B. Van der Bruggen

Influence of molecular size, polarity and charge on the retention of organic molecules by nanofiltration

Because of the growing interest in nanofiltration for industrial use, a better insight in the retention mechanisms in nanofiltration is needed, which will make it possible to understand membrane performances for specific applications. In this paper, the retention of a series of organic molecules by four nanofiltration membranes was studied. The membranes that were used are NF70 (Dow/FilmTec), NTR 7450 (Nitto-Denko), UTC-20 (Toray Ind.) and Zirfon® (VITO). In order to correlate the retention with the size of the molecule, which is the main factor that determines the retention, use was made of different parameters for the molecular size: molecular weight, the Stokes diameter, the equivalent molar diameter, and a diameter obtained with energy minimisation calculations. For each size parameter, the correlation with retention in nanofiltration experiments was calculated. For the Zirfon® membrane, retentions were too low to obtain a good correlation. For the three other membranes, a good correlation with retention was found for each of the size parameters. Two other factors were found to have an influence on retention of organic molecules: the polarity of the molecule, and the charge of the molecule. The importance of these factors depends on the molecules as well as on the type of membrane.

Nanofiltration as a treatment method for the removal of pesticides from ground waters

Abstract A major problem in the drinking water production is the presence of pesticides in raw water, not only in surface waters, but also in ground waters. Concentrations of several μg/l have been found. Traditionally, pesticides are removed using granular or powdered activated carbon, which is an expensive process. The cost is enhanced by competition of pesticides and natural organic matter (NOM) for adsorption sites on the activated carbon. Nanofiltration (NF) can be a valuable alternative for the removal of pesticides. Moreover, pesticides and hardness can be removed in one step. In this paper, the retention of four pesticides (atrazine, simazine, diuron, isoproturon) was determined experimentally for four NF membranes: NF-70 (Dow/Film Tec), NF-45 (Dow/Film Tec), UTC-20 (Toray Inc.) and NTR-7450 (Nitto-Denko). The experimental retentions indicate that NF-70 is a suitable membrane for removal of pesticides. The experimental retentions were related to structural properties of the molecules. The size of the molecule is the most important parameter, and may be represented by molecular weight, or more accurately by a molecular diameter calculated from the molecular structure. Furthermore, a comparison with retentions of a series of saccharides confirms the earlier found effect of the dipole moment: a higher dipole moment results in a lower retention. No effect of the pesticide concentration was found. The matrix of the ground waters caused an increase of the pesticide retention, together with a decrease of the water flux through the membrane.

Modelling of the retention of uncharged molecules with nanofiltration.

In this paper, a model is developed for the retention of organic molecules with a given nanofiltration membrane at different pressures as a function of the molecular weight. The Spiegler-Kedem transport equations were used to derive the reflection coefficient, the maximal retention that would theoretically be obtained at infinite transmembrane pressure, from experimental retention values for a large set of molecules with the effective diameter of the molecule as a size parameter. Secondly, the pore size distribution of the membrane is derived from the experimental reflection coefficients. This allows to calculate the reflection coefficient for a molecule with a given effective diameter. Since this parameter is not readily available, a correlation between the effective diameter and the molecular weight has been established and introduced in the model equations. Subsequently, the contribution of diffusion in the transport of molecules through the membrane was evaluated by introducing a membrane diffusion parameter, which was determined experimentally for the membranes NF70, NTR 7450 and UTC-20. Finally, the pore size distribution, the diffusion parameter and the experimental water flux through the membrane were used to calculate the retention as a function of the molecular weight and pressure for the same three membranes. This allows to determine retention curves at different pressures, and to calculate the variation of the MWC with pressure.

Mechanisms of retention and flux decline for the nanofiltration of dye baths from the textile industry

Removal of dye compounds from colour baths used in the textile industry is a possible application of nanofiltration. However, the mechanisms involved in this process are not clearly understood and the practical application of the process is facing many problems such as fouling and flux decline. The mechanisms of retention and flux decline were examined using two different approaches. Firstly, synthetic dye baths were prepared according to manufacterers recipes. Retention of two reactive dyes (reactive blue 2 and reactive orange 16) was studied in separate baths and with different concentrations of Na2SO4, Na2CO3, NaOH and a surfactant. Different nanofiltration membranes (UTC-60, NF70 and NTR 7450) were used. The water flux in each of the experiments was monitored. It was found that the retention of ions decreased with the ion concentration due to a decrease of the Donnan potential. The retention of the dyes was high and was not influenced by the dye concentration or the ionic strength. The water flux was dependent of the ion concentration in the feed solution: high ion concentrations caused a dramatic decrease of the water flux. The dye concentration in the bath was found to have only a minor influence, whereas surfactants did not change water flux or dye retention. A theoretical explanation for these effects is given. The phenomenon of flux decline limits strongly the applicability of nanofiltration for direct treatment of dye baths. In a second step, an industrial wastewater from a textile factory was treated biologically in an active sludge system. The effluent was used as feed for nanofiltration with the same membranes as in the first step. The overall results for these experiments were satisfactory. The ion concentration was much lower than in the earlier experiments due to mixing of different feed streams. Therefore, the water fluxes were not considerably lower compared with the clean water fluxes. The retentions were sufficiently high to make recirculation of the treated water possible, thereby providing a considerable saving of water.

Application of computer modelling to predict the leaching behaviour of heavy metals from MSWI fly ash and comparison with a sequential extraction method

Abstract Combustion residues in general and MSWI fly ashes in particular form a major environmental problem as they are polluted with heavy metals. The heavy metals can be removed from the fly ash by leaching with e.g. the acid wastewater obtained in the wet scrubber of the air pollution control system. The remaining fly ash can be landfilled or valorized e.g., a construction material. An understanding of the leaching reactions and of the factors that influence leaching is very important for the treatment of fly ash. Therefore, the composition of the fly ash was determined and the influence of pH was examined. In this paper, the results of experimental leaching tests of fly ash are compared with computer calculations of the thermodynamic equilibrium of the leaching solution–fly ash system. The computer program MINTEQA2, used for this purpose, allows to predict the metal concentrations in the leaching solution, the minerals that precipitate, and the pH of the leaching solution at equilibrium. Also a sequential extraction procedure was performed on the fly ash. The procedure can be used to investigate which leaching conditions are necessary to obtain a good extraction of the heavy metals. Comparison of the experimental and calculated leaching data with the results from the sequential extraction procedure allowed verifying the accuracy of the sequential extraction procedure.

Flux decline in nanofiltration due to adsorption of organic compounds

Abstract The water flux for two nanofiltration membranes (UTC-20 and NF70) was measured using aqueous solutions of 11 organic compounds in different concentrations. The flux declined, compared to the pure water flux, by more than 50% for solutions containing less than 1 g/l of some organic compounds. Flux decline as a function of the concentration of the organic compound showed the typical shape of a Freundlich isotherm. Comparison to adsorption experiments indicated that flux decline, for the solutions used, was related to adsorption on the membrane material. The dipole moment, the octanol–water partition coefficient and the water solubility were examined as parameters to explain adsorption on nanofiltration membranes. A clear correlation was found between the octanol–water partition coefficient and adsorption; adsorption also appeared to be related to the dipole moment and the water solubility. This shows that both the surface charge and hydrophobicity of the membrane can play a role in the adsorption. When flux decline is related to octanol–water partition coefficient, the molecular size (molecular weight) should be taken into account as well. A subset of molecules with a molecular weight in a narrow range, somewhat below the cut off value of the membranes, was selected. Within this subset, a clear correlation between the octanol–water partition coefficient and flux decline was observed.

Fluxes and rejections for nanofiltration with solvent stable polymeric membranes in water, ethanol and n-hexane

Abstract The solvent flux and rejection of 2,2′-methylenebis-(6-tert-butyl-4-methyl phenol) was studied in water, ethanol and n -hexane for four solvent-stable polymeric nanofiltration membranes (N30F, NF-PES-10, MPF-44 and MPF-50). Solvent fluxes were analysed with the Machado model to determine hydrophobicity/hydrophilicity of the membranes. The results are consistent with N30F, NF-PES-10 and MPF-44 being hydrophilic membranes, and MPF-50 being a hydrophobic membrane. The solvent flux for MPF 50 increases with decreasing polarity of the solvent: for n -hexane, a flux as high as 1600 l / m 2 h was obtained at a 20 bar transmembrane pressure. Much lower fluxes were found with ethanol; water fluxes were almost zero. For the other three membranes, the flux decreases in the order water>ethanol> n -hexane, i.e. with decreasing polarity. Pretreatment of the membranes by immersion for 24 hours in the solvent used for the flux measurements reduced the ethanol flux and n -hexane flux for MPF-50, but increased the solvent fluxes for the hydrophilic membranes. According to the Machado model, a possible explanation is that changes in the membrane structure cause a decrease of the hydrophobicity or hydrophilicity of the membranes. Rejections of 2,2′-methylenebis-(6-tert-butyl-4-methyl phenol) (insoluble in water) in ethanol and n -hexane were always below 10%. Rejection of maltose in water was much higher for all membranes, although maltose is a component with a comparable molecular weight (insoluble in ethanol and n -hexane). The enhanced mobility of polymeric chains in organic solvents might cause these low rejections by increasing the effective pore size available for transport through the membrane. Moreover, in aqueous solution a component such as maltose may have a water shell increasing the effective size of the molecule; this effect is absent in organic solvents. The difference in rejection was also obtained with components with higher molecular weight: raffinose in water and DL -alpha-tocopherol hydrogen succinate in ethanol and n -hexane.

Influence of fermentation by-products on the purification of ethanol from water using pervaporation

Pervaporation is claimed to be a promising separation technique for the purification of ethanol from fermentation broths during bio-ethanol production. In this study, influence of fermentation by-products on the purification of ethanol from water during hydrophobic pervaporation was investigated. Sugars and salts were found to increase the membrane performance. Reason for this was a change in vapor/liquid equilibrium. 2,3-butanediol decreased the ethanol flux and selectivity factor, while glycerol exhibited no effect. This was explained by a strong sorption of butanediol into PDMS and no sorption of glycerol. Due to the presence of carboxylic acids, hydrophobicity degree of the Pervap 4060 membrane decreased, which resulted in an irreversible increase in water flux and decrease in separation performance. These observations suggested the presence of silicalite-based fillers in the membrane. When the pH was raised to a value above the dissociation constant, no changes in hydrophobicity degree and membrane performance were found.

Performance of PDMS membranes in pervaporation: effect of silicalite fillers and comparison with SBS membranes.

Laboratory-made silicalite filled PDMS membranes were tested by means of concentration and temperature influence on the membrane performance in removal of ethanol from ethanol/water mixtures. This allowed studying the applicability of solution-diffusion model in the transport mechanism description. Experiments were performed by varying the ethanol concentration in the feed and temperature. Two types of fillers were incorporated into the PDMS network: commercial zeolite silicalite (CBV 3002) and laboratory-made colloidal silicalite-1. Obtained results were then compared with data gathered for unfilled PDMS membranes to examine the effect of fillers incorporation. Moreover, the comparison with novel block co-polymer based porous and dense SBS membranes was done. It was found that the solution-diffusion model was a good representation of ethanol transport through both filled and unfilled PDMS membranes, whereas the water flux did not obey this model due to the swelling effects. Incorporation of the fillers increased membrane stability and improved the selectivity. Performance of the SBS membranes characterized by a dense structure was found to be similar to the performance of filled PDMS membranes.

Evaluation of parameters describing flux decline in nanofiltration of aqueous solutions containing organic compounds

Abstract One of the major drawbacks for the introduction of membrane technology is the possible occurrence of fouling. Especially in newer processes such as nanofiltration, understanding of the mechanisms of flux decline and fouling is limited. Water fluxes obtained with nanofiltration of pure water are usually different from those obtained with real feed solutions. Depending on the feed composition, a flux decline ranging from a few percent to a complete loss of water flux can be found. For aqueous solutions containing organic components, in the absence of e.g. suspended solids or high concentrations of ions that may cause scaling, adsorption of organic material on the membrane surface is the major fouling mechanism. Identification of the parameters that play a role in the process of adsorption on the membrane surface should lead to a better understanding of the mechanism that results in a hindrance of the water flux, and eventually to pore blocking. In this study, the following parameters were selected for a detailed investigation of the adsorption process: the dipole moment, the polarisability, the dielectric constant, the solubility in water, the octanol—water partition coefficient, the contact angle membrane/water, the Small number, the modified Small number, molecular size, pKa, and the Taft parameter. All of these parameters were evaluated in the framework of adsorption on nanofiltration membranes from an aqueous solution. The use of each parameter for describing adsorption will be discussed. The pKa was rejected on theoretical grounds; other parameters such as the solubility in water proved to be impractical. Eventually, the dipole moment, the octanol—water partition coefficient, and the molecular size were selected. Correlations between adsorption on nanofiltration membranes and these interaction parameters show that there is a clear influence on adsorption.