Selective permeation and separation of steam from water–methanol–hydrogen gas mixtures through mordenite membrane

Separation properties of a mordenite membrane for water–methanol–hydrogen mixtures were studied in the temperature range from 423 to 523 K under pressurized conditions. The mordenite membrane was prepared on the outer surface of a porous alumina tubular support by a secondary-growth method. It was found that water was selectively permeated through the membrane. The separation factor of water/hydrogen and water/methanol were 49–156 and 73–101, respectively. Even when only hydrogen was fed at 0.5 MPa, its permeance was as low as 10⁻⁹ mol m⁻² s⁻¹ Pa⁻¹ up to 493 K, possibly suggesting that water pre-adsorbed in the micropores of mordenite hindered the permeation of hydrogen. The hydrogen permeance dramatically increased to 6.5 × 10⁻⁷ mol m⁻² s⁻¹ Pa⁻¹ at 503 K and reached to 1.4 × 10⁻⁶ mol m⁻² s⁻¹ Pa⁻¹ at 523 K because of the formation of cracks in the membrane. However, the membrane was thermally stabilized in the presence of steam and/or methanol.     

Inlet characteristics of bioaerosol samplers

The inlet sampling characteristics of several commercial bioaerosol samplers operating in indoor and outdoor environments have been analyzed by use of available and newly developed equations for sampling efficiency. With a focus on the physical aspects of sampling efficiency, the aspiration and transmission efficiencies have been calculated for the bioaerosol particle size range 1–30 μm, which represents single bacteria, bacteria aggregates, bacteria carrying particles, fungal spores, yeast, and pollen. Under certain sampling conditions, the bioaerosol concentration was found to be significantly over- or underestimated. At wind velocities between 0 and 500 cm s⁻¹, calculations show that the AGI-30 would sample 1–10 μm particles with an inlet sampling efficiency of 20–100%. The entrance efficiency of the 6-stage Andersen viable sampler is 90–150% when sampling isoaxially with respect to horizontal aerosol flows, and 8–100% when oriented vertically at a right angle to the horizontal aerosol flow. For the Burkard portable air sampler, an even wider range of deviation may occur. The bioaerosol samplers used for large particles such as pollen are even less accurate: e.g. 10 times the ambient concentration of Lycopodium spores has been calculated to be aspirated by the Lanzoni sampler when operated at 0.5 1 min⁻¹ facing the wind at wind velocity of about 500 cm s⁻¹.

The actual bioaerosol concentration can be calculated from the measured data by use of the indicated procedures. The sampling efficiency graphs presented can be used to bracket the sampling conditions that enable the investigator to avoid or minimize significant sampling biases for each sampler. The findings can also be used for the design of new samplers or for improving commercially available samplers.     

Variation of the aerosol charge neutralization coefficient in the entire particle size range

The rate at which the mean charge on aerosol particles relaxes to its steady-state value under bipolar charging is characterized by the neutralization rate constant, β (s⁻¹). It is an important parameter for fixing the nt product in charge neutralizers as well as in the theory of charging-induced diffusion. Here we compute the neutralization coefficient, β/n (where n is the mean ion density), as a function of particle size through the use of ion-particle combination coefficients provided by the recent theories. The results indicate that β/n decreases from a continuum limit value of 3.1 × 10⁻⁶ cm³ s⁻¹, to a free molecular limit value of 1.4 × 10⁻⁶ cm³ s⁻¹. The changeover occurs rapidly in the transitional regime (10–100 nm). This clearly indicates that the nt product required to attain steady state is higher for nano particles than for larger ones. The paper also presents the variations of the mean square variance of charge, the coefficients of charging-induced drift and diffusion, as a function of particle size.     

Hydrodynamic studies of an advanced fluidized-bed bioreactor for direct interaction with coal

Fine coal particles fluidized by the upflow of a liquid medium containing a dissolved biocatalyst undergo size reduction as the reaction progresses. Three aspects of the design of such a reactor were examined:

1. (1)the use of force balances to describe pressure drop for the segregated bed;

2. (2) measurement of liquid-phase dispersion coefficients; and

3. (3) fluorescent tagging of particles to track size distribution.

Hydrodynamic data were obtained for a liquid-solid fluidized bed of coal particles in the size range 30–150 μm. Illinois No. 6 coal was ball-milled, sieved into four fractions, and suspended in a 0.1% aqueous solution of Tween 80. A sample with a bimodal particle size distribution centred on 49 and 63 μm was placed in a glass column and fluidization and pressure-drop data were compared with a new model developed to describe particle segregation. Measured liquid-phase dispersion coefficients varied from 0.034 to 0.283 cm² s⁻¹ as the flow varied from 0.005 to 0.0159 cm s⁻¹. A technique was also developed for coating coal particles with a fluorescent paint which may allow direct measurement of the change in the fraction of marked particles of known size along the axis of a fluidized bed.     

On mass transfer effect due to electrolytically evolved gas

Based on the Ibl penetration model mass transfer equations for gas-evolving electrodes were derived and compared to the effect of forced convection. Experimental studies were conducted in a rectangular flow channel with the working electrode facing downward. The variables were linear bubble velocity, linear electrolyte velocity, nature of the gas and electrode position. Up to bubble velocities (U_x) of 2 and 6 cm s⁻¹ for O₂ and H₂ gases respectively, the thickness of the Nernst diffusion layer (δ_av) was described well by the equation δ_av = [Dd_eL/(U_x)_av]^1/3. Intermediate slopes between − 1/3 and − 1 were observed for O₂ bubble velocities between 2 and 6 cm s⁻¹. A theoretical derivation suggests that in the absence of bubble coalescence, the mass transfer effect due to laminar flow induced by electrolytically evolved gas exceeds that due to forced external laminar flow for all practical channel designs.     

Effect of physical parameters on the microfiltration of wine on a flat polymeric membrane

This paper investigates the microfiltration of unfiltered red and white wines through organic polyvinylidene difluoride membranes in plate and frame systems. Various pore sizes between 0.1 and 3 μm were first tested using a small laboratory unit and it was found that 0.4 μm pores gave the best comprise between turbidity and flux requirements. Tests with membrane areas between 1 and 2 m² were carried out i a modular plate and frame system in which both gasket thickness and parallel vs. series arangements of compartments can be modified. With red wine the permeate flux is almost independent of fluid velocity but increases linearly with transmembrane pressure, reaching 50 1 h⁻¹ m⁻² at 3 bar. The permeate flux is larger in a turbulent regime than in a laminar one at the same velocity. With white wine, the permeate flux is higher and increases almost linearly with velocity reaching 170 1h⁻¹ m⁻² at 2 bar and 3.6 m s⁻¹. The turbidity of the permeate was generally below 0.5 NTU for both white and red wines while initial turbidity was above 150 NTU for red and above 50 NTU for white. When several membrane compartments are placed in series, the output of the downstream compartment is reduced by 10%–15% compared with that of the upstream one owing to the drop in transmembrane pressure.     

Effect of ammonium polyacrylate on rheology of anatase TiO2 nanoparticles dispersed in silicon alkoxide sols

Ammonium polyacrylate (NH₄PA) was introduced into powdered mixtures consisting of anatase-structured TiO₂ nanoparticles and silicon alkoxide precursors at the sol level, and the rheological behavior of the mixtures was examined under various solid loadings (φ=0.05–0.13 in volumetric ratios), shear rates (  s⁻¹) and NH₄PA concentrations. The alkoxide precursors were mixtures of tetraethyl orthosilicate (TEOS, Si(OC₂H₅)₄), ethyl alcohol (C₂H₅OH), H₂O and HCl in a constant [H₂O]/[TEOS] ratio of 11. The nanoparticle–sol mixtures generally exhibited a pseudoplastic flow behavior over the shear-rate regime examined. The NH₄PA appeared to serve as an effective surfactant which facilitates the suspension flow by reducing the flow resistance at low NH₄PA concentrations. At φ=0.10, a viscosity reduction ca. 85% was found at  s⁻¹ when the NH₄PA concentration was held at 2.5 wt.% of the solids. As the NH₄PA exceeded a critical level, e.g., [NH₄PA]≥3.0 wt.%, the NH₄PA acted as a catalyst which quickly turned the TiO₂–silica sol mixtures (φ=0.10) into a gelled structure, resulted in a pronounced increase of mixture viscosity. The maximum solids concentration (φ_m) of the mixtures was experimentally determined from a derivative of relative viscosity, i.e., (1−η_r^−1/2)–φ dependence. The estimated φ_m increased from 0.127 to 0.165 when NH₄PA of 0.5 wt.% was introduced into the TiO₂–silica sol mixtures.     

Separation and enrichment of carbon dioxide by capillary membrane module with permeation of carrier solution

A novel facilitated transport membrane for gas separation using a capillary membrane module is proposed in which a carrier solution is forced to permeate the membrane. Both a feed gas and a carrier solution are supplied to the lumen side (high pressure side, feed side) of the capillary ultrafiltration membrane and flow upward. Most of the carrier solution which contains dissolved solute gas, CO₂ in the present case, permeates the membrane to the permeate side (low pressure side, shell side), where the solution liberates dissolved gas to form a lean solution. The lean solution is circulated to the lumen side. This type of capillary membrane module was applied to the separation of CO₂ from model flue gases consisting of CO₂ and N₂. Monoethanolamine (MEA), diethanolamine (DEA) and 2-amino-2-methyl-1-propanol (AMP) were used as carriers or absorbents of CO₂. The feed side pressure was atmospheric and the permeate side was evacuated at about 10 kPa. CO₂ in the feed gas was successfully concentrated from 5–15% to more than 98%. The CO₂ permeance was as high as 2.7×10⁻⁴ mol m⁻² s⁻¹ kPa⁻¹ (8.0×10⁻⁴ cm³ cm⁻² s⁻¹ cmHg⁻¹) when the CO₂ mole fraction in the feed was 0.1 and temperature was 333 K. The selectivity of CO₂ over N₂ was in the range from 430 to 1790. The membrane was very stable over a discontinuous one-month testing period.     

The Porto Torres biodepyritization pilot plant: Light and shade of one year operation

The nominal 50 kg h⁻¹ dry coal biodepyritization pilot plant was built at Porto Torres (Sassari, Italy) in the area of EniChem S.p.A. Chemical Works, with financial support of the Commission of European Communities and with the participation of the DMT of Essen (Germany), the Mining and Mineral Processing Department of the University of Cagliari (Italy), the Technical University of Delft (The Netherlands) and the Warren Spring Laboratory of Stevenage (UK). The plant went on stream in early September 1992. Test runs were carried in the range 6.5–41.5% solids concentration. For every solids concentration the steady state was attained in about 10 d. For all the test runs the assays showed that more than 90% pyrite removal is achieved in the first five bioreactors; for a pulp flow rate of 6.94 × 10⁻² dm³ s⁻¹, (250 l h⁻¹) and useful bioreactor volume of 7.5 m³, this corresponds to a residence time of 540 345 s, i.e., 6.254 d, and to a pyritic iron biosolubilization rate of 36 mg dm⁻³ h⁻¹. The pyrite biosolubilization rate constant is 1.53 × 10⁻² h⁻¹. The power input per bioreactor operating on a 40% solids pulp is 4 kW with cos Φ = 0.76. Hence, the power required for processing 100 kg h⁻¹ coal at 40% solids in the bioreactor section is 4 × 5 = 20 kW h, i.e. 200 kW h per tonne dry coal.     

Controlled wet erosive wear of polycrystalline alumina

Controlled wet erosive wear tests were performed on polycrystalline alumina specimens of mean grain size G = 1.2, 3.8 and 14.1 μm and on sapphire specimens. The tests were performed by using an apparatus consisting of a jet body that rotates immersed in the slurry medium (SiC grits). Impacts are normal to the target surface. The construction and calibration of the apparatus are described. The impacting velocity used was 2.7 m s⁻¹. The weight loss of polycrystalline alumina and sapphire specimens increased linearly with impacting time. The wear rate of polycrystalline alumina specimens increased with grain size. Wear rates of 1.83, 8.36 and 11.3 nm s⁻¹ correspond to G = 1.2, 3.8 and 14.1 μm, respectively. For sapphire specimens the wear rate was 9.56 nm s⁻¹. Worn surfaces of both polycrystalline alumina and sapphire specimens were analysed by scanning electron microscopy.     

Particle characteristics of a stable fluidized bed aerosol generator

An aerosol generator consisting of a vibrating system for feeding dust into a fluidized bed was developed and tested to determine its dust output characteristics. The dust feed unit can produce 0–40 g min⁻¹ of coal dust and shows constant output up to 3 h operation durations. These correspond to mass concentrations of 0–101 g m⁻³ of coal particles for an air flowrate of 395 l min⁻¹ through the aerosol generator. The aerosolized coal particles show constant particle size distribution with time for up to h of testing under varied operation parameters. The normalized particle size distribution remains almost identical for a given feed material for a range of dust loadings. The time required to reach steady state aerosol generation is negligible for the sizes of coal particles used in this investigation.     

Electrochemical hydrodynamics in a magnetic field with laser interferometry

An unusual magnetic field effect on the concentration profiles during electrochemical deposition of copper in an unstirred electrolyte solution has been observed during a galvanostatic electrolysis which was followed using laser interferometry. The study was done on the Cu/0.1 N CuSO₄/Cu system at its natural pH and at different c.d.'s with plane parallel electrodes in the vertical position (V) and in the horizontal position of cathode over anode (C/A) and anode over cathode (A/C). The field produces a marked distortion of the interference fringes which indicate concentration gradients. The gradients are proportionately higher at higher c.d's in the range of c.d's of 0.12 mA cm⁻² to 12.0 mA cm⁻² with vertical (V-position) electrodes. In the C/A position the occurrence of a smooth increase in fringe bend with time suggests no hydrodynamic flow. Flow velocities observed using suspended particles range from 2 cm s⁻¹ to 13 cms⁻¹ at a magnetic field strength of 0.10–0.60 T. The shape of the interference fringes in an interferogram varies with electrode separation distance in the vertical position. The paramagnetism of the cupric ion which results in a magnetic fluid is also used in the explanation of the effects.     

Chlorination kinetics of pyrite mineral in two Turkish lignites

The kinetics of the chlorination of pyrite in two Turkish lignites in water and water-carbon tetrachloride media at ambient pressure ( 610 mm Hg) are investigated. The effects of speed of stirring (5–20 s⁻¹), particle size (74–88, 150–180 and 250–425 μm), temperature (13–70°C) and reaction time (0–18 000 s) were studied. The experimental data were analyzed on the basis of the unreacted shrinking core model. The fine pyrite particles are assumed to be embedded inside the coal particles. The rate-controlling step was found to be diffusion of chlorine through the ash (the coal matrix). The activation energies were calculated as 25.1 kJ mol⁻¹ for Dada i coal in water medium and 25.0 kJ mol⁻¹ for Mengen coal in water-carbon tetrachloride medium.     

Mechanism of gas flow through coal

Coals are fossilized plant material plus inorganic silt deposited in irregular layers and containing 1–20% void space which provides a medium which is porous to gas flow. Gas flows have been measured using discs of coal cut from several coal seams. Observed flow phenomena include molecular diffusion (low gas pressure, small pores) and bulk diffusion (higher gas pressure, larger pores). In the examples investigated, methane flows ranged from 1.2 × 10⁻¹⁰ cm² s⁻¹ atm⁻¹ for Pittsburgh-seam attrital coal to 2.0 cm² s⁻¹ atm⁻¹ for Oklahoma Hartshorne coal. Flow characteristics have been compared with gas flows observed by Knudsen through glass capillaries. This information can be applied to mine safety and to coal utilization studies.     

Oxidative coupling of methane in a mixed-conducting perovskite membrane reactor

Ionic-electronic mixed-conducting perovskite-type oxide La_0.6Sr_0.4Co_0.8Fe_0.2O₃ was applied as a dense membrane for oxygen supply in a reactor for methane coupling. The oxygen permeation properties were studied in the p_O2-range of 10⁻³−1 bar at 1073–1273 K, using helium as a sweeping gas at the permeate side of the membrane. The oxygen semi-permeability has a value close to 1 mmol m⁻² s⁻¹ at 1173 K with a corresponding activation energy of 130–140 kJ/mol. The oxygen flux is limited by a surface process at the permeate side of the membrane. It was found that the oxygen flux is only slightly enhanced if methane is admixed with helium. Methane is converted to ethane and ethene with selectivities up to 70%, albeit that conversions are low, typically 1–3% at 1073–1173 K. When oxygen was admixed with methane rather than supplied through the membrane, selectivities obtained were found to be in the range 30–35%. Segregation of strontium was found at both sides of the membrane, being seriously affected by the presence of an oxygen pressure gradient across it. The importance of a surface limited oxygen flux for application of perovskite membranes for methane coupling is emphasized.     

Dissolved oxygen concentration in an undivided rotating cylinder electrode reactor

The steady state mass balance for dissolved oxygen in an undivided rotating cylinder electrode reactor was established. Cases involving the absence or presence of diffusion controlled oxygen reduction reaction were treated. The results show that in the absence of the oxygen reduction reaction and under low electrolyte velocities employed industrially, the rotating cylinder electrode reactor, exhibiting a continuous stirred tank reactor behaviour, yields lower dissolved gas concentrations than a plug flow reactor. Therefore, the rotating cylinder electrode reactor leads to lower redissolution rates of the detached metallic particles by oxygen corrosion. The effect of the dissolved gas concentration in the inlet electrolyte also demonstrates the benefits of a prior degassing step in the flow sheet. The simpler equation derived in the absence of oxygen reduction may be used in its presence for inlet dissolved oxygen concentrations below saturation and industrially relevant electrolyte velocities ranging from approx. 0 to 0.3 m s⁻¹ if metal deposition current efficiency is higher than 87% or, conversely, if the dissolved metal concentration to dissolved oxygen concentration ratio is greater than 27. An even simpler equation based on the added hypothesis of total desorption of the gas produced can also be used with electrolyte velocities below 0.025 m s⁻¹. Finally, the derived equations allow optimization of the rotating cylinder electrode reactor.     

Helping to choose operating parameters for a coating fluid bed process

The feasibility of solid particles coating in a fluid bed with a Wurster tube is studied for several types of particles and aqueous coating solutions. The model products are wheat semolina, beads of glass, alumina, resin polystyrene, plastic PMMA, with a size range between 125 and 1250 μm and densities between 500 and 2500 kg m⁻³. The chosen coatings are representative of those used for the food products, such as maltodextrin, acacia gum, and sodium chloride in aqueous solution.

The air flow rate suitable for a regular circulation of particles in the reactor is determined for each particle type. For each coating solution, the flow rate leading to agglomeration is considered as the maximal limit flow rate to use for coating. Then comparative coating experiments were realized.

For a similar initial load of particles, the same mass of coating was atomized (13.5 g min⁻¹) at 50 °C. The mass of coating deposit on particle surface is increased linearly during an atomization sequence lasting 33 min. For example, for every 100 g of alumina particles, the rates are 0.48, 0.51, and 0.53 g min⁻¹ for sodium chloride, maltodextrin, and acacia gum, respectively. We then obtain a coating efficiency between 87% and 98%.

In the specific case of sodium chloride on glass beads, the deposit of crystallized salt was linear during 10 min then stopped. Addition of acacia gum (50%) to the NaCl coating solution leads again to a linear deposit over 65 min.     

Group VI transition metal carbides as alternatives in the hydrodechlorination of chlorofluorocarbons

The synthesis of transition metal carbides of tungsten and molybdenum has been carried out via temperature programmed reactions (TPRs) of metal oxides or passivated nitrides. Their specific chloropentafluoroethane conversion rates were at best one order of magnitude less than that of a reference Pd based catalyst. The intrinsic rates range from 4.7 to 14.7 nmol m⁻² s⁻¹ and decrease as follows: Mo₂C>WC>W₂C≈WC_1−x>MoC_1−x. The group VI carbide samples catalyse hydrodehalogenation and dehydrofluorination. WC appears to be as selective towards pentafluoroethane (HFC-125) as the Pd based catalyst. Then the selectivity decreases in the following sequence: W₂C>Mo₂C>WC_1−x>MoC_1−x. All the carbide catalysts deactivate at the early stages of the reaction. Based on the XPS results and the product distribution of the reaction, the deactivation has been mainly attributed to a site blocking phenomenon due to a strong deposit of polymeric carbon and of hydrofluorocarbon polymers. Polymerisation of detected unsaturated compounds take place on acidic sites probably generated by fluoride and/or chloride in the course of the reaction.