Search

This study aims at applying and verifying the MFI-UF method to predict particulate fouling in RO plants. Two full-scale RO plants treating surface water, with average capacity of 800–2000 m3/h, were studied. Firstly, the MFI-UF of RO feed and concentrate was measured using 5–100 kDa membranes at same flux applied in the RO plants (20–26 L/m2.h). Subsequently, the particle disposition factor (Ω) was calculated to simulate particle deposition in RO cross-flow filtration. Finally, particulate fouling rates were predicted based on MFI-UF and Ω, and compared with the actual fouling rates in the plants. For plant A, the results showed that the fouling rates predicted using MFI-UF measured with 100 kDa membrane have the best agreement with the actual fouling (with 3–11 % deviation). For plant B, the fouling rates predicted based on both 10 and 100 kDa membranes agree well with the actual fouling (with 2 % and 15 % deviation, respectively). However, the fouling predicted based on 5 kDa membrane is considerably overestimated for both plants, which is attributed to the effect of the low surface porosity of 5 kDa membrane. More widespread applications of MFI-UF in full-scale RO plants are required to demonstrate the most suitable MFI-UF membranes for fouling prediction.

For assessing the particulate fouling of water, the modified fouling index (MFI0.45) is a superior test to the silt density index (SDI). There is a need to compare both tests in terms of sensitivity, how they are affected by the filter material and the type of support plate and also illustrate their use for monitoring of seawater quality over time. In this work, we studied seven different filter holders with different filter support plates and three different 0.45 μm filter materials, and we applied the tests for monitoring of North Sea water quality. The results illustrated that the type of support plate of the filter holder greatly influences the measured MFI0.45 values and thus, a correction for the effective membrane area may be needed when carrying out an MFI0.45 test. An attempt to normalize differences in MFI0.45 due to filter material with a Formazin solution was tested but proven not successful. When monitoring the seawater, the MFI0.45 was much more sensitive than SDI to water quality variations in particular during algal growth. As the SDI and MFI0.45 tests can be measured with help of the same equipment, more alignment in the ASTM protocols for both methods is recommended.

A direct method for measuring adenosine-triphosphate (ATP) in seawater was developed recently, in which commercial reagents are added directly to seawater. However, calibration is required if seawater quality changes (such as changes in salinity, pH, Mg2+, Fe3+) as the seawater matrix interferes with ATP measurement. In this research, a 0.1 μm filtration process is introduced to eliminate such interferences. In addition, a filter rinsing step with sterilized artificial seawater is proposed to eliminate interference of free ATP. The ATP-filtration method is fast (2 = 0.72, n = 100) with intact cell concentration. Microbial ATP concentration measured using the ATP-filtration method and the ATP-direct method were comparable. Microbial ATP measured along the treatment train of a full-scale seawater reverse osmosis (SWRO) plant decreased from 530 in the raw seawater to 10 ng-ATP/L after pre-treatment and to 0.5 ng-ATP/L in the SWRO permeate. The method was also applied to monitor bacterial growth potential (BGP) across the pre-treatment train of a (pilot) seawater desalination plant, where the removal of BGP through the media filtration and ultrafiltration was 44% and 7%, respectively.

Various bacterial growth potential (BGP) methods have been developed recently to monitor biofouling in seawater reverse osmosis (SWRO) systems such as assimilable organic carbon and bacterial regrowth potential. However, the relationship between these methods and biofouling in SWRO desalination plants has not yet been demonstrated. In this research, an attempt is made to investigate if a correlation exists between BGP of SWRO feed water and the chemical cleaning frequency in SWRO plants using an ATP-based BGP method employing an indigenous microbial consortium. Using ATP-based BGP method at 5 different seawater locations showed low variations of bacterial yield. The BGP method was applied to assess the pretreatment performance of three full-scale SWRO plants with different pretreatment processes. Dual media filtration (DMF) showed the highest BGP removal (>50%) in two SWRO plants. Removal of BGP and hydrophilic organic carbon in dissolved air floatation combined with ultrafiltration was similar to the removal achieved with DMF in combination with inline coagulation. For the three SWRO plants investigated, a higher BGP in SWRO feed water corresponded to a higher chemical cleaning frequency. However, more data is required to confirm if a real correlation exists between BGP and biofouling in SWRO plants.

Pretreatment with microfiltration (MF) or ultrafiltration (UF) membranes has been proposed for seawater reverse osmosis (SWRO) plants to address operational issues associated with algal blooms. Here, we investigated the MF/UF rejection and fouling potential of algal organic matter (AOM) released by common species of bloom-forming marine ( Alexandrium tamarense and Chaetoceros affinis ) and freshwater ( Microcystis sp.) algae. Batch culture monitoring of the three algal species illustrated varying growth pattern, cell concentration, AOM released and membrane fouling potential. The high membrane fouling potential of the cultures can be directly associated (R 2 > 0.85) with AOM such as transparent exopolymer particle (TEP) while no apparent relationship with algal cell concentration was observed. The AOM comprised mainly biopolymers (e.g., polysaccharides and proteins) and low molecular weight organic compounds (e.g., humic-like substances). The former were largely rejected by MF/UF membranes while the latter were poorly rejected. MF (0.4 μm and 0.1 μm pore size) rejected 14%–56% of biopolymers while conventional UF (100 kDa) and tight UF (10 kDa) rejected up to 83% and 97%, respectively. The retention of AOM resulted in a rapid increase in trans-membrane pressure (ΔP) over time, characterised by pore blocking followed by cake filtration with enhanced compression as illustrated by an exponential progression of ΔP.

Reliable methods for measuring and predicting the fouling potential of reverse osmosis (RO) feed water are important in preventing and diagnosing fouling at the design stage, and for monitoring pre-treatment performance during plant operation. The Modified Fouling Index Ultrafiltration (MFI-UF) constant flux is a significant development with respect to assessing the fouling potential of RO feed water. This research investigates (1) the variables influencing the MFI-UF test at constant flux filtration (membrane pore size, membrane material, flux rate); and (2) the application of MFI-UF into pre-treatment assessment and RO fouling estimation. The dependency of MFI on flux, means that to assess accurately particulate fouling in RO systems, the MFI should be measured at a flux similar to a RO system (close to 20L/m2/h) or extrapolated from higher fluxes. The two studied membrane materials showed reproducible results; 10% for PES membranes and 6.3% for RC membranes. Deposition factors (amount of particles that remain on the surface of membrane) were measured in a full-scale plant ranging between 0.2 and 0.5. The concept of “safe MFI” is presented as a guideline for assessing pre-treatment for RO systems.

This article reviews the occurrence of HABs in seawater, their effects on the operation of seawater reverse osmosis (SWRO) plants, the indicators for quantifying/predicting these effects, and the pretreatment strategies for mitigating operational issues during algal blooms. The potential issues in SWRO plants during HABs are particulate/organic fouling of pretreatment systems and biological fouling of RO membranes, mainly due to accumulation of algal organic matter (AOM). The presence of HAB toxins in desalinated water is also a potential concern but only at very low concentrations. Monitoring algal cell density, AOM concentrations and membrane fouling indices is a promising approach to assess the quality of SWRO feedwater and performance of the pretreatment system. When geological condition is favourable, subsurface intake can be a robust pretreatment for SWRO during HABs. Existing SWRO plants with open intake and are fitted with granular media filtration can improve performance in terms of capacity and product water quality, if preceded by dissolved air flotation or sedimentation. However, the application of advanced pretreatment using ultrafiltration membrane with in-line coagulation is often a better option as it is capable of maintaining stable operation and better RO feed water quality during algal bloom periods with significantly lower chemical consumption.

In supersaturated solutions the period preceding the start of ‘measurable’ crystallization is normally referred to as the ‘induction time’. This research project aimed to investigate the induction times of CaCO3 in the presence of Mg2+ and SO42−. The prepared synthetic solutions have the same ionic strength values found in the Gulf of Oman SWRO concentrates at 30% and 50% recovery. The results showed a significant increase in the induction time by 1140%, 2820%, and 3880% for a recovery of 50%, when adding SO42− only, Mg2+ only, or both Mg2+ and SO42−, respectively, to synthetic SWRO concentrate compared to that obtained in the absence of Mg2+ and SO42− at an initial pH of 8.3. The increase in the induction time in the presence of SO42− was more than likely to be due to nucleation and growth inhibition while the presence of Mg2+ affected the nucleation and growth through both complexation and inhibition. After a 5-month solution stabilization period, ESEM and XRD analyses showed aragonite in solutions containing Mg2+. On the contrary, calcite was the final crystal phase formed in solutions with no Mg2+. This suggests that magnesium may play an important role in inhibiting the formation of calcite.

The pH of seawater reverse osmosis plants (SWRO) is the most influential parameter in determining the degree of supersaturation of CaCO3 in the concentrate stream. For this, the results of pH measurements of the concentrate of a seawater reverse osmosis pilot plant were compared with pH calculations based on the CO2―HCO3−―CO32− system equilibrium equations. Results were compared with two commercial software programs from membrane suppliers and also the software package Phreeqc. Results suggest that the real concentrate pH is lower than that of the feed and that none of the used programs was able to predict correctly real pH values. In addition, the effect of incorporating the acidity constant calculated for NaCl medium or seawater medium showed a great influence on the concentrate pH determination. The HCO3− and CO32− equilibrium equation using acidity constants developed for seawater medium was the only method able to predict correctly the concentrate pH. The outcome of this study indicated that the saturation level of the concentrate was lower than previously anticipated. This was confirmed by shutting down the acid and the antiscalants dosing without any signs of scaling over a period of 12 months.

Polyaluminium chloride (coagulant) is a very complex compound. The change in the ratio HCO 3 /CO 2 during coagulation cause a change in pH in the subsequent RO unit (because CO 2 is not rejected by the RO membrane), and it was found that this change is constant for any given recovery, e.g. at 75% recovery, the pH change between the RO feed water and concentrate water is 0.6. The simulation of aluminium solubility and concentration along RO systems were performed assuming the presence of three aluminium species, Al 13 7+ , Al(OH) 4 − , Al(OH) 2+ . The simulation of aluminium solubility was performed at two temperatures, 20°C to represent the summer period and 5°C to represent the winter period. Results showed that for winter (5°C) an RO feed water pH of at least 7.2 and for warm weather (20°C), a feed water pH of at least 6.7 are required respectively, to avoid scaling of aluminium on the RO membranes. In conclusion, theoretical simulation of aluminium solubility and results of deposition factor showed that aluminium more than likely plays a role in fouling of the RO membranes and could be a reason for the frequent cleaning.

In order to ensure stable treated water quality and to reduce chemical costs in any treatment plant it is necessary to study and optimize the coagulant dosing control (CDC). This research focused on the affects of coagulant in the integrated membrane (UF & RO) system employed for industrial water production. The dose of coagulant (PACl) might be associated (partly) with the frequency of cleaning in the RO units, and a value of 20 abs/m in the UF permeate is the control of the dosage process. This could suggest that organic fouling is directly and indirectly (inducing biofouling as well) the cause of fouling of the RO membranes. Nevertheless, high doses of PACl could produce scaling of aluminium and, in this work minimizing PACl to prevent operational problems in RO membranes was focused. The approach involved the study of the treatment processes, determination of the optimum dose of coagulant, evaluation of the removal efficiency of UV and DOC by coagulation. The available data, which comprised monthly/weekly measurements for a period of six years of operation, was studied and analyzed and an attempt was made to draw some conclusions for the plant regarding the coagulant dosage and the link with UV absorbance as control. The study of the coagulant dosing control revealed that the use of simple and robust online sensors like UV measurement allows an automatic dosing control although this parameter is not found to be sufficient to fully characterize nor predict fouling during membrane operation and there is no link between UV after the UF and the cleaning frequency of the RO. Parameters, as the added value allowed to verify the doses efficiencies in UV and DOC removal. Coagulant dose depends on the level of UV absorbance in the UF permeate and it should never be higher than 20 abs/m to ensure a RO cleaning frequency as long as possible. However, it was found that the target value of 20 abs/m produced substantial over dose of coagulant — 90% of the time – when UV removal is considered only. In the same way, for coagulant doses in excess of 5 mg/L, the additional removal of UV is less than 2.5% per mg coagulant/L, which suggests that the current dose is on the high side. Furthermore, a coagulant dose in excess of 5 mg/L did not produce significant additional UV removal, and thus a reduction from 14 mg/L (2005 dose) to ca. 7 mg/L could be considered.

Natural organic matter (NOM) plays a significant role in fouling of ultrafiltration membranes in drinking water treatment processes. The aim of this study was to obtain a better understanding of the interactions between the fractional components of NOM and a hydrophilic PES/PVP hollow fiber ultrafiltration membrane (150–200 kDa). NOM was fractionated into hydrophobic, transphilic and hydrophilic acid fractions according to the XAD-8/4 resin method proposed by Aiken et al. [1]. UF filtration tests were performed with diluted NOM fractions (1.5 mg/l DOC) isolated from surface water (Lake Ijssel). From NOM fractionation results, Lake IJssel water comprised about 53–55% hydrophobic, 20–22% hydrophilic and 25% transphilic NOM acids. Filtration of three NOM fractions (hydrophobic, hydrophilic and transphilic) suggested that the fouling potential was in the order: hydrophilic > hydrophobic > transphilic. The reversibility of fouling due to the hydrophilic fraction was very poor, and decreased from ca. 50% in the first cycle to 20,000 Da) in the LC-OCD chromatogram of the UF permeate, suggested that polysaccharides (> 20,000 Da) were rejected by the UF membrane and may have contributed to fouling.

The MFI-UF, based on cake filtration, was developed to measure and predict the particulate fouling potential of feedwater to membrane filtration installations. The MFI-UF is determined in constant pressure filtration with the flux deceasing during the test. However, many membrane systems, e.g., reverse osmosis (RO), operate at constant flux with pressure increasing when fouling occurs. As both pressure and flux contribute to cake compression, determining the MFI-UF in constant flux with correction to the flux of an RO system is expected to more closely simulate particulate fouling Therefore, this research investigated the development of the MFI-UF test in constant flux filtration applying low (tap water) and high fouling (diluted canal water) feedwater. Preliminary experiments were promising; the fouling index (I) (and hence the MFI-UF) of all feedwater could be determined within 2 h under constant flux filtration. Cake filtration was demonstrated as (1) a minimum in the fouling indexvs time plot and (2) by linearity of the fouling index with feedwater particulate concentration. The fouling index increased with increasing applied flux due to cake compression. Further investigation at higher and lower applied flux is required to identify a reference test flux and to develop a method to correct the fouling index to the reference test flux and/or the flux of a membrane filtration system The fouling index can then be applied in a model to predict fouling.

Membrane fouling in a side stream biomass separation MBR pilot plant was investigated. Constant flux filtration (18–721/M2h) was employed. Air was continuously supplied to the MBR system with the feed (sludge) to flush the membrane surface, and backwashing was applied every 5–10 min for 8 s to control membrane fouling. Although the duration of pore blocking was generally short (completed in 8 s at a flux of 52 llm2h), blocking resistance (mainly irreversible blocking resistance) was the main cause of membrane fouling. However, the resistance of the filter cake also played an important role, particularly when the backwashing interval was extended to 10 min. In terms of fouling reversibility, blocking resistance was not completely reversible by backwashing, especially at higher fluxes (e.g. 69 I/m2h), and frequent chemical cleaning (once every week at 401/m2h) was required. However, cake filtration was easily reversible via a combination of backwashing and sludge/air flushing of the membrane surface. Finally, a simple method to identify both irreversible and reversible blocking resistance and filter cake resistance was proposed.

A set of coherent tools for (i) analysing the fouling of NF and RO membranes and (ii) determining the fouling potential of feed water is presented in this paper. These tools—developed for diagnosis, prediction, prevention and control of fouling — are applied in practice and have proven their value in controlling fouling. A database obtained with the described tools at plants with and without operational problems is available for the interpretation of observations in pilot plants and in practice.

Desalination, along with wastewater reuse and water importation, can provide a means of increasing the available fresh water in regions of the world where water is scarce. If desalination is to be successfully implemented in developing countries to increase the supply of fresh water for municipal and industrial purposes, capacity building is urgently needed at both the undergraduate and postgraduate level. However, courses dealing with the technical aspects of desalination technology alone are not the solution to the problem. Undergraduate and postgraduate programmes involving an integrated approach to sustainable water management are needed; and desalination, wastewater reuse and irrigation management should be part of these programmes.

Desalination technology is finding new outlets in supplying water to meet growing municipal domestic consumption needs in water scarce countries with a per capita availability below 1,000 m3/y. An expansion of the current municipal water desalination market was related to the population growth and the groundwater scarcity in the coming 25 years in various regions of the world: Europe, The Caribbean, South East and Western Asia, GCC States and North Africa. First, the current impact of desalination on the renewable groundwater resources in these selected areas was determined. Results indicated, that the desalination capacity exceeds 2–10 times the renewable groundwater resources in Qatar, Kuwait, Malta and Saudi Arabia, 10–50% in Libya and Barbados, and less than 0.5% in Jordan, Yemen and Singapore. In the future, a population growth from 51–116 million, 1995–2025, was assumed to be the driving force determining the need for desalination in order to maintain the current urban municipal domestic water consumption (an average of 0.265 m3/cap/d) in these countries. By 2000, a total sea and brackish water desalination capacity of 7.3 million m3/d was installed for municipal purposes in these countries. This indicated a growth in the desalination capacity of 1.9 million m3/d, 35%, between 1995 and 2000. By 2025, the growth in the municipal water desalination market will need to reach 14.8 million m3/d, 200%, to maintain the current urban municipal domestic water needs and to prevent any decline in renewable groundwater resources in the 10 water scarce countries selected in this study.

In nanofiltration and reverse osmosis membrane systems 75–90% of the feed water is conversed to product, and as a consequence dissolved compounds are concentrated in the membrane concentrate. Subsequently, sparingly soluble inorganic compounds such as calcium carbonate and barium sulphate become supersaturated and may precipitate in a membrane element (scaling). Scaling leads to an increase in energy consumption and chemical cleaning frequency. This is undesirable and, therefore, an accurate prediction and monitoring of the risk of scaling is important. In this study a model is presented, called Supersaturation Prediction Model, which is able to calculate the supersaturation ratios of sparingly soluble compounds at the membrane surface. The model is a combination of a concentration polarisation model from literature and a model for calculation of the supersaturation ratio of sparingly soluble compounds. The concentration polarisation model from literature was fitted to measured average concentration polarisation factors of magnesium sulphate in a 4″×40″ spiral wound reverse osmosis membrane element operated with different permeate fluxes and linear flow velocities of the membrane concentrate. The fitted concentration polarisation model was also able to predict the average concentration polarisation factors of sodium chloride and magnesium chloride in a 2.5″×40″ spiral wound nanofiltration membrane element. With the aim to detect scaling in an early stage a scaling monitor has been developed and tested. The scaling monitor, named ScaleGuard®, measures the normalised flux of a small membrane element fed by a part of the concentrate of the full-scale plant. This scaling monitor was tested at a pilot plant treating anaerobic groundwater for the production of drinking water. The monitor demonstrated to be able to detect scaling before it took place in the last stage of the pilot plant.

The feasibility of ultra and microfiltration depends strongly on the achieved net flux. Direct treatment of surface water frequently results in low net fluxes and high cleaning frequencies. This is largely determined by the quality of the feed water and the mechanisms that determine the fouling and the fouling rate of the membranes. Controlling and preferably reducing the rate of fouling is a major challenge. Research has been carried out to reduce fouling to achieve a stable operation at an economical feasible net flux. This research investigates decreasing the membrane contact area for suspended and colloidal matter by a dynamic filter, i.e. Enhanced Pre-Coat Engineering (EPCE). The general applicability and the effects on the membrane performance of this new approach, has been determined on different water sources and with different pre-coat materials. Pre-coating results initially in a higher fouling rate, however this fouling rate stabilises after several filtration cycles. The stabilisation is caused by a better restoration of the permeability after a backwash and an enhanced backwash. From this research it can be concluded that EPCE is a promising technique to increase membrane performance. Long term effects, costs and full-scale implementation need to be demonstrated.

In reverse osmosis (RO) and nanofiltration silica may precipitate on the membrane surface causing the decline in mass transfer. The appearance of deposits will depend on the different forms of silica, namely monomer, polymer and colloidal present in the feed water and concentrate. Exceeding the solubility of monomer silica is commonly considered to result in the formation of silica scale, which is difficult to remove. As a consequence, deposition of silica scale is prevented avoiding supersaturations, which results in limited conversions with silica rich feedwaters, and/or more recently the addition of antiscalants. The removal of silica in a separate process may be a successful alternative to scale control if the monomer silica concentration is either removed below the solubility level or the remaining silica supersaturation will not cause the deposition in next stage of a RO system. For this purpose, silica deposition from supersaturated solutions (200–700 mg/l) was tested in seeded silica gel (size 63–200 μm) batch experiments (pH = 7, 25°C). Results indicated that monomer silica deposit on seeds and form silica polymers in solution. The formation of latter strongly negatively affects the removal of silica by silica gel seeds. Addition of iron (III) in comparable experiments had a profound effect on removal of monomer and polymer silica. Removal mechanisms most likely include the formation of sparingly soluble iron silicates resulting in un-saturated amorphous silica solution and coagulation of polymer silica.

Existing fouling indices to measure the particulate fouling potential of reverse osmosis (RO) feedwater, the SDI and MFI0.45 do not include smaller colloidal particles. Therefore, the MFI using ultrafiltration (UF) membranes was developed (1000–100,000 Da pore size) to incorporate and measure smaller particles. A 13,000 Da membrane was selected as the most promising reference membrane for application in the MFI-UF test. This research investigates its application to (1) measure the particulate fouling potential of RO feedwater and (2) evaluate the efficiency of pretreatment processes on particle removal at two RO pilot plants. Where possible a comparison was made with the SDI, MFI0.45, and particle counts. Both RO plants use conventional pretreated surface water. Subsequently, the Rhine River plant uses ozonation, biological activated carbon filtration, and slow sand filtration while the Ijssel Lake plant uses ultrafiltration (150–200,000 Da). The MFI-UF of the influent feedwater was about 400–1400 higher than the corresponding MFI0.45 and SDI, due to the retention of smaller particles. A reduction in the particulate fouling potential, > -80%, was found by MFI-UF with pretreatment at both plants. For the larger particles the MFI0.45 gave a 90–≈100% reduction at the Rhine River plant, while the SDI gave more variable results: 70 > 90%.

In nanofiltration and reverse osmosis membrane systems 75–90% of the feed water is conversed to product, and as a consequence dissolved salts are concentrated in the membrane concentrate. Subsequently, sparingly soluble inorganic compounds such as calcium carbonate and barium sulphate become supersaturated and may precipitate in a membrane element (scaling). Scaling leads to an increase in energy consumption and chemical cleaning frequency. From an economic and environmental point of view this is not desired and early detection of scaling is therefore necessary. With the aim to detect scaling in an early stage a scaling monitor has been developed and tested. The scaling monitor, named ScaleGuard®, measures the normalised flux of a small membrane element fed by a part of the concentrate of the full-scale plant. This scaling monitor was tested at a pilot plant treating anaerobic groundwater for the production of drinking water. The monitor demonstrated to be able to detect scaling before it took place in the last stage of the pilot plant.

Within the framework of the project ‘Integrated multi-objective membrane systems for control of microbials and DBP precursors, funded by AWWARF and USEPA, four integrated membrane systems (IMSs) for surface water treatment were investigated. The identified IMSs are based on pretreatments including slow sand filtration, bank filtration, coagulation-sedimentation-rapid filtration (CSF) and ultrafiltration respectively. The IMSs were tested on three locations in the US and three locations in The Netherlands. One of the primary objectives of the project was to compare membrane performance, a conventional NF system applied to ground water was used as a reference. Membrane performance was studied by normalising process data towards the mass transfer coefficient (MTC) and normalised feed channel pressure drop (NPD). Results show that a soil passage is to be preferred, especially when the water can be extracted under anaerobic conditions; such an IMS leads to a low cleaning frequency at low costs. When soil passage is not possible due to local circumstances, ground water conditions can be approximated by slow sand filtration after a CSF pretreatment. Since slow sand filtration is a space consuming technique, an IMS was investigated based on ultrafiltration after CSF pretreatment. After optimization of both CSF pretreatment and ultrafiltration no significant fouling of the RO membranes was observed.

Scaling in reverse osmosis (RO) and nanofiltration (NF) systems is generally controlled by limiting the conversion and/or the addition of antiscalants. An alternative approach might be the desupersaturation of membrane concentrate before e.g. the last stage of an RO and/or NF system. A desupersaturation unit (DU) is aimed at reducing the degree of supersaturation in membrane concentrate stream by forced precipitation of sparingly soluble inorganic compounds on seed crystals. Barium sulfate precipitation was examined in a laboratory scale (DU) using supersaturated synthetic (no organic matter) and natural RO and NF concentrates. Effective barium removal for at least 100 h was observed in the case of synthetic concentrate. Early break-through of barium sulfate crystallization occurred in NF concentrate. This was attributed to the adsorption of organic matter (TOC = 22 mg/l) and subsequent poisoning of barium sulfate seed crystals. Two RO concentrates from surface water treatment pilot plants were compared. Barium sulfate removal was not affected (100 h) by organic matter (TOC = 12 mg/l) in River Rhine water pretreated by ultrafiltration. However, break-through was observed from River Rhine water with organic matter (TOC = 8 mg/l) pretreated by ozonation. Irreversible poisoning of seed crystals was attributed to adsorption of organic compounds, e.g. low molar mass acids produced by ozone pretreatment of RO feedwater. Inactivated barium sulfate crystals were restored by backwashing with acidified (pH = 2) water.

Deposition of particulates, especially colloids, during membrane filtration causes a decline in permeate flux. The existing modified fouling index (MFI) uses a microfilter membrane as a quic test of feedwater quality. The MFI is based on ca e filtration, and thus a model can be developed for flux decline prediction. However, this MFI is not sensitive to the presence of smaller particles. Therefore, more recently the MFI using ultrafiltration membranes (MFI-UF) was developed. This research investigates various criteria of the MFI-UF test for use as a water quality indicator; stability of the MFI-UF over time, linearity of the index with particulate concentration, and reproducibility (1) of the test (reusability of a UF module) and (2) module manufacture. The pressure dependency of the MFI-UF was also examined. The aforementioned criteria were examined using a polyacrylonitrile module of 13,000 molecular-weight-cut-off for low fouling (tap and process water) and high fouling waters (diluted canal water). The MFI-UF was stable over time and directly related to colloidal concentration. The MFI-UF test was reproducible for one module with repeated testing and reproducible module manufacture was found for 80% of tested modules. Ca e compression occurred at higher applied test pressures and using the compressibility factor determined in this research, accurate pressure correction was found.

The successful operation of dead-end ultrafiltration systems depends on the effectiveness of cleaning (hydraulic and/or chemical cleaning). In this study, backwashing conditions such as pressure and duration were studied in order to maximize the net flux per cycle. Monitoring the backwash flux and turbidity of the backwash water showed that the efficiency of backwashing was more dependent on backwashing time than pressure. The effectiveness of backwashing was significantly improved by the use of crossflushing. In general, 99.9±1% flux restoration could be achieved when backwashing was preceded by crossflushing while 94.5±1% could be achieved by backwashing alone. Crossflushing was more effective when the crossflush velocity in the fiber was in the turbulent regime (v ≅ 1.6 m/s). Crossflushing and pressure pulsing can be employed to increase the efficiency of dead-end, hollow fiber ultrafiltration systems because the frequency of backwashing and chemical cleaning, permeate consumption and system down time are reduced.

A desupersaturation unit (DU), placed between the stages of either RO or NF systems is aimed at reducing the degree of supersaturation in the concentrate stream thus preventing membrane scaling in the last stage of a membrane filtration system while allowing conversion to be maximized up to design limits (>90%). Barium sulfate desupersaturation in an up-flow fixed bed reactor was demonstrated for ca. 200 h (Concentrate A). However, barium removal in the DU was shown to be source dependent and early break-through of the DU occurred with Concentrate B after ca. 30 h. Calculations showed that barium sulfate deposition in the DU was affected by natural organic matter (NOM). Blocking of seed crystal surface by rapid adsorption of certain NOM fractions hindered crystal growth and was believed to be the main reason for break-through in the DU. Recycling of these naturally occurring organic molecules as “natural antiscalants” to maximize the conversion of RO (NF) systems may be possible if they can be isolated and characterized.