Table of contents

The oxygen incorporation mechanism at the growth interface during Czochralski (CZ) silicon single crystal growth has been studied using melt quenching technique developed by the double-layered Czochralski (DLCZ) process. A slow crystal rotation rate of 0.5 rpm was chosen to investigate the influence of melt convection in relation to the source of oxygen on the oxygen inhomogeneity in the grown crystal. Micro-Fourier transform infrared spectroscopy (micro-FTIR) measurement and preferential etching revealed that the oxygen variation in the crystal was determined by the balance of the bulk melt with a high oxygen content and the oxygen-depleted melt from the free surface. The major cause of the periodic intake of the oxygen-depleted melt flow during one crystal rotation was probably the inhomogeneous radial temperature gradient. This suggests that thermal asymmetry in the melt is a fatal factor in oxygen inhomogeneity in the crystal. We have postulated that the equilibrium oxygen segregation coefficient is not much smaller than unity, but is rather close to it.

The binding energy, E _b, between a vacancy and a P atom or an As atom in Si is large. Therefore the pair diffusion model (PDM) and the decrease in quasi-vacancy formation energy are applicable to P and As diffusions. With the decrease in quasi-vacancy formation energy, E _b also decreases. Fermi level obtained from the Boltzmann statistics is used. Anomalous diffusion of P consists of two components, a tail and a plateau. The tail has been attributed to excess vacancies, the generation of which is a characteristic of PDM. The plateau has been explained by PDM and the decrease in quasi-vacancy formation energy. The reason for this is given for the first time. It is emphasized that the large value of the phosphorus effective diffusion coefficient at the plateau is not anomalous but normal.

Low-temperature Si epitaxial growth has been achieved by irradiating electrons and Si₂H₆ molecules simultaneously onto Ge(100) and Si(100) surfaces. The growth takes place exclusively on the electron-irradiated area. The electron stimulation of the surface enables growth rate enhancement and improvement of crystallinity. Si growth is governed by surface reactions induced by electron beam irradiation.

An emitter edge-thinning design is used for a lattice-matched InGaAs/InP double-heterostructure-emitter bipolar transistor (DHEBT) grown by low-pressure metalorganic chemical vapor deposition (MOCVD). Using the emitter edge-thinning technique, the surface recombination current is reduced and the current gain is improved. This structure reveals a typical current gain as high as 120 at a collector current density of 241 A/cm², along with an offset voltage as low as 45 mV. Furthermore, the problem of surface recombination current is also discussed in terms of the emitter size effect on current gain. The Gummel plots are shown to explain the influence of the emitter edge-thinning process on the base current ideality factor.

The semi-empirical tight-binding method is used to investigate band structures of wurtzite III-V nitride semiconductors. The tight-binding band structures of GaN and InN are first obtained by fitting the data to experiments and more accurate calculations. To obtain better description of the band structures, second-nearest-neighbor s and p state interactions are included and Ga 3d and In 4d electrons are treated as band states. Then, the band structure of the alloy Ga_1-xIn_xN is predicted based on the obtained tight-binding parameters and a pseudobinary alloy model. The predicted alloy band gaps are in good agreement with experimental ones. The electron effective masses of the alloy are also presented. The effects of the second-nearest-neighbor interactions and nearest-neighbor s-d and p-d state interactions on the band structures are discussed in detail.

Closed-type solutions of the basic model for photoinduced metastable transformation of the EL2 centers in SI GaAs are studied. Conditions for the existence of maximum in the optical absorption α, photocapacitance C _d, and electron paramagnetic resonance (EPR) transients are determined. Several methods for the estimation of the initial neutral EL2 fraction (occupancy f), and cross-section for the metastable transformation σ^* from the dynamics of "fingerprint" transients are proposed, and the accompanying ready-to-use nomograms computed and presented.

GaAs surfaces were passivated with selenium using Se/NH₄OH solution. Selenium powder of 99.8% purity was dissolved in NH₄OH and the GaAs substrates were immersed in this solution. The extent of passivation was studied by electrical measurements (current-voltage characteristics) of the Schottky diodes. The experimental results indicated that the surface properties were substantially improved without requiring a succeeding Na₂S treatment. The degradation of the ideality factor, n, of the Schottky diodes was efficiently retarded by this selenium passivation. The surface chemistry of the passivated GaAs surface was investigated with X-ray photoelectron spectroscopy (XPS). The XPS data indicated that both the oxidation of the GaAs surface (formation of Ga₂O₃ and As₂O₃) and segregation of elemental arsenic ( As⁰) at the surface were suppressed or retarded by this passivation. This corresponds well to retardation of degradation of the n value. As₂Se₃ was observed in the Se/NH₄OH-passivated surface, which suggests that the segregated As was removed by chemical reaction between As and Se to produce As₂Se₃. There is a possibility of the formation of Ga₂Se₃ (Ga-Se bond), though it is not observed in our XPS data. The formation of As-Se and/or Ga-Se bonds is suggested as the reason for suppression of oxidation of the GaAs surface. When the Se-passivated surface was exposed to air ambient, the total amount of selenium and the amount of As₂Se₃ decreased. This may be due to oxidation of As₂Se₃ to form As₂O₃. The extent of Se/NH₄OH passivation and the degree of degradation due to air exposure are similar with those of the ( Na₂Se/NH₄OH+Na₂S) passivation reported by Sandroff et al. [J. Appl. Phys. 67 (1990) 586].

Influence of light-soaking temperature on the formation of photocreated neutral dangling bonds (DBs) in hydrogenated amorphous silicon is investigated by light soaking at 77 K, room temperature (RT) and 90°C. Distributions of thermal annealing activation energies for the DBs photocreated at the three temperatures are obtained. The effect of light-soaking temperature on the increase rate of DBs and on the distribution of thermal annealing activation energies can be self-consistently fitted using two different rate equations. One contains a photocreation term, a thermal annealing term and a light-induced annealing term. The other has a photocreation term and a thermal annealing term modified by the light-induced annealing effect. The former rate equation can explain the very high density of photocreated DBs, while the latter can explain the recent result of light-induced annealing.

The microcrystallinity of hydrogenated amorphous silicon films deposited by the conventional radio-frequency plasma-enhanced chemical vapor deposition (rf-PECVD) method and its dependence on chamber pressure are discussed. In a wide range of pressure at which the microcrystalline film can be formed, a critical pressure (500 mT) is found. Films deposited at this critical pressure possess the highest crystalline volume fraction and the smallest grain size. An ion-bombardment-assisted model is proposed to explain the experimental results. Concerning the applications of microcrystalline films to thin-film transistors (TFTs), the subthreshold swing and the field effect mobility are studied, both of which are found to be smaller than those of the hydrogenated amorphous silicon (a-Si:H) TFTs.

A ZnSe buffer layer has been applied as an attractive alternative to a CdS buffer layer in the development of polycrystalline Cu(InGa)Se₂ (CIGS) thin-film solar cells, thus eliminating entirely the use of cadmium by employing the ZnO/ZnSe/CIGS structure. Moreover, we propose the use of a new deposition method for ZnSe buffer layers, the atomic-layer deposition (ALD) method. This method is basically the same as an "atomic-layer epitaxy" method but is applied to polycrystalline materials. Currently the best efficiency of CIGS thin-film solar cells with an about 10-nm-thick ZnSe buffer layer is 11.6%. Applying irradiation with a solar simulator under one-sun (AM-1.5, 100 mW/cm²) conditions, the efficiency of these cells was improved from about 5% to over 11% due to increased open-circuit voltage and fill factor with no change in short-circuit current density even after six-hour irradiation.

The use of 1.35 eV amorphous silicon-germanium (a-SiGe:H) alloy as the second/third intrinsic layer along with 1.85 eV front layer in double/triple tandem solar cells is believed to be the best combination for the maximum power output for multijunction cells. In this study high quality low-band-gap (1.36 eV) a-SiGe:H alloy has been developed by RF glow discharge optimizing the deposition parameters and helium dilution of source gases. It has been observed that the structural, electronic properties and defect densities of alloy films developed under the deposition condition which is the transition from low-discharge-power to high-discharge-power regime, become optimum. In the present case this deposition condition is a combination of chamber pressure 0.8 Torr and RF power 60 mW/cm². The properties of the alloy films developed under helium dilution improve and defect density decreases with the increase of deposition rate up to 120 Å/min. The 1.36 eV alloy film prepared under this condition has very low defect density ( 3.2×10¹⁶ cm^-3 eV^-1). The analysis of spectral response of Pd/a-SiGe:H Schottky barrier solar cells reveals that the hole transport properties improve due to increase of RF power from 15 to 60 mW/cm² and also due to increase of growth rate from 51 to 120 Å/min.

The precursors of a-Si:H films in electron cyclotron resonance plasma chemical vapor deposition have been investigated by analyzing the deposition profile on a trench. The profile of the film prepared from radicals produced in the gas phase by electron collision with SiH₄ is simulated by a Monte Carlo method using a sticking probability of 0.8 for Si, SiH and SiH₂ radicals and of 0.1 for SiH₃ radicals. A comparison between the experimental result and the simulation has shown that the ratio of the total flux of SiH_x (x=0-2) to the flux of SiH₃ in the radicals reaching the surface is 0.82 at 3 mTorr. The flux of radicals which maintain the momentum of SiH₄ emitted from a nozzle is also compared with that of radicals with a random direction of momentum.

Dehydrogenation characteristics have been investigated in detail for excimer-laser annealing of hydrogenated amorphous silicon (a-Si:H) films. Only the XeF excimer-laser light has a negligibly small photonic-dissociation rate of the Si-H bonds, and thus hydrogen atoms remain in the annealed film, resulting in highly conductive poly-Si film. The XeF excimer-laser annealing has been confirmed to be useful for fabricating high-performance a-Si:H thin-film transistors.

The real (ε₁) and imaginary (ε₂) parts of the complex dielectric function, ε(E)=ε₁(E)+ iε₂(E), of SnTe have been measured by spectroscopic ellipsometry (SE) in the 1.15-5.5 eV photon-energy range at room temperature. The measured SE spectra reveal distinct structures at energies of E₁, E₂ and E₃ critical points. These data are analyzed using two theoretical models, namely, the model dielectric function (MDF) and standard critical-point (SCP) model. It is found that both the MDF and SCP models successfully explain the first-derivative spectra of ε(E) [ dε(E)/ dE]. The MDF also shows excellent agreement with the experimental ε(E) spectra, but the SCP does not. Dielectric-related optical constants, such as the complex refractive index (n^*=n+ ik), absorption coefficient (α) and normal-incidence reflectivity (R), of SnTe are also presented.

We attempted to control the arrangement of the native Ga vacancies in Ga₂Se₃ films on (100)GaAs substrates by molecular beam epitaxy. When the GaAs substrates were heat-treated at 550° C before growth, the vacancy-ordered superstructure was formed in the [01] direction, which is different from the ordering direction in Ga₂Se₃ on (100)GaP. On the other hand, when the heat treatment was not carried out, the ordering direction was [011] at growth temperatures below 500° C. These results indicate that the initial stage of the growth affects the ordering direction in Ga₂Se₃ on (100)GaAs. We have successfully controlled the ordering direction by introducing a Ga₂Se₃ buffer layer grown at low temperature. Furthermore, we investigated the effects of the growth conditions on the ordering of the Ga vacancies in Ga₂Se₃ films on (100)GaAs, and it was found that the vacancy ordering was highly developed at high VI/III ratio and low growth temperature.

Hot electrons were generated in n-doped GaAs/AlAs multiple quantum wells by optical intersubband excitation and their energy distributions were studied at 10 K. By measuring photoluminescence (PL) spectra, the hot-electron temperature T _el was quantitatively evaluated, taking into account such effects as the degenerate distribution of electrons, impurity-induced level broadening, and the distribution of holes. It was found that, when the intensity P of the excitation was 30 kW cm^-2, i.e. a few percent of the onset intensity for absorption saturation, T _el increased from 10 K to 120 K while the lattice temperature remained unchanged. When P was between 10 W cm^-2 and 30 kW cm^-2, the rise of T _el was well explained by the balance between the photon energy absorption and the energy loss due to the LO phonon emission inside the ground subband.

The surge-handling capability of bi-directional thyristors is estimated for four kinds of surge waveforms: 8/20, 1.5/30, 15/100 and 10/1000 µ s. The surge-handling capability increases as the surge waveform becomes shorter. The failure modes of the device are classified into three modes: soft breakdown in the forward direction, a short mode in the forward direction and a bi-directional short mode. The occurrence rate of these modes depends on the applied surge waveforms. A melt hole with many cracks appears in the bi-directional short mode sample. Furthermore, electrode metal diffuses into the silicon substrate near the melt hole and the cracks. The diffused elements and their concentrations are identified by X-ray microanalysis. The likely cause of the melt hole and the diffusion of electrode metal is discussed considering temperature increase caused by surge energy.

Recent studies of etching technology using plasma indicate that serious problems are induced by a charge build-up. The charge build-up causes radiation damage and distortion of etched profiles. These problems are induced by a non-neutrality of charge supply to the surface. By controlling the electric potential of surface material with a dc power supply, dependences of damage and pattern distortions on charge-up are investigated. Keeping the net current to the wafer zero is effective for reducing the radiation damage. The profile distortions occur due to the potential difference between mask and etched material. The profile is improved by setting the applied dc potential equal to floating potential.

Chemical states of 6H polytype crystalline silicon carbide (6H-SiC) surfaces were investigated using X-ray photoelectron spectroscopy (XPS). Surface contaminants such as C-C, C-H and C-O species were evaluated from C1s photoelectron spectra after treatment under various conditions. Clean SiC surfaces were found to appear after the chemical etching of a thermal oxide ( SiO₂) film using a buffered HF solution. Furthermore, to clarify the chemical etching characteristics of SiO₂ formed on 6H-SiC substrates, the depth profiling of the SiO₂ was also performed using XPS. The C1s peak at a binding energy of about 286.2 eV was detected on the surfaces of the 6H-SiC substrates at the moment when only the SiO₂ was completely removed, and the C1s peak was characterized as due to the C-O bonds formed near the SiO₂/SiC interface.

Rapid thermal technique was used in the post-metallization annealing (PMA) of thin gate oxide devices. A suitable choice of the rise rate, the setting temperature, and the hold time in the rapid thermal PMA (RTPMA) process is helpful to improve the oxide quality. It was found that the samples subjected to appropriate RTPMA conditions exhibit almost the same initial characteristic in flatband voltage V _FB and midgap interface trap density D _itm as those subjected to conventional furnace PMA (FPMA). However, the RTPMA samples exhibit longer time-to-breakdown t _BD and higher time-zero-dielectric-breakdown (TZDB) field E _BD than the FPMA ones. In addition to the known spiking effect caused by aluminum penetration into silicon, which seriously degrades the breakdown property, formation of aluminum oxide near the Al/SiO₂ interface in the early stage and then aluminum silicon alloy in the later stage was proposed to explain the experimental observation.

The temperature dependent variation of capacitance in amorphous silicon solar cells is found to have different behaviours at low frequency (100 Hz). The variation at high temperature has been shown to arise due to the interface states at p-i junction. The density of states at p-i junction and in the whole i-region have been estimated to be 7×10¹⁶ eV^-1 cm^-3 and 5×10¹⁶ eV^-1 cm^-3, respectively, in as-deposited samples. The effect of the reverse biased annealing on the interface states has been observed.

We examine whether there is a correlation between the morphologies at the cap SiO₂ layer/top Si layer interface and the top Si layer/buried SiO₂ layer interface of SIMOX (Separation by IMplanted OXygen) (100) wafers during high-temperature annealing. The morphologies were evaluated by scanning a marked position with an atomic force microscope (AFM). From the AFM observation, although some similarities were observed in the mosaic morphologies, squares were not positioned at corresponding locations at the two interfaces. Therefore, it is concluded that there is no correlation of the morphologies between the two interfaces. The squares on each interface grow independently.

Bi_2.4-xPb_xSr₂CaCu₂O_y crystals with x=0.05, 0.1 and 0.2 were irradiated with ⁶⁰Co γ-rays of 0.6 MR/h up to a dose of 64.8 MR at room temperature. For all the irradiated samples with x=0.05, 0.1, the superconducting transition temperature T _c increased after γ-ray irradiation, while the T _c of irradiated samples with x=0.2 increased at lower doses, and then decreased at 64.8 MR. X-ray photomission spectroscopy results reveal that higher oxidation states of Bi and Cu occurred in the sample irradiated with 64.8 MR. The dependence of T _c on γ-ray irradiation dose can be explained in terms of the change of valence state of Bi and Cu.

Temperature dependent critical current density and flux-creep measurement were carried out in magnetically aligned HgBa₂Ca₂Cu₃O_y. For 20 K<T<60 K, we observed quasi-exponential behavior of the critical current density J _c. The normalized relaxation rate S vs T curve has an N-like shape with the minimum around T=15 K and the peak at 50 K. The theoretical U(J) given by the collective flux creep model describes the experimental result well.

Single crystals of SmBa₂Cu₃O_7-x (Sm123) have been successfully grown by a modified top-seeded solution growth (TSSG) method in 1% oxygen partial pressure atmosphere ( P(O₂)=0.01 atm). Solid Sm₂BaCuO₅ (Sm211) solute and Ba-Cu-O solvent with Ba to Cu ratio of 3:5 were used in a Sm₂O₃ crucible. The Sm123 as-grown crystals were confirmed by X-ray structure analysis to have the SmBa₂Cu₃O_7-x tetragonal structure, and the substitution of Sm ion to the Ba site was not detected within experimental accuracy. The transport properties of the grown crystals annealed in an oxygen gas flow were measured by the direct-current four-contact method, both in the ab-plane and along the c-axis. The superconductive transition temperature was 93 K and the transition width was 0.3 K.

In situ annealed high temperature superconducting YBa₂Cu₃O_7-δ thin films have been deposited on an MgO (100) substrate from a single stoichiometric target using DC magnetron sputtering. The films were characterized by X-Ray diffraction (XRD) and scanning electron microscopy (SEM) techniques. The effect of varying substrate temperature, T _s, and O₂/Ar ratio on lattice parameters and on the degree of orientation of the films were examined. Both c- and a-lattice parameters decreased with increasing T _s. The reduction of c and a-lattice parameters as well as the oxygen deficiency in the films, δ, obey general (T _s - T₀)^-4 behavior. We develope a new method to measure a more accurate way to find the degree of preferrential orientation along c and a-axis of the deposited films, (ΔV₀₀₆/ΔV₂₀₀), at different T _s by using X-ray diffraction theory and JCPDS files to obtain ∣F₍₀₀₆₎/F₍₂₀₀₎∣². At T _s=735°C, the volume fraction along the c-axis was found to be ΔV_c≈5.5 ×ΔV_a corresponding to 85 grains having preferred orientation along c-axis. In addition, we have also measured FWHM of the (006) and (200) peaks by varying T _s. The thickness of the grains were estimated at different substrate temperature using the Scherrer formula.

Proximity coupling between superconducting filaments in NbTi multifilamentary wires with Cu or Cu–0.5 wt investigated. Critical current densities of these matrices J _cp were estimated from measured twist-pitch dependence of magnetization. The upper critical field of matrix B _c2p is obtained by applying the scaling law to J _cp data. At T=4.0 K, B _c2p's of Cu matrix are 1.4–4.6 T for interfilamentary spacing d _N=0.20–0.59 µm and are larger than those of Cu–0.5 wt of the matrix ξ _N estimated from the temperature dependence of B _c2p are 0.18–0.35 µm at T=4.2 K and show the clean limit characteristics, being proportional to T^-1 for Cu matrix wire. For Cu–0.5 wt T=4.2 K and show the dirty limit characteristics, proportional to T^-1/2.

We have examined the structure and magnetic properties in terms of the temperature dependence of magnetic permeability and magnetization for the Fe₇₄Si_xB_22-xCu₁Nb₃ (x=8-18) alloys. The temperature dependence of the permeability of the alloys shows various characteristic behaviors depending on the alloy composition (x=8-18) and annealing temperature (450-630° C). The permeability of the samples having coexisting α-Fe(Si) and amorphous phases decreases abruptly near the Curie temperature of the amorphous materials, while the permeability of the samples containing precipitated Fe-B compounds is not appreciably affected by the magnetism in the residual amorphous region. The precipitation of Fe-B compound depends on the composition of the mother alloys. The Fe-B compounds precipitated in the amorphous region upon annealing at high temperature are mainly Fe₂₃B₆ for x=14, Fe₃B for x=18, and Fe₂B for the composition x less than 12.

In this study we suggest a new method to prepare an anisotropic bond magnet. In order to achieve an anisotropic magnet we take advantage of the magnetic anisotropy of the tetragonal Nd₂Fe₁₄B crystals. A dispersion of a melt spun Nd–Fe–B magnet powder with epoxy resin was hardened in a high frequency (HF) magnetic field (ν= 230 kHz) to align the easy magnetization axis (c-axis in Nd₂Fe₁₄B) with the external field. Magnetization measurements and an X-ray analysis showed that an alignment of the powder due to the treatment in a HF magnetic field was achieved. A Mößbauer analysis confirmed the alignment.

Switching fields of hexagonal particles with nonuniform magnetic properties are investigated using computer simulation. The switching field of a hexagonal particle with a low anisotropy energy constant (K _u) region at a corner of all three thickness layers is lower than that of the uniform particle when the field is applied almost perpen-dicular to the hexagonal plane. However, when the field is applied almost parallel to the hexagonal plane, the switching field of the nonuniform hexagonal particle is similar to that of the uniform particle. A hexagonal particle with a low K _u region at the center also switches in a lower field than the uniform particle does when the field is applied perpendicular to the hexagonal plane. The switching field of a hexagonal particle with low K _u region in the surface layer is almost the same as that of a uniform particle.

Mössbauer spectra have been investigated for Fe clusters precipitated in a sputter (SP)-deposited nonequilibrium Fe₁₁Cu₈₉ alloy, and compared with one produced by cluster beam (CB) deposition. With annealing time of the SP-deposited alloy the number of isolated atoms (monomer) does not change at the first stage, but, the number of small clusters decreases and that of large clusters (γ-Fe) increases. This is attributable to Ostwald ripening. Moreover, the size of Fe clusters is more uniform in the CB-deposited alloy than in the SP-deposited alloy after annealing at 573 K for 98.8 ks.

Electron paramagnetic resonance (EPR) and Mössbauer spectra of Japanese traditional Bizen pottery and its constituent clays have been measured to study the relationship between the color of pottery surface and the relevant states of iron ions ( Fe³⁺ and Fe²⁺). Hyperfine signals of Mn²⁺, presumably in carbonates, and a broad signal at g=2.0 similar to that of hematite ( Fe₂O₃) were observed for good-quality clay, while a signal at g>9 similar to that of magnetite ( Fe₃O₄) was observed for poor-quality clay. In pottery, the apparent g-factor of g=4.3 due to a large orthorhombic distortion E(S_x²-S_y²) and g=6 due to a large axial field DS_z² were observed in addition to the broad signal around g=2 due to oxidation of iron into Fe₂O₃. Subtle change of colors resulted in the change of EPR spectra. Mössbauer spectra indicatcd that almost all of the iron ions at the surface of pottery are strongly oxidized into Fe³⁺ when the pottery is fired in oxidizing atmosphere, while those inside the pottery and at the surface fired at reducing atmosphere are not strongly oxidized into Fe³⁺.

We report very uniform characteristics of 0.85 µ m vertical-cavity surface-emitting lasers (VCSELs) array grown by metal organic chemical vapor deposition (MOCVD) at a high growth rate. For a 16 µ m diameter pixel, the threshold current is 3.29±0.25 mA and the emission wavelength is 850.9±0.6 nm at 6 mA along the 8×8 pixels. Considering the offset gain and Al content of the spacer layer, the threshold voltage is 2.1 V. By growing VCSELs on an AlGaAs substrate, we were able to acheive a bottom-emitting 0.85 µ m VCSEL array where all elements operated successfully. We also discuss the modulation speed limit and the thermal interference problem of VCSEL arrays. The modulation limit is shown to depend on the capacitance and resistance (CR) time constant of electrodes, and the maximum temperature increase is about 15° C around the center of the chip when the other pixels are driven at the threshold current.

High-efficiency operation at two micrometers for a very short (less than 30 mm long) Tm³⁺, Ho³⁺ co-doped AZF glass fiber laser pumped by a Ti:sapphire laser has been demonstrated near room temperature. A cw output power of up to 150 mW with a slope efficiency of 37% was achieved using a 24.5-mm-long, Tm³⁺, Ho³⁺ co-doped AZF glass fiber at 15° C.

Heteroepitaxial growth of LiNbO₃ film is expected to provide a novel hybrid integration of optical and electronic components and devices. In this paper, we present, for the first time, to our knowledge, characterization of LiNbO₃ optical-waveguide film grown on c-cut sapphire by ArF excimer laser ablation. Almost droplet-free LiNbO₃ film was deposited only by increasing the target-to-substrate distance d to 40 mm, where Li-rich LiNbO₃ ceramics with the Li/Nb ratio of 2.1 was used as the target. The deposited LiNbO₃ film also exhibited a strong (006) peak in the X-ray diffraction spectrum, showing good orientation along the c axis. Optical properties were characterized based on excitation of guided modes in the film with a prism coupler. The propagation loss of the TM₀ mode was lowered to be 5 dB/cm at λ=0.633 µ m. The film indices were exactly determined, followed by evaluation of the chemical composition of the film from the measured index n _e of the extraordinary wave. Lack of Li₂O in the film became more significant as the distance d increased. A Bragg light deflector was then fabricated to evaluate the electrooptic coefficient r₃₃ in the film deposited under the condition that d=40 mm. r₃₃ was found to be as large as 4.9 pm/V, which was nearly one-sixth of the single-crystal value, irrespective of nonstoichiometry of the film.

A signal-regenerating optical-fiber loop memory with a side-injection-light-controlled bistable laser diode exhibiting thresholding gain characteristics is demonstrated. This system digitally regenerates memorized signals bit by bit. Output signals with a high on/off ratio of around 20 dB and stable intensity levels with less than 7% fluctuation are obtained even after 100 circulations.

The amplitude and phase of the chirped pulse leaving an optical fiber in a chirped-pulse amplification laser system are derived using the perturbation method. They facilitate understanding of actions of linear and nonlinear portions of the chirped pulse on the pulse compression. The contrast ratio of the compressed pulse can also be evaluated by using spectral and temporal filterings.

It was found that the photocurrent quantum efficiency of photoelectrochemical cells with p-CuSCN photo-cathode sensitized with double dye layers such as Rhodamine- C₁₈ and Methylviolet- C₁₈ is greatly enhanced (∼15%), when the double dye system is prepared using Langmuir-Blodgett (LB) technique. An ordered arrangement of dye molecules in LB films provides a suitable condition to transfer photogenerated holes and electrons between two dye layers, so that photogenerated holes upon excitation can transfer effectively into the valence band of p-CuSCN.

Minority-carrier diffusion, instead of direct penetration of pump light, can be a dominant factor which determines the probed region in photoreflectance measurements of epitaxial GaAs wafers. Appearance of Franz-Keldysh oscillations (FKOs) originating from epitaxial layer/substrate interfaces is found to critically depend on the penetration depth of pump light relative to the surface depletion-layer width. This relationship indicates the importance of generation and diffusion of minority carriers within the neutral region, which can induce photoreflectance from unexpectedly deep locations. By inserting a potential barrier (a doped AlAs layer) into the path for minority-carrier diffusion, such a FKO component is shown to be eliminated. The experiment confirms that the sample under study is indeed in the diffusion-limited regime, and at the same time, provides a simple method which allows an accurate evaluation of the FKOs from the doped region (surface depletion layer), which usually is of interest from technological aspects.

Ferroelectric Pb(Zr, Ti)O₃ (PZT) thin films were successfully deposited on PbTiO₃/ITO/Corning glass and PbTiO₃/(100)MgO single-crystal substrates by atmospheric-pressure metal-organic chemical vapor deposition. The PZT thin films deposited at 400° C showed poor optical transparency, while PZT thin films showing high transmittance were deposited above 450° C. The PZT thin films deposited at 500° C showed an optical transparency of 80% and a good quadratic electro-optic effect, of which the coefficient was 1.0×10^-18 m²/V².

We formulated and applied a new algorithm based on the continuum theory of nematic liquid crystals to the characterization of the electrooptic response of homogeneous nematic liquid crystal cells. Appropriate selection of the device parameters ensures predictions of the electrooptic response which are in agreement with results of optical measurements to a precision of better than 1°. The effect of induced molecular twist and the limitations of decoupling between orthogonal polarizations are discussed.

Zinc oxide ceramic varistors with simplified compositions of ZnO+Bi₂O₃+Co₃O₄+M₂O (M=K or Na) show nonlinearity coefficients (α) of 40-75. The electron paramagnetic resonance spectra and optical reflectance spectra show that there is a direct interdependence between the oxidation state of transition metals and the alkali ions. The X-ray photoelectron spectra indicate that the alkali ions preserve a higher oxidation state of cobalt, Co(III), in the grain boundary regions than in the grain interiors having more Co(II). Admittance spectroscopy shows that, while the nature of traps remains unaltered, the trap density increases with the concentration of alkali ions near the interface. The observed defect states are associated with the grain bulk than with the grain boundary interfaces, as indicated by the isothermal capacitance transient signals.

High-quality (Pb_1-xLa_x)Ti_1-x/4O₃ (x=0.05) [PLT] thin films were fabricated successfully by sol-gel processing onto Pt/Ti/SiO₂/Si and p-Si substrates. The optimum route and the conditions for preparing homogeneous and crack-free thin films were systematically investigated. Ferroelectric perovskite phase was observed in the PLT thin films annealed at 600° C for 30 min. We studied the electrical properties of the PLT film capacitors with the metal-ferroelectric-metal (MFM) and the metal-ferroelectric-semiconductor (MFS) structure for use in ferroelectric memories. The dielectric constant and dissipation factor measured for MFM capacitor were 225 and 0.018 at 100 kHz, respectively. The remanent polarization (P _r) and coercive field (E _c) of the PLT film were 8.5 µC/cm² and 110 kV/cm, respectively. Also the film with 600 nm thickness showed current density as low as 10^-6 A/cm² at the electric field of 200 kV/cm, which corresponds to the electric resistivity of ∼10¹² (Ω·cm) and represents excellent insulating properties. The MFS capacitor showed ferroelectric switching properties in C-V characteristics and the memory window of the hysteresis loop was about 1 V in the operation of ±10 V.

Relaxation behaviors of BEK, a new engineering plastic with high heat resistance (T _g: 181° C, T _m: 364° C), were investigated. Dielectric relaxation peaks were observed at -83° C and 51° C in 9.8 Hz, and their activation energies were 7 kcal/mol and 19 kcal/mol, respectively. The relaxation peaks were also observed in mechanical measurements. The `chemical structure-relaxation behavior' relationship in BEK and its related polymers suggests that the origins of the high temperature relaxation peak of BEK are crankshaft type chain motions while those of the low-temperature relaxation peak are wigglings of benzonitrile or phenyl groups.

Stresses in polyimide films coated on silicon wafers were measured in situ during the curing and cooling cycles. The films had a single rodlike molecular skeleton. Stress measurements were performed on films of uniform thickness under several different heating rates during the curing cycle. After the stress measurements, the cured polyimide films were examined by thermal mechanical analysis. The stress after the curing cycle increased with increasing heating rate during the curing cycle. Similarly, the thermal coefficient of expansion of the cured polyimide film increased with increasing heating rate. This phenomenom was thought to be caused by the decrease in the degree of in-plane orientation of the polyimide molecular chain with increased heating rate. The in-plane orientation was presumed to be influenced by the evaporation rate of the residual solvent.

The critical crystallite size for the cubic (c-) to tetragonal (t-) phase transformation of ultrafine BaTiO₃ crystallites was examined. Crystallites were prepared by calcination of barium titanyl oxalate tetrahydrate (BTOT) at various temperatures as to obtain crystallites with different sizes. The c- BaTiO₃ crystallite which is thermodynamically considered as a high-temperature form can exist at room temperature if the crystallite is smaller than certain size. The critical size for the c→t phase transformation is ∼30 nm. The size may become larger if the crystallite suffers higher lattice strain which can be induced by different powder preparation methods.

Ni(III) oxides with the K₂NiF₄-type structure, (A, A^')₂M_0.5^'Ni_0.5O₄ ( A^'=alkaline earth, A=rare earth, and M^'=Mg, Zn, Ti, Al, Ga) have been prepared under high oxygen pressure ( ∼150 MPa). The evolution of the local distortion of NiO₆ octahedra in such lattices was studied using X-ray diffraction analysis (XRD), electron paramagnetic resonance (EPR) and infrared (IR) spectroscopies. The observed distortions are discussed on the basis of covalency and relative strength of the axial [ (A^', A)-O] and equatorial ( M^'-O) competing bonds surrounding the NiO₆ octahedron in the lattice. These results are related to the structural tolerance factor proposed by Poix.

Nylon 6 film quenched from the molten state into liquid nitrogen exhibits an electric displacement vs electric field hysteresis loop at room temperature with the remanent polarization of 26 mC/m², as well as an exothermic anomaly around 340 K in the specific heat due to the heating process, which is attributed to the independent dipole rotation in the nematic-type unstable structure with loose molecular packing and imperfect hydrogen bonding.

Sub-atomic layer growth of SiC has been achieved using diethylsilane ((C₂H₅)₂SiH₂) by resolving elemental kinetics of its chemical vapor deposition. The growth rate of 0.1 monolayer/cycle was obtained over the temperature range between 590°C and 675°C. The detailed growth characteristics, the composition of the grown film, and the surface morphology have been presented. The initial growth and the coverage of the grown film were also investigated by X-ray photoelectron spectroscopy.

A semi-microscale 3-dimensional film growth simulation code using a simple Monte Carlo method was developed. This code predicts the step coverage on a trench and a hole of arbitrary shapes and requires much smaller computer memory size and less calculation time than the Direct Simulation Monte Carlo (DSMC) method. The simulation code was evaluated by comparing the results with experimental results of a zirconia ( ZrO₂) film grown on a hole. The experiments and/or 3-dimensional simulations indicated that the film grown on the side and bottom walls of a hole is thinner than that of a 2-dimensional trench, and complete occlusion of a hole is more difficult compared with a trench with opening width equal to the hole diameter. The surface reaction rate constant is the most important factor in the occlusion process. When the reaction rate constant is small, the hole is occluded with a thin film. However, when the reaction rate constant is large, there remain a void inside and a small unfilled pinhole through the thick film grown on the top.

CuPc films with good mechanical and chemical resistance were prepared by the plasma-activated evaporation method and the electrical properties of the films were measured in pure N₂ and in 100 ppm NO_x environments. In N₂ environment, the dc conductivity of the film depends on the temperature as e^-E/kT while the ac conductivity is independent of temperature and proportional to ωⁿ with the exponent n∼0.95. Adsorption of NO_x gas affects mainly the dc conductivity by forming acceptor states to generate holes, but negligibly affects the ac conductivity. The increment of dc conductivity is related to the amount of adsorbed NO_x gas molecules, which depends on the temperature and the film morphology. The impedance of CuPc thin film can be reasonably expressed as that of an equivalent circuit consisting of two resistor-capacitor pairs in series corresponding to the grain and the grain boundary. As the exposure time in NO_x environment increases, the resistances of both grain and grain boundary decrease, while the capacitances do not vary with the exposure time.

Various surface treatments, including wet chemical etching, heat treatment in H₂ and sputter etching with Ar ion, were applied to CdTe(111) surfaces. The induced surface layer and morphology were studied with angle-resolved X-ray photoelectron spectroscopy (ARXPS), Auger electron spectroscopy (AES) and atomic force microscopy (AFM). For wet chemical etching, both Br/methanol and KOH/methanol resulted in a Te-enriched surface, while the Te-rich layer induced by Br/methanol etching can be at least partially removed by treating with reducing etchants including hydrazine and dithionite. A hillock like morphology was the common feature of these wet-chemically etched surfaces. After hydrogen heat treatment, both Cd(111) and Te(111) surfaces were Cd-rich and a distinct difference in microscopic morphology between the Cd(111) and the Te(111) surfaces was observed. The surface morphology developed by hydrogen heat treatment can be used to identify the crystal polarity. The sputter etching caused compositional changes in the surface of CdTe(111), resulting in Te enrichment at the outermost layer but Te depletion in the subsurface region. Also, an Ar-ion-energy-dependent surface morphology was observed.

Photoeffects in ferrocene solid samples with physically adsorbed ethanol vapor have been studied in detail at different sample temperatures by monitoring the change in the electrical current in sandwich cells of the material in powder form. The photoeffects in the ethanol-adsorbed ferrocene sample resulted in the decrease of measured currents in the sample cells. Such experiments have indicated photoinduced desorption of ethanol vapors from the ferrocene solid surface. Previously, desorption of adsorbed vapor in the dark from the ferrocene surface was studied by flushing with dry nitrogen gas, but photoinduced desorption of adsorbed vapors from ferrocene is reported here for the first time. The photoeffects in the vapor-adsorbed ferrocene sample cell were also studied using different vapors. It was found that the photoeffects in the vapor-adsorbed ferrocene sample were very selective, and photoinduced desorption was not observed in all the vapors studied. In some vapors, photoeffects in the vaporadsorbed state resulted in enhanced photocurrents of the material. The above results are discussed.

Epitaxial growth of In layer on Si(111)-√3×√3-Ga was observed using an ultrahigh-vacuum scanning electron microscope (UHV-SEM). √3×√3-Ga (1/3 ML) and √3×√3-Ga [(1/3+2/3) ML] surfaces were prepared. The latter was formed by deposition of 2/3 ML of Ga on the √3×√3-Ga (1/3 ML) surface at room temperature. After 2 ML of In were deposited on the √3×√3-Ga (1/3 ML) surface, √7×√7 and √3×√3 structures appeared. These structures were clearly observed in the SEM images. Around the steps, ditches were observed. At the In thickness of 6 ML, flat islands with various thicknesses were observed. When 2 ML of In were deposited on the √3×√3-Ga [(1/3+2/3) ML] surface, √3/2×√3/2 structure appeared. All the terraces showed the same number of atomic layers without steps. For further In deposition the terraces had various thicknesses which differed stepwise by one atomic layer.

Scanning tunneling microscopy (STM) in liquid was applied in order to prevent the influence of oxidation in air. Using 1,2,4-trichlorobenzene as a liquid for our STM studies, an image of a Cu (001) surface was taken with high reproducibility and an atom image was also obtained. The surface geometry of Cu (001) differently polished from bulk crystals was investigated.

Film preparation was carried out by plasma-enhanced chemical vapor deposition using triethoxysilane (TES)/O₂, TES/N₂, and TES/Ar systems. Films deposited at 50° C showed IR bands due to silica networks and organic groups such as Si-H, Si-OEt and C=O. For the TES/O₂ system, the intensities of the signals due to organic groups decreased with an increase in substrate temperature, resulting in conventional silica spectra above 200° C. The organic groups still remained at 200° C for films prepared from TES/N₂, and Si-H groups were observed even at 350° C. The films exhibited blue photoluminescence at room temperature. Relative intensity of the photoluminescence was related to IR absorption intensity of the Si-H groups, suggesting that O-Si-H complexes or defects induced by Si-H may be the cause of photoluminescence.

Field-effect microtips are more coherent electron sources than hot filaments, and are well suited for electron optics applications such as low-energy electron interference. We propose to use this physical phenomenon in an integrated vacuum voltage comparator whose structure is derived from the classical Möllenstedt-type electron biprism. Our setup is separated into two independent parts, and acts both as a biprism and as a deflector for the fringe pattern. The deflection, proportional to the voltage we want to measure, is known with high precision thanks to the fine structures of the pattern. We conclude that it is possible to fabricate a device with a precision of 25 meV, a response time of 110 ps and a very high input resistance.

Irradiation effects of 1-MeV electrons on InP-related materials such as InP, InGaP and InGaAsP have been examined in comparison with those of GaAs. Superior radiation-resistance of InP-related materials and their devices compared to GaAs has been found in terms of minority-carrier diffusion length and properties of devices such as solar cells and light-emitting devices. Moreover, minority-carrier injection-enhanced annealing of radiation-induced defects in InP-related materials has also been observed.

The time behavior of the electric characteristic resistance and inductance in a pulsed gas discharge is revealed through a procedure which exploits only the voltage waveform. This can be achieved combining step by step the waveform with the differential equation governing the performance of the system. This method is described analytically in the text. Its application showed that the resistance drops sheerly in the first few ns from a very high value while the inductance presents a peak in the "formation phase" of the discharge (first 40 ns). These are due to the reaction of the system to the oncoming sudden variation.

The influence of the electrode material on O₂ radio-frequency discharge structure is investigated using the relaxation continuum (RCT) model. The discharges considered in this study are self-sustained between parallel plates with spacing of 2 cm at 13.56 MHz. The spatiotemporal discharge structure for the pressure of 1.0 Torr and the sustaining voltage of 100–300sin ωt V is calculated. The results are discussed for two types of material differing in the surface recombination probability of atomic oxygen γ. When γ is changed from 0.015 to 0.001, the number density of atomic oxygen increases and the distribution has less spatial dependence. As a result, the rate of the electron detachment caused by the collision between O^- and O increases, and the plasma density and the density ratio of the negative to positive ions decrease. The change of the material surface also affects the fluxes of positive ions and atomic oxygens to the electrode at a fixed sustaining voltage. The influence of the surface on the dissipated power is also discussed.

A Monte Carlo simulation model which combines ion backscattering and sputtering of solids with transport of emitted particles in the scrape-off layer plasma is developed for the study of the plasma-wall interaction in magnetic confinement fusion devices. The emphasis is put on the precise understanding of ionization and redeposition processes of backscattered and sputtered particles. In the model, Maxwellian ion-solid interaction, ionization of backscattered and sputtered neutral particles in the plasma, and gyromotion of the ionized particles in an oblique magnetic field are included. The heavier masses (lower charge states and lower energies) the particles have, the more they are redeposited. Due to impact of carbon ions on carbon and tungsten, therefore, less backscattered carbon is redeposited, whereas most of the sputtered tungsten atoms are redeposited. The redeposition causes substantially lower effective sputtering yields and a decrease in low-energy component of the sputtered atoms.

70 GHz slow wave propagation characteristics in a transversely magnetized partially filled ring-form solid-plasma waveguide have been studied experimentally employing n-type InSb ring at liquid nitrogen temperature as the plasma material. Obtained results were compared with those for the previously studied straight parallel-plate configuration. It is clearly shown that the surface wave resonance corresponding to the one observed in the straight plasma waveguide takes place also in the curved structure, and the propagation characteristics are quite similar to those of the straight configuration. Reduction of the resonant magnetic field is found with increasing the curvature of the waveguide.

For a planar or cylindrical triple probe which is contaminated with metal or gas, errors in the determination of electron temperature are numerically studied. As factors of error, we focus on i) relative contact voltage due to contaminants between two probes of the triple probe and ii) ion current parameter represented by a constant of proportionality between the square of ion current and probe bias. It is found that errors due to relative contact voltage and ion current parameter can be minimized by selecting the DC bias of the triple probe for the two probe types.

Electroluminescent (EL) characteristics in EL devices made of vacuum-sublimed dye films and spin-coated polymer films were compared. Low-molar-mass dye, 9,10-bis[4-(N,N-diphenylamino)styryl]anthracene (dye-BSA), for the preparation of vacuum-sublimed films, and polymer with 9,10-bis[4-(N,N-diphenylamino)-styryl]anthracene chromophore linked with alkylether groups (polymer-BSA) were employed. Single-layer devices, indium-tin oxide (ITO)/dye-BSA/MgAg and ITO/polymer-BSA/MgAg were prepared, and EL performances were compared. Double-layer devices which have an oxadiazole derivative (OXD-7) electron transport layer, ITO/dye-BSA/OXD-7/MgAg and ITO/polymer-BSA/OXD-7/MgAg, were also prepared. The current density-voltage relationships between dye-BSA devices and polymer-BSA devices were considerably different mainly due to the poor film quality of polymer-BSA. The comparisons of the luminance-current density relationships of the devices with two classes of BSA films showed that the polymer-BSA devices exhibited similar EL characteristics as the dye-BSA devices in the region of current density higher than 10 mA/cm². The possibility of the use of common material design concept for low-molar-mass materials and polymers was discussed.

Simultaneous depth analysis of deuterium and helium in the surface layer of deuterium-annealed He⁺-implanted yttrium iron garnet (YIG) was achieved by means of elastic recoil depth analysis (ERDA) and the results were compared with those of state analysis of deuterium and helium extracted by thermal desorption mass spectrometry (TDS). As a result, it was shown that both deuterium and helium moved to the strained lattice sites due to implantation via thermal annealing of the samples. By annealing the critically helium-saturated sample, helium atoms remained at the stable sites in the lattice, while deuterium occupied the less stable sites. Stabilization of helium sites at annealing temperature of around 350° C was observed for unsaturated and critically saturated samples. In the case of an oversaturated sample, three kinds of sites for helium were found, and it was observed that helium formed bubbles and that deuterium concentrated at sites corresponding to those of helium via annealing.

Gate oxide breakdown phenomena in a magnetron plasma were investigated from the viewpoint of the effect of magnetic field distribution on the radiation damage. It was found that strong parallel and normal magnetic field regions simultaneously present on a wafer induced surface potential difference along the wafer, and could cause gate oxide breakdown. An optimized magnetic field, consisting of flux lines parallel to the wafer, was found to reduce the degradation of the gate oxide.

A new high-density plasma etching system has been developed using a dipole-ring magnet (DRM). The system utilizes a parallel magnetic field up to 600 G with excellent uniformity extending over 250 mm in diameter. The nonuniformity of plasma was compared with that of a conventional permanent-magnet-enhanced plasma using a gate oxide integrity test. The plasma generated using DRM produced no gate oxide degradation, while the conventional magnetron plasma produced some gate oxide degradation under the most highly accelerated conditions. Si etch rate is shown to depend strongly on magnetic field strength, increasing from 1.3 µm/min at 120 G to 2.1 µm/min at 600 G.

Thin-film formation of a positive photoresist was investigated experimentally in terms of weight loss and temperature shift of the photoresist solution on a glass wafer during the spinning process. The former is analyzed by gravimetry and the latter by thermometry. The results show that weight loss of the photoresist solution is dominated mainly by outflow and is influenced only slightly by the evaporation of the solvent during the spinning process. Furthermore, film uniformity has been improved by a large quantity of radial flow, during which the solution behaved as a Newtonian fluid, at the middle stages of the spinning process. Film uniformity is degraded by non-Newtonian flow and the temperature difference in the glass wafer at the later stages of the spinning process. Consequently, a most suitable spinning time (minimum thickness deviation) exists for each spin velocity.

Simulation procedures of transverse gas-flow laser amplifiers for power extraction are intrinsically two-dimensional. The power development in the laser amplifier, which should be fully coupled to the flow-field variables, depends on a number of hydrodynamic and geometric parameters. In order to simplify the optimization of them, a quasi-two-dimensional model is proposed.

The effect of heat treatment on the taste of milk was investigated using a taste sensor. The transducer is composed of seven electrodes with different kinds of lipid membranes. Sensory evaluations by humans were made in terms of three taste characteristics of "richness (koku)," "cooked flavor" and "deliciousness" together with a measurement of whey protein denaturation. This study provided a quantitative description of the taste change caused by heat treatment of milk, because the output showed high correlations with richness and the degree of protein denaturation.