Andrea I. Schäfer

Ultrafiltration of natural organic matter

Increasingly stringent regulations for drinking water quality have stimulated the application of ultrafiltration to water treatment. In addition to removing particulate materials from water (including microorganisms, bacteria and viruses), the use of membrane treatment also meets purification requirements. However, irreversible fouling curtails the economic viability of such a process. Experiments in stirred-cells were conducted to evaluate the effects of surface water composition on rejection and fouling of two ultrafiltration membranes with different molecular weight cut-offs (10 and 100 kDa). Experimental solutions consisted of natural organic matter or humic substances in a background electrolyte. The effect of calcium concentration decreased rejection of humic acid under certain circumstances. This is believed due to reduced molecular size with an initial increase in calcium concentration. However, at about 2.5 mM CaCl2, IHSS humic acid aggregates. This aggregation increased rejection, and also caused irreversible fouling of the 100 kDa membrane, presumably as a result of pore size reduction due to internal deposition of aggregates. This was confirmed by blocking law analysis. The variation of transmembrane pressure indicated the importance of a ‘critical flux’ effect. The organics and their various fractions showed differences both in rejection and flux decline. The larger and more UV-absorbing fraction of humic acid was shown to be responsible for irreversible pore adsorption and plugging. The fulvic acid and the hydrophilic fraction showed a smaller and mostly reversible flux decline.

Nanofiltration of Natural Organic Matter: Removal, Fouling and the Influence of Multivalent Ions

The presence of calcium and humic substances or natural organic matter (NOM) in surface waters can cause severe fouling of nanofiltration (NF) membranes. Conditions of fouling were studied using a stainless steel stirred cell of volume 185 ml and a membrane area of 21.2×10−4 m2 at a transmembrane pressure of 5 bar. Deposition of organic matter was determined by mass balance in feed and concentrate samples. Electron microscopy and X-ray photoelectron spectroscopy (XPS) were used to study the morphology and composition of the fouling layer. During permeate recycle experiments, which were used for fouling studies, it was found that calcium concentration (as a representative of multivalent ions) and the type of organic play a major role in fouling. The calcium forms complexes with some of the organics and deposits on the membrane surface. Depending on the solution conditions the organic or calcite (on which organics adsorb) precipitate. Factors, which influence the concentration of organics and ions at the membrane surface such as stirring, flux and transmembrane pressure, influenced the deposition of organic matter significantly. Irreversible fouling occurred with all membranes at high calcium concentrations, although the cellulose acetate membrane showed an overall better performance, possibly due to its low salt rejection and smooth surface. IHSS humic acid is the organic which deposits most easily and comparison of UV absorbance and DOC data showed that the fraction which absorbs UV strongest, and is more hydrophobic, deposits preferentially on the membranes. These substances also have the lowest solubility stressing the importance of concentration polarisation effects.

FOULING EFFECTS ON REJECTION IN THE MEMBRANE FILTRATION OF NATURAL WATERS

Membrane processes in drinking water applications are micro- (MF), ultra- (UF) and nanofiltration (NF). These processes remove turbidity and bacteria (MF), viruses and macromolecules (UF) and small molecules and hardness (NF). Of particular concern in water treatment is the removal of natural organic matter (NOM) which contains potential disinfection by-product precursors. The presence of colloids, multivalent ions and organics in surface waters may cause substantial fouling of membranes. A study was carried out which looked at the rejection abilities of a range of membranes targeting hematite colloids (40–500 nm), NOM and cations, fouling conditions and cost of treatment of these processes with consideration of chemical pretreatment with ferric chloride [1]. In this paper the effect of membrane fouling on rejection is presented. The study was based on experiments with two MF membranes (GVWP, GVHP, 0.22 μm, Millipore), six UF membranes (1, 3, 5, 10, 30, 100 kDa, regenerated cellulose, Millipore), and four organic NF membranes (TFC-SR, TFC-S, TFC-ULP, CA-UF, Fluid Systems, US). Three different types of organics (IHSS humic acid, IHSS fulvic acid and an Australian concentrated NOM) in a carbonate buffer containing calcium chloride and a background electrolyte were used. Experiments were carried out in perspex (MF, UF) and stainless steel (NF) stirred cells of a volume of 110–185 mL and a membrane area of 15.2–21.2×10−4 m2 at transmembrane pressures of 1, 1–3, and 5 bar for MF, UF, and NF, respectively. UF removes 10–95% of NOM depending on the molecular weight cut-off (MWCO) of the membrane. Pore sizes of <6 nm are required to remove about 80% of NOM, where a 6 nm pore size corresponds to a MWCO of about 10 kDa. Colloids are fully rejected. NF removes NOM effectively (70–95% as dissolved organic carbon (DOC) and 85–98% as UV absorbance). Cation rejection is very membrane dependent and varies for the investigated membrane types between 13 and 96% for calcium and 10–87% for sodium. Fouling was also dependent on pore size and was caused by large colloids (250 nm) or coagulant flocs in MF, small colloids, organic-calcium flocs and aggregates with a dense structure (formed slowly) in UF, and by a calcium-organic precipitate in NF. The fouling influenced the rejection of colloids in MF and that of NOM in UF and NF. If a highly charged layer was deposited on the NF membranes, cation rejection was also influenced. The characterisation of permeate organics revealed that low molecular weight acids passed through the NF membranes and that the rejection of these acids was also dependent on the deposit on the membrane. The mechanisms which can explain such an increase in rejection are different for the three membrane processes. In MF, pore plugging and cake formation was found responsible for fouling. This reduces the pore size and increases rejection. In UF, internal pore adsorption of calcium-organic flocs reduces the internal pore diameter and subsequently increases rejection. In NF, the key factor appears to the charge of the deposit. This was investigated with the deposition of a ferric chloride precipitate. If the precipitate was of high positive charge, the rejection of cations increased and that of negatively charged low molecular weight acids decreased compared to more neutral or negative precipitates. In essence, the rejection characteristics of membranes depend more on the fouling state of the membranes and the nature of the foulants than on the initial membrane characteristics.

Cost Factors and Chemical Pretreatment Effects in the Membrane Filtration of Waters containing Natural Organic Matter

This paper compares the membrane processes available for water treatment. Membranes have the advantage of currently decreasing capital cost, a relatively small footprint compared to conventional treatment, generally a reduction in chemicals usage and comparably low maintenance requirements. Three membrane processes applicable to water treatment, micro- (MF), ultra- (UF), and nanofiltration (NF), are compared in terms of intrinsic rejection, variation of rejection due to membrane fouling and increase in rejection by ferric chloride pretreatment. Twelve different membranes are compared on the basis of their membrane pore size which was calculated from their molecular weight cut-off. A pore size of < 6 nm is required to achieve substantial (> 50%) organics removal. For a fouled membrane this pore size is about 11 nm. UV rejection is higher than DOC rejection. Coagulation pretreatment allows a higher rejection of organics by MF and UF and the cut-off criterion due to initial membrane pore size is no longer valid. A water quality parameter (WQP) is introduced which describes the product water quality achieved as a function of colloid, DOC and cation rejection. The relationship between log (pore size) and WQP is linear. Estimation of membrane costs as a function of WQP suggests that open UF is superior to MF (similar cost at higher WQP) and NF is superior to tight UF. Chemical pretreatment could compensate for the difference between MF and UF. However, when considering chemicals and energy costs, it appears that a process operated at a higher energy is cheaper at a guaranteed product quality (less dependent on organic type). This argument is further supported by environmental issues of chemicals usage, as energy may be provided from renewable sources.

Micropollutant sorption to membrane polymers: a review of mechanisms for estrogens.

Organic micropollutants such as estrogens occur in water in increasing quantities from predominantly anthropogenic sources. In water such micropollutants partition not only to surfaces such as membrane polymers but also to any other natural or treatment related surfaces. Such interactions are often observed as sorption in treatment processes and this phenomenon is exploited in activated carbon filtration, for example. Sorption is important for polymeric materials and this is used for the concentration of such micropollutants for analytical purposes in solid phase extraction. In membrane filtration the mechanism of micropollutant sorption is a relatively new discovery that was facilitated through new analytical techniques. This sorption plays an important role in micropollutant retention by membranes although mechanisms of interaction are to date not understood. This review is focused on sorption of estrogens on polymeric surfaces, specifically membrane polymers. Such sorption has been observed to a large extent with values of up to 1.2 ng/cm(2) measured. Sorption is dependent on the type of polymer, micropollutant characteristics, solution chemistry, membrane operating conditions as well as membrane morphology. Likely contributors to sorption are the surface roughness as well as the microporosity of such polymers. While retention-and/or reflection coefficient as well as solute to effective pore size ratio-controls the access of such micropollutants to the inner surface, pore size, porosity and thickness as well as morphology or shape of inner voids determines the available area for sorption. The interaction mechanisms are governed, most likely, by hydrophobic as well as solvation effects and interplay of molecular and supramolecular interactions such as hydrogen bonding, π-cation/anion interactions, π-π stacking, ion-dipole and dipole-dipole interactions, the extent of which is naturally dependent on micropollutant and polymer characteristics. Systematic investigations are required to identify and quantify both relative contributions and strength of such interactions and develop suitable surface characterisation tools. This is a difficult endeavour given the complexity of systems, the possibility of several interactions taking place simultaneously and the generally weaker forces involved.

Adsorption and Transport of Trace Contaminant Estrone in NF/RO Membranes

Despite of their low concentration, the impact of steroid estrogens such as estrone, 17 α-estradiol, and ethinylestradiol are often more serious than other synthetic endocrine-disrupting chemicals ...

Adsorptive interactions between membranes and trace contaminants

Widespread occurrences of endocrine-disrupting chemicals (EDCs) in waterways have attracted a great attention of the scientific community. While scientific evidence associated with human health is restricted due to the long-term effects, impacts of EDCs on trout at the common concentration encountered in sewage effluent have been confirmed by both in vitro and in vivo studies. The impacts of steroid estrogens such as estrone, estradiol (natural hormones) and ethinylestradiol (a synthetic hormone) are often more serious than other synthetic EDCs as they have far higher endocrine-disrupting potency, despite of their low concentration. This paper investigates retention and adsorptive behavior of the natural hormones estrone by two commercial reverse osmosis membranes TFC-S and X-20, using dead end stirred cell systems. While an adsorptive process that reaches a breakthrough governs the retention of estrone by the TFC-S membrane; a sieving mechanism is responsible for the high removal of estrone using the X-20 membrane.

Nanofiltration of Hormone Mimicking Trace Organic Contaminants

Abstract The removal mechanisms of three hormone mimicking organic compounds by nanofiltration (NF) membranes have been examined. Two NF membranes having different pore sizes were used in laboratory‐scale nanofiltration experiments with feed solutions spiked with a hormone mimicking compound—nonylphenol, tert‐butylphenol, or bisphenol A. Retention of the compounds was determined at various solution chemistries, namely aqueous solution pH, ionic strength, and presence of natural organic matter. The nanofiltration behavior of the selected hormone mimicking compounds appears similar to that of natural hormones as reported in our previous work. While the solution pH can dramatically influence the retention of hormone mimicking compounds by a loose NF membrane, ionic strength does not affect the nanofiltration of such contaminants. However, in the presence of natural organic matter in the feed solution, ionic strength appears to play a significant role in solute‐solute and solute‐membrane interactions, resulting in increased retention due to partitioning of the hormone mimicking compounds onto organic matter at a higher ionic strength.

The Importance of Dehydration in Determining Ion Transport in Narrow Pores

The transport of hydrated ions through narrow pores is important for a number of processes such as the desalination and filtration of water and the conductance of ions through biological channels. Here, molecular dynamics simulations are used to systematically examine the transport of anionic drinking water contaminants (fluoride, chloride, nitrate, and nitrite) through pores ranging in effective radius from 2.8 to 6.5 Å to elucidate the role of hydration in excluding these species during nanofiltration. Bulk hydration properties (hydrated size and coordination number) are determined for comparison with the situations inside the pores. Free energy profiles for ion transport through the pores show energy barriers depend on pore size, ion type, and membrane surface charge and that the selectivity sequence can change depending on the pore size. Ion coordination numbers along the trajectory showed that partial dehydration of the transported ion is the main contribution to the energy barriers. Ion transport is greatly hindered when the effective pore radius is smaller than the hydrated radius, as the ion has to lose some associated water molecules to enter the pore. Small energy barriers are still observed when pore sizes are larger than the hydrated radius due to re-orientation of the hydration shell or the loss of more distant water. These results demonstrate the importance of ion dehydration in transport through narrow pores, which increases the current level of mechanistic understanding of membrane-based desalination and transport in biological channels.

pH dependence of steroid hormone—organic matter interactions at environmental concentrations

The interaction of estradiol, estrone, progesterone and testosterone with environmentally relevant concentrations of Aldrich humic acid, alginic acid and tannic acid was studied using solid-phase microextraction (SPME). Since bulk organic matter and certain hormones such as estradiol and estrone contain dissociable functional groups, the effect of pH on sorption was investigated as this will influence their fate and bioavailability. For humic acid and tannic acid, sorption was strongest at acidic pH when the bulk organic matter was in a non-dissociated form and decreased when they became partially negatively charged. At acidic and neutral pH the strength of partitioning was influenced by hormone functional groups content, with the strongest sorption observed for progesterone and estrone. At alkaline pH conditions, when the bulk organics were dissociated, sorption decreased considerably (up to a factor of 14), although the non-dissociated hormones testosterone and progesterone indicated greater sorption to humic acid at pH 10 compared to the partially deprotonated estradiol and estrone. This study demonstrates that SPME can be used to assess organic matter sorption behaviour of a selected range of micropollutants and at environmentally relevant organic matter concentrations.