Outstanding Paper Awards Subcommittee of Journal of Chemical Engineering of Japan has assessed the 120 papers published in volume 50 in 2017, and the editorial board finally selected the five papers for JCEJ Outstanding Paper Awards of 2017; these are the papers on “Effect of Co-products on Pd Membrane Performance in Membrane Reforming of Desulfurized Kerosene,” “Classification Characteristics of a Cyclone Type Classifier with Improved Collection Boxes for Separating Particles near the Wall Surface,” “The Solubility Characteristics of Aluminum Sulfate in Sulfuric Acid Aqueous Solution in the Presence of Oxalic Acid,” “Starting Torque of Vertical Paddle Impellers,” and “Separation of Aromatic Compounds from Hydrocarbon Mixtures by Vapor Permeation Using Liquid Membranes with Ionic Liquids.”

Robust and efficient modeling of industrial processes is vital in realizing stable and economical process design, operation, and control. However, inherent uncertainty of the processes makes the modeling very challenging. Uncertainty is of three dimensions: location, level, and nature. The location of uncertainty identifies a source of uncertainty, the level of uncertainty places the source of uncertainty within the range between absolute determinism and total ignorance, and the nature of uncertainty classifies the source of uncertainty into reducible or non-reducible uncertainty. This study focuses on identification of these dimensions of uncertainty in process modeling. In this regard, models of a steelmaking process and a naphtha reforming process are taken as case studies, and all the three dimensions of uncertainty are elaborated with relevant examples from the two process models. Also, methods of uncertainty analysis are discussed. This review aims to provide a platform for model developers to identify the sources of uncertainty and efficiently model its effect on decision making in process design, operation, and control.

The metastable phase equilibrium of the quaternary system K⁺, Mg²⁺, Ca²⁺//Cl⁻–H₂O at 298.2 K is done using an isothermal method. The equilibrium liquid phase compositions, densities and refractive indexes of the equilibrated solutions are measured at the same time experimentally. The crystalloid forms of the solid phase are determined using an X-ray diffraction method. Experimental results show that this is a complex system at 298.2 K with two double salts (carnallite KCl·MgCl₂·6H₂O and tachyhydrite 2MgCl₂·CaCl₂·12H₂O). The metastable phase diagram of this system under the investigated temperature contains three invariant points, five crystallization fields and seven univariant curves. With the changes of J(CaCl₂), the water content, densities and refractive indexes change regularly.

Data are presented on the solubilities, densities, and refractive indices of the aqueous ternary systems NH₄⁺, Sr²⁺ (Ca²⁺)//Cl⁻–H₂O at 298 K by using the isothermal dissolution method. Stable phase diagrams and the diagrams of the physicochemical properties (densities and refractive indices) depending on the composition were constructed on the basis of measured data. The solid phase was determined using Schreinemakers’ wet residue method and X-ray diffraction method. Results show that these two systems which consist of one invariant point, two univariant curves and two regions of simple salts’ crystallization, belong to simple type, without solid solution or double salt formed. In the stable phase diagram of Ca²⁺, NH₄⁺//Cl⁻–H₂O, the region of CaCl₂·6H₂O crystallization is smaller than that of NH₄Cl. Salt CaCl₂·6H₂O has a salting out effect on NH₄Cl. In the stable phase diagram of Sr²⁺, NH₄⁺//Cl⁻–H₂O, the crystallization region of SrCl₂·6H₂O is smaller than that of NH₄Cl, and salt SrCl₂ has salting out effect on NH₄Cl.

In this paper, the agitation characteristic curve of medium light particles in stirred tank is proposed based on numerical simulation study. The agitation characteristic curve contains the power number curve as the input of the stirring system and the particle dispersion curve as the output. The standard k–ε turbulence model together with the volume of fluid model (VOF) and the discrete phase model (DPM) were used to simulate the particle distribution in an unbaffled tank with free surfaces. Based on the standard deviation, dispersion degree curves in both axial and radial directions (σ_la and σ_lr) were proposed to quantify the discrete particle distribution in the tank. Compared with the instantaneous fluid field, the dispersion degree curves represent well the particle distribution for different rotation speeds and immergences. Besides, the power number (N_p) was used to quantify the power consumed during the drawdown process of floating particles. This work shows that the agitation characteristic curve would serve as a benchmark so that stirring systems with different configurations could be compared and optimized.

An experimental and supplementary numerical model for ceramic filtration with considering the effect of particle deposit is proposed. A comparison between the model and experiments indicates that the early stage of initial model agrees well with the experiments but the deviation is increasingly growing as time passing because of the continuous detachment of dust. To ensure predictive accuracy, the Fourier fitting method was applied to acquire the variation of specific deposit in the detaching stage (R-square≥0.9997). A modified model consisting of three stages was proposed by using initial model on depositing stage without detachment, employing Fourier fitting equations on depositing stage with detachment. Then, the influence of specific deposit and mean porosity (of ceramic-cake) on pressure drop was studied. It suggests that heavy specific deposit, small mean porosity and large wall thickness significantly help the improvement of pressure drop. At last, the time dependence of pressure drop was established by using the modified model of specific deposit. The results show that the pressure drop of TCP-LG100 increases rapidly because of heavy specific deposit and large wall thickness, and those of TCP-LG80, and TCP-LG50 increase slowly, and then rapidly, and then slowly, and eventually tend to constant values.

The gas fluidization of three binary powder mixtures, viz. SiO₂–ZnO, SiO₂–Fe₂O₃, and ZnO–Fe₂O₃ were studied in a mechanically vibrating fluidized bed. The effects of the powder particle characteristics and their agglomeration (cohesiveness) properties on segregation in the fluidized bed were investigated. For SiO₂–ZnO, large agglomerated particles with high ZnO contents were observed, which segregated to the bottom of the fluidization column. The underlying mechanism was found to be the following: low-cohesion and low-density SiO₂ particles migrate to the upper layer through the channels because of gas flow. On the other hand, the large agglomerated particles with high ZnO contents are formed in the bottom layer because of the interparticle collisions due to vibrations, followed by sedimentation and segregation to the bottom. A similar phenomenon was observed for SiO₂–Fe₂O₃. The ZnO–Fe₂O₃, which exhibited similar agglomeration properties and densities, formed agglomerated particles that contained the components in uniform amounts and ratios and did not segregate to the bottom. These results highlight the possibility of separation of component on a vibrating fluidized bed based on the differences in their agglomeration properties.

In order to effectively recover lithium from underground brine, a novel solvent extraction system [tri-isobutyl phosphate (TIBP)+butyl acetate (BA)+FeCl₃+sulfonated kerosene (SK)] is studied experimentally. The effects of organic phase composition, molar ratio of Fe³⁺ and Li⁺ [n(Fe³⁺/Li⁺)], phase ratio [R(O/A)], temperature and pH in the stage of lithium extraction, as well as NaOH concentration and phase ratio in the stage of back-extraction were investigated. The single-stage extraction rate of lithium at room temperature was about 72% with the solvent extraction system (50% TIBP+10% BA+40% SK) under the optimum extractive condition of n(Fe³⁺/Li⁺)=2.0, pH=2–3, phase ratio R(O/A)=1.0, and the single-stage back-extraction rate was about 82% under the optimum back-extraction condition 1.3 mol·L⁻¹ NaOH at the phase ratio R(O/A)=2.0. These results indicate that lithium ion in underground brine can be efficiently recovered by employing a multistage continuous extraction method using centrifuge or mixer settler at these conditions.

A new method has been developed for determining the average porosity of filter cake overlying the membrane in dead-end ultrafiltration from only the flux decline behavior observed in two ultrafiltration tests without using any specially designed filtration cells. The validity of the method was examined for dead-end ultrafiltration of bovine serum albumin solutions. The method was based on a comparison of the ultrafiltrate flux data obtained from a concentrated solution with that obtained from a dilute solution. The average specific cake resistance was determined over a wide range of pressures from a single step-up pressure dead-end ultrafiltration test of dilute protein solution. The pressure dependence of the average cake porosity was evaluated by comparing these data with the infinite specific cake resistance data determined from a variable pressure ultrafiltration test of the concentrated solution. The porosity vs. pressure data were in reasonable agreement with those obtained from a series of constant pressure dead-end ultrafiltration tests performed at several pressures using a specially designed filtration cell capable of a sudden reduction in the effective filtration area. Furthermore, the flux decline behavior in constant pressure dead-end ultrafiltration experiments carried out at different pressures and solution concentrations was well evaluated based on the compressible cake filtration model, using the porosity data together with the specific cake resistance data. The simplified method presented here for easily estimating the porosity vs. pressure data from the flux decline behavior in dead-end ultrafiltration has important implications for analyzing the properties of filter cake formed in dead-end ultrafiltration.

In chemical looping combustion (CLC), CO₂ is captured inherently during fuel combustion because the oxygen for fuel combustion is supplied from oxygen carrier (OC) in a pure oxygen form. NiO/Al₂O₃ OC particles tend to agglomerate when NiO is reduced to Ni. Even though MgO addition alleviates the agglomeration phenomenon, excess MgO addition over some quantity reduces the oxygen transport reactivity before satisfactory prevention of the agglomeration. In this work, Mg–Al layered double hydroxide (LDH) was tested as a new starting raw support material to develop a spray-dried NiO OC containing sufficient amount of Mg, with achieving less performance loss. This new NiO OC with Mg content of 0.054 g/g, or 0.09 g/g as MgO, showed much faster oxygen transport rate compared with previous spray-dried NiO OCs made from MgO (0.084 g/g) and α-Al₂O₃ or γ-Al₂O₃. Furthermore, high mechanical strength could be obtained at a lower calcination temperature. The reason of performance improvement was explained with the mixed effect of the improved physicochemical properties. This work indicates that the use of Mg–Al LDH as a support for NiO OC can overcome several shortcomings found in the Mg added NiO/Al₂O₃ OC prepared from MgO and α-Al₂O₃ or γ-Al₂O₃ such as bigger grain growth caused by high calcination temperature, agglomeration easily occurring in OC with high mechanical strength, and low limit of the amount of MgO addition due to the reactivity decrease with increasing MgO content.

The present study investigates the chemical reaction kinetics for InN metal organic vapor phase epitaxy. Important surface reactants of the InN process were investigated similarly to those of the GaN process. The structures and vibrational frequencies of reactants and transition states were determined by density functional theory, and reaction pathways were evaluated by comparison with transient energies. It is suggested that some species play a key role in the surface reaction of InN MOVPE.

Methods to control the particle sizes of active pharmaceutical ingredients were investigated using sodium ecabet (Na-ECA) hydrate, which tends to generate large plate-like crystals, as a model compound. In batch-cooling crystallization in H₂O, the particle size decreased as the cooling rate was increased. Although the particle size reached 97 µm at a cooling rate of 40°C/h, this may be too large for formulation without further physical processing. A series of semi-batch crystallization experiments was conducted by adding aqueous Na-ECA solution to aqueous NaCl. In these experiments, the particle sizes decreased to less than 7 µm as the NaCl concentration was increased to >1.2% w/w. The particles yielded by semi-batch crystallization showed improved particle size distributions compared to those obtained by the batch-cooling crystallization method. The solubility of Na-ECA hydrate decreased to less than 10% of that in pure water as the NaCl concentration was increased to 3.0% w/w. This solubility change enabled the production of a highly supersaturated environment, allowing facile generation of microcrystals with a narrow particle size distribution.

Encapsulation of fullerene C60 in CD-MOF was achieved by crystallization from the C60/γ-CD complex in KOH solution. C60 was located in hydrophobic nanopores composed of a pair of γ-CD tori, whereas hydrophilic porphyrin (TCPP) was located in the spherical nanopores at the center of the (γ-CD)₆ unit. The isolated C60 and TCPP molecules in CD-MOF emitted green and orange fluorescence in the solid state, respectively. The fluorescence intensity of the C60/TCPP/CD-MOF was greatly weakened compared to that of TCPP/CD-MOF due to electron injection from the excited state of TCPP to C60.

Mesoporous titania was prepared employing a modified sol–gel method using titanium(IV) tetraisopropoxide (TTIP) in a deep eutectic solvent (DES) consisting of choline chloride and phenol. Monodisperse mesopores formed because of the low hydrolysis rate of TTIP in the highly viscous DES and the slow diffusion of water vapor into the DES. However, the removal of the DES from the titania gel using methanol washing caused the titania framework to shrink during calcination, resulting in washed titania with low pore volume. By contrast, the DES in the unwashed precursor contributed to preventing the shrinking of the framework. As a result, the pore volume and Brunauer–Emmett–Teller (BET) specific surface area of the unwashed titania were higher than those of the washed titania. The adsorption separation of gold nanoparticles (GNPs) using unwashed and washed titania calcined at 500°C was investigated. Washed titania lacked the ability to adsorb GNPs due to its low pore volume. High pore volume unwashed titania could capture GNPs with sizes of 10 nm or smaller, while the 15 nm GNPs were not adsorbed due to the sieve effect of the monodisperse mesopores.

The control of crystal shapes by melting are attempted, and melting phenomena of the organic compound crystals in a melt with the secondary components are discussed. The melting rates of an m-chloronitrobenzene (CNB) mother crystal in an m-/p-CNB melt were measured for various p-CNB compositions in a melt, and the shapes of a melted m-CNB mother crystal were observed. The melting rates of an m-CNB mother crystal for various directions were not zero and fluctuated with experimental duration. The ratios of the melting rate of an m-CNB mother crystal for the y direction relative to x direction were varied with the p-CNB compositions in an m-/p-CNB melt. Adsorption and desorption of p-CNB occur on the surfaces of the m-CNB mother crystal in the m-/p-CNB melt during melting, and the difficulty of adsorption and desorption of p-CNB changes according to the respective faces of the m-CNB mother crystal.