Search

Abstract: This study investigates the possibility of using a composite-supported polymeric liquid membrane (CSPLM) configuration for the Membrane Aromatic Recovery System (MARS). The membrane was prepared by impregnating the liquid membrane phase into the pores of a composite flat sheet membrane consisting of a microporous support and a thin non-porous layer. A polyvinylidenfluoride (PVDF) microporous support coated with a thin non-porous polydimethylsiloxane (PDMS) layer was chosen as a composite membrane due to its superior combination of mass transfer rates and chemical resistance among the membranes tested and polypropylene glycol (PPG) was used as a liquid membrane phase due to its high affinity for phenol. The resulting membrane showed uniform selectivity and operational stability under the continuous MARS operating conditions for more than two months. The mass transfer rates () were 5 times higher than those of the silicone rubber tubing () which is used in commercial scale MARS technology. The CSPLM also exhibited reduced water flux and low sodium ion transfer. The partition coefficient of phenol between PPG/water was measured as 84. A marked effect of ionic strength on partition coefficient of phenol was observed with partition coefficient increase to 134 when measured in the presence of 20wt% KCl in phenol aqueous solution. [Copyright &y& Elsevier]

Abstract: This paper describes an experimental investigation into mass transfer enhancement by chemical reaction in the membrane aromatic recovery system (MARS). The MARS process has recently been commercialised for recovery of cresols from aqueous process streams. In this paper, mass transfer enhancement as a function of the pH and organic concentration of the stripping solution is examined for constant bulk solution conditions. The application of two models of mass transfer enhancement (the Olander and Hatta models) is evaluated by determining the model parameters through independent experiments, limiting cases, and from literature. Model predictions are then compared to experimental data. Three organic compounds are used in this work: phenol, 4-chlorophenol (4CP), and triethylamine (TEA), which span a range of properties of typical molecules recovered by the MARS process. It is concluded that for compounds such as phenol, for which membrane mass transfer resistance dominates the system, there is no reaction enhancement. For compounds such as 4CP and TEA, enhancement effects are important. These were found to be best described through application of the Olander model, which allows for a second-order instantaneous reversible reaction, although the Hatta model, allowing for a second-order instantaneous irreversible reaction, was found to give a good description in some cases as well. [Copyright &y& Elsevier]

Abstract: This work investigates three different models describing mass transfer enhancement by a reversible and instantaneous second-order chemical reaction. The three models are applied to the study of mass transfer phenomena occurring in a membrane process for recovery of organic chemicals, the Membrane Aromatic Recovery System (MARS). Typical MARS operating conditions are used as model inputs, and the results obtained are used to assess the degree of complexity that should be taken into account in describing the mass transfer phenomena. The most complex model derived (N–P model) accounts for chemical reaction reversibility and the Nernst–Planck effect created by ionic species and is solved numerically. Following Olander model for a second order reversible instantaneous reaction model is proposed, for which we derive an analytical solution in terms of bulk solution properties. Finally, the simplest model follows the analysis of Hatta, assuming irreversible chemical reaction and neglecting the Nernst–Planck effect. The reversibility of the reaction is shown to be important, while N–P effects are negligible. The Olander model is recommended for use in describing the mass transfer phenomena. The models developed can be applied further to other processes of similar type. [Copyright &y& Elsevier]

This paper presents an experimental study on organic and water mass transfer through thin film composite (TFC) membranes, made of a 2 μm thick polydimethylsiloxane (PDMS) layer coated onto a microporous support of polyethersulfone. These were used for the extraction of various organic molecules from aqueous solutions with different ionic strengths. Independent measurements of organic and water transfer rates were performed using lab-scale single- and multi-tube mass exchangers.The rate of organic transfer across TFC membranes is shown to be independent of the affinity of the organic molecules for the PDMS layer, owing to negligible mass transfer resistance in this thin layer. As a consequence, the overall mass transfer coefficient measured for aqueous–aqueous extraction of hydrophilic compounds (i.e. with low affinity for PDMS) was up to one order of magnitude larger for TFC membranes than for 350 μm thick PDMS membranes operated under the same hydrodynamic conditions.The effect of osmotic pressure on water transport through TFC membranes was assessed by liquid–liquid extraction experiments using saline solutions. Water flux increases with increasing difference in salt concentration across the membrane, due to increased activity gradient. For the range of salt concentration and PDMS layer thickness used in this work, the osmotic water flux across the membrane has only a marginal effect on the rate of organic extraction. On the other hand, the ionic strength of feed and extractive phases has a direct influence on the organic transfer, due to the dependence of transport and equilibrium properties (i.e. diffusion and partition coefficients) on salt concentration. As a result, TFC membranes exhibit an asymmetric behaviour depending on operating mode. For the same hydrodynamic conditions and ionic strength, organic flux is higher when a saline solution is on the coated side than when it is on the porous side of the membrane. This has obvious implications in the design and operation of TFC membrane processes for the extraction of organics from saline solutions. [Copyright &y& Elsevier]

A new class of composite tubular membranes, steel braided membranes, has been tested in high flow velocity membrane systems for the extraction of organic pollutants from dilute aqueous solution. These membranes consist of a tubular membrane made of rubber material (e.g. polydimethylsiloxane or ethylene propylene diene monomer) contained in a netlike shell (braid), made of stainless steel threads, which supports the rubber tube.Overall mass transfer coefficients have been measured through steel braided and bare PDMS membrane tubes, using air, distilled water and polypropylene glycol (PPG) as extractive phases under a range of hydrodynamic conditions. The transport mechanism has been described by a resistance-in-series model approach. This has allowed the contributions of the braided side of the membrane tubes to the overall mass transfer resistance to be elucidated, and has shown that in aqueous-aqueous extraction operations the braid is the main resistance to mass transport across the membrane. Mass transfer coefficients measured through bare membrane tubes (1.4×10−5 to 2.5×10−5 m s−1) were 40% higher than those measured through braided membrane tubes (1.0×10−5 to 1.5×10−5 m s−1) under similar hydrodynamic conditions. This limitation has been overcome by using an extractive phase with high affinity for toluene (PPG). This enables high mass transport rates (overall mass transfer coefficients ranging from 4.3×10−5 to 5.2×10−5 m s−1) and allows an efficient use of braided membranes for extraction in modules with high fluid velocity. [Copyright &y& Elsevier]

Aqueous–aqueous extraction experiments were performed in a double tube mass exchanger (DTMEx) in order to investigate mass transfer characteristics of steel-braided non-porous membranes in a contained liquid membrane (CLM) module. Toluene was the permeating molecule, polydimethylsiloxane (PDMS) and ethylene propylene diene monomer (EPDM), were used as non-porous membrane materials, and silicone oil and polypropylene glycol (PPG) as liquid membranes. The liquid membrane contribution to the overall mass transfer resistance was assessed for different module configurations (e.g. varying membrane separation distance) and operating modes (e.g. stationary and mobile liquid membrane). The liquid membrane was found to be the dominant resistance when the module was operated with a stationary liquid membrane, a large separation distance between the braided membranes, and non-porous membranes with high affinity for the permeating molecules. When the liquid membrane was circulated through the module this resistance became negligible compared to the cumulative resistances in the aqueous hydrodynamic boundary layers (HBLs) and membrane at the feed and extractive side, resulting in overall mass transfer coefficients (OMTCs) up to 2.7×10−5 m s−1. CLMs with braided membranes and mobile liquid membrane may therefore be advantageously used for aqueous–aqueous separation of hydrophobic molecules in modules operating with high liquid velocity in the membrane tubes. [Copyright &y& Elsevier]

During the past few decades organic solvent nanofiltration has received a great deal of attention and a growing number of studies has been reported on development and optimisation of solvent resistant membranes and their transport mechanism. However, most of these studies have used flat sheet membranes. On the other hand, many researchers studied fluid dynamics and mass transfer in spiral-wound membrane modules, almost exclusively in aqueous solutions. This paper reports the performance of four spiral-wound membrane modules tested in 0–20 wt% solutions of sucrose octaacetate in ethyl acetate under various pressures and retentate flowrates. These modules were made of two different types of membranes (a commercial membrane, PuraMem® S600, and a development product, Lab-1, from Evonik Membrane Extraction Technology Limited) and covered three module sizes (1.8″×12″, 2.5″×40″ and 4.0″×40″). All modules had the same feed and permeate spacers. The classical solution diffusion model was applied to describe the transport of solute and solvent through the membrane and regress the unknown model parameters from flat sheet data. Correlations for characterising the fluid dynamics and mass transfer in the spiral-wound membrane modules, as well as the parameters describing the feed and permeate channels, were determined by performing the regression of experimental data of a 1.8″×12″ PuraMem® S600 membrane module. The classical solution–diffusion model, combined with the film theory, was then successfully applied to predict the performance of other modules of larger size (such as the 2.5″×40″ and 4.0″×40″ module sizes) and/or made of a different membrane material (such as Lab-1). The procedure proposed in this paper predicts the performance of a specific module by obtaining a limited number of experimental data for flat sheets and a 1.8″×12″ spiral-wound membrane module only (necessary to obtain the fitting parameters characteristic of the membrane and the module). Furthermore, with this procedure, it is not necessary to know a priori the spacer geometry, because the necessary information about the spacer geometry will be also obtained by regression of few experimental data. [ABSTRACT FROM AUTHOR]

Abstract: Due to their excellent resistance to a range of solvents, integrally skinned polyimide membranes have been used to achieve selective separations in a range of solvent-based industrial and lab-scale chemical operations. These include: homogeneous catalyst recycle, petrochemical dewaxing, solvent exchange and chiral resolutions. However, despite the widening scope of use of these membranes, there is still little understanding of how the microstructure defines their separation performance. As a first step towards determining this, integrally skinned nanofiltration membranes were fabricated by phase inversion using Lenzing P84 polyimide. The microstructures of these membranes, dry and wetted in solvent, were investigated by SEM, TEM and ESEM (where appropriate). SEM and TEM imaging of dry membranes revealed that this type of polyimide membrane has three microstructurally distinct polyimide layers, not the two indicated in prior literature. Furthermore, TEM images reveal nano-sized pore-like features in the polyimide structure, which indicate that the transport mechanism is probably neither only solution-diffusion nor only pore flow. ESEM imaging showed that when saturated in ethanol at the working pressure of the ESEM (5.50Torr), the microstructure of the membranes changes; it is wispy and thus quite different to the more solid polymer nodules and interconnected polymer network observed in the dry membranes. Thus, transport and separation mechanisms based on the structure of the dry membranes may not be accurate. Overall, these results indicate that the current theory used to describe polyimide membrane mass transfer and separation performance (where it is based on dry membrane microstructures), most likely needs to be rethought. [Copyright &y& Elsevier]

Abstract: The membrane aromatic recovery system (MARS) is a new membrane technology which recovers aromatic acids and bases. The first industrial installation has been operating at a Degussa site in the UK recovering cresols since 2002. The state of the art MARS technology employs a tubular silicone rubber membrane. However, this places some limitations on the process due to relatively low mass transfer rates and limited chemical resistance. In this paper, flat sheet composite membranes were investigated for application to the MARS process. In particular for recovery of compounds, such as 1,2-benzisothiazolin-3-one (BIT) which show low mass transfer rates through the current membrane. These composite membranes are comprised of a thin nonporous PDMS selective layer coated on a microporous support layer cast from polyacrylonitrile, polyvinylidene fluoride, polyetherimide or polyphenylenesulphone. The membranes have been characterised using SEM and gas permeation. The mass transfer of BIT through the composite membranes with no chemical reaction enhancement was an order of magnitude higher than through tubular silicone rubber membranes (10−7 ms−1 versus 10−8 ms−1). With chemical reaction enhancement, the mass transfer increased by another order of magnitude to 1.6×10−6 ms−1 for BIT through a PVDF supported composite membrane. Mass transfer through the composite membrane was described well using analysis based on the resistance in series theory with chemical reaction. However, when a high osmotic pressure was applied across the membrane (molarity∼3M), significant water transport occurred across the membrane. [Copyright &y& Elsevier]