The concept of a MSMPR crystallizer, in which all crystals are assumed to disperse perfectly in suspension regardless of the crystal size, has mostly been used in the study on crystallization in order to simplify the discussion. However, there is a possibility that the spacial suspension density distribution in a crystallizer affects the crystal size distribution of the product. In this work, the effect of the spatial suspension density distribution in the crystallizer on the crystal size distribution of the product has been studied experimentally by using two types of continuous crystallizers. For expressing the mixing degree of all crystals of all sizes in the crystallizer, the multi-component mixedness defined by Ogawa was used after the definition was expanded. Experimental results show that the suspension density and the crystal size distribution are different at different heights in the crystallizers. The variation of the mixedness with impeller rotational speed is made clear. It is clarified that the mean crystal size of product crystals increases with decreasing mixedness. The usefulness of mixedness as an indicator of the classification in the crystallizer is made clear.
In this study, bubble formation in high pressure systems is investigated. The volume and shape of bubbles formed at an orifice submerged in liquid under high pressure are experimentally measured. To elucidate the bubble formation mechanism and to estimate the bubble volume in high pressure systems, a non-spherical bubble formation model is used. When the volume and shape of the bubble are calculated by the model under highly pressurized conditions, surface tension should be considered as a function of system pressure. The volume and shape of bubble are estimated relatively well by the model, including the change in surface tension with the system pressure.
The effect of two-dimensional progressive waves on the surface of a flowing liquid upon gas absorption is investigated theoretically. The convective-diffusion equation for mass transfer is solved on the basis of a boundary layer approximation. An analytical solution for the dimensionless average mass transfer coefficient is obtained as a function of three dimensionless numbers; the dimensionless streamwise distance, the relative wave velocity ratio (c – us)/us0, which is defined by the ratio of the difference between wave velocity c and time-average surface velocity us to the amplitude of surface velocity fluctuation us0 in the streamwise direction, and the surface velocity ratio us/us0. The solution shows that the effect of progressive wave motion relative to the moving liquid surface on gas absorption is very significant.
Low-energy electrons generated in a corona-discharge reactor are captured by electronegative impurities, producing negative ions. The ions migrate in the electric field to the anode (reactor wall) and are removed at the wall. The effects of the reactor structure, namely, the cathode diameter, the anode (reactor) shape and the number of cathodes on the removal efficiency with respect to three dilute gaseous pollutants, methyl iodide, chlorofluorocarbon and acetaldehyde, are investigated experimentally. The results reveal that the thicker the cathode diameter, the higher the removal efficiency. In contrast, the smaller the reactor diameter among three equivolume reactors, the higher the removal efficiency. As for the number of cathodes in a single reactor vessel, the single-cathode reactor exhibits higher removal efficiency than the 5-cathode one.
Ultra fine TiO2 particles which dispersed in various organic solvents were synthesized through hydrolysis of a titanium alkoxide to form colloidal TiO2 particles, followed by modification of the colloidal particles with a hydrophobic coupling agent. The organic solvent solutions of the synthesized TiO2 particles were transparent and had high colloidal stability. Photochemical properties of the TiO2 particles in organic solvents were studied: namely how Ti3+ species were formed in the particles and accumulated in them; and how the particles were photocatalytically active for oxidation of I– and reduction of methylene blue and Ag+. From comparative studies of the photocatalytic reactions, the TiO2 particles are found to have higher photocatalytic activity than commercially available anatase and rutile powders.
The effects of composition and reaction conditions on product distribution are investigated over various Fe promoted Cu-based catalysts to improve performance for synthesis of hydrocarbons from hydrogenation of carbon dioxide. The formation of carbon monoxide is suppressed and the formation of hydrocarbons increases with the increase in reaction temperature. Synthesis of hydrocarbons is expressed by the Schulz-Flory distribution function. A synergetic effect between copper and iron is required for hydrocarbon synthesis.
The effects of feed S/Cd ratio and feed thiophenol (PhSH) concentration on a PhSH modification reaction and the redispersion property of CdS ultrafine particles prepared in situ in reverse micellar systems are investigated on the basis of the stoichiometric balance of the reactants. The concentration of PhSH connected with the surface of CdS ultrafine particles (CSHcon) is estimated as a function of the feed S/Cd ratio, y. The experimental relationship between particle redispersion ratio R and y can be explained by assuming that R is a function of CSHcon. The effect of decrease in the feed PhSH concentration on R is investigated using the calculated values of CSHcon. The calculated R profile agrees very well with the previous experimental data taken without adding pyridine. Pyridine addition is likely to be effective for increasing the value of R by preventing the irreversible coagulation of CdS particles.
Erythromycin, a macrolide antibiotic, permeates through liquid membranes without a carrier. The contribution of local resistances to the overall permeation resistance is conceived from the permeation mechanism. According to measured distribution equilibria of erythromycin between aqueous and oil phases and previously reported dissociation equilibrium of erythromycin in the aqueous phase, the distribution coefficient of free erythromycin is constant for each oil phase and independent of proton concentration. In a supported liquid membrane, a high degree of concentration is attained in the case of a 1-decanol membrane. Not only the membrane phase but also the two aqueous phases contribute to overall resistance. For the heptane membrane, the controlling step is predicted to be in the membrane phase. An emulsion liquid membrane also concentrates erythromycin in the stripping phase, however, it gives a low degree of concentration which is ascribed to undesirable reduction of the overall permeation coefficient. For a m-xylene membrane, the controlling resistance is in the membrane and stripping phases, and for the heptane membrane, it is in the membrane phase.
Different kinds of negatively charged membranes, namely poly (acrylonitrile-co-acrylic acid) (AN-AAc), sulfonated polysulfone (SPSF) and sulfonated polyethersulfone (SPES) membranes, were prepared by changing the contents of carboxyl and sulfonate groups in the membrane polymers. Using the prepared membranes, sorption and pervaporation experiments were conducted at 333 K in the mixtures of water and ethanol. In evaluating the sorption and pervaporation properties of these membranes, the values of ΔHWM and ΔHEM, which are the difference in solubility parameter between water and membrane polymer and that between ethanol and membrane polymer, respectively, are defined. Irrespective of the kind of membrane examined, a linear relation is observed between logarithmic values of selective sorption factor of water to ethanol, βW/E, and (ΔHEM/ΔHWM), as well as between those of separation factor, αW/E, and (ΔHEM/ΔHWM). It is found that the values of βW/E for cellulose acetate and polyion complex membranes can be evaluated from the relationships obtained in the AN-AAc, SPSF and SPES membranes.
This paper presents a new approach to the analysis of unstable phenomena in a continuous crystallizer from the viewpoint of picture processing. Suspension in the NaCl-water-ethanol crystallizer was two dimensionally observed. Unstable crystal suspension pictures corresponding to sustained oscillation were obtained in the experiment when mean residence time becomes larger.
The pervaporation (PV) of an aqueous solution by a porous hydrophobic membrane is considered as simple distillation in a membrane pore. It’s separation factor (α) is generally smaller than the equilibrium one (αeq) by simple distillation, but we obtained larger separation factors than the equilibrium one under limited experimental conditions in the previous work. In order to investigate this phenomena, we carried out adhesion experiments using two types of membranes, hydrophobic and hydrophilic membranes. Hydrophobic membranes exhibited ethanol-selective adsorption properties. With the results of the adhesion experiments, we rearranged the previous data and obtained the following results. In the case that the downstream (permeate side) pressure P2 was small enough and close to vacuum, PV was controlled by simple distillation. Permeate flux (JT) was proportional to (P* – P2) and α was smaller than αeq. In the case that P2 became large and was close to the saturated pressure P*, PV performance was significantly affected by the hydrophobicity of the membrane. JT was small and α became large and exceeded αeq because the contact time of solution with hydrophobic membrane became long and the concentration near the surface on the membrane increased by the selective ethanol adsorption properties on the hydrophobic membrane.
Batch cooling crystallization of potassium alum was performed in the presence of an azo dye, Bismarck Brown G, as a crystal-growth suppressor. The crystal size of the product was caused to increase unexpectedly by the presence of the growth suppressor. This is closely related to the decrease in the number of crystals in the crystallizer and the resultant increase in supersaturation. Such a suppressor effect is pointed out to be important from an industrial point of view.
When a microparticle slurry is continuously dried in a powder-particle fluidized bed, agglomerates of microparticles and coarse particles are produced at high drying efficiency, and continuous operation is not achievable. To avoid agglomeration, it is very important to fluidize the coarse particles vigorously in the bed. For the same purpose, the continuous drying of microparticle slurry is investigated in a powder-particle spouted bed. In this dryer, the microparticle slurry can be continuously dried with higher drying efficiency comparing with an ordinary powder-particle fluidized bed. Limited drying efficiency is defined as the maximum drying efficiency for continuous operation at a certain gas velocity. An empirical equation for estimating the limiting drying efficiency is obtained.
Hold-up of cohesive fine particles in a powder-particle spouted bed under continuous fines feeding was investigated. Nine powders, aluminium hydroxide and aluminium oxide materials, ranging in particle size from 0.51 μm to 3.5 μm (Group C powders in Geldart’s classification) were used. Silica sand particles of 920 μm in average diameter (Group D powders in Geldart’s classification) were used as coarse particles. Cohesive force of the fines was determined by the tensile strength method. The effects of superficial gas velocity, the feed rate of fines, the size of fines and cohesive force among fines on the hold-up of fines in the bed at a steady state were examined experimentally. It is found that fines hold-up is dominated by cohesive force and the effect of operating conditions upon the fines hold-up could be explained from the view point of the cohesive force.
Based on the concepts of clustering suspension and core-annulus flow in the riser, a pressure balance model for the entire loop of a circulating fluidized bed system was developed, where the gas flow between the riser and the downcomer was first taken into account. For the present model, the following correlation for the decay factor of the voidage profile was developed within ±30% accuracy:
The model was successfully validated by a comparison of the predicted pressure loop distribution, the maximum solid flux and the dense region height with the previous experimental results. By taking into account gas flow through the loop seal between the riser and the downcomer, it became possible to predict quantitatively and rigorously the effect of solid inventory on the axial voidage profiles in the riser reported by Weinstein et al. (1984) and Mori et al. (1989). It was also found that there is an operable extension of solid inventory for a given set of superficial gas velocity and solid circulating flux.
An attempt is made to study the fluidization and the surface-to-bed heat transfer coefficient in a fluidized bed of very fine (∼10 μm) hydrogen-absorbing alloy (HAA) powder. Considering the difficulty in handling HAA powder fluidized directly by hydrogen, experiments were carried out using helium and nitrogen. In addition, two different types of nickel powders with similar size range were also employed for the investigation using hydrogen as well as helium and nitrogen as fluidizing gases. Bed pressure drop and voidage were measured with variation of gas velocity. It is found that the three powders are similar in fluidization characteristics, although some differences can be seen in shape and surface roughness of the primary particles. Surface-to-bed heat transfer coefficients were measured by immersed probes with cylindrical surfaces heated electrically. The measurement results were compared with predictions by a very simple model based on the application of the equation for single-phase flow, in which the dimensionless Nusselt, Reynolds and Prandtl numbers describe the effective properties of gas-solid suspension calculated by empirical correlations in literature. It is shown that the predictions agree with the experimental very well when modification is made to the effective velocity.
The coating of ultra-fine particles inside their agglomerates was performed by chemical vapor deposition. Mathematical models are adopted to describe the internal deposition process by chemical reaction and interparticle diffusion, and the porous structural change with the deposition progress. The effects of kinetic, diffusional and agglomerate-geometrical and structural parameters are analyzed for the radial distribution of the deposit. The experimental deposition profiles for the coating of Si3N4 ultra-fine particles with AlN are quantitatively compared with simulated ones for different reaction temperatures and agglomerate sizes. As a result, uniform deposition of AlN is proven to be obtained at comparatively low temperature with small loosely packed agglomerates.
The continuous measurement of the number concentration of growing crystals in the crystallization of fructose was attempted by using a commercial laser diffraction particle size analyzer. It is based on the relation that the maximum diffracted light intensity at the detector is proportional to the crystal number concentration and the square of mean crystal size, which is formulated for the non-growing particles. The crystal number concentration measured by the laser diffraction particle size analyzer almost agrees with that by microscopic observation unless generation of small crystals such as secondary nuclei occurs. The generating point of small crystals is detected approximately by monitoring the ratio of mean diffracted light intensity at the whole detector to the maximum one.
This paper develops a learning-type multi-loop control system for interacting multi-input/multi-output industrial process systems. The recently developed single neural controller (SNC) is adopted as the decentralized controller. With a simple parameter tuning algorithm, the SNC in each loop is able to learn to control a changing process by merely observing the process output error in the same loop. To circumvent loop interactions, static decouplers are incorporated in the presented scheme. The only a priori knowledge of the controlled plant is the process steady state gains, which can be easily obtained from open-loop test. Extensive comparisons with decentralized PI controllers were performed. Simulation results show that the presented decentralized nonlinear control strategy appears to be a simple and promising approach to interacting multivariable process control.
A new convenient method to synthesize optimal control for distributed parameter systems is presented. Merging the two conventional optimal control synthesis schemes based on partial differential and integral equations, the new method generates the optimal control when applied to an unsteady state tubular reactor model. For this type of optimal control problems, we found that the new method is more efficient than the conventional methods in computing procedure and in computing time.
An appropriate structure for a reduced-order model of transfer function type is first proposed for an ideal heat integrated distillation column (HIDiC) based on the process dynamics. The reduced-order model could well represent the process not only in steady state but dynamic state as well. A multivariable internal model control (IMC) system is designed and applied to the ideal HIDiC. The simulation results demonstrate that the IMC controller can deal with the interactive nature of the ideal HIDiC more effectively than PI controllers, indicating the IMC controller could be a better means for operation of the ideal HIDiC in the case where high system performance is required. However, the ideal HIDiC is extremely sensitive to changes in operating conditions and this makes it necessary to adopt an on-line model adaptation mechanism to facilitate the performance of the IMC controller.
A glucose sensor, using glucose oxidase immobilized in an electrically conductive polymer membrane, is developed for transient-mode measurement by on-off reaction control in the present work. An implantable glucose sensor loses its stability under the skin by fouling resulting from fibroblast adhesion on the glucose oxidase-immobilized membrane. Change in glucose distribution inside the membrane and in the vicinity of the surface of the membrane resulting from the fouling makes the measurement of the glucose concentration unstable. A technique of measuring transient current of the glucose sensor after activation of glucose oxidase may hardly be affected by the fouling as oxidation of coenzyme of glucose oxidase is rapid and the transient current is sensitive to glucose concentration. A glucose sensor with glucose oxidase immobilized in an electrically conductive polypyrrole membrane was fabricated. The activity of the glucose oxidase is switched on and off by a stepwise change of potential of the polypyrrole membrane from 0.15 V to 0.45 V versus a saturated calomel electrode. The two conditions described above can be confirmed by measuring transient current. Current increased momentarily after activation of glucose oxidase and decreased moderately after that. The oxidation and reduction of glucose oxidase is sufficiently rapid, and current intensity depends on glucose concentration in the test solution.
The protection (stabilization) effect of various hydrophobic ligands on the denaturation and aggregation of carbonic anhydrase from bovine (CAB) has been quantitatively investigated under various heat stress conditions. In a limited temperature range (40∼60°C), where the protein was only partially denatured and the local hydrophobicities (LH) of CAB were positive effective stabilization of the protein is achieved by the addition of various ligands. The importance of balance between hydrophobic head and hydrophilic tail of the ligands is hypothesized.
Adsorption of Hg(II) on polyaminated highly porous chitosan (PEI-CS) from washing solutions of refuse incinerator flue gases is technically feasible. Equilibrium isotherms for adsorption Hg(II) on PEI-CS were measured by a batch method. The experimental data correlate well with the Langmuir equation in the range of about 5 to 70 mol/m3 Hg(II) aqueous solution. When HCl exists in Hg(II) solution, the saturation capacity and equilibrium constant for adsorption of Hg(II) on PEI-CS decrease with increasing concentration of HCl, CHCl. When NaCl is in Hg(II) solution, the equilibrium constant is constant, but the saturation capacity decreases with an increase in the concentration of NaCl. Breakthrough curves for adsorption of 10 mol/m3 Hg(II) were measured using a PEI-CS packed column at Re′ = 5. The intraparticle effective diffusivities Deff of Hg(II) are determined from the experimental breakthrough curves. When HCl exists in Hg(II) solution, Deff increases with increasing C0, HCl when C0, HCl < 180 mol/m3. When C0, HCl > 180 mol/m3, it shows a constant value, which is about 8 times larger than that measured in pure Hg(II) aqueous solution. When NaCl is in Hg(II) solution, the shapes of the breakthrough curves are similar to those in pure Hg(II) solution. Deff is kept constant in the range of 0 ≤ C0, NaCl ≤ 1000 mol/m3.
Recovery of mercury(II) adsorbed on highly porous polyaminated chitosan (PEI-CS) appears to be technically feasible. Desorption experiments of Hg(II) from a PEI-CS packed column were carried out using sulfuric acid solutions of different concentrations at different temperatures. The peak of the elution curve is 250 mol/m3 when the concentration of H2SO4 in the eluent is 1000 mol/m3 at 308 K. The elution curve in PEI-CS is compared with that of Unicellex UR-120H, which is considered the best commercial chelate resin available in Japan for removal of Hg(II). The peak of the elution curve in UR-120H is about one eighth at the same concentration of PEI-CS. The equilibrium isotherm and the breakthrough curve for adsorption of Hg(II) are not affected after repeated cycles of adsorption and desorption.
Cellulose, a major component of woody biomass, was reacted in hot-compressed water under catalyst-free conditions at different reaction temperatures from 200 to 350°C. The product distribution was analyzed at different temperatures and/or holding times. A simplified reaction scheme is proposed, in which char-like residues are produced from cellulose through water-soluble products and oils, as intermediates. Experiments with glucose and cellulose feedstock resulted in the same product distribution, suggesting that cellulose decomposition starts with a first hydrolysis step. Levoglucosan and 5-(hydroxymethyl)furfural (HMF) were also detected. The role of alkali catalysts in cellulose liquefaction is discussed.
In order to apply a seeding method to the removal of silica from geothermal brines, the effects of temperature, initial seed concentration and initial silica concentration on the silica removal were experimentally examined. When silica gel seeds are added to a supersaturated silica solution, amorphous silica is precipitated on the pore surface of the seeds and the silica concentration in the solution decreases. The silica removal ratio, which means the ratio of the removed silica to the excess silica, increases with increasing initial seed concentration and increasing temperature. A high silica removal ratio is obtained for relatively low initial silica concentrations (≤0.7 kg-SiO2·m–3). When the initial silica concentration is higher, the silica removal ratio is expected to be enhanced by increasing initial seed concentration. The seeding method is suggested to be effective for the removal of silica in actual geothermal brines.
The behavior of solute transfer in sodium dioleylphosphate (SDOLP)/n-heptane and SDOLP-AOT/n-heptane systems was elucidated by focussing on the extraction process in Winsor II and the solute-exchange process between microemulsion droplets. The extraction of tryptophan (Trp) between the organic and aqueous phases is rather fast for the simultaneous extraction of water and Trp in the SDOLP and SDOLP/AOT systems, whereas in the extraction of Trp without water transfer between the microemulsion droplet and the aqueous phase, the rate slows down significantly with the addition of SDOLP in the AOT system. The rate of the interdroplet exchange process is insensitive to the presence of SDOLP. The reason for the difference in these two processes is discussed with the help of a model of the exchange processes.