Read Microsoft Word - WISA-MTD 2007 Abstract booklet.doc text version



Symposium and Workshop

Mabalingwe Nature Reserve Limpopo Province 18 - 20 March 2007


The expertise of the entire South African membrane research base should be harnessed and coordinated to optimise the management of South Africa's limited water resources. Future research and development on synthetic membranes should be directed at the development of new, innovative membranes and their cost-effective application in certain pre-selected niches in the broad fields of water purification to potable water standards, the recycling of industrial wastewater and the protection of the general water environment, in order to ensure the future well-being of all South Africans.


Research on synthetic membranes will aim to develop, implement and use cost-effective membranes and membrane systems to purify or treat water for potable use, industrial reuse or the abatement of environmental pollution, in the interests of all South Africans.


To provide a forum to facilitate the exchange of information and experience on membrane technology for water and wastewater management. Organise conferences, symposia, demonstrations and plant visits. Develop educational courses and workshops for academics and membrane end users. Identify research needs and current problem areas. To stimulate awareness of the potential of membrane technology in the user community. Membership is open to all stakeholders in membranes for water and effluent management. Membership is free if you are a WISA member, otherwise WISA membership fees are payable. A conference/workshop is held every two years during which papers, posters and courses are presented. Students and private companies and practitioners are especially encouraged to participate. At the workshop an Annual General Meeting is held, during which the new MTD Management Committee for the following two years is normally elected.




Yet again, we come together to talk about our passion ­ membranes. To first-timers at this symposium, welcome to our membrane family, and I trust that you will be enriched by the experience. One of the key objectives of the WISA MTD Symposium and Workshop is to equip delegates with the necessary tools to make decisions regarding the future of membrane technology in their environment. This symposium will continue to strive to achieve this objective. The opening of the symposium coincides with National Water Week. The theme for this year's National Water Week is "Water is life - Protect our scarce resources". Aligned with this theme, the symposium will endeavour to promote the application of membrane technology in order to conserve water and reduce the pollution of water, thereby lessening the burden on our water resources. Since the last symposium, there have been exciting developments in membrane technology in both the research and the industrial arenas. We can certainly say that the application of membrane technology is soaring. We have an exciting program lined up for the 7th WISA MTD Symposium & Workshop, with a number of international speakers, the latest developments from the research institutions and exciting application reports from the industry. As is customary, the symposium also offers a workshop and a session for poster presentations. This year's workshop is entitled "Membrane Processes in Environmental Technology". The symposium hopes to provide you with an opportunity to network with your peers in the field of membrane technology and to cement relationships. The 7th WISA MTD Symposium & Workshop will maintain the informal style that has become the hallmark of these meetings. I think you will agree that Mabalingwe Nature Reserve provides



a perfect setting for us to focus on our passion, membranes, in the tranquil wilderness overlooking the undulating African bush. At this point, I would like to extend a warm welcome to the WISA President and the WISA Chief Executive to our symposium. I would like to thank our presenters for their efforts in putting together top quality papers. I also wish to thank our sponsors for their support. Their contribution has made it possible for us to keep the symposium reasonably priced. And last, but not least, I would like to thank the WISA MTD committee members for their selfless volunteer efforts to make this symposium a success. Have a great Symposium and Workshop!




Sunday 18 March 2007 15:00 - 18:00 18:00 ­ 19:30 19:30 Early Registration Icebreaker Snacks and Cocktails Meet & Greet Dinner

Monday 19 March 2007 Session Chair: Stéphanie Marais

07:30 08:30

Registration Welcome and Opening Address Marlene Van der Merwe-Botha - WISA President Plenary Lecture 1 Understanding fouling in a biofilm membrane bioreactor for municipal wastewater treatment TorOve Leiknes (EMS/NTNU, Norway)


Session 1: Fouling Session Chair: Stephan Roux


Paper 1

The cleaning of the membranes in a small spiral wrap filter plant, using infrasound pulses D. McLachlan (University of Stellenbosch) R.O. Biofouling control : Case histories describing the successful transitioning from oxidising to non-oxidising icrobicides P.J. Allison (Buckman Laboratories)



Paper 2



Paper 3

Membrane fouling, visualization and infrasound cleaning G. Offringa (Water Research Commission) (In memory of Prof. F. Reineke)

10:40 ­ 11:00

Tea Break

Session 2: Fouling / Desalination Session Chair: Stephan Roux


Paper 4

The cleaning of various foulant layers from a number of membranes in a flat bed cell, using magnetic beads D. McLachlan (University of Stellenbosch) Fouling control in a submerged flat sheet membrane system: Bubbling and hydrodynamic effects. V. Ndisini (Durban university of technology) Economic evaluation of a new ultrafiltration membrane for pretreatment of seawater reverse osmosis F. Knops (X-Flow, Netherlands)


Paper 5


Paper 6




Workshop Membrane processes in environmental technology Japie Schoeman (University of Pretoria) AGM / Tea Break Poster Presentations Depart for game drive (The game drive will end at the dinner venue) Symposium dinner and Entertainment (In the bush at Kalahari Oasis, the relic of TV film set)

15:00-15:30 15:30-16:30 16:45

18:30 for 19:00



Tuesday 20 March 2007

Session Chair: Stéphanie Marais

05:45­07:30 08:30-08:45 08:45-09:30

Game Drive Start of day's proceedings / registration Plenary Lecture 2 Membrane technology for the separation of micropollutants from salts Wouter Pronk (Eawag, Switzerland)

Session 3: Membrane bioreactors Session Chair: Mesuli Mbanjwa


Paper 7

The characterisation of anaerobic feed solutions for flat sheet membranes S. Pillay (University of Kwazulu-Natal) Oxygen mass transfer kinetics for an immobilised biofilm of Phanerochaete chrysosporium on a polysulphone capillary membrane S.K.O. Ntwampe (Cape Peninsula University of Technology) Momentum transfer inside a vertical orientated capillary membrane bioreactor B. Godongwana (Cape Peninsula University of Technology) Dutch innovations and lessons learnt in MBR facility design for industrial and municipal wastewater treatment A. Giesen (DHV, Netherlands)


Paper 8


Paper 9


Paper 10


Tea Break



Session 4: Membrane materials/Wastewater treatment Session chair: Chris Swartz


Paper 11

Developing a new biopolymeric material for use as an electrolyte separation in fuel cells A. Masala (Vaal University of Technology) Influence of hydrophilic modification via blending of a branched, block copolymer on polysulphone ultrafiltration membrane structure, performance and anti-fouling characteristics S. Roux (University of Stellenbosch) Membrane-assisted solvent extraction components from agro-industrial wastewaters C. Garcin (University of Cape Town) of valuable


Paper 12


Paper 13


Paper 14

Immersed membrane filtration : Local developments L. Pillay (Durban University of Technology) Production of demineralised water from cooling tower blowdown J. Van Aartsen (VWS Envig)


Paper 15

13:00­14:00 LUNCH

Session 5: Wastewater treatment /General Session Chair: Gerhard Offringa


Paper 16

Recovery of cooling tower blow-down using Cold Lime Softening, Ultrafiltration and Reverse Osmosis ­ A chemical treatment perspective H.W.J. Els (Buckman Laboratories) Nanofiltration and reverse osmosis in soft drink production W. Taljaard (VWS Envig) Data normalization as a tool to evaluate the field performance of reverse osmosis membranes S. Peter (Dow, Germany)


Paper 17

14:40 - 15:00

Paper 18



15:00 - 15:20

Paper 19

Evaluation of diffusion dialysis for acid recovery from spent solutions J.J. Schoeman (University of Pretoria) Upgrading Tutuka power station's desalination plant M. Pryor (Keyplan)

15:20 - 15:40

Paper 20

15:40 - 16:00 16:00

Discussion on international conference CLOSING OF SYMPOSIUM BY WISA-MTD CHAIRPERSON




Poster 1:


J. Ali, F. J. Reineke and R. D. Sanderson UNESCO Associated Center for Macromolecules & Materials, Department of Chemistry and Polymer Science, University of Stellenbosch, P. Bag X1, Matieland 7602, South Africa

Poster 2:


S. Pillay *, K.M. Foxon and C.A. Buckley Pollution Research Group, School of Chemical Engineering, Howard College, University of KwaZulu-Natal, Durban, South Africa, 4041 * Corresponding author: Tel.: +27 31 260-3561; Fax: +27 31 260-3241 Email address: [email protected] (Sudhir Pillay)

Poster 3:


P. Swart, ; S. Govender ; M.W. Bredenkamp and E.P. Jacobs

1 2 1 1 2 2

Dept. of Biochemistry, Dept of Chemistry and Polymer Science, University of Stellenbosch, Stellenbosch, 7600

Poster 4:


Z. Allie , N. Kolar , E.P. Jacobs and P. Swart

1 2 1 1 2 1

Department of Biochemistry & Department of Chemistry (Polymer Science), University of Stellenbosch, Private Bag X1, Matieland, 7602, South Africa * Corresponding author. Tel: +27218085883/8085877; Fax: +27218085863; E-mail: [email protected]

Poster 5:


A. Lourens and I.W. vd Merwe Proxa (Pty) Ltd, PO Box 43, Paarl, 7622, Tel +27-27-872-0089 ([email protected])

Poster 6:


G. Rencken , T. Els , B. Sauvet-Goichon

1 1 1 2

VWS Envig (Pty) Ltd, S04 Modderfontein House, High Street, Modderfontein, South Africa, Tel. +27 11 663 3600; Fax +27 11 608 4772 2 Veolia Water Technical Division, St.Maurice, France



Poster 7


B. Hendry Chemical Engineering, CPUT, P.O. Box 1906, Bellville campus, 7535 Tel: +27 21 959 6045, Fax: +27 21 959 6165, Cell: +27 82 200 6898; email: [email protected]

Poster 8


V. Mishra Chemistry Department, UWC, Moderdam Road, Bellville, 7535, Tel: +27 21 959 3878, email: [email protected]





TorOve Leiknes NTNU-Norwegian University of Science and Technology, Department of Hydraulic and Environmental Engineering, S.P. Andersensvei 5, N-7491 Trondheim, Norway, Corresponding author: [email protected], +47 7359 4758

An alternative treatment scheme to the AS-MBR is combining a biofilm reactor with membrane filtration for enhanced biomass separation (BF-MBR). A critical and inherent feature in any membrane process is fouling of the membrane. Understanding and controlling membrane fouling will therefore be a central issue in developing the process. Fouling is caused by different substances (i.e. solids, particles, dissolved constituents) and the mechanisms are rather complex and interrelated. The objective of this study has been to gain better understanding of membrane fouling in a BF-MBR. Pilot plant studies have been conducted to investigate the performance of the membrane unit as function of operating conditions and to make an assessment of the dominant fouling mechanisms. Fouling rates were found to be affected by the operating conditions of the treatment scheme; low organic loading rates appear to give a better performance compared to high organic loading rates. Characterization of the foulants show that fouling correlates with the submicron particle fraction in the membrane reactor. Investigations are being undertaken to determine the variation of these particles as a function of operating conditions and their composition in more detail. This paper will present the assessment of membrane fouling in a BF-MBR and thus contribute to a better understanding of membrane fouling in MBRs in general.



Session 1 ­ Paper 1


D.J. Koen, F.J. Bester, D.S. McLachlan and R.D. Sanderson Department of Chemistry and Polymer Science, University of Stellenbosch

Introduction The routine shutting down of filtration plants for cleaning, using chemicals, is a slow and costly procedure. The disposal of the chemicals can also present a problem. Therefore it is desirable to use physical methods of cleaning the membranes, which does not involve shutting down the plant. This paper will describe a system in which successful experiments have been performed , using back pulsing into the permeate space, which holhs the system at over 905 of its clean water value and will clean an already fouled membrane. Experimental Apparatus and Procedures A small spiral wrap filter plant which can contain up to 3 elements, in series or parallel, is described. The plant has a Mono cloudy water 10m per hour pump and digital flow and pressure meters in the feed and permeate lines. There are also pressure transducers at the outlet of the permeate space, where the infrasound pulses are introduced, and at the input of the feed line, where the pulse that passes thought the membrane can be monitored. The two transducers are coupled to a dual beam oscilloscope, the output of which can be recorded to give the pulse shapes. The pulses come from a pressurized reservoir and are released using a fast acting magnetic solenoid valve driven from a function generator into the permeate space. Results and Discussion The results of a number of runs using a 2&1/2" Polysulphone element, under various conditions, are given below. In all cases a Dextrin (0.5 g/l) in RO water solution with feed pressure of 100 Kpa is used. After each run the spiral wrap filter is chemically cleaned back to a clean water flux of about 80 L/h. The best results so far are with 350 Kpa in the pressure reservoir and pulses, just over 100 milliseconds long, which are applied every second. As can be seen in Fig. 1, the permeate flow rate stabilizes at 92%




of its clean water value. Fig 1 also shows that if the pulsing is switched off the permeate flow rate rapidly drops towards its "no pulsing" equilibrium value of 22%. However, when the pulsing is recommenced the permeate flow rate rapidly rises to 82% and reaches 92% again after 50 minutes.



Run on Dextrin

Sol.valve 350kPa





Flux (L/h)






0 0 100 200 300 400 Time (min) 500 600 700 800 900

Figure 1. Permeate flux for the 350kPa cleaning pulse at 1 Hz run Further details will be given in the presentation, but it appears that. from the preliminary results using a dextrin solution and a polysulphone membrane, that it maybe be possible to optimize this "infrasonic" process in a spiral wrap plant to give net flow rate over several hours, days, weeks or months of better than 90% of its clean water value. Further tests with different `non model" solutions and membranes remain to be made to see if the initial promise shown by this system is fulfilled.



Session 1 ­ Paper 2


P.J. Allison , A. Maartens , E. Rava , S. Steenekamp , M.P. Augustyn , P.G. Boshoff , and P.J. Scurr

1 2 3 4 1 1 1 1 2

Buckman Laboratories, Hammarsdale, South Africa; Environmental Science and Technology,

3 4

Sasol Secunda; Wastewater Desalination, Sasol Synfuels, Sasol Secunda; Effluent Treatment, Columbus Stainless, Middelburg

Background Biofouling in reverse osmosis plants is practically inevitable and can be directly linked to as much as 56 to 74% of costs of membrane operation (Kelle Zeiher and Phillip, 2000). Chlorine-based biocides have been used to reduce the microbiological fouling potential in Reverse Osmosis (RO) systems. Reference literature revealed that increased biofouling tendencies, contribution to membrane deposition/ blinding, and higher cleaning in place (CIP) costs were some of the adverse effects that have been reported with chlorination/dechlorination (Baker and Dudley, 1999). This paper presents case histories of successful conversion from chlorine to non-oxidising microbicides. Case Histories Rapid development of biofilm resulted in the need to change the conventional chlorination/dechlorination disinfection programme of a 8 Ml/day RO plant at a refinery in South Africa. High sessile counts of 10

2 6

colony forming units (CFU) per cm of membrane were measured. A non-oxidizing microbicide treatment programme, with proven compatibility with polyamide membranes, was recommended based on a successful case history in the USA (Ortega, et al., 1997). Microbial populations were monitored using ATP measurements, heterotrophic plate counts (HPC), population surveys, bacterial identifications and microbicide kill studies. In addition, a side-stream flat sheet membrane monitoring unit (Maartens et al., 2002) was installed to permit early detection of biofouling. A key factor that contributed to the excellent results achieved included the use of alternating biocides with different active ingredients to broaden the spectrum of control and minimise microbial resistance. In addition, regular CIP's and good operational monitoring and control are essential supplements to the anti- microbial chemicals.



Chlorination was necessary to control algae and bacterial growth in the RO supply water used to treat effluent from a stainless steel production plant in South Africa. Excessive blinding of the pre-filtration system and poor operational stability were being experienced. Based on the success at the refinery described above, a proposal was submitted to replace chlorination with the combination of non-oxidizing microbicides that had given favorable results. Once again, biofouling control was better with the replacement programme. Conclusions Application of non oxidizing biocides for feedwater disinfection has numerous benefits over chlorination and dechlorination. References Baker, J. S. and Dudley, L. Y., Water Supply (1999), 17(1), 173-181 Kelle Zeiher, E.H. and Philip Y.U., Ultrapure Water, March (2000), 55 ­ 64 Maartens, A. Jacobs E.P. Peet, D. and Augustyn, M.P., US Patent 2002/0134716 A1, September, 26, (2002). Ortega, C. Rossing, A. Cornstock, D. and Condon,P.G., Ultrapure Water, October (1997) 31-36.



Session 1 ­ Paper 3


F.J. Reineke, J. Ali, M.B. Mbanjwa, R.D. Sanderson Department of Chemistry and Polymer Science, University of Stellenbosch

Introduction Brown-colored surface water found in the southern coastal belt of South Africa (between Cape Town and Port Elizabeth) is a potential source for the supply of drinking water. However, this colored water is viewed as unacceptable for potable use for a number of reasons. The color is unacceptable for aesthetic reasons. The complexes formed after combining organic matter with heavy metals and pesticides may be potentially carcinogenic. The higher dosages of oxidants required for effective disinfection and formation of disinfection by-products (such as organo-chlorine compounds upon chlorination) are conceived to be health hazards. Corrosion of metal pipes and other materials occurs because of the low pH of this water. Using the organic compounds as the nutrient source, there is rapid growth of microorganisms that reduces the water quality in purification plants and distribution systems. The color of natural brown water is produced mainly by the presence of natural organic matter (NOM), a complex mixture of organic compounds, derived from the plant residues of bacterial and fungal origin, soil, peat bogs and sediments. The major fraction of NOM is composed of humic substances, mainly humic acid (soluble in water at pH>2) and fulvic acid (soluble in water at any pH). Humic substances represent over 50% of the total dissolved organic carbon in brown water. They comprise stable ring structures and carboxylic acid containing phenolic functionality. Membrane filtration of aqueous effluents is a widely used technique in water purification, in the food industry and other industrial processes. The fouling of membrane filters during the filtration process of aqueous effluents containing organic matter is a known phenomenon. The fouling of the membranes causes a decrease in flux and eventually the filtration process becomes inefficient. Experimental Apparatus and Procedure The fouling of the filters was monitored in situ by ultrasonic time domain reflectrometry using flatbed Nylon filter membranes. This technique was applied for NOM water and beer effluent. Beer effluent was used as a reference substance for the NOM water because the beer brewing process is temperature and pressure



controlled. Since beer is only brewed from barley, hops, yeast and water the organic contents are well controlled. For the filtration experiments beer effluent was used that still contained life yeast cells. The application of pulses in the infrasound regime during the filtration is applied in order to avoid the fouling partially. This leads to better flux characteristics and hence to a longer life time of the filters. The filtration was conducted under dead end conditions, i.e. the entire effluent had to pass through the membrane. A comparison of the contents of the feed effluent and the permeate were analyzed in order to get a better understanding about the fouling process itself. Results The results that will be presented show a flux enhancement of up to 60% compared to dead end filtration without pulsing.



Session 2: Paper 4


D.S. McLachlan, F.J. Bester and G. Sauti Department of Chemistry and Polymer Science, University of Stellenbosch

Introduction The routine shutting down of filtration plants for cleaning, using chemicals, is a slow and costly procedure. The disposal of the chemicals can also present a problem. Therefore it is desirable to use physical methods of cleaning the membranes, which does not involve shutting down the plant. This paper describes the use of polymer Hematite beads, with a permants magnetic moment, which are moved using magnetic fields so as to produce a cleaning action on the surface of a fouled or fouling membrane. Apparatus and Experimental Procedures The magnetic material encapsulated in the polymer beads was 1-3 µm Hematite (Fe3O2) powder, which when encapsulated and poled at a field in excess of 7000 Gauss produced strong permanent magnetic beads. Note that if the beads are to fit through even the larger spacer cloths of spiral wrap filters they must be 50 µm or less. In practice we took 20 µm to be the lower limit in order that the beads had a sufficiently large magnetic moment to have an effective cleaning action. The first method used to produce the beads (PS spheres) was polymerization of polystyrene in a reactor in which the Hematite was kept suspended by the stirrer. This method gave only a limited fraction of spheres which were hard to separate from the cluster like powder particles. A second method tried was to blend the magnetite powder into polyethylene and to extrude the product. This was successfully done and the spaghetti like product was cryogenically ground down. A large fraction of the granular product had the correct volume but we were unable to satisfactorily convert the rough granules into spheres (PE granules). The most successful method was to use phase inversion. For this purpose the Hematite was first coated with silane and then blended into a Poly Acryl Nitride solution. The solution was sprayed, through a fine nozzle, into a bath of 75% acetone and 25% water and the spheres (PAN spheres) formed as they entered the solution and precipitated to the bottom..



The experiments were initially carried out in the flat cells used in another project. This consisted of an inlet and outlet manifold which passed the feed water over a flat membrane which rested on a sintered bronze section. On the other side of this is the permeate outlet. This design proved unsatisfactory at high flow rates and the design (to be illustrated) was modified. Polysulfone ultra filters and nylon (0.45 micron) micro filters are used in this investigation. The foulants used are alumina powders, with narrow specific size ranges, and commercial yeast which has a cell size of 30 microns. However, as this contains some nano protein and sugar molecules on its surface, the yeast was washed to (partially?) remove these nano particles. . Initially it was thought that if the magnetized spheres were rotated in a uniform AC field a sufficient cleaning action would be achieved. For this purpose a Helmholtz coil was constructed and the AC field provide by a function generator driving a 2500Watt power amplifier at 10 Hz. Continuous peak to peak field of just over 100 Gauss were produced. However it soon became clear that because the density of the spheres was just over one, when in water there was not sufficient downward force to produce a cleaning action. We first tried a combination of homogeneous (Helmholtz coil) AC fields and inhomogeneous field< to produce a downward force, using permanent magnets. This was not successful because when the inhomogeneous field was strong enough it prevented the beads for being rotated by the available homogeneous field. We therefore moved to AC inhomogeneous fields above a rectangular block (face 60x70 mm, length 200mm.) of SiFe transformer material around which a rectangular coil of Copper wire, which carried the AC current, was wound. The idea was that the dipoles would be attracted towards the membrane by the inhomogeneous field and rotated by the homogeneous component. Therefore in the experiments to be described an AC inhomogeneous field, with a frequency of 10 Hz, is used to both pull the particles onto the membrane and to rotate them to give a cleaning action. At first we only achieved partial success the fields generated in this way were not sufficiently large due to the large demagnetization coefficient of the short SiFe blocks. This has been overcome in a design which is essentially that of a standard transformer but with a split central core (to be illustrated). The results form this system will be reported at the conference. In the experiments described below the magnetic beads were circulated by catching them on a filter, place near the output side of the feed, from where they were circulated back into the feed water using a small pump or with high flow rates the venturi effect. Results The first cleaning experiments were done in the original design cells with no flow across the membrane, which meant that the beads did not have to be recycled. The experiments were done using unwashed yeast and a SiFe Ac magnet. After a foulant layer was formed using the appropriate feed water and a trans-membrane pressure of 40kPa, the cell was isolated, the beads were activated. The yeast layer was observed to be removed in that the water in the cell became cloudy, which was a hopeful sign. In the next experiments, in the original cell, the water flowed over the membrane at 4 cm/sec and while a bead re-circulation system was being evolved, the beads were kept over the membrane by slowly swept

WISA-MTD 2007 20

inhomogeneous field form a strong permanent magnet. With this system we found that we could "visually" remove an unwashed yeast foulant layer, but that the flow rate did not improve. We concluded that this was due to the nano particles on the surface of the yeast blocking the pores of the membrane. For this reason we obtained alumina grinding powder with narrow size distributions centered on 0.1 µm, 0.3 µm and 1 µm. The experiments with the Alumina using the above procedure were very disappointing. The flow rate dropped (e.g. from 60 to 16 ml/min in 15 mins.) but did not improve when the system was switched to clean water. If the trans-membrane pressure was removed the flow rate improved from 16 to 22 ml/min during which period the Alumina was observed to come off the membrane. . We then did experiments where the beads were continually circulated (small pump) and activated from the time that we switched from clean to feed water. It was hoped that these steps would aid the cleaning action of the beads. Unfortunately this did not prove to be the case as the beads became trapped in the Alumina (all three sizes) with the flow rate dropping continuously (e.g. from 66ml/min. to 4 ml/min in 15 minutes). When the system was switched to clean water there was no improvement of the flow, although there was a continual circulation of the alumina as evidenced by the milky colour in the recirculation tubes, Experiments at higher flow rates (about 20 cm/sec) with the foulant introduced with the beads already activated did not improve matters. Using washed yeast and a trans membrane pressure of 40 kPa., the flow dropped from 50 to 15 ml/min in 15 minutes and improved to 18/ml/min. when clean water was introduced. However, using Alumina (all sizes) the flow rate dropped (e.g. 64 to 15 ml/min at 40kPa.), but never recovered. Conclusion The experiments must be repeated in higher AC fields.



Session 2: Paper 5


N.V. Ndinisa , A.G. Fane and D.E. Wiley

1 2 2

1. Department of Chemical Engineering, Durban Institute of Technology, Durban, South Africa, 2. UNESCO Centre for Membrane Science and Technology, University of New South Wales, NSW 2052, Australia

KEYWORDS Submerged membrane bioreactor; Aeration; Flat sheet membrane; Membrane baffles; Microfiltration; Intermittent filtration; Wastewater treatment

ABSTRACT Submerged flat sheet membranes are mostly used in membrane bioreactors for wastewater treatment. The major problems for these modules are concentration polarization and subsequent fouling. By using gas-liquid two-phase flow, these problems can be ameliorated. This study aimed to optimize the use of gas-liquid two-phase flow as a fouling control mechanism for submerged flat sheet membrane bioreactors. The effect of various hydrodynamic factors such as airflow rate, nozzle size, intermittent filtration, channel gap width, feed concentration, imposed flux and the use of membrane baffles were investigated. Experiments conducted on model feeds showed that fouling reduction increased with air flow rate up to a given value and beyond this flow rate no further enhancement was achieved. The effect of bubbling was also found to increase with nozzle size at constant airflow. Using intermittent filtration as an operating strategy was found to be more effective than continuous filtration and it also reduced energy requirements. The study showed the importance of the size of the gap between the submerged flat sheet membranes. As the gap was increased from 7 mm to 14 mm, the fouling became worse and the degree of fouling reduction by two-phase flow decreased by at least 40% based on suction pressure rise (dTMP/dt). This is the first study which has reported the effects of baffles in improving air distribution across a flat sheet submerged membrane. It was found that baffles could decrease the rate of fouling by at least a factor of 2.0 based on the dTMP/dt data, and could significantly increase critical flux.



Session 2: Paper 6


F. Knops , S. van Hoof and R. Te Lintelo

1 2 1 2 3

X-Flow B.V. (The Netherlands) [email protected]; NORIT Process Technology B.V. (The Netherlands) [email protected]; NORIT X-Flow (The Netherlands) [email protected]


KEY WORDS desalination, sea water, pretreatment, economics, ultrafiltration, capillary membrane

ABSTRACT Numerous tests around the world have proven that ultrafiltration (UF) provides optimum pretreatment for seawater desalination based on reverse osmosis membranes (SWRO). Ultrafiltration will remove all suspended solids and will provide a substantial reduction in micro biological activities. Plugging of RO spacers is completely eliminated and the RO cleaning frequency can be substantially reduced. The main obstacle against use of ultrafiltration membranes for SWRO pretreatment has always been the higher operating cost of ultrafiltration when being compared with conventional pretreatment. Up until now this higher cost has prevented implementation of UF in all but a few SWRO plants. A new membrane has been designed with the aim of tailoring it towards lowest whole of life cost when compared to historical methods of pre-treatment. Typical operating conditions have been used to quantify the following parameters for a typical large scale desalination system (UF + SWRO): · Amortization of investment in UF membranes and equipment. · Operating costs of the UF system. · Reduction in operating costs of the SWRO desalination plant, when being compared against historical pretreatment + SWRO system. · The increased output of the SWRO desalination plant due to higher availability and shorter construction time.



The total whole of life cost of a UF based SWRO desalination plant has been determined (expressed in A $ per m of water produced). Taking all factors into account the total whole of life cost of a dual membrane desalination plant (Seaguard UF + SWRO) will be 2 to 7% lower than the total whole of life cost for a SWRO plant based on historical pretreatment. This paper will describe the development efforts of the new ultrafiltration membranes and will provide the latest information of plants in operation in Turkey and the Middle East.






Wouter Pronk SWW Eawelg, Sietzerland, Tel: + 41 44 823 5381, Fax: +41 44 823 5389, email: [email protected]

In the drinking water as well as the waste water sector, several membrane applications exist where the separation of salts from micropollutants is required. For example, in the treatment of drinking water, reverse osmosis leads to removal of micropollutants, but the salts are removed simultaneously which affects the taste and the nutritional quality of the water. In waste water treatment, recycling of micropollutants can be advantageous in order to increase the retention time and thus the degradation rate in the activated sludge, but a separation step is required which separates the salts from the micropollutants in order to prevent salt accumulation in the bioreactor. A third type of application concerns the treatment of source separated urine, which gains increasing attention in Europe because of the potential to control micro- and macropollutants at the source. Special sanitary equipments and dual waste water systems are applied in order to collect urine. In general, the urine-based products are intended to be applied as fertilizers in agricultural applications. This requires the removal of micropollutants from the salts. For the latter application, the processes of Electrodialysis and Nanofiltration were evaluated. Several nanofiltration membranes were screened, and the influence of pH and the matrix was evaluated. Long term-experiments were performed with electrodialysis, showing that membrane adsorption and retention are dependent from the compound structure and the run time. Furthermore, results will be presented from a pilot plant for the treatment of urine including electrodialysis and ozonation.



Session 3: Paper 8


S. Pillay , V. Yvenat , C. Cabassud , K.M. Foxon and C.A. Buckley

1 1 2 2 1 1

Pollution Research Group, Chemical Engineering, Howard College, University of KwaZulu-Natal (UKZN), Durban, South Africa, 4041


Institu National des Sciences Appliquées (INSA), Toulousse, France

* Corresponding author: Tel.: +27 31 260-3561; Fax: +27 31 260-3241; Email address: [email protected] (Sudhir Pillay)

Introduction This study forms part of a larger European Union project, EUROMBRA, which seeks to develop sustainable solutions for municipal wastewater treatment based on membrane bioreactor (MBR) technology. The focus of this study was to evaluate the fouling potential of anaerobic feed solutions, with specific emphasis on the fluid from an anaerobic baffled reactor (ABR), the reactor type chosen for membrane integration. A test cell technique standardised among EUROMBRA partners was used to assess the fouling potential of different biological solutions and involves the filtration of a biological supernatant in the dead-end mode under a range of constant transmembrane pressure (TMP). The advantage of this approach is that it is short in duration, easy to replicate, utilises a small proportion of test solution and requires a small area of membrane, which is representative of a large flat-sheet membrane . Materials and Methods Source of feed solutions Sludge (2.5 L) was collected from three sources; (1) secondary anaerobic digester from a municipal wastewater treatment plant (WWTP) treating wastewater composed primarily of domestic wastewater, (2) compartment 2, 3 and 4 of an anaerobic baffled reactor treating domestic wastewater only, and (3) aeration tank from a municipal WWTP treating wastewater composed primarily of domestic wastewater. Fouling ability of flat sheet membrane (FAFS) apparatus A fouling ability of flat sheet membrane (FAFS) apparatus was used to characterise the nature of the biological fluid. The experimental set-up consists of a pressure gauge, a pressurised vessel and a dead1



end stirred cell (AMICON model 8400). A polyethersulfone membrane (AMICON PM10) was used as the test membrane. The supernatant obtained after ultracentrifugation (10 000 G for 15 min) of sludge was used as the test solution. The FAFS apparatus was operated under a range of constant transmembrane pressure (TMP), with each step lasting 10 min. The membrane resistance before and after each run were determined using mQ water, with chemical (0.5M NaOH for 1 h) and mechanical cleaning (rinsing with mQ water) between each test. A total of two membranes were used for the study; one for anaerobic digester sludge tests (done in triplicate) and the other for the remaining supernatant tests (done in duplicate, total of eight tests). The resistance-in-series model (Darcy's law) was then used to evaluate the fouling characteristics. Standardised physico-chemical analyses were used to characterise the feed solution . Results and Discussion The relationship between permeate flux and TMP for the different tests solutions is presented in Figure 1. The initial membrane resistance and final membrane resistance for this set of experiments is presented in Table 1. Table 1: Initial membrane resistance and final membrane resistance for each test




Membrane resistance ( Rm ) [(m ) x 10 ] C2 test 1 C2 test 2 C3 test 1 4.6 x 10 1.9 x 10 4.1 x 10 5.1 x 10 6.3 x 10 6.5 x 10 4.2 x 10 7.5 x 10 1.0 x 10

12 -1 12

Final resistance [(m ) x 10 ]

-1 12

30 Permeate flux (L/m^2.h)



2.7 x 10 4.0 x 10 5.2 x 10 7.3 x 10 7.8 x 10 8.0 x 10 7.6 x 10 1.3 x 10 1.3 x 10






12 12


0 0 20 40 60 TMP

C2 test 1 C3 test 1 C4 test 1 AD sludge Act. sludge 2 C2 test 2 C3 test 2 C4 test 2 Act. sludge 1

C3 test 2

80 100 120 140



C4 test 1 C4 test 2 AD sludge Act. sludge test 1 Act. sludge test 2





Figure 1: Permeate flux rates for different test solutions. C2, compartment 2; C3, compartment 3; C4, compartment 4; AD, anaerobic digester sludge. Error bars represent standard error.







, mean value of 3 tests



The supernatant from compartment 2 was found to have the lowest fouling potential whilst the supernatant from the secondary anaerobic digester had the highest [as evaluated by permeate flux and fouling velocity rates (not shown)]. The fouling potential of compartment 4 and activated sludge tests could not be properly assessed as the initial membrane resistance was higher than those reported from previous tests. The increase in the initial membrane resistance was hypothesised to be the result of irreversible fouling as time progressed, despite chemical and mechanical cleaning between each test. The results suggest that the membranes used in this study have a limited lifespan, which from this set of experiments, was estimated to be a maximum of three fouling tests per membrane. Despite the limited lifespan of the membranes, the experiments in this study were relatively easy replicate (provided the membrane was `fresh') and provided useful insight into the fouling characteristics of a particular fluid intended for flatsheet membrane treatment. References 1. JIMENEZ DB, NARBAITZ RM AND MATSUURA (2004) Membrane fouling test: Apparatus evaluation. Journal of Environmental Engineering, 130, 90-99. 2. APHA-AWWA-WEF (1998) Standard Methods for the Examination of Water and Wastewater. 20 U.S.A. p. 2.26­9.74. 3. CABASSUD C, GUIGUI C, AND POLLET S (2006) Confidential EUROMBRA report, D10: Common methodologies. INSA, Toulouse.


Edition. (Edited by Clesceri L.S., Greenberg A.E. and Eaton A.D.). United Book Press Inc., Baltimore,



Session 3: Paper 9


S.K.O. Ntwampe, M.S. Sheldon* and H. Volschenk Department of Chemical Engineering, Cape Peninsula University of Technology, P.O. Box 652, Cape Town, 8000, South Africa. Corresponding author. Tel.: +27 21 460 3430; Fax: +27 21 460 3282, Email: [email protected]


Keywords: mass transfer, oxygen microsensor, diffusion, membrane bioreactor, biofilm, Phanerochaete chrysosporium

Abstract Submerged, batch fungal systems are widely used for the production of secondary metabolites, with applications ranging from bioremediation to pharmaceutical industries. A novel membrane gradostat reactor system was developed for continuous production of secondary metabolites and was shown to be effective in this regard as it mimics the environment in which most microorganisms live. Even though oxygen mass transfer is widely studied in biofilms, it has not been studied within the membrane gradostat bioreactor system. Single fibre capillary membrane bioreactor systems were used to cultivate P. chrysosporium and were operated for 264 hrs. During the experiments, bioreactor systems were taken off at time intervals of 72, 120, 168, 216 and 264 hrs to measure the distribution of oxygen at a specific point within the cultivated fungi. A Clark oxygen microsensor was used to determine the dissolved oxygen distribution at depth intervals of 10 m. The profiles were then used to quantify mass transfer parameters in relation to biofilm thickness incubation time. Fungal biofilm thickness increased from 912 µm after 72 hrs, to 2246 µm after 264 hrs of bioreactor operation. Dissolved oxygen penetration was determined at an average of 310 µm to 390 µm. A nutrient film was clearly evident when the dissolved oxygen profiles were measured and it was determined to be in the range of 10 µm to 30 µm. An increase in the anaerobic zone was clearly



established in the immobilised biofilms. The dissolved oxygen penetration ratio decreased from an average of 0.34 after 72 hrs to 0.13 after 264 hrs. A decrease in the immobilised biofilms porosity of the biofilms was also determined.



Session 3: Paper 10


B. Godongwana, M.S. Sheldon and D. Solomon

* *

Department of Chemical Engineering, Cape Peninsula University of Technology, Cape Town Campus, P.O. Box 652, Cape Town, 8000, South Africa. * Corresponding author: Tel: +27(21) 460 3160; Fax: +27(21) 460 3282; Email: [email protected]

Abstract Considerable research has been done in developing economically viable membrane bioreactor systems for the cultivation of fungi for potential industrial applications. This research, however, has been hampered by the lack of a thorough understanding of the kinematics as well as the dynamics of the flow through these devices. Previous analyses of momentum transfer in membrane bioreactors have been based on steady-state conditions, assuming a horizontally orientated bioreactor and ignoring the different modes of operations of membrane bioreactors. In this article, analytical and numerical solutions to the Navier-Stokes equations were developed for the description of hydrostatic pressure and velocity profiles in a single fiber capillary membrane bioreactor (SFCMBR). The mathematical expressions developed were proposed as solutions to transient-state, laminar, incompressible, viscous and isothermal flow through a membrane with a constant hydraulic permeability. These profiles were developed for the lumen and shell sides of the SFCMBR, taking into account the osmotic pressure effects that occur when solute particles are rejected on the surface of the membrane. These expressions were based on a similarity solution since the wall Reynolds number, Rew<1. The transient-state models developed are applicable to different orientations as well as different reactor configurations. A numerical scheme was also developed to complement the simplified analytical models. The partial differential equations were solved in steps of time using a revised form of the SIMPLE algorithm. The models presented in this paper give similar results to those in previous publications when ignoring gravitational and osmotic effects. The transient-state solutions converge to give profiles which are in agreement with experimental results. Also, in the limiting case of the hydraulic permeability





flow resembles that of a straight circular tube of constant cross section (Hagen Poiseuille flow). Convergence of the numerical solution is influenced by a number of factors and one of the major factors was found to be the grid spacing.



Session 3: Paper 11


A. Giesen*, A. van Bentem* and G. Gademan** *DHV B.V., Box 1132, 3818 EX Amersfoort, The Netherlands (phone: +31 33 4682497, fax: +31 33 4682300, e-mail: [email protected]), **SSI, P O Box 5195, Tygervalley, Cape Town, South Africa (phone: +27 21 9454114, fax: +27 21 9451270)

The Dutch are currently internationally acknowledged for their extensive contributions to Membrane Bio Reactor (MBR) developments. This was achieved as spin-off of industrial MBR design and operation experiences and specific knowledge gained following a multi-year development programme initiated by STOWA (the dedicated research company for the Dutch Waterboards, e.g. comparable to DWAF). This development programme included not only extensive pilot tests with many commercially available membranes, but also the design and operation of the first full scale demonstration MBR for municipal wastewater in the Netherlands. Many recent innovations and considerations for MBR design were initiated based on experience gained through this programme. The aim of the Dutch development approach was to optimise the operating costs of MBR to make the process economically feasible for relatively diluted wastewater, i.e. municipal sewage of a non-industrial nature. The full-scale demonstration MBR plant for Sewage Treatment Plant (STP) Varsseveld was part of this development programme and is now considered to be one of the best performing and most costeffective plants in the world. DHV was the leading consultant in this development programme on behalf of STOWA. DHV is still one of the prime driving forces behind the Dutch MBR developments and has been responsible for the design of many industrial MBR's. This experience with industrial and municipal MBR design and operation learned that if consideration is given to all different aspects of a MBR unit e.g. sewerage/feed system, influent characteristics, effluent requirements (and potential uses) as well as the environmental impact, a reliable MBR system with excellent membrane performance, good effluent qualities with significant energy and cost savings can be achieved.



MBR Varsseveld

Industrial MBR

It is proposed that the following topics will be covered in the presentation: 1. Status of Dutch MBR experience for industrial and municipal experiences a. Several industrial MBR-units have been up and running for many years in The Netherlands, using this experience as a stepping stone a development program was shaped to investigate the applicability and more importantly to develop the technology potential, for treating diluted municipal wastewater. The first phase of this development program started with a pilot research at Beverwijk WWTP. This was in cooperation with the membrane suppliers Kubota, Mitsubishi, X-flow, Zenon, Memfis, Seghers-Keppel and Huber-VRM, various MBR pilot systems with capacities up to 20 m /h were tested, in the period from 2000 until 2004. In 2002 two other pilot projects were initiated, a Zenon MBR pilot installation with a capacity of 20 m /h at Maasbommel and the second, a Kubota MBR pilot with a capacity of 5 m /h at Hilversum WWTP. Both of these pilot projects were directed to the feasibility of meeting future expected discharge limits (2.2 mg TN/l and 0.15 mg TP/l) and carried out with great success. The next phase focused on the construction of three demonstration plants including one fullscale MBR (Varsseveld WWTP, 755 m /h) and a hybrid installation (Heenvliet WWTP, 100 m /h) which were started up in 2004 and 2006 respectively. In a hybrid installation, the MBR is integrated in a conventional activated sludge system. A third demonstration installation (also a hybrid installation) is under construction at Ootmarsum WWTP (150 m /h) and will be started up in the beginning of 2007. Varsseveld, Heenvliet and Ootmarsum are equipped with Zenon (capilar) membranes, Toray (flat sheet) membranes and X-Flow (tubular) membranes respectively. 2. Review performance and operation experience with industrial units and the Varsseveld STP a. The performance of the industrial plants is excellent have confirmed, as expected, long membrane lifetimes.

3 3 3 3 3 3



b. The Varsseveld STP complies with the MTR-quality (TN = 2.20 mg/l, TP = 0.15 mg/l) and is considered to be one of the world's most cost-effective MBR's and the membranes still have similar high performance as in the start up period (in 2004). 3. Review lessons learnt in design and operating of industrial and municipal with focus on a. MBR configuration i. Separate or hybrid ii. Segmentation of bioreactor and membrane b. Pre-treatment i. Far-going, pre-treatment is of greatest importance for reliable membrane performance. The selection of and required capacity of the pre-treatment equipment depends strongly on wastewater characteristics. c. Membrane cycle operation i. The selected membrane cycle operation significantly impacts the overall energy consumption of the plant. It was proven that by optimising the cycle, substantial energy reductions could be achieved. ii. The selected membrane cycle also showed significant impact in increasing the efficiency of the MBR and both the performance and chemical consumption for the membranes. d. Energy consumption i. Up-to-date information showing total energy consumption and the individual energy consuming processes will be presented. e. Effluent quality i. Detailed data on achieved effluent quality will be presented. At this point of time, it is foreseen to prepare a speech that is of interest to both industrial and municipal participants, however, depending on the actual participants background focus can be shifted to one them if desired.



Session 4: Paper 12


A. Masala; P.O. Osifo and H. van der Merwe Department of Chemical Engineering, Vaal University of Technology, Private Bag X021, Vanderbijlpark 1900, South Africa; Tel: +27 12 5223371; Fax: + 27 12 522 8042 ; email: [email protected]

1. Introduction The World today is completely dependent on the use of energy which is produced from various technologies e.g. nuclear power and internal combustion engines. Such technologies are not environmentally friendly since they emit greenhouse gases and are also too costly. The transportation industry has been identified as one of the major sources of air pollution and in view of this, a large potential market for vehicle applications have developed a strong interest in the use of a direct methanol fuel cell (DMFC). The DMFC can be operated at low temperatures (25 - 130 C) and its use of methanol in vehicles limits pollution with about 100 times below current emissions and within the national standards (Singh et al., 1999). The major drawbacks for using a DMFC is its low power density and low fuel utilization which are related to poor reaction kinetics and methanol permeation through the polymer electrolyte membrane (PEM). Methanol permeation causes a mixed potential hence contributing to a major fuel loss and as a result the DMFC performance decreases. The commercial PEM which is currently used in DMFC is Nafion 117 membrane. It has the ability to provide high current densities, at relatively low temperature and pressure but its methanol permeability is very high hence reducing the fuel cell performance. Recently, researchers have focused on the use of chitosan as membrane material for fuel cell. Chitosan is a an N-de-acetylated derivative of chitin and chitin can be obtained in large amounts from fishery waste like crabs, lobsters and quills (Muzzarilli, 1977) thus can be obtainable at a very low cost. In this paper chitosan membranes are manufactured, characterized and fabricated with Pt/C and PtRu/C catalyst layers for performance evaluation in the DMFC. 2. Experimental Chitosan membranes were prepared by dissolving chitosan flakes (Sigma Aldrich and Biopolymer, USA) in acetic acid. The solution was cast on a container and then transferred to an oven to form a membrane. Prior to tests the membranes were cross linked in sulfuric acid and neutralized in sodium hydroxide.




2.1. Membrane Characterization The Degree of deacetylation for the two chitosan membranes was determined. Chitosan and Nafion 117 membranes were characterized in terms of thermal stability, water uptake, membrane resistance and methanol permeability. Thermal stability was determined using a Differential Scanning Calorimetry (DSC) and thermogravimetry (TG). The membrane resistance and methanol permeability was determined by using a two compartment diffusion cell similar to the one used by Pivovar et al. (1999). 2.2 Membrane fabrication and Single Cell Testing Chs-I, Chs-II and Nafion 117 membranes were fabricated by hot pressing the anode with Pt/Ru/C and the cathode with Pt/C. The hot press temperature was 60 C for chitosan and 110 C for Nafion 117

o o

membranes. The fabricated chitosan membranes are shown in Figure 1. The performance was tested in a single cell air breathing DMFC.

Figure 1: Fabricated Chs-I and Chs-II membranes 3. Results and Discussion The TG, DTG and DSC thermograms for Chs-I and Nafion 117 membranes under nitrogen indicated that chitosan membranes are stable to 230 C and Nafion 117 is stable to 320 C. Water uptake values of 65% and 60% for Chs-I and Chs-II membranes were found. Chitosan is highly hydrophilic with a hydroxyl (OH ) and the amino (NH2) groups on its backbone. The hydrophilic groups cause the swelling on chitosan membranes (Wan et al., 2003). It was also established that Chs-I membranes posses a higher water uptake than Chs-II membranes with a difference of about 5%. Wan et al. (2003) associated the lower water uptake in high DDA membranes with a strong intermolecular hydrogen bond between the hydroxyl group and the amine group. The proton flux (J) across the membrane was determined at different speed and also the comparison of flux through Chs-I, Chs-II and Nafion 117 membranes are shown in Figure 4 which indicates that Nafion 117 membranes have lower flux than chitosan membranes. This is attributed to its high overall mass transfer resistance.

o o



3 e-6

2 e-6

2 Flux (mol/cm .s)

2 e-6

2 e-6

1 e-6 C h s-I C h s-II N a f io n 1 1 7

5 e-7

0 0 2000 4000 6000 8000 10000 12000

T im e (se c )

Figure 2: The change in flux (J) with time though Chs-I, Chs-II and Nafion 117 membranes at 20 C and 410 rpm The membrane resistance () to proton flow was determined for each membrane and the values of 196.7, 205.7, 284s/cm were found for Chs-I, Chs-II and Nafion 117 membranes respectively. The high of Nafion 117 membranes compared to chitosan membranes can be attributed to the membranes' lower water uptake which minimizes the mobility of protons through the membrane. Methanol permeability results are given in Table 1. Methanol permeability was found to increase with an increase in concentration and also increased with an increase in temperature for all membranes. The permeability of Nafion 117 membranes was high when compared to chitosan membranes. membranes showed high methanol permeability when compared to Chs-I membrane at 20 C Table 1: Comparison of methanol permeability properties of Chs-I, Chs-II and Nafion 117 membranes at 20 C Membrane Type Thickness (cm) Chs-I Chs-II Nafion 117 0.020 0.020 0.018 8M 1.40 3.09 3.90 Permeability (cm /s) ·10 12.5M 1.75 4.22 5.03

2 -6 o o



18M 2.92 3.92 8.26

Figure 5 gives a comparison of cell voltage and power density against current density for a DMFC with Chs-I MEA, Chs-II MEA and Nafion 117 MEA. The DMFC performance using Chs-I MEA slightly exceeds that of Chs-II MEA. The low cell performance in Chs-II MEA is attributable to its high membrane resistance to proton flow, high methanol permeability, low water uptake and possibly the amount of PtRu/C catalyst loading on the membrane. Therefore, the Chs-I MEA has better properties than the Chs-II membrane but Nafion 117 membranes still exhibits a high performance with results of the same order magnitude as Chs MEA. The maximum power density for commercial Nafion 117 membranes was found

WISA-MTD 2007 38

to be 0.698 W/cm , which is higher than that of the Nafion 117 membranes fabricated in the laboratory. This result suggest that either the fabrication technique or the amount of catalyst loading was not sufficient but an exact explanation for this behavior is yet to be found.


0.5 Chs-IMEA Chs-IIMEA Nafion 117-MEA



Power Density (W/cm2)


Cell Voltage (V)



Chs-I MEA Chs-II MEA Nafion 117 membrane




0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6

0.00 0.0 0.1 0.2 0.3 0.4 0.5 0.6

Current Density (A/cm 2)

Current Density (A/cm2)

Figure 3: Comparison of cell voltage and power density against current density for a DMFC with Chs-I MEA, Chs-II MEA and Nafion 117 MEA

4. Conclusions The degree of deacetylation was determined by Infrared Spectroscopy with Chs-II exhibiting a high DDA. The TG and DSC thermogram have shown that Nafion 117 membranes have better thermal properties than chitosan membranes since they only begin to decompose at 330 C without losing a lot of weight. Chitosan membranes water uptake was higher than Nafion 117 membranes because chitosan membranes are very hydrophilic. The degree of deacetylation of the two-chitosan membranes significantly affected the water uptake. Chs-II membranes were found to have a high DDA which enhanced its crystalinity thus preventing the water from entering the crystalline region resulting in decreased water uptake. The study of membrane resistance to proton flow has shown that Nafion 117 membranes has higher resistant to proton flow than chitosan membranes. Moreover, the resistance of Chs-II membranes was higher than that of Chs-I membrane due to their different water uptake values. The permeability through chitosan membranes was much lower than that of Nafion 117 membranes. The effect of feed concentrations and temperature on methanol permeability was also studied, and permeability was found to increase by increasing either the methanol concentration or the operating temperature. 5. References PIVOVAR, B.S., WANG, Y. & CUSSLER, E.L. 1999. Pervaporation membranes in direct methanol fuel cells, Journal of Membrane Science, Vol. 154: 155 - 162. WAN, Y., CREBER, K.A.M., PEPPLEY, B. & BUI, V.T. 2003. Ionic conductivity of chitosan membrane, Polymer, Vol. 44 (4): 1057 - 1065.

WISA-MTD 2007 39


SINGH, D., LU, D.M. & DJILALI, N. 1999. A two-dimensional analysis of mass transport in proton exchange membrane fuel cells, International Journal of Engineering Science, Vol. 37: 431 - 451. MUZZARILLI, R.A.A. 1977. Natural Chelating Polymers: International series of monographs in analytical chemistry, Vol. 35: 83 -138.



Session 4: Paper 13


S.P. Roux, E.P. Jacobs*, A.J. van Reenen and M. Meincken Department of Chemistry and Polymer Science, University of Stellenbosch, Private Bag X1, Matieland 7602, South Africa, Fax: +27218084967, email: [email protected]

Introduction Adsorptive fouling decreases membrane performance, increasing the cost of membrane dependent applications. Hydrophilic modification of polysulphone (PSU) ultrafiltration (UF) membranes decreases adsorptive fouling [1, 2, 3]. Blending the hydrophilic copolymer [1, 4] into membrane casting solutions affects membrane morphology, flux and separation characteristics, which complicates direct comparison between fouling data for modified and unmodified membranes. The preparation of hydrophilic modified membranes with similar characteristics as the unmodified control is vital to demonstrate that any decrease in solute fouling for the more hydrophilic membranes is due only to the modification process. Brief description of project The blending of a branched poly(ethylene oxide) -block-PSU copolymer into the membrane formulation renders all polymer surfaces hydrophilic [4, 5] but this method also affects membrane solution viscosity and membrane separation characteristics. These changes were investigated to facilitate the preparation of hydrophilic modified membranes similar in performance characteristics to the unmodified PSU control. Adsorptive fouling for both modified and unmodified membranes was investigated in relation to the difference in the hydrophilic/hydrophobic nature of the membrane surfaces. Results and discussion The addition of the hydrophilic copolymer to the membrane formulation impacted significantly on the adsorptive fouling of the modified membrane, indicating that this membrane modification process may help to reduce the cost of membrane operations in future.



References 1. Hancock, L.F., Fagan, S.M. and Ziolo, M.S., (2000) Hydrophylic, semipermeable membranes fabricated with poly(ethylene oxide)-polysulfone block copolymer, Biomaterials 21, 725-733. 2. Kingshott, P., Wei, J., Bagge-Ravn, D., Gadegaard, N. and Gram, L., (2003) Covalent attachment of poly(ethylene glycol) to surfaces, critical for reducing bacterial adhesion, Langmuir 19, 69126921. 3. Park, J.Y., Acar, M.H., Akthakul, A., Kuhlman, W. and Mayes, M.A., (2006) Polysulfone-graftpoly(ethylene glycol) graft copolymers for surface modification of polysulfone membranes, Biomaterials 27, 856-865. 4. Hancock, L.F., Fagan, S.M. et al., (2001) Highly branched block copolymers, US Patent 6,172,180 B1. 5. Roux, S.P., Jacobs, E.P., Van Reenen, A.J., Morkel, C. and Meincken M., (2006) Hydrophilisation of polysulphone ultrafiltration membranes by incorporation of branched PEO-block-PSU copolymers, Journal of Membrane Science 276 8-15.



Session 4: Paper 14


C.J. Garcin Chemical Engineering, University of Cape Town, South Africa, Tel: +27 21 650 5523, Fax: +27 21 650 5501, Email: [email protected]

Low molecular weight biophenolic compounds with high antioxidant activity were selectively extracted from wastewater arising from the production of table olives. This was achieved using a hydrophobic hollow fiber membrane module as contactor between the aqueous wastewater phase and an organic extracting phase. In this manner only the desired solutes were extracted through the membrane. The method has advantages over conventional liquid/liquid extraction in that the bulk phases do not mix and thus subsequent phase separation is not required. The experimental system used is shown in Figure 1. Wastewater was recycled through the membrane A slight lumens while solvent was recycled through the shell space. over-pressure (75kPa) was applied to the wastewater stream to stabilise the membrane interface and prevent solvent breakthrough.

Figure 1: Experimental membrane extraction system



Concentrations of desired biophenols in the two streams were measured over time by HPLC. The overall mass transfer coefficient was evaluated as a function of flow rates, solvent-to-wastewater ratios and pH. Figure 2 shows results of some extraction experiments.

Organic phase hyrdoxytyrosol concentration


Aqueous phase hydroxytyrosol concentration

3500 3000


Concentration (mg/L)

1 Hz 5 25 Hz 35 Hz 45 Hz

Concentration (mg/L)

2500 2000 1500 1000 500 0



1 Hz 5


25 Hz 35 Hz 45 Hz


0 0 5 10 15 20 25 30 35









Tim e (m in.)

Tim e (m in.)

Figure 2: Variation of extraction rates at different flow rates There was no significant variation of extraction rates at the various flow rates investigated (aqueous Re 2 ­ 6, organic Re 0.15 ­ 5), indicating that boundary layer mass transfer resistance was minimal, and transfer through the membrane itself was the rate limiting step. A mathematical model was applied to determine the overall mass transfer coefficients for different experimental conditions. The mass transfer coefficient was determined to be around 1.3x10 m.s for all experiments. It was possible to increase the amount of biophenols extracted by increasing the wastewater-to-solvent ratio. As this ratio increases, the distribution coefficient favours increased equilibrium concentrations in the solvent phase. An maximum extraction rate of approximately 2 g.m .hr irrespective of solvent-to-wastewater ratio. The process compares very favourably with conventional liquid/liquid extraction, and has numerous additional advantages which include: no moving parts, independent control of phase flow rates without the dangers of washout or flooding, aseptic operation and easy (linear) scale-up. The interfacial surface area available for mass transfer in the hollow fiber modules exceeds that of directly dispersed systems, resulting in high mass transfer coefficients and smaller plant footprint compared to conventional systems. In addition, foaming and emulsions are avoided as the bulk fluids are not directly mixed. As well as producing a highly valuable by-product that can be used as a natural ingredient in the food, cosmetic or nutraceutical industries, extraction of the biophenols from olive-derived wastewaters also reduces the COD and antimicrobial properties of the remaining wastewater, resulting in easier treatment and biological degradation thereof. The experimental system is currently being evaluated for long-term membrane performance, however to date membrane fouling has not been an issue as this is a diffusion-based (not filtration) process.

WISA-MTD 2007 44

-2 -1 -5 -1

was generally observed,

Problems encountered included aqueous breakthrough into the solvent due to membrane wet-out, and solvent loss due to slight solubility of ethyl acetate in water. Future work will include cross-linking the membrane pores with a hydrogel (polyvinyl alcohol) in order to stabilise the membrane interface between the two phases.



Session 4: Paper 15


V.L. Pillay and E. P. Jacobs

1 1 2

Dept. of Chemical Engineering, Durban University of Technology, Mansfield Rd, Durban, 4001, [email protected]


Dept of Chemistry and Polymer Science, University of Stellenbosch, Matieland, 7602, [email protected]

In Immersed Membrane Filtration (IMF), the membrane modules are submerged into the feed suspension and are operated in an outside-in configuration. Clear liquid permeates the membrane wall, driven by a gravity head or a suction pump, and is recovered as the product. Contaminants are retained on the outside of the membrane module and form the fouling layer. Usually, air scouring is used to reduce the build-up of the fouling layer, and maintain higher fluxes. IMF can offer various advantages over conventional membrane modules. In particular, the energy

consumption in IMF systems is significantly lower than conventional membrane systems. Accordingly, IMF systems are increasing in popularity internationally, particularly in membrane bioreactor (MBR) applications, and increasingly in potable water production. At present two IMF modules, viz. the capillary module produced by Zenon, and the flat-sheet module produced by Kubota, dominate the international market. The Dept of Chemistry and Polymer Science, US, and the Dept. of Chemical Engineering, DUT, are engaged on a collaborative project to develop a local flat-sheet IMF module for applications in developing economies. The Water Research Commission of South Africa is sponsoring this work. The basis of the approach is to exploit the locally produced woven fiber microfiltration (WFMF) fabric. In investigations to date the WFMF system has exhibited significant advantages over the current commercial IMF modules. Firstly, the WFMF fabric never requires a chemical clean. Secondly, the WFMF fabric can be allowed to dry out without destroying its separation capability. Thirdly, the WFMF fabric is extremely robust, and can be easily repaired if accidently damaged. These make the WFMF system very attractive for applications in developing economies. To date two types of modules have been developed, viz. a rigid module and a flexible module. The rigid module is more suitable for immersed membrane bioreactor (IMBR) applications, and industrial



separations. The flexible module is considerably cheaper to produce and would be more suited to potable water production. This paper will discuss the module designs, their advantages and disadvantages, and the performance results obtained to date in potable water treatment and IMBR applications.



Session 5: Paper 16


J. van Aartsen , B. Steytler , H. Grant and M. Govender

1 1 2 4 3

VWS Envig (Pty) Ltd, S04 Modderfontein House, High Street, Modderfontein, South Africa Tel. +27 11 663 3600; Fax +27 11 608 4772


TB Process Technologies (Pty) Ltd; Sasol Synfuels (Pty) Ltd, Secunda, South Africa


Keywords: Reverse osmosis, ultra-filtration, cooling tower blowdown, reuse

Introduction The construction of the Sasol Turbo plant in Secunda resulted in a significant increased demand for utilities on the existing Sasol Synfuels site. Demineralised (polished) water will be used as make-up water for the new waste heat boiler and the existing water treatment plants on site could not supply the additional demand. In addition to the limited production capacity of the existing systems, no additional allocation of raw water was available from the Vaal River system. Alternative water sources had to be found and blowdown water from the Utility Cooling Towers was identified as the most suitable source. Brief Description of Project Various challenges such as the variability of water quality from cooling towers, degradation of feed water sources and impact of possible future feed sources had to be overcome at design stage. A new desalination plant utilising clarification, softening, ultra-filtration, double pass reverse osmosis (RO) and mixed bed ion exchange systems was constructed to achieve the required demineralised water quality. The new plant was designed to produce the following treated water:

3 3

· ·

RO permeate: Polished water:

350 m /h 128 m /h

Chemical Softenening Cooling tower blowdown (800 m /h) is treated in a Multiflo sidestream softener to remove hardness and suspended solids. Excess lime and soda-ash is utilised to ensure maximum chemical softening and




typical outlet hardness of < 50 mg/l CaCO3 is achieved in this process step. The pH of the clarified water from the softener is reduced with sulphuric acid prior to the ultra-filtration (UF) system. Ultra-Filtration The UF system provides the final solids removal step to remove all suspended solid particles from the clarified water and to produce filtered water with a silt density index of less than 3. This level is required for satisfactory operation of the downstream ion exchange and RO systems. The UF membranes function on an outside-in basis, with air scouring on the outside surface of the membranes to reduce fouling of the membrane surface. Each train has a dedicated blower for air scouring. The membranes are also regularly backpulsed (reversal of flow) to extend membrane runtime between maintenance cleans. The recovery rate of the UF system is 90% and the reject water is recycled to the inlet of the plant to maximize the overall water recovery. Ion Exchange Softeners The ion exchange (IX) softeners remove residual hardness from the UF filtrate by absorption on IX resin. Packed bed technology is used for this application and counter-current regenerartion with 10% sodium chloride solution ensures that effluent volumes are minimised. Removal of most of the Ca the Sasol Site. Double Pass Reverse Osmosis The spiral reverse osmosis system is fed with softened water from the IX softeners. The normal mode of operation is two trains in operation and the third on standby or in cleaning mode. Each RO train that is in operation produces 200 m /h of permeate.

3 2+

and Mg



are required due to a stringent hardness specification for the RO retentate (brine), which is recycled within

The RO recovery rate will normally be 80%.


provision is made for the recovery to be varied between 70 and 80%. In order to meet the silica specifications after the mixed bed polisher and to avoid a too high TOC load in the feed to the polisher, the Reverse Osmosis plant is executed as a two-pass system where the permeate of the first pass is used as feed water for the second pass. The retentate (brine) of the first RO pass can be used as make-up to the Process Cooling Towers. Permeate of the second pass is fed to the mixed bed polishers as well as to the existing condensate tanks and pumps at Unit 044. A facility is also provided to transfer permeate to Sasol East via the transfer line during a total West shutdown. The retentate of the second pass is returned to the feed of the first pass to maintain a high over-all recovery of the system. Mixed Bed Polishers The mixed bed polishers remove traces of TDS from the RO permeate by absorption on IX resin. The polished water system has two parallel trains, one in operation and the second on regeneration / standby. The normal production rate per train is 167 m /h, but the flow can be turned down to a minimum of 148

WISA-MTD 2007 49


m /h. The polished water product of System 05 is supplied to the Client (mainly Turbo SCC with the balance to Unit 044). The mixed bed polisher vessels are filled with a combination of strong acid cation resin and strong base anion resin. The resin is regenerated with diluted sulphuric acid (5 %) and caustic soda (4 %). Plant Details The plant has been in operation since August 2006 and preliminary operational data will be presented for both the pre-treatment and reverse osmosis systems.




Session 5: Paper17


H.W.J. Els , A. Maartens , F.J. van Heerden , P.J. Allison , M. Govender , L. van Wyk , H.W. Grant and A.J. Steytler

1 2 4 1 1 1 1 2 2 3

Buckman Laboratories, Hammarsdale, South Africa, Utilities Operations, Sasol Secunda, Sasol Technology, Sasol Secunda, TB Process Technologies (Pty) Ltd., Kempton Park.



Background Large quantities of raw water can be saved in the industrial sector when blow-down water from large cooling systems is recovered and re-used (Grant, 2006). Cooling tower blow-down is softened through a high rate clarifier. The supernatant is fed to a submerged ultrafiltration plant for final clarification before further softening by ion exchange. The softened water is feed into a reverse osmosis plant and the final permeate subsequently polished to boiler feed water by mixed bed polishers. This paper presents an overview of the lessons learned during commissioning and optimisation of the cold lime softener, the treatment and monitoring of biological growth in the feed streams to the reverse osmosis plant. Case Study As in many other documented cases it was not foresee that biological growth would not be a concern (Flemming et al., 1997). The original biocide programme consisted of only an oxidising biocide (chlorine) dosed upstream of the submerged ultrafiltration plant. The oxidising effect of this biocide was then neutralised by the addition of sodium meta-bisulphite before the ion exchangers. This chlorine dosing was found not to be very effective and lead to severe blocking of cartridge filters used to protect the down stream reverse osmosis membranes. In addition, biological growth was found to be a great challenge down stream of the sodium meta-bisulphite dosing point. Changing to target bromination resulted in an increase of about 60% in the available halogen.



In addition, to protect the rest of the system downstream from the sodium meta-bisulphite dosing point, a non oxidising biocide was introduced. This non oxidising biocide is dosed just before the reverse osmosis plant. An improvement was immediately noticed by a reduction in the rate at which the differential pressure across the cartridge filters increase. Conclusion Monitoring the efficiency of the disinfection programme was done by using adenosine tri-phosphate measurements, conventional plate counts and autopsies. The subsequent results were: Improved disinfection of ultrafiltration membranes, Reduction in the replacement frequency of cartridge filters, no evidence of bio-fouling on reverse osmosis membranes. Future Options To move the non oxidising biocide further upstream to increase contact time and thereby the killing efficiency. References: Grant, H.W., Reclaiming cooling tower blow down as feed for polished water plant, Proceedings of the Water Institute of South Africa, Biennial Conference & exhibition, Durban, (2006). Flemming, H.C. Schaule, G. Griebe, T. Schmitt, J. and Tamachkiarowa, J., Biofouling- the Achilles heel of membrane processes, Desalination, 113 (1997) 215-225.



Session 5: Paper 18


M. Chaumaz , W. Taljaard , P. Mouton , W. Manyana and J. du Toit

1 2 1 1 2 2 2

VWS Envig, S04 Modderfontein House, High Street, Modderfontein 2065; Retail Brand Interafrica, 1 Guthrie Road, Wadeville 1422

Introduction Retails Brands Interafrica (RBI) produces carbonated soft drinks, alcoholic beverages and iced tea products. Since February 2004, RBI have been owned by Pioneer Foods who have signed a franchise agreement with PepsiCo International to produce Pepsi products for the African market. This integration into the PepsiCo International Group allows RBI to offer a wider range of soft drinks but the result is that in order to manufacture these products more stringent water requirements have to be met. RBI through VWS Envig has installed a new water treatment plant which is able to deliver a variable flow rate with a constant water quality. The quantity and the quality of the effluent generated by the treatment plant was also a major concern for RBI in terms of the environmental impact that the produced brine stream would have. To allow RBI to achieve the required objectives VWS Envig supplied a nanofiltration and a reverse osmosis system. Project Description The water treatment plant installed at RBI was conceptualized and designed to meet the following stringent treated water criteria set by RBI: Flowrate [m /h] Microbiological count [cfu/ml] Total Alkalinity [mg/l as CaO3] Total hardness[mg/l as CaO3] Free chlorine [mg/l]


80 (peak flowrate, highly variable) 0 < 35 < 10 0

An important design aspect of the water treatment plant was the minimisation of effluent discharge and as such the required water recovery was set at 85 %.



Potable water from the municipality still contains hardness and alkalinity which can alter the taste of the soft drinks. The objective of the treatment plant was to lower the concentration of these constituents. Membrane technology was particularly suitable for this application. In terms of process, the pre-treatment is a typical beverage process flowsheet consisting of disinfection, deep bed sand filtration and carbon filtration. Following the pretreatment process, a nanofiltration unit has been installed to decrease the concentration of total dissolved solids in the treated water. In particular the hardness and alkalinity have to be removed partially to minimise the potential taste implications associated with the removal of these ions completely. Nanofiltration was selected due its characteristic selectivity towards divalent and large ionic species. Additionally membrane technology was selected over alternative technologies such as ion exchange to ensure that a constant water quality was achieved. Ion exchange technology often produces variable water quality through the treatment cycle and following regeneration steps. Additionally ion exchange processes produce difficult acid and caustic waste streams which have to be disposed of. To further increase the recovery on the water treatment system, the brine from the nanofiltration plant was further treated in a reverse osmosis unit. The brine sent to the RO skid is further "squeezed" to produce additional permeate while also reducing the volume of brine requiring disposal. The design of this unit was critical as the high salt concentration in the brine being fed to the RO unit has a high scaling and fouling potential. Results and Discussion The results achieved on the water treatment plant over the first few months of operations have been favourable. · Flow rates of between 0 and 80 m /h have been achieved without the need for shutting down of the plant and without detrimental effect on water quality or plant operations. · The treated water microbiological count has been zero for the entire duration of operation of the plant as compared to the conventional treatment process previously utilised in which excessive microbiological counts were often encountered · · Total alkalinity concentrations have been maintained at a constant level of slightly below 10 mg/l. Total hardness has been maintained at < 1 mg/l


Additionally and most importantly an overall water plant recovery of 90% has been achieved. The plant has been operational without a CIP for in excess of two months and no scaling or fouling effects are evident at this stage. The effluent discharge characteristics from the RBI facility have also improved. A better quality effluent which is now less detrimental to the municipal system is currently being discharged.

WISA-MTD 2007 54

Conclusion The process, now operational for a number of months has met and in some cases exceeded the treated water and efficiency requirements of RBI. Despite the recovery already being high on the process, the possibility of further increasing the recovery on the RO plant is achievable as the plant is operating at optimum levels without any of the normal problems associated with scaling and fouling in high recovery operations. Further changes, such as backwash recovery and CIP effluent neutralisation are in the pipeline and these initiatives are being driven by RBI and VWS Envig to further decrease the discharge volumes while maintaining an improved wastewater quality. The continual improvement in the plant water recovery, while still maintaining an optimum water quality, is evidence of RBI's commitment to sustainable development within the South African beverage industry.



Session 5: Paper 19


P. Sehn Dow Water Solutions, Germany, Tel: +49 722 79 13752, Fax: +49 722 79 13708, email: [email protected]

Any reverse osmosis plant is designed to provide a certain permeate flow rate of a certain quality at a certain recovery from a given feed water source at a certain feed pressure. The operating conditions however are normally not stable. Therefore the membrane performance cannot be measured directly. Normalization means translating actual performance data measured under actual and varying conditions into theoretical performance data under defined and stable reference conditions. A plot of these normalized performance data shows the trend of the membrane performance independent on changing operating conditions. It is useful to determine the cleaning requirements and to judge the system performance with respect to the design values. Different methods can be used to normalize field data so that they can be compared against a reference case. One method is projecting the RO performance with RO design software such as ROSA (Reverse Osmosis System Analysis) software under actual conditions and comparison with the measured numbers. For a continuous analysis and graphical display of trends, software which is able to normalize data online such as FTNORM is better suited. This software is based on a commonly accepted industry standard, ASTM D4516: the normalized permeate flow rate is calculated from the actual flow rate by multiplication with the ratio of the actual versus the reference net driving pressure and the ratio of the respective temperature correction factors. The normalized salt passage is the actual salt passage corrected to the reference conditions and the reference pressure. If the normalized flow performance decreases because of fouling, the normalized salt passage will also decrease as a result of the lower flux. In practice, most plants are operated not at constant pressure but at constant flow rate. Therefore, it might be more useful to calculate the normalized salt passage with reference to a constant system flow rate rather than to a constant pressure. This way, the result is directly proportional to the salt permeability of the membrane and does not depend on operating parameters.



The reference condition is normally defined at design conditions after 24 hours of continuous operation since start-up. This provides the baseline for the operator to decide when to clean. For best long-term performance we recommend to clean when the normalized flow has declined by 10%. Some operators prefer to clean less frequently in order to save labor and plant down time. This may impact the long term membrane performance and reduce the membrane life time. Another option is to use the results of the ROSA projection as the reference. This can be the start-up performance, the long-term performance or the warranted performance. For systems with more than one train it makes sense to use a common reference that allows comparing the trains with each other. Case history Elbe A large reverse osmosis plant provides process water for a paper mill in Germany. The river water is pretreated with flocculation and open sand filtration; the membranes are FILMTECTM BW30-400FR. The flow started better than projected, but decreased in the following months. Any cleaning attempts failed; only a strong nitric acid treatment could restore the flow. A surfactant of a polymer emulsion used as flocculant was the cause of the flux decline. When this was replaced by solid polymer it was possible to recover the flow performance after a series of standard cleanings.

Normalized System Permeate Flow


400 projected initial performance


300 projected long-term performance (after cleaning) 250


200 150 100



0 1-Jan-04 1-Mar-04 1-May04

1-Jul-04 31-Aug- 31-Oct- 31-Dec- 2-Mar-05 2-May04 04 04 05

1-Jul-05 31-Aug- 31-Oct- 31-Dec- 2-Mar-06 2-May05 05 05 06



31-Oct- 31-Dec06 06

Case History Infra-Zeitz Another industrial RO plant with BW30-365FR membranes is running on river water. FTNORM was used to determine the optimum timing of cleaning. The membranes performed without deterioration in flow and salt rejection for more than five years.



Normalized System Permeate Flow Rate

100 90 80 70

projected initial performance

Flow Rate (m3/h)

60 50 40 30 20 10 0 01-May-01 01-May-02 01-May-03 30-Apr-04 01-May-05 01-May-06

projected long-term performance (after cleaning)

RO1 RO2 RO3 Reference



Session 5: Paper 20


J.J. Schoeman and S. van Zyl University of Pretoria, Department of Chemical Engineering, Water Utilisation Division, Pretoria Tel. +27 (12) 420 3569; Fax (12) 362 5089; email: [email protected]

Abstract Diffusion dialysis using anionic or cationic membranes is a concentration driven process that can be used for acid or caustic soda recovery from spent solutions. This technology, however, is not used in South Africa despite the fact that it is a very promising technology for acid and caustic soda recovery from spent solutions for chemical recovery for reuse. Therefore, diffusion dialysis was evaluated for acid and caustic recovery from spent industrial (pickling, electroplating, and other processes) solutions for acid and caustic recovery. It was found that diffusion dialysis functions effectively for recovery of nitric and hydrofluoric acids from spent pickling acid solutions. Nitric acid and hydrofluoric acid recovery ratios of approximately 80 and 60% could be obtained, respectively. Higher percentage nitric acid recovery (approximately 100%) is possible. Optimum acid recovery ratios were obtained at a feed liquor and feed water flow rates of approximately 1 l/h.sq m. Metals can be removed effectively from spent pickling effluent with diffusion dialysis. Iron was reduced in one case from 45 g/l in the feed liquor to 1.31 g/l in the recovered acid. Chromium and nickel could be reduced from 6.6 and 7.0 g/l in the feed liquor to 0.1 and 0.3 g/l in the recovered acid, respectively. Consequently, acids could be recovered effectively with diffusion dialysis for reuse in the pickling process. The dialysate (waste acid) contains high concentration levels of metals (approximately 40 to 50 g/l), nitrates (approximately 40 to 50 g/l HNO3) and fluoride (approximately 6 to 12 g/l HF). However, it would be possible to treat the dialysate effectively with lime for metal removal and acid neutralisation prior to final disposal. The clarified water, however, will contain high concentration levels of calcium (approximately 9 000 mg/l), nitrate-nitrogen (approximately 9 500 mg/l), and fluoride (approximately 11 mg/l).The clarified water should be disposed of safely.



References 1. 2. 3. J.D. Edwards and M.M. Benjamin (1990). Diffusion dialysis for recovery of acid from concentrated S. Koter and A. Warszawski (2000). Electromembrane processes in environmental protection. J. WiNniewski, G. WiNniewska and T. Winnicki (2004). Application of bipolar electrodialysis to process solutions: The importance of chemical speciation. Environ. Sci. Technol. (1990), 24, 880 - 885. Polish Journal of Environmental Studies, 9(1), (2000), 45 - 56. the recovery of acids and bases from water solutions. Desalination 169 (2004), 11 - 20.




M J Pryor , M van der Walt , M Hughes

1 2 1 2 2

Keyplan (Pty) Ltd., Eskom, Tutuka Power Station

In December 2005 a tender was awarded to Keyplan to upgrade and increase the capacity of the desalination plant located at the Tutuka Power Station. During 2006, the existing spiral wound reverse osmosis (SRO) plant was extended to a capacity of 25 megalitres per day, making it one of the largest membrane desalination plants for industrial water in the country. Background New Denmark Colliery is a mine that has been dealing with excess quantities of underground water for some time. As it is one of the deepest mines in the area, total abstraction is used. When mining is completed in a section, the roof collapses, which in turn breaks the aquifers causing old water to flow into the mine. The mine is continuously expanding and therefore there is a constant influx of water into the mine, creating an increasing demand for water to be pumped out, as the water cannot be discharged into the natural environment due to its high salinity. The volume of water in the mine has exceeded the capacity of the compartments and the mine is unable to cope with additional water influx. As a result Tutuka Power Station is faced with an increase in salt load and in order for the mine to continue its operations, the SRO plant's capacity needed to be increased from 12.5 to 25 M/day. SRO plant before upgrade Number of SRO trains: three Design capacity: Plant availability: Plant capacity: Feed to SRO plant: 12.5 M/d 90% 11.25 M/d (10.8 for own cooling water desalination) Concentrated cooling water (3.85 M/day) Mine water (max 7.4 M/day) Upgrade of the SRO plant Number of microfiltration (MF) trains: Number of SRO trains: Design capacity: Plant availability: Feed to SRO plant: five five 25 M/d 90% Concentrated cooling water (6 M/day) Mine water (19 M/day available)



Process Outline Water is pumped from the lime softening clarifiers to the new MF plant by three 550 m /hr (50%) pumps. The water is dosed with acid prior to a self-cleaning 200 micronfilter. The water is then dosed with chlorine to control biological activity. The microfiltration plant is based on the Pall Microza UNA membrane module in Polyvinylidene Fluoride (PVDF). PVDF is resistant to chlorine, making biogrowth easy to control. The membrane is a capillary operating in outside filtration mode with a cut off pore size of 0.1 micron (µm), which has proven to be ideal on this water. It also results in excellent water quality. Therefore, Keyplan was able to remove the existing cartridge filters before the RO thereby providing a significant saving during operation and the replacement of cartridges. Keyplan has successfully installed an additional two RO skids, each capable of treating 208 m /hr of feedwater at a design recovery of 87%. As with the old SRO system the RO design is based on a three bank array with interstage pumping between Banks 2 and 3. The existing three RO trains have been upgraded from 173 m /hr to 208 m /hr. This is possible as a result of the cleaner feedwater quality and subsequent, more consistent, lower pressure drop through the units. The old cartridge filters and feed flow control valves have been removed and the control of the RO feed pumps is now performed using variable speed drives. This has resulted in the following benefits: power saving, reduced water hammer during start-up and shutdown, significantly lower noise levels and expected reduced maintenance. The RO permeate from Bank 1 will have a lower total dissolved solids (TDS) than from Banks 2 and 3. Permeate back pressure is being used to balance the flux, therefore a pressure sustaining valve has been included to maintain the correct pressure thereby ensuring that the low TDS product is always available for use at the demin plant. Achievements and Conclusions


3 3 3 3

All the new equipment was installed into the existing building within the bounds of keeping the existing plant operating. The benefits of microfiltration and improved SDI has been noticeable in improving the pressure drop through the old SRO trains. No CIP has been required for over two months. The removal of flow control valves and operation of the SRO feed pumps using VSD's has reduced the noise levels in the SRO building. From the awarding of the tender, the plant was commissioned within a one year period."






7th WISA MTD Symposium and Workshop



Microsoft Word - WISA-MTD 2007 Abstract booklet.doc

63 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate


You might also be interested in

Microsoft Word - Farhang Shadman CV
Carbon dioxide sensors for intelligent food packaging applications
Microsoft Word - WISA-MTD 2007 Abstract booklet.doc
Microsoft Word - JournalPapersonUnsaturatedSoils.docx