Kyu-Jin Kim

The performance of ultrafiltration membranes pretreated by polymers

Abstract The fouling of ultrafiltration(UF) membranes by protein has been reduced by surface pretreatment with various polymers. In most cases the treatment provided an increase in initial UF flux and a slower flux decline. The ultrafiltration performance of the treated membranes appears to be influenced by the molecular configuration and size of the polymer as well as its interfacial structure. Methylcellulose(MC) was found to be the most effective of polymers tested in enhancing UF flux, showing an average flux advantage of 30–40% for first usage. For multiple cycles of usage the benefits increased up to 100% after 5 cycles, due to the reduced amount and reversibility of the fouling protein deposit.

Evaluation of electroosmosis and streaming potential for measurement of electric charges of polymeric membranes

Abstract Electrical properties for a range of commercial membranes have been determined by electroosmosis and streaming potential measurements under identical conditions. For both techniques, the pH range was 3–7 and the electrolyte used was 10 −3 M KCl at 25°C, as the charges on the membrane are strongly dependent upon the pH and the ionic strength of the solution passing through the membrane. For both ultrafiltration (UF) and microfiltration (MF) membranes the absolute values of the apparent zeta potential determined from electroosmosis were generally greater than those from streaming potential measurements with a greater difference at pH ≤ 4. The apparent zeta potentials obtained from electroosmosis were negative for all membranes studied, even at pH below the isoelectric point where streaming potentials were positive. A substantial increase of the negative zeta potential at low pH (∼ pH 4) was interpreted in terms of the effect of the applied electric field on distribution of ions and differences in mobility of ions, particularly H + compared to the other ions. The magnitude of the zeta potentials determined from electroosmosis exhibited an apparent difference due to change in orientation of the membrane for asymmetric UF membranes, whereas the streaming potential measurements showed less directional-effect for both UF and MF membranes.

Effect of calcium ion on the fouling of nanofilter by humic acid in drinking water production

Abstract The influence of the charge of humic acid (HA) on the fouling of nanofilters was investigated as a function of pH and concentration of calcium ions. The charge of the humic acid as well as the zeta-potential of the membrane surface were measured to elucidate the mechanism of humic acid deposition on the membrane surface. The negative charge of humic acid as well as the negative zeta-potential of the membrane surface fouled with humic acid increased with higher pH. As a result the further deposition of humic acid on the membrane surface was expected to decrease with higher pH because of larger repulsive forces. With moderate calcium ion concentration, however, the adsorption of humic acid onto the membrane surface decreased until neutral pH and then increased again. It was attributed to calcium ion bridging between two free functional groups of humic acids. Calcium ion rejection decreased in the presence of humic acid deposited on membrane surface. This could be explained by the effect of humic acid on the net charge of the membrane. The addition of calcium chelating agent, EDTA, improved flux, especially at alkaline conditions. The diagram of Ca–EDTA vs pH was used to explain the flux improvement.

Characteristics of microfiltration membranes in a membrane coupled sequencing batch reactor system

Factors affecting filtration performance were investigated in a sequencing batch reactor (SBR) coupled with a submerged microfiltration module. Special bioreactors for aerobic and anoxic phases were specifically designed in order to differentiate the effect of dissolved oxygen (DO) from that of mixing intensity, on membrane filterability. When the filterability of a submerged microfilter was examined at each SBR phase, DO concentration, as well as mixing intensity proved to have a major influence on the membrane performance regardless of the SBR phase. A higher DO concentration resulted in a slower rise in TMP, corresponding to less membrane fouling, which was investigated in depth through a series of analyses including resistance measurements and compressibility of the cake layer as well as particle sizes as a functions of DO for both aerobic and anoxic phases in SBR.

Quantitative microscopic study of surface characteristics of ultrafiltration membranes

This paper presents the first high-resolution field emission scanning electron micrographs of the top skin layer of various ultrafiltration membranes at a low accelerating voltage. It was found that the fine structure on the top surface of PTHK, SKIP, ETNA 20A and MPS membranes was network-like. All the membrane surfaces observed were inhomogeneous, with wide variations in the size and shape of the pores, in surface roughness, and in appearance. The surface pore characteristics of the membranes have been determined quantitatively from the micrographs using image analysis. In most cases the pore density ranged from ca. 1010–1011 pores/cm2 and the surface porosity was less than 10%. The pore spread factor (diameter for 20% greater than/diameter for 80% greater than) ranged from 2.2–6.6. The present results from field emission scanning electron micrographs compare reasonably with those obtained from transmission electron micrographs.

A comparative study of techniques used for porous membrane characterization: pore characterization

A range of commerical UF membranes have been characterized by thermoporometry, biliquid permporometry and molecular weight cut-off experiments. A comparison of results from these three independent techniques for the same types of membrane shows an indication of the strength and weakness of the methods. MWCO values determined from actual rejection values using PEG and dextran were significantly lower than the manufacturer supplied data. The data obtained using the biliquid permporometry and solute rejection tests produced contrasting results for Amicon polysulfone (PM30) and regenerated cellulose (YM30) membranes. While MWCO determination resulted in sharper cut-off curves, the biliquid permporometry offered a broader size distribution with the PM30 and vice versa with the YM30. The pore sizes obtained by thermoporometry were significantly larger than those by the biliquid permporometry. The biliquid permporometry and thermoporometry give significantly higher values than the MWCO method. The closest comparison is obtained between the EM values and the MWCO method. This suggests that the controlling pore dimension for separation is the surface skin dimension.

The cleaning of ultrafiltration membranes fouled by protein

The relationship between membrane fouling and cleaning was investigated in terms of flow conditions, transmembrane pressures, pH, membrane properties, and cleaning agents using a stirred batch cell and aqueous albumin solution. Fouling was less at the pH extremes than at the isoelectric point for both retentive and partially permeable membranes. Membranes with partial permeability showed a greater tendency for fouling and were less responsive to cleaning than retentive membranes. The results in the stirred cell were shown to be similar to those for a crossflow module under similar operating conditions.

Characterisation of ultrafiltration membranes by impedance spectroscopy. I. Determination of the separate electrical parameters and porosity of the skin and sublayers

The response of synthetic polymer membranes to alternating electric fields has been studied. From the experimental results presented for two types of polysulphone membranes, a determination was made of the individual capacitances and conductances of both the skin layer and porous sublayer. The capacitance of the skin layer was found to be 0.2 mF/m2 in 0.2 mM KCl solution for both PM30 and PTTK membranes. The conductance of the skin layer was 4–6 S/m2 and 5–9 S/m2 for the PM30 and PTTK membranes respectively. For both membranes a porosity of 2–5% for the skin layer was obtained assuming a skin layer thickness of 0.2–0.3 μm.

Chemical and electrical characterization of virgin and protein-fouled polycarbonate track-etched membranes by FTIR and streaming-potential measurements

Abstract Electrical properties for a range of virgin and protein-fouled polycarbonate track-etched (PCTE) membranes have been determined by streaming-potential measurements at pH 4–7 and 10 −3 M KCl at 25°C. The apparent zeta potentials generally increased with increasing pore size above pH 5 and decreased with pore size below pH 5. For membranes with smaller pores (PCTE 0.01, 0.03 and 0.05 μm) the zeta potentials became constant (the surface is fully charged) at about −6 mV at pH ≥ 5, whereas those of larger pores (PCTE 0.1 and 0.2 μm) showed constant zeta potentials of ca. −10 to −13 mV at pH > 6. In contrast, the protein-fouled membranes showed very similar apparent zeta potential vs. pH profiles for all pore sizes. This suggests that the measured zeta potentials are determined by the protein properties. The loss of strong pore-size dependence on zeta potentials upon protein fouling of membranes implies that membrane zeta potentials could be controlled by the conditions at pore entry. This ‘model’ requires protein deposition around the pore openings for all pore sizes; this was in agreement with electron microscope observations. While the isoelectric points (IEP) of the virgin membranes determined from the pH dependence of streaming-potential measurements were between pH 3.7 – 4.4 with a trend to higher IEP for the larger pore membrane, all protein-fouled membranes showed IEP values at pH 4.6 – 4.8, which is close to the IEP of the BSA protein used. The chemical properties of the PCTE membranes investigated had FTIR spectra similar to those of polycarbonate from bis(4-hydroxyphenyl)-2-propane, bisphenol A polycarbonate. No BSA peak was revealed in the FTIR spectra of the small-pore membranes fouled with protein, possibly due to the insufficient instrument detection limit. For the membranes with larger pores, however, the BSA peaks increased with increasing pore size. The depth profiling of the fouled membranes indicated a predominant BSA adsorption on the surface or near the pore entries rather than throughout the membrane thickness, confirming the zeta potential ‘model’ and electron microscope observations.

The effect of Langmuir-Blodgett layer pretreatment on the performance of ultrafiltration membranes

Abstract Ultrafiltration membranes pretreated with non-ionic surfactant monolayers using the Langmuir-Blodgett (LB) technique exhibited flux improvement for the stirred ultrafiltration (UF) of albumin proteins. Typical flux enhancement of 30% relative to untreated membranes was obtained for the first usage. Surfactant coating resulted in a higher initial UF flux and reduced flux decline and protein deposition. The hydrophilicity/hydrophobicity of the modified surface appeared to play a minor role in the fouling process. This effect was confirmed in separate experiments with strongly hydrophobic LB-layers of stearic acid. Results are presented for the effects of LB-layer deposition conditions, surfactant type and number of layers on flux increase.