Surfactant fouling of nanofiltration membranes: measurements and mechanisms.

Fouling of nanofiltration membranes is studied during filtration of aqueous surfactant solutions under different conditions. To this purpose, four typical nanofiltration membranes (Desal51HL, NF270, NTR7450 and NFPES10) and three typical surfactants (nonionic neodol, anionic SDBS and cationic cetrimide) are selected. Fouling is studied as a function of the surfactant concentration, with and without addition of an electrolyte (NaCl), at different pH and when filtering a mixture of surfactants. Adsorption experiments and hydrophobicity measurements (to study the orientation of the surfactants on the membrane surface) are also performed under the different conditions. The least membrane fouling is found for the anionic surfactant SDBS, while for the cationic surfactant cetrimide very low relative fluxes are observed. Neodol shows an intermediate degree of fouling. Both hydrophobic and electrostatic interactions (in the case of ionic surfactants) between the membrane surface and the surfactant explain the degree of adsorption and hence fouling, as membrane fouling is correlated with the amount of adsorbed surfactant. The difference between cetrimide and SDBS becomes especially visible when changing the pH: increasing the pH leads not only to an opposite orientation of the adsorbed surfactants, but also to an opposite trend in adsorbed amount and membrane fouling. This study permits selection of an optimal nanofiltration membrane to recycle wastewater containing surfactants in the carwash industry. The optimal choice would be a hydrophilic membrane with a low molecular weight cut-off and a small negative surface charge at neutral pH. Cationic surfactants in the wastewater should also be avoided as much as possible.     

Streaming potential measurements as a characterization method for nanofiltration membranes

The streaming potentials of two different nanofiltration membranes were studied with several electrolyte solutions to investigate the influence of salt type and concentration on the zeta potential and kinetic surface charge density of the membranes. The zeta potentials decreased with increasing salt concentration, whereas the kinetic surface charge densities increased. The kinetic surface charge densities could be described by Freundlich isotherms, except in one case, indicating that the membranes had a negligible surface charge. The kinetic surface charge density observed was caused by adsorbed anions. Salt retention measurements showed different mechanisms for salt separation for the two investigated membranes. One membrane showed a salt retention that could be explained by a Donnan exclusion type of separation mechanism, whereas for the other membrane the salt rejection seemed to be a combination of size and Donnan excluion. Comparing the results obtained by the streaming potential measurements with those of the retention measurements, it could be concluded that the membrane with the highest kinetic surface charge density showed the Donnan exclusion type of separation, whereas the membrane with the lower surface charge density showed a separation mechanism that was not totally determined by Donnan exclusion, size effects seemed to play a role as well.     

Thermoosmosis of various electrolyte solutions through anion-exchange membranes

Thermoosmosis through anion-exchange membranes was measured for 10^-3 to 2 mol/kg of aqueous KCl, LiCl, and NH₄Cl and for 10^-3 to 3 x 10^-1 mol/kg of aqueous KIO₃ and K₂SO₄. For all electrolytes used the direction of thermoosmosis was from the cold side to the hot side over the whole range of concentrations. For KCl and LiCl the experimental results were analyzed with an extension of a previously published theory, using additional data for transport numbers of ions in membranes and for electroosmosis. The agreement between theory and experiment is satisfactory. The absolute value of thermoosmosis for KIO₃ is larger than that for other electrolytes because the pore volume fraction of the membrane for KIO₃ is larger than that for the other electrolytes.     

Contribution of convection,diffusion and migration to electrolyte transport through nanofiltration membranes

Transport mechanisms through nanofiltration membranes are investigated in terms of contribution of convection, diffusion and migration to electrolyte transport. A Donnan steric pore model, based on the application of the extended Nernst-Planck equation and the assumption of a Donnan equilibrium at both membrane-solution interfaces, is used. The study is focused on the transport of symmetrical electrolytes (with symmetric or asymmetric diffusion coefficients). The influence of effective membrane charge density, permeate volume flux, pore radius and effective membrane thickness to porosity ratio on the contribution of the different transport mechanisms is investigated. Convection appears to be the dominant mechanism involved in electrolyte transport at low membrane charge and/or high permeate volume flux and effective membrane thickness to porosity ratio. Transport is mainly governed by diffusion when the membrane is strongly charged, particularly at low permeate volume flux and effective membrane thickness to porosity ratio. Electromigration is likely to be the dominant mechanism involved in electrolyte transport only if the diffusion coefficient of coions is greater than that of counterions.     

Fouling of nanofiltration membranes used for separation of fermented glycerol solutions

In this study, the glycerol solutions were fermented using Lactobacillus casei bacteria. The broths were pre-treated by microfiltration, followed by a further separation with nanofiltration. The latter process was carried out in two stages, using the NF270 and NF90 membranes, respectively. The concentrates thus obtained were enriched with citric acid (first stage) and then with lactic acid and glycerol (second stage). By means of SEM and AFM microscopy, as well as ATR-FTIR analysis, the intensity of membrane-fouling was studied. The colloidal fouling and bio-fouling caused a more than two-fold decrease in the permeate flux during microfiltration of the broth. This pre-treatment stage was effective, and a permeate turbidity of less than 0.2 NTU was obtained. However, the nanofiltration membranes exhibited a 30 % flux decline over the course of the process, mainly due to the organic fouling.     

Dielectric analysis of nanofiltration membrane in electrolyte solutions: influences of electrolyte concentration and species on membrane permeation

Dielectric properties of a nanofiltration membrane immersed in dilute aqueous electrolyte solutions were measured, and frequency dependence of capacitance and conductance of the systems was analyzed, based on the interfacial polarization theory, giving values of permittivity and conductivity of the membrane and the solutions. Permittivity, epsilon m, of the membrane slightly decreased whereas conductivity, km, of the membrane increased with increasing electrolyte concentration, as a result of entrance of ions into the membrane. The ratio of membrane/solution conductivity, km/kw, also depended on the electrolyte concentration, showing that distribution of ions in the membrane and in solutions follow Donnan equilibrium, due to the presence of negative fixed charges in the membrane. New expressions were derived from Donnan equilibrium principle to explain this phenomenon, and negative fixed charge concentration ce of the membrane was obtained; thus the Donnan potential, DeltaPhi Don, of the membrane in solutions at various concentrations could be calculated. The new expressions could be expected to be usable to analyze ion permeation property through membrane.     

Colloidochemical properties of ultra-and microfiltration membranes in electrolyte solutions

Dependences of the structural, electrokinetic, and adsorption characteristics on solution pH and background electrolyte (NaCl) concentration are extensively studied for Sartorius and Vladisart cellulose acetate microfiltration membranes with pore sizes of 0.45 and 0.2 μm and a Vladisart ultrafiltration membrane with the rejection of 20 kD. It is revealed that effective hydrodynamic pore radii and maximum pore radii of the microfiltration membranes are 1.5-to 2-and 2.5-to 4-fold, respectively, larger than those presented in the catalog, which is related to the membrane calibration relative to the sizes of rejected particles. For the ultrafiltration membrane, it is shown that, when the pressure increased from 0.5 to 8.0 atm, filtration factor of a liquid and streaming potential substantially decrease owing to the contraction of the polymer network. Measurements of membrane conductivity by the difference and contact methods suggest that a structural anisotropy is virtually absent in the microfiltration membranes and that the ultrafiltration membrane has a nonuniform structure. Negative electrokinetic potentials, whose absolute values increase with the pH and dilution of a background electrolyte solution, are observed for all studied membranes. Isoelectric points of the ultrafiltration and microfiltration membranes are observed at pH ≤ 3 and 2.1 ± 0.2, respectively.     

Separation of aqueous electrolyte solutions with asymmetric membranes containing one charged layer

The Nernst-Planck equation and fine-pore membrane model are applied to describe the ultra- and nanofiltration of electrolyte solutions through a inhomogeneous membrane containing one charged layer. Concentration and electric potential distributions, as well as dependences of electrolyte rejection coefficient (selectivity) and streaming potential on system parameters are determined. Asymmetry effect is revealed with respect to the rejection coefficient and streaming potential at different orientations of the selective charged layer relative to the direction of the filtration flow. The cases of 1: 1 and 1: 2 electrolytes are investigated in detail. Theoretical calculations demonstrate that the rejection coefficient of a bi-layer membrane rises in the following series of binary electrolytes: 1: 2 < 1: 1 < 2: 1, when the first layer is positively charged, and in the opposite series of these electrolytes, when the first layer is negatively charged.     

Streaming potential across cation-exchange membranes in methanol-water electrolyte solutions

Streaming potential measurements across charged membranes separating two equal solutions have been carried out. Two cation-exchange membranes with different cross-linked and swelling properties (Ionics and Nafion membranes) and methanol-water electrolyte solutions of KCl have been used in the experiments. The obtained results show that the streaming potential is higher for the Ionics membrane and that the values depend on the methanol content of the solutions. A different behavior is found in the dependence of the streaming potential on the methanol percentage for each membrane. The study of the relaxation times in the decay of electrokinetic steady states of streaming potential has been carried out from the time dependence of the streaming potential when the pressure difference through the membrane is suppressed. The results show the existence of two different parts or partial relaxations, mechanical and electric. A different behavior of the mechanical relaxation time with the methanol percentage has been found for the two membranes, but any significant difference between their electric relaxation times. These differences have been explained in terms of the different degree of swelling of the membranes used.     

Estimated ionicB-coefficients from NMR measurements in aqueous electrolyte solutions

¹⁷O-NMR spin-lattice relaxation timesT ₁ of D₂O molecules were measured at 5–85°C in D₂O solutions of alkali metal halides (LiClCsCl, KBr, and KI), DCl, KOD, Ph₄PCl, NaPh₄B, and tetraalkylammonium bromides (Me₄NBrAm₄NBr) in the concentration range 0.1–1.4 mol-kg^–1 TheB-coefficients of the electrolytes obtained from the concentration dependence of relaxation ratesR ₁=1/T₁ were divided into the ionicB-coefficients by three methods: (i) the assumption ofB (K⁺)=B(Cl^–), (ii) the assumption ofB(Ph₄P⁺)=B(Ph₄B^–), and (iii) the use ofB(Br^–) obtained from a series ofB(R₄NBr). It was found that Methods (ii) and (iii) resulted in an abnormal temperature dependence of theB-coefficients of alkali metal ions and a negative values of rotational correlation times _c at lower temperatures for hydroxide and halide ions. These results suggest that the methods based on the van der Waals volume are not adequate for the ionic separation of NMRB-coefficients. From the analysis using the assumption ofB(K⁺)=B(Cl^–), it was found that D₃O⁺, OD^–, and Me₄N⁺ ions are the intermediates between structure makers and breakers, and that the hydrophobicity of phenyl groups is weaker than that of alkyl groups due to the interactions between water molecules and -electrons in phenyl groups.     

Specific interactions in mixed electrolyte solutions from solubility measurements

The solubility of thallium I chloride in a wide range of dilute electrolyte solutions is interpreted using a new specific-interaction equation for activity coefficients. This equation is shown to be consistent with the solubility data, but the Guggenheim specific-interaction equation is not. The relation of interaction coefficients to ion association is discussed.     

On the prediction of permeate flux for nanofiltration of concentrated aqueous solutions with thin-film composite polyamide membranes

The nanofiltration of binary aqueous solutions of glucose, sucrose and sodium sulfate was investigated using thin-film composite polyamide membranes with different molecular weight cut-off's. The NF experiments, in total recycle mode, were performed in a plate-and-frame module Lab 20 (AlfaLaval), at 22 °C and with a flowrate of 8.2 L/min, using the membranes NF90, NF200 and NF270 from FilmTec (Dow Chemical), for transmembrane pressures between 1 and 6 MPa and with aqueous solutions with osmotic pressures of between 0.5 and 3.0 MPa. The permeate flux was predicted by the osmotic pressure model, using the membrane hydraulic resistance and the solution viscosity inside the membrane pores, and computing the concentration polarization with recourse to a mass-transfer correlation specific for the plate-and-frame module used. The flux predictions, using the pure water viscosity, agree reasonably with the experimental data only for low transmembrane pressures and with the most diluted solutions. For higher transmembrane pressures and for higher solute concentration the predicted fluxes can be as far as 2.5, 4.1 and 9.6 times higher than the experimental one, for the aqueous solutions of Na₂SO₄, glucose and sucrose, respectively. These deviations are strongly reduced when the pure water viscosity is replaced by the solution viscosity adjacent to the membrane. In this case, the maximum deviation between predictions and experiments occurs also for higher transmembrane pressures and for higher solute concentration, but the maximum ratio between predicted values and the experiments were reduced now to 1.8, 2.1 and 2.9, for the aqueous solutions of Na₂SO₄, glucose and sucrose, respectively. Even using the solution viscosity adjacent to the membrane, and for the systems investigated, the osmotic pressure model must used with caution for design purposes because it may over predict the permeate flux by a factor of about 2 when the solute concentration is high.     

Asymmetric layer-by-layer polyelectrolyte nanofiltration membranes with tunable retention

Nanofiltration (NF) membrane processes are attractive to remove multivalent ions. As ion retention in NF membranes is determined by both size and charge exclusion, negatively charged membranes are required to reject negatively charged ions. Layer-by-layer assembly of alternating polycation (PC) and polyanion layers on top of a support is a versatile method to produce membranes. Especially the polyelectrolyte (PE) couple polydiallyldimethylammoniumchloride and poly(sodium-4-styrenesulfonate) (PDADMAC/PSS) is extensively investigated. This PE couple cannot form highly negatively charged membrane surfaces, due to interdiffusion and charge overcompensation of PDADMAC into the PSS layers, which limits the operational window to tailor membrane properties. We propose the use of asymmetric layer formation and show how combining two charge densities of one PC can produce negatively charged NF membranes. Starting from hollow fiber ultrafiltration supports coated with base layers of PDADMAC/PSS, they are coated with PDADMAC/PSS or poly(acrylamide-co-diallyldimethylammoniumchloride), P(AM-co-DADMAC)/PSS layers. P(AM-co-DADMAC) has a charge density of only 32% compared to 100% for PDADMAC. The particular novel membranes coated with P(AM-co-DADMAC) have a highly negatively charged surface and high permeabilities (7–19 L/[m²hbar]), with high retentions for Na₂SO₄ of up to 95%. These values position the developed membranes in the top range compared to commercial and other layer-by-layer membranes.     

Effect of membrane pretreatment on performance of solvent resistant nanofiltration membranes in methanol solutions

This paper reports the effect of membrane pretreatment using different organic solvents on the performance of polyamide, polyimide and polydimethylsiloxane (PDMS) membranes in methanol solutions. Membrane pretreatment using acetone, methanol and toluene results in significant changes of membrane flux and rejection for polyamide- and polyimide-based membranes (Desal-DK and STARMEM 228) due to membrane swelling. The Performance of a polydimethylsiloxane (PDMS)-based membrane (MPF-50) in methanol solutions was not significantly affected by membrane pretreatment.     

Study on transmembrane electrical potential of nanofiltration membranes in KCl and MgCl2 solutions

The transmembrane electrical potential (TMEP) across two commercial nanofiltration membranes (ESNA1-K and Filmtec NF) was investigated in KCl and MgCl(2) solutions. TMEP was measured in a wide range of salt concentrations (1-60 mol·m(-3)) and pH values (3-10) at the feed side, with pressure differences in the range of 0.1-0.6 MPa. A two-layer model based on the Nernst-Planck equation was proposed to describe the relation between TMEP and permeation flux. From the pattern of these curves, the information of membrane structure could be deduced. In the concentration range investigated, TMEP in KCl solutions was always positive and decreased as the salt concentration increased. The contribution of the membrane potential to the TMEP decreased. TMEP was greatly affected by the feed pH. When the feed pH increased, the mobility of cations increased, which indicated that the charges of NF membranes were more negative. The zero point of TMEP and the minimum of rejection in KCl solution were consistent and occurred at the isoelectric point of NF membranes, while in MgCl(2) solution the zero point of TMEP located at a higher pH value. The TMEP in MgCl(2) solutions changed its sign at a given concentration, and by calculating the transport number the location of the minimum rejection could be determined.     

Dielectric characterization of a nanofiltration membrane in electrolyte solutions: its double-layer structure and ion permeation

Dielectric spectroscopy (DS) was applied to a nanofiltration (NF) membrane to detect its double-layer structure and ion permeation. Dielectric measurements were carried out on the systems composed of the NF membrane NTR7450 and dilute solutions of eight electrolytes, LiCl, NaCl, KCl, NH(4)Cl, MgCl(2), CaCl(2), BaCl(2), and CuCl(2). Two relaxations were observed in the frequency range from 40 Hz to 4 MHz for each system. On the basis of characteristics of the dielectric spectra and the Maxwell-Wagner interfacial polarization theory, the low-frequency relaxation was attributed to inhomogeneity of the membrane structure itself, whereas the high-frequency relaxation was attributed to interfacial polarization between the membrane and the solution. A multiphase dielectric model previously developed by one of the authors and co-workers was adopted to present systems to analyze the dielectric spectra, and electric parameters, i.e., capacitance and conductance, of the two layers composing the membrane were obtained. The electric properties estimated for the two layers were different and changed with the environment in a different manner. Further analyses suggest that the two layers had a different separation mechanism due to their difference in materials, looseness, and fixed charge content. The fixed charge density of one layer was estimated, and the ion permeation difficulties in both layers was compared. This research revealed that DS was by far an effective method to obtain detailed electric parameters about the inner multilayer structure of the NF membrane and to elucidate separation mechanisms of each layer.     

Evaluating the charge of nanofiltration membranes

In this study, several methods were used to determine the charge of commercially available nanofiltration membranes, and were compared. First the ion-exchange capacity was determined by titration, this method is able to distinguish between positively and negatively charged functional groups on the membrane. Secondly, measurements of the streaming potential gave a value for the charge density at the exterior membrane surface; the effect of the pH of the solution on the membrane charge was studied. Finally, measurements of the membrane potential allowed to evaluate the total membrane charge density.The results of the three methods were used to compare the membrane charge of nanofiltration membranes mainly in a qualitative way. It is shown that measurements of the membrane potential are preferred for the evaluation of the membrane charge.     

Physico-chemical characterization of nanofiltration membranes.

This study presents a methodology for an in-depth characterization of six representative commercial nanofiltration membranes. Laboratory-made polyethersulfone membranes are included for reference. Besides the physical characterization [molecular weight cut-off (MWCO), surface charge, roughness and hydrophobicity], the membranes are also studied for their chemical composition [attenuated total reflectance Fourier spectroscopy (ATR-FTIR) and X-ray photoelectron spectroscopy (XPS)] and porosity [positron annihilation spectroscopy (PAS)]. The chemical characterization indicates that all membranes are composed of at least two different layers. The presence of an additional third layer is proved and studied for membranes with a polyamide top layer. PAS experiments, in combination with FIB (focused ion beam) images, show that these membranes also have a thinner and a less porous skin layer (upper part of the top layer). In the skin layer, two different pore sizes are observed for all commercial membranes: a pore size of 1.25-1.55 angstroms as well as a pore size of 3.20-3.95 angstroms (both depending on the membrane type). Thus, the pore size distribution in nanofiltration membranes is bimodal, in contrast to the generally accepted log-normal distribution. Although the pore sizes are rather similar for all commercial membranes, their pore volume fraction and hence their porosity differ significantly.     

Ion transport modelling through nanofiltration membranes

The aim of this work is to study the transport mechanism of ions through nanofiltration membranes. A model based on extended Nernst–Planck and film theory equations is reported. This model can be characterized by three transport parameters: the water permeability L_p, the salt transmittance Φ and the effective salt transfer coefficient K_eff. The knowledge of the feed and permeate concentration and of the permeate volumetric flux enable us to calculate these transport parameters. The model is used to estimate cadmium salts rejection by a NANOMAX 50 membrane. Experimental and calculated results are shown to be in good agreement. The model is then successfully extended to experimental data reported in the literature.     

Polypyrrole modified solvent resistant nanofiltration membranes

New solvent resistant nanofiltration (SRNF) membranes with polypyrrole (PPy) modified toplayer were prepared on different types support by in situ pyrrole polymerization. The morphology of the membranes was studied by SEM. The PPy modified membranes were applied in the filtration of organic solvents. All the PPy modified membranes showed a very high retention of the negatively charged RB in different solvent systems, comparable to those of the MPF-50 and STARMEM 122 commercial membranes, but at much higher flux. The extended filtration experiment in strong aprotic DMF of PPy modified membranes showed a clearly stable permeability and retention over 30 h. In addition, the PPy modified membranes showed a much higher flux in THF systems than for earlier reported crosslinked poly(imide) membranes.