J.A.M.H. Hofman

Removal of pesticides and other micropollutants with cellulose-acetate, polyamide and ultra-low pressure reverse osmosis membranes

Abstract In 1995 several membrane manufacturers started to sell ultra low-pressure reverse osmosis membranes. The specifications of these membranes indicated that they have rejections for dissolved salts comparable to “conventional” composite (polyamide) membranes, while the required feed pressure to realize a specific production capacity is 30–40% less. This article describes the results of a preliminary study on the performance of these new membranes. The results showed that the rejections and productivity as specified by the manufacturers could be reproduced. Also the results of the spiking experiments with organic micropollutants showed that the ultra low pressure membranes have excellent properties to remove these compounds from water. NV PWN Water Supply Company of North Holland selected these membranes for their new water treatment plant Heemskerk when previously it was determined that cellulose acetate membranes were not suitable for the removal of organic micropollutants.

RO treatment: selection of a pretreatment scheme based on fouling characteristics and operating conditions based on environmental impact

Amsterdam Water Supply investigated the possibilities of increasing the capacity of the River Dune production plant from 70 to 83 million m3/year. Two possibilities concern the direct treatment of pretreated Rhine River water (C/S/F) by reverse osmosis, without soil passage. In scheme 1, the RO feedwater is pretreated by ozonation, biologically activated carbon filtration and slow sand filtration. In scheme 2 only slow sand filtration is applied as pretreatment of the RO feedwater. To avoid scaling problems with BaSO4 and CaCO3, the RO feedwater was acidified with HCl up to a LSI in the concentrate of 0.0 and the RO units were operated with a recovery of 85%. The use of an anti-scalant (Flocon 100) in combination with H2SO4 dosage (LSI in the concentrate 2.0) and a recovery of 90% was also successful to control scaling, but this operation mode resulted in severe biofouling as the anti-scalant acted as a nutrient for microbial regrowth in the membrane elements. Both pretreatment schemes resulted in an excellent RO feedwater quality with respect to MFI, DOC, AOC and BFR (biofilm formation rate) with an accompanying very stable operation of the RO units. Over a period of 11 months the MTC of the RO units only showed a decrease of 16% and 23% for pretreatment schemes 1 and 2, respectively, while the differential pressure remained constant. Pretreatment scheme 1 is preferred above pretreatment scheme 2 as it results in additional capacity related to disinfection and organic micropollutant removal. Also the RO concentrate is easier to be disposed of from an environmental point of view. Possibilities are mentioned to further optimize scheme 1 with respect to chemical costs and environmental impact.

The use of Biological Activated Carbon Filtration for the Removal of Natural Organic Matter and Organic Micropollutants from Water

Amsterdam Water Supply (AWS) uses Biological Activated Carbon Filtration (BACF) for the removal of natural organic matter in general and the removal of organic micropollutants in particular. In order to minimize costs and environmental burden, it is important to know whether successive reactivations of carbon reduces its effectivity, and whether pesticides are effectively removed after prolonged running times of the carbon filters. The first aspect avoids the necessity of carbon replacement (i.e. costs), while the second aspect reduces the reactivation frequency (i.e. environmental burden). In a future extension scheme, AWS considers the use of an Integrated Mebrane System (IMS), and it is important to know whether the application of BACF is beneficial in the IMS. Six years of operation of BACF in the River-Lake Waterworks (31 million m3/year) have shown that successive reactivations do not affect the DOC removal capacity of the carbon. Three years of operation of BACF in the River-Dune Waterworks (70 million m3/year) have shown that the carbon retains its pesticide removal capacity. The use of BACF in an IMS shows important perspectives in minimizing the fouling of reverse osmosis membranes and in minimizing the organic carbon content in the membrane concentrate.

Integrated multi-objective membrane systems for surface water treatment:Pretreatment of reverse osmosis by conventional treatment and ultrafiltration

Within the scope of this project funded by AWWARF and USEPA three very promising integrated membrane systems (IMS) were identified for surface water treatment. This paper will cover some highlights of the research carried out by the Water Supply Company of North Holland and Kiwa on the combination ultrafiltration/ultra low pressure reverse osmosis. Productivity of the ultrafiltration proved to be stable for the first 6000 h. Then, at the end of summer, the backwash regime was not able to prevent accumulation of biofouling. This biofouling coincides with an increase in bacterial counts in the ultrafiltrate. Also the ultra low pressure reverse osmosis unit showed no significant fouling up to 6000 hours. Then a rapid decrease of the second stage MTC started caused by BaSO 4 precipitation, coinciding with an increase in feed-concentrate pressure drop caused by a severe biofouling. Ultrafiltration reduced the indicator organisms E. coli and spores of sulphite reducing Clostridia to under the detection limit. To determine the elimination capacity the feed of UF and ULPRO were spiked with MS-2 phages. Generally UF showed a complete (>5 log) removal. Once only a 2.7 log removal was found caused by defects in UF-membranes. ULPRO showed a high (4.7-4.9 log) although no complete removal. The results show that integrity measurement and monitoring of both membrane steps are essential. UF-integrity is measured by a pressure hold test and monitored by particle counting. Integrity of ULPRO is measured by vacuum testing and monitored by in situ conductivity measurement and particle counting. ULPRO gives a very high precursor removal. Potential measurements (24 h, 20°C, 1-2 mg/l Cl 2 -residual) showed >95% removal of THMFP and HAAFP. IMS with UF and ULPRO are exceptional tools for control of microbials and DBP-precursors. Crucial for application may be control of biofouling in the system.

Retention of herbicides and pesticides in relation to aging of RO membranes

Abstract Amsterdam Water Supply (AWS) intends to increase the capacity of the Leiduin production plant. In the existing plant (capacity 70×10 6 m 3 /y) pretreated Rhine River water was infiltrated in the dune area west of Amsterdam for artificial recharge and after a residence time of approximately 100 days extracted and posttreated to achieve drinking water quality. In the extension (capacity 13×10 6 m 3 /y) the pretreated (coagulation, sedimentation, filtration) Rhine River water was not infiltrated in the dunes, but was treated directly without soil passage. First, an additional pretreatment (ozonation, biological activated carbon filtration and slow sand filtration) was carried out, and finally reverse osmosis (RO) was used for desalination, hardness removal, disinfection and removal of pesticides and others micro-pollutants. Former research on removal of pesticides has already showed a removal of >99.5% with the ozonation/BACF preceding RO. To prove high retention of RO membranes as a second barrier and to examine the influence of aging of the membranes, several dosing experiments were carried out by AWS and Kiwa. During the period before March 1997, Toray SU 710L membranes were used. From then on Fluid Systems 4821 ULP membranes have been used in the RO pilot plant. The RO feed flow is 9 m 3 /h and the recovery is 85%. To compare the removal of pesticides and herbicides with the two different applied RO membranes, six dosing experiments were carried out: two with the Toray membranes and four with the Fluid Systems membranes. A cocktail of pesticides was dosed with a feed concentration of approximately 5 μg/l. The results of the test showed an equal retention for bentazon, DNOC and pirimicarb for both types of membranes. The removal of metamitron and metribuzin was substantially higher with the Fluid Systems membranes. As a result, it was concluded that RO is a second barrier for pesticides in this treatment concept as biological activated carbon filtration in the first barrier. Retention of the Fluid Systems membranes is higher than the retention of the Toray membranes. After 3 years of operation with the Fluid Systems membranes, no pesticide retention decline was observed.

Scaling control of RO membranes and direct treatment of surface water

Abstract Since 1992 Amsterdam Water Supply has been conducting pilot-plant (2 × 9 m3/h) research on the use of reverse osmosis (RO) for direct surface water treatment from the Rhine River. The investigations were made as a feasibility study for an extension with 13 Mm3/y of the existing treatment plant (70 Mm3/y). Research was carried out with ultra-low-pressure membranes and an organophosphonate antiscalant (Permatreat 191). With this antiscalant no biofouling was introduced upon its use. Because of its low environmental impact, sulphuric acid was preferred above hydrochloric acid, but the barium sulfate scaling risk was increased. Experiments were conducted with both types of acid in increasing recovery steps from 80% up to 90%. While starting experimental difficulties with dosing the antiscalant occurred and BaSO4-scale formation was a fact, especially in the installation with sulphuric acid, a severe MTC decrease was seen. Successive cleaning was possible with an alkali cleaner containing a complex builder. From the barium mass balance over this 3-month period, the precipitated barium amount was calculated. Barium analysis of the cleaning solution showed an almost complete recovery of the barium from the membrane elements. Moreover, the MTC was restored completely. From this it was concluded that the precipitated barium sulphate was washed out completely. Recoveries up to 87% are possible without BaSO4-scale formation and run times up to 1 year. The working inhibition time of the antiscalant in concentrate from increasing recoveries was determined experimentally.

Water and energy as inseparable twins for sustainable solutions.

Although the water cycle is only a minor contributor to the energy demand in society, it is a matter of good housekeeping to minimize the energy need within a sustainable water cycle. Wastewater treatment should not only be applied to purify the water, but also recover the energy present in this water, as well as to recover essential elements like nitrogen and phosphorus. From an energy analysis of the Dutch water cycle it is concluded that creating an energy neutral water cycle by using the heat content or by making use of the organic load of wastewater is within hands.

Evaluation of different disinfection calculation methods using CFD

Computational Fluid Dynamics combined with a particle tracking technique provides valuable information concerning residence times and contact times in chemical reactors. In drinking water treatment, for example an accurate estimation of the disinfection is important to predict the microbial safety. Ozone contactors are widely used for disinfection, but the complex geometry of the system causes suboptimal hydraulics and requires optimizations of the flow. This results in a lower ozone dosage, which may reduce the formation of unwanted disinfection-by-products and the consumption of energy. To that end disinfection needs to be calculated precisely, accounting for the complex hydraulics. Several calculation methods estimating the disinfection performance of ozone contactors were evaluated using Computational Fluid Dynamics. For an accurate disinfection prediction, the full distribution of ozone exposures (CT values) is needed, only a mean CT value or residence time distribution provides insufficient information for an accurate disinfection prediction. Adjustments to the geometry of the ozone contactor that reduce the short-circuit flows resulted in an increase in disinfection capacity, whereas the mean CT value remained the same. A sensitivity analysis with respect to the kinetics was conducted. The gain in disinfection capacity obtained by optimizing the hydraulics was significant for typical values used in practice.

Electrodialysis as an alternative for reverse osmosis in an integrated membrane system

Abstract Amsterdam Water Supply investigated the possibilities of increasing the capacity of the production plant “Leiduin” from 70 to 83 million m3/y. Three different integrated membrane systems (IMS) are evaluated for the extension of 13 million m3/y. In IMS 1 pretreated Rhine River water (C/S/F) is treated sequentially by ozonation, biologically activated carbon filtration, slow sand filtration and reverse osmosis (RO). In IMS 2 pretreated Rhine River water is treated sequentially by ozonation, biologically activated carbon filtration, slow sand filtration and electrodialysis reversal (EDR). In IMS 3 pretreated Rhine River water is treated sequentially by EDR, ozonation, biologically activated carbon filtration and slow sand filtration. The IMSs were compared for temperature dependency, scaling and fouling, disinfection strategy and bromate control, dual-barrier concept, and energy and chemical consumption. All IMSs performed well with respect to scaling and fouling. The IMSs using EDR were much more affected by temperature as compared with IMS 1 using RO. IMS 2, in which EDR was used as the final process unit, cannot comply with the bromate standard of 5 μg/l in the finished water. Hence, the choice is between IMS 1 and IMS 3. IMS 1 using RO has the advantage that a dual barrier is present for disinfection and removal of organics (a.o. pesticides). IMS 3 using EDR is characterized by a lower energy consumption (factor 2) and lower chemical consumption (factor 3). The final selection of the optimum process scheme depends on the criteria mentioned above and on investment costs and environmental impact.

Permeability reduction of porous media on transport of emulsions through them

Abstract Permeability reductions, effected in a porous medium consisting of 40–60 μm glass spheres by passage of emulsions, are reported for Na oleate-stabilized emulsions at different electrolyte concentrations, and for an emulsion containeing, Aerosol OT (Na dioctyl sulphosuccinate; AOT) at low electrolyte concentration. All emulsions show a pronounced permeability reduction, but for Na oleate-stabilized emulsions, in the absence of added electrolyte, substantial amounts of disperse phase have to be passed through in order to reduce the permeability to about 35% of its initial value. Addition of electrolyte above the critical coagulation concentration (CCC) leads to more rapid permeability reduction. AOT-stabilized emulsions, even in the absence of added electrolyte, reduce the permeability of the porous medium in a way comparable to that found for Na oleate-stabilized emulsions with [NaCl] > CCC. This is ascribed to coagulation of the AOT-containing emulsion even in the absence of added electrolyte. This coagulation is indeed observed on passing the AOT-stabilized emulsions through the porous medium. The reduced stability of the AOT-stabilized emulsion can be accounted for by the low interfacial tension in lhis emulsion, permitting emulsion destabilization through droplet deformation.