Greg Leslie

Cake resistance and solute rejection in bacterial microfiltration: The role of the extracellular matrix

Abstract No common mechanism for flux limitation in bacterial crossflow microfiltration has as yet been identified in spite of the many theoretical hypotheses and experimental investigations reported in the literature. Current models are based on questionable assumptions about cell compressibility and rely primarily on particle bed pressure drop analysis. The influence of cell surface properties on microfiltration has been largely ignored. The purpose of this study was to investigate the role of the bacterial extracellular matrix in resistance and rejection. The gram-negative marine bacterium SW8 was grown in chemostats and characterised for particle size, surface hydrophobicity, dry weight concentration and with transmission and scanning electron microscopy. Modification of the extracellular matrix was carried out with a proteolytic enzyme (pronase) and a chelating agent (ethylenediaminetetraacetic acid). Controlled cake heights of the bacteria were deposited on Anopore 0.2 μm, Millipore 0.22 μm and Nuclepore 0.2 μm membranes at 100 kPa in batch filtration cells. Significant differences were observed in cake resistance and solute rejection (of the protein bovine serum albumin) independently of cell size and cake height for bacteria with modified cell surfaces. Furthermore, deposition of just one layer of cells resulted in fluxes as low as 5 l-m−2-hr−1 and protein rejections of 70%. The mechanism by which cake resistance and solute rejection occurs was shown to depend on the extracellular matrix of the organisms. The implications for future studies on crossflow microfiltration of biomass and biofluids are significant.

Mixing characterisation of full-scale membrane bioreactors: CFD modelling with experimental validation

Membrane Bioreactors (MBRs) have been successfully used in aerobic biological wastewater treatment to solve the perennial problem of effective solids-liquid separation. The optimisation of MBRs requires knowledge of the membrane fouling, biokinetics and mixing. However, research has mainly concentrated on the fouling and biokinetics (Ng and Kim, 2007). Current methods of design for a desired flow regime within MBRs are largely based on assumptions (e.g. complete mixing of tanks) and empirical techniques (e.g. specific mixing energy). However, it is difficult to predict how sludge rheology and vessel design in full-scale installations affects hydrodynamics, hence overall performance. Computational Fluid Dynamics (CFD) provides a method for prediction of how vessel features and mixing energy usage affect the hydrodynamics. In this study, a CFD model was developed which accounts for aeration, sludge rheology and geometry (i.e. bioreactor and membrane module). This MBR CFD model was then applied to two full-scale MBRs and was successfully validated against experimental results. The effect of sludge settling and rheology was found to have a minimal impact on the bulk mixing (i.e. the residence time distribution).

Comparison of treatment options for removal of recalcitrant dissolved organic matter from paper mill effluent.

Recycling paper mill effluent by conventional water treatment is difficult due to the persistence of salt and recalcitrant organics. Elimination of dissolved organic matter (DOM) from paper mill effluent was studied using three treatment options, ion exchange resin (IER), granular activated carbon (GAC) and nanofiltration (NF). The removal efficiency was analysed based on hydrophobicity, molecular weight and fluorogenic origin of the DOM fractions. For IER, GAC and NF treatments, overall removal of dissolved organic carbon was 72%, 76% and 91%, respectively. Based on the hydrophobicity, all the three treatment methods majorly removed hydrophobic acid fractions (HPhoA). Further, IER acted on all fractions, 57% of HPhoA, 44% of transphilic acid and 18% of hydrophilics, substantiating that the removal is by both ion exchange and adsorption. Based on the molecular weight, IER and GAC treatments acted majorly on the high molecular weight fractions, whereas NF eliminated all molecular weight fractions. After GAC adsorption, some amount of humic hydrolysates and low molecular weight neutrals persisted in the effluent. After IER treatment, amount of low molecular weight compounds increased due to resin leaching. Qualitative analysis of fluorescence excitation emission matrices showed that the fulvic acid-like fluorophores were more recalcitrant among the various DOM fractions, considerable amount persisted after all the three treatment methods. Three treatment methods considerably differed in terms of removing different DOM fractions; however, a broad-spectrum process like NF would be needed to achieve the maximum elimination.

Removal Efficiency and Integrity Monitoring Techniques for Virus Removal by Membrane Processes

Membranes are a mechanical form of disinfection that works by physical separation of the target pathogen. In theory, an intact membrane is a barrier to pathogens that are larger in size than the pore size of the membrane. However, in practice the pore size distribution or molecular weight cutoff will provide an indication of the separation efficiency and in many cases complete rejection of the virus is not demonstrated by various membrane processes. The general mechanism of pathogen removal by membrane processes is predominantly achieved by size exclusion, influenced by the physicochemical properties of the membrane, surface properties of the pathogens, and the solution environment. However, in order to evaluate the intrinsic performance of the membranes for a specific objective, a standard protocol with defined set of operating conditions is required. Any anomaly with the membrane surface (e.g., abnormally big pores, compromised glue line, holes) and the filtration system (e.g., compromised O-rings, broken mechanical seals) will result in microbial contamination risk of the product water. Therefore, monitoring membrane treatment system integrity is essential for the protection of public health from microbial risk. The demand for real-time monitoring is increasing as the usage of membrane-processed water is on the rise. Monitoring membrane integrity can be achieved by regular noncontinuous direct integrity testing performed on the membrane itself, or continuous indirect tests carried out on the filtrate. These monitoring techniques provide ongoing assurance that the water quality objectives are continually met. Existing integrity monitoring techniques have been shown in various instances to be reliable and sensitive for detecting contaminants of 1.0 μm and larger, while enteric virus particles are in the size range 0.01–0.04 μm. Some of the indirect monitoring techniques are capable of detecting a membrane breach less than 1 μm, but the practical challenge resolution is typically around 1–2-log. Similarly, the process control monitoring parameters lack sensitivity and therefore virus breakthrough occurs even before a loss of integrity is detected. At present it is difficult to demonstrate virus removal without conducting challenge tests due to a lack of sufficiently sensitive process monitoring. On the other hand, regulators are under increasing pressure to grant log removal credits for enteric viruses to membrane treatment processes due to the inability of present integrity tests to verify ongoing removal of virus-sized particles. Therefore, finding reliable and sensitive integrity tests for viruses is a high-priority issue in the water treatment sector. The authors aim to review the present state of knowledge on the process efficiencies of virus removal by membrane processes and the associated membrane integrity monitoring practices for virus-sized particles. Some of the knowledge gaps and research needs associated with the existing guidelines and standing protocols for integrity monitoring are also identified.

Fouling of a microfiltration membrane by two Gram-negative bacteria

Abstract In this study a method is described to assess fouling of microfiltration membranes based on aggregation and redispersion of bacteria from filter cakes. Fouling by the hydrophobic Gram-negative marine bacterium SW8, grown in continuous culture under conditions of carbon- and nitrogen-limitation was compared to fouling by the hydrophilic, batch grown Escherichia coli. Effects of ionic strength and dissolved protein on cake stability were investigated. Cake redispersion was found to depend on the solution ionic strength and on the presence of dissolved proteins. Redispersion was also found to depend on bacterial physiology. However, it was not possible to discriminate between carbon- and nitrogen-limited bacteria based on measurements of surface charge and cell surface hydrophobicity. This work shows that studies which rely on bacterial surface charge and cell surface hydrophobicity to describe membrane surface fouling overlook bacterial surface interactions which control cake stability.

Microfiltration of biomass and biofluids: Effects of membrane morphology and operating conditions

Abstract Fundamental studies carried out at the Centre for Membrane and Separation Technology (Australia) considered the effect of membrane surface properties and ionic strength using biomass microfiltration on controlled suspensions of Escherichia Coli . Anopore, Nuclepore and Millipore (hydrophilic and hydrophobic) membranes of similar pore size (0.2 μm) were characterised and subjected to flux tests with E.Coli suspension at different ionic strengths and transmembrane pressures. Membrane recovery after shear was also monitored and analysis using Scanning Electron Microscopy was performed on the cake formation. It was found that both ionic strength and membrane surface properties affected flux significantly and influenced recovery. The more porous and smooth hydrophilic Anopore membrane returned higher fluxes than the less porous Nuclepore and the rough Millipore membranes. High ionic strength solutions resulted in lower fluxes which can be explained using the colloid double-layer theory. Membrane recovery was easier for low pressure operation and low ionic strength suspensions.

Numerical simulation of bubble induced shear in membrane bioreactors: Effects of mixed liquor rheology and membrane configuration

A CFD model, incorporating an empirically determined rheology model and a porous media model, was developed to simulate bubble induced surface shear in membrane bioreactors configured with hollow fibre membranes with outer diameters ranging from 1.3 to 2.4 mm, arranged in vertically orientated modules with packing density from 200 to 560 m(2)/m(3). The rheology model was developed for mixed liquor suspended solids (MLSS) concentrations of 3 to 16 gL(-1) in the presence and absence of coagulant (generated by addition of a ferrous salt) for shear rates ranging from 0 to 500 s(-1). Experimentally determined particle relaxation times for the biological flocs in the mixed liquor, both in the absence and presence of iron, were negligible, consistent with an environment where positive buoyancy forces were greater than negative settling forces thereby allowing the sludge mixture to be modelled as a single continuous phase. The non-Newtonian behaviour of the mixed liquor was incorporated into the CFD simulations using an Ostwald-de Waele rheology model. Interactions between mixed liquor and hollow fibre membranes of different fibre size and packing density were described using a porous media model that was calibrated by empirical measurement of inertial loss coefficients over a range of viscosities (0.8 × 10(-3) to 2.1 × 10(-3) Pa.s) and velocities (0 to 0.35 m/s) typically encountered in full scale MBRs. Experimental results indicated that addition of iron salts resulted in an increase in MLSS and sludge viscosity. Shear stress is affected by both velocity and viscosity. The increase in sludge viscosity resulted in an increase in resistance to flow through the hollow fibre membrane bundles and, as a result, decreased the liquid flow velocities. CFD simulations provided insight on the effects of point of coagulant addition and MLSS concentration on bubble-induced shear over a range of industrially relevant conditions. A 12% increase in shear stress was observed when ferrous salts were added to the membrane filtration zone compared to addition to the primary anoxic zone. The presence of iron salts also improved the distribution of shear stress especially at the lower zone of the membrane module. The CFD models developed here were validated using Particle Image Velocimetry (PIV) with the average difference between simulated liquid velocities and PIV measured velocities found to be 5.5%.

Diagnosis of dissolved organic matter removal by GAC treatment in biologically treated papermill effluents using advanced organic characterisation techniques

Granular activated carbon (GAC) exhaustion rates on pulp and paper effluent from South East Australia were found to be a factor of three higher (3.62cf. 1.47kgm(-3)) on Kraft mills compared to mills using Thermomechanical pulping supplemented by Recycled Fibre (TMP/RCF). Biological waste treatment at both mills resulted in a final effluent COD of 240mgL(-1). The dissolved organic carbon (DOC) was only 1.2 times higher in the Kraft effluent (70 vs. 58mgL(-1)), however, GAC treatment of Kraft and TMP/RCF effluent was largely different on the DOC persisted after biological treatment. The molecular mass (636 vs. 534gmol(-1)) and aromaticity (5.35 vs. 4.67Lmg(-1)m(-1)) of humic substances (HS) were slightly higher in the Kraft effluent. The HS aromaticity was decreased by a factor of 1.0Lmg(-1)m(-1) in both Kraft and TMP/RCF effluent. The molecular mass of the Kraft effluent increased by 50gmol(-1) while the molecular mass of the TMP/RCF effluent was essentially unchanged after GAC treatment; the DOC removal efficiency of the GAC on Kraft effluent was biased towards the low molecular weight humic compounds. The rapid adsorption of this fraction, coupled with the slightly higher aromaticity of the humic components resulted in early breakthrough on the Kraft effluent. Fluorescence excitation-emission matrix analysis of the each GAC treated effluent indicated that the refractory components were higher molecular weight humics on the Kraft effluent and protein-like compounds on the TMP/RCF effluent. Although the GAC exhaustion rates are too high for an effective DOC removal option for biologically treated pulp and paper mill effluents, the study indicates that advanced organic characterisation techniques can be used to diagnose GAC performance on complex effluents with comparable bulk DOC and COD loads.

Accelerated seeded precipitation pre-treatment of municipal wastewater to reduce scaling.

Membrane based treatment processes are very effective in removing salt from wastewater, but are hindered by calcium scale deposit formation. This study investigates the feasibility of removing calcium from treated sewage wastewater using accelerated seeded precipitation. The rate of calcium removal was measured during bench scale batch mode seeded precipitation experiments at pH 9.5 using various quantities of calcium carbonate as seed material. The results indicate that accelerated seeded precipitation may be a feasible option for the decrease of calcium in reverse osmosis concentrate streams during the desalination of treated sewage wastewater for irrigation purposes, promising decreased incidence of scaling and the option to control the sodium adsorption ratio and nutritional properties of the desalted water. It was found that accelerated seeded precipitation of calcium from treated sewage wastewater was largely ineffective if carried out without pre-treatment of the wastewater. Evidence was presented that suggests that phosphate may be a major interfering substance for the seeded precipitation of calcium from this type of wastewater. A pH adjustment to 9.5 followed by a 1-h equilibration period was found to be an effective pre-treatment for the removal of interferences. Calcium carbonate seed addition at 10 g l(-1) to wastewater that had been pre-treated in this way was found to result in calcium precipitation from supersaturated level at 60 mg l(-1) to saturated level at 5 mg l(-1). Approximately 90% reduction of the calcium level occurred 5 min after seed addition. A further 10% reduction was achieved 30 min after seed addition.

Evaluation of ion exchange resins for the removal of dissolved organic matter from biologically treated paper mill effluent

In this study, the efficiency of six ion exchange resins to reduce the dissolved organic matter (DOM) from a biologically treated newsprint mill effluent was evaluated and the dominant removal mechanism of residual organics was established using advanced organic characterisations techniques. Among the resins screened, TAN1 possessed favourable Freundlich parameters, high resin capacity and solute affinity, closely followed by Marathon MSA and Marathon WBA. The removal efficiency of colour and lignin residuals was generally good for the anion exchange resins, greater than 50% and 75% respectively. In terms of the DOM fractions removal measured through liquid chromatography-organic carbon and nitrogen detector (LC-OCND), the resins mainly targeted the removal of humic and fulvic acids of molecular weight ranging between 500 and 1000 g mol(-1), the portion expected to contribute the most to the aromaticity of the effluent. For the anion exchange resins, physical adsorption operated along with ion exchange mechanism assisting to remove neutral and transphilic acid fractions of DOM. The column studies confirmed TAN1 being the best of those screened, exhibited the longest mass transfer zone and maximum treatable volume of effluent. The treatable effluent volume with 50% reduction in dissolved organic carbon (DOC) was 4.8 L for TAN1 followed by Marathon MSA - 3.6L, Marathon 11 - 2.0 L, 21K-XLT - 1.5 L and Marathon WBA - 1.2 L. The cation exchange resin G26 was not effective in DOM removal as the maximum DOC removal obtained was only 27%. The resin capacity could not be completely restored for any of the resins; however, a maximum restoration up to 74% and 93% was achieved for TAN1 and Marathon WBA resins. While this feasibility study indicates the potential option of using ion exchange resins for the reclamation of paper mill effluent, the need for improving the regeneration protocols to restore the resin efficiency is also identified. Similarly, care should be taken while employing LC-OCND for characterising resin-treated effluents, as the resin degradation is expected to contribute some organic carbon moieties misleading the actual performance of resin.