. The phase compression in the bunch length was observed in the rf gun. A sub-picosecond electron bunch was observed experimentally in the rf gun at the low injection phase, such as 0.7±0.2 ps at 15° and 0.5±0.1 ps at 10°. The dependences of the transverse emittance, the bunch length and the energy spread on the rf phase of a booster linear accelerator (linac) downstream of the rf gun were investigated experimentally.

For pulsed periodical signals with a low duty cycle and a constant height, a carelsess use of a digital oscilloscope may lead to inaccurate waveforms. This is the case particularly when pulse timing is affected by jittering. An exemplary waveform, as previously reported [N. Harba and S. Hayashi: Jpn. J. Appl. Phys. 42 (2003) 2971], is re-examined in terms of a stochastic model using colored Gaussian noise.

In this investigation, we explore the flow field of an annular channel between two horizontal concentric rotating cylinders, i.e., between a rotating inner cylinder and a stationary outer cylinder. Resulting from interactions between centrifugal force, viscous force and different boundary conditions, flow fields in an annular channel probably develop groups of opposite Taylor vortices when the Taylor number is higher than a critical value. Geometrical parameters of rotating cylinder channels, such as the channel widths and circumferential ribs are also affected by the flow field. Four types of rotating inner cylinder are available in this experiment: smooth walled (Model A), and circumferential ribs of three different sizes (Models B–D). The aspect ratios for circumferential ribs are 5/3, 5/2, and 10/3, which generate periodically embedded cavities of 10, 15, and 20 mm. The radius ratios between the inner and outer cylinders were η_s=0.89, and η_rib=0.94, respectively. Taylor numbers ranged between (8.565–1312.943)×10³, and centrifugal force arising from the rotation of Model A was F_s=0.22–3.3 N. The centrifugal forces arising from the inner cylinders with embedded circumferential ribs were F_rib=0.23–3.49 Nt. Because the wavelength of the Taylor vortices was subjected to the influence of different geometrical conditions, the flow field structure of Model A was different at both ends of the cylinder. Conversely, various forms of Taylor vortices occurred between annular channels with circumferential ribs in the case of Models B–D, and the vortex evolved from the edge of the circumferential ribs in a more stable manner than in Model A. The wavelengths of the Taylor vortices were λ_A=30 mm, λ_B and λ_D=20 mm, and λ_C=15 mm. Experimental results of flow visualization demonstrated it to be well suited for benchmarking engineering designs for heat transfer, cooling, and tribology of rotating machinery.

The tantalum plasma flash X-ray generator is useful for performing high-speed enhanced K-edge angiography using cone beams because K-series characteristic X-rays from the tantalum target are absorbed effectively by gadolinium-based contrast media. In the flash X-ray generator, a 150 nF condenser is charged up to 80 kV by a power supply, and flash X-rays are produced by the discharging. The X-ray tube is a demountable cold-cathode diode, and the turbomolecular pump evacuates air from the tube with a pressure of approximately 1 mPa. Since the electric circuit of the high-voltage pulse generator employs a cable transmission line, the high-voltage pulse generator produces twice the potential of the condenser charging voltage. At a charging voltage of 80 kV, the estimated maximum tube voltage and current were approximately 160 kV and 40 kA, respectively. When the charging voltage was increased, the K-series characteristic X-ray intensities of cerium increased. The K lines were clean and intense, and hardly any bremsstrahlung rays were detected. The X-ray pulse widths were approximately 100 ns, and the time-integrated X-ray intensity had a value of approximately 300 µGy at 1.0 m from the X-ray source with a charging voltage of 80 kV. Angiography was performed using a filmless computed radiography (CR) system and gadolinium-based contrast media. In the angiography of nonliving animals, we observed fine blood vessels of approximately 100 µm with high contrasts.

A CR-39 plastic nuclear track detector was used as a linear energy transfer (LET) detector for carbon ion radiotherapy. We compared dose-averaged LET distributions in water obtained using the CR-39 detector for a monoenergetic beam and spread-out Bragg peak beam by calculations using the one-dimensional heavy-ion transport code used in the current heavy-ion treatment planning. We confirmed that the CR-39 detector could measure the high LET particles that are dominant contributors to dose-averaged LET. On the other hand, the CR-39 result was overestimated in the tail region of the distal edge in depth-dose distributions, due to its detection limit for lower LET particles. However, physical dose in the region is quite small. Namely, the effect of this difference on the biological dose distribution is also small. These results demonstrate that the CR-39 detector is a useful detector for measuring the LET distribution in carbon ion radiotherapy.

A 10 MHz pulsed Doppler ultrasound was applied to measure both Doppler power and Doppler velocity from stirred porcine blood of various hematocrits for assessing variations in blood properties during blood coagulation and clot formation. For each measurement, blood was recalcified by adding calcium chloride solution. Results obtained from original blood at hematocrits of 25, 35, 45, and 55% showed that the mean Doppler power and Doppler velocity were respectively equal to 40.2, 38.5, 38.1, and 37.6 dB, and 24.6, 21.4, 20.0, and 19.6 cm/s. The variations in blood properties during blood coagulation caused Doppler power and velocity to fluctuate markedly. As the clot was formed, Doppler power was increased by approximately 5.5 dB and velocity was decreased to approximately 5.2 cm/s. These studies validated the suggestion that Doppler ultrasound is a feasible and sensitive means to quantify blood properties during blood coagulation and clot formation.

Laurdan is a useful fluorescence dye for observation of physical properties of membranes. An optical system was constructed in order to image generalized polarization (GP), that characterizes the change of the spectrum of laurdan. The fluorescence images through the optical system were recorded on a digital video recorder and the images were processed to GP images, which made it possible to measure the spatial distribution of membrane fluidity quantitatively at video rate. Using this imaging instrument, the change of membrane fluidity at every moment of phase transition was observed in a giant liposome of pure dimyristoylphosphatidylcholine and the dynamics of phase separation was observed in a giant liposome composed of dimyristoylphosphatidylcholine and dimyristoylphosphatidylethanolamine after an annealing process. A clear phase separation was observed in the giant liposome in the cooling process from 43 to 40°C. The GP value of the phase-separated region on the liposome was related to the shape of the same region. The linear part of the liposome had a high GP value and the hinge part had a low GP value.

A three-pronged microchannel is fabricated in resin by UV laser ablation. A number of heat-hardened resin films are layered on a soda glass. A laser fabricates a part of the channel on each film for every lamination. The channels are 45 µm in depth and 50–150 µm in width. Through holes are made in the laminate film with a laser. Inlet pipes with reservoirs are inserted into the holes. A blood cell can be transported by gravity applied to it. The blood is led into the center of the channel with a sheath flow of solution from both side channels. Blood cells are then focused at the junction by controlling the height difference between a blood reservoir and the solution reservoirs.

Natural oxidation processes of hydrogenated Si nanocrystallites were investigated to clarify effects of surface oxidation on photoluminescence wavelength. Hydrogenated Si nanocrystallites were prepared by pulsed laser ablation in hydrogen gas ambient. The Si–H bonds on the surface of the nanocrystallites enable us to estimate the local configuration of Si–O bonds using infrared frequency shifts. The natural oxidation process was investigated by measuring the density and local configuration of Si–O bonds. The oxidation process can be classified into first and second stages. The first stage is due to the diffusion of oxygen molecules in the nanocrystallites through voids in the porous structure, and the second stage is due to the oxidation of each nanocrystallite from the top surface to the sub-surface. The configurations of Si–O bonds in the first and second stages are silicon-rich and oxygen-rich compositions, respectively. The photoluminescence wavelength was blue-shifted with increasing Si–O bond density. This PL peak shift was not continuous, but three PL peak regions at 800, 600–700, and 400–500 nm were observed. This result indicates that the origin of this PL peak shift is not due to quantum confinement because of decreased diameter of Si nanocrystallites, but is due to the existence of surface oxide. A photoluminescence peak at 800 nm was observed in fresh specimens, and those at 600–700 and 400–500 nm were observed from the first and second stages of oxidation, respectively.

The carrier density and Hall mobility of indium tin oxide (ITO) films uniformly deposited with single layers of monodispersive nanoparticles of fluorine-doped tin oxide (FTO) show intriguing variations as functions of the average size of FTO particles (D_av). For D_avs smaller than about 10 nm, the bilayer has conduction-electron densities lower than bare ITO films, accompanied by increased Hall mobilities. Surprisingly, for D_avs of a few tens of nanometers, the carrier density in the bilayer is higher by at least 30% than that of bare ITO films. A classical double-Schottky band model of an n-n isotype heterojunction can account for these behaviors. This model assumes high-density local states at the ITO-film/FTO-nanoparticle interface, which trap conduction electrons, while FTO nanoparticles act as electron injectors. It is pointed out that the ability to control the optoelectronic properties simply by optimizing the film structure is useful in applications to practical transparent conducting electrodes.

We have studied femtosecond laser processing using subwavelength thin metal slits patterned onto Si substrates. A thin Ta film (200 nm thickness) with a periodic-structured slit array (200–1000 nm width) was formed on the Si substrates. The ablation thresholds of the Ta and Si were investigated. Near-field enhancement effects were observed at the edge of the metal slit arrays. Ridge structures were created in the slits on the Si surfaces using a single laser shot. In addition, an interesting processing effect was observed that was dependent on the polarization of the beam; only a p-polarized beam could create grooves perpendicular to the slits on the Si substrate. The grooves were formed under the metal layer. Our experimental results concerning the enhancement of the electrical field at the edge of the slits were consistent with the results from finite-difference time-domain simulations.

The effects of charge injections into a single 4,4''-terphenyldithiol molecule were investigated using density-functional calculations. It is shown that the atomic structures of the molecule are remarkably modified by electron or hole injections into it. Strengthening and weakening of the C–C and C–S π-bonds brought about by the charge injections are closely associated with these structural modifications. Analyses of the wave functions of the highest occupied molecular orbital (HOMO) and the lowest un-occupied molecular orbital (LUMO), especially those on the arrangements of the nodes and loops of them, help our understanding of such modulations of the bond orders. It is also shown that the energies required for charging the molecule are considerably affected by these deformations. These results suggest the importance of the electron–lattice interactions in the current conduction due to the single-electron tunneling through the 4,4''-terphenyldithiol molecular single-electron island.