Achieving a uniform film thickness profile over a substrate has been difficult in CVD-Al₂O₃ coating processes. Although this problem can be solved by H₂S-doping to the source gas mixture, AlCl₃/CO₂/H₂, the role of H₂S in the mechanisms of the CVD coating have not been clarified. In this study, the effects of H₂S-doping on the improvement of the uniformity of a coated layer is studied, focusing on the particles generated in the gas phase. It is found in the measurement of gasborne particles that the number concentration of particles larger than 200 nm in diameter is reduced dramatically by the H₂S-doping. The concentration is over 10⁸ particles/m³ when no H₂S is doped, while it is less than 10⁶ particles/m³ at 0.20% doping. The improvement of the coating by H₂S-doping is concluded to be caused by a size reduction of particulate matters of the Al₂O₃ precursors that leads to an increase of the diffusivity.

View full abstract

Pretreatments before epitaxial furnace insertion were successfully performed for protective oxidation by ozonized ultrapure water and hydrogen termination by diluted HF. Before and after pretreatment, the surfaces were evaluated for surface cleanliness and chemical oxide thickness by X-ray photoelectron spectroscopy, and surface micro-roughness by atomic force microscopy. For both processes, a clean and smooth surface was obtained by removing the protective oxide and oxygen residues in heat treatment at 950°C in hydrogen. Pretreatment time varied for both processes, 5min for protective oxidation and approximately 1min for hydrogen termination. Attaining both pretreatment, a high quality epitaxial layer was obtained at conditions of 950°C, 1.01×10⁵ and 4.00×10⁴Pa.

View full abstract

Chemical vapor deposition (CVD) films are widely used to make Very Large Scale Integrated Circuits (VLSIs). For optimum VLSI performance, precise control of both film thickness and composition is required. For this purpose, we carried out three-dimensional computer simulation of WSix CVD processes using WF₆-SiH₄ source gases in a cold-wall and single wafer type reactor. For this simulation we used the reaction model, the radical chain reaction model for CVD film formation developed by using analysis of the reaction kinetics from the results in the simplified reactor. By this simulation, agreement between simulated and measured values was approximately 5%. On the basis of this result, we investigated the effect of gas outlet number and outlet height position on the radial distribution of film deposition rate, composition and uniformity. When we selected the suitable outlet height position, the variation was improved remarkably less than 4%. This tendency might be explained by the gas velocity and gas composition distribution in the reactor. These results demonstrate how CFD simulations can be used to optimize CVD reactor design.

View full abstract

In this study, alumina thin films were prepared by chemical vapor deposition from aluminum chloride, hydrogen and carbon dioxide, and growth kinetics were investigated. The deposition experiments were carried out in a non-isothermal tubular CVD reactor. The deposition rate of alumina and temperatures at inner surface, and pressure near the inlet and exit of the reactor were measured. To clarify the growth kinetics, the numerical model, in which the axial distributions of temperature and pressure in the reactor were estimated from the measured values by the fourth-order and linear polynomial interpolation, respectively, has been developed. To confirm these estimations, the distributions of temperature and pressure were calculated using a two-dimensional numerical analysis. The measured and estimated distributions of the temperature and pressure agreed well with calculated results in He and H₂ streams. In an Ar stream, the measured temperature did not agree with the calculated gas temperature near the inlet and outlet of the reactor where the temperature was drastically changed in the axial direction. Therefore, the deposition experiments were performed in a He stream. The oxidation of H₂2 by CO₂, which is the rate-controlling step for alumina CVD, mainly occurrs in the gas-phase, and its activation energy is found to be 235kJ/mol.

View full abstract

We proposed a new method to evaluate the interaction between a substrate and metal clusters based on computational chemistry with a reduced calculation time. This method is applied to study the interaction between a SiO₂ surface and Au clusters to elucidate the potential energy as well as the energetics of Au clusters on the SiO₂ surface. The method comprises three steps. The first step is evaluation of the interaction between the SiO₂ surface and Au atom using ab initio calculation. Secondly, we determine the parameters that are used in two-body potential functions to reproduce the interaction evaluated by ab initio calculation. In the third step, we investigate the interaction between a SiO₂ surface and Au clusters using the two-body potential functions. This study reveals that the atomic-scale roughness of the substrate surface predominantly determined the most stable point of the Au clusters on the SiO₂ substrate. We also find that the shape and orientation of the Au clusters affect the interaction.

View full abstract

Size distribution measurements and classification performance tests under low pressure conditions ranging from 123 to 300 Torr were systematically carried out for nano-meter sized silver particles formed in a nitrogen gas flow at atmospheric pressure using a new type of differential mobility analyzer (LPDMA) based on the Vienna-type design concept which has a short classification zone. Although the values of the volumetric average diameter, D_{V, DMA}, are approximately 7% larger than those of the volumetric average diameter, DV, TEM, obtained from the transmission electron microscope (TEM) observation results of the silver particles collected at the LPDMA inlet, both the DMA and TEM size measurements are found to have good correlation in the range from 6 to 15nm. When the ratio of the sheath gas flow rate and aerosol gas flow rate is set at 5: 1, the theoretical sizes calculated from the classification voltage applied to the LPDMA at 160 Torr are also found to be in good correlation with those of D_{V, TEM} obtained from the TEM observation results of the silver particles collected after the classification by the LPDMA in the range from 6 to 25nm, although the theoretical sizes set by the LPDMA are approximately 14% larger than those of D_{V, TEM} measured from the TEM observation results after the classification by the LPDMA.
The geometric standard deviations of the size distributions obtained from the TEM images after classification range from 1.08 to 1.17, and they are approximately equal to the theoretical value (1.093) of the particles classified by the LPDMA, which is derived from the assumption that the broadening effects by Brownian diffusion are ignored, whereas the geometric standard deviations of the size distributions obtained from the TEM images before the classification range from 1.34 to 1.50 and are greater than those after classification. Thus, we experimentally demonstrate the validity of our LPDMA system in size measurements and classification of the nm-sized particles under low pressure conditions.

View full abstract

The roles of reflectors and a substrate in the circular lamp heating furnace are systematically evaluated, for the first time, using the direct application model using the ray trace simulation. The rays emitted from the lamp are concentrated and distributed in the region of the concentric circle formed by the cylindrical reflectors, in order to improve the heat intensity and heat uniformity. With increasing reflectivity of the substrate surface and with increasing the distance between the light source and the substrate, the ray intensity absorbed by the substrate decreases and its profile tends to be flatter. The rays emitted from the heated substrate surface can reach near its emission site. The heat loss at the outer region of heated substrate is evaluated to be larger than that at the inner region.

View full abstract

The effect of gas composition on the plasma structure is investigated using a self consistent continuous fluid plasma model, which assumes that the local field approximation is attained in the RF plasma. Boltzmann analysis was performed for the estimation of swarm parameters for the various amount of O₂ containing Ar, where the oxygen content was from 1% to 100%. Positive ion density, negative ion density and electron density in the Ar/O₂ mixture 1D rf plasma are calculated using a self -consistent continuous fluid plasma model. Though the calculation results show that swarm parameter is not greatly affected by the amount of oxygen, the plasma structure is drastically changed even for the 1% oxygen containing. It was concluded that negative ion in the oxygen plasma play an important role in the mixture gas plasma structure, and the analysis for the mixture processing gas should be done by paying attention to negative charged particles formation in the plasma.

View full abstract

CFD simulation of a hot wall type CVD reactor with a shower head gas supply is carried out as numerical experiments for a wide range of reaction conditions and rate constants. The results of these numerical experiments are summarized into equations of dimensionless numbers, and these empirical equations can be applied to quick estimation of growth rate and its dependency on operating conditions.
For the reaction scheme of direct deposition of source gas, growth rate, G.R., is expressed as follows,
whereδ, the thickness of the concentration boundary layer is
In the case of deposition of intermediate species produced in gas phase reaction, G.R., is expressed as follows,
whereδR, the thickness of heating zone above the substrate is
Here, W is the distance between substrate and shower head, Re is Reynolds number based on W, Sc is Schmidt number, Le is Lewis number, T_s is substrate temperature, T₀ is inlet gas temperature, ks is surface reaction rate constant, D and D_s are diffusion coefficient at T₀ and Ts respectively, k_g0 and Ea are pre-exponential factor and activation energy of gas phase reaction respectively, R_g is gas constant and C₀ is inlet source gas concentration.

View full abstract

The effects of pressure and flow rate of gas in a CVD reactor are studied on the GaN epitaxial thin film preparation by a metalorganic chemical vapor deposition (MOCVD) method using trimethylgallium (TMG) and ammonia as the source gases. The thickness profiles of GaN films are measured for those grown on a substrate heated at 1373K in a single-substrate, vertical, cold-wall reactor. A model to describe the gas-phase chemical reactions leading to the formation of adducts and polymers from TMG and ammonia is introduced to simulate the GaN film preparation process. The growth rate profile of GaN films is then predicted by numerically calculating the distributions of flow velocity, temperature and concentration of species in the reactor. The calculated results reproduce well the measured change in growth rate with a total pressure change from the atmospheric to its 1/8, and a flattening of growth rate profile at an increased flow rate and a reduced pressure. It is concluded from the simulation that the changes of growth rate profile are caused by the influence of recirculating flow in the reactor on the GaN precursor concentration distribution above the substrate.

View full abstract

The effect of plasma formed gas on the plasma structure is investigated using a self consistent continuous fluid plasma model, which assumes that the local field approximation was attained in the RF plasma. The swarm parameter was obtained by Boltzmann analysis for O₂ gas which contains various amounts of O radicals. Positive ion density, negative ion density, and electron density in the O₂ RF plasma including O radical generation were calculated by self-consistent one dimensional continuous fluid plasma model. The calculation results show the dominant positive charged particle became O+ from O₂+ with increasing amount of O radical in O₂. The plasma positive densities are not greatly affected if O radical content did not exceed 1% of O₂. It was concluded that O₂ plasma structure can be evaluated by source gas plasma calculation neglecting radical formation, since O radical formation in the O₂ plasma is estimated as less than 0.2% of O₂.

View full abstract

The authors carried out nanofiltration of skim milk and analyzed the salt rejection data using the extended Nernst-Planck equation.
The following observed rejections of major constituents in salts were obtained. Na and K: 50 to 60%, Cl: -5 to 15%, Ca, Mg and P: >97%. It is considered that since Ca and Mg are mostly combined with phosphate and citrate in milk, they were rejected by a sieving effect, thus producing high rejections. Electroneutrality in permeate, and no-electric-current condition in solute flux are observed, respectively.
In the subsequent analysis, the authors used a set of proposed assumptions, 1) a complete rejection of solute except Na, K and Cl by the membrane, 2) existence of a fixed charge in skim milk due to impermeable ions of proteins, etc. Rejection of Na, K and Cl calculated from the analysis are in good agreement with those of the experimental data.

View full abstract

An iron oxide bubbling fluidized bed reactor in the direct causticizing process, which constitutes the main part of the process, is developed. Its optimum operating conditions and reaction characteristics were investigated through a 500mm sq. bench scale test. In this process, high-grade iron oxide, whose diameter is 0.2-1.0mm was used as causticizing agent as well as the bed material. Concentrated black liquor was fed over the bubbling iron oxide bed by means of a spray nozzle at the temperature of 1, 173-1, 273K. Reacted sodium ferrite was hydrolyzed to recover sodium hydroxide and residue iron oxide was recycled. During the reaction in the bed as well as hydrolysis, iron oxide particle is gradually attritted so that the operation of fluidized bed becomes difficult. A portion of the iron oxide fine particles was pelletized by compaction before it was recycled to the reactor. The minimum ratio of fines to be pelletized is found to be 50% to maintain the bed as well as to recover sodium from black liquor through 40hrs' of continuous operation of the bench scale test.

View full abstract

A pulsated fluidized bed is operated with two sequential duration designated as the on-period with injecting fluidization air, and the off-period without injecting fluidization air. The heat transfer coefficient between a vertically immersed heater and bed in a pulsated fluidized bed is measured under various pulse cycles and fluidized particles. The obtained results are compared with those in a normal fluidized bed with continuous fluidization air injection. The relation between heat transfer coefficient and bubble characteristics, evaluated by using a digital video camera, was investigated. For certain fluidized particles and operated pulse cycles, fluidization of particles and increment of heat transfer coefficient can be obtained under lower mean air velocity based on a pulse cycle duration than minimum fluidization air velocity of the normal fluidized bed. Under the pulse cycles where a static bed is formed in off -period duration, the improved heat transfer rate in contrast with the normal fluidized bed can be measured. This is because of the large bubbles formation. As the heat transfer in the pulsated fluidized bed is obstructed with increasing time to keep a static bed due to the excess off-period duration, it is indicated that there is an optimum off-period duration based on the heat transfer rate.

View full abstract

A macroscopic momentum equation for flows through porous media of dual structure was derived by integrating Navier-Stokes equations over a local control volume. The smaller scale-fluid motions associated with the “secondary” structure were modeled first. Then, the larger scale-fluid motions within the “primary” structure were modeled successively, to obtain the macroscopic momentum equation for porous flows through porous media of dual structure. Both pressure measurement and numerical experiment conducted for the flows in porous media of dual structure substantiate the validity of the present macroscopic momentum equation.

View full abstract

The behavior of pesticides (dymron, mefenacet, bensulfuron-methyl) was measured and simulated in a paddy field with rapid water penetration. In the field experiment, the concentration of each pesticide in paddy water reached a maximum shortly after its application, and then diminished rapidly to a non -detectable level within three days. In paddy soil, the concentrations of dymron and mefenacet in the topsoil (from soil surface to 5cm, 0-5cm) were much higher than those in the subsoil (5-10cm) four days after application, and their concentrations in both soils became almost the same after six days. On the other hand, a significant difference in the concentration of bensulfuron-methyl between the topsoil and the subsoil could not be observed. The behavior of pesticides in the paddy field was predicted well using a mathematical simulation model. It was found that transfer of pesticide with dissolution and penetration of water significantly contributed to the behavior of pesticides in paddy water, and the spatial distribution of the pesticides in paddy soil depended on the adsorbability of each pesticide. Dymron and mefenacet were shown to be spatially less mobile than bensulfuron-methyl because of their higher adsorbability.

View full abstract

A novel separation system was demonstrated for protein purification. The pore surface of the porous substrate is covered with thermosensitive grafted polymer, and hydrophobicity of the pore surface can be drastically varied by small temperature change. A feed solution containing hydrophobic and hydrophilic proteins is continuously supplied to the feed side, and the membrane temperature is changed stepwise below and above a lower-critical solution temperature for the thermosensitive polymer. Hydrophobic proteins adsorb on the hydrophobic pore surface above the critical temperature, and only the hydrophilic protein can pass to the permeate side. When the temperature is changed to below the critical point, the pore surface becomes hydrophilic and the adsorbed proteins will be desorbed from the membrane, the the hydrophobic proteins might be condensed in permeate side. N-isopropyl acrylamide (NIPAM) was employed as the thermosensitive polymer, and the NIPAM grafted polymer was fixed in the pores of the porous base membranes by plasma-graft filling polymerization technique. The grafted polymer formation profile was controlled by changing plasma treatment and graft polymerization conditions. Above 32°C, the membrane adsorbed the globulin due to its hydrophobic interaction, and the solutes were desorbed from the pore surface below 32°C due to its hydrophilic nature. Solute adsorption and desorption took place each cycle, and the protein was concentrated in the permeate at the desorption step. The results show the potential of this separation system for protein purification.

View full abstract

In order to design the microstructure of ceramics with desired dielectric property, an estimation method of dielectric constant of ceramics taking into account the characteristics of microstructure of the ceramics is proposed. In the estimation model, the microstructure of ceramics is represented by the assembly of unit cells comprising of grain, pore and grain boundary. The sizes of grain and pore and the thickness of grain boundary in each unit cell were determined exactly according to their size and thickness distributions in a real ceramic. The dielectric constant of the assembly can be calculated on the basis of equivalent circuit theory.
The estimated values of dielectric constant of ceramic BaTiO₃ using the proposed estimation method agree well with experimental ones. The dependence of characteristics of microstructure on the dielectric constant was clarified by the estimation of dielectric constants for the assemblies of unit cells with different characteristics. The microstructure of ceramic BaTiO₃ with desired dielectric properties can be designed by using the proposed estimation method.

View full abstract

Unstructure models were developed for a mixed culture where one microorganism assimilates the metabolite produced by another microorganism. The model system was a mixed culture using Lactobacillus delbrueckii and Ralstonia eutropha where the former assimilates glucose and produces lactic acid, and the latter assimilates lactate and produce poly-β-hydroxy butyrate (PHB). Performance improvement is shown for the mixed culture over a single culture for chemostat using the model developed. Since the optimal dissolved oxygen (DO) concentration is different for each microorganism, we developed another model which takes into account the effect of DO concentration on the dynamic behavior. Then the optimal DO concentration is obtained for chemostat. Moreover, we developed another model which takes into account NH₃ concentration on the cell growth and PHB production by R. eutropha. Then the optimal time-profile for NH₃ concentration is derived using the Maximum principle. It is found that high PHB production could be attained even if NH₃ concentration is not controlled if initial NH₃ concentration is appropriately selected.

View full abstract

The oxidation reaction of squalene is investigated using a bubble column reactor under the various conditions. For oxidation, two stages, which are an induction period with a slow increase in hydroperoxide and an accelerated period with rapid increase, are observed. The peroxide concentration increases with the progress of the oxidation and reaches a maximum value in the last part of the accelerated period, and then decreases again due to the decomposition of peroxide. The peroxide produced is analyzed by a HPLC system equipped with chemiluminescence and UV/VIS detectors. The main component of the decomposition of peroxide was confirmed to be carbonyl compounds based on the results of the hydroxylamine hydrochloride titration, IR, ¹³C-NMR and LC-MS analyses. The formation rate equation of peroxide is derived from the chain reaction model of lipid radical and the experimental results were analyzed using the rate equation. The formation rate of the peroxide is independent on the peroxide concentration in the induction period, and is expressed by the first order for the peroxide concentration in the accelerated period.

View full abstract

We investigated the phosphate ion removal ability of an adsorbent, which was prepared by baking volcanic ash soil mixed with iron (II) sulfate at 500°C. It is found by batch experiments that the removal rate of phosphate ion is initially rapid, and falls gradually. Several tens of days are required to achieve equilibrium. The lower the solution pH, the higher the removal amount of phosphate ion. Regardless of the initial pH, the solution pH is closer to the neutral along with time. It is thought that the pH rise of acidic solution is due to the release of OH- attached to Al3+ and Fe³⁺ by ion exchange with phosphate ion, and the pH drop of alkali solution is due to the formation of H+ by the hydrolysis of Al³⁺ and Fe³⁺ dissolved from the surface of baked soil. A column experiment was conducted with PO4-P concentration of 3mg/dm³ and SV=2 (1/hr). The effluent PO₄-P concentrations remained less than 1mg/dm3 for 65 days for influent pH of 6.00 and for 35 days for that of 7.55, respectively.

View full abstract

The capabilities of Candida rugosa lipase, which was immobilized into organically modified silicates formed on kieselguhr such as Celite 545 are investigated in esterification of (-) -or (+) -menthol with butyric acid. In this study, we optimized the preparation conditions such as a molar ratio of precursors. From these investigations, we find that the optimized hybrid gel-immobilized lipase is derived from a mixture of dimethyldimethoxysilane and tetramethoxysilane at a molar ratio of 4: 1 and its enantioselectivity, which is defined as a ratio of the initial reaction rate of (-) -menthol to that of (+) -menthol, is enhanced up to 41 at 75°C. In addition, we observe the temperature dependence of these immobilized lipases. Consequently, the optimized hybrid gel-immobilized lipase shows higher activity than deposited lipase on Celite at all temperatures, and it is suggested that this activation results from a remarkable enhancement of the V_max value.

View full abstract

The solubility of triolein in supercritical carbon dioxide (SCCO₂) was measured by a continuous flow system at various temperatures (313, 323, 333 and 343K) and pressures (15, 20 and 25MPa). The solubility of triolein increases with experimental pressure. The obtained data were compared with previous investigations. At 15MPa, the solubility is considerably governed by temperature (343-313K), i. e. 1010²-fold change of solubility. On the other hand, at 25MPa, which is fully mature supercritical conditions, the influence of temperature on the solubility become weaker. Under constant pressure conditions, solubility of triolein decreases gradually with increasing temperature. Solubility is correlated well with the estimated density, according to Bender (1970).

View full abstract

The authors investigated the performance of atomization with a spinning disk in a spray dryer for suspensions behaving as a non-Newtonian fluid under high-shear rate. Droplet size under various operating conditions was compared with those of previous works for Newtonian fluid. It is found that the mean diameter of the droplet for suspensions can be calculated from Friedman's equation, using the viscosity given by Mooney's equation for suspension in complete dispersion.

View full abstract

A Microchannel (upper width=280μm, lower width=138μm, depth=100μm, length=27mm) was formed on a (100) silicon wafer by means of wet chemical etching, and a platinum layer was then coated on the microchannel walls by sputtering. The resulting channel was sealed with a glass cover by an anodic bonding technique. Cyclohexane vapor, carried by a stream of nitrogen, was then introduced into the microreactor at 400°C, and the concentrations of both the reactant, and the products of the ensuing dehydrogenation reaction over the platinum catalyst, were determined by means of a micro gas chromatograph. Thus, a series of procedures for manufacturing and testing a microreactor, such as lithography of a microchannel, the formation of a catalytic Pt film, the introduction of a reactant into the covered microchannel, or an analysis of reactants and products, was established and verified.

View full abstract

The possibility of particle coating in a solution is to be expected under such a condition that causes only heterogeneous nucleation.
In a glass-water-CaCO₃ system, it is proven that there is a certain range of the solution concentrations for heterogeneous nucleation by the precipitation method. It is found to be important in practice to come up with some ideas of the way of supplying and mixing of solution, because its range is Limited to a narrow range.

View full abstract