Asymmetric Cellulose Acetate Membranes Research Articles

Effect of poly(vinyl pyrrolidone) concentration and coagulation bath temperature on the morphology, permeability, and thermal stability of asymmetric cellulose acetate membranes

AbstractCellulose acetate (CA) is widely used in membrane processes. In this study, CA (weight‐average molecular weight = 52,000) was mixed with poly(vinyl pyrrolidone) (PVP; weight‐average molecular weight = 15,000) as an additive in 1‐methyl‐2‐pyrrolidone as a solvent. The phase‐inversion method was used for the preparation of flat‐sheet membranes. The effects of PVP concentration and coagulation bath temperature (CBT) on the morphology, pure water permeation flux, and thermal stability of the prepared membranes were studied and are discussed in this article. The solute rejection of the developed CA membranes was quantified with an insulin protein solution. The results showed that an increase in the CBT levels from 0 to 23°C along with an increase in the PVP concentration in the cast film from 0 to 1.5 wt % resulted in an increase in the macrovoid formation in the membrane sublayer, an increase in the pure water flux (PWF), and a decrease in insulin rejection. Further increases in the PVP concentration from 1.5 to 3, 6, and 9 wt % resulted in gradual suppression of the macrovoid formation, a decrease in PWF, and an increase in insulin rejection. Higher PVP concentrations and lower CBT levels also appeared to result in higher glass‐transition temperatures. © 2008 Wiley Periodicals, Inc. J Appl Polym Sci, 2009

Effect of preparation variables on morphology and pure water permeation flux through asymmetric cellulose acetate membranes

In this study, cellulose acetate (CA) ultrafiltration (UF) membranes were prepared using the phase inversion method. Effects of CA and polyethylene glycol (PEG) concentrations in the casting solution and coagulation bath temperature (CBT) on morphology of the synthesized membranes were investigated. Based on L 9 orthogonal array of Taguchi experimental design 18 membranes were synthesized (with two replications) and pure water permeation flux through them were measured. It was found out that increasing PEG concentration in the casting solution and CBT, accelerate diffusional exchange rate of solvent 1-methyl-2-pyrrolidone (NMP) and nonsolvent (water) and consequently facilitate formation of macrovoids in the membrane structure. Increasing CA concentration, however, slows down the demixing process. This prevents instantaneous growth of nucleuses in the membrane structure. Hence, a large number of small nucleuses are created and distributed throughout the polymer film and denser membranes are synthesized. Rate of water flux through the synthesized membranes is directly dependent on the size and number of macrovoids in the membrane structure. Thus, maximum value of flux is obtained at the highest levels of PEG concentration and CBT (10 wt.% and 23 °C, respectively) and the lowest level of CA concentration (13.5 wt.%). Analysis of variance (ANOVA) showed that all parameters have significant effects on the response. However, CBT is the less influential factor than CA and PEG concentrations on the response (flux).

Effect of Membrane Pore Size and Membrane-Solute Interactions on Lysozyme Ultrafiltration

A model based on steric hindrance mechanisms [1] is used to determine the pore sizes of two ultrafiltration (UF) membranes. The lysozyme rejection coefficients of those membranes are predicted through the same model after modification of the pore size and solute radius by taking into account the development of electric double layers. Two asymmetric cellulose acetate membranes M1 and M2 were prepared and characterized. Membrane M1 has an hydraulic permeability of 2.1x10-6 m/s/bar, a molecular weight cut-off (MWCO) of 30,000 Da and an average pore radius of 2.6 nm. Membrane M2 has an hydraulic permeability of 5.9x10-6 m/s/bar, a molecular weight cut-off (MWCO) of 60,000 Da and an average pore radius of 5.3 nm. Aqueous solutions of lysozyme containing a NaCl concentration of 0.1M were ultrafiltrated through membranes M1 and M2. The predicted lysozyme rejections considering the development of electric double layers on the protein and membrane pore surfaces, are in good agreement with the experimental results.

Reverse osmosis of binary organic solute mixtures in the presence of strong solute-membrane affinity

Reverse osmosis performance for certain low molecular weight organic solutes in water, such as phenol and its derivatives, is markedly different from that of simple aqueous salt solutions. The unusual solute-membrane affinity reverse osmosis behaviour of phenol-water-cellulose acetate system has been observed. This strong solute-membrane affinity has caused the separation decrease with increasing pressure and permeate flux reduction in the system. This is due to the strong solute-membrane affinity through cellulose acetate (CA) membranes. This work deals with the separation of a mixed organic solute system. A systematic study has been conducted to demonstrate the applicability of reverse osmosis fractionation of diluted aqueous solutions, namely phenol, p-nitrophenol and 3-chlorophenol by asymmetric cellulose acetate membranes. Besides, the effects of operating pressure and feed concentration on the performance of separation are investigated. Results show that the separation can be positive, zero, or negative, depending on the operating pressure, solute concentration as well as the coupling solutes.

Reverse osmosis of binary organic solute mixtures in the presence of strong solute-membrane affinity

Reverse osmosis performance for certain low molecular weight organic solutes in water, such as phenol and its derivatives, is markedly different from that of simple aqueous salt solutions. The unusual solute-membrane affinity reverse osmosis behaviour of phenol-water-cellulose acetate system has been observed. This strong solute-membrane affinity has caused the separation decrease with increasing pressure and permeate flux reduction in the system. This is due to the strong solute-membrane affinity through cellulose acetate (CA) membranes. This work deals with the separation of a mixed organic solute system. A systematic study has been conducted to demonstrate the applicability of reverse osmosis fractionation of diluted aqueous solutions, namely phenol, p-nitrophenol and 3-chlorophenol by asymmetric cellulose acetate membranes. Besides, the effects of operating pressure and feed concentration on the performance of separation are investigated. Results show that the separation can be positive, zero, or negative, depending on the operating pressure, solute concentration as well as the coupling solutes.

In situ reparation of defects on the skin layer of reverse osmosis cellulose ester membranes for pervaporation purposes

The in situ treatment of reverse osmosis (RO) membranes of Loeb–Sourirajan type (asymmetric cellulose acetate (CA) membranes) by means of aqueous acetone solutions was studied by different techniques (permeations, FTIR spectroscopy and atomic force microscopy). It appears that the skin layer of the asymmetric membranes had pores of molecular size which make them much less selective in pervaporation (PV) than the dense membranes made of the same materials. A mechanism was proposed for the “defects” reparation of the membrane skin layer by this method of post-treatment. Under the action of a solvent medium that increases sufficiently the polymer chain mobility, the reparation of chains at the boundary of the polymer nodules in the active-layer seals the inter-nodule gaps (i.e. the defects). The active-layer healing is sufficient for the “tight” asymmetric RO membranes to be quite selective in PV.

Experimental study of the osmotic behaviour of reverse osmosis membranes for different NaCl solutions and hydrostatic pressure differences

The osmotic behaviour of five different present-day reverse osmosis membranes was studied, using different NaCl solutions (<6 wt.%) and applying different hydrostatic pressure differences (<5.2 bar) between the high and low concentration sides. The osmotic water flux through the membranes was found to increase non-linearly with respect to the increasing concentration difference over the membrane, leading to non-constant values for apparent water permeability. The polyamide composite membranes studied resulted only in very weak osmosis for even relatively small hydrostatic pressure differences and for strong NaCl solutions. Asymmetric cellulose acetate membranes showed somewhat better osmotic behaviour. However, all the osmotic water fluxes measured were still far too small for practical application in the production of energy.

Mass transfer dynamics of the evaporation step in membrane formation by phase inversion

AbstractA ternary diffusion model has been developed for the evaporation step of the phase inversion process. The model is applied to the analysis of mass transfer dynamics of the evaporation step for the methanol–acetone–cellulose acetate (CA) ternary casting system. The combined analysis of quantitatively computational results from the ternary evaporation model and qualitative dynamic results during the quench process has shown that the evaporation step is essentially necessary to prepare the defect‐free, ultrathin skinned asymmetric CA membrane for the separation of CO2/CH4. The skin layer of high CA concentration obtained by evaporation has an ability to suppress liquid–liquid phase separation. And the skin layer with high tensile strength can resist the interfacial tension caused by spinodal decomposition from the substructure. Although the CA concentration in the skin layer increases considerably because of the evaporation step and the following delay time during the quench process, the substructure can still induce the spinodal decomposition because the strong coagulant, methanol, can diffuse rapidly across the ultrathin skin layer. Hence the defect‐free, ultrathin‐skinned asymmetric membrane for gas separation can be prepared from methanol–acetone–CA casting system by evaporation step and the wet phase inversion. © 2002 Wiley Periodicals, Inc. J Appl Polym Sci 86: 1564–1571, 2002

Tailoring the Properties of Asymmetric Cellulose Acetate Membranes by Gas Plasma Etching

Cellulose triacetate (CTA) ultrafilters and cellulose acetate blend (CAB) desalination membranes were treated with a radiofrequency gas plasma (tetrafluoromethane (CF4) or carbon dioxide (CO2), 47–49 W, 0.04–0.08 mbar). Treatment times were varied between 15 s and 120 min. The plasma-treated top layer of the membranes was characterized by scanning electron microscopy, X-ray photoelectron spectroscopy, and contact angle measurements to obtain information about surface structure, chemistry, and wettability, respectively. The membrane properties (e.g., permeability, selectivity, fouling) were studied by waterflux measurements, molecular weight cutoff measurements, and fouling experiments with bovine serum albumin. CO2 plasma treatment resulted in gradual etching of the membrane's dense top layer. Permeation and selectivity changed significantly for treatment times of 0–15 min for CTA and 5–60 min for CAB membranes. Moreover, CTA membranes were hydrophilized during CO2 plasma treatment whereas CF4 plasma treatment led to hydrophobic surfaces due to strong fluorination of the top layer. This study shows that gas plasma etching can tailor the properties of asymmetric cellulose acetate membranes by simultaneously modifying the chemistry and structure of the top layer. The low fouling properties of CTA membranes were thereby largely maintained.

Separation of linear hydrocarbons and carboxylic acids from ethanol and hexane solutions by reverse osmosis

In this study, asymmetric cellulose acetate membranes with moderate NaCl rejection (85.5%) were prepared and used to study the influence of the chemical nature of organic solutes in different organic solvents. The solute rejection and the solvent flux of linear hydrocarbons ( M w=226–563 g/mol) and linear carboxylic acids ( M w=228–340 g/mol) in ethanol and hexane were studied as a function of the molecular weight, the feed concentration and the transmembrane pressure. The ethanol flux was three times higher than the hexane flux. The rejection coefficients for both types of solute were quire acceptable ( R=60–90%), when ethanol was the solvent. In hexane the linear hydrocarbons showed a rejection of 40–60%, while all carboxylic acids reached a negative rejection of −40 to −20%. This negative “observed” rejection can be attributed to accumulation of carboxylic acid at the membrane; the solute concentration at the membrane becomes much higher than in the bulk solution, due to a higher affinity of the solute with the membrane in hexane than in ethanol. Sorption experiments support this hypothesis. Furthermore, it was found that the rejection increases with increasing molecular weight and the rejection and flux are hardly affected by the feed concentration.

Effect of hydrolysis on porosity of cellulose acetate reverse osmosis membranes

AbstractThe alkaline hydrolysis of asymmetric cellulose acetate membranes was investigated. The changes of cellulose triacetate and cellulose diacetate membrane performances caused by hydrolysis of the membrane material were measured, and the data were analyzed in terms of pore size distributions and effective number of pores in the upper membrane layer before and after hydrolysis. The resulting membranes rejection factors decrease of 10–25%, and a noticeable permeation rates increase were caused by closing of a part of the smallest “network” membrane pores and by increasing the size of the other pores. © 2001 John Wiley &amp; Sons, Inc. J Appl Polym Sci 81: 1768–1775, 2001

Study of the structure of asymmetric cellulose acetate membranes for reverse osmosis using electron spin resonance (ESR) method

Electron spin resonance (ESR) technique was used to study the structure of the skin layer of asymmetric cellulose acetate membranes for reverse osmosis. TEMPO (2,2,6,6-tetramethyl-1-piperridinyloxy-free radical) was used as a spin probe that was brought into the membranes either by reverse osmosis experiments with feed solutions involving TEMPO, or by blending TEMPO into casting solutions. It was found that the mobility of TEMPO in the asymmetric membrane shrunk at 90°C was the same as TEMPO in a dense homogeneous membrane prepared from the same casting solution as that used for the preparation of the asymmetric membranes. On the other hand, TEMPO was more mobile in the asymmetric membranes when they were shrunk at lower temperatures and less mobile in a dense homogeneous membrane when the latter is prepared from a casting solution without pore former (magnesium perchlorate and water). Reverse osmosis experiments were also performed using feed aqueous solutions of sodium chloride and/or TEMPO.

Theoretical and Experimental Study of the Voltage–Current Characteristic of an Asymmetric Cellulose Acetate Membrane

The voltage–current characteristic of an asymmetric cellulose acetate (KP96) membrane in NaCl solutions (0.005–0.1 M) has been studied both experimentally and theoretically. The theoretical analysis was carried out within the context of a bilayer model, thereby taking into account ion transport across the interfaces of the active layer and the effect of concentration polarization at the selective layer/nonselective layer boundary. The transport process was treated mathematically by applying the Nernst–Planck–Poisson equations. Image forces affecting ion transport across the surfaces of the active layer have been incorporated into the theoretical description by means of a mathematical simulation of the profile of the image force potential. Most of the parameters characterizing the model membrane were estimated from literature data, from the known value of the final rejection coefficient of an asymmetric KP96 cellulose acetate membrane, and from our experimental results on the NaCl concentration dependency of the small-current resistance of an asymmetric KP96 membrane. The thickness of the active layer, however, was considered as an adjustable parameter in our analysis. The theoretical predictions, obtained by numerically solving the Nernst–Planck–Poisson equations, have been compared with the experimental results to obtain insight into the factors underlying the observed nonlinear and asymmetric features of the voltage–current characteristic. For the case of a relatively thick active layer, the nonequilibrium state of the interfaces was also investigated theoretically. Our analysis indicates that ion transport across the interfacial regions of the active layer plays a crucial role in the explanation of the observed phenomena, whereas the influence of concentration polarization in the porous support layer is of minor importance.

Atomic force microscopy of dense and asymmetric cellulose-based membranes

The surface structures of dense and integrally skinned cellulose acetate (CA) and cellulose acetate butyrate (CAB) membranes, prepared by phase inversion under different casting conditions, are investigated by tapping mode atomic force microscopy (TM AFM). The results obtained show that: (i) The top and bottom surfaces of the dense CA membrane were quite uniform in comparison with the corresponding faces of asymmetric CA and CAB membranes. Despite the casting conditions the active and support layers of the asymmetric membranes display large differences on the roughness parameters. (ii) The asymmetric membranes prepared with an organic system as a non-solvent pore-former (method IV) display smaller nodule aggregates and lower values of the roughness parameters than the ones prepared using an inorganic system as swelling agent (method I). This is more pronounced for the CA membranes than for the CAB membranes. (iii) In the active layer of asymmetric CA membranes casted at longer evaporation times, the measured values of surface roughness parameters tend to decrease. Also, for these CA membranes, as the evaporation time increases the average size of the depression areas observed on the surface decreases. The laboratory-made CA and CAB membranes display a wide range of nanofiltration and reverse osmosis permeation characteristics. These characteristics are correlated to surface roughness parameters of the active layers.

Effect of humic substances and anionic surfactants on the surface charge and performance of reverse osmosis membranes

Streaming potential measurements were used to determine the surface zeta potential of two reverse osmosis (RO) membranes and to investigate the effect of solution chemical composition on membrane surface charge. Performance measurements were performed to evaluate the effect of membrane surface chemistry on membrane flux and rejection. The RO membranes evaluated were a thin-film composite polyamide membrane and an asymmetric cellulose acetate membrane. The solution chemistries investigated include Suwannee River humic acid and sodium dodecyl sulfate (an anionic surfactant). Zeta potential measurements revealed that in the presence of an indifferent electrolyte, both membranes had a positive zeta potential in the low pH range, passed through an isoelectric point between pH 2 and 3, and had a negative zeta potential in the mid to high pH range. Suwannee River humic acid and sodium dodecyl sulfate were found to readily adsorb to the membrane surface and markedly influence the membrane surface charge. Suwannee River humic acid had a significant influence on salt rejection at low pH where adsorption of the organic macromolecules caused a change in sign of the zeta potential (from positive to negative) and therefore a change in co-ion exclusion effects. The effects of the sodium dodecyl sulfate, also more apparent at low pH, were attributed to the formation of hemimicelles which caused decreased flux and increased salt rejection.

Influence of quench medium on the structure and gas permeation properties of cellulose acetate membranes

Integrally skinned asymmetric cellulose acetate membranes made by the wet phase inversion for removal of CO2 from natural gas were investigated. The membrane was cast with the membrane-forming systems of cellulose acetate–acetone and quench media, such as methanol, ethanol and isopropanol, respectively, without heat-treating and multistage exchange process. By means of evaluation on separating characteristics of the membrane for CO2/CH4, observation of morphologies by scanning electron photomicrographs and analysis of the phase diagrams on the membrane-forming systems, it has shown that the membrane-forming system of cellulose acetate–acetone–methanol is quite suitable to prepare integrally skinned asymmetric cellulose acetate membranes for gas separation with good selectivity CO2/CH4 = 30 and flux coefficient = 2.4 × 10−5 cm3/cm2 − s − cm Hg. © 1998 John Wiley & Sons, Inc. J Appl Polym Sci 68: 1269–1276, 1998

Dielectric dispersion of a membrane consisting of a large number of solid bilayers containing mobile ions

The complex permittivity and the impedance of a lamellar membrane consisting of a repetition unit of two conducting dielectrics is calculated using the linearized Nernst–Planck–Maxwell electrodynamics previously developed for isolated interfaces, lamellar films with identical cation and anion mobility inside each type of lamella and with identical Nernst distribution coefficients and for homogeneous polymer films with ionic impurities. Any intrinsic dielectric dispersion of the two (ion-free) dielectrics is neglected and periodic boundary conditions are assumed in a infinite array of identical bilayers. The anion and the cation may be of any charge type and their diffusion coefficients may have any values in the two dielectrics. The Nernst distribution coefficient may be different for the cation and the anion within limits determined by the validity of the linearisation procedure. The calculated complex permittivities of the composite material may differ substantially from the permittivities calculated by a simple Maxwell–Wagner–Sillars (MWS) procedure because of the dynamical electrical double layers formed at 1–2 Debye lengths from the interfaces. These polarisation layers give rise to the so-called excess impedance. The MWS results are, however, always exact in the high-frequency limit. Furthermore, the MWS approach is correct for the dielectric loss at low frequencies. A frequency weighted ‘Cole–Cole plot’ of the real and imaginary components of the excess impedance has a characteristic ‘apple shape’ starting in the origin at very low frequencies and moving clockwise around the origin with increasing frequency. In the high-frequency limit, the origin is always approached in the fourth quadrant along a line with slope-1. The low-frequency asymptote of the frequency weighted Cole–Cole plot for the excess impedance is sensitive to the special conditions, however. The Cole–Cole plots of the total membrane impedance (MWS+excess) exhibit sometimes two partially overlapping semi-circular arcs, one for each type of lamella. When the thickness of the lamellae becomes of the same order of magnitude as the Debye length, the distinction between the two relaxations is wiped out and only a single, flattened arc is seen. A model lamellar ‘cellulose acetate’ (CA) membrane, consisting of alternating layers of ‘wet CA phase’ and ‘aqueous microphase’ with NaCl dissolved, is investigated. The parameters of the membrane are chosen quite realistically (except the one-dimensional geometry) as those pertaining to dense CA membranes, in accordance with previous emf and impedance spectroscopical studies of dense and asymmetric cellulose acetate membranes. The usual electrochemical ‘equivalent circuit’ approach, connecting in parallel the membrane resistance and the membrane capacitance, is put to a test. From the time constant related to the maximum negative reactance the so-called ‘geometric capacitance’ is calculated. From this, the ‘geometric’ membrane permittivity of the membrane is found. At low salt concentrations, the geometric permittivity coincides with the low-frequency limit of the membrane permittivity, but with increasing salt concentrations the geometric permittivity increases less rapidly with concentration than does the low-frequency membrane permittivity. The geometric membrane permittivity seems to be a complicated mean value between the low-frequency membrane permittivity and the high-frequency membrane permittivity (independent of concentration), since the characteristic frequency for the impedance relaxation is many orders of magnitude higher than the characteristic frequency of the dielectric relaxation. The former is related to the diffusional relaxation time of the CA layer, the second to the relaxation time in the aqueous microphase.

Diffusion coefficients of conductive ions in a copolymer of vinylidene cyanide and vinyl acetate obtained from dielectric measurements using the model of Trukhan

The dielectric dispersion measurements of Furukawa et al. are treated in the light of a previously proposed model of Trukhan. The latter describes the influence of mobile ions on the dielectric dispersion in a slab of material placed between polarizable electrodes. It is shown that the so-called ‘‘constant phase element’’ is just a crude approximation to the predictions of the theory of Trukhan, an approximation not valid at very low frequencies. At low frequencies macropolarizations appear analogous to the ones observed in asymmetric cellulose acetate membranes by Malmgren-Hansen et al. The polarizations are much larger in the present case, and this indicates that there are no microheterogeneities in the polymeric film of Furukawa et al. The diffusion coefficient of the most rapidly diffusing ion (presumably H+) may be found as a function of temperature within some uncertainty. The Arrhenius plot shows clearly the change in activation energy around the glass transition temperature (182 °C). Below the glass transition the activation energy for diffusion is much larger (∼50000 K) than above. The diffusion coefficients increase from ∼10−17 m2/s at 170 °C to ∼5.10−16 m2/s at 195 °C. The concentration of electrolyte present in the polymer increases from ∼15 to ∼200 mol m−3 in the same temperature interval.

Optical Polarizing Studies of Cellulose Acetate Membranes Prepared by Phase-Inversion

Abstract Anisotropic spherical cells dispersed in an isotropic medium were observed, using polarizing microscopy, in asymmetric cellulose acetate (CA) membranes. Membranes CA-400 and CA-316 were prepared from ternary casting solutions of CA-acetone-formamide and CA-acetone-magnesium perchlorate/water respectively. The content of nonsolvent, formamide, was varied to yield membranes CA-400-22, CA-400-27 and CA-400-32 that show a decreasing number of larger anisotropic cells with the increase of nonsolvent content. The membranes CA-316 were subjected to a heat treatment—annealing—at varying temperatures yielding membranes CA-316-50, CA-316-68 and CA-316-86. The crystallinity observed was dependent on the temperature of this heat post-treatment. The selective permeation properties are correlated with the ratio of isotropic to anisotropic phases at the membrane skin surface. This ratio is in turn dependent on the casting solution composition-type and concentration of nonsolvent and on the annealing treatment. ...

Field tests of membrane modules for the separation of carbon dioxide from low-quality natural gas

A commercial-scale, single-stage, spiral-wound membrane system has been operated for approximately 20 months to upgrade low-quality natural gas from a well in East Texas. Throughout the test period the retentate product (“sales” gas) met pipeline specifications. Data were obtained on two membrane modules containing two types of asymmetric cellulose acetate membranes, one “standard” and the other one of higher density. A summary of the field test data shows the effects of the operating variables of pressure, feed flow rate, and CO2 concentration in the feed (from 3 to 25 mole percent). Concentrations greater than 6 mole percent were obtained by adding pure CO2 to the feed gas. In addition, computer models for the separation of gases under “perfect mixing” and cross-flow conditions were applied to the analysis of the field data. In general, the field test data were consistent with a flow regime which was intermediate between perfect mixing and cross-flow.