Emile Cornelissen

Water recovery from sewage using forward osmosis.

This research is part of the Sewer Mining project aimed at developing a new technological concept by extracting water from sewage by means of forward osmosis (FO). FO, in combination with a reconcentration system, e.g. reverse osmosis (RO) is used to recover high-quality water. Furthermore, the subsequent concentrated sewage (containing an inherent energy content) can be converted into a renewable energy (RE) source (i.e. biogas). The effectiveness of FO membranes in the recovery of water from sewage has been evaluated. Stable FO water flux values (>4.3 LMH) were obtained with primary effluent (screened, not treated) used as the feed solution. Fouling of the membrane was also induced and further investigated. Accumulated fouling was found to be apparent, but not irreversible. Sewer Mining could lead to a more economical and sustainable treatment of wastewater, facilitating the extraction of water and energy from sewage and changing the way it is perceived: not as waste, but as a resource.

Trace organic solutes in closed-loop forward osmosis applications: Influence of membrane fouling and modeling of solute build-up

In this study, trace organics transport in closed-loop forward osmosis (FO) systems was assessed. The FO systems considered, consisted of an FO unit and a nanofiltration (NF) or reverse osmosis (RO) unit, with the draw solution circulating between both units. The rejection of trace organics by FO, NF and RO was tested. It was found that the rejection rates of FO were generally comparable with NF and lower than RO rejection rates. To assess the influence of fouling in FO on trace organics rejection, FO membranes were fouled with sodium alginate, bovine serum albumin or by biofilm growth, after which trace organics rejection was tested. A negative influence of fouling on FO rejection was found which was limited in most cases, while it was significant for some compounds such as paracetamol and naproxen, indicating specific compound-foulant interactions. The transport mechanism of trace organics in FO was tested, in order to differentiate between diffusive and convective transport. The concentration of trace organics in the final product water and the build-up of trace organics in the draw solution were modeled assuming the draw solution was reconcentrated by NF/RO and taking into account different transport mechanisms for the FO membrane and different rejection rates by NF/RO. Modeling results showed that if the FO rejection rate is lower than the RO rejection rate (as is the case for most compounds tested), the added value of the FO-RO cycle compared to RO only at steady-state was small for diffusively and negative for convectively transported trace organics. Modeling also showed that trace organics accumulate in the draw solution.

Sorption and biodegradation of organic micropollutants during river bank filtration: A laboratory column study

This study investigated sorption and biodegradation behaviour of 14 organic micropollutants (OMP) in soil columns representative of the first metre (oxic conditions) of the river bank filtration (RBF) process. Breakthrough curves were modelled to differentiate between OMP sorption and biodegradation. The main objective of this study was to investigate if the OMP biodegradation rate could be related to the physico-chemical properties (charge, hydrophobicity and molecular weight) or functional groups of the OMPs. Although trends were observed between charge or hydrophobicity and the biodegradation rate for charged compounds, a statistically significant linear relationship for the complete OMP mixture could not be obtained using these physico-chemical properties. However, a statistically significant relationship was obtained between biological degradation rates and the OMP functional groups. The presence of ethers and carbonyl groups will increase biodegradability, while the presence of amines, ring structures, aliphatic ethers and sulphur will decrease biodegradability. This predictive model based on functional groups can be used by drinking water companies to make a first estimate whether a newly detected compound will be biodegraded during the first metre of RBF or that additional treatment is required. In addition, the influence of active and inactive biomass (biosorption), sand grains and the water matrix on OMP sorption was found to be negligible under the conditions investigated in this study. Retardation factors for most compounds were close to 1, indicating mobile behaviour of these compounds during soil passage. Adaptation of the biomass towards the dosed OMPs was not observed for a 6 month period, implying that new developed RBF sites might not be able to biodegrade compounds such as atrazine and sulfamethoxazole in the first few months of operation.

Experimental studies and modeling on concentration polarization in forward osmosis

Concentration polarization (CP) is an important issue in forward osmosis (FO) processes and it is believed that the coupled effect of dilutive internal CP (DICP) and concentrative external CP (CECP) limits FO flux. The objective of this study was to distinguish individual contribution of different types of DICP and CECP via modeling and to validate it by pilot studies. The influence of DICP/CECP on FO flux has been investigated in this study. The CP model presented in this work was derived from a previous study and evaluated by bench-scale FO experiments. Experiments were conducted with drinking water as the feed and NaCl/MgSO(4) as draw solutions at different concentrations and velocities. Modeling results indicated that DICP contributed to a flux reduction by 99.9% for 0.5 M NaCl as a draw solution although the flow pattern of both feed and draw solutions was turbulent. DICP could be improved via selection of the draw solution. The modeling results were well fit with the experimental data. It was concluded that the model could be used for selection of the draw solution and prediction of water flux under similar situation. A draw solution with greater diffusion coefficient or a thinner substrate of an asymmetric FO membrane resulted in a higher flux.

Preliminary study of osmotic membrane bioreactor: effects of draw solution on water flux and air scouring on fouling.

Preliminary study on a novel osmotic membrane bioreactor (OMBR) was explored. Objective of this study was to investigate the effects of draw solution on membrane flux and air scouring at the feed side on fouling tendency in a pilot OMBR system composing the anoxic/aerobic and forward osmosis (FO) processes. Domestic sewage was the raw feed, FO membrane from HTI and NaCl/MgSO4 draw solutions were used in the experiments. Fluxes of 3 l/m2/h (LMH) and 7.2 LMH were achieved at osmotic pressure of 5 and 22.4 atm, respectively. No significant flux decline was observed at 3 LMH over 190 h and at 7.2 LMH over 150 h when air scouring was provided at the feed side of the membrane. However, without air scouring, the flux at 22.4 atm osmotic pressure declined by 30% after 195 h and then levelled off. The potential advantages of the fouling reversibility with air scouring under the operating conditions of the pilot OMBR and better water quality in OMBR over the conventional MBR were preliminarily demonstrated.

Influence of biofouling on pharmaceuticals rejection in NF membrane filtration.

The effects of biomass attachment and growth on the surface characteristics and organic micropollutants rejection performance of nanofiltration membranes were investigated in a pilot installation. Biomass growth was induced by dosing of a readily biodegradable carbon source resulting in the formation of a biofouling in the investigated membrane elements. Surface properties and rejection behaviour of a biofouled and virgin membrane were investigated and compared in terms of surface charge, surface energy and hydrophobicity. The last two were accomplished by performing contact angle measurements on fully hydrated membrane surfaces, in order to mimic the operating conditions of a membrane in contact with water. Compared to a virgin membrane, deposition and growth of biofilm did slightly alter the surface charge, which became more negative, and resulted in a higher hydrophilicity of the membrane surface. In addition, the presence of the negatively charged biofilm induced accumulation of positively charged pharmaceuticals within the biomass layer, which probably also hindered back diffusion. This caused a reduction in rejection efficiency of positively charged solutes but did not alter rejection of neutral and negatively charged pharmaceuticals. Pharmaceuticals rejection was found to positively correlate with the specific free energy of interaction between virgin or biofouled membranes and pharmaceuticals dissolved in the water phase. The rejection values obtained with both virgin and biofouled membranes were compared and found in good agreement with the predictions calculated with a solute transport model earlier developed for high pressure filtration processes.

Quantitative assessment of the efficacy of spiral-wound membrane cleaning procedures to remove biofilms

Cleaning of high pressure RO/NF membranes is an important operational tool to control biofouling. Quantitative information on the efficacy of cleaning agents and protocols to remove biomass is scarce. Therefore, a laboratory cleaning test to assess the efficiency of cleaning procedures to remove attached biomass was developed. The major components of the test are (i) production of uniform biofilm samples, (ii) the quantification of the biomass concentrations with robust parameters and (iii) a simple test procedure with optimal exposure of the biofilm samples to the chemicals. The results showed that PVC-P is a suitable substratum for the production of uniform biofilm samples. ATP and carbohydrates (CH) as major components of the biofilm matrix for nucleotides (living bacterial cells) and extracellular polymeric substances EPS, respectively, were selected as robust biomass parameters. The removal of ATP and CH with the NaOH/Sodium Dodecyl Sulfate (SDS) mixture, selected as a standard treatment at pH 12.0, was reproducible. The resistance of the EPS matrix against chemical cleaning was demonstrated by a low CH removal (32.8 ± 6.0%) compared to the ATP removal (70.5 ± 15.1%). The inverse relationship of biomass removal with the CH to ATP ratio (μg/ng) of the biofilms demonstrated the influence of the biomass characteristics on cleaning. None of the 27 chemicals tested (analytical-grade and commercial brands) in single step or in double-step treatments were significantly more effective than NaOH/SDS. Oxidizing agents NaOCl and H(2)O(2), the latter in combination with SDS, both tested as common agents in biofilm control, showed a significantly higher efficiency (70%) to remove biofilms. In the test, simultaneously, the efficiency of agents to remove precipitated minerals such as Fe can be assessed. Validation tests with Cleaning in Place (CIP) in 8 and 2.5-inch RO membrane pilot plant experiments showed similar ranking of the cleaning efficiency of cleaning protocols as determined in the laboratory tests. Further studies with the laboratory test are required to study the effect of cleaning conditions such as duration, temperature, shear forces as well as chemical conditions (concentrations, alternative agents or mixtures and sequence of application) on the efficiency to remove attached biomass.

Effect of anionic fluidized ion exchange (FIX) pre-treatment on nanofiltration (NF) membrane fouling

Anionic Fluidized Ion Exchange (FIX) is used to improve the performance of downstream Nanofiltration (NF). The research is divided in three parts: (i) NOM removal by FIX, (ii) the effect of FIX treatment on NF fouling and (iii) FIX treatment in relation to biological stability. Pre-treated anaerobic groundwater was (i) fed directly to a 4-inch membrane element and (ii) fed to another 4-inch membrane element after anionic FIX treatment. The operational parameters of the membrane set-up were monitored during 42 days, followed by a membrane autopsy study in which accumulated biological, organic and inorganic fouling was determined. Parallel to this experiment, two small ion exchange (IEX) resin and glass beads filled columns were operated to study the effect of FIX on the biomass concentration of the feed water. FIX operated satisfactory and selectively removed humic substances (>90%) and hydrophobic organic carbon (HOC) (>80%) from the feed water. Furthermore, iron was substantially removed (71%) which was explained by complexation with humic substances. Removal of NOM by FIX did not reduce membrane fouling problems; the Membrane Transport Coefficient (MTC) decreased and the Normalized Pressure Drop (NPD) increased more rapidly for the NF membrane after FIX compared to the membrane without FIX pre-treatment. NOM removal by FIX did not reduce adsorption of organic matter onto the downstream membrane element, since predominantly humic substances were removed which did not adsorb to the membrane surface. FIX treatment resulted in higher biomass densities (400%) and slightly less iron deposition (20%) onto the membrane surface. Fouling of the membrane element after FIX treatment was dominated by biofouling and fouling of the reference membrane element experienced more colloidal iron fouling compared to the membrane element after FIX, both resulting in an increase in NPD. The microbiological water quality deteriorated after anionic FIX treatment, as was observed by an increase in ATP content. Growth of biomass onto the IEX resins was observed which was caused by both IEX materials and feed water components, such as NOM fractions.

Threshold concentrations of biomass and iron for pressure drop increase in spiral-wound membrane elements

In a model feed channel for spiral-wound membranes the quantitative relationship of biomass and iron accumulation with pressure drop development was assessed. Biofouling was stimulated by the use of tap water enriched with acetate at a range of concentrations (1-1000 μgCl(-1)). Autopsies were performed to quantify biomass concentrations in the fouled feed channel at a range of Normalized Pressure Drop increase values (NPD(i)). Active biomass was determined with adenosinetriphosphate (ATP) and the concentration of bacterial cells with Total Direct Cell count (TDC). Carbohydrates (CH) were measured to include accumulated extracellular polymeric substances (EPS). The paired ATP and CH concentrations in the biofilm samples were significantly (p<0.001; R(2)=0.62) correlated and both parameters were also significantly correlated with NPD(i) (p<0.001). TDC was not correlated with the pressure drop in this study. The threshold concentration for an NPD(i) of 100% was 3.7 ng ATP cm(-2) and for CH 8.1 μg CH cm(-2). Both parameters are recommended for diagnostic membrane autopsy studies. Iron concentrations of 100-400 mg m(-2) accumulated in the biofilm by adsorption were not correlated with the observed NPD(i), thus indicating a minor role of Fe particulates at these concentrations in fouling of spiral-wound membrane.

Assessing the effect of surface modification of polyamide RO membrane by l-DOPA on the short range physiochemical interactions with biopolymer fouling on the membrane

Theoretical predictions of interaction energies for membrane-biopolymer foulant pairs were used to compare the fouling tendencies of a virgin commercial polyamide reverse osmosis (RO) membrane with a amino acid 3-(3,4-dihydroxyphenyl)-l-alanine (l-DOPA) coated RO membrane. Lifshitz-van der Waals (LW) and Lewis acid-base (AB) surface tension components of the membranes were determined based on contact angle results using the van Oss approach. From these values, the LW and AB components of the free energy of adhesion between membrane and foulants were calculated. Electrostatic (EL) double layer interaction energies between the membrane and foulants were also estimated using the measured surface charge data of the membranes and fouling agents. Bovine serum albumin (BSA) and alginic acid sodium salt (alginate) were used as model biopolymers causing membrane fouling. Based on the calculated adhesion free energies, acid-base interactions were found to have the strongest impact on the adhesion of both BSA and alginate to the either membranes surfaces. It was found that l-DOPA modification has significantly lowered acid-base interaction affinity toward the adhesion of both foulants studied. On the basis of calculated free energies of adhesion, lower fouling tendency of the l-DOPA modified membrane was expected. The accelerated fouling tests indicated a lower flux decline rate for the modified membrane and confirmed the results obtained from theory.