纳滤膜分离机理及其应用研究进展

综述了纳滤膜的分离机理及其应用研究现状和进展,纳滤膜分离过程是一个不可逆过程,其分离机理可以运用电荷模型（空间电荷模型和固定电荷模型）和细孔模型,以及近年才提出的静电排斥和立体阻碍模型等来描述。纳滤膜应用研究现状的介绍包括低聚糖分离和精制、果汁的高浓度浓缩、多肽和氨基酸的分离、抗生素的浓缩与纯化、牛奶及乳清蛋白的浓缩、农产品的综合利用以及纳滤膜生化反应器的开发等。     

Determination of concentration-dependent transport coefficients in nanofiltration: Experimental evaluation of coefficients

To characterize solute transport in nanofiltration (NF) the Spiegler–Kedem equation requires that two coefficients be determined for two-component solutions (a solute in water), solute permeability ω and reflection coefficient σ. For salts both coefficients strongly and in a complex way depend on concentration, which greatly complicates their evaluation from experiments. For this reason, the parameters are usually assumed constant for a given feed and the concentration dependence is assessed from flux–rejection curves for several feeds. This procedure however ignores the fact that the solute concentration and hence the coefficients significantly vary across the membrane. One way to overcome this inconsistency and address concentration dependence is to use physical models explicitly introducing exclusion mechanism(s) and fitting relevant membrane-specific parameters, such as fixed charge or dielectric properties. This procedure often fails to produce unique values of parameters for a given membrane and different salts. In the present study a new phenomenological approach is proposed and critically analyzed, based on the assumption of a similar concentration dependence of ω and 1 − σ, previously shown to be valid under fairly general conditions, thereby the Peclét coefficient A = (1 − σ)/ω may be assumed to be independent of concentration. The coefficients and their concentration dependence for a given solute may be directly and consistently evaluated by fitting flux–rejection data for several feeds and fluxes to numeric solution of the modified transport equations without the need to invoke specific physical models. The values of transport parameters deduced in this way for representative membranes and salts allow important conclusions regarding the transport mechanism. In particular, the roles of different mechanisms in overall salt exclusion could be addressed directly from the variation of ω or 1 − σ with concentration. On the other hand, the value of the Peclét coefficient, free of the effect of salt partitioning, may be analyzed in terms of hindered transport. Using the proposed method, this value was found to be very small for studied thin-film composite membranes, which may significantly simplify the transport equations.     

Concentration polarization in a reverse osmosis/nanofiltration plate-and-frame membrane module

The flow structure and solute concentration distribution in a nanofiltration/reverse osmosis plate-and-frame module with radial thin feed channels that have considerable entrance and outlet effects was determined by computational fluid dynamics (CFD). Simulations were performed for binary aqueous solutions, Reynolds (Re) numbers in the range of 64–570 (based on the channel height) and Schmidt (Sc) numbers between 450 and 8900.     

Modelling the separation by nanofiltration of a multi-ionic solution relevant to an industrial process

The current work focuses on the application of nanofiltration (NF) to the isolation of a pharmaceutical product, clavulanate (CA⁻), from clarified fermentation broths, which show a complex composition with five main identified ions (K⁺, Cl⁻, NH₄⁺, SO₄²⁻ and CA⁻). Our aim is to predict the rejection rates of these five ions, with the NF membrane Desal-DK, which may influence the separation of CA⁻ and play a role in the whole downstream process.     

Prediction of the concentration polarization in the nanofiltration/reverse osmosis of dilute multi-ionic solutions

The aim of this work was to develop a simple and accurate model for predicting the concentration polarization index in the nanofiltration (NF)/reverse osmosis (RO) of dilute multi-ionic solutions. On the grounds of this model, the total flux of the ion i at the feed-solution/membrane interface consists of the sum of the diffusion, convection and migration fluxes, the former of which is determined by conventional mass-transfer correlations duly corrected to take into account the permeation through the membrane (suction effect). The coupling of the ionic fluxes is enforced by the electroneutrality requirement at the feed-solution/membrane interface. The model developed dispenses with the arbitrary assumption of the thickness of a film layer in the vicinity of the membrane surface.

Assessing the accuracy/validity of this model with multi-ionic solutions would be rather harsh, thus the model accuracy and ranges of validity were ascertained for a benchmark case: NF/RO of single salt solutions. The model predicts approximate concentration polarization indexes of the salts A⁺B⁻, A⁺₂B²⁻ and A⁺₃B³⁻ (or A²B⁻₂ and A³⁺B⁻₃) with positive deviations lower than 10% with respect to the benchmark concentration polarization index, for ions diffusivities ratios, D₁/D₂ (or D₂/D₁), in the range 0.16–5.5 and ≡J_v/k_c<3, where J_v is the permeation flux and k_c is the mass-transfer coefficient of the salt for vanishing mass-transfer rates at impermeable walls. The main assumption of the model – the individual mass-transfer coefficients of the ions are independent of each other – appears to hold in a broad range of conditions, for single salt solutions.

The model developed was expeditely applied to predict the concentration polarization in the nanofiltration of solutions containing Na⁺, Cl⁻ and a dye³⁻ (experimental data of Bowen and Mohammad [AIChE J. 44 (8) (1998) 1799–1812]), and its predictions are in fair agreement with the predictions of the extended Nernst–Planck equations in the film layer of the “slowest” ion.     

Influence of operating conditions on the rejection of cobalt and lead ions in aqueous solutions by a nanofiltration polyamide membrane

The potential use of nanofiltration polyamide membrane for removing cobalt and lead ions from wastewater was investigated. Rejection experiments were conducted with Pb(NO₃)₂ and Co(NO₃)₂ in both single-salt solutions and mixtures. Experimental rejection rates were corrected for concentration polarization phenomenon by means of film theory. The structural features of the membrane (pore radius and thickness-to-porosity ratio) were first estimated from the fitting of glucose rejection rates. Its surface charge properties were then investigated in single-salt solutions at pH values between 3 and 7. Rejection of both heavy metal ions was found to be influenced by operating conditions such as permeate flux, solution pH and feed salt concentration. In single-salt solutions, rejection of lead was higher than that of cobalt at pH ≥ 5. This behavior may be explained by (i) higher normalized volume charge density in the Pb(NO₃)₂ than in the Co(NO₃)₂ solution and (ii) lower ionic strength of the Pb(NO₃)₂ solution as compared with the Co(NO₃)₂ solution. At pH < 5, the dielectric exclusion would be more important for Co(NO₃)₂ than for Pb(NO₃). Lead rejection was almost the same in both single-salt solutions and ternary mixtures, whereas cobalt rejection was strongly affected by the presence of lead. Cobalt was found to be rejected much more than lead in mixtures at equal mass concentrations, the difference between rejections of the two cations being greater as pH increased.     

Experimental analysis, modeling, and theoretical design of McMaster pore-filled nanofiltration membranes

A side-by-side comparison of the performance of McMaster pore-filled (MacPF) and commercial nanofiltration (NF) membranes is presented here. The single-salt and multi-component performance of these membranes is studied using experimental data and using a mathematical model. The pseudo two-dimensional model is based on the extended Nernst–Planck equation, a modified Poisson–Boltzmann equation, and hydrodynamic calculations. The model includes four structural properties of the membrane: pore radius, pure water permeability, surface charge density and the ratio of effective membrane thickness to water content. The analysis demonstrates that the rejection and transport mechanisms are the same in the commercial and MacPF membranes with different contributions from each type of mechanism (convection, diffusion and electromigration). Solute rejection in NF membranes is determined primarily by a combination of steric and electrostatic effects. The selectivity of MacPF membranes is primarily determined by electrostatic effects with a significantly smaller contribution of steric effects compared to commercial membranes. Hence, these membranes have the ability to reject ions while remaining highly permeable to low molecular weight organics. Additionally, a new theoretical membrane design approach is presented. This design procedure potentially offers the optimization of NF membrane performance by tailoring the membrane structure and operating variables to the specific process, simultaneously. The procedure is validated at the laboratory scale.     

Prediction of physical properties of nanofiltration membranes using experiment and theoretical models

Two commercial nanofiltration (NF) membranes, viz., Desal-HL and NF 700 MWCO were investigated experimentally using neutral and charged solutes, viz., glucose, sodium chloride and magnesium chloride. Effect of pH was studied for sodium chloride rejection and isoelectric point of the membrane was deduced. Experimental results were analyzed using Donnan steric pore and dielectric exclusion models. Dielectric exclusion arises due to the difference in dielectric constant between the bulk and the nano-pore. Born dielectric effect was used as dielectric exclusion phenomena in the present investigation. Stokes–Einstein, Born effective and Pauling radii were used for theoretical simulation, which accurately predicted different charge densities. Empirical correlations were proposed between charge density, concentration and pH for each radius. Charge density decreased drastically when dielectric exclusion term was included in the theoretical model, which showed the real physical characteristics of the membranes employed. Charge density and radius of pore was found to be an important surface parameter in predicting the separation effects in NF membranes.     

Theory of pressure-driven transport of neutral solutes and ions in porous ceramic nanofiltration membranes

Hindered transport theory and homogeneous electro-transport theory are used to calculate the limiting, high volume flux, rejection of, respectively, neutral solutes and binary electrolytes by granular porous nanofiltration membranes. For ceramic membranes prepared from metal oxides it is proposed that the membrane structural and charge parameters entering into the theory, namely the effective pore size and membrane charge density, can be estimated from independent measurements: the pore radius from the measured hydraulic radius using a model of sintered granular membranes and the effective membrane charge density from the hydraulic radius and the electrophoretic mobility measurements on the ceramic powder used to prepare the membrane. The electro-transport theory adopted here is valid when the membrane surface charge density is low enough and the pore radius is small enough for there to be strong electrical double layer overlap in the pores. Within this approximation the filtration streaming potential is also derived for binary 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.     

Identification of dielectric effects in nanofiltration of metallic salts

Transport of four metallic salts (CuCl₂, ZnCl₂, NiCl₂ and CaCl₂) through a polyamide nanofiltration (NF) membrane has been investigated experimentally from rejection rate and tangential streaming potential measurements. Rejection rates have been further analyzed by means of the steric, electric and dielectric exclusion (SEDE) homogeneous model with the effective dielectric constant of the solution inside pores as the single adjustable parameter.     

Critical flux in NF of high molar mass polysaccharides and effluents from the paper industry

High molar mass polysaccharides (locust bean gum and karaya gum) and effluents from a mechanical pulp mill and a paper mill were nanofiltered with commercially available nanofiltration (NF) membranes. The effect of the filtration conditions on the flux (critical flux), retention, and the fouling of the membranes was studied. The experiments were conducted by increasing and decreasing the pressure and measuring the flux thus obtained.

The critical flux was observed to increase with increasing flow velocity and decreasing concentration. An increase in pH increased the electrostatic repulsion between the membrane and the dissociated (charged) components in the paper mill effluents. As a result, a higher critical flux was obtained and also the retentions of the charged substances improved. Only a weak form of the critical flux was observed with the mill effluents. The permeate flux deviated from the pure water flux even at the lowest pressure, but increased linearly with pressure until the weak form of the critical flux was exceeded. The small decrease in flux immediately after filtration was started was probably caused by the plugging of the free spaces in the membranes or by the adsorption of foulants onto the membrane surface.

In the filtrations with the high molar mass polysaccharides, a strong form of the critical flux as well as a weak form were observed. The significant irreversible fouling of the most hydrophobic membrane was due to adsorption of the model substances by hydrophobic interaction. A precleaning of the membranes with an alkaline cleaning agent improved the pure water fluxes by up to 30%, but it had only a small effect on the critical or the limiting flux. The pure water fluxes of precleaned membranes after filtration were still higher than the pure water fluxes of the untreated membranes before filtration.     

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.     

Modeling the separation performance of nanofiltration membranes for the mixed salts solution

A new model is proposed to evaluate the separation performance of nanofiltration (NF) membranes for the mixed salts solution. In the model, the observed transmission of an ion through a NF membrane is applied to express the separation performance of the membrane for the ion in the mixed salts solution, which has a relationship with the total concentration, the equivalent fraction and the species of each ion in the mixed salts solution. The verification of the model was carried out in the permeation experiments of some mixed salts solutions ((1) Na⁺, Cl⁻ and F⁻; (2) Na⁺, K⁺ and Cl⁻; (3) Na⁺, F⁻, Cl⁻ and NO₃⁻; (4) Na⁺, Cl⁻, NO₃⁻ and SO₄²⁻) through three commercial NF membranes (ESNA 1-LF, ESNA 1 and LES 90). According to the permeation experiments of three NF membranes for some binary salts solutions, the competition coefficients of ions were obtained. The model evaluation results agreed quite well with the experimental data. Finally, the model was applied to evaluate the observed transmission of each ion in the mixed salts solution (Na⁺, F⁻, Cl⁻, NO₃⁻ and SO₄²⁻) through three NF membranes. The agreement between the model evaluation results and the experimental data indicated that the model is suitable for evaluating the separation performance of three NF membranes for the mixed salts solution.     

Effect of colloids on salt transport in crossflow nanofiltration

The role of colloid deposition on the performance of a salt-rejecting NF membrane was evaluated by modeling salt transport using a two-layer transport model, which quantified the relative contributions of advection and diffusion in the cake and the membrane layers, and the effects of flux on the membrane sieving coefficient. The model was able to accurately describe how the measured permeate concentration, rejection, osmotic pressure, and flux decline varied with time. The two-layer model confirmed that the Peclet number in the cake layer was about an order of magnitude higher than that in the membrane layer, leading to significant concentration polarization at the membrane surface, as shown by others. However, the cake layer also increased overall resistance, which resulted in flux decline during constant pressure operation. Flux decline caused an increase in the actual sieving coefficient, leading to higher solute flux, lower observed rejection, and thus lower the bulk concentration. These coupled phenomena tended to mitigate the increase in concentration polarization caused by the cake. Therefore, as predicted by the model and verified by experiment, the osmotic pressure does not increase monotonically as the cake grows, and in fact can decrease when the cake layer is thick and the flux decline is significant. In our experimental system, the pressure drop across the cake layer, which was proportional to the cake thickness, was significant under the conditions studied. The effects of cake-enhanced osmotic pressure analyzed here are lower than those observed in previous studies, possibly because the transport model employed explicitly accounts for the effect of flux decline due to cake growth on the membrane sieving coefficient, and possibly because we used a somewhat different methodology to estimate cake porosity.     

Mass diffusion-based separation of sugars in a microfluidic contactor with nanofiltration membranes

Processes such as chromatographic separation and nanofiltration can remove low molecular weight sugars from liquid mixtures of oligosaccharides. As an alternative for the separation of such liquid mixtures, we studied mass diffusion separation of such sugars in a microfluidic device with incorporated nanofiltration membranes. This separation method is based on differences between diffusivities of components and does not require high transmembrane pressures. The effects of channel depth and flow rate were studied in experiments. The key parameters selectivity and rejection increased with increasing channel depth due to increased external mass transfer limitations. Among the studied membranes, the obtained selectivities and rejections correlated to the specified retention values by the manufacturers. Compared to more conventional nanofiltration where high pressure forces solutes through membranes, we obtained corresponding selectivities and fluxes of only an order of magnitude smaller. Simulated results indicated that with optimized microchannel and membrane dimensions, the presented separation process can compete with currently available separation technologies.     

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.     

Solvent-induced swelling of membranes — Measurements and influence in nanofiltration

This paper describes improvements to an apparatus for in-situ determinations of swelling where a linear inductive probe and electronic column gauge with an overall resolution of 0.1 μm was used for measurements of seven variants of polyacrylonitrile (PAN)/polydimethylsiloxane (PDMS) composite nanofiltration membranes in a range of alkane, aromatic and alcohol solvents. The unswollen membranes incorporated PDMS layers between 1 and 10 μm nominal thickness and were manufactured with a radiation and/or thermal crosslinking step.

The tested membranes exhibited a range of swelling dependent on the degree of crosslinking, the initial PDMS layer thickness and the type of solvent. With no applied pressure the PDMS layer on some radiation cross-linked membranes swelled as much as 170% of the initial thickness whilst other membranes were restricted to a maximum swelling of 80%. When a pressure up to 2000 kPa was applied to a membrane then swelling could be reduced to 20% of the value obtained at zero applied pressure. By vertically stacking up to three membrane samples it was possible to determine the swelling of PDMS layers as thin as 1 μm, although higher imposed pressures rendered some results unreliable as the measurement resolution of the apparatus was approached. The results of the swelling experiments are contrasted with crossflow nanofiltration performance in terms of solvent flux and solute rejection.     

3D topography design of membranes for enhanced mass transport

This work describes the process of fabrication of 3D topography membranes and the fully quantitative characterisation of their topography using atomic force microscopy (AFM), small-angle light scattering (SALS), scanning electron microscopy (SEM) and polarizing optical microscopy (POM). The use of these membranes and the impact of the 3D membrane topography on the enhancement of mass transport during solute recovery (hexyl acetate) from a viscous room temperature ionic liquid 1-n-butyl-3-methyl-imidazolium tetrafluoroborate ([C₄mim] [BF₄]) by organophilic pervaporation is presented and discussed.